Wednesday, 23 October 2019



Before commencing clinical studies, batches of the vector must be made under FDA current good manufacturing practice (GMP) and satisfy rigorous lot release criteria. As with all drugs, the identity (now including full DNA sequencing), purity, and potency of the vector must be verified in validated assays. For safety reasons, the absence of RCA is a major concern and vector destined for human usage must contain <1 RCA per dose. As dose increases, the challenge of making vector free of RCA in the 293 cell line becomes more problematic. Vectors must also be verified for the absence of not only human viruses but also animal viruses including porcine and bovine viruses that may be carried through from reagents used in cell culture. Taken together, these studies impose a high hurdle of cost and time before having a vector suitable for human administration on hand. Most production is now done in dedicated facilities with high overhead to maintain the required trained personnel, facility integrity, and quality control units.

Human clinical studies also require prior toxicology studies in experimental animals assessing the effects of both vector and transgene expression. There is now considerable human data on the effects of administration of first-generation Ad5based vectors, so the effects of the transgene are usually the focus. Depending on the novelty of the application and the phase to which clinical studies have progressed, toxicology studies may need to conform to FDA good laboratory practice and may require nonhuman primates in addition to rodents. The information derived from toxicology studies is necessarily limited due to biological differences between experimental animals and humans, especially as it relates to the innate and acquired immune response. Experimental animals are naive to human Ads so the immune response to vector is primary while most humans have preexisting immunity to Ad5 and therefore the immune response is secondary. There are differences in antigen processing and presentation between inbred animals and humans and the effects of some E3 genes are limited to human MHC proteins. In addition, the potential effects of RCA are different since human Ads do not replicate in rodent cells but some human Ad is actually oncogenic in rodents but not in humans.

The final step of commencing a clinical study is obtaining regulatory approval which is also complex. Gene therapy, in general, and Ad, in particular, are generally perceived as being particularly hazardous and therefore multiple levels of review are required. The local institutional review board (IRB) and institutional biosafety committee (IBC) must approve as well as the recombinant DNA advisory committee of the National Institutes of Health. The investigational new drug application represents the last phase of approval. Each group has slightly different concerns and the final result is a protocol that is conservative yet likely to be safe, ethical, and informative. All revisions to the protocol must be approved by all regulatory groups and therefore keeping the protocol current needs well-organized management. The actual clinical study must be performed under good clinical practice (GCP) with timely reporting of all adverse events to the FDA, RAC, IRB, and IBC.

In addition, many institutions have added a data safety monitoring board to further ensure the safety of gene therapy studies. Data on models of disease in experimental animals provided a sufficient basis to proceed to Ad vector-mediated gene transfer in humans. The first human studies were commenced in 1993 and a total of 228 studies are listed in the current update of the database of the RAC. These have generally been small phase I or phase II studies. While the first studies were directed at CF, the currently active studies are focused on cancer and cardiovascular disease. In total, 75% of the listed studies involve cancer with the same vector being studied in a number of different tumors or different patient subsets or concurrently with different concurrent therapies. In general, capsid modified vectors and second-generation vectors have not been used in humans and it is unknown if the advantages these new vectors sometimes offer in experimental animals apply in humans.

In general, gene transfer of Ad vectors has been safe and well-tolerated. The notable exception is that intravenous administration results in dose-dependent toxicity that has resulted in one death. This reaction is rapid and is believed to be related to the innate immune system. In addition to showing safety, several studies have shown that there is also effective delivery of the vector to the patient and subsequent expression of the therapeutic gene. Demonstrations of actual therapeutic benefit will require critical placebo-controlled testing. If an Ad vector does prove to be effective in a large study, there remain substantial production and support issues that will require a significant investment of time and money to provide the reliable supply of a marketed drug.



