CA2132015A1 - Tissue implant systems and methods for sustaining viable high cell densities within a host - Google Patents

Abstract

Implant assemblies (10) and methodologies provide immuno-protection for implanted allografts, xenografts, and isografts. The assemblies and methodologies establish an improved boundary (34) between the host and the implanted cells. The boundary (34) has a pore size, an ultimate strength, and a metabolic transit value that assures the survival of the cells during the critical ischemic period and afterward. The boundary (34) allows the fabrication and clinical use of implant assemblies (10) and methodologies that can carry enough cells to be of therapeutic value to the host, yet occupy a relatively small, compact area within the host.

Description

WOg3/19700PCT/US93/0266~

2 1 ~201~;

~:

: ~:
:
:: ::
: SYSTEMS AND MFI HODS FOR CELL IMPLANTS

. el~ted Applio4tion::~
J ~ : S ' This~ application is a continuation in part of copending~U~S.~Application Serial No. 735,401 enti-led :"Close Vascularization Implant Material" ~iled July 24, 1991~
ield o~ th~ InYentions ~:
10~ The~inventions relate~to systems and methods ;~ : for~implanting living cells:withi~n a host.
Baokground~o~ :the~Inventio~
For~:severa~ years, r~esear~hers have been ~ rying to surgica~lly implant:living c211s in a host to `~ 15 ~ ~treat various~c:ell:and ~molecular~::deficiPncy diseases.
In theory, the~implanted:cells:wil:l generate biologi-cal:pr~ducts ~hat the h~st, because of disease or in-jury, cannot produce for itsel~f. For example, ~he : implant:~ :assembly: ~ can ~contain pancreatic cells 2~0 :~ (cluste~s of which are:called l'isletsl'), which gener-ate~insulin th~at~a diabetic host:lacks.
Yet,~in practice, conventional implant as-: : semblies and methodologies usually fail to keep theimplanted cells alive long enough to provide the in-,: ~ : - ~

~: :
~::

WOg3/197~ PCTJUS93/02665 21 32 01-~ ~ 2 -J
: , tended therapeutic benefit. For example, pancreatic cells implanted for the treatment of diabetes usually die or become dysfunctional within a f w days or weeks after implantation.
SFor a period after implantation, the region of the host tissue next to the implant assembly can be character.ized as ischemic. "Ische~ic" means that there is not a sufficient flow of blood in the tissue ; region closely surrounding the implant assembly. Usu-lOally, this ischemic condition exists during the first two weeks of implantation. Most implanted cells fail to live through this~period.
During the ischemic period, a foreign body capsule ~orms~around;the implanted cells. The capsule lS~ ~consis~ts of;~fla~ttened m~crophages, foreign body giant cells~, and~fibroblasts.~Conventional hypotheses blame the~ foreign~body;~capsule for causi~g implanted cells `~ to~die~;or become dysfunctional during the ischemic ~ per~iod~
;~ 20 ~ Th~e~inventors have ~discovered that these widely~held~hypotheses are wrong. The inventors have dlscovered~that~ the cells~ do not die because of the intervention ~of~the~ foreign~body~ capsule. I~stead, the~ce~ls~die~bec~ause~conventional implant assemblies ~ 2~5~ and~methodol~gies~themselves lack the lnnate capacity ; ~ to~support~the;implanted ceIls' ongoing life processes durin~ the~ critical ischemic period, when the host's vascular ~structures are not nearby. Because of this, ; the ~implanted~ cells~ perish before the host can grow ` 30 ~;new~vascular~structures~close~enough to sustain them.
When~implanted cells die during the ischemic period, a~classical foreign~ body capsule inevitably forms around the~implant. The persistent presence of this~ capsule~ led pre~ious researchers to the false conclusion that the~ho~st's foreign body reaction was ~; : :

' WO93/197~ P~T/US93/02~S
..
~ 3 ~ 213201~) the cause of implanted cell death, rather than its ; result.
The invention corrects these and other problems in existing implant assemblies and methodo~ogies.
Many previous implant assemblies have also failed t~ be useful in a clinical setting, because they Gannot be practically implanted and tolerated by : the host without danger or discomfort.
; For ex~ample~ an implant assembly that ~oused ; cells within hollow ~ribers was recently used by Cyt~Therapeutics~to~ successfully treat diabetes in ~ ~rats. The assembly consisted of 7 fibers, each being .~ 2 cm long and;0:.~0:73 cm in diameter. The pancreatic lS~ ~ ~cells;were~present~within the ~ibers at a density of a~bout 25,:000:~;cells per cm3. For this assembly to be , ~
olinically useful for the treatment of diabetes in : humans, ~ t ~would~ have to contain at least about 2~50,000 pancrea~tis~islets ~each islet ~ontains about 2~0~ l0~00~c~ells)'.~ This~means~;~that, to hold enough pancre-atic~cells~.;to.treat human~diabetes, the assembly would ha~e to be~about ll7:~feet:long~. :This makes tXe assem-bly~unusable~for~clinical~use in humans.
Recently,~ c~ells have also been ~ncapsulated .~ 25-~ in tlny~hydro~gel~vessel~s,~called~microcapsules. These ;::tiny~:~vessels~cannot~;be~implanted within the host's ;so~ft:tissuès~ because:~they lack th~e physical strengt~
*o withstand~the~physiological stresses normally en-countered c:lose~to:the host tissue. Instead, the 30~ microcapsules~ are~ suspende~d in a free floating state , ~
~ ; wi:thin:a~solution: that: is infused into the host's .~ :: perit:one~al'cavity.~
:In:real~ity, the~microcapsules have only lim-;~ ted clinical~application~. Not all persons can toler-35~ ~a~te~ their :injection:free of danger or discomfort.

. ~ -, ~ WO93/197~ PCT/US93/02665 r~ ~ 3 ~ ~ 4 ~

~icroScapsules are non-adhesive, and they do not sticX
to organs. Instead, they settle in large masses at the bottom of the peritoneal cavity. And, if imp}anted directly within the host's tissue, the microcapsules will rupture, and the contained cells would perish. ~or these reasons, microcapsules fail to provide a widely usable clinical solution to the problems surrounding the therapeutic implantation of cells.
The inventions have as an important objec-tive the design of implant assemblies and methodologies that combine effectiveness and prac-~ ticality required for widespread clinical use.
:~ 8u~mary~of:the Inventions lS . To meet these and other objectives, the in-ventions provide improved implant assemblies and ~ methodologies: that can carxy enou~h cells to be of `~ therapeutic value to the host, yet occupy a relatively small, compact area within the~hos~. The implant as-semblies and methodologies ~that the inventions provide also establish an improved boundary b~tween the im-~:~ ; planted~ cells~an~d the host. The improYed boundary sustains::the~viabil:ity of the imp~anted cells, both before:and:~after:the growth:of va~cular structures by 2~5 the host~
To assure the long term surYival and functionality~;of implanted cells, the host must grow new vascu:lar structures to serve them. The inventors have discovered that an animal host will not naturally provide thesé~new ~ascular structures. ~t must be ; stimulated to do so.
: The implant assembly itself must provide this crucial:stimulatinn to the host. Otherwise, new vascular structures will not form close to the bound-ary. The implanted cells will die or will not : :
.

WO93/l9~00 PCT~USg3/02665 , - s - 21 320 1 ) function as expected~
The inventors have found that some cells implanted for therapeutic reasons, like pancreatic islets, naturally secrete angiogenic material. "Angio-~; 5 genic" identifies a type of material th~t has the characteristic of stimulating the growth of new vas-~ cular structures by the host close to the boundary ,~ that separates the~ implanted cells from the host.
, "Close" means that the vascular structures generally , lO lie within about one cell layer away from the bound-ary, which is usually~less than about 15 microns.
These angiogenic source cells, if implanted, create their own~ stimulation for close neovascular ~ ,growth. Yet, other cells do not naturally secrete an-; ~ 15 ~ giogenic materials. These cells, if implanted alone, will~not induce vascular~ization. If these cells are imp~lanted~ the~ implant assembly should include a separate,angiogenic~ source for them.
,~ Still, the~presence of an angiogenic source ,~ 2~0~ d~oes~not~assure~cell survival during the ischemic pe-riod~,~before the~close vascular structures form. Even cells that~naturally secrete angiogenic material often ~ die~ or~become; dysfunct~ional~soon into the ischemic ,~ perlo~ Their,~re~lease~ of angiogenic material stops, `~ 25 ~ too~ ringing,vascularization~to a halt.
Th~e~in~entors~have discovered that implanted cells~peri~sh~;~during the ischemic period, because the assemblies housing~them lack the intrinsic capacity to bri~ng in;enou,gh~nut~rients and let out enough wastes to 30~ support the~ir~ ong~oing metabolic processes whe~ the host's~ vascular~structures~a~re absent. This capacity will be referred~to-as~"metabolic transit."
It~is ~the ~;~lack of sufficient metabolic transit innate~ in prior~ implant assemblies and methodoIogies,~a~nd~not~ the formation of the foreign :~ ~

~ ~ : :: ',.

WO93/197~0 PCT/US93/02665 ~ody capsule, that causes the implanted cells to ex-pire or become dysfunctional during the ischemic peri-od. It is the lack of suffîcient metabolic transit by the ~oundary that stymies the formation of close vas-:~ S cular structures and causes the implant to fail.
The inventors have discovered that an implant assembly will support the ongoing metabolic -processes of implanted cells during the ischemic peri-od, even when a foreign body capsule forms, if the : 10 assembly has a s~fficient metabolic transit value to support these pr~cesses in the absence of close vascu-lar structures. With their metabolic processes sup-ported, the cells survive the ischemic period. When the assembly includes implanted angiogenic source cells, they also release their a~giogenic materials to ~ stimulate~ new~vascular structures. Formation of the .~ new ~ascular structures, in turn, marks the end of the ischemic period. A sufficient metabolic transit value sustains and promotes all these complementary process-:2~0 es.~
ne~ aspect; of the inventions provides ` ; : impla:nt assemblies and methodologies that present an improved~boundary between the hast tissue and the im-planted cel~ls~ The:boundary~is characteriæed in terms :5:~ o~ its~pore sizei its:ultimate physica1 strength; and ts~metabo}i~c~tran~sit value. The metabolic transit value i5, in turn,~characterized in terms of the per-meability and porosit~ of the boundary.
The~pore~size and ul~imate physical strength : characteristios~serve to isolate the implant tissue cells from:~the: immune response of the host during the ischemic period~and afterward. The metabolic transit value serves to sustain viabllity of the implanted cells:during the ischemic period and afterward, even when a foreign body capsule forms.
'': ~ :
:~
;~' W093/lg700 PCTJUS93/Ot~5 21~201 ~
- 7 ~

In a preferred arrangement, the boundary has a surface conformation that also supports and fosters the growth of the new vascular structures that the ~: i~proved implant assemblies and methodologies stimulate.
~; Another aspect of the inventions provides a ~ methodology to derive and use a therapeutic loading :~ factor to characterize and predict the clinical efec-tiveness of a given implant assembly for a given cell 10type. The~therapeutic loading factor takes into ac-count the~number of cells that are required to be implanted to achieve the desired therapeutic effect;
the effective area of the boundary between the im-planted cells:and~host that the host can be reasonably lS ~e.Ypected to~tolerate; and the metabolic transit value needed : to sustain cell viability. Using the therapeutic loading factor, a practitioner can provide an implant~assembly that combines the benefits of com-pact size~;with ~the~ability to sustain the requisite -;20~therapeutiGal number of cells.
The~;inventions provide impl~nt assemblies and~methodologies~havi:ng:significantly improved per-ormance~characteristics. The~improved characteris-tics~sustain~ high~density ~cell populations within a 2~5~compact~area~ within::a::~host. Assemblies and methodologies~that:embody the features of the inven-~:: tionslsuppor:t:as~many~as ~ times more implanted cells . . in a given:volume than prior a~semblies and methodolo-gies.

