High-dimensional Analysis Research Articles

Abstract 129: Hypercholesterolemia-induced Lxr Signaling In Smc Contributes To Atherosclerotic Lesion Remodeling And Regulates Vascular And Visceral Smc Function

Vascular smooth muscle cells (VSMCs) play an important role during atherogenesis, regulating vasotone and lesion remodeling. Recent advances in combing VSMC lineage tracing and scRNA-seq have generated high-dimensional analysis of VSMC profiles of atherosclerotic lesion, delineating multiple VSMC-derived transitional states in vascular lesions, which can further adopt plaque stabilizing or plaque destabilizing cell states. Despite the identification of several transcription factors that govern SMC fate, the signaling events that promote SMC transitional state remain poorly understood. Our trajectory analysis of VSMCs gene expression reveals the activation of liver X receptor (LXR) signaling, a master regulator of lipid metabolism, during the progression of atherosclerosis, suggesting a relevant role of this transcription factor in controlling VSMC fate decision and lesion remodeling. To dissect the specific contribution of LXR signaling in VSMC during atherogenesis, we generated and characterized mice in which the LXRα and β genes were conditionally ablated in SMCs in the adult vasculature and visceral tissues. Deficiency of LXR in SMCs under hypercholesterolemic condition influenced lesion remodeling by altering the fate of de-differentiated SMC and promoting the accumulation of VSMC-derived transition cells. This phenotypic switching is accompanied with increased features of plaque instability characterized by larger necrotic cores and reduced fibrous cap thickness. Moreover, LXR SMC deficiency impaired vascular function and caused visceral myopathy characterized by bladder maladaptive remodeling and altered gut lipid absorption. Together, we uncovered an essential role of LXR-regulated signaling that contributes to atherosclerotic lesion remodeling, and vascular/ visceral SMC function.

Bispecific T cell engager therapy for refractory rheumatoid arthritis.

Bispecific T cell engagers (BiTEs) kill B cells by engaging T cells. BiTEs are highly effective in acute lymphoblastic leukemia. Here we treated six patients with multidrug-resistant rheumatoid arthritis (RA) with the CD19xCD3 BiTE blinatumomab under compassionate use. Low doses of blinatumomab led to B cell depletion and concomitant decrease of T cells, documenting their engager function. Treatment was safe, with brief increase in body temperature and acute phase proteins during first infusion but no signs of clinically relevant cytokine-release syndrome. Blinatumomab led to a rapid decline in RA clinical disease activity in all patients, improved synovitis in ultrasound and FAPI-PET-CT and reduced autoantibodies. High-dimensional flow cytometry analysis of B cells documented an immune reset with depletion of activated memory B cells, which were replaced by nonclass-switched IgD-positive naïve B cells. Together, these data suggest the feasibility and potential for BiTEs to treat RA. This approach warrants further exploration on other B-cell-mediated autoimmune diseases.

P134 Developing a potential composite pharmacodynamic serum marker of skin severity in systemic sclerosis

Abstract Background/Aims Previously, through multilevel and high dimensional analysis, we derived a composite biomarker for modified Rodnan Skin Score (mRSS) in systemic sclerosis (SSc) patients using a derivation cohort (BIOPSY). This identified 4 blood proteins (COMP, COL4A1, TNC and SPON1) that independently correlated with mRSS and also with skin gene and dermal blister expression at a separate 12-month timepoint. These proteins likely reflect overlapping aspects of SSc skin pathobiology that are not overly influenced by other disease compartments. In the present study, we undertake technical and biological validation of this novel composite marker in well characterised SSc patients. Methods COMP (DCMP0, R&D Systems), COL4A1 (CSB-EL005741HU, Cusabio), TNC (ab213831, Abcam) and SPON1 (CSB-EL022599HU, Cusabio) were examined in serum samples from diffuse cutaneous SSc (dcSSc) cases in two independent cohorts of patients, BIOPSY (n = 33) and MODERNISE (n = 37). The BIOPSY cohort was previously used to derive the composite marker and MODERNISE was a novel validation cohort. Later, early dcSSc patients from both cohorts were combined into a third analysis group (n = 36). Multiple regression analysis derived a formula for a composite biomarker score which predicted mRSS based on serum ELISA results. Results Serum concentration positively correlated with mRSS for COMP and TNC, particularly in early dcSSc patients (COMP, r = 0.10, p = 0.05; TNC, r = 0.13, p = 0.03). No significant correlation was observed for COL4A1 (r = 0.03, p = 0.34). Interpretable data could not be obtained for SPON1 due to technical limitations of the ELISA. Confirming previous results multiple regression analysis derived a formula for the prediction of mRSS using the three remaining serum proteins (r = 0.15, p = 0.009). mRSS = 9.896 + 0.01719(COMP) - 0.006481(COL4A1) - 0.002318(TNC). Patients receiving mycophenolate mofetil showed numerically lower serum concentrations of COMP (p = 0.16), COL4A1 (p = 0.32), TNC (p = 0.15) and a lower composite biomarker score (p = 0.22) with increasing treatment duration in early dcSSc cases (Table 1). Conclusion Our results confirm this novel three-protein composite serum marker predicts skin severity in SSc better than each individual analyte. After further validation, this blood-based approach may help evaluate treatment response to standard immunosuppression. Further work will establish utility of this composite measure as a predictive or prognostic biomarker. Disclosure E. Roblin: None. K.E.N. Clark: None. C. Beesley: None. V.H. Ong: None. C.P. Denton: Consultancies; C.P.D. has received consulting fees from Arxx Therapeutics, Roche, Janssen, GlaxoSmithKline, Bayer, Sanofi, Galapagos, Boehringer Ingelheim, CSL Behring and Acceleron. Honoraria; C.P.D. has received honoraria from Janssen, Boehringer Ingelheim, and Corbus. Grants/research support; C.P.D. has received research grants to the institution from Servier, Horizon, Arxx Therapeutics and GlaxoSmithKline.

