Gene Therapy (1998) 5, 778–788 1998 Stockton Press All rights reserved 0969-7128/98 $12.00 http://www.stockton-press.co.uk/gt IL-1/IL-3 gene therapy of non-small cell lung cancer (NSCLC) in rats using ‘cracked’ adenoproducer cells MC Esandi1,2, GD van Someren1, A Bout3, AH Mulder4, DW van Bekkum3, D Valerio1,3 and JL Noteboom1,5 1Section Gene Therapy, Department of Molecular Cell Biology, Leiden University; 3IntroGene BV, Leiden; and 4Pathologisch Laboratorium, Dordrecht, The Netherlands Cytokine gene therapy was studied in established L42 tumour responses. These were due to local release of cyto- tumours in syngeneic rats. L42 is a transplantable non- kines, not to systemic effects. Growth retardation also immunogenic non-small cell lung cancer (NSCLC). Genes occurred in contralateral tumours which were not injected. coding for human interleukin-1␣ and for rat interleukin-3␤ When rats carrying established tumours were vaccinated were transferred by injecting producer cells of recombinant with lysates of tumours collected during treatment with adenovirus vectors into the tumour in attempts to achieve ‘cracked’ producer cells, significant tumour growth retar- high concentrations of the cytokines inside the tumor with- dation was obtained. We speculate that both cytokines, if out systemic toxicity. Limited tumour growth delay was produced at sufficiently high concentrations in tumours, obtained with viable producer cells. For logistic reasons induce inflammation which in turn initiates an immune stocks of pooled frozen producer cells allowed intensive response against tumours growing at a distant site. These treatment of groups of tumour bearing rats. The cells were findings seem to justify further exploration of IL-1 and IL-3 lysed by thawing before administration. Ten daily injections gene transfer for the treatment of cancers. of such ‘cracked’ producer cells induced reproducible Keywords: gene therapy; interleukin-1; interleukin-3; lung cancer; adenovirus; therapeutic vaccination Introduction Accordingly, our focus has been on inducing a strong inflammatory reaction by gene transfer of IL-1 or IL-3. Cytokine gene therapy has been proposed as a promising IL-1 is one of the most pleiotropic cytokines capable of new approach to immunotherapy of cancer, because of potentiating both specific (humoral and cellular) as well 1,2 impressive tumour responses in experimental animals. as non-adaptive immune responses.17,18 There are two In several studies, implantation of tumour cells geneti- distinct IL-1 genes. Their corresponding proteins, IL-1␣ cally engineered to produce cytokines evoked tumour- and IL-1␤ bind to the same cell surface receptor and have specific immunity, as evidenced by the rejection of sub- identical biological properties. The interspecies similarity sequently injected parental cells. Such effects have been is more than 70% for both proteins and cross-species 3,4 5 observed with transduced cells expressing IL-2, IL-4, bioactivity was demonstrated.19 Apte et al10 reported a ␥ 6 7 8 9 interferon- , IL-7, GM-CSF, IL-12 and several other direct correlation between the constitutive expression of 10 11 cytokines including IL-1 and IL-3. IL-1␣ and reduced tumorigenicity of a murine fibrosar- Fewer reports are available of significant responses of coma. They have found that IL-1␣ induces enhanced established tumours as a result of treatment with cyto- helper T cell activity which provides auxiliary signals for 12,13 kine expressing cells. In another approach, in vivo the development and proliferation of cytotoxic T lympho- gene transfer using adenovirus vectors harbouring the IL- cytes (CTL). Non-adaptive effector cells, activated locally 2 gene into established tumours has resulted in by IL-1 expressing fibrosarcoma cells, were also shown 14,15 regression. to contribute to the eradication of the tumours. The introduction of cytokine genes was originally con- The cross-species reactivity of IL-3 is poor.20 IL-3 exerts ceived as a means of inducing immunostimulatory sig- a broad spectrum of biological properties, including nals inside tumours to enhance the host’s antitumour stimulation of effector functions of monocytes, eosino- immune response. During local inflammation, invading phils and basophils, which initiate and enhance inflam- host cells may destroy cells and release tumour antigens mation and allergic reations.21–23 Urticaria-like lesions in 16 in a form that can be presented to T lymphocytes. the skin as well as arthritis were observed in rhesus mon- keys following treatment with high doses of rhesus recombinant IL-3 protein.24 These lesions contained many Correspondence: DW van Bekkum, IntroGene bv PO Box 2048, 2301 CA mast cells which – if also accumulating in a tumour – Leiden, The Netherlands 2 might add a new component to inflammatory carcino- Current addresses: Department of Biologya y Bioquimica Universidad 25 Nacional del Sud, Bahia Blanca, Argentina; and 5Department