InForm software: a semi-automated research tool to identify presumptive human hepatic progenitor cells, and other histological features of pathological significance

Spatial omics techniques and data analysis for cancer immunotherapy applications

In-depth profiling of cancer cells/tissues is expanding our understanding of the genomic, epigenomic, transcriptomic, and proteomic landscape of cancer. However, the complexity of the cancer microenvironment, particularly its immune regulation, has made it difficult to exploit the potential of cancer immunotherapy. High-throughput spatial omics technologies and analysis pipelines have emerged as powerful tools for tackling this challenge. As a result, a potential revolution in cancer diagnosis, prognosis, and treatment is on the horizon. In this review, we discuss the technological advances in spatial profiling of cancer around and beyond the central dogma to harness the full benefits of immunotherapy. We also discuss the promise and challenges of spatial data analysis and interpretation and provide an outlook for the future.

Deep learning on tertiary lymphoid structures in hematoxylin-eosin predicts cancer prognosis and immunotherapy response

Tertiary lymphoid structures (TLSs) have been associated with favorable immunotherapy responses and prognosis in various cancers. Despite their significance, their quantification using multiplex immunohistochemistry (mIHC) staining of T and B lymphocytes remains labor-intensive, limiting its clinical utility. To address this challenge, we curated a dataset from matched mIHC and H&E whole-slide images (WSIs) and developed a deep learning model for automated segmentation of TLSs. The model achieved Dice coefficients of 0.91 on the internal test set and 0.866 on the external validation set, along with intersection over union (IoU) scores of 0.819 and 0.787, respectively. The TLS ratio, defined as the segmented TLS area over the total tissue area, correlated with B lymphocyte levels and the expression of CXCL13, a chemokine associated with TLS formation, in 6140 patients spanning 16 tumor types from The Cancer Genome Atlas (TCGA). The prognostic models for overall survival indicated that the inclusion of the TLS ratio with TNM staging significantly enhanced the models' discriminative ability, outperforming the traditional models that solely incorporated TNM staging, in 10 out of 15 TCGA tumor types. Furthermore, when applied to biopsied treatment-naïve tumor samples, higher TLS ratios predicted a positive immunotherapy response across multiple cohorts, including specific therapies for esophageal squamous cell carcinoma, non-small cell lung cancer, and stomach adenocarcinoma. In conclusion, our deep learning-based approach offers an automated and reproducible method for TLS segmentation and quantification, highlighting its potential in predicting immunotherapy response and informing cancer prognosis.

Neoantigen-specific CD8 T cells with high structural avidity preferentially reside in and eliminate tumors

The success of cancer immunotherapy depends in part on the strength of antigen recognition by T cells. Here, we characterize the T cell receptor (TCR) functional (antigen sensitivity) and structural (monomeric pMHC-TCR off-rates) avidities of 371 CD8 T cell clones specific for neoantigens, tumor-associated antigens (TAAs) or viral antigens isolated from tumors or blood of patients and healthy donors. T cells from tumors exhibit stronger functional and structural avidity than their blood counterparts. Relative to TAA, neoantigen-specific T cells are of higher structural avidity and, consistently, are preferentially detected in tumors. Effective tumor infiltration in mice models is associated with high structural avidity and CXCR3 expression. Based on TCR biophysicochemical properties, we derive and apply an in silico model predicting TCR structural avidity and validate the enrichment in high avidity T cells in patients' tumors. These observations indicate a direct relationship between neoantigen recognition, T cell functionality and tumor infiltration. These results delineate a rational approach to identify potent T cells for personalized cancer immunotherapy.

MIAMI: mutual information-based analysis of multiplex imaging data

MOTIVATION

Studying the interaction or co-expression of the proteins or markers in the tumor microenvironment of cancer subjects can be crucial in the assessment of risks, such as death or recurrence. In the conventional approach, the cells need to be declared positive or negative for a marker based on its intensity. For multiple markers, manual thresholds are required for all the markers, which can become cumbersome. The performance of the subsequent analysis relies heavily on this step and thus suffers from subjectivity and lacks robustness.

RESULTS

We present a new method where different marker intensities are viewed as dependent random variables, and the mutual information (MI) between them is considered to be a metric of co-expression. Estimation of the joint density, as required in the traditional form of MI, becomes increasingly challenging as the number of markers increases. We consider an alternative formulation of MI which is conceptually similar but has an efficient estimation technique for which we develop a new generalization. With the proposed method, we analyzed a lung cancer dataset finding the co-expression of the markers, HLA-DR and CK to be associated with survival. We also analyzed a triple negative breast cancer dataset finding the co-expression of the immuno-regulatory proteins, PD1, PD-L1, Lag3 and IDO, to be associated with disease recurrence. We demonstrated the robustness of our method through different simulation studies.

