Prediction Analysis Of Microarrays Research Articles

Abstract PO2-16-04: Proteogenomic characterization of non-luminal A versus luminal A breast cancer tumor cells enriched by laser microdissection

Abstract Introduction: Breast cancer (BC) is classified into four widely-accepted intrinsic subtypes based on PAM (Prediction Analysis of Microarray) 50 gene expression profiles: Luminal A (LumA), Luminal B (LumB), Her2-enriched (Her2) and basal-like (Basal). Recent multi-omic studies of human BC have identified many potential therapeutic biomarkers for these subtypes and have also revealed the extensive molecular heterogeneity of the disease. Patients with LumA tumors generally have better outcomes because these tumors are typically slower-growing and responsive to hormone therapy. However, there is still considerable variation in the clinical outcomes of patients with non-LumA tumors. Here, we strive to understand the proteogenomic characteristics of intrinsically-defined non-LumA tumors in reference to LumA tumors. Methods: A total of 117 retrospectively-collected, untreated primary breast tumor specimens, with a focus on non-LumA subtypes, were selected from the Clinical Breast Care Project and consented using a HIPAA-compliant, IRB-approved protocol. The study cohort had a median patient follow-up of 9.3 years, enabling clinical outcome-related analyses. Immunohistochemistry (IHC) subtyping was used to enrich the cohort with non-LumA subtypes. Breast tumors were embedded in OCT (Optimal Cutting Temperature) compound and processed by laser microdissection (LMD) to enrich for tumor cells. DNA, RNA, and protein were simultaneously extracted from each tumor using the illustra triplePrep kit. Paired-end RNA sequencing and whole genome sequencing (WGS) were performed for 117 and 99 tumors, respectively, using the Illumina HiSeq platform. Quantitative global proteomics and phosphoproteomic analyses were performed on 112 and 50 tumors, respectively, using isobaric TMT 6-plex labeling with the “universal reference” strategy. Results: We observed significantly lower stromal, immune and microenvironment scores in non-basal-like tumors, especially the LumA subtype from LMD-processed samples, compared to that of the same tumors but bulk-processed in TCGA-breast cancer study. There was also significantly lower stromal gene expression in LMD LumA compared to TCGA LumA. Unlike a recent report on proteomics clustering of bulk-processed tumors, we did not observe a stromal-enriched cluster, probably because the use of LMD minimized stromal components. Many common patterns of somatic mutations were observed among non-LumA tumors, such as dominant TP53 mutations and 5q deletion, suggesting a potentially common cell-of-origin for non-LumA tumors, such as an ER negative cancer stem cell or progenitor cell for these tumor subtypes. Cell proliferation-associated genes were amplified and proliferation pathways strongly enriched in non-LumA tumors. In addition to this, we identified two distinct phosphoproteomic profiles for relapsed and relapse-free basal-like BC. Moreover, we also identified 17 differentially expressed phosphopeptides that could identify cases of relapsed and relapse-free basal-like BC with a significant difference in progression-free interval of surviving cases. Conclusion: This study provides an integrated proteogenomic characterization of non-LumA vs LumA tumor specimens enriched for cancer cells. We identified many common features of non-LumA tumors and also identified phosphopeptides that could serve as potential biomarkers for less aggressive basal-like BC and possibly guide treatment selections. Disclaimer: The contents of this publication are the sole responsibility of the author(s) and do not necessarily reflect the views, opinions or policies of USUHS, HJF, the DoD or the Departments of the Army, Navy or Air Force or the DOE or PNNL. Mention of trade names, commercial products, or organizations does not imply endorsement by the U.S. Government. Citation Format: Praveen Kumar Raj Kumar, Xiaoying Lin, Tao Liu, Lori Sturtz, Marina Gritsenko, Vladislav Petyuk, Tyler Sagendorf, Brenda Deyarmin, Jianfang Liu, Anupama Praveen-Kumar, Guisong Wang, Jason McDermott, Anil Shukla, Ronald Moore, Matthew Monroe, Bobbie-Jo Webb-Robertson, Jeffrey Hooke, Leigh Fantacone-Campbell, Brad Mostoller, Leonid Kvecher, Jennifer Kane, Jennifer Melley, Stella Somiari, Patrick Soon-Shiong, Richard Smith, Richard Mural, Karin Rodland, Craig Shriver, Albert Kovatich, Hai Hu. Proteogenomic characterization of non-luminal A versus luminal A breast cancer tumor cells enriched by laser microdissection [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-16-04.

Ensemble Learning for Higher Diagnostic Precision in Schizophrenia Using Peripheral Blood Gene Expression Profile.

Stigma contributes to a significant part of the burden of schizophrenia (SCZ), therefore reducing false positives from the diagnosis would be liberating for the individuals with SCZ and desirable for the clinicians. The stigmatization associated with schizophrenia advocates the need for high-precision diagnosis. In this study, we present an ensemble learning-based approach for high-precision diagnosis of SCZ using peripheral blood gene expression profiles. The machine learning (ML) models, support vector machines (SVM), and prediction analysis for microarrays (PAM) were developed using differentially expressed genes (DEGs) as features. The SCZ samples were classified based on a voting ensemble classifier of SVM and PAM. Further, microarray-based learning was used to classify RNA sequencing (RNA-Seq) samples from our case-control study (Pune-SCZ) to assess cross-platform compatibility. Ensemble learning using ML models resulted in a significantly higher precision of 80.41% (SD: 0.04) when compared to the individual models (SVM-radial: 71.69%, SD: 0.04 and PAM 77.20%, SD: 0.02). The RNA sequencing samples from our case-control study (Pune-SCZ) resulted in a moderate precision (59.92%, SD: 0.05). The feature genes used for model building were enriched for biological processes such as response to stress, regulation of the immune system, and metabolism of organic nitrogen compounds. The network analysis identified RBX1, CUL4B, DDB1, PRPF19, and COPS4 as hub genes. In summary, this study developed robust models for higher diagnostic precision in psychiatric disorders. Future efforts will be directed towards multi-omic integration and developing "explainable" diagnostic models.

