Search

Patients with colorectal liver metastasis (CLM) present with heterogenous clinical outcomes and improved classification is needed to ameliorate the therapeutic output. Macrophages (Mϕ) hold promise as prognostic classifiers and therapeutic targets. Here, stemming from a single-cell analysis of mononuclear phagocytes infiltrating human CLM, we identified two Mϕ markers associated with distinct populations with opposite clinical relevance. The invasive margin of CLM was enriched in pro-inflammatory monocyte-derived Mϕ (MoMϕ) expressing the monocytic marker SERPINB2, and a more differentiated population, tumor-associated Mϕ (TAM), expressing glycoprotein nonmetastatic melanoma protein B (GPNMB). SERPINB2+ MoMϕ had an early inflammatory profile, whereas GPNMB+ TAMs were enriched in pathways of matrix degradation, angiogenesis, and lipid metabolism and were found closer to the tumor margin, as confirmed by spatial transcriptomics on CLM specimens. In a cohort of patients, a high infiltration of SERPINB2+ cells independently associated with longer disease-free survival (DFS; P = 0.033), whereas a high density of GPNMB+ cells correlated with shorter DFS (P = 0.012) and overall survival (P = 0.002). Cell–cell interaction analysis defined opposing roles for MoMϕ and TAMs, suggesting that SERPINB2+ and GPNMB+ cells are discrete populations of Mϕ and may be exploited for further translation to an immune-based stratification tool. This study provides evidence of how multi-omics approaches can identify nonredundant, clinically relevant markers for further translation to immune-based patient stratification tools and therapeutic targets. GPNMB has been shown to set Mϕ in an immunosuppressive mode. Our high dimensional analyses provide further evidence that GPNMB is a negative prognostic indicator and a potential player in the protumor function of Mϕ populations.

Supplementary Figures 1-4, Tables S1 and S5

Supplementary Table S4

Patients with colorectal liver metastasis (CLM) present with heterogenous clinical outcomes and improved classification is needed to ameliorate the therapeutic output. Macrophages (Mϕ) hold promise as prognostic classifiers and therapeutic targets. Here, stemming from a single-cell analysis of mononuclear phagocytes infiltrating human CLM, we identified two Mϕ markers associated with distinct populations with opposite clinical relevance. The invasive margin of CLM was enriched in pro-inflammatory monocyte-derived Mϕ (MoMϕ) expressing the monocytic marker SERPINB2, and a more differentiated population, tumor-associated Mϕ (TAM), expressing glycoprotein nonmetastatic melanoma protein B (GPNMB). SERPINB2+ MoMϕ had an early inflammatory profile, whereas GPNMB+ TAMs were enriched in pathways of matrix degradation, angiogenesis, and lipid metabolism and were found closer to the tumor margin, as confirmed by spatial transcriptomics on CLM specimens. In a cohort of patients, a high infiltration of SERPINB2+ cells independently associated with longer disease-free survival (DFS; P = 0.033), whereas a high density of GPNMB+ cells correlated with shorter DFS (P = 0.012) and overall survival (P = 0.002). Cell–cell interaction analysis defined opposing roles for MoMϕ and TAMs, suggesting that SERPINB2+ and GPNMB+ cells are discrete populations of Mϕ and may be exploited for further translation to an immune-based stratification tool. This study provides evidence of how multi-omics approaches can identify nonredundant, clinically relevant markers for further translation to immune-based patient stratification tools and therapeutic targets. GPNMB has been shown to set Mϕ in an immunosuppressive mode. Our high dimensional analyses provide further evidence that GPNMB is a negative prognostic indicator and a potential player in the protumor function of Mϕ populations.

Supplementary figure legends

Supplementary Table S3

Supplementary Table S2

PDF file, 113K.

