Originally published as JCO Early Release 10.1200/JCO.2004.12.964 on January 15 2004

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Albelda, S. M. Articles by Sterman, D. H. Search for Related Content PubMed PubMed Citation Articles by Albelda, S. M. Articles by Sterman, D. H. Related Articles Related Article Social Bookmarking                            What's this?

TNFerade to the Rescue? Guidelines for Evaluating Phase I Cancer Gene Transfer Trials

Thoracic Oncology Program, University of Pennsylvania Medical Center, Philadelphia, PA

In this issue of the Journal of Clinical Oncology, Senzer et al [1] describe the results of their phase 1 gene transfer trial of TNFerade. The goal of this trial was to take advantage of the well-known antitumor effects of tumor-necrosis factor alpha (TNF-), while avoiding the considerable systemic toxicities that currently limit its use. This was accomplished by injecting tumors with a replication deficient adenovirus encoding the human TNF- gene driven by a radiation-inducible promoter, followed by treatment with external beam radiation. The trial showed that repeated intratumoral injection of TNFerade in combination with external beam radiation was safe. A significant proportion (43%) of patients demonstrated objective tumor responses, although it is unclear how many responses would have been achieved with radiation alone. In some cases, it was possible to compare the size of "control" lesions that were only radiated with those treated with TNFerade plus radiation in the same patient. For the most part, lesions treated with TNFerade and radiation showed more response than these "control" lesions.

Before discussing this study, we should comment on our choice of describing this study as a gene "transfer" trial rather than gene "therapy" trial. The US Food and Drug Administration (FDA) and the National Institutes of Health have concerns that references to gene therapy trials, particularly in the phase I stage, implies a therapeutic benefit from the transfer of genetic material that has not yet been proven. This may pose a problem for the consent form in these studies, particularly for patients with advanced cancers who are desperate for novel and more effective treatments. Accordingly, we recommend the use of the terminology gene transfer or gene delivery, until well-designed controlled studies have clearly demonstrated therapeutic efficacy.

Is this a good phase 1 cancer gene transfer study? Although, for the most part, this trial should be evaluated like any other standard phase I trial of an experimental antineoplastic agent, there are additional criteria that should be considered when evaluating a cancer gene transfer trial.

Any good clinical study must meet certain well-accepted standards. Patient inclusion and exclusion criteria must be clearly stated and explained. The trial protocol must be logical and understandable. Dose-limiting toxicities should be defined. A reasonable number of patients should be studied, and the numbers of patients who dropped out of the study and the reasons for these dropouts should be stated. All regulatory and informed consent issues must be addressed. In gene transfer trials, in addition to standard institutional approvals, review by the Recombinant DNA Advisory Committee and the FDA is required. The TNFerade trial performed well in all these areas.

Good clinical studies need to begin with a well-reasoned hypothesis and a thorough understanding of the limitations of the therapeutic agent. Given the current limitations of cancer gene therapy vectors, this latter point is especially important. All of the currently available vectors (ie, adenovirus, herpes simplex virus, retrovirus, lentivirus, adeno-associated-virus, and liposomes) have the ability to transduce only a small minority of cells within a tumor, and none is currently effective in targeting tumor after systemic administration. To achieve any sort of efficacy, these limitations require that gene therapy agents induce strong "bystander effects." This term refers to the ability of a transgene, expressed in transduced cells, to affect other nontransduced cells, both neighboring and distant. Examples of transgenes with bystander potential include: "suicide" genes, such as cytosine deaminase or herpes simplex thymidine kinase, in which a toxic product is secreted or moves from cell to cell; genes encoding for secreted proteins that induce death of tumor cells or tumor vasculature, such as antiangiogenic compounds; genes that stimulate immunologic effects, such as cytokines; or genes encoding for molecules that, if expressed on the tumor, could trigger a larger overall immune response, such as T-cell costimulatory molecules or CD40-ligand. The transgene used in this trial, TNF-, clearly has the potential for bystander effect from a number of mechanisms. The use of gene therapy to treat local tumors by intratumoral injection is also reasonable and consistent with the limitations of early generation adenoviral vector systems.

