On the Horizon:
Gene and Virus Therapies for Sarcomas

An ESUN Article by Timothy P. Cripe, MD, PhD

Introduction

Can you fight fire with fire? Can one human scourge be used to treat another? Can the enemy of my enemy be my friend? When it comes to biological therapies for cancer, there is a growing number of scientists, doctors, and biomedical companies hoping the answer to these questions is "Yes!" People have been thinking about using infections to treat cancer for a long time. In 1893, William B. Coley, MD, an assistant surgeon to the "Hospital for Ruptured and Crippled" in the Medical School of New York, published a paper in the American Journal of Medical Sciences entitled "The Treatment of Malignant Tumors by Repeated Inoculations of Erysipelas, with a Report of Ten Original Cases" (1). In the report Dr. Coley described his experience using direct injection of bacterial cultures into sarcoma tumors. The bacteria had been grown from patients suffering from the Erysipelas infection, a skin infection caused by streptococcus bacteria. Several of the patients experienced cures of their cancers. Although he did not know the exact cause of the tumor regression, Dr. Coley postulated it was due to bacterial toxins, and the therapy became known as "Coley’s toxin."

Over 113 years later we still don’t know the exact mechanisms of tumor regression in Dr. Coley’s cases. Nevertheless, we now know that infectious agents such as bacteria, yeast and viruses can directly infect and kill tumor cells. We also know that these agents can cause stimulation of the immune system and sensitize it to tumor-specific antigens, causing an antitumor immune response. Based on Dr. Coley’s early encouraging results, over the past century physicians have attempted to use bacteria and viruses in order to treat patients with cancer in various venues and in fact have had some tumor responses. But much of the work in the middle of the 20th century was abandoned due to toxicities from these infectious agents. Incomplete understanding of the nature of infections, lack of knowledge about the immune response, and an inability to tinker with either infections or host hampered progress. Over the past several decades, however, advances in our understanding of infectious disease and the reaction to infection by the host immune system as well as technical advances in our ability to manipulate genes and change the genetic makeup of viruses and bacteria have led to a strong resurgence of physicians and scientists examining these potential types of cancer biotherapy.

What is Gene and Virus Therapy?

Gene therapy is the use of nucleic acids to modify a cell. This can include the use of Deoxyribonucleic Acid (DNA) or Ribonucleic Acid (RNA). The DNA or RNA can be comprised of an entire virus genome (i.e., the entire genetic makeup of the virus, be it DNA for a DNA-type virus, or RNA for a RNA-type virus; these can range from 5,000 to 150,000 genetic building blocks, or base pairs), a small plasmid of DNA (a short, circular molecule usually in the range of 1,000 to 10,000 base pairs), or just a chemically synthesized series of base pairs (oligonucleotides) which can be made of DNA, RNA, or RNA:DNA hybrids (ranging in the 10’s to 100’s of base pairs). Virus therapy is the use of a virus to lyse (kill) a cell. Because this type of therapy is usually used to infect and kill cancer cells, this process has been dubbed "oncolysis" (onco meaning cancer, lysis meaning break open).

There are several methods of getting genes into cells in order to change the cells’ behavior. These include the use of viruses, liposomes, glycoproteins or polycations that are able to bind up DNA, direct injection of "naked DNA" that has not been wrapped with or bound to proteins or liposomes, and finally the gene gun, which uses gold beads coated with nucleic acid. Virus-mediated delivery is quite popular because it is very efficient: viruses have evolved over millions of years to very efficiently insert their genetic makeup into cells. Many different virus types have been explored for gene delivery, most commonly retrovirus, adenovirus, adeno-associated virus, and herpes simplex virus.

The advantages of using genes as therapy are many. First, it is relatively easy to mass produce DNA or viruses, because the factories are billions of bacteria or yeast or eukaryotic (primate or human) cells grown in a culture that make the copies, instead of requiring extensive vats of chemicals for synthesis of a small molecule drug. In general, DNA and viruses are also easier to purify than proteins. Also, it is relatively straightforward to alter the DNA sequence using genetic engineering, so that different versions can be created that have new properties. The transfer of genes can also result in prolonged production of a new protein, obviating the need for frequent drug injections.

Why for Cancer?

