CASE FOR ADOPTIVE IMMUNOTHERAPY IN CANCER

1190

,

the received oxygen concentrations at the "tracheal" end of an O2 delivery system. The importance of knowing the precise Fio2 has often been stressed, since too weak a concentration may not alleviate hypoxaemia and too much can induce oxygen toxicity or respiratory depression. For calculation of the alveolar-arterial O2 tension difference, it is important that the Fio2 be known.’ Venturi masks allow a constant Fio2 to be delivered uninfluenced by the pattern of breathing.4,5 These masks are known as fixed-performance systems-by contrast with standard face masks in which the Fio2 varies from patient to patient and breath to breath. 8,9 Although mathematical models are available that allow prediction of the Fio2, when the peak flow of the patient exceeds that delivered by the mask these models assume that volume and frequency of breathing are constant-rarely the case in clinical settings." Our circuit in a simulated patient allowed us to alter one variable at a time and ensured that circumstances would be identical on different days. With this mechanical model, we observed important differences between delivered and received oxygen concentrations. This difference increased at high breathing frequencies and increased tidal volumes and could be lessened by increasing the delivered gas flow rate. We attributed the differences between delivered and received oxygen concentrations to the mixing of incoming oxygen with ambient room air entrained by the mask. As the delivered gas flow rate was increased, the amount of ambient air entrained diminished. We suppose that, when CO2 is exhaled, a further reduction in Fio2 will occur. There was no difference between the still environment and the turbulent environment. However, there was a noticeable interaction between delivered gas flow and environment. At low flows (28 -5 l/min), the mean concentration difference under still conditions was 5 - 3% and turbulent conditions 6-9%. Under conditions of high flow (60 1/min), the concentration difference in a still environment was 1 -7% and in a turbulent environment it was 2 -0%. Thus, the overall similarity between the still and turbulent environment is explained by the interaction between flow and environment. Wexler et al. 10 have tried to improve the administration of oxygen by modifying oxygen masks with plastic shields. 10 We recreated their proposed modification of the face mask but there was no improvement in oxygen delivery even under the turbulent conditions regarded as ideal for the plastic shield. What are the clinical implications of these findings? Oxygen delivery by face mask is known to be altered by breathing pattern: the deeper and faster the rate ofbreathing, the greater will be the disparity between delivered and received O2,11 As tachypnoea and hyperventilation are frequently present in patients whose need for oxygen is great, it is unfortunate that the differences increase as Fio2 increases. Our clinical experience is that when patients require an Fio2 in excess of 0 - 4, the staff frequently change from "safe" Venturi masks to standard face masks whose limitations are described here. Our study has shown that increasing the flow rate is an effective means of lessening the differences between delivered and received oxygen. When Fio2 is to be measuredin a patient breathing from a standard face mask, the sampling port should ideally be placed in the trachea, oropharynx, or mouth rather than in the mask itself study was supported by West Park Hospital Foundation and by Medigas Limited, We thank Mr V. Ramcharan, Mrs A. Csima, and Mrs S. This

Little for expert technical assistance. We thank Prof. A. A. Scott for his encouragement.

Correspondence should be addressed to R. S. G., Respiratory Medicine, Hospital, 82 Buttonwood Ave, Toronto, Ontario, Canada M6M2J5. West Park

REFERENCES 1. 2. 3. 4. 5. 6. 7.

8. 9.

10.

11.

