Abstract 3766: Utilizing whole-transcriptome digital spatial profiling for glioblastoma clinical biomarker discovery

Intratumoral heterogeneity promotes tumor progression, treatment resistance, and recurrence in glioblastoma (GBM), and the cellular tumor core and infiltrating edge represent transcriptionally distinct cellular ecosystems. In this study, we characterize and validate the molecular differences across the core-edge spatial axis in GBM by applying whole-transcriptome GeoMx Digital Spatial Profiling (DSP) to 5 IDH-wildtype tumors. Three 3-micron tissue cores were extracted from each tumor and arranged into a tissue microarray (TMA). Core and edge regions of interest (ROIs) were chosen through coregistration of H&E and immunofluorescent TMA scans. Following DSP, data for 44 ROIs (30 core and 14 edge) and 8530 RNA targets were bioinformatically analyzed. Single-sample gene set enrichment analysis aligned the DSP core and edge regional signatures to those of publicly annotated datasets, substantiating our profiling methodology. Unsupervised analyses revealed ROI clusters unique to each tumor specimen for both pre- and post-normalized data. To mitigate a potential patient-level effect in subsequent differential gene expression and pathway analyses, we leveraged a linear mixed-effect model using core versus edge as a fixed effect and tumor specimen as a random effect. Our analyses showed upregulation of epigenomic markers in core ROIs and neuronal markers in edge ROIs. Further, core and edge signatures were incongruent with the canonical GBM subtypes (Mesenchymal, Classical, Proneural) and malignant cell states (OPC-like, NPC-like, AC-like, MES-like). Cellular deconvolution analysis highlighted differential enrichment of neoplastic cells in core ROIs and neuroglia (astrocytes and oligodendrocytes) in edge ROIs. In summary, we leverage spatial profiling to further validate the geographic molecular heterogeneity of GBM, which is characterized by divergent neuronal programs within the infiltrating edge.

Cutaneous squamous cell carcinoma (cSCC) and basal cell carcinoma (BCC) are the most prevalent types of nonmelanoma skin cancer (NMSC) and exhibit significant inter- and intra-tumor heterogeneity. cSCC has a higher metastatic potential than BCC, accompanied by a considerable mortality rate. However, the detailed mechanisms of tumor evolution in cSCC have not yet been described. We performed single-cell RNA sequencing (scRNA-seq) and T cell receptor (TCR) clonal analysis of skin biopsies from five BCCs, three squamous cell carcinomas insitu (SCCIS), and two invasive squamous cell carcinomas (SCC). Independent SCC specimens were used for spatial transcriptomic (ST) analysis using GeoMx Digital Spatial Profiler (DSP). Using scRNA-seq, we analyzed a total of 117,663 cells. We distinguished cancer cells using copy number variation and identified SCC-specific genes that potentially contribute to tumor progression. Analysis of tumor clones revealed SCC-specific COL6A1+/ITGA5+ carcinoma cells which produce CXCL16. We also annotated CXCR6+ regulatory T cells (Tregs) which potentially move toward the tumor site by CXCL16, shaping the immunosuppressive TME. ST analysis supported these clones were located at the invasion site of SCC. We suggest COL6A1 and ITGA5 promote the invasive and metastatic property of SCC. We also uncovered how SCC recruits Tregs via the CXCL16/CXCR6 axis to create a TME favorable for its survival. These molecules can be used as potential therapeutic targets for treatment of SCC.

Intrinsic intra-tumor heterogeneity (ITH) has been linked to worse patient outcomes. The development of spatial profiling technology has enabled the deciphering of ITH with multiple analysis readouts. Advanced ovarian clear cell carcinoma (OCCC), known to harbor ITH, is chemoresistant, poor prognostic, and possesses distinct molecular and histological characteristics. However, detailed spatial information of the nature of ITH within OCCC remains unclear. Here, we utilized the NanoString Digital spatial profiling (DSP) GeoMx platform to perform multiplex protein expression analysis on tumor samples of primary and colonic metastatic sites from one advanced OCCC patient. The spatial resolution revealed the existence of an epithelial-mesenchymal (EM) gradient within the metastatic tumor but not the primary tumor, and similar EM gradient was not observed within the primary tumor. The EM gradient exhibited a distinct pattern from the periphery to the core of the metastatic tumor. Compared to tumor cells at the primary site, there was an intermediate zone in between the tumor periphery and the tumor core in the colonic metastasis with differential expression patterns of pan-cytokeratin (PanCK), fibronectin (FN), smooth muscle actin (SMA), neural cell adhesion molecule (NCAM), integrin alpha X (ITGAX), and Ki-67. Our study provides the first spatially resolved in situ evidence of intermediate or hybrid EM states within the tumor samples of similar morphology. This not only demonstrates the promising applications of spatial profiling in precision medicine but also provides an unprecedented view of the EM gradient during the progression of cancer such as OCCC. Citation Format: Duncan Y.T. Wang, Tuan Zea Tan, Ya-Ting Tai, Jieru Ye, Wei-Chou Lin, Lin-Hung Wei, Ruby Y. J. Huang. Digital spatial profiling of metastatic clear cell carcinoma reveals intra-tumor heterogeneity in epithelial-mesenchymal gradient [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5173.

