Phenotype prediction from single-cell RNA-seq data using attention-based neural networks.

MOTIVATION: A patient's disease phenotype can be driven and determined by specific groups of cells whose marker genes are either unknown or can only be detected at late-stage using conventional bulk assays such as RNA-Seq technology. Recent advances in single-cell RNA sequencing (scRNA-seq) enable gene expression profiling in cell-level resolution, and therefore have the potential to identify those cells driving the disease phenotype even while the number of these cells is small. However, most existing methods rely heavily on accurate cell type detection, and the number of available annotated samples is usually too small for training deep learning predictive models. RESULTS: Here, we propose the method ScRAT for phenotype prediction using scRNA-seq data. To train ScRAT with a limited number of samples of different phenotypes, such as coronavirus disease (COVID) and non-COVID, ScRAT first applies a mixup module to increase the number of training samples. A multi-head attention mechanism is employed to learn the most informative cells for each phenotype without relying on a given cell type annotation. Using three public COVID datasets, we show that ScRAT outperforms other phenotype prediction methods. The performance edge of ScRAT over its competitors increases as the number of training samples decreases, indicating the efficacy of our sample mixup. Critical cell types detected based on high-attention cells also support novel findings in the original papers and the recent literature. This suggests that ScRAT overcomes the challenge of missing marker genes and limited sample number with great potential revealing novel molecular mechanisms and/or therapies. AVAILABILITY AND IMPLEMENTATION: The code of our proposed method ScRAT is published at https://github.com/yuzhenmao/ScRAT.

Establishment and validation of preclinical models of SMARCA4-inactivated and ARID1A/ARID1B co-inactivated dedifferentiated endometrial carcinoma.

OBJECTIVE: Dedifferentiated endometrial cancer (DDEC) is an uncommon and clinically highly aggressive subtype of endometrial cancer characterized by genomic inactivation of SWItch/Sucrose Non-Fermentable (SWI/SNF) complex protein. It responds poorly to conventional systemic treatment and its rapidly progressive clinical course limits the therapeutic windows to trial additional lines of therapies. This underscores a pressing need for biologically accurate preclinical tumor models to accelerate therapeutic development. METHODS: DDEC tumor from surgical samples were implanted into immunocompromised mice for patient-derived xenograft (PDX) and cell line development. The histologic, immunophenotypic, genetic and epigenetic features of the patient tumors and the established PDX models were characterized. The SMARCA4-deficienct DDEC model was evaluated for its sensitivity toward a KDM6A/B inhibitor (GSK-J4) that was previously reported to be effective therapy for other SMARCA4-deficient cancer types. RESULTS: All three DDEC models exhibited rapid growth in vitro and in vivo, with two PDX models showing spontaneous development of metastases in vivo. The PDX tumors maintained the same undifferentiated histology and immunophenotype, and exhibited identical genomic and methylation profiles as seen in the respective parental tumors, including a mismatch repair (MMR)-deficient DDEC with genomic inactivation of SMARCA4, and two MMR-deficient DDECs with genomic inactivation of both ARID1A and ARID1B. Although the SMARCA4-deficient cell line showed low micromolecular sensitivity to GSK-J4, no significant tumor growth inhibition was observed in the corresponding PDX model. CONCLUSIONS: These established patient tumor-derived models accurately depict DDEC and represent valuable preclinical tools to gain therapeutic insights into this aggressive tumor type.

NOTCH Signaling Limits the Response of Low-Grade Serous Ovarian Cancers to MEK Inhibition.

