One of the main goals for all cancer therapies is the selective targeting and killing of tumor cells, thereby increasing the therapeutic ratio. Both chemotherapy and radiotherapy causes dose limiting toxicities in normal tissues, which eventually reduce their clinical effectiveness29
. Unlike conventional therapeutic approaches, in cancer gene therapy suitable strategies can be employed to target the therapeutic transgene directly to tumor cells, thereby avoiding normal tissue toxicity.
Controlled regulation of transgene expression is now playing a major role in targeted cancer gene therapy strategies. Tumor-specific promoters are ideal in this regard as they should be highly active in tumor cells with little or no activity in normal cells. They are more generalizeable and safer than tissue specific promoters, as in the case of the latter, gene expression can occur in normal tissues such as brain, liver etc causing catastrophic effect. The promoter of the human Survivin gene has emerged as a favorable promoter of choice for gene therapy applications as it possesses the much coveted “Tumor on” and “Liver off” profile3,7,30–32
. Survivin promoter targeted infectivity enhanced conditionally replicative adenoviruses have shown tumor regression in mouse models of human cancers like breast, mesothelioma, cholangiosarcoma etc8,33,34
. In our hands, Survivin promoter provided the maximum specificity in non-tumor tissues compared to the widely employed constitutive human CMV promoter or tumor-specific human Telomerase promoter (hTERT) (data not shown). However, Survivin promoter was also a poor activator of transcription when compared to the CMV promoter, a benchmark promoter in the field. Hence the TSTA system was employed to enhance activity of the Survivin promoter.
There are many different strategies to increase promoter strength, but they each have their advantages and disadvantages20
. Strategies such as eliminating extra sequences from a natural promoter or introducing activating point mutations in the promoter sequences are useful but cannot be universally employed. Chimeric promoters designed by combining the transcription regulatory elements from different promoters or enhancer elements from the CMV promoter or other stronger viral promoters are often employed in these approaches. However, it was reported that these transcriptional regulator elements can often destroy promoter specificity in non-target tissues35
. Recombinant transcriptional activator approaches like the TSTA systems are by far more generalizable and has shown amplification of gene expressions from a wide variety of promoters without affecting their specificity.
The TSTA system employs the transactivator fusion protein Gal4-VP16 to amplify expressions of tissue specific promoters23,24,36,37
. The potent transactivation domain of VP16 enhances transcription from the promoter by multiple ways- 1) by recruiting38
the transcriptional machinery on the promoter by interacting with different transcriptional co-activators like TBP, TFIID, TFIIA, TFIIH40–42
and with the Mediator (TRAP-SMCC-ARC) complex43
2) by relieving nucleosome-mediated repression44
and iii) by facilitating chromatin remodeling45
. When the chimeric activator binds to Gal4-binding sites located upstream of the TATA box of a promoter they stimulate transcription in a synergistic fashion45,46
. Earlier studies with uni-directional TSTA system showed that a combination of 5-Gal4-binding sites and two activation domains of VP16 protein is optimum for achieving maximum amplification22,47
. Accordingly, in the current study the uni-directional TSTA system also amplified expression of both the genes 15–20-fold from the human Survivin promoter. The robust expression was now 4.3–11.5% of that obtained from the CMV promoter in the same cells. Interestingly, increasing the number of Gal4-binding sites from 5 to 8 resulted in a reduction in the level of expression. Despite earlier reports of a potential synergism between the different units of Gal4-VP16 bound to multiple Gal4-binding sites, studies also showed that multimerizing the VP16-activation domains can inhibit the degree of activation due to steric hindrance causing decreased interaction between the domains and their targets. In fact, linkers that relaxed the steric hindrance by spacing apart the VP16 activation domains achieved a higher level of activation39
. Therefore it is possible that multiple VP2 units recruited on closely spaced 8-Gal4 binding sites interfered with each other, thereby diminishing the level of expression compared to the 5-Gal4-binding site templates. However, in case of the bi-directional TSTA system, 5-Gal4 binding sites showed a distinct directional bias. Although strong (~15–20-fold), the amplification effect was concentrated in one direction only. This is in accordance with our previous results where we showed that only one of the genes was expressed strongly by the bi-directional TSTA system26
