WO1992016645A1

WO1992016645A1 - PROTEIN PRODUCTION FROM IMMORTAL CELLS GROWN $i(IN VIVO)

Info

Publication number: WO1992016645A1
Application number: PCT/GB1992/000459
Authority: WO
Inventors: David James Graham White; Johan Beyers Van Den Bogaerde
Original assignee: Imutran Limited
Priority date: 1991-03-15
Filing date: 1992-03-16
Publication date: 1992-10-01
Also published as: JPH06505628A; FI934026A0; AU1372192A; EP0579621A1; CA2104787A1; FI934026A; GB9105532D0

Abstract

Antibodies or other protein derived from a given species can be produced in a discordantly related host animal of choice by culturing a hybridoma or other immortal protein-producing cell line in the peritoneal cavity of the host animal, if both the humoral immune response and the cell-mediated immune response of the host animal are compromised. The humoral immune response can be compromised congenitally; so for example C3-deficient animals can be used. The cell-mediated immune response is preferably compromised by administering cyclosporin A. This enables human-derived antibodies to be produced in vivo and in general it allows the scaling up of in vivo production methods by using larger animals than mice, which are generally currently used.

Description

PROTEIN PRODUCTION FROM IMMORTAL CELLS GROWN IN VIVO

This invention relates to a method of producing proteins. In particular, the invention enables the production of large amounts of monoclonal antibodies, which can then be purified, giving a much higher yield of antibody than conventional methods used at present. The technology for both the production and purification of monoclonal antibodies is well described; this invention in part relates to a method of making monoclonal antibodies more cheaply than conventional methods. Additionally, the invention enables the production, in vivo, of monoclonal antibodies which have hitherto only been producible in vitro.

The method of generating monoclonal antibody-producing cell lines was first described in 1975 (Koehler and Milstein, Nature 256:495 and European Journal of Immunology 6:511, 1976). In 1977 monoclonal antibodies against T cell subsets were produced (Galfre et al Nature 266:550, 1977). Some workers produced monoclonal antibodies to MHC antigens (Lemke et al , Nature 271:249, 1978, Oi et al , Current Topics Microbiol . Immunol. 81:115, 1978, Moller Immunol . Rev. 41:1, 1979). It became clear that monoclonal antibodies could be produced against a wide range of antigens, and that these antibodies were very specific. Several workers used monoclonal antibodies against T cell subsets in transplantation studies to prolong graft survival and some even succeeded in inducing long term tolerance to allografts across a full MHC mismatch (Cobbold et al Transplantation 41:119, 1986, Madsen et al , Transplantation 44:849, 1987, Herbert et al , Transplantation, 46:128s, 1987, Qin et al, «7. of _Sxp.

Medicine 169:779, 1989). The immunosuppressive effects of monoclonal antibodies appear to depend on depletion of

T cell subsets, and/or blocking of antigens which are essential to T cell activation. The use of monoclonal antibodies in clinical transplantation, and in human autoimmune disease, is largely experimental, but the potential clinical applications are diverse (reviewed by

Waldmann, Annual Reviews of Immunology 7:407, 1990).

A major problem is that for these antibodies to be effective large doses would have to be injected. To produce 99% depletion of rat CD4-positive subsets, the dose of antibody must be at least 20mg of antibody per kilogram (Bogaerde et al, Transplantation January 1991, Herbert et al , Transplantation 46:128s, 1988). This would represent 1.4g of antibody, injected twice a week, if equivalent dosage regimes were followed in a 70kg human.

The production of these amounts of monoclonal antibody, using conventional methods of monoclonal antibody production, would be very expensive. In use today, there are two basic ways of producing monoclonal antibodies:

1. In vitro, by tissue culture techniques.

2. In vivo by growing the hybrid cells which produce the monoclonal antibody, in the peritoneal cavity of laboratory animals which are genetically identical to the parental cells and therefore not re ected by the host immune system. The in vitro method is useful because any cell type can be grown if the correct culture medium is used, but the concentration of antibody in the supernatant fluid is very low (depending on the cell type and the in vitro system used) . For example, the concentration in supernatant fluid of cultured cells may be approximately 100 to 200 μg/ml) . Purification of the antibody is expensive and time consuming, all the more so when it is realised that the culture medium generally has to be supplemented with irrelevant proteins, such as, for example, with 5% foetal calf serum. It would also be difficult to purify the antibody in such concentrations that the injection of the correct amount of antibody would not lead to fluid overload problems in the clinical setting.

