CA2170523C - Cloned glutamic acid decarboxylase - Google Patents

Abstract

Isolated polypeptides useful in ameliorating GAD-associated autoimmune disease as well as diagnostic and therapeutic methods of using the peptides are disclosed.

Description

Sep-25-00 05:15pm From-S6B/Wo, +2328440 T-172 P.06/32 F-845 =. r95ro7992 -i~{', ~ .= ~ E } ~ ~ -' ~ i .~ pC7'lUS94/p9479 'i F:~i' Til~'..~ :> 1 ~:1 CLQNLD [3I.II~'~1Hxo_ 714`I!) 1)ECAREO1tYi A~S
r 5 BACKGROHMD OF THE IN'R .Pi'PIOH
The present irtvent4on was suppoXted by orant N322z55 Irom the Nltio1'lgl instj,ttItes of H4alth. The Uttited States Gevernment has certain rights in this invention.
1. FzELD. oF THE SNVVNTICli The present invention relates to glutamic acid decarboxylaseo (GAD65) polypeptides and methods +bf tlsing GADe poXypeptides diagnostically and therapeutically in autoiramune disease.

2. B C R
Insulin-dependent diabetes nlellltus (IDDM; type I
diabetes) is one of the tabst common metaboZ3.c disorders. in the United States, IDDM affectu approxamataly one in 300 to 400 people, and epic7emiological studies suggest that its inoidenoe is increasing. The disease resulta fro-m the autoimmune deatructian of the insulin-pr4duci.ng P-cells of the panoreas. More specifically, the preonset stage is CharaCtsri,ked by "insLtlitis , in which lyaaphocyt.e;e infiltrate the pancreatic islets and saiectively dastrQy the ~3-cells. insulitis may be present for inany years before the onset of clinical symptoms. The typical IDDM presentation of hyperglycemia appears only after at least 80% of the insulin-producing P-celis are lost. The remaining P-COlls are destroyed during the next few years.
Although insulin therapy allowq most IppM patients to lead normal liver., this reglacement is imperfect and does not completelx restore metabolic hoxneostasis.
'>'hus, severe complicatiahs which result in dysfunctioncs of the eye, kidney, heart, and other organs are common in IDDM patients undergoing insulin therapy. Because of this, it is highly desirable to extend the latency period and prevent progression (e.g., through administration of immunosuppressant drugs to interfere with the autoimmune process and insulin to achieve better control of the effects of sustained hypoglycemia) between the start of 0-cell destruction and the actual requirement of insulin replacement (i.e., when 80% of the fl-cells are destroyed).
Therefore, a diagnostic test which determines the beginning of 0-cell destruction would allow the clinician to administer immunosuppressant drugs (Silverstein, et al., New England Journal of Medicine, 319:599-604, 1988) or prophylactic insulin therapy (Keller, et al., Lancet, 341:927, 1993) to extend this latency period and thus significantly delay the onset of insulin replacement side effects.
Many IDDM patients have sera which contain antibodies to a 64kD molecule (Baekkeskov, et al., J.CIin.Invest., 79:926-934, 1987; Atkinson, et al., Lancet, 335:1357-1360, 1990), to islet cell cytoplasmic (ICA) molecules or islet cell surface (ICSA) molecules (Bottazzo, et al, Lancet, 1:668-672, 1980), or to insulin (Palm(ar, et al., Science, 222:1137-1139, 1983;
Atkinson, et al., Diabetes, 35:894-898, 1986).
Atkinson and coworkers (Atkinson, et al., Lancet, 335:1357-1360, 1990) have demonstrated that the presence of antibodies to the 64kD molecule in human sera appears to be the earliest and most reliable indicator that onset of IDDM symptoms will eventually occur.
Recently, Baekkeskov and coworkers established that the 64kD molecule and glutamic acid decarboxylase (GAD) have several antigenic epitopes in common and thus they may be identical or very similar molecules. Although this identification is an important finding, the use of this information as a diagnostic tool for predicting IDDM is quite cumbersome and limited unless knowledge of the molecular biology of GAD is known. Studies by Kaufman, et al., (J. Clin. Invest., 89:283, 1992) established that the 64kD molecule was intact GAD65. Consequently, the cloning and subsequent production of large quantities of GAD65 or a GAD molecule which is antigenically substantially identical to the GAD65 molecule or fragments of the GAD65 molecule, both of which can be easily purified, will allow the development of a diagnostic kit designed to predict IDDM
as well as effective therapeutic modalities. The present invention provides a means for accomplishing these results.

SZTNIlKARY OF THE INVENTION

According to one aspect of the present invention, there is provided a polypeptide fragment of the GAD65 as described herein selected from the group consisting of:
KPCSCSKVDVNYAFLHATDL; TAGTTVYGAFDPLLAVADICKK;

EYLYNIIKNREGYEMVFDGK; IPPSLRYLEDNEERMSRLSK;
SRLSKVAPVIKARMMEYGTT; EYGTTMVSYQPLGDKVNFFR;
ATHQDIDFLIEEIERLGQDL; LAFLQDVMNILLQYVVKSFDRS;
EEILMHCQTTLKYAIKTGHP; DERGKMIPSKLERRILEAKQ;
KHYDLSYDTGDKALQCGRHV; AALGIGTDSVILIKCDERGK;
GLLMSRKHKWKLSGVERANS; LEAKQKGFVPFLVSATAGTT; and VNFFRMVISNPAATHQDIDF.

According to another aspect of the present invention, there is provided a method for detecting autoantibody to GAD65 in a patient specimen which comprises contacting the specimen with a GAD65 fragment and determining whether autoantibody binds to the fragment, wherein the fragment is the polypeptide defined above.

According to still another aspect of the present - 3a -invention, there is provided use of a fragment of GAD65 to ameliorate a GAD65 associated autoimmune disorder in a patient, wherein the fragment is the polypeptide defined above.

According to yet another aspect of the present invention, there is provided a system for ameliorating a GAD65 associated autoimmune disorder in a patient comprising a fragment of GAD65 and a means for delivery of the fragment of GAD65 to a patient, wherein the fragment is the polypeptide defined above.

According to a further aspect of the present invention, there is provided a method for detecting the status of a GAD65 associated autoimmune disorder in a patient which comprises contacting a T-cell of the patient with at least one polypeptide fragment defined above and detecting the response of the T-cell to the peptide.

According to yet a further aspect of the present invention, there is provided a kit useful for the detection of antibody to the polypeptide defined above in a specimen suspected of containing said antibody, the kit comprising carrier means compartmentalized to receive in close confinement therein one or more containers comprising a container containing the polypeptide defined above.

According to still a further aspect of the present invention, there is provided a kit useful for determining the status of a GAD65 associated disorder in a specimen of a patient, the kit comprising carrier means compartmentalized to receive in close confinement therein one or more containers comprising a container containing the polypeptide fragment defined above.

- 3b -According to another aspect of the present invention, there is provided a method for detecting a GAD65 associated autoimmune disorder in a patient having or at risk of having the disorder, which comprises administering to the patient a diagnostically effective amount of a diagnostically effective fragment of GAD65, wherein the diagnostically effective fragment is the polypeptide defined above.

The present invention arose out of the discovery that recombinant DNA technology could be used to produce eukaryotic GAD65 polypeptide and that GAD65 polypeptide could be used in the diagnosis and therapy of patients with autoimmune disease. Particularly relevant is the use of eukaryotic GAD65 polypeptide in the diagnosis and therapy of patients having, or at risk of having, GAD65-associated autoimmune disorders such as insulin-dependent diabetes mellitus (IDDM) or stiff man disease.

A major advantage of the present invention is that it provides the art with a ready source of eukaryotic GAD65 polypeptide corresponding to that purified from natural sources, while avoiding the problems associated with the isolation of naturally occurring eukaryotic GAD65 polypeptide when separating it from other eukaryotic non-GAD65 polypeptides. This absence of other eukaryotic non-GAD65 polypeptides is significant in that it allows the development of test systems which will only detect antibodies specifically reactive with GAD65 polypeptides ! O ~ ~
WO 95/07992 G..1PCT/US9-1/09478 Another advantage of providing eukaryotic GADO
polypeptide in host cells is that by so doing, it is possible to obtain much larger quantities of the polypeptide than are currently practicably available ~
from natural sources. As a consequence, not only is it possible to use the polypeptide of the invention to ~
more accurately classify and treat patients with such autoimmune diseases as IDDM, but it is also now possible to provide commercially useful quantities of GAD polypeptide for use in diagnostic systems and pharmaceutical compositions.
DESCRIPTION OF THE DRAWINGS
FIGURE 1 Cloning strategy for obtaining GAD65 and GAD67 spec if ic cDNA probes.
FIGURE 2 DNA sequence and corresponding amino acid sequence for rat GADO.
FIGURE 3 DNA sequence and corresponding amino acid sequence for human GADbs.
FIGURE 4 Comparison of rat GAD65 and human GAD65 amino acid sequences.
FIGURE 5 GAD65 and GAD67 cDNAs hybridize to different size RNAs.
FIGURE 6 Southern blots hybridized with CDNA
probes specific for GAD65 and GAD67.
FIGURE 7 Immunological identification of GAD65 and GAD67.
FIGURE 8 Proliferative T-cell responses of NOD
mice to J3 cell antigens.
FIGURE 9 Characterization of GAD specific T-cell response of NOD mice as primed Thl cells by enhanced clonal size (a) and cell surface markers (b) and IFNy secretion.
FIGURE 10 Intramolecular spreading of T-cell autoimmunity within the GAD molecule.
FIGURE 11 Delay of onset of IDDM following immunization with GADO.

~

DETAILED DESCRIPTION OF THE INVENTION
The present invention relates to the manipulation of genetic materials by recombinant DNA procedures which make possible the production of polypeptides possessing part or all of the primary structural conformation for one or more of the epitopes for binding autoantibodies to glutamic acid decarboxylase6s (GAD65) and for polypeptides that bind to MHC receptors to block T-cell recognition. These polypeptides are highly useful for the immunological detection of autoantibodies reactive with them, since such autoantibodies are pre-diagnostic and indicative of autoimmune diseases such as insulin dependent diabetes mellitus and "stiff man" syndrome. These polypeptides can also be used for purposes of screening drugs, such as those that alter GAD function, and for generation of polyclonal and monoclonal antibodies which, in turn, can be used diagnostically to detect GADO.
The development of specific DNA sequences encoding eukaryotic GADO polypeptide for splicing into DNA
vectors can be accomplished using a variety of techniques. For example, alternative methods which can be employed include (1) the isolation of a double stranded DNA sequence from the genomic DNA of the eukaryote; (2) the chemical manufacture of a DNA
sequence to provide the necessary codons for the polypeptide of interest; and (3) the in vitro synthesis of a double stranded DNA sequence by reverse transcription of mRNA isolated from a eukaryotic donor cell. In the latter case, a double stranded DNA
complement of MRNA is eventually formed which is generally referred to as CDNA.
The manufacture of DNA sequences is frequently the method of choice when the entire sequence of amino acid residues of the desired polypeptide product is known.
When the entire sequence of amino acid residues of the desired polypeptide is not known, the direct manufacture of DNA sequences is not possible and the method of choice is the formation of cDNA sequences.

Among the standard procedures for isolating cDNA
sequences of interest is the formation of plasmid-carrying cDNA libraries which are derived from reverse transcription of mRNA which is abundant in donor cells 5 that have a high level of genetic expression. When used in combination with polymerase chain reaction technology, even rare expression products can be cloned. In those cases where significant portions of the amino acid sequence of the polypeptide are known, the production of labeled single or double stranded DNA
or RNA probe sequences duplicating a sequence putatively present in the target cDNA may be employed in DNA/DNA hybridization procedures which are carried out on cloned copies of the cDNA which have been denatured into a single stranded form (Jay, et al., Nucleic Acid Research, 11:2325, 1983).
Hybridization procedures are useful for the screening of recombinant clones by using labeled mixed synthetic oligonucleotide probes wherein each is potentially the complete complement of a specific DNA
sequence in the hybridization sample which includes a heterogeneous mixture of denatured double stranded DNA.
For such screening, hybridization is preferably performed on either single stranded DNA or denatured double stranded DNA. These procedures are particularly useful in the detection of cDNA clones derived from sources where an extremely low amount of mRNA sequences relating to the polypeptide of interest are present.
In other words, by using stringent hybridization conditions directed toward avoidance of non-specific binding, it is possible, for example, to allow the autoradiographic visualization of a specific cDNA clone by the hybridization of the target DNA to that single probe in the mixture which is its complete complement (Wallace, et al., Nucleic Acid Research, 9:879, 1981).
In addition, a GAD cDNA library can be screened by injecting the various cDNAs into oocytes, allowing sufficient time for expression of the cDNA gene products to occur, and testing for the presence of the .,~:.;~.;....

desired cDNA expression product, for example, by using antibody specific for GAD65 polypeptide, by using functional assays for GAD65 enzymatic activity, or by measuring the ability of the expression product to stimulate pathogenic T-cells.
Alternatively, a cDNA library can be screened indirectly for GAD65 peptides having at least one epitope using antibodies to GAD65 (Chang and Gottlieb, J.Neurosci., 8:2123, 1988). Such antibodies can be either polyclonally or monoclonally derived and used to detect expression product indicative of the presence of GAD65 cDNA. Preferred are antibodies directed to an epitope found in the first 100 amino acids of the N-terminal portion of GADO.
Of the three above-noted methods for developing specific DNA sequences for use in recombinant procedures, the use of genomic DNA isolates, is the least common. This is especially true when it is desirable to obtain the microbial expression of mammalian polypeptides because of the presence of introns.
The present invention provides novel polypeptides of GAD65 which have part or all of the primary structural conformation, that is, a continuous sequence of amino acid residues, having at least one epitope for antibodies to GAD65 or at least one determinant for T-cell recognition. It is possible to use the polypeptide fragments of the invention rather than intact GAD to detect autoantibodies to GAD. The term "polypeptide," as applied to GAD polypeptide, includes any sequence of amino acids having an epitope for autoantibodies to GAD or binds to a T-cell MIIC
receptor. Thus, the polypeptide fragments of GAD
encompassed by the invention possess a biological activity such as the ability to induce and/or bind autoantibodies to GAD, bind to T-cell MHC receptors (especially receptors on pathogenic T-cells) and the like.

The polypeptides resulting from microbial expression of the DNA sequences of the invention or from other synthetic techniques, such as solid-phase peptide synthesis, can be further characterized by their freedom from association with other eukaryotic polypeptides or other contaminants which might otherwise be associated with GAD in its natural cellular environment or in such extracellular fluids as plasma or urine.
Studies by the present inventors unequivocally establish that GAD65 and GAD67 are encoded by distinct genes and are not produced, for example, by post-transcriptional or post-translational modification of a common genomic sequence. Evidence proving that GAD65 and GAD67 are encoded by different genes include:
(a) the largest contiguous sequence of exact identity between GADO and GAD67 cDNAs is only 17 nucleotides in length, (b) cDNAs from GADO and GAD67 do not cross hybridize with each other's or with each other's mRNA
under low stringency conditions (2.0 x SSC, 0.01% SDS, 23 C), and (c) GADO and GAD67 cDNAs do not cross hybridize with isolated genomic clones encoding GAD67 and GAD65, respectively.
The term "host" includes not only prokaryotes, but also such eukaryotes as yeast, filamentous fungi, plant and animal cells, as well as insect cells which can replicate and express an intron-free DNA sequence of eukaryotic GADO. However, prokaryotes are preferred as the host organism for screening purposes while eukaryotic cells, especially insect cells, are preferred for expression.
The term "prokaryotes" includes all bacteria which can be transformed or transfected with the gene for the expression of GAD65. Prokaryotic hosts may include gram negative as well as gram positive bacteria such as, for example, E. co1i, S. typhimurium, Serratia marcescens and Bacillus subtilis.

