Since 2001 GIST treatment has benefited from blocking a single mutant protein, c-KIT. Targeting c-KIT with Gleevec and Sutent has led to years of extended survival for many GIST patients. Unfortunately cancer is an adaptive disease. Adaptive c-KIT mutations can bypass current therapies in one common mechanism of GIST resistance. As researchers observe this phenomenon, the need arises to seek and block new proteins in the GIST cell signaling pathway.

PI3K proteins have been identified in crucial signaling paths of multiple cancers. They can become over-active via a variety of mechanisms including upstream signaling. PI3Ks have recently been identified as active downstream signal points in the c-KIT pathway in GIST. A November 2007 paper authored by researchers at Brigham and Women’s Hospital (Bauer et al.) reported that imatinib resistant GIST cell lines underwent substantial cell death when treated with a PI3K inhibitor, but not with PI3K inhibition was proposed to be a promising therapeutic strategy in imatinibresistant GIST.

c-KIT is a tyrosine kinase. It’s often depicted as a spark-plug protruding through the cells surface sending signals down into the cell nucleus. Two c-KIT proteins pair up with each other (homodimerize) and are normally activated when the part outside the cell attaches to a circulating growth factor. This leads to activation of the part of the KIT protein inside the cell, which then attaches phosphates to tyrosine residues on multiple downstream proteins. PI3Ks are part of these downstream proteins.

PI3K is short for “Phosphoinositide 3- kinase”. PI3Ks come in two parts, a regulatory component and a catalytic component. The two must pair up (heterodimerize) in order to work. PI3Ks do not protrude through the cell membrane, instead they are entirely within the cell and migrate between the cell surface and the cell interior. When activated, PI3Ks work by attaching a phosphate to proteins called phosphatidylinositols (PtdIns or PI) at the cell surface. Activation of the PI PIP3 leads to the activation (phosphorylation) of AKT at the cell membrane. Activation of AKT then leads to alterations in protein translation, metabolism, cell survival, and proliferation.

PI3Ks differ from c-KIT in ways that have implications for targeted therapy.

PI3K’s comprise a group of about 15 proteins with very similar structures. They are organized into three closely related classes. The Class I PI3Ks are the only group to activate AKT downstream and are the main focus of GIST researchers. As a group, Class I PI3Ks are involved in a variety of lifesustaining functions, including immune system response and glucose metabolism. For instance, PI3Ks are key components in insulin regulation and diabetes. Some Class I PI3Ks are ubiquitous (found in every cell type) which presents a challenge to researchers designing and selecting Class I PI3K inhibitors for testing. It is preferable to have a drug that targets only one isoform (a form similar with small differences) of PI3K to avoid “collateral damage” and to unambiguously assess the effects of pathway inhibition.

Until recently, clinical researchers have only had compounds that inhibited all Class I PI3Ks. These have been called pan-PI3K inhibitors. Most have poor drug qualities or are too toxic. Using molecular design and novel drug delivery techniques, researchers have recently been able to design newer PI3K inhibitor candidates that more selectively inhibit Class I PI3Ks and/or deliver a pan-PI3K inhibitor selectively to tumor tissues. These drugs are just now entering Phase I trials for solid tumors.

PI3Ks: A Short History

PI3K’s were first identified in the literature in 1990, four years after the discovery of c-KIT. For many years they have been the subject of basic science and in the last three years, interest has grown greatly. Figure 2 charts the scientific papers referring to either c-KIT or PI3Ks that have been entered into Pub- Med.

Inhibitors for PI3Ks have been available in the lab since 1993 when wortmannin, a metabolite of a penicillin fungus, was discovered. Wortmannin is a highly effective inhibitor of PI3K and related proteins, but don’t look for it on the pharmacy shelf. Wortmannin is highly unsuited as a drug; it is toxic, with a poor solubility and a ten minute half-life in tissue samples. However, the availability of wortmannin and a synthetic version, LY294002, opened up the study of the PI3K pathway in the lab. These inhibitors and others enabled researchers to better understand the signal components and the pathways.

In the late 1990s, researchers were able to identify different isoforms or versions of PI3K proteins. In 1999, the crystal structure of one PI3K catalytic component, p110γ (gamma) was documented. Throughout this period, PI3K-related proteins were being identified as active in different cancers. There are now 15 regulatory and catalytic components Since 2004, PI3K mutations have been associated with a wide range of cancers. This discovery leveraged the significant body of science that had been building around the PI3K family of proteins since they were first documented in 1990.

Mutation is one of several reasons for that increased PI3K signaling. In 2004 and 2005, multiple papers identified the gene PIK3CA (See Figure 3) as generally mutated in cancer. PIK3CA encodes the PI3K catalytic subunit p110α. This subunit is responsible for cell survival and glucose metabolism.
The first PI3K inhibitor trials in humans began in early to mid 2007. Interest in PI3Ks in GIST began with the publishing of a June 2007 identifying PI3K proteins as promising therapeutic targets in resistant GIST signaling.

