Background: Molecular signatures identified from high-throughput transcriptomic studies often have poor reliability and fail to reproduce across studies. One solution is to combine independent studies into a single integrative analysis, additionally increasing sample size. However, the different protocols and technological platforms across transcriptomic studies produce unwanted systematic variation that strongly confounds the integrative analysis results. When studies aim to discriminate an outcome of interest, the common approach is a sequential two-step procedure; unwanted systematic variation removal techniques are applied prior to classification methods. Results: To limit the risk of overfitting and over-optimistic results of a two-step procedure, we developed a novel multivariate integration method, MINT, that simultaneously accounts for unwanted systematic variation and identifies predictive gene signatures with greater reproducibility and accuracy. In two biological examples on the classification of three human cell types and four subtypes of breast cancer, we combined high-dimensional microarray and RNA-seq data sets and MINT identified highly reproducible and relevant gene signatures predictive of a given phenotype. MINT led to superior classification and prediction accuracy compared to the existing sequential two-step procedures. Conclusions: MINT is a powerful approach and the first of its kind to solve the integrative classification framework in a single step by combining multiple independent studies. MINT is computationally fast as part of the mixOmics R CRAN package, available at

Triple-negative breast cancer (TNBC) is an aggressive subtype of breast cancer that has few effective treatment options due to its lack of targetable hormone receptors. Whilst the degree of tumour infiltrating lymphocytes (TILs) has been shown to associate with therapy response and prognosis, deeper characterization of the molecular diversity that may mediate chemotherapeutic response is lacking. Here we applied targeted proteomic analysis of both chemotherapy sensitive and resistant TNBC tissue samples by the Nanostring GeoMx Digital Spatial Platform (DSP). By quantifying 68 targets in the tumour and tumour microenvironment (TME) compartments and performing differential expression analysis between responsive and non-responsive tumours, we show that increased ER-alpha expression and decreased 4-1BB and MART1 within the stromal compartments is associated with adjuvant chemotherapy response. Similarly, higher expression of GZMA, STING and fibronectin and lower levels of CD80 were associated with response within tumour compartments. Univariate overall-survival (OS) analysis of stromal proteins supported these findings, with ER-alpha expression (HR=0.19, p=0.0012) associated with better OS while MART1 expression (HR=2.3, p=0.035) was indicative of poorer OS. Proteins within tumour compartments consistent with longer OS included PD-L1 (HR=0.53, p=0.023), FOXP3 (HR=0.5, p=0.026), GITR (HR=0.51, p=0.036), SMA (HR=0.59, p=0.043), while EPCAM (HR=1.7, p=0.045), and CD95 (HR=4.9, p=0.046) expression were associated with shorter OS. Our data provides early insights into the levels of these markers in the TNBC tumour microenvironment, and their association with chemotherapeutic response and patient survival.

The composition and activation status of the cellular milieu contained within the tumour microenvironment (TME) is becoming increasingly recognized as a driving factor for immunotherapy response. Here, we employed multiplex immunohistochemistry (mIHC), and digital spatial profiling (DSP) to capture the targeted immune proteome and transcriptome of tumour and TME compartments from an immune checkpoint inhibitor (ICI)‐treated (n = 41) non‐small cell lung cancer (NSCLC) patient cohort. We demonstrate by mIHC that the interaction of CD68+ macrophages with PD1+, FoxP3+ cells is enriched in ICI refractory tumours (p = 0.012). Patients responsive to ICI therapy expressed higher levels of IL2 receptor alpha (CD25, p = 0.028) within their tumour compartments, which corresponded with increased IL2 mRNA (p = 0.001) within their stroma. In addition, stromal IL2 mRNA levels positively correlated with the expression of pro‐apoptotic markers cleaved caspase 9 (p = 2e−5) and BAD (p = 5.5e−4) and negatively with levels of memory marker, CD45RO (p = 7e−4). Immuno‐inhibitory markers CTLA‐4 (p = 0.021) and IDO‐1 (p = 0.023) were suppressed in ICI‐responsive patients. Tumour expression of CD44 was depleted in the responsive patients (p = 0.02), while higher stromal expression of one of its ligands, SPP1 (p = 0.008), was observed. Cox survival analysis also indicated tumour CD44 expression was associated with poorer prognosis (hazard ratio [HR] = 1.61, p = 0.01), consistent with its depletion in ICI‐responsive patients. Through multi‐modal approaches, we have dissected the characteristics of NSCLC immunotherapy treatment groups and provide evidence for the role of several markers including IL2, CD25, CD44 and SPP1 in the efficacy of current generations of ICI therapy.