The in vivo method has some advantages over the above in vitro system. The concentration of antibody is approximately an order of magnitude higher in the ascitic fluid than in the tissue culture fluid. Most murine hybridomas grown up in mice give antibody concentrations of 5 to lOmg/ml in ascitic fluid. The disadvantage of this method is that mice are small animals and each animal produces a maximum of about 10ml of ascites. A more important problem is that only wholly mouse-derived mouse hybridomas can be grown in mice. Human-mice hybridomas do not grow in the mouse peritoneal cavity, presumably because they are rejected. Since many of the clinically important monoclonal antibody-producing cell lines are of non-murine origin, the latter problem is significant. Further, there is a need for a host animal to produce human monoclonal antibodies. In summary, the use of monoclonal antibodies presents a new frontier in clinical medicine, but until a source of cheap, highly purified monoclonal antibodies becomes available the clinical applications of this type of therapy will remain experimental.

This invention relates in part to solving the problem of producing monoclonal antibodies in vivo in an animal of choice. By appropriate selection of the animal, the invention can be used to produce large amounts of monoclonal antibodies. If hybrid cells could be grown in animals larger than mice, such as rabbits or guinea pigs, much more ascites would be produced. The problem is that the hybrid cells would need to grow in the peritoneal cavity of a host animal which is of a different species. This is essentially the same problem as growing dual- species hybridomas (such as human-mouse) in any animal and producing humanised or other hybrid antibodies. A closely related problem in the production of other proteins from cells in an animal, where either the protein or the cell is not recognised by the animal as 'self.

According to the present invention there is provided a method of producing protein-producing the method comprising culturing an immortal antibody-producing cell line in vivo in a host animal, at least one of the protein and the cell being wholly or partially of a different species discordantly related to the host animal, wherein the humoral immune response and the cell- mediated immune response of the host animal are compromised. The protein (which term includes glycoproteins and other proteins which have been modified post-translationally) can be any protein which it is desired to produce. The protein should not have a significantly adverse effect on the host animal and should have a sufficiently long half life in the host animal for recovery to be a practicable proposition; apparent difficulties with either of these considerations, though, may be overcome by a number of ways, such as producing the protein in a precursor or inactivated form, which subsequently could be converted into the protein of interest. Generally, then, the protein can be any protein which can be secreted from the cell line, the capabilities of recombinant DNA technology make the type of protein to which the invention can apply virtually limitless. Examples of proteins which may be produced by the invention include α-fetoprotein, blood factors such as Factor VIII, erythropoietin and insulin, each of which may be from practically any species, but will often be of human origin.

The cell line can be any line which can be established in vivo in a host animal and which secretes the protein of interest. By virtue of the invention, the species of the host animal does not have to be matched to that of the cell line. suitable cell lines include BHK, CHO and, especially, COS cells, as well as many others, suitable transformation techniques are well known in the art.

One of the most important embodiments of the invention, though, relates to the production of antibody.

An immortal antibody-producing cell line will for preference be a hybridoma cell line, which will usually originally have been formed by fusing an antibody-forming cell with an appropriate tumour line. However, it may be possible to transform antibody-producing cells in other ways, for example with Epstein-Barr (EB) virus. The cell line need not be indefinitely immortal, although this may be preferred. Immortality is only necessary to enable enough antibody to be accumulated within the host animal prior to harvesting.

The immortal ceil line is cultured in vivo. The most suitable location is within the peritoneal cavity. After an appropriate culture period, antibody may then be harvested, and purified as appropriate, for example from the ascitic fluid.

The host animal can be of any suitable species. If the aim is to produce large amounts of the monoclonal antibodies, animals larger than mice can be used. Suitable animals may include guinea pigs, rabbits or even such larger animals as sheep, pigs and cattle. However, if the invention is being used not primarily to produce large quantities of antibodies, but rather to provide a suitable host animal for a hybrid cell or hybrid antibody (as discussed below) , any suitable species including such small animals as mice can be used.

