BACK TO THE DRAWING BOARD-THE NEED FOR MORE REALISTIC MODEL SYSTEMS FOR

IMMUNOTHERAPY PETER ALEXANDER

In experimental animals the growth of tumors which display strong immunoge- nicity can be slowed by immunological maneuvers that increase the magnitude of the host response to the specific tumor antigens. Such immunogeneic tumors do not, in general, cause distant metastases and may, therefore, not he relevant to the treatment of disseminated disease in man. This may explain why the current experience with immunotherapy based on such animal models has, with very few exceptions, been disappointing. Animal tumors which are not immunogenic by standard transplantation tests frequently disseminate and it seems likely that clinically useful immunotherapy has to be based on pro- cedures which are effective against such tumors. The lack of immunogenicity detectable by transplantation may be due to the absence of tumor-specific transplantation antigens (TSTAs), in which case if there is to be any immuno- therapy it will have to be directed at boosting some innate type of host resist- ance. Alternatively, the lack of immunogenicity may be attributable to the intervention of “escape mechanisms” which pervert the immunologically spe- cific host response to TSTAs. In this case, the immunological maneuvre should be directed at overcoming the escape problem and not at boosting the magni- tude of the specific host reaction to the TSTAs.

Cancer 40:467-470, 1977.

H E TERM IMMUNOTHERAPY HAS BEEN MUCH T abused and is not infrequently applied to a miscellany of procedures which lack a clear sci- entific rationale and involve ill-defined products of natural origin. In this paper I shall use the term immunotherapy for treatments that aug- ment actively or passively specific reactions di- rected against tumor-specific antigens which can act as targets for graft rejection because they are situated in the membrane of tumor cells (i.e. against tumor-specific transplantation type anti- gens-TSTAs). The presence of macromole- cules which are within membranes of tumor cells but absent on cells from normal adult tis- sues which are targets for immunological graft rejection processes was established within the last 20 years in most chemically, virally or phys- ically induced animal tumors.

The discovery of these TSTAs provides a ra-

Presented at the American Cancer Society and National Cancer Institute National Conference on Cancer Research and Clinical Investigation.

From the Chester Beatty Research Institute, Surrey, Eng- land.

Address for reprints: Peter Alexander, Chester Beatty Research Institute, Clifton Avenue, Belmont, Sutton, Sur- rey, England.

Accepted For publication, January 21, 1977

tionale for attempting to manipulate the growth of tumors by immunological procedures. Yet in an authoritative review of tumor immunology in 1964 Old and Boyse stated: “Despite the evident antigenicity of a variety of experimental tu- mours, it is a fact that no immunological ma- neuvre is known that will cause the rejection of an established malignant tumor, primary or transplanted, regardless of size. ” However, in the same year, my colleagues and I were able to demonstrate that the growth of autochthonous (i.e. primary) sarcomata induced in rats by ben- zpyrene could be slowed-and occasionally cures were achieved-by two different immu- nological procedures which showed a high de- gree of specificity. Ones was active immuno- therapy and boosted the immuno-reactivity mounted by the host to the TSTA and the other’ passive in the sense that the pool of cytotoxic lymphocytes was increased by cell transfer from specifically immunized donors. Since 1964 other immunological treatments have been found to be of limited therapeutic value in carefully cho- sen animal systems and the list of immunothera- peutic procedures showing active, albeit seldom dramatic, effects in some animal tumors is now quite large. Yet, 12 years after the first successes

467

with primary animal tumors neither the pro- cedures developed then nor others which are directed against TSTAs has been convincingly demonstrated (i.e. in properly conducted clini- cal trials using contemporaneous controls) to be therapeutically effective in disseminated human malignant disease. There are two possible ex- ceptions: one is acute myelogenous leukemia where some prolongation of life, but no cures, may have been obtained using immunotherapy in conjunction with chemotherapy. The other is the highly promising ongoing trial for Stage 1 lung cancer where the intrapleural injection of B.C.G. looks as if it will be of substantial thera- peutic benefit.

I have deliberately excluded as immuno- therapy the destruction of human skin tumors at sites of idammation caused either by local in- jection of BCG or by induction of delayed hyper- sensitivity reactions to common antigens. Such procedures can be locally effective but there is no reason to believe that they constitute im- munotherapy as I have defined it. The contribu- tion of immunity in these procedures is to create sites of inflammation in the midst of which tu- mor cells fare badly for reasons that are not yet clearly understood.

There are several and not mutually exclusive explanations for the lack of success within the last decade to translate immunotherapy from rodent to man. An optimistic view is that the conditions for success for immunotherapy are highly critical and that they were achieved by serendipity in the rat experiments and still re- main to be determined for human cancer. Some support for this interpretation comes from ani- mal investigations in which it was shown that the anti-tumor effect obtained by auto-immu- nizing with the irradiated tumor cells depends critically on the number of cells inoculated. lo,ll

Excessive numbers facilitate tumor growth and promote the development of metastasis possible because in following such treatments, TSTAs circulate in a soluble form and negate the host response against the tumor.' However, I am in- clined to dismiss failure to optimize treatment procedures as an important factor in the failure of immunotherapeutic procedures which are ef- fective in some animal experimental systems and useful in clinical practice.