NCAPH drives breast cancer progression and identifies a gene signature that predicts luminal a tumour recurrence.

Luminal A tumours generally have a favourable prognosis but possess the highest 10-year recurrence risk among breast cancers. Additionally, a quarter of the recurrence cases occur within 5 years post-diagnosis. Identifying such patients is crucial as long-term relapsers could benefit from extended hormone therapy, while early relapsers might require more aggressive treatment. We conducted a study to explore non-structural chromosome maintenance condensin I complex subunit H's (NCAPH) role in luminal A breast cancer pathogenesis, both in vitro and in vivo, aiming to identify an intratumoural gene expression signature, with a focus on elevated NCAPH levels, as a potential marker for unfavourable progression. Our analysis included transgenic mouse models overexpressing NCAPH and a genetically diverse mouse cohort generated by backcrossing. A least absolute shrinkage and selection operator (LASSO) multivariate regression analysis was performed on transcripts associated with elevated intratumoural NCAPH levels. We found that NCAPH contributes to adverse luminal A breast cancer progression. The intratumoural gene expression signature associated with elevated NCAPH levels emerged as a potential risk identifier. Transgenic mice overexpressing NCAPH developed breast tumours with extended latency, and in Mouse Mammary Tumor Virus (MMTV)-NCAPHErbB2 double-transgenic mice, luminal tumours showed increased aggressiveness. High intratumoural Ncaph levels correlated with worse breast cancer outcome and subpar chemotherapy response. A 10-gene risk score, termed Gene Signature for Luminal A 10 (GSLA10), was derived from the LASSO analysis, correlating with adverse luminal A breast cancer progression. The GSLA10 signature outperformed the Oncotype DX signature in discerning tumours with unfavourable outcomes, previously categorised as luminal A by Prediction Analysis of Microarray 50 (PAM50) across three independent human cohorts. This new signature holds promise for identifying luminal A tumour patients with adverse prognosis, aiding in the development of personalised treatment strategies to significantly improve patient outcomes.

The Prediction Analysis of Microarray 50 (PAM50) Gene Expression Classifier Utilized in Indeterminate-Risk Breast Cancer Patients in Hungary: A Consecutive 5-Year Experience.

Breast cancer has been categorized into molecular subtypes using immunohistochemical staining (IHC) and fluorescence in situ hybridization (FISH) since the early 2000s. However, recent research suggests that gene expression testing, specifically Prosigna® Prediction Analysis of Microarray 50 (PAM50), provides more accurate classification methods. In this retrospective study, we compared the results of IHC/FISH and PAM50 testing. We also examined the impact of various PAM50 parameters on overall survival (OS) and progression-free survival (PFS). We analyzed 42 unilateral breast cancer samples, with 18 classified as luminal A, 10 as luminal B, 8 as Human epidermal growth factor receptor 2 (HER2)-positive, and 6 as basal-like using PAM50. Interestingly, 17 out of the 42 samples (40.47%) showed discordant results between histopathological assessment and the PAM50 classifier. While routine IHC/FISH resulted in classification differences for a quarter to a third of samples within each subtype, all basal-like tumors were misclassified. Hormone receptor-positive tumors (hazard rate: 8.7803; p = 0.0085) and patients who had higher 10-year recurrence risk scores (hazard rate: 1.0539; p = 0.0201) had shorter OS and PFS. Our study supports the existing understanding of molecular subtypes in breast cancer and emphasizes the overlap between clinical characteristics and molecular subtyping. These findings underscore the value of gene expression profiling, such as PAM50, in improving treatment decisions for breast cancer patients.

A machine learning method integrating ECG and gated SPECT for cardiac resynchronization therapy decision support.

Cardiac resynchronization therapy (CRT) has been established as an important therapy for heart failure. Mechanical dyssynchrony has the potential to predict responders to CRT. The aim of this study was to report the development and the validation of machine learning models which integrate ECG, gated SPECT MPI (GMPS), and clinical variables to predict patients' response to CRT. This analysis included 153 patients who met criteria for CRT from a prospective cohort study. The variables were used to model predictive methods for CRT. Patients were classified as "responders" for an increase of LVEF ≥ 5% at follow-up. In a second analysis, patients were classified as "super-responders" for an increase of LVEF ≥ 15%. For ML, variable selection was applied, and Prediction Analysis of Microarrays (PAM) approach was used to model response while Naïve Bayes (NB) was used to model super-response. These ML models were compared to models obtained with guideline variables. PAM had AUC of 0.80 against 0.72 of partial least squares-discriminant analysis with guideline variables (p = 0.52). The sensitivity (0.86) and specificity (0.75) were better than for guideline alone, sensitivity (0.75) and specificity (0.24). Neural network with guideline variables was better than NB (AUC = 0.93 vs. 0.87) however without statistical significance (p = 0.48). Its sensitivity and specificity (1.0 and 0.75, respectively) were better than guideline alone (0.78 and 0.25, respectively). Compared to guideline criteria, ML methods trended toward improved CRT response and super-response prediction. GMPS was central in the acquisition of most parameters. Further studies are needed to validate the models.