Low-grade serous ovarian cancer (LGSOC) is a rare subtype of epithelial ovarian cancer with high fatality rates in advanced stages due to its chemoresistant properties. LGSOC is characterized by activation of MAPK signaling, and recent clinical trials indicate that the MEK inhibitor (MEKi) trametinib may be a good treatment option for a subset of patients. Understanding MEKi-resistance mechanisms and subsequent identification of rational drug combinations to suppress resistance may greatly improve LGSOC treatment strategies. Both gain-of-function and loss-of-function CRISPR-Cas9 genome-wide libraries were used to screen LGSOC cell lines to identify genes that modulate the response to MEKi. Overexpression of MAML2 and loss of MAP3K1 were identified, both leading to overexpression of the NOTCH target HES1, which has a causal role in this process as its knockdown reversed MEKi resistance. Interestingly, increased HES1 expression was also observed in selected spontaneous trametinib-resistant clones, next to activating MAP2K1 (MEK1) mutations. Subsequent trametinib synthetic lethality screens identified SHOC2 downregulation as being synthetic lethal with MEKis. Targeting SHOC2 with pan-RAF inhibitors (pan-RAFis) in combination with MEKi was effective in parental LGSOC cell lines, in MEKi-resistant derivatives, in primary ascites cultures from patients with LGSOC, and in LGSOC (cell line-derived and patient-derived) xenograft mouse models. We found that the combination of pan-RAFi with MEKi downregulated HES1 levels in trametinib-resistant cells, providing an explanation for the synergy that was observed. Combining MEKis with pan-RAFis may provide a promising treatment strategy for patients with LGSOC, which warrants further clinical validation.

Framework of Intrinsic Immune Landscape of Dormant Prostate Cancer.

Androgen deprivation therapy (ADT) is the standard therapy for men with advanced prostate cancer (PCa). PCa often responds to ADT and enters a dormancy period, which can be recognized clinically as a minimal residual disease. However, the majority of these patients will eventually experience a relapse in the form of castration-resistant PCa with poor survival. Therefore, ADT-induced dormancy is a unique time window for treatment that can provide a cure. The study of this well-recognized phase of prostate cancer progression is largely hindered by the scarcity of appropriate clinical tissue and clinically relevant preclinical models. Here, we report the utility of unique and clinically relevant patient-derived xenograft models in the study of the intrinsic immune landscape of dormant PCa. Using data from RNA sequencing, we have reconstructed the immune evasion mechanisms that can be utilized by dormant PCa cells. Since dormant PCa cells need to evade the host immune surveillance for survival, our results provide a framework for further study and for devising immunomodulatory mechanisms that can eliminate dormant PCa cells.

Modeling Androgen Deprivation Therapy-Induced Prostate Cancer Dormancy and Its Clinical Implications.

Treatment-induced tumor dormancy is a state in cancer progression where residual disease is present but remains asymptomatic. Dormant cancer cells are treatment-resistant and responsible for cancer recurrence and metastasis. Prostate cancer treated with androgen-deprivation therapy (ADT) often enters a dormant state. ADT-induced prostate cancer dormancy remains poorly understood due to the challenge in acquiring clinical dormant prostate cancer cells and the lack of representative models. In this study, we aimed to develop clinically relevant models for studying ADT-induced prostate cancer dormancy. Dormant prostate cancer models were established by castrating mice bearing patient-derived xenografts (PDX) of hormonal naïve or sensitive prostate cancer. Dormancy status and tumor relapse were monitored and evaluated. Paired pre- and postcastration (dormant) PDX tissues were subjected to morphologic and transcriptome profiling analyses. As a result, we established eleven ADT-induced dormant prostate cancer models that closely mimicked the clinical courses of ADT-treated prostate cancer. We identified two ADT-induced dormancy subtypes that differed in morphology, gene expression, and relapse rates. We discovered transcriptomic differences in precastration PDXs that predisposed the dormancy response to ADT. We further developed a dormancy subtype-based, predisposed gene signature that was significantly associated with ADT response in hormonal naïve prostate cancer and clinical outcome in castration-resistant prostate cancer treated with ADT or androgen-receptor pathway inhibitors. IMPLICATIONS: We have established highly clinically relevant PDXs of ADT-induced dormant prostate cancer and identified two dormancy subtypes, leading to the development of a novel predicative gene signature that allows robust risk stratification of patients with prostate cancer to ADT or androgen-receptor pathway inhibitors.

