bioRxiv Science⌕ Search

Biology subjects

Dorward, D. A.

Publications and source records attributed to Dorward, D. A..

5 recordsLinked to original sources

Beyond Navigation - Tissue-contacting Fluorescent Lifetime Imaging reveals a pathology-linked lung cancer phenotype at the point of biopsy

Background: Accurate sampling of suspected peripheral lung cancers depends on access to the lesion and confirmation that the biopsy tool is in contact with target tissue. Current bronchoscopic navigation and imaging techniques can guide instruments to a target but do not provide real-time biological confirmation at the point of sampling. Fluorescence lifetime imaging microscopy (FLIM) provides molecular contrast by measuring fluorescence decay - how long photons continue to be emitted from fluorescent molecules. In the Precision Lung clinical study (ISRCTN15093468), the Prothea Imaging System (Generation 1) identified a candidate tumour-associated phenotype of spatially overlapped low fluorescence lifetime and low intensity (LLLI) from in-vivo imaging. We used this observation as the basis for a reverse-translational study to determine whether the LLLI phenotype is linked to cancer pathology; reproducible with the Imaging System (Generation 2); and distinguishable from normal lung tissue. Methods: Previously reported Precision Lung findings were used as the clinical starting observation and were not re-analysed. Validation was then performed using: (i) pathology linked benchtop FLIM of early-stage non-small-cell lung tissue microarrays encompassing malignant cell clusters of approximately 300 um2, matched to the EoT imaging scale; (ii) five sequential fresh lung-cancer resections imaged at tumour and comparator regions, including visibly blood-rich contact sites, using the (Generation 2) Imaging System; and (iii) systematic mapping of two ventilated non-cancer donor lungs, one from a smoker and one from a non-smoker, across all available lobes. The LLLI phenotype was defined as spatial co-localisation of low intensity and short lifetime. Results: Using a real time fibre based FLIM system, capable of deployment through a working channel of a bronchoscope, the LLLI tumour phenotype was optically identified in freshly resected tumour tissue. The same phenotype was identified in fixed tissue samples with known pathology, and with images taken in the Precision Lung clinical study. Whole human lung controls did not show evidence of the tumour phenotype. Conclusions: This evidence forms a reverse-translational chain that supports the concept of the Prothea Imaging System - as a platform that confirms that the tool is in contact with a region of cancer in the lesion, while preserving continuous access for biopsy or intervention.

cancer biology↗

Self-supervised AI reveals a lethal discohesive phenotype in lung adenocarcinoma

Applications of artificial intelligence (AI) to histopathology are now common, but most require supervision which inherently limits their scope. By using self-supervised learning (SSL), we discover and quantify the full range of histopathological appearances in a disease, and associate them with clinicopathological ground truths such as prognosis. We used this approach to discover under-appreciated morphologies of lung adenocarcinoma (LUAD), using a highly characterised resected tumour cohort of over 4000 slides from over 1000 patients. By constructing an authoritative lexicon of recurrent LUAD appearances, we ab initio discovered several stromal morphologies strongly predictive of outcome. With multimodal data integration and external dataset validation, we propose that epithelial discohesion is lethal, but only in the context of immunologically cold stroma. Both these morphological features are independent of current prognostic schema. Crucially, we describe these features in the context of real-world diagnostic histopathology, giving them immediate clinical translatability. Statement of SignificanceHistopathology relies heavily on epithelial morphology, often neglecting the stroma. We use self-supervised AI to identify under-appreciated morphologies in LUAD linked to poor outcome and validate these observations in an external cohort, demonstrating the utility of self-supervised AI as a powerful biological discovery tool.

cancer biology↗

Subpopulations of cancer-associated fibroblasts expressing fibroblast activation protein and podoplanin in non-small cell lung cancer are a predictor of poor clinical outcome