An active learning method using deep adversarial autoencoder-based sufficient dimension reduction neural network for high-dimensional reliability analysis

Reliability analysis often requires time-consuming evaluations, especially when dealing with high-dimensional and nonlinear problems. To address this challenge, surrogate model methods are frequently employed. One way to improve the efficiency of surrogate model methods involves selecting informative samples that significantly enhance the accuracy of the surrogate model. This paper introduces a novel approach to facilitate the construction of surrogate models and selection of informative samples in high-dimensional reliability analysis, through an active learning method based on a deep adversarial autoencoder-based sufficient dimension reduction (AAE-SDR) neural network. The AAE-SDR neural network serves as a surrogate model, transforming complex high-dimensional variables into tractable, low-dimensional embeddings relevant to the target. These embeddings are Gaussian-distributed with a distinct latent limit state boundary. A new sampling strategy is proposed to select informative misclassified samples by iteratively identifying candidate samples near the latent limit state boundary and uniformly sampling from the candidate sample dataset based on the latent Gaussian distribution. The effectiveness of the proposed approach is demonstrated through two high-dimensional numerical examples and a cable-stayed bridge case study. Results show that the proposed method simplifies complex high-dimensional reliability problems and provides a relatively accurate estimated failure probability with a limited number of samples.

In vivo single-cell high-dimensional mass cytometry analysis to track the interactions between Klebsiella pneumoniae and myeloid cells.

In vivo single-cell approaches have transformed our understanding of the immune populations in tissues. Mass cytometry (CyTOF), that combines the resolution of mass spectrometry with the ability to conduct multiplexed measurements of cell molecules at the single cell resolution, has enabled to resolve the diversity of immune cell subsets, and their heterogeneous functionality. Here we assess the feasibility of taking CyTOF one step further to immuno profile cells while tracking their interactions with bacteria, a method we term Bac-CyTOF. We focus on the pathogen Klebsiella pneumoniae interrogating the pneumonia mouse model. Using Bac-CyTOF, we unveil the atlas of immune cells of mice infected with a K. pneumoniae hypervirulent strain. The atlas is characterized by a decrease in the populations of alveolar and monocyte-derived macrophages. Conversely, neutrophils, and inflammatory monocytes are characterized by an increase in the subpopulations expressing markers of less active cells such as the immune checkpoint PD-L1. These are the cells infected. We show that the type VI secretion system (T6SS) contributes to shape the lung immune landscape. The T6SS governs the interaction with monocytes/macrophages by shifting Klebsiella from alveolar macrophages to interstitial macrophages and limiting the infection of inflammatory monocytes. The lack of T6SS results in an increase of cells expressing markers of active cells, and a decrease in the subpopulations expressing PD-L1. By probing Klebsiella, and Acinetobacter baumannii strains with limited ability to survive in vivo, we uncover that a heightened recruitment of neutrophils, and relative high levels of alveolar macrophages and eosinophils and the recruitment of a characteristic subpopulation of neutrophils are features of mice clearing infections. We leverage Bac-CyTOF-generated knowledge platform to investigate the role of the DNA sensor STING in Klebsiella infections. sting-/- infected mice present features consistent with clearing the infection including the reduced levels of PD-L1. STING absence facilitates Klebsiella clearance.

TSPAN8+ myofibroblastic cancer-associated fibroblasts promote chemoresistance in patients with breast cancer.

