NEW YORK – An international team led by investigators in China has demonstrated the use of pharmacogenomic profiling on patient-derived cell models to find new treatment strategies and response biomarkers in head and neck squamous cell carcinoma (HNSCC), a difficult-to-treat form of head and neck cancer.
"[O]ur study shows that this HNSCC cell resource, as well as the resulting pharmacogenomic profiles, is effective for biomarker discovery and for guiding precision oncology therapies in HNSCCs," co-senior and co-corresponding authors Shuyang Sun and Zhiyuan Zhang, oncology researchers affiliated with Shanghai Jiao Tong University, and their colleagues wrote, noting that prior molecular analyses of HNSCC have highlighted recurrent mutations in tumor suppressor genes such as TP53.
As they reported in Science Translational Medicine on Wednesday, the researchers initially attempted to come up with patient-derived cell (PDC) models for 130 HNSCC cases, which they successfully established for 56 of these. They then included these PDCs, along with 18 immortalized cell lines, in their "HNSCC Cell Model Repository" (HNCR) — a resource that was subsequently characterized with pharmacogenomic approaches and available clinical data from corresponding HNSCC cases.
The authors noted that "our HNSCC PDCs recapitulate the histological and molecular features of the original HNSCC tumors and conserve the genomic landscape of [human papillomavirus-negative] HNSCCs, thereby representing a useful resource for pharmacological profiling."
With a three-stage high-throughput drug screening strategy, the team tracked select HNCR cell model responses to more than 2,200 small molecule compounds, analyzing the drug response profiles in relation to exome sequence, RNA sequence, and other data for the PDCs, tumor samples, patient-derived xenograft samples, and matched normal samples.
Along with analyses that highlighted dozens of compounds with significant effects on the growth of HNSCC cell models, the researchers flagged molecular features that seemed to impact responses to the small molecule candidates.
"The HNCR models displayed distinct drug sensitivities, and hierarchical clustering of drug sensitivities … identified three distinct clusters of models, which we termed cluster R (resistant), cluster M (moderate), and cluster S (sensitive)," the authors explained, noting that different drug types elicited distinct effects in cell models from each group.
Within a subset of PDCs marked by particularly low expression of the keratin 18-coding gene KRT18, for example, the researchers saw enhanced response to a JAK2 inhibitor being used as a treatment for adult myelofibrosis. On the other hand, a group of models involving increased IL6-receptor expression appeared to be amenable to treatment with a topoisomerase inhibitor such as mitoxantrone.
On the drug resistance side, meanwhile, the team saw signs that enhanced expression of SOD1 coincided with afatinib ERBB inhibitor drug resistance, while higher-than-usual expression of the BTG3 gene downstream of the TP53 tumor suppressor coincided with more frequent resistance to EGFR inhibitors.
The researchers also linked docetaxel resistance to expression of the ITGB1 gene — a proposed biomarker backed up by additional data from a Phase II clinical trial of chemotherapy and radiation after surgery in HNSCC patients.
"We retrospectively confirmed ITGB1 as a prognostic biomarker for patients undergoing postoperative docetaxel treatment in a clinical study," the authors reported. "Carrying these insights closer toward clinical relevance, we are now enrolling patients with HNSCC in a prospective validation cohort and are using matched PDC models to screen for docetaxel sensitivity to facilitate clinical trials based on an integrative pharmacogenomic approach."