AU700766B2 - Methods employing microporous macrocapsules - Google Patents

Description

7

AUSTRALIA

PATENTS ACT 1990

ORIGINAL

COMPLETE DIVISIONAL SPECIFICATION 40 A, p 4.*A p.

4 6.

Name of Applicant: Address of Applicant: Actual Inventors: Brown University Research Foundation 42 Charlesfield Street, Providence, Rhode Island 02906, United States of America Frank T. GENTILE; Patrick A. TRESCO; Tyrone HAZLETT; Thomas FLANAGAN; Edward J. DOHERTY; David REIN; and Laura HOLLAND 06.0 4*

A

S p PSt P nov Address for Service: DAVIES COLLISON CAVE, Patent Attorneys, 1 Little Collins Street, Melbourne, 3000.

Complete Specification for the invention entitled: METHODS EMPLOYING MICROPOROUS MACROCAPSULES The following statement is a full description of this invention, inctuding the best method of performing it known to us: -1r~ i -12- -1 P:\OPER\MRO\687371.DIV 8/15198

-IA-

METHODS EMPLOYING MICROPOROUS MACROCAPSULES This application is a divisional application of Australian Patent No. 687371, which corresponds to International Patent Application No. PCT/US93/11232, the contents of which are incorporated herein by way of reference. Throughout this specification and the claims that follow, unless the context requires otherwise, the word "comprise", or variations such as "comprises" or "comprising" will be understood to imply the inclusion of a stated element or step or integer or group of elements or steps or integers but not the exclusion of any other element or step or integer or group of elements or steps or integers.

BACKGROUND TO THE INVENTION The present invention is directed to methods of determining the viral retentivity of a So macroporous macrocapsule or external hollow fibre jacket thereon and, more particularly, the viral retentivity of an implantable permselective macrocapsule or external hollow fibre jacket thereon, by measuring the escape of a microorganism of known physical characteristics, such as known size and shape, from said macrocapsule or external hollow S: fibre jacket. Preferably, the external hollow fibre jacket tested may be that which is described in Australian Patent No. 687371. The inventive method is particularly useful in determining whether any particular macrocapsule or external hollow fibre jacket thereon is capable of preventing the escape of a virus particle from the inside of the jacket to the outside, such as, but not limited to, a virus particle which may be present in an encapsulated cell. The advantages of the present invention are of considerable utility when the immunoisolatory capabilities of a macrocapsule or external hollow fibre jacket are assessed.

The methods of implanting living cells and/or tissue into the body to provide therapeutically useful substances cell therapy) is currently the focus of a number of research efforts.

Among others, therapeutic diabetes and neural degenerative diseases such as Alzheimer's Disease, Parkinson's Disease and epilepsy. Additionally, cells have also been shown to have great therapeutic potential for the removal of detrimental substances from the body. For example, hepatocytes have been implanted for the treatment of high cholesterol levels as shown by Wang et al Transplantatin Proceedings. 23:894-895 (1991).

-2- In one form of cell therapy, the cells that are implanted into the patient have been genetically modified (transduced) in vitro with exogenous genetic material so as to enable the cells to produce a desired biological substance that is useful as a therapeutic agent. There are variety of mechanisms for transducing cells. Retroviral vectors have been of interest due to the ability of some vectors to transduce a very high percentage of target cells. Replication of target cells is necessary for proviral integration to occur using these vectors. In theory, cells transduced with retroviruses are not able to spread virus to other cells.

However, several safety issues surround the use of these systems.

Risks include the potential during production of the retroviral vector .preparations for contamination with-infectious viruses or pathogens (Cornetta, Human Gene Therapy Vol 2:5-14. 1991).

1 Adenoviruses are also used as gene transfer vectors. They are double stranded DNA viruses capable of infecting post-mitotic cells, and their successful transfection of host cells can result in the 'i expression of large amounts of gene product. Adenoviruses are a minor pathogen in humans and are not normally associated with 5 malignancies. While the adenovirus vector generally remains episomal and does not undergo replication, studies have shown that gene expression can persist for a significant period, opening up the :oo possibility that the vectors are at least replication-competent.

While present cell therapy methods show great therapeutic potential, they have several limitations. The injection of cells into a patient can create a severe immune reaction. An immune reaction to a first injection of cells will most likely preclude a second injection, thus limiting the benefit which can be gained from such treatment. Some candidate cell types may shed virus particles which could be

I

I

i 0O I

A

detrimental to the host and therefore may not be considered suitable for implantation. In the case of cells that have been genetically modified to produce a desired biological substance using viral techniques, the cells may harbor viral particles, thus allowing the possibility of infecting a patient's cells. Another disadvantage with the current cell therapy methods is that many cells that might be suitable for such therapies are known to migrate in situ, glial cells). The inability to retain implanted cells in a fixed location may make them unsuitable as therapeutic agents. Likewise, cells which are autologous or even allogeneic to the host may continue to divide unchecked and produce tumors.

Current methods of cell therapy do not readily allow termination o adjustments to, the cell therapy protocol once the cells are implanted.

This is because cells implanted into a patient's body are not well isolated from the patient's own tissue and thus cannot be readily retrieved or manipulated. This creates a real fear when the cells have been genetically modified using retroviral particles because implanted cells could migrate in situ and potentially result in proviral integration into the host germ line cells, thus passing the provirus onto offspring.

There are a number of specific circumstances where it may be critical to terminate or remove implanted therapeutic cells. These include: 1. One or more of the implanted cells has become oncogenic or tumor forming, or has induced an adverse immune reaction.

2. A dose adjustment may require a reduction in the number of implanted cells at specific times.

3. Genetically engineered cell populations may have a small number of cells with improper incorporation of genetic material leading to detrimental effects.

0* 99 o 99 9 9 -4- 4. The therapy may have a defined end point growth hormone treatment) at which the point the continued presence of the therapeutic cells is unwanted or detrimental.

Genetically modified cells may develop an adverse response to concurrently administered pharmacological agents.

6. The development of an improved therapeutic option may warrant a change in therapy requiring removal of the transplanted cells.

Incorporation of inducible suicide genes into cells for implantation has IQ been suggested as a potential means to terminate cell therapy i.e. by killing the implanted cells. One disadvantage of this approach, however, is that since cell division is a stochastic process, there is always a chance that the suicide mechanism in one or a few cells might become inactivated. Undesirable effects could be produced even if only one of the cells continued to divide. This approach also may have undesirable results caused by the dumping of secretory products such as growth factors or proteases or the release of viruses, etc., due to the large scale destruction of implanted cells.

Cell encapsulation methods have been used to isolate cells while allowing the release of desired biological materials. Two techniques have been used, microencapsulation and macroencapsulation.

Typically, in microencapsulation, the cells are sequestered in a small permselective spherical container, whereas in macroencapsulation the cells are entrapped in a larger non-spherical membrane.