Harrison RA, Walton JR. Clinical application of blood gases. Chicago: Yearbook Medical Publishers, 1977: chap 15. Egan DF. Fundamentals of respiratory therapy. St Louis: C Mosby, 1977. Kory RC, Bergmann JC, Sweet RD, Smith JR. Comparative evaluation of oxygen therapy techniques. JAMA 1962; 179: 767-72. Campbell EJM, Gebbie T. Masks and tent for providing controlled oxygen concentrations. Lancet 1966; i: 468-69. Flenley DC, Hutchison DCS, Donald KW. Behaviour of apparatus for oxygen administration. Br Med J 1963; ii: 1081-88. Sykes MK, McNicol MW, Campbell EJM. Respiratory failure. Oxford: Blackwell Scientific Publications, 1976: 139-50. Leigh JM. Variation in performance of oxygen therapy devices. Anaesthesia 1970; 25: 210-22. Leigh JM. Variation in performance of oxygen therapy devices. Ann Roy Coll Surg Engl 1973; 52: 234-53. Wexler HR, Levy H, Cooper JD, Aberman A. Mathematical model to predict inspired oxygen concentration: Description and validation. Can Anaesth SocJ 1975; 22: 410-16. Wexler HR, Levy H, Cooper JD, Aberman A. Measurement of intratracheal oxygen concentrations during face mask administration of oxygen: a modification for improved control. Can Anaesth Soc J1975; 22: 417-31. Bethune DW, Collis JM. The evaluation- of oxygen masks. Anaesthesia 1970; 22: 43-54.

Shapiro B,

Hypothesis CASE FOR ADOPTIVE IMMUNOTHERAPY IN CANCER ARTHUR BERKEN

Department of Medicine, State University of New York, Stony Brook, New York, U.S.A. Patients with normal immune systems may be unable to mount effective defences against solid tumours because of (1) the generation of suppressor T cells in the low zone tolerance response elicited by the low concentrations of antigen furnished by slow growing solid tumours; (2) the ineffectiveness of the cytolytic T-cell response when the tumour cell membrane lacks the major histocompatibility gene products required for linkage to tumour antigens; and (3) the hindrance of antibodydependent cellular cytotoxicity by antitumour antibodies when the precise requirements for the reaction cannot be fulfilled in the sites occupied by solid tumours. Recent immunological advances suggest that it should be possible to isolate antigens from cancer cells, produce antibodies against these antigens, bind the antibodies to the patient’s macrophages and K lymphocytes, and reinject the bound cells into the patient to stimulate lymphokine synthesis and antibody-dependent cellular cytoxicity.

Summary

INTRODUCTION

IMMUNE control of the invasion of tissue by cancer cells surely differs from that of invasion by bacterial cells. Although most protocols for cancer immunotherapy continue to invoke the precepts of antimicrobial immunity,

concepts about the function of the immune system in the of cancer are made possible by knowledge recently gained in the laboratory. (For reviews see Benacerraf,1 new

treatment

Burakoff, and Greened) When introduced into man bacterial cells multiply extremely rapidly, with doubling times ofafew minutes. They bear many antigenic sites on their membranes. In a short time the infected host is challenged by enough bacterial antigen to

1191 an antibody response, which often disposes of a sufficient number of bacteria to prevent lethal infection. The antibodies bind to the offending bacteria, and the complex formed provides receptors which permit complementmediated lysis and receptors which attract mononuclear and polymorphonuclear phagocytes. The story for cancer cells is different. These cells multiply more slowly than bacterial cells, usually with doubling times of days to weeks or months. Whatever antigenic sites they bear are sparsely distributed. Ultimately the tumour cells may increase in number to the point at which the number of antigens reach the threshold necessary for eliciting an immune response. The number of cells necessary to provide the threshold tumour concentration of-antigens depends on the density of antigen on the tumour cell and the growth rate of the tumour. Unlike’ most animal models which use designated numbers of transplanted cancer cells to evaluate immune responses, human cancer probably starts from a single cell. Richter’s concept of the network theory4suggests that when antigen concentration is too low to provoke a normal immune response, a low zone tolerance response will occur. The low concentration of antigen permits the expression of an antiidiotype response to the initial immune reaction to the antigen. The anti-idiotype response consists of anti-idiotype antibodies and more significantly of anti-idiotype suppressor T cells.4 Application of this theory to spontaneous human cancer suggests that in many nascent cancers the tumour antigen could induce only low zone tolerance. Under such conditions populations of suppressor T cells for humoral and cellular antitumour defence mechanisms would be generated.

