Gene expression analysis to detect disseminated tumor cells in the bone marrow of triple-negative breast cancer patients predicts metastatic relapse

Abstract

PurposeDisseminated tumor cells (DTCs) in the BM of breast cancer patients predict early disease relapse, but the molecular heterogeneity of these cells is less well characterized. Expression of a 46-gene panel was used to detect DTCs and classify patient BM samples to determine whether a composite set of biomarkers could better predict metastatic relapse.MethodsUsing a high-throughput qRT-PCR assay platform, BM specimens collected from 70 breast cancer patients prior to neoadjuvant therapy were analyzed for the expression of 46 gene transcripts. Gene expression was scored positive (detectable) relative to a reference pool of 16 healthy female control BM specimens. To validate findings from a subset of 28 triple-negative breast cancer (TNBC) patients in the initial 70 patient cohort, an independent set of pre-therapeutic BM specimens from 16 TNBC patients was analyzed.ResultsExpression of each of the 46 gene transcripts was highly variable between patients. Individual gene expression was detected in 0–84% of BM specimens analyzed and all but two patient BM specimens expressed at least one transcript. Among a subset of 28 patients with TNBC, positivity of one or more of eight transcripts correlated with time to distant relapse (p = 0.03). In an independent set of 16 triple-negative patient BM samples, detection of five of these same eight gene transcripts also correlated with time to distant relapse (p = 0.03) with a positive predictive value of 89%.ConclusionsWe identified a set of gene transcripts whose detection in the BM of TNBC patients, prior to any treatment intervention, predicts time to first distant relapse, thus identifying a TNBC patient population which requires additional treatment intervention. Because these genes are presumably expressed in populations of DTCs and many encode proteins that are known therapeutic targets (e.g., ERBB2), these results also suggest a potential approach for targeted DTC therapy to mitigate distant metastases in TNBC.Electronic supplementary materialThe online version of this article (10.1007/s10549-019-05405-7) contains supplementary material, which is available to authorized users.

A microfluidic-based filtration system to enrich for bone marrow disseminated tumor cells from breast cancer patients

“…Because of our on-going interest in detecting and classifying DTCs based upon their molecular (gene expression) signatures, we next evaluated whether the enrichment documented with tumor cells spiked into patient BM samples translated into more sensitive detection of native BM DTCs using molecular biomarkers. Accordingly, we assessed quantitative levels of gene expression for seven transcripts that have been previously associated with the presence of BM DTCs in BC patients [ 32 ]— EPCAM , ERBB2 , PD-L1 , PDGFRB , STEAP1 , TWIST1 , and WNT5A - by droplet digital PCR ( ddPCR ) pre- and post-Parsortix enrichment. 10 BM and 9 peripheral blood specimens were analyzed.…”

“…Unsupervised cluster analysis of ~9,500 cell libraries revealed two distinct clusters of cells that were enriched for two marker transcripts which we and others have previously associated with DTC-positive BM from BC patients: CAV1 and EpCAM [ 32 ]. In whole BM gene expression analyses, expression of CAV1 is frequently associated with early distant disease relapse [ 32 ]. In fact, this patient experienced rapid disease progression 9 months after diagnosis.…”

“…Each pre-amplification reaction was then subjected to digital droplet PCR ( DDPCR ) amplification using gene-specific primer / probes and the QX200 droplet generator / reader system (Biorad). Primer / probe details used for gene-specific amplification of DTC-associated transcripts ( EPCAM , ERBB2 , PDL-1 , PDGFRB , STEAP1 , TWIST1 , WNT5A ) are provided in S1 Table [ 32 ]. Droplet counts from duplicate reactions were averaged and normalized to counts from GAPDH primer / probes in each sample to account for differences in input cell number and RNA yield and quality between pre- and post-filtration samples, and between patients.…”

Combined Analysis of Disseminated Tumor Cells (DTCs) and Circulating Tumor DNA (ctDNA) in a Patient Suffering from Triple Negative Breast Cancer Revealed Elevated Risk

“…However, analysis of the bone marrow revealed a positive DTC status and hence an elevated risk for recurrence. The number of detected DTCs was 85 cells among 4 million bone marrow cells and hence unusually high, keeping in mind that 1-10 DTCs per 1 million immunocytes were reported to be average [11]. Subsequently, adjuvant treatment with bisphosphonates, which might eradicate micro-metastasis from the bone marrow and improve prognosis, was applied.…”

Selection of endogenous control genes for normalising gene expression data derived from formalin-fixed paraffin-embedded tumour tissue

“…Measuring gene expression is increasingly important for a diverse range of clinical applications [23][24][25][26] . Purification of RNA, an essential prerequisite for qPCR, removes other cellular components and data must be normalised based on the stable expression of endogenous control genes.…”