When gene therapy is the conceived addition of a good copy of a defective gene, it is natural that the initial focus was on genetic diseases such as CF. As described above, the feasibility of intratracheal administration to the lung was demonstrated in animals and the first clinical trial of an Ad vector was to the airway epithelium of CF patients. The initial study was a dose-escalating safety study in which vector administered to the bronchi in 20 mL of fluid. It became clear that this volume was not well tolerated and so subsequent studies used smaller volumes, or a spray of an aerosolized vector into the bronchi. The relative accessibility of the site of administration allows that samples of respiratory epithelium can be recovered by bronchial brushing and the presence of vector and therapeutic gene expression can be assessed repetitively. By sensitive quantitative PCR methods, the expression of the CFTR gene delivered by the vector is seen at the site of administration at vector doses of 5 × 10 8 pfu and greater. The level of vector-derived CFTR mRNA is approximately 5% of the level of expression of the endogenous CFTR gene and this is believed to be sufficient for therapeutic effect. However, this level of expression is only achieved for a period of a few days and expression rapidly declines to baseline by 30 days. Interestingly, the administration of the vector to the airway does not lead to a significant immune response against Ad reflected in either neutralizing antibodies nor Adspecifi c T-cell proliferation.

Since the initial safety of vectors expressing CFTR was demonstrated, a study with repetitive administration has been completed. The important result of this study is that expression is reduced or eliminated in subsequent administrations as expected from the data from experimental animals, presumably from the immune response to the first dose. To date, only one study of adenoviral vectors for any metabolic disease has been initiated. This is not surprising since all animal data suggests only a short time of expression of genes delivered by Ad vectors to most tissues, and the rationale for a human study was not firmly established. Ornithine transcarbamylase (OTC) deficiency is a recessive metabolic disorder of nitrogen metabolism. An E1 − , E4 − deleted Ad vector expressing the cDNA for OTC was constructed and administered by the intrahepatic route to adults with partial OTC deficiency and safety parameters and the efficiency of gene transfer is currently being assessed. During this study, it became apparent that a large dose of the intravenous vector could be fatal. The use of a helper-dependent vector for the treatment of hemophilia by intravenous administration was stopped when toxicity was observed.



Due to the unknown safety profile of Ad vectors, it is generally easier to design the early human trials for life-threatening diseases. Of the protocols listed by the RAC, 75% involve cancer. Four basic approaches can be identified: local prodrug activation, tumor suppressor genes, immunotherapy, and oncolytic viruses.

One of the first strategies of human gene therapy for cancer was to locally deliver novel enzymes that metabolize prodrugs into the active chemotherapy agent. The general concept of the studies is that local activation of the prodrug in the tumor will concentrate the active agent in the tumor, thus limiting the systemic toxicity from the active drug. Two genes have been used in human clinical trials: the HSV-TK and the E. coli CD gene. The HSV-TK protein activates the prodrug ganciclovir to ganciclovir monophosphate, an inhibitor of DNA polymerase. For CD, the prodrug is 5-fl Uorocytosine, which is activated by CD into the active chemotherapeutic agent 5-fluorouracil. For both agents and activating enzymes, a theoretical benefit is a bystander effect in which active drug would be excreted from the vector-infected cells to kill the neighboring cells of the tumor. Thus, it is not essential to infect every cell of the tumor with the Ad vector.

Currently, active protocols apply the prodrug strategy to many types of cancer including prostate cancer, CNS malignancies, ovarian cancer, mesothelioma, hepatic metastases of colon cancer, and squamous cell carcinoma of the head and neck. Most studies involve phase I/II study with intratumoral injection of escalating doses of vector prior to chemotherapy with the prodrug and subsequent scheduled surgery. Tumor removal provides samples for analysis of vector levels, for expression of the therapeutic gene, and for analysis of the activation of prodrug and histological studies for cell death and inflammation. The primary endpoint of these studies is safety, which has been established in some cases.

Tumor suppressor genes have also been used in human clinical studies: p53 (for ovarian cancer, prostate cancer, squamous cancer of the head and neck, breast cancer, nonsmall cell lung cancer, hepatic carcinoma, and hepatic metastases), retinoblastoma susceptibility gene (for bladder cancer), mda7 gene for melanoma, and p16 gene for prostate cancer. The concept is that tumor cells have defective tumor suppressor genes that cannot limit cell division, but the restoration of the wild-type gene will limit cell division. The theoretical limitation of using anti-proliferative genes for tumor therapy is that they will only inhibit the proliferation of the cell they infect and have no cis effect on neighboring cells. The trial designs are generally similar to those for the prodrug strategy.