Lim, US Patent Nos. 4,409,331 and 4,352,883, discloses the use of microencapsulation methods to produce biological materials generated by ce~lls in vitro, wherein the capsules have varying permeabilities depending upon tne biological materials of interest being produced.

Wu et al, Int. J. Pancreatology, 3:91-100 (1988), disclose the transplantation of insulin-producing, microencapsulated pancreatic islets into diabetic rats. Aebischer et al., Biomaterials, 12:50-55 (1991), disclose the macroencapsulation of dopamine-secreting cells.

Various polymeric materials have been used in the art for filtering viruses from liquids. Anazawa et al. patent no. 5,236,588) teach an ultrafiltration membrane prepared by irradiating a monomer to produce a membrane having communicating pores. Allegrezza U.S.

10 patent no. 5,096,637) reports an asymmetric composite ultrafiltration membrane for isolating viruses from protein containing solutions.

DiLeo patent no. 5,017,292) teaches asymmetric skinned membranes comprising a porous substrate, an ultrafiltration surface S. skin and an intermediate zone free of voids which form a break in the 15 skin and which cause fluid to communicate directly with the porous substrate. The membranes are used for isolating viruses from protein containing solutions. U.S. patents nos. 4,808,315 and 4,857,196 teach a hydrophilic hollow fiber membrane for removing virus from a protein containing solution.

None of the above references describe virally retentive, permselective, biocompatible membranes that can be implanted into a patient for celltherapy purposes. There are many situations where it would be useful to provide cell therapy for the in vivo delivery of biological materials, while substantially preventing release of detrimental viruses that may be shed from the encapsulated cells. There are also many situations where easy termination of cell therapy treatment would be desirable.

However, effective means for such treatments are lacking in the art of cell therapy.

Oalr rRliM~ P:\OPER\MRO\687371.DIV 815/98 Those skilled in the relevant art will be aware that the ultrafiltration characteristics of j microcapsule membranes are limited by the nature and amount of the substance whi, i is capable of crossing the encapsulating membrane. In this regard, the inventive microporous macrocapsules described herein combine the advantages of holding greater amounts of therapeutic substances or secretory cells therein whilst exhibiting a selectively permeable membrane capable of permitting the passage of greater amounts of therapeutic substances, Sbased on a larger nominal molecular weight exclusion level than is possible using the ultrafiltration membranes known previously.

S 10 The selectively permeable external jacket may be made such as to permit the passage of larger Smolecules than is possible using ultrafiltration membranes, whilst still effectively prohibiting the passage of very large molecules or organisms, such as virus particles. The need to adjust the selective permeability of macrocapsular jackets for different therapeutic applications, whilst retaining the ability of said jacket to retain a virus particle that might be present in the encapsulated cells, has produced the concomitant need for a means of testing a macrocapsular jacket for the ability to prevent the escape of virus particles from the inside of the jacket to the outside.

In this regard, as the selective permeability of a macrocapsule is significantly different from the selective permeability of a microcapsule utilising an ultrafiltration membrane, the technical problem of testing macrocapsular jackets was not a problem associated with ultrafiltration membranes.

P:\OPER\MRO\687173A.DIV 8/5/98 -6- SUMMARY OF THE INVENTION One aspect of the present invention provides a method of determining the viral retentivity of an external jacket of an implantable permselective macrocapsule, said jacket comprising a hollow fiber having a top and bottom end, said method comprising selecting a bacteriophage having dimensions similar to a virus for which said external jacket is being tested for retention, loading said bacteriophage into said hollow fiberjacket, sealing said top and bottom ends, placing said jacket into a bath solution, inoculating a lawn of bacteria susceptible to said bacteriophage with said bath solution, and calculating viral retentivity based on the number of plaques formed on said lawn of bacteria.

o In an alternative embodiment, the present invention provides a method of determining the viral retentivity of an external hollow fibre jacket on an implantable permselective macrocapsule, said method comprising at least the steps of: selecting a bacteriophage having similar physical characteristics to said virus; S° 15 (ii) loading said bacteriophage into said hollow fibre jacket and sealing said jacket; (iii) placing said jacket into a bath solution; (iv) inoculating a lawn of bacteria capable of being infected by said bacteriophage with said bath solution; and calculating viral retentivity based on the number of plaques formed on said 20 lawn of bacteria.

According to this embodiment, the physical characteristics of the virus may be any physical characteristic which contributes to, or is related to, the ability of the virus to pass through a permeably-selective membrane, such as, but not limited to, any one or more of those characteristics selected from the list comprising size, shape, flexibility, molecular weight, Stokes rad 4 us and diffusivity in water.

The bacteriophage used in performance of the inventive method may be any bacteriophage having known physical characteristics making it suited for assessing the therapeutic efficacy of a particular permselective macrocapsule or external jacket thereof, such as, but not limited P:\OPER\MRO\687173A.DIV -8/5/98 -7- A to bacteriophage selected from the list comprising M13, phi X, phi X174 (X174) and lambda gtlO (XgtlO), amongst others. Wherein the bacteriophage is ,gtlO or a derivative or equivalent thereof having substantially the same physical characteristics as Xgtl0, and the bacteria are E.coli, the calculated retentivity reflects the permeability of the permselective macrocapsule to one or more animal viruses. Wherein the bacteriophage is 4X174 or a derivative or equivalent thereof having substantially the same physical characteristics as 4X174, the calculated retentivity reflects the permeability of said microporous macrocapsule to a parvovirus or rhinovirus.

Preferably, the step of placing the jacket into the bath solution is performed for a time and under conditions sufficient for any bacteriophage particles excluded by said hollow fibre jacket to move into said bath solution.

In a particularly preferred embodiment, the implantable permselective macrocapsule is one S 15 which is described in Australian Patent No. 687371 which is incorporated herein by way of reference and more preferably, an implantable permselective macrocapsule which is defined "in the claims of Australian Patent No. 687371 which is incorporated herein by way of reference.

Li..

P:\OPER\MRO\687173A.DIV 815198 t-y i~" S-7Ai: BRIEF DESCRIPTION OF THE DRAWINGS t Figure 1 shows the log-order reouction in release of phage from thew microporous macrocapsules.

Figure 2 shows the testing cartridge used to measure the viral retentivity of the microporous macrocapsules.

i Figures 3 and 4 show the percent rejection of various sized dextrans from the microporous macrocapsules.

SD ETAILED DESCRIPTION OF THE INVENTION °I In one embodiment, the instant invention provides permselective, biocompatible macroencapsulation devices and methods of using same for restraining cells, retaining certain detrimental viruses shed from restrained cells or preventing entry of detrimental viruses from host tissue through the encapsulation device. The macrocapsules also allow for the efficient termination of cell therapy regimes.

The invention further provides, in another aspect, a method of determining the viral retentivity of an external jacket of an implantable permselective macrocapsule which me be intended for any particular therapeutic application.