provoke

CELLULAR IMMUNE MECHANISMS AGAINST TUMOURS

Unlike the usual situation in antibacterial therapy, cellular immune mechanisms must assume a major role in antitumour immunotherapy. Tumour cells cannot be as efficiently attacked by antibody and complement as are bacterial cells, (because they have fewer antigenic sites on their membranes), nor can they be as easily opsonised and phagocytosed (because of their large size). Non-specific cellular immunity may be mediated by activated macrophages and natural killer (NK) lymphocytes. Specific antitumour activity has been attributed to macrophages and killer (K) lymphocytes which bind Fc-receptor-bearing antibodies to produce antibodydependent cellular cytotoxicity (ADCC). Effector cells possessing more clearly defined specificity are the cytolytic T lymphocytes (CTLs). The early studies by Zinkernagel and’ Doherty5 on cells infected by virus and later studies on tumour cells have shown that CTLs from immunised animals are directed against cell membrane targets which are combinations of major histocompatibility loci (MHC) coded antigens and the specific viral or tumour antigens. The degree of intimacy of the linkage between MHC antigen and specific antigen is uncertain. It is not known whether the CTL acts by a single receptor for the combination of antigens or by two receptors, one for each component. It is known that CTLs, human or animal, will not lyse target cells unless at least one MHC antigen is shared.5 However, studies with virus-infected human target cells have shown that the sharing of certain MHC

antigens by

CTLs and the virus-infected

target cell will not permit significant lysis by the CTLs.6 The production of specific CTLs after viral infection seems to be favoured by the presence of certain MHC antigens in linkage with specific viral antigen on the target cell membrane.7 These observations imply that in some individuals ineffective control of viral infection may result from failure

to

generate

CTLs

cells. This failure may be related to the absence of the required MHC antigens genetic, on the cell membrane. By analogy, in some individuals ineffective control of cancer may be related to the failure to generate anti-tumour CTLs because of the absence of the MHC antigens necessary for linkage to the tumour antigen in

against virus-infected

question. The other type of potentially specific cell-mediated cytotoxicity depends on the presence of Fc-receptor bearing antibodies. This receptor which provides cytophilic properties to the immunoglobulins is borne by antibodies comprising the major portion of human and animal IgG. These antibodies will bind to Fc binding sites on macrophages and on K lymphocytes. The binding is easily reversible at physiological temperatures if the antibodies are not aggregated or combined with antigen. Specific antibodies must compete with a myriad of other circulating antibody molecules with similar affinities for the Fc binding sites. Only low temperatures will prevent unaltered cytophilic antibodies from being competitively displaced from the mononuclear cell membrane sites to which they were bound.8 It is not likely that at usual body temperatures specific antibodies would be capable of permanently adhering to and arming mononuclear cells to provide the vehicle for ADCC. In-vitro studies, with cells from patients immunised against influenza by the use of mononuclear effector cells bereft of antibody have shown that ADCC can occur so long as antibody-producing cells were present in the reaction mixture with the infected target cells.9 ’Clearly, if ADCC is more than an in-vitro phenomenon, it requires the simultaneous presence of target cell, antibody, and mononuclear effector cell at the reaction site in the body in an intimate mixture and in the required proportions. This situation is possible in cancers highly exposed to the blood or lymph streams (leukaemia and possibly lymphoma), but is unlikely in most solid tumours. Indeed in extravascular sites where K cells or macrophages are not immediately available for binding to the Fc receptors of the antibodies which have reacted with the tumour cell, the tumour-cell antigens may be shed as antigen-antibody complexes, temporarily stripping the cell of its antigenicity. 10 .

REASONS FOR

INADEQUATE IMMUNITY AGAINST TUMOURS

Specific active antitumour immunisation is under any of of the following circumstances. 1. When suppressor cells are tumour-antigen concentrations.

generated

likely

fail

because of low

2. When the required MHC gene products obviating the production of, effective

T-lymphocytes.

to

are

absent,

cytolytic

_

,

3. When the antibody-dependent cellular cytotoxicity is required to be effective in extravascular sites (in solid tumours). Although non-specific active immunity may activate NK lymphocytes and macrophages, it may not protect against tumours because it may stimulate antibody formation to injected adjuvant substances and to other antigens which chance to’be present during adjuvant inoculation. The result would be a plethora of antibody-antigen complexes unrelated to the tumour but capable of binding to K lymphocytes and macrophages and making these cells unavailable for antitumour ADCC. Support for this possibility can be found in data given by Theophilopoulos and Dixon. 11I

