ProteoMixture: A cell type deconvolution tool for bulk tissue proteomic data

Numerous multi-omic investigations of cancer tissue have documented varying and poor pairwise transcript:protein quantitative correlations, and most deconvolution tools aiming to predict cell type proportions (cell admixture) have been developed and credentialed using transcript-level data alone. To estimate cell admixture using protein abundance data, we analyzed proteome and transcriptome data generated from contrived admixtures of tumor, stroma, and immune cell models or those selectively harvested from the tissue microenvironment by laser microdissection from high grade serous ovarian cancer (HGSOC) tumors. Co-quantified transcripts and proteins performed similarly to estimate stroma and immune cell admixture (r ≥ 0.63) in two commonly used deconvolution algorithms, ESTIMATE or ConsensusTME. We further developed and optimized protein-based signatures estimating cell admixture proportions and benchmarked these using bulk tumor proteomic data from over 150 patients with HGSOC. The optimized protein signatures supporting cell type proportion estimates from bulk tissue proteomic data are available at https://lmdomics.org/ProteoMixture/.

Abstract 5782: The dynamic immune behavior of primary and metastatic tumors of ovarian cancer

Background: Despite advances in cancer diagnosis and therapy, high-grade serous ovarian cancer (HGSC) is often diagnosed when spread to multiple intraperitoneal areas; it is prone to metastasize to adipose-rich tissues such as the omentum. To gain a deeper understanding of the molecular determinants of the HGSC milieu, we carried out a single-cell analysis of primary and metastatic tumors of HGSC. Thus, our aim is to identify molecular mechanisms that lead to immunosuppressive mechanisms of HGSC in primary and metastatic tumors of HGSC.

Methods: Fresh HGSC surgical samples from 19 patients were collected right after surgery, dissociated, and then frozen. For single-cells analysis, cells were sorted by a viability dye and CD45+/- populations followed by Single cell 10 × 3’v3 protocol (10X Genomics) and sequenced using the NovaSeq6000 S2 sequencer. Cell Ranger toolkit v3.1.0 (10x Genomics) was applied for data processing, followed by further downstream analysis using multiple packages from R Package Seurat.

Results: To elucidate the cellular heterogeneity of HGSC, we analyzed 100,480 cells, including epithelial, lymphoid, and myeloid populations; these were identified and represented across all patients. We explored the epithelial compartment further, including a total of 21,144 cells. Given the importance of metastatic lesions for the treatment and outcome of patients with HGSC, we examined expression programs among primary and omentum tumors as well treated and untreated tumors. The major differences between primary and omental metastatic tumors included enrichment of EMT pathways and angiogenesis, as well as a decrease in the IFNα and IFNγ response in the omentum. IFNα and IFNγ response pathways were also upregulated in primary treated (P-NACT) tumors, when compared with primary untreated (P-UT) tumors. To explore the immune compartment, we clustered the immune cells across patients in 28 sub-clusters, including 19 sub-clusters of T-cells and Natural Killer cells. Altogether, the immune infiltration on P-UT indicated an immune infiltrate environment and the composition of P-UT was highly enriched in lymphoid cells with a late stage of differentiation. This finding was also demonstrated by the analysis of T cell trajectories, with a clear definition that CD8 and CD4 cells from P-UT tumors are in a late/final and exhausted stage of differentiation, while post-NACT tumors (independent of tissue type) are in early stages (naïve/central memory). The measurement of a dysfunctional score showed that P-UT tumors have the highest scores, indicating a potential increase in tumor reactivity in P-UT tumors.

Conclusions: Collectively, these data indicate that HGSC primary and omentum tumors are very distinct niches for immune cells, with primary being much more dysfunctional than omentum tissue, which could indicate that the application of immunotherapies would have different impacts in variable niches of HGSC.

Citation Format: Elaine Stur, Fuduan Peng, Pang-ning Teng, Emine Bayraktar, Min Hu, Sara Corvigno, David J. Brown, Sanghoon Lee, Kathleen N. Moore, Nicholas W. Bateman, Kathleen M. Darcy, George L. Maxwell, Thomas Conrads, Nicole Fleming, Nicholas Navin, Linghua Wang, Anil K. Sood. The dynamic immune behavior of primary and metastatic tumors of ovarian cancer. [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 5782.

