Search

Transduction of the murine interferon-alpha (IFN-alpha) gene into various malignant mouse tumor cells has resulted in the loss of tumorigenicity and an acquired capacity to induce long-lasting antitumor immunity following their injection into immunocompetent syngeneic mice. In the present study, we investigated the effectiveness of IFN-alpha-producing tumor cells in the therapy of mice with established mouse tumors. In DBA/2 mice bearing subcutaneous (s.c.) Friend erythroleukemia cell (FLC) tumors, we found that to achieve some antitumor response (i) it was necessary to inject high numbers of IFN-alpha-producing FLC, which occasionally lead to the formation of slowly growing tumors; and, that (ii) repeated injections of irradiated IFN-alpha-FLC did not result in any antitumor effect. The therapeutic potential of IFN-alpha-producing FLC rendered sensitive to ganciclovir (GCV), by transfer of the herpes simplex virus thymidine kinase (tk) gene, was investigated. Complete tumor rejection and cure was observed in > or = 70% of the animals after injection of high numbers (10(7)) of IFN-alpha-producing tk-expressing tumor cells followed 4 days later by repeated GCV treatments, whereas only a slight increase in survival time was obtained after administration of control tk-expressing tumor cells (not producing IFN) and GCV. Tumor rejection was associated with a dramatic destruction of tumor tissue and with the subsequent development of a potent and long-lasting antitumor immunity. No therapeutic effect was observed in immunosuppressed nude mice. These data indicate that this approach may represent an effective and safe therapeutic strategy for antitumor cytokine gene therapy.

Peritumoral injection of relatively low doses of either mouse interferon (IFN)-alpha/beta (10,000-20,000 units/injection) or of recombinant human interleukin-1 (IL-1) beta (125-250 ng/injection) in mice transplanted s.c. with Friend erythroleukemia cells (FLC) resulted in some inhibition of primary tumor growth, inhibition of liver and splenic metastases and increased survival time. A synergistic anti-tumor effect was observed in mice injected with both IL-1 and IFN-alpha/beta. Highly purified mouse IFN-beta also exerted a synergistic anti-tumor effect when combined with IL-1-beta in mice injected with FLC. The anti-tumor action of IL-1/IFN was markedly reduced in mice treated with antibodies to CD4 antigens. Antibodies to asialo-GM1 also diminished the anti-tumor effect by the combined cytokine treatment. The combined IL-1/IFN therapy was effective in NK-deficient bg/bg mice, although the extent of the anti-tumor response in these mice was less than that observed in bg/+mice.

Friend erythroleukemia cells (FLC) passaged in mice are highly tumorigenic and multiply extensively in the livers of suckling DBA/2 mice without differentiating. In contrast, in vitro passaged FLCs injected intravenously were of low tumorigenicity, multiplied to a limited extent in the livers of suckling mice, and underwent marked differentiation from the proerythroblast to the orthochromatic erythroblast stage in the liver. The presence of characteristic C-type virions budding from the cell surface in various stages of erythroid differentiation served as a marker of the injected FLCs. When the same in vitro passaged FLCs that differentiated in the liver were injected subcutaneously in suckling mice, they formed large subcutaneous tumors consisting of sheets of undifferentiated tumor cells. It is concluded that the tumorigenicity of FLCs depended on the site of tumor growth and that there is an inverse correlation between the tumorigenic capacity and the capacity to differentiate.

