Combined gene delivery has been widely adopted in gene therapy to increase therapeutic efficacy. However, some gene products are very toxic to normal tissues, which limit effective clinical application. To overcome this obstacle, the expression of one or more genes in the combined delivery should be regulated. Gene therapy utilizing a combination of IL-12 and GM-CSF has been previously established [4, 5]. In this strategy, GM-CSF is able to stimulate proliferation, maturation, and function of antigen representation cells, while IL-12 is able to enhance T-helper 1 cell’s immunity, increase cytotoxicity of T-lymphocytes, and inhibit angiogenesis [5, 8]. Another benefit of this strategy is that IL-12 can counteract the negative regulation of GM-CSF on Tc cells . However, high toxicity was observed with this combination due to the consistently high IL-12 expression. To overcome the high toxicity, we constructed an adenovirus to constitutively express human GM-CSF while controlling IL-12 expression via a heat-inducible promoter. After viral infection, heat stress induced a pulse-like expression of hIL-12 and a high constitutive expression of hGM-CSF in vitro and in vivo.
Consistent with previous reports, constitutive hIL-12 expression was very low in both the A549 and Hep3B cells under no heating. Heat stress induced 15 to 19 fold increases in hIL-12 expression in cultured cells, while it induced a 16.9 fold increase in Hep3B tumor tissues after a second heat treatment. This suggests that hsp70 promoter is highly inducible with low background activity. Consistent with our previous findings, heat-induced hIL-12 expression peaked at 24 hrs and began to decline at 48 hrs post heat treatment . This pattern can reduce the consistently high IL-12 expression-induced toxicity. In addition, we found that the second heat treatment is more effective than the first heat treatment in inducing hIL-12 expression, but the third heat treatment is less effective than the second heat treatment. The lower efficacy of the third heat treatment in inducing gene expression may suggest that one injection of non-replicating adenovirus can only support a limited number of heat treatments that induce gene expression. In addition, high virus dose could produce high hIL-12 expression under heat stress. However, low dose infection produced relatively higher amplification rate in hIL-12 expression due to the existence of low leak in hsp promoter activity. This observation supports the idea that the virus dose can be selected for clinical application. We acknowledge that we didn’t test the temperature-dependent effect of IL-12 expression and that is a weakness to this study. However, previous studies demonstrated a temperature-dependent effect in hsp70 promoter controlled gene expression [19, 20]. The second weakness is that the activity and toxicity of inducible human IL-12 cannot be tested in the animal model because human IL-12 shows no activity in animals and the nude mice used in this study are immunodeficient.
In this study, the adenovirus was constructed with a CMV-IE promoter to control human GM-CSF expression. The CMV promoter should produce highly constitutive hGM-CSF expression. However, heat treatment at 45°C increased hGM-CSF expression by 1-1.5 folds in A549 cells and 2-3 folds in Hep3B cells. In addition, the in vivo heating study revealed that although the first heat treatment induced a 2-fold increase in hGM-CSF expression, the second and third heat treatments didn’t increase hGM-CSF expression as much as the first treatment. In contrast, hGM-CSF expression was stable for 6 days after first heat treatment and declined on day 7. This observation suggests that the heat-inducible hGM-CSF expression is not heat-dependent but time-dependent. We also noted that heat induction of hGM-CSF expression is more obvious in Hep3B cells than in A549 cells, suggesting cell type dependence. Recently, the stimulating effect of heat stress on CMV promoter activity has been studied [21, 22]. Although the possible mechanisms might be complex, a considerable homology to the heat stress element core consensus (GA–TCC) within 18 bp elements in IE enhancer might be the most reasonable explanation [21, 23]. Heat stress might regulate CMV-IE activity directly and indirectly through heat-activated transcription factors. Heat stress inducing various transcriptional factors, including those activating the CMV-IE promoter, has been reported [21, 22]. Therefore, the cell type dependence might reflect the high specificity of the signaling pathway and transcription factors.
In this study, we established constitutive high expression of human GM-CSF and heat-induced expression of human IL-12 with a single adenoviral vector. The heat-induced hIL-12 expression has a pulse like shape with a peak at 24 hrs post heat stress that is maintained for 24 hrs in tumor tissues. Repeated heat treatments are effective but limited by the clearance of non-replicating adenovirus. Together with the low background activity of hsp70 promoter, heat induced gene expression enables a fairly strict control of gene expression, which diminishes the cytotoxicity of toxic cytokines . We also observed that the CMV-IE promoter driven constitutive high expression of hGM-CSF could be stimulated by heat stress in a cell type dependent manner. However, the CMV-IE promoter activity cannot be regulated by heat stress. Our study provided solid evidence for the feasibility of heat-induced regulation of gene expression in a combined gene delivery vector.