Abstract 2852: Identifying circulating biomarkers of acute response and resistance to clinical ATR and Chk1 inhibitors

Objectives: Clinical ATR kinase inhibitors (ATRi) are an emerging class of therapeutics being used to treat diverse solid tumor malignancies, including ovarian cancers. The development of circulating biomarkers indicating response or resistance to clinical ATRi will support the ongoing development of this emerging class of therapeutics. This study is focused on establishing isogenic models of ATRi-resistance in ovarian cancer cell lines and identifying conserved secreted protein alterations correlating with acute response or resistance to ATRi treatment in these models. Methods: We generated isogenic cell line models of ATRi-resistant, CCNE1-amplified (OVCAR3) and non-CCNE1 amplified (OV90) ovarian cancer cell by metronomic treatment of cell lines with the ATR inhibitor AZD6738 (Astra Zeneca). We characterized sensitivity of models to AZD6738 as well as inhibitors of checkpoint kinase 1 (Chk1, LY2606368), ataxia-telangiectasia mutated (ATM, KU55933) poly (ADP-ribose) polymerase (PARP, BMN-673) and cisplatin combination by dose response assay and further assessed impact of ATRi treatment on cell cycle. Lastly, we analyzed conditioned media by LC-MS/MS-based proteomic analyses from ATRi-resistant OVCAR3 cells or following acute treatment with ATRi. Results: Metronomic treatment of ovarian cancer cells with ATRi induced resistance to ATRi irrespective of CCNE1 status. ATRi-resistant cells were also resistant to Chk1i, but not to ATM, PARP inhibitors or combinations of ATRi and cisplatin. Cell cycle analyses revealed that ATRi-sensitive cells arrest in S-phase and continue active DNA replication, whereas ATRi-resistant cells arrest in G1/S phase and cease DNA replication following ATRi treatment. Comparison of proteins identified in conditioned media revealed > 30 proteins bearing signal peptide sequences potentially mediating cellular secretion as significantly elevated in the cellular secretomes of ATRi-resistant or ATRi sensitive cells treated with ATRi, with only a single protein co-altered between these sample, i.e. alpha-N-acetylgalactosaminidase (NAGA). Conclusions: Our findings show that metronomic treatment of ovarian cancer cells with ATR inhibitors induces resistance to ATRi as well as an inhibitor of Chk1, an immediate downstream effector of ATR. Our analyses show that ATRi-resistant cells remain sensitive to inhibitors of orthogonal DNA damage response signaling pathways, i.e. ATM and PARP, as well as to cisplatin sensitization induced by ATRi co-treatment. ATRi-resistant cells further exhibit a conserved G1/S-phase cell cycle arrest response to ATRi treatment. Lastly, our preliminary secretome analyses provide proof of concept data that identification of secreted biomarkers specific for acute response or resistance to ATRi (Chk1i)-treatment is feasible.

Citation Format: Nicholas Bateman, Wei Ao, Domenic Tommarello, Kelly Conrads, Pang-ning Teng, Kathleen Darcy, Chad Hamilton, G. Larry Maxwell, Christopher Bakkenist, Thomas Conrads. Identifying circulating biomarkers of acute response and resistance to clinical ATR and Chk1 inhibitors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2852.

Abstract 4631: AKAP12 is elevated in paclitaxel-resistant ovarian cancer cells and correlates with poor ovarian cancer patient survival

Objective: Although most ovarian cancer patients respond initially to the standard of care involving carboplatinum/paclitaxel-based chemotherapy, the majority will recur with chemotherapeutically resistant disease. Unfortunately, no clinical biomarker(s) for predicting or monitoring chemotherapeutic resistance are available. We sought to identify shed and/or secreted proteins from paclitaxel-resistant human ovarian cancer cells and identify whether any of these correlate with ovarian cancer patient outcome.

Methods: A proteomic analysis of secretomes from high grade serous ovarian cancer (HGSOC) cells (OV90) was compared to that from a syngeneic paclitaxel-resistant variant (OV90-TR) and a cancer cell line developed from a chemorefractory HGSOC patient. Associations between the identified candidates and patient outcome were assessed from publicly available transcript expression data (n = 545 ovarian cancer patients) and validated independently (n = 795 ovarian cancer patients).

Results: Among the 81 differentially abundant proteins identified (q<0.05) from paclitaxel-sensitive vs -resistant HGSOC cell secretomes, AKAP12 was verified to be elevated at the protein and transcript level in all models of paclitaxel-resistant HGSOC. Further, elevated AKAP12 transcript expression was found to significantly correlate with poor progression-free (Hazard Ratio = 1.487 (± 0.148), p = 0.0085 (± 0.002) and overall (Hazard Ratio = 1.22 (± 0.08), p = 0.009 (± 0.013) survival.

Conclusions: These findings suggest that elevation and efflux of AKAP12 is characteristic of paclitaxel-resistant HGSOC cells and that elevated AKAP12 expression correlates with poor ovarian cancer patient survival.

