Human adenoviruses commonly cause acute infections of the respiratory and GI tracts, the eye, and various other tissues. They also form long-term persistent infections, and their emergence is a major problem in immunocompromised patients. Recently, there has been a great deal of interest in using adenoviruses as genetically-engineered “vectors” for specialized vaccines, for gene therapy, and for cancer gene therapy (adenovirus as a drug to treat cancer).
Aside from their interest as agents of disease, adenoviruses are one of the premier model systems to study the molecular and cellular biology of human cells. Adenoviruses are easy to work with, they are very well characterized, and they pose no risk to the researcher. They are classified as "DNA tumor viruses" because they have a DNA genome, they transform cultured cells to a malignant state, and they can induce tumors in experimental animals (but not in humans). Thus, adenoviruses are excellent models to understand cancer. Adenovirus genes are expressed in the cell nucleus using primarily host cell mechanisms, and many of the adenovirus gene products usurp key regulatory aspects of the cell and convert the cell into a factory for efficient virus replication. Many fundamental discoveries have been made using adenoviruses, including pre-mRNA splicing (Nobel prize awarded in 1993) and identification of novel transcription factors and cellular proteins that control the cell cycle and are the key regulators of malignancy. There is a long tradition of adenovirus research at SLU dating back to the 1950's, and it can be argued that SLU is the birthplace of adenovirus molecular virology.
I have studied the molecular aspects of human adenovirus replication for more than three decades. Much of our research has been focused on the seven protein products of the E3 transcription unit. We were the first to identify six of the seven E3 proteins (named E3-14.7K, E3-RIDα, E3-RIDβ, Adenovirus Death Protein, E3-6.7K, and E3-12.5K), and we and our collaborators were the first to determine the function of the E3-14.7K, E3-RIDα, E3-RIDβ, and Adenovirus Death Protein proteins (Fig. 1). In general, the E3 proteins function to protect adenovirus-infected cells from destruction by host immune killer cells. We showed that the E3 RID, 14.7K, and 6.7K proteins prevent cells from apoptosis induced by the so-called “death” ligands (tumor necrosis factor, Fas ligand, and TRAIL) that are expressed on immune killer cells. We believe that these E3 proteins enable adenovirus to evade the host’s anti-viral defenses.
Fig. 1. Adenovirus Proteins in the E3 transcription unit that counteract killing of adnovirus-infected cells by cells of the host immune system.
One of the E3 proteins discovered and characterized in my laboratory is named the “Adenovirus DeathProtein” (ADP). In a series of studies we showed that ADP functions at the culmination of infection to facilitate lysis of the infected cell and the release of progeny adenovirus from the cell.
In the late 1990’s I began research to develop and characterize oncolytic adenovirus vectors for cancer gene therapy. Oncolytic adenovirus vectors destroy cancer cells through the natural process of virus replication (Fig. 2). Our approach to constructing a potent anti-cancer vector was based on the hypothesis that if adenovirus could be engineered to “overexpress” ADP, then the adenovirus vector would have enhanced ability to lyse cancer cells and to spread from cell to cell in the tumor, thereby destroying the tumor. This approach proved to be successful: our ADP-overexpressing vector named VRX-007 spreads from cell-to-cell much more efficiently than wild-type adenovirus (serotype 5), and it is very effective in suppressing the growth of tumors in experimental animals following intratumoral injection of the vector.
Fig. 2. Oncolytic adenovirus vectors kill cancer cells through the natural process of virus replication. The figure shows infection of a cancer cell with a single adenovirus particle (upper left), replication to produce progeny virus particles, and the lysis (disruption) of the cell and the release of progeny virus.