3~0~Other~features~and~advantages of the inven-t~ions will~become:apparent upon~review of the follow-ing speciflcation,~drawings, and claims.
Brief Description~of the Drawi~
Pig~ l is ~a perspective view of an implant ~: 35assembly that~embodies the~features of the invention :
I
:~
'~

I' W0'93/lg700 PCr/U5g3/02665 ~ . ' ` ' 3 ~ 8 -being held in the hand of a practitioner;
Fig. 2 is an enlarged perspective view of the implant assembly shown in Fig. 1;
` Fig. 3 is an enlarged and exploded perspective view of the implant ~ssembly shown in Fig.
2;
~: Fig. 4 is a side section view of the implant assembly taken qenerally along line 4-4 in Fig. 2;
Fig. 5 is an enlarged and exploded lO : perspective view of another implant assembly that em-;;~ bodies the~Seatures~of the invention, showing the pra-ctitioner loading~implanted cells into the assembly;
Fig. 6 is an enlarged assembled view of the assembly shown in Fig. 5, be~ore the formation of a peripheral seal;
Fig.~;7 is an enlarged view of the assembly shown in~Pig.~6,~partially peeled apart to show the nterior;
Fig.~:8~is an enlarged assembled view of the ;~ Z:O ~ : assembly;shown~in`Fig.;~5 after the formation of a pe-ripheral~;s:eal;~
Fig.:~9~;is~a side section view of a portion of~the:~sealèd~assembly~t~akèn~generally along line 9-9 Fig.~l:O~is:a~side s~ection view of the assem-bIy betore~s-aling,~ taken~g~enerally along line 10-10 :Fig. ~ is a perspective view of a amination ;~slide~;holding the bottom layer of the lO~laminated~ boundary~ s;tructure that embodies the features of~thè~invention; `:
Fig.;~ 12~ is~;a: side section view of the :lamination~slide~taken generally along line 12-12 in ;~: 35~ ~Flg.::13 is a perspecti~e view of several , ~
,:~

_~

WO~3/197~ PCT~USg3/02665 ;' ,"''"

. lamination slides laid side by side for the ap-plication of adhesive filaments in the process of making the laminated boundary structure;
~ Fig. 14 is a side section ~iew of the : S laminated boundary structure with its top layer laid over the cement filaments applied in ~ig. 13;
Fig. 15 is a side section view of the laminated boundary structure clamped between two lam-:;: ination slides while the cement filaments cure;
Fig. 16 is a perspective view of individual bsundary wall elements being cut from the laminated structures made~following the steps shown in Figs. 11 to lS;
: Fig.:`17~is a diagrammatic depiction of an 15 ~ implant~assembly that embodies the features of the inve~tion~after hav~ing been:surgically implanted in host~tissue;~
Fig. ~18 i5 a diagrammatic depiction of the ` implant ass~embly~ duri~g the ischemic period, after 20 ~ about ~one~:or~two~ days of~implantation, showing the surrounding~wound~area fi~lled with~exudate;
Flg. l9 is~a diagr:ammatic depiction of the ;implant ~a~ssembly~ after~: about two weeks of implantat:ion~ ;showing~;:the :forma~ion of vascular ~ ;2~5 :~ structures olose;to~the~boundary, ending the ischemic `~ period; ~
Fig.~ 20; is a diagrammatic depiction of a section of the~implant assembly in which the implanted cell;s have:~survived~the~ i`schemic period, showing the 30~ formation of~va;scul~ar~structur~es close to the ~oundary and:the~result~lng:alteration of the foreign body cap-sule;
Fig.;21:is a ~diagrammatic depiction of a : section of th~e:implant assembly in which the implanted 35 ~ ~ cells have not survived the~ischemic period, showing :~
~;:
' ~

,~ WO93~19700 `2 1 3 ~ O 1 5 PCT/US93/02665 . ,.~

the lack of vascular structures close to the ~oundary and the resulting intervention of the foreign body capsule;
Fig. 22 is a graph showing the therapeutic . S loading ourve for pancreatic cells derived in ac-cordance with the invention.
Before explaining the preferred embodiments, it is to be und~erstood that the inventions are not limited in use to the details of construction or meth-; 10 odologies~ there set ~orth or as illustrated in the drawings. The;~ inYentions are capable of other embodi-ments and of being practiced and carried out in vari-ous ways.
Description of~the Preferre~Embodiment~:
; ~ 15 ~ ~ Figs.~l to 4 show an implant assem~ly 10 that~embodies~the~eatures of the invention.
The~assembly 10 can carry preselected types of living cells 12~for implanting within the soft tis-sue of a~host~ The~implanted cells 12 generate -;20 ~ biologica~l~products~that~the host, because of disease or~injury~ cannot produce for itself.
For~example, the implant assembly 10 ~an carry clusters~of pancreatic ~ells (called "islets"), ; which~generate~insulin~ for~release into and use by a 25 ~ dia~betic hos~t.~
he~ a~s~sembly 10 forms a porous, life sus-taining boundary~between the implanted cells 12 and the host. The porous boundary isolates the implanted cells 12 from~attack and~destruction by certain bio-3~0 ~logi~cal mechani~sms of the~host. At the same time, the porous boundary~associates;with the host's biological system closely~enough to~ransfer nutrients and wastes ; in sùpport of~the;biological processes of the implant-ed cells 12. ;The porous boundary also transfers the therapeutic~products generated by the implanted cells :~ :
: ~ .

~ WO93~197~ P~T/US93~02665 ~ - 2`1320:~

l2 to the host.
In the embodiment shown in Figs. l to 4, the assembly lO includes a hoop-like housing ll. The housing ll includes a first hoop element 14 and a 5second hoop element 16 that nests within the first hoop element 14. The assembly lO also forms a cell chamber 18 within the h~op-like housing ll.
The ~irst hoop element 14 has an upstanding cylindrical side wall 20 that peripherally defines an lOopen area. First and second central openings 22 and 24 lead into the open area. The first central opening 22 is smaller than the second central opening 24.
;~ This forms an interior step or ledge 26 next to the first opening 22~. ~
15Thé~ second hoop element 16 also has a central opening 28~. The~seCond hoop element }6 has an outer diameter~that is slightly greater than the inner diameter of~ the~open~;~area of the first hoop element 14~.` The peripheral edge of the second central opening 20~6~contains;a~s1ight chamfer~30 to receive the second `~ hoop element~l6.~Wh~en;assembled, the second hoop ele-`~ ment~ nests~snugly~ in; an interference press fi~
wit~hin~the~open~area~of the~first hoop element 14 (see ~ Fig. 2).
-~ 2~The~first~hoop element 14 and the second hoop~-element~ 16~are~made~of à durable biocompatible ceramic or~metallic~material, like titanium. Like titanium, the selected~material should als~ preferably be subject~to~detection within the host tissue by flu-30 ~oroscopy~ x-ray~ and~the~like.
The~specific dimensions of the hoop-like housing 11 may~vary according to its intended use and ` the volume aP cells~12 it contains.
In~one preferred embodimen~, the side wall 35 ~of ~the fir~st hoop~element l~ is about .055 inch in WO93/197~ ' r PCT/US93/02665 . 21~201~ t height and has an outer diameter of about .375 inch.
The open area has an inner diameter of about .325 inch where it joins the in~er edge of the chamfer 30 of the second central opening 24. The second central opening S 24 has an inner diameter of about .326 inch around the outer edge of the chamfer 30. The first central open-ing 14 has an inner diameter of about .275 inch and a depth of about .015 inch, where it joins the interior ledge 26.
; 10 I~ this embodiment, the associated second hoop element 16 has a height of about .025 inch; an outer diameter of about .326; and an inner diameter (for its central opening 28) of about .250 inch. The range of interference necessary to snugly join the ~ lS second hoop element 16 within~the first hoop element `~ 14 wi11 of course depend~upon the nature of the mate-rials selected.
The chamber includes a first porous wall element~32, a;second porous wall element 34t and a ~ sealing~gasket; or~ring 36 that is sandwiched be~ween them. The sealing ring 36 is made of a mesh polyester ` material. ~
: ~
Th~e~wall elements 32 and 34 and sealing ring 36~are sized t~ fit snugly within the confines of the ~hoop-like~ housing 1:1. And, as will be described in greater detail later, at least one (and preferably `~ both3 porous wall elements 32 and 34 have certain physical characteristics selected to protect and sus-tain~the viability of the cells 12 within the host.
~The ring 36 has a central open region 38.
The open ring~reg~ion 38, together with the overlying ~first and second porous wall elements 32 and 34, cre-ate the chamber 18 to hoId the implanted cells 12 (see ig. 4)-In making the assembly 10 shown in Figs. 1 ::

