Sex differences in symptomatology and immune profiles of Long COVID (preprint)

Strong sex differences in the frequencies and manifestations of Long COVID (LC) have been reported with females significantly more likely than males to present with LC after acute SARS-CoV-2 infection1-7. However, whether immunological traits underlying LC differ between sexes, and whether such differences explain the differential manifestations of LC symptomology is currently unknown. Here, we performed sex-based multi-dimensional immune-endocrine profiling of 165 individuals8 with and without LC in an exploratory, cross-sectional study to identify key immunological traits underlying biological sex differences in LC. We found that female and male participants with LC experienced different sets of symptoms, and distinct patterns of organ system involvement, with female participants suffering from a higher symptom burden. Machine learning approaches identified differential sets of immune features that characterized LC in females and males. Males with LC had decreased frequencies of monocyte and DC populations, elevated NK cells, and plasma cytokines including IL-8 and TGF-{beta}-family members. Females with LC had increased frequencies of exhausted T cells, cytokine-secreting T cells, higher antibody reactivity to latent herpes viruses including EBV, HSV-2, and CMV, and lower testosterone levels than their control female counterparts. Testosterone levels were significantly associated with lower symptom burden in LC participants over sex designation. These findings suggest distinct immunological processes of LC in females and males and illuminate the crucial role of immune-endocrine dysregulation in sex-specific pathology.

Distinguishing features of Long COVID identified through immune profiling (preprint)

SARS-CoV-2 infection can result in the development of a constellation of persistent sequelae following acute disease called post-acute sequelae of COVID-19 (PASC) or Long COVID 1–3 . Individuals diagnosed with Long COVID frequently report unremitting fatigue, post-exertional malaise, and a variety of cognitive and autonomic dysfunctions 1–3 ; however, the basic biological mechanisms responsible for these debilitating symptoms are unclear. Here, 215 individuals were included in an exploratory, cross-sectional study to perform multi-dimensional immune phenotyping in conjunction with machine learning methods to identify key immunological features distinguishing Long COVID. Marked differences were noted in specific circulating myeloid and lymphocyte populations relative to matched control groups, as well as evidence of elevated humoral responses directed against SARS-CoV-2 among participants with Long COVID. Further, unexpected increases were observed in antibody responses directed against non-SARS-CoV-2 viral pathogens, particularly Epstein-Barr virus. Analysis of circulating immune mediators and various hormones also revealed pronounced differences, with levels of cortisol being uniformly lower among participants with Long COVID relative to matched control groups. Integration of immune phenotyping data into unbiased machine learning models identified significant distinguishing features critical in accurate classification of Long COVID, with decreased levels of cortisol being the most significant individual predictor. These findings will help guide additional studies into the pathobiology of Long COVID and may aid in the future development of objective biomarkers for Long COVID.

Post-infectious inflammatory disease in MIS-C features elevated cytotoxicity signatures and autoreactivity that correlates with severity (preprint)

Multisystem inflammatory syndrome in children (MIS-C) is a life-threatening post-infectious complication occurring unpredictably weeks after mild or asymptomatic SARS-CoV2 infection in otherwise healthy children. Here, we define immune abnormalities in MIS-C compared to adult COVID-19 and pediatric/adult healthy controls using single-cell RNA sequencing, antigen receptor repertoire analysis, unbiased serum proteomics, and in vitro assays. Despite no evidence of active infection, we uncover elevated S100A-family alarmins in myeloid cells and marked enrichment of serum proteins that map to myeloid cells and pathways including cytokines, complement/coagulation, and fluid shear stress in MIS-C patients. Moreover, NK and CD8 T cell cytotoxicity genes are elevated, and plasmablasts harboring IgG1 and IgG3 are expanded. Consistently, we detect elevated binding of serum IgG from severe MIS-C patients to activated human cardiac microvascular endothelial cells in culture. Thus, we define immunopathology features of MIS-C with implications for predicting and managing this SARS-CoV2-induced critical illness in children.