_

1192 ADOPTIVE IMMUNISATION AGAINST TUMOURS

Recent immunological advances make adoptive immunisation a distinctly practical measure when active immunisation is inadequate. The three components necessary for constructing an effective system are (1) appropriate antigens, (2) selected antibodies, and (3) effector cells which can be armed. Human tumour cells bearing oncofetal antigens or high concentrations of ferritin are accessible to injected radiolabelled xenogenic antibodies to these ’substances. 12,13 Breast and stomach cancers express antigens which are unique among the other cells of the individual, but which seem to be related to common human antigens such as blood group substances. 14 Examples include the T (ThomsenFriedenreich) antigen in human breast cancer and heterophile antigen in human stomach cancer. This information suggests that it may be possible to isolate or synthesise common antigens which, although not unique to the tumour cells are, by virtue of their concentration, specific for an individual’s cancer cells. New monoclonal techniques15 indicate that it would be possible not only to obtain human antitumour antibodies but to select only Fc-receptor-bearing (cytophilic) immunoglobulins. These antibodies when aggregated should bind strongly to Fc binding sites of macrophages and K-lymphocytes harvested from the patient. When these bound cells are reinjected into the patient, lymphokine synthesis should be stimulated.16 The effector cells will remain armed because the aggregated antitumour antibodies cannot be readily displaced from the cell membrane Fc binding sites by circulating non-specific cytophilic antibodies. These antitumour antibody-armed effector cells should be able to reach tumour cells in extravascular locations and should be able to provoke ADCC even in the presence of specific T-cell suppressors and in the absence of the MHC gene-product required for CTL

activity. There are several advantages of such adoptive immunity. The treatment can be used during chemotherapy without compromise of its effectiveness. This is especially important during adjuvant chemotherapy when the potential for cure exists. If it is assumed that there is persisting postoperative subclinical disseminated tumour load of 108 or 109 cells, and that 1 cell per 106 cells becomes resistant to the chemotherapeutic agents,17 102 or 103 cancer cells may remain viable. Mononuclear cells can be harvested from the patient before chemotherapy, and when armed can be reinjected during the course of chemotherapy. These armed effector cells should be injectable in large enough numbers.to effect complete destruction of the remaining cancer cells. Another benefit is the suitability of this technique for treatment of tumours with antigenic heterogeneity. ISEffector cells can be armed with antibodies against different tumour-associated antigens to provide the potential for simultaneous ADCC against such

heterogeneous tumour cells.

testing of this system in patients with cancer does not inappropriate, especially after colorectal carcinoma resection. Monoclonal anti-CEA antibodies of the IgG 1 or IgG 3 human subclasses must be isolated. These antibodies should then be aggregated and reacted with mononuclear cells harvested from the patient and prepared by gradient centrifugation. Specific isolation of Fc-receptor bearing macrophages and K-lymphocytes should not be necessary. These cells, now armed and prepared for ADCC, may be reinjected into the patient during the period of adjuvant chemotherapy or radiotherapy. The risks to the patients’ The

seem

comfort and survival of this kind of treatment should be no greater than those of adjuvant immunotherapy programmes

already in use.

The benefits may be

considerably greater.

Correspondence should be addressed to A. B., 4277 Hempstead Turnpike, Bethpage, New York, 11714, U.S.A. REFERENCES

Baruj. Cellular interactions. In: Dorf ME, ed. The role of the major in immunobiology. New York: Garland S.T.P.M. Press 1981; 255-69. 2. Burakoff SJ. Specificity of cytolytic T-cell responses. In: Dorf ME, ed. The role of the major histocompatibility complex in immunobiology. New York: Garland ST. P.M. Press, 1981: 343-71. 3. Greene MI. Tumour immunity and the MHC. In: Dorf ME, ed. The role of the major histocompatibility complex in immunobiology. New York, Garland S.T.P.M. 1. Benacerraf