The inventive methods all rely on permselective macroencapsulation devices which, depending upon the nature cf the cells to be implanted and the site of implantation, may or

JA

-8may not have immunoisolatory characteristics. Thus in some Sinstances, the cells are capable of producing a desired function without the need of a physical barrier between themselves and the immune |system of the recipient. Such cells are many of the same cells believed useful for classic unencapsulated cell therapy.

The macrocapsules of the present invention are biocompatible. The term "biocompatible" as used herein refers collectively to both the intact macrocapsule and its contents. Specifically, it refers to the capability of the implanted intact macrocapsule and its contents to avoid detrimental effects of the body's various protective systems, such as the immune system or foreign body fibrotic response, and remain functional for a significant period of time. In addition, ."biocompatible" also implies that no specific undesirable cytotoxic or Ssystemic effects are caused by the vehicle and its contents such as would interfere with the desired functioning of the vehicle or its contents.

The macrocapsules are also "permselective", and they can have their permeability characteristics tailored for their specific usage. For '4 example, macrocapsuies that are used to encapsulate xenogene, cells, which may harbor detrimental viruses, may be designed to minimize the release of the viruses. The methods of the present invention allow one to know prior to the implantation of a particular macrocapsule, its viral retentivity/retardatior' characteristics. Viral retentivity (how much the membrane retards viral transport) can be determined by calculating where D is the diffusion coefficient of the substance through the membrane. This compares the resistance of transport through the membrane with that of an equivalent distance of water. For"viruses of the 18 120 nm sizes, -9- Dwae, is 2.4 x 10 7 cm 2 /sec to 3.8 x 10' 8 cm 2 /sec. The diffusion coefficients of various viruses through membranes were measured based upon the data initially and at six weeks. The diffusion coefficient is a measure of resistance to transport through the membrane and is given by the equation: J D(OC)/(OX) where J is the flux through the membrane, 0C is the concentration difference between the inside and the outside of the membrane, and OX is the membrane wall thickness. D is a constant and is .0 independent of concentration. It is a proportionality constant between OC and the resulting flux, J. The viral retentivity of a membrane is preferably less than about 0.5, more.preferably less than 0.1, and "most preferably less than about 10' 2 Preferably the membrane, once implanted into the patient, will maintain its viral retentivity/retardation 15 properties at least 1 week, preferably at least 2 weeks, and more preferably at least 6 weeks, It is most preferable for the membrane to retain a given viral retentivity/retardation value for the duration of the therapy.

r The permselective characteristic of a macrocapsule can also be 20 defined in terms of its molecular weight cutoff. The term "molecular weight cutoff" (MWCO) refers to the size or molecular weight of particles which are retained by a particular membrane under convective or pressure driven conditions. Since semipermeable membranes generally have a Gaussian distribution of pore sizes the term MWCO as used herein refers to the size of a marker molecule when tested with a specific membrane, 90% of which is retained by the membrane under the test condition. Specifically, while the majority of pores in a membrane may be large enough to retain a given particle, some pores may exist which are large enough for the molecule to pass through. Thus, the MWCO of a given membrane is not an absolute and may differ somewhat when measurea using different tests. While particles of a particular size, if large enough, may be completely retained for a given period of time, there is a range of particle sizes where a portion of particles will be retained and a portion will be released. If, for example, 60% of particles having a molecular weight of about 50 kD are retained by a given membrane, then a greater percentage of particles that are larger than 50 kD will be retained by the same membrane given the same time period. The term "nominal MWCO" (nMWCO) refers to the size of particles to retained by the macrocapsule at 90% levels. Thus, for example, if 100 of the macrocapsule is 100 kD.

There are several different methods for measuring nMWCO, which can 15 give different results. Tests may be diffusive or convective.

Convective methods measure ,ransmembrane diffusion under .pressure. The convective nMWCO is a good indicator of how quickly S. molecules will passively diffuse from the membrane. When testing nMWCO, it is important to realize that the shape and nature of the I 20 molecule used in testing may influence the measured values. Thus, Swhen comparing data from various sources, it is important to consider the test method employed to obtain the values.

The macrocapsules can be designed to retain cells that have been genetically modified with viruses to secrete a desired biological substance that is useful as a therapeutic agent. The nMWCO is, selected so that the biological substance is substantially released from the capsule, but residual viral particles, which may be harbored by the modified cells, is substantially retained. The phrase "genetically -11- modified with virus" includes any standard method used in the art to introduce exogenous DNA by viral vectors (retrovirus, modified herpes viral, herpes-viral, adenovirus, adeno-associated virus, and the like) to the cells that are to be encapsulated.

In some embodiments, the macrocapsules of the invention will be permeable to substances of greater molecular weight than the Clq component of complement which has a molecular weight of about 440 kD. Thus, these embodiments will be substantially nonimmunoisolatory. Such capsules will generally be implanted at 0 10 immuno-privileged sites, such as the brain. The capsules have the primary purpose of serving as cell retrieval devices. The relatively high MWCOs of such devices will enable relatively large therapeutically useful agents to be released from the cells and delivered to the site of implantation. Macrocapsules that serve 15 primarily as cell retrieval devices may have MWCOs in the 1,000 3,000 kD range. The relatively high molecular weight cutoffs (MWCOs) of the macrocapsules can be achieved by methods standard in the art including those taught by Strathmann in "Production of Microporous Media by Phase Inversion Processes," Materials Science 20 of Synthetic Membranes, ACS Symposium Series 269, American Chemical Society, Washington D.C. (1985). Additionally, various methods for altering the MWCO of the macrocapsules are described below.

The permselective macrocapsules can be manufactured by the same methods and using the same materials disclosed in copending PCT ii application US92/03327. Briefly, the capsule is comprised of a core which contains isolated cells, either suspended in a liquid medium or immobilized within a hydrogel matrix, and a biocompatible

YC

o a o a .4 K Pr -12surrounding or peripheral region ("jacket") of permselective matrix or membrane which is substantially free of the isolated cells.

The core of the macrocapsule is constructed to provide a suitable local environment for the particular cells isolated therein. In some embodiments, the core comprises a liquid medium sufficient to maintain the cells. Liquid cores are particularly suitable for maintaining transformed cells, such as PC12 cells. In other embodiments, the core comprises a hydrogel matrix which immobilizes and distributes the cells, thereby reducing the formation of dense cellular agglomerations.

Cores made of a hydrogel matrix are particularly suitable for maintaining primary cells which tend to form agglomerates, such as the cells in islets of Langerhans, or adrenal chromaffin cells.

Optionally, the core of the. instant vehicle can contain substances 15 which support or promote the function of the isolated cells. These substances include natural or synthetic nutrient sources, extracellular matrix (ECM) components, growth factors or growth regulatory substances, or a population of feeder or accessory cells.