At least one of the antibody and the cell is wholly or partially of a different species, which is discordantly related to the host animal. The antibody can be "partially" of a different species if it is a hybrid, that is to say if it has different regions being derived from different species, and one of the species is different from the host animal. "Humanised" antibodies are an example. The cell can be "partially" of a different species if, for example, it is a hybridoma derived from two different species. A human-mouse hybridoma would be an example. -5

The different species is discordantly related to the host animal. This is a term derived from the field of tissue grafts across species barriers. Rejection of tissue across species barriers is known as xenogeneic rejection. 0 Species combinations are either discordant or concordan . A concordant species difference means that vascularised organs transplanted from one species into another are rejected in days and not minutes or hours (Calne Transplant Proc 2:550, 1970). An example of a concordant 5 species difference is the hamster to rat combination. Discordant species are species combinations in which vascular organs are rejected in minutes, as in the rat to guinea pig combination. The presumed mechanism of discordant rejection is by antibody complement mediated 0 mechanisms (Reviewed by Auchincloss, Transplantation 46:1-20, 1988). Injection of human-mouse, or a mouse- mouse, antibody producing hybrid cells into the peritoneal cavity of guinea pigs would represent a discordant species difference, and the mechanism of 5 rejection of these cells would be similar to rejection of vascularised tissue grafts. The proviso to this statement is that the preformed antibodies which bind to antigens of the vascularised graft are present on the surface of the hybrid cells. If these antigens are not 0 present then an antibody complement mediated mechanism can still be important if the injected cells lead to the formation of antibodies to antigens on their surface. These cells would not have complement inhibiting proteins on their cell surface which could deactivate the complement components of the host organism, and any antibody binding to the cell surface would lead to binding and activation of host complement and subsequent lysis of the cells (Atkinson and Farries Immunol . Today 8:212, 1987) . It is also possible that in the absence of any antibody binding, the discordant cells might stimulate the alternate complement pathway and be lysed. There is some evidence that discordant cells could be destroyed by the alternative complement pathway alone

(White et al , in press, Miyagawa et al , Transplantation

46:825, 1988). The antibody complement mediated rejection of hybrid cells in discordant hosts would have to be addressed if these cells were to grow and produce antibodies in discordant hosts.

These hybrid cells would also be recognised as foreign by the host T cells, and cellular rejection would also have to be inhibited. There is some proof that T cells do not recognise discordant antigen presenting cells directly, and that cellular rejection of discordant grafts might be easier to inhibit than cellular rejection of allogeneic grafts (Yoshizawa et al , Journal of Immunol . 132:2820, 1984, Alter et al , J. Exp. Med. 171:333, 1990, Bogaerde et al , Transplantation January 1991, Pierson et al , J. Exp. Med. 170:991, 1989).

In general, discordantly related species are only distantly related to each other, and the mechanisms of rejection of a xenograft from one to another are predominantly humoral (Bogaerde, and White, Mechanism of Xenograft Rejection, in press) . The humoral immune response is that which causes cell destruction either as a result of activation by antibodies of the classical pathway of complement or as a result of direct activation via the alternative pathway.

The humoral immune response is compromised, in this invention, so as to prevent antibody complement mediated hyperacute rejection of discordant tissue. This can be achieved in several ways. Antibody depletion is one method; it has been achieved by plasmaphoresis, ex vivo perfusion of donor organs with recipient blood and by using protein A columns (reviewed in "Xenograft 25", Editor Mark Hardy, Elsevier 1989) . None of these depletion regimes achieved long term survival of discordant organ grafts. Another, and preferable, approach is to deplete the complement components by administration (for example by injection) of a complement inhibitor such as Cobra Venom Factor (CoF) , which leads to extension of discordant graft survival. The complement inhibitor may either prevent the activation of complement or remove it, or one or more of its components. Preferably, the complement inhibitor will inhibit both the classical and alternative pathways of complement activation, so it is preferred that the complement inhibitor act to block the pathway at C3, whether by inhibition or removal. It will be appreciated, therefore, that antibodies (for example monoclonal antibodies) against C3 may be used as the complement inhibitor in the present invention. Other useful complement inhibitors include soluble decay activating factor (DAF) and soluble membrane cofactor protein (MCP) . The most preferred complement inhibitor. however, is cobra venom factor. It has been shown that long term survival of hamster hearts in rats is possible by continued injection of CoF in spite of high levels of anti-hamster antibody (antibody titre 1/16000) . The anti-hamster antibody in these animals was capable of causing hyperacute rejection when injected into rats with beating hamster hearts which had not received CoF. Depletion of complement was therefore successful in inducing long term survival of vascular organ grafts in the face of elevated antibody levels.