Cancer-associated fibroblasts (CAFs) are the dominant cell type in the stroma of solid organ cancers, including non-small cell lung cancer (NSCLC). Fibroblast heterogeneity is widely recognised in many cancers, with subpopulations of CAFs being identified and potentially being indicative of prognosis and treatment efficacy. Here, the subtypes displayed by CAFs isolated from human NSCLC resections are initially identified by flow cytometry, using the markers FAP, CD29, SMA, PDPN, CD90, FSP-1 and PDGFR{beta}, showing five distinct subpopulations, CAF-S1-S5. Our findings show that when comparing fibroblasts from tumour tissue with that from adjacent lung tissue, CAF-S2 and CAF-S3 are found in the normal tissue and marker expression suggests a less activated phenotype whereas CAF-S1, CAF-S4 and CAF-S5 are predominantly found in the tumour tissue and are positive for a combination of markers of fibroblast activation. We focus on these subtypes most associated with fibroblast activation, primarily focussing on a previously unreported CAF-S5 subtype, and comparing to the previously identified CAF-S1. Both these subsets express FAP and PDPN as markers of fibroblast activation, but CAF-S5 lacks expression of the common activation marker SMA. The spatial relevance of these subtypes in a cohort of 163 NSCLC patients was then investigated by multiplex immunofluorescence on a tumour micro-array of patient samples, revealing CAF-S5 are found further from tumour regions than CAF-S1. To understand the functional role of CAF-S5, scRNA sequencing data was used to compare the subset to the previously identified CAF-S1, finding that CAF-S5 displays an inflammatory phenotype, whereas CAF-S1 displays a contractile phenotype. We demonstrate that presence of either the CAF-S1 or CAF-S5 subtype is correlated to worse survival outcome in NSCLC, highlighting the importance of the identification of CAF subtypes in NSCLC.

cancer biology↗

Location of CD39+ T cell sub-populations within tumours predict differential outcomes in non-small cell lung cancer.

An improved mechanistic understanding of immunosuppressive pathways in NSCLC is important to develop novel diagnostic and therapeutic approaches. Here, we reveal that the prognostic significance of the rate limiting ectonucleotidases in adenosine production CD39 and CD73 requires knowledge of cell type specific expression and localisation within tumours. In a cohort of early treatment naive NSCLC patients, high stromal expression of CD39 and CD73 predicts poor outcome. CD39 expression amongst T cells identifies CD39+CD4+ Tregs which predict poor outcome and CD39+CD103+ CTL which confer a survival benefit if high densities are observed inside of the tumour nest. Bulk RNA Seq shows that the TME of NSCLC upregulates regulatory pathways in CD4+ T cells and exhaustion in CD8+ T cells. Analysis of single-cell RNASeq datasets illustrates that CD39+CD4+ Tregs are enriched in Treg signature gene sets, and CD39+CD103+ CTL show gene signatures indicative of an exhausted cytotoxic phenotype with an upregulated expression of CXCL13. Combined knowledge of patterns of distribution and location are required to understand the prognostic impact of CD39+ T cell populations in NSCLC. This study provides an improved understanding of the spatial and functional characteristics of CD39+ T cells and illustrates their significance to patient outcome.

cancer biology↗

Quantitative proteomics identifies tumour matrisome signatures in patients with non-small cell lung cancer

The composition and remodelling of the extracellular matrix (ECM) are important factors in the development and progression of cancers, and the ECM is implicated in promoting tumour growth and restricting anti-tumour therapies through multiple mechanisms. The characterisation of differences in ECM composition between normal and diseased tissues may aid in identifying novel diagnostic markers, prognostic indicators and therapeutic targets for drug development. Using tissue from non-small cell lung cancer (NSCLC) patients undergoing curative intent surgery, we characterised quantitative tumour-specific ECM proteome signatures by mass spectrometry, identifying 161 matrisome proteins differentially regulated between tumour tissue and nearby non-malignant lung tissue. We defined a collagen hydroxylation functional protein network that is enriched in the lung tumour microenvironment. We validated two novel putative extracellular markers of NSCLC, the collagen cross-linking enzyme peroxidasin and a disintegrin and metalloproteinase with thrombospondin motifs 16 (ADAMTS16), for discrimination of malignant and non-malignant lung tissue. These proteins were up-regulated in lung tumour samples, and high PXDN and ADAMTS16 gene expression was associated with shorter survival of lung adenocarcinoma and squamous cell carcinoma patients, respectively. These data reveal tumour matrisome signatures in human NSCLC. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=19 SRC="FIGDIR/small/510064v1_ufig1.gif" ALT="Figure 1"> View larger version (8K): org.highwire.dtl.DTLVardef@b34045org.highwire.dtl.DTLVardef@40a5a9org.highwire.dtl.DTLVardef@3c150corg.highwire.dtl.DTLVardef@89c6f2_HPS_FORMAT_FIGEXP M_FIG C_FIG

cancer biology↗