WO1993020185A1

WO1993020185A1 - Method for in vitro proliferation of dendritic cell precursors and their use to produce immunogens

Info

Publication number: WO1993020185A1
Application number: PCT/US1993/003141
Authority: WO
Inventors: Ralph M. Steinman; Kayo Inaba; Gerold Schuler
Original assignee: Steinman Ralph M; Kayo Inaba; Gerold Schuler
Priority date: 1992-04-01
Filing date: 1993-04-01
Publication date: 1993-10-14
Also published as: AU4046193A; AU687733B2

Abstract

A method for producing proliferating cultures of dendritic cell precursors is provided. Also provided is a method for producing mature dendritic cells in culture from the proliferating dendritic cell precursors. The cultures of mature dendritic cells provide an effective means of producing novel T cell dependent antigens comprised of dendritic cell modified antigens or dendritic cells pulsed with antigen, including particulates, which antigen is processed and expressed on the antigen-activated dendritic cell. The novel antigens of the invention may be used as immunogens for vaccines or for the treatment of disease. These antigens may also be used to treat autoimmune diseases such as juvenile diabetes and multiple sclerosis.

Description

METHOD FOR IN VITRO PROLIFERATION OF DENDRITIC CELL PRECURSORS AND THEIR USE TO PRODUCE IMMUNOGENS

This invention was made with United States Government support under NIH grant AI13013 awarded by the National Institutes of Health. The United States Government has certain rights in this invention. The making of this invention was also supported by the Austrian Government through grants NB 4370 (Austrian National Bank) and P 8549 (Austrian Science Foundation) .

This application is a continuation-in-part of United States Patent Application 07/981,357 filed November 25, 1992 which in turn is a continuation-in-part of United States patent application 07/861,612 filed April 1, 1992.

TECHNICAL FIELD OF THE INVENTION This invention relates to a method of culturing cells of the immune system. In particular a method is provide f°^r culturing proliferating dendritic cell precursors an for their maturation in vitro to mature dendritic cells.

This invention also relates to dendritic cell modifie antigens which are T cell dependent, the method of makin them, and their use as immunogens. Vaccines, methods o immunizing animals and humans using the mature dendriti cells of the invention, and the modified antigens are als described.

BACKGROUND OF THE INVENTION The immune system contains a system of dendritic cell that is specialized to present antigens and initiate severa T-dependent immune responses. Dendritic cells ar distributed widely throughout the body in various tissues. The distribution of dendritic cells has been reviewed i (1) . Dendritic cells are found in nonlymphoid organs eithe " close to body surfaces, as in the skin and airways, or in interstitial regions of organs like heart and liver. Dendritic cells, possibly under the control of the cytokine granulocyte macrophage colony-stimulating factor, (hereinafter GM-CSF) , can undergo a maturation process that 5 does not entail cell proliferation (2,3). Initially, the dendritic cells process and present antigens most likely on abundant, newly synthesized MHC class II molecules, and then strong accessory and cell-cell adhesion functions are acquired (4-7) . Dendritic cells can migrate via the blood ^and lymph to lymphoid organs (8-10) . There, presumably as the "interdigitating¹* cells of the T-area (8,11-13), antigens can be presented to T cells in the recirculating pool (14) . However, little is known about the progenitors of dendritic cells in the different compartments outlined above.

The efficacy of dendritic cells in delivering antigens in such a way that a strong immune response ensues i.e., "immunogenicity", is widely acknowledged, but the use of these cells is hampered by the fact that there are very few 0 in any given organ. In human blood, for example, about 0.1% of the white cells are dendritic cells (25) and these have not been induced to grow until this time. Similarly, previous studies (20, 21) have not reported the development, in culture, of large numbers of dendritic cells from bone 5 marrow. A more recent report described the development of dendritic cells in GM-CSF supplemented marrow cultures, however no documentation as to the origin of the dendritic cells or use of proliferating aggregates as an enriched source of dendritic cells was observed. Scheicher et al. J___ _Q Immunol. Method. 154:253-264 (1992). While dendritic cells can process foreign antigens into peptides that immunologically active T cells must recognize (4,6,7,14) i.e., dendritic cells accomplish the phenomenon of "antige presentation", the low numbers of dendritic cells prohibit c their use in identifying immunogenic peptides. ° Dendritic cells in spleen (15) and afferent lymph (16,17) are not in the cell cycle but arise from a proliferating precursor. Ultimately, dendritic cells emanate from the bone marrow (15,16,18,19), yet it has been difficult to generate these cells in culture except for two 5 reports describing their formation in small numbers (20,21) . Although a bone marrow precursor cell has been reported, conditions have not been reported that direct its proliferation in culture. Steinman, R. "The Dendritic Cell System and Its Role In Immunogenicity", Ann. Rev. Immunol.. 0 9:271-96 (1991). Identification of proliferating dendritic cells in bone marrow, in contrast to blood, is difficult because there are large numbers of granulocytes that develop in response to GM-CSF and these crowd the immature dendritic cell cultures, preventing maturation of the dendritic 5 precursors. The use of cell surface markers to enrich bone marrow dendritic cell precursors has been reported to result in only modest increases because the markers are also expressed by numerous non-dendritic bone marrow cells. Bowers, W.E. and Goodell, "Dendritic Cell Ontogeny" Res. 0 Immunol. 140:880-883 (1989).

Relatively small numbers of dendritic cells have also been isolated from blood. Vakkila J. et al. "Human Peripheral blood-derived dendritic cells do not produce interleukin lα, interleukin 13, or interleukin 6" Scand. J. 5 Immunol. 31:345-352 (1990); Van Voorhis W.C. et al., "Human Dendritic Cells", J.EXP. Med.. 1172-1187 (1982). However, the presence in blood of dendritic cell precursors has not been reported and as recently as 1989 the relationship between blood dendritic cells and mature dendritic cells in _Q other tissues was uncertain. Furthermore, it was recognize that dendritic cells are "rare and difficult to isolate and have not as yet been shown to give rise to DC [dendriti cells] in peripheral tissues." MacPherson G.G. "Lymphoi Dendritic cells: Their life history and roles in immun 5 responses", Res. Immunol. 140:877-926 (1989). ° Granulocyte/macrophage colony-stimulating factor (GM- CSF) is a factor which modulates the maturation and function of dendritic cells. Witmer-Pack et al, "Granulocyte/macrophage colony-stimulating factor is essential for the viability and function of cultured murine epidermal Langerhans cells". J.Exp.Med. 166:1484-1498 (1987) . Heufler C. et al. , "Granulocyte/macrophage colony- stimulating factor and interleukin 1 mediate the maturation of urine epidermal Langerhans cells into potent immunostimulatory dendritic cells", J. Exp. Med. 167:700-705 (1988) . GM-CSF stimulated maturation of dendritic cells in vitro suggests that the presence of GM-CSF in a culture of dendritic cell precursors would mediate maturation into immunologically active cells, but the important goal of achieving extensive dendritic cell growth has yet to be solved.

T-dependent immune responses are characterized by the activation of T-helper cells in the production of antibody by B cells. An advantage of T-dependent over T-independent immune responses is that the T-dependent responses have 0 memory, i.e. cells remain primed to respond to antigen with rapid production of antibody even in the absence of antigen and the immune response is therefore "boostable". T- independent immune responses are, in contrast, relatively poor in children and lack a booster response when a T- 5 independent antigen is repeatedly administered. The im unologic memory of T cells likely reflects two consequences of the first, "primary" or "sensitizing" limb of the immune response: (a) an expanded number of antigen- specific T cells that grow in response to antigen-bearing _Q dendritic cells, and (b) the enhanced functional properties of individual T cells that occurs after dendritic cell priming [Inaba et al. , Resting and sensitized T lymphocytes exhibit distinct stimulatory (antigen presenting cell) requirements for growth and lymphokine release; J.Exp.Med. 5 160:868-876 (1984); Inaba and Steinman, "Protein-specific ^ϋ helper T lymphocyte formation initiated by dendritic cells". Science 229: 475-479 (1985); Inaba et al., "Properties o memory T lymphocytes isolated from the mixed leukocyt

_ reaction", Proc.Natl.Acad.Sci. 82:7686-7690 (1985)].

Certain types of antigens characteristically elicit T 5 cell dependent antibody responses whereas others elicit a T cell independent response. For example, polysaccharide generally elicit a T-cell independent immune response. There is no memory response and therefore no protection t subsequent infection with the polysaccharide antigen. 0 Proteins, however, do elicit a T-cell dependent response i infants. The development of conjugate vaccines containin a polysaccharide covalently coupled to a protein convert the polysaccharide T-independent response to a T-dependen response. Unfortunately, little is known concerning th 5 sites on proteins which confer their T-cell dependen character, therefore hampering the design of more specifi immunogens.

As stated above, dendritic cells play a crucial role i the initiation of T-cell dependent responses. Dendriti o cells bind and modify antigens in a manner such that th modified antigen when presented on the surface of th dendritic cell can activate T-cells to participate in th eventual production of antibodies. The modification o antigens by dendritic cells may, for example, includ 5 fragmenting a protein to produce peptides which have region which specifically are capable of activating T-cells.

The events whereby cells fragment antigens int peptides, and then present these peptides in associatio with products of the major histocompatibility complex, (MHC Q are termed "antigen presentation". The MHC is a region o highly polymorphic genes whose products are expressed on th surfaces of a variety of cells. MHC antigens are th principal determinants of graft rejection. Two differen types of MHC gene products, class I and class II MH 5 molecules, have been identified. T cells recognize foreig ° antigens bound to only one specific class I or class II MHC molecule. The patterns of antigen association with class I or class II MHC molecules determine which T cells are stimulated. For instance, peptide fragments derived from extra cellular proteins usually bind to class II MHC 5 molecules, whereas proteins endogenously transcribed in dendritic cells generally associate with newly synthesized class I MHC molecules. As a consequence, exogenously and endogenously synthesized proteins are typically recognized by distinct T cell populations. 0 Dendritic cells are specialized antigen presenting cells in the immune response of whole animals (14,31). Again however, the ability to use dendritic cells to identify and extract the^"immunogenic peptides is hampered by the small numbers of these specialized antigen presenting 5 cells.

Particle uptake is a specialized activity of mononuclear and polymorphonuclear phagocytes. Dead cells, immune complexes, and microorganisms all are avidly internalized. Following fusion with hydrolase-rich o lysosomes, the ingested particles are degraded (60,61). This degradation must be to the level of permeable amino acids (62,63) and saccharides, otherwise the vacuolar apparatus would swell with indigestible materials (64,65). Such clearance and digestive functions of phagocytes 5 contribute to wound healing, tissue remodeling, and host de ense.

Another consequence of endocytosis, the processing of antigens by antigen presenting cells (APCs) , differs in many respects from the scavenging function of phagocytosis. _Q First, processing requires the generation of peptides at least 8-18 amino acids in length (66,67), while scavenging entails digestion to amino acids (62,63). Secondly, presentation requires the binding of peptides to MHC class II products (6,68), whereas scavenging does not require MHC 5 products. Thirdly, antigen presentation can function at a ° low capacity, since only a few hundred molecules of ligand need to be generated for successful stimulation of certain T-T hybrids (69,70) and primary T cell populations (71). During scavenging, phagocytes readily clear and destroy hundreds of thousands of protein molecules each hour (63) . 5 Lastly, antigen presentation is best carried out by cells that are rich in MHC class II but show little phagocytic activity and few lysosomes, i.e., dendritic cells and B cells, while phagocytes (macrophages and neutrophils) often have low levels of class II and abundant lysosomes. These 0 observations, together with the identification of antigenic specializations within the endocytic system of dendritic cells and B cells, have lead to the suggestion that the machinery required for antigen presentation may differ from that required for scavenging, both quantitatively and 5 qualitatively (31) .

In the case of dendritic cells, there have been indications that these APCs are at some point during their lifetime capable of phagocytic activity. Pugh et al. noted Feulgen-stained inclusions in some afferent lymph dendritic o cells and suggested that phagocytosis of other cells had taken place prior to entry into the lymph (16) . Fossum noted phagocytic inclusions in the interdigitating dendritic cells of the T cell areas in mice that were rejecting allogeneic leukocytes (71) . Reis e Sousa et al. (74) found 5 that freshly isolated epidermal Langerhans cells, which are immature but nonproliferating dendritic cells, internalize small amounts of certain particulates. Neither report, however, demonstrates or suggests the occurrence of phagocytosis when particles are administered to cultures of _Q proliferating dendritic cells.

Injection of dendritic cells pulsed with pathogenic lymphocytes into mammals to elicit an active immune response against lymphoma is the subject of PCT patent application WO 91/13632. In addition, Francotte and Urbain, Proc. Nat'l. 5 Acad. Sci.. USA 82:8149 (1985) reported that mouse dendritic cells, pulsed in vitro with virus and injected back into mice, enhances the primary response and the secondary response to the virus. Neither the report by Francotte and Urbain and patent application WO 91/13632 provide a practical method of using dendritic cells as an adjuvant to activate the immune response because both of these methods depend on dendritic cells obtained from spleen, an impractical source of cells for most therapies or immunization procedures. In addition, neither report provides a method to obtain dendritic cells in sufficient quantities to be clinically useful.

SUMMARY OF THE INVENTION This invention provides a method of producing a population of dendritic cell precursors from proliferating cell cultures. The method comprises (a) providing a tissue source comprising dendritic cell precursors; (b) treating the tissue source from (a) to increase the proportion of dendritic cell precursors to obtain a population of cells suitable for culture in vitro; (c) culturing the tissue source on a substrate in a culture medium comprising GM-CSF, or a biologically active derivative of GM-CSF, to obtain proliferating nonadherent cells and cell clusters; (d) subculturing the nonadherent cells and cell clusters to produce cell aggregates comprising proliferating dendritic cell precursors; and (e) serially subculturing the cell aggregates one or more times to enrich the proportion of dendritic cell precursors.

This invention also provides a method of producing in vitro mature dendritic cells from proliferating cell cultures. The method comprises (a) providing a tissue source comprising dendritic cell precursor cells; (b) treating the tissue source from (a) to increase the proportion of dendritic cell precursors in order to obtain a population of cells suitable for culture in vitro; (c) culturing the tissue source on a substrate in a culture medium comprising GM-CSF, or a biologically active derivative of GM-CSF, to obtain non-adherent cells and cell clusters; (d) subculturing the nonadherent cells and cell clusters to produce cell aggregates comprising proliferating dendritic cell precursors; (e) serially subculturing the cell aggregates one or more times to enrich the proportion of dendritic cell precursors; and (f) continuing to culture the dendritic cell precursors for a period of time sufficient to allow them to mature into mature dendritic cells. To reduce the proportion of non-dendritic precursor cells, the tissue source may be pretreated prior to culturing the tissue source on a substrate to obtain the non-adherent cells or during the early stages of the culture. Preferred tissue sources for the practice of the invention are bone marrow and, in particular, blood.

This invention also provides a method of increasing the proportion of dendritic cells present in the tissue source by pretreating the individual with a substance to stimulate he atopoiesis. When bone marrow is used as the tissue source the pretreatment step comprises killing cells expressin antigens which are not expressed on dendritic precursor cells by contacting the bone marrow with antibodies specific for antigens not present on dendritic precursor cells in a medium comprising complement. Removal of undesirable non- dendritic cell precursors may also be accomplished b adsorbing the undesirable non-dendritic or their precurso cells onto a solid support.

This invention also provides dendritic cell precursor and dendritic cells in amounts which may be use therapeutically and which also may be used to prepare ne therapeutic antigens. In addition, the dendritic cel precursors and dendritic cells prepared according to th method of this invention are also provided. Another embodiment of the invention are antigen activated dendritic cells prepared according to the method of the invention in which antigen-activated dendritic cells have been exposed to antigen and express modified antigens for presentation to and activation of T cells.

This invention also provides novel antigens which are produced by exposing an antigen to cultures of dendritic cells prepared according to the method of the invention in which the antigen is modified by the dendritic cells to produce modified antigens which are im unogenic fragments of the unmodified or native antigen and which fragments activate T cells.

These novel antigens may be used to immunize animals and humans to prevent or treat disease.

This invention also provides a method of preparing antigens from dendritic cell precursors comprising providing precursor dendritic cells from a population of precursor cells capable of proliferating, contacting the precursor cells with antigen for a period of time sufficient to allow the dendritic cell precursors to phagocytose the antigen and obtain antigen-containing dendritic cell precursors; culturing the antigen containing-dendritic cell precursors under conditions and for a period of time sufficient for the antigen to be processed and presented by dendritic cell precursors.

The antigens processed by the dendritic cell precursors as a result of phagocytosis may themselves be used alone or in combination with adjuvants including dendritic cell precursors to evoke an immune response in an individual to the antigen.

Also provided are compositions and methods for increasing the number of myeloic dendritic progenitor cells in blood in those individuals.

In a further embodiment, the yield of dendritic cell precursors is increased by culturing the precursors in a sufficient amount of GM-CSF and other cytokines to promote proliferation of the dendritic cell precursors. Other " cytokines include but are not limited to G-CSF, M-CSF, TNF- α, Interleukin-3, and Interleukin-lα, Inteukin-1/S, Interleukin 6 and stem cell factor.

In another embodiment, the invention provides self- peptide antigens produced by pulsing the dendritic cells of

5 the invention with a protein to which an individual has developed an immune response and extracting the relevant self-peptide or autoantigen.

This invention also provides a method of treating autoimmune diseases by treating an individual with 0 therapeutically effective amounts of self-peptides produced according to the method of the invention to induce tolerance to the self-proteins.

The treatment of autoimmune diseases comprising administering to an individual in need of treatment a 5 therapeutically effective amount of antigen-activate dendritic cells where the antigen is a self-protein o autoantigen is also provided.

The use of the compositions and methods of the invention to treat autoimmune diseases selected from th o group of juvenile diabetes, myasthenia gravis, and multipl sclerosis is also provided.

This invention also provides treatment for inflammator diseases in which the pathogenesis involves exaggerated cell mediated immune responses such as those present i 5 atopic dermatitis and contact dermatitis.

This invention also provides a method for providing a antigen to a host comprising exposing an antigen to culture of dendritic cells prepared according to the metho of this invention to produce antigen-activated dendriti _Q cells followed by inoculating the host with the antigen activated dendritic cells.

This invention further provides a method of activatin

T cells comprising the use of proliferating dendritic cell for capturing protein, viral, and microbial antigens in a 5 immunogenic form in situ and then presenting these antigen ⁰ in a potent manner to T cells either in vitro or in situ.

This invention additionally provides a method comprising the use of mature and precursor dendritic cells to present MHC class I and II products with antigen peptides. 5 This invention also provides a method for making antigenic peptides that are specific for an individual's MHC products thereby increasing the number of specialized stimulatory antigenic presenting cells available to provide an immunogenic response in an individual. 0 Also provided are compositions and methods to treat infectious diseases, including but not limited to diseases caused by mycobacteria including tuberculosis, bacteria, and viruses.

Compositions and methods for using dendritic cells or 5 dendritic cell precursors as vehicles for active immunization and immunotherapy in situ are also provided.

Vaccines comprised of any of the antigens or antigen- activated dendritic cells described above are also provided as are the methods of immunizing against disease in humans o or animals comprising administering any of the compositions of the invention.

An object of this invention is to provide a method of culturing dendritic cell precursors in vitro so that they evolve into mature dendritic cells suitable for use as 5 immunogens or adjuvants when combined with an antigen.

It is also an object of this invention to provide dendritic cell precursors capable of phagocytosing antigenic material to be processed and presented by the dendritic cell precursors. _Q Another object of this invention is to provide a convenient and practical source of sufficient quantities of dendritic cells and dendritic cell precursors to be useful in the treatment or prevention of disease.

Another object of this invention is to provide novel 5 immunogens comprising the dendritic cells or dendritic cell precursors of this invention which have been exposed to antigen and express modified antigen on their surface.

Another object of this invention is to provide antigens which have been modified through their exposure to dendriti cell precursors or dendritic cells and which modifie antigens are effective as T-cell dependent antigens.

