Supplementary materials and methods

AZ’1569 combined with chemotherapy or cetuximab does not induce apoptosis in KRASG12C MT colorectal cancer cells. A, CTG assays in colorectal cancer cells treated with no drug (control), 5-FU, AZ’1569, or 5-FU in combination with AZ’1569 for 72 hours. CI values were calculated using the method of Chou and Talalay. CI values <1, >1, and equal to 1 indicate synergy, antagonism, and additive effects for the drug combinations, respectively. Dashed lines indicate CI values of 0.3, 0.7, and 1. Representative results of at least three experiments (left). Absolute cell viability for different combinations (right). Dashed line indicates 50% cell viability. B, Colorectal cancer cells were cotreated with AZ’1569 and 5-FU, oxaliplatin, or SN-38 for 48 hours. Top: PARP, caspase 9, caspase 3, caspase 8, and KRAS levels were determined by WB. Bottom: Apoptosis was assessed by caspase 3/7 activity assay. * = nonspecific band. C, CTG cell viability assays in KRASG12C MT colorectal cancer cells cotreated with AZ’1569 and cetuximab for 120 hours. CI values were calculated to evaluate the nature of interaction. D,KRASG12C MT colorectal cancer cells were cotreated with AZ’1569 and cetuximab for 48 hours, apoptosis was assessed by WB analysis for PARP (top) and caspase-3/7 activity assays (bottom). A two-way ANOVA was used to analyze statistical significance.

Background:Identification of the consensus molecular subtypes (CMS) opened significant potential for understanding the tumor biology and intertumoral heterogeneity of colorectal cancer (CRC). However, molecular subtyping in CRC traditionally relies on bulk transcriptomics, therefore, lacks spatial and single-cell level aspect. Methods:We constructed tissue microarrays using tumor cores from 222 CRC patients. Arrays were stained and imaged using 54 cell identity and cancer hallmark markers, delivering spatially resolved protein profiles of >2 million cells. RNA sequencing data and CMS classification were also available for these patients. After segmentation of cancer, stromal and immune cells, we investigated intratumoral heterogeneity within CMS subtypes using spatially resolved single-cell protein profiling (>2 million cells). We compared cell types, their spatial organization and their expression of cancer hallmark-related proteins in CMS 1-4 subtypes. Results:We revealed tissue atlases illustrating the cell types/states, spatial heterogeneity, cellular neighborhoods, cellular network, and single-cell protein profiles of CMS tumors. CMS1 tumors had more CD3+, CD8+, and PD1+ immune cells that were found in the epithelial layer frequently. CMS1 was also associated with higher levels of metabolic reprogramming markers such as upregulated glycolysis. CMS2 showed immune segregation, reactive stroma patterns and higher levels of apoptotic and proliferative signaling proteins. CMS3 exhibited clustered cancer cells with high RIP3 levels, suggesting a pro-inflammatory microenvironment. CMS4 displayed stromal-centric and immune-evasive tumors characterized by decreased HLA-1 levels and regulatory T-cell exclusion from epithelium. Conclusion:We present a spatial protein atlas of CRC at single-cell resolution and demonstrate novel aspects of CMS tumour structures.