Abstract 1904: Somatic copy number alterations and gene expression signatures are associated with aggressive tumor biology in African Americans and European Americans with renal cell carcinoma

Abstract Background: Renal cell carcinoma (RCC) is the deadliest of all urologic cancers in the US. RCC survival is highly dependent on histology. Survival rates range between 55-60% for clear cell RCC (ccRCC) and 80-90% for papillary RCC (pRCC) patients. African Americans (AAs) have lower survival rates when compared to their European American (EA) counterparts for both histologies (ccRCC 67.1 vs 72.6, pRCC 80.5 vs 87.5). RCC survival disparities may be accredited to population differences in tumor biology, such as somatic copy number alteration (SCNA) profiles of tumor suppressor and oncogenes, and differential gene expression signatures with associated cellular pathway dysregulation. SCNAs are the most common structural change in the human genome and ccRCC and pRCC-specific SCNAs have been associated with clinical outcomes. ClearCode34, a 34 gene expression signature, was developed to classify ccRCC tumors into two distinct ccRCC subtypes (ccA and ccB). The ccB subtype is associated with worse survival and features like increased tumor size, grade, and metastasis rates. We hypothesize there are more aggressive tumor biology phenotypes in AA RCC patients compared to their European counterparts. Methods: Comparative genome-wide, chromosomal, and gene SCNA analyses were performed in AA and EA ccRCC and pRCC patients in TCGA using Partek Genomics Suite 7 (1-way ANOVA, P<0.05). Prediction analysis of microarray in R was used to classify ccRCC patients into ccA and ccB subtypes by race based on the ClearCode34 gene expression signature (t-test, P<0.05). Differentially expressed genes (DEGs) between AA and EA patients were identified using Partek Genomics Suite 7 (1-way ANOVA, P<0.05). Gene Set Enrichment Analysis (GSEA) was performed using a pre-ranked gene list (Fold changes >2 or <-2, P<0.01) and two Molecular Signatures Database collections (C6 oncogenic signatures, C7 immunologic signatures). Results: AA-enriched SCNVs were identified on chromosomes 1, 3, 6, 7, and 12 in ccRCC patients, and chromosomes 1, 4, 6, 7, 8, 10, 16 and 20 in pRCC patients. Six cytobands and genes experienced a SCNV across both histologies (1p21.1, 1p21.3, 6p12.1, 6p12.3, 6p21.1, 6q12), BMP5 (on 6p12.1), CD2AP (on 6p12.3), ENPP4 (on 6p21.1), and SLC44A3-AS1 (on 1p21.3). AAs and EAs had significantly different frequencies of ccA (AA 33.9% vs EA 52.2%) and ccB subtypes (AA 66.1% vs EA 47.8%). Unique DEGs were found between AA and EA ccA and ccB patients (131 ccA only, 165 ccB only), which corresponded with up- and down-regulation in kidney cancer with oncogenic KRAS (C6) and various immunologic pathways (C7). Conclusions: Population-specific SCNAs, ccB subtypes, and pathways may help explain differences in RCC tumor biology amongst AA and EA patients. Future Directions: Replicate these findings in the Cooperative Human Tissue Network/Geisinger validation cohort. Citation Format: Samantha Greenberg, Alex Bart, Destiny Ortiz Fernandez, Heinric Williams, Khadijah A. Mitchell. Somatic copy number alterations and gene expression signatures are associated with aggressive tumor biology in African Americans and European Americans with renal cell carcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1904.

Tumor biomarkers and efficacy in patients treated with trastuzumab emtansine + pertuzumab versus standard of care in HER2-positive early breast cancer: an open-label, phase III study (KRISTINE)

BackgroundKRISTINE is an open-label, phase III study of trastuzumab emtansine + pertuzumab (T-DM1 + P) versus docetaxel + carboplatin + trastuzumab + pertuzumab (TCH + P) in patients with HER2-positive, stage II–III breast cancer. We investigated the association of biomarkers with clinical outcomes in KRISTINE.MethodsPatients were randomized to receive neoadjuvant T-DM1 + P or TCH + P and assessed for pathologic complete response (pCR; ypT0/is, ypN0). HER2 status (per central assessment), hormone receptor status, PIK3CA mutation status, HER2/HER3 mRNA levels, tumor-infiltrating lymphocyte levels, PD-L1 status, and NanoString data were analyzed. pCR rates by treatment arm were compared across biomarker subgroups. Analyses were descriptive.ResultsBiomarker analyses included data from all 444 patients (T-DM1 + P, n = 223; TCH + P, n = 221) enrolled in KRISTINE. Biomarker distribution was balanced across treatment arms. All subgroups with higher HER2 amplification/expression and immune marker levels showed numerically higher pCR rates in both arms. Mutated versus non-mutated PIK3CA tumors were associated with numerically lower pCR rates in the T-DM1 + P arm but not in the TCH + P arm. In a multivariate analysis, Prediction Analysis of Microarray with the 50-gene classifier (PAM50) HER2-enriched subtype, HER2 gene ratio ≥ 4, and PD-L1-positive status positively influenced the pCR rate. Biomarkers associated with lower pCR rates (e.g., low HER2 levels, positive hormone receptor status, mutated PIK3CA) were more likely to co-occur. Dynamic on-treatment biomarker changes were observed. Differences in the treatment effects for T-DM1 + P versus TCH + P were similar to those observed in the intent-to-treat population for the majority of the biomarker subgroups.ConclusionsAlthough our biomarker analysis did not identify a subgroup of patients that benefited from neoadjuvant T-DM1 + P versus TCH + P, the data revealed that patients with higher HER2 amplification/expression and immune marker levels had improved response irrespective of treatment arm. These analyses confirm the role of HER2 tumor biology and the immune microenvironment in influencing pCR in the neoadjuvant setting and reaffirm the molecular diversity of HER2-positive breast cancer.Trial Registration: ClinicalTrials.gov NCT02131064. Registered 06 May 2014.