A recombinant DNA molecule coding for the GAD, polypeptides can be used to transform or transfect the host using any of the techniques commonly known to those of ordinary skill in the art. Especially preferred is the use of a plasmid or a virus containing the GAD65 coding sequence for purposes of prokaryotic transformation or transfection, respectively.
Alternatively, liposomes containing the DNA of interest can be used to obtain expression in the host (Zhu, et al., Science, 261:209, 1993) Methods for preparing fused, operably linked genes and expressing them in bacteria are well-known in the art (Maniatis, et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989). The genetic constructs and methods described therein can be utilized for expression of GAD65 in prokaryotic hosts.
In general, expression vectors containing promotor sequences which facilitate the efficient transcription of the inserted eukaryotic genetic sequence are used in connection with the host. The expression vector typically contains an origin of replication, a promoter, and a terminator, as well as specific genes which are capable of providing phenotypic selection of the transformed cells. The transformed prokaryotic hosts can be grown in fermentors and cultured according to techniques known in the art to achieve optimal cell growth. The polypeptides of the invention can then be isolated from the grown medium, cellular lysates, or cellular membrane fractions.
The isolation and purification of the expressed polypeptides of the invention may be by any conventional means such as, for example, preparative chromatographic separations and immunological separations such as those involving the use of monoclonal or polyclonal antibody.
By having provided the sequence of amino acid residues of GAD65r the present invention provides for the manufacture of DNA sequences which code for the host expression of polypeptide analogs or derivatives of GAD65 which differ from naturally-occurring forms in terms of the identity or location of one or more amino acid residues and which share some or all of the epitopes of naturally-occurring polypeptide forms.
The novel DNA sequences of the invention include all sequences useful in providing the expression in prokaryotic or eukaryotic host cells of polypeptides which have at least a part of the primary structural conformation for one or more epitopes capable of reacting with autoantibodies to GADO which are comprehended by: (a) the DNA sequence as set forth in Figures 2 or 3 or their complementary strands; (b) DNA
sequences which hybridize to DNA sequences defined in 15. (a) or fragments thereof; and (c) DNA sequences which, but for the degeneracy of the genetic code, would hybridize to DNA sequences defined in (a) and (b) above. Specifically comprehended in (b) are genomic DNA sequences which encode allelic variant forms of GAD... Part (c) specifically comprehends the manufacture of DNA sequences which encode GAD6s, GADes fragments, and GAD65 analogs wherein the DNA sequences thereof may incorporate codons which facilitate translation of mRNA in non-vertebrate hosts.
Since the cDNA sequence of the invention encodes essentially the entire human or rat GAD6S molecule, it is now a matter of routine to prepare, subclone, and express smaller polypeptide fragments of cDNA from this or a corresponding cDNA sequence which would encode as few as one epitope for autoantibodies to human or rat GAD65. The presence of such an epitope on a cloned polypeptide can then be confirmed using, for example, serum from a patient with autoantibodies to GAD65. An example of such a smaller peptide is the first approximately 100 amino acids from the N-terminus of GAD6S (shown in Figure 3). This amino acid sequence is essentially absent from GAD67. Other examples of specific peptides of the invention are shown in Table 7 as well as the approximate carboxy-terminal two-thirds of GAD from about amino acid 224 to about amino acid 585. Especially preferred in the carboxy-terminal two-thirds of GAD is the amino acid segment from about amino acid 224 to about amino acid 398.
The present invention further relates to monoclonal antibodies which are specific for the polypeptides of the invention as well as the diagnostic and therapeutic use of these monoclonal antibodies. This specificity enables the monoclonal antibody, and like monoclonal antibodies with like specificity, to be used to bind the polypeptide of the invention when the polypeptide, or amino acids comprising the polypeptide, are present in specimens or a host, such as a human.
Numerous techniques can be utilized to produce the monoclonal antibodies of the invention without resorting to undue experimentation. To a great extent, the products of such monoclonal antibodies is rendered routine because of the highly defined nature of the polypeptides of the invention. Thus, whether the polypeptides of the invention are used for immunization and/or screening, the very limited number of immunogenic determinants on the polypeptides greatly simplifies the identification of cell lines producing monoclonal antibodies of the invention, for example, by limiting the repertoire of clonal expression possible.
One very useful type of cell line for expression of the monoclonal antibodies of the invention is the hybridoma. The general method used for production of hybridomas producing monoclonal antibody is well known (Kohler and Milstein, Nature, 256:495, 1975). The resulting hybridomas were then screened for production of monoclonal antibodies capable of binding to the polypeptides of the invention.
The techniques of sensitization and/or immunization, cell fusion, ascites production, selection of mixed hybridomas, or subcloning of monoclonal hybridomas are generally well known in the art. Attention is directed to Koprowski, et al., U.S.

Patent No. 4,172,124, Koprowski, et al., U.S. Patent No. 4,196,265, or Douillard, J.Y. and Hoffman, T., Basic Facts about Bybrzdomas, in Compendium of Immunology, Vol. II, L. Schwartz, ed. (1981).
In general, the purified peptides can be modified to have a cystine attached at the C-terminus to permit unidirectional attachment of the synthetic peptide to an immunogenic protein through a connecting bridge, for example, maleimidobenzoylated (MB)-keyhole limpet hemocyanin (RLH). Other immunogenic conjugates can also be used, for example, albumin, and the like. The resulting structure may have several peptide structures linked to one molecule of protein.
somatic cells derived from a host immunized against the synthetic peptides can be obtained by any suitable immunization technigue. The host subject is immunized by administering the antigen, usually in the form of a protein conjugate, as indicated above, by any suitable method, preferably by injection, either intraperitoneally, intravenously, subcutaneously, or by intra-foot pad. Adjuvants may be included in the ir-munization protocol.
The initial immunization with the protein bound antigen can be followed by several booster injections given periodically at intervals of several weeks. The antibody contained in the plasma of each host can then be tested for its reactivity with the immunizing polypeptide of the invention. The host having the highest response is usually most desirable as the donor of the antibody secreting somatic cells used in the production of hybridomas. Alternatively, hyperimmunization can be effected by repeatedly injecting additional amounts of peptide-protein conjugate by intravenous and/or i-ntraperitoneal route.

~ . :.:.

-The isolation of hybridomas producing monoclonal antibodies of the invention can be accomplished using routine screening techniques which permit determination of the elementary reaction pattern of the monoclonal antibody of interest. Thus, if a monoclonal antibody being tested binds with a polypeptide of the invention, then the antibody being tested and the antibody produced by the hybridomas of the invention are equivalent.
Alternatively, since the invention teaches polypeptides or amino acid sequences which are specifically required for binding of the preferred monoclonal antibodies of the invention, it is now possible to use these peptides for purposes of 15. immunization to produce hybridomas which, in turn, produce monoclonal antibodies specific for the polypeptide. This approach has the added advantage of decreasing the repertoire of monoclonal antibodies generated by limiting the number of antigenic determinants presented at immunization by the polypeptide. The monoclonal antibodies produced by this method can be screened for specificity using standard techniques, for example, by binding polypeptide to a microtiter plate and measuring binding of the monoclonal antibody by an ELISA assay.
It is also possible to determine, without undue experimentation, if a monoclonal antibody has the same specificity as a monoclonal antibody of the invention by ascertaining whether the former prevents the latter from binding the polypeptide of the invention. If the monoclonal antibody being tested competes with the monoclonal antibody of the invention, as shown by a decrease in binding by the monoclonal antibody of the invention, then it is likely that the two monoclonal antibodies bind to the same, or a closely related, epitope.
Still another way to determine whether a monoclonal antibody has the specificity of a monoclonal antibody of the invention is to pre-incubate the monoclonal f y -7 WO 95/07992 2170523 PCT/US9 i/09478 antibody of the invention with the polypeptide of the invention with which it is normally reactive, and then add the monoclonal antibody being tested to determine if the monoclonal antibody being tested is inhibited in its ability to bind the antigen. If the monoclonal antibody being tested is inhibited then, in all likelihood, it has the same, or a closely related, epitopic specificity as the monoclonal antibody of the invention.
The GADbj of the invention is particularly suited for use in immunoassays in which it can be utilized in liquid phase or bound to a solid phase carrier. In addition, GAD65 used in these assays can be detectably labeled in various ways.
Examples of immunoassays which can utilize the GAD6s of the invention are competitive and non-competitive immunoassays in either a direct or indirect format.
Examples of such immunoassays are the radioimmunoassay (RIA), the sandwich (immunometric assay) and the Western blot assay. Detection of antibodies which bind to the GADO of the invention can be done utilizing immunoassays which run in either the forward, reverse, or simultaneous modes, including immunohistochemical assays on physiological samples. The concentration of GAD6S which is used will vary depending on the type of immunoassay and nature of the detectable label which is used. However, regardless of the type of immunoassay which is used, the concentration of GAD65 utilized can be readily determined by one of ordinary skill in the art using routine experimentation.
The GAD and GAD fragments of the invention can be bound to many different carriers and used to detect the presence of antibody specifically reactive with the polypeptide. Alternatively, the carrier-bound GAD and GAD fragments can be used therapeutically for extracorporeal absorption of autoimmune antibodies in patients having, or at risk of having, GAD-associated disorders. Examples of well-known carriers include alass, volvstvrene. Dolvvinvl chloride, Dolvpronvlene, ~ WO 95/07992 2170523 PCT/US94/09475 , polyethylene, polycarbonate, dextran, nylon, amyloses, natural and modified celluloses, polyacrylamides, agaroses, and magnetite. The nature of the carrier can be either soluble or insoluble for purposes of the invention. Those skilled in the art will know of other suitable carriers for binding GAD65, or will be able to ascertain such, using routine experimentation.
There are many different labels and methods of labeling known to those of ordinary skill in the art.
Examples of the types of labels which can be used in the present invention include enzymes, radioisotopes, colloidal metals, fluorescent compounds, chemiluminescent compounds, and bioluminescent compounds.
Alternatively, the polypeptide of the invention which comprises the GAD enzymatic domain can be used to detect antibodies to GAD by measuring GAD enzymatic activity. For example, GADO and a specimen suspected of having antibodies to GAD65 can be incubated for a period of time and under conditions sufficient to allow binding to occur between GADO and the antibodies. The reaction product is precipitated and then tested for GAD enzymatic activity.
For purposes of the invention, the antibody which binds to GAD65 of the invention may be present in various biological fluids and tissues. Any sample containing a detectable amount of antibodies to GAD, can be used. Normally, a sample is a liquid such as urine, saliva, cerebrospinal fluid, blood, serum and the like, or a solid or semi-solid such as tissue, feces and the like.
The materials for use in the assay of the invention are ideally suited for the preparation of a kit. Such a kit may comprise a carrier means-being compartmentalized to receive in close confinement one or more container means such as vials, tubes and the like, each of the container means comprising one of the separate elements to be used in the method. For examDle. orie of the container means mav comDrise GAM, : .~ 2170523 bound to a carrier. A second container may comprise soluble, detectably-labeled second antibody, in lyophilized form or in solution.
In addition, the carrier means may also contain a plurality of containers each of which comprises different, predetermined amounts of GAD65. These latter containers can then be used to prepare a standard curve into which can be interpolated the results obtained from the sample containing the unknown amount of autoantibodies to GAD65.
In using the kit all the user has to do is add, to a container, a premeasured amount of a sample containing a measurable, yet unknown amount of autoantibodies to GADO to be detected, a premeasured amount of carrier-bound GAD65 present in the first container, and a premeasured amount of the detectably labeled second antibody present in the second container. Alternatively, the non-detectably labeled GAD63 can be provided attached to the container to which the sample and the detectably labeled second antibody are added. After an appropriate time for incubation, an immune complex is formed and is separated from the supernatant fluid, and the immune complex or the supernatant fluid are detected, as by radioactive counting or addition of an enzyme substrate, and color development.
In an alternative embodiment, a kit comprising the GAD polypeptide of the invention can be used to detect the stage of GAD-associated autoimmune disease in a patient. As further shown herein, Applicants have discovered that certain GAD peptides or fragments are associated with different levels of progression in the autoimmune disease and that the level of disease process can be ascertained by looking at immune cell proliferative response, such as that of the pathogenic T-cell of the patient.
The term "ameliorate" denotes a lessening of the detrimental effect of the autoimmune response in the patient receivinct therapv= The term 11therapeuticallv ~ WO 95/07992 17 - 2170523 PCT/US94/09478 -effective" means that the amount of GAD65 polypeptide used is of sufficient quantity to ameliorate the cause of disease due to the autoimmune response.
The GAD65 polypeptides, including whole GADO, of the invention can be used therapeutically in patients having, or at risk of having, an autoimmune response associated with GAD65. Such therapy can be accomplished, for example, by the administration of GAD65 polypeptide to induce tolerance to GAD. Such administration can utilize unlabeled as well as labeled GAD65 polypeptide. When unlabeled GADO polypeptide is utilized advantageously, it would be in a form wherein, for example, the GADO polypeptides are in fragments which are too small to stimulate an immune response, but large enough to bind, or block, the continuance of the autoimmune response. For example, GADO could be digested enzymatically into epitope-sized peptides (typically 5-12 amino acids in length) and thereby bind to Fab binding portions present in the body fluids, or on the surface of immune cells, of the patient with autoimmune disease. Alternatively, peptides having at least one determinant for binding to T-cell MHC
receptor can be similarly produced or chemically synthesized.
Alternatively, the GADO polypeptides of the invention can be administered labeled with a therapeutic agent. These agents can be coupled either directly or indirectly to the GADO polypeptides of the invention. One example of indirect coupling is by use of a spacer moiety. These spacer moieties, in turn, can be either insoluble or soluble (Diener, et al., Science, 231:148, 1986) and can be selected to enable drug release from the GAD65 polypeptide at the target site. Examples of therapeutic agents which can be coupled to the GAD65 polypeptides of the invention for immunotherapy are drugs, radioisotopes, lectins, and toxins.