AZ’1569-acquired resistant cells exhibit increased PD-L1 expression and a proinflammatory phenotype. A, Left: RW7213 parental and AZ’1569-resistant clones (No. 2, No. 3, and No. 4) were treated for 72 hours with indicated concentrations of AZ’1569 or sotorasib, and cell viability was determined using CTG assays. Right: Lysates from RW7213 parental and AZ’1569-resistant clones were analyzed by WB for KRAS, pEGFRY1068, EGFR, pMETY1234/1235, MET, pERK1/2T202/Y204, ERK1/2, pS6S235/6, S6, pAKTS473, AKT, pEphA2S897, pEphA2Y588, pEphA2Y772, and EphA2. Active Raf1-bound Ras was isolated from RW7213 parental and resistant clones using an RAS-GTP assay and basal GTP-bound and total KRAS levels assessed by WB. LE = longer exposure. KRAS mRNA was quantified using RT-PCR. Raw values were normalized to ACTB and GAPDH expression and were analyzed using the ΔΔCT method. A one-way ANOVA was used to calculate statistical significance. Data are representative of three independent experimental repeats. Results of NGS of RW7213 Par and AZ’1569-R clones are shown. B, RW7213 parental and resistant cells were treated with SHP-099, BI-3406, 5-FU, SN-38, oxaliplatin, crizotinib, cetuximab, dasatinib, trametinib, ulixertinib, capivasertib, PF-4708671, AZD1480, ONC206, ABT-737, sabutoclax or entinostat for 72 hours, at the indicated concentrations and cell viability was assessed using CTG assays. Heatmap represents cell viability relative to control. Data are representative of three independent experimental repeats. C, Top left: Human cytokine array using conditioned medium of RW7213 parental and resistant clones. Right: Mean spot pixel density was analyzed using Image J, rZ (relative to parental cells) were calculated using densitometry data and presented in a heatmap. Bottom left: CXCL1, CD274, and IL8 mRNA in parental and resistant clones were quantified using RT-PCR. Raw values were normalized to the expression of housekeeping genes ACTB and GAPDH and were analyzed using the ΔΔCT method. CXCL1 protein levels in the culture media of parental (Par) and resistant subpopulations were measured by ELISA. A one-way ANOVA was used to calculate statistical significance. Data are representative of three independent experimental repeats. D, Left: Dose–response curves for AZ’1569 in RW7213 cells, incubated with conditioned media from parental cells or drug-resistant clones No. 2, No. 3, or No. 4. Cells were treated for 72 hours and cell viability was determined using CTG assay. IC50 values were calculated using a Prism software package. Dashed line indicates 50% cell viability. A representative of three independent experiments is shown. Right: A 24-well 5-μm polycarbonate Transwell insert-plate system was used. 2.5 × 105 PBMCs were resuspended in 2% FCS-supplemented DMEM and were added to the top chamber. The bottom chamber was filled with conditioned medium (medium = 2% FCS-supplemented DMEM) obtained from RW7213 parental and resistant cells. Cells were incubated for 4 hours, following which CellTiter-Glo was used to measure PBMC migration to the bottom chamber (RLU = relative luminescence). Serum-free DMEM was used in the bottom chamber as a negative control (neg CT). Data are representative of three independent experimental repeats.

Response to AZ’1569 in KRASG12C MT colorectal cancer cells. A,KRASG12CMT colorectal cancer cells were treated with increasing concentrations of AZ’1569 for 120 hours and cell viability determined using CTG assay. IC50 values were calculated using Prism software package. Dashed line indicates 50% cell viability. Representative of three independent experiments is shown. B, Colorectal cancer cells were treated with AZ’1569 for 48 hours. PARP, cleaved C3, and KRAS were determined by WB (top), caspase-3/7 activity levels were measured with values presented as a percentage of their respective controls. Significance was analyzed using an unpaired t test (bottom). (Cl = cleaved). C, Signaling analysis upon AZ’1569 treatment. KRASG12C MT colorectal cancer cell lines were treated with 1 μmol/L AZ’1569 for the indicated times, and protein lysates were used for WB analysis for the KRAS downstream effectors. Densitometry on WB images was quantified using ImageJ software and normalized to the respective untreated control. Dashed lines on the graphs represent a value of 1.

Combined KRASG12C and Bcl-xL inhibition results in reduction in growth of KRASG12C MT colorectal cancer in vivo. A, Growth rate (left) and mouse weight (right) of SW1463 xenografts in NOD/SCID mice treated with vehicle, AZ’8037, navitoclax, or AZ’8037 in combination with navitoclax. Differences in growth were determined using a one-way ANOVA with Tukey test for multiple comparisons. WB analysis for pERK1/2, ERK1/2, and KRAS in tumor samples collected at day 15. B, Growth rate (left) and mouse weight (right) of SNU1411 xenografts in NOD/SCID mice treated with vehicle, AZ’8037, navitoclax, or AZ’8037 in combination with navitoclax.

