Targeting Targeted Therapies

KEYWORDS::

• 5T4
• cancer vaccine
• companion diagnostic
• molecular targeted therapy
• trophoblast glycoprotein

Historically, cancer therapeutics have targeted the phenotypic features of cancer cells, principally their rapid cell division, which has resulted in cytotoxic agents becoming the mainstay of cancer treatment. However, their lack of specificity often results in a range of toxicities, which can restrict the maximum tolerated dose to levels below those required for complete efficacy. Indeed, it is a sad fact that the treatment of cancer has often been associated with some of the lowest response rates of any disease.

More recently, a better understanding of the genetic and biochemical make-up of cancer has led to the identification of numerous molecular targets that differentiate tumor cells from normal cells, either as a primary component of the neoplastic transformation or as a secondary result of that transformation. Molecular targeted therapies are designed to attack the cancer cell through these molecular Achilles‘ heels. This class of rationally designed drugs includes small molecules, monoclonal antibodies and therapeutic vaccines that target tumor-associated antigens. While these drugs can be very effective, their targeted nature means that they are only likely to work when used in the correct context, which challenges the traditional assumptions that all patients with a specific cancer type will respond similarly. In order to maximize the clinical benefit of each patient, there is increased focus on the development of molecular diagnostics and predictive biomarkers to ensure that the drug is targeted to patients who are most likely to respond.

Perhaps the most straightforward example of a targeted therapy is imatinib (Gleevec^®; Novartis, Basel, Switzerland), a small molecule inhibitor of the Abelson tyrosine kinase (ABL), used for the treatment of chronic myeloid leukemia (CML). CML is driven almost exclusively by a chromosomal translocation that gives rise to the breakpoint cluster region–ABL fusion protein, and constitutive activation of ABL. Results from Phase II clinical trials in CML demonstrated that imatinib produced a complete hematological response in over 90% of patients who were previously refractory for interferon treatment [1]. A small subset of patients developed secondary mutations in the ABL kinase domain, which prevent drug binding; however, in the majority of patients, imatinib is an extremely effective targeted therapy that sets a high benchmark and expectation for targeted therapies. However, CML is unique amongst cancers in that the chronic phase is almost exclusively characterized by a single mutation [2].

Unlike CML, most cancers are very heterogeneic, both in etiology and genetics. Breast cancer, for example, is one such highly heterogeneous cancer, which can be characterized by the differential expression of various receptors, such as the human EGF receptor (HER)2, progesterone receptor and estrogen receptor. HER2 is a receptor tyrosine kinase involved in signaling pathways that regulate growth and differentiation. HER2 is overexpressed in 15–30% of breast cancers as a result of gene amplification, and is associated with poor prognosis, accelerated metastasis and low treatment response rates [3]. Trastuzumab (Herceptin^®; Genentech, CA, USA) is a humanized HER2-specific monoclonal antibody that has demonstrated impressive clinical results in a subset of breast cancer patients. A pivotal Phase III study recruited women with metastatic breast cancer who had high levels of HER2 expression on their tumors. This trial demonstrated that the addition of trastuzumab to chemotherapy resulted in significantly improved clinical efficacy compared with chemotherapy alone [4]. Trastuzumab was approved in 1998 alongside the first ‘companion diagnostic‘, Hercep-Test™ (Dako, Glostrup, Denmark), an immunohistochemistry assay used to identify patients with HER2-positive metastatic breast cancer. In addition, a fluorescence in situ hybridization assay has been approved to detect gene amplification. Many thousands of breast cancer patients have benefited from treatment with trastuzumab; however, it is sobering to consider that the drug probably would not have been granted approval had the pivotal trial also included patients whose tumors showed low levels of HER2 expression.