Peritumoral injection of recombinant human interleukin 1 beta (IL-1 beta) in mice transplanted subcutaneously with Friend erythroleukemia cells (FLC) resulted in a marked increase in survival time and inhibition of metastatic tumor growth in liver and spleen. In contract, IL-2 treatment alone did not significantly inhibit the development of FLC metastases. A synergistic antitumor effect was observed after combined IL-1/IL-2 therapy of these mice. The antitumor action of IL-1/IL-2 treatment was abolished or markedly reduced in mice treated with antibodies to CD4 or CD8 antigens, whereas antibodies to asialo-GM1 were ineffective. A clear-cut increase in the percentage of CD4+ cells was observed in the spleens of cytokine-treated mice on days 17 and 23. On day 23 of cytokine therapy, CD8+ cells were increased in both spleens and lymph nodes. On day 17, infiltrates of host-reactive cells (i.e., lymphocytes, granulocytes, and monocytes) were observed in both spleen and liver from FLC-injected mice treated with IL-1/IL-2, in association with tumor cells. On days 17 and 23, spleen cells and cells recovered from mesenteric lymph nodes of IL-1/IL-2-treated mice exerted a potent antitumor effect as determined by Winn assay experiments. This antitumor activity was abolished by preincubation of spleen cells with anti-CD8 antibody, but not by treatment with antibodies to asialo-GM1; antibodies to CD4 exerted only a slight effect. Combined IL-1/IL-2 therapy was more effective on established (i.e., 6-7-d) FLC tumors than on early (i.e., 1-d) tumor-transplanted mice. IL-1/IL-2 treatments were also highly effective in increasing survival time of mice from which the subcutaneous primary tumors were excised 7 d after FLC injection. These data indicate that in mice injected with FLC, the antitumor effects of IL-1/IL-2 are mediated by CD4+ and CD8+ cells (but not NK cells), and suggest that this combined cytokine treatment may be effective against established metastatic tumors.

Administration of alpha/beta-interferon (IFN) exerts a marked antitumor effect in DBA/2 mice given injections i.v. of large numbers of IFN-alpha/beta-resistant erythroleukemia cells (FLC). To investigate the possible mechanisms of FLC tumor inhibition in the liver of interferon-treated mice, we developed an in vitro model consisting of a coculture of IFN-alpha/beta-resistant 3Cl8 FLC and syngeneic mouse hepatocytes. Whereas IFN-alpha/beta did not inhibit the multiplication of 3Cl8 FLC cultivated alone, it effectively inhibited the multiplication of 3Cl8 FLC in coculture with hepatocytes. The inhibitory effect was directly proportional to the amount of IFN-alpha/beta added to the cocultures, and more than 90% inhibition of FLC multiplication was noted with 1.6 x 10(5) IU/ml of IFN-alpha/beta on Day 3 of coculture. When FLC were separated from the monolayer of hepatocytes by a pored membrane (0.4 microns), the inhibitory effect on FLC proliferation was unchanged, indicating that a soluble factor(s) released from IFN-treated hepatocytes was most important in the inhibition of FLC multiplication. An inhibitory activity of FLC multiplication was detected only in the conditioned medium of IFN-treated hepatocytes but not in the conditioned medium of control hepatocytes nor in extracts of IFN-treated or control hepatocytes. The inhibitory factor(s) in the conditioned medium of IFN-treated hepatocytes was retained by an ultrafiltration membrane (Mr cut off 10,000), and its activity was completely abrogated by trypsin digestion. Its stability to treatment with 1 M acetic acid as well as lack of correlation between the antiproliferative effect and the amount of L-arginine in the medium distinguished this factor(s) from liver arginase which was also found to be a potent inhibitor of FLC multiplication in vitro. The inhibitory factor(s) was also distinguishable in its biological activity from IFN gamma, interleukin 1 alpha and beta, and transforming growth factor beta 1 and beta 2. These results suggest the possibility that the inhibitory effect of IFN-alpha/beta on the development of 3Cl8 FLC in the livers of IFN-treated mice may be mediated by an IFN-induced inhibitor of FLC multiplication.

Lipid analyses were carried out on transplantable murine Friend leukemia cell tumors, 6 h after intratumoral administration of tumor necrosis factor-alpha (TNF). The levels of the major phospholipid classes were uniformly decreased to about 70% of control values; free fatty acids were increased to about 170%; diacylglycerol was decreased to about 50% and triacylglycerol, the main lipid component, was not significantly altered. These results analysed in the light of concomitant alterations in the levels of phospholipid precursors and catabolites (determined in previous 31P NMR studies) and histological modifications demonstrated that at early stages of TNF-induced inhibition of tumor growth (a) phospholipid catabolism was significantly enhanced; (b) morphological changes were apparently correlated with alterations in the levels of phosphatidylcholine and its catabolic products.