W093/19700 2 ~ 3 2 01 ~ PCTIUS93/02665 ~ A

..
to 4, the practitioner lays one wall element 32 ontQ
the ledge 26 formed in the first hoop element 14. The practitioner then lays the sealing ring 36 upon the : wall element 3~. Next, the practitioner inserts the desired amount of cells 12 to be implanted into the open region 38 of the ring 36. The amount of : implanted cells 12 is sufficient to induce the expect-ed therapeutic response in the host.
: The practitioner next lays the other wall element 34 over the first wall element 32, sealing ~;~ ring 36, and inserted cells 12. To complete the as-sembly 10, the practitioner presses the second hoop element 16 through the second central opening 24 into pressing engagement~against the adjacent wall element 15 ~ 34. This:seals the peripher~ of the cell holding chamber 18,~ which now snugly rests within the open - area of the hoop-like housing 11.
;~ Once assembled, one wall element 32 rests against the;interior l:edge 26 and is there exposed 20~ ~ through the first central openin~ 2~. The other wall element 34 rests aga;inst the second hoop element 16 and :is there~exposed through its central opening 28.
Figs.~5 to 10 show another implant assembly ~ 10' that embodies the~ features of the invention.
:~ 25~ Like,: the implant a~ssembly 10 pre~iou~ly described, the assembly lO'~includes a cell cham~er 18' formed by ~ first and:second porous wall elements 32' and 34' and :~ an intermediate sealing ring 36'.
Unlike the first described implant assembly 30~ lo, the assembly 10' does not rely upon the hoop-like housing 11 to hold and seal the chamber 18'. Instead, a preformed~ peripheral weld 40 bonds and seals the edges of the porous wall elements 32' and 34' to the ~; interior ring 36'.
In making the assembly 10' shown in Figs. 5 :
:: :
:

~ W093/19700 PCT/US93/02665 2~ ~ -. 1 ~201S~ , '`'`

to 10, the practitioner lays the sealing ring 36' upon one wall element 32' and inserts the desired amount of cells 12 to be implanted into the open region 3~' of the ring 36' ~see Fig. 5). The practitioner overlays ~ 5 the other wall:element 34' (as Fig. 6 shows). The :~ practitioner then forms the weld 40 to seal the periph~ral edges of the first and second wall elements 32' and 34' to~the ring 36' (as Fig. 8 shows). The weld compresses the peripheral edge of the assembly ; : 10 10' together, as Fig. 9 shows.
~ -:; The pra:ctitioner selects a sealing technique ;~ that does not damage the cells 12 within the chamber '. For example, the inventors find that sonic weld-ing:can;be~:used~without damage to the inserted tissue : 11 ce s.
In ~ ~ a ~preferred embodimen~ (using the laminated~structure~72 made as shown in Figs. 11 to : 16,:as~wil~ be~de~scribed later), the practitioner uses a~Branson:sonic welder. The welder is ~perated at 40 2~0;:~ Khz,~with 941AES~:actuator, 947m power supply, and glC
power contr~ller.~ :The horn amplitude is about 1.4 mils~ and~ s~ operated~at a hold time of about 0.3 s~c-: onds; a weld~time~:of~about .:20 seconds; a pressure of about:~50~PSI;~a~ trigger force o~ about 20 pounds; and :~: a~down:speed;~of~about 1.~2~5 (machine setting).
These:are typical operating ranges for mak-: ing the sonic weId and can vary according to the mate-rials used and degree of cell loading within the cham-ber.:~
30~ ~ ~The~ int~egral ~assembly 10' formed in this manne~ can be:implanted directly within host tissue, ;~ without use~:of an exterior housing.
Preferably, as ~ig. 8 shows, the assembly 10' :includes ~an attached clip 42 made of a material that can ~be~ detected~within the host tissue by : :
,~
: ' ~ :

WO93/19700 2 1 3 2 0 1 i PCT/US93/02~5 .~. ' ,, fluoroscopy, x-ray, and the like. In this way, the practitioner can easily locate the assembly 10' within the host, if required.
Like the first described embodiment, the specific dimensions of the assembly 10' may vary a~-cording to its intended use. And, like the first de-scribed embodiment, at least one (and preferably both) porous wall elements 32' and 34' have certain physical characteristics selected to protect and sustain the ~: 10 viability of the cells within the host.
~ Regardless of the assembly used, the : : practitioner surgically implants it in the soft tissue 44 of the host (see Fig. 17). During surgery, the practitioner~positions the assembly 10 so that the ex-posed first ~and second wall elements 32 and 34 rest cl:ose to the surrounding host tissue 44. I~ Figs. 17 ; to 21, assembly lO:also encompasses assembly 10'.
The first and second.wall elements 32 and 34 thereby together form the desired boundary 46 between 20 ~ ~:;the biological: system of the host tissue 44 li~ing : outside the~:chamber 18 and the biological system of the implant~t:is-ue cells 12;1iving within the chamber For a~period of time after i~plantation, the 25 ~ ~ region of the::host~tissue 44 immediately surrounding the implant ~assembly lO:is ischemic (see Fig. 18).
The region is ~ischemic, because the host treats the :: assembly 10 as a foreign body.
T~e host forms a wound area 48 around the 30 ~ ~ assembly l~O~(see:Fig. 18). The wound area 48 has spaces that bec~ome filled wi*h wound exudate 50. The wound exudate;50~keeps this area 48 ischemic.
.Soon after implantation, host inflammatory cells enter and occupy the exudate area 48.
"Inflammatory~cells" include macrophages, foreig~ body WO93/197~ rCT/U5~3/U~65 ~,~ 3? ~

giant cells, and fibroblasts~
The inflammatory cells try to remove the foreign implant assembly. Macrophages ~rom the host try to ingest the foreign implant assembly 10. In some cases, the macrophages coalesce to form multinucleated giant cells. Fibroblast layers form to create a fi-~: brous sa~ of cells and collagen around the foreign implant assembly 10, commonly called the foreign body capsule 52 (see Fig. 20).
~ ~ 10 The inventors have discovered that it is not `~ the foreign body capsule 52 that most threatens the ~:~ viability of the implanted cells during the ischemic period. Rather, the existence of the cells is most th~eatened:during the ischemic period when the bound-15 : ary 46 itself fails to allow enough extracellula~ nu-trients like~glucose an~ other metabolic support com-~ pounds pres;ent;~at~ the boundary 46 to pass to the :~ : cells. W~ithout~met~abolic support, the implanted oells ; become dysfunctional or:perish.
~ As~Fig. 1;8 shows, the wound exudate 50 forms : a~ flu:id~bar:rier~ between the vascular system of the ;:~ host~and~the~ boundary 46.~ This barrier hinders ~he ;~ ex~trac~llular::passage~of nutrients from the host vas-cular system~to~the boundary 46. The concentrations 25 ~:of~nutrients~:decrease as they transit the exudate bar-riér to reach the~:boundary 46.
The hostis inflammatory ce}ls that in time enter the wqund exudate region 50 also create a metabolic~sink~. These ~cells compete for and further extract more:~of:~the host~s extracellular nutrients : before they~reach;~the boundary.
If :the~host is stimulated to grow new vas-" ~
: cular structures S4 close to the boundary 46, host en-dothellal cells~will also enter the region 48. These cells begin the crucial process of forming the new ~ :
: :
~ ~ :

: :~

WO g3/1g700 2 1 3 2 - P~US93,02665 OIJ
... . ..

:
vascular structures 54. Still, their presence further contrib~tes to the metabolic sink effect. The host's endothelial cells further reduce the availability of ~utrients for the implanted cells.
The ischemic period will end, if enough neovascular structures 54 from the host grow within the exudate region 50 close to the boundary 46 of the ; assembly lO (as Figs. l9 and 20 show). The close vas-~ : cular structures 54 shorten the extracellular path .~ 10 that nutrients must travel to reach the boundary 46.
The close vascular~structures 54 provide nutrients in : higher concentrations to the implanted cells. Close vascularization~ also transports the therapeutic products generated by the implanted cells 12 to the 15:~ host.~
`;~ However,~all these desired benefits accxue only i~f thé implanted~ cells 12 survive the critical i:schemic period.i~
The~inventor~s have~di~scovered that the di-2Q~ :minished concentrati~ons o~ nutrien~s present at the :boundary~;4~6,~a~1though;~significaDtly :reduced by the ~ exudat~e;~barrier~and~:métabolic:~sink effects, are still `~ enough~to:suatain~the~implanted~cel~ls~. This is true, even in the~pr~ésence~of:~a~foreign body capsule.
`~ 25;~ Still~ the célls will~die, if the boundary :46 itself~:1acks!~the~:capacity~to~let enough of the re-maining nu~rients~through to the cells at a suffi-ciently high rate. The inventors refer to this capac-ity:as the me;tabolic transit~value.
30~ :The~i~nventors~have~discovered that the bo-undary;:46 itsel~f~can~also present another significant I .
barrier~to:the:passaqe of~nutrients. The added bar-rier effect of~ithe:boundary 46 can further reduce the already:diminished côncentration of nutrients, until 3 5: : there is e~sseneially nothinq left to sustain the ~ ~ :

WO93/l97~D 2 13 2 0 1'~ PCT/US93/0266S

cells.
The series barriers to the extracellular passage of ~utrients (the wound exudate 50, the bound-: ary 46, and the metabolic sink effect) also inhibit 5the reverse passage metabolic wastes from the îm-planted cells~
:.The inventors have discovered that two prin-cipal factors threaten the survival of the implanted cells during the ischemic period. The first factor : 10(which is conventionally recognized) is the failure to isolate the~cells from the natural immune response of the host. The second ~actor (which is not convention-: ally reco~nized~ is the undesirable additional barrier effect of the boundary 46 that impedes the essential 15 ~ ~ flux of already~ scarce nutrients to the implanted cells before` close vascularization fully develops.
: The same barrier effect impedes the flux of metabolic waste products away ~rom the implanted cells to the host.
Z0 ~the:boundary 46 does not support the on-~ ; ::going metabol~ic processes~of the implanted cells while `~ isolati~g: t;h;em ~rom the:immune response of the host ;~ duri~ng the: ischemic~period, the implanted cells wi11 not live~long~enough to derive the benefits of close 25~~vascularization~ if~it occurs.
According~to this aspect o~ the invention, then, th~ porous:boundary 46 is charac~erized in texms , of its pore size; its ultimate physical strength; and its metabolic transi:t value. The first two ~harac-30teristics serv~ to isolate the implant tissue cells rom the immune~response of the host. The last char-acteristic ~serves:~to :transfer nutrients and waste :: products in support of the metabolic processes of im-planted cells during the ischemic period/ before close ~ 35vascularization occurs. The last characteristic sus-'~