Similar to trastuzumab in breast cancer, it was expected that cetuximab (Erbitux^®; ImClone Systems, NY, USA) and panitumumab (Vectibix^®; Amgen, CA, USA), both monoclonal antibodies that target the EGF receptor (EGFR) would demonstrate efficacy in metastatic colorectal cancer that express EGFR. Indeed, cetuximab was initially approved for treatment in patients with EGFR-expressing tumors, although subsequent evidence revealed that patients with nonexpressing tumors also benefited from this treatment [5]. Following approval of the drug, postmarketing studies demonstrated that a significant proportion of patients (those with K-RAS mutations) were refractive to cetuximab/panitumumab treatment and the labels were subsequently restricted to patients without K-RAS mutations at the request of the licensee. K-RAS testing in metastatic colorectal cancer, generally by PCR, has recently been recommended by both the American Society of Clinical Oncology (ASCO) and the National Comprehensive Cancer Network (NCCN) and is now considered standard practice [6,101].

Monoclonal antibodies and small molecule drugs have highly specific targets and direct modes of action, which make it possible to rationalize the design of a predictive biomarker (e.g., does the cancer express target molecule X? Does the cancer harbor mutation Y?). However, in the case of immunotherapeutic cancer vaccines, while there is often a discrete target, the mode of action is indirect, relying on activation and interaction of various elements of the immune system in order to achieve therapeutic efficacy. This indirect mode of action means that defining the ‘right patient‘ and identifying a predictor of treatment benefit is likely to prove challenging. It is not only overexpression of a particular tumor antigen or presence of a specific mutation that determines treatment suitability, but also the responsiveness of patient‘s immune system as a whole and the status of the tumor microenvironment.

Cancer vaccines or antigen-specific cancer immunotherapy (ASCI) hit the headlines recently with the approval of sipuleucel-T (Provenge^®; Dendreon) for the treatment of patients with hormone-refractory prostate cancer [7]. This long-awaited breakthrough finally provides proof of principle that the patients‘ immune systems can be harnessed to delay or prevent tumor progression. It is hoped that this news will be the ‘shot in the arm‘ that stimulates further investment and success in this under-resourced field.

Numerous companies have ASCI products in late-stage clinical development. Arguably the most ambitious development program is GlaxoSmithKline‘s (GSK; Brentford, UK) MAGE-A3-targeted approach, which is currently recruiting patients to two large Phase III trials in non-small-cell lung cancer (NSCLC; n = 2270 patients) and melanoma (n = 1300 patients). In developing their MAGE-A3 therapy, GSK has invested heavily in a gene expression profiling program during early-/mid-phase development. Pretreatment tumor biopsies were assessed by gene microarray with the aim of identifying biomarkers predictive of patient response to the MAGE-A3 vaccine in both NSCLC and melanoma. Gene signatures that correlated with treatment benefit in MAGE-A3-treated patients were identified and consisted of 33 gene probes in melanoma and 25 gene probes in NSCLC. Interestingly, the majority of genes were related to immune response/function and, therefore, reflective of the immunological status of the tumor microenvironment prior to treatment [8,9]. In a Phase IIb study in NSCLC, patients treated with MAGE-A3 showed a nonsignificant improvement in disease-free interval (hazard ratio = 0.75) compared with the control arm. However, patients with the predictive gene signature who were treated with the MAGE-A3 vaccine showed greater clinical benefit (hazard ratio = 0.57), whereas patients in the control arm with the gene signature showed no increase in clinical benefit [10]. In July 2009, GSK and Abbott announced that they would develop a companion diagnostic for the MAGE-A3 ASCI using Abbott‘s m2000 automated real-time PCR system. This methodology is currently being developed alongside Phase III clinical trials as a companion diagnostic [102]. The availability of gene profiling data on such a large scale will undoubtedly provide valuable insights into the impact of the tumor genotype on the tumor microenvironment and, potentially, on the relative efficacy of ASCI-type approaches.