IntroductionImmunotherapies, such as immune checkpoint inhibitors (ICI) have shown durable benefit in a subset of non-small cell lung cancer (NSCLC) patients. The mechanisms for this are not fully understood, however the composition and activation status of the cellular milieu contained within the tumour microenvironment (TME) is becomingly increasingly recognised as a driving factor in treatment-refractory disease.MethodsHere, we employed multiplex IHC (mIHC), and digital spatial profiling (DSP) to capture the targeted immune proteome and transcriptome of tumour and TME compartments of pre-treatment samples from a 2nd line NSCLC ICI-treated cohort (n=41 patients; n=25 responders, n=16 non-responders).ResultsWe demonstrate by mIHC that the interaction of CD68+ macrophages with PD1+, FoxP3+ cells is significantly enriched in ICI refractory tumours (p=0.012). Our study revealed that patients sensitive to ICI therapy expressed higher levels of IL2 receptor alpha (CD25, p=0.028) within the tumour compartments, which corresponded with the increased expression of IL2 mRNA (p=0.001) within their stroma, indicative of key conditions for ICI efficacy prior to treatment. IL2 mRNA levels within the stroma positively correlated with the expression of pro-apoptotic markers cleaved caspase 9 (p=2e-5) and BAD (p=5.5e-4) and negatively correlated with levels of memory T cells (CD45RO) (p=7e-4). Immuno-inhibitory markers CTLA-4 (p=0.021) and IDO-1 (p=0.023) were also supressed in ICI-responsive patients. Of note, tumour CD44 (p=0.02) was depleted in the response group and corresponded inversely with significantly higher stromal expression of its ligand SPP1 (osteopontin, p=0.008). Analysis of differentially expressed transcripts indicated the potential inhibition of stromal interferon-gamma (IFNγ) activity, as well as estrogen-receptor and Wnt-1 signalling activity within the tumour cells of ICI responsive patients. Cox survival analysis indicated tumour CD44 expression was associated with poorer prognosis (HR=1.61, p=0.01), consistent with its depletion in ICI sensitive patients. Similarly, stromal CTLA-4 (HR=1.78, p=0.003) and MDSC/M2 macrophage marker ARG1 (HR=2.37, p=0.01) were associated with poorer outcome while levels of apoptotic marker BAD (HR=0.5, p=0.01) appeared protective. Interestingly, stromal mRNA for E-selectin (HR=652, p=0.001), CCL17 (HR=70, p=0.006) and MTOR (HR=1065, p=0.008) were highly associated with poorer outcome, indicating pro-tumourigenic features in the tumour microenvironment that may facilitate ICI resistance.ConclusionsThrough multi-modal approaches, we have dissected the characteristics of NSCLC and provide evidence for the role of IL2 and stromal activation by osteopontin in the efficacy of current generations of ICI therapy. The enrichment of SPP1 in the stroma of ICI sensitive patients in our data is a novel finding, indicative of stromal activation that may aid immune cell survival and activity despite no clear association with increased levels of immune infiltrate.

Mesenchymal stromal cells (MSC) are widely used for the study of mesenchymal tissue repair, and increasingly adopted for cell therapy, despite the lack of consensus on the identity of these cells. In part this is due to the lack of specificity of MSC markers. Distinguishing MSC from other stromal cells such as fibroblasts is particularly difficult using standard analysis of surface proteins, and there is an urgent need for improved classification approaches. Transcriptome profiling is commonly used to describe and compare different cell types; however, efforts to identify specific markers of rare cellular subsets may be confounded by the small sample sizes of most studies. Consequently, it is difficult to derive reproducible, and therefore useful markers. We addressed the question of MSC classification with a large integrative analysis of many public MSC datasets. We derived a sparse classifier (The Rohart MSC test) that accurately distinguished MSC from non-MSC samples with >97% accuracy on an internal training set of 635 samples from 41 studies derived on 10 different microarray platforms. The classifier was validated on an external test set of 1,291 samples from 65 studies derived on 15 different platforms, with >95% accuracy. The genes that contribute to the MSC classifier formed a protein-interaction network that included known MSC markers. Further evidence of the relevance of this new MSC panel came from the high number of Mendelian disorders associated with mutations in more than 65% of the network. These result in mesenchymal defects, particularly impacting on skeletal growth and function. The Rohart MSC test is a simple in silico test that accurately discriminates MSC from fibroblasts, other adult stem/progenitor cell types or differentiated stromal cells.