Heterogeneity is a hallmark of tumors and has a crucial role in the outcome of the malignancy, because it not only confounds diagnosis, but also challenges the design of effective therapies. There are two types of heterogeneity: inter-tumor and intra-tumor heterogeneity. While inter-tumor heterogeneity has been studied widely, intra-tumor heterogeneity has been neglected even though numerous studies support this aspect of tumor pathobiology. The main reason has been the technical difficulties, but with new advances in single-cell technology, intra-tumor heterogeneity is becoming a key area in the study of cancer. Several models try to explain the origin and maintenance of intra-tumor heterogeneity, however, one prominent model compares cancer with a tree where the ubiquitous mutations compose the trunk and mutations present in subpopulations of cells are represented by the branches. In this review we will focus on the intra-tumor heterogeneity of glioblastoma multiforme (GBM), the most common brain tumor in adults that is characterized by a marked heterogeneity at the cellular and molecular levels. Better understanding of this heterogeneity will be essential to design effective therapies against this devastating disease to avoid tumor escape.

Profiling lamina specific pyramidal neurons using postmortem human formalin fixed paraffin embedded frontal cortex tissue in combination with digital spatial profiling.

<h3>Background</h3> Most patients with advanced non-small-cell lung cancer (NSCLC) fail to derive significant benefit from programmed cell death protein-1 (PD-1) axis blockade, and new biomarkers of response are needed. In this study, we aimed to discover and validate spatially resolved protein markers associated with sensitivity to PD-1 axis inhibition in NSCLC. <h3>Methods</h3> We initially assessed a discovery cohort of 56 patients with NSCLC treated with PD-1 axis inhibitors at Yale Cancer Center. Using the GeoMx Digital Spatial Profiling (DSP) system, 71 proteins were measured in spatial context on each spot in a tissue microarray. We used the AQUA method of quantitative immunofluorescence (QIF) to orthogonally validate candidate biomarkers. For external independent validation, we assessed whole tissue sections derived from 128 patients with NSCLC treated with single-agent PD-1 axis inhibitors at the 12 de Octubre Hospital (Madrid) using DSP. We further analyzed two immunotherapy untreated cohorts to address prognostic significance (n=252 from Yale Cancer Center; n=124 from University Clinic of Navarra) using QIF and DSP, respectively. <h3>Results</h3> Using continuous log-scaled data, we identified CD44 expression in the tumor compartment (pan-cytokeratin (CK)+) as a novel predictor of prolonged progression-free survival (PFS) (multivariate HR=0.68, p=0.043) in the discovery set. We validated by QIF that tumor CD44 levels assessed as continuous QIF scores were associated with longer PFS (multivariate HR=0.31, p=0.022) and overall survival (multivariate HR=0.29, p=0.038). Using DSP in an independent immunotherapy treated cohort, we validated that CD44 levels in the tumor compartment, but not in the immune compartment (panCK–/CD45+), were associated with clinical benefit (OR=1.22, p=0.018) and extended PFS under PD-1 axis inhibition using the highest tertile cutpoint (multivariate HR=0.62, p=0.03). The effect of tumor cell CD44 in predicting PFS remained significant after correcting for programmed death-ligand 1 (PD-L1) Tumor Proportion Score (TPS) in both cohorts. High tumor cell CD44 was not prognostic in the absence of immunotherapy. Using DSP data, intratumoral regions with elevated tumor cell CD44 expression showed prominent (fold change&gt;1.5, adjusted p&lt;0.05) upregulation of PD-L1, TIM-3, ICOS, and CD40 in two independent cohorts. <h3>Conclusions</h3> This work highlights CD44 as a novel indicative biomarker of sensitivity to PD-1 axis blockade that might help to improve immunotherapy strategies for NSCLC.