The development of TG4010 (Transgene, Strasbourg, France), a modified vaccinia Ankara (MVA)-MUC1/IL-2-based cancer vaccine for NSCLC, has also been accompanied by a ‘large-scale biomarker discovery program‘ based on immunology, proteomics, transcriptomics and genomics. Data from a Phase IIb trial in NSCLC identified a subpopulation of patients who appeared to derive significant benefit from TG4010 treatment coupled with chemotherapy, compared with chemotherapy alone. This subgroup consisted of approximately 70% of the patient population who had normal levels of activated natural killer cells at baseline (pretreatment). Patients with this ‘favorable‘ pretreatment factor demonstrated a statistically significant increase in median survival (17.1 months for TG4010 treated vs 11.3 months in the control arm) [11]. Response rate, time to progression and progression-free survival data also reportedly confirmed the identification of activated natural killer cells as an appropriate predictive biomarker for treatment benefit of TG4010 in combination with chemotherapy in NSCLC. This observation is now being validated prospectively in future trials.

Oxford BioMedica (Oxford, UK) has also utilized the attenuated MVA vaccine vector to deliver the tumor-associated antigen 5T4 (MVA-5T4; TroVax^®). Results from nine open-label Phase I and II studies in renal, colorectal and prostate cancer demonstrated that TroVax was well tolerated and very efficient at inducing an immune responses against the 5T4 tumor target. Importantly, such immune responses were associated with signs of clinical benefit [12–20]. Recently, TroVax was tested in a Phase III trial in 733 patients with renal cancer, but, unfortunately, the trial did not meet the predefined primary end point of increased survival. Despite this, a strong association between 5T4 antibody response and enhanced survival was detected [21]. Furthermore, exploratory analyses have helped identify several pretreatment hematological factors, which seem to identify groups of patients who are more likely to mount the strongest 5T4-specific antibody responses and show improved survival when treated with TroVax. These factors include platelets, monocytes and hemoglobin levels, all of which are measured as standard at the start of any study. Therefore, it is relatively straightforward to include or exclude patients based on such standard measurements. The challenge is to now turn these retrospective observations into prospective analyses that help to select the patients most likely to benefit from this class of targeted therapy.

In conclusion, we would suggest that some of the failures that have hindered the development of cancer therapeutics may have been the result of an imperfect understanding of the disease harbored by the patients being treated. While there is a long way to go to fully understand all of the molecular and physiological transitions occurring in individual patients during cancer development, the community is becoming increasingly aware that drugs can fail if we do not understand this process, at least partially. The indirect mechanism of action of immunotherapy products may mean that we will need an even better understanding of the patient‘s disease to be able to identify those who are likely to receive the most benefit. Therefore, it is perhaps not surprising that factors related to immunoresponsiveness, in the periphery or at the tumor site, are already being reported as potential markers of enhanced efficacy. In the near future, it is hoped that we will be able to identify patients who are likely to significantly benefit from the targeted therapy based on a number of readily measured factors, which are likely to include:

• Target antigen: the presence and expression level of the target antigen (ideally on the tumor being targeted);

• Peripheral immune status: a panel of pretreatment hematological factors that provide a good indication of the patient‘s ability to mount a potent immune response to the target protein(s);

• Tumor immune status: genotypic/phenotypic data taken from a tumor biopsy that provide a picture of the immunoregulatory/immunostimulatory balance within the tumor.

By combining one or more of these factors with rational clinical trial design – including a move away from the cytotoxic drug-development paradigm that has plagued immunotherapy – it is hoped that Dendreon‘s cancer vaccine will be the first of many targeted immunotherapy products to be approved by the US FDA. Although it would be naive to conclude that biomarker programs are cheap and straightforward, it is likely to be a price worth paying in order to maximize the chance of success in the tortuous process of cancer drug development, and, ultimately, to provide better value for both patients and healthcare providers.

Financial & competing interests disclosure

William Shingler and Richard Harrop are employees of Oxford BioMedica (UK) Ltd. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.

Additional information

Funding

William Shingler and Richard Harrop are employees of Oxford BioMedica (UK) Ltd. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. No writing assistance was utilized in the production of this manuscript.