Cancer-associated fibroblasts (CAFs) are abundant stromal cells in the tumor microenvironment that promote cancer progression and relapse. However, the heterogeneity and regulatory roles of CAFs underlying chemoresistance remain largely unclear. Here, we performed a single-cell analysis using high-dimensional flow cytometry analysis and identified a distinct senescence-like tetraspanin-8 (TSPAN8)+ myofibroblastic CAF (myCAF) subset, which is correlated with therapeutic resistance and poor survival in multiple cohorts of patients with breast cancer (BC). TSPAN8+ myCAFs potentiate the stemness of the surrounding BC cells through secretion of senescence-associated secretory phenotype (SASP)-related factors IL-6 and IL-8 to counteract chemotherapy. NAD-dependent protein deacetylase sirtuin 6 (SIRT6) reduction was responsible for the senescence-like phenotype and tumor-promoting role of TSPAN8+ myCAFs. Mechanistically, TSPAN8 promoted the phosphorylation of ubiquitin E3 ligase retinoblastoma binding protein 6 (RBBP6) at Ser772 by recruiting MAPK11, thereby inducing SIRT6 protein destruction. In turn, SIRT6 down-regulation up-regulated GLS1 and PYCR1, which caused TSPAN8+ myCAFs to secrete aspartate and proline, and therefore proved a nutritional niche to support BC outgrowth. By demonstrating that TSPAN8+SIRT6low myCAFs were tightly associated with unfavorable disease outcomes, we proposed that the combined regimen of anti-TSPAN8 antibody and SIRT6 activator MDL-800 is a promising approach to overcome chemoresistance. These findings highlight that senescence contributes to CAF heterogeneity and chemoresistance and suggest that targeting TSPAN8+ myCAFs is a promising approach to circumvent chemoresistance.

Joint analysis of multivariate longitudinal, imaging, and time-to-event data

Abstract Alzheimer’s (AD) is a progressive neurodegenerative disease frequently associated with memory deficits and cognitive decline. Despite its irreversible once onset, some discoveries revealed the existence of a certain percentage of people who are non-susceptible to AD. This study proposes a joint analysis of multivariate longitudinal data, survival data with a non-susceptible fraction, and ultrahigh-dimensional imaging data. The proposed model comprises three major components. The first component is a mixture proportional hazards cure model with images to examine the potential predictors of the non-susceptible probability and hazards of interest. The second component is a dynamic factor analysis model with images to characterize group-specific latent factors through multiple observed variables. The last component is a semiparametric trajectory model to reveal the change patterns of the dynamic latent factors in the ‘non-susceptible’ and ‘susceptible’ groups. A two-stage approach is developed for statistical inference. The first stage manages the imaging data through high-dimensional functional principal component analysis. The second stage develops a Bayesian approach coupled with penalized splines, data augmentation, and Markov chain Monte Carlo techniques to perform estimation. The application to the Alzheimer’s Disease Neuroimaging Initiative dataset sheds new insight into the pathology of AD.

High-dimensional covariate-augmented overdispersed poisson factor model.

The current Poisson factor models often assume that the factors are unknown, which overlooks the explanatory potential of certain observable covariates. This study focuses on high dimensional settings, where the number of the count response variables and/or covariates can diverge as the sample size increases. A covariate-augmented overdispersed Poisson factor model is proposed to jointly perform a high-dimensional Poisson factor analysis and estimate a large coefficient matrix for overdispersed count data. A group of identifiability conditions is provided to theoretically guarantee computational identifiability. We incorporate the interdependence of both response variables and covariates by imposing a low-rank constraint on the large coefficient matrix. To address the computation challenges posed by nonlinearity, two high-dimensional latent matrices, and the low-rank constraint, we propose a novel variational estimation scheme that combines Laplace and Taylor approximations. We also develop a criterion based on a singular value ratio to determine the number of factors and the rank of the coefficient matrix. Comprehensive simulation studies demonstrate that the proposed method outperforms the state-of-the-art methods in estimation accuracy and computational efficiency. The practical merit of our method is demonstrated by an application to the CITE-seq dataset. A flexible implementation of our proposed method is available in the R package COAP.

High-dimensional Statistical Analysis and Its Application to an ALMA Map of NGC 253

In astronomy, if we denote the dimension of data as d and the number of samples as n, we often find a case with n ≪ d. Traditionally, such a situation is regarded as ill-posed, and there was no choice but to discard most of the information in data dimensions to let d < n. The data with n ≪ d is referred to as the high-dimensional low sample size (HDLSS). To deal with HDLSS problems, a method called high-dimensional statistics has rapidly developed in the last decade. In this work, we first introduce high-dimensional statistical analysis to the astronomical community. We apply two representative methods in the high-dimensional statistical analysis methods, noise-reduction principal component analysis (NRPCA) and automatic sparse principal component analysis (A-SPCA), to a spectroscopic map of a nearby archetype starburst galaxy NGC 253 taken by the Atacama Large Millimeter/submillimeter Array (ALMA). The ALMA map is an example of a typical HDLSS data set. First, we analyzed the original data, including the Doppler shift due to the systemic rotation. High-dimensional PCA can precisely describe the spatial structure of the rotation. We then applied to the Doppler-shift corrected data to analyze more subtle spectral features. NRPCA and R-SPCA were able to quantify the very complicated characteristics of the ALMA spectra. Particularly, we were able to extract information on the global outflow from the center of NGC 253. This method can also be applied not only to spectroscopic survey data, but also to any type of data with a small sample size and large dimension.