Adult DBA/2 mice were given injections s.c. with either interferon-sensitive (745) or -resistant (3Cl-8) Friend erythroleukemia cells (FLC). After tumor nodules had developed, mouse interferon-alpha/beta was injected daily into the tumor. 31P-Nuclear magnetic resonance (NMR) spectroscopy examinations were undertaken on freshly dissected tumors at different days of treatment with either interferon or control preparations. Analysis of 745 FLC tumors in untreated mice at different days of tumor growth (day 8 to 13 after tumor implantation) showed marked increases in the levels of phosphorylcholine (PCho), glycerophosphorylethanolamine (GroPEtn) and glycerophosphorylcholine (GroPCho). In contrast high levels of PCho, GroPEtn and GroPCho were already detectable in the 3Cl-8 FLC tumors on day 8, and no significant changes were observed during subsequent tumor growth. The intracellular pH value remained practically constant in both FLC tumors. Daily intratumoral administration of either partially purified (10(7) IU/mg of protein) or highly purified (10(9) IU/mg of protein) mouse interferon-alpha/beta to both cell tumors resulted in decreases in the levels of PCho, GroPEtn and GroPCho and in increases in the intracellular pH with respect to tumors treated with control preparations or left untreated. Two days of daily treatment of mice with interferon sufficed to induce these metabolic changes which preceded the appearance of necrosis in the tumors. Treatment of FLC tumors with X-rays on day 12 of tumor growth did not result in any comparable metabolic changes 2 days after irradiation. Changes in the levels of phospholipid metabolites were not observed when 745 or 3Cl-8 cells were cultivated in the presence of interferon. As interferon induced these changes in both interferon-sensitive and -resistant tumors we conclude that interferon treatment results in host-mediated effects on the biosynthesis and/or catabolism of tumor cell phospholipids.

We have used binding of radioactive lectins (i.e. Concanavalin A (ConA), Wheat Germ Agglutinin (WGA) and Ricinus communis agglutinin I (RCAI)) to membrane glycoproteins separated in SDS gel electrophoresis, to detect specific carbohydrate changes in plasma membrane proteins of in vivo passaged Friend erythroleukemia cells (FLC). These cells are highly metastatic to the liver, whereas the original in vitro passaged tumor cells do not metastasize. Marked qualitative differences in the high molecular weight region of the gels (100-200 kD) were observed between the WGA binding glycoproteins of metastatic in vivo passaged FLC and nonmetastatic in vitro passaged FLC. Furthermore, the binding of WGA to plasma membrane proteins of in vivo passaged FLC was much greater than the binding of WGA to plasma membrane proteins of in vitro passaged FLC. Lectin binding experiments after sialic acid removal by in situ mild acid hydrolysis of FLC glycoproteins indicated that an increased sialylation of the 120 and 145 kD glycoproteins was responsible for the increased WGA reactivity of in vivo passaged FLC plasma membranes. Besides the increased sialylation, other changes in glycosylation of the 100-200 kD glycoproteins of in vivo passaged FLC were observed: (1) qualitative differences between the WGA binding patterns of the two cell types were restored after treatment of the gels with mild acid and subsequent Smith degradation; (2) after chemical removal of sialic acid residues from the gels, qualitative differences in the RCA binding patterns to the glycoproteins of the two cell types were apparent.