The use of gene and virus therapy has shown great promise in affecting cancer cells in several ways. First, these strategies may modify the cancer cell directly by the expression of new genes in the cell (or by killing the cell, in the case of oncolytic viruses). They can also induce the cellular machinery for programmed cell death (termed apoptosis), or induce differentiation of tumor cells so that they are no longer malignant. Second, they can be a means to modify other cells around the infected cell. For example, gene or virus therapy can stimulate immunity and induce an anticancer immune response. In addition, genes can be used to protect normal cells from cytotoxic therapies, thereby allowing an increase in doses of these drugs. Third, because cancer is fundamentally a genetic disease, gene-based therapy might be able to correct the genetic defects. For example, many cancers are defective for a protein called p53, and the re-introduction of a normal p53 gene using adenovirus has been tested in clinical trials. With the use of genes and viruses it is possible to exploit the genetic alterations to selectively target cancer cells compared with normal cells. For further review of possible ways to modify genes and possible gene targets in a variety of different cancer types, please see the review by Cripe and Mackall, 2001 (2).

How Do You Hit Every Cell?

One of the limitations of the use of the gene transfer for cancer is the need to affect every single tumor cell. If one tumor cell is left unaffected the tumor may regrow and not be cured. This is difficult with "single pass" gene transfer technology, where the only cells affected are those initially exposed to the gene or virus. This limitation has been overcome using two major strategies: (i) a bystander effect and (ii) the use of replicative viruses. The most common way to induce a "bystander effect" has been to use a gene that results in a toxic product that can be transferred from one cell to the next. For example, the thymidine kinase gene from herpes simplex virus is a gene that activates the drugs acyclovir and ganciclovir. When this gene is introduced into a cell and the drug is administered, that cell is killed after the activation of the drug. The activated drug can also pass to neighboring cells via intercellular channels known as gap junctions. In this way, cells surrounding the one that initially received the gene or virus (bystander cells) can be affected. There are other examples as well where cell-to-cell transfers via secretion of the metabolite to adjacent cells can result in this bystander effect. In addition, the inhibition of what’s known as a paracine pathway, such as insulin-like growth factor (IGF)-2 can potentially affect more cells than just the cell that was hit directly. Another example is the induction of an immune response to tumor antigens that can then affect other cells expressing those antigens. Finally, the last example of a bystander effect is the use of genes or viruses that affect the new blood vessel formation found in cancers (angiogenesis). By decreasing angiogenesis, other cells that rely on the new blood vessels can also be affected.

The second major technique used to overcome the hurdle of needing to hit every cell is the use of replicative viruses. These oncolytic viruses propagate themselves after infecting a cell and spread to the next cell and the next cell and the next and so on (Fig. 1). They have been engineered to selectively replicate and produce themselves in tumor cells but not in normal cells. These intriguing viruses are discussed in more detail below.

Viruses as Cancer Therapy

Viruses are proving to be useful in multiple ways as agents for cancer therapy. As mentioned above they can be used as a vehicle for gene delivery. In addition, it’s been shown in a number of animal models they are capable of inducing an antiviral or anticancer immunity. Just the direct injection into a tumor can sometimes sensitize the immune system to tumor specific antigens, thereby functioning as an "in situ" cancer vaccine. Some virus proteins, such as adenovirus E1A, have also been shown to sensitize cells to chemotherapy. Therefore, their use in combination with chemotherapy may be beneficial. Finally the direct infection of cells via virus replication appears to be a major use of viruses for treating cancer. Advantages of direct cell lysis are that in most cases these avenues of cell death are not subject to the usual mechanisms of resistance that are found for chemotherapy. The virus can spread from cell-to-cell and be made selective for cancer cells. Finally, these viruses can also be "armed" with other therapeutic genes to affect the cancer growth in other ways than just virus infection. These viruses utilize both gene and virus therapy in a single package.

How Can Viruses be Made to be Selective for Cancer Cells Compared with Normal Cells?

There are a number of different mechanisms of cancer cell selectivity for virus infection.

First and foremost is the natural metabolic environment of a cancer cell, which is very favorable to virus replication. This is particularly true of large double stranded DNA viruses such herpes simplex viruses (HSV). These viruses require a large machinery to replicate their DNA over and over and over again. All this machinery is well expressed and up regulated in most cancer cells, but not often present in normal, non-dividing (quiescent) cells.