-The drugs with which can be conjugated to the GAD65 polypeptides of the invention include compounds which are classically referred to as drugs such as mitomycin C, daunorubicin, and vinblastine.
In using radioisotopically conjugated GAD65 polypeptides of the invention for immunotherapy, certain isotopes may be more preferable than others depending on such factors as leukocyte distribution as well as stability and emission. Depending on the autoimmune response, some emitters may be preferable to others. In general, a and 0 particle-emitting radioisotopes are preferred in immunotherapy.
Preferred are short range, high energy a emitters such as 212 Bi. Examples of radioisotopes which can be bound to the GAD65 polypeptides of the invention for therapeutic purposes are 121I, 131I, 90Y, 67Cu, 212Bi, 211At, 212pb, 47Sc, 109Pd and lasRe.
Lectins are proteins, usually isolated from plant material, which bind to specific sugar moieties. Many lectins are also able to agglutinate cells and stimulate lymphocytes. However, ricin is a toxic lectin which has been used immunotherapeutically. This is accomplished by binding the a-peptide chain of ricin, which is responsible for toxicity, to the antibody molecule to enable site specific delivery of the toxic effect.
Toxins are poisonous substances produced by plants, animals, or microorganisms that, in sufficient dose, are often lethal. Diphtheria toxin is a substance produced by Corynebacterium diphtheria which can be used therapeutically. This toxin consists of an a and (3 subunit which under proper conditions can be separated. The toxic A component can be bound to GAD, polypeptide and used for site specific delivery to a leukocyte expressing a receptor for GAD65 polypeptide.
Other therapeutic agents which can be coupled to the GAD, polypeptides of the invention, as well as ex vivo and in vivo therapeutic protocols, are known, or can be easily ascertained, by those of ordinary skill in the art.
The present invention also relates to a polypeptide which can be administered therapeutically to ameliorate, or utilized diagnostically to identify, the disease process in patients having, or at risk of having, this disease. The conventional single-letter code used to represent the various amino acids relates as follows:

Phe: F Leu: L Ile: I Met: M
Val: V Ser: S Pro: P Thr: T
Ala: A Tyr: Y His: H Gln: Q
Asn: N Lys: K Asp: D Glu: E
Cys: C Trp: W Arg: R Gly: G

A polypeptide sequence of the invention was identified by comparing the amino acid sequences of human GADO, human GAD67, and the P2-C protein of the picornavirus, coxsackie virus. The P2-C polynucleotide plays a role in the virus membrane bound replication complex. These analyses established the presence of an extensive sequence similarity between both GADO
molecules and the coxsackie virus. A core polypeptide of six contiguous amino acid residues of the GAD65 and P2-C polypeptide are identical in amino acid sequence.
Indeed, of the 24 amino acids in the polypeptide, 19 are identical or conserved. In addition, there also exists a high charge density and the presence of a proline residue which would render this region highly antigenic (see Table 2).

,..; 2170523 WO 95/07992 PCT/US94/09 t78 Ul O
N

r-1 U
QI
d =

-r1 4.i r-1 -~i U
b =~
dA
.0 go a vo >1 4J .0 ~o a a w a~
>
>+

m U 7~ w a a x c ~i > c O+
pq I-I E-4 fs1 W =~l $4 E 44 14 o ~ > >

W W H U Rf ~~
O
w w ~to =
O
~ - a m +J
>4 w 0 3 a a w ri IA

w -r1 XI N w, H ~', =ri V~1/
rl co O G O 1~ tOA O
-~ C7 C~ U a .OC ~
d O ~ af t11 H cd k x 0 >

~

In Table 2, the solid line encloses identical amino acids whereas the dashed line encloses amino acid residues with similar charge, polarity, or hydrophobicity.
The discovery of this common polypeptide region = supports an etiologic role for "molecular mimicry" in the precipitation of diabetes. Thus, where a patient genetically susceptible to IDDM is infected by a coxsackie virus, the immune response to the similar GAD
sequence in the patient's (3-cells. The immunological response is maintained by the antigenically similar GAD
polypeptides resulting in the eventual destruction of the fl-cells and the subsequent presentation of IDDM.
At present, it is believed that the destruction of pancreatic fl-cells in IDDM is mediated by a cellular autoimmune response. As described herein, a polypeptide of the invention can ameliorate the autoimmune response to GAD. Because of the complexity of autoimmune disease, it is possible to envision numerous possible therapeutic modalities which would allow the polypeptides of the invention to be used to ameliorate such diseases. In one embodiment, it appears that the polypeptides of the invention can be utilized to block recognition by a specific T cell.
receptor (TCR) or an MHC receptor presenting an autoimmune antigen on the surface of an antigen presenting cell (APC). The inhibition of such recognition might occur, for example, by providing the patient with the polypeptide of the invention which, in turn, can displace the autoimmune antigen being presented in the antigen-cleft of the MHC receptor.
However, although not wanting to be bound to a particular theory, it is believed that the polypeptides of the invention probably act to induce or restore a tolerogenic state by direct interaction with the appropriate TCR on the surface of a GAD specific pathogenic T-cell. This latter therapeutic approach of direct interaction with the TCR is supported by the examples and suggests that suppression of the autoimmune response can be achieved through induction of high-zone tolerance by use of high concentrations of polypeptide, preferably soluble. Another possible mechanism is that the polypeptide of the invention may play a role in anergizing pathogenic T cells by binding to the T cell MHC receptor, thereby preventing the appropriate costimulatory signal.
Alternatively, the polypeptides of the invention could be used to stimulate a T-suppressor cell population in order to restore self-recognition and, thereby, ameliorate the autoimmune disease.
Stimulation of T-suppressor cell populations could be achieved, for example, by use of a bi-specific antibody having one variable region specific for an epitope present on the autoimmune antigen residing in the cleft of the MHCII receptor and, a second variable region specific for an epitope present on the CD8+ receptor.
The production of antibody specific for the polypeptide of the invention is a matter of routine to those of skill in the art, as is the preparation of bi-specific antibodies having specificity for 2 or more epitopes.
Polypeptide analogs of the present invention may be designed which will compete for recognition of self-antigens at the level of antigen presentation or induce anergy in T cells, due to a lack of a costimulatory signal. Since MHC molecules contain a single peptide binding site, it is possible to design polypeptides which will bind with high affinity to disease-associated MHC molecules, but will not activate disease-causing T-helper cells. Such polypeptides act as antagonists for self-antigen recognition. In the present invention, support for this mechanism is found in the examples, especially Example 7. Precedent for such an approach arises from observation that a mouse lysozyme polypeptide, itself non-immunogenic, can ~

compete for MHC binding with an immunogenic polypeptide from hen-egg white lysozyme and thereby reduce T cell activation by that polypeptide (Adorini, et al., Nature, 334:623-625, 1988) as well as studies using T-cell receptor peptides to block formation of complex = between T-cells, autoantigen and MHC (Howell, et al., Science, 246:668, 1989). Similarly, such a therapeutic approach for screening effective polypeptide analogs has been utilized in such autoimmune diseases as experimental autoimmune encephalomyelitis (EAE) (Wraith, et al., Cell, 59:248, 1989; Urban, et al., Cell, 59=257, 1989).
The single-letter symbols used to represent the amino acid residues in the polypeptides of the present invention are those symbols commonly used in the art.
The peptides of the invention include not only the natural amino acid sequences, but also peptides which are analogs, chemical derivatives, or salts thereof.
The term "analog" or "conservative variation" refers to any polypeptide having a substantially identical amino acid sequence to a polypeptide provided herein and in which one or more amino acids have been substituted with chemically similar amino acids. For example, one polar amino acid, such as glycine or serine, may be substituted for another polar amino acid; or one acidic amino acid, such as aspartic acid may be substituted for another acidic amino acid, such as glutamic acid;
or a basic amino acid, such as lysine, arginine, or ,histidine may be substituted for another basic amino acid; or a non-polar amino acid such as alanine, leucine, or isoleucine may be substituted for another non-polar amino acid.
The term "analog" or "conservative variation" also means any polypeptide which has one or more amino acids deleted from or added to a polypeptide of the present invention, but which still retains a substantial amino acid sequence homology to such peptide. A substantial -sequence homology is any homology greater than 70%, preferably at least about 80%, and more preferably at least about 90%. The term "fragment" also means any shorter version of the polypeptides identified herein having at least 6 amino acid residues, wherein the fragment possesses biological activity, or is a fragment capable of inhibiting the stimulation of T-cells by a stimulating polypeptide fragment or substantially full-length molecule.
The term "chemical derivative" means any polypeptide derived from a polypeptide of the present invention and in which one or more amino acids have been chemically derivatized by reaction of the functional side groups of amino acid residues present in the polypeptide. Thus, a "chemical derivative" is a polypeptide that is derived from the sequences or polypeptides identified herein by one or more chemical steps. Such derivatized molecules include, for example, those molecules in which free amino groups have been derivatized to form amine hydrochlorides, P-toluene sulfoamides, benzoxycarboamides, T-butyloxycar-boamides, thiourethane-type derivatives, trifluoroacet-ylamides, chloroaceamides, or formamides. Free carbox-yl groups may be derivatized to form salts, methyl and ethyl esters or other types of esters or hydrazides.
Free hydroxyl groups may be derivatized to form 0-acyl or 0-alkyl derivatives. The imidazole nitrogen of histidine may be derivatized to form N-im-benzylhis-tidine. Also included as chemical derivatives are those polypeptides which contain one or more naturally occurring amino acids derivatives of the 20 standard amino acids. For example, 4-hydroxyproline may be substituted for proline; 5-hydroxylysine may be substituted for lysine; 3-methylhistidine may be substituted for histidine; homoserine may be substituted for serine, and ornithine may be substituted for lysine.

~ WO 95/07992 It should be understood that the present invention is not limited to the illustrative polypeptides depicted in Table 2 and Table 9, instead, a polypeptide falling within the scope of this invention may extend outside of, or comprise less than, the region between amino acid 28 and amino acid 50 of coxsackie virus P2-C, or between amino acid 250 and amino acid 273 of GADO, or between amino acid 258 and amino acid 281 of GAD67, as well as the region between amino acid 78 and amino acid 97, or between amino acid 247 and amino acid 266, or between amino acid 335 and amino acid 356, or between amino acid 479 and amino acid 498, or between amino acid 509 and amino acid 528, or between amino acid 524 and amino acid 543, or between amino acid 539 and amino acid 556, or between amino acid 564 and amino acid 583 of GADO, as long as a substantial part of a given polypeptide is characterized by an amino acid sequence from that region, or segments or combinations thereof, and the polypeptide demonstrates the desired immunological or biological activity against autoimmune disease. In addition, polypeptides according to this invention include those having amino acid sequences which are longer or shorter in length than those of the polypeptides illustrated in Table 2 and Table 9, or which comprise segments or combinations thereof, as long as such polypeptides consist substantially of the region between the amino acids illustrated in Table 2 and Table 9 and demonstrate immunological or biological activity. All polypeptides of the invention should not stimulate or enhance the autoimmune disease.
Accordingly, it should be understood that the specific'selection of any one polypeptide within the polypeptides of the invention does not involve undue experimentation. Such a selection,may be carried out by taking a number of polypeptides and testing them for their immunological and biological activity in ameliorating the autoimmune disease or for detecting antibody. The NOD mouse represents an excellent and well characterized model for screening polypeptides of the invention capable of ameliorating or preventing diabetes. Example 7 illustrates an acceptable procedure for routine screening of candidate polypeptides with biologic activity.
The polypeptides according to the present invention may be prepared by recombinant techniques or by conventional synthesis using known polypeptide synthetic methods, including synthesis on a solid support. An example of a suitable solid phase synthetic technique is that described by Merriweather (J.Am.Chem.Soc., 85:2149, 1963). Other polypeptide synthetic techniques may be found, for example, in Bodanszky, et al., Peptide Synthesis, John Wiley &
Sons, 2d ed., 1976, as well as other references known to those skilled in the art. A summary of polypeptide synthesis techniques can be found in Stewart, et al., Solid Phase Peptide Synthesis, Pierce Chemical Company, Inc., Rockford, I11., 1984. The synthesis of polypeptides by solution methods may also be used, for example, as described in The Proteins, Vol. II, 3d ed., Neurath, et al., eds., 105, Academic Press, New York, NY, 1976. Appropriate protective groups for use in such synthesis can be found in the above references as well as in J. McOmie, Protective Groups in Organic Chemistry, Plenum Press, New York, NY, 1973.
The polypeptides of the invention may also be prepared in an appropriate host transformed with DNA
sequences that code for the desired polypeptide. For example, a polypeptide may be prepared by the fermentation of appropriate hosts that have been transformed with and which express a DNA sequence encoding the polypeptide. Alternatively, a DNA
sequence encoding several of the polypeptides of this invention may be linked together and those sequences may then be used to transform an appropriate host to permit the expression of polypeptides involved in the autoimmune disease.
The dosage ranges for the administration of the GAD
polypeptides of the invention are those large enough to produce the desired effect in which the symptoms or cellular destruction of the autoimmune response are ameliorated. The dosage should not be so large as to cause adverse side effects, such as unwanted cross-reactions, anaphylactic reactions, and the like.
Generally, the dosage will vary with the age, condition, sex, and extent of the disease in the patient and can be determined by one of skill in the art. The dosage can be adjusted by the individual physician in the event of any counterindications.
Dosage can vary from about 0.lmg/m2 to about 2000mg/m2, preferably about 0.lmg/m2 to about 500mg/m2/dose, in one or more dose administrations daily, for one or several days.
The GAD polypeptides of the invention can be administered parenterally by injection or by gradual perfusion over time. The GAD polypeptides of the invention can be administered intravenously, intraperitoneally, intra-muscularly, subcutaneously, intracavity, transdermally, intranasally, or enterally.
Preparations for parenteral administration include sterile aqueous or non-aqueous solutions, suspensions, and emulsions. Examples of non-aqueous solvents are propylene glycol, polyethylene glycol, vegetable oils such as olive oil, and injectable organic esters such as ethyl oleate. Aqueous carriers include water, alcoholic/aqueous solutions, emulsions or suspensions, including saline and buffered media. Parenteral vehicles include sodium chloride solution, Ringer's dextrose, dextrose and sodium chloride, lactated Ringer's, or fixed oils. Intravenous vehicles include fluid and nutrient replenishers, electrolyte replenishers (such as those based on Ringer's dex-WO 95/07992 21 7~1 523 PCT/US94/09-178 U

trose), and the like. Preservatives and other additives may also be present such as, for example, antimicrobials, anti-oxidants, chelating agents, and inert gases and the like.
The invention also relates to a method for preparing a medicament or pharmaceutical composition comprising the GAD65 polypeptides of the invention, the medicament being used for therapy of autoimmune response to GAD6s.
The above disclosure generally describes the present invention. A more complete understanding can be obtained by reference to the following specific examples which are provided herein for purposes of illustration only and are not intended to limit the 15, scope of the invention.