The levels of expression of histocompatibility antigens on the cell membrane and their gene expression in non-metastatic and in highly metastatic Friend leukemia cells (FLC) were measured and the levels of expression of these antigens were correlated with the different in vivo behaviour of the tumor cells. Highly metastatic in vivo passaged FLC (either interferon-sensitive 745 or interferon alpha/beta-resistant 3Cl-8 cells) expressed higher levels of class I H-2K and H-2D antigens on their cell membrane with respect to the non-metastatic in vitro passaged counterparts. The increased expression of H-2 class I antigens was associated with an increased transcription of H-2K and H-2D genes. As both in vitro and in vivo passaged FLC have been shown to be resistant in vitro to the natural killer (NK) cell activity, we tried to correlate the levels of expression of histocompatibility antigens with the in vivo clearance of [125I]UDR-labeled FLC. However, no correlation was found between the levels of expression of H-2 antigens and the in vivo clearance of tumor cells. In fact, in vivo passaged FLC (tested either after 1 or after 15 in vitro passages) expressed virtually identical levels of H-2 antigens; however, the freshly explanted in vivo passaged FLC exhibited markedly lower levels of clearance from the lung, spleen and liver (when injected i.v. in DBA/2 mice) with respect to the corresponding FLC cultivated for several passages in vitro. Pretreatment of in vitro passaged 745 FLC with either interferon alpha/beta or interferon gamma resulted in the acquisition of some metastatic potential of FLC to the liver when interferon-treated FLC were subsequently injected i.v. in DBA/2 mice; such in vitro treatments resulted in a 2-3-fold increase in the expression of H-2K antigens versus the control untreated FLC. We suggest that such increases could represent some advantages for the homing properties of tumor cells and/or for the tumor progression, by mechanisms different from the resistance to the NK cells.

A 31P NMR study has been carried out on Friend erythroleukemia cells (FLC) induced to undergo erythroid differentiation in vitro. Significant levels of glycerophosphorylcholine (GroPCho) and phosphorylcholine (P-Cho) were identified both in the untreated cells and in their PCA extracts. In FLC treated 4 days in vitro with either dimethylsulfoxide (DMSO) or hexamethylenebisacetamide (HMBA), the intracellular concentration of P-Cho was markedly increased, whereas that of GroPCho appeared to be significantly reduced. HMBA was more effective than DMSO in producing this effect. The concomitant modulations of GroPCho and P-Cho in differentiated FLC suggest the hypothesis that erythroid differentiation involves modifications of the regulatory mechanisms controlling biosynthesis and catabolism of phospholipids.

Morphologic and metabolic studies have been carried out on Friend leukemia cell (FLC) tumors (grown in DBA/2 mice) or fibrosarcomas (grown in C3H/HeN or C3H/HeJ mice) shortly after peritumoral injection of recombinant mouse tumor necrosis factor (TNF) alpha. Marked vascular congestion and focal extravasation of erythrocytes were observed as soon as 1 hr after injection of either FLC tumors or fibrosarcomas with TNF. Focal areas of disaggregation of tumor cells were observed 1 hr after injection of TNF. Intraluminal thrombi (composed of degranulated platelets and fibrin) were detected 3 and 6 hr after TNF treatment, and were associated with areas of depletion of endothelial cell cytoplasm. To correlate these morphologic changes in the tumor with alterations in tumor metabolism, NMR spectroscopy and biochemical studies were undertaken on freshly dissected FLC tumors and fibrosarcomas shortly after injection of TNF. The earliest metabolic alterations observed after 1 hr in TNF-treated FLC tumors of fibrosarcomas were: (i) increase in the average intratumoral pH; (ii) decrease in the levels of ATP. These phenomena were not associated with a reduced glycolytic capacity of TNF-treated tumors as, at these early times after injection, the levels of lactic acid were virtually the same for TNF-treated or control treated tumors. Alterations in the levels of some products of phospholipid degradation (GroPCho, GroPEtn, GroP and Cho) also occurred in these tumors as early as 3 hr after TNF treatment. These metabolic changes were not observed in ascitic FLC tumors after TNF treatment. We suggest that TNF induces alterations in tumor blood vessels which subsequently lead to changes in tumor metabolism and tumor degeneration.