Intranodal Injection of Immune Activator Demonstrates Antitumor Efficacy in an Adjuvant Approach.

The tumor-draining lymph nodes (tdLN) are the initial site of metastases and are the prime site for generating robust antitumor responses. In this study, we explored the efficacy of a universal immune activator (ImmAct) targeted to the tdLN. This approach can be viewed as an attempt to turn a cold, unresponsive tdLN into a hot, responsive site. The adjuvant antitumor efficacy of our novel intranodal injection was evaluated in an aggressive metastatic mammary carcinoma murine model. The cancer cells were inoculated subcutaneously in the lower quadrant of the mouse to provoke the tdLN (inguinal lymph node). The study encompasses a range of methodologies, including in vivo and in vitro assays and high-dimensional flow cytometry analysis. Our findings demonstrated that intranodal administration of ImmAct following the dissection of the primary tumor led to improved tumor-free survival and minimized weight loss. ImmAct led to both local and systemic alterations in the cellular and humoral immunity. Additionally, after ImmAct treatment, non-responders showed a higher rate of exhausted CD8+ T cells compared to responders. Indeed, our innovative approach surpassed the gold standard surgery of sentinel lymph node excision. Overall, intranodal administration of ImmAct yielded a robust antitumor immune response, offering protection against micrometastases and relapse.

High-Dimensional Analysis for Generalized Nonlinear Regression: From Asymptotics to Algorithm

Overparameterization often leads to benign overfitting, where deep neural networks can be trained to overfit the training data but still generalize well on unseen data. However, it lacks a generalized asymptotic framework for nonlinear regressions and connections to conventional complexity notions. In this paper, we propose a generalized high-dimensional analysis for nonlinear regression models, including various nonlinear feature mapping methods and subsampling. Specifically, we first provide an implicit regularization parameter and asymptotic equivalents related to a classical complexity notion, i.e., effective dimension. We then present a high-dimensional analysis for nonlinear ridge regression and extend it to ridgeless regression in the under-parameterized and over-parameterized regimes, respectively. We find that the limiting risks decrease with the effective dimension. Motivated by these theoretical findings, we propose an algorithm, namely RFRed, to improve generalization ability. Finally, we validate our theoretical findings and the proposed algorithm through several experiments.

Abstract 1073: AI-assisted high-dimensional cytological single cell analysis distinguishes normal and malignant urine samples using cell morphology images with high accuracy at the single cell level

Abstract Background: Urothelial cancer (UCC) is the 4th most common malignancy in men. The current clinical problem in UCC management is the high interobserver variability of tissue UCC grading and the low sensitivity of urine cytology for early detection of recurrences and progression. We hypothesize that assessing urine samples and UCC tissues with detailed quantitative image analysis based on artificial intelligence has strong potential in both urine-based early detection of tumor recurrence and progression as well as tissue-based pathological grading. Methods: Here, we utilized the Deepcell® Human Foundation Model (HFM), a self-supervised deep learning model, to assess a cohort of urine samples from 54 normal healthy cases, 19 atypical cases, and 4 UCC cases. High resolution images of label-free single cells were captured and analyzed using the HFM to identify distinct morphological characteristics of cells from different urine origins. The urine samples were loaded onto a single-use microfluidic cartridge as part of the Deepcell® REM-I platform and brightfield images of single cells were captured in real-time. For data analysis, images of all cells run through the REM-I instrument are automatically sent to the Deepcell® Axon Suite, which provides an analysis platform to visualize cell images from specific runs, generate UMAPs of the embedding space, and run additional deep neural network models on samples in silico. Results: High-dimensional analysis of 115 deep learning and morphometric features reveals 12 morphotypes among the pooled urine samples, as represented by morphology UMAP plots. Morphology analysis reveals distinct morphotypes in normal compared to pathological urine samples. Confusion matrix showed deep learning model performance in classifying atypical versus UCC with 89.1% (Atypical (n = 8642), UCC (n = 12637)). This indicates high model performance accuracy for AI prediction of atypical versus UCC samples at the single cell level. Conclusions: The Deepcell® REM-I platform performs well with urine samples and identified numerous cell clusters among normal and pathological urines. AI-assisted high-dimensional morphology analysis distinguishes atypical urine samples from UCC cases with 89.1% accuracy at the single cell level. Future applications of this study include early detection or monitoring of UCC cases using non-invasive patient samples such as urine. Citation Format: Thierry P. van den Bosch, Geert J.L.H. van Leenders, Siamala Sinnadurai, Zhouyang Lian, Ryan Carelli, Matthew Nakaki, Vivian Lu, Manisha Ray, Janifer Cruz, Kelly R. Miller, Mahyar Salek, Maddison Masaeli, Peter J. van der Spek. AI-assisted high-dimensional cytological single cell analysis distinguishes normal and malignant urine samples using cell morphology images with high accuracy at the single cell level [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 1073.