Nearly all labs developing oncolytic adenovirus vectors perform anti-tumor efficacy studies in the human tumor xenograft-immunodeficient mouse model. Human cancer cells injected into the mice grow as tumors because the mice are immunodeficient. The adenovirus vectors replicate in the tumors and suppress their growth. The problem with this model is that tissues of the mouse are not permissive for human adenovirus replication and that the mice are immunodeficient; thus one cannot address the possible replication of the vector in non-tumor tissue or the role of immunity in vector toxicity and anti-tumor efficacy. We used the xenograft-nude mouse model in our early studies, but in 2006 we developed a new animal model, the Syrian hamster, to study the anti-tumor efficacy of VRX-007. The virtues of this model are that the hamster is immunocompetent and the hamster tumors and normal tissues are permissive for replication of VRX-007 and Ad5. Thus, the hamster may be a good model for what might happen when cancer in a human is treated with VRX-007. In the hamster, we showed that VRX-007 effectively suppresses the growth of subcutaneous tumors following intratumoral injection (Fig. 3). We further showed that the host immune response to VRX-007 reduces the vector anti-tumor activity, because when the hamsters were suppressed using cyclophosphamide then the vector suppressed tumor growth more effectively (Fig. 3). VRX-007 also suppresses the growth of intraperitoneal pancreatic tumors following intraperitoneal injection of VRX-007. The hamster has also been very useful in studying the toxicity of VRX-007 and Ad5 and the tissue biodistribution of these viruses following intravenous injection of the viruses.
Fig. 3. The oncolytic adenovirus vector VRX-007 suppresses the growth of subcutaneous tumors in Syrian hamsters following intratumoral injection of the vector. Some hamsters were fully immunocompetent and some were immunosuppressed using cyclophosphamide (CP). Tumor sizes were measured using calipers at different days postinjection of tumors with VRX-007.
Our studies in the nude mouse-xenograft and Syrian hamster models have provided compelling anti-tumor efficacy and low toxicity data that led to the approval by the FDA of a Phase I clinical trial for cancer using our vector named VRX-007. One patient with a chemotherapy refractory and non-resectable 3 cm diameter squamous cell carcinoma on the tongue has been treated so far by intratumoral injection of a low dose of VRX-007. No adverse events were observed. The tongue tumor nearly disappeared and it became fibrotic tissue instead. The patient no longer needs morphine for oral pain.
We have also pioneered the Syrian hamster as a unique animal model to study the pathogenesis of human adenoviruses and to evaluate the efficacy and toxicity of drugs to treat adenovirus infections. As mentioned, most tissues of the Syrian hamster are permissive for replication of adenovirus. Further, when the hamster is immunosuppressed, replication continues for long periods (weeks) in the liver, lungs, and other organs. We have used this model to demonstrate that the drug hexadecyloxypropyl-cidofovir (named CMX001) inhibits Ad5 replication in immunocompetent and immunosuppressed hamsters (Toth, K., et al., PNAS, 105, 7293-7297, 2008) (Fig. 4). Further studies are underway to evaluate the anti-adenovirus activity of a variety of other drugs. We anticipate that our studies will help advance these drugs through clinical trials and then eventually be used to treat adenovirus infections in humans.
Fig. 4. CMX001 decreases adenovirus serotype 5 (Ad5) replication and Ad5-induced lesions in the liver. Livers of hamsters were sacrificed at day 6 after intravenous injection of Ad5 and then subjected to histopathological and immunohistochemistry (IHC) evaluation to detect Ad5 replication in the liver. Animals infected with Ad5 and not treated with CMX001 exhibited extensive coagulation necrosis throughout the liver (A) and widespread replication of Ad5, demonstrated by IHC staining for fiber (an Ad5 protein) (B). Treatment of Ad5-infected hamsters with CMX001 resulted in a significant reduction in hepatocellular injury (C) and greatly reduced IHC staining for fiber (D). Compare the very extensive brown staining of cells in Panel B indicative of an active Ad5 infection in the liver versus the nearly complete absence of brown staining in Panel D indicative of inhibition of Ad5 infection by CMX001. The arrows indicate intranuclear inclusion bodies. (Scale bars: 200 _m for the larger images and 50 _m for Insets.) N, necrosis.