Multiomics Characterization of Low-Grade Serous Ovarian Carcinoma Identifies Potential Biomarkers of MEK Inhibitor Sensitivity and Therapeutic Vulnerability.

Low-grade serous ovarian carcinoma (LGSOC) is a rare tumor subtype with high case fatality rates in patients with metastatic disease. There is a pressing need to develop effective treatments using newly available preclinical models for therapeutic discovery and drug evaluation. Here, we use multiomics integration of whole-exome sequencing, RNA sequencing, and mass spectrometry-based proteomics on 14 LGSOC cell lines to elucidate novel biomarkers and therapeutic vulnerabilities. Comparison of LGSOC cell line data with LGSOC tumor data enabled predictive biomarker identification of MEK inhibitor (MEKi) efficacy, with KRAS mutations found exclusively in MEKi-sensitive cell lines and NRAS mutations found mostly in MEKi-resistant cell lines. Distinct patterns of Catalogue of Somatic Mutations in Cancer mutational signatures were identified in MEKi-sensitive and MEKi-resistant cell lines. Deletions of CDKN2A/B and MTAP genes were more frequent in cell lines than tumor samples and possibly represent key driver events in the absence of KRAS/NRAS/BRAF mutations. These LGSOC cell lines were representative models of the molecular aberrations found in LGSOC tumors. For prediction of in vitro MEKi efficacy, proteomic data provided better discrimination than gene expression data. Condensin, minichromosome maintenance, and replication factor C protein complexes were identified as potential treatment targets in MEKi-resistant cell lines. This study suggests that CDKN2A/B or MTAP deficiency may be exploited using synthetically lethal treatment strategies, highlighting the importance of using proteomic data as a tool for molecular drug prediction. Multiomics approaches are crucial to improving our understanding of the molecular underpinnings of LGSOC and applying this information to develop new therapies. SIGNIFICANCE: These findings highlight the utility of global multiomics to characterize LGSOC cell lines as research models, to determine biomarkers of MEKi resistance, and to identify potential novel therapeutic targets.

Nontransformed and Cancer Cells Can Utilize Different Endocytic Pathways To Internalize Dendritic Nanoparticle Variants: Implications on Nanocarrier Design.

Three hyperbranched polyglycerol nanoparticle (HPG NP) variants were synthesized and fluorescently labeled for the study of their cellular interactions. The polymeric nanoparticle that contains a hydrophobic core and a hydrophilic HPG shell, HPG-C10-HPG, is taken up faster by HT-29 cancer cells than nontransformed cells, while similar uptake rates are observed with both cell types for the nanoparticle HPG-C10-PEG that contains a hydrophobic core and a polyethylene glycol shell. The nanoparticle HPG-104, containing neither the hydrophobic core nor the polyethylene glycol shell, is taken up faster by nontransformed cells than HT-29 cells. Importantly, cancer and normal cells can utilize different endocytic mechanisms for the internalization of these HPG NPs. Both HPG-C10-HPG and HPG-C10-PEG are taken up by HT-29 cells through clathrin-mediated endocytosis and macropinocytosis. Nontransformed cells, however, take up HPG-C10-HPG and HPG-104 through macropinocytosis, while these cells utilize both clathrin-mediated endocytosis and macropinocytosis to internalize HPG-C10-PEG.

Design Considerations for Developing Hyperbranched Polyglycerol Nanoparticles as Systemic Drug Carriers.