Abstract 1543: 40-color spectral flow cytometry for comprehensive immunophenotyping of murine cancers

Abstract Introduction: High-dimensional immunoprofiling is essential for studying host response to cancer immunotherapy. Synergistic agents often induce systemic changes in leukocyte abundance/function due to targeting, dosage, or delivery, necessitating multi-tissue analyses to decipher immune response mechanisms. We set out to dissect leukocyte composition in lymphoid tissues and tumors by using spectral cytometry, a multiplexed, high-throughput technology. Methods: We engineered a 40-color spectral flow cytometry panel to immunophenotype all major leukocytes and their functional subsets. Naïve and tumor-burdened C57BL/6 mice treated with combinations of agonist CD40 (aCD40) and checkpoint inhibitors were sacrificed, and tissues were processed. Stained cells were acquired on a 5-laser CyTEK Aurora. Unmixing accuracy and resolution was validated via 780 NxN plots, and high-dimensional analysis was performed. Results: Spectral cytometry affords resolution of 43+ parameters compared to 26 or less in conventional systems. We first designed a 25-marker backbone panel to identify all immune lineages. Additional channels were assigned to functional markers, allowing for profiling of migration, maturation, and activation in tumor-infiltrating leukocytes (TILs). In naïve secondary lymphoid tissues (SLTs), 35 distinct CD45+ populations were resolved. Germinal center B cells (17% vs. ~0.3% of CD45+ in other SLTs) and T follicular cells (31% vs. ~1-5% of CD4+ T cells in other SLTs) were mostly found in the Peyer’s patches, representing a hypothesized model tissue for studying tertiary immune structures in tumors. Peripheral blood contained the highest frequency of NK cells (4%), monocytes (6%), basophils (0.3%), eosinophils (4%), and neutrophils (7%) as a percentage of CD45+ cells. The spleen held the largest reservoir of CD4/8 central/effector memory cells (5% of CD45+) while ~80% of all T cells in the lymph nodes were naïve. In primary lymphoid tissues (PLTs), 12 distinct thymocyte and 7 B cell populations were resolved. Wanderlust trajectory analysis revealed CD24 and CD127 as key developmental regulators in lymphocyte differentiation. Autofluorescence extraction was essential for myeloid phenotyping. With this approach, we found monocytes comprised ~60% of TILs following aCD40 treatment (~8% in non-treatment and checkpoint only controls) and had altered macrophage differentiation fates. aCD40-treated tumors also contained higher frequencies of activated CD40/80/86+ DCs (~80/86/60% vs. ~60/51/50% of DCs in NTCs). Conclusion: This work describes a novel tool that efficiently resolves leukocytes on the axes of time, tissue, and treatment. Applications range from serial monitoring of circulating blood, endpoint analysis, validation of multiplexed-IHC, and cell sorting. PLT and SLT data has been made publicly available to serve as a training tool and reference proteomic atlas in future work. Citation Format: Aris John Kare, Lisa Nichols, Ricardo Zermeno, Marina N. Raie, Spencer K. Tumbale, Katherine W. Ferrara. 40-color spectral flow cytometry for comprehensive immunophenotyping of murine cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 1543.

Abstract 3429: Using metabolomics to identify biomarkers mediating the associations of adiposity in childhood and early adulthood with mammographic breast density in premenopausal women

Abstract Introduction: Mammographic breast density (MBD), a strong predictor of breast cancer risk, is highly influenced by body mass index (BMI) in childhood and early adulthood, but the mechanisms elucidating this relationship are undetermined. To our knowledge, no study has explored the mediating role of biomarkers on the relationship between early-life adiposity and MBD. The goal of this study is to discover metabolites that mediate the relationship between BMI at ages 10 and 18 with MBD in premenopausal women. Methods: This study includes premenopausal women who had their screening mammogram at Washington University in St. Louis, MO, and provided a fasting blood sample. Metabolon performed untargeted metabolomic profiling, detecting 1,074 metabolites. Metabolites missing &amp;gt;300 observations were excluded, leaving 828 metabolites; the other missing values were imputed using the nearest neighbor method. To mitigate batch effect, we normalized the metabolite peak area data using ComBat. Volumetric percent density (VPD) was calculated in 700 women using Volpara software. BMI at age 10 was estimated using the Stunkard pictogram, and BMI at age 18 was calculated from self-reported weight at 18 and height at study initiation. To assess the mediating role of the 828 metabolites, we performed high dimensional mediation analysis using the hima R package adjusting for potential confounders. Missing values in the covariates and BMI measures included in the mediation analysis were imputed using multivariate imputation by chain equations. Associations were considered significant if FDR p-value &amp;lt;0.1. Results: Four metabolites (glutamate, beta-cryptoxanthin, phytanate, and cortolone glucuronide (1)) mediated the relationship between BMI at age 10 and VPD; and 2 metabolites (glutamate, beta-cryptoxanthin) mediated the relationship between BMI at age 18 and VPD. Glutamate and beta-cryptoxanthin significantly mediated the relationship between both BMI measures and VPD, but glutamate was the strongest mediator across both time points. Glutamate mediated 6.7% (FDR p-value=0.06) and 9.3% (FDR p-value=0.008) of the total effect between BMI at age 10 and 18, respectively, on VPD. Beta-cryptoxanthin mediated 4.1% (FDR p-value=0.06) and 6.3% (FDR p-value=0.04), of the total effect between BMI at age 10 and 18, respectively, on VPD. Additionally, phytanate mediated 2.6% (FDR p-value=0.06) and cortolone glucuronide (1) mediated 2.0% (FDR p-value=0.07) of the total effect between BMI at age 10 and VPD. Conclusion: Amino acid, lipid, cofactor/vitamin, and xenobiotic metabolites mediated the associations of BMI in early-life and VPD in premenopausal women. This innovative study offers insight into the biological mechanisms underlying the link between early-life adiposity and MBD, which can support future research into breast cancer prevention. Citation Format: Kayla R. Getz, Myung Sik Jeon, Lili Liu, Lei Liu, Chongliang Luo, Jingqin Luo, Adetunji T. Toriola. Using metabolomics to identify biomarkers mediating the associations of adiposity in childhood and early adulthood with mammographic breast density in premenopausal women [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3429.