The jacket of the macrocapsule is made of a material which may be the same as that of the core or may be different. In either case, the material used results in a surrounding or peripheral region which is permselective and biocompatible. The biocompatibility of the jacket means that it does not elicit a detrimental immune response sufficient to result in rejection of the implanted macrocapsule or to render it inoperable. It also means that the jacket does not elicit unfavorable tissue responses. In addition, the external surface can be selected or designed in such a manner that it is particularly suitable for -13implantation at the selected site. For example, the external surface can be smooth, stippled or rough, depending on whether attachment by cells of the surrounding tissue is desirable. The shape or configuration can also be selected or designed to be particularly appropriate for the implantation site chosen. Typically the jacket will comprise permse.lective membranes which have been formed into hollow fibers or flat sheets. A hollow fiber membrane is an annulus consisting of an inner cylindrical surface, a wall structure for support, and an outer cylinder surface. One or both of the surfaces can be .1.0 selective for molecules of varied molecular weight. A flatsheet is a S" planar composition of a hollow fiber.

The surrounding or peripheral region of the jacket can be made of a hydrogel matrix or of a different material, such as a thermoplastic membrane or hollow fiber. It can also be made of a matrix-membrane composite, such that a permselective thermoplastic membrane having matrix-filled pores, such as hydrogel filled pores, is formed.

Suitably, the external jacket may be formed of a thermoplastic material known to be biocompatible, such as the ones described herein. In addition, other jackets which have been used in the microcapsule field may also be used herein, such as alginate, suitably cross-linked with a multivalent ion such as calcium.

The jacket may be directly cross-linked to the core matrix, eliminating the need for an intermediate linking layer such as poly-L-lysine (PLL).

The elimination of an intermediate PLL layer is advantageous in that PLL is believed to be fibrogenic. Also, the jacket in this device can be controlled for permselectivity better than those made with PLL.

'1 \i -14- The macrocapsule can be formed by coextrusion or in a step-wise fashion. Techniques for coextrusion, which can be used to form the macrocapsules of the present invention, are taught in U.S. Patent No.

5,158,881. With the coextrusion method, the macrocapsules are formed by coextruding, from a nested-bore extrusion nozzle, materials which form the core and surrounding or peripheral regions, under conditions sufficient to gel, harden, or cast the matrix or membrane precursor(s) of the surrounding or peripheral region (and of the core region). A particular advantage of this coextrusion embodiment is that the cells in the core are isolated from the moment of formation of the S vehicle, ensuring that the core materi:.ls do not become contaminated or adulterated during handling of the vehicle prior to implantation. A :i 1 further advantage of the coextrusion process is that it ensures that the surrounding or peripheral region is free of cells and other core materials.

The macrocapsule can also be formed step-wise. For example, if the capsule being made includes a hydrogel core containing the isolated cells,. the core can be formed initially, and the surrounding or peripheral matrix or membrane can be assembled or applied subsequently.' Conversely, the surrounding or peripheral matrix or membrane can be preformed, and then filled with the preformed isolated-cell containing core material or with materials which will form the core core precursor materials). The capsule is sealed in such a manner that the core materials are completely enclosed. If a core precursor material is used, the vehicle is then exposed to conditions which result in formation of the core.

The surrounding or peripheral region is produced in such a manner' that it has pores or voids of a predetermined range of sizes; as a result, the vehicle is permselective. The permeability and biocompatibility characteristics of the surrounding or peripheral region are determined by both the matrix or membrane precursor materials used, and the conditions under which the matrix or membrane is I 5 formed.

The molecular weight cutoff (MWCO) selected for a particular vehicle will be determined by the molecular size of the largest substance to be allowed to pass into and/or out of the vehicle. In some instances, where viral release is not a concern, the pores of the macrocapsule 3.0 will be as large as possible while still retaining cells, about from 0.1 10 pm, depending on the type of cell encapsulated.

Hollow fiber membranes can be prepared in a wide range of MWCOs for the purposes of exclusion or delivery of various biological agents while preventing significant passage either into or out of the capsule of 15 detrimental viruses. Alteration of the MWCO of the macrocapsule can be accomplished by a variety of methods. The basic methodology of alteration of membrane cutoff is to alter the phase separation properties of a given set of membrane casting solutions and coagulation baths. The thermodynamics of phase separation for each 20 material/solvent combination is different. In general, the longer it takes for the phase inversion process to take place, the larger the pore-size of the resulting membrane. The particular membrane material selected will affect the MWCO. Material/solvent systems such as poly(sulfone)/N-methyl pyrrolidone (NMP) and poly(vinylidene diflucride) (PVDF)/tetrahydrofuran (THF) are more easily made into higher MWCO membranes, due to the strength of the polymer/solventnonsolvent interactions, than materials such as poly(acrylonitrile) (PAN)/dimethyl sulfoxide (DMSO). Permeability properties of membranes can be selected by choosing appropriate -16material systems. For example, using a system where the polymer is somewhat soluble in the precipitating liquid will greatly slow down the phase inversion process.

The addition of a low molecular weight additive to the membrane casting solution in which the main membrane polymer is not soluble will also affect the MWCO of the membrane (Cabasso, 1976). Low molecular weight, non-soluble additives make the casting solution less stable. Such solutions can phase invert much faster resulting in potential lower MWCO membranes. Conversely, higher molecular weight additives can result in slower phase separation which will cause 0000. relatively large pores to form. An example of this is when .d a d) poly(acrylonitrile-co-vinyl chloride) (PAN/PVC) is blended with high molecular weight polyethylene oxide (PEO) (MW 100 kD). Also, higher molecular weight or macroscopic additives will perhaps diffuse ^15 away from the membrane matrix and leave behind large pores.

The resulting molecular weight of a membrane can also be altered by the addition of a membrane polymer solvent to both the inner bore and /outer coagulation solutions. Addition of solvent greatly slow- down the .precipitation, which can result in much higher cutoff membranes.

o Changing the temperature during the membrane formation process can also affect the MWCO. This method of altering MWCO is usually used in combination with solvent or non-solvent additions because temperature change alone does not always alter MWCO. For example, solutions of polysulfone, PEG-200 and (NMP) can be made into fibers of differing MWCOs by adjusting the temperature of the coagulation bath. This system combination is termed a "lower critical solution temperature system" (LCST) in which the stability of the V3.

-17solution goes down with increasing temperature. Thus, higher temperatures give rise to higher MWCOs. Changing the temperature of the casting and the coagulation solutions during the course of membrane formation will also alter the MWCO. Combinations of any of the above techniques can also be used to alter MWCO. For example, temperature change in combination with the addition of a high or low molecular weight additive can be used to affect MWCO.

Because of the moderate to high MWCO characteristics of some .10 embodiments of the present invention, the macrocapsules will have high permeabilities and resultant enhanced diffusional characteristics across the outer device wall. The improved transport properties mean the capsules will support higher level densities of viable cells than immunoisolatory capsules, which have lower MWCOs in order to 15 prevent immunological rejection from the transplant recipient. Higher 4* cell densitirsu may allow the use of smaller devices as compared to immunoisolatory capsules since they provide more therapeutic substance per unit volume. By the term "immunoisolatory capsule" it co is meant that the permselectivity of the capsule is such that it protects 20 cells in its core from the immune system of the individual in whom the vehicle is implanted. It does so by preventing harmful substances of the individual's body from entering the core of the vehicle, by minimizing contact between the individual and inflammatory, antigenic, or otherwise harmful materials which may be present in the core and by providing a spatial barrier sufficient to prevent immunological contact between the isolated cells and the individual's immune system.