Complement inhibitors such as cobra venom factor will generally be given parenterally and for this purpose will generally be sterile. Cobra venom factor itself may be formulated in phosphate buffered saline (without azide) or any other suitable solvent or carrier. It may be that oral formulations can be developed. The concentration of complement inhibitor in a pharmaceutical formulation will again depend on the ultimate intended dose, which will be under the control of the experimenter. For guidance, however, a preparation of cobra venom factor in azide- free PBS can be made at a concentration ranging from 0.05 mg/ml up to the limit of solubility, which will be dependent on its purity. Compositions having a concentration up to 5 mg/ml may be useful in practice, for example those having a concentration of from 0.1 to 2 mg/ml, for example about 0.5 mg/ml.

CoF is however very expensive, and this form of therapy may not be commercially viable in the bulk production of monoclonal antibodies. The most preferred way to compromise the antibody- mediated immune response is for the host animal to be congenitally compromised. Fortunately congenitally complement deficient animals already exist and others may be identified. C3, C4 and C2 deficient guinea pigs have been successfully bred, as well as C5 deficient rabbits. C3 deficient dogs are known to exist. These animals would be unable to activate the classical pathway or, in the preferred case of the C3 deficient animals such as guinea pigs or dogs, both the classical and he alternative pathway.

The cell-mediated immune response is that which causes cell destruction as a result of action of cells of the haematopoietic system either directly or as a result of targeting by antibody or non-specific factors. There are many ways of inhibiting or otherwise compromising the cellular response of the host animal. Procedures can be borrowed from the treatment of recipients of transplanted tissue. These include irradiation, thymectomy, the use of pharmacological agents such as cyclosporines or azathioprine, and monoclonal or polyclonal antibody therapy. One of the most effective and easiest methods is by administering (for example by injection) cyclosporin A. Other compounds having cyclosporin A-like activity can be used. Such compounds will be well known to those skilled in the art and include cyclosporins other than cyclosporin A, such as cyclosporin C and cyclosporin G. Eicosanoids also have cyclosporin A-like activity and include PGI₂ (prostacyclin) , PGE₂, PGD₂ and FK 506 (Fujisawa) . Compound FK 506 is one of those that is preferred for use in the present invention, but the most preferred is cyclosporin A itself. Cyclosporin A is available from Sandoz under the trademark SANDIMUN. Administration of cyclosporin A has the advantage of being economic.

■5 Cyclosporin A and other immunosuppressants with cyclosporin A-like activity will usually be administered parenterally in the practice of the present invention. In such cases, the preparation containing the immunosuppressant will be sterile. However, cyclosporin IC A itself, and some other immunosuppressants, may be administered orally, in which case sterility is not essential. Cyclosporin A itself is only sparsely soluble in water and for practical purposes it is preferred to dissolve it in an oil, such as a vegetable oil. Olive 5 oil has been found to be acceptable in practice. Dissolution in oil is only necessary when dealing with an immunosuppressant which is not soluble in water, and in general terms, the immunosuppressant may be dissolved in any appropriate physiologically acceptable carrier. In 0 some cases, dissolution may not even be necessary. The concentration at which the immunosuppressant is prepared will of course depend on the nature of the immunosuppressant itself and the intended dose, which will generally be under the guidance of the experimenter. 5 As a general guide, however, it has been found appropriate to prepare formulations of the cyclosporin at a concentration of from 1 to 200 mg/ml, for example from 5 to 150 mg/ml, typically from 20 to 100 mg/ml.

0 Reference has been made above to parenteral injection. The preferred site of injection is muscle, because of the depot effect resulting from an intramuscular injection. Intravenous injection, however, may be appropriate in some circumstances and subcutaneous injection may have some if not all of the advantages of intramuscular injection. Intraperitoneal injection could be used if circumstances dictate.

It is to be understood that when an immune response is said to be "compromised" in this specification, it is not necessary that the response be totally absent. Rather, the response is at least sufficiently attenuated for the acceptable working of the present invention.

