SummaryInoculated immunogenic cancer cells after initial growth are potentially rejected by specific host immunity; however, the outcome of the interaction between host and inoculated cancer cells is a function of multiple factors including the route of inoculation, the number of cells, the density of antigens on the injected cancer cells, and the state of the immune system of the host. In the present study, we have examined a different kind of variable: the stroma that inoculated tumor cells initially reside in. The impetus to examine this factor arises from observations that cancer cells from several lines inoculated as fragments of solid tumors often grow progressively, whereas the same number or more than 10-fold larger numbers of identical type cells injected as a suspension are rejected, even though fragments or suspended cells are both tumorigenic at the same doses in nude mice. In the present studies, we found that: (a) indeed, cancer cells inoculated as fragments were more tumorigenic than cancer cells in suspension; (b) the tumorigenicity of suspended cancer cells was increased by injection of the cells into polyurethane sponge implants; (c) cancer cells were more tumorigenic embedded in syngeneic stroma than in transgenic antigenic stroma expressing the K 2t6 major histocompatibility complex class I antigen; and (d) antigenic, bone marrow-derived, stromal components (presumably passenger leukocytes) were suflident to cause rejection of immunogenic but antigenically unrelated cancer. Together, these studies show that the stromal milieu helps determine the outcome of the complex interaction between host immunity and cancer cells and suggests a new mode of immunotherapy based upon targeting of antigenic leukocytes to the tumor site to serve as stimulators of immunological rejection of cancer.

A murine solid tumor was transfected to express various levels of an allogeneic major histocompatibility complex class I gene (K216), in order to test the effect of the level of antigen expression on immunogenicity and sensitivity to lysis by cytotoxic T lymphocytes (CTL). The growth rates of clones of tumor cells expressing different levels of the transfected gene were similar in vitro and in nude mice. Although all tumor cells, including cells freshly isolated from growing tumors, were equally sensitive to lysis by specific CTL, only tumor cells expressing the highest level of the K216 antigen stimulated CTL and were rejected by normal mice. In contrast, tumor cells expressing lower levels of antigen failed to immunize for CTL and grew progressively in normal mice, despite retaining expression of the transfected gene and remaining fully sensitive to CTL-mediated lysis; thus, the threshold of antigen needed to stimulate CTL responses was considerably higher than that needed to lyse tumor cells. Reduction of K216 antigen expression from 100-fold to 40-fold above background, impaired significantly the ability of the tumor cells to induce a K216-specific immune response, while tumor cells expressing K216 at levels 2-fold above background were as susceptible to CTL-mediated lysis as tumor cells expressing 50-fold more antigen. The important implication of these findings is that some tumors occurring in nature may not be immunogenic but nevertheless express antigens which are potential targets for immune therapy.