The COVID-19 PHARMACOME: Rational Selection of Drug Repurposing Candidates from Multimodal Knowledge Harmonization (preprint)

The SARS-CoV-2 pandemic has challenged researchers at a global scale. The scientific community's massive response has resulted in a flood of experiments, analyses, hypotheses, and publications, especially in the field drug repurposing. However, many of the proposed therapeutic compounds obtained from SARS-CoV-2 specific assays are not in agreement and thus demonstrate the need for a singular source of COVID-19 related information from which a rational selection of drug repurposing candidates can be made. In this paper, we present the COVID-19 PHARMACOME, a comprehensive drug-target-mechanism graph generated from a compilation of several disease maps and experimental data focused on SARS-CoV-2 / COVID-19 pathophysiology. By applying a systematic approach, we were able to predict the effect of drug pairs on SARS-CoV-2 infection. Experimental validation of our results demonstrate that our graph can be used to not only explore the involved mechanistic pathways, but also to identify novel combinations of drug repurposing candidates.

A neutrophil activation signature predicts critical illness and mortality in COVID-19 (preprint)

Pathologic immune hyperactivation is emerging as a key feature of critical illness in COVID-19, but the mechanisms involved remain poorly understood. We carried out proteomic profiling of plasma from cross-sectional and longitudinal cohorts of hospitalized patients with COVID-19 and analyzed clinical data from our health system database of over 3,300 patients. Using a machine learning algorithm, we identified a prominent signature of neutrophil activation, including resistin, lipocalin-2, HGF, IL-8, and G-CSF, as the strongest predictors of critical illness. Neutrophil activation was present on the first day of hospitalization in patients who would only later require transfer to the intensive care unit, thus preceding the onset of critical illness and predicting increased mortality. In the health system database, early elevations in developing and mature neutrophil counts also predicted higher mortality rates. Altogether, we define an essential role for neutrophil activation in the pathogenesis of severe COVID-19 and identify molecular neutrophil markers that distinguish patients at risk of future clinical decompensation.

Generating hard-to-obtain information from easy-to-obtain information: applications in drug discovery and clinical inference (preprint)

In many important contexts involving measurements of biological entities, there are distinct categories of information: some information is easy-to-obtain information (EI) and can be gathered on virtually every subject of interest, while other information is hard-to-obtain information (HI) and can only be gathered on some of the biological samples. For example, in the context of drug discovery, measurements like the chemical structure of a drug are EI, while measurements of the transcriptome of a cell population perturbed with the drug is HI. In the clinical context, basic health monitoring is EI because it is already being captured as part of other processes, while cellular measurements like flow cytometry or even ultimate patient outcome are HI. We propose building a model to make probabilistic predictions of HI from EI on the samples that have both kinds of measurements, which will allow us to generalize and predict the HI on a large set of samples from just the EI. To accomplish this, we present a conditional Generative Adversarial Network (cGAN) framework we call the Feature Mapping GAN (FMGAN). By using the EI as conditions to map to the HI, we demonstrate that FMGAN can accurately predict the HI, with heterogeneity in cases of distributions of HI from EI. We show that FMGAN is flexible in that it can learn rich and complex mappings from EI to HI, and can take into account manifold structure in the EI space where available. We demonstrate this in a variety of contexts including generating RNA sequencing results on cell lines subjected to drug perturbations using drug chemical structure, and generating clinical outcomes from patient lab measurements. Most notably, we are able to generate synthetic flow cytometry data from clinical variables on a cohort of COVID-19 patients—effectively describing their immune response in great detail, and showcasing the power of generating expensive FACS data from ubiquitously available patient monitoring data. Bigger Picture Many experiments face a trade-off between gathering easy-to-collect information on many samples or hard-to-collect information on a smaller number of small due to costs in terms of both money and time. We demonstrate that a mapping between the easy-to-collect and hard-to-collect information can be trained as a conditional GAN from a subset of samples with both measured. With our conditional GAN model known as Feature-Mapping GAN (FMGAN), the results of expensive experiments can be predicted, saving on the costs of actually performing the experiment. This can have major impact in many settinsg. We study two example settings. First, in the field of pharmaceutical drug discovery early phase pharmaceutical experiments require casting a wide net to find a few potential leads to follow. In the long term, development pipelines can be re-designed to specifically utilize FMGAN in an optimal way to accelerate the process of drug discovery. FMGAN can also have a major impact in clinical setting, where routinely measured variables like blood pressure or heart rate can be used to predict important health outcomes and therefore deciding the best course of treatment.