A further objective of the invention is to provide a method of immunizing individuals with T-cell dependen antigens for the prevention and treatment of disease.

FIGURE LEGENDS

Fig. 1. Flow plan for inducing dendritic cell "colonies."

Fig. 2. FACS analyses bf dendritic cells released fro proliferating aggregates. Several mAbs which recogniz various cell surface determinants on dendritic cel precursors (23,24,28) are shown. Except for MHC class I an II products, the phenotype of the released cells i homogeneous. The staining with no primary mAb was identica to RB6 and RA3.

Fig. 3. FACS analyses of dendritic cell precursors tha could be dislodged by Pasteur pipetting of proliferatin aggregates, and dendritic cells released spontaneously i culture. The mAb are: Ml/42 anti-MHC class I [ATCC # TI 126]; NLDC145 anti-interdigitating cell (13); M5/114 anti MHC class II [ATCC # TIB 120]; 33D1 anti-dendritic cel [ATCC # TIB 227]; B5-5 anti-thy-1. The staining with anti MHC mAbs is bimodal, but the released cell fraction o dendritic cells is richest in expression of MHC class I an II.

Fig. 4. MLR stimulating activity of populations isolate from the GM-CSF stimulated mouse blood cultures [see text] " Fig. 5. Progressive development of MLR stimulating activity in bone marrow cultured in the presence of GM-CSF.

Ia-negative precursors, B and T cell-depleted marrow cells were cultured in GM-CSF with 3/4 of the medium being replaced every 2d. At each time point, the cells were dislodged by gently pipetting. After irradiation, graded doses of marrow cells were applied to 3 x 10⁵ allogeneic

[C57BL/6, left] or syngeneic [BALB/C x DBA/2 Fl] T cells and cultured for 4 days in the MLR. 3H-TdR uptake was measured at 80-94h [values are means of triplicates with standard error bars] .

Fig. 6. Physical properties of the MLR stimulating cells that develop in GM-CSF supplemented bone marrow cultures [see text] . A. Cultures similar to those in Fig 5 were separated into nonadherent [open symbols] and loosely adherent fractions [closed symbols] , the latter being cells that could be dislodged by gently pipetting over the monolayer. For the d4 separations, loosely adherent cells [mainly 0 granulocytes] were rinsed away at d2, and for the d6 separation, granulocytes were rinsed away at d2 and d4. The cells were irradiated and applied in graded doses to allogeneic T cells as in Fig. 5.

B. At the indicated time points, free cells and cell 5 aggregates were dislodged from the stromal monolayer and separated by lg sedimentation. The aggregates were cultured for 1 day to provide released cells. These cells were irradiated and tested as MLR stimulators, as were firmly adherent cells that were dislodged in the presence of 10 mM Q EDTA [open squares].

Fig. 7. Cell cytofluorometry of the development of la- positive cells from aggregates within bone marrow cultures supplemented with GM-CSF. 5 ^u GM-CSF stimulated, bone marrow cultures [left, unfractionated] were compared with loosely attached cel aggregates [middle] and cells released from the aggregate after overnight culture [right] . The cells were taken a day 4 or day 6, so that the released cells were analyzed a

5 day 5 and day 7. The cells were stained with no primary mA [no lry] , or with mAb to granulocytes [RB-6] or MHC class I products [B21-2] followed by FITC-mouse anti-rat Ig.

Fig. 8. Detailed cell cytofluorometric phenotype analysi 0 of the la-positive cells released from the growing dendriti cell aggregates. Contaminating, la-negative granulocyte were gated out on the basis of lower forward light scatter so that one could examine the expression of many surfac antigens on the larger cells using rat and hamster anti 5 mouse mAbs (7,17) as indicated.

Fig. 9. Quantitation of developing cells that bear th dendritic cell restricted granule antigens 2A1 and M342.

Dendritic cells contain intracellular granules tha o react with the mAb such as M342 and 2A1 (34) mAbs. Ia negative nonlymphocytes from mouse marrow were cultured i GM-CSF, and the loosely adherent granulocytes rinsed away a d2 and d4. The data on day 2 and 4 represent cells tha could be dislodged by pipetting, while the data on d3 an 5 d5-8 were cells released from the monolayer. At each of th indicated time points, at least 500 cells were counted i cytospins prepared and stained. [See text]. When culture are started at 5 x 10⁵ cells/cm² and fed with 3/4 volum fresh medium every 2 days, the yields of total and Ia⁺ cell Q were at d2, 1.05 x 10⁶ and 2.1 x 10⁴, at d4 1.81 x 10⁶ an 2.12 x 10⁵, and at d6, 1.54 x 10⁶ and 3.21 x 10⁵.

Fig. 10. Progenitor-progeny relationships in growin dendritic cells. Growing aggregates were separated at d 5 from bone marrow cultures and pulsed with ³H-TdR at 0. μCi/ l, 3 x 10⁵ cells/well, for 12h. All wells were replaced with fresh medium and returned to culture for 1, 2, or 3 days of chase. The yields of released cells during the chase were 2.0, 2.9, and 3.0 x 10^s respectively per well. The content of Ia⁺ cells was 28% after the pulse, and 47%, 55%, and 62% on days 1, 2, and 3 respectively. The data are shown as percentage of cells that were radiolabeled, with the filled in bars being cells that express the 2A1 granule cell antigen of mature dendritic cells.

Fig. 11. Diagram of the proposed pathway of dendritic cell development in marrow cultures supplemented with GM-CSF. A proliferating aggregate forms from a precursor that either attaches to the cell stroma or is itself adherent. During dendritic cell differentiation, which is evident at the periphery of the aggregate and in cells released therefrom, there is a progressive increase in cell processes, MHC class II, NLDC-145 surface antigen, and M342 and 2A1 intracellular antigen [see text] and a progressive decrease in adherence to plastic.

Fig 12: Diff-Quick stains of developing dendritic cells that have been exposed to latex and carbon.

A. An aggregate of developing dendritic cells cytospun after a 2Oh exposure to 2u latex spheres. Many cells in the aggregate are labeled with the uniform latex particles [arrows] .

B. Same as A, but the cultures were chased for a day to allow the production of mature single dendritic cells. Many of the released dendritic cells contain the uniform and lucent latex spheres arranged around a clear cut centrosphere [arrows] .

C. Same as A and B, but the aggregates were pulsed with colloidal carbon and then chased for a day in carbon-free medium. The centrosphere of some of the mature dendritic cells that release from the aggregate contain small but clear cut endocytic granules of black, indigestible phagocytic tracer [arrows].

D. Mature dendritic cells were exposed to carbon after they had been produced from proliferating aggregates. Carbon deposits are not evident.

Fig 13: Uptake of BCG into developing dendritic cells using two-color labels for acid fast bacilli and dendritic cell antigens. Clusters of developing dendritic cells [6d marrow cultures induced with GM-CSF] were exposed for 2Oh to BCG. The monolayers were washed and chased in medium with GM-CSF for 2d. The cells were dissociated, labeled with FITC-anti- I-A mAb, and the class II-rich cells were isolated by cell sorting [most of the cells in the culture are class II-rich as shown previously (16)]. The sorted cells were cytospun, stained with auramine-rhodamine to visualize the cell- associated BCG, and double labeled with a different mAb an immunoperoxidase.. The left and right panels of each pair are phase contrast and acid fast views respectively. Arrows on the left indicate the location of the bacilli on th right. The label for class II, [I-A and I-E, M5/114] outlines the cell processes better than the dendritic cell restricted NLDC-145 antibody.

Fig 14: Electron microscopy of BCG in dendritic cells. As in Fig 2, BCG was added to GM-CSF stimulated d6 bon marrow cultures for a day. After washing and 2 more days o culture, the released cells were processed for electro microscopy. A,B. Low power views to show the typical dendriti cells with numerous processes and a few phagocytosed BC [arrows] .

C,D. Higher power views to show phagosomal membrane against the BCG, as well as organelles of the dendritic cel centrosphere including endocytic vacuoles [E], Golg apparatus [GA] , and small vesicles with a dense core [*]. Fig 15: Antigen presentation to CFA primed/IFA primed T cells.

T cells were purified from lymph nodes that drain paws that had been primed with complete [CFA] or incomplete [IFA]

Freunds adjuvant. The different APCs are listed. Mature dendritic cells are d8 bone marrow cultures, and immature dendritic cells are from d5-6 cultures.

Fig 16: Antigen presentation to naive lymph node T cells in situ. Growing cultures of bone marrow dendritic cells were pulsed with BCG at d5-6, and used immediately or after a 2d chase culture to activate T cells. The populations were injected into the paws of naive mice without artificial adjuvants. Five days later the draining lymph nodes were taken and stimulated in vitro with graded doses of PPD or BSA (the dendritic cells had been grown with fetal calf serum) , the BSA to serve as a nonparticulate antigen. Data are means and standard errors for groups of 5 mice, each studied separately. Control lymph nodes not exposed to BCG pulsed dendritic cells did not respond to PPD or to BSA (<2000 cpm) .

Fig 17: Antigen presentation to naive spleen cells in situ. Growing cultures of bone marrow dendritic cells were pulsed with BCG at d5-6 (immature) , at d7-8 (mature) , or at d5-6 followed by a 2d chase. 10⁶ cells of each group were injected i.v. into groups of mice. 5 or 10 days later, the spleen cells were cultured in vitro with graded doses of PPD or BSA as antigen. Since the dendritic cells were cultured in FCS, the use of BSA serves as control to ensure that all dendritic cell populations were comparably immunogenic in vivo. Unprimed spleen did not respond to either BSA or PPD. Figs 18A, B and C: Mixed Leukocyte Reaction (MLR) assay of human dendritic cells produced according to the metho described in Example 6. Graded doses of irradiated cell (30 to 30,000 in serial 3 fold dilutions) were added to 2 10⁵ accessory cell-depleted T cells. The T cell response o cells that had been cultured the absence of added cytokin (X) ; and in the presence of GM-CSF (o) ; GM-CSF + IL-lα (•) ; GM-CSF + TNF-α (□) ; GM-CSF + TNF-α + IL-lα (■) ; GM-CSF + IL 3 (Δ) ; and GM-CSF + IL-3 + IL-1 (▲ ) was measured with a 16 pulse of ³H-thymidine on the 5th day. The response of non dendritic cells is also shown in C, ( ♦ ) . Three differen experiments, A, B, and C are presented. Patients providin cells for experiments A & B were pretreated with G-CSF; patient in experiment C was pretreated with GM-CSF. Cytokines were used at the following concentrations: rh GM-CSF, 400 or 800 U/ml; rhu IL-lα, 50 LAF units/ml (IL-l was present in cultures only during the last 24 hours prio to harvesting the cells) ; rhu TNF-α 50 U/ml; and rhu IL- 100 U/ml. The values on the X axis represent the number o dendritic cells except for X where dendritic cells wer absent and the number is equivalent to total cell number. Standard deviations of triplicate cultures were <10% of th mean, and are not shown.

DETAILED DESCRIPTION OF THE INVENTION This invention relates to a method of producin cultures of proliferating dendritic cell precursors whic mature in vitro to mature dendritic cells. The dendriti cells and the dendritic cell precursors produced accordin to the method of the invention may be produced in amount suitable for various immunological interventions for th prevention and treatment of disease.

The starting material for the method of producin dendritic cell precursors and mature dendritic cells is tissue source comprising dendritic cell precursors whic precursor cells are capable of proliferating and maturing i vitro into dendritic cells when treated according to the method of the invention. Such precursor cells are nonadherent and typically do not label with mAb markers found on mature dendritic cells such as Ia antigens, 2A1 and M342 antigens (34, 44) and the NLDC145 interdigitating cell source antigen (13) . Preferably such tissue sources are spleen, afferent lymph, bone marrow and blood. More preferred tissue sources are bone marrow and blood. Blood is also a preferred tissue source of precursor cells because it is easily accessible and could be obtained in relatively large quantities.

To increase the number of dendritic precursor cells in animals, including humans it is preferable to treat such individuals with substances which stimulate he atopoiesis. Such substances include G-CSF, GM-CSF and may include other factors which promote hematopoiesis. The amount of hematopoietic factor to be administered may be determined by one skilled in the art by monitoring the cell differential of individuals to whom the factor is being administered. Typically, dosages of factors such as G-CSF and GM-CSF will b^e similar to the dosage used to treat individuals recovering from treatment with cytotoxic agents. Preferably, GM-CSF or G-CSF is administered for 4 to 7 days at standard doses prior to removal of source tissue to increase the proportion of dendritic cell precursors. (Editorial, Lancet. 339: March 14, 1992, 648-649). For example, we have determined that dosages of G-CSF of 300 micrograms daily for 5 to 13 days and dosages of GM-CSF of 400 micrograms daily for 4 to 19 days have resulted in significant yields of dendritic cells. Fetal or umbilical cord blood, which is also rich in growth factors is also a preferred source of blood for obtaining precursor dendritic cells.

According to a method of the invention, the tissue source may be treated prior to culturing to enrich the proportion of dendritic precursor cells relative to other cell types. Such pretreatment may also remove cells whic may compete with the proliferation of dendritic precurso cells or inhibit their proliferation or survival Pretreatment may also be used to make the tissue source mor suitable for in vitro culture. The method of treatment wil likely be tissue specific depending on the particular tissu source. For example, spleen or bone marrow if used as tissue source would first be treated so as to obtain singl cells followed by suitable cell separation techniques t separate leukocytes from other cell types. Treatment o blood would involve cell separation techniques to separat leukocytes from other cells types including red blood cell (RBCs) which are toxic. Removal of RBCs may be accomplishe by standard methods known to those skilled in the art. I addition, antitoxins such as anti-erythroid monoclonal VIE 64 antibody which bind RBCs may be used to facilitat binding of RBC to a substrate for removal using a pannin technique.

According to a preferred method of this invention, whe bone marrow is used as the tissue source, B cells ar removed prior to culturing of bone marrow in GM-CSF. Whil B cells and pre-B cells do not grow in response to GM-CSF they represent approximately 50% of the initial marro suspension and thereby preclude the use of staining wit anti-la monoclonal antibodies to quickly enumerate dendriti cells. Additionally, granulocytes are GM-CSF responsive an readily proliferate in the presence of GM-CSF. As such, th B cells and granulocytes mask the presence of dendritic cel precursors. B cells can express the M342 and 2A1 granula antigens that are useful markers for distinguishin dendritic cells from macrophages and granulocytes Moreover, granulocytes have a tendency to overgrow th cultures and compete for available GM-CSF. The mos preferred method under this invention is to remove th majority of nonadherent, newly-formed granulocytes from th bone marrow cultures by gentle washes during the first 2- days in culture.

Preferably, in one form of pretreatment cells which compete and mask the proliferation of precursor dendritic cells are killed. Such pretreatment comprises killing cells expressing antigens which are not expressed on dendritic precursor cells by contacting bone marrow with antibodies specific for antigens not present on dendritic precursor cells in a medium comprising complement. Another form of pretreatment to remove undesirable cells suitable for use with this invention is adsorbing the undesirable precursor cells or their precursors onto a solid support using antibodies specific for antigens expressed on the undesirable cells. Several methods of adsorbing cells to solid supports of various types are known to those skilled in the art and are suitable for use with this invention. For example, undesirable cells may be removed by "panning" using a plastic surface such as a petri dish. Alternatively, other methods which are among those suitable include adsorbing cells onto magnetic heads to be separated by a magnetic force; or immunobeads to be separated by gravity. Non adsorbed cells containing an increased proportion of dendritic cell precursors may then be separated from the cells adsorbed to the solid support by known means including panning. These pretreatment step serves a dual purpose: they destroy or revives the precursors of non-dendritic cells in the culture while increasing the proportion of dendritic cell precursors competing for GM-CSF in the culture.

In addition, la-positive cells, i.e. B cells and macrophages preferably are killed by culturing the cells in the presence of a mixture of anti la-antibodies, preferably monoclonal antibodies, and complement. Mature dendritic cells which are also present in bone marrow are also killed when the cells from the bone marrow are cultured in the presence of anti la-antibodies, however, these mature dendritic cells occur in such low quantities in the blood ^ϋ and bone marrow and possess such distinct antigenic markers from dendritic cell precursors that killing of these mature dendritic cells will not significantly effect the proliferation and yield of dendritic cell precursors. T and B cells as well as monocytes which also may be present in the bone marrow may be killed by including antibodies directed against T and B cell antigens and monocytes. Such antigens include but are not limited to CD3, CD4, the B cell antigen B220, thy-1, CD8 and monocyte antigens. The remaining viable cells from the bone marrow are then cultured in medium supplemented with about 500-1000 U/ml GM- CSF and cultured as described below. It should be noted that CD4 and CD8 antigens may be present on young dendritic cell precursors, therefore, antibodies directed to these antigens may deplete the dendritic cell precursor populations.

When blood is used as a tissue source, blood leukocytes may be obtained using conventional methods which maintain their viability. According to the preferred method of the invention, blood is diluted into medium (preferably RPMI) 0 containing heparin (about 100 U/ml) or other suitable anticoagulant. The volume of blood to medium is about 1 t 1. Cells are pelleted and washed by centrifugation of the blood in medium at about 1000 rpm (150g) at 4°C. Platelets and red blood cells are depleted by suspending the cell 5 pellet in a mixture of medium and ammonium chloride. Preferably the mixture of medium to ammonium chloride (final concentration 0.839 percent) is about 1:1 by volume. Cells are pelleted by centrifugation and washed about 2 more times in the medium-ammonium chloride mixture, or until 0 population of leukocytes, substantially free of platelet and red blood cells, is obtained.

Any isotonic solution commonly used in tissue cultur may be used as the medium for separating blood leukocyte from platelets and red blood cells. Examples of suc isotonic solutions are phosphate buffered saline. Hank ° balanced salt solution, or complete growth mediums including for example RPMI 1640. RPMI 1640 is preferred.

Cells obtained from treatment of the tissue source are cultured to form a primary culture on an appropriate substrate in a culture medium supplemented with GM-CSF or a 5 GM-CSF derivative protein or peptide having an amino acid sequence which sequence maintains biologic activity typical of GM-CSF. The appropriate substrate may be any tissue culture compatible surface to which cells may adhere. Preferably, the substrate is commercial plastic treated for use in tissue culture. Examples include various flasks, roller bottles, petri dishes and multi-well containing plates made for use in tissue culture. Surfaces treated with a substance, for example collagen or poly-L-lysine, or antibodies specific for a particular cell type to promote 5 cell adhesion may also be used provided they allow for the differential attachment of cells as described below. Cells are preferably plated at an initial cell density of about 7.5 X 10⁵ cells per cm². At this dose, the surface is not fully covered by cells, but there are no big spaces (2-3 0 cell diameters) either.

When bone marrow which has been treated to reduce the proportion of non-dendritic cell precursors is cultured, aggregates comprising proliferating dendritic cell precursors are formed. The la-negative marrow 5 nonlymphocytes comprising dendritic cell precursors are preferably cultured in high numbers, about 10⁶/well (5 x 10⁵ cells/cm²) Liquid marrow cultures which are set up for purposes other than culturing dendritic cell precursors are typically seeded at 1/lOth this dose, but it is then 0 difficult to identify and isolate the aggregates of developing dendritic cells.

The growth medium for the cells at each step of the method of the invention should allow for the survival and proliferation of the precursor dendritic cells. Any growth medium typically used to culture cells may be used according to the method of the invention provided the medium is supplemented with GM-CSF. Preferred medias include RPMI 1640, DMEM and α-MEM, with added amino acids and vitamins supplemented with an appropriate amount of serum or a defined set of hormones and an amount of GM-CSF sufficient to promote proliferation of dendritic precursor cells. Serum-free medium supplemented with hormones is also suitable for culturing the dendritic cell precursors. RPMI 1640 supplemented with 5% fetal calf serum (FCS) and GM-CSF is preferred. Cells may be selected or adapted to grow in other serums and at other concentrations of serum. Cells from human tissue may also be cultured in medium supplemented with human serum rather than FCS. Medias may contain antibiotics to minimize bacteria infection of the cultures. Penicillin, streptomycin or gentamicin or combinations containing them are preferred. The medium, or a portion of the medium, in which the cells are cultured should be periodically replenished to provide fresh nutrients including GM-CSF.