High-throughput drug screen reveals that pharmacologic inhibition of Bcl-xL synergizes with KRASG12C inhibition in KRASG12C MT colorectal cancer. A, SW837 and SNU1411 cells were cotreated with 1 μmol/L AZ’1569 alone or combined with a panel of 45 small-molecule inhibitors for 72 hours and cell viability assessed using the CTG assay. Three concentrations per drug were tested (Supplementary Table S2). Cell viability was analyzed using a CTG assay. Scatter plot showing rZ for each compound concentration used in the drug screen. Negative rZ-scores indicate agents that sensitize to AZ’1569, and vice versa. Dashed lines on graphs indicate values of 1.5 and −1.5. Venn diagram indicates number of compounds (past a threshold of rZ = −1.5) that resulted in sensitization to AZ’1569 in both cell lines. B, CTG cell viability assays in KRASG12C MT colorectal cancer cells cotreated with AZ’1569 and ABT-737 for 72 hours. CI values were calculated to evaluate the nature of interaction. Absolute cell viability for AZ’1569/ABT-737 combinations in SW847 and SNU1411 cell lines are also shown. Dashed lines on graphs represent 50% cell viability. C, PARP, cleaved C9, cleaved C8, cleaved C3, and KRAS expression levels in KRASG12C MT colorectal cancer cells cotreated with AZ’1569 and ABT-737 for 48 hours (24 hours for RW7213, SW1463, and V481 cells). CM = combination. D,KRASG12C MT colorectal cancer cells were treated with AZ’1569 alone or combined with cetuximab or ABT-737 (0.25 μmol/L for C106, LIM2099, SNU1411; 0.5 μmol/L for SW837, V481, RW7213 and 2.5 μmol/L for SW1463) for 48 hours (24 hours for C106 cells) and PARP, cleaved C9, cleaved C8, and cleaved C3 determined by WB.

Bcl-xL regulates intrinsic resistance to KRASG12C inhibition in KRASG12C MT colorectal cancer. A, Targeted siRNA screen in SW837 and SNU1411 cells. Top: SW837 and SNU1411 cells were reverse transfected with 10 nmol/L ON-Targetplus siRNA's targeting 42 genes in the absence or presence of 1 μmol/L AZ’1569 for 72 hours and cell viability was evaluated using the CTG assay. Scatter plot showing rZ for siRNA screen in SW837 and SNU1411 cells. Positive scores indicate potential mediators of sensitivity to AZ’1569, while negative scores indicate mediators of resistance to AZ’1569. Dashed lines indicate rZ = 0, 1.5, and −1.5; cut-off thresholds of ±1.5 were applied to the data. Bottom: The siRNA approach and analysis. XY graph illustrates hits resulting in sensitization or resistance to AZ’1569 in both cell lines. Data show average rZ-scores from three independent experiments. B, Colorectal cancer cells were transfected with 10 nmol/L on-target SMARTpool siRNA against BCL2L1 and cotreated with 1 μmol/L AZ’1569 (0.25 μmol/L AZ’1569 for RW7213 and C106 cells) for 24 hours (48 hours for SNU1411) and apoptosis assessed by WB for PARP and cleaved caspase 8 and 3 (top) and caspase 3/7 activity assay (bottom). n.d denotes not detected. A two-way ANOVA was used to evaluate significance. C, Expression of PARP, cleaved caspase 9, caspase 8, Myc-tag, and KRAS in SW837 and RW7213 cells transiently transfected with 1 μg of Myc-tagged Bcl-xL for 24 hours, followed by treatment with 1 μmol/L AZ’1569 (AZD) for the indicated times. Caspase-3/7 activity on cell lysates was also determined. A two-way ANOVA was used to evaluate significance.

ObjectiveInducing tumour cell apoptosis is a primary objective of chemotherapy but, to date, there are no validated biomarkers of apoptosis sensitivity or resistance. Our objective was to image multiple apoptosis pathway proteins at single cell level and determine multi-protein associations with recurrence risk and chemotherapy response in patients with stage II colorectal cancer (CRC).Methods and analysisMultiplexed imaging of 16 proteins in the intrinsic and extrinsic apoptosis pathways at single cell resolution on resected tissue from 194 patients with stage II CRC who either received adjuvant chemotherapy (n=108) or were treated with surgery only (n=86). K-means clustering of >600 000 cancer cells and cell level intensities of APAF1, procaspase-9, procaspase-3, XIAP, SMAC, BAX, BAK, BCL2, BCL-XL, MCL-1, procaspase-8, BID, FADD, FLIP, RIP3 and CIAP1 identified distinct cell cluster profiles.ResultsChemotherapy-treated patients with a higher percentage of cell clusters with low procaspase-3 and high XIAP had a higher risk of recurrence. This was validated in an independent cohort of adjuvant chemotherapy-treated high-risk patients with stage II CRC. We also applied two established system models of apoptosis initiation and execution to estimate cellular apoptosis sensitivity and show that these cell clusters do not appear to have impaired mitochondrial outer membrane permeabilisation sensitivity, but downstream procaspase-3 cleavage is compromised. This represents a key characteristic of drug-tolerant ‘persister’ cells.ConclusionThis study represents the most comprehensive analysis to date of apoptosis protein distribution at single cell level in CRC tumours. Our study identifies a subgroup of patients with stage II CRC with an apoptosis-resistant ‘persister’ cell profile who do not benefit from adjuvant chemotherapy.