1/ For ease of retrieval, the macrocapsules will generally have one or more tethers to allow location and grasping of the device without -18damage to it. Additionally, the tether can be used to find the implanted macrocapsule when it is desirous to terminate therapy.

Tethers are made according to the methods of Aebischer et al. in copending PCT application US92/05369. The macrocapsules can also be made to contain substances that enhance imaging to aid in the implantation of the device to the target site.

When the permselectivity of the macrocapsules is such that they are substantially non-immunoisolatory, the cells to be encapsulated should be compatible with the desired therapy or function. They will be cells that survive, with or without immunoisolation, long enough to have therapeutic value to the recipient. In general, choice of cells for specific functions, methods to insure implant survival, and various other considerations concerning cell viability in cell therapy situations are the same as those outlined by Gage et al., US Patent No.

15 5,082,670.

S. In some instances, it may be possible to provide cell therapy using the transplant recipient's own cells in the case of tumor, endothelial or epithelial biopsies etc.). Cells that are syngeneic to that of the recipient may also be used. In these situations, the lack of p immunoisolation provided by the macrocapsule is of little concern.

However, when xenogeneic and/or allogeneic cells are used, immunosuppressive techniques may be required. Local immunosuppression methods are disclosed by Gruber, Transplantation 54:1-11 (1992), and by Rossini, US Patent No. 5,026,365. General reviews and citations for the use of recombinant methods to reduce antigenicity of donor cells are disclosed by Gruber (supra). Exemplary approaches to the reduction of immunogenicity of transplants by cell surface modification are disclosed by Faustman WO 92/04033 (1992).

-19- Weiss, WO 93/14767, discloses methods of making antigen depleted cells from target cells. Sims, W093/02188, discloses genetically modified universal donor cells.

When the cells to be encapsulated are xenogeneic to the recipient, the need for immunosuppression is the greatest. Generally some method of reducing or eliminating the immune response to the encapsulated cells needs be employed. Thus recipients will often be immunosuppressed, either through the use of immunosuppressive drugs such as cyclosporin, or through local immunosuppression strategies employing locally applied immunosuppressants.

Alternatively the immunogenicity of the cells may be reduced by o preparing cells from a donor with reduced antigenicity, such as transgenic animals which have altered or deleted MHC antigens.

When cells that are allogeneic to that of the recipient are used, most S. 15 often tissue typing will be used in an effort to most closely match the histocompatibility type of the recipient.

An approach to insuring the viability of implanted cells as an alternative to such methods as immunomodulation, cell surface C modification, and general and local immunosuppression, is the 20 placement of cells within an immunoisolatory device prior to placement in the permselective macrocapsule. Suitable immunoisolatory devices would include microspheres or alginate noodles. Alginate has been demonstrated to be a useful immunoisolant as long as it contains a cell free region around the perimeter.

For cell therapy purposes, the macrocapsules of the present invention additionally make useful as therapeutic agents some cells which are currently not useful for unencapsulated cell therapy because such cells have specific disadvantageous properties such as the shedding of detrimental viruses or unwanted migratory characteristics.

In cell therapy methods, it is desirable to protect the host from viruses which may be carried by donor cells. Oftentimes, xenogeneic cells are used. While xenogeneic tissue is preferably sourced from herds of animals having carefully controlled heredity and diet, viral contamination is still possible. While most animal viruses will not be transmitted to humans, there are a number of zoonotic diseases which can be transmitted from animals to humans by a virus. For example, viruses which can be transmitted from cattle to humans include Bovine Diarrhea Virus, Infectious Bovine Rhinotracheitis Virus, Para-lnfluenza 3 Virus, Bovine Adeno Virus, Bovine Parvo Virus, Bovine Reovirus.

ce e The permeability of the encapsulation devices can be tailored so that the host is protected from these and other adventitious agents which could be released from xenogeneic cells.

Even in tha case of allografts, tissue may contain pathogens. While it is possible to screen for many detrimental agents, it is not possible to do so for all. For example, the diagnosis of HIV-free tissue may not be guaranteed because antibodies to the virus are generally not present until several weeks after infection. Such agents are 000" transmissible and use of macrocapsules that retard the transport of viruses add an additional measure of safety.

It may also be desirable to provide protection to the encapsulated tissue from viral attacks from an infected host. While there are many diseases which are clearly viral in origin, there are others, such as multiple sclerosis, amyotrophic lateral sclerosis and schizophrenia, -21where viral etiologies have been hypothesized. By having healthy cells encapsulated within a membrane which retards viral transmission, an infected host may be treated by cell therapy.

Table 1 lists various diseases and conditions for which the permselective macroeacapsulation cell therapy methods of the present invention may be useful, and the corresponding therapeutic substances that are released from the encapsulated cells.

C 0 *oo oa o o0 0 ap 1 0

I

I.

1

U

il 3 -22- TABLE 1 Disease/Condition Therapeutic Substance Deliverable by Genetically Engineered Cells Leukemia Kaposi's Sarcoma Hepatitis B a-interferon B-interferon y-interferon Hepatitis C

M.S.

Renal carcinoma Chronic granulomatosis Congenital leukopenia ,t I -i a re 3 r a s 9 a a o a j a r granulocyte colony stimulatory factor Hemophilia Factor VIII, Factor IX 15 B Cell lymphoma Autoimmune disease Arthritis Gaucher's disease Elevated cholesterol Tumors Angioplasty Growth hormone deficiency Nerve degeneration Bone repair Idiotypic antibodies antibody-secreting cells 111-RA Glucocerebrosidase macrophage colony stimulating factor tumor necrosis factor tPA growth hormone nerve growth factors bone morphogenic peptide Because viruses exist in a wide variety of shapes and sizes, various assays may be used to determine whether a given membrane will have the desired properties relative to a given virus. The present invention provides a method to test fibers for their properties relative to viruses so that their jiral release rate may be approximated upon I Ik: I i implantation. This method includes plaquing techniques which are known to be particularly sensitive to viruses. In many cases, the particular virus of interest may be tested directly by loading a given concentration of virus into the lumen of a fiber, securely sealing the ends, and placing the fiber in a bath solution. The bath solution can then be tested at desired time in ervals and assayed for viruses using standard plaquing procedures [Davis, Microbiology: including Immunology and Molecular Genetics, 3rd ed., Harper Row (1980)].

In cases where it may be undesirable to test the virus of interest .1p directly pathogenic viruses) a model of the virus of interest can be used. Bacteripphages of similar dimensions of the virus of interest serve as useful models due to their. viral nature. A variety of phages are available from the ATCC, and are also listed in Molecular Cloning: 'A Laboratory Manual (Sambrook, J. et al., Cold Spring Harbor 15 Laboratory Press, 1989). A single virus may infect a target bacterium.