The invention will now be illustrated by the following non-limiting examples. The examples refer to the drawings, in which:

FIGURE 1 shows a Fluorescence-Activated Cell Sorting (FACS) profile showing mouse R73 antibody-producing cells having been grown in ascitic fluid in the peritoneal cavity of C3 deficient guinea pigs;

FIGURE 2 shows a FACS profile of the negative control, namely the absence of R73 antibody- producing cells grown in normal guinea pigs; and

FIGURE 3 shows a FACS profile of the positive control, namely the presence of R73 antibody- producing cells in tissue culture supernatant.

Example 1

The murine cell line which produces the monoclonal antibody R73 (available from Dr. Thomas Hunig, University of Oxford, was used. This antibody stains all rat T cell receptor positive rat T cells. Guinea pigs which were congenitally depleted of the C3 component of complement were used as the recipients. (Guinea pigs and rats are discordantly related; rat hearts transplanted into guinea pigs are rejected hyperacutely (Adachi et al, Transplant. Proc. 19:1145 (1987)).) Five days before the injection of the cell line the guinea pigs received an intraperitoneal injection of 3ml pristane. Incomplete Freunds Adjuvant (IFA) can also be used. The injection of pristane (2,6,10,14- ecramethylpentadecane, Sigma) or IFA is used as a standard procedure to prepare the peritoneal cavity of mice^'before ascites producing cells are introduced into the peritoneal cavity. Five days after the injection of pristane, 10⁷ R73 producing cells which had been growing in tissue culture were injected into the peritoneal cavity of the animals. Cyclosporin A was also injected intramuscularly on the same day at a dose of lOOmg/kg, and cyclosporin A injections were given every day for 2 weeks in the same dose.

At the end of three weeks the animals were sacrificed and an autopsy performed. Growth of tumour cells was seen in the peritoneal cavity of all animals. Ascites fluid was present and a sample of this fluid was tested by Fluorescence-Activated Cell Sorting (FACS) , and compared with tissue culture supernatant of R73-producing cells; there was no difference in the staining pattern. This shows that functional^" ascitic fluid can be produced in the peritoneal cavity of discordant host animals.

These results are shown graphically in Figures 1 to 3. Figure 1 shows that in accordance with the invention murine R73 antibody-producing cells can be grown in ascites fluid in C3 deficient guinea pigs. The same results are obtained using hybridoma tissue culture supernatant (Figure 3) but not when an attempt is made to grow the cells in normal guinea pigs (Figure 2) .

Example 2

The procedure of Example l was repeated, except that C4- deficient guinea pigs were used. Growth of tumour cells was again seen in the peritoneal cavity of all animals. There was no macroscopically obvious difference in tumour growth between the C3 and the C4 deficient animals.

Claims

1. A method of producing protein, the method comprising culturing an immortal protein-producing cell line in vivo in a host animal, at least one of the protein and the cell being wholly or partially of a different species discordantly related to the host animal, wherein the humoral immune response and the cell-mediated immune response of the host animal are compromised.

2. A method as claimed in claim 1, wherein the protein is antibody.

3. A method as claimed in claim 2, wherein the immortal antibody-producing cell line is a hybridoma cell line.

4. A method as claimed in claim 1, 2 or 3, wherein the host animal is a guinea pig or rabbit.

5. A method as claimed in any one of claims 1 to 4, wherein the antibody is a hybrid antibody.

6. A method as claimed in claim 5, wherein the hybrid antibody is a humanised antibody.

7. A method as claimed in any one of claims 1 to 6, wherein the humoral immune response is compromised by antibody depletion or by administration of a complement inhibitor.

8. A method as claimed in any one of claims 1 to 6, wherein the humoral immune response is compromised by the host animal being congenitally compromised.

9. A method as claimed in claim 8, wherein both the classical and the alternative pathway of complement activation are compromised.

10. A method as claimed in any one of claims 1 to 9, wherein the cellular response of the host animal is compromised by irradiation, thymectomy, the use of a pharmacological agent or by monoclonal or polyclonal antibody therapy.

11. A method as claimed in claim 10, wherein the pharmacological agent has^' cyclosporin A-like activity.

12. A method as claimed in claim 10, wherein the pharmacological agent is cyclosporin A.