~ WO 93/197 PCr/~;99Y02665 ,~ , ' 2132O1~J

_ ,9 _ tains the viability of the implanted cells during the ischemic period, even as a foreign body capsule forms.
According to another aspect of the inven-tion, the assembly also includes an angiogenic material. The presence of an angiogenic material stimulates the neovascularization re~uired close to the boundary 46 to bring an end to the ischemic period.
According to yet another aspect of the in-vention, the porous boundary 46 includes an interface 47 ~with the host tissue that is characterized by a conformation that supports and fosters the growth of vascular structures by the host close to the boundary 46. ~ ~
lS ~ ; Further details of the beneficial character-istics of ~the boundaxy 46 and its associated host ; interface 47 will~now be individually described.
Bound~ry Pore ~ize ; The~boundary 46~has a pore size sufficient `~ O~ to~isolate~;the impl~ant tissue cells from the immune response of the ho~t.
As used~in~this Specification, "pore size"
refers~ to~the~max~imum por~e si~ze of the material. The pra~ctitioner~ determines psre~;size using conventional 2 ~ bubble~point~methodolQgy, as ~descri~ed in Phar-maceuti~cal~Technology, May 1983, pages 36 to ~2.
As a threshold requirement, the pore size selected must make the boundary 46 impermeable to the ~ vascùlar structure~that forms close to the boundary -~ 3~0~ 46. Penetration~of the pores by the vascular structure breaches~ the integrity of the boundary 46, exposing the implanted cells to the complete immune response of the hos~t~ Generally speaking, pore sizes less than about~2 microns will block the ingress of vascular structures.~

::
:~

WO 93/19700 P~/US93/û2665 2~32!~l5 ` 20 -The ultimate pore size selected also depends upon the species of the host and the biologic rela-tionship between the host and the donor o~ the implant tissue cells~
When the implanted cells are from another animal species (i.e., xenografts), the pore size must be sufficient to prevent the passage of both ~: inflammatory cells and molecular immunogenic factors from the host into the implant tissue chamber. As used in this Specification, "molecular immunogenic factors" refers to molecules such as antibodies and complement.
Pore sizes sufficient to block passage of both inflammatory ells and molecular immunogenic fac-tors in humans lie in the range of about .015 micron.
Of course, these~pore sizes are also impermeable to vascular structures.:
When the implanted cells are fxom the same animal species but having a different geneti~ make up 20~ (i.e, allografts), the pore size usually must be suf-ficient to prevent~the~passage of only inflammatory : cells ~r~om the host into the implant cell chamber. .In allografts~, molecular immunogenic factors do not seem to advers~ely: affect the viability of the implanted 2:5 ~cells:. Still,: some degree o~ tissue matching may be : required f~or~complete protection.
Pore sizes sufficient to block passage of inflammatsry cells in humans lie in the range of below about 0.:8:micron. :::~These pore sizes, too, are imper-meable ~o vascular: structures When~the implanted cells are isografts (au-: tologous implants of geneti~ally engineered cells~, the pore size;must~be sufficient only to prevent the ;;: isografts from entering the host. Still, with isografts, the pore size selected must also prevent ~W093/19700 2 1 3 2 0 1~ PCT/US93/~2~5 !: ' ~, - 21 -ingress of vascular.structures.
.~.Boundary ~trenqth The boundary 46 has an ultimate strength value that is sufficient to withstand, without : 5 rupture, the growth of new vascular structures, the growth of new cells within the chamber 18/~8', and :~ other phy$iological stresses close to the host tissue~
Keeping the;: boundar:y 46 secure assures isolation of the implanted cells from both the immunogenic factors and :inflammatory cells of the host.
These physiological stresses are caused when the host moves about in carrying out its normal life functions. The proliferation of implanted cells and the ~rowth of vascular structures 54 also contributes to the physiological stresses close to the boundary 46. The stresses~:~challenge the physical integrity of the boundary ~4~6:by stretching or otherwise deforming it.
`~ Absent à~sufficient:~ultimate strength value, . ~ 20~ normal~ physiological stresses~can rupture the boundary 46,~exposing~the~implanted cells to the full effect of the: host's immune:~;and inflammatory systems.
The~inventors presently believe that ulti-mate~ strength :~values:; sufficient to withstand 25~ physio~logica~ stresses~clos:e to:~:the host tissue ;~ without~rupture:~:~in~animals lie above about 100 pounds , ~
per~square inch~(PSI)~ In~comparison, the ultimate ~ strength value for PVA:hydrogel microcapsules is only ;~ ~ ~about 2 to 2;~.5~PSI. ~
`~ 30 ~ The~ultimate~strength values are determined by measuri~ng:~the ~tensile strength of the material.
Tensile~strength~îs measured by ASTM D-412.
Metabolic ~r~n it V~lue Th~e:boundary 46 a~lso has a metabolic transit ~ :: 35 value that sustains a flux of nutrients into the :~

W0~3/~97~ PCr/usg3/o266s ~ .
chamber 18 and waste products from the chamber 18 suf-ficient to sustain the viability of the implanted cells during the ischemic period.
The metabolic transit value takes into ac-count the permeability value (P) and the porosity val-ue (PORE) of the boundary 46.
The_Permeability_Value The permeability value ~P) i5 the measure of : the amount of solute that travels through the boundary ~10 per unit time and unit surface area, given some fixed ;~ external solute concentration (measured in cm/sec in ~:~ this Speci~ication). Example 1 sets forth a methodology for determining the permeability value according to this~aspect of the invention.
15 ~ The~:Porositv Value The porosity value~ (PORE) represents the space in ~the boundary 46 that does not contain material, or is~empty, or is composed of pores. Ex-pressed as~a~percentage, the porosity value (PORE) 20::~ measures; the %:~volume~of the boundary 46 that is not occupi~ed~by~oundary~material.
To derive the porosity value PORE (in %) for materials~having~a~PORE~:egual to sr greatex than 10%, the~practitioner:uses the folIowing formula:
25 ~ PORE~= ~lOO~ (pb/pm) where~
Pb is the density of the boundary as ;~ determined from its weight and volume, and : ::Pm ls the density of the boundary mate rial.
To derive the psrosity value PORE (in ~) for materials having~a PO~E less than 10%~ the practitio-ner uses using a scanning electron microscope to sb-tain the number of pores and their average diameter on the boundary. PORE is then derived according to the :

~ WO93/19700 PCTJUS93/02~
,.,......... ., . , 1~2ol,.~,.

following furmula:
:
PORE ~ N~(d2/4) where:
N;is the pore den~ity and e~uals (pn/a), ~: 'Pn is the number of pores in the boundary, a is the total area of the boundary (in cm ), and n is the transcendental constant 3.1416... , d is the average diameter of the pores (in cm)-The inventors have found that, above a 15; ~ threshold minimum porosity value, the permeability value:~is the~ principal influence upon the o~erall m~tabolic transit ~value. Still, below the threshold ;~ minim~m porosity value, the metabolic transit value must;~also~ ake~i~nto account the~porosity value and the 20~ physical structure:of the porous boundary 46. These conslderations will be discusséd later in greater de-; To~simpli~fy~the~ election o~ an boundary 46, he~invento~s~:r~commend the use o~ boundaries having 5~ a~po~os~ity~alue:~(PORE3 greater~than the observe~ min-imum threshold~valu~e.~ Then, metabolic transit Yalue and:the~permeability value can be treated as the same~
As the following Example l shows, the inven-tors~:~ have~discovered that there is a direct cor-relation: bctween ~:the metabolic transit value and mplanted cell survival during the ischemic period.
EX~MPLE~1 Embryonic lungs enclosed in membrane ~ham-bers having dlfferent permeability values were im-35. planted in~subcutaneous~sites in rats.

~:~
~: :

~ ' WO g3/19700 P~/USg3/~2665 ~ 2 1 3 2 0 1 ;~

1. Permeability The permeability values for the membrane chambers were obtained for insulin diffusion in a con-ventional benc~top dif~usion chamber/ made by Crown ~lass Company, Somerville, New Jersey tPart Number DC-100) t using radioactively labeled (125 I~ insulin as ; ~ the solute (obtained from ICN Biochemicals). The dif-fusion chamber had two ch~mbers (which will be called Chambers A and B~, each with a volume of 3 ml. The diffusion chamber presented a membrane surface area between the:two chambers (where diffusion occurs) of ~: 0.7 cm2.
The practitioner cuts the membrane material to be tested to a predetermined, known size.
lS If the membrane is hydrophobic, the practitioner wets the membrane before conducting the permeability:~test, using conventional wetting tech-niques. ~ .
The practitioner places the membrane in the 20~ di~fusion~ ch;amberO The assembly of the diffusion : chamber loc~ates the membrane between the two chambers of equal volume:,~called:~Chamber A and Chamber B. 'In his~way,~the~prac~i:tioner:also fixes the cross sec-tion~al:area~:~(A) of~thé membrane. The diffusion cham-~ ber :is~ uniformly heated~to a temperature of about 37 degrees C during:the test. ~
The practitioner loads e~ual am~unts of buf-fer solution into Chamber A and Chamber B~ The buffer solution can vary. ~In~this Example, the practitioner ~ ca~ use phosphate bu~fered saline, 0.5% BSA as the buffex solution.
: The~practi~tioner then loads equal amounts of unlabeled (non-radioactive~ insulin (about 3 . 4 micro : unitslml) into Chamber A and Chamber B. Porcine pan-: 35~ creas insulin purchased from Sigma with an acti~rity of ': ~
:

WO93~1g700 PCT/US93/02~5 ~ 213201~

26.1 units/ml, or comparable material, can be used.
The unlabeled insulin occupies any adsorption sites that may be present.
The practitioner uniformly stirs the fluids . within the chamber at about 600 RPM, using a magnetic stir plate :and magnetic stir rods (about 1 cm in length) placed in each Chamber A and ~. The prac-titioner allows the system to equilibrate for about ;~ one hour.
~; 10 The practitioner then removes a selected volume of buffer solution from Chamber A and adds back n equal volume of radioactlve insulin. The radioac-tive insulin suspension is filtered before use to re-move free 125}
;15~ While ~stirring the fluids within Chamber A
and Chamber B, the practitioner draws equal aliquots ,, ~
of fluld from ea~ch Chamber A:and B (e.g. about 15 uL) at 2, 4, 6~ 8~, ~10,~15, and 30.minute intervals.
The::~pract~itioner then counts the radioac-20~ tivi~ty levels~in:the samples using a gamma counter.
The~pr~actitioner:determines ~he change in the-counts~ .e.~ insulin concentration) in Chambers A~:and;~ per~unit~of time,:suitably corrected for back-m ground:noise~
:25~ The~practitioner graphs the c~unt and time pairs~o~each~Chamber:in terms of time versus the coun~s (with the~ counts :being: the Y-coordinates and time being ithe X-coordinates), restricting the analysis to:points~for which the counts in Chamber B
~ :are~less than;about 10% of the initial coun~s in Cham ber A. :The~practitioner then derives a linear equa-~
tion, fitting~the range of counts (yj over the set of times (x~ ~or:each Chamber according to the following equations~
or Cham er A:

':