GM-CSF has surprisingly been found to promote the proliferation in vitro of precursor dendritic cells. Cells are cultured in the presence of GM-CSF at a concentration sufficient to promote the survival and proliferation of dendritic cell precursors. The dose depends on the amount of competition from other cells (especially macrophages and granulocytes) for the GM-CSF, or to the presence of GM-CSF inactivators in the cell population. Preferably, the cells are cultured in the presence of between about 1 and 1000 U/ml of GM-CSF. More preferably cells from blood are cultured in the presence of GM-CSF at a concentration of between about 30 and 100 U/ml. This dose has been found to be necessary and sufficient for maximal responses by cells obtained from mouse blood. Most preferably, cells ar cultured in the presence of GM-CSF at a concentration o about 30 U/ml. GM-CSF at a concentration of between abou 400-800 U/ml has been found to be optimal for culturin proliferating human dendritic cells from blood. Cells from bone marrow require higher concentrations of GM-CSF because of the presence of large numbers of proliferating granulocytes which compete for the available GM-CSF, therefore, doses between about 500-1000 U/ml are preferred for cultures of cells obtained from marrow.

When suspensions of mouse bone marrow are cultured in the presence of GM-CSF, three types of myeloid cells expand in numbers. (1) Neutrophils predominate but do not adhere to the culture surface. Neutrophils have a characteristic nuclear morphology, express the RB-6 antigen, and lack MHC class II products. (2) Macrophages are firmly adherent to the culture vessel, express substantial levels of the F4/80 antigen, and for the most part express little or no MHC class II [but see below] . When mouse or human blood leukocytes are cultured in GM-CSF at 30 U/ml or 400-800 U/ml, respectively, the cultures develop a large number of aggregates or cell balls from which typical dendritic cells are eventually released. In the absence of GM-CSF, no colonies develop. Cytologic criteria may be used to initially detect the dendritic cells which characteristically extend large, sheet-like processes or veils (25-27) .

GM-CSF may be isolated from natural sources, produced using recombinant DNA techniques or prepared by chemical synthesis. As used herein, GM-CSF includes GM-CSF produced by any method and from any species. "GM-CSF" is defined herein as any bioactive analog, fragment or derivative of the naturally occurring (native) GM-CSF. Such fragments or derivative forms of GM-CSF should also promote the proliferation in culture of dendritic cell precursors. In addition GM-CSF peptides having biologic activity can be identified by their ability to bind GM-CSF receptors on appropriate cell types.

It may be desirable to include additional cytokines in the culture medium in addition to GM-CSF to further increase the yield of dendritic cells. Such cytokines include granulocyte colony-stimulating factor (G-CSF) , monocyte- acrophage colony-stimulating factor (M-CSF) , interleukins 1 α and 1 3, 3 and 6 (IL-1 α, IL-1/3, IL-3 and IL-6, respectively) , tumor necrosis factor α (TNFα) , and stem cell factor (SCF) . Cytokines are used in amounts which are effective in increasing the proportion of dendritic cells present in the culture either by enhancing proliferation or survival of dendritic cell precursors. Preferably, cytokines are present in the following concentrations: IL- lα and β, 1 to 100 LAF units/ml; TNF-α, 5-500 U/ml; IL-3, 25-500 U/ml; M-CSF, 100-1000 U/ml; G-CSF, 25-300 U/ml, SCF, 10-100 ng/ml; and IL-6, 10-100 ng/ml. More preferred concentrations of cytokines are: IL-lα, 50 LAF units/ml; TNFα, 50 U/ml; IL-3, 100 U/ml; M-CSF, 300 U/ml; and G-CSF, 100 U/ml. Preferred cytokines are human proteins. Most preferred cytokines are produced from the human gene usin reco binant techniques (rhu) . (TNFα) at concentrations fro about 10-50 U/ml may be used to increase dendritic cell yields several fold. The primary cultures from the tissue source are allowe to incubate at about 37°C under standard tissue cultur conditions of humidity and pH until a population of cell has adhered to the substrate sufficiently to allow for th separation of nonadherent cells. The dendritic cel precursor in blood initially is nonadherent to plastic, i contrast to monocytes, so that the precursors can b separated after overnight culture. Monocytes an fibroblasts are believed to comprise the majority o adherent cells and usually adhere to the substrate withi about 6 to about 24 hours. Preferably nonadherent cells ar separated from adherent cells between about 8 to 16 hours. Most preferably nonadherent cells are separated at about 1 hours. Any method which does not dislodge significan quantities of adherent cells may be used to separate th adherent from nonadherent cells. Preferably, the cells ar dislodged by simple shaking or pipetting. Pipetting is most preferred.

To culture precursor cells from human blood from this primary culture, cells which have been depleted of cells that are not dendritic cell precursors are cultured on a substrate at a density of preferably about 5 X 10⁵ cells per cm². After 5 days, with feedings every other day, cell aggregates appear (also referred to as "balls") . These aggregates may then be treated as described below.

The nonadherent cells from the primary culture are subcultured by transferring them to new culture flasks at a density sufficient to allow for survival of the cells and which results in the development over time of clusters of growing cells that are loosely attached to the culture surface or to the firmly adherent cells on the surface. These clusters are the nidus of proliferating dendritic cell precursors. As used herein "culture flasks" refers to any vessel suitable for culturing cells. It is desirable to subculture all of the nonadherent cells from the primary culture at a density of between about 2 X 10⁵ cells and 5 X -⁵ cells per cm². Preferably at about 2.5 X 10⁵ per cm². Cells are incubated for a sufficient time to allow the surface of the culture dish to become covered with a monolayer of tightly adherent cells including macrophages and fibroblasts affixed to which are aggregates of nonadherent cells. At this time, any nonadherent cells are removed from the wells, and the cellular aggregates are dislodged for subculturing. Preferably the cells from the aggregates are subcultured after about 10 days or when the number of aggregated cells per cm² reaches about 3 to 4 X 10⁵.

For serially subculturing the aggregated cells, the aggregated cells are dislodged from the adherent cells and the aggregated cells are subcultured on a total surface area of preferably between about 2 to 5 times that of the surface area of the parent culture. More preferably the cells are subcultured on a surface area that is about 3 times the surface area of the parent culture. Cells having sheet-like processes typical of dendritic cells appear in the culture at about 4-7 days. Between about day 10 and day 17 of culture the number of single cells that can be recovered from a given surface area doubles. Both dendritic cell precursors and mature dendritic cells are present in the aggregates.

For producing dendritic cell from bone marrow, preferably the distinctive aggregates of proliferating, less mature dendritic cells are separated away from the stroma at about after about 4-6d of culture. Large numbers of dendritic cells are released and it is this released population that expresses the cardinal features of mature dendritic cells. Because bone marrow initially contains a greater proportion of dendritic cell precursors than blood, only about 4-6 days of culture of the cells obtained from bone marrow are necessary to achieve about the same number of cells which are obtained after about 10 to 25 days of culture of cells obtained from blood. To further expand the blood derived population of dendritic cells, cell aggregates may be serially subcultured multiple times at intervals which provide for the continued proliferation of dendritic cell precursors. Preferably, aggregates are subcultured prior to the release into the medium of a majority of cells having the dendritic cell morphology, for example between about 3 and 30 days. More preferably aggregates of cells are subcultured between about 10 to 25 days in culture, and most preferably at 20 days. The number of times the cells are serially subcultured depends on the number of cells desired, the viability of the cells, and the capacity of the cultures to continue to produce cell aggregates from which dendritic cells are released. Preferably, cells can be serially subcultured for between about 1 to 2 months from when the nonadherent cells were subcultured or between about one to five times. More preferably cells are serially subcultured about two to three times. Most preferably cells are serially subcultured twice.

According to a preferred method, to serially subculture the cells of the primary and subsequent cultures, cells are dislodged by pipetting most of the aggregates of growing dendritic cells as well as some cells in the monolayer of growing macrophages and fibroblasts. Pipetting usually disrupts the aggregates, particularly the peripheral cells of the aggregates which are more mature. With time in culture, e.g., at 2 weeks, the aggregates of the growing dendritic cells become more stable and it is possible to dislodge the aggregates for separation by lg sedimentation.

Alternative approaches may be used to isolate the mature dendritic cells from the growing cultures. One is to remove cells that are nonadherent and separate the aggregates from cells attached to substrate and single cells by lg sedimentation. Dendritic cells are then released in large numbers from the aggregates over an additional 1-2 days of culture, while any mature dendritic cells can be isolated from other single cells by floatation on dense metrizamide as described (Freudenthal and Steinman, Proc. Natl. Acad. Sci. USA 87:7698-7702, 1990). The second method, which is simpler but essentially terminates the growth phase of the procedure, is to harvest all the nonadherent cells when the aggregates are very large, leave the cells on ice for about 20 minutes, resuspend vigorously with a pipette to disaggregate the aggregates and float the mature dendritic cells on metrizamide columns.

Typically the contents of five 16 mm wells are applied to a 6 ml column of 50% FCS -RPMI 1640 in a 15 ml conical tube [Sarstedt, 62.553.002 PS]. After at least 20 min, the applied medium and top 1 ml of the column are removed. RPMI is added, the aggregates are pelleted at 1000 rpm at 4° for 5 min, and the cells are suspended gently for subculture in fresh medium. Various techniques may be used to identify the cells present in the cultures. These techniques may include analysis of morphology, detecting cell type specific antigens with monoclonal antibodies, identifyin proliferating cells using tritiated thymidin autoradiography, assaying mixed leukocyte reactions, an demonstrating dendritic cell homing.

The dendritic cells besides being identified by thei stellate shape may also be identified by detecting thei expression of specific antigens using monoclonal antibodies. A panel of monoclonal antibodies may be used t identify and characterize the cells in the GM-CSF expande cultures. The monoclonal antibodies are reviewed elsewher (23, 24 which are incorporated herein by reference).

Among the specific monoclonal antibodies suitable fo identifying mature dendritic cells are: l) those which bin to the MHC class I antigen (Ml/42 anti-MHC class I [ATCC TIB 126]); 2) those which bind to the MHC class II antige (B21-2 anti-MHC class II [ATCC # TIB 229]; M5/114 anti-MH class II [ATCC # TIB 120]); 3) those which bind to hea stable antigen (Ml/69 anti-heat stable antigen [HSA, ATC #T1B 125]); 4) 33D1 anti-dendritic cell antibodies [ATCC TIB 227]; 5) those which bind to the interdigitating cel antigen (NLDC145 anti-interdigitating cell (13) ; and 6) those which bind to antigens in granules in the perinuclea region of mature dendritic cells (monoclonal antibodies 2A and M342, (23) Agger et al.). Other antigens which ar expressed by the dendritic cells of the invention and whic may be used to identify mature dendritic cells are CD4 (identified with monoclonal antibody 2D2C) , and CDll (identified with monoclonal antibody Ml/70. The Ml/69, Ml/70, Ml/42 monoclonal antibodies are described i Monoclonal antibodies. NY, Plenum 1980, ed. R. Kennett e al. pages 185-217 which is incorporated herein by reference. Those of skill in the art will recognize that othe antibodies may be made and characterized which are suitabl for identifying mature dendritic cells. Similarly, the production of dendritic precursor cells also facilitates the production of antibodies specific for dendritic precursor cells.

To identify and phenotype the proliferating cells and their progeny, cultures may be labelled with tritiated thymidine to identify the cells in the S phase of mitosis. In addition to labelling the cells with a mitotic label, cells may also be co-labelled with monoclonal antibodies to determine when markers associated with mature dendritic cells are expressed. The distinctive phenotype of the dendritic cell precursors is stable so that for example, the dendritic cell progeny do not become macrophages even when maintained in macrophage colony stimulating factor (M-CSF) .

Another index of dendritic cell maturity is the ability of mature dendritic cells to stimulate the proliferation of T-cells in the mixed leukocyte reaction (MLR) . The ability of dendritic cells to migrate to lymph nodes, i.e., dendritic cell homing is another index of dendritic cell maturation which may be used to assess the maturity of the cells in culture.

The criteria that have become evident for identifying dendritic precursor cells according to the invention enables the identification of proliferating progenitors of dendritic cells in other organs. It is known that proliferating precursors give rise to the rapidly turning over populations of dendritic cells in spleen (15) and afferent lymph (16) . The proliferation of leukocytes [other than T cells] occurs in the bone marrow, but it may be that for dendritic cells, the marrow also seeds the blood and other tissues with progenitors which then proliferate extensively as shown here. By being able to prepare the otherwise trace dendritic cell in large numbers according to the method of this invention, other previously unexplored areas of dendritic cell function may now be determined. Specifically, growing dendritic cells will facilitate molecular and clinical studies on the mechanism of action of these APCs, including their capacities to capture and retain antigens in an immunogenic form and act as adjuvants for the generation of immunity in vivo.

There is an increased interest in the use of constituent proteins and peptides to modulate T cell responses to complex microbial and cellular antigens in situ. Typically artificial adjuvants such as alum are required to produce a maximum immunogenic effect. Several antigens are known to be immunogenic when administered in association with dendritic cells but in the absence of additional adjuvants (1) . The immunogenicity of dendritic cells in situ has been shown with for example contact allergens (45) , transplantation antigens (46-49) , and more recently foreign proteins (31,50,51). Other types of antigens include but are not limited to microbial, tumor and viral antigens. Dendritic cells serve directly as APCs in s_itu, because the T cells that are primed are restricted to recognize only antigens presented by the particular MHC class of the immunizing dendritic cells rather than host APCs (14,31,50,51). These observations, when coupled wit data that dendritic cells are efficient at capturing protein antigens in an immunogenic form in situ (52-54) , allow these APCs to be considered "nature's adjuvant". This invention therefore enables the utilization of dendritic cells by disclosing methods and compositions suitable for providin sufficient quantities of dendritic cell precursors in orde to take advantage of their unique antigen presentin capabilities in clinical and therapeutic practices.

Dendritic cells are capable of processing comple antigens into those peptides that would be presented by sel MHC products. Among the preferred embodiments of ou invention is a method for using dendritic cells whereby th dendritic cell precursors internalize particulates during a early stage in their development from proliferatin progenitors. We have established that stimulation of bon marrow suspensions with GM-CSF leads to the production of clusters of proliferating dendritic cell precursors. The cells that pulse label with 3H-thymidine in the clusters lack many of the characteristic markers of dendritic cells, e.g., stellate shape and antigenic features like NLDC-145 antigen and high levels of MHC class II. In pulse chase experiments, 3H-thymidine-labeled progeny with all the features of dendritic cells are released. We have found that cells within the aggregate also are phagocytic, and that in analogous pulse chase protocols, the progeny dendritic cells are clearly labeled with the phagocytic meal. When the particles are BCG organisms such as those causing tuberculosis, mycobacterial antigens associatedwith the dendritic cells are presented in a potent manner to T cells in vitro and in situ. Foreign and autoantigens are processed by the dendritic cells of the invention to retain their immunogenic form. The immunogenic form of the antigen implies processing the antigen through fragmentation to produce a form of the antigen that can be recognized by and stimulate T cells. Preferably, such foreign or autoantigens are proteins which are processed into peptides by the dendritic cells. The relevant peptides which are produced by the dendritic cells may be extracted and purified for use as immunogens.

Peptides processed by the dendritic cells may also be used as toleragens to induce tolerance to the proteins processed by the dendritic cells or dendritic cell precursors. Preferably when used as toleragens, the processed peptides are presented on dendritic cells which have been treated to reduce their capacity to provoke an immune response as by inhibiting their accessory function by blocking accessory molecules such as B7 present on the dendritic cells.

The antigen-activated dendritic cells of the invention are produced by exposing antigen, in vitro, to the dendritic cells prepared according to the method of the invention. Dendritic cells are plated in culture dishes and exposed t antigen in a sufficient amount and for a sufficient perio of time to allow the antigen to bind to the dendritic cells. The amount and time necessary to achieve binding of th antigen to the dendritic cells may be determined b i munoassay or binding assay. Other methods known to thos of skill in the art may be used to detect the presence o antigen on the dendritic cells following their exposure t antigen.

Without being bound by theory, the information a present suggests that the development of dendritic cell proceeds by the following pathway [Fig. 11]. The dendriti cell precursors in both blood and marrow lack MHC class I antigens as well as B and T cell and monocyte markers [B220 CD3, thy-l, CD4/8], and the precursors are nonadherent. Th precursors attach to the stroma and give rise to aggregate of class II positive cells. Perhaps the growing aggregate arise from a subset of strongly class II-positive cells tha are found in the firmly adherent monolayer even at late time points. However, these firmly adherent, class II ric cells lack the MLR stimulatory activity of dendritic cell and may express substantial levels of FC receptors and th F4/80 antigen. The final stage of development is that th loosely attached aggregate releases mature, nonproliferatin dendritic cells. The latter have even higher levels of MH class II [Fig. 2-3] and can attach transiently to plastic much like many of the dendritic cells released from splee (25) . As development occurs in the aggregate, there seem to be a reduction in the levels of cytoplasmic staining fo FC receptors and F4/80 antigen, and an increase in granul [M342, 2A1] and surface antigens [33D1, NLDC145] that ar characteristic of dendritic cells. Lastly, accessor function for primary T-dependent immune responses increase as cells are released from the growing aggregates.

Mature dendritic cells, while effective in sensitizin T cells to several different antigens, show little or n phagocytic activity. To the extent that endocytosis is required for antigen processing and presentation, it was not previously evident how dendritic cells would present particle-associated peptides. Based on our work, it is now evident tha progenitors to dendritic cells which this invention provides can internalize such particles for processing and presentation. The types of particles which may be internalized by phagocytosis include bacteria, viral, mycobacteria or other infectious agents capable of causing disease. Accordingly, any antigenic particle which is internalized and processed by the dendritic cell precursors of this invention is also suitable for making the various immunogens, toleragens and vaccines described as part of this invention. Processing of antigen by dendritic cells or dendritic cell precursors includes the fragmentation of an antigen into antigen fragments which are then presented.

Phagocytoses of particulate matter by dendritic cell precursors may be accomplished by culturing the dendritic cell precursors in the presence of particulate matter for a time sufficient to allow the cells to phagocytose, process and present the antigen. Preferably, culturing of the cells in the presence of the particles should be for a period of between 1 to 48 hours. More preferably, culturing cells in the presence of particulate matter will be for about 20 hours. Those of skill in the art will recognize that the length of time necessary for a cell to phagocytose a particle will be dependent on the cell type and the nature of the particle being phagocytosed. Methods to monitor the extent of such phagocytosis are well known to those skilled in the art. Cells should be exposed to antigen for sufficient time to allow antigens to be internalized and presented on the cell surface. The time necessary for the cells to internalize and present the processed antigen may be determined using pulse-chase protocols in which exposure to antigen is followed by a wash-out period. Once the minimum time necessary for cells to express processed antigen on their surface is determined, a pulse-chase protocol may be used to prepare cells and antigens for eliciting immunogenic responses.

