[CANCER RESEARCH 48, 3607-3612, July 1. 1988| Increased Antitumor Effect of Immunoconjugates and Tumor Necrosis Factor in Vivo Mark J. Smyth, Geoffrey A. Pietersz, and Ian F. C. McKenzie Research Centre for Cancer and Transplantation, Department of Pathology, University of Melbourne, Parkville, Victoria, 3052, Australia ABSTRACT The potential of specifically targeting antineoplastic drugs and toxins to tumors with the use of monoclonal antibodies (MoAbs) reactive with tumor-associated antigens is currently being examined. JV-Acetyl-mel- phalan-MoAb (N-AcMEL-MoAb) conjugates have previously been shown to have greater antitumor activity than N-AcMEL, melphalan, or MoAb alone against both subcutaneous and ascites murine thymomas in mice (1). Although this conjugate is also a highly selective tumor inhibitor in vitro, it may not reach all the tumor cells in a high concentration, and consequently larger tumors (>0.4 cm2) cannot be eradicated. This con jugate is representative of many drug-MoAb conjugates in that they are unable to gain adequate access to the tumor site to exert their cytotoxic effect. To potentiate the antitumor effect of the N-AcMEL-MoAb con jugate, studies were undertaken to analyze its action in combination with recombinant human tumor necrosis factor a (rTNF-a), a monokini-, capable of causing acute necrosis of syngeneic tumor transplants in mice. Treatment of mice with murine thymomas (0.4 to 0.6 cm2 in size) demonstrated that 30% of the tumors in mice receiving conjugate and rTNF-a partially or completely regressed, while no regressions were observed in the tumors of mice receiving N-AcMEL-anti-Ly-2.1 conju gate or rTNF-a alone. This and other experiments indicated that the antitumor effect and tumor localization of N-AcMEL-MoAb conjugates can be enhanced in vivo by rTNF-a, thereby enabling successful eradi cation of lai HITestablished subcutaneous murine tumors. INTRODUCTION There are many problems associated with the use of immu- noconjugates, such as the complexity of coupling relatively Hydrophobie and bifunctional drugs to antibody (1), the potency (2) and nonspecific toxicity (3) of drug/toxin-MoAb1 conju gates, and the heterogeneity of tumor cells (4, 5), which have all been addressed. Successful treatment of subcutaneous tu mors by intratumor immunotoxin therapy (6) suggests that the relative inaccessibility of solid tumor to drug/toxin-MoAb con jugates is a major limitation of the more widely applicable i.v. route of administration. One method of partially overcoming this problem is to use vasoactive agents. By their selective action on normal blood vessels, vasoactive drugs can alter the tumor/ normal tissue perfusion ratio, thereby enhancing the access of drug-MoAb conjugates to tumors and increasing the effective ness of tumor therapy (7). However, it is clear from these studies that other methods have to be found to get more antibody out of the circulation and into the tumor. It is apparent that antibodies produced in response to bacterial infection require a local inflammatory response (with vasodilation and increased permeability of vessels) to permit the antibodies to reach their target, and on the basis of these observations, we have used TNF-a, a monokine capable of inducing hemorrhagic necrosis and a subsequent inflammatory response in tumors, to Received 12/15/87; revised 3/23/88; accepted 3/31/88. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1The abbreviations used are: MoAb, monoclonal antibodies; N-AcMEL, N- acetyl-melphalan; MEL, melphalan; rTNF-a, recombinant tumor necrosis factor a; rll-N -), recombinant -y-interferon; PBS, phosphate-buffered saline; DME, Dulbecco's modified Eagles medium; TNF-a, tumor necrosis factor a; TNF, tumor necrosis faeton %T/C, therapeutic index. increase the potential antitumor effect and tumor access of immunoconjugates. In partially purified form, TNF-a is cytostatic or cytotoxic for a variety of tumor cells in vitro (8) while having a number of noncytotoxic effects on normal cells (9). Recently, rTNF-a protein (10) has confirmed the selective cytotoxic effects in vitro on tumor cells previously observed with the natural TNF-a. TNF-a binds to specific receptors on the membrane of target cells (11) and is subsequently internalized and degraded (12), although the presence or absence of TNF receptors, however, is not correlated with sensitivity or resistance to the cytolytic effects of TNF-a (13). The overall spectrum of the antitumor activity of TNF-a is not yet understood, and in addition to a direct cytostatic or cytocidal effect on tumor cells, TNF-a is also capable of eliciting an antitumor effect by a direct cytotoxic effect on capillary endothelial cells, thereby leading to vascular damage and subsequent hemorrhagic tumor necrosis (14). Studies analyzing the effect of TNF-a in combination with other lymphokines (murine -y-interferon) and immunochemo- therapy in an in vivo model system are also a prerequisite for exploring the full potential of TNF-a given that TNF-a has a very narrow therapeutic range. We now present evidence that the antitumor effect of N-AcMEL-MoAb conjugates can be enhanced by recombinant TNF-a, enabling improved therapy of larger established murine thymomas. MATERIALS AND METHODS Tumor Growth. The E3 clonal variant of the murine thymoma ITT(1)7SNS (1) was maintained in vitro in DME, supplemented with 10% heat-inactivated newborn calf serum (Flow Laboratories, Sydney, Australia), 2 niM glutamine (Commonwealth Serum Laboratories, Mel bourne, Australia), 100-11' of penicillin/ml (Commonwealth Serum Laboratories), and 100 ¿tg/mlof streptomycin (Glaxo Laboratories, Melbourne). For in vivo experiments E3 was maintained by serial passage in the ascites form in CS7BL/6 x BALB/c !•', (hereafter called B6CF|) mice; cells from the ascites fluid were washed and centrifugea (400 x g, 5 min) twice in DME and PBS (pH 7.3), resuspended in PBS, and injected s.c. into mice. Tumor cells were injected s.c. into the abdominal wall and were allowed to develop into palpable tumors before commencing treatment. Mice were then subjected to a series of i.p. and i.v. treatments, and the size of the tumors was measured daily with a caliper square measuring along the perpendicular axes of the tumors; the data were recorded as the mean tumor size [product of two diameters ±SE (<5%)]. Experimental groups of 10 to 20 mice, all of the same sex and age, were used while a number of parameters were recorded during the course of experiment for each group, including: (a) the number of deaths due to rTNF-a toxicity (determined by histopa- thology); (b) %T/C tumor size = the mean tumor size of a treated group of mice relative to the mean tumor size of PBS treated mice on Day 20; (c) the proportion of tumors which remained completely regressed 100 days after tumor inoculation; and (d) the proportion of partial (>50% reduction in original tumor size) and complete tumor regres sions combined. Mice. B6CFi mice were produced in the Department of Pathology, University of Melbourne. MoAb. The MoAb used in this study was anti-Ly-2.1, a murine MoAb reactive with the murine Ly-2.1 specificity (IgG2a) (15). This MoAb was isolated from ascitic fluid by precipitation with 40% am- 3607 Research. on October 17, 2021. © 1988 American Association for Cancer cancerres.aacrjournals.org Downloaded from