Abstract 8: Deep Immunomics pipeline for discovery and validation of novel cancer-specific T cell receptors

Abstract Background: T cell receptor-engineered T cell (TCR-T) therapy has been studied as a high potential approach for cancer treatment. Beyond cell-surface tumor antigens, TCR-Ts can recognize HLA-presented intracellular epitopes which allows them to address a wide range of cancer-specific targets. Existing TCR-T therapies focus on a limited number of epitopes, many of which are restricted to HLA allele HLA-A*02:01. Identification of novel anti-tumor TCRs covering a broad range of HLA alleles is necessary to address unmet clinical needs. Natural T cells undergo thymic selection and present a favorable safety profile for mining such anti-tumor TCRs for therapeutic development. However, discovery of natural anti-tumor TCRs has been challenging due to i) the absence of high-throughput methods for identifying tumor-reactive TCRs in limited quantity human samples and ii) lack of sensitivity in detecting low frequency, naturally occurring tumor-reactive T cells. Methods: We developed TargetScape® as a part of our Deep Immunomics platform for efficient discovery of natural tumor-specific TCRs in healthy donors and cancer patients. Cells from blood and tissue samples were stained with metal-barcoded peptide-MHC tetramers, containing 500+ tumor-specific antigens, alongside metal-conjugated antibodies, allowing rapid identification and high dimensional phenotypic analysis of CD8+ T cells at the millions of cells scale. Single-cell sequencing revealed paired TCR sequences. Selected TCRs were introduced into luminescent Jurkat reporter cells for validation screening with peptide-pulsed targets and the relevant HLA restriction. TCRs of interest were further integrated into primary T cells and evaluated for their effector function. Results: We utilize our sensitive, high-throughput TargetScape® platform to detect and characterize putative cancer-specific TCRs from blood and tissue samples in frequencies as low as ~0.003% of total CD8+ cells. We have functionally validated over 100 TargetScape® derived TCRs with luminescent Jurkat reporters. Validated TCRs cover high prevalence HLA alleles and target a broad range of cancer antigen classes, including tumor associated antigens (TAA), Human endogenous retroviruses (HERVs), shared splice variants, and frameshift mutations. Several natural TCRs demonstrate recognition of low concentrations of peptide, indicating high functional avidity. Further evaluations of specificity and activity of TCRs of interest transduced into primary T cells are currently in progress. Conclusions: Our Deep Immunomics platform allows us to identify and confirm activity of novel, naturally-derived TCR candidates against therapeutically relevant cancer antigens presented on six common HLA alleles. These findings validate our discovery workflow leading to a portfolio of potential clinically relevant candidates that are currently under preclinical development. Citation Format: Hannah Fields, Juliana Velez Lujan, Kan Xing Wu, Natalie Epstein, Nazihah Beevi, Florence Chioh, Nicholas Tan, Neeraja Kulkarni, Noella Mukobo, Srikanthi R, Yovita Purwanti, Evan Newell, Alessandra Nardin, Michael Fehlings, Katja Fink, Daniel MacLeod. Deep Immunomics pipeline for discovery and validation of novel cancer-specific T cell receptors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 8.