The phagocytic dendritic precursor cells are obtained by stimulating cell cultures comprising dendritic precursor cells with GM-CSF to induce aggregates of growing dendritic cells. These dendritic precursor cells may be obtained from any of the source tissues containing dendritic cell precursors described above. Preferably, the source tissue is bone marrow or blood cultures. Cells within these aggregates are clearly phagocytic. If the developing cultures are exposed to particles, washed and "chased" for 2 days, the number of MHC-class II rich dendritic cells increases substantially and at least 50% contain internalized particles such as BCG mycobacteria or latex particles. The mycobacteria-laden, newly developed, dendritic cells are much more potent in presenting antigens to primed T cells than corresponding cultures of mature dendritic cells that are exposed to a pulse of organisms. A similar situation pertains when BCG-charged, dendritic cells are injected into the footpad or bloo stream of naive mice. Those dendritic cells that hav phagocytosed organisms induce the strongest T cell responses to mycobacterial antigens in draining lymph node and spleen. The administration of antigens to GM-CSF induced, developin dendritic cells — by increasing both antigen uptake an cell numbers — will facilitate the use of these APCs fo active immunization in situ. The production of such stron immunogenic responses due to the presentation of antigen b the dendritic cells makes these cells and this syste particularly desirable as adjuvants useful for producin immunogenic responses in individuals. Such immunogeni responses and the development of antibodies to the presente antigens may be used to treat ongoing infections or preven future infections as with a vaccine. The use of dendriti cells to produce a therapeutic or prophylactic immune response in an individual may be particularly useful to treat or prevent infection by drug resistant organisms, such as, for example, the BCG mycobacterium causing tuberculosis. Immunogenicity of ingested particles can be obtained with BCG mycobacteria (Fig 12-13) . In any inoculum of the BCG vaccine, there are live bacilli [approximately 50% of the bacilli act as colony forming units], dead bacilli, and probably a number of mycobacterial proteins. The phagocytosed pool of BCG is being presented to T cells by dendritic cells. This is evident after comparing the presentation of mycobacterial antigens with bovine serum albumin (BSA) , a component of the serum in which the dendritic cells are grown. All the APC populations were comparable in presenting BSA, but dendritic cells that had phagocytosed the most BCG were the most effective APCs for mycobacteria (Fig 12 and 13, ♦). BCG particle uptake, therefore, accounts for the bulk of the mycobacterial priming by the dendritic cell precursors.

Another embodiment of this invention is therefore to pulse dendritic cell precursors with mycobacteria tuberculosis bacteria antigen, including for example BCG antigen, to induce host resistance to mycobacteria infection, a matter of importance given the need to develop better vaccination and treatment protocols for tuberculosis, including the drug resistant variety (78) .

In effect, the pulse and chase protocol which may be used to charge developing dendritic cells with organisms according to our invention allows the two broad components of immunostimulation to take place sequentially. These components are a) antigen capture and presentation, here the capture of particulates by immature dendritic cells, and b) development of potent accessory or immunostimulatory functions during the chase period. The situation is comparable to that seen in the handling of soluble proteins (4,6) and particles (74) by epidermal Langerhans cells. Each of the two broad components of APC function entails many subcomponents. For example, immature dendritic cells not only are more phagocytic but display other features needed for antigen presentation such as active biosynthesis of abundant MHC class II molecules and invariant chain (6,7) and numerous acidic endocytic vacuoles (36) .

The capacity to charge APCs with antigens using pulse chase protocols may be a special feature of dendritic cells. Prior studies with macrophages and B cells had suggeste that T cell epitopes are short-lived (75) . The results described here and elsewhere (6,14,71) indicates tha immunogenic peptides can be long lived on dendritic cells at least 2 days prior to injection into mice. This retention capacity should enable dendritic cells to migrate an sensitize T cells in draining lymphoid tissues over a perio of several days (14,50,51).

An important feature of the dendritic cells of thi invention is the capacity to efficiently present microbial and other antigens on both class I and II products. In th case of BCG, the bulk of the primed cells are CD4+ T cells, most likely because the antigenic load is handled by th endocytic pathway and MHC class II products (76) . In th case of influenza, it has been found that the class pathway for inducing CD8+ cytotoxic T lymphocytes (CTL) requires adequate delivery of antigen (infectious virus) into the cytoplasm, whereas the purely endocytic pathwa delivers noninfectious virions for presentation only to CD4 helpers (77) . Developing dendritic cell cultures provide an opportunity for charging MHC class I products wit peptide, since cell proliferation allows various methods o gene insertion (as with retroviral vectors) to be applied.

According to this further embodiment of the invention, the proliferating dendritic cells may be injected with vector which allows for the expression of specific protein by the dendritic cells. These viral proteins which ar expressed by the dendritic cell may then be processed an presented on the cell surface on MHC I receptors. The viral antigen-presenting cells or the processed viral antigens themselves may then be used as immunogens to produce an immunogenic response to the proteins encoded by the vector. Vectors may be prepared to include specific DNA sequences which code and express genes for proteins to which an immunogenic response is desired. Preferably, retroviral vectors are used to infect the dendritic cells. The use of retroviral vectors to infect host cells is known to those skilled in the art and is described in WO 92/07943 published May 14, 1992 and in Richard C. Mulligan, "Gene Transfer and Gene Therapy:Principle, Prospects and Perspective" in Enolocry of Human Disease at the DNA Level. Chapter 12. J. Linsten and A. Peterson, eds. Rover Press, 1991 which are both incorporated herein by reference. By using developing dendritic cells to charge MHC class I and/or II products, several desirable components of T cell modulation in situ can be achieved. Antigen uptake and presentation by immature progenitors, allows the APC to tailor the peptides that are appropriate for an individual's MHC products, and increases the number of specialized stimulatory APCs. These properties of dendritic cell progenitor populations meet many of the demands for using cells as vehicles for active immunization and immunotherapy in situ. The present invention provides for the first time a method of obtaining dendritic cells in sufficient quantities to be used to treat or immunize animals or humans with dendritic cells which have been activated with antigens. In addition, dendritic cells may be obtained in sufficient quantities to be useful as reagents to modify antigens in a manner to make the antigens more effective as T-cell dependent antigens.

To use antigen-activated dendritic cells as a therapeutic or immunogen the antigen-activated dendritic cells are injected by any method which elicits an immune response into a syngeneic animal or human. Preferably, dendritic cells are injected back into the same animal o human from whom the source tissue was obtained. Th injection site may be subcutaneous, intraperitoneal, intramuscular, intradermal, or intravenous. The number o antigen-activated dendritic cells reinjected back into th animal or human in need of treatment may vary depending o inter alia, the antigen and size of the individual. A ke feature in the function of dendritic cells in situ is th capacity to migrate or home to the T-dependent regions o lymphoid tissues, where the dendritic cells would be in a optimal position to select the requisite antigen-reactive cells from the pool of recirculating quiescent lymphocyte and thereby initiate the T-dependent response.

According to the preferred method of stimulating a immune response in an individual, a tissue source from tha individual would be identified to provide the dendritic cel precursors. If blood is used as the tissue sourc preferably the individual is first treated with cytokine t stimulate hematopoieses. After isolation and expansion o the dendritic cell precursor population, the cells ar contacted with the antigen. Preferably, contact with th antigen is conducted in vitro. After sufficient time ha elapsed to allow the cells to process and present th antigen on their surfaces, the cell-antigen complexes ar put back into the individual in sufficient quantity to evok an immune response. Preferably between 1 X 10⁶ and 10 X 10 antigen presenting cells are injected back into th individual.

The novel antigens of the invention are prepared b combining substances to be modified or other antigens wit the dendritic cells prepared according to the method of th invention. The dendritic cells process or modify antigen in a manner which promotes the stimulation of T-cells by th processed or modified antigens. Such dendritic cel modified antigens are advantageous because they can be mor specific and have fewer undesirable epitopes than non- modified T-dependent antigens. The dendritic cell modified antigens may be purified by standard biochemical methods. For example, it is known to use antibodies to products of the major histocompatibility complex (MHC) to select MHC- antigenic peptide complexes and then to elute the requisite processed peptides with acid [Rudensky et al., Nature 353:622-7 (1991); Hunt et al., Science 255: 1261-3 (1992) which are incorporated herein by reference] .

Antigen-activated dendritic cells and dendritic cell modified antigens may both be used to elicit an immune response against an antigen. The activated dendritic cells or modified antigens may by used as vaccines to prevent future infection or may be used to activate the immune system to treat ongoing disease. The activated dendritic cells or modified antigens may be formulated for use as vaccines or pharmaceutical compositions with suitable carriers such as physiological saline or other injectable liquids. The vaccines or pharmaceutical compositions comprising the modified antigens or the antigen-activated dendritic cells of the invention would be administered in therapeutically effective amounts sufficient to elicit an immune response. Preferably, between about 1 to 100 micrograms of modified antigen, or its equivalent when bound to dendritic cells, should be administered per dose. The present invention also provides a method and composition for treating autoimmune disease. Such autoimmune diseases include but are not limited to juvenile diabetes, multiple sclerosis, myasthenia gravis and atopic dermatitis. Without being bound by theory, it is believed that autoimmune diseases result from an immune response being directed against "self-proteins", i.e., autoantigens that are present or endogenous in an individual. In an autoimmune response, these "sel -proteins" are being presented to T cells which cause the T cells to become "self-reactive". According to the method of the invention. dendritic cells are pulsed with the endogenous antigen t produce the relevant "self-peptide". The relevant self peptide is different for each individual because MH products are highly polymorphic and each individual MH molecules might bind different peptide fragments. Th "self-peptide" may then be used to design competing peptide or to induce tolerance to the self protein in the individua in need of treatment.

Because dendritic cells can now be grown fro precursors according to the methods and principle identified here, and because dendritic cells can modif antigens to produce killer T cells, the compositions of thi invention are particularly useful as vaccines toward viruses and tumor cells for which killer T cells migh provide resistance.

EXAMPLES

Example 1 Production of Mouse Dendritic Cells In Vitr

From Proliferating Dendritic Cell Precursor From Blood

MATERIALS

A. Mice: BALB/C, BALB/C X DBA/2 Fl, BALB/C x C57BL/6 Fl, C57BL/6 x DBA/2 Fl, and C57BL/6 males and females, 6-8 week of age were purchased from Japan SLC Ine [Shizuoka, Japan], the Trudeau Institute [Saranac Lake, NY], and Charles Rive Wiga [Sulzberg, FRG] . Four preparations of rGM-CSF wer evaluated with similar results, the yield of dendritic cell reaching a plateau with 30-100 U/ml. The preparations wer from Dr. S. Gillis, Immunex Corp, Seattle WA; Genetic Institute [supernatant from COS cells transfected with mGM CSF; used at 30U/ml or greater] ; and Dr. T. Sud [supernatant from CHO cells transfected with the expressio vector, pHSmGM-CSF (22), and E.Coli expressed material]. B. Blood Preparation: Blood was obtained by cardiac puncture or from the carotid artery. The blood was diluted in, or allowed to drip into, RPMI-1640 with 100 u/ml heparin [about 2 ml/mouse]. Blood cells were pelleted at 1000 rpm at 4°, resuspended in RPMI 1640, and sedi ented again. The pellet was suspended in 1 ml RPMI 1640 per mouse and mixed with an equal volume of 1.66% ammonium chloride in distilled water to lyse the red cells. After 2 min at room temperature, the suspension was spun at 1000 rpm at 4°. The pellet, which still contained red cells, was resuspended again in 0.5 ml RPMI and 0.5 ml NH₄C1 for 2 min, diluted in RPMI, and sedimented again. After 2 more washes, most platelets and red cells had been depleted and a population of blood leukocytes, had been obtained.

C. Aggregates of proliferating dendritic cells from blood supplemented with GM-CSF

Blood leukocytes, usually from CxD2 Fl mice, were cultured in 16 mm tissue culture wells [24 well dishes, Costar, #25820] in medium (1 ml per well) supplemented with GM-CSF at 30 U/ml and at 1.5 x 10⁶ cells/well. The medium was RPMI 1640 supplemented with 5% fetal calf serum [JRH Biosciences, Lenexa, KA] , 50 uM 2-ME, 20 ug/ml gentamicin, and recombinant mouse GM-CSF. After overnight culture, many monocytes adhered and the nonadherent cells were transferred to new 16 mm wells. The adherent cells did not develop dendritic cell colonies, but during the next week, the nonadherent populations exhibited three changes. First, most of the lymphocytes and granulocytes died or could be removed by washing. Second, the surface of the well became covered with a monolayer of tightly adherent cells that included macrophages and fibroblasts. Third, affixed to scattered sites on the monolayer, there developed small aggregates of cells. The cultures were fed with GM-CSF (30 u/ml) at day 6-7 and then every 3 days by aspirating 0.5- 0.75 ml of the medium and adding back an equal volume of fresh medium with GM-CSF. The aggregates continued t expand in number and size. At about day 10, the cells wer ready to be subcultured. Any residual loose cells could b rinsed off prior to dislodging the aggregates into fres medium and GM-CSF. About 0.8-1 million dislodged cells pe original well were divided into 3 subculture wells.

Most of the aggregates disassembled during this firs subculture, while the bulk of the adherent monolaye remained attached to the original well. Upon transfer, mos of the cells in the dislodged aggregates adhered as singl cells to the new culture well but over a period of 2-3 days aggregates reappeared. The aggregates again were affixed t adherent stromal cells, but these adherent cells were muc less numerous than the dense monolayer in the origina culture. Over the next 4-7 days, aggregates filled th wells. These colonies were often larger than those of th original wells and were covered with many sheet-lik processes typical of dendritic cells. It was more difficul to count cells at this point, since many of the aggregate contained a core of tightly associated cells. However, th number of single cells that could be recovered per wel expanded about 2 fold between days 10 and 17 of culture.

If the cultures were allowed to overgrow, some cell with the morphology of dendritic cells were released. Mor typically, the cells were not allowed to overgrow and th aggregates were dislodged and subcultured again at about 2 days. Prior to subculture, the aggregates could be purifie from free cells by lg sedimentation. Such separations wer more easily performed with longer periods of culture, i.e. it was easier to isolate intact aggregates at 3 vs. 2 vs. week of culture. With additional subculturing, the numbe of aggregates that were produced per well was progressivel reduced. However colonies of growing cells, as confirmed b 3H-TdR labeling and autoradiography [below] , could b generated in subcultures for 1-2 months. Followin subculturing at 2-3 weeks, typical single dendritic cell were now released into the medium. By direct observation with video recording, these released cells had the active motility of dendritic cells, continually extending and retracting large veils or sheet-like processes. In the presence of continued GM-CSF, one observed both free dendritic cells as well as expanding colonies. In the absence of GM-CSF, only free dendritic cells were released and the aggregates essentially fell apart and did not reform in the medium and colonies of aggregates did not develop. The yields of free dendritic cells per subculture ranged from 0.3 - 2.5 x 10⁵.

In summary, from a starting blood mononuclear culture of 1.5 x 10⁶ cells, where dendritic cells were difficult to detect, we on average obtained 5-10 subcultures each with at least 3-10 x 10⁴ released dendritic cells at 3 weeks, as well as many aggregates capable of further proliferation. Therefore aggregates of growing cells were developing in mouse blood supplemented with GM-CSF, and these aggregates were covered with dendritic cells many of which could be released spontaneously into the medium.

D. Phenotype of the cell aggregates and dendritic cells released therefrom

Cytospin preparations were made in a Shandon cytocentrifuge using 3-10 x 10⁴ cells. The slides were stored with desiccant prior to fixation in acetone and staining with mAb followed by peroxidase mouse anti-rat lg [Boehringer Mannheim Bioche icals , #605-545] or rabbit anti- hamster lg [Accurate Chemical & Scientific Corp, # JZY-036- 003]. The preparations were stained with Giemsa and mounted in Permount for bright field analysis. For cytofluorography [FACScan, Becton Dickinson], aliquots of cells were stained with primary rat or hamster mAb followed by FITC mouse anti-rat lg [Boehringer, #605-540] or biotin rabbit anti-hamster lg [Accurate, JZY-066-003] and FITC- avidin. Cytospin preparations of 2-3 week cultures were examined with a panel of mAb and an immunoperoxidase method. The released cells, and many of the cells that could be dislodged from the periphery of the aggregate, were similar in their stellate shape and phenotype. Most of the cells stained strongly with mAb to MHC class II, the CD45 leukocyte common antigen, CR3 receptor CDllb, and heat stable antigen (HSA) , and CD44. Staining with mAbs to the Fc receptor [2.4G2] and macrophage F4/80 antigen (MAC) was weak or undetectable in >95% of the cells. The cultures contained only rare B cells [B220 mAb, RA-3], T cells [thy-1 mAb, B5-5], or granulocytes [GRAN, mAb RB6] . Some cells at the periphery of the aggregate, and many of the cells that were released fro the aggregates, were stained with two markers that are largely restricted to dendritic cells. The interdigitating cell antigen [mAb NLDC 145 (13), IDC], which also binds to thymic epithelium, stained many but not all of the dendritic profiles. Virtually all of the dendritic profiles stained with mAbs 2A1 and M342 stain granules in the perinuclear region of mature dendritic cells, B lymphocytes, as well as interdigitating cells in sections through the T areas of lymphoid organs. Macrophages from many sites [blood monocytes; peritoneal cavity macrophages; macrophages in sections of lymph node, thymus, spleen] do not contain 2A1 or M342-reactive granules. Cytofluorography was used to gain semi-quantitative information on the expression of antigens at the cell surface. A panel of mAb were applied to two populations: cells that could be dislodged from the aggregates by Pasteur pipetting, and cells that were released spontaneously when the aggregates were subcultured for 1 day. These "dislodged" and "released" populations were identical in their dendritic shape and in phenotype but for some exceptions that are considered below. The phenotype of the released cells is shown in Fig 2, and the few differences between aggregated and released cells are in Fig 3. Virtually all the dendritic cells developing in and from the aggregates expressed high levels of the leukocyte common [CD45, mAb Ml/9.3] and heat stable [mAbs Ml/69 and Jlld] antigens, as well as high levels of CD44 and CDllb [mAb Ml/70] . Low levels of the following antigens were detected on the cell surface: the dendritic cell antigen 33D1, the macrophage marker F4/80, the Fcγ receptor antigen 2.4G2, the p55 IL-2 receptor CD25 antigen 3C7, and the CDllc integrin N418 [Fig 2] . These antigens were noted on all cells by FACS even though many of the antigens like F4/80 and 2.4G2 were weak or absent in the cytoplasm with an immunoperoxidase method. Several antigens were absent: RB6 granulocyte, RA3 B cell, B5-5 thy-1, GK 1.5 CD4, and SER-4 marginal zone macrophage^' [Fig 2].

Expression of class I and II MHC products by the dendritic cells in these cultures was very high but nonetheless bimodal [Figs 2 and Fig 3]. Most of the dendritic cells that were dislodged from the aggregates had somewhat lower levels of MHC class I and II, while dendritic cells that were released from the aggregates had very high levels of MHC products. The other marker that was different in the released and loosely attached dendritic cells was NLDC 145 which was higher in the released population. [Fig 3, top panels]. We conclude that the phenotype of the cells that arise from the proliferating aggregates is very much like that seen in cultured dendritic cells from skin, spleen, and thymus (24,28) with the exception that the Ml/70 CDllb marker is more abundant.

E. 3H-TdR autoradiography to verify growth of dendritic cell precursors

After 2 and 3 weeks in liquid culture, the wells contained numerous expanding aggregates of cells, and in some cases were already releasing nonadherent dendritic cells in large numbers. Cultures were labeled with 3H- thymidine to identify and phenotype the proliferating cells and their progeny. For pulse labeling, 3H-TdR was added to the cultures [6 Ci/mM, 1 uCi/ml final]. 2h later, the medium was replaced with 3H-TdR free medium, and the cultures were separated into nonadherent released cells and residual adherent aggregates for examination on cytospin preparations [Shandon Ine, Pittsburgh PA, #59900102]. The cytospin cells were stained for specific antigens with mAb and immunoperoxidase as above. Also, the slides were dipped in photographic emulsion [Kodak autoradiography emulsion type NTB2 #165-4433] for exposure [5 days] prior to development, staining with Giemsa, and mounting in Permount. For pulse chase experiments, a lower dose of 3H-TdR was used to maintain cell viability, but the cells were handled similarly otherwise. The pulse was applied at 0.1 uCi/ml for 2h or for 16h, the latter to provide higher initial labeling indices. The cells were washed and chased for 1-3 days prior to harvesting and analysis as above with immunoperoxidase, autoradiography, and Giemsa staining.

The 2 and 3 week cultures were exposed to 3H-TdR and examined for proliferative activity. The labeled cells were washed, spun onto slides, and the cytospins stained with mAb and an immunoperoxidase method prior to dipping and exposure to photographic emulsion. Important markers were mAbs 2A1 and NLDC-145 which recognize intracellular granules and a cell surface antigen in mature dendritic cells respectively. When cultures were labeled with a 2h pulse of 3H-TdR, it was apparent that the labeling index in the aggregates was very high, at least 10-15% of the profiles in the aggregates being in S phase. In contrast, if 3H-TdR was applied to cultures that were releasing typical nonadherent dendritic cells, the released fraction contained only rare labeled profiles. If GM-CSF was removed, 3H-TdR labeling ceased within a day. Virtually all the 3H-TdR labeled cells in the aggregate failed to label with mAb to markers found on mature dendritic cells i.e., 2A1 and NLDC145. The level of staining with anti-MHC class II mAb was less on the cells in S-phase than in the released dendritic cell populations [not shown] .