ENST00000535926 is an unfavorable prognosis-related and tumor-promoting transcript of the CHPF gene in luminal A and B breast cancer

Chondroitin sulfate synthase 2 (CHPF) is characterized as an oncogenic and poor prognosis-related gene in breast cancer. However, this gene has alternative splicing products encoding proteins of different lengths. Breast cancer is a group of heterogeneous tumors with distinct clinical and genomic characteristics. In this study, we explored the expression profile and prognostic value of the two transcripts of CHPF using data from The Cancer Genome Atlas (TCGA)-BRCA. The functional regulation of the two transcripts was also studied in MCF-7 and BT-474 cells. Among the two transcripts of CHPF, ENST00000535926 expression was significantly upregulated in the tumor samples and was the dominant isoform. ENST00000535926, but not ENST00000243776 upregulation, was associated with significantly worse progression-free survival (PFS) and disease-specific survival (DSS) in luminal A/B cases. However, no significant association was observed in PFS or DSS in other Prediction Analysis of Microarray 50 (PAM50) subgroups. CHPF isoform 2 protein (encoded by ENST00000535926) significantly elevated the expression of P3H1 and RCN3 at the mRNA and protein levels in MCF-7 and BT-474 cells. The effect of ENST00000535926 was significantly stronger than ENST00000243776 in promoting tumor cell colony formation. The expression of P3H1 and RCN3 was negatively correlated with CD8+ T cell infiltration but was positively correlated with cancer-associated fibroblast infiltration in luminal A/B tumors. In summary, this study revealed that ENST00000535926 is an unfavorable prognosis-related and tumor-promoting transcript of the CHPF gene in luminal A/B breast cancer.

Analysis of Tumor Microenvironment Heterogeneity among Breast Cancer Subtypes to Identify Subtype-Specific Signatures.

Breast cancer is one of the most frequent malignancies in women worldwide. According to 50-gene signature, Prediction Analysis of Microarray 50 (PAM50), breast cancer can be categorized into five molecular subtypes, and these subtypes are highly heterogeneous in different molecular characteristics. However, the landscape of their tumor microenvironment (TME) heterogeneity has not been fully researched. Using the multi-omics dataset of breast cancer from the METABRIC cohort (n = 1699), we conducted extensive analyses of TME-related features to investigate TME heterogeneity in each breast cancer subtype. We then developed a cell-based subtype set enrichment analysis to identify the subtype-specific TME cells, and further evaluate their prognostic effects. Our results illustrate that different breast cancer subtypes exhibit different TME patterns. Basal-like and HER2-enriched subtypes are associated with high immune scores, expression of most immune regulatory targets, and immune cell infiltration, suggesting that these subtypes could be defined as "immune hot" tumors and suitable for immune checkpoint blockade (ICB) therapy. In contrast, Luminal A and Luminal B subtypes are associated with low immune scores and immune cell infiltration, suggesting that these subtypes could be defined as "immune cold" tumors. Additionally, the Normal-like subtype has relatively high levels of both immune and stromal features, which indicates that the Normal-like subtype may be suitable for more diverse treatment strategies. Our study reveals the breast cancer tumor microenvironment heterogeneity across subtypes. The comprehensive analysis of breast cancer TME-related characteristics may help us to adopt a tailored treatment strategy for different subtypes of patients.

COL11A1 serves as a biomarker for poor prognosis and correlates with immune infiltration in breast cancer.

Breast cancer is the malignant tumor with the highest incidence rate at present, and its incidence rate ranks first in the female population. COL11A1 is an important component of collagen XI and is considered to play an important role in a variety of connective tissue diseases. Recent studies have shown that COL11A1 is associated with the occurrence and development of many kinds of malignant tumors. However, its prognostic value in breast cancer and its correlation with immune cell infiltration in tumor tissue are not clear. In this paper, we reveal the prognostic value of COL11A1 in breast cancer and its tumor immune-related function through in-depth bioinformatics analysis. The expression of COL11A1 is abnormally upregulated in breast cancer and is significantly related to the poor prognosis of breast cancer. In the analysis of the clinical characteristics of the patients, we found that the expression level of COLL11A1 was closely related to lymph node metastasis, PAM50 (Prediction Analysis of Microarray 50) expression, clinical stage and so on. Gene Ontology (GO) and Kyoto Encyclopedia of Gene and Genome (KEGG) all suggest that COL11A1 is related to tumor immunity. Further study found that the COL11A1 expression was significantly correlated with the degree of immune infiltration and the expression of a variety of immune cell markers in tumor tissue. More importantly, COL11A1 can affect the prognosis of breast cancer patients by participating in the regulation of tumor immune infiltration. Therefore, we believe that COL11A1 is a very potential target for diagnosis and treatment of breast cancer.

Different thresholding methods on Nearest Shrunken Centroid algorithm

This article considers the impact of different thresholding methods to the Nearest Shrunken Centroid algorithm, which is popularly referred as the Prediction Analysis of Microarrays (PAM) for high-dimensional classification. PAM uses soft thresholding to achieve high computational efficiency and high classification accuracy but in the price of retaining too many features. When applied to microarray human cancers, PAM selected 2611 features on average from 10 multi-class datasets. Such a large number of features make it difficult to perform follow up study. One reason behind this problem is the soft thresholding, which is known to produce biased parameter estimate in regression analysis. In this article, we extend the PAM algorithm with two other thresholding methods, hard and order thresholding, and a deep search algorithm to achieve better thresholding parameter estimate. The modified algorithms are extensively tested and compared to the original one based on real data and Monte Carlo studies. In general, the modification not only gave better cancer status prediction accuracy, but also resulted in more parsimonious models with significantly smaller number of features.