histocompatibility complex

Press 1981: 373-95. 4. Richter PH. The network idea and the immune response. In: Bell GI, Perelson AS, Pimbley GH, Sr, eds. Theoretical immunology. New York: Marcel Dekker, 1978: 539-69. 5. Zinkernagel RM, Doherty PC. Possible mechanisms of disease susceptibility association with major transplantation antigens. In: HLA and disease. Dausset J, Svejgaard A. Copenhagen: Munksgaard. 1977: 256-68. 6. McMichael A. HLA restriction of human cytotoxic T lymphocytes specific for Med 1978; 148: 1458-67. influenza virus. Exp J 7. Shaw S, Shearer GM, Biddison WE. Human cytotoxic T cell responses to type A and type B influenza viruses can be restricted by different HLA antigens. J Exp Med 1980; 151: 235-45. 8. Berken A, Benacerraf B. Properties of antibodies cytophilic for macrophages. J Exp

Med 1966; 123: 119-44. LH, Zinkernagel RM, Oldstone MBA. Immune response in humans after vaccination with vaccinia virus - specific cytotoxic activity by human peripheral

9. Perrin

lymphocytes. J Exp Med 1977; 146:

949-69.

10.

Colafat J, Hilgers J, VanBlitterswijk WJ, Verbeet M, Hageman PC. Antibody-induced modulation and shedding of mammary tumor virus antigens on the surfaces of GR ascites leukemia cells as compared with normal antigens. J Natl Cancer Inst 1976;

11.

Theophilopoulos AN, Dixon FJ. Immune complexes associated with neoplasia. In: Herberman RB, McIntyre RK, Eds. Immunodiagnosis of cancer, part 2. New York:

12.

Goldenberg DM, DeLand F, Kim E, et al. Use of radiolabelled antibodies to CEA for the detection and localization of diverse cancers by external photoscanning. N Engl J

56: 1019-29.

Marcel Dekker, 1978: 896-937.

Med 1978; 298: 1384-88. 13. Leibel SA, Klein JL, Sgagias M, Leichner P, Order S. The integration of tumor associated antigens in cancer management. Sem Oncol 1981; 8: 92-102. 14. Levine P. Blood group and tissue genetic markers in familial adenocarcinoma: potential specific immunotherapy. Sem Oncol 1978; 5: 25-45. 15. Olsson L, Kaplan HS. Human-human hybridomas producing monoclonal antibodies of predefined antigenic specificity. Proc Natl Acad Sci USA 1980; 77: 5429-31. 16. Neville ME, Lischner HW. Activation of Fc-receptor-bearing lymphocytes by immune complexes. II. Killer lymphocytes mediate Fc ligand-induced lymphokine production. J Exp Med 1981; 154: 1868-80. 17. Elliot EW. Genetics of drug resistance. In: Mihich E, ed. Drug resistance and selectivity. New York: Academic Press. 1973: 41-71. 18. Albino AP, Lloyd KO, Houghton AN, Oettgen HF., Old LJ. Heterogeneity in surface antigen and glycoprotein expression of cell lines derived from different melanoma metastases of the same patient. Implications for the study of tumour antigens. J Exp Med 1981; 154: 1764-78.

"Preparing medical students for the habits of good medical practice requires more elan and imagination than transiently showing and telling in clinics or in hospital wards. The job calls for continuous exposure to simple, reliable record keeping and a genuine concern for the comfort of the patient. My teachers were nurses. As student and house officer I stood in awe of them. They had an enormous influence. They used it to bring peripatetic young physicians back to earth and to the patient’s bedside. Nurses taught me to give injections, showed me aseptic technique, how to scrub up for the operating room (and threw me out ifIdidn’t scrub up properly), and what to do with bedpans. Even when I was not interested or was distracted, I could not escape. Of course, they had allies-the great clinicians who assiduously read all of the nurses’ notes and made it a point to discuss the patient’s course with the nurse every day. It did not take us long to understand that a principal ingredient of clinical excellence was learning from nurses. Sadly, these true realists of medicine have succumbed to the illusions of biomedical scientism. Nursing has copied the world of academic and institutional medicine with its pursuit of degrees, certificates, clinical specialities, and in its increasing distance from the sickbed. For both medicine and nursing, the cares of personal biology have become a ’scut’ job."-RiCHARD V. LEE. The Generalist: A Jaundiced View. Am J Med 1982; 73: 465-66.