PEGylation is commonly used to increase the plasma residence time of anticancer drug nanocarriers. However, PEGylation may trigger antibody production and lead to accelerated blood clearance in subsequent administrations. Moreover, the presence of PEG shells on nanocarriers may also hamper endosomal escape and decrease drug payload release. To avoid these shortcomings, we synthesized and evaluated a non-PEGylated, hyperbranched polyglycerol nanoparticle (HPG NP) with a hydrophobic core and a hydrophilic HPG shell, HPG-C10-HPG, as a candidate for systemic delivery of anticancer drug. In vitro studies with primary human cell lines revealed that HPG-C10-HPG possesses low cytotoxicity. The presence of long chain alkyl groups (C1o) in the core as the hydrophobic pocket in the NP enabled the binding and sustained release of the hydrophobic drug docetaxel. Remarkably, the docetaxel-loaded HPG-C10-HPG formulation also confers preferential protection to primary cells, when compared to cancer cells, potentially widening the therapeutic index. HPG-C10-HPG, however, accumulated at higher levels in the liver and spleen when administered intravenously in mice. Comparing the biodistribution patterns of HPG-C10-HPG, PEGylated HPG-C10-PEG, and unmodified HPG in a xenograft model reveals that the accumulation pattern of HPG-C10-HPG was attributed to insufficient shielding of the hydrophobic groups by the HPG shell. Our results revealed the influence of the nature of the hydrophilic shell and the presence of hydrophobic groups on the tumor-to-tissue accumulation specificities of these HPG NP variants. Therefore, the present study provides insights into the structural considerations of future HPG NP designs for systemic drug delivery.

Investigation of hydrophobically derivatized hyperbranched polyglycerol with PEGylated shell as a nanocarrier for systemic delivery of chemotherapeutics.

We report the synthesis and characterization of a polymeric nanoparticle (NP) based on hyperbranched polyglycerol (HPG) containing a hydrophobic core and a hydrophilic shell, and assessed its suitability to be developed as a systemic anticancer drug carrier. HPG NP displayed low toxicity to primary cell cultures and were well-tolerated in mice after intravenous administration. When tested in mice tumor xenograft models, HPG NP accumulated significantly in the tumors with low accumulation in the liver and the spleen. In vitro studies demonstrated that HPG NP was capable of hydrophobically binding docetaxel and releasing it in a controlled manner. The HPG NP formulation of docetaxel conferred a preferential protective effect on primary non-cancerous cells while effectively killing cancer cells, indicating great potential for widening its therapeutic index. Taken together, these data indicate that HPG NP is a highly promising nanocarrier platform for systemic delivery of anticancer drugs. FROM THE CLINICAL EDITOR: The use of polyethylene glycol on nano-carriers as "stealth" to deliver intravenous drugs is well known. Here, the authors developed polymeric nanoparticle (NP) with hyperbranched polyglycerol (HPG) and tested its efficacy in delivering docetaxel. The results showed that this formulation could preferentially killed cancer cells with a high therapeutic index. It seems that this platform could have a great potential in cancer therapy.

Heterogeneity of breast cancer stem cells as evidenced with Notch-dependent and Notch-independent populations.

Studies have suggested the potential importance of Notch signaling to the cancer stem cell population in some tumors, but it is not known whether all cells in the cancer stem cell fraction require Notch activity. To address this issue, we blocked Notch activity in MCF-7 cells by expressing a dominant-negative MAML-GFP (dnMAML) construct, which inhibits signaling through all Notch receptors, and quantified the effect on tumor-initiating activity. Inhibition of Notch signaling reduced primary tumor sphere formation and side population. Functional quantification of tumor-initiating cell numbers in vivo showed a significant decrease, but not a complete abrogation, of these cells in dnMAML-expressing cells. Interestingly, when assessed in secondary assays in vitro or in vivo, there was no difference in tumor-initiating activity between the dnMAML-expressing cells and control cells. The fact that a subpopulation of dnMAML-expressing cells was capable of forming primary and secondary tumors indicates that there are Notch-independent tumor-initiating cells in the breast cancer cell line MCF-7. Our findings thus provide direct evidence for a heterogeneous cancer stem cell pool, which will require combination therapies against multiple oncogenic pathways to eliminate the tumor-initiating cell population.

CD44-mediated elongated T cell spreading requires Pyk2 activation by Src family kinases, extracellular calcium, phospholipase C and phosphatidylinositol-3 kinase.