of Clinical lysis. Pulaski and colleagues showed that expression of Oncology, Leiden University Hospital, The Netherlands IL-3 in a mouse lung carcinoma inhibits growth by stimu- + + Received 23 May 1997; accepted 13 January 1998 lation of CD8 lymphocytes through a CD4 -dependent Cytokine gene therapy NSCLC in rats MC Esandi et al 779 mechanism and enhanced expression of major histocom- in a proliferation assay on D-10 cells as described by patibility complex (MHC) class I molecules in the trans- Hopkins and Humphries.29 fected tumour cells. Recently, these investigators have As an ELISA is not available for rat IL-3, we used a demonstrated that IL-3 enhanced CTL development by proliferation assay to establish if L42 cells are capable of increasing the number of macrophage-like antigen producing and secreting IL-3. The cells were infected in presenting cells (APC) within the tumour.26 McBride et culture with IG.Ad.CMV.rIL-3␤ virus. Seventy-two hours al11 have found that IL-3 was effective in enhancing the later, supernatants were tested by limiting dilution for immunogenicity of irradiated tumour cell vaccines in a IL-3 activity in a proliferation assay on FDC-P1 cells. The murine fibrosarcoma model. Protection was tumour spe- supernatants of three different cultures were tested separ- cific and spleen cells from immunized mice could adop- ately and the standard curve with mouse receiving IL-3 tively transfer immunity. They also found that MHC class was made in triplicate. In all cases the triplicate dilution I expression is increased on IL-3-transduced tumour cells. versus proliferation curves were closely similar. Super- We investigated whether high levels of IL-1 or IL-3 in natants of cells infected with IG.Ad.CMV.Luc virus did the tumour would induce regression of established not stimulate proliferation, indicating that the activity tumours and generate tumour-specific immunity. The observed in supernatants of cells infected with tumour employed is a non-immunogenic transplantable IG.Ad.CMV.rIL-3␤ virus was specific for IL-3. As the rat bronchial carcinoma (L42). In the exploratory phase of species specificity of rat IL-3 is not known quantitatively our study, we observed regression of tumours following and since the test cells and the standard IL-3 were both repeated intra-tumour injections of extremely large doses of murine origin, one can only derive comparative values (5 ␮g daily for 10 days) of human IL-1␤, as well as from this assay. In supernatants of L42 cells infected at regression of an untreated second tumour growing 100 MOI the activity was equivalent to that of 750 U/ml simultanously in the same rat. Moreover, the cured rats of mouse IL-3. In supernatants of similarly infected rejected subsequent inoculates of L42 cells. However, sys- human A 549 cells the activity was found to be about 15 temic toxicity occurred during the treatment with IL-1, times higher, which supports the validity of the assay. and lower doses did not induce tumour responses, sug- The higher transduction of human cells is due to their gesting that the bulk of the injected protein was flushed higher susceptibility to infection by adenovirus vectors out by the blood circulation. To avoid systemic toxicity as was previously reported.30,31 and yet maintain high concentrations of the cytokine in the tumour, we have explored adenovirus-mediated gene Intra-tumour injection of purified recombinant adenovirus transfer. L42 tumour cells are relatively permissive to harbouring the IL-1␣ precursor cDNA adenovirus as compared with other rat lung tumour cell Established s.c. L42 tumours (n = 5) were injected on day lines. For logistic reasons we decided to treat the tumours 10 and day 20 after implantation with 109 infectious par- with adenovirus producer cells. For repeated adminis- ticles of IG.Ad.MLP.hIL-1␣ virus in a volume of 300 ␮l. tration to groups of rats, producer cells from multiple cul- No effect on the growth rate was observed. This is most tures were pooled and cryopreserved to be used after likely due to insufficient secretion of cytokines by the thawing – which lysed the cells – for treatment. In the transduced cells. The volume at the start of the treatment present study, we demonstrate the feasibility of this was about 50 mm3 (5 × 5 × 5 mm) that is equal to roughly delivery system to inhibit tumour growth in vivo and to 5 × 107 cells. If we consider an ideal situation in which induce systemic immunity against remote tumours. the virus particles are equally distributed over the tumour, the in vivo MOI corresponding to the injection of 109 infectious virus particles was about 20. Assuming Results a linear relation between MOI and cytokine production, we estimate from the in vitro data that between 1 and 2 Cytokine production by L42 cells after in vitro infection ng of IL-1 is produced per day in the tumours after the with recombinant adenovirus harbouring the hIL-1␣ or recombinant adenovirus injection.