Citation Format: Nicholas Bateman, Elizabeth Jaworski, Guisong Wang, Elizabeth Dubil, Charlotte Marcus, Kelly Conrads, Pang-ning Teng, Brian Hood, Chad Hamilton, Larry Maxwell, Kathleen Darcy, Thomas Conrads. AKAP12 is elevated in paclitaxel-resistant ovarian cancer cells and correlates with poor ovarian cancer patient survival. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4631. doi:10.1158/1538-7445.AM2015-4631

Elevated AKAP12 in paclitaxel-resistant serous ovarian cancer cells is prognostic and predictive of poor survival in patients

A majority of high-grade (HG) serous ovarian cancer (SOC) patients develop resistant disease despite high initial response rates to platinum/paclitaxel-based chemotherapy. We identified shed/secreted proteins in preclinical models of paclitaxel-resistant human HGSOC models and correlated these candidate proteins with patient outcomes using public data from HGSOC patients. Proteomic analyses of a HGSOC cell line secretome was compared to those from a syngeneic paclitaxel-resistant variant and from a line established from an intrinsically chemorefractory HGSOC patient. Associations between the identified candidate proteins and patient outcome were assessed in a discovery cohort of 545 patients and two validation cohorts totaling 795 independent SOC patients. Among the 81 differentially abundant proteins identified (q < 0.05) from paclitaxel-sensitive vs -resistant HGSOC cell secretomes, AKAP12 was verified to be elevated in all models of paclitaxel-resistant HGSOC. Furthermore, elevated AKAP12 transcript expression was associated with worse progression-free and overall survival. Associations with outcome were observed in three independent cohorts and remained significant after adjusted multivariate modeling. We further provide evidence to support that differential gene methylation status is associated with elevated expression of AKAP12 in taxol-resistant ovarian cancer cells and ovarian cancer patient subsets. Elevated expression and shedding/secretion of AKAP12 is characteristic of paclitaxel-resistant HGSOC cells, and elevated AKAP12 transcript expression is a poor prognostic and predictive marker for progression-free and overall survival in SOC patients.

Abstract 842: Proteomics-based comparative analysis of mitochondria isolated from human cisplatin-sensitive and resistant ovarian carcinoma cell lines

Epithelial ovarian cancer (EOC) is the leading cause of death among women with gynecologic malignancies and the fifth leading cause of cancer death in women in the United States. Cisplatin is one of the most effective and active agents in the treatment of advanced EOC, however, development of chemoresistance is a major obstacle. A recurrent theme is emerging that mitochondria and mitochondrial dysfunction are centerpiece in contributing to an aggressive carcinogenic phenotype. Similarly, we hypothesized that changes in the mitochondrial proteome are required to support development of cisplatin resistance in human EOC. We applied an organellar proteomics approach to elucidate quantitative alterations in protein abundance in mitochondria enriched from an isogenic pair of cisplatin sensitive (A2780) and resistant (A2780-CP20) human EOC cells. Protein isolates from mitochondria were analyzed by high resolution liquid chromatography-tandem mass spectrometry (LC-MS/MS) and quantitative determination of changes in the relative abundance level of identified proteins was accomplished by spectral counting. This analysis confirmed that there are substantial alterations in the mitochondrial proteomic composition in cisplatin resistant EOC cells. Differential proteins selected for western blot validation include hypoxia up-regulated protein 1 (HYOU1), activated leukocyte cell adhesion molecule (ALCAM), isoform 1 of dynamin-like 120kDa protein (OPA1), A kinase anchoring protein 12 (AKAP12), and ferredoxin reductase (FDXR). Our proteomic analysis demonstrated overexpression of HYOU1, ALCAM, AKAP12, and OPA1 in the resistant EOC cells. This panel of protein candidates have been implicated in other tumor types and are involved in relevant cell mechanisms to include evasion of apoptosis, increased tumor invasiveness, tumor hypoxia, cellular respiration, and mitochondrial fragmentation.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 842. doi:1538-7445.AM2012-842

Abstract 799: Comprehensive proteomic analysis of cisplatin resistance in ovarian cancer