Pulse chase experiments were then done to establish that labeled cells in the aggregate were giving rise to typical dendritic cells. Cultures were first exposed to a low dose of 3H-TdR, either for 2h or for 16h, the latter to label a larger percentage of the cells in the aggregates. The wells were washed free of radiolabel, and then the aggregates were dislodged and separated from free cells by lg sedimentation. The aggregates were transferred to fresh medium without radiolabel, and over the next 1-3 days of culture, many dendritic cells were released into the medium. When the "chased" cultures were examined, several findings were apparent. The labeling index remained high, i.e., most of the progeny of cells that were proliferating in the aggregates were not being lost from the cultures. Second, the grain counts were diluted several fold from those apparent in the original pulse. Third, cells expressing the markers of mature dendritic cells [NLDC145, the 2A1 granular antigen, high levels of MHC class II] were now radiolabeled. Therefore the cellular aggregates that GM-CSF was inducing in cultured mouse blood were actively proliferating and releasing nonproliferating progeny with many of the typical cytologic and antigenic features of mature dendritic cells including the 2A1 granular antigen, the NLDC145 marker, and high levels of MHC class II.

F. Accessory cell function for T cell proliferative responses

MLR stimulating activity was monitored in the GM-CSF treated blood cultures. Cells from the blood cultures were exposed to 1500 rads [137Cs] and applied in graded doses to 3 x 10⁵ purified syngeneiσ or allogeneic T cells in 96 well, flat-bottomed microtest wells. The T cells were nylon wool nonadherent, spleen and lymph node suspensions that were treated with anti-la plus Jlld mAbs and complement to remove residual APC. 3H-TdR uptake was measured at 72-86 h [6 Ci/mM, 4 uCi/ml final].

Initially there was little or no MLR stimulatin activity [Fig 4, iV] . Some stimulating activity was noted a day 1 of culture [Fig 4, O] . An examination of cytospi preparations revealed that these 1 day nonadherent bloo cells had a low [<0.3%] but clear subset of la-rich, dendritic profiles. By day 7, when the proliferatin aggregates were first evident on the monolayer, th stimulating activity of the dislodged aggregates ha increased further, but was still 100 times less in specifi activity than typical dendritic cells [Fig 4, compare Δ an

•] even though most of the cells at day 7 and subsequen time points were MHC class II positive. By day 14, at whic time typical nonadherent dendritic cells were just beginnin to be released from the aggregates, the nonadheren population had considerable MLR stimulating activity, [Fi

4, V]. After 3 weeks, typical mature dendritic cells ha become abundant, and these indeed stimulated comparably t their splenic counterparts [Fig 4, compare 0 and •] . Othe cells in the culture, such as those dislodged from th aggregates, were about 10 fold less active than dendriti cells [Fig 4, ♦] . We conclude that the aggregates o proliferating dendritic cells have some MLR stimulatin activity but that it is the mature released cells that ar fully potent, some 100-300 times more active on a per cel basis than the populations in the starting culture at 1- days. During day 7-20 of culture, total cell numbers als expanded at least 5-10 fold.

G. Homing activity of dendritic cells in vivo

A second specialized feature of dendritic cells i their capacity to home to the T areas of peripheral lymphoi tissues (8,10). Dendritic cells or other cell types wer labeled at 2-10 x 10⁶/ml with carboxyfluorescein for 10 mi on ice [Molecular Probes C-1157; 30 uM final concentration in Hanks balanced salt solution (HBSS) with 5% FCS], washed in RPMI 1640, and injected in a volume of 50 ul RPMI-1640 into the foot pads. One day later, the draining popliteal lymph nodes were removed, frozen in OCT medium, and sectioned [10 μ] in a cryostat. To sample the entire node, we took duplicate specimens at regular intervals. The sections were applied to multiwell slides [Carlson Scientific microslides #111006], stored at -20°C, dried in a desiccator 30' prior to use [or left at room temp overnight], fixed in acetone, and stained with a peroxidase conjugated rabbit anti-FITC antibody [Dakopatts, P404]. To verify that the dendritic cells in the lymph node were in the T-dependent areas as described (8) , we added appropriate mAb to B cell, T cells, macrophages, or dendritic cells and visualized the latter with alkaline phosphatase conjugated mouse anti-rat lg [Boehringer Mannheim, #605-5357] plus a chromogen kit [Biomeda Corp, Foster City CA #S04]. We then blocked endogenous peroxidase with "Endo Blocker" [Biomeda Corp, #M69] followed by the peroxidase anti-FITC as above. Blood leukocytes, even when given at a dose of 10⁶ cells per footpad, failed to home to the lymphoid organ. When we tested dendritic cells that had been generated with GM-CSF from blood, homing to the T area was observed with injections of 200,000 cells. The selective localization to the T areas was confirmed by double labeling the specimens with mAb that stain B cells or T cells. Therefore dendritic cells produced in culture have the key functional features of this lineage: homing to the T-dependent regions and strong accessory activity.

H. Requirements for generating dendritic cell colonies from blood

The surface phenotype of the blood cell that gives rise to the dendritic cell colonies was assessed by treating the starting population with antibodies and complement. Treatment with either 33D1 anti-dendritic cell, anti-MH class II, or anti thy-1 did not eliminate the colony formin unit [not shown] . Instead, removal of thy-l⁺ or Ia⁺ cell enriched colony numbers several fold. CSF's other than GM CSF were also tested, either at the start of the 1-3 wee culture, or upon transfer of 2-3 week old aggregates to for veiled cells. None of the CSF's tested, i.e., IL-3, M-CSF G-CSF, SCF, supported the formation of colonies or matur dendritic cells. Therefore the growing dendritic colonie are very much dependent upon GM-CSF. In an effort to identify proliferating precursors t the dendritic cell system, we set up cultures from severa tissues that lacked mature dendritic cells and supplemente these with different growth factors particularly the CSF' [M-CSF, G-CSF, IL-3, GM-CSF, IL-1, and SCF]. Dendritic cel precursors were not observed from neonatal epidermis, whi contains mainly la" Langerhans cells (29) . To avo overgrowth of granulocytes in bulk bone marrow cultur which may make the identification of typical cell coloni or large numbers of dendritic cells difficult, it preferred to remove the nonadherent, proliferati granulocytes on days 2 and 4. Blood, which has few typic dendritic cells in the mouse (30) , proved to be ve effective for obtaining dendritic cell precursors. Growi cell aggregates appeared after about 6 days in culture, a these were often covered with profiles having the unusu and motile processes of dendritic cells. With time, typic nonadherent dendritic cells were released. The latter h the morphology and movement of dendritic cells as previous described in cultured mouse spleen, mouse skin, lymph fr several species, and human blood (25-27) . Therefore identify proliferating dendritic cells, it seems critical begin with an appropriate starting population, preferab blood, and to supplement the culture with GM-CSF.

Without wishing to be bound by any theory, we thi that the initial aggregates that appeared in the cultur ° represented clones, since very small groups of 4-6 cells were observed early on e.g., day 5. We tried to prove that the aggregates were clonal by mixing blood cells from strains that were distinguished with markers to polymorphic antigens like CD44 and MHC class II. However we could not 5 complete the experiments since we found that mouse strains differed in the number and speed with which colonies developed. BALB/C and DBA [and Fl strains derived therefrom] were the most active; B6 and BIO were several times less active; and strains like CBA/J, C3H/He, and A/J 0 were poor sources of proliferating, dendritic cell aggregates.

The precursors to the aggregates of proliferating dendritic cells were not typical monocytes or dendritic cells, because the number of aggregates that developed could 5 be increased substantially if one depleted monocytes by adherence or la-positive cells with antibody and complement. Without wishing to be bound by theory, we tentatively conclude that blood contains an la-negative precursor that forms a proliferating aggregate. In the aggregate, o dendritic cells mature and are released as nonproliferating progeny.

The formation of aggregates of dendritic cells required exogenous GM-CSF. If the aggregates were placed in macrophage or granulocyte-restricted CSF's [M-CSF, G-CSF], 5 proliferation ceased and neither macrophages nor granulocytes were formed. Because the cultures contained macrophages and some stromal cells, in addition to the dendritic cell aggregates, it was possible that other cytokines were being produced that were critical to the 0 formation of dendritic cells. It appears however that the cells in the aggregates have lost responsiveness to M- and G-CSF, and that dendritic cells represent a distinct myeloid pathway of development. Perhaps, without wishing to be bound by theory, the pathway originates from a common 5 precursor in which the dendritic cell lineage is an offshoot that no longer responds to macrophage and granulocyte restricted CSF's.

Labeling with 3H-thymidine, using pulse and pulse-chase protocols, was important in establishing the precursor- product relationships that were taking place in these liquid cultures. In a 2h pulse, virtually every labeled cell lacked two typical markers of mature dendritic cells, i.e., the NLDC-145 interdigitating cell surface antigen (13) and the recently identified 2A1/M342 granular cytoplasmic antigens (34) . These mAb do not stain most macrophage populations that we have examined either as isolated cells [blood, spleen, peritoneal macrophages] or in sections [thymic cortex, spleen red pulp, lymph node medulla] . In pulse chase protocols, large numbers of labeled progeny were released from the aggregates, and these released cells were nonadherent, motile, and strongly stimulatory in the MLR. After combined autoradiography and immunoperoxidase labeling, the labeled progeny carried the granular antigens, the NLDC-145 antigen, and very high levels of MHC class II. Each of these cytologic and antigenic markers are largely restricted to dendritic cells.

Without wishing to be bound by theory, we believe that maturation to typical nonproliferating dendritic cells occurred within the aggregate. The aggregates were covered with cells with the sheetlike or veiled processes of dendritic cells. Cells with markers of mature dendritic cell markers [high MHC class II, 2A1 positive granules, NLDC antigen] were also observed at the periphery of the cell aggregates. However, it was difficult to isolate the aggregate intact, i.e. , without dislodging these more mature cells. The mechanism whereby dendritic cells matured and left the aggregate was not clear. Maturation was enhanced in older cultures [>2 weeks] or by removing adherent stroma cells. Both proliferation and maturation was blocked if the cultures contained too many fibroblasts.

The functional maturation that occurred in the ° proliferating aggregate is striking. The dendritic cells that were generated in culture were potent MLR stimulators. 100 dendritic cells induced a much stronger primary MLR than 100,000 blood leukocytes. The increase in stimulating activity per la-positive cell was at least 2 logs between the time that the aggregates first appeared and the time that typical dendritic cells were released in large numbers. Over this time period, cell recovery increased 5-10 fold. Also the dendritic cell progeny homed in a precise way to the T cell area of lymph node, another functional property that was not detectable in blood cells [data not shown].

Example 2 Generation of Large Numbers of Dendritic

Cells From Mouse Bone Marrow Cultures Supplemented With GM-CSF

5 MATERIALS

A. Mice: Female BALB/C, male DBA/ 2 , and female C57BL/6 mice, 7 wks old, were purchased from Japan SLC [Hamamatsu, Shizuoka, Japan]. BALB/C x DBA/2 Fl, of both sexes 7-10 wks old, were from Japan SLC and the Trudeau Institute, Saranac 0 Lake, NY.

Reagents: The culture medium was RPMI-1640 [Nissui, Tokyo, Japan; GIBCO, Grand Island, NY] supplemented with 5% FCS, 50 μM 2-Mercaptoethanol, and 20 μg/ml genta icin. Murine rGM- CSF [10⁸U/mg protein] was kindly provided by Kirin Brewery 5 Co [Maebashi, Gumma, Japan]. A panel of rat and hamster mAbs to mouse leukocyte antigens is described elsewhere (23, 24) . FITC- and peroxidase-conjugated mouse anti-rat IgG were purchased from Boehringer Mannheim [Indianapolis, IN] and FITC- and peroxidase-conjugated goat anti-hamster lg [γ 0 and L-chain] were from Jackson Immunoresearch Lab [Westgrove, PA] and Caltag [San Francisco, CA] respectively.

B. Bone marrow cultures: After removing all muscle tissues with gauze from the mouse femurs and tibias, the 5 bones were placed in a 60 mm dish with 70% alcohol for 1 min, washed twice with PBS, and transferred into a fresh dish with RPMI-1640. Both ends of the bones were cut with scissors in the dish, and then the marrow was flushed out using 2 ml of RPMI-1640 with a syringe and 25G needle. The tissue was suspended, passed through nylon mesh to remove small pieces of bone and debris, and red cells were lysed with ammonium chloride. After washing, lymphocytes and la- positive cells were killed with a cocktail of mAbs and rabbit complement for 60 min at 37°C. The mAbs were GK 1.5 anti-CD4, HO 2.2 anti-CD8, B21-2 anti-la, and RA3-3A1/6.1 anti-B220/CD45R all obtained from the ATCC [TIB 207, 150, 229, and 146 respectively]. 7.5-10 x 10⁵ cells were placed in 24 well plates [Nunc, Naperville, IL] in 1 ml of medium supplemented with 500-1000 U/ml rGM-CSF. The cultures were usually fed every 2d for about 2 to 10 days, by gently swirling the plates, aspirating 3/4 of the medium, and adding back fresh medium with GM-CSF. An object of these washes was to remove nonadherent granulocytes without dislodging clusters of developing dendritic cells that were loosely attached to firmly adherent macrophages. o enrich for growing dendritic cells, we utilized a procedure similar to that described for the mouse blood cel cultures of Example 1. Briefly, the aggregates of attache cells were dislodged with Pasteur pipettes and applied to 6 ml columns of 50% FCS-RPMI 1640. Residual granulocytes i the cultures, often in aggregates as well, were easil dissociated at this step. Upon lg sedimentation of th dislodged cells, clusters moved to the bottom of the tub and single granulocytes were left at the top. Th aggregates were subcultured at 2-3 x 10⁵/ml in fresh mediu with GM-CSF, typically for 1 day in 16 mm wells. Afte overnight culture, large numbers of typical dendritic cell were released. Adherent macrophages also expanded in thes cultures, but most remained firmly adherent to the cultur surface. ° C. Cytological Comparison of Dendritic Cell Precursors and Ia-ne ative. Bone Marrow Nonlymphocytes

To compare the released [dendritic-cell enriched; top] and adherent [macrophage-enriched; bottom] .fractions of 7 day bone marrow cultures, la-negative, bone marrow 5 nonlymphocytes were cultured in GM-CSF. At days 2 and 4, nonadherent cells were gently washed away and at day 6, the loosely attached cell aggregates were isolated by lg sedimentation. After a day in culture, the cells that were released from the aggregates were cytospun onto glass slides 0 and stained with different mAbs plus peroxidase anti-Ig as well as Giemsa and nonspecific esterase. The firmly adherent cells in the original cultures were dislodged with EDTA and also cytospun. Many dendritic profiles are in the released fraction [a hand lens is useful to detect cell 5 shape and contaminating granulocytes, in the Giemsa stain], while the adherent cells are for the most part typical vacuolated macrophages. Strong MHC class II expression occurs on all released cells but for a few typical granulocytes. Only a subset of the firmly adherent cells o express class II. Most, released cells express the 2A1 endocytic vacuole antigen, while the adherent cells are 2A1 weak or negative.

D. Cell surface and intracellular antigens: Cell surface 5 staining utilized cytofluorography [FACScan; Becton

Dickinson, Mountain View CA]. Staining with primary rat or hamster mAbs was followed by FITC-conjugated mouse anti-rat or goat anti-hamster Ig's as described in Example ID. A panel of mAbs to cell surface (23, 24) and to intracellular 0 antigens (33, 34) was tested on cytospin preparations. We studied both adherent and nonadherent populations, the former being dislodged in the presence of 10 mM EDTA [the adherent cells were rinsed twice with PBS and once with

EDTA-PBS, and then incubated with EDTA-PBS for 20 min at 5 37°C] . The cytospins were fixed in acetone and stained with mAbs followed by peroxidase conjugated anti-rat or anti- hamster lg. The peroxidase was visualized with diaminobenzidine, and the nuclei counterstained with Giemsa.

E. Cytologic assays: Giemsa stains were performed on cytospin preparations as was the case for the nonspecific esterase [α-naphthyl acetate as substrate] stain using standard methods (35) except that the cytospin preps were fixed with 2% glutaraldehyde in Hanks medium instead of buffered acetone formalin. Phase contrast observations, usually of living cells, were made with inverted microscopes [Nikon Diaphot] at a final magnification of 100 and 400X. Transmission electron microscopy (36) and ³H-thymidine autoradiography were performed on developing dendritic cells as described in Example IE.

F. Mixed leukocyte reactions: Cells from the bone marrow cultures were exposed to 15 Gy of X-ray irradiation and applied in graded doses to 3 x 10⁵ syngeneic or allogeneic T cells in 96 well flat bottomed culture plates for 4d. The cells were prepared by passing spleen and lymph node suspensions through nylon wool and then depleting residual APCs with anti-la plus Jlld mAbs plus complement. 3H- thymidine uptake was measured at 80-94h after a pulse of 4 uCi/ml [222 GBq/mmol; American Radiolabeled Chemicals, Ine, St.Louis, MO].

G. Aggregates of proliferating dendritic cells from mouse bone marrow supplemented with GM-CSF.

Prior to culture, we treated the marrow suspensions with a cocktail of mAbs to B cells, T cells, and MHC class II antigens plus complement. This pretreatment of bone marrow cells which reduces the number of B cells and granulocytes, is necessary to identify growing dendritic cells in bone marrow because B cells and granulocyte are also GM-CSF responsive and proliferate and mask the presence of dendritic cell precursors.

Accordingly, at d2 and d4 of culture, we gently swirled the plates to remove loosely adherent cells which proved to be granulocytes typical in morphology and expression of the RB6 antigen [see below] . With these steps, we recognized by day 4 cellular aggregates attached to a layer of adherent cells. Some of the profiles in the aggregates had the veil or sheet-like processes of dendritic cells. The aggregates could be dislodged by gentle pipetting and separated by lg sedimentation. Within 3h of replating, many spiny adherent cells emigrated from the clusters and had the appearance of fresh splenic adherent cells (13) . After another day of culture, these adherent cells came off the surface and many typical dendritic cells were seen floating in the culture medium. Optimal yields of dendritic cells were obtained when the aggregates were harvested on day 6 and then cultured overnight. The capacity of bone marrow to generate dendritic cells is striking, >5 x 10⁶ from the 4 major hind limb bones in a week.

Attached to the surface of the culture wells were cells with the cytologic features of macrophages, and these also expanded in numbers during the first week of culture. These cells could be dislodged by pipetting after incubation at 37°C in the presence of 10 mM EDTA.

If the cultures were maintained in M-CSF, large numbers of macrophages grew out and were firmly attached to the plastic surface. However, no dendritic cells or dendritic cell aggregates were apparent. If a mixture of M-CSF and GM-CSF was applied, then colonies of adherent macrophages as well as aggregates of growing granulocytes and dendritic cells were noted.

H. Development of potent MLR stimulator cells in bone marrow cultures

It is known that suspensions of mouse bone marrow are not active as MLR stimulators (38) and do not contain detectable dendritic cells (30) . Given the cytologic observations above, we cultured la-negative, bone marrow nonlymphocytes for 6d and checked MLR stimulating activity at daily intervals. As long as the cultures were supplemented with GM-CSF, strong MLR stimulating activity developed [Fig 5]. The increase was progressive and by day 6, as few as 100 of the marrow cells induced MLRs with stimulation indices of 20 or more.