SRY-box transcription factor 10 is a highly specific biomarker of basal-like breast cancer.

Basal-like breast cancer is an aggressive molecular subtype associated with younger age and early relapse. Most cases lack expression of oestrogen receptor (ER), progesterone receptor, and human epidermal growth factor receptor 2, limiting targeted therapeutic options. Basal-like breast cancer is defined by the expression of genes in the outer/basally located epithelial layer of mammary glands, including those encoding cytokeratin (CK) 5 and CK14, and epidermal growth factor receptor (EGFR). SRY-box transcription factor 10 (SOX10), for which there is a readily available immunohistochemical stain, is expressed in a subset of breast cancers, particularly triple-negative carcinomas. In this study, we sought to: (i) assess the association between SOX10 expression and intrinsic molecular subtypes as defined by Prediction Analysis of Microarray 50 (PAM50) gene expression; and (ii) compare the performance of SOX10 with that of other surrogate markers of the basal-like subtype, including CK5, EGFR, nestin, and inositol polyphosphate 4-phosphatase type II (INPP4B). SOX10 immunostaining was performed on tissue microarrays constructed from a contemporary series enriched for ER-negative and weakly ER-positive cancers that had also undergone PAM50 gene profiling. A total of 211 cases were informative for both SOX10 immunohistochemistry and PAM50 subtype, including 103 basal-like cancers. Staining for SOX10 was positive in 73 of 103 basal-like cancers and in only two of 108 cancers of other subtypes (P < 0.001), resulting in a sensitivity of 70.9% and a specificity of 98.1%. SOX10 was more specific than the other tested basal markers, and the results were independent of ER status. SOX10 is a moderately sensitive, but highly specific, immunohistochemical biomarker for the basal-like intrinsic subtype of breast cancer, which, unlike other commonly used immunohistochemical biomarkers, is independent of hormone receptor status.

Announcement of the 2022 winners of the Spychala and Women in Science Awards

Announcement of the 2022 winners of the Spychala and Women in Science Awards

Exploratory Profiling of Extracellular MicroRNAs in Cerebrospinal Fluid Comparing Leptomeningeal Metastasis with Other Central Nervous System Tumor Statuses.

The diagnosis of leptomeningeal metastasis (LM) is often difficult due to the paucity of cancer cells in cerebrospinal fluid (CSF) and nonspecific findings on neuroimaging. Investigations of extracellular microRNAs (miRNAs) in CSF could be used for both the diagnosis and study of LM pathogenesis because they reflect the activity of disseminating cancer cells. We isolated CSF extracellular miRNAs from patients (n = 65) of different central nervous system tumor statuses, including cancer control, healthy control, LM, brain metastasis (BM), and primary brain tumor (BT) groups, and performed miRNA microarrays. In unsupervised clustering analyses, all LM and two BM samples showed unique profiles. Among 30 miRNAs identified for LM-specific biomarkers via a Prediction Analysis of Microarrays, miR-335-5p and miR-34b-3p were confirmed in both the discovery and validation samples (n = 23). Next, we performed a significance analysis of the microarray (SAM) to extract discriminative miRNA profiles of two selected CSF groups, with LM samples revealing a greater number of discriminative miRNAs than BM and BT samples compared to controls. Using SAM comparisons between LM and BM samples, we identified 30 upregulated and 6 downregulated LM miRNAs. To reduce bias from different primary cancers, we performed a subset analysis with primary non-small cell lung cancer, and 12 of 13 upregulated miRNAs in LM vs. BM belonged to the upregulated miRNAs in LM. We identified possible target genes and their biological processes that could be affected by LM discriminative miRNAs in NSCLC using the gene ontology database. In conclusion, we identified a unique extracellular miRNA profile in LM CSF that was different from BM, suggesting the use of miRNAs as LM biomarkers in studies of LM pathogenesis.

Inherited GATA3 genetic Variants Are Associated With Childhood BCR-ABL1-Like Acute Lymphoblastic Leukemia and Increased Risk Of Relapse