There is a totally pessimistic interpretation and that is the great majority of human tumors do not carry TSTAs and hence the whole con- cept of immunotherapy as defined above has to be abandoned. While such a view cannot be excluded since the existence of TSTAs has yet to

be unambiguously demonstrated in any human tumor, there is much data which suggest there is some immune reaction by man against tumors.

T o me the most plausible reason why im- munotherapy has not been successfully trans- lated from the experimental to the clinical situa- tion is that the animal models on which they were based were not realistic. The core problem of clinical cancer is disseminated disease. Con- trol of primary and local spread can in most cases be achieved by surgery and radiotherapy and treatment failure due to local recurrence is a relatively rare situation. The challenge of cancer is to eliminate distant metastases. Unfortu- nately, our own experimental studies initiated 14 years ago and the experimental studies of many others were directed to the control in ex- perimental animals of a local tumor by immu- nological procedures. In most instances, includ- ing our models, these tumors are surgically curable and all we did was to substitute a com- plex immunological procedure for simple surgi- cal excision. Indeed, we and others showed that in animals immunotherapy was effective for those tumors which did not require new forms of therapy as they were localized. In our hands the immunological procedures which were effective in retarding the growth of non-metastasising primary sarcomata in rats were without effect on sarcomata which had the capacity to me- tastasize to distant sites. We' now know that in experimental animal tumors there is a correla- tion between the capacity to metastasize and the immunogenicity of the tumor. The more im- munogenic the tumor the less likely it is to me- tastasize.

More recent investigations lead to the propo- sition that tumors which are highly immuno- genic respond to immunotherapy but that these are the very tumors that are curable by local treatment such as surgery or radiotherapy. The tumors which present the clinical problem of distant dissemination are tumors which are not very immunogenic and which do not respond to the immunotherapeutic procedures so far stud- ied in experimental animal systems. Indeed we might go further and claim that surgery or local radiotherapy is the best type of immunotherapy we know. Because tumors invariably shed ma- lignant cells and dissemination therefore always occurs, such disseminated cells are destroyed by the host's immune reaction when the tumor is immunogenic. Hence the tumor is cured by sur- gery since the surgeon enables the host response to deal with the disseminated tumor cells by removing the bulk of tumor mass. In other

No. 1 MORE REALISTIC MODEL SYSTEMS FOR IMMUNOTHERAPY Alexander 469

words, immunotherapy works in a situation where surgery or radiotherapy is simpler and more effective but fails in situations where sur- gery or radiotherapy does not cure because of the presence of metastatic disease.

Those tumors which metastasize and where removal of the primary tumor does not lead to cure, i.e. the situation where new therapies are needed, arises when the disseminated tumor cells cannot be destroyed by the host reaction even after removal of the primary mass.

If the failure to destroy shed tumor cells in a host rendered free of microscopic tumor were due to the fact that the tumors did not have TSTAs, then of course the role of immuno- therapy to human cancer would be bleak. How- ever, animal data suggest that the lack of immu- nogenicity and the failure of shed tumor cells to be destroyed may not be due to the absence of TSTAs but that the host's reaction to them is negated by an efficient escape mechanism which allows tumor cells carrying TSTAs to avoid de- struction by the immune defenses of the body. The available data are consistent with the view that a contributory factor to metastatic spread is that the tumor cells which are shed have devel- oped the capacity to by-pass host reactions di- rected against TSTAs on their surface in much the same way as many bacterial infections per- sist in the presence of an active immune re- sponse. In animal models we have found that those tumors which are able to release TSTAs into the environment in a soluble form are able to escape from immune destruction in uiuo.' Moreover, when tumors are passaged they fre- quently change from a non-metastasizing to a metastasizing form. This progression is not nec- essarily associated with the loss of TSTAs but also occurs because a change in the physical properties of the membrane has taken place such that TSTAs become more readily shed.

If efficient escape is associated with the capac- ity to metastasize, then immunotherapeutic pro- cedures which rely on increasing the magnitude of the reaction against the tumor either by spe- cifically stimulating the host or by passively ad- ministering effector processes such as cytotoxic cells or antibodies are unlikely to be useful for the control of metastasis.