To correlate the development of MLR stimulating activity with the appearance of dendritic cells in these heterogenous cultures, we first separated the cultures into nonadherent and loosely adherent fractions [Fig 6A] . The nonadherent cells, which were mainly granulocytes in the first 4 days, were obtained by gently swirling the plates and harvesting the cells. The loosely adherent cells, which contained the aggregates of presumptive dendritic cell precursors and dendritic cells at day 4 and later times, were dislodged by pipetting over the surface of firmly adherent stromal cells. At d2 and at d4, the most potent stimulating activity was in the adherent fraction. By d6, the nonadherent fraction was very active. If one tested firmly adherent macrophages, there was no MLR stimulating activity [Fig 6B, open squares].

As mentioned above, in the presence of GM-CSF the cultures developed aggregates of growing cells that release typical dendritic cells between d4-8 of culture. These aggregates could be isolated by gentle pipetting over the monolayer followed by lg sedimentation. When the aggregates were returned to culture, populations enriched in dendritic cells were released, and these released cells proved to have the very strong MLR stimulating activity that is characteristic of dendritic cells [Fig 6B] .

I. Cell surface markers — cytofluorography

By cytofluorography, two populations of cells were readily distinguished in the nonadherent or easily dislodged ° cells. One population had a low forward light scatter, high levels of the RB6 antigen, and low levels of MHC class II. The other population was larger and had the reciprocal phenotype. The aggregated cells were enriched relative to unfractionated cultures in MHC class II positive cells [Fig 5 8, compare left and middle], and the level of MHC class II on individual cells increased when the aggregates were cultured overnight to release highly enriched populations of dendritic cells [Fig 8, compare middle and right] . More MHC class II rich, RB6 antigen negative cells were seen in day 0 6 verses day 4 cultures [Fig 8]. None of the cells reacted with the mAbs to the B220 antigen of B cells or the SER-4 antigen of macrophages [not shown] .

More detailed FACS studies were performed on cells that had been released from the aggregates. The granulocytes 5 were gated out on the basis of lower forward light scattering. The larger, dendritic cells had uniformly high levels of MHC class I and II as well as CD44 and CDllb [Mac- 1; Ml/70]. Intermediate level staining was noted for the heat stable antigen [HSA; Ml/69], CD45 [Ml/9.3], and CD18 o [2E6]. Lower level staining was evident for the low affinity IL-2 receptor [CD25, 7D4], interdigitating cell antigen [NLDC-145] , Fc-γ receptor [2.4G2], dendritic cell antigen [33D1] , macrophage antigen [F4/80], and CDllc pl50/90 02-integrin [N418]. Several antigens were not 5 detectable including phagocyte [SER-4 marginal zone macrophage, RB6 granulocyte] and lymphocyte [RA3-6.1 B lymphocyte; thy-1, CD3,4,8 T lymphocyte] markers. This phenotype is similar in many respects to that seen in splenic and epidermal dendritic cells (24, 27, 28). The one 0 exception is the high level in the marrow-derived cells of CDllb, an integrin that helps mediate emigration of myeloid cells from the vasculature.

5 ° J. Cytospin preparations

Cytospins were prepared to further compare the released dendritic cells with the firmly adherent stromal population. By Giemsa stain, the cells that had released from the aggregates had the typical stellate shape of dendritic 5 cells, while the adherent cells were for the most part vacuolated macrophages. Many of the dendritic cells had a perinuclear spot of nonspecific esterase stain, while the more adherent populations had abundant cytoplasmic esterase. The released cells stained strongly for MHC class II 0 products, except for the contaminants with typical granulocyte nuclei. The strongly adherent cells contained a subpopulation of class II positive cells. Recently antigens have been described that are primarily localized in intracellular vacuoles of dendritic cells and B cells but 5 not mononuclear phagocytes. The antibodies are termed M342 (34) and 2A1. Many of the dendritic cells had strong 2A1 stain, and a smaller number expressed M342. The adherent cells had a few profiles with weak 2A1.

The development of la-positive cells, and cells o expressing granular intracellular antigens, was quantitated on cytospins. [Fig 10], MHC class II antigens were expressed first, followed by the 2A1 and M342 granular antigens. [Fig 10], By day 8, the majority of the cells were dendritic and had high levels of MHC class II products 5 and 2A1 antigen. If granulocytes were not removed from the cultures, the yield of nonadherent cells was much larger but the highest percentage of MHC class II positive cells that we detected was 30%, and it was difficult to identify and isolate the aggregates that were the site of dendritic cell 0 growth.

When the cytospins were stained for other myeloid antigens, the released cells stained weakly and sometimes not at all above background with monoclonals to the Fcγ receptor [2.4G2] and macrophage restricted antigen [F4/80] . 5 Most of the firmly adherent cells in contrast stained strongly for both antigens. This suggests that while low levels of 2.4G2 and F4/80 are found on the surface of the released dendritic cells, synthesis and expression are probably being downregulated much as occurs hen epidermal dendritic cells are placed in culture (27) . On day 4, some 30-50,000 la-positive cells were floating in the cultures, while on both day 6 and on day 8, another 50-100,000 la-positive cells were harvested. The quantitative data indicated that each well produced some 200,000 or more la-positive cells in a week. Since we obtain about 20-30 wells of the starting la-negative marrow cells from two tibia and two femurs, the total yield of la- positive cells is 5 x 10⁶ or more, exceeding the total estimated number of Langerhans cells in the skin of a mouse (27).

K. 3H-thymidine pulse chase experiments

To further document the proliferation and differentiation of dendritic cells in these cultures, clusters of cells were isolated on day 4, exposed to a 12h pulse of ³H-thymidine, and examined by autoradiography immediately or after 1, 2 and 3 days of chase in ³H- thymidine free medium. The majority of cells in the aggregate were labeled initially, and almost all cells released from the aggregates were labeled. During the chase, increasing percentages of the released progeny expressed the 2A1 granule antigen of mature dendritic cells.

L. Electron microscopy

The released cells had many large veils or lamellipodia extending from several directions of the cell body. The cytoplasm had many mitochondria, few electron dense granules and lysosomes, but several electron lucent vesicles some with the cytologic features of multivesicular bodies. The numerous cell processes extending from the dendritic cells were evident in the semi-thin sections of our preparations. A bone marrow-derived dendritic cell at d5 of culture shows many cytoplasmic veils. A close up of the perinuclear region shows profiles of smooth reticulum and vacuoles. There are few lysosomal or phagocytic structures.

Example 3 Mouse Dendritic Cell Progenitors Phagocytose

Particulates Sensitizing Mice to Mycobacterial Antigens In Vivo

MATERIAL AND METHODS

A. Mice: BALB/C x DBA/2 Fl, C57BL/6 x DBA/2 Fl, and BALB/C male and female mice were purchased from the Trudeau Institute [Saranac Lake, NY] and Japan SLC [Hamamatsu] and used at 6-10 weeks of age.

B. Bone marrow cultures: As described in Example 2 above, bone marrow was flushed from the femus and tibias, depleted of red cells with 0.83% ammonium chloride, and cultured in 24 well plates [Nunc, Napaville, IL and Corning #25820, Corning NY] at 10° cells/well in 1 ml of RPMI-1640 supplemented with 5% fetal calf serum, 20 ug/ml gentamicin, and 1000 U/ml of recombinant murine GM-CSF [Kiren Brewery, Maebashi, Gumma, Japan; 9.7 x 10⁷ U/mg] . At d2, 0.75 ml of medium and the nonadherent cells were removed, and replaced with fresh medium. This was repeated at d4-5, thereby removing most of the developing granulocytes and leaving behind clusters of proliferating dendritic cells adherent to a stroma that included scattered macrophages. The culture medium was then supplemented with particulates of BCG mycobacteria [described in greater detail below] , and phagocytosis was allowed to proceed for 20-24h usually on d5-6. At this point the cultures were rinsed free of loose cells and particles, and the cells analyzed immediately for particle uptake. Alternatively cells in the washed cultures were dislodged and 3-4 x 10⁶ cells transferred to a 60 mm Petri dish for a 1 or 2 day "chase" period in particle-free, fresh, GM-CSF supplemented medium. Class Il-rich, mature dendritic cells developed during the chase as described in Example 2, and these were isolated by cell sorting [below]. To compare the phagocytic activity of developing and mature dendritic cells, particles were also administered to 7-8d bone marrow cultures that are rich in single nonproliferating mature dendritic cells.

C. Particulates: BCG mycobacteria [Trudeau Institute, 1.5-2.5 x 10⁸ CFU/ml; Kyowa Pharmaceutical Industries, Tokyo] were administered at approximately 10⁷ live BCG per 16mm diameter well. Uptake was assessed following an "acid fast" stain using an auramine-rhodamine procedure that is more sensitive than Ziehl^" Neelsen and facilitates organism counts. Colloidal carbon [Pellikan Ink, Hannover, Germany] was added at 1:2000 dilution. The carbon was identified as a black granular stain in specimens stained with Diff-Quik^R [Baxter Healthcare Corp, Miami, FL] . Suspensions of 2u latex particles [0.5% v/v; Seradyn, Indianapolis, IN] were applied to the cultures at 50 ul/well, a dose which covers the surface of the culture well with beads.

D. Isolation of mature dendritic cells by cell sorting: As noted before in Example 2, the dendritic cells that are produced in GM-CSF stimulated bone marrow cultures express very high levels of surface MHC class II products [monoclonals B21-2, TIB 227 and M5/114, TIB 120 from the ATCC] as well as moderate levels of a dendritic cell- restricted antigen recognized by monoclonal NLDC-145. Immediately after the pulse with BCG, or after an additional 2 days of "chase" culture, the cells were stained with biotin B21-2 and FITC-streptavidin [Tago, Burlingame, CA]. Class Il-rich cells then were sorted [FACStar Plus, Becton Dickinson, Mountainview, CA] and cytospun onto glass slides [Shandon Inst. Sewicky, PA] . The sorted cells were stained with Diff Quick® which outlines the stellate shape of dendritic cells in cytospins and allows enumeration of profiles containing perinuclear depots of internalized colloidal carbon or latex spheres. To visualize BCG, the cytospins were fixed in absolute acetone for 10 min at room temperature and stained with M5/114 anti-class II, NLDC-145 anti-dendritic cell, or RA3-6B2 anti-B220 or anti-B cell [the latter as a control] followed by POX conjugated mouse anti-rat lg [Boehringer Mannheim, Indianapolis,IN] and diaminobenzidine tetraHCl [Polyscience Ine, Warrington, PA]. The preparations were then double labeled for acid-fast bacilli with auramine rhodamine. Virtually all the cells in the preparation were rich in NLDC-145 and MHC class II products. The number of BCG bacilli in at least 400 cells were enumerated.

E. Electron microscopy: To prove that cell-associated BCG were all internalized, the dendritic cells produced in pulse chase protocols [above] were fixed in 2.5% glutaraldehyde and processed for EM as described in Example 2.

F. Antigen presentation in vitro: Mice were primed with complete Freunds' adjuvant [CFA, SIGMA, St.Louis, MO; 25 ul in the fore and rear paws] or as a control, mycobacteria- free incomplete Freunds' [ICFA]. 7-14d later, the draining lymph nodes were dissociated into a single cell suspension and depleted of APCs with mAbs to MHC class II, B220, and heat stable antigens [M5/114 anti-la, RA3-6B2 anti-B220, and Jlld anti-HSA; TIB 120, 146, and 183 from the ATCC respectively] and rabbit complement. 3xl0⁵ of these APC- depleted, primed T cells were cultured in 96 well flat- bottomed microtest wells [Corning #25860] in RPMI-1640 medium supplemented with 0.5% mouse serum and 50 uM 2- mercaptoethanol. Graded doses of BCG-pulsed, bone marrow or spleen APCs were added. 1 uCi of 3H-thymidine [NEN, Boston, MA; 20 Ci/mmol; 4 uCi/ml] uptake was added to monitor DNA synthesis at 72-88h. Data shown are means of triplicates in which standard errors were <15% of the mean.

G. Antigen presentation in vivo: APCs that had been pulsed with antigen in vitro were administered in vivo to unprimed CxD2 Fl mice. To prime T cells in draining lymph node, 2x10^s dendritic cells were injected into the paws, and lymph node cells were prepared 5d later. To prime T cells in spleen, 10⁶ cells were injected i.v. , and splenocytes were prepared 5 or lOd later. To measure T cell priming, bulk lymph node or spleen cells were cultured as above and challenged with graded doses of protein antigens, either purified protein derivative [PPD, from Statenserum Institute, Copenhagen, Denmark, or from Dr. Ichiro Toida, Research Institute for BCG in Japan, Kiyose, Tokyo] or bovine serum albumin [Sigma] and 3H-thymidine measured at 72-88h. To characterize the proliferating cells, the populations were treated with antibodies and complement prior to measuring 3H-thymidine uptake.

H- Phagocytosis of latex particles within clusters of developing dendritic cells: pulse and pulse chase protocols

When mouse bone marrow or blood is stimulated with GM- CSF, proliferating cell aggregates appear, and these give rise to large numbers of typical immunostimulatory dendritic cells. In bone marrow, which was used for the experiments described below, the proliferating aggregates are best identified by washing away the majority of nonadherent granulocytes that are also induced by GM-CSF in the cultures. At d5-6, the time point when the aggregates were first sizable [5-10 cells wide] , we applied different particles over a 20-22h period.

Following administration of 2u latex spheres, heavy labeling was noted in scattered macrophages on the monolayer. In addition, some clear labeling occurred within the developing dendritic cell aggregates [Fig 12A] . Aggregates that had been exposed to particles were recultured an additional 2 days. During this time, large numbers of cells were released into suspension. These primarily were mature dendritic cells with characteristic stellate shapes and high levels of MHC class II and NLDC-145 antigens. When the released cells were examined by light microscopy, many contained latex spheres and often around a clear perinuclear zone or centrosphere [Fig 12B] . We also studied colloidal carbon uptake in a similar manner. When aggregates were pulsed with colloid and mature dendritic cells allowed to form during a chase period, some of the released cells had a centrosphere with small but clear cut carbon deposits [Fig 12C] . In contrast, when latex or carbon was offered to mature dendritic cells, little uptake occurred [Fig 12D] .

I. BCG mycobacteria uptake by developing dendritic cells — acid fast stains

Live BCG mycobacteria were administered as the phagocytic meal over a 20-22h period using the protocol for administering latex particles described above. Cell- associated bacilli were visualized by a sensitive fluorescent acid-fast stain. Following the 2Oh pulse, the developing dendritic cell aggregates contained many organisms. To isolate the more mature dendritic cells from the cultures, the cells were resuspended and sorted those cells with high levels of MHC class II products. Immediately after the BCG pulse, about 20% of the sorted cells contained acid fast bacilli [Table 1] . The majority of MHC class II-weak cells were not studied further because of excessive stickiness during cell sorting.

Companion cell cultures were then studied after 2 days of a chase culture. Because many mature dendritic cells formed during the chase period, the number of la-rich progeny had increased four fold [Table 1]. TABLE 1

Frequency of dendritic cells with phagocytosed BCG organisms in GM-CSF stimulated mouse bone marrow cultures

Quantitative data of dendritic cells containing BCG. Mouse bone marrow cultures were stimulated in 16mm wells for 5d with GM-CSF, washed, and exposed to BCG organisms for 20h. The cultures were washed again and either examined immediately, or pooled and transferred to a 60 mm dish for an additional 2d chase culture. The dendritic cells in the cultures were selected as la-rich cells using a fluorescent activated cell sorter and then cytospun onto glass slides for staining for acid fast bacilli. During the chase period, the percentage of la-rich cells in the cultures increased 2-2.5 fold, and the total number of cells increased 2 fold, resulting in a 4-5 fold increase in the number of la-rich cells. The percentage of dendritic cells containing BCG also rose to 50% [Table 1, Fig. 13]. Double labeling experiments verified that cells with acid fast bacilli expressed MHC class II and the dendritic cell-restricted NLDC 145 antigen Fig. 13. Because the total number of MHC class II and NLDC- 145 positive cells had increased 4-fold in just 2d, it is likely that these BCG-laden dendritic cells were derived from less mature but phagocytic progenitors in the aggregates.

J. Electron microscopy of BCG pulsed APCs

The perinuclear location of the cell-associated particles by light microscopy indicated that organisms had been internalized. The matter was verified by electron microscopy. About 50% of the dendritic cell profiles contained internalized BCG, although the number of organisms per profile was small, usually one but only up to four. Figs. 14 A, B. Each organism seemed to occupy its own vacuole. It appeared that a phagosomal membrane closely approximated most bacilli, Fig. 14 C, D.

K. Presentation in vitro of mycobacterial antigens to primed T cells

To test the presenting function of dendritic cells that had been pulsed or pulse-chased with BCG organisms, we first prepared antigen-responsive T cells from the draining lymph nodes of mice that had been injected with CFA [complete Freund's adjuvant, which contains heat-killed mycobacteria] or with incomplete Freund's adjuvant [IFA as control; see Methods] . When dendritic cells were added to IFA-primed T cells, a syngeneic mixed leukocyte reaction was observed. This was comparable whether or not the APCs had been exposed to BCG. [Fig 15, right]. However, when dendritic cells had been pulsed with BCG and added to CFA-primed T cells, strong proliferative responses were induced [Fig 15, left]. If dendritic cells were tested immediately after the one day pulse, or after an additional 2 day chase period, the chased population was much more potent. [Fig 15, left; compare ♦ and T] . As few as 100 BCG pulse-chased, dendritic cells elicited sizable T cell responses in vitro [Fig 15, left ♦]. The BCG pulse-chased populations also were 5-10 times more potent in inducing responsiveness to mycobacterial antigen than mature dendritic cells freshly exposed to either PPD or BCG. [Fig 15 left, compare ♦ with #, A] . Therefore, it appeared that the extent of phagocytosis correlated with the efficacy of presentation, as the pulse chased populations were the most active APCs and contained the most intracellular BCG. [Table 1] .

L. Presentation in vivo of mycobacterial antigens to unsensitized mice Comparable populations of BCG-pulsed, and BCG-pulsed and chased, APCs were tested for the capacity to present mycobacterial antigens to unprimed mice. Following injection into the footpads, strong responsiveness to PPD was observed. [Fig 16] . Again the dendritic cells were the most potent if tested after a 2d chase [Fig 16; compare O and Δ] , and this chase period greatly increased the total yield of dendritic cells.

To test if the increased antigen presenting function of BCG pulse-chased dendritic cells was related to the increased number of APCs carrying BCG, the primed populations were also tested for responsiveness to bovine serum albumin [BSA] , since the dendritic cells had been grown in the presence of fetal calf serum. All the dendritic cell populations, regardless of the details of the exposure to BCG, primed mice similarly to BSA. [Fig 16, filled symbols] . This indicates that each population was comparably efficient in immunizing to a soluble protein, whereas the dendritic cells that had phagocytosed BCG were more effective in eliciting responses to mycobacterial antigens. The surface markers of the primed cells were tested by antibody and complement mediated lysis of the populations prior to measuring 3H-thymidine uptake [data not shown] . The proliferating cells were positive for thy-1, but negative for MHC class II, heat stable antigen, and B220. Anti-CD4 hybridoma culture supernatant blocked proliferation more than 85% i.e., the primed cells were helper-type T cells.

Priming was also observed when spleen T cells were tested after an intravenous infusion of BCG-pulsed and BCG- pulse chased dendritic cells [Fig 17]. The cells were more responsive at 5 versus 10 days after injection [compare Figs 17 A and C] . Again dendritic cells that had been cultured ["chased"] for 2 days after exposure to BCG were the most potent [Fig 12 compare ♦ with A; but all populations primed the spleen cells similarly to BSA [Fig 17B] . We conclude that dendritic cell progenitors capture and retain mycobacterial antigens in a manner that is highly immunogenic in vivo.

Example 4

Antigen activated dendritic cells as immunogens. Dendritic cells prepared according to the method described in Example 1 are plated at a concentration of approximately 1 x 10^s cells per well of a 24 well plastic culture plate. The cells are incubated in RPMI 1640 containing 5% fetal calf serum and GM-CSF (30 u/ml) . Antigen is added to the dendritic cell cultures and the cultures are incubated with antigen for approximately 4 hours or for sufficient time to allow the dendritic cells to handle the antigen in an immunologically relevant form, or in a form that can be recognized by T cells. Such handling of the antigen by the dendritic cells involves the dendritic cells 1) acquiring, 2) processing, and 3) presenting the antigen to the T cells in a form which is recognized by the T cells. Following binding of the antigen to the dendritic cells the cells are collected from the culture and used to immunize syngeneic mice. The activated dendritic cells are injected subcutaneously into the mice in an amount sufficient to induce an immune response to the antigen.