The virus will replicate and be released to surrounding bacteria. Thus very low levels of virus may be identified by plaque formation on a i lawn of bacteria.

Table 2 gives the sizes of some representative animal viruses. They S 20 range from 18-26 nm for the parvoviruses to 150-300 nm for the pa:amyxo (parainfluenza) viruses.

I

-24- Table 2 Sizes of Representative Animal Viruses Virus Family Virus Diameter (nrn) Parvoviridae

(DNA)

Adenoviridae

(DNA)

Picornaviridae

(RNA)

Reoviridae

(RNA)

Orlhcimyxoviridae

(RNA)

Paramyxoviridae

(RNA)

Retroviridae Parvo Adeno Rhino 18-26 70-90 24-30 Reo Influenza Paramyxo (para influenza) oncovirus (Type C. B, Q) 80-120 150-100 100 p. 90 p *4 0* Table 3 gives the molecular weight, Stokes radius, and diffusivity of various markers and viruses.

Table 3 Molecular Weight, Stokes Radius, and Diff usivity of Various Markers Marker p I flew I. o Molecular Weight (9lmole) Stokes Radius (Angstroms) Diffusivity in Water 37 C (cm 2 /sec) Glucose Vitamin 812 Cytochrome C Myoglobin oi-chymotrypsinogen Ovalbumnin

BSA

IgG Apoferritin Parvovirus Orthomyxoviridae 186 1.300 13,400 16,900 24,500 43,500 67,000 155,000 440,000 -3.5 -7.7 16.5 15.2 22.5 27.6 51.3 59.3 -180 -:1200 9.24 x 10* 6 5.00 x 10*6 1.78 x 104 1.57 x 10'6 1.44 x 10'6 1.02 x 10-6 9.64 X 10, 7 6.29 x 6.11 x 1* -2.4 x 1*

~I-

As the nMWCO of a membrane increases from the ultrafiltration to the microporous range, the probability of virus passing through the membrane as a result of a concentration driving force increases. This ability to pass through a membrane is a function of the characteristic size of the viral particle and of the viral shape. Figure I shows the logorder reduction in phage dffision across two membranes of different nMWCOs. The ultrafiltration grade PAN/PVC membrane with a nMWCO of 100 kD (0.002 pm) is completely permeable to particles up to 440 kD when left over a 2 week period of time. Nevertheless, this membrane provides a 9 log-order reduction to the passage of phiX174 phage (-23 nm) over 2 weeks. This increases to a 12 log-order reduction for lambda GF10 (head 95 nm x 65 nm; tail 115 nm x 17 nm) even after 6 weeks. When the nMWCO of the membrane is increased into the microporous range -a polysulfone membrane with 15 a 0.1 pm cutoff- the log-order reduction of phiX174 is eliminated over a 3 day period of time and lambda GT10 is reduced to a 6 log-order reduction at 6 weeks.

In adr n-n to overall size, the shape of the virus is important in determining passage through the membrane. Phage M13 (5 nm x 200 nm) was completely blocked by the ultrafiltration membrane (log-order reduction 14 after 2 weeks). It is hypothesized that some configurations of virus such as long flexible viral particles may become entangled within the membrane pores contributing to a higher logorder reduction than size alone would account for. Proper matching of membrane properties with particular virus properties, particularly size, shape, and flexibility, will allow for selective retention of viral sized particles.

-26- Convective MWCO data for a range of particle sizes (proteins and dextrans) is used to characterize the fiber pore size. The MWCO of each fiber batch can be measured on a bundle of fibers potted in a cartridge. The cartridge is a hollow tube-like structure having capped ends. The fibers are threaded through the caps and sealed at the ends. The protein MWCO test involves ultrafiltering a protein solution (bovine serum albumin, Immunoglobulin G, myoglobin) at 5 pounds per square inch gauge (psig) transmembrane pressure. The reservoir and filtrate protein concentrations are measured on a spectrophotometer and the rejection coefficient is calculated from their ratio as R 1 Cf/Cr, where Cf is the filtrate concentration and Cr is the reservoir concentration.

The convective dextran MWCO test differs from thp diffusive protein "*MWCO test by controlling the filtrate volumetric flow rate in addition to 15 controlling the retentate volumetric flow rate. The dextran test reagent is a polydisperse solution varying in molecular weight from 2,000 g/mole up to 4,000,000 g/mol (from Pharmacia). Dextran concentration across a range of molecular weights is measured both in the reservoir and in the filtrate with gel permeation chromatography.

20 Rejection coefficients as a function of dextran radius are then calculated from the ratios of filtrate to reservoir concentrations. Either S: the full rejection profile can be reported or the dextran radii at 5 0 and 90% rejection.

In addition to cell therapy uses, permselective macrocapsules can also be used to hold polymeric implants such as bioerodible polymers containing therapeutically useful substances. Such implants without a j carrier may become brittle or fragile. Without a capsule carrier they may be difficult to retrieve after b6ing implanted. As erodible polymers r L I ii -27break down and release their target products, they may also become dislodged and may move into unintended sites. Microporous macrocapsules add the additional benefits of safety, retrievability, and tissue biocompatibility to polymeric delivery systems.

In order that the invention described herein may be more fully understood, the following examples are set forth. It should be understood that these examples are for illustrative purposes only and are not to be construed as kmiting the scope of the invention in any manner.

10 Example 1 Implantation of Allogeneic Fetal Brain Tissue Microporous fibers were prepared by extruding a 10% polyethylene oxide (MW=100K), 10% PAN/PVC (MW=100K), 80% Dimethyl formamide solution through the outer lumen of a coaxial extrusion nozzle. Water was used as a coagulant and was coextrudeu with the polymer through the center bore of the nozzle. Nominal MWCO of these fibers was 2,000 kD. The fiber had an inner diameter of 0.8 mm and an outer diameter of 1 mm.

o Fetal brain tissue was prepared according to the method of Tresco, et o 20 al., Society for Neuroscience Abstracts, 18:393.2 (1992). Tissue was loaded into these fibers in the presence of Matrigel (Collaborative Research) as a growth substratum and implanted in the striatum of healthy adult male Sprague Dawley rats within three hours of tissue isolation. Rats were sacrificed at 2 (n 3) and 4 (n 4) weeks, and the implants were histologically analyzed. After two and four weeks, viable cells were present that displayed morphological phenotypes similar to cultures maintained in vitro and isolated at the same time.

-28- Some of the cells stained positive for tyrosine hydroxylase, indicating neuronal survival in vivo for at least four weeks.

Example 2 Preparation of Macrocapsules Using Low Molecular Weight Additives The following casting solutions were prepared for hollow fibers: Solution 1 contained 12.5% PAN/PVC, 50.0% polyethylene glycol (MW 200) (PEG-200) and 37.5% N-methyl pyrrolidone (NMP). PEG-200 is a low molecular weight additive in which the PAN/PVC is not soluble.