Purpose: Epithelial ovarian tumors (EOT) are among the most lethal of malignancies in women. We have previously identified ZIC2 as expressed at a higher level in samples of a malignant form (MAL) of EOT than in samples of a form with low malignant potential (LMP). We have now investigated the role of ZIC2 in driving tumor growth and its association with clinical outcomes.Experimental Design:ZIC2 expression levels were analyzed in two independent tumor tissue collections of LMP and MAL. In vitro experiments aimed to test the role of ZIC2 as a transforming gene. Cox models were used to correlate ZIC2 expression with clinical endpoints.Results:ZIC2 expression was about 40-fold in terms of mRNA and about 17-fold in terms of protein in MAL (n = 193) versus LMP (n = 39) tumors. ZIC2 mRNA levels were high in MAL cell lines but undetectable in LMP cell lines. Overexpression of ZIC2 was localized to the nucleus. ZIC2 overexpression increases the growth rate and foci formation of NIH3T3 cells and stimulates anchorage-independent colony formation; downregulation of ZIC2 decreases the growth rate of MAL cell lines. Zinc finger domains 1 and 2 are required for transforming activity. In stage I MAL, ZIC2 expression was significantly associated with overall survival in both univariate (P = 0.046) and multivariate model (P = 0.049).Conclusions:ZIC2, a transcription factor related to the sonic hedgehog pathway, is a strong discriminant between MAL and LMP tumors: it may be a major determinant of outcome of EOTs. Clin Cancer Res; 18(16); 4313–24. ©2012 AACR.

PDF file - 330K, The clinical stratification of the 76 selected patients used for the integrative analysis and the MAGIA2 complete results.

PDF file - 127K, miR-192 and miR-194 expression values stratified according to tumor grade and survival analysis (OS and PFS). Data shown refers to the entire cohort of 67 stage I mucinous EOC patients.

Supplemental Table 1: Table of all antibodies used in studies including the company, catalog numbers and dilutions used to perform western blots.

PDF file - 144K, Resampling strategy to identify miRNA markers.

PDF file, 62K, A. List of primer pair sequences used for RT-qPCR analysis and respective annealing temperatures. B. List of siRNA sequences used. C. Details of deletion constructs (see Fig. 4A).

PDF file - 126K, Survival analysis methodology and results for each histotype.

Purpose: Stage I epithelial ovarian cancer (EOC) represents about 10% of all EOCs and is characterized by good prognosis with fewer than 20% of patients relapsing. As it occurs less frequently than advanced-stage EOC, its molecular features have not been thoroughly investigated. We have demonstrated that in stage I EOC miR-200c-3p can predict patients' outcome. In the present study, we analyzed the expression of long non-coding RNAs (lncRNA) to enable potential definition of a non-coding transcriptional signature with prognostic relevance for stage I EOC.Experimental Design: 202 snap-frozen stage I EOC tumor biopsies, 47 of which relapsed, were gathered together from three independent tumor tissue collections and subdivided into a training set (n = 73) and a validation set (n = 129). Median follow up was 9 years. LncRNAs' expression profiles were correlated in univariate and multivariate analysis with overall survival (OS) and progression-free survival (PFS).Results: The expression of lnc-SERTAD2-3, lnc-SOX4-1, lnc-HRCT1-1, and PVT1 was associated in univariate and multivariate analyses with relapse and poor outcome in both training and validation sets (P < 0.001). Using the expression profiles of PVT1, lnc-SERTAD2-3, and miR-200c-3p simultaneously, it was possible to stratify patients into high and low risk. The OS for high- and low-risk individuals are 36 and 123 months, respectively (OR, 15.55; 95% confidence interval, 3.81–63.36).Conclusions: We have identified a non-coding transcriptional signature predictor of survival and biomarker of relapse for stage I EOC. Clin Cancer Res; 23(9); 2356–66. ©2016 AACR.