CLONING AND EXPRESSION OF GAD
A. RECOMBINANT DNA PROCEDURES
In order to obtain cDNA probes specific for GAD65 and GAD67, total RNA was extracted from adult rat brain by guanidine isothiocyanate-cesium gradient using the method of Chirgwin, et al. (Biochemistry, 18:5294, 1979). Poly (A) RNA was purified on oligo dT
cellulose, using the protocol by Bethesda Research Laboratories (BRL). First strand synthesis was performed by using MMLV-reverse transcriptase (BRL), with conditions suggested, except that poly d(N6)-mers (Pharmacia) were used as primers. This cDNA-RNA
mixture was heat inactivated at 65 C for 15 min and stored at -20 C. For PCR, 1/50 of the sample was added to the 100 l reaction. Degenerate oligonucleotides were synthesized (Applied Biosystems) to encode the underlined common amino acid sequences of feline (from cDNA) (Kobayashi, et al., J.Neurosci., 7:2768, 1987) and rat (from peptides) (Chang and Gottlieb, Sep-25-00 05:15pm From-SB,B/FBCo, +2328440 T-172 P.07/32 F-845 WO 95/07992 PCT/US94r04476 - ~9 -,T.NeurPsc,t=, .Q:2123, 1988) GAD (Figl,ixe 1). The 51-end sequenaa of each degenerata oligoriuGlevtida Cont.ained one strand of the DNA sequence recngr:ized by either SatY and 8indrIZ (5' oligo) or Sstl and SstII (3'-end oliga). The6e primers were usad fox selective amplification by polymerase chain reaction ot the generated cI1NA template as described by Could, et al.
(Px`oc.Nati.ACad.sai..U,sA, BEa:I934e 19$9), ricR products were subcloned into EiirdlYl/Sstl double digested Bluescript SR!vector (Stratagene), transformed into D115 (BRI,), and plated by standard methods (ManiatS,s, at ai., Maiecular Cloning: A Laboratory MsnuaZ, cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989).
colony hybridization wae done with . ari 5' ^9ZF end labeled o],igonualBOtide specific to feliAa GA-p,7 (EOk+a.yashi, et al., J'.NCuxoscj., 2_ 2768, 1987). End labeling of oligonucleotids, =hybridization conditions, and washing ooriditions were done as described (Wallace, at al., in Gujde to Molecular Cloning Techniques, aexger, et al., Zds. ixa, ldethods of F.t:zymoiogy; Abelson, at al., Eds. Academic Press, .rnc., San Diego, 432-442, 1987), except that the nitY-Qeellulos filter6 were washed at 5d C for 15 min. Colonies which were positive and negative in the hybr.fi,dization were individually picked and grown Pvernight in Terrific BXoth (Tartof, at al., FoCUs,=9:12, 1987). DNA was isolated using a boiliDg me'thod (Maniatis, et al., molacular Cloniag: A Laboratory Manual, Cold spring [iarboY`iLaboratory, Cold spring Harbor, NY, 1989) and templates i0ere denatured by 0.2N NaDH and purified kry Sephacryl S400* spun aol-upns (Fharmacia) . Sequericibg oP
= dene,tured daubl9 stranded te:nplate was by the chain-termination method (Sanger, at al _, ,P.roc.N'atl.Acad.3c.i.,USA, 2A:5453, 1977) using the T7-saquencing kit (Phhrmacia).

* Trade-mark WO 95/07992 2170523 PC'I'/US94/09478 As shown in Figure 1, PCR-generated rat GAD65 and GAD67 cDNAs were used as probes to screen a lambda ZAP
(Stratagene) rat hippocampus library provided by S.
Heinemann (Salk Institute) by standard techniques (Maniatis, et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989). A 2400 nucleotide GADO cDNA (the largest clone) was isolated and subcloned by "zapping"
as described by Stratagene. When a rat GAD67 cDNA was obtained which was smaller than a 3.2kb rat GAD67 cDNA
clone already on hand, the larger cDNA was sequenced.
Exo III deletions (Henikoff, Gene, 28:351, 1984) were made in both directions for GAD65 and GAD67 and templates were prepared and sequenced as described above.
Anchored PCR (Frohman, et al., Proc.Natl.Acad.Sci.,USA, 85:8998, 1988) was done to clone the remaining 5'-ends of GAD65 and GAD67 mRNAs which were not represented in the original cDNA clones isolated in the library screening. Sequencing of these clones revealed that neither GAD65 nor GAD67 mRNAs contained any further initiation codons (AUGs) in frame with the previously designated initiation codons of the original cDNA
clones.

CHARACTERIZATION OF CLONED GAD6s A. NORTHERN BLOT HYBRIDIZATION
Two PCR-derived cDNA probes were hybridized to Northern blots containing rat brain RNA in order to determine whether the GAD67 and GADO cDNAs were derived from two different mRNAs. RNA was extracted as described in Example 1. Poly (A) RNA was separated by electrophoresis in formaldehyde and transferred onto Biotrans (ICN) membranes, and hybridization was performed as described by Well, et al. (J.Neurosci., 16:311, 1986), except that 100 l/ml of poly (A) was Sep-25-00 05:16pm From-S&B/F&Co, t2328440 T-172 P.08/32 F-845 WO +95/07992 PGT/i7s94J04478 - 31 ~
addad. Probes were labeled to approximataly 109 dpm/Ag by the oligolAbelirig procedure of Feinberg and voqeYstein (Ana1 .8~.acl~em. rMs 6, 1993), Identioal . results were stxbsequantly obtained with fqll-lehgth S clo3las of GAD6$ and GAD47 cDNAs.
. As shown in Figura 5, lanes 1 and 2 contain lpg af poly (A) selected RNA extracted from rat cerebellum.
Lane 1 Was hybridized to a aDNA probe for the rat cognate of feline GADi, (Kobayashi, st &1., J.Neurasci., -7:2765, 1987) and lane 2 with a GDNA pI`Qbe for the 7Cat peptide sequence (whiokl aorresponds to GAD6$).
The CpAIA pt-obe for the rat peptide sequence hybridixed to a 5.7kb RNA, while tha cpNA probe for the rat cognate of feline GAD67 cDNA, hybridized to a 3.7kb RNA. This demonstrates that GADgd and GA1367 are not derived from the eame DINA.

8. GENOMZC RYARIDYZATION OF PAD-AND dAbe In order ta investigate the paseibility that GA067 and GAD65 arise from 6eparate genes, aDNp.S of both GAag7 and GAI7O were hybridized to DNA blots eontainirig genomic Dk=tA.
For Southern blots, genflmic pNA was extracted from rat liver as described (Kaiser, et al -, in DNA Cloning, vol.T, A Practical Approach, D.M. Glover ed., YRL
PrAss, Oxford, pp. 38=40, 1985). DNA (lbAg/sample) was digested to caxnpletlon with EcoRZ and Hindzli using conditions racomuaended by the $uppliers (SitL, Gaithersburg, MD). ANA fragments were separated by electrophoresis at 1.5v/cm for 16 hrs in 0.8% sgarose.
The DNA was then trattaferred to Zeta-Probe*tuembranes (Sio-Rad), hybridized, and washed, as tlesoribed by Gatti, et al, (Bl4techn.iques, 2:148, 1984), axcept that 5ug/rcl OaXnatiQn dried milk was substituted for Denhard't'6 solution. Probes i`or Southoxn blots were ].abeled as dascribed in Example 1, above.

* Trade-mark Sup-25-00 05:16pm From-SaB/F&Ca, +2328440 T-172 P.09/32 F-845 WO 9SNi7992 PGTlUS94107J7$

As ehown in Figure 6, genomic DNA digested with HindilT and EcoRT -are in lanes ]. and 3 and 1aYies 2 and 4, respective].y. GADQ7 cDNA was hybridized to laries i and 2, whereas GADbs cDNA was hybrj,dized to lanea 3 and 4. Ni4mber6 along the side of the gel are t}j$ DNA
fragment sizes in kiloriases.
This data shtiws that the two oDNAs hybridixe to gehotaiG fragments of different sizes. In addition, the greatest cQntiguou5 stretch=of identical nucleotide sequence of cADO and GAD+7 CDNAs is only 17 nqCleotide bases in length. Thus, CAa67 and GADas are encoded by two distinct genes.

C. ENZYHAT2C COHP ARXSON or GADC ANp GADg Studies ware done cprnparing the effect of pI,F on the activity of GAD67 and GADM. In s4 doing, both cDNAs Were subcloned int0 Vectors that allowed their expression in bacteria (Studier, at al., 18911i3, 1986). 4vero7cpression of "fusjonless" GADss ancl GAD,, Was accomplished by subcic-ning GAAsS oDNA into the Nco2 site of pET-8C and GAD67 oDNA into the Nhei ' site of pBx-5c vectors (Studier, et a.Tõ J.MOZ.B.iol., 189:i1~, 198b).
To obtain compatible sticky ends for correct ir-frame subc].onikig of both cDNAs, seleetive amplif.ication was peri'aY-fied by PCR using conditions suggested by United States Bioahemical (u98), With 200MM dNTPS and 1.5mM MgC12 in the mixture snd annealing at 55 C With 20 cycles to decrease infidelity of AxapliTAQ* (USS).
Primers Specific for GADO and GA0,, Contaihed one Dt4A, str,and of the Nool and Spe= reoognition slte.g, respective0.y. since theYe is a NheY restrictYon site within the coding region of GAp17, SpaZ (which is oompatible with Nhel) was used.
PCR prodtlcts were subcloned into their rempsctive pET vectors, transforme.d into DiiS and plated as described (Maniatis, et a2., ,Molecular Cloning: A

* Trade-mark . .. ,,; t...

~
- -Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989). Colonies were picked and grown overnight in LB broth with 50 g/ml ampicillin.
Subclones with correct orientation were transformed into BL21(DE3) strain (Studier, et al., J.MoI.Biol., = 189:113, 1986) for overexpression. As a negative control, the pET-8C vector with no insert was transformed and subsequently induced. Single colonies were picked, grown, induced by 1mM isopropyl-B-D-thiogalacto-pyranoside (IPTG), and analyzed on SDS-PAGE
gels as described (Sambrook, et al., Molecular Cloning a Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 17.15-17.16, 1989).
To measure GAD activity, we induced lOml cultures of bacteria at OD6w-0.5 with 1mM IPTG. Two hours after induction, bacteria was spun down and resuspended and sonicated in iml of homogenizing buffer (imM
phenylmethylsulfonyl fluoride (PMSF), 1mM 2-aminoethylisothiouronium bromide (AET), and 60mM
potassium phosphate, pH 7.1). After sonication, cell debris was removed by centrifugation and protein concentration was measured (Bradford, Anal.Biochem., 72:248, 1986) in the supernatant (supernatant was stored in aliquots at -70 C). Brain homogenates were prepared as described (Legay, et al., J.Neurochem., 46:1478, 1986). GAD activity was measured as described (Krieger, et al., J.Neurochem., 33:299, 1984) with 0.2mM PLP present or absent and 20 l of brain ,homogenate or bacterial lysate in the incubation mixture. Production of 14CO2 in bacterial lysates was linear with respect to time of incubation and protein concentration.

~

GAD SPecific Activity Fold Increase gource - PLP + PLP in induction BL21(DE3) + pET-8c 12 0.4 9 1 --BL21(DE3) + pET-GADw 115 3 773 61 6.7 BL21(DE3) + pET-GADc 160 2 389 8 2.4 Rat Brain 131 5 216 2 1.6 cpms of 14COZ/ g protein/hr of triplicates S.E.M.

As shown in Table 3, bacterial lysates containing GAD65 or GAD67 catalyze the conversion of [ 1-14C] -glutamate to GABA and 14C02.
PLP stimulates the enzymatic activity of GADO more than GADO. This greater stimulation probably reflects the faster cycling of GADO through the inactivation cycle proposed by Martin and coworkers (Martin, Cell.Mol.Neurobiol., 7:237, 1987). This faster cycling suggests that GAD6S contributes more to the pool of apo-GAD that exists in vivo (Miller, et al., Brain Res.Bull., 5(Supp1.2):89, 1980). Thus, in vivo, PLP
appears to regulate GAD65 activity more than GAD67 activity.
GAD63 activity in bacterial lysates is similar to the five-fold PLP stimulation of GAD activity found in synaptosomes prepared from rat substantia nigra (Miller, et al., J.Neurochem., 33:533, 1979). Because both GADs are more dependent upon added PLP in bacteria than is the GAD activity in crude rat brain homogenates, the endogenous PLP concentration of bacteria lysates may be less than rat brain homogenates.

~
- -D. IMMUNOLOGICAL IDENTIFICATION OF GAD6; AND GAD~-, Rat brain homogenates and bacterial lysates were extracted as described above. Equal volumes of loading buffer were added to each sample as described (Harlow, et al., Antibodies, A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1988).
Proteins were separated by electrophoresis in a 10%
acrylamide gel in SDS and electrophoretically transferred to nitrocellulose (Harlow, et al., Antibodies, A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1988). The unreacted sites were blocked with a phosphate buffered saline (PBS) solution containing 2% bovine serum albumin (fraction V), 1% gelatin, and 1% Triton-X-100 at 42 C for one hr. After washing, the nitrocellulose filter was then cut into three sections and incubated with the following primary antibodies: lanes 1 to 4 with a 1/2000 dilution of the antiserum of Oertel, et al. (Neuroscience, 6:2689, 1981), which recognizes both GAD67 and GADO; lanes 5-8 with a 1/2000 dilution of K-2 antiserum, which recognizes only GAD67; lanes 9-12 with a 1/2000 dilution of GAD-6 monoclonal antibody, which is specific for GAD65 (Chang, et al., J.Neurosci., 8:2123, 1988). All filters were extensively washed and appropriate secondary antibodies were incubated and washed. Bound antibodies were detected with luI-labeled protein A and autoradiography. Each lane contained the following: lanes 1, 5, and 9 are BL21(DE3) + pET-GAD67;
lanes 2, 6, and 10 are BL21(DE3) + pET-GAD65i lanes 3, 7, and 11 are rat brain homogenate; and lanes 4, 8, and 12 are BL21(DE3) + pET-8c.
The immunoblots of bacterially produced GAD65 and GAD67 demonstrated that GAD65 indeed corresponds to the smaller GAD in brain extracts, and GAD67 to the larger form (Figure 7). Previous work has demonstrated the WO 95/07992 36 - 2 i 7 0 5 2 3 pCT/US94/09478 -correspondence of GAD67 to the larger GAD for feline GAD67, and for mouse GAD67 (Katarova, et al., Eur.J.Neurosci., 2:190, 1990; 235, 1987). The mobilities of bacterially produced GAD65 and GAD67 (as detected with the antiserum of Oertel, et al.
(Neuroscience, 6:2689, 1981) are identical to the immunoreactive doublet seen in rat brain homogenate.
The smaller molecular weight and larger molecular weight forms of GAD in rat brain are thus identical in antigenicity and size to the products of GAD6S and GAD67 cDNAs, respectively. Consequently, the two GADs in rat brain are GAD65 and GAD67. From these data it can also be concluded that the molecular identity of the reported PLP-dependent and PLP-independent GADs by Tapia (Bayon, et al., J.Neurochem., 29:519, 1977) are GAD65 and GAD67, respectively. Martin and coworkers (Spink, et al., Brain Res., 421:235, 1987) have reported the existence of four kinetically different forms of rat brain GAD. However, immunoblotting experiments (with the antisera used here) of these forms have not been reported.