At room temperature, the solution is a homogeneous gel. The casting solution is heated to 50°C and phase inverted into a bath at ambient temperature. A polymer results in wvhich the nMWCO is approximately 120 kD. The hollow fiber resulting from a solution without the PEG- 200 additive has a nMWCO of about 80 kD.

15 Solution 2 contains 13% PAN/PVC, 15% glycerol and NMP. It is a solid at room temperature and is spun fabricated in the manner described above. Membranes with nMWCOs ranging from approximately 200 kD to the microporous region, to even cellpermeable, may be fabricated.

20 Example 3 Q:n Preparation of Macrocapsules Using Solvent Addition in the Bore and Bath A 12.5% PAN/PVC, 87.5% NMP solution is prepared. When this solution is coagulated in water, a standard ultrafiltration membrane results. WheLi the same solution is precipitated in a solvent/water mixture, the MWCO increases greatly to the very high end of the ultrafiltration range (200 kD).

~1 -29- Example 4 Retrieval of Immunoisolated Cells: Alginate Noodles A sterile suspension of pancreatic islets was prepared from human pancreases by the method of Scharp et al., US Patent No. 4,868,121.

A 2% solution of sodium alginate in physiological saline (PS; 150 mM NaCI) was prepared under sterile conditions, which was then diluted 1:1 with the prepared cells, for a final concentration of 1% alginate in the islet suspension, mm cylindrical "noodles" of alginate w/alginate, Pronova")containing human pancreatic islets surrounded by a cell-free perimeter region were prepared as follows. The islet suspension was loaded a into the inner chamber of a nested dual-bore coextrusion device of the .c configuration described in copending U.S. Patent Application Serial No. 07/461,999, the inner bore of which has a diameter of 500 microns 15 and the peripheral bore of which has a diameter of 600 microns. The outer chamber of this device was loaded with a so'ution of sterile 1% sodium alginate in PS.

:The tip of the nozzle was immersed in a bath containing a sterile solution of 1% CaCI, in PS, which induces the hardening or gelling of 20 alginate by cross-linkage of alginate polyions. The materials loaded into the chambers were coextruded into this bath, generating a continuously forming alginate cylinder containing a core region of alginate matrix-immobilized islets and a surrounding region of alginate matrix free of islets. The outer diameter of the jacket was 1.2 mm.

The inner diameter of the core was 1.0-1.05 mm. The total islet volume of the core was 0.8 mm 3 (200 islets). The total core volume was 25.98 mm 3 The volume of the islets was 3% of the total core volume. The alginate of the core was cross-linked with the alginate of the jacket.

The relative thickness of the surrounding region was modified by adjusting the speeds at which the materials were extruded from the core and peripheral bores of the nozzle. In general as flow in the core increased, wall thickness decreased. As flow in the peripheral bore was increased wall thickness increased, Ranges of flow rates used were the same for both the core and periphery (0.3 1.5 ml/min.).

The ends of the cylinder were sealed by dipping the cylinder first in a sterile 2% sodium alginate bath, then in a sterile 1% CaC, bath. The macrocapsules so formed was maintained in sterile tissue culture medium prior to implantation.

00 These islet-containing "noodles" are cut to 2 cm. lengths using a sterile scalpel and loaded into 1.5 mm inner diameter PEO/PAN/PVC 15 capsules and implanted subcutaneously into diabetic recipients.

a Following 4 weeks, capsules are retrieved and islets assayed for viability.

o *4C Example Retrieval of Immunoisolated Cells: Microspheres Is" 20 Microspheres containing living cells are prepared by two separate methods. Thermoplastic microspheres containing PC12 cells are prepared according to the method of Sefton US Patent No. 4,353,888.

Alginate:polylysine microspheres are prepared according to the method of Lim US patent No. 4,352,883. In both instances, the microspheres are of a diameter no greater than 500 pm.

Microspheres are loaded into the thermoplastic restraining devices i 1 7 -t ,vU Sr So Sr 0 -31described in Example 1. All devices are tested to have nominal MWCOs of greater than 200 kD.

The PC12 containing macrocapsules are implanted into rat striata and left for 2 weeks. The macrocapsules are recovered, inspected for evidence of fibromatous overgrowth and assessed for cell viability.

Example 6 Implantation of Adrenal Chromaffin Cells Human adrenal chromaffin cells are prepared from human cadavers by the method of Sagen, U.S. Patent No. 4,753,635, and placed in permselective macrocapsules prepared as described in Example 1.

Capsules are implanted intrathecally into the lumbar regions of human patients suffering from chronic pain. Following three weeks capsules are removed and found to be intact and containing viable cells.

Example 7 Bacteriophage Retentivity Testing A test was developed to measure the transport retardation of lambda provided by a PAN/PVC membrane. This phage was chosen to serve as a model system to predict flux retardation for encapsulation of similarly-sized or larger animal viruses. Three different fibers were studied as shown in Table 4: Table 4 Type of Material Pore Skin Inner Wall Membrane Size Diameter thickness (mm) (mm) ultrafiltration PANIPVC -0.01" double 0.779 0.097 ultrafiltration PAN/PVC -0.01" single 0.682 0.086 microporous polysulfone 0.2 none 0.948 0.029 'corresponds to a nominal MWCO of 100 kD S o be *r o S 2o i ii -32- Lambda gt10 phage stocks were grown on lawns of NM514 coli from ATCC). Selected plates were reconstituted with 5 ml SM buffer (prepared using the methods of Sambrook J. et al., supra). These plates were gently shaken for 4 hours, the liquid pipetted into a centrifuge tube and 100 pl chloroform was added. After incubating for an hour, the stock was spun down at 3700 RPM and the supernatant recovered. Phage titer was determined prior to fiber loading.

NM514 were grown in NZCYM.media supplemented with 0.2% maltose to an optical density (OD) of 2. After the OD was established, bacteria were spun down at 3700 RPM and resuspended in 10 mM MgSO 4 (made in sterile water) using 37 0 C shaker bath until the pellet was fully suspended.

The testing device consisted of a 12 cm long cartridge with an internal volume of 9 ml filled with SM medium (Figure Two or three fibers 15 were potted into the male taper luers of the testing cartridge using minute epoxy. After potting the fibers into the devices, they were deglycerinated by ultrafiltration with Milli-Q water (100x fiber volume).

The PAN/PVC fibers were conditioned by ultrafiltering HL1 media at 2x 0 the fiber volume through the fibers. This treatment simulates the loading conditions used for encapsulating cells. Pore size of the membranes is decreased by protein adsorption to the membrane surface. In contrast, unconditioned polysulfone, microporous fibers were used to evaluate the increase in phage diffusion with increased membrane pore size. Loading the fiber with bacteriophage was accomplished by injecting 0.1 ml of bacteriophage stock into the fiber lumens through the luer using a 1 ml disposable syringe.

ends were capped and the devices placed in a 37 OC incubator. Baths were sampled at 1, 3 and 6 weeks and assayed for bacteriophage concentration using standard plaquing procedures.