Example 5

Dendritic cells prepared as described in Example 1 are pulsed with a protein antigen for a time sufficient to allow the dendritic cells to acquire, process and present the modified antigen on the surface of the dendritic cells. The dendritic cells are then collected from the culture for extraction of the modified antigen.

For extraction of the modified antigen, the dendritic cells are solubilized with detergent to extract the modified antigen bound to MHC molecules. The MHC molecules bound to modified antigen are purified by precipitation with antibodies which bind the MHC molecules such as MH2. The modified antigens are extracted from the precipitate for analysis.

Example 6 Preparation of Dendritic Cells from Human

Blood

A. Patients

Seventeen experiments were performed with blood from human patients undergoing consolidation chemotherapy (15 with leukemias/lymphomas in full remission, 2 with solid tumors) followed by treatment with G-CSF. Three experiments were performed with blood from patients after chemotherapy (1 (acute myeloic) leukemia, 2 solid tumors) and GM-CSF treatment. The results of three experiments, two from the G-CSF treated group of patients, A and B, and one from the GM-CSF treated group of patients, C, are presented. B. Rationale

Results of procedures described in Example 1 relating to mouse blood and Example 2 relating to bone marrow (J. EXP. Med. 175:1157-1167, 1992 and J. EXP. Med. 176:1693- 1702, 1992), identified several features of dendritic cell growth and development: (a) dendritic cell progenitors do not express the MHC class II antigens that are typical of mature immunostimulatory progeny and of many other cell types (B cells, monocytes) ; (b) dendritic cell progenitors require GM-CSF and perhaps other cytokines that can be provided by the cells in culture or as supplements to proliferate and mature; (c) critical steps in dendritic cell growth and development take place in distinctive aggregates that are loosely adherent to standard tissue culture surfaces; (d) by monitoring the appearance of these aggregates, one can evaluate the numerous variables that are pertinent to the generation of dendritic cells, a trace but specialized type of antigen presenting cell that operates in a potent fashion to induce T cell immunity and tolerance in situ (Ann.Rev.Immunol. 9:271-296, 1991).

Protocol

1. Blood mononuclear cells were isolated by sedimentation in standard dense media, here Lymphoprep (Nycomed, Oslo) .

2. The isolated mononuclear cells were depleted of cells that were not dendritic cell progenitors. These contaminants were coated with monoclonal antibodies to CD3 and HLA-DR antigens and depleted on petri dishes coated with affinity-purified, goat anti-mouse IgG ("panning") .

3. 10⁶ cells in 1 ml of culture medium were plated in 16 mm diameter plastic culture wells (Costar, Rochester, N.Y.). The medium was RPMI-1640 supplemented with 50 uM 2- mercaptoethanol, 10 mM gluta ine, 50 ug/ml gentamicin, 5% serum from cord blood (without heat inactivation) or 5% ^c fetal calf serum (with inactivation) , and 400 U/ml human recombinant GM-CSF. Every 2nd day thereafter and for a total of 16 days, the cultures were fed by removing 0.3 ml of the medium and replacing this with 0.5 ml of fresh medium supplemented with the cytokines. Cells were cultured under the following conditions: 1) without presence of additional cytokines; 2) GM-CSF, 400 or 800 U/ml; 3) GM-CSF, 400 or 800 U/ml, plus IL-lα, 50 LAF units/ml for the last 24 h of culture; 4) GM-CSF, 400 or 800 U/ml, plus TNFα, 50 U/ml; 5) GM-CSF, 400 or 800 U/ml, plus TNF-α, 50 U/ml, plus IL-lα, 50 LAF units/ml for the last 24 h of culture; 6) GM-CSF, 400 or 800 U/ml, plus IL-3, 100 U/ml; 7) GM-CSF, 400 or 800 U/ml, plus IL-3, 100 U/ml, plus IL-lα, 50 LAF units/ml for the last 24 h.

In experiment C, non-dendritic cells which sank in dense metrizamide were also tested.

4. Characteristic proliferating dendritic cell aggregates (hereafter termed "balls") appeared by the 5th day, as evident upon examination with an inverted phase contrast microscope. These balls expanded in size over the course of a week (day 5-11) . Some balls appeared in the original wells (steps 3 and 4) , but typically these did not enlarge to the same extent as the nonadherent wells (step 4). The wells must be subcultured, e.g., 1 well split into 2-3 wells, as cell density increases. 5. Two alternative approaches were used to isolate the mature dendritic cells from the growing cultures. One method consisted of removing cells that were nonadherent and separate the balls from nonballs by lg sedimentation. Dendritic cells were then released in large numbers from the balls over an additional 1-2 days of culture, and mature dendritic cells were isolated from the nonballs by floatation on dense metrizamide as described (Freudenthal and Steinman, Proc. Natl. Acad. Sci. USA 87:7698-7702, 1990) . The second method is simpler but essentially 5 terminates the growth phase of the procedure. According to ^c the second procedure, the nonadherent cells were harvested when the balls were very large. The cells were left on ice for 20 minutes, resuspended vigorously with a pipette to disaggregate the balls, and the mature dendritic cells were floated on metrizamide columns. 6. To demonstrate the immunostimulatory activity of the dendritic cell progeny, graded doses of irradiated cells (30 to 30,000 in serial 3 fold dilutions) were added to accessory cell-depleted T cells (200,000 for the mixed leukocyte reaction assay, MLR; 150,000 for the oxidative mitogenesis assay, OXMI) . The T cell response was measured with a 16h pulse of 3H-thymidine on the 5th (MLR) or 2nd day

(OXMI) . T cell-stimulation experiments (oxidative mitogenesis and mixed leukocyte reaction) were performed in the presence of 1 microgram/ml indomethacin. Data from three MLR experiments are presented in Figs. 18A, B, and C.

D. Results

1. GM-CSF is an essential cytokine. G-CSF, M-CSF, IL-3, or no cytokine do not permit the development of dendritic cell balls. GM-CSF at 400-800 U/ml is optimal, irregardless of whether donors had been treated with either GM-CSF or G-CSF to expand the number of myeloid progenitor cells in blood. Addition of TNFα at 10-50 U/ml usually but not always increased dendritic cell yields up to two-fold (cf. Caux et al.. Tumor necrosis factor alpha strongly potentiates interleukin-3 and granulocyte-macrophage colony- stimulating factor-induced proliferation of human hematopoietic CD34+ progenitor cells. Blood 2292-2298, 1990) . As evident from the representative experiments described in Figs. 18A, B and C, TNFα supplementation also substantially improves the function of the dendritic cell progeny. rhu IL-lα (50 LAF units/ml) in some experiments proves a further increase in function, when added during the 5 last 24h of the culture. Experiments with tissue from ° patients with solid tumors or leukemias/lymphomas gave comparable results with regard to the generation of dendritic cells.

2. Starting from 60 ml of blood, and after culturing in the presence of GM-CSF only, the yield of typical mature immunostimulatory dendritic cells was 6-12 x 10⁶ cells, representing 40-80% of the total cells. This yield is at least 20 times greater than the yield of mature dendritic cells in 60 ml of fresh blood which would be at most 5% (3-6 x 10^s) of this (Proc. Natl. Acad.Sci. 87:7698-7702, 1990). 3. The phenotype of the dendritic cells generated by this method included the fact that the cells were strongly positive for HLA-DR, MHC class II products but negative for CDla, CD14, and B cell markers.

4. The development of granulocytes in the cultures reduces the purity of the dendritic cells. Typically, these granulocyte balls are more adherent and are left behind at the day 2 transfer step of the protocol. If these adherent granulocyte colonies reappear, simply transfer the growing dendritic cells may be transferred to another well.

Example 7

Other sources of dendritic cell progenitors have been tested according to the method described in Example 6: a) For two patients, a small sample of bone marrow was also provided. When the above procedure was applied, the dendritic cell balls and mature immunostimulatory dendritic cells were formed in large numbers. b) Blood from 7 normal donors has been evaluated using the method described in Example 6. The number of balls proved to be much less (10-20/well of 2xl0⁶ cells) , but the use of normal blood is obviously simpler and has the advantage that granulocyte colonies do not form as noted before (comment 5) in comparing mouse blood and marrow, 5 J.Exp. Med. 175:1157-1167, 1992 vs. J.Exp. Med^'. 176:1693- 1702, 1992) . c) Fetal or umbilical cord blood was also tested, because it too contains more progenitor cells than adult blood. Since the number of CD34+ progenitors is still very small (about 1%) , we tested the simpler method above in which CD34+ cells are not purified initially. DC balls ar readily induced, except that red blood cells which are toxic were depleted. By adding the anti-erythroid monoclonal VIE- G4 (provided by Dr. W. Knapp, Vienna) to the panning step (step 2) , and using an additional floatation on Lymphoprep (step 1) after panning. The yields of dendritic cells from cord blood are roughly comparable to that described in the method (1-5 x 10⁶ dendritic cells, representing 20-40% of the total cells from 40 ml cord blood without a metrizamide floatation step) . The balls are more adherent, and the dendritic cells express CDla, in contrast to adult blood.

^u REFERENCES

1. Steinman, R.M. 1991. The dendritic cell system and its role in immunogenicity. Ann . Rev. Immunol . 9:271.

2. Witmer-Pack, M.D., W. Olivier, J. Valinsky, G. Schuler, and R.M. Steinman. 1987. Granulocyte/macrophage colony-stimulating factor is essential for the viability and function of cultured murine epidermal Langerhans cells. J. Exp. Med. 166:1484.

3. Heufler, C. , F. Koch, and G. Schuler. 1987. Granulocyte-macrophage colony-stimulating factor and interleukin-1 mediate the maturation of murine epidermal Langerhans cells into potent immunostimulatory dendritic cells. J. Exp. Med. 167:700.

4. Romani, N. , S. Koide, M. Crowley, M. Witmer-Pack, A.M. Livingstone, C.G. Fathman, K. Inaba, and R.M. Steinman. 1989. Presentation of exogenous protein antigens by dendritic cells to T cell clones: intact protein is presented best by immature, epidermal Langerhans cells. J. Exp. Med . 169:1169.

5. Inaba, K. , N. Romani, and R.M. Steinman. 1989. An antigen-independent contact mechanism as an early step in

T-cell-proliferative responses to dendritic cells. J. Exp. Med. 170:527.

6. Pure', E. , K. Inaba, M.T. Crowley, L. Tardelli, M.D. Witmer-Pack, G. Ruberti, G. Fathman, and R.M. Steinman. 1990. Antigen processing by epidermal Langerhans cells correlates with the level of biosynthesis of major histocompatibility complex class II molecules and expression of invariant chain. J. Exp. Med. 172:1459.

7. Kampgen, E., N. Koch, F. Koch, P. Stoger, C. Heufler, G. Schuler, and N. Romani. 1991. Class II major histocompatibility complex molecules of murine dendritic cells: synthesis, sialylation of invariant chain, and antigen processing capacity are downregulated upon culture. Proc. Natl . Acad. Sci . USA 88:3014. 5 ^υ 8. Austyn, J.M. , J.W. Kupiec-Weglinski, D.F. Hankins, and P.J. Morris. 1988. Migration patterns of dendritic cells in the mouse. Homing to T cell-dependent areas of spleen, and binding within marginal zone. J^". Exp. Med. 167:646. 9. Larsen, C.P., P.J. Morris, and J.M. Austyn. 1990. Migration of dendritic leukocytes from cardiac allografts into host spleens: a novel pathway for initiation of rejection. J . Exp. Med . 171:307.

10. Austyn, J.M. , and C.P. Larsen. 1990. Migration patterns of dendritic leukocytes. Transpl . 49:1.

11. Veldman, J.E., and E. Kaiserling. 1980. Interdigitating cells. In The Reticuloendothelial System,Morphology. I. Carr, and W.T. Daems, editors. Plenum Publishing Corp., New York. 381-416. 12. Witmer, M.D., and R.M. Steinman. 1984. The anatomy of peripheral lymphoid organs with emphasis on accessory cells: light microscopic, immunocytochemical studies of mouse spleen, lymph node and Peyer's patch. Am. J. Anat . 170:465. ¹³> Kraal, G. , M. Breel, M. Janse, and G. Bruin. 1986. Langerhans cells, veiled cells, and interdigitating cells in the mouse recognized by a monoclonal antibody. J. Exp. Med. 163:981.

14. Inaba, K. , J.P. Metlay, M.T. Crowley, and R.M. Steinman. 1990. Dendritic cells pulsed with protein antigens in vitro can prime antigen-specific, MHC-restricted T cells in situ. J. Exp. Med . 172:631.

15. Steinman, R.M. , D.S. Lustig, and Z.A. Cohn. 1974. Identification of a novel cell type in peripheral lymphoid organs of mice. III. Functional properties in vivo. J. Exp. Med. 139:1431.

16. Pugh, C.W., G.G. MacPherson, and H.W. Steer. 1983. Characterization of nonlymphoid cells derived from rat peripheral lymph. J. Exp. Med . 157:1758. ' 17. Fossum, S. 1989. The life history of dendritic leukocytes [DL] . In Current Topics in Pathology. O.H. Ivessen, editor. Springer-Verlag, Berlin. 101-124.

18. Katz, S.I., K. Tamaki, and D.H. Sachs. 1979. Epidermal Langerhans cells are derived from cells originating in bone marrow. Nature 282:324.

19. Hart, D.N.J. , and J.W. Fabre. 1981. Demonstration and characterization of la-positive dendritic cells in the interstitial connective tissues of rat heart and other tissues, but not brain. J. Exp. Med. 154:347. 20. Bowers, W.E., and M.R. Berkowitz. 1986. Differentiation of dendritic cells in cultures of rat bone marrow cells. J. Exp. Med. 163:872.

21. Reid, C.D.L., P.R. Fryer, C. Clifford, A. Kirk, J. Tikerpae, and S.C. Knight. 1990. Identification of hematopoietic progenitors of macrophages and dendritic Langerhans cells - [DL-CFU] in human bone marrow and peripheral blood. Blood 76:1139.

22. Kajigaya, S., T. Suda, J. Suda, M. Saito, Y. Miura, M. Iizuka, S. Kobayashi, N. Minato, and T. Sudo. 1986. A recombinant murine granulocyte/macrophage (GM) colony-stimulating factor derived from an inducer T cell line (IH5.5) . Functional restriction to GM progenitor cells. J. Exp. Med. 164:1102.

23. Agger, R. , M.T. Crowley, and M.D. Witmer-Pack. 1990. The surface of dendritic cells in the mouse as studied with monoclonal antibodies. Int. Rev. Immunol . 6:89.

24. Crowley, M. , K. Inaba, M. Witmer-Pack, and R.M. Steinman. 1989. The cell surface of mouse dendritic cells: FACS analyses of dendritic cells from different tissues including thymus. Cell . Immunol . 118:108.

25. Freudenthal, P.S., and R.M. Steinman. 1990. The distinct surface of human blood dendritic cells, as observed after an improved isolation method. Proc. Natl . Acad. Sci . USA 87:7698. ° 26. Drexhage, H.A. , H. Mullink, J. de Groot, J. Clarke, and B.M. Balfour. 1979. A study of cells present in peripheral lymph of pigs with special reference to a type of cell resembling the Langerhans cells. Cell . Tisε . Res . 202:407. 27. Schuler, G., and R.M. Steinman. 1985. Murine epidermal Langerhans cells mature into potent immunostimulatory dendritic cells in vitro. J. Exp. Med . 161:526.

28. Romani, N. , M. Witmer-Pack, M. Crowley, S. Koide, G. Schuler, K. Inaba, and R.M. Steinman. 1991. Langerhans cells as immature dendritic cells. CRC Press, Boston. 191-216.

29. Romani, N. , G. Schuler, and P. Fritsch. 1986. Ontogeny of la-positive and Thy-1 positive leukocytes of murine epidermis. J. Invest . Dermatol . 86:129.

30. Nussenzweig, M.C., R.M. Steinman, M.D. Witmer, and B. Gutchinov. 1982. A monoclonal antibody specific for mouse dendritic cells. Proc . Natl . Acad . Sci . USA 79:161.

31. Inaba, K. , J.P. Metlay, M.T. Crowley, M. Witmer-Pack, and R.M. Steinman. 1990. Dendritic cells as antigen presenting cells in vivo. Int . Rev. Immunol . 6:197.

32. Barclay, A.N. , and G. Mayrhofer. 1981. Bone marrow origin of la-positive cells in the medulla of rat thymus. J. Exp. Med . 153:1666. 33. Rabinowitz, S.S., and S. Gordon. 1991. Macrosialin, a macrophage-restricted membrane sialoprotein differentially glycosylated in response to inflammatory stimuli. J. Exp. Med . 174:827.

34. Agger, R. , M. Witmer-Pack, N. Romani, H. Stossel, W.J. Swiggard, J.P. Metlay, E. Storozynsky, P. Freimuth, and R.M. Steinman. 1992. Two populations of splenic dendritic cells detected with M342, a new monoclonal to an intracellular antigen of interdigitating dendritic cells and some B lymphocytes. J. Leuk . Biol . 52:34. ^ύ 35. Kaplow, L.S. 1981. Cytochemical identification of mononuclear phagocytes, in "Manual of macrophage methodology" eds H.B. Herscowitz, H.T. Holden, J.A. Bellanti, and A. Ghaffer. Marcel Dekker,Inc, New York. 199-207. 36. Stossel, H., F. Koch, E. Kampgen, P. Stoger, A. Lenz, C. Heufler, N. Romani, and G. Schuler. 1990. Disappearance of certain acidic organelles [endosomes and Langerhans cell granules] accompanies loss of antigen processing capacity upon culture of epidermal Langerhans cells. J. Exp. Med. 172:1471.

37. Steinman, R.M. , and Z.A. Cohn. 1973. Identification of a novel cell type in peripheral lymphoid organs of mice. I. Morphology, quantitation, tissue distribution. J^". Exp. Med. 137:1142. 38. Steinman, R.M. , and M.D. Witmer. 1978. Lymphoid dendritic cells are potent stimulators of the primary mixed leukocyte reaction in mice. Proc. Natl . Acad. Sci . USA

75:5132. 39. Scheicher, C. , M. Mehlig, R. Zecher, and K. Reske. 1992. Dendritic cells from mouse bone marrow: in vitro differentiation by low doses of recombinant granulocyte-macrophage-CSF. J. Immunol . Methods In Press:

40. Crowley, M.T., K. Inaba, M.D. Witmer-Pack, S. Gezelter, and R.M. Steinman. 1990. Use of the fluorescence activated cell sorter to enrich dendritic cells from mouse spleen. J. Immunol . Methods 133:55.

41. Naito, K. , K. Inaba, Y. Hirayama, M. Inaba-Miyama, T. Sudo, and S. Muramatsu. 1989. Macrophage factors which enhance the mixed leukocyte reaction initiated by dendritic cells. J. Immunol . 142:1834.

42. Steinman, R.M. , G. Kaplan, M.D. Witmer, and Z.A. Cohn. 1979. Identification of a novel cell-type in peripheral lymphoid organs of mice. V. Purification of spleen dendritic cells, new surface markers, and maintenance 5 in vitro. J. Exp. Med. 149:1. ° 43. Goud, T.J.L.M., C. Schotte, and R. van Furth. 1975. Identification and characterization of the monoblast in mononuclear phagocyte colonies grown in vitro. J^". Exp. Med. 142:1180.

44. Inaba, K. , Steinman, R.M. , Witmer-Pack, M. , K. Aya, M. Inaba, T. Sudo, S. Wolpe, and G. Schuler. 1992.

Identification of proliferating dendritic cell precursors in mouse blood. J. Exp. Med . 175:1157.

45. Britz, J.S., P.W. Askenase, W. Ptak, R.M. Steinman, and R.K. Gershon. 1982. Specialized antigen-presenting cells: Splenic dendritic cells, and peritoneal exudate cells induced by mycobacteria, activate effector T cells that are resistant to suppression. J. Exp. Med . 155:1344.