E. DISTRIBUTION OF GAD., and GAD IN RNAs IN BRAIN
TISSUE
Experiments were done to determine the distribution of GAD65 and GAD67 in RNAs in cerebellum using in situ hybridization.
Transcripts of, respectively, 3.2kb and 2.3kb from GAD65 and GAD67 cDNAs, were radiolabeled with 35S
according to Wuenschell, et al.
(Proc.Natl.Acad.Sci.,USA, 83:6193, 1986) procedure.
Hydrolyzed fragments of 200 bp were hybridized to coronal sections of a rat cerebellum. Animals were anesthetized under halothane and decapitated. The brain was rapidly frozen in dry ice and coronal frozen sections (12 m) were fixed for 30 min in freshly prepared 4% formaldehyde in phosphate-buffered saline Sap-25-00 05:16pm From-S&B/FBCo, +2328440 T-172 P.10/32 F-645 WO 95/07992 PCI`IUS941QF478 (p$S; 130A1M NdCI, 10mM Na phosphate, pH 7.0). The tissue was dahydrated through graded ethanol solutians and stored at -70 C.
zn order to increase tissue pemenbi].ity, the 8 sections were submitted tta the follawinq pretre$ttponts:
rehydrat,ion through graded ethanol solutions (5 min each in 95%, 85%, 70$, 50%, and 30% ethanol); PHS (5 tain); 0.02N Hcl (10 min); PHS (5 min); 0.01$ Triton N4 101* an PBS (1 min) ; pP6 (2 x 5 min) ; 1pg/ml prateirl$s$
R('I.5 min); and glycine (tp inhibit px-ateinase K) in PBS (3 x 5 min). Proteinase K was digested fcrr 3o min at 37 c bafare use. Sections=ware then incubated at 37 C in 50% formamide, 750mM NaCI, 25mM EpTA, 0.2% SDS, 0.02t BSA, 0,002t F,icoll, 0.02V polyvinylpyrrvlidone, 25Rpg/m1 yeast tRNA, 250~Cg/ml poly 1-, and 25m1d PPES (pH
1'or the hylbridi3ation, 100]nM DTT, 10% deXtran suZfate,.s hd sense or antisense 35S-RNA were added to the prehybridization solution. An aliquot (sbAl) of the hybridizatioh solution contairsing abotit 3 ng (104 cpm) of probB (sense or antieense) waa added onto the slides. Each slide was caverslipped and 'incubated for 16 hra at 50 c, i'ollowirig which the si.}.iconfxed coverslips were removact by brief washing i,n 4 x SSC (1 x SSC ~ 15omM NaCI, 60mM Na citrate, pH 7.0).
Sectiqns were thetl treated with ribonucle3se A
(50y:g/nl in 0.5M NaC1, 10mM Na thipSulfa'te, imM EDT'A, lAmM TrisHCL, pH 8.0) for 20 7nin at 37 C and rinse.d ior 2 hrs at room temperature,in 2 x SSC, 10mM Na thiosults te, for 30 3rlin at 550 C. Sectiong were dehydrated in e;thanol, delipldated in xylene, caated with Kodak HTS2 emulsion and exposad for 10 days at 4"c. The enlulsion waS developed with, Kodak D19* and the tissue counteY-stained with cresyl violet.
Autoradiographic grains were detected using reflected polarized light and grain numbers, densities, nd cell areas were determined with an Analytic imaging * Trade-mark { i . a.

Concepts image analyzer system. Due to the low background level, the criteria for defining a cell "labeled" was based on the presence of more than 5 clustered grains. The GAD labeled cells were found scattered throughout the brain, enabling the measurement of the number of grains over individual cells. The boundary of the cell and the area covered by a grain allowed the calculation of the number of grains per cell. This analysis was done at a high magnification (800X), using both reflected polarized light and transmitted light to simultaneously visualize the stained cell and the superimposed grains. Numbers are means S.E.M. of "n10 cells.

GRAINS/CELL
CELL TYPE GAD.,mRNA GAD~mRNA GADr~ s GAD~
Purkinje 172 t 34 (87)` 43 2 (70) 4.0 Golgi II 96 8 (80) 64 9 (65) 1.5 Basket 61 12 (102) 16 1 (57) 3.8 Stellate 55 15 (65) 18 3 (37) 3.1 ' t S.E.M.(n) In all neuronal types GAD67 mRNA levels are greater.
The observations with in-situ hybridization are consistent with previous findings (Nitsch, J.Neurochem., 34:822, 1980; Denner, et al., J.Neurochem., 44:957, 1985; Itoh, et a.Z., Neurochem.
Res. 6:1283, 1981) that the ratio of PLP dependent to independent GAD activities in the cerebellum is one of the lowest in brain regions tested. In addition, as shown in Table 3, the order of amounts for GAD67 mRNA is Purkinje > Golgi II > Basket > Stellate cells; in contrast, for GAD65 mRNA, this order is Golgi II >
Purkinje > Basket > Stellate cells.
The expression of GAD65 and GAD67 mRNAs thus differs among classes of neurons. The contribution of each to ~

total GAD activity in turn affects how GABA production is regulated. For example, the substantia nigra contains one of the highest ratios of PLP-dependent to = PLP-independent GAD activities (Nitsch, J. Neurochem., 34:822, 1980). Increasing GABA concentration in the substantia nigra by local injection of inhibitors of GABA catabolism is especially effective in reducing seizure susceptibility (Gale, Fed. Proc., 44:2414, 1985). Experimental animals undergoing seizures induced by PLP-antagonists may therefore be unable to inhibit seizure propagation because of inhibition of GAD65 particularly in nerve terminals within the substantia nigra.

F. SIIBCELLIILAR LOCATION OF GADO, AND GAD67 The distribution of GAD65 and GAD67 was evaluated in the S2 and synaptosome subcellular fractions. S2 is a high speed supernatant consisting of the cytosol of all cells in the brain, while the synaptosomal fraction consists primarily of nerve endings (Gray, et al., J.
Anat., Lond, 96:79, 1962). For these studies, whole rat brain fractionation was performed as described by Booth and Clark (Booth, et al., Biochem. J., 176:365, 1978). Protein concentrations were determined by Schaffner and Weissman (Schaffner, et al., Anal.
Biochem. .56:502, 1973). Samples were prepared as described (Kaiser, et al., DNA Cloning, Vol. I, A
Practical Approach, D.M. Glover ed. (IRL Press, Oxford, 1985, pp. 38-40), and immunoblotting was done as described above using GAD-6 monoclonal antibody and K-2 antiserum. Equal amounts of protein (16 g) were added to each lane. Autoradiography showed a linear response of increasing amount of 125I-protein A bound to antibody with protein concentrations of 1, 3, 10, 30, l00 gs with both K-2 antiserum and GAD-6 monoclonal antibody (data not shown).

WO 95/07992 2170'" 23 PCT/US94/09478 The results showed that GADO was present in equal amounts in both fractions. Since the S2 fraction contains the cytosolic proteins of glial (as well as other non-neuronal) and neuronal cells, the concentration of GAD67 must be greater in neuronal cell bodies than in nerve endings. In contrast, the concentration of GADO was greater in synaptosomes than in S2. These subcellular fractionation experiments suggest that, in contrast to GADO, a much greater fraction of GAD67 is present in cell bodies of neurons than in nerve terminals. Thus, subcellular fractionation, like immunohistochemistry, shows that GAD65 and GAD67 have different subcellular distributions.
in vivo experiments utilizing inhibitors of GABA
synthesis and degradation have suggested that the GABA
pool in neuronal cell bodies is different from that in the nerve terminals (Iadarola, et al., Mol. Cell.
Biochem., 39:305, 1981). GABA produced by GAD67 may be involved more in cellular metabolism (for example, in the GABA shunt) and in dendrodendritic synapses. The dendrites of granule cells in the olfactory bulb; which form dendrodendritic synapses with mitral dendrites (Shepard, Physiol. Rev., 52:864, 1972) and probably release GABA (McLennan, Brain Res., 29:177-184, 1971), label intensely with K-2 antiserum. While not shown here, it has also been found greater GAD67 than GAD6s mRNA levels (2-3 fold) in the olfactory bulb. This distribution is consistent with the reported finding that most GAD activity in the olfactory bulb is present in S2 and Pi (crude nuclear pellet) and not in synaptosomes (Quinn, et al., J. Neurochem., 35:583, 1980).
The differing subcellular distributions of GAD65 and GAD67 could result from cytoskeletal'anchoring or from some unknown protein targeting mechanism. Some cytoskeletal proteins have distributions that resemble GAD65 and GAD67. For example, in cultured sympathetic ~

neurons Peng, et al. (J Cell. Biol., 102:252, 1986), demonstrate that 84% of tau is in axons while 100% of MAP-2 is in cell bodies and dendrites. In addition, 43kd protein, a cytoskeletal protein, is thought to anchor the acetylcholine receptor to the underlying membrane cytoskeleton (Flucher, et al., Neuron, 3:163, 1989).

DETECTION OF GAD AUTOANTIBODIES IN CLINICAL SPECIMENS
A. MATERIALS AND METHODS
1. Patient Specimens. Sera from four groups of individuals were selected from a previous study by Atkinson and co-workers (Atkinson, et al., Lancet, 335:1357-1360, 1990). These groups consisted of:
Group (1), 1 new onset IDD patients diagnosed according to the established National Diabetes Data Group (NDDG) criteria (Gleichman, et al., Diabetes, 36:578-584, 1987) that had been referred to the University of Florida, Diabetes Clinics; Group (2), 5 randomly selected islet cell cytoplasmic antibody (ICA) negative non-diabetic controls without any known family history of autoimmune disease; Group (3), 13 individuals whose sera had been collected 3 to 66 months prior to their documented clinical onsets of IDD; Group (4), non-diabetic controls and relatives, and those who were studied prior to their onsets of IDD; and Group (5), 3 patients at risk for IDDM, but where onset has not yet occurred. This latter group had been ascertained through ongoing prospective ICA screening studies of more than 5000 first degree relative of IDD probands, and 8200 individuals from the general population (of which 4813 were school children).

-2. islet Cell Autoantibodies. ICA were assayed by indirect immunofluorescence on blood group 0 cryocut pancreatic (Atkinson, et al., Lancet, 335:1357-1360, 1990). All results were interpreted on coded samples, with control negative and positive sera in each batch.
The degrees of ICA positivity were analyzed with the guidelines established by the Immunology Diabetes Workshop (IDW) for the standardization of ICA
(Gleichman, et al., Diabetes, 36:578-584, 1987). All positive sera were titered by end point dilution, and the Juvenile Diabetes Foundation (JDF) units were determined by reference to a standard serum previously calibrated to the international JDF standard of 80 units. In the studies reported here, a positive ICA
result was defined by replicate titers of 10 JDF units or greater.
3. HLA DR Typing. HLA DR typing was performed as adapted from the method described by Van Rood and Van Leuwen (Nature, 262:795-797, 1976), using DR trays (One Lamda Laboratories, Los Angeles, CA).

4. Human Islet Cells. Human pancreatic islets were isolated from cadaveric pancreases and maintained in vitro as previously described (Ricordi, et al., Diabetes, 37:413-420, 1988). The islet cells were metabolically labeled with 35S methionine (Amersham, Arlington Heights, IL) in vitro (95% air/5%C02).
5. Islet Cell Extractions and Immunoprecipitations.
Islet cells were extracted as previously described by Atkinson, et al. (Lancet, 335:1357-1360, 1990) with the following modifications. For immunoprecipitation studies, the islet cell lysates were precleared twice by incubation (2h, 4 C) with either control, IDD serum (100 l), or GAD-6 (Chang, et al., J.Neuro, 8:2123-2130, 1988) (l l in 99 l of Tris buffer (Atkinson, et al., Lancet, 335:1357-1360, 1990) for every 1000 islets.
Immune complexes were then absorbed (lh 4 C) with an Sep-25-00 05:1Tpm From-SBB/Wo, +2328440 T-172 P.11/32 F-845 Wo 95/07092 P+CT/US94109479 exce3s of protein A Sepharose CI,-4H (P.harmacia, NJ).
Aliquot volumes repreaenting 1000 islet cells containing unbo"nd (precleared) lysate were then inaubated (i.xh, 40C) with either IDD or control sera (25M1) , or GALI-8 (Chang, et a,Z., J.Neuz'o, 8:2123-2I30, 1988) (3/tl in 25jA]. Tris bufZer). Following another incabation with protein A Sepbarose*CYr-4B (lh, 46c) , the oomplexes wer$ then washed 5 timep with 5bm2i Tria HCI, (pH 7.4) with 0.1% SAS, 1.04; Triton X-114, and 2m13 RDfiA, and then washed again one time in double distilled water. The protein A Sepharose CL-4B was then boiled in I,aem7ali sample buffer (Laemmli, Natura, 222: G8 0-685, 1970), and the samples were sub j ecteri to SD8-IyAGE and fluor4radiography (Kodak, X-omat l.RS) 25 using Erihance (NeW England Nuclear). Alternatively, the autoradi.ograPhs wexe analyzed by a BETAGEN (BoSton, MA) analyzer. Bo" 44KA posit.ive and negative sera wexc us e3 in each assay, to serve as interassay controls. All fluororadiagraphs wera analyzed and rated as positive or negative after comparison with the known intex'atsay controls. positivs serum samples were designated as 1 when a sample resulted in imamunoprecipitatiQn of a low intensity 64,000 M* barid, 2 if a moderate intensity band was observed and 3 if the intensity of the immunoprecipitated protein was high.
A similar rating procedure was employed for the intansity of bar-ds corrasponding to 9Wuunoprecfpxtated "STGAD,s and 'sS-GAUd,.
6. 2mmunonreeipiti}tions. Satmunopreoipitation of bacterial iysates containing 31S-CA176j or 3sS-CiAD67, and GAD Froso human brain homogeriate, was comploted as described above in immunopreoipitafiion studies of human islet ce7,1 e:;tractions_ 7. G Assava. Human brain homogenates were incubated with patient sera as described above'in human islet ce1].s. After absorption and washes, the protea.n A
ag4rose slurry was aliquoted into three equal voZumes * Trade-mark ~
-and GAD activity was measured as described (Krieger, et al., Neurochem. 33:299, 1984). Briefly, Protein A
agarose beads were incubated with (1-14C)-glutamate (Amersham) in a designated incubation mixture (Krieger, et al., J. Neurochem. 33:299, 1984) and production of laCOZ was quantitated by a liquid scintillation counter.
8. Production of 3$S-GAD,, and 35S-GAD4,. Rat GAD65 and GAD67 cDNAs were subcloned into a bacterial expression system as previously described. Labeling of 35S-GADs was completed by pulsing IPTG induced bacterium (growing in Minimal Media) for 15 minutes with TRAN 35S-label (ICN). Cultures were then spun down and resuspended and sonicated in iml of homogenizing buffer (1mM phenylmethylsulfonyl fluoride (PMSF), 1mM 2-aminoethylisothiouronium Bromide (AET) and 60mM
potassium phosphate, pH 7.1). After sonication, cell debris was removed by centrifugation and protein concentration was measured (Bradford, Anal.Biochem., 72:248, 1986) in the supernatant (supernatant was stored in aliquots at -70 C).

B. IMMUNOREACTIVITY OF IDDM SPECIMENS
Sera from patients with IDDM were tested for the ability to precipitate GAD from human brain homogenates.

~-' ~ WO 95/07992 PCT/US94/09478 ~
N ON O O -0 O tf1 Ifl I11 O
co .0 co co co Nr, l, tll a ri N N m N N
Ai 1+f e I
rq m J..1 -'i Q~ voooo00000 "4 t0 N co H

w m ~
m a .~

Mf Q
wfI M M M N N O C C O O
y '-i -.i Il~ La"
t." C'.
J..1 .rl .-i 00 0 a.-iinO~i ro rtf ro ~j ro@ ~
qp.f N n n G." C7 L: ~ A C7 N~ bf ro (tf W
41 m'~3 -4 4.) m w ~
=r1 A U UI
~ N -.i fd m W ~ W ~ ro + + o ro cd ro U m N''~
N 1) N W
GL
m 4.) -4=~x ro m -.i m m y~ m u.m ro.,+ o H
z a z~ z a b WO 95/07992 2170523 PCT/US9~3/09-175 As shown in Table 5, the sera of four (out of five) at risk for IDDM or IDDM patients bound significantly greater amounts of enzymatically active GAD of human brain extracts than sera from control patients. In addition, sera from one of the patients was drawn in a pre-IDDM period, thus autoantibodies to GAD are present prior to the onset of IDDM symptoms (see C below).
Further experiments (results not presented) showed that the sera of two at risk IDDM patients (DA, DC) immunoprecipitated recombinantly produced 35S-GAD6s whereas recombinantly produced 35S-GAD67 was only recognized by sera of patient DA (and to a lesser degree than 35S-GAD65) .
Additional studies using patient DA sera showed the presence of antibodies which recognize specific polypeptides produced in human pancreatic islet cells.
Electrophoretic analysis of the bound polypeptides demonstrated the presence of autoantibodies to a 64kD
component, as previously shown by others in human IDDM
(Baekkeskov, et al., Nature, 298:167-169, 1982) and in animal models (Baekkeskov, et al., Science, 224:1348-1350, 1984; Atkinson, et al., Diabetes, 37:1587-1590, 1988). Prior absorption of these sera with GAD-6 monoclonal, which recognized GAD65 but not GAD67, or with bacterially produced GAD65, abolished the ability of the sera to recognize the 64kD pancreatic polypeptide. The epitopes recognized by autoantibodies to the 64kD
autoantigen are thus present in GADO, indicating that the 64kD autoantigen is indeed GAD65. In order to investigate the predictive value of GAD65, sera drawn from patients prior to onset of clinical manifestation of IDDM were tested for autoantibodies to GAD65.