The proline-rich tyrosine kinase 2, Pyk2, is a focal adhesion related kinase expressed in T cells that is tyrosine phosphorylated and activated by integrin, chemokine or T cell receptor stimulation. Ligation of the cell adhesion molecule CD44 also induces Pyk2 phosphorylation and T cell spreading, and this is negatively regulated by the protein tyrosine phosphatase CD45. Here, we identify the activation requirements for Pyk2 and demonstrate its requirement for CD44-mediated elongated T cell spreading. Upon CD44-mediated cell spreading, Pyk2 was recruited to CD44 clusters in both CD45(+) and CD45(-) T cells, yet was more strongly phosphorylated in T cells lacking CD45. In these cells, Pyk2 phosphorylation was dependent on Src family kinase activity and required actin polymerisation, phosphatidylinositol-3 kinase and phospholipase C activity as well as extracellular calcium. Inhibition of any of these events prevented Pyk2 phosphorylation and T cell spreading. Transfection of a truncated form of Pyk2 lacking the kinase domain, PRNK, inhibited CD44-mediated cell spreading, demonstrating an important role for Pyk2. However, inhibition of microtubule turnover by Taxol prevented elongated T cell spreading but did not affect Pyk2 phosphorylation, indicating that microtubule reorganisation is downstream, or independent, of Pyk2 phosphorylation. Together this demonstrates that multiple factors are required for CD44-induced Pyk2 activation, which plays a critical role in CD44-mediated elongated T cell spreading.

CD45 down-regulates Lck-mediated CD44 signaling and modulates actin rearrangement in T cells.

The tyrosine phosphatase CD45 dephosphorylates the negative regulatory tyrosine of the Src family kinase Lck and plays a positive role in TCR signaling. In this study we demonstrate a negative regulatory role for CD45 in CD44 signaling leading to actin rearrangement and cell spreading in activated thymocytes and T cells. In BW5147 T cells, CD44 ligation led to CD44 and Lck clustering, which generated a reduced tyrosine phosphorylation signal in CD45(+) T cells and a more sustained, robust tyrosine phosphorylation signal in CD45(-) T cells. This signal resulted in F-actin ring formation and round spreading in the CD45(+) cells and polarized, elongated cell spreading in CD45(-) cells. The enhanced signal in the CD45(-) cells was consistent with enhanced Lck Y394 phosphorylation compared with the CD45(+) cells where CD45 was recruited to the CD44 clusters. This enhanced Src family kinase-dependent activity in the CD45(-) cells led to PI3K and phospholipase C activation, both of which were required for elongated cell spreading. We conclude that CD45 induces the dephosphorylation of Lck at Y394, thereby preventing sustained Lck activation and propose that the amplitude of the Src family kinase-dependent signal regulates the outcome of CD44-mediated signaling to the actin cytoskeleton and T cell spreading.

Differential function of PTPalpha and PTPalpha Y789F in T cells and regulation of PTPalpha phosphorylation at Tyr-789 by CD45.

CD45 is a major membrane protein tyrosine phosphatase (PTP) expressed in T cells where it regulates the activity of Lck, a Src family kinase important for T cell receptor-mediated activation. PTPalpha is a more widely expressed transmembrane PTP that has been shown to regulate the Src family kinases, Src and Fyn, and is also present in T cells. Here, PTPalpha was phosphorylated at Tyr-789 in CD45(-) T cells but not in CD45(+) T cells suggesting that CD45 could regulate the phosphorylation of PTPalpha at this site. Furthermore, CD45 could directly dephosphorylate PTPalpha in vitro. Expression of PTPalpha and PTPalpha-Y789F in T cells revealed that the mutant had a reduced ability to decrease Fyn and Cbp phosphorylation, to regulate the kinase activity of Fyn, and to restore T cell receptor-induced signaling events when compared with PTPalpha. Conversely, this mutant had an increased ability to prevent Pyk2 phosphorylation and CD44-mediated cell spreading when compared with PTPalpha. These data demonstrate distinct activities of PTPalpha and PTPalpha-Y789F in T cells and identify CD45 as a regulator of PTPalpha phosphorylation at tyrosine 789 in T cells.