Wnt signaling is a major driver of stemness and chemoresistance in ovarian cancer, yet the genetic drivers that stimulate its expression remain largely unknown. Unlike other cancers, mutations in the Wnt pathway are not reported in high-grade serous ovarian cancer (HGSOC). Hence, a key challenge that must be addressed to develop effective targeted therapies is to identify nonmutational drivers of Wnt activation. Using an miRNA sensor-based approach, we have identified miR-181a as a novel driver of Wnt/β-catenin signaling. miR-181ahigh primary HGSOC cells exhibited increased Wnt/β-catenin signaling, which was associated with increased stem-cell frequency and platinum resistance. Consistent with these findings, inhibition of β-catenin decreased stem-like properties in miR-181ahigh cell populations and downregulated miR-181a. The Wnt inhibitor SFRP4 was identified as a novel target of miR-181a. Overall, our results demonstrate that miR-181a is a nonmutational activator of Wnt signaling that drives stemness and chemoresistance in HGSOC, suggesting that the miR–181a–SFRP4 axis can be evaluated as a novel biomarker for β-catenin–targeted therapy in this disease.Significance:These results demonstrate that miR-181a is an activator of Wnt signaling that drives stemness and chemoresistance in HGSOC and may be targeted therapeutically in recurrent disease.

Immunohistochemistry with anti-CD-31 antibody

PDF file, 198K, Forced over-expression of ZIC2 in NIH-3T3 cells produces increased cell proliferation and knockdown of ZIC2 protein inhibits proliferation of ovarian cancer cell lines.

Heatmap of SCNA profiles of tumor tissues and plasma. Genome-wide patterns distribution of SCNA in plasma (PL-1) and matched tumor tissues from 35 patients at diagnosis (T0). X- axis depicts the loci of the 23 chromosomes, while y-axis indicates the copy number calls for each sample, grouped by patient. Copy number gains are represented in red; copy number losses, blue. Gray, neutral. White cells are the centromeres of each chromosome. The name of each plasma samples is shown in red on the left side of the figure.

The levels of ctDNA of 15 HGS-EOC patients are shown in relation to the patient's clinical history. The TF% in cfDNA is reported for each plasma sample (purple line) withdrawn against the CA-125 levels (green line). Red bars refer to the tumor volume calculated from TC analysis as described in Supplementary methods (). * indicates patients for whom TC scan is not available at time of diagnosis and/or at recurrence. Figures S11 and S12 specifically show cases where the TF% in cfDNA does not anticipate tumor progression.

Patients' database

PDF file, 96K, Role in transformation of individual domains of ZIC2 protein assessed by deletion analysis.

Purpose: Gene expression profile was analyzed in 68 stage I and 15 borderline ovarian cancers to determine if different clinical features of stage I ovarian cancer such as histotype, grade, and survival are related to differential gene expression.Experimental Design: Tumors were obtained directly at surgery and immediately frozen in liquid nitrogen until analysis. Glass arrays containing 16,000 genes were used in a dual-color assay labeling protocol.Results: Unsupervised analysis identified eight major patient partitions, one of which was statistically associated to overall survival, grading, and histotype and another with grading and histotype. Supervised analysis allowed detection of gene profiles clearly associated to histotype or to degree of differentiation. No difference was found between borderline and grade 1 tumors. As to recurrence, a subset of genes able to differentiate relapsers from nonrelapsers was identified. Among these, cyclin E and minichromosome maintenance protein 5 were found particularly relevant, as their expression was inversely correlated to progression-free survival (P = 0.00033 and 0.017, respectively).Conclusions: Specific molecular signatures define different histotypes and prognosis of stage I ovarian cancer. Mucinous and clear cells histotypes can be distinguished from the others regardless of tumor grade. Cyclin E and minichromosome maintenance protein 5, whose expression was found previously to be related to a bad prognosis of advanced ovarian cancer, appear to be potential prognostic markers in stage I ovarian cancer too, independent of other pathologic and clinical variables.