~ WO93/19700 . PCT/US93/02665 21320 ~

Ya = Ylntercep~ a * X) where YlnterCep~ is the count value where the graph intersects the Y axis, and ~a is ~he slope of the Chamber A graph.
For Chamber_B:
Yb - YJn~ercept + (Nb * X) where Y~n~erCep~ is the count value where the graph intersects the Y ax}s, and Nb is the slope of the Chamber B graph.
The practitioner preferably uses a commer-cially available computer program to simplify the der-ivation process des~ribed above.
: 15 The practitioner then derives the per-meability value (P~ according to the general expres-sion: :

V~O * d~cb - M ~'MA ~

: where:
~ Vb is the volume of Chamber B
;~ 20 ~ ~ dMb/dT;is the c~ange in counts in ~ham-; : ber ~ per unit time, which is the slope of the B graph derived above ~Nb),;
:~:: P is the permeability Yalue~
.j A is the area of the boundary tested, and Ma - Mb is the mass gradient of insulin across the membrane.
The practitioner knows Vb and A, which re-main constant throughout the test. The practitioner also knows dMb/dT, the slope of the graph fvr Chamber ~ s (~b~ from the linear equation derived for Chamber R.
:

WO~3/197~ 1 3 2 0 1~ P~T/US93/026~5 < 2 The practitioner converts the units of Nb ~counts per ; min/min) into counts per minute/sec by dividing by 60 (the number of seconds in a minute).
The practitioner calculates Ma by solving 5the linear equat~o~ derived for Chamber A for y when : t = 15 minutes (i.e., the mid point time for the : test~. By us~in~ the mid:point time for the test, the practitioner obtains an a~erage value for the period of the test. The practitioner similarly calculates Mb ~l0by solving the~first order linear equation derived for : Chamber B ;for~y when t = 15 minutes. From these val-ues, the practitioner calculates Ma ~ Mb.
:The~;practitioner can now derive the per-meability:value:(in~cm/sec~) a;s follows:

~ p = ~ VbNb 60A~

15~Actually~ the permeability value derived ~ :also~includ~es:the~boundary layer effect~ that are as-`~ : ;sociated~with~ineYi:ta~le~stagnate fluîd layers at the membrane sur;face~in~Chambers A and B during the test.
~ To~:arr:ive~at~the~"true" ~intrinsic permeability vaiue ;~ 20 ~ ~for~the:boundary,~:the~practitioner would have: t~ ad-just~or th:e-~boundary layer~efects. However, for the purpos~es~of~thi~s~invention, a knowledge of the inher-ent: membran~e~permeability is not essential, because it will be proportional to the experimental permeability 25~value determined~ following ~he methodology detailed above~
Yet,~ the~ practitioner can follow the foregoing ~methodology to quantify the relative per-meability va:lues~::for ~selected boundaries, since 30boundary:layer effects will remain c~nstant as long as the sti~rring~method used remains the same ~ , : :

WO93/197~ ~ PCT/US~3/02665 2 ~ 3 ~ O 1 The disclosed methodology can be used to as-sess whether a given boundary fits the criteria es-tabli~hed for the permea~ility value according to this aspect of the inYention.
2. Porosit~
The porosity values ~PORE) of the boundaries tested ranged ~rom less than about 15% to greater than about 70~.
3. Determining Cell Survival lO~mbryonic lungs were removed from Lewis rat embryos between days 13.5 and 17.5 of development.
The lungs were kept on ice in Dulbecco's Modified Eagle's Medium (DMEM), 20% fetal bovine serum. The lungs were minced until they were approximately 1 mm2.
15Minced lung tissue (5-10 ~l) was placed into implant as~semb~ies like those shown in Figs. 1 to 4. The lung tissoe was encapsu:lated within~test membranes having varlous permeabilities, porosities, and pore sizes.
The implant assemblies: were pla ed in D~EM (20% fetal 20~bovine serum~at 37~degrees C until s~rgery, which occurred within 2 hours. :~he implant assemblies were ::implanted;~in sub~cutaneous`or epididymal fat sites,in : male~ewi~ra~ts for 3 weeks.~
After ~hree~weeks of implantation, the as-:25~:s:emblies were explanted, trimmed of excess fat, and ixed~ with 2~glutaraldehyde in Sorensen's huffer.
Sections of the assemblies: were stained with ~, hematoxylin and eosin.
Cell~survival was scored based upon histo-30logical appearance of the implanted cells. Tissues were scored as "éxcellent" i~ they had normal charac-teristics of lung:tissue, such as epithelial tubules, ilia, and formed cartilage. Tissu~s were scored as "good" if the tissue were:still alivel but not well dlfferentiated (for example, a high number of :~ :
!

WO 93/19700 2 i 3 2 01 j P~/US93/0266~
,~, mesenchymal cells). The tissues were scored as "poorl' if no or few cells remained alive.
In othex histology studies using implanted pancreatic ~::ells, survival assessment would involve analyzing the differentiated funr tion of the pancreatic cells in terms of their insulin release in the resporlse to a glucose challenge.
Table l shows the permeability value for those boundaries having a porosity value (PORE) greater than 70%, correlated with the survival of the implanted lung tissues.
Table l: Mem~ranes with PORE > 15%
Membrane Pore Size or MW Perme- Tissue ability' Survival cel!ulose acetate1 ~ unknown 9 excellent cellulose acetate~ unknown 5.3 exce11ent Biopore~ ~ ~ 0.45 ~m 2.6 excellent oiyvinyl difluoride1 unknown 2.5 ~ood ~ ce11ulose mixed ~ 1.2 ,um 2.0 poor; ;~ 20 ester2 po1yviny1 dif1uoride~ ;unknown ~ 1.7 ~ood po1ypropy1ene3~ 0.075 pm ~ 1.4 poor ; ce11u10se~ acetate 1 ~ ~ unknown 1.3 poor . ~
- ~ cellulose mixed~ 0 4~ pm ~ 0.9 poor 25~ ~ester2 ~
~ po1ye1hy1ene3 ~ 0 08 ~m 0.9 poor ; ce11u10se4 300 kD 0.6 poor , ~~ cellulose4 ~ 50 kD Q.2 poor 30~ ~*X 104 cm/s~
Baxter Healthcare Corporation ~D~ ~ield, Il) : 2:Mi1lipore Corporation (Bedford, Ma~
~ Hoechst Celanese (Charlotte, NC) :~:;: 4 Spectrum~Medical Instruments (Los Angeles, Ca) :`:
:: ~
: ~

~ W0~3/l~7~ ` ` PCT/U5~J~

. ' Table 2 shows the per~eability ~alue of those boundaries having a porosity value ~PORE) less than 15%, correlated with the survival of the implant~
ed cells.
S Table 2: Membranes with PORE ~ 15%
Perme~
embrane* Pore Size ability~ Survival Nucl~pore1 0.8 4.4 Fair Nuclepore 0.4 3.1 Poor Nuclepore 0.22 2.3 Poor :~ : Poretics2 : 0.1 2.2 Poor Poretics 0.08 0.5 Poor : Poretics 0.05 1.2 Poor Poretics 0.03 0.~ Poor ~~Poretics . 0.01 0.2 Poor ~ :
* polycarbonate X 104 cm/s (1) Nuclepore Corporati:on (P~leasanton, Ca~
20 ~~(2) Poretic Corporation (Livermore, Ca) Tables~ and ~2 demonstrate the direct relati~onship ~between the~metaboli~ transit value~of the~boundary~and implanted cel1 survival. More par-ticularly,~the Tables show that implanted cell survi-;~ 25~al signific~ntly improves when the pexmeability ~alue o~ the boundary;increases.~
~For the type of cells studied in Example 1, boundarie~ ha~iny a permeability value f~r insulin less than about 1.~5 x~10~cm/sec, as determined using 0the described~methodology, consistently did not sup-! ~ port cell su~vival, regardIess of the porosity value.
Yet, bounda~ies ha~ing a permeability value for insu lin greater than abo~ut 1.5 x 10~ cm/sec and a porosity value greater than about 15% uniformly suppor~ed vig-WOg3/lg700 PCr/US~ 665 - 31 - 213~01:j :~ , orous cell survival.
Boundaries having a lower porosity value (less than about 15%) also supported cell survival (see ~able 2). S~ill, the metabolic transit value ~or these less porous boundaries requires a higher rela-tive permeability value. For the type of cells ~ ;; studied in Example 1, boundaries havi~g a lower poros-1;~ ity ~alue (less than about 15~) supported cell surviv-~: al when the:permeability value for insulin was greater than about 4.0 x 10~ cm/sec.
: : The inventors believe that, when considering less porous boundaries, their specific physical ~: structure must also be taken into account. The less porous interfaces used in Example 1 were track-etched 15: ~ ~membranes;.~ These~membranes have uniform cylindrical pores separated~by~ relatively large, nonporous re-gions.
The;~ poor ~tissuè survival using the low ~ porosity~ boundaries could~ ~e due to uneven .~ 20~ ~ localization~of areas~of high~permeability, or due to constraints~ produced:: by cells on the particular physical ~properties~of the track-etched membra~es.
For~ example~ the~cells ~may be more efficient at `~ plugging~up~the~cylindrical pores of the track- etched membranes~ ;either:~:with cell extensions or cell secretions.~Thus~ al~though th:e track-etched mem~ranes have high permeability values in vitro, the response of the cèlls~in viVo may prevent the attainment of suf:ficient ~metabolic transit to support the graft 3:0 cells.: :~
Example~ emonstrates a methodQlogy ~hat can be followed to idéntify for other cell types the :~ applicable metabolic transit ~alue tha~ assures cell survival:during the ischemic period after implantat-~:1 35 : ion.
I -I
I
I
I ~ ~ :
l ,~ , 1 ~
,, WO93/1~700 PCr~USg3/02665 . . ~, ! ,"

The absolute per~eability and porosity val-ues that constitute a given metabolic transport value will depend upon the type of cell and the methodologies of determining permeability and porosity. Different condition~ will give different absolute values. Still, regardless of the test con-::~ ditions, ~he relative differences in permeability and ~ porosity ~al:ues derived under ronstant, stated condi-.~ tions will servé as an indicator of the relative capabilities~ of the boundaries to support implanted cell viability.
Tables 1 and 2 also show that good tissue survival occurs even with membrane materials that are ;~ ~ subject to the fvrmation of an avascular fibrotic re-15~ ~sponse (the so-called "foreign body capsule"). The ;fact that ~t;hese~;~membrane materials create this response has,:~in:the::past, led to the widely held view that the formation ~f:the foreign body capsule caused ~ poor~d~iffuS~ion~of:~nutrients. ~Example 1 shows the er-.~ ~ ~ 2~0 :~ ~ror of thi~s~conventional wisdom.
: As ~Table~:l shows, ~he use of relative thicker cellulose~ acetate~membranes with 0.45 micron re~:size~ 3~o~microns thick)~having an insulin perme-abi~lity o~O~.9~x: 10~ ~cm/ ec: results in poor tissue 2s~ survival.: On~the other ;hand, the use of relatively thinner~cellulose~ acetate~membranes with the same : approximate~:pore size~:(10 microns thick) and having a greater permeability o~f 5.3 x 10~ cm/sec rPsults in excellent tissue~survival. ~
3-0 : ::~ ~: The~thickness of:the membrane does not alter the foreign:body response;; a~f~reign body capsule will form~whether the membrane is relati~ely thick or thin.
;~ However, membrane thickness does alter the per-;~ meability value~
~ 35 : Thus, the cells died when the. thicker :~
I ~
~ :
I ' ~ `

WOg3/~97~ 2 1 3 2 0 1~ PCT~US93/~2~5 ., boundary was used, not because of the formation of the foreign body capsule, but because of poor nutrition and poor waste removal due to the low permea~ility of the thicker boundary. The tissue survived when the thinner boundary is used, because the higher per-meability provided improved cell ~utrition and im-proved waste removal to support cell metabolism, e~en when the same foreign body capsule form~.