Recent genomic profiling of acute lymphoblastic leukemia (ALL) has identified a novel high-risk subtype with a gene expression signature similar to that of BCR-ABL1-positive (Ph+) ALL and a poor prognosis. This novel BCR-ABL1-like ALL subtype is characterized by genomic aberrations targeting lymphoid development, cytokine receptor, and tyrosine kinase signaling pathways, e.g., IKZF1 deletion, CRLF2 rearrangements, and JAK mutations. However, the role of the inherited genetic variation in the pathogenesis of BCR-ABL1-like ALL remains unknown.Taking a genome-wide association study (GWAS) approach, we sought to examine germline single-nucleotide polymorphisms (SNPs) for their association with the predisposition to BCR-ABL1-like ALL. In the discovery GWAS, we compared germline SNP genotype frequencies between ALL cases with (N=75) vs. without (N=436) the BCR-ABL1-like expression signature enrolled on the high risk Children's Oncology Group (COG) AALL0232 protocol, and between BCR-ABL1-like ALL (N=75) vs. non-ALL controls (N=6,661). BCR-ABL1-like ALL was identified by prediction analysis of microarrays. After adjusting for genetic ancestry, we observed a single susceptibility locus on 10p14 signified by two SNPs within the GATA3 gene with genome-wide significant association (P=1.05×10-8 and P=2.62×10-7 [BCR-ABL1-like ALL vs. non-BCR-ABL1-like ALL]; P=2.17×10-14 and P=4.94×10-12 [BCR-ABL1-like ALL vs. non-ALL controls]). In a replication cohort of high-risk ALL, namely COG P9906, the disease risk alleles at both GATA3 polymorphisms were consistently over-represented in BCR-ABL1-like ALL (N=32) compared with 139 non-BCR-ABL1-like ALL (P=0.01 and P=0.004, respectively) and compared with an independent set of 5,755 non-ALL controls (P=3.69×10-5 and P=0.0001, respectively). This GATA3 SNP was also associated with GATA3 expression in lymphoblastoid cell lines (N=56, P=0.034) and in primary ALL blasts (COG AALL0232, N=511, P=9.2×10-8; COG P9906, N=173, P=3.6×10-6), indicating that this variant may function as a cis-acting element regulating GATA3 transcription. Ectopic overexpression of GATA3 in ALL cell lines led to global changes in gene expression, with a particular enrichment of genes within the BCR-ABL1-like ALL expression signature.In the discovery cohort (COG AALL0232), the GATA3 SNP was also significantly associated with CRLF2 rearrangements (P=9.64×10-7), JAK mutations (P=2.06×10-5), and IKZF1 deletion (P=0.0002), which was validated in COG P9906 (P=0.08, 0.008, and 0.002, respectively) and in a third independent cohort, COG P9905 (P=0.009, 0.05, and 0.0002, respectively). The GATA3 risk allele was further enriched among patients with multiple “BCR-ABL1-like ALL-related” lesions concurrently, with the highest frequency observed in cases with three genomic lesions, followed by those with one or two lesions, and lowest among patients without any of the three lesions (COG AALL0232, P=6.09×10-5; COG P9906, P=0.0005; and COG P9905, P=7.6×10-5). Interestingly, the association of the GATA3 SNP with BCR-ABL1-likeness remained significant after adjusting for all three genomic lesions (P=0.001).We further examined the relationships between GATA3 SNP genotype and ALL relapse. In the COG P9906 cohort, the GATA3 allele linked to BCR-ABL1-like ALL was associated with a higher risk of relapse (N=222, P=0.002) and poor early treatment response as measured by minimal residual disease (N=193, P=9.8×10-5), which was validated in COG P9905 (N=777, P=0.007 and N=710, P=0.039, respectively). Of interest, the GATA3 allele associated with relapse and BCR-ABL1-like ALL was more frequent in Hispanics (40%) than in populations of European descent (12%), consistent with the inferior treatment outcome for Hispanic children with ALL.In conclusion, we have identified common GATA3 germline variants associated with BCR-ABL1-like ALL, somatic molecular lesions characteristic of this subtype, and importantly the risk of ALL relapse. These results suggest plausible interactions between host and tumor genomic variations in ALL pathogenesis and drug response. Disclosures:Evans:St. Jude: In accordance with institutional policy, St. Jude allocates a portion of the income it receives from licensing inventions and tangible research materials to those researchers responsible for creating this intellectual property. Patents & Royalties, Under this policy, I and/or my spouse have in the past received a portion of the income St. Jude receives from licensing patent rights related to TPMT polymorphisms as clinical diagnostics., Under this policy, I and/or my spouse have in the past received a portion of the income St. Jude receives from licensing patent rights related to TPMT polymorphisms as clinical diagnostics. Other. Relling:St. Jude: In accordance with institutional policy, St. Jude allocates a portion of the income it receives from licensing inventions and tangible research materials to those researchers responsible for creating this intellectual property. Patents & Royalties, Under this policy, I and/or my spouse have in the past received a portion of the income St. Jude receives from licensing patent rights related to TPMT polymorphisms as clinical diagnostics., Under this policy, I and/or my spouse have in the past received a portion of the income St. Jude receives from licensing patent rights related to TPMT polymorphisms as clinical diagnostics. Other.

Abstract 669: A comprehensive single cell analysis of cancer-associated fibroblasts (CAFs) reveals subtypes with prognostic impact in pancreatic ductal adenocarcinoma

Abstract Introduction The presence of heterogeneity in cancer-associated fibroblasts (CAFs) of pancreatic ductal adenocarcinoma (PDAC) is well-known. The presence of mouse inflammatory CAFs (iCAFs) and myofibroblastic CAFs (myCAFs) were published earlier. Recently, we identified the first patient-derived sub-populations of CAFs (pCAFassigner) with different phenotypic and pathological characteristics. We demonstrated the presence of intra-tumoral heterogeneity in patients, differential sensitivity of CAF subtypes in co-cultures with cancer cells to existing PDAC treatment options and prognostic impact of these subtypes. The presence of antigen-presenting CAFs (apCAFs) in humans and mouse was introduced recently. However, a comprehensive, unbiased analysis is required to reconcile different CAF classifications and understand the prognostic significance of these subtypes for future clinical translation. Methods Mouse and human PDAC single-cell RNA sequencing (scRNA-seq)-based suite of analyses were performed to assign CAFs into pCAFassigner and other subtypes. Enrichment (hypergeometric) and differential gene expression (DGE) analyses were performed to identify sample- and gene-wise similarities between classifications, as well as develop and validate a new pCAFassigner gene signature. For prognostic impact, a published single sample classifier was used to identify a dominant CAF subtype in 1071 whole tumor samples, for which clinical information was available for 618 samples. The overall survival (OS) analysis was performed using Kaplan-Meier survival analysis. Results As expected, our analysis revealed the synonymity between subtype-A and myCAF (91 %), subtypes-B and -C with iCAF (69 %) and subtype D with apCAF (91 %, FDR &amp;lt; 0.001). We revised our pCAFassigner gene signature from a prediction analysis of microarrays (PAM) centroid to an 82 gene signature unique to each subtype. Given the similarities between these separate classifications, we sought to identify association of iCAF (subtype-B/C), apCAF (subtype-D) and myCAF (subtype-A) with survival. We observed a significant difference in OS with apCAF- and myCAF-dominant samples having the poorest prognosis, while patients with dominant iCAF had prolonged survival, questioning the understanding of iCAFs being pro-tumorigenic given its secretory nature and consequently, its presumed association with poor survival. Conclusions Our results indicate that while methodologies to derive each CAF classification is distinct, ultimately, they have proven to be similar. Further, we reveal prognostic impact of iCAFs, apCAFs and myCAFs, supporting the use of these CAF subtypes as independent prognostic indicators for resected PDAC and will aid better management of the disease. Citation Format: Aasia Hussain, Jiahui Ji, Prashant K. Srivastava, Anguraj Sadanandam. A comprehensive single cell analysis of cancer-associated fibroblasts (CAFs) reveals subtypes with prognostic impact in pancreatic ductal adenocarcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 669.