~ -47-~~ .-i O O O N . - - 1 .--1 O O ~ - i O.-1 o.-i H
- A .
N ri N r-I f7 r-1 tn N e-1 O('7 ci m QI O O O O O O O O O O O Q ro N to a a i ~10 m m 01 m O m Or m m ai -,i m m ~ ~ O q 11 a e} 1D 1D tn lA N C1D O% e! O
[+1 t0 NIV ei rl AA H
Ab~y ~
4..1 r- co `n ~N~~~~~(n`~
~ .
ld M c+~
m ~Fy m .ri f6 N t!~ -I a! t0 ri A c=1 c~1 ,~e'i 10 m N
~ 3 m e/ N N`~T. V 10 t~1 f'f f7 ~ V 4.) H A .~

m b H
mi O 4) N H
Q m G

ri C'+

H
AbY1 r N 11 .., .- . , .

WO 95/07992 - 48 217 0 5 2 3 pCT/US94/09-178 -As shown in Table 6, 9 out of 12 specimens (75%) were immunoreactive with 35S-GAD,5. In addition, two patients (JA and VC) were immunoreactive to GAD67, but not GADO under these conditions. Therefore, in combination, autoantibodies to GADO and GAD67 were present in 11 out of 12 (91%) of these patients sera.
This finding suggests that although autoantibodies to GADO are more common than autoantibodies to GAD67, the use of both recombinant GADs (GAD65 and GAD67) in an assay would allow for greater predictability of IDDM.
Previous tests of these sera (Atkinson, et al., Langet, 335:1357-1360, 1990) demonstrated that 11 out of 12, or 92%, immunoreacted with the 35S-64kD molecule from human pancreatic islet cells. The serum which contained detectable autoantibodies to the 64kD molecule and not GAD65 was a serum which contained the lowest titer (or 111") for the 64kD molecule. Thus, the false negative obtained was due to a lack of sensitivity in this assay. Furthermore, this assay predicted IDDM in one patient (BR) who was negative for 64K.
These results show that the 64kD molecule identified in fl-cells of human pancreas is identical in size and antigenicity to rat GADO. Furthermore, sera drawn from patients prior to IDDM onset contain autoantibodies to GADO. Consequently, the GADO
recombinant molecule is of great utility as a diagnostic tool for predicting IDDM. The ability of a physician to diagnose IDDM prior to actual symptoms may .result in a greater extension of time before insulin therapy is needed. The sensitivity of such immunoassays will improve with the use of a recombinant GADO of human origin which represents the GAD form present in f3-cells of the pancreas.

WO 95/07992 PCT/US94/09-t78 IMMUNE PROLIFERATIVE RESPONSE TO POLYPEPTIDE
Polypeptides were synthesized using an automatic instrument (Applied Biosystems) and standard conditions. These polypeptides were then tested to compare their relative ability to stimulate proliferation of splenic lymphocytes and islet infiltrating T lymphocytes (IITLs). In this study, polypeptides derived from the GADO core sequence and from the homologous region of polio virus were compared. Appropriate cells were cultured for 5 days with the respective polypeptide in the presence of 5 X
104 irradiated spleen cells. 3H-thymidine was added during the last 16 hours of culture.

E
a ~
O S Ln g tV r) d O N N
a OW
V
Z_p LU =
Z
p~i =
cr) O
~
x } Q
Y

p J
~
V LU
Q
a Z Q- LL
aW 2 Y d Y LL cm 3:
~ Q c`v 4) Z L c o > o X
= ~ a~
o p Q z ~ V A

in these studies, there was no significant difference in the proliferative activity of cultures of spleen lymphocytes exposed to either the polio or the GAD65 polypeptides. However, both polypeptides stimulated a T cell response which was higher than that found in the media control. The lack of difference in proliferation in the spleen cell population may be due to a lower frequency of GAD polypeptide specific T
cells.
14 The IITL population, when evaluated in the same manner, showed a marked difference in cell proliferation. In this system, the response to the GAD65 polypeptide was 9-fold greater than that of either the culture media or the polio polypeptide. This data strongly suggests that the GADw is an important antigen for T cell responses in the IITL population. This data suggests that molecular mimicry plays a role in the pathogenesis of diabetes.

ERAMPLE S
GAD INDUCES PROLIFERATION OP SPLEEN CELLS OF NOD MICE
Proliferative T-cell responses to /3-cel3 antigens (#CA) develop spontaneously in the nonobese diabetic (NOD) mouse model in a defined chronological order.
The NOD mouse experimental model is considered the most analogous in vivo system available for studying IDDM in humans. This example describes studies on the antigen-induced blastogenesis of spleen cells from newborn to 5 month old fe.male NOD mice when exposed to GAD and other peptides.
The SCAs tested included one of the two forms of GAD (Kaufman, et a1., Science, 232:1138-1140, 1986;
Erlander, et al., Neuron, 7:91-100, 1991; Kaufman, et al., Trends in Pharm. Sci. 14(4) Trends Pharmacol Sci.
107-109 (Apr. 1993) ), (GlZ1D65r previously known as the 64K
autoantigen (Baekkeskov, et al., Nature, 298:167-169, 1981; Baekkeskov, et al., Nature, 347:151-156, 1990), carboxypeptidase H (CPH) .~ ,.

(Castano, et al., J. Clin. Endoctrinol. Metab., 78:1197-1201, 1991), insulin (Palmer, Predicting IDDM, Diabetes Reviews, 1:104-115, 1993) and a peptide of hsp which has been shown to be the immunodominant determinant recognized by NOD T-cells (Elias, Proc.
Natl. Acad. Sci., 88:3088-3091, 1991). GAD in particular, is a good candidate for the initial target antigen in IDDM since autoantibodies to GAD arise early in the natural history of the disease (Baekkeskov, supra; Atkinson, et al., Lancet, 335:1357-1360, 1990;
Kaufman, et al., J. Clin. Invest., 89:283-292,1992).
Furthermore, unlike the ubiquitous hsp, GAD is expressed primarily in P-cells and the immunologically privileged central nervous system (CNS) and gonads. As control antigens, irrelevant prototype foreign and self antigens including hen eggwhite lysozyme (HEL), human serum albumin (HSA), E. coli. P-galactosidase (,8-gal) and murine myelin basic protein (MBP) were used.
NOD (Taconic farms) and BALB/c mice (Jackson Laboratories) were kept under specific pathogen free conditions. The mice were sacrificed at the ages indicated and the spleen cells were tested directly ex vivo for their proliferative recall response to antigen. Single cell suspensions of spleen cells were plated at 1 x 106 cells per well in 96 well microtiter plates in 200 l serum free HL-1 medium (Ventrex) that was supplemented with 2mM glutamine with or without 10 g/ml antigen (or 7 M peptide) in triplicate cultures.
During the last 16h of the 72h culture period, 1 Ci[3H]-thymidine was added per well. Incorporation of label was measured by liquid scintillation counting.
Both human GAD65 (Bu, et al., Proc. Nat1. Acad.
Sci., 89:2115-2119, 1992) and E. coli 0-gal (control) were purified from recombinant bacteria on the basis of a hexahistidine tag which allows their rapid affinity purification by metal affinity chromatography (Hochuli, et al., Bio/Technology, 6:1321-1325, 1988). Bovine CPH

was the generous gift of L. Fricker (Albert Einstein Col. Med.) and human insulin was purchased from Eli Lilly.
As illustrated in Figure 8, while proliferative T-cell. responses were not detected at any time point to the control antigens, a response to GAD arose at 4 weeks of age in NOD mice, concurrent with the onset of insulitis in the colony. The blastogenesis induced by GAD increased during the next four weeks and then declined to background levels by week 16. At 6 weeks of age, near the peak of anti-GAD reactivity, T-cell responses to hsp appeared and increased until week 15 and then diminished as well (Figure 8). In all NOD
mice tested, hsp reactivity was preceded by an anti-GAD
response, suggesting that the former reactivity developed as a secondary event during the autoimmune process. Similarly, while no response was detected to CPH at 4 weeks of age, a strong anti-CPH response was observed by week 8. In some mice, a weak response to insulin was observed at 12 weeks, which became more prevalent at 15 weeks of age (Figure 8 and Table 8).
None of the antigens induced proliferation in T-cells from age-matched control BALB/c or (NOD x BALB/c) F1 mice, both of which do not develop insulitis or IDDM.
T-cell reactivity subsequently arises to other flCAs, consistent with the inter-molecular diversification of the autoimmune response. Thus, the autoimmune response to GAD was the first to occur among the autoantigens tested. In view of this, tolerization to GAD should prevent the spread of autoimmunity to other flCAs and insulitis. If this were not the case then tolerization to GAD should have no effect on the response to these other antigens.
Blastogenesis provides an apprQximation of the relative clonal sizes of antigen-specific CD4+ T-cells (Corradin, et al., J. Immunol., 119:1048-1053, 1977).
The data in Figure 8 shows that GAD reactive T-cells -"spontaneously" undergo clonal expansion concurrent with the onset of insulitis. These findings are consistent with an endogenous priming event.

INDUCTION OF TOLERANCE WITH GAD
This example describes a study which shows that induced tolerance to GAD can ameliorate IDDM.
1. In these experiments female NOD mice were intravenously injected at 3 weeks of age with 50 g GAD, fl-galactosidase, mycobacterial hsp65 (m-hsp) or 0.1 g of the immunodominant hsp peptide (hsp-p), in PBS. At 12 weeks of age, mice were examined for insulitis and autoantigen reactive T-cells. At this age both indications are established in untreated NOD mice.
Pancreatic tissue sections were stained by immunoperoxidase techniques for insulin and were counterstained with hematoxylin. Insulitis was scored in a blinded manner by examining 54 to 87 islets on 5 interrupted tissue sections from each pancreas.
Proliferative splenic T-cell responses induced by various antigens were performed as described above in Example 4. Data in Table 8 are expressed as the average [3H] -thymidine label (cpm) incorporated in triplicate cultures.

~ WO 95/07992 -55- 2170523 PCTIUS94/09478 m ~ N N
{{
iLE O1J O11 T T
~J r~ O
z ~ a 1 ~rl LO
o ,l, H N
N cn II
L1) ~- O 0 N
O O ci ii ia ii fl LO

N u o ~
-H

N N ~ N a W C
O 0 O O CD L, ii ii H
Le {{ '0 C m +I E
N t u CO cn ~ a O O O O E O T
~
y~ +i +1 1 i ~ ItI A
~, II~, ~
O ~ C) 0) 2 0 0 ~= ~
~r m 41 -H r- $ o a O O C p~0 .0 .
M n t O =~
cr) N
O 0 cq -H +I
~ ~p m N {~ p V
+f a Cc =m ch uO , =g N ~ O O ~ ~ o 0 0 o c o 76 +I -H +f +1 +I
o at d r ~ ~ C
t1) tA O !l) t1) Z La ~
ZI N m eh tC! C n O
O O O O O O~ ydy ~
-H -H m ii +l E C
-st m f* 00 w m N (N
O ~ =- ~^ W =
m U.) e' a C m `1 m ., O tA ch cn Z
m C
+r a y ~ QZ V r = E
= J V

o 0 0 0 z z z LO
tV

p ~
CL co ~ v a- ~' =
~

r m .2 ~aui +i o o cn -H
a - cc m E
a ,~ o W Cl? U?
.}{ 0 0 ~ v a ._ C cn C
- "' Q) m ? ~ vf LO
~ V p W N IS
~ 0 p~ w * -H N m N L O

Q C '~p N G
~ LO
ii cs 10 C
~ -H

~ o~
cc ~.. ~õ

~ +i ii iH

_ rp cc ~- C
~'n N
zI
c~ c A: c.4t m ~
Lq +i LO E
N n *-C : 0 0 ^ ~ ~
~ V m O ;-- ~
m LO
m U _ 0 m cn cr) cn z a m Q O =..
m tA
~
n aLL au-~ ~ ( a+ LL + s a u 2, 7 O527 PCT/US94/09478 ~

d o 0 0 c) z z z LO
U~
N
II
' Q) 00 c') N
~ . .
M p LO

L hi .
W cc am .= p E 41 +1 +1 o _ o ~
CrA a~c cd ~rii C? C 0 orn r_ q m o c, p ~ . ao 0 p m ^
t p ID
~ N V
tA C tt~ C
N) t'r) ~ m +I +I ~ a ti C