AVAILABILITY AND IMPLEMENTATION

The associated R package can be found here, https://github.com/sealx017/MIAMI.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

FIBER-ML, an Open-Source Supervised Machine Learning Tool for Quantification of Fibrosis in Tissue Sections

Pathologic fibrosis is a major hallmark of tissue insult in many chronic diseases. Although the amount of fibrosis is recognized as a direct indicator of the extent of disease, there is no consentaneous method for its quantification in tissue sections. This study tested FIBER-ML, a semi-automated, open-source freeware that uses a machine-learning approach to quantify fibrosis automatically after a short user-controlled learning phase. Fibrosis was quantified in sirius red-stained tissue sections from two fibrogenic animal models: acute stress-induced cardiomyopathy in rats (Takotsubo syndrome-like) and HIV-induced nephropathy in mice (chronic kidney disease). The quantitative results of FIBER-ML software version 1.0 were compared with those of ImageJ in Takotsubo syndrome, and with those of inForm in chronic kidney disease. Intra- and inter-operator and inter-software correlation and agreement were assessed. All correlations were excellent (>0.95) in both data sets. The values of discriminatory power between the pathologic and healthy groups were <10^-3 for data on Takotsubo syndrome and <10^-4 for data on chronic kidney disease. Intra-operator agreement, assessed by intra-class coefficient correlation, was good (>0.8), while inter-operator and inter-software agreement ranged from moderate to good (>0.7). FIBER-ML performed in a fast and user-friendly manner, with reproducible and consistent quantification of fibrosis in tissue sections. It offers an open-source alternative to currently used software, including quality control and file management.

A SIMPLI (Single-cell Identification from MultiPLexed Images) approach for spatially-resolved tissue phenotyping at single-cell resolution

Multiplexed imaging technologies enable the study of biological tissues at single-cell resolution while preserving spatial information. Currently, high-dimension imaging data analysis is technology-specific and requires multiple tools, restricting analytical scalability and result reproducibility. Here we present SIMPLI (Single-cell Identification from MultiPLexed Images), a flexible and technology-agnostic software that unifies all steps of multiplexed imaging data analysis. After raw image processing, SIMPLI performs a spatially resolved, single-cell analysis of the tissue slide as well as cell-independent quantifications of marker expression to investigate features undetectable at the cell level. SIMPLI is highly customisable and can run on desktop computers as well as high-performance computing environments, enabling workflow parallelisation for large datasets. SIMPLI produces multiple tabular and graphical outputs at each step of the analysis. Its containerised implementation and minimum configuration requirements make SIMPLI a portable and reproducible solution for multiplexed imaging data analysis. Software is available at “SIMPLI [https://github.com/ciccalab/SIMPLI]”.

Current high-dimension imaging data analysis methods are technology-specific and require multiple tools, restricting analytical scalability and result reproducibility. Here the authors present SIMPLI, a software that overcomes these limitations for single-cell and pixel analysis of multiplexed images at spatial resolution.

Assessment of patient-to-patient and intra-individual human abdominal skin immune cell variability

INTRODUCTION

The objective of the study was to characterize the baseline intra-individual and inter-individual variability of immune cell subsets within abdominoplasty skin specimens.

METHODS

Abdominoplasty biopsies were taken from 5 patients and analysed using the Vectra 3 automated quantitative pathology imaging system with inForm software.

RESULTS

Adjacent skin regions demonstrated intra-patient variability in immune subset counts ranging from 1- to 5-fold. Inter-variability between patients was approximately 2- to 7-fold for most subsets, except for HLA-DR⁺ antigen presenting cells, which varied 19-fold.

CONCLUSIONS

Our data highlight the importance of including multiple patients and multiple patient samples when designing dermatological studies that utilise abdominoplasty skin.

Human Amnion Epithelial Cells Produce Soluble Factors that Enhance Liver Repair by Reducing Fibrosis While Maintaining Regeneration in a Model of Chronic Liver Injury

Human amnion epithelial cells (hAECs) exert potent antifibrotic and anti-inflammatory effects when transplanted into preclinical models of tissue fibrosis. These effects are mediated in part via the secretion of soluble factors by hAECs which modulate signaling pathways and affect cell types involved in inflammation and fibrosis. Based on these reports, we hypothesized that these soluble factors may also support liver regeneration during chronic liver injury. To test this, we characterized the effect of both hAECs and hAEC-conditioned medium (CM) on liver repair in a mouse model of carbon tetrachloride (CCl₄)-induced fibrosis. Liver repair was assessed by liver fibrosis, hepatocyte proliferation, and the liver progenitor cell (LPC) response. We found that the administration of hAECs or hAEC-CM reduced liver injury and fibrosis, sustained hepatocyte proliferation, and reduced LPC numbers during chronic liver injury. Additionally, we undertook in vitro studies to document both the cell-cell and paracrine-mediated effects of hAECs on LPCs by investigating the effects of co-culturing the LPCs and hAECs and hAEC-CM on LPCs. We found little change in LPCs co-cultured with hAECs. In contrast, hAEC-CM enhances LPC proliferation and differentiation. These findings suggest that paracrine factors secreted by hAECs enhance liver repair by reducing fibrosis while promoting regeneration during chronic liver injury.