Abstract Background: Invasive lobular carcinoma (ILC) is the second most pervasive subtype after invasive ductal carcinoma (IDC), accounting for approximately 10-15% of all breast cancers. It is characterized by loss of E-cadherin expression and non-adherent tumor cells that invade the stroma in a “single-file” pattern. Women with ILC are typically diagnosed at an older age and later stage with ER-positive disease. ILC is more likely to exhibit late recurrence and metastasize to the gastrointestinal tract and urogenital tract compared with IDC. It is routinely treated with anti-endocrine therapy and chemotherapy, however, while not entirely chemo-refractory, displays a poor response to chemotherapy compared with IDC. As such, options for recurrent disease are limited and there is an urgent need to develop tailored therapy for ILC, especially for those patients that recur. The family of bromodomain and extra-terminal domain (BET) proteins, comprising BRD2/3/4/T, are epigenetic readers that bind to acetylated lysine residues on histones and recruit transcription factors to drive the expression of oncogenes. Previously. we discovered that BRD3 is a marker of poor prognosis in ILC and there is emerging evidence that BET inhibitors (iBET) are effective in diverse types of breast cancer. Here, we investigated the therapeutic potential of iBET in ILC, alone and in combination with FGFR inhibitors. Methods: IC50s for a panel of iBET using two typical ILC cell lines MDA-MB-134VI (MM134) and SUM44PE (SUM44) were determined. RNA-sequencing and Genexplain analysis was applied to reveal transcriptional networks, master regulators and potential resistance mechanisms. BRD3 and FGFR3 were knocked down using siRNA to evaluate their function in ILC cell lines. Furthermore, we utilized ILC cell-derived xenograft (CDX) models in SCID-beige mice established by mammary intraductal (MIND) implantation to evaluate the therapeutic potency of iBET alone or in combination with fibroblast growth factor receptor (FGFR) inhibitor in vivo. Results: We demonstrated that iBET significantly inhibited ILC cell growth in both 2D and 3D culture, with the greatest potency demonstrated by JQ1 and Mivebresib (ABBV-075). RNA-sequencing revealed dysregulated pathways in cell cycle division, DNA damage, apoptosis and MAPK signaling following iBET treatment. Reverse engineering of transcriptional profiles using Genexplain revealed that FGFR3 is a significantly upregulated master regulator (MTR) among 142 MTRs across both cell lines and both iBETs. Upregulation of FGFR3 after iBET treatment was verified at the protein level by Western blotting. We also show that BRD3 and FGFR3 knockdown significantly inhibited cell growth, which supports the key role both play in ILC progression. Further, we analyzed the iBET therapeutic effect when combined with the FGFR inhibitor, erdafitinib, as a strategy to overcome potential resistance due to FGFR upregulation post iBET treatment. This revealed that the combination of iBET and erdafitinib could inhibit ILC cell growth more effectively compared to using either agent alone. Furthermore, our in vivo study showed that JQ1 could inhibit tumor growth in a SUM44-MIND model and alleviate metastasis to peritoneum, bone and ovary compared with the vehicle group. Moreover, we also assessed the combination of mivebresib and erdafitinib in vivo. This revealed that iBET and an FGFR inhibitor work synergistically to decrease tumour burden and metastatic potential in both MM134-MIND and SUM44-MIND models. Conclusion: Our results provide evidence that iBET, either alone or in combination with erdafitinib, is remarkably effective at inhibiting ILC growth, both in vitro and in vivo and represents a rational therapeutic strategy for recurrent ILC patients in the future. Citation Format: Binbin Gao, Elspeth Ward, Anna Blümel, Emer Conroy, Rachel Moore, Grainne Cremin, Rachel Bleach, Kathryn Haley, Tríona Ní Chonghaíle, Andreas Lindner, Jochen Prehn, Yi Zhang, Idalia Cruz, Leena Hilakivi-Clarke, Georgios Sflomos, Cathrin Brisken, William Gallagher, Darran O'Connor. Rational therapeutic combination of Bromodomain and Extra-Terminal domain (BET) inhibitor and Fibroblast Growth Factor Receptor (FGFR) inhibitor for treatment of invasive lobular carcinoma [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 PO3-18-08.