Introduction Glioblastoma (GBM) is the most common primary adult brain tumour with a median overall survival (OS) of 12–15 months. Molecular characterization of multiple immunooncology targets in GBM may help target novel immunotherapeutic strategies. We used NanoString GeoMx® Digital Spatial Profiling (DSP) to assess multiple immunooncology protein targets in methylated versus unmethylated IDH-wild-type glioblastoma. Methods NanoString GeoMx® DSP technology uses multiple primary antibodies conjugated to indexing DNA oligos with a UV photocleavable linker. Tissue regions of interest (ROIs) are selected with bound fluorescent antibodies; oligos are released via a UV-mediated linker and quantitated. We used DSP multiplex analysis of 31 immunooncology proteins and controls (CD4, CD14, CD68, CD8A, B7-H3, PD-L1, CD19, FOXP3, CD44, STAT3 (phospho Y705), CD45, Pan Cytokeratin, MS4A1/CD20, CD45RO, PD1, CD3, beta-2 microglobulin, VISTA, Bcl2, GZMB, PTEN, beta-catenin, CD56, Ki-67, STAT3, AKT, p-Akt, S6, Histone H3, IgG Rabbit control, and Mouse IgG control) from ROIs in a cohort of 10 IDH-wild-type glioblastomas (5 methylated and 5 unmethylated). An nCounter platform allowed quantitative comparisons of antibodies between ROIs in MGMT methylated and unmethylated tumours. Mean protein expression counts between methylated and unmethylated GBM were compared using technical and biological replicates. Results The analysis showed 10/27 immunooncology target proteins were significantly increased in methylated versus unmethylated IDH-wild-type glioblastoma tumour core (false discovery rate (FDR) <0.1 by Benjamini–Hochberg procedure). Conclusions NanoString GeoMx® DSP was used to analyse multiple immunooncology protein target expression in methylated versus unmethylated IDH-wild-type glioblastoma. In this small study, there was a statistical increase in CD4, CD14, CD68, CD8A, B7-H3, PDL-1, CD19, FOXP3, CD44, and STAT3 protein expression in methylated versus unmethylated GBM tumour core; however, this requires larger cohort validation. Advanced multiplex immunooncological biomarker analysis may be useful in identifying biomarkers for novel immunotherapeutic agents in GBMs.

Glioblastoma (GBM) is one of the most aggressive primary brain tumors in adults. Despite various strategies, including differential single-cell gene expression analyses between GBM and healthy tissues, none of the resulting findings could benefit the development of promising therapies, yet. Various macrophage populations coevolve in the tumor microenvironment of GBMs and not only support tumor immune evasion but also tumor spreading throughout the surrounding tissue. To distillate unique immune cell features of GBMs, we compared their immune cell composition with other central nervous system (CNS) tumor entities, in particular meningiomas (MNGs), non-GBM gliomas, and metastases (MTS). We could identify a macrophage population characterized by the coexpression of CX3CR1 and Siglec10, representing a potential indicator population of GBMs. This GBM-specific macrophage subset might be supported by a glioma-characteristic lipid mediator (LM) milieu of enriched docosahexaenoic acid (DHA) and its lipoxygenase (LOX)-derived metabolites. Moreover, the GBM tumor microenvironment (TME) is enriched in arachidonic acid (AA)-derived cyclooxygenase (COX) products prostaglandins (PG) E2 and F2α in combination with enhanced CD24 expression. By our comparative approach, the data hint toward a pro-tumorigenic Siglec10+CX3CR1+ macrophage population depending on the characteristic tumor microenvironment (TME) of highly malignant GBMs.

BackgroundThe canonical WNT-β-catenin signaling pathway is vital for development and tissue homeostasis but becomes strongly tumorigenic when dysregulated. and alter the transcriptional signature of a cell to promote malignant transformation....