A third general approach to Ad gene therapy for cancers has been immunostimulatory genes Several different genes have been used in human studies, including CD40 ligand for chronic lymphocytic leukemia, granulocyte-macrophage colony-stimulating factor for melanoma and non-small cell lung cancer, interleukin-2 for neuroblastoma, MART-1 (a melanoma-specific antigen) and B7 (CD80) for melanoma. The concept of immunostimulatory gene therapy is to promote the natural immune surveillance and elimination of tumors, which express abnormal antigens by giving a general boost to the cellular immune system (e.g., with IL-2) or with a tumor-specific antigen (e.g., MART-1) and anti-erbB-2 single-chain antibodies (for ovarian cancer).

A novel anti-proliferative approach has been used in human studies using conditionally replication-competent viruses. As described above, the E1B gene is essential to protect Ad infected cells from apoptosis and its mode of action is through interaction with p53. It follows that E1B function would only be effective in p53 positive cells, but not in p53 deficient tumor cells, i.e., E1B positive viruses would induce apoptosis only in p53 negative tumor cells, while normal cells should be protected by the p53 gene. In this context, E1A negative, E1B positive viruses have been demonstrated in animal models to show selective cytolytic effects against tumors. The same viruses have been used in phase I and phase II studies of human ovarian cancers, pancreatic cancer, and head and neck cancer with direct intratumoral injection in conjunction with chemotherapy. These studies are now being extended to phase III testing.

The simplest approach to therapy is to directly administer the vector to the tumor. This is complex due to the need to protect the cells cultured and infected ex vivo from adventitious agents while they are outside the patient. Three approaches have been used. The most common is to infect the tumor cells with an immunostimulatory vector (e.g., Ad expressing a cytokine) expecting to evoke an anti-tumor immune response upon returning the cells to the patient. Alternatively, dendritic cells, a potent antigen-presenting cell, have been infected ex vivo with Ad expressing a known tumor antigen to evoke an immune response upon returning to the donor. Finally, bone marrow-derived stem cells have been purged of possible contaminants by infection by an antiproliferative vector prior to returning to a donor.



With the observation that there is only short-term gene expression from Ad vectors, the question arose as to which medical applications might benefit from the transient expression of a therapeutic gene. The general area of tissue repair and engineering emerged as a good candidate where secreted growth factors would initiate the desired cascade of tissue remodeling which, once initiated, would not require the continuous presence of the therapeutic gene. For example, expression of vascular endothelial growth factor (VEGF) after injection of an Ad vector expressing VEGF into rat retroperitoneal fat pad is brief, reverting to baseline after 10 days. By contrast, the VEGF protein induces an angiogenic response that persists long after the stimulus has disappeared.


FIGURE. Time course of gene expression and anatomical

response after administration of Ad expressing VEGF. The retroperitoneal fat pad of rats was injected with 5 × 10 8 pfu of either AdVEGF (a first-generation E1 − E3 − vector expressing the 165 amino acid form of human VEGF, solid symbols) or the control vector expressing no transgene (AdNull, open symbols). At intervals, animals were anesthetized, a laparotomy performed, and the fat pad photographed. The number of vessels crossing a circle of 1 cm diameter centered on the injection site was measured (left axis: ®, n). The fat pad was also homogenized and the level of VEGF determined by ELISA (right axis: °, l.)

On the basis of extensive preclinical testing in animal models of angiogenesis, a number of groups performed clinical studies of adenoviral gene transfer to the ischemic heart or ischemic limb in an effort to induce the growth of new vessels and to increase blood flow. While there is no rigorous efficacy data, there is preliminary evidence of improved cardiac function in a number of patients receiving either Ad expressing FGF-4 in the coronary arteries or Ad expressing VEGF121 in the myocardium. Moreover, when injected into the muscles of patients with peripheral vascular disease Ad expressing VEGF resulted in marginal improvements in some parameters of blood flow to that limb.


The early Ad gene therapy trials demonstrated that, while effective gene transfer could be achieved, the persistence of expression is clearly a problem for Ad vectors. Although there was clearly an immune response to the vector and possibly the transgene itself, the biology of that response is not well understood. Animal models, that is particularly those involving inbred mice, have limited utility in predicting the immune response in humans. To assess the human–host response to Ad vectors, two trials with normal subjects have been performed using intradermal or intratracheal administration of an Ad vector expressing the E. coli CD gene. The intent of these trials is to describe the immune response in humans to an E1 − , E3 − Ad vector to provide a background to assess more advanced vectors on a rational basis.

No comments:

Post a comment