BACKGROUND: Mucinous rectal cancer is associated with a higher incidence of microsatellite instability and a poorer response to neoadjuvant chemoradiotherapy compared to other subtypes of rectal adenocarcinoma. Immune checkpoint inhibitors are an emerging family of anticancer therapeutics associated with highly variable outcomes in colorectal cancer. Although the immune landscape of mucinous rectal cancer has not been fully explored, the presence of mucin is thought to act as a barrier preventing immune-cell infiltration. OBJECTIVE: The aim of this study was to determine the immune properties of mucinous rectal cancer and investigate the degree of lymphocyte infiltration in this cohort. DESIGN: This is a retrospective cohort study that involved multiplexed immunofluorescence staining of tumor microarrays. SETTINGS: Samples originated from a single university teaching hospital. PATIENTS: Our cohort included 15 cases of mucinous and 43 cases of nonmucinous rectal cancer. MAIN OUTCOME MEASURES: Immune cells were classified and quantified. Immune-cell counts were compared between mucinous and nonmucinous cohorts. Immune marker expression within tumor epithelial tissue was evaluated to determine the degree of lymphocyte infiltration. RESULTS: Cytotoxic ( p = 0.022) and regulatory T cells ( p = 0.010) were found to be overrepresented in the mucinous cohort compared to the nonmucinous group. Programmed cell death protein 1 expression was also found to be significantly greater in the mucinous group ( p = 0.001). CD3 ( p = 0.001) and CD8 ( p = 0.054) expressions within the tumor epithelium were also higher in the mucinous group, suggesting adequate immune infiltration despite the presence of mucin. In our analysis, microsatellite instability status was not a predictor of immune marker expression. LIMITATIONS: The relatively small size of the cohort. CONCLUSIONS: Mucinous rectal cancer is associated with an immune-rich tumor microenvironment, which was not associated with microsatellite instability status. See Video Abstract at http://links.lww.com/DCR/C65. IMÁGENES DE INMUNOFLUORESCENCIA MULTIPLEXADAS REVELAN UN MICROAMBIENTE TUMORAL RICO EN INMUNIDAD EN EL CÁNCER RECTAL MUCINOSO CARACTERIZADO POR UNA MAYOR INFILTRACIÓN DE LINFOCITOS Y UNA EXPRESIÓN MEJORADA DE PD-1 ANTECEDENTES: El cáncer rectal mucinoso se asocia con una mayor incidencia de inestabilidad de microsatélites y una peor respuesta a la quimiorradioterapia neoadyuvante en comparación con otros subtipos de adenocarcinoma rectal. Los inhibidores de puntos de control inmunitarios son una familia emergente de tratamientos contra el cáncer asociados con resultados muy variables en el cáncer colorrectal. Aunque el panorama inmunitario del cáncer rectal mucinoso no se ha explorado completamente, se cree que la presencia de mucina actúa como una barrera que previene la infiltración de células inmunitarias. OBJETIVO: El objetivo de este estudio fue determinar las propiedades inmunes del cáncer de recto mucinoso e investigar el grado de infiltración de linfocitos en esta cohorte. DISEÑO: Este es un estudio de cohorte retrospectivo que involucró la tinción de inmunofluorescencia multiplexada de micromatrices tumorales. AJUSTES: Las muestras se originaron en un solo hospital docente universitario. PACIENTES: Nuestra cohorte incluyó 15 casos de cáncer de recto mucinoso y 43 casos de cáncer de recto no mucinoso PRINCIPALES MEDIDAS DE RESULTADO: Las células inmunitarias se clasificaron y cuantificaron. Se compararon los recuentos de células inmunitarias entre cohortes mucinosas y no mucinosas. Se evaluó la expresión del marcador inmunitario dentro del tejido epitelial tumoral para determinar el grado de infiltración de linfocitos. RESULTADOS: Se encontró que las células T citotóxicas ( p = 0,022) y reguladoras ( p = 0,010) estaban sobrerrepresentadas en la cohorte mucinosa en comparación con el grupo no mucinoso. También se encontró que la expresión de PD-1 era significativamente mayor en el grupo mucinoso ( p = 0,001). La expresión de CD3 ( p = 0,001) y CD8 ( p = 0,054) dentro del epitelio tumoral también fue mayor en el grupo mucinoso, lo que sugiere una infiltración inmunitaria adecuada a pesar de la presencia de mucina. En nuestro análisis, no se encontró que el estado de inestabilidad de los microsatélites sea un predictor de la expresión del marcador inmunitario. LIMITACIONES: El tamaño relativamente pequeño de la cohorte. CONCLUSIONES: El cáncer rectal mucinoso se asocia con un microambiente tumoral rico en inmunidad, que no se asoció con el estado de inestabilidad de microsatélites. Consulte el Video del Resumen en http://links.lww.com/DCR/C65. (Traducción— Dr. Yesenia Rojas-Khalil )