Abstract 5520: β2 adrenergic receptor signaling promotes colorectal cancer liver metastasis

Abstract Liver is the most common site for cancer metastases, which may be attributed to the hepatic immunosuppressive microenvironment. Recent studies have highlighted the immunomodulatory and tumor-supportive roles of neurotransmitter-receptor signals in various cancers. Given that various neurotransmitter receptors are expressed in the liver, we aimed to investigate their possible involvement in the metastatic-prone liver microenvironment. Single-cell RNA sequencing analysis identified ADRB2, which encodes β2-adrenergic receptor (β2-AR), as the top-upregulated neurotransmitter receptor gene in colorectal cancer (CRC) and liver metastasis in patients with CRC liver metastasis (CRLM). Co-immunofluorescence (co-IF) staining confirmed the colocalization of β2-AR with myeloid cell marker CD11b instead of T cell marker CD3 and NK cell marker CD56 in patients with CRLM. Consistently, we observed the upregulation of CD11b/β2-AR double-positive cells in the liver tissues from mice bearing CRC or CRLM compared to naïve mice through co-IF staining. High-dimensional flow cytometry analysis further revealed that β2-AR was mainly expressed in macrophages and myeloid-derived suppressor cells (MDSCs). Treatment of CRC-tumor bearing mice with pan β-AR inhibitor propranolol and β2-AR specific inhibitor showed a potential to reduce liver metastatic incidence without suppressing primary CRC. Mechanistically, β2-AR activation may upregulate the expression of S100a8/9 to enhance the immunosuppressive activity of macrophages and MDSCs, thereby promoting CRC liver metastasis. Taken together, our findings reveal the potential of targeting β2-AR signaling for the prevention and treatment of liver metastasis. This project is supported by the Hong Kong University Grants Committee through the General Research Fund (14104820, 14107622, 24110323) and the Strategic Seed Funding for Collaborative Research Scheme (SSFCRS) from the Chinese University of Hong Kong. Conflict of interest: The authors declare no conflict of interest. Citation Format: Yixuan Zhang, Jintian Chen, Thomas Ting Hei Chan, Huanyu Wang, Jiahuan Lu, Jingqing Li, Jingying Zhou. β2 adrenergic receptor signaling promotes colorectal cancer liver metastasis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5520.

Abstract 6934: Integrating deep learning morphology profiling, cell sorting, and scRNASeq reveals sample diversity and enriched sub-populations

Abstract Advances in single cell technologies have ushered in a new multi-omics era that has provided more granular information on cell types and functions at unprecedented resolution. Morpholomics (the information encoded by cell morphology) has traditionally been used as a readout of cell identity, state, and function; but scalability, interpretation, and sorting based on morphology has remained challenging. The Deepcell® platform, REM-I, performs high-dimensional morphology analysis of unlabeled, single cells using deep-learning on high-resolution bright-field images captured in microfluidic flow to profile and sort target cells in real-time and at scale. Morphology analysis of single cells is achieved by a self-supervised deep learning foundation model, ‘Human Foundation Model’ (HFM), which combines deep learning features with morphometrics. HFM determines cell identities using morphological features associated with distinct processes characterized by granules, vesicles, cell size, pigmentation. and others. To demonstrate the ability of HFM modeling to distinguish and characterize a diversity of cell types representative of tumor microenvironments, we processed various sample types on the REM-I platform. Human melanoma, immune cells, stromal cells, and dissociated biopsies were combined to recapitulate heterogeneous tumor samples. HFM analysis enabled morphology-based exploration of the resulting bright-field images, revealing the presence of various morphologically distinct populations within each sample. Tumor cells were morphologically distinct and clustered separately from non-tumor cells. Immune cells exhibited subtle but separable morphologies, including T cells of distinct activation states. To demonstrate the morphology-based sorting capabilities of REM-I, we applied it to a ‘reference sample’ consisting of several fixed cell lines, at proportions ranging from 1% to &amp;gt;90%. Using only brightfield images of single cells analyzed through HFM, we used the platform to enrich each cell line from the ‘reference sample’ with high yield and purity, and identified cell types constituting a small percentage (~1%) of the total population. We next showed that unlabeled cells enriched on REM-I that are characterized as unperturbed and viable are amenable to downstream molecular analyses. These cell populations were sorted and the retrieved live cells were analyzed by scRNA-Seq (10x Chromium), demonstrating enrichment of various cell types of the tumor microenvironment in each cell population relative to the pre-sorted sample. These results illustrate: 1) the ability to perform downstream molecular analyses on sorted cells from REM-I, 2) the platform can be used to parse the tumor microenvironment, and 3) the link between morphology, cell identity, and molecular signatures. Citation Format: Andreja Jovic, Christian Corona, Kiran Saini, Tiffine Pham, Ryan Carelli, Vivian Lu, Senzeyu Zhang, Stephane C. Boutet, Maddison Masaeli. Integrating deep learning morphology profiling, cell sorting, and scRNASeq reveals sample diversity and enriched sub-populations [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6934.