We have conducted a comprehensive proteomic analysis of an ovarian cancer cell line (A2780) and its cisplatin resistant daughter line (A2780-CP20) to identify key regulators and signaling pathways that confer cisplatin resistance. Quantitative proteomic profiling was conducted of defined biological compartments including that from global lysates, the nuclear fraction, the phosphoproteome and the secretome. Total cell lysates and nuclear fractions were harvested from A2780 and A2780-CP20 cultured in the presence or absence of cisplatin (3 µM) for 72 h. The secretome fractions were collected from A2780 and A2780-CP20 cultured in serum-free, phenol red-free media in the presence or absence of cisplatin (3 µM) for 24 h. Phosphopeptides were enriched from peptide digests of total lysates of A2780 and A2780-CP20 cultured in the presence or absence of cisplatin (3 µM) for 15 min using TiO2. Sample digests were analyzed by high-resolution LC-MS/MS that resulted in the identification of 2703, 2286, 1815, and 1362 proteins from the global, nuclear, secretome, and phosphoproteome fractions, respectively. Spectral counting was utilized to quantify the relative abundance of proteins identified. Nuclear proteins identified at elevated abundance in the cisplatin resistant A2780-CP20 including an array of chromatin remodeling proteins, such as SATB1, SATB2, coilin, AT-rich interactive domain-containing protein 3A (ARID3A), ARID3B, and ARID3Cwere validated by Western blot and qPCR in A2780/A2780-CP20 as well as other gynecologic cancer cell lines. Modulation of the abundance level of these candidates and the phenotypic impact on cisplatin resistance utilizing shRNA knockdown will be described. Differential proteins selected from the secretome analysis for western blot validation include Fras1-related extracellular matrix protein 2 (FREM2), stanniocalcin-1 (STC1), angio-associated migratory cell protein (AAMP), plasminogen activator inhibitor 1 (SERPINE1), high mobility group protein B1 (HMGB1), and chloride intracellular channel protein 2 (CLIC4), stromelysin-2 (MMP10), interstitial collagenase (MMP2), and CD166 antigen. Validated candidate protein targets from the secretome analysis will be tested by IHC on formalin fixed tissue sections as well as ELISA and multiple reaction monitoring (MRM) by mass spectrometry in serum from recurrent and non-recurrent ovarian cancer patients.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 799. doi:1538-7445.AM2012-799

Restraint stress and stress hormones significantly impact T lymphocyte migration and function through specific alterations of the actin cytoskeleton

Stress triggers complex response mechanisms designed to recognize and adapt to perturbations in homeostasis. The immune system is highly responsive to stress, although the complete mechanisms linking stress and immune mediators including T lymphocytes, are not fully understood. Stress exerts its effects on immune effectors through two primary pathways: the sympathetic-adrenal-medullary pathway, and the hypothalamic-pituitary-adrenal pathway which modulate adaptive immunity and lymphocyte migration. In this report we show that stress via release of stress hormones induces early T cell activation and greatly impacts the cytoskeleton by modulating numerous actin-regulating proteins. In particular, proteomic profiling revealed significant decreases in numerous key actin-binding proteins including moesin. Although confocal microscopy showed that moesin and actin were uniformly distributed on the surface of resting T cells, a remarkable polarization and redistribution of moesin and actin was observed following treatment with stress hormones with moesin localizing at the distal pole complex. In addition, the alteration in moesin localization and eventual decrease in expression were accompanied by a loss of CD43; a receptor involved in negatively regulating T cell activation. In conclusion, we have defined a novel molecular mechanism whereby stress hormones negatively impact T cell activation and migration through regulation of key cytoskeletal and plasma membrane factors.

Assessment of buffer systems for harvesting proteins from tissue interstitial fluid for proteomic analysis

Tissue interstitial fluid (TIF) bathes cells in tissues, and it is hypothesized that TIF proximal to a developing tumor may contain an enriched population of tumor-specific shed and secreted proteins relative to peripheral blood. Extraction of TIF proteins is typically accomplished through passive incubation of surgically resected tissues in phosphate buffered saline (PBS); however, its influence on cellular activity and viability has not been fully explored. The present investigation sought to characterize whether different buffer systems influence the recovered TIF proteome. Five TIF buffer systems were investigated including PBS, Dulbecco's modified Eagle medium (DMEM), and three organ transplantation preservative solutions: Celsior solution S (CS), histidine-tryptophan-ketoglutarate (HTK), and University of Wisconsin (UW). Kidney tumor, adjacent normal kidney, and ovarian tumor tissues were incubated in each of the buffer systems, and the harvested TIF proteins were analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Although the present results indicate that no significant differences exist in the recovered proteins from these two neoplasms between the five solution groups, additional sample preparative steps are required prior to LC-MS/MS for TIF proteins harvested from DMEM, UW, CS, and HTK. These data support that PBS is a suitable and convenient solution for harvesting TIF proteins for MS-based proteomics.

Advances in proximal fluid proteomics for disease biomarker discovery

Although serum/plasma has been the preferred source for identification of disease biomarkers, these efforts have been met with little success, in large part due the relatively small number of highly abundant proteins that render the reliable detection of low abundant disease-related proteins challenging due to the expansive dynamic range of concentration of proteins in this sample. Proximal fluid, the fluid derived from the extracellular milieu of tissues, contains a large repertoire of shed and secreted proteins that are likely to be present at higher concentrations relative to plasma/serum. It is hypothesized that many, if not all, proximal fluid proteins exchange with peripheral circulation, which has provided significant motivation for utilizing proximal fluids as a primary sample source for protein biomarker discovery. The present review highlights recent advances in proximal fluid proteomics, including the various protocols utilized to harvest proximal fluids along with detailing the results from mass spectrometry- and antibody-based analyses.