Results Table 5 shows the phage concentration and Dmembrane at time 0 and 6 weeks.

Table Distribution of Lambda 0 Treatment Phage Concentration (pfu/ml) 15 cartridge 0 weeks 6 weeks 6 weeks Dwrcune *lumen lumen bath PAN/PVC 1 1.0e15 2.0el 1 10 7x10- 2 double skin 20 HL 1 2 1.0e15 2.0e9 10 7x10'" conditioned 0 0. PAN/PVC 1 1.0e15 2e10 10 7x10"' single skin HL1 2 1.0e15 3e9 1e3 8x10 conditioned 3 1.0e15 2e11 4e5 3x10 0 Polysulfone 1 1.0e15 6e10 7e7 5x10-' 6 not conditioned 2 1.0e15 5e10 5e7 3.8x10 6 3 1.0e15 le10 8e7 6x10- 6 As seen by the difference in the phage concentrations in the lumen and the bath at 6 weeks, the membrane provided significant resistance for phage diffusion into the bath. There was a decrease in the total

I

-34phage activity after 6 weeks of encapsulation. Controls indicated that phage stock held at 37°C at high (1 x 10" pfu/ml) and low (1 X 10 3 pfu/ml) are stable after 6 weeks. If the phage adsorbed to the membrane surfaces, this could account for the measured decrease in activity over the 6 weeks. Consequently, phage retention at 6 weeks was calculated in two ways: the total amount of phage loaded into the capsules was divided by the total phage concentration in the bath at 6 weeks or the total amount of phage present in the lumen of the capsule at 6 weeks was divided by the phage concentration in the bath at 6 weeks. These calculated log-order reductions in phage Idiffusion is shown in Table 6.

l Table 6 Log-order Reduction in Phage Diffusion' 15 Treatment Log Order Reduction Cartridge Method A Method B PAN/PVC double 1 9 12 skin HL1 conditioned 2 7 12 PAN/PVC single 1 7 skin HL1 conditioned 2 8 25 3 6 12 Polysulfone 1 3 8 unconditioned 2 3 6 3 3 6 'Method A log-order reduction lumen conc at 6 wks/bath conc at 6 wks Method B log order reduction lumen conc at 0 wks/100 bath conc at 6 wks Because the average dimensions for the lambda gt10 studied are 95 x nm for the head and 17 x 115 nm for the tail, the results of this experiment show that PAN/PVC membranes can retain adventitious agents larger than 95 nm by a factor of at least six orders of magnitude.

Example 8 Phage Phi-X Retentivity Testing The same types of membranes and methods of Example 7 were used to determine the transport retardation of phage phi-X. This phage is about 24-30 nm in diameter and is essentially spherical. The ultrafiltration PAN/PVC membrane was preconditioned with PC1 media, and the polysulfone microporous fiber was untreated. The samples were taken at 3 days for the microporous fiber and 1 and 2 Sweeks for the ultrafiltration fiber. LB broth (prepared using the methods of Sambrook J. et al., supra) was used in place of the SM media. The phi-x phage has a lower log order reduction at a shorter time than was seen for the larger lambda gt10 phage (Table 6 v. Table S O °oea

C

°«f l^ 'ft 6; a hri:l

C

o -36- Table 7 Log-order Reduction in Phage Diffusion' Treatment Log Order Reduction Cartridge' Time Method A Method B 2 weeks 4 9 1 2 2 weeks 7 3 1 week 6 9 4 3 days 1 3 days 1 6 3 days 1 'Cartridges 1,2,3: PAN/PVC conditioned with PC1 media.

Cartridges 4,5,6: Polysulfone untreated fibers Z ra~ 09 0 998.

Ih

I

Example 9 Protein Retentivity Testing The fibers studied in Example 7 were further characterized by measurement of percent rejection for several standard proteins and a blend of polydisperse dextrans. The rejection for the proteins are shown in Table 8. The rejection for the polydisperse dextrans are shown in Figures 3 and 4.

Table 8 Fiber Protein Rejection ~96.2 PAN/PVC double skin HL1 conditioned Bovine Serum Albumin Ovalbumin Myoglobin Bovine Serum Albumin Ovalbumin, 96.2% 26% PAN/PVC single skin HL1 conditioned 79% -37- All documents referred to herein are hereby expressly incorporated by reference.

*0 0

Claims (7)

1. A method of determining the viral retentivity of an external hollow fibre jacket on an implantable permselective macrocapsule, said method comprising at least the steps of: selecting a bacteriophage having similar physical characteristics to said virus; (ii) loading said bacteriophage into said hollow fibre jacket and sealing said jacket; (iii) placing said jacket into a bath solution; (iv) inoculating a lawn of bacteria capable of being infected by said bacteriophage with said bath solution; and calculating viral retentivity based on the number of plaques formed on said lawn of bacteria. a

2. The method according to claim 1 wherein the plhsical characteristics are selected from the list comprising size, shape and/or flexibility.

3. The method according to claim 1, wherein the physical characteristics are selected from the list comprising molecular weight, Stokes radius and/or diffusivity in water.

4. The method according to any one of claims 1 to 3 wherein the bacteriophage is selected from the list comprising M13, phi X, and lambda gtlO. The method according to any one of claims 1 to 4 wherein the bacteriophage is lambda gtlO XgtlO), the bacteria are E.coli and the calculated retentivity reflects the permeability of the permselective macrocapsule to one or more animal viruses.

6. The method according to any one of claims 1 to 4, wherein the bacteriophage is phi X174 (X174) and the calculated retentivity reflects the permeability of said microporous macrocapsule to a parvovirus or rhinovirus.

7. The method according to any one of claims 1 to 6, wherein the implantable j P:WOPERMRO\S67 1301 V.CLM 815198 -39- permselective macrocapsule is the implantable permselective macrocapsule defined by any one of claims 1 to 22 of Australian Patent No. 687371.

8. The method according to any one of claims 1 to 7 substantially as hereinbefore described with reference to the Figures and/or Examples. DATED this I11TH day of MAY, 1998 Brown Uniyersity Research Foundation by DAVIES COLLISON CAVE Patent Attorneys for the Applicants P:\OPER\MRO\687371B.DIV 8/5/98 ABSTRACT The present invention is directed to a method of determining the viral retentivity of a macroporous macrocapsule or external hollow fibre jacket thereon and more particularly, to the viral retentivity of an implantable permselective macrocapsule or external hollow fibre jacket thereon, by measuring the escape therefrom, of a microorganism of known physical characteristics, such as size, shape, flxibility, Stokes radius, molecular weight and diffusivity in water, amongst others. The inventive method is particularly useful in determining whether any particular macrocapsule or external hollow fibre jacket thereon is capable of preventing I 10 the escape of a virus particle from the inside of the jacket to the outside. *t* 1 or S *l *facts