N M = -e o 0 0 0 a 9 -H -H +1 Sm et N p ~
m =
O
m LO
d r- N
C
C C ~
0 ~ a E m c ~ .
n N
G> CD
m ~ mII O
#' tn r) cnC7 z m a t17~
a C + LL
W r' _ V) a = a Seventy five percent of the GAD treated mice, but none of the controls, showed no T-cell reactivity to GAD
(indicating complete tolerization) or to other flCAs.
These mice were also completely free of insulitis (score 0.0). If there were another effector T cell population in the islets, specific for an unknown #CA, that preceded the anti-GAD response, the release of cytokines by this population should have promoted T-cell responses to (3CAs and insulitis (Sarvetnick, et al., Nature, 346:844, 1990; Heath, et al., Nature, 359:547, 1992). Twenty five percent of the GAD-treated mice were not completely tolerized to GAD, as evidenced by a weak residual GAD reactivity (SI of about 3) and displayed very limited peri-insulitis. In contrast, while tolerization to both of the hsp antigens was complete, these treatments reduced, but did not prevent, the development of T cell responses to other flCAs or insulitis. Thus, while the inactivation of GAD-reactive T cells prevented fl cell autoimmunity, hsp tolerization only partially reduced it, as would be expected if a secondary element was removed from the amplifactory cascade.
In ongoing experiments examining the effects of GAD
tolerization on diabetes incidence, all of the GAD
treated mice (n=17, presently 37 weeks old) have normal glucose levels, while 70% of the mice receiving control antigens developed hyperglycemia by 19 weeks of age (n=20). Five GAD treated mice were sacrificed at 30 weeks of age. All were free of detectable OCA reactive T cells. Of these five animals, four mice were completely free of insulitis and one mouse displayed very limited peri-insulitis. These data show that inactivation of GAD reactive T-cells prevents the long term development of insulitis and diabetes.
2. In a second set of experiments neonatal female NOD mice were injected intraperitoneally IFA with peptide 11, a mixture of peptides 34 and 35 plus IFA, -or with IFA alone and at 12 weeks of age the mice were examined for insulitis and autoantigen reactive T-cells as in Example 6.1. Proliferative splenic T-cell responses induced by the various antigens were performed as in Example 4, and data in Table 8B are expressed as the average [3H]-thymidine label (cpm) incorporated in triplicate cultures.
The data in Table 8B show that tolerization with control peptide 11 did not prevent auto antibody response to GAD or to GAD peptides 17, 34 or 35. Nor was response to hsp peptide prevented by tolerization with peptide 11. IFA alone was somewhat effective at suppressing immune response. By contrast, tolerization with the mixture of peptides 34 and 35 suppressed the 15, autoimmune response of spleen cell proliferation to all species tested: fl-gal, GAD, GAD peptides 17, 34 and 35, and hsp peptide. In addition, tolerization to GAD
peptides 34 and 35 greatly reduces insulitis but does not completely prevent it as whole GAD65 does.
3. In a third set of experiments female NOD mice were intravenously injected at three weeks of age with peptide 11 (control), a mixture of peptides 34 and 35 plus IFA, or with HEL peptides and at 12 weeks of age the mice were examined for insulitis and autoantigen reactive T-cells as in Example 6.1. Proliferative splenic T-cell responses induced by the various antigens were performed as in Example 4, and data in Table 8C are expressed as the average [3H]-thymidine label (cpm) incorporated in triplicate cultures.
The data in Table 8C show that tolerization with control peptide 11 did not prevent auto antibody response to GAD, to GAD peptides 17, 34 or 35 or to hsp although the response was not as great as in Example 6.2. Nor was response to hsp peptide prevented by immunization with peptide 11. By contrast, immunization with the mixture of peptides 34 and 35 plus IFA suppressed the autoimmune response of spleen cell proliferation to all species tested: #-gal, GAD, GAD peptides 17, 34 and 35, and hsp peptide.
E%AMPLE 7 CHARACTERIZATION OF GAD-REACTIVE T-CELLB
This example describes studies on GAD-Reactive T-Cells for additional properties that distinguish activated/memory from resting/naive lymphocytes.
In one series of experiments 7 interferon (IFNy) was measured by ELISA in culture supernatants (CSN) of spleen cells of 6-9 week old mice after challenge with GAD or control antigens HEL and MBP. Additionally, the frequency of antigen specific, IFNy-producing cells was determined by an ELISA spot technique (T. Taguchi, et al., J. Immunol., 145:68-77, 1990). Frequency of antigen-induced, spot forming cells (SFC) among 103 spleen cells is represented in Figure 9(a). Values are the mean + SEM from 5 individual female NOD mice, each tested in triplicate cultures with or without antigen.
Results from a single experiment are shown. These are representative of 3 separate experiments.
In performing these experiments, freshly isolated spleen cells were cultured with or without antigen as described in Example 4. CSN were taken after 48 h and the concentration of IFNy was determined by ELISA
(Macy, et al., FASEB J., 3003-3009, 1988). IFNy specific monoclonal antibody (mAb) R4-6A2 (Pharmingen) was used as the capturing reagent and biotinylated mAb XMG 1.2 (Pharmingen, also specific for IFN7) was used in conjunction with streptavidin-alkaline phosphatase (Zymed) and p-nitrophenol for detection of bound lymphokine. Recombinant murine IFNry (Pharmingen) was used as a standard. ELISA spot assays for the detection of antigen-specific, IFNy-producing cells were performed as described (Taguchi, et al., J.
Inununol., 145:68-77, 1990). After a 24h pre-activation culture of spleen cells with our without antigen, cells were transferred by serial dilution to 96 well ~

microtiter plates (Millipore) that had been pre-coated with mAb R4-6A2. After 24h, the cells were removed and IFNT spots were visualized using XMG 1.2-biotin in conjunction with nitroblue terazolium-bromochloroindolyl phosphate substrate (Sigma). Spots were counted visually and the frequency of antigen specific cells was determined from the difference between the number of spots seen with and without antigen.
As shown in Figure 9(a), when freshly isolated T-cells from 6-9 week old NOD mice were challenged with GAD or control antigens, high concentrations of IFN7 were detected only in cultures containing GAD, suggesting that the GAD specific T-cells had been pre-activated in vivo, since only pre-activated T-cells (Thi) produce IFNy within 48 hours after antigen recognition (Ehlers, et al., J. Exp. Med., 17,3:25-36 1991; Croft, et al., J. Exp. Med., 176:1431-1437, 1992). In contrast, T-cells from age matched BALB/c mice did not respond to GAD or to control antigens by IFNy production (data not shown).
Results of the ELISA spot assay to measure directly the frequency of GAD-specific T-cells showed that while in 6-9 week old NOD mice, T-cells reactive to control antigens constituted approximately 1 in 105 cells in the spleen, the frequency of GAD-reactive T-cells was about two orders of magnitude higher, ranging from 90-291 cells per 105 cells (Figure 9(a), confirming the data ,obtained by proliferation assays (Figure 8) that these cells had been clonally expanded in vivo.
In another series of experiments, GAD specific T-cells were characterized for expression of the cell surface marker L-selectin, since murine T-cells convert from an L-selectin+ (L-sel+) to an L-selectin (L-sel-) phenotype upon activation (Bradley, et al., J.
Immunol., 148:324-331, 1992).

-To perform these studies, pooled spleen cells from 3 to 4 age matched mice were panned on plates coated with goat-anti-mouse Ig (Zymed) to remove adherent macrophages as well as B cells. Next, CD8+ cells were coated with mAb 58.6-72 (ATCC) and removed by panning over plates coated with goat-anti-rat Ig (Zymed). The non-adherent CD4+ cell fraction was labeled with anti-L-selectin mAb MEL-14 (ATCC) and panned on goat-anti-rat Ig coated plates. Both the adherent (CD4+ L-sel+) and non-adherent (CD4+,L-sel') fractions were sampled.
Purity of the cell fraction was assessed by FACS
analysis; cells were )90% CD4+ and )95% enriched for the L-sel- or L-sel+ phenotype. The purified cell fractions were tested for GAD reactivity by seeding them at 2 x 105 cells per well in 96 well microtiter plates with or without antigen. Irradiated (3000 rad), unseparated spleen cells of 3 week old NOD mice were added at 5 x 105 cells per well as a source of antigen presenting cells. Supernatants of triplicate cultures were taken 48h later and their IFNy content was determined by ELISA.
The results of this study showed that by 2-3 weeks of age, GAD reactive T-cells could not be detected in either the L-sel+ or the L-sel- population, consistent with a low frequency of antigen reactive precursors at this time point. However, by 6 weeks of age high levels of IFNT were induced by GAD (but not by control antigens) in the L-sel- (but not the L-sel+) subpopulation of CD4+ cells (Figure 9(b)).
The increase in clonal size of GAD reactive T-cells, their production of IFN7 and their L-sel"
phenotype provide three independent lines of evidence that a potentially pathogenic (Ando, et al. Cell Immunol., 124:132-143, 1989) Thl type T-cell response is spontaneously primed to GAD in vivo early in NOD
development.

,, ..
~ WO 95/07992 63 - 2170523 PCT/US94/09478 -CHARACTERIZATION OF GAD SPECIFIC T-CELL
DETERMINANT RECOGNITION
The fine specificity of the anti-GAD T-cell response was mapped using a set of 38 peptides (numbered successively from the N-terminus) that were 20-23 amino acids (aa) long and span the entire GAD6s (Bu, et al., Proc. NatZ. Acad. Sci., 89:2115-2119, 1992) sequence with 5aa overlaps (Figure 10).
Spleen cells were tested from 4 (Figure 10a), 5 (Figure 10b) and 7 (Figure lOc) week old NOD mice for proliferative responses (as described in Example 4) to the GAD peptides. Peptides were present in cultures at 7 M and the label was added during the last 16 hours of a 5-day culture. The peptides were synthesized using standard Fmoc chemistry and purified by reverse phase HPLC (Advanced Chemtech). The sequence of stimulatory peptides are shown below in Table 9.

Pentide Number GAD Region Amino Acid sequence KPCSCSKVDVNYAFLHATDL

34 509-528 _IPPSLRYLEDNEERMSRLSK

Murine and human GAD65 are 95% identical at the 35 amino acid level (555/585) and are 98% conserved, with most of the differences localized near their N-termini.
The underlined amino acid in the stimulatory peptide sequences above are conservatively substituted in murine GAD6S. In separate experiments, the murine form of key peptides (#17 and #34) were tested and produced similar results.
As shown in Figure 10, peptides that triggered stimulation indices )3 are indicated as black bars.
These peptides did not induce proliferation in T-cells from NOD mice (3 or )16 weeks in age, or from control (BALB/c x NOD)F1 mice (data not shown). The data are represented as the mean SI standard error calculated from 3-6 individual mice tested twice in each age group. Characteristic results for peptide induced blastogenesis in individual mice are shown in Table 6.
The first detectable response, at 4 weeks of age, was confined to the carboxy-terminal region of GAD, and involved two adjacent peptides (aa 509-528 and 524-543, peptides #34 and #35, respectively, Figure 10a). At 5 weeks of age, responses to an additional determinant (aa 247-266, peptide #17, which contains a region of sequence similarity with Coxsackievirus (Kaufman, et al., J. Clin. Invest., 89:283-292, 1992) (Figure lOb) were regularly recorded. During the next two weeks, responses to peptide #17 (aa247-266) increased and T-cell autoimmunity spread to two additional peptides at the carboxy terminus (aa 479-498 and 539-558; peptides #32 and #36 respectively, Figure 10c). Subsequently, reactivity to the GAD peptides declined (data not shown), paralleling the loss of response to the whole protein (Figure 8). It is unclear why the initial T-cell response to flCAs fades in NOD mice. Possible explanations include: a) immune regulatory mechanisms;
b) exhaustion of the response due to the continuous stimulation by the endogenous antigen; and c) induction of anergy in specific T-cells owing to their recognition of the autoantigen on "non professional"
antigen presenting cells such as the fl cells themselves (Markmann, et al., Nature, 336:476-479, 1988).

WO 95/07992 217" `' 23 PCTIUS94/09478 ~

The gradual diversification of the primed autoreactive T-cell repertoire that was observed in this naturally occurring autoimmune disease parallels the shifts in T-cell recognition recently observed in experimentally induced autoimmunity to the CNS where autoreactivity spreads both intra- and intermolecularly among CNS proteins (Lehmannn, et al., Nature, 358:155-157, 1992; Perry, et al., J. NeuroimmunoZ., 33:7-15, 1991; Watanabe, et al.,Nature, 305:150-153, 1983;
Liebert, et al., J. Neuroimmuno1.,17:103-118, 1988).
Apparently, lymphokine secretion by the first wave of autoantigen specific T-cells in the target organ results in up-regulation of antigen presentation and creates a microenvironment that favors priming of additional autoreactive T-cells (Lehmann, et al., Immunol. Today, 14:203-208, 1993; Sarvetnick, et al., Nature, 346:844-847, 1990; Heath, et al., Nature, 359:547-549, 1992). Since hsp reactive CD4+ T-cells are capable of inducing IDDM (Elias, et al., Proc.
Natl. Acad. Sci., $7:1576-1580, 1990; Elias, et al., Proc. Nati. Acad. Sci., 88:3088-3091, 1991), their recruitment into the activated T-cell pool, along with T-cells reactive to other QCAs, probably reflects an amplificatory cascade that eventually leads to fl cell destruction.
In summary, the data above establish GAD as a critical target antigen in the pathogenesis of IDDM in NOD mice. The results show that T-cell responses to flCAs diversify both intramolecularly and intermolecularly as the disease progresses, consistent with a dynamic autoimmune repertoire (Lehmann, et al., ImmunoZ. Today, 14:203-208, 1993). However, interference with the early autoreactive T-cell population can prevent the recruitment of additional autoantigens into the primed repertoire thereby halting a cascade of autoimmune responses that eventually leads to fl cell destruction. As a similar autoimmune . .,~ .

~
- -progression is also likely to occur during the development of human IDDM (Palmer, J.P., Predicting IDDM, Diabetes Reviews, 1:14-115, 1993; Atkinson, et al., Lancet, 339:458-459, 1992), these findings suggest that peptide-based immunotherapeutic agents would be useful in predicting and ameliorating human IDDM.

AIITOANTIBODY REACTIVITY WITH GAD FRAGMENTS
This example describes a study which examined the variability in recognition of epitopes in human GAD6s polypeptides by IDDM autoantibodies in sera of human patients.
Portions of human GAD63 cDNA were amplified by the polymerase chain reaction (PCR; Saiki, et al., Science, 15. 239:487, 1988) to produce DNA segments encoding three polypeptide segments: amino acid residues 1-224 (segment A); 224-398 (segment B); and 398-585 (segment C). Each construct also contained a T7 promoter, a consensus sequence for the initiation of translation and an initiating methionine codon (Korak, M., J. Cell Biol., 108:229, 1989). Each PCR product was then trascribed in vitro with T7 RNA polymerase and translated in vitro in a rabbit reticulocyte cell-free system in the presence of 35S-methionine, using conditions recommended by the supplier (Amersham Corp., Arlington Heights, IL). Each test serum (30 l) was incubated with the resulting 35S labeled-polypeptides.
The bound peptides were isolated with PAS and analyzed by SDS-PAGE in 12% polyaacrylamide and autoradiography.

~
- -IDDM PATIENT SERA REACTIVITY WITH GAD SEGMENTS
SEGMENT
PATIENT A B C
Control (N=7) - - -052 - + +

705 - + +
UC2 - + +
N.L. - - -L.I. - - -T.L. - - -P.T. - + -J.D. - - -B.Y. - + +
M.C. - - -R.S. - - -K.O. - - -T.B. - - -S.M. - - -A.W. - + -J.B. - + +
J.A. - - -P.C. - + +
L.R. - - -J.M. - + -G.A. - - -Sep-25-00 05;17pm From-SaB/F&Co, +2328440 T-172 P.12/32 F-845 Wo 95/07992 pCT/U59410947$

As shown in Table 10, none of the gTsecimans had detectable levele of antibodi.Bs to the amino terminal third (segment A) tif GAD whereas 9 patients (41t) had antibodies reactive with the lal2ldle third (seqment 8) and 6 patiebts (27$) had an'tibodies to the carboacyl,-tokiuinal third (segment C) of GAD.
E1CA,Kp;jB 1 D
V)%EDICT20N O~ 4jaCYPIENT IDDM DY C~_"
EPYTDP$ RECGG27IT1oN pAMRR21 The ].ncroa-sing likelihood of an IDDM interventive therapy and the (recently ac]cnowledggd) benefits of 3nanaged glucbse homeostasxs in preventing IDDM
asSociated Ccmplications makes the early detecti4n of ~
cell autoimmunity before clinical IDC7M onset and in 75 NIDDIK patiolnts (10$ of whom eventually convert to IDDM) a crucial goal. Autoariitbodies to CAD may pxpVide the earliest and most reliable 3tia7Ckor of impending IDDM
among the molecularly defibed IDI]M associated aLttoantigens. To determine whether GAD peptides will bind to IDbM associated autoantibQdies the followiDg study was conducted.
A bct of pept3des (20-23 s,mino acids in length, with 5 aa overlaps) that span the human GAD65 molecttle were mxnthesi2@d to determine whether sera from most indiViduals at risk, prc-IDAM and with SpDM (in contrast to healthy controls) do in fact produce antibodies that difterentially recognize GAD65 linear epitopes distributed throughout the moleCule.
Patient sera and most control sera were those used in a previous study (ICaufiman, et al., J, Clin.
Investa.gation, supra) All samples wera codad and tested in a blind manner. Peptidea were synthesized using an automatic instrument (Applied 8iosystems, Pqster City, CA) and standard conditiQns. Peptides were dissolved in 60 IIIM &odium bicarbonate buffer (pti 9.6) at 20 ug/ml and lon ul of each was added to duplicated wea,ls of a 96 well Nund-I3R]puno Plate. Pepticles were Allowed to ' Trade-mark S p-25-00 05;lTpm From-S8B/F&Ca, +2328440 T-172 P.13/32 F-845 r. ~ 95107992 3rCTl7J594/D9479 pi.nd at +{=C overnight. The platea were then washed three timea wi.th PBS + 0.14 Tween 20 (wash buffer), after which the plates were pre-absorlaed With 3.1 )BgA in sodium bicarbonate buffer for 0.5 hours at 37 C, or at Xoom temperature overnight. The plataa were then washed 5 tizas with the h.bove wash blsffer. 100 ul Qf earum at a 1/300 dilution in PBS + 0.1% Tween 20 azid 1t BSA waS added to each well and antibadies were allowed to bind for 1 hour at 37PC. The pl8tea ware washed 5 times With wash ]auffer. 100 Ul of a1/600 dilution of HRp-goat ant~-human TgG (sRi., Caithersberg, MD) was added to each well and allowQd tb bind for i hour at 37 G. The plates were then wa;shed 7 times and lp0 ul of substrate buffer was added to a,Gh well for 30 minutes at room temperature. The color development was measured at 410 nns using an r4L=SA plate reader (2CN, Diomedi,cal6, Gosta Masa, CA). Positive sera were defined as: 0D,ia of the sa2tlple/negative cantrol 2,3Ø

