Federal Agency Roles in Cancer Drug Development from Preclinical Research to New Drug Approval: The National Cancer Institute and the Food and Drug Administration (2005)

The National Cancer Institute (NCI), since it began work in the 1930s, has played a central role in cancer research worldwide, including the research and development enterprise for new cancer drugs in the United States. Pharmaceutical companies were reluctant to enter the oncology market until the drugs developed by NCI began to demonstrate that cancer was, in fact, treatable and, in some cases, curable, although they played their usual roles in producing and marketing cancer drugs developed by NCI. Today, even though a few large pharmaceutical companies and a host of smaller pharmaceutical and biotechnology companies are engaged in cancer drug development, the role of the NCI remains large, commensurate with its current $4–5 billion annual budget (see NCI section below).

Over much of the same span of time, the Food and Drug Administration (FDA) has been responsible for approving anticancer drugs and (later) biologics, for safety and efficacy. With advances in cancer science and commensurate advances in cancer treatment, the FDA’s contributions to standards for preclinical and clinical testing and approval of therapeutics for cancer patients have grown. Encouraged by the promise of new cancer therapies (and to some extent by external influences), the agency has sharpened its focus on cancer organizationally as well as scientifically (see FDA chapter below).

Prior to the identification of cancer genes (oncogenes), tumor suppressors, and other factors that act in complex molecular pathways to control cell physiology and proliferation, cancer drug discovery occurred in at least three important ways. It could be fortuitous, stemming from unexpected findings in the course of studies not directed toward controlling the disease; it was often highly empirical, relying on the screening of large numbers of compounds and complex biochemical extracts derived at random (or nearly so) from myriad synthetic, microbial, botanical, marine, and other sources; or it was grounded in concepts of cell cycle, drug resistance, and biochemical pharmacology before such efforts shifted to seeking agents that blocked key molecular pathways that had been identified as causing cancer (Chabner and Roberts, 2005).

These approaches have uncovered many effective anticancer agents that form the core of contemporary chemotherapy. However, most of these agents have significant adverse side effects that cannot easily be eliminated without decreasing therapeutic effectiveness because, while they may have a high affinity for cancer cells, they do not target vulnerabilities unique to cancer cells. Consequently, the development of a promising lead compound into a therapeutic agent that is safe, effective, and has tolerable side effects is typically a slow and expensive process, and most of the time and resources spent during this process yield little of value because the majority of lead compound derivatives fall by the wayside before they ever reach the clinic.

Today, with the complete mouse and human genome sequences in hand, and an increasingly sophisticated view of the molecular and cellular biology of cancer at the disposal of researchers, we are poised to enter an era of rational cancer drug development in which the targets of promising therapies, and the methods for hitting these targets, are more readily identifiable than ever before. Scientists in government, industry, academia, and non-profit research institutes—working independently and collaboratively—have already used the human genome sequences to pinpoint several new genes that are deleted or amplified in human cancers, or that are aberrantly switched on or off in cancer cells (Sherwood, 2003).

Concerted efforts to find effective anticancer agents began after World War II, following observations that personnel exposed to mustard gas during World War I had experienced severe bone marrow depression, and that this feature of mustard gas derivatives was a dominant theme in molecules synthesized through the World War II era. Nitrogen mustard was subsequently found to display significant effects against lymphoma and became a starting point for the synthesis of many related drugs that damage DNA and, as a result, are preferentially toxic to proliferating cells.

The therapeutic successes of these drugs, called alkylating agents, were paralleled by similar postwar outcomes with other drugs, called antimetabolites. These compounds interfere with DNA synthesis and hence disrupt cell proliferation by inhibiting the production of the building blocks of DNA.

In 1956, Bristol Laboratories contracted with the NCI to identify potential anticancer antibiotics and formed a similar research alliance with Japanese researchers. This work led to the discovery of a number of anticancer antibiotics. Other important examples were discovered by Italian researchers.

Many chemotherapy agents were later derived from plant sources, some from the extensive NCI natural products screening program which was discontinued in 1982, and others from a new natural products program begun in 1986, which produced a number of drugs from marine sources.

Clearly, the compounds that emerged from the empirical approaches to cancer drug discovery act through specific pathways—molecular pathways—even though they were not selected or developed with precise foreknowledge of the targets or processes that they would be affecting. Empirical drug discovery still continues, but there is a shift toward using the detailed information emerging from basic research about the molecular abnormalities that underlie cancers and the peculiarities of cancer cells to look for agents with specificity for those targets.