ImmunoAIzer: A Deep Learning-Based Computational Framework to Characterize Cell Distribution and Gene Mutation in Tumor Microenvironment

Spatial distribution of tumor infiltrating lymphocytes (TILs) and cancer cells in the tumor microenvironment (TME) along with tumor gene mutation status are of vital importance to the guidance of cancer immunotherapy and prognoses. In this work, we developed a deep learning-based computational framework, termed ImmunoAIzer, which involves: (1) the implementation of a semi-supervised strategy to train a cellular biomarker distribution prediction network (CBDPN) to make predictions of spatial distributions of CD3, CD20, PanCK, and DAPI biomarkers in the tumor microenvironment with an accuracy of 90.4%; (2) using CBDPN to select tumor areas on hematoxylin and eosin (H&E) staining tissue slides and training a multilabel tumor gene mutation detection network (TGMDN), which can detect APC, KRAS, and TP53 mutations with area-under-the-curve (AUC) values of 0.76, 0.77, and 0.79. These findings suggest that ImmunoAIzer could provide comprehensive information of cell distribution and tumor gene mutation status of colon cancer patients efficiently and less costly; hence, it could serve as an effective auxiliary tool for the guidance of immunotherapy and prognoses. The method is also generalizable and has the potential to be extended for application to other types of cancers other than colon cancer.

Immune Checkpoint Blockade Improves Chemotherapy in the PyMT Mammary Carcinoma Mouse Model

Despite the success of immune checkpoint blockade in cancer, the number of patients that benefit from this revolutionary treatment option remains low. Therefore, efforts are being undertaken to sensitize tumors for immune checkpoint blockade, which includes combining immune checkpoint blocking agents such as anti-PD-1 antibodies with standard of care treatments. Here we report that a combination of chemotherapy (doxorubicin) and immune checkpoint blockade (anti-PD-1 antibodies) induces superior tumor control compared to chemotherapy and immune checkpoint blockade alone in the murine autochthonous polyoma middle T oncogene-driven (PyMT) mammary tumor model. Using whole transcriptome analysis, we identified a set of genes that were upregulated specifically upon chemoimmunotherapy. This gene signature and, more specifically, a condensed four-gene signature predicted favorable survival of human mammary carcinoma patients in the METABRIC cohort. Moreover, PyMT tumors treated with chemoimmunotherapy contained higher levels of cytotoxic lymphocytes, particularly natural killer cells (NK cells). Gene set enrichment analysis and bead-based ELISA measurements revealed increased IL-27 production and signaling in PyMT tumors upon chemoimmunotherapy. Moreover, IL-27 signaling improved NK cell cytotoxicity against PyMT cells in vitro. Taken together, our data support recent clinical observations indicating a benefit of chemoimmunotherapy compared to monotherapy in breast cancer and suggest potential underlying mechanisms.

The role of neoadjuvant chemotherapy in the management of low-grade serous carcinoma of the ovary and peritoneum: Further evidence of relative chemoresistance

OBJECTIVE

Low-grade serous carcinoma of the ovary/peritoneum (LGSC) is relatively chemoresistant in the adjuvant, neoadjuvant, and recurrent settings. We sought to expand our prior work and evaluate response rates of women with LGSC to neoadjuvant chemotherapy (NACT) compared to women with high-grade serous carcinoma of the ovary/peritoneum (HGSC).

METHODS

Thirty-six patients with LGSC who received NACT were matched to patients with HGSC. A single radiologist re-reviewed pre- and post-NACT imaging for response using RECIST 1.1. Pre- and post-NACT CA-125 values were compared using paired t-tests. Kaplan-Meier estimates of progression free survival (PFS) and overall survival (OS) were performed.

RESULTS

All patients received neoadjuvant platinum-based regimens. LGSC patients received a median of 5 cycles (range 3-9), HGSC patients received a median of 4 cycles (range 3-9). Interval cytoreductive surgery was performed in 29/36 (81%) of LGSC and 32/36 (89%) HGSC patients. Complete cytoreduction was reported and achieved in 11/29 (38%) of LGSC patients and 24/32 (75%) of HGSC patients (p = 0.002). Median pre- and post-treatment CA-125 levels for LGSC patients were 295.5 U/mL and 144 U/mL (52% decrease) (p < 0.001). The median pre- and post-treatment CA-125 levels for HGSC patients were 767.5 and 35.6 (96% decrease) (p < 0.001). For LGSC patients, 4/36 (11%) had partial response (PR), 30/36 (83%) had stable disease (SD), and 2/36 (6%) had progressive disease (PD). In HGSC patients, 27/36 (75%) had PR, and 9/36 (25%) SD. Median PFS for LGSC patients was 18.5 months and median OS was 47.4 months.