Single-Cell Omics Reveals Dyssynchrony of the Innate and Adaptive Immune System in Progressive COVID-19 (preprint)

A dysregulated immune response against the SARS-CoV-2 virus plays a critical role in severe COVID-19. However, the molecular and cellular mechanisms by which the virus causes lethal immunopathology are poorly understood. Here, we utilize multi-omics single-cell analysis to probe dynamic immune responses in patients with stable or progressive manifestations of COVID-19, and assess the effects of tocilizumab, an anti-IL-6 receptor monoclonal antibody. Coordinated profiling of gene expression and cell lineage protein markers reveals a prominent type-1 interferon response across all immune cells, especially in progressive patients. An anti-inflammatory innate immune response and a pre-exhaustion phenotype in activated T cells are hallmarks of progressive disease. Skewed T cell receptor repertoires in CD8 T cells and uniquely enriched V(D)J sequences are also identified in COVID-19 patients. B cell repertoire and somatic hypermutation analysis are consistent with a primary immune response, with possible contribution from memory B cells. Our in-depth immune profiling reveals dyssynchrony of the innate and adaptive immune interaction in progressive COVID-19, which may contribute to delayed virus clearance and has implications for therapeutic intervention.

Circulating Markers of Angiogenesis and Endotheliopathy in COVID-19 (preprint)

Despite over 9.3 million infected and 479,000 deaths, the pathophysiological factors that determine the wide spectrum of clinical outcomes in COVID-19 remain inadequately defined. Importantly, patients with underlying cardiovascular disease have been found to have worse clinical outcomes,1 and autopsy findings of endotheliopathy as well as angiogenesis in COVID-19 have accumulated.2,3 Nonetheless, circulating vascular markers associated with disease severity and mortality have not been reliably established. To address this limitation and better understand COVID-19 pathogenesis, we report plasma profiling of factors related to the vascular system from a series of patients admitted to Yale-New Haven Hospital with confirmed diagnosis of COVID-19 via PCR, which demonstrate significant increase in markers of angiogenesis and endotheliopathy in patients hospitalized with COVID-19.

Neuroinvasive potential of SARS-CoV-2 revealed in a human brain organoid model (preprint)

Although COVID-19 is considered to be primarily a respiratory disease, SARS-CoV-2 affects multiple organ systems including the central nervous system (CNS). Reports indicate that 30-60% of patients with COVID-19 suffer from CNS symptoms. Yet, there is no consensus whether the virus can infect the brain, or what the consequences of infection are. Following SARS-CoV-2 infection of human brain organoids, clear evidence of infection was observed, with accompanying metabolic changes in the infected and neighboring neurons. Further, no evidence for the type I interferon responses was detected. We demonstrate that neuronal infection can be prevented either by blocking ACE2 with antibodies or by administering cerebrospinal fluid from a COVID-19 patient. Finally, using mice overexpressing human ACE2, we demonstrate in vivo that SARS-CoV-2 neuroinvasion, but not respiratory infection, is associated with mortality. These results provide evidence for the neuroinvasive capacity of SARS-CoV2, and an unexpected consequence of direct infection of neurons by SARS-CoV2.