Viruses can also be mutated or can be manipulated to become more selective for a cancer cell. As an example, HSV encodes an number of proteins that are required to replicate its DNA, such as thymidine kinase, ribonucleotide reductase, and DNA polymerase. Because these proteins are not normally up regulated in normal, non-dividing cells, the virus must provide them. Deletion of these genes from the virus therefore renders these viruses very ineffective in infecting normal cells. However, when these viruses enter a cancer cell where the DNA replication machinery is already in place the viruses can replicate well. Viruses containing this form of simple genetic deletions or mutations of viral genes to make them cancer selective are called type 1 viruses. Examples of type 1 viruses are those gene mutations that actually exploit defects in the cancer cell pathway. For example, in adenovirus the E1A protein binds and inactivates the retinoblastoma protein in a cell in order to have the virus replicate. Deletion of the E1A gene makes the virus nonreplicative because it is no longer able to inactivate retinoblastoma protein. In cancers that have retinoblastoma protein deleted, however, the virus can replicate without difficulty. Similarly the E1B gene of adenovirus inactivates the p53 protein in order to achieve maximum virus replication. Again in cells that are defective for p53 adenoviruses containing E1B deletions can replicate because they no longer need to inactivate p53.

Finally, critical virus defense pathways are often defective in cancer cells. For example, the RNA induced protein kinase "PKR" is activated upon virus infection and shuts down protein synthesis in a normal cell, which prevents the virus from making more of itself and spreading to the next cell (Fig. 2). The herpes simplex virus has devised a mechanism to counteract that defense pathway using the so-called ICP35.5 protein. When the gene encoding this protein is deleted from a virus it can no longer infect normal, quiescent cells. It turns out that the PKR pathway is frequently abnormal in cancer cells. Therefore the ICP35.5 protein is not needed for virus replication in cancer cells. Thus ICP34.5 deleted herpes simplex viruses are often able to replicate quite well and kill cancer cells but not normal, quiescent cells.

The other major category of viruses that have been engineered to be cancer selective are so called type 2 viruses. These are viruses where critical viral genes have been placed under the control of a cancer selective promoter such as alpha-fetoprotein carcinoembryonic anogen or other genes that are known to be selectively upregulated in tumor cells but not in normal cells. Thus when a virus enters a normal cell, the promoter is not turned on and the virus is not able to replicate, but when it enters a tumor cell the promoter is active, the critical protein is produced and the virus is able to replicate.

Oncolytic HSV

One of the most promising viruses for sarcomas is herpes simplex virus. We have shown that adenovirus does not infect most sarcomas well due to a poor expression of the receptors for adenovirus (3), with the exception of Ewing sarcoma cells (4). Recent studies suggest that different strains and newly engineered adenoviruses are more effective for sarcoma cells, renewing hope that these "next generation" versions may be useful for soft tissue sarcomas (5). In contrast to our results with adenovirus, we found that herpes simplex virus is able to enter and lyse most sarcoma cells quite readily. See these references for more detail (6-10).

Herpes simplex viruses have been shown to have a safe clinical profile by a variety of routes of infection in rodents and primates. Many different cancer types have been shown to be susceptible to HSV infection in addition to sarcomas (liver, breast, colon, pancreas, prostate, squamous cell carcinoma, ovary, and bladder cancers). In many models, direct intratumoral injection appears to induce an antitumor effect and in several model types it has been shown to act as an in situ vaccine, preventing the development of subsequent tumors (11). Intravenous therapy can also result in an anticancer effect at distant sites, suggesting that these agents may be useful for the treatment of metastatic disease (12, 13). Most people have been exposed to HSV type 1 and therefore have immunity to the virus. This preexisting immunity does not seem to be a barrier for delivery of virus to tumors and for its anitumor effect.

There have been several clinical trials demonstrating the use of these viruses as safe and potentially effective as treatment for cancer (Fig. 3).

The most extensively studied Herpes Simplex Virus mutant is HSV1716, which was created by S. Moira Brown and is being tested by Crusade Laboratories in England. In their first phase I study (14) nine patients aged 22-65 were treated with increasing doses of HSV1716 directly injected into their brain tumors. No patients experienced clinical signs or symptoms of encephalitis and there was no reactivation of latent HSV. The same group also injected five patients with stage IV melanoma directly into their tumor. Two patients received one injection, two patients received two injections and one patient received four injections. Again no patients experienced signs or symptoms of HSV infection or reactivation of latent HSV. In the patient who received four injections, two of which were in two different nodules, both nodules flattened and one of those that was resected showed a striking degree of tumor kill. Therefore it appears that HSV is safe not only to directly inject into brain tumors but also into peripheral tumors in the skin. In another study (15), patients with brain tumors ages 38-64 were given a single injection. Again the injections appeared to be safe and there was evidence in resected tissues of replication of HSV. Finally (16), 12 patients ages 33-66 with brain tumors were given direct injection into the tumors and there was no evidence of toxicity associated with the virus. There was evidence of clinical improvement in some patients and at the time of the report 5 of the 12 patients were still alive at time points beyond what would otherwise have been expected. These patients were immunocompromised, based on prior therapy, and there was no evidence of toxicity from the virus. These data suggest that viruses such as HSV1716 are safe for injection into tumors in people. In fact, there is now an ongoing phase III clinical trial in Europe for patients with brain tumors (www.crusadelabs.co.uk). The main challenge will be how to use these new therapies in combination with other standard cancer therapies (Fig. 4). As described in the legend to Fig. 4, the most likely use will be as a form of "biologic surgery" whereby the viruses would help shrink tumors that are not amenable to surgery or radiation. It is also possibly they will be useful for treating metastatic disease as has been shown in some animal models.