I feel that new procedures of immunotherapy need to be developed that concentrate on devis- ing ways of by-passing the escape mechanism rather than boosting further an already devel- oped host response. Essentially this means that experimental immunotherapy has to go back to the drawing board and that new ideas have to be

developed in animal models which mimic a clin- ically meaningful situation. In my view the in- travenous injection of tumor cells to give lung tumors does not constitute a valid model for the treatment of distant metastases. Intravenous in- jection giving rise to lung tumors is in essence a local phenomena since it is in the lung that the tumor cells become arrested. There are now plenty of models in both mice and rats where a tumor may be transplanted and after surgical removal of the initial transplant the animals die of distant metastases months later. I think it is important that models be chosen where there is a significant interval between the appearance of metastases and the surgical removal of the origi- nal implant. Models such as the Lewis lung tumor in which tumor cells are shed from the primary implant in vast numbers and in which animals die from lung metastases within two or three weeks of the implantation at a sub- cutaneous site are not realistic. However, good models do exist and the need now is to develop procedure for perverting excape mechanisms which result in distant disseminated cells devel- oping into metastases.

Logically the clinical application of any of the new generation of immunotherapy ought to await much more detailed study of the mecha- nisms of escape. However, we are impatient and empirical experiments are being carried on in parallel with the more systematic studies. Thus, several investigator^^*^*^ (Proctor et al., 1973; Milas et al., 1974; Suit et al., 1976) using metastasizing animal tumors have found that stimulation of the reticulo-endothelial system, particularly with CoTynebacterium paruum, reduces distant metastases but that this treatment is rel- atively ineffective against the primary implant. This is the exact reverse of the first generation of immunotherapy where the principal effect in experimental animals was on the local tumor and little or no effect was obtained against dis- seminated disease. A possible rationale for the anti-metastatic effect of R.E. stimulation is that it increases the rate at which soluble TSTAs are cleared from the circulation. Removal of circu- lating TSTAs impedes those escape mechanisms which depend on circulating TSTAs in a soluble form inhibiting the effector arms of the host response.

The emphasis in immunotherapy must change from an approach predicated to increas- ing the magnitude of the host response to at- tempts to render ineffective the escape mecha- nisms developed by tumors that metastasize. The clinical application of such concepts is in

470 CANCER July Supplement 1977 Vol. 40

the future; initial efforts should be concentrated on by-passing escape in clinically realistic ani- mal models. Only when successes have been obtained in such models is it justifiable to test such procedures in man.

I must plead guilty that 10 years ago our group made a rash extrapolation and immedi- ately and unsuccessfully attempted to translate to man an immunotherapeutic procedure which

worked in an animal system. This was done without considering whether the problem suc- cessfully attacked in the animal system was the problem which needed to be attacked in man. It is important that the second generation of im- munotherapy is not initiated prematurely and that we are quite sure that the animal models on which it is based truly mimic the clinical prob- lem which needs to be solved.

REFERENCES

1. Alexander, P.: Escape from immune destruction by the host through sheeding of surface antigens: Is this a charac- teristic shared by malignant and embryonic cells? Cancer Res. 34:2077-2082, 1974.

2. Alexander, P., Currie, G . A,, and Thomson, D. M. P.: The presence of soluble tumor-specific antigens in the blood of patients and of experimental animals with tumors and their role in immunotherapy. I n 26th Annual Symposium on Fundamental Cancer Research at University of Texas, M. L). Anderson Hospital. Baltimore, Williams and Wilkins, 1973.

3. Delorme, E. J., and Alexander, P.: Treatment of primary fibrosarcoma in the rat with immune lymphocytes. Lancet 11: 117-121, 1964.

4. Eccles, S. A., and Alexander, P.: Macrophage content of tumours in relation to metastatic spread and host immune reaction. Nature 250:667-669, 1974.

5 . Haddow, A., and Alexander, P.: An immunological method of increasing the sensitivity of primary sarcomas to local irradiation with x-rays. Lancet I:452-457, 1961.

6. Milas, L., Hunter, N., and Withers, R. H.: Corynebac-

terium granulosum-induced protection against artificial pul- monary metastases of a syngeneic fibrosarcoma in mice. Cancn Res. 34:613-620, 1174.

7. Old, L. J., and Boyse, E. A,: Immunology of experi- mental tumours. Annu. Rev. Med. 15:167-186, 1964.

8. Proctor, J , , Rudenstam, C. M., and Alexander, P.: Increased incidence of lung metastases following treatment of rats bearing hepatomas with irradiated tumour cells and the beneficial effect of Corynebacterium parvum in this sys- tem. Biomedicine 19:248-252, 1973.

9. Suit, H. D., Sedlacek, R. S., Silobrcic, V., and Ling- good, R. M.: Radiation therapy and Corynebacterium par- vum in the treatment of murine tumors. Cancer

10. Vaage, J.: Specific de-sensitization of resistance against a syngeneic Methylcholanthrene-induced sarcoma in C3HF mice. Cancer Res. 32:193, 1972.

11. Vanwijik, R. R., Godrick, E. A., SMith, H. G., Gold- weitz, J., and Wilson, R. E.: Stimulation or suppression of metastases with graded doses of tumour cells. Cancn Rcs. 31:1559, 1971.

37:2573-2579, 1976.