In an experiment, the practitioner grew RAT-2 fibroblasts (ATCC CRL 1764) in 20% Fetal Bovine Serum, 2 mM l-glutamine, and D~EM (Sigma) (high glucose) until 100% confluent. The RAT-2 cells were split l:2 in~the above media, 16 to 24 hours before ;;~ 15 surgery.
On the day of surgery, the cells were washed with 15 ml of~HBSS~(no ions) and trypsinized off the culture ~iask. ~ The practitioner n~utralized the trypsin by adding 5 ml of the above media The ~ pract~itioner pe1l~eted the cells by centrifugation lO00 rpm, lO~mi~nutes, at 22 degrees C).
The~ p~lleted cells were counted and resuspended~-in~media~ in;~three concentrations: 5.3 x 103 cells/10~1; 5.~8:x 105 cells/lO ~l; and S.8 x lO6 ~ cel1s/10~
Impl~a~nt~assemblies~like that shown in Figs.
1 to 4 having~boundaries~ of differing permeability ~; values were made. The permeability values ranged from 0.2 x lO~ cm/sec~to 9 x 104 cm/sec (see Tables l and ;30 2~ to fo11Ow~ The ~tota1 boundary area for each assembly was~about .77 cm2.~
The ~arious cell concentrations were loaded into the assemblies.~ The practitioner implanted the assembli s both subcutaneously and within the 3~ epididymal fatpad of bost rats.

WO93/197~ PCT/US93/02665 ~ .
After 3 weeks, the assemblies were explanted ~: and examined histologically, as described previously.
s The inventors observed that assemblies load-ed with 5.8 x.103 ceIls and 5.8 x lO5 cells displayed : 5 excellent results, given sufficient boundary per-meabîlity values. A~ter 3 weeks of implantation, the initial load of 5.8 x io5 cells proliferated to ap--proxima~ely 2.0 x 107 cells. The inventors observed that assemblies having higher initial loads of 5.8 x . ~ lO 106 cells displayed poorer results.
Lower initial loads (less than 5 X lo6) were : able to survive the ischemic period and even proliferate 30:to;3000 fold. The final cell counts in the assemb~l~ie~s ~with lower initial loads were three lS~ ~times higher~th~an the initial load of th~ assemblies th.at:failed~ because of higher:initial loads. Thus, ` ~ :;high loads~of~cells (greater:than 5 x 106) are unable ~ to: survive~during the ischemic period, yet the same -~ cell~ loads :are~:able to~survive after the ischemic ~` 20~ period as~proyeny:~of~the :oells fxom lower initial loads.
,,, Clo~e~Va~cularization at the Boundar~
;Presencè~of_Anqio~en c Material ``~ Neovascularization~;close to the boundary is ;25 ~ essential to~thé~long~term~:~survi~al of the implanted cells~within:the~host.~:~ The inventors have found that .j~ the host ~will ~ot grow new vascular stxuctures 54 close to the~boun~ary (as Figs. 24 and 25 show), un~less~ it~;is:~stimulated to do so. Without proper : :stimulation, ~the :`ischemic period never ends, ~ecause a classical foreign:body~reactiQn occurs.
~ : The~assembl~y::lO :therefore includes an an-;~ giogenic ma~erial 56 for :stimulating neoYas~ulari zation clo e to the boundary.

: ~

~ WO93/1970Q rCT~U593/~2~-5 21 32 01 ~

The specific identity of the angiogenic ma-terial 56 is ~ot known. Still, the inventors have determined that the presence of certain cells stimulate neovascularization, while others do not.
For example, the presence of lung tissues;
pancreatic islets; adult pancreatic ducts; and cul-tured cell lines of fibroblast~, mammary gland, and smooth muscle cells induces or stimulates neovascularization, when compared to the vas-cularization on control grafts where these cell types were not present.
: :
o~ the other hand, the presence of primary skin fi~roblasts and microvascular endothelial cells do not induce ~neovascularization.
15 ~ The inventors believe that certain cells in-duce or~ stimulate ~neovascularization by secreting angiogenic factors.~Bec~au;se the stimulus crosses mem-~ branes that; are~ impermeable to cells, it must be a `~ molecular;signa~l~that the living cell generates. This 20~ further~underscores~the need to support the implanted cells~ during ~the ischemic p~eriod. If angiogenic source cells~ perish,~the mo~ecular signal stops, a~d the neovascularization process~comes to a halt.
According to this~ aspect of the invention, 25 ~ when~ cells~ are~ implanted ~that have a desired therapeutic;~effect, but do not secrete angiogenic material, ~the;~a~ssembly 10 includes a separate an-, ~ g`iogenic source cell or~material 56.
FolIawing the invention, the practitioner ;30~ ~ ~;sele~cts~an~boundary~46~having a sufficient metabolic transit value~ to support the viability of th~
i;mplanted cells;,~ .e.~,;the an~iogenic source cells and other non-angiogenic, therapeutic cells (when present) implanted with them~. The practitioner also sele~ts a pore slze and ultimate physical strength to make the W093/19700 PCT/US93/0266~
' 2 i 3 2 0 1 "~3 ' ~ ti _ _ boundary 46 impermeable to the neovascular growth that the angiogenic source cells stimulate.
Alternatively, the practitioner may coat the exterior of the boundary 46 itself with an angiogenic material 56. Of course, the coated boundary 46 still must have sufficient pore size, ultimate strength, and metabolic transit value to sustain the cells 12 iso-lated behind the boundary 46.
~: Because the new vascular structures 54 can-not penetrate the boundary 46, and because the angiogenic signal to the host continues, the new vas-culatur proliferates~close to the boundary 46.
As Fig. 21 shows, when the cells 12 die during the ischemic period~, and close vascularization 15~ is~not stimulated,~the~fibroblasts of the foreign body capsule 52 become closely packed and dense. However, as Fig.~20~shows,~when~the cells 12 survive the is-chemic period,~and~the process of close vasculariza-tion is~stimulated,~the fibroblasts of the foreign ~20~ body~capsu~le 52 is ~altered to~form a less dense and ;more dispers~ed~stru;cture.~
2~)~ Confo~maeio tor Close Vascularization n~the~ preferred~;embodiment, the porous boundary ~46;~includes~an interface 47 with the host 25~ tissUe~ that~is~characterized~by a structural confor~
mation that;further enhances the growth of vascular structures~by~ the~host~close to the boundary.
To~ach~ieve this result, each wall element ; 32/32' and 34t34'~ of the assemblies 10/10' includes a ~ 3~0~ first porous~region 5a~ and a different second porous `~ region 60.~The~first por~ous~region 58 comprisesm the boundary 46 previously~described. The second porous region 60~comprises~ the interface 47.
The fIrst ~porous region 58 faces the ~ 35 implanted cells 12 ~see ~ig. 20~. The first porous `~

~ W0~3/l9700 2 1 3 2 0 lJ PCT/US93/02~5 "`', .

, ~ 37 -region 58 has the boundary characteristics, above de-scribed, of pore size; ultimate physical strength; and metabolic~transit value. It is this region 58 that ates the ~implanted cells from the immune ; 5 mechanisms of~the~host,~while sustaining their viabil-ity through~the flux of nutrients and wastes during the ischemic~period.~ ` ~
The~second~porous region 60 faces the host tissue 44 and forms~the interface 47 with it (see Fig.
10 ~ 20). The~se~cond porous region 60~has an architecture that~ènhances~the~formation of vascular structures 54 close~ to the~boundary ~46. The formation of these vascular ~structures~ 54 ~within the second region 60 mark the~end~;of~the~ischemic period. Vascularization 15~ in~the second~region~60~sustains the viability of the imp1anted;~cells~12~after~the ischemic period~ ends.
A~foreign body;capsule~52 still forms about the ~ implàDted~ a;ssembly~ o.~ ~HQwever, close vas-ar~ization~within~the~second~ porous region 60 can 2~0~ alter ~the nor.ma~ configuration of the foreign body capsule 52~ s ~ig 20 shows, a life sustaining vas-boundary ~;46~ eeping ~fl~attened~macrophages, foreign body ~gi~ant ~ 11s,~ and~ fibroblasts from pressing 25~ against~and~block~inq~the~boundary 4~.
B~ec;ause~of~the~pore~si~e, strength, :and per-meability cha~racteristics of the porous first region `'~ 58,~it~is impermeable~to~the neovasculature 54 formed in ~the~seco ~reglo ;~6~0~
30;~ The~ nventors ;~ believe that close vascularization~ occurs~ if~ the~ three dimensional co`nformation~of~second~region~60 creates certain host :, ~
inflammatory~;ce11 behavior.~
The~ inventors have o~served by light and 35~ ~ ~electron microscopy~that close vascularization occurs :~
, ~

