Tissue distribution and tumor localization of effector cells in adoptive immunotherapy of cancer.

In adoptive immunotherapy (AIT) of cancer, lymphocytes are isolated from the patient's blood and activated in vitro by the cytokine interleukin-2 (IL-2). In response to the IL-2 the lymphocytes proliferate vigorously and their cytotoxic potential increases several fold. After 5-10 days in culture, the cells-now called lymphokine-activated killer (LAK) cells-are injected back into the patient together with IL-2. The many positive results from preclinical animal models justified the rapid transit of AIT into the clinic, but the clinical results have far from fulfilled expectations. Many cancer centers have concluded that AIT in its present configuration is not cost-effective given that the average response rate is as low as 20-30%. Since a significant group of patients has shown complete responses after AIT, the challenge is to elucidate the conditions leading to optimal efficacy of AIT. It is generally accepted that the antineoplastic effect of LAK cells requires a close contact between the LAK cells and tumor cells. A central question in analyses of the mechanisms behind AIT is the ability of the LAK cells to localize to the malignant tissues. The earliest studies of the tissue distribution of 51Cr- and 111In-labeled LAK cells indicated that LAK cells, upon intravenous (i.v.) injection, are initially retained in the lungs, but redistribute to liver and spleen during the following 16-24 hours. However, our studies of the traffic and fate of i.v. injected tumor cells have shown that the use of 51Cr and 111In as cell labels often results in an over-estimation of the traffic of cells to liver and spleen and leads to falsely high predictions as to the survival of the injected cells, due to non-specific accumulation of 51Cr and 111In in liver and spleen after their release from dead cells. Use of 125IUdR, which does not accumulate in liver and spleen following release from dead cells, shows that the traffic of LAK cells into these organs was much lower than previously thought. These experiments have now been repeated using other cell labels (such as fluorescence dyes and immunohistochemistry) and they confirm that only few LAK cells redistribute from the lungs to the liver and spleen and that most die within the first 24 hours following injection. Thus, the circulatory potential of LAK cells is very low and chances that i.v. injected LAK cells will be able to localize into tumors and metastases located in other organs than the lungs, seems small. Indeed, while fluorescence-labeled LAK cells selectively localize into pulmonary metastases following intravenous injection, no infiltration of extrapulmonary metastases is seen. Furthermore, quantitative analyses have shown that even though the localization of LAK cells into pulmonary metastases is highly specific (5-10 fold higher numbers of LAK cells are often found in the metastases compared to the surrounding normal lung tissue), only 5% of the injected cells reach the malignant tissues. It is therefore reasonable to assume that the efficacy of AIT can be improved if the in vivo survival of the LAK cells can be prolonged and if their ability to infiltrate tumors regardless of their location, can be augmented. Previous studies in murine models have shown that i.v. injected tumor cells are sequestrated in the lungs and that only few of them reach other organs. However, when the tumor cells were injected into the left ventricle of the heart (bypassing the lung capillaries), significant numbers of tumor cells were found in the liver. It therefore seemed reasonable to speculate that LAK cells injected into the left ventricle of the heart or directly into the arteries supplying the tumor-bearing organ would have better chances of localizing to the malignant tissue. This seemed to be correct in that 10 fold higher numbers of LAK cells were found in the liver following intraportal injection compared to intravenous injection.