Masha G. Savelieff, PhD
Precision medicine is the selection of treatment options tailored to the specific features of a patient’s disease. Early successes hailed the promise of precision medicine to find treatments for cancer. For example, the tyrosine kinase inhibitor imatinib was designed to inhibit a constitutively active form of the tyrosine kinase Abelson (Abl), the chimera Bcr-Abl, which results from a chromosomal translocation that generates the so-called Philadelphia chromosome. Imatinib yielded a durable response in chronic myelogenous leukemia (CML) that were Philadelphia chromosome-positive (Ph+). It also proved beneficial to the majority of CML patients, since more than 90 percent of CML cases are Ph+ and express Bcr-Abl. Since then, a number of clinical trials have been established to match a tumor’s vulnerability to a targeted treatment, hoping to duplicate the success with imatinib and other next-generation kinase inhibitors for CML. While some ongoing trials are adopting a functional approach, trials using next generation sequencing (NGS) are yielding a faster selection of targeted treatments.
MATCHing a tumor’s genetic alteration to targeted therapy
A number of clinical trials using NGS platforms have been launched, including the National Cancer Institute’sNCI-MATCH program. This phase II trial is actively recruiting participants without restriction on cancer type, which can range from advanced solid tumors to hematological lymphomas or myeloma, in addition to rare cancers other than breast, prostate, colorectal, or non-small cell lung carcinoma. The trial is using NGS to sequence a patient’s tumor biopsy to identify “actionable” mutations with known inhibitors that are either FDA-approved for other indications or that have a demonstrated safety profile in other ongoing clinical trials. Another program, limited to lung squamous cell carcinoma (SCC) patients called Lung-MAP, is presently in phase II/III clinical trials and is using NGS to pinpoint actionable genetic anomalies. The trial will then randomize the specific targeted therapy versus the standard-of-care for SCC.
The NGS approach to precision oncology enjoys many benefits, particularly from a relatively low cost as sequencing prices continue to decrease. It is also a relatively straightforward approach since many actionable mutations with established inhibitors are already known, so assignment of a targeted therapy can proceed directly from the genetic alterations identified by NGS, although prioritization may be necessary in cases where multiple actionable mutations are present. This is simultaneously a disadvantage of the technique, however. If an actionable genetic alteration is not found, the tumor cannot be matched to a targeted therapy. Alternatively, a mutation may be present in an otherwise actionable gene, but which has not been functionally validated. The biopsy, which only samples a small part, may also not be representative of the entire tumor, which is frequently heterogenous. In addition, even if a match is possible, biochemical pathways are frequently redundant, and targeting one genetic alteration may be compensated by the tumor via alternative pathways.
Functional testing for phenotypic features
A tumor’s genetic makeup affects all subsequent layers of tumor biology, including the transcriptome, epigenome, proteome, and metabolome. While genomics platforms for precision oncology can assign targeted therapies in a straightforward manner, they do not account for all the pathways downstream of the genotype, whose cumulative effect produces the tumor phenotype. Therefore, another approach to tailored therapy selection is through functional testing of the tumor’s phenotype. This type of platform does not require an understanding of the tumor’s underlying biology or mutations, but rather tests its overall susceptibility to a panel of drugs or drug combinations in an empirical screen. Functional screens can be performed in vivo, such as in patient-derived xenografts (PDXs), or in vitro, on patient-derived cancer cells, both of which offer distinct advantages.
PDXs: A cancer patient’s avatar
PDXs are generated by implanting a small piece of tissue from a patient’s biopsy, either subcutaneously or orthotopically, into immune deficient mice. The mice cannot reject the human tumor tissue, thus allowing it to grow. PDX mice can be propagated to produce a larger number on which to test drugs. The most effective drug identified in PDX mice is then administered to the patient. A number of early clinical trials and studies have been registered, including for triple negative breast cancer, head and neck cancer, sarcomas, osteosarcoma, and a variety of other cancers. The small number of trials that have already been completed have shown a good level of concordance between the drug selected in PDXs and its efficacy in the patient. In a phase II trial of pancreatic and other aggressive cancers, PDXs were generated for 14 patients with refractory advanced disease, which were used to screen 63 drugs in 232 therapy regimens. The screen was successful for 12 patients, which identified 17 potential therapies, of which 15 produced a durable partial remission. A later study of PDXs from sarcomas revealed a similar level of concordance, with 13 out of 16 patients that responded to the treatment regimen indicated by their PDXs.
PDXs possess several advantages, such as the ability to test the tumor phenotype, which is the cumulative effect of all genetic mutations, not just the effect of actionable mutations present. In addition, the approach can test combination treatments that don’t need to be designed to specifically target a mutation; therefore, the method is amenable to a broader range of drugs and to tumors lacking known actionable mutations. Finally, unlike cell-based functional assays, testing is performed in vivo, which may more accurately reflect the tumor microenviroment. Unfortunately, PDXs are relatively expensive and time consuming, a significant problem in patients with aggressive tumors who may succumb before the PDX screen yields results. Genetic drift in the tumor is also possible as the PDX is propagated, and the small biopsy sample size may not represent the whole tumor. Moreover, the mouse microenviroment, even though in vivo, does not accurately reflect the tumor environment in humans and certain human biopsy tissue sometimes fails to grow in mice. Since the mice are immunodeficient, it is not possible to test immunotherapies, although newer humanized mouse models are aiming to address this shortcoming.
In vitro systems accelerate screens: Novel advanced cell culture platforms
The comparatively higher cost and length of time required for functional testing in PDXs can be surmounted by using in vitro tests on cell cultures. Although more difficult to obtain or expand from patients with solid tumors, obtaining patient-derived cancer cells from hematological malignancies is relatively straightforward. One such phase II clinical trial is using an in vitroscreen of blood cancer cells to identify possible drugs. More advanced culture formats are also in development, such as spherical suspension cultures, which are cultured as undifferentiated, stem cell-like, 3-D aggregates. A phase II trial of such “spheroids” or “tumorspheres” for metastatic colorectal cancer is in progress. Drug testing can also be performed using “organoids,” patient-derived cancer cells formed into miniature, layered tissue assemblies. Clinical trials for this drug-testing platform are also in effect for breast cancer, metastatic pancreatic cancer, non-small cell lung carcinoma, esophageal, and rectal cancer.
Ultimately, the best prediction for targeted cancer therapy may come from a combination approach that uses both NGS genomics platforms and functional testing. Challenges remain for either method, such as the possibility that a small biopsy sample may not be representative of the entire tumor. However, it is anticipated technological advances already in development, such a liquid biopsy NGS, that may help address present shortcomings.