In our laboratory studies we have found that HSV1716 efficiently infects and lyses sarcoma cells in culture. We tested cells derived from rhabdomyosarcoma, Ewing’s sarcoma, osteosarcoma, neuroblastoma, and malignant peripheral nerve sheath tumors. Our data on sensitivity comparing this virus with two other herpes simplex virus mutants that are currently in clinical trials is shown in Table 1. In the table, each "+" sign represents a 10-fold increase in susceptibility of the cancer cells to killing by the respective virus. This is measured by determining the concentration of virus that inhibits growth of cells to 50% compared to an uninfected cell culture (inhibitory concentration 50, IC50). Our hope is to be able to test this virus in patients with sarcomas, though reaching that goal has so far proven elusive.

Day 6 IC50 MOI >1 (-), 0.1-1 (+), 0.01-0.1 (++), 0.001-0.1 (+++), 0.0001-0.001 (++++).

Translating Virus Therapy From the Laboratory to the Clinic

There are a number of steps required to launch a clinical trial, particularly of a novel biologic therapeutic (Fig. 5).

First, one needs a "clinical grade" product produced under so-called "current good manufacturing procedures" (cGMP). This type of production usually requires a biotechnology or pharmaceutical company, but some academic centers have built cGMP facilities. The cost for production of cGMP materials is usually on the order of $1 million or more, even for a small clinical trial. The National Institutes of Health also sponsor two programs that will produce cGMP vector for investigators, for those who are chosen from a series of applicants. Next, the cGMP product must undergo formal toxicology testing, which includes giving the agent at different doses to a variety of different animals, testing the effects on their growth, behavior, and blood counts, and the effects on all their organs by close examination of tissues under the microscope. The cost of toxicology testing is usually in the $250,000-$500,000 range. The third step involves the regulatory approvals, encompassing review by local and national committees. Local approval usually includes a scientific review committee (SRC) as well as the institutional biosafety committee (IBC) and the institutional review board (IRB). Each of these committees typically find issues or concerns that need to be addressed, often prompting changes to the clinical trial. Changes must go back to the other committees for their approval as well, so the process is multiply reiterative. At the federal level, the National Institutes of Health Recombinant Advisory Committee (RAC) must review all protocols that involve gene delivery. In addition, an investigational new drug (IND) application must be filed with the Food and Drug Agency (FDA) in order to launch a clinical trial. The requirements by the FDA to prove safety in animal models prior to testing in humans is quite extensive, having been increased over the past several years due to some adverse events in early clinical trials. (Although the FDA requirements are grounded in a true desire to assure safety, my personal opinion, based on my own experience, is that the requirements can sometimes be unreasonable. The bar should be lowered considerably when considering trials for patients who have exhausted all other options.) For the typical new drug, these steps are usually performed by a pharmaceutical company. For target diseases with a reasonably high projected market value, companies are willing to incur these expenses with the expectation of eventual profit. In the absence of a company, however, these steps often represent major hurdles for investigators at academic centers. Thus the clinical translation of findings in the laboratory often takes years, and sometimes these hurdles prove too burdensome even for the most determined investigator in the absence of adequate resources.

Conclusions

The use of viruses and genes holds great promise for future cancer therapy. There are many different exciting and novel approaches being tested in a variety of cancer models, including sarcomas, by many different investigators. The use of a virus has potential for multiple different mechanisms of action against cancer in a single agent. The hurdles for getting these new therapies into the clinic for testing in human patients are many, some of which are arguably unduly excessive and needlessly inhibit progress. Nevertheless, there are several ongoing trials currently sponsored by drug or biotechnology companies (though none yet for patients with sarcomas). My firm belief is that clinical trials testing genes and viruses are just beginning, and these new treatments – whose origins date back to the 1800’s -- will play an important role in near-future cancer therapy.