Lenalidomide-Mediated Erythroid Improvement in Non-Del(5q) Myelodysplastic Syndromes Is Associated with Bone Marrow Immuno-Remodeling

Introduction: The main therapeutic goal for patients with International Prognostic Scoring System (IPSS) low or intermediate risk MDS is to improve anemia. Lenalidomide yields red blood cell (RBC) transfusion independence (RBC-TI) in 60-75% of del(5q) MDS, and 40-60 % achieve complete or partial cytogenetic response.Lenalidomide also yields RBC-TI ≥8 weeks in ~25% of RBC transfusion-dependent (TD) non-del(5q) MDS patients after resistance to erythropoiesis-stimulating agent (ESA). Although reasons for lenalidomide being less effective in non-del(5q) MDS patients are unclear, it has been shown that those who responded to lenalidomide had a poor expression of erythroid genes in their bone marrow (BM).Addition of epoetin β to lenalidomide increased the erythroid response (HI-E) compared to lenalidomide alone in the GFM-LEN-EPO-2008 clinical trial (clinicaltrials.gov NCT01718379, EudraCT number 2008-008262-12), but the combination could be less efficient if epoetin β is delayed after response failure to lenalidomide alone. We have shown that the number and size of the clones within the hematopoietic stem and progenitor cell (HSPC) compartment are diminished in responding patients while the loss of response is associated with the re-increase of allele burden and/or the positive selection of clones with additional mutations. Therefore, lenalidomide efficacy may be due to a direct antitumor activity on clonal HSPCs. Interference with leukemic and stromal cells interaction and/or delivery of an immunomodulatory signal through the activation of cytotoxic T cells and NK cells have also been suggested.Methods: Appended to the GFM-LEN-EPO2008 clinical trial, we performed a transcriptomic analysis of bone marrow samples at baseline and after 4 cycles or more than 9 cycles of treatment to investigate molecular markers of response using Affymetrix U133Plus 2.0 microarrays or RNA-seq.Results: Genes differentially expressed between responders and non-responders at baseline were considered for Gene Set Enrichment Analysis (GSEA) and comprised (i) upregulated genes implicated in cell division, epigenetic modifications and translation, and (ii) downregulated genes involved in lymphocyte activation. A supervised class-prediction analysis performed by using Prediction Analysis for Microarrays (PAM) software in R revealed a 50-gene signature predictive of the response. After 4 cycles of treatment, a unique geneset encompassing B progenitor markers (B cell response) or T lymphocyte signaling (T cell response) was up-regulated in responding patients, suggesting that induction of immune cell markers reflected a different sensitivity to lenalidomide treatment.To obtain further insights on the modulation of microenvironment by lenalidomide, we took advantage of the transcriptome-based computational method called Microenvironment Cell Populations (MCP)-counter in R package to estimate the abundance of immune (T cells, CD8 T cells, cytotoxic lymphocytes, NK cells, B cells, monocytes, myeloid dendritic cells) and non-immune stromal (neutrophils, endothelial cells and fibroblasts) cell populations from RNA-sequencing data. Compared to controls, the relative abundance of T, cytotoxic, B, NK lymphocytes and myeloid dendritic cells was significantly lower in patients irrespective of the response to lenalidomide. After 4 cycles of treatment, the abundance of cytotoxic T cells significantly increased while other populations NK cells, monocytes, neutrophils, myeloid dendritic cells, endothelial cells and fibroblasts did not vary significantly. Patients with a T cell response under therapy had less abundant monocyte, and trend to less abundant B cell and endothelial cells compared to those with a B cell response. Consistent with the induction of a T cell response, IFNγ level increased significantly after treatment in responders. Altogether, these analyses demonstrate that immune cell populations including cytotoxic T cells are recruited as potential effectors of immune response upon lenalidomide.Conclusion: Our results show that lenalidomide induces an adaptive immune response that may participate in the erythroid improvement in non-del(5q) low/int-1 risk MDS. Future investigations will help understanding the architecture of immune cells and inflammation in the bone marrow microenvironment that may influence the prognosis. DisclosuresFenaux:Amgen: Honoraria, Research Funding; Astex: Honoraria, Research Funding; Novartis: Honoraria, Research Funding; Janssen: Honoraria, Research Funding; Celgene: Honoraria, Research Funding; Celgene: Honoraria, Research Funding; Amgen: Honoraria, Research Funding; Astex: Honoraria, Research Funding; Janssen: Honoraria, Research Funding; Novartis: Honoraria, Research Funding. Fontenay:Celgene: Research Funding.