Abstract Apoptosis is regulated by intrinsic and extrinsic signaling orchestrated through multiple proteins that initiate or inhibit apoptosis. Inducing tumor cell apoptosis is a primary objective of chemotherapy, but to date, biomarkers of apoptosis sensitivity or resistance have shown mixed results. To understand the interplay between these proteins and whether the balance of apoptosis regulators influences chemotherapy responses and patient outcomes, we conducted multiplexed imaging of 16 proteins in the intrinsic and extrinsic apoptosis pathways at single cell resolution on resected tissue from 194 stage II colorectal cancer (CRC) patients who either received adjuvant chemotherapy (n=108) or were treated with surgery only (n=86). K-means clustering of >400,000 individual cancer cells and cell-level intensities of Apaf-1, procaspase-9, procaspase-3, XIAP, SMAC, BAX, BAK, Bcl-2, Bcl-xL, MCL-1, procaspase-8, BID, FADD, FLIP, RIP3 and cIAP1 identified distinct cell cluster profiles and demonstrated profound patient-to-patient heterogeneity. In chemotherapy-treated stage II CRC patients, those with a higher percentage of cell clusters with low procaspase-3 and high XIAP had a higher risk of recurrence. The high risk cell cluster (low-procaspase 3, high XIAP) was validated in an independent cohort of adjuvant chemotherapy-treated high-risk stage II CRC patients. To further interrogate the apoptosis sensitivity of the cell clusters, we also applied two established systems models of apoptosis initiation and execution, the BCL-2 pathway (DR_MOMP) and the caspase activation pathway (APOPTO-CELL). Here we showed that cell clusters associated with increased recurrence risk do not appear have impaired MOMP sensitivity, but downstream procaspase-3 cleavage is compromised. This represents a key characteristic of drug-tolerant ‘persister’ cells. Our study represents the most comprehensive, integrated analysis to date of apoptosis protein distribution at single-cell level in CRC tumors and identifies a subgroup of stage II patients with an apoptosis resistant, ‘persister’ cell profile who do not benefit from adjuvant chemotherapy.

Abstract Apoptosis is regulated by intrinsic and extrinsic signaling orchestrated through multiple proteins that initiate or inhibit apoptosis. Inducing tumor cell apoptosis is a primary objective of chemotherapy, but to date, biomarkers of apoptosis sensitivity or resistance have shown mixed results. To understand the interplay between these proteins and whether the balance of apoptosis regulators influences chemotherapy responses and patient outcomes, we conducted multiplexed imaging of 16 proteins in the intrinsic and extrinsic apoptosis pathways at single cell resolution on resected tissue from 194 stage II colorectal cancer (CRC) patients who either received adjuvant chemotherapy ( n= 108) or were treated with surgery only (n=86). K-means clustering of >400,000 individual cancer cells and cell-level intensities of Apaf-1, procaspase-9, procaspase-3, XIAP, SMAC, BAX, BAK, Bcl-2, Bcl-xL, MCL-1, procaspase-8, BID, FADD, FLIP, RIP3 and cIAP1 identified distinct cell cluster profiles and demonstrated profound patient-to-patient heterogeneity. In chemotherapy-treated stage II CRC patients, those with a higher percentage of cell clusters with low procaspase-3 and high XIAP had a higher risk of recurrence. The high risk cell cluster (low-procaspase 3, high XIAP) was validated in an independent cohort of adjuvant chemotherapy-treated high-risk stage II CRC patients. To further interrogate the apoptosis sensitivity of the cell clusters, we also applied two established systems models of apoptosis initiation and execution, the BCL-2 pathway (DR_MOMP) and the caspase activation pathway (APOPTO-CELL). Here we showed that cell clusters associated with increased recurrence risk do not appear have impaired MOMP sensitivity, but downstream procaspase-3 cleavage is compromised. This represents a key characteristic of drug-tolerant ‘persister’ cells. Our study represents the most comprehensive, integrated analysis to date of apoptosis protein distribution at single-cell level in CRC tumors and identifies a subgroup of stage II patients with an apoptosis resistant, ‘persister’ cell profile who do not benefit from adjuvant chemotherapy.