The data shown in Table ii establish that a number of GAD psptides were recognized by patients previously shown to be 64K positive, but riot by control sera.
Each patient showed a differant pattern of GAD epitope recogniti.ori. P2ptides 20, 21 and 25, Were each recogn,ized by 6/8 patients, and nonc of the controlss-with the exception of popti,de 25 wriich was reCagnizad by 1 out of 13 controls. Based on immunoreactivtty to 2 of these peptideB (t2o and 21) 7/8 (9E*) of the patientic (and none of the controls) could be identified as possessing GAD autoantipodies. peptidas 3, 6, 22, 25 and 37 were each recHgnized by only 2fi-37% of the patients (and none of the co2ttrol sera); but taken tvqether. 75t of the patients recognized at least one of these. Peptides 5, 9 and 24 were Qft.cn positive for 3.mmtxnorsactivity by both control at-e7 patient sara.
This levei of sensitivity is comparable to the best currently available asSays Using whole GAp65 purified * Trade-mark . .. ...P^ 6 4r; ~ ~ ~ ~
WO 95/07992 ir1PCT/US94/09-178 from brain or recombinant organisms. Besides avoiding laborious antigen purification, peptide based autoantibody screening, together with PCR based HLA
typing, may reveal epitope recognition patterns associated with progression or lack of progression to IDDM and its associated complications. Individuals determined to be at high risk could then consider therapeutic intervention.
It should also be noted that the GAD peptides recognized by autoantibodies were different from those recognized by NOD GAD reactive T cells in Example 6.

WO 95/07992 _ 71 _ PCTIUS94/09478 on v cn en cn N
cn A A
ri rn re m N iC
co c14 N A
N
Ln N >C A .-~ A 9C
N iC ~ 9C ?C aC DC aC aC AC eC eC ~C aC iC ~C aC 'Xi N
CV
N
ri N

N
O aC 'Q
a z ri ~
O OD

rA
z o A
Ln a A
O
N
ri el YG A
ri a> A A A>C N A >C
co ye r ~ ~
Ln xDG DC x8C ?C A 9C A X?C 9C !C !C x cn N
u) 0 u) O u) rq - N N
O

0 a E L, ~D frt R'i ~' t0 Q INf1 N ln a u, ln a v w w~-1 w w F. o ao M,.a WO 95/07992 _72_ PCTIUS94/09478 D A DC .-i M
N
M
M A A !C .-~
M
cn ri AA A ~ ~
N
~
rn cn N ae x A A ~e x a OD
N
~ aC .1 N
t0 N
N >C A?C A-C i4 A iC 9C DC 0G m N JC aC eC ?C x9C x9C N. 9C 9C aC xx'x. >C
fh O
N =
f~f cy 'IC x N
N
c~~t OC XA A .Ci x aC x >C t~ 0 N
c~v PCPCx A A A aCiCx x r U
0% rl rq co r1 O =
N
r1 If1 Ci =
1[1 +1 N
.-1 W R
m a 'C! 'C7 ri (! O
>+ Of cv a.)x .-1 =.1 >
1~-4 AAAA x AA N u U A
o O
rq M
x AA AA x 0 ao A A A A DC
-.4 m A A .t~ W N
r Ø0.0 A ~C N
~o ~
m tn ?C 9C 9C 9C A 9C DC DC aC 9C A~C ~C iC aC ?C ~ e~
41 m x ~+ m w O

fC XA A X X A ?C A?C Ln N x ,0 LA 0 LO Lf) 0 N N C') ~i ~i N ri N r1 N ri N~ i N m .....v~...vvv~... M
IUH do N O = Ot a !- W F7 R: R~ rJ t- V) N O

Sep-25-00 05:18pm Fram-S&B/F&Cu, +2328440 T-172 P.14/32 F-845 ,95107992 3fCT/US94/09479 -EXAlLPLE ix GU =MMON12ATION PROTheTS NOD iKrf,:B FROM IbbM
The availability af cpNAS encoding GAD65 aZlowa th+e testing of this molecule in new intexventive therapies s desigrigc3 to interfere with G4XD-MPeGifiF T cells. Testa weY'e avnducted to xamine the ability of GADB5 immunixation to protect NOD 1ri#,Ce bt S' weeks of age, A
time at vhich T cell responses to a number af b ce1.1 antigens and insulitis is already well established. zt s7D immunotherapy was effective at this stage, it would hold promise for treatment in humans in which the autoimmurne process has already been established.

,p.ratigens as An IPTG inducible T7 expression veutor was used to express bor_h human GAD55 and E. ao-11 b gale-otosidase (A-ga].). Xn IFTO induced recoea'binant. E. coIj, GAb and A-gal constitute about 10-20% of the total bacterial protein. Howevgr, almost all of the GAD was in inclusi8h bodies, which could be isolated and extensive3,y waahed to obtain material that 3s about 8ot GAD. We then did affinity puXificati.ons of GAD and ga2 on the basis of a hexa-histidine "tag" rdhich was attached to GAD during the raubc2o7friilig process. These 2s extr,a hi9tidine residues allow the rapid affinity purifiGatron (Novagen) of GAD by metal affinity ch,tomatography (Iiochuli, at al., Rip Techl7ology, 1:1321~1325~ 1988). The inclusion boely material is solubilized in 6M guanidirle hydraChloride (GH'CL),].Dtattt P--mercaptoethanol and 1t triton X-100. After binding to the column, the coluzh was extensively washed with GHCL a.nd 8M urt+a in phosphate buffers. Only the central peak GAD fraction was utilized for subsequent studies. Eiuman GAD65 shares 96t amino acid sequence identity with mur].ns GAU65, with m:ost of the aIDino dc3,d differences being conservative substitutions, * Trade-mark ~
- -The GAD preparation appeared to be free of immunologically detectable contaminants. It also appeared to be free of bacterial contaminants on overloaded silver stained gels. Analysis by a national reference laboratory found (0.06ng LPS/ug GAD. Human GAD65 did not induce T cell proliferation in (4 or )16 week old NOD or control BALB/c or (NOD/BALB/c) Fl spleen cells. The results using synthetic GAD peptides (Figure 10) precisely parallel the data using whole recombinant GAD (Figure 8). Other antigens described herein elsewhere that are not involved in IDDM (such as the beta galactosidase) did not induce NOD T cell responses. After immunizing mice with GAD, we were unable to detect cross reactive T cell responses in recall experiments with other proteins that were purified from recombinant E. coli by the same metal affinity chromatography procedure. Amino acid sequence analysis of GAD and (j-gal each gave a single expected amino acid N-terminal sequence. If there had been appreciable endotoxins, heat shock proteins, or other contaminants present in the GAD preparation, spleen, PBMC (Atkinson, et al.), and T cell proliferation responses that were not disease specific would have been expected.
Breeder mice were purchased from Taconic Farms and housed under specific pathogen-free conditions. Only female NOD mice were used in this study. The average age of IDDM onset in unrelated females in the colony was 22 weeks. Insulitis is generally observed beginning at 4 weeks of age. T cell responses to GAD, HSP, CPH were found by 6 weeks in age. The incidence of IDDM in female mice is 70-90% by one year of age.
Immunizations At 8 weeks of age, 25 ug GAD or control 0-gal. was injected intraperitoneally (ip) in 100 ul of incomplete Freunds adjuvant (IFA). Because there may be a requirement for continual antigen presentation c . ^

~

-(Ramsdell, et al., Science, 257:1130-1133, 1992) mice were treated again every 6 weeks. Urine glucose levels were monitored twice weekly. After observing above normal glucose in urea, blood glucose levels were monitored twice weekly. Two consecutive blood glucose level readings of 300 mg/ml was considered as IDDM
onset, after which the mice were sacrificed and spleen cells were tested as described above in Example 6 for evidence of spleen cell proliferation.
Immunization of 8 week old NOD mice produced a clear delay in the onset of IDDM compared to control ~-gal immunized mice (Fig. 11). While two of the GAD
immunized mice (open circles) developed IDDM at about the normal age of onset (20 weeks), the other 8 GAD
i5 immunized mice showed no signs of hyperglycemia until 36 weeks in age. Four of the GAD treated mice developed IDDM between 37 and 40 weeks in age. Four of the GAD treated mice currently remain disease free (at 52 weeks of age). In contrast, the majority of 0-gal injected mice (closed circles) had hyperglycemia by 22 weeks of age and 6/10 developed IDDM by 27 weeks in age. At 52 weeks of age, 2 of the /3-gal treated mice remain disease free. This experiment shows that GAD
immunization significantly delayed ((0.02) or prevented diabetes of NOD mice in which 0 cell autoimmunity has already significantly progressed.
0 cell autoimmunity is already well established at 8 weeks of age, and it is likely to also be in individuals determined to be at risk for IDDM on the basis of circulating autoantibodies. Although the mechanism of this protection is not clear, periodic injections of GAD have a profound moderating effect on the induction of disease.

24 DICKKYKIWNIIiVDAAWGGGLLMS

The invention now being fully described, it will be apparent to one of ordinary skill in the art that many changes and modifications can be made without departing from the scope of the invention.

Claims (34)

CLAIMS:

1. A polypeptide fragment of the GAD65 of Fig. 2 or Fig. 3 selected from the group consisting of:
KPCSCSKVDVNYAFLHATDL;

TAGTTVYGAFDPLLAVADICKK;
EYLYNIIKNREGYEMVFDGK;
IPPSLRYLEDNEERMSRLSK;
SRLSKVAPVIKARMMEYGTT;
EYGTTMVSYQPLGDKVNFFR;
ATHQDIDFLIEEIERLGQDL;
LAFLQDVMNILLQYVVKSFDRS;
EEILMHCQTTLKYAIKTGHP;
DERGKMIPSKLERRILEAKQ;
KHYDLSYDTGDKALQCGRHV;
AALGIGTDSVILIKCDERGK;
GLLMSRKHKWKLSGVERANS;
LEAKQKGFVPFLVSATAGTT; and VNFFRMVISNPAATHQDIDF.

2. An isolated polynucleotide sequence encoding the polypeptide fragment of claim 1.

3. The polynucleotide of claim 2 which is DNA.

4. A method for detecting autoantibody to GAD65 in a patient specimen which comprises contacting the specimen with a GAD65 fragment and determining whether autoantibody binds to the fragment, wherein the fragment is the polypeptide of claim 1.

5. The method of claim 4, wherein the specimen is blood.

6. The method of claim 4 or 5, wherein the polypeptide is detectably labeled.

7. The method of claim 6, wherein the detectable label is selected from the group consisting of a radioisotope, a fluorescent compound, a colloidal metal, a chemiluminescent compound, a bioluminescent compound, a phosphorescent compound, and an enzyme.

8. The method of any one of claims 4 to 7, wherein the polypeptide is bound to a solid phase.

9. An antibody to the polypeptide of claim 1.

10. A vector containing the polynucleotide of claim 2.

11. A host cell transformed with the polynucleotide of claim 2.

12. Use of a fragment of GAD65 to ameliorate a GAD65 associated autoimmune disorder in a patient, wherein the fragment is the polypeptide of claim 1.

13. Use of a fragment of GAD65 in the preparation of a medicament to ameliorate a GAD65 associated autoimmune disorder in a patient, wherein the fragment is the polypeptide of claim 1.

14. The use of claim 12 or 13, wherein the disorder is Insulin-Dependent Diabetes Mellitus (IDDM).

15. The use of claim 12 or 13, wherein the disorder is Stiff Man's Disease.

16. The use of claim 12 or 13, wherein the fragment of GAD65 is adapted for enteral administration.

17. The use of claim 12 or 13, wherein the fragment of GAD65 is adapted for oral administration.

18. The use of claim 12 or 13, wherein the fragment of GAD65 is adapted for parenteral administration.

19. The use of claim 12 or 13, wherein the fragment of GAD65 is adapted for subcutaneous, intramuscular, intraperitoneal, intracavity, transdermal, intranasal, or intravenous injection.

20. The use of claim 12 or 13, wherein the fragment of GAD65 is adapted for administration at a dosage of about 10.01 mg/kg/dose to about 2000 mg/kg/dose.

21. The use of claim 12 or 13, wherein the fragment of GAD65 is therapeutically labeled.

22. The use of claim 21, wherein the therapeutic label is selected from the group consisting of a radioisotope, a drug, a lectin, and a toxin.

23. A system for ameliorating a GAD65 associated autoimmune disorder in a patient comprising a fragment of GAD65 and a means for delivery of the fragment of GAD65 to a patient, wherein the fragment is the polypeptide of claim 1.

24. The system according to claim 23, wherein the means for delivery is by subcutaneous, intramuscular, intraperitoneal, intracavity, transdermal, intranasal, or intravenous injection.

25. A method for detecting the status of a GAD65 associated autoimmune disorder in a patient which comprises contacting a T-cell of the patient with at least one polypeptide fragment of claim 1 and detecting the response of the T-cell to the peptide.

26. A method according to claim 25, wherein the disorder is IDDM or Stiff Man's Disease.

27. The method of claim 25 or 26, wherein the response of the T-cell is stimulation.

28. The method of claim 27, wherein the stimulation is measured by cellular uptake of radiolabeled nucleoside.

29. The method of claim 28, wherein the nucleoside is 3H thymidine.

30. A kit useful for the detection of antibody to the polypeptide of claim 1 in a specimen suspected of containing said antibody, the kit comprising carrier means compartmentalized to receive in close confinement therein one or more containers comprising a container containing the polypeptide of claim 1.

31. A kit useful for determining the status of a GAD65 associated disorder in a specimen of a patient, the kit comprising carrier means compartmentalized to receive in close confinement therein one or more containers comprising a container containing the polypeptide fragment of claim 1.

32. A use of a diagnostically effective fragment of GAD65 for detecting a GAD65 associated autoimmune disorder in a patient having or at risk of having the disorder, wherein the fragment is the polypeptide of claim 1.

33. The use of claim 32, wherein the disorder is IDDM.

34. The use of claim 32, wherein the disorder is Stiff Man's Disease.