Follicular Lymphoma (FL) is the most common indolent Non-Hodgkin Lymphoma. Despite generally favorable survival outcomes, 20% of FL patients experience 'Progression of Disease within 24 months' (POD24) and subsequently have poor long-term overall survival (OS) (Casulo, JCO 2015). Unfortunately, POD24 has limited clinical value because it cannot guide up-front clinical decisions. Accurate pre-therapy prognosticators are vital for clinical trial design and are also increasingly being mandated by funding agencies for stratification of patients to emerging front-line treatments. The new 'state-of-the-art' prognosticators 'm7-FLIPI' and POD24-PI' (Pastore, Lancet Oncol 2015; Jurinovic, Blood 2016) supplement clinical parameters with genetic mutational status. However, their applicability to population based cohorts including early-stage and asymptomatic patients remains unknown. Furthermore, there is significant heterogeneity of outcome within these prognostic groupings. The established biological and prognostic importance of the tumor microenvironment (TME) in FL suggests that prognosis would be enhanced by incorporating information on host immunity (Scott, Nat Rev Can 2014). Forty-five pre-treatment FL biopsies were categorized into 'hot' or 'cold' immune nodes by multiplex immunofluorescent imaging and respectively characterized by concordant high or low expression of multiple immune effector and checkpoint-associated proteins. (Fig 1A). Consistent with these findings, gene expression using the Nanostring platform showed that immune effectors (CD4/CD8/TNFa/CD137/CD56) positively correlated with immune checkpoints (PD-1/PD-L1/PD-L2/TIM3/LAG3/CD163/CD68) indicative of an adaptive immune response. Additionally, high-throughput unbiased TCRb sequencing showed the intratumoral TCR repertoire was more clonal in 'hot' compared to 'cold' FL samples (p=0.024), indicative of a skewed T-cell immune response (Fig 1B). We then applied these findings to an independent population-based cohort of 175 cases of FL from the rituximab era with long-term follow-up (median ~7 years), including advanced (n=137) and localized cases (n=38). The aims were to: a) identify new targetable immune parameters of prognostic importance in the rituximab-era; and b) compare and contrast these with published prognostic tools: FLIPI, FLIPI-2, m7-FLIPI, POD24-PI and 'immune survival score' ('ISS', Dave, NEJM 2004). OS was not only inferior in those experiencing POD24 (HR 4.88, p<0.0001, Fig 1C) but these patients had a >2-fold increase in 5-year patient health costs. Hence, POD24, as well as FFS and TT2T were therefore chosen as the primary outcome measures. M7 mutation frequencies were similar to those previously published (Pastore, Lancet Oncol 2015). However, the prognostic utility of the m7-FLIPI could not be demonstrated, whereas the FLIPI, FLIPI-2, and POD24-PI retained their prognostic value. The POD24-PI was most predictive of FFS (p<0.0001, HR=3.54) and was most specific in identifying cases that experience POD24 (Sp=68%). The prognostic utility of the TME was then tested. Notably both the ISS (p=0.024, HR=1.74) and multiple immune genes not represented in the ISS including PD-L2, TIM3, LAG3, CD137, TNF and CD4 predicted FFS. PD-L2 demonstrated the strongest association with FFS (p<0.0001, HR=3.74, Fig 1D). It not only out-performed the ISS but was independent to the FLIPI and POD-24-PI. The prognostic significance of PD-L2 was validated in an independent population based cohort of uniformly R-CVP treated patients from an in-silico dataset with gene expression quantified using the Illumina DASL platform (Pastore, Lancet Oncol 2015). We have validated the TME in predicting outcome in a population based cohort of FL patients with long-term follow-up treated in the rituximab era. Furthermore, we describe the role of PD-L2 as well as several additional pertinent, clinically-actionable markers of the TME which predict survival to conventional therapies in FL. Low expression of PD-L2 appears to be a surrogate of a broadly co-ordinated downregulation of the intratumoural response. These immune scores are independent of and additive to additive to the FLIPI and POD24-PI. Development of new prognostic models require the incorporation of host immunity along with clinico-genetic features to further improve the specificity, and to accurately risk stratify FL patients. Disclosures No relevant conflicts of interest to declare.

Available transcriptomes of the mammalian kidney provide limited information on the spatial interplay between different functional nephron structures due to the required dissociation of tissue with traditional transcriptome-based methodologies. A deeper understanding of the complexity of functional nephron structures requires a non-dissociative transcriptomics approach, such as spatial transcriptomics sequencing (ST-seq). We hypothesize that the application of ST-seq in normal mammalian kidneys will give transcriptomic insights within and across species of physiology at the functional structure level and cellular communication at the cell level. Here, we applied ST-seq in six mice and four human kidneys that were histologically absent of any overt pathology. We defined the location of specific nephron structures in the captured ST-seq datasets using three lines of evidence: pathologist's annotation, marker gene expression, and integration with public single-cell and/or single-nucleus RNA-sequencing datasets. We compared the mouse and human cortical kidney regions. In the human ST-seq datasets, we further investigated the cellular communication within glomeruli and regions of proximal tubules–peritubular capillaries by screening for co-expression of ligand–receptor gene pairs. Gene expression signatures of distinct nephron structures and microvascular regions were spatially resolved within the mouse and human ST-seq datasets. We identified 7,370 differentially expressed genes (padj < 0.05) distinguishing species, suggesting changes in energy production and metabolism in mouse cortical regions relative to human kidneys. Hundreds of potential ligand–receptor interactions were identified within glomeruli and regions of proximal tubules–peritubular capillaries, including known and novel interactions relevant to kidney physiology. Our application of ST-seq to normal human and murine kidneys confirms current knowledge and localization of transcripts within the kidney. Furthermore, the generated ST-seq datasets provide a valuable resource for the kidney community that can be used to inform future research into this complex organ.

Spatial transcriptomics (ST) measures and maps transcripts within intact tissue sections, allowing the visualization of gene activity within the spatial organization of complex biological systems. This review outlines advances in genomic sequencing technologies focusing on in situ sequencing–based ST, including applications in transplant and relevant nontransplant settings. We describe the experimental and analytical pipelines that underpin the current generation of spatial technologies. This context is important for understanding the potential role ST may play in expanding our knowledge, including in organ transplantation, and the important caveats/limitations when interpreting the vast data output generated by such methodological platforms.