46. Lechler, R.I. and J.R. Batchelor. 1982. Restoration of immunogenicity to passenger cell-depleted kidney allografts by the addition of donor strain dendritic cells. J. Exp. Med. 155:31.

47. Boog, C.J.P. , W.M. Kast, H.Th.M. Timmers, J. Boes, L.P. De Waal, and C.J.M. Melief. 1985. Abolition of specific immune response defect by immunization with dendritic cells. Nature 318:59.

48. Faustman, D.L., R.M. Steinman, H.M. Gebel, V. Hauptfeld, J.M. Davie, and P.E. Lacy. 1984. Prevention of rejection of murine islet allografts by pretreatment with anti-dendritic cell antibody. Proc . Natl . Acad. Sci . USA 81:3864.

49. Iwai, H. , S.-I. Kuma, M.M. Inaba, R.A. Good, T. Yamashita, T. Kumazawa, and S. Ikehara. 1989. Acceptance of murine thyroid allografts by pretreatment of anti-la antibody or anti-dendritic cell antibody in vitro. Tranεpl . 47:45.

50. Havenith, C.E.G., A.J. Breedijk, M.G.H. Betjes, W. Calame, R.H.J. Beelen, and E.C.M. Hoefsmit. 1992. T cell priming in situ by intratracheally instilled antigen-pulsed dendritic cells. Am. J. Resp. Cell & Molec . Biol . Submitted. ° 51. Liu, L.M. and G.G. MacPherson. 1993. Antigen acquisition by dendritic cells: Intestinal dendritic cells acquire antigen administered orally and can prime naive T cells "in vivo". J. Exp. Med. In Press:

52. Holt, P.G., M.A. Schon-Hegrad, and J. Oliver. 1987. MHC class II antigen-bearing dendritic cells in pulmonary tissues of the rat. Regulation of antigen presentation activity by endogenous macrophage populations. J^". Exp. Med. 167:262.

53. Bujdoso, R. , J. Hopkins, B.M. Dutia, P. Young, and I. McConnell. 1989. Characterization of sheep afferent lymph dendritic cells and their role in antigen carriage. J^". Exp. Med. 170:1285.

54. Crowley, M. , K.^* Inaba, and R.M. Steinman. 1990. Dendritic cells are the principal cells in mouse spleen bearing immunogenic fragments of foreign proteins. J. Exp. Med. 172:383.

55. Shimonkevitz, R. , J. Kappler, P. Marrack, and H. Grey. 1983. Antigen recognition by H-2-restricted T cells. I. Cell-free antigen processing. J^". Exp. Med. 158:303. ⁵⁶« Ziegler, H.K. and E.R. Unanue. 1982. Decrease in macrophage antigen catabolism caused by ammonia and chloroquine is associated with inhibition of antigen presentation T cells. Proc. Natl . Acad. Sci . USA 79:175.

57. Pancholi, P., A. Mirza, V. Schauf, R.M. Steinman, and N. Bhardwaj. 1993. Presentation of mycobacterial antigens by human dendritic cells: lack of transfer from infected macrophages. Infection and Immunity submitted:

58. Inaba, K. , M. Inaba, N. Romani, H. Aya, M. Deguchi, S. Ikehara, S. Muramatsu, and R.M. Steinman. 1992. 0 Generation of large numbers of dendritic cells from mouse bone marrow cultures supplemented with granulocyte-macrophage colony stimulating factor. J. Exp. Med. 176:1693.

5 ° 59. 1993. Manual of methods for general bacteriology. In P. Gerhardt, R.G.E. Murray, R.E. Cogtillo, E.W. Nester, W.A. Wood, N.R. Krieg, and G.B. Philips, editors.

60. Cohn, Z.A. 1963. The fate of bacteria within phagocytic cells. I. The degradation of isotopically labeled bacteria by polymorphonuclear leucocytes and macrophages. J. Exp. Med. 117:27.

61. Steinman, R.M. and Z.A. Cohn. 1972. The interaction of particulate horseradish peroxidase (HRP)-anti HRP immune complexes in mouse peritoneal macrophages in vitro. J. Cell Biol . 55:616.

62. Ehrenreich, B.A. and Z.A. Cohn. 1967. The uptake and digestion of iodinated human serum albumin by macrophages in vitro. J. Exp. Med. 126:941.

63. Steinman, R.M. and Z.A. Cohn. 1972. The interaction of soluble horseradish peroxidase in mouse peritoneal macrophages in vitro. J. Cell Biol . 55:186.

64. Cohn, Z.A. and B.A. Ehrenreich. 1969. The uptake, storage and intracellular hydrolysis of carbohydrates by macrophages. J. Exp. Med . 129:201. ⁶⁵* Ehrenreich, B.A. and Z.A. Cohn. 1969. The fate of peptides pinocytosed by macrophages in vitro. J. Exp. Med . 129:227.

66. Hunt, D.F., H. Michel, T.A. Dickinson, J. Shabanowitz, A.L. Cox, K. Sakaguchi, E. Appella, H.M. Grey, and A. Sette. 1992. Peptides presented to the immune system by the murine class II major histocompatibility complex molecule I-Ad. Science 256:1817.

67. Rudensky, A.Y., P. Preston-Hurlburt, S.-C. Hong, A. Barlow, and CA. Janeway Jr.. 1991. Sequence analysis of peptides bound to MHC class II molecules. Nature 353:622.

68. Jensen, P.E. 1988. Protein synthesis in antigen processing. J. Immunol . 141:2545.

69. Harding, C.V. and E.R. Unanue. 1990. Quantitation of antigen-presenting cell MHC class II/peptide complexes 5 necessary for T-cell stimulation. Nature 346:574. ° 70. Demotz, S., H.M. Grey, and A. Sette. 1990. The minimal number of class II MHC-antigen complexes needed for T cell activation. Science 249:1028.

71. Bhardwaj, N. , J.W. Young, A.J. Nisanian, J. Baggers, and R.M. Steinman. 1992. Small amounts of superantigen on dendritic cells are sufficient to initiate T cell responses. Submitted

72. Breel, M. , R.E. Mebius, and G. Kraal. 1987. Dendritic cells of the mouse recognized by two monoclonal antibodies. Eur. J. Immunol . 17:1555. 73. Fossum, S. and B. Rolstad. 1986. The roles of interdigitating cells and natural killer cells in the rapid rejection of allogeneic lymphocytes. Eur. J. Immunol . 16:440.

74. Reis e Sousa, C. , P.D. Stahl, and J.M. Austyn. 1993. Phagocytosis of antigens by langerhans cells in vitro.

J. Exp. Med. In Press.

75. Harding, C.V. , R.W. Roof, and E.R. Unanue. 1990. Turnover of la-peptide complexes is facilitated in viable antigen-presenting cells:Biosynthetic turnover of Ia vs. peptide exchange. Proc. Natl . Acad. Sci . USA 86:4230.

76. Brodsky, F.M. and L.E. Guagliardi. 1991. The cell biology of antigen processing and presentation. Ann . Rev. Immunol . 9:707.

77. Nonacs, R. , C. Humborg, J.P. Tarn, and R.M. Steinman. 1992. Mechanisms of mouse spleen dendritic cell function in the generation of influenza-specific, cytolytic T lymphocytes. J. Exp. Med. 176:519.

78. Freiden, Thomas R. , T. Sterling, A. Pablos-Mendez, J. Kilburn, G. Cauthen and S.W. Dooley, 1993. The emergence of drug-resistant tuberculosis in New York City. No Eng. J. of Med. 328:521.

While we have hereinbefore described a number of embodiments of this invention, it is apparent that the basic constructions can be altered to provide other embodiments 5 which utilize the methods and compositions of this invention. Therefore, it will be appreciated that the scope of this invention is defined by the claims appended hereto rather than by the specific embodiments which have been presented hereinbefore by way of example.

Claims

WE CLAIM:

1. A method of producing a population of dendritic cell precursors from proliferating cell cultures comprising: a) providing a tissue source comprising dendritic 5 cell precursors; b) treating the tissue source from (a) to increase the proportion of dendritic cell precursors to obtain a population of cells suitable for culture in vitro: c) culturing the tissue source on a substrate in a 0 culture medium comprising GM-CSF to obtain nonadherent cells and cell clusters; d) subculturing the nonadherent cells and cell clusters to produce cell aggregates comprising proliferating dendritic cell precursors; 5 e) serially subculturing the cell aggregates at least one time to enrich the proportion of dendritic cell precursors.

2. The method according to claim 1 wherein the tissue o source is blood or bone marrow and GM-CSF is present in the medium at a concentration of about 1-1000 U/ml.

3. The method according to claim 2, further comprising that when the tissue source is bone marrow 5 treatment step (b) comprises killing cells expressing antigens which are not expressed on dendritic precursor cells by contacting the bone marrow with antibodies specific for antigens not present on dendritic precursor cells in a medium comprising complement. 0

4. The method according to claim 3, wherein the tissue source is bone marrow and the antibodies are directed against at least one antigen selected from the group consisting of Ia antigen, antigens present on T cells, and 5 antigens present on mature dendritic cells. 5. The method according to claim 4, wherein the bone marrow is cultured with rGM-CSF at a concentration of about 500-1000 U/ml.

6. The method according to claim 5 wherein Ia- negative marrow nonlymphocytes are cultured at a concentration of about 5 x 10⁵ cells/cm².

7. The method according to claim 6, wherein the anti- la antigen antibodies and anti-T cell, B cell and monocyte antibodies are selected from the group consisting of GK 1.5 anti-CD4, Ho 2.2 anti-CD8, B21-2 anti-la, and RA3-3A1/6.1 anti-B220/CD45R.

8. The method according to claim 3 wherein the cell aggregates of step (e) are serially subcultured one to five times.

9. The method according to claim 8 wherein the cell aggregates are serially subcultured two to three times.

10. The method according to claim 9 wherein the cell aggregates are serially subcultured two times.

11. The method according to claim 3 wherein the nonadherent cells and cell clusters of step (c) are subcultured after from about 0.3 to 1 day and the cell aggregates are serially subcultured every 3 to 30 days.

12. The method according to claim 11 wherein the cell aggregates are serially subcultured every 10 to 20 days.

13. The method according to claim 12 wherein the cell aggregates are serially subcultured every 20 days. ° 14. The method according to claim 3 wherein the tissue source is blood or bone marrow, the nonadherent cells and cell clusters of step (c) are subcultured after about 0.3 to 1 day, the cell aggregates are serially subcultured one to five times every 3 to 30 days.

15. The method according to claim 14 wherein the nonadherent cells and cell clusters of step (c) are subcultured after about one half day and the cell aggregates are twice serially subcultured after 20 days.

16. The method according to claim 3 wherein the culture medium is selected from the group consisting of RPMI 1640, DMEM, and α-MEM and wherein the culture medium is supplemented with serum.

17. The method according to claim 16 wherein fetal calf serum is present in the culture medium in an amount of about 1 to 15%.

18- The method according to claim 17 wherein the fetal calf serum is present in the culture medium in an amount of about 10%.

19. The method according to claim 14 wherein the tissue source is blood and the concentration of GM-CSF in the medium is about 30-100 U/ml.

20. The method according to claim 14 wherein the tissue source is bone marrow and the concentration of GM-CSF 0 in the medium is about 500-1000 U/ml.

21. A method of producing a population of mature dendritic cells from proliferating cell cultures comprising: a) providing a tissue source comprising dendritic 5 cell precursors; ^u b) treating the tissue source from (a) to increase the proportion of dendritic cell precursors to obtain a population of cells suitable for culture in vitro; c) culturing the tissue source on a substrate in a culture medium comprising GM-CSF to obtain nonadherent cells and cell clusters; d) subculturing the nonadherent cells and cell clusters to produce cell aggregates comprising proliferating dendritic cell precursors; e) serially subculturing the cell aggregates one or more times to enrich the proportion of dendritic cell precursors; and f) continuing to culture the dendritic cell precursors for a period of time sufficient to allow them to mature into mature dendritic cells.

22. The method according to claim 21 wherein the tissue source is blood or bone marrow and GM-CSF is present in the medium at a concentration of about 1-1000 U/ml.

23. The method according to claim 21, further comprising that when the tissue source is bone marrow the pretreatment step comprises killing cells expressing antigens which are not expressed on dendritic precursor cells by contacting the bone marrow with antibodies specific for antigens not present on dendritic precursor cells in a medium comprising complement.

24. The method according to claim 23, wherein the tissue source is bone marrow and the antibodies are directed against at least one antigen selected from the group consisting of Ia antigen, antigens present on T cells, and antigens present on mature dendritic cells.

5 ° 25. The method according to claim 24, wherein the bone marrow is cultured with rGM-CSF at a concentration of about 500-1000 U/ml.

26. The method according to claim 25 wherein Ia- negative marrow nonlymphocytes are cultured at or concentration of about 5 x 10⁵ cells/cm².

27. The method according to claim 25 wherein the anti- la antigen antibodies and anti-T cell , B cell and monocyte antibodies are selected from the group consisting of GK 1.5 anti-CD4, Ho 2.2 anti-CD8, B21-2 anti-la, and RA3-3A1/6.1 anti-B220/CD45R.

28. The method according to claim 21 wherein the cell aggregates of step (e) are serially subcultured one to five times.

29. The method according to claim 28 wherein the cell aggregates are serially subcultured two to three times.

30. The method according to claim 29 wherein the cell aggregates are serially subcultured two times.

31. The method according to claim 21 wherein the nonadherent cells and cell clusters of step (c) are subcultured after from about 0.3 to 1 day and the cell aggregates are serially subcultured every 3 to 30 days.

32. The method according to claim 31 wherein the cell 0 aggregates are serially subcultured every 10 to 20 days.

33. The method according to claim 32 wherein the cell aggregates are serially subcultured every 20 days.

5 34. The method according to claim 31 wherein the cell aggregates are serially subcultured one to five times.

35. The method according to claim 29 wherein the nonadherent cells and cell clusters of step (c) are subcultured after about one half day and the cell aggregates are twice serially subcultured after 20 days.

36. The method according to claim 21 wherein the culture medium is selected from the group consisting of RPMI 1640, DMEM, and α-MEM and wherein the culture medium is supplemented with serum.

37. The method according to claim 36 wherein fetal calf serum is present in the culture medium in an amount of about 1 to 15%.

38. The method according to claim 37 wherein the fetal calf serum is present in the culture medium in an amount of about 10%.

39. The method according to claim 21 wherein the tissue source is blood and wherein GM-CSF is present in the medium at a concentration of about 30-100 U/ml.

40. The method according to claim 21 where the tissue source is bone marrow and wherein the CM-CSF is present in the medium at a concentration of about 500-1000 U/ml.

41. A method for providing an antigen to a host comprising exposing an antigen to a culture of dendritic cells obtained according to the method of any one of claims 21, 39 or 40 to produce antigen-activated dendritic cells followed by inoculating the host with the antigen-activated dendritic cells. ^c 42. The method according to claim 41 wherein the host is human.

43. A composition comprising dendritic cells prepared according to the method of any one of claims 21, 39 or 40.

5

44. A composition comprising antigen-activated dendritic cells wherein dendritic cells prepared according to claim 21 are pulsed with an antigen and wherein the dendritic cells process the antigen to produce a modified 0 antigen which is expressed by the dendritic cells.

45. A composition comprising a dendritic cell modified antigen wherein a substance to be modified is exposed to a culture of dendritic cells prepared according to any one of 5 claims 21, 39 or 40 and whereby the substance is modified by the dendritic cells to produce the modified antigen.

46. A method of immunizing against disease in humans or animals comprising, administering a vaccine comprising o "the composition of claim 44.

47. A vaccine comprising the composition of claim 44.

48. A method of immunizing against disease in humans or animals comprising, administering a vaccine comprising the composition of claim 45.

49. A vaccine comprising the composition of claim 45.

0 50. Amethod of treating autoimmune disease comprising administering to a person in need of treatment a therapeutically effective amount of the composition of claim 44 and wherein the antigen to be modified is a self-protein.

5 ^ϋ 51. A method of treating autoimmune disease comprising administering to a person in need of treatment a therapeutically effective amount of the composition of claim

45 wherein the substance to be modified is a self-protein.

5 52. The method of claim 50 wherein the autoimmune disease is selected from the group consisting of multiple sclerosis myasthenia gravis, atopic dermatitis and juvenile diabetes.

0 53. The method of claim 51 wherein the autoimmune disease is selected from the group consisting of multiple sclerosis and juvenile diabetes.

54. Dendritic cell precursors prepared according to 5 the method of claim 1.

55. The dendritic cell precursors according to claim

46 wherein the tissue source is blood or bone marrow.

o ⁵⁶« The dendritic cell precursor according to claim 55 wherein the tissue source is blood.

57. The dendritic cell precursor according to claim 47 wherein the tissue source is bone marrow. 5

58. The method according to claim 2 wherein the tissue source is human blood and GM-CSF is present in the medium at a concentration of about 400 to 800 U/ml.

0 59. The method according to claim 58 wherein at least one factor selected from the group consisting of TNF-α, G- CSF, IL-1 and IL-3 is present in the culture medium.

5 " 60. The method according to claim 14 wherein the tissue source is human blood and GM-CSF is present in the medium at a concentration of about 400 to 800 U/ml.

61. The method according to claim 60 wherein at least one factor selected from the group consisting of TNF-α, G-

CSF, IL-1 and IL-3 is present together with GM-CSF in the culture medium.

62. The method according to claim 22 wherein the tissue source is human blood and GM-CSF is present in the medium at a concentration of about 400 to 800 U/ml.

63. The method according to claim 62 wherein at least one factor selected from the group consisting of TNF-α, G- 5 CSF, IL-1 and IL-3 is present together with GM-CSF in the culture medium.

64. The dendritic cell precursors according to claim 56 wherein the cells are obtained from human blood and are 0 cultured in the presence GM-CSF at a concentration of about 400 to 800 U/ml.

65. A method of preparing an antigen fragment from an antigen comprising contacting the antigen with cells 5 selected from the group consisting of dendritic cells and dendritic cell precursors, incubating the cells with the antigen for sufficient time to allow the cells to process the antigen into the fragments and present the antigen fragment on the cell surface. 0

66. The method according to claim 65 wherein the dendritic cells or dendritic cell precursors are derived from blood or bone marrow.

5 ^c 67. The method according to claim 66 wherein the dendritic cells or dendritic cell precursors are derived from blood.

68. The method according to claim 67 wherein the dendritic cell or dendritic cell precursors are from human blood and are cultured in the presence of GM-CSF at a concentration of about 400 to 800 U/ml.

69. The method according to claim 65 wherein the antigen is phagocytosed by the dendritic cell precursors.

•70. The method according to claim 69 wherein the antigen is selected from the group consisting of mycobacterial, bacterial and viral antigens.

71. The method according to claim 70 wherein the antigen is a mycobacteria tuberculosis bacteria.

72. The method according to claim 71 wherein the mycobacteria tuberculosis bacteria is BCG.

73. An antigen fragment prepared according to the method of claim 65.

5 74. The antigen fragment according to claim 73 wherein the antigen is phagocytosed.

75. The antigen fragment according to claim 74 wherein the antigen is selected from the group consisting of 0 mycobacterial, bacterial or viral antigens.

76. The antigen fragment according to claim 75 wherein the antigen is a mycobacteria tuberculosis bacteria antigen.

5 ° 77. The antigen fragment according to claim 76 wherein the mycobacteria tuberculosis bacteria is BCG.

78. The antigen fragment according to claim 75 wherein the antigen is a gene product expressed by a viral vector phagocytosed by the dendritic cell precursors.

79. A method of treating tuberculosis comprising administering a therapeutically effective amount of the antigen fragment of claim 77.

80. The method according to clam 1 wherein the tissue source is obtained from an individual who has been pretreated with a substance to stimulate hematopoiesis prior to removal of the tissue source from the individual.

81. The method according to claim 80 wherein the hematopoietic substance is selected form the group consisting of GM-CSF and G-CSF.

5

0

5