Abstract 7040: Mutations in cytoskeleton genes SPTAN1/SPTBN1 serve as novel predictive markers for immunotherapy against bladder cancer

Abstract Cancer progression is often driven by somatic mutations. Despite several well-known driver mutations, more evidence has shown that genetic alterations in cytoskeleton organization enhance tumor aggressiveness and promote cancer progression, suggesting their roles as potential drivers. In this study, pan-cancer genetic analyses by using TCGA cohorts, more frequent mutation events were found in cytoskeleton-related genes as compared to total annotated genes, especially in cancers derived from hollow organs. High-dimensional analysis of molecular alterations in cancer (HD-MAC) suggested gene mutations in SPTAN1 and its partner SPTBN1 as potential prognostic biomarkers for clinical outcomes of patients with bladder cancer. Gene knockdown in either SPTAN1 or SPTBN1, mimicking silencing mutations in cancer tissues, caused cell morphology changes, DNA breaks, and alternations in cytokine profiles via activating cGAS-STING mediated interferon signaling. Xenograft tumor models also indicate accelerated tumor growth in SPTAN1- or SPTBN1-silenced cells, suggesting a more advanced cancer type. Based on immune staining in both human and mouse samples, increased CD8+ T cell infiltration was in tumor lesions with mutations or truncations in SPTAN1/SPTBN1 genes. In addition, several known substitution mutations were found as potent neo-antigens based on MHC II binding prediction, which could be utilized as targets for immune checkpoint therapy. This study provides a novel insight into how the cytoskeleton can contribute to cancer progression and find out predictive biological indicators with therapeutic values. Citation Format: Brian Yu-Ting Kuo. Mutations in cytoskeleton genes SPTAN1/SPTBN1 serve as novel predictive markers for immunotherapy against bladder cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7040.

Abstract 6031: Fasting associated microbiota contributes to small intestinal radioprotection by modulating the epigenome

Abstract Normal tissue toxicity often restricts the administration of effective radiation doses for optimal tumor cell elimination. We identified a novel host-protective effect conferred by short-term fasting when treating mice with high dose radiation. We showed that fasting mice for 24h induces epigenetic and transcriptional changes in small intestinal epithelial cells (siECs) and confers protection to small intestinal stem cells from high dose radiation. Here we report that fasting induces changes in the gut microbiome favoring bacteria producing metabolites, specifically short-chain fatty acids, that induce epigenetic changes in siECs. 16S rRNA sequencing of feces isolated from the small intestines of fed and fasted mice revealed a marked enrichment of Akkermansia muciniphila (AKK) in fasted relative to fed mice. Specific depletion of AKK through antibiotic treatment impaired the ability of fasting to provide organismal radioprotection whereas both organismal and OLFM4+ stem cell protection was restored in fasted animals upon reintroduction of AKK. AKK produces short chain fatty acids (SCFAs) including propionate which is able to induce histone modifications and our metabolomic analysis of feces isolated from the ileums of mice demonstrated a significant increase in propionic acid levels in fasted mice compared to fed mice. Importantly, we showed that AKK conditioned medium as well as its metabolite (propionate) are able to induce histone acetylation in Lgr5+ stem cell enriched epithelial spheroids cultured in vitro. We are currently employing high-dimensional analysis techniques, including single cell analysis, Cut&amp;Tag and RNAseq to determine how the epigenetic and transcriptional changes induced by fasting provide organismal and SI radioprotection. Citation Format: Praveen Barrodia, Emre Arslan, Sabrina Jeter-Jones, Robert R. Jenq, Kunal Rai, Helen Piwnica-Worms. Fasting associated microbiota contributes to small intestinal radioprotection by modulating the epigenome [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6031.

Abstract 232: Multi-dimensional patient-derived organoid analysis highly correlates with progression-free survival of pancreatic cancer patients in a retrospective clinical trial

Abstract Patient-derived tumor organoids (PDOs) are promising cancer models, since they preserve the clonal heterogeneity, mutational landscape, and histological architecture of the originating patient tumor1-2. Recent studies have even provided first evidence that PDOs responded similarly to the corresponding patient when treated with the same standard of care therapy, but these studies were only able to predict clinical responses in a subset of patients3,4. This restriction was largely due to the limitation of current analysis methods. The gold-standard analysis method for PDO-based research, CellTiter-Glo 3D,5,6 fails to account for the heterogeneity of PDOs by relying on a bulk lysis approach, and thus only extracts a fraction of the clinically-relevant information that PDOs could provide7. Therefore, we hypothesized that using higher-dimensional analysis methods, will further unlock the predictive performance of PDOs and facilitate translation of research. In this study, we developed an artificial intelligence-driven analysis platform for live-PDO imaging. By combining drug screening metrics with dynamic quantification of organoids on a single-organoid resolution, we unraveled patient-specific tumor heterogeneity and PDAC aggressiveness in response to therapies8. Using a fully characterized a PDAC organoid panel (n=8) we matched our PDO analysis with retrospective clinical patient response to standard of care therapies (gemcitabine-paclitaxel and FOLFIRINOX). Our results demonstrated that our PDO analysis method identified patient-specific sensitives to therapy that were in-line with clinical outcomes. Moreover, our PDO readouts highly correlated with progression-free survival of matched patients (R=0.97), which was a significant improvement to the current, conventional drug response readouts. Taken together, our approach increased the clinical translatability of PDOs by more accurately recapitulating and measuring the complexity of human tumors. This work highlights the potential applications (extendable to other tumor types) and clinical translatability of our approach in drug discovery and clinical care, in the emerging era of personalized medicine.