Strong preclinical data are necessary for a gene transfer trial. In addition to defining the mechanisms and limitations of the approach in relevant animal models (which allows much better planning of the phase I and later trials), showing efficacy in animal tumor models is required to convince regulatory agencies to approve and funding agencies or industry to finance the trial. This trial receives high marks in this area. It was well supported by a number of good preclinical studies showing efficacy in multiple human xenograft tumor models. The one deficiency in the background preclinical data described by the investigators is the lack of efficacy testing in a syngeneic, immunocompetent animal tumor model. For a transgene like TNF- with significant potential for immunostimulation, in addition to its antiangiogenic effects, testing in an immunocompetent animal model would have been interesting in further determining the likelihood of efficacy in humans, as well as elucidating mechanisms of antitumor effects.

As in any other phase I trial, evaluation of safety and establishment of a maximally tolerated dose are primary goals. This requires preclinical toxicology data in an appropriate animal species. For practical reasons, such as trying to separate effects of vector from those of the tumor, most of these studies are usually performed in normal rodents, not tumor-bearing animals. Depending on how novel or how potentially dangerous a specific approach might be, the FDA may or may not require studies in nonhuman primates. For this trial, the FDA did not require such studies. The TNFerade trial investigators have published their preclinical toxicology studies [2]. These studies revealed a large therapeutic window with low serumTNF- levels and dose-limiting toxicity (local ulceration and necrosis) at a relatively high dose of virus. Thus, the investigators had a solid basis to choose their dose escalation schema and to predict toxicities.

The investigators of the trial carefully spelled out the dose escalation schedule and dose limiting toxicities, which, for the most part, were characteristic of a standard phase I trial design. However, establishment of a maximally tolerated dose (MTD) can be difficult in a gene transfer trial for a number of reasons. First, as in this trial, MTD may not be reached at the highest dose planned. Doses may also be limited by reaching a "maximal affordable dose" if large amounts of virus are being used. A second issue is that the highest dose administered may not be the most effective dose. This is especially relevant when using a biologically active transgene, such as TNF-. This may have been a problem in this study, as we will discuss later, as evidenced by the fact that the investigators took the unusual step of adding two lower doses of vector after completion of the five initial dose levels.

In terms of safety, cancer gene transfer trials are probably held to a higher standard than conventional chemotherapy trials. The reasons for this are multifactorial, but may relate to a general public unease with "genetic engineering" and a few highly publicized deaths in other gene transfer trials. This close scrutiny makes conduct of phase I studies especially difficult, because the cancer patients enrolled onto most cancer gene transfer trials have advanced disease and the concomitant medical problems that accompany end-stage cancer. In this trial, however, toxicity was very low. Nonetheless, the authors do a very thorough job of listing every adverse event and toxicity.

Although dose finding and safety are the stated goals of phase I trials, in actuality, the best phase I cancer gene transfer trials need to do more. Given our very limited knowledge about the efficiency of gene transfer in humans, gene transfer trials should try to incorporate methods to assess transgene expression. Without knowing if a vector has effectively transduced the intended cells and produced the transgene in clinically useful amounts, it is very difficult to predict if the approach chosen will work as hypothesized and how the next phase I or phase II trial can be designed. This information is especially important given that vectors such as adenoviruses have transgene-independent biologic effects that may not be readily discernible in the absence of "empty-vector" control groups. Although serum TNF- levels were assessed in the TNFerade trial, there was no data quantifying the presence and amount of transgene produced in the vector-injected tumors. This determination would have been somewhat invasive, likely requiring biopsies for subsequent immunohistochemistry or molecular analysis. However, without these data, we cannot answer key questions: (1) Did viral transduction of tumor cells or other cells in the tumor stroma occur? (2) Was there also gene expression? (3) If so, how much and where? (4) what was the mechanism of the observed antitumor effect? Was there increased apoptosis, were inflammatory cells recruited, or was there vascular necrosis? (5) Did tumor response correlate with gene transfer? It is easy to see how this information would have strengthened this study and would have been very useful in planning future approaches. Given that repeat tumor biopsy for gene transfer studies in cancer increases the risk of the protocol, several investigators are evaluating noninvasive means to evaluate and confirm gene transfer, including the use of positron-emission tomography.