. That this was not gene specific was proved by switching around the orientation of the genes 180°. A possible reason could be the asymmetric positioning of the two E4TATA promoters around the 5X-Gal4 binding sites. While one is only 21 bp apart, the other one is 60 bp downstream of the Gal4-binding sites. Thus it is possible that the VP16 activation domains are unable to activate a promoter farther away and hence the distant promoter failed to show any gene amplification. Although earlier studies have demonstrated that VP16 can activate transcription strongly from promoters as far as 77 bp downstream39,48
, there is also substantial evidence that indicates that position of the TATA box from the Gal4-binding sites do effect VP16 mediated transactivation44,49
. This also explains why the introduction of 3 or 4 additional Gal4-binding sites in the vicinity of the compromised promoter that equalized the distances between the two TATA boxes (~28-bp apart) abolished the directional bias. However future studies should be performed to confirm this. The two minimal promoters seem to share the available transcriptional machinery, since an increase in expression from one direction dropped the expression from the other side, till they reached equilibrium.
In the current study, a replication incompetent adenoviral vector was developed to deliver the transgenes intratumorally. Adenoviral vectors are one of the most popular vehicles for gene transfer in clinical trials. However, the therapeutic outcome is often marred by the dose-related toxicity effects related to the virus. Intratumoral administration helps to maximize the infection ability of adenoviruses and limits exposure of the virus to the systemic circulation. As demonstrated in mouse xenograft models, the newly developed TSTA adenovirus results in efficient gene expression from the weak Survivin promoter following intratumoral administration. In living animals, the expression level was almost as high as that obtained from the very strong CMV promoter. This data indicated that the amplification system will be immensely helpful in achieving desirable therapeutic outcome with a lower viral dose in clinical trials. Unfortunately, although gene expression could be followed over time by this process, it was difficult to compare the results due to huge variability in viral retention in the tumors following intratumoral administration. Viruses are known to disseminate quickly from the site of injection into systemic circulation following injection50
. Adenovirus vectors in systemic circulation are well known for their natural hepatotropism which causes majority of the delivered viral dose to get sequestered in the liver. We therefore investigated whether amplification of the promoter activity by the TSTA system will negatively affect the specificity of the promoter. Although, the Survivin promoter-targeted TSTA adenovirus, when delivered systemically, showed a basal expression in the tumor-free liver, the expression was not significantly different compared to the promoter only system. Therefore, the TSTA system was capable of amplifying the strength of the promoter without affecting its specificity. We are currently testing the vector in animals with hepatic metastases to assess the ability of the targeted virus to distinguish between normal and tumor tissues in the same organ following systemic delivery (Ray et al. unpublished data).
The therapeutic protein TRAIL is known to have a soluble form. Therefore, TRAIL concentration in tumor lysates may not be a true representation of the absolute expression level of the cytokine. Measurements can be further confounded in living animals by the presence of endogenous TRAIL. Despite all these factors, we were able to achieve a good correlation between TRAIL concentration in crude tumor lysates and firefly luciferase gene expression. This very high degree of correlation is an important finding of the current paper because it lays the foundation for using similar vectors in the future to monitor the expression of any therapeutic gene in human gene therapy trials. In the future we can couple the therapeutic gene to a PET reporter gene such as Herpes Simplex Virus Thymidine kinase51
or the Sodium Iodide Transporter52
or a multimodality reporter gene for human applications.
In conclusion, we demonstrated the potential of a novel and readily generalizable gene therapy vector system that will not only allow one to monitor the level of expression of a delivered therapeutic gene, but can also amplify the expression of the transgene in a tumor-specific pattern. Such a vector system will be instrumental in achieving high therapeutic efficacy in future gene therapy applications.