CONCLUSIONS

This study provides further evidence of relative chemoresistance of LGSC in patients treated with NACT.

PTEN-deficient prostate cancer is associated with an immunosuppressive tumor microenvironment mediated by increased expression of IDO1 and infiltrating FoxP3+ T regulatory cells

BACKGROUND

Accumulating evidence shows that tumor cell-specific genomic changes can influence the cross talk between cancer cells and the surrounding tumor microenvironment (TME). Loss of the PTEN tumor suppressor gene is observed in 20% to 30% of prostate cancers (PCa) when first detected and the rate increases with PCa progression and advanced disease. Recent findings implicate a role for PTEN in cellular type I interferon response and immunosuppression in PCa. However, the way that PTEN inactivation alters antitumor immune response in PCa is poorly understood.

MATERIALS AND METHODS

To investigate the changes associated with PTEN loss and an immunosuppressive TME in PCa, we used CIBERSORT to estimate the relative abundance of 22 immune-cell types from 741 primary and 96 metastatic tumors. Our in silico findings were then validated by immunohistochemical analysis of immune cells and IDO1 and PDL1 checkpoint proteins in a cohort of 94 radical prostatectomy specimens.

RESULTS

FoxP3+ T regulatory cells (Tregs) were significantly increased in PTEN-deficient PCa in all three public domain cohorts. Loss of PTEN in bone metastases was associated with lower CD8+ T-cell abundance, but in liver metastasis, FoxP3+ Tregs were present at higher levels. PTEN-deficient lymph node metastasis had a distinct profile, with high levels of CD8+ T cells. Moreover, we found that metastatic PCa presents higher abundance of FoxP3+ Treg when compared to primary lesions. Since PTEN-deficient tumors are likely to be immunosuppressed as a consequence of increased FoxP3+ Tregs, we then evaluated the localization and expression of IDO1, PDL1 immune checkpoints, and the corresponding density of FoxP3+ Treg and CD8+ T cells using our validation cohort (n = 94). We found that IDO1 protein expression and FoxP3+ Treg density were higher in neoplastic glands compared with benign adjacent tissue. Moreover, higher densities of FoxP3+ Treg cells in both stromal (P = 0.04) and tumor (P = 0.006) compartments were observed in PTEN-deficient tumors compared to tumors that retained PTEN activity. Similarly, IDO1 protein expression was significantly increased in the tumor glands of PTEN-deficient PCa (P < 0.0001). Spearman correlation analysis showed that IDO1 expression was significantly associated with FoxP3+ Treg and CD8+ T-cell density (P < 0.01).

CONCLUSIONS

Our findings imply that PTEN deficiency is linked to an immunosuppressive state in PCa with distinct changes in the frequency of immune cell types in tumors from different metastatic sites. Our data suggest that determining PTEN status may also help guide the selection of patients for future immunotherapy trials in localized and metastatic PCa.

Carnosic acid alleviates brain injury through NF‑κB‑regulated inflammation and Caspase‑3‑associated apoptosis in high fat‑induced mouse models

High fat diet (HFD) is a risk factor for various diseases in humans and animals. Metabolic disease-induced brain injury is becoming an increasingly popular research topic. Carnosic acid (CA) is a phenolic diterpene synthesized by plants belonging to the Lamiaceae family, which exhibits multiple biological activities. In the present study, a mouse model of HFD-induced metabolic syndrome was generated. The body weight, liver weight, daily food intake, daily caloric intake, serum TG, serum TC, serum insulin and serum glucose of animals treated with CA were recorded. Additionally, the gene and protein expression levels of inflammatory cytokines, NF-κB signaling componnts, and caspase-3 were evaluated in the various CA treatment groups via immunohistochemical analysis, western blotting, reverse transcription-quantitative PCR. CA treatment significantly decreased HFD-induced metabolic syndrome by decreasing the serum levels of triglycerides, total cholesterol, insulin and glucose. Furthermore, CA served a protective role against brain injury by inhibiting the inflammatory response. CA significantly decreased the protein expression levels of various pro-inflammatory cytokines in serum and brain tissues, including interleukin (IL)-1β, IL-6 and tumor necrosis factor-α, regulated by the NF-κB signaling pathway. In addition, CA was revealed to promote the expression levels of anti-apoptotic Bcl-2, and to decrease the expression levels of pro-apoptotic Bax and matrix metallopeptidase 9. The present results suggested that CA was able to alleviate brain injury by modulating the inflammatory response and the apoptotic pathway. Administration of CA may represent a novel therapeutic strategy to treat metabolic disease-induced brain injury in the future.