Intra-tumoural heterogeneity is a common driver of failure for novel IDH-wildtype glioblastoma (GBM) treatments, and many preclinical studies rely on cell lines established from the tumour core. Therefore, they fail to characterise the residual infiltrative cells which remain post-surgery and ultimately drive tumour recurrence. We hypothesise that plasma membrane proteins on the surface of these infiltrative cells offer attractive targets for chemotherapeutic agents as they are most visible to circulating drugs. The membrane proteome of patient-derived GBM cell lines isolated from the tumour infiltrative margin (GIN8, GIN28, and GIN31), which are a proxy for residual disease post-surgery, were characterised. The identified membrane proteome was compared to that of the healthy astrocyte membrane proteome. Any proteins ubiquitously expressed on GBM invasive margin cells and not detectable on healthy astrocytes have the potential to reduce off target toxicity. We combined targeted plasma membrane protein with total protein using proteomics techniques to uncover potential therapeutic targets most amenable for drug repurposing. Eighteen membrane proteins including EDIL3, DYSF, ROBO1, SERPINE2, LOXL1 and CD70 were upregulated in the GBM cell lines when compared to astrocytes, indicating putative roles associated with GBM progression. Furthermore, nine of these ubiquitously upregulated proteins were undetectable in astrocyte cells. Analysis of correlation between protein abundance ratio and survival time in GBM clinical samples retrieved from the CPTAC study on cBioPortal showed that ROBO1, LOXL1 and DYSF conferred a negative correlation with survival, which suggests that these membrane proteins could exhibit a functional role in GBM survival and progression. These findings suggest that targeting membrane proteins offers promise for developing effective GBM therapies predicated on the most clinically accurate intra-tumour region. Future work utilises molecular docking techniques to identify suitable repurposed drugs against the membrane proteins observed.

INTRODUCTION: Intratumour heterogeneity (ITH) may account for therapy failure and tumour recurrence in adult glioblastoma multiforme (GBM). Low availability of fresh frozen tissue has hindered characterisation of ITH in paediatric GBM. Here, we propose the application of desorption electrospray ionisation mass spectrometry (DESI-MS) to assess spatial lipidomic profiles of formalin-fixed paraffin-embedded (FFPE) tissue of paediatric GBM. METHOD: Optimisation of DESI-MS has been conducted on FFPE tissue of normal adult brain prior to transfer of the technology to paediatric GBM tissue. Immunohistochemical analyses for proteins associated with lipid biosynthesis delineated distinct intratumour regions. Spatial characterisation of lipids synthesised by these metabolic proteins will permit optimisation of DESI-MS, prior to unbiased lipidomic analyses. RESULTS: Paediatric GBM cell lines grown under lipoprotein-deficient conditions demonstrated greater resilience compared to an adult cell line, indicating distinct lipid metabolism that will be revealed through lipidomic analyses. Proteins associated with lipid biosynthesis demonstrated evidence of ITH in paediatric GBM, with 69% of patients displaying heterogeneous expression of fatty acid synthase that correlated with areas of high cellularity. CONCLUSION: DESI-MS has the potential to characterise early indications of paediatric GBM ITH through characterisation of lipid markers associated with distinct intratumour regions. Proteins associated with the synthesis of identified lipid markers may represent potential therapeutic targets. We anticipate the establishment of an optimised DESI-MS method that will facilitate lipidomic studies of other paediatric brain tumours.

BackgroundEpigenome-wide association studies (EWAS) and differential gene expression analyses are generally performed on tissue samples, which consist of multiple cell types. Cell-type-specific effects of a trait, such as disease, on the omics expression are of interest but difficult or costly to measure experimentally. By measuring omics data for the bulk tissue, cell type composition of a sample can be inferred statistically. Subsequently, cell-type-specific effects are estimated by linear regression that includes terms representing the interaction between the cell type proportions and the trait. This approach involves two issues, scaling and multicollinearity.ResultsFirst, although cell composition is analyzed in linear scale, differential methylation/expression is analyzed suitably in the logit/log scale. To simultaneously analyze two scales, we applied nonlinear regression. Second, we show that the interaction terms are highly collinear, which is obstructive to ordinary regression. To cope with the multicollinearity, we applied ridge regularization. In simulated data, nonlinear ridge regression attained well-balanced sensitivity, specificity and precision. Marginal model attained the lowest precision and highest sensitivity and was the only algorithm to detect weak signal in real data.ConclusionNonlinear ridge regression performed cell-type-specific association test on bulk omics data with well-balanced performance. The omicwas package for R implements nonlinear ridge regression for cell-type-specific EWAS, differential gene expression and QTL analyses. The software is freely available from https://github.com/fumi-github/omicwas

Outcomes of Isocitrate Dehydrogenase Wild Type Glioblastoma after Re-irradiation.

BackgroundBladder cancer is the fourth most common cancer in American men with chances of 1 in 27 developing this form of cancer. Despite the progress in treating these patients with...

Glut3 addiction is a druggable vulnerability for a molecularly defined subpopulation of glioblastoma