MicroRNA expression profiling with a droplet digital PCR assay enables molecular diagnosis and prognosis of cancers of unknown primary.

Metastasis is responsible for the majority of cancer‐related deaths. Particularly, challenging is the management of metastatic cancer of unknown primary site (CUP), whose tissue of origin (TOO) remains undetermined even after extensive investigations and whose therapy is rather unspecific and poorly effective. Molecular approaches to identify the most probable TOO of CUPs can overcome some of these issues. In this study, we applied a predetermined set of 89 microRNAs (miRNAs) to infer the TOO of 53 metastatic cancers of unknown or uncertain origin. The miRNA expression was assessed with droplet digital PCR in 159 samples, including primary tumors from 17 tumor classes (reference set) and metastases of known and unknown origin (test set). We combined two different statistical models for class prediction to obtain the most probable TOOs: the nearest shrunken centroids approach of Prediction Analysis of Microarrays (PAMR) and the least absolute shrinkage and selection operator (LASSO) models. The molecular test was successful for all formalin‐fixed paraffin‐embedded samples and provided a TOO identification within 1 week from the biopsy procedure. The most frequently predicted origins were gastrointestinal, pancreas, breast, lung, and bile duct. The assay was applied also to multiple metastases from the same CUP, collected from different metastatic sites: The predictions showed a strong agreement, intrinsically validating our assay. The final CUPs' TOO prediction was compared with the clinicopathological hypothesis of primary site. Moreover, a panel of 13 miRNAs proved to have prognostic value and be associated with overall survival in CUP patients. Our study demonstrated that miRNA expression profiling in CUP samples could be employed as diagnostic and prognostic test. Our molecular analysis can be performed on request, concomitantly with standard diagnostic workup and in association with genetic profiling, to offer valuable indications about the possible primary site, thereby supporting treatment decisions.

A novel metabolic gene signature-based nomogram to predict overall survival in breast cancer.

BackgroundBreast cancer risk prediction is often based on clinicopathological characteristics despite the high heterogeneity derived from gene expression. Metabolic alteration is a hallmark of cancer, and thus, the integration of a metabolic signature with clinical parameters is necessary to predict disease outcomes in breast cancers.MethodsMetabolic genes were downloaded from the Gene Set Enrichment Analysis (GSEA) dataset. Genes with statistical significance in the univariate analysis were applied in the least absolute shrinkage and selection operator (LASSO) analysis to build a gene signature in the GSE20685 dataset. Clinicopathological characteristics and risk scores with prognostic significance were incorporated into the nomogram to predict the overall survival (OS) of patients. The Cancer Genome Atlas (TCGA) and GSE866166 datasets were used as the validation datasets. Time-dependent receiver operating characteristic (tROC) curves and calibration plots were used to assess the accuracy and discrimination of the model.ResultsA 55-gene metabolic gene signature (MGS) was constructed, and was significantly related to OS both in the discovery (P<0.001) and validation (P<0.001) datasets. The MGS was an independent prognostic factor and could divide patients into high- and low-risk groups regardless of their different prediction analysis of microarray 50 (PAM50) subtypes. Time-dependent ROC curves indicated that the risk scores based on the MGS [area under the ROC curve (AUC): 0.931] were superior to the those based on the American Joint Committee on Cancer (AJCC) stage (AUC: 0.781) and PAM50 (AUC: 0.675). A nomogram based on the AJCC stage and risk score could predict OS, and the calibration curves showed good agreement to the actual outcome, indicating that the nomogram may have practical utility. Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analysis indicated that this MGS was primarily enriched in amino acid pathways.ConclusionsOur results demonstrated that the MGS was superior to existing risk predictors such as PAM50 and AJCC stage. By combining clinical factors (AJCC stage) and the MGS, a nomogram was constructed and showed good predictive ability for OS in breast cancer.

Screening and Identification of Key Common and Specific Genes and Their Prognostic Roles in Different Molecular Subtypes of Breast Cancer.

Breast cancer is one of the most common cancers. Although the present molecular classification improves the treatment effect and prognosis of breast cancer, the heterogeneity of the molecular subtype remains very complex, and the applicability and effectiveness of treatment methods are still limited leading to poorer patient prognosis than expected. Further identification of more refined molecular typing based on gene expression profile will yield better understanding of the heterogeneity, improving treatment effects and prolonging prognosis of patients. Here, we downloaded the mRNA expression profiles and corresponding clinical data of patients with breast cancer from public databases and performed typical molecular typing using PAM50 (Prediction Analysis of Microarray 50) method. Comparative analyses were performed to screen the common and specific differentially expressed genes (DEGs) between cancer and corresponding para-cancerous tissues in each breast cancer subtype. The GO and KEGG analyses of the DEGs were performed to enrich the common and specific functional progress and signaling pathway involved in breast cancer subtypes. A total of 38 key common and specific DEGs were identified and selected based on the validated results, GO/KEGG enrichments, and the priority of expression, including four common DEGs and 34 specific DEGs in different subtypes. The prognostic value of these key common and specific DEGs was further analyzed to obtain useful potential markers in clinic. Finally, the potential roles and the specific prognostic values of the common and specific DEGs were speculated and summarized in total breast cancer and different subtype breast cancer based on the results of these analyses. The findings of our study provide the basis of more refined molecular typing of breast cancer, potential new therapeutic targets and prognostic markers for different breast cancer subtypes