The first hints that cancer patients could be restored to health by modulating a specific biological process emerged from studies of the hormonal control of prostate tumors in the 1940s at the University of Chicago. The large-scale transition of cancer drug discovery that continues to evolve was made possible by numerous subsequent lines

of inquiry—many with origins in basic research funded by the NIH in the late 1960s and early 1970s which explored the biology not only of human cells, but also of other seemingly unlikely yet ultimately telling cellular sources. This work, carried out primarily at universities and non-profit research institutes, revealed the fundamental mechanisms of growth control and programmed cell death and led to the identification of a large number of molecules and pathways that regulate the proliferation of cells of higher animals. These findings eventually enabled researchers directly interested in cancer to pinpoint promising molecular targets and growth control pathways for the development of novel cancer therapies.

The discovery in the late 1970s and early 1980s of the cellular cancer genes (oncogenes) and later of tumor suppressor genes spawned the detailed molecular description of cancer that now prevails and helped explain the root causes of cancer. A landmark publication by Hanahan and Weinberg (Hanahan and Weinberg, 2000) outlines six “hallmarks of cancer” (to which evading immune system surveillance and impairment of DNA repair might be added), all of which have known molecular components that are being investigated as therapeutic targets.

According to Hanahan and Weinberg, the complexities of cancer physiology revealed over the past three decades can be understood in terms of a small number of underlying principles, or hallmarks, of cancer. One hallmark of cancer cells is their escape from control of extracellular growth factor signals. Instead, owing to mutations or overactivity in growth factor receptors or any of a number of other signal transduction components or transcription factors, cancer cells generate their own growth signals. They therefore, can grow independently. Today, small molecules and monoclonal antibodies that target aberrant growth factor signals or receptors are prominent examples of targeted cancer therapies aimed at this kind of abnormality.

To maintain the integrity of tissues (such that they comprise neither too many nor too few cells) and to guide the differentiation of cells into specialized types, several antiproliferative “don’t grow” signals operate within and between cells, which either may cause cells to permanently exit the cell division cycle and adopt specialized properties, or alternatively divert cells from the cell division cycle and lead them into a quiescent state from which they may resume proliferating if future conditions dictate. In normal cells, anti-growth signals are relayed in large part by a pathway that is controlled by a prototypical tumor suppressor protein. A second hallmark of cancer is the loss of this suppressor function in cancer cells which can lead, for example, to the unopposed activation of a powerful growth stimulating transcription factor. This factor is a promising target for attack by small molecule drugs under development.

A third hallmark of cancer is disruption of the normal cellular process of programmed cell death (apoptosis). Surplus cells in developing or adult organisms receive extracellular signals to kill themselves, and pre-cancerous cells can detect that they are abnormal and similarly commit suicide through this process. Disruption of programmed cell death through mutations in tumor suppressor genes or increased levels

of activity of inhibitors of the process allows cancer cells to survive and proliferate. Treatments under investigation to address these abnormalities are aimed at targeting the changes in cellular mechanisms that result from defective tumor suppressor genes or overactive inhibitors of programmed cell death, among others.

A fourth hallmark of cancer relates to the normal process in which the tips of chromosomes in most human cells become progressively shorter in each cycle of cell proliferation. With the exception of stem cells (which can proliferate indefinitely), normal human cell types, after 60 or 70 doublings, sense that their chromosome tips have become too short to support additional cell division without compromising chromosome integrity, and they either cease proliferating or die. In stem cells and in cancer cells, an enzyme actively maintains the DNA sequences located at the tips of chromosomes, and provides stem cells and cancer cells limitless replicative potential. One strategy for cancer therapy involves the development of inhibitors of the enzyme which would block the capacity for endless proliferation of cancer cells.

Another hallmark of cancer involves the ability of tumors to capture the usually tightly regulated process of new blood vessel formation (angiogenesis), and, in so doing, to provide themselves with increasing supplies of oxygen and nutrients as they grow. Cutting off the blood supply to tumors by blocking this process with various inhibitors has therefore emerged as a promising cancer therapy.

Metastasis, or the ability to develop pioneer cells that leave the primary tumor and invade distant tissues, is a sixth, and deadly, hallmark of cancer, accounting for 90 percent of cancer deaths. Tissue invasion and metastasis by cancer cells involve the severing of normal linkages between cells and their intercellular matrix. The activation of extracellular protease enzymes enables migrating cancer cells to slip through the normally impassable spaces between cells in tissues and blood vessel walls. Blocking tissue invasion and metastasis is predicted to significantly improve the prognosis of cancer patients. However, these processes are arguably the least well understood among the prominent hallmarks of cancer, and progress in developing therapies directed against the molecules involved in tissue invasion and metastasis has been slow.