WO93/!97~ PCT/US93/02665 ~ 1 3 ~ O 1 ~.~ ,"

if, in the initial period of implantation, at least some macrophages entering the material are not ~; activated. Activated macrophage are characterized by cell flattening.
The inventors o~serve close vascularization in regions of an implant where the macrophages that have entered the cavities of the ma~erial retain a r~unded appearance when viewed through light microscopy (:~ 400x). At 3000x (TEM) the rounded : ~ 10 ~ macrophage is observed to have substantially conformed to the contours of the material. Although there is a correlation with macrophage sXape, it is not clear : that macrophages control the observed response.
However, it is clear that invasion of the structure by 15: host cells is reguired~. ~Although the bulk of the cells appear`to~be~macrophages, it is possible that other inflammatory cells ~control the response, therefore the~inventors refer~:to the invading cells as `~ "infiammatory~:;cells," which include but are not 20~:~ limited~ta macropha~ges. ~
On~the other: hand, foreign body capsule formation ~occurs ~when,~ in~ the~ initial period of implantation~ inflammatory~cells in contact with the imp~lant mater~ial fl~tten against those portions of the 25~ mater}a1 ~which~ present:~ an~:area amenable to such flattening~behavior::by an inflammatory cell.
: The material for the second region 60 that res~lts in formation of close~ vascular structures is a~polymer;~membrane~having an average nominal pore size of ~approximately:~Q.6 to about 20 ~m, using ~ : conventional methods for determlnation of pore size in ;~ the~trade. Pref~erably, at least approximately 50% of ; the pores of:the~membrane have an average size of approximately 0.~ to about ~0 ~m.
The s~ructural elements which provide the ~:

~.", ~, ` ;"i ,~L~ . "~ , j,", ;, ~ ` ; " ~ " ~ ~, WO93/197~ PCT/US93/02~5 2 1 3 2 0 1 ., three dimensional conformation may include fibers, strands, glo~ules, cones or rods of amorphous or uniform geometry which are smooth or rough. These elements, referred to generally as "strands," have in general one dimension larger than the other two and the smaller dimensions do not exceed five microns.
In one arrangement, the material consists of strands that define "apertures" formed by a frame of the interconnected strands. The apertures have an average size of no more than about 20 ~m in any but the longest~dimension. The apertures of the material form a framework of interconnected apertures, defining "cavities" that are no greater than an average of about 20 ~m in any~but the longest dimension.
In this arrangement, the material for the second region~has at least some apertures having a sufficient~ si~ze to allow at least some vascular stxuctures to be ~created within the cavities. At least~some of~these apertures, while allowing vascular 20~ ~ structures~to~form~ within the cavities, prevent connective~tissue from forming therein because of size restrictions.~
Further details of the material are set orth~in copending~U.~S. Application Serial No. 735,401 ~entitled "Close~Vasculari~ation Implant Material"
- filed July~24, 1991, which is incorporated into this Specification by reference.
;~ Mankin~ ~ Bounda~
~ Figs~.~ll to 16 show a method of making a `~ 30 preferred embod~iment of the wall elements 32 and 34 that forms the~boundary. The method integrally joins material selected for the first region 5~ to another material selected for the second region 60. The two joined materials form the composite, or lami~ated, structure 72 shared by both wall elements 32 and 34.
~:~

2 ~ 3 o I PCT/US~/02~
-40- ' ~ .
The laminated structure 72 joins the interface 47 to the boundary 46.
In the illustrated embodiment, a porous PTFE
membrane material having a thickness of about 35 mi-crons and a~pore size of about .4 micron is selected for the first region 58. This material is commercial-ly available from Millipo;re Corporation under the t:radename BioporeTM.
The~porous material selected for the first '~ ;10 ~ region 58~has'~a~ thi~cknes~s of about 30 microns and an ultimate (tensile)~strength value of at least 3700 PSI, which is~we11 a~bove~the desired minimum value.
,~ ' The selected~ m~aterial~has pore~size of .35 microns, ~ which~blocks~the~passage~o inflammatory cells. The ;~ 15 selected materia1~has a~permeability value for insulin ,~ of~2.~6~x 104~cm/sec;~,and'~a~porosity value of greater than 70%~ ~The membrane there~fore meets the metabolic '"''~ transit value~requirements.~
It~s,hould~be~appreciated that other, compa-',,~ ,2'0~ rable~ materials~can~meet the~ stated requirements for the~`~firàt ~regio~ 58~ For ~example, polyethylene, 'polypropylene,~c~e1~lulose~aceta~te, cellulose nitrate, polyaarbonaté,;~po1yester,~nylon,~;ahd polysulfone mate-,i,~ "~ rials~ can~ be,~ used.~ Mixed~ esters of cellulose, 25~ `polyvinyl'ide~ne;,~ d;ifluoride, ~ silicone, and p1oyacry1Onitri~1e~can~a1so~be~used~.
In~ th~e~ lu'stra~ted;~embodiment, ,a membrane ,materlal~;~made~by~W.~L~. ~Gore~and~Asssciates (Elkton, Hary'land) undér'~ è~tradename Gore-TexTM is selected ~;,30~ for~the se~condl'~reg~ion ~60. The Gore-TexTM material ;~ ,compr~ises a~m'icroporou~s;~membrane,made from PTFE. The , `~ mèmbrane is~1~5~microns~thick a~nd has a pore size of 5 microns.~ Po1yest~er~strands~61;join the PTFE membrane ,~ to form a ba;ckin~ for it.~ The backing has a depth of 35~ about 1~20 microns~

I , ~

I ~ ~
I ~
:: :

WO93/19700 2 1 3 2 0 1~ PCT~VS93/02665 The Gore-TexTM material also has an ultimate strength value well above the desired minimum value.
The conformation of the polyester strands 61 also meets the criteria, set forth earliert for promoting the growth of neovascular structures.
.In Step 1 (see Figs. 10 and 11), the practitioner secures the edges of a strip of the Gore-Tex matexial ~second region 60) to a lamination slide 62, with the polyester backing 61 facing the slide 62.
In Step 2 (see Fig. 13), the practitioner places 2 or 3 lamination slides 62 side-by-side on a ~: work surface. Using a syringe 64, the practitioner applies cement or adhesive in continuous filaments 66 in a: back and forth pattern acxoss the lamination slides 62. The practitioner touches the syringe tip 64 to the work~surface at the end of each filament 66 to:begin a new filament 66.
Step~2 forms a criss-crossi~g pattern of cement~filaments 66 across the~secured strips of the second:region:~material, as Fig. 13 shows.
: The:~cement selected can vary. For example, the cemènt~can be~cellulose acetate or similar epoxy ;material. ~::In:~the:~illustrated embodiment, the cement comprises a~mixture of~Vynathene EY 90500 EVA resin and toluene~(made~by Mal:linckrodt).
In forming the EVA cement mixture, the pra-ctitioner~adds~ about 30 grams o~ resin and an equal amount:~of~toluene to a~bottle. The practitioner seals ~ the bottle to~àllow~the:resin to dissolve. The bottle :~ 30 may be periodically shaken to accelerate this process.
The relative amounts of resin and toluene may have~ to :be ;slightly~adjusted to arrive at the ri~ht consisten~y: for the cement. If the- cement is too thin to form continuous filaments when applied, ~ 35 use less toluene. :I~f the cement is to viscous to be :~

~t~t~
WO93/197~ P~TJUS93/02665 3 ~ 0 ~ 42~ s.
expressed from the syringe, use more toluene. Small changes i~ the amount of toluene added result is sig-nificant changes in the viscosity of the cement.
In Step 3 (as Fig. 14 shows), the practitioner places preformed strips of the BioporeTM
membrane material (first region 58) upon the cement filaments 66 applied in Step 2. In the illustrated embodiment, the practitioner precuts the BioporeTM
membrane material into disks having the diameter desired for the wall elements 32 and 34.
I~ Step 4 (as Fig. 15 shows), the ~; practitioner lays a strip of release material 68 (like Patapar) over the first region material S8 and covers the:layered structure with another lamination slide ;~ 15 70. The practitioner clamps the lamination slides 62 and 70 together, bringing the membrane layers into intimate contact.~
In Step 5, the practitioner places the clamped lami~nation slides 62 and 70 in an oven for about 5 to lO~ minutes at a temperature of about 80 :degrees C. The heat melts the EVA cP-ment.
Step 6:, the heated lamination slides 62 and~70 are~a~l~low~ed to cool ~o room temperature. Upon ~ cooling and~solidification, the filaments 66 secure,ly `~ 25 :~join~the BioporeTM membrane material to the Gore-TexTM
membrane material.~The:prac~itioner then unclamps the lamina~tion~slides~62 a~d 70 a~d removes t~e finished composite structure:72 (in strips).
In:~ Step 7 :(as Fig. 16 shows), the ~, 30 practitioner lays the composite structure 72 strips on a polypropylene cutting ~slab 74. The practitioner ; aligns a presized punch 76 over each precut disk, striking the punch with~a hammer. The practitioner ~ thereby frees the wall elements 32 or 34 formed of the ;~ 35 composite structure o~ the desired dimensions. Small :
~: ~

' W~93/197~ PCT/USg3/02665 ""'"' ' ~43~ ' 2 13201'j "
' scissors may be used to snip any adherent polyester strands not cut by the die.
Implant assemblies 10/10' are made using the wall elements in the manner previously described.
It should be appreciated that the first region material 58 can be applied to the second region material 60 by various aIternative means to form the laminated structure 72. For example, the first region material 58 can be extruded in place upon the second region material 60.

~ Assemblies like that shown in Figs. 1 to 4 ;~ and constructed according to the foregoing process have been successfully used to accomplish complete correction of diabetes in partially pancreatectomized and streptozotocin-treated rat hosts. The animals were correct;ed up to 293 days. Vpon removal vf the mplants, the~ animals reverted to a diabetic state.
Histology of the implants revealed the presence of 2 0 vascular structures close to the boundary.
Thes'e~assèmblies presented a boundary area of:about .77~cm2~. Each assembly sustained an initial cell~load~ of ~about ~OO~pancreatic islets (or about `~ 60~,000~pancr'éatic~cells).
2~5 ~When~ implanted, the as~emblies sustained cell~dens~ities~ôf~about ZOO~,OOo islets/cm3. These as-~ ; semblies,'mad~e~and used i~ accordance with the in~en-;~ tion,~ supported~8~times~ more pancreatic islets in a given volume~'than ';the CytoTherapeutics assemblies , 30 ~ (~aving cell qensities of only 25,000 islets/cm3).
Deriving~a Therapeuti~ Loadinq_Factor As earller described, one aspect of the in-vention provides the abili~y to identify a metabolic trans~it~ value as;soc~iated with a giYen cel-l type.
~ 35 Xnowing the re~uired metabolic transit value, in turn, .
.