Impact of each predictor included in the Random Forest classifier on the probability of recurrence alone and in combination with the others.

The frequency of lymphocytes infiltrating tumors is a known prognostic in estrogen receptor (ER) negative cancers. ER+ disease is putatively believed to be immune cold, however, there exists a subset of ER+ tumors with high immune infiltrate and with a significant spatial heterogeneity. The clinic impact of such infiltrate - especially between the Oncotype Dx Recurrence Score risk categories - remains unclear. Moreover, the distribution of tumor and stromal tissues, while noted as significantly heterogenous, is still ill-defined and not yet clinically used prognostically, despite evidence to support its utility. Using a cohort (n=450) of serial sections taken from early-stage, ER+/HER2- breast tumors of the Irish arm of the TAILORx clinical trial, we aimed to investigate the tumor architecture and spatial distribution of tumor immune infiltrate and proliferating tumor cells using digital image analysis. Antibodies against Ki67 (proliferation marker) and CD45 (leukocyte common antigen), and a routine Haematoxylin and Eosin stain were applied to serial sections of 450 full-face tumors via chromogenic immunohistochemistry, as outlined previously [1]. Digital image analysis was performed using open-source software, QuPath [2]. Pixel classifiers were trained and validated against an expert pathologist in order to define observed lymphocytes as tumor or stromal-infiltrating, and to establish a classifier to quantify the tumor-stroma ratio (TSR) and infiltrating tumor area. Distances of CD45-positive cells from tumor were computed, along with autocorrelation statistics [3,4] of CD45 and Ki67 hotspots; firstly in order to quantify spatial heterogeneity, and secondly to examine whether Ki67 as a component gene in the Oncotype Dx assay has a foundation in tumor biology or is being confounded by potentially Ki67-positive lymphocytes. Subdividing by Oncotype Dx risk categories, no significant difference in TSR was observed (p=0.09799), neither for intermediary risk patients receiving hormone therapy alone or in combination with chemotherapy (p=0.3873). While there was an observed trend overall (p=0.092), no significance was found for recurrence between intermediary risk subcategories (HT alone: p=0.393, HT+CT: p=0.288). However, in the cohort as a whole, median TSR was 0.3215 (range 0 - 5.023), with statistically significant differences in recurrence risk observed (cohort high v low by median TSR. HR: 6.356, 95CI: 2.263-17.84, p<0.0001). Citation Format: Zak Kinsella, Anna Blümel, Mairi Lucas, Andreas Lindner, Claudia A. Gonzalez, Arman Rahman, Joanna Fay, Tony O'Grady, Verena Murphy, John Crown, Cathy Kelly, William Gallagher, Darran O'Connor. Modelling the spatial heterogeneity of CD45-positive tumor infiltrating lymphocytes in early-stage, estrogen receptor-positive breast 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 5787.

Detailed description of the patients data handling and inclusion criteria for downstream analyses of the discovery, expansion and validation cohorts.

PDF file - 66K, Temporal protein profiles for HeLa cells.

Kaplan-Meier estimates for disease-free and overall survival for n=120 stage III patients of the discovery cohort categorized based on TN staging, tumor location and lymphovascular invasion. Additional exploratory analyses investigated the prognostic value of the APOPTO-CELL-PC3 signature within each sub-group identified by the clinical features.