The complete human genome sequences provide new opportunities for discovering genes altered in cancer cells or expressed at significantly increased or decreased levels. Moreover, by grouping numbers of genes—some with known and others with unknown functions—into co-regulated sets, researchers can obtain clues to the function of previously uncharacterized genes. Other analyses reveal which few genes are the linchpins of a particular process in tumors, how they are functionally related to each other, and what approaches might be taken toward developing therapies based on this information.

These advances have become the basis in preclinical research for a variety of high-throughput screens to identify candidate targets for drugs; for determining the extent to which lead compounds affect the molecular pathways of interest; and for

characterizing potential adverse side effects of agents under development by determining whether the agents affect a molecular pathway or pathways associated with adverse side effects. Much of the past progress and future discovery of the kind outlined above has come about since the advent of DNA microarray analysis (which can analyze the expression of thousands of genes in a cancer sample) and related technologies that continue to be developed (Clarke et al., 2001).

Identification of the many individual proteins in large multiprotein complexes and determining the functional relationships among these proteins and how such protein complexes are altered in cancer cells or in response to therapy, that is, proteomics, has resulted from technological advances in protein mass spectrometry and from the marriage of this technology to complete genome sequence information. Accordingly, results that previously required years of effort to obtain (that is, by the cloning and sequencing of the genes encoding such large numbers of proteins) now take only days or weeks, and such results provide researchers valuable information much more rapidly than traditional methods for establishing functional relationships among proteins (Petricoin et al., 2004).

Progress in both genomics and proteomics research has been spurred by parallel developments in a blend of molecular biology and computer science called bioinformatics. Biologists, programmers, and others engaged in bioinformatics research are developing increasingly sophisticated analytical software, powerful statistical methods, databases, and user interfaces for the management, manipulation, and mining of experimental genomic and proteomic data and other allied information (such as human knowledge and the vast amount of published information about a particular topic). The organization and analysis of large amounts of experimental data and other biological information is crucial for enabling researchers to explore the landscape of the human genome and proteome, reveal new sequence elements within DNA or functional domains within proteins, and correlate these features with cancer biology.

To identify or validate therapeutic targets, and to assess drug efficacy and toxicity in cell culture or in animal models, researchers may employ animals with precisely-defined genetic backgrounds, including the absence of a particular relevant gene—a gene knockout. Using new technology, called RNA interference, scientists can reliably create these knockouts in a wide variety of organisms from mice and other mammals to fungi, fruit flies, and plants in a stably inherited form, and thus examine experimental animals without a gene (and function) of interest under a variety of experimental conditions. A knockout shows researchers what would happen if an agent against that target were completely effective.

Cancer immunotherapy is an approach to treating the disease or preventing recurrences by encouraging the immune system to recognize cancer cells as foreign and attack them. New insights into the cells, molecules, and signaling pathways that regulate immune responsiveness are reinvigorating this field (Pardoll, 2002). Overall, one can conclude that tumors that reach the stage of being clinically detectable are likely to have done so either by generating tolerance in the immune system or by developing ways of resisting immune recognition. In terms of cancer treatment, immunotherapeutic

approaches are geared toward identifying ways to break tolerance or circumvent resistance mechanisms.

Recently, research has been focusing on designing antibodies to act on cell types and molecules necessary for tumor growth and on using antibodies as vehicles to carry a toxic agent to a tumor site (Dillman, 2001). Numerous approaches to cancer vaccines are also being tested (Fearon et al., 1990, Pardoll, 2002a), and the future may see more widespread use of immunotherapy in combination with other modalities: surgery plus radiation plus immunotherapy, for example. However, a major challenge in cancer vaccines remains finding strategies to break self tolerance and generate immunity through manipulating both the vaccine target (antigen) and the delivery system (Wang and Wang, 2002).

The increasingly rapid identification of specific therapeutic targets—as outlined above—and improvements in validating these targets through refined in vitro systems and more sophisticated animal models of cancer provide an important foundation for developing small molecules and other agents with the potential to be highly specific, potent, and non-toxic. And immunologic approaches are zeroing in on ways to engage the immune system very specifically against cancers.

The future development of successful agents will involve both traditional and modified high-throughput screening of very large combinatorial libraries of compounds as well as in silico molecular modeling (so-called “rational drug design”), both of which will be supported by improved and accelerated methods for determining structure-activity relationships of drug candidates.

Progress in every stage of cancer drug discovery and therapy—from target identification, to compound development, to rational treatment regimens based on precise molecular diagnoses of individual patients—is poised to provide the next generation of cancer patients, and certainly the generation after that, with far better options for eliminating or controlling their disease than exist today. The next chapters of this background paper discuss the programs of two key federal agencies for realizing those better options for cancer drug development.