Abstract
Introduction: The mechanism of action and resistance to chemotherapy is poorly understood and measures of efficacy typically rely on clinical outcome data. Recent advances suggest that prospective gene expression profiling (GEP) can be used to more accurately define not only short-term but lasting treatment benefits. We recently reported that baseline tumor cell gene signatures, encompassing 70 and as few as 17 genes, can discriminate risk groups of myeloma patients both in the untreated and previously treated settings. However, a subset of predicted low-risk cases followed an aggressive clinical course accompanied by a shift from 70-gene-defined low-to high-risk over time, either reflecting clonal evolution or outgrowth of aggressive clones present, but undetectable, at diagnosis. Accurately identifying this patient population is a first step in preempting such transformation. We hypothesized that changes in gene expression patterns of tumor cells following a short term in-vivo challenge with a single agent chemotherapeutic might expose these latently aggressive cells. Unlike in-vitro testing, clinical drug administration also allows for assessing tumor cell perturbation in the context of host interactions. Building on recent observations that clinical outcome in myeloma patients could be correlated with 48hr GEP changes induced in vivo following single agent administration of thalidomide, lenalidomide, and dexamethasone, we now examined whether such short-term tumor-cell GEP and proteomic alterations could fine-tune clinical outcome prediction beyond the well established 70-gene-based baseline prediction model.
Methods: Affymetrix U133Plus2.0 microarray analysis and mass spectrometry were performed on CD138-enriched tumor cells prior to and 48hr following a single test-dose application of bortezomib at 1.0mg/m2 in 142 newly diagnosed patients with MM. A total of 1051 genes (P < .005) were differentially expressed at 48 hours. Both change in expression, adjusting for baseline expression, and post-drug expression levels of each gene were examined for correlations with event-free survival in a Cox proportional hazards model. Post-drug expression was chosen and 113 genes were retained (p <= .05). The difference of the mean log2 expression of genes with hazard ratios (HR) of < 1.0 (favorable) and genes with HR >=1 (unfavorable) was used to create a score which, in the context of running log rank statistics, was used to classify patients into high- and low-risk groups. The independent prognostic power of the score for event-free and overall survival was investigated, together with baseline prognostic variables, by multivariate analysis. This method was tested in a 10-fold cross-validation procedure using the same data set. The model is currently being validated in an independent set of 100 cases and results will be reported.
Results: Changes in the expression of proteasome genes, and their related proteins, predominated a list of 113 outcome-related genes. A high-risk score, associated with upregulation of proteasome genes, seen in 24% of cases, was associated with median survival of less than 24 months, dramatically contrasting with a 3-year survival estimate of greater than 80% in the 76% in whom proteasome genes were not activated (p<0.0001). The cross-validated post-bortezomib score was an independent predictor of outcome in multivariate analysis of standard and genetic variables, including the well established and validated baseline 70-gene risk score. Importantly, 12% of patients in the 70-gene model-defined low-risk category were upgraded to high-risk by the 113-gene post-bortezomib model, with poor outcomes resembling those in the 70-gene-defined high-risk baseline model. Moreover, the 113-gene post-bortezomib score alone accounted for an unprecedented 57% of outcome variability by R2 statistics, with hazard ratio values of 5.45 for overall and 7.84 for event-free survival.
Conclusions: The rapid activation of proteasome genes and their corresponding proteins in MM cells within 48hr of a single bortezomib test-dose exposure as an indicator of poor clinical outcome suggests a novel and perhaps central mechanism of resistance to this new class of cancer therapeutics and perhaps standard genotoxic agents, which were part of the overall treatment. We are now testing the hypothesis whether the high-risk associated with the post-bortezomib proteasome activation can be overcome by higher doses of bortezomib or the addition of agents targeting other critical pathways.
Mathematical modelling of COVID-19 vaccination strategies in Kyrgyzstan