Purpose: Epithelial ovarian cancer (EOC) is one of the most lethal gynecologic diseases, with survival rate virtually unchanged for the past 30 years. EOC comprises different histotypes with molecular and clinical heterogeneity, but up till now the present gold standard platinum-based treatment has been conducted without any patient stratification. The aim of the present study is to generate microRNA (miRNA) profiles characteristic of each stage I EOC histotype, to identify subtype-specific biomarkers to improve our understanding underlying the tumor mechanisms.Experimental Design: A collection of 257 snap-frozen stage I EOC tumor biopsies was gathered together from three tumor tissue collections and stratified into independent training (n = 183) and validation sets (n = 74). Microarray and quantitative real-time PCR (qRT-PCR) were used to generate and validate the histotype-specific markers. A novel dedicated resampling inferential strategy was developed and applied to identify the highest reproducible results. mRNA and miRNA profiles were integrated to identify novel regulatory circuits.Results: Robust miRNA markers for clear cell and mucinous histotypes were found. Specifically, the clear cell histotype is characterized by a five-fold (log scale) higher expression of miR-30a and miR-30a*, whereas mucinous histotype has five-fold (log scale) higher levels of miR-192/194. Furthermore, a mucinous-specific regulatory loop involving miR-192/194 cluster and a differential regulation of E2F3 in clear cell histotype were identified.Conclusions: Our findings showed that stage I EOC histotypes have their own characteristic miRNA expression and specific regulatory circuits. Clin Cancer Res; 19(15); 4114–23. ©2013 AACR.

Purpose:This study was aimed at investigating whether the PPARγ agonist pioglitazone—given in combination with trabectedin—is able to reactivate adipocytic differentiation in myxoid liposarcoma (MLS) patient-derived xenografts, overcoming resistance to trabectedin.Experimental Design:The antitumor and biological effects of trabectedin, pioglitazone, and the combination of the two drugs were investigated in nude mice bearing well-characterized MLS xenografts representative of innate or acquired resistance against trabectedin. Pioglitazone and trabectedin were given by daily oral and weekly i.v. administrations, respectively. Molecular studies were performed by using microarrays approach, real-time PCR, and Western blotting.Results:We found that the resistance of MLS against trabectedin is associated with the lack of activation of adipogenesis. The PPARγ agonist pioglitazone reactivated adipogenesis, assessed by histologic and gene pathway analyses. Pioglitazone was well tolerated and did not increase the toxicity of trabectedin. The ability of pioglitazone to reactivate adipocytic differentiation was observed by morphologic examination, and it is consistent with the increased expression of genes such as ADIPOQ implicated in the adipogenesis process. The determination of adiponectin by Western blotting constitutes a good and reliable biomarker related to MLS adipocytic differentiation.Conclusions:The finding that the combination of pioglitazone and trabectedin induces terminal adipocytic differentiation of some MLSs with the complete pathologic response and cure of tumor-bearing mice provides a strong rationale to test the combination of trabectedin and pioglitazone in patients with MLS.

Dynamic changes of TF and CNI in patients receiving front line chemotherapy. Bar plots of average percentage distribution of TF (A, B) or CNI (C, D) values in plasma of HGS-EOC at T=0 and at different time points of the "treatment window" (A0 to A1, and N0 to N4). Data are separated according to the different protocols they received: panels A, and C depict patients receiving CT while panels B and D depict patients receiving NACT protocol. Within each box, the horizontal line represents the median distribution, the box margins the 25% and 75% percentiles, and the whiskers the 5% and 95% percentiles. Blue circles, Pt-ps patients. Red circles, Pt-r patients. Green circles, Pt-s patients. *, p< 0.05; ** p

Supplementary Data from Analysis of Gene Expression in Early-Stage Ovarian Cancer

PDF file - 162K, Human Ovarian Surface Epithelium (HOSE) experiments and analyses.

Purpose:High-grade serous epithelial ovarian cancer (HGS-EOC) is defined by high levels of somatic copy-number alterations (SCNA) with marked spatial and temporal tumor heterogeneity. Biomarkers serving to monitor drug response and detect disease recurrence are lacking, a fact which reflects an unmet clinical need.Experimental Design:A total of 185 plasma samples and 109 matched tumor biopsies were collected from 46 patients with HGS-EOC, and analyzed by shallow whole-genome sequencing (sWGS). The percentage of tumor fraction (TF) in the plasma was used to study the biological features of the disease at the time of diagnosis (T0) and correlated with patients' survival. Longitudinal analysis of TF was correlated with CA-125 levels and radiological images to monitor disease recurrence.Results:Gain in the clonal regions, 3q26.2 and 8q24.3, was observed in the 87.8% and 78.05% of plasma samples, suggesting that plasma sWGS mirrors solid biopsies. At T0, multivariate analysis revealed that plasma TF levels were an independent prognostic marker of relapse (P < 0.022). After platinum (Pt)-based treatment, circulating tumor DNA (ctDNA) analysis showed a change in the heterogeneous pattern of genomic amplification, including an increased frequency of amplification, compared with before Pt-based treatment in the 19p31.11 and 19q13.42 regions. TF in serially collected ctDNA samples outperformed CA-125 in anticipating clinical and radiological progression by 240 days (range, 37–491).Conclusions:Our results support the notion that sWGS is an inexpensive and useful tool for the genomic analysis of ctDNA in patients with HGS-EOC to monitor disease evolution and to anticipate relapse better than serum CA-125, the routinely used clinical biomarker.See related commentary by Dhani, p. 2372