~ WO93/197~ 2 1 3 2 0 1 ~ PCT/USg3/~2665 makes it possible to identify the clinically practical region of operation, where compact implant assemblies can sustain therapeutically large volumes of c~ells.
This aspect of the invention provides the methodology to deriYe and use a therapeutic loading factor (L) to characterize and predict the clinical effectiveness of a given implant assembly for a given cell type.
The therapeutic loading factor (L) takes into account the number of cells (N) that are required ; to be implanted to achieve the desired therapeutic ef-fect; the effective area (A) of the boundary between ~ the implanted cells and host that the host can be ; : : reasonably expected to tolerate; and the metabolic transit value (T~ needed to sustain cell viability.
~; The tberapeutic loading factor for a given implant assembly and given implanted cell type can be expressed~as:follows::
c = (A/Nc) * Tmin ;20~: ~ where c is the given:cell typ , Lc is the therapeutic loading actor for :~ :the given cell type~, : A is~ the area~of boundary between the 2~5 ~ implanted~cells and;~the host offered by the given ;~ : implant assembly, : ~
Nc is~the number of cells supported by the boundary area ~A), and~ ~
, ., ~
~ Tmjn is the minimum metabolic transit :~ , 30 value that will support cell survival during the isch-~ ~ emic perlod, deter~ined according the methodology set `~ forth in Example:l.
}f the practitioner selects boundaries having a porosity value of greater than 15%, then the : 35 ~permeability value ~P) alone can be used to express : :
~ .

WO93/19700 PCT/USg3/02665 ,,~ 45- 213201 , the metabolic transit value (T~. The therapeutic load factor can then be expressed:
c = (A/Nc) * Pmin where Pmjn is the minimum permeability value that will support cell survival during the ischemic period.
In the assemblies described in Example 3, the observed ratio between the boundary area and the number of implanted cells tA/NC) for the successful im-I0 ~ plantation of pancreatic cells was 128 ~m2/pancreatic cell~ The inventors:believe that a somewhat larger : ratio of about 150 ~m2/pancreatic cell will provide a satisfactory ~;m~argin: for variance among different hosts~
~ As earlier discussed, given a boundary porosity ra~lue~of greater than 15%, ~ permeability value~ tP)~gre~ater;than about 1.5 x lO~ cm/sec for nsulin should~:be provided a metabolic transit value that~will sustain cell survival during the ischemic 20 ~ ~ period~and afterward.~ :
.~ Flg.~ 22;~:shows~the therapeutic loading curve for:pancreatic~cells generated based upon the above considerations~ The~ curve displays the predicted ~ re~ion~of:::c~ell;~survival~in~terms~of~the boundary area-;~ 25~ : to-cell number~ratio: A/N ~(x-coord:inate) and per-meabil~ity;::~va~lue`:~P~:(y-:coordinate) (given a porosity : value~ o f greater~than about 1~5%).
e~:: Fig~ 22~predicts that as-emblies operating to the righ:t of the therapeutic loading curve will -~ 30 ~ sustain implanted:pancreatic cells. Fig. 22 pxedicts that assemblies~ope~ating to the le~t o~ the therapeu-tic l:oading curve will:not:.
~ : The~inventors believe that a human diabetic ;:~ will require~the transplan~ation of about 250,000 :~

`: :~

W0 93/19700 PC~/US93/02665 n ~ V l ~ -46 pancreat-c islets (or about 250 million pancreatic cells) to derive a therapeutic benefit. With this in mind, one can calculate a range of sizes for an implant assembly based upon the A/N ratio.
The equation for calculating the side dimension tL) in cm of a square implant assembly based upon the A/N ratio is as follows:

(250,000~1000) -: L ~ 2--- N * 1o-8 :where: the factor 10-8 converts micron2 to ~cm2.~
The:equation for calc~ulating the diameter (D) in cm~o~a round:implant assembly based upon the A/N~:ratio is~as follows~

2(250,000~1000) A
D~ N ~ l o-8 `.,~J:~ : where~ the::factor:108 converts micron2 to abl;e~3~ l}sts~a~;range of L~s and D's ;t di~f:ferént A/N~:~ratios~ for~an lmplant assembly~ho~ding ` ~ 250,~000~pa~nc~eat:ic:~islets~
L~ A ~f~cm2~/side~ LLml D(cm~
28~ 160::~ 12.6 14.3 ~. ~
150~ ~ ~ 188 13.7 15.5 ~ 200 ~ :2~50 ::~ 15.8 17.8 ii~ 328~ 4~10 ~ 20.:2 22.8 -~ 25 ~ 46~3 :~ 579~ 24.0 24.1 ~ Based upon~ ~the~ foregolng considerations, the "'`~
,::~
,~

W093~197~ PCT/US93/02665 ~47- 2 1 32 O1~j . '' invent~rs believe that A/N ratios less than about 200 ~m2/pancreatic cell define the operating reg~on of implant asse~blies that offer compart, clinically practical implant boundary areas. Fig. 22 shows this preferred region.
As Fig. 22 a~so shows, a practitioner can provide ; an implant assembly that combines the benefits of com-pact size with the ability to sustain the requisite therapeutical number of cells, by selecting a perme-ability v~lue for the boundary that achieves a reqion of operation to the right of the therapeutic loading curve. The practitioner also selects the prescribed pore size and ultimate physical strength determined ln accordance with the invention.
Fig. 22 shows that the prior art hollow fiber implant assembly made by CytoTherapeutics (described in the Background section of this Specification) falls well outside the preferred region of clinically practical operation. This assembly offers an A/N
ratio of about 328 ~m2/pancreatic cell, about 1.5 ~times the A/N ratio of the ~invention.
Fig. 22 also shows a prior art hollow fiber implant assembly made by W.R. Grace and Co.
Lexin~ton, Ma~ as~reported by Proc._Natl. Acad Sci.
U.S.A., Vol. 88,~ ~pp.~11100-11104 (December 1991).
Each~hollow f~iber~had a length of 2-3 cm, and an inside diameter~of 0.177 cm. There were 200 to 40n pancreatic islets loaded into each fiber for implanation. Ta~ing an average length of 2.5 cm ~nd an average celliload of 300 islets, the associated A/N
ratio is 46~3,~more than twice the A/N ratio of the `~ invention.
The foregoing establishes a methodology to deri~e and use a therapeutic loa~ing factor (L) for pancreatic islets. This methodology can be followed ~ .
~ .
~:

~;. WOg3/19700 PCT/US93/02665 213~Ql!~ 48-. ,.
to identify a therapeutic loading factor for other cell types and other ranges of metabolic transit . values~ The absolute value of the therapeutic loading factor derived will of course depend upon the type of cell and the methodologies used to determine permeability and porosity. Different conditions will give different absolute values for the therapeutic loadinq factor.
Still, regardless of the test conditions, the ~:: 10 relative differences in the A/N ratios, permeability values, and porosity values derived under constant, stated conditions~ will serve as a means to ~:~ characterize and predict the clinical effectiveness of a given implant assembly for a given cell type.
The following claims further define the features ~ and benefits of the invention.

;'~
,, ~ ~

~:;
. ~ ~
~:

~:
:::
,

Claims (19)

WE CLAIM:

1. An implant assembly for a host tissue comprising wall means defining a chamber for holding cells for implantation, the wall means including means for forming a porous boundary between the host tissue and the implanted cells in the chamber, the porous boundary being characterized by an ultimate strength value sufficient to withstand physiological stresses and vascularization of the host tissue close to the boundary without rupture;
a pore size sufficient to isolate the implanted cells from the immune response of the host tissue; and a metabolic transit value that sustains a flux of nutrients from the host tissue to the implanted cells and waste products from the implanted cells to the host tissue in the absence of close vascular structures sufficient to sustain the viability of the implant tissue cells during the ischemic period of implantation.

2. An assembly according to claim 1 and further including an angiogenic material that, when implanted, stimulates the growth of vascular structures by the host tissue close to the boundary to bring an end to the ischemic period.

3. An assembly according to claim 1 or 2 and further including a second material that overlies the porous material at the boundary and forms an interface with the host tissue, the second material having a structural conformation that, when implanted, supports the growth of vascular structures by the host close to the boundary.

4. An assembly according to claim 1 or 2 wherein the pore size blocks the passage of molecular immunogenic factors from the host to the implanted cells.

5. An assembly according to claim 1 or 2 wherein the pore size blocks the passage of inflammatory cells from the host to the implanted cells.

6. An assembly according to claim 1 or 2 wherein the metabolic transit value includes a permeability value for insulin that is greater than about 1.5 x 104 cm/sec.

7. An assembly according to claim 6 wherein the metabolic transit value includes a porosity value that is greater than about 15%.

8. An assembly according to claim 1 or 2 wherein the ratio between the area of the boundary and the number of implanted cells is less than about 200 µm2/cell.

9. An assembly according to claim 1 or 2 wherein the ultimate strength value is greater than about 100 pounds per square inch (PSI).

10. An assembly according to claim 1 or 2 implanted within host tissue.

11. An implant assembly for a host tissue comprising:
wall means defining a chamber for holding implanted cells, the wall means including means for forming a porous boundary between the host tissue and the implanted cells in the chamber, the porous boundary being characterized by an ultimate strength value sufficient to withstand physiological stresses and vascularization close to the host tissue without rupture;
a pore size sufficient to isolate the implanted cells from the immune response of the host;
a porosity value greater than about 15%;
a permeability value to insulin greater than about than about 1.5 x 10-4 cm/sec.; and a ratio between the area of the boundary and the number of implanted cells that is less than about 200 µm2/cell.

12. An assembly according to claim 11 and further including an angiogenic material that, when implanted, stimulates the growth of vascular structures by the host tissue close to the boundary to bring an end to the ischemic period.

13. An assembly according to claim 11 and further including a second material that overlies the porous material at the boundary and forms an interface with the host tissue, the second material having a structural conformation that, when implanted, supports the growth of vascular structures by the host close to the boundary.

14. An assembly according to claim 11 wherein the pore size blocks the passage of molecular immunogenic factors from the host to the implanted cells.

15. An assembly according to claim 11 wherein the pore size blocks the passage of inflammatory cells from the host to the implanted cells.

16. An assembly according to claim 11 implanted in host tissue.

17. A method of implanting cells comprising the steps of surrounding at least a portion of the implanted cells with a porous boundary that has an ultimate strength value sufficient to withstand physiological stresses and vascularization close to the host tissue without rupture; a pore size sufficient to isolate the implanted cells from the immune response of the host; and a metabolic transit value that sustains a flux of nutrients from the host tissue to the implanted cells and waste products from the implanted cells to the host tissue in the absence of close vascular structures sufficient to sustain the viability of the implant tissue cells during the ischemic period of implantation, and implanting the porous boundary within the host tissue so that nutrients pass from the host tissue to the implanted cells sufficient to maintain their viability while vascular structures of the host tissue form close to the boundary.

18. A method according to claim 17 and further including the step of exposing the boundary to an angiogenic substance that stimulates the formation of vascular structures by the host close to the boundary.

19. A method according to claim 17 or 18 and further including the step of, before implanting the boundary, overlaying on the boundary a second material that has a structural conformation that supports the growth of vascular structures by the host tissue close to the boundary.