Finally, although not the stated purpose, phase I cancer gene transfer trials are almost always scrutinized for some evidence of antitumor efficacy. This is somewhat difficult, becasue responses are sought in small, nonrandomized, noncontrolled trials in highly selected end-stage patients. However, if no evidence of efficacy is demonstrated (at least at the higher doses), it is unlikely that phase II trials will be funded and performed. In most cases, investigators make comparisons of response rates based on historical controls, while looking for consistent partial or complete response rates. In this trial, there were a number of responses, including some with complete resolution of the radiated tumor. Were these responses due to the TNF gene transfer? One factor that might argue against this was the lack of any clear relationship between the dose of vector and response. This important point was not mentioned by the authors and probably should have been. In favor of possible efficacy of TNFerade were the five patients that served as their own "internal controls," with multiple lesions that were radiated, but with only one lesion receiving TNFerade and radiation. In four of five patients, the lesion treated with gene therapy and radiation responded better than the lesions treated only with radiation.

Was this a perfect gene transfer trial? The answer is, probably not. The actual amount of gene transfer and transgene expresssion was not assessed, leaving questions about how well the radiation-inducible promoter actually worked. A clear relationship between gene transfer and dose was not established. No true MTD was reached. On the other hand, there were many positive aspects of this trial. A relatively large number of patients were successfully studied, safety and feasibility were established, and some interesting responses (especially among the "internal control" patients) were seen. Certainly, the approach was safe. Whether there is enough evidence to justify moving to larger, controlled phase II trials is a question that the readers, the investigators, and, most importantly, the sponsors, will have to answer. Given the low level of toxicity, lack of alternative therapies, and intriguing hints of efficacy, we believe further investigation (hopefully with good controls and with gene transfer assessment) is warranted.

Authors' Disclosures of Potential Conflicts of Interest

The authors indicated no potential conflicts of interest.

REFERENCES

1. Senzer N, Mani S, Rosemurgy A, et al: TNFerade biologic, an adenovector with a radiation-inducible promoter, carrying the human tumor necrosis factor alpha gene: A phase I study in patients with solid tumors. J Clin Oncol 22:592-601, 2004[Abstract/Free Full Text]

2. Rasmussen H, Rasmussen C, Limpicki M, et al: TNFerade Biologic: Preclinical toxicology of a novel adenovector with a radiation-inducible promoter, carrying the human tumor necrosis factor alpha gene. Cancer Gene Ther 9:951-957, 2002[CrossRef][Medline]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

Related Article

• TNFerade Biologic, an Adenovector With a Radiation-Inducible Promoter, Carrying the Human Tumor Necrosis Factor Alpha Gene: A Phase I Study in Patients With Solid Tumors
  Neil Senzer, Sridhar Mani, Alexander Rosemurgy, John Nemunaitis, Casey Cunningham, Chandan Guha, Natalia Bayol, Michelle Gillen, Karen Chu, Camilla Rasmussen, Henrik Rasmussen, Donald Kufe, Ralph Weichselbaum, and Nader Hanna
  JCO 2004 22: 592-601 [Abstract] [Full Text]

This article has been cited by other articles:

				A Agrawal, E Gutteridge, J M W Gee, R I Nicholson, and J F R Robertson Overview of tyrosine kinase inhibitors in clinical breast cancer Endocr. Relat. Cancer, July 1, 2005; 12(Supplement_1): S135 - S144. [Abstract] [Full Text] [PDF]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Albelda, S. M. Articles by Sterman, D. H. Search for Related Content PubMed PubMed Citation Articles by Albelda, S. M. Articles by Sterman, D. H. Related Articles Related Article Social Bookmarking                            What's this?