The levels of ctDNA of 15 HGS-EOC patients are shown in relation to the patient's clinical history. The TF% in cfDNA is reported for each plasma sample (purple line) withdrawn against the CA-125 levels (green line). Red bars refer to the tumor volume calculated from TC analysis as described in Supplementary methods (). * indicates patients for whom TC scan is not available at time of diagnosis and/or at recurrence. Figures S11 and S12 specifically show cases where the TF% in cfDNA does not anticipate tumor progression.

Consort Diagram. The flow diagram depicts how the total 46 HGS-EOC cases were retrospectively selected in this study to generate genome wide somatic copy number profiles using shallow whole genome sequencing analysis.

Genomic plots showing potential genes of interest in the 11q13.3 region (A), 19p13.11 region (B), and 19q13.42. The red square indicates the position of the region within the chromosomes. Coordinates refer to the hg38 genome assembly.

Correlation index of SCNA profiles before and after chemotherapy in plasma samples. Box plots and points represent the distribution of the Pearson correlation index of SCNA profiles obtained at different time points in the "therapeutic window" (see Figure 1 in the main text) with the one measured at baseline (T=0). Horizontal lines within each box indicate median values, while box whiskers are estimated as 1.5 times the standard deviations.

PDF file - 176K, miRNA expression analysis methodology, the complete lists of differentially expressed genes for each subtype comparison and the relative resampling score.

The levels of ctDNA of 15 HGS-EOC patients are shown in relation to the patient's clinical history. The TF% in cfDNA is reported for each plasma sample (purple line) withdrawn against the CA-125 levels (green line). Red bars refer to the tumor volume calculated from TC analysis as described in Supplementary methods (). * indicates patients for whom TC scan is not available at time of diagnosis and/or at recurrence. Figures S11 and S12 specifically show cases where the TF% in cfDNA does not anticipate tumor progression.

Heatmap of key genes in adipogenesis process.

List of genes encoded on cytoband 19q13.42 within the significant focal regions.

Supplementary data, material and results. Figure S1.1. Overall analysis schema Table S1.1 REMARK form. Figure S1.2 a) Kaplan Meier curves of patients divided by tumor grades (OS and PFS model). The table below summarizes the models used. Figure S1.2 b) Kaplan Meier curves of patients divided in patients who received chemotherapy (Yes) and patients who did not undergo chemotherapy (No). Figure S1.2 c) Kaplan Meier curves of patients divided in FIGO substages. The table below summarizes the models used.

Supplementary Results

PDF file - 78K, Tissue sample collection.

Supplemental Figures 1-8.

List of genes encoded on cytoband 19p13.11 within the significant focal regions.

PDF file - 217K, The clinical stratification of the 76 selected patients used for the integrative analysis and the MAGIA2 complete results.

PDF file, 82K, Univariate analysis. Linear regression analysis between histopathological features and ZIC2 expression. Units are 103 multiplied.

Supplementary Methods

PDF file, 87K, Main characteristics of patients and tissues. A, clinical and histopathological parameters of the 232 EOT biopsies enrolled in the study and subdivided into tumor tissue collection A and B. UK unknown. n, number of patients enrolled. NA, not available.

Antitumor activity of trabectedin, pioglitazone and their combination in ML004 PDX.