Reference List
Bauer, S., et al. “KIT oncogenic signaling mechanisms in imatinib-resistant gastrointestinal stromal tumor: PI3-kinase/AKT is a crucial survival pathway.” Oncogene. . (2007). Corless, Christopher L. and Michael C. Heinrich. “Molecular Pathobiology of Gastrointestinal Stromal Sarcomas.” Annual Review of Pathology: Mechanisms of Disease 3.1 (2008): 557-86. Karakas, B., K. E. Bachman, and B. H. Park. “Mutation of the PIK3CA oncogene in human cancers.” Br.J Cancer 94.4 (2006): 455-59.

PI3K drugs in phase I trials

Phase I trials usually determine pharmacological and metabolic actions of drugs administered for the first time in humans. They are also called safety studies because they try to determine the “maximum tolerable dose” (MTD) of the new drug. This is true of all six PI3K drugs on our list.

Additional similarities for all six:
• Are sponsored by the manufacturer
• Specify either cancer or solid tumors
• Do not specify GIST
• Are for adults only
• Do not exclude prior therapies (they may, but they do not specify)
The main differences are in the PI3K isoforms these drugs inhibit and how well. These are important because:
• A multi-PI3K inhibitor is more likely to block a pathway relevant in GIST. All trial drugs here inhibit multiple Class I PI3K isoforms. Most manufacturers have published data on which isoforms are inhibited and how strongly. Some have not.
• A multi-PI3K inhibitor may also be more risky. PI3K’s are heavily involved in the immune system and in glucose metabolism; inhibiting multiple PI3Ks may risk inhibition of these vital functions. These risks would be worth inquiring about prior to trial participation with any of the drugs here.
mTOR is a closely related PI4K protein that is inhibited by some of the drugs in these Phase I trials. mTOR inhibitors have now been approved for use in transplant rejection and renal cell cancer. The prescribing information for these drugs warns about hyperglycemia, secondary cancers and opportunistic infections. These risks may also be worth exploring when evaluating trials that include an mTOR inhibitor.
PI3K inhibitor based strategy differs is several respects from c-KIT inhibition. PI3Ks are part of the PI3K/PDK/Akt/mTOR pathway that promotes cell survival in many cancers. By inhibiting this pathway researchers hope to increase the effectiveness of existing therapies that can lead to cell death. PI3K inhibitors may therefore be combined with one or more other therapies in future trials.

In the order of the date the trial started:

BEZ235: This oral drug from Novartis targets Class I PI3Ks and mTOR. It is one of two PI3K inhibitors from Novartis currently in phase I. This is a combined Phase I/I trial, with an accrual goal of 80 patients. The trial design will admit solid tumors in Phase I. However, Phase II is limited to breast cancer patients. It is currently available at The Nevada Cancer Institute, Las Vegas, Nev. and at Sarah Cannon Research Institute in Nashville, Tenn.

SF1126: Semafore Pharmaceuticals, a small company in Indianapolis, Ind. has developed a broad spectrum PI3K inhibitor based on the lab drug LY294002. SF1126 is a Prodrug that inhibits Class IA PI3Ks and DNA-PK and mTOR. A Prodrug is given in one form and converted into a different, active form by metabolic action in the body. Semafore has designed SF1126 to have an affinity for the vascular forming areas of tumors. This design is intended to concentrate the effect of a broad spectrum inhibitor within the tumor. Of the six PI3K inhibitors listed here, SF1126 is the only one administered intravenously. Current trial sites include Arizona Cancer Center, Tucson, Ariz. and Indiana University, Indianapolis, Ind.

XL147: A second PI3K inhibitor from Exelixis. XL147 is a more selective PI3K inhibitor. It inhibits Class IA and IB PI3Ks. It is also administered orally. Trial sites currently include Dana Farber (under Jeffrey Shapiro, MD), Mary Crowley Medical Research Center, Dallas, T.X. and Hospital Universitario Vall d’Hebron in Barcelona, Spain.

XL 765: Exelixis is a drug development company based in San Francisco, Calif.. Exelixis has 13 drugs in Phase II or earlier stages of development. XL765 is one of two oral PI3K targeted drugs in development. XL765 inhibits Class 1A and 1B PI3Ks as well as DNA-PK and mTOR and is administered orally. Trial sites currently include Karmanos in Detroit, Mich., South Texas Accelerated Research therapeutics (START) in San Antonio, T.X. and Hospital Universitario Vall d’Hebron in Barcelona, Spain.

BGT226: This is a second PI3K inhibitor from Novartis, which inhibits Class I PI3Ks. This trial is classified as a combined Phase I/II. Initially patients with solid tumors will be accepted, but the phase II expansion part will enroll only advanced breast cancer patients and Cowden syndrome patients. Cowden syndrome is an inherited disease in which the

PTEN gene is defective leaving the survival signal from the PI3K pathway un-attenuated. The only current trial site is Nevada Cancer Institute, Las Vegas, Nev.

GDC-0941: This drug was initially developed by Piramed in the United Kingdom. GDC-0941 selectivity has not yet been published, so this information is unknown. We have been told informally that this oral drug inhibits multiple class I PI3K’s. It is currently in trial at Dana Farber in Boston, Mass. under George Demetri, MD.