Gaining Insight into SARS-CoV-2 Infection and COVID-19 Severity Using Self-supervised Edge Features and Graph Neural Networks (preprint)

A molecular and cellular understanding of how SARS-CoV-2 variably infects and causes severe COVID-19 remains a bottleneck in developing interventions to end the pandemic. We sought to use deep learning to study the biology of SARS-CoV-2 infection and COVID-19 severity by identifying transcriptomic patterns and cell types associated with SARS-CoV-2 infection and COVID-19 severity. To do this, we developed a new approach to generating self-supervised edge features. We propose a model that builds on Graph Attention Networks (GAT), creates edge features using self-supervised learning, and ingests these edge features via a Set Transformer. This model achieves significant improvements in predicting the disease state of individual cells, given their transcriptome. We apply our model to single-cell RNA sequencing datasets of SARS-CoV-2 infected lung organoids and bronchoalveolar lavage fluid samples of patients with COVID-19, achieving state-of-the-art performance on both datasets with our model. We then borrow from the field of explainable AI (XAI) to identify the features (genes) and cell types that discriminate bystander vs. infected cells across time and moderate vs. severe COVID-19 disease. To the best of our knowledge, this represents the first application of deep learning to identifying the molecular and cellular determinants of SARS-CoV-2 infection and COVID-19 severity using single-cell omics data.

Self-supervised edge features for improved Graph Neural Network training (preprint)

Graph Neural Networks (GNN) have been extensively used to extract meaningful representations from graph structured data and to perform predictive tasks such as node classification and link prediction. In recent years, there has been a lot of work incorporating edge features along with node features for prediction tasks. One of the main difficulties in using edge features is that they are often handcrafted, hard to get, specific to a particular domain, and may contain redundant information. In this work, we present a framework for creating new edge features, applicable to any domain, via a combination of self-supervised and unsupervised learning. In addition to this, we use Forman-Ricci curvature as an additional edge feature to encapsulate the local geometry of the graph. We then encode our edge features via a Set Transformer and combine them with node features extracted from popular GNN architectures for node classification in an end-to-end training scheme. We validate our work on three biological datasets comprising of single-cell RNA sequencing data of neurological disease, \textit{in vitro} SARS-CoV-2 infection, and human COVID-19 patients. We demonstrate that our method achieves better performance on node classification tasks over baseline Graph Attention Network (GAT) and Graph Convolutional Network (GCN) models. Furthermore, given the attention mechanism on edge and node features, we are able to interpret the cell types and genes that determine the course and severity of COVID-19, contributing to a growing list of potential disease biomarkers and therapeutic targets.

Genome-wide CRISPR screen reveals host genes that regulate SARS-CoV-2 infection (preprint)

Identification of host genes essential for SARS-CoV-2 infection may reveal novel therapeutic targets and inform our understanding of COVID-19 pathogenesis. Here we performed a genome-wide CRISPR screen with SARS-CoV-2 and identified known SARS-CoV-2 host factors including the receptor ACE2 and protease Cathepsin L. We additionally discovered novel pro-viral genes and pathways including the SWI/SNF chromatin remodeling complex and key components of the TGF-{beta} signaling pathway. Small molecule inhibitors of these pathways prevented SARS-CoV-2-induced cell death. We also revealed that the alarmin HMGB1 is critical for SARS-CoV-2 replication. In contrast, loss of the histone H3.3 chaperone complex sensitized cells to virus-induced death. Together this study reveals potential therapeutic targets for SARS-CoV-2 and highlights host genes that may regulate COVID-19 pathogenesis.

