Detecting subtle transcriptomic perturbations induced by lncRNAs knock-down in single-cell CRISPRi screening using a new sparse supervised autoencoder neural network.

Single-cell CRISPR-based transcriptome screens are potent genetic tools for concomitantly assessing the expression profiles of cells targeted by a set of guides RNA (gRNA), and inferring target gene functions from the observed perturbations. However, due to various limitations, this approach lacks sensitivity in detecting weak perturbations and is essentially reliable when studying master regulators such as transcription factors. To overcome the challenge of detecting subtle gRNA induced transcriptomic perturbations and classifying the most responsive cells, we developed a new supervised autoencoder neural network method. Our Sparse supervised autoencoder (SSAE) neural network provides selection of both relevant features (genes) and actual perturbed cells. We applied this method on an in-house single-cell CRISPR-interference-based (CRISPRi) transcriptome screening (CROP-Seq) focusing on a subset of long non-coding RNAs (lncRNAs) regulated by hypoxia, a condition that promote tumor aggressiveness and drug resistance, in the context of lung adenocarcinoma (LUAD). The CROP-seq library of validated gRNA against a subset of lncRNAs and, as positive controls, HIF1A and HIF2A, the 2 main transcription factors of the hypoxic response, was transduced in A549 LUAD cells cultured in normoxia or exposed to hypoxic conditions during 3, 6 or 24 h. We first validated the SSAE approach on HIF1A and HIF2 by confirming the specific effect of their knock-down during the temporal switch of the hypoxic response. Next, the SSAE method was able to detect stable short hypoxia-dependent transcriptomic signatures induced by the knock-down of some lncRNAs candidates, outperforming previously published machine learning approaches. This proof of concept demonstrates the relevance of the SSAE approach for deciphering weak perturbations in single-cell transcriptomic data readout as part of CRISPR-based screening.

Spatial Distribution of Non-Immune Cells Expressing Glycoprotein A Repetitions Predominant in Human and Murine Metastatic Lymph Nodes.

Several types of cancer spread through the lymphatic system via the sentinel lymph nodes (LNs). Such LN-draining primary tumors, modified by tumor factors, lead to the formation of a metastatic niche associated with an increased number of Foxp3+ regulatory T cells (Tregs). These cells are expected to contribute to the elaboration of an immune-suppressive environment. Activated Tregs express glycoprotein A repetitions predominant (GARP), which binds and presents latent transforming growth factor beta 1 (TGF-ß1) at their surface. GARP is also expressed by other non-immune cell types poorly described in LNs. Here, we mapped GARP expression in non-immune cells in human and mouse metastatic LNs. The mining of available (human and murine) scRNA-Seq datasets revealed GARP expression by blood (BEC)/lymphatic (LEC) endothelial, fibroblastic, and perivascular cells. Consistently, through immunostaining and in situ RNA hybridization approaches, GARP was detected in and around blood and lymphatic vessels, in (αSMA+) fibroblasts, and in perivascular cells associated with an abundant matrix. Strikingly, GARP was detected in LECs forming the subcapsular sinus and high endothelial venules (HEVs), two vascular structures localized at the interface between LNs and the afferent lymphatic and blood vessels. Altogether, we here provide the first distribution maps for GARP in human and murine LNs.

Identification of oncolytic vaccinia restriction factors in canine high-grade mammary tumor cells using single-cell transcriptomics.

Mammary carcinoma, including triple-negative breast carcinomas (TNBC) are tumor-types for which human and canine pathologies are closely related at the molecular level. The efficacy of an oncolytic vaccinia virus (VV) was compared in low-passage primary carcinoma cells from TNBC versus non-TNBC. Non-TNBC cells were 28 fold more sensitive to VV than TNBC cells in which VV replication is impaired. Single-cell RNA-seq performed on two different TNBC cell samples, infected or not with VV, highlighted three distinct populations: naïve cells, bystander cells, defined as cells exposed to the virus but not infected and infected cells. The transcriptomes of these three populations showed striking variations in the modulation of pathways regulated by cytokines and growth factors. We hypothesized that the pool of genes expressed in the bystander populations was enriched in antiviral genes. Bioinformatic analysis suggested that the reduced activity of the virus was associated with a higher mesenchymal status of the cells. In addition, we demonstrated experimentally that high expression of one gene, DDIT4, is detrimental to VV production. Considering that DDIT4 is associated with a poor prognosis in various cancers including TNBC, our data highlight DDIT4 as a candidate resistance marker for oncolytic poxvirus therapy. This information could be used to design new generations of oncolytic poxviruses. Beyond the field of gene therapy, this study demonstrates that single-cell transcriptomics can be used to identify cellular factors influencing viral replication.