Development and validation of the COVID-19 severity index (CSI): a prognostic tool for early respiratory decompensation (preprint)

ObjectiveThe goal of this study was to create a predictive model of early hospital respiratory decompensation among patients with COVID-19. DesignObservational, retrospective cohort study. SettingNine-hospital health system within the Northeastern United States. PopulationsAdult patients ([≥] 18 years) admitted from the emergency department who tested positive for SARS-CoV-2 (COVID-19) up to 24 hours after initial presentation. Patients meeting criteria for respiratory critical illness within 4 hours of arrival were excluded. Main outcome and performance measuresWe used a composite endpoint of critical illness as defined by oxygen requirement (greater than 10 L/min by low-flow device, high-flow device, non-invasive, or invasive ventilation) or death within the first 24 hours of hospitalization. We developed models predicting our composite endpoint using patient demographic and clinical data available within the first four hours of arrival. Eight hospitals (n = 932) were used for model development and one hospital (n = 240) was held out for external validation. Area under receiver operating characteristic (AU-ROC), precision-recall curves (AU-PRC), and calibration metrics were used to compare predictive models to three illness scoring systems: Elixhauser comorbidity index, qSOFA, and CURB-65. ResultsDuring the study period from March 1, 2020 to April 27,2020, 1,792 patients were admitted with COVID-19. Six-hundred and twenty patients were excluded based on age or critical illness within the first 4 hours, yielding 1,172 patients in the final cohort. Of these patients, 144 (12.3%) met the composite endpoint within the first 24 hours. We first developed a bedside quick COVID-19 severity index (qCSI), a twelve-point scale using nasal cannula flow rate, respiratory rate, and minimum documented pulse oximetry. We then created a machine-learning gradient boosting model, the COVID-19 severity index (CSI), using twelve additional variables including inflammatory markers and liver chemistries. Both the qCSI (AU-ROC mean [95% CI]: 0.90 [0.85-0.96]) and CSI (AU-ROC: 0.91 [0.86-0.97]) outperformed the comparator models (qSOFA: 0.76 [0.69-0.85]; Elixhauser: 0.70 [0.62-0.80]; CURB-65: AU-ROC 0.66 [0.58-0.77]) on cross-validation and performed well on external validation (qCSI: 0.82, CSI: 0.76, CURB-65: 0.50, qSOFA: 0.59, Elixhauser: 0.61). We find that a qCSI score of 0-3 is associated with a less than 5% risk of critical respiratory illness, while a score of 9-12 is associated with a 57% risk of progression to critical illness. ConclusionsA significant proportion of admitted COVID-19 patients decompensate within 24 hours of hospital presentation and these events are accurately predicted using bedside respiratory exam findings within a simple scoring system.

Single-cell longitudinal analysis of SARS-CoV-2 infection in human bronchial epithelial cells (preprint)

SARS-CoV-2, the causative agent of COVID-19, has tragically burdened individuals and institutions around the world. There are currently no approved drugs or vaccines for the treatment or prevention of COVID-19. Enhanced understanding of SARS-CoV-2 infection and pathogenesis is critical for the development of therapeutics. To reveal insight into viral replication, cell tropism, and host-viral interactions of SARS-CoV-2 we performed single-cell RNA sequencing of experimentally infected human bronchial epithelial cells (HBECs) in air-liquid interface cultures over a time-course. This revealed novel polyadenylated viral transcripts and highlighted ciliated cells as a major target of infection, which we confirmed by electron microscopy. Over the course of infection, cell tropism of SARS-CoV-2 expands to other epithelial cell types including basal and club cells. Infection induces cell-intrinsic expression of type I and type III IFNs and IL6 but not IL1. This results in expression of interferon-stimulated genes in both infected and bystander cells. We observe similar gene expression changes from a COVID-19 patient ex vivo. In addition, we developed a new computational method termed CONditional DENSity Embedding (CONDENSE) to characterize and compare temporal gene dynamics in response to infection, which revealed genes relating to endothelin, angiogenesis, interferon, and inflammation-causing signaling pathways. In this study, we conducted an in-depth analysis of SARS-CoV-2 infection in HBECs and a COVID-19 patient and revealed genes, cell types, and cell state changes associated with infection.