OpenCyto: an open source infrastructure for scalable, robust, reproducible, and automated, end-to-end flow cytometry data analysis

Tertiary lymphoid structure-related immune infiltrates in NSCLC tumor lesions correlate with low tumor-reactivity of TIL products

Adoptive transfer of tumor infiltrating lymphocytes (TIL therapy) has proven highly effective for treating solid cancers, including non-small cell lung cancer (NSCLC). However, not all patients benefit from this therapy for yet unknown reasons. Defining markers that correlate with high tumor-reactivity of the autologous TIL products is thus key for achieving better tailored immunotherapies. We questioned whether the composition of immune cell infiltrates correlated with the tumor-reactivity of expanded TIL products. Unbiased flow cytometry analysis of immune cell infiltrates of 26 early-stage and 20 late-stage NSCLC tumor lesions was used for correlations with the T cell differentiation and activation status, and with the expansion rate and anti-tumor response of generated TIL products. The composition of tumor immune infiltrates was highly variable between patients. Spearman's Rank Correlation revealed that high B cell infiltration negatively correlated with the tumor-reactivity of the patient's expanded TIL products, as defined by cytokine production upon exposure to autologous tumor digest. In-depth analysis revealed that tumor lesions with high B cell infiltrates contained tertiary lymphoid structure (TLS)-related immune infiltrates, including BCL6+ antibody-secreting B cells, IgD+BCL6+ B cells and CXCR5+BLC6+ CD4+ T cells, and higher percentages of naïve CD8+ T cells. In conclusion, the composition of immune cell infiltrates in NSCLC tumors associates with the functionality of the expanded TIL product. Our findings may thus help improve patient selection for TIL therapy.

Specific CD4+ T cell phenotypes associate with bacterial control in people who 'resist' infection with Mycobacterium tuberculosis

A subset of individuals exposed to Mycobacterium tuberculosis (Mtb) that we refer to as 'resisters' (RSTR) show evidence of IFN-γ- T cell responses to Mtb-specific antigens despite serially negative results on clinical testing. Here we found that Mtb-specific T cells in RSTR were clonally expanded, confirming the priming of adaptive immune responses following Mtb exposure. RSTR CD4+ T cells showed enrichment of TH17 and regulatory T cell-like functional programs compared to Mtb-specific T cells from individuals with latent Mtb infection. Using public datasets, we showed that these TH17 cell-like functional programs were associated with lack of progression to active tuberculosis among South African adolescents with latent Mtb infection and with bacterial control in nonhuman primates. Our findings suggested that RSTR may successfully control Mtb following exposure and immune priming and established a set of T cell biomarkers to facilitate further study of this clinical phenotype.

MetaGate: Interactive analysis of high-dimensional cytometry data with metadata integration

Flow cytometry is a powerful technology for high-throughput protein quantification at the single-cell level. Technical advances have substantially increased data complexity, but novel bioinformatical tools often show limitations in statistical testing, data sharing, cross-experiment comparability, or clinical data integration. We developed MetaGate as a platform for interactive statistical analysis and visualization of manually gated high-dimensional cytometry data with integration of metadata. MetaGate provides a data reduction algorithm based on a combinatorial gating system that produces a small, portable, and standardized data file. This is subsequently used to produce figures and statistical analyses through a fast web-based user interface. We demonstrate the utility of MetaGate through a comprehensive mass cytometry analysis of peripheral blood immune cells from 28 patients with diffuse large B cell lymphoma along with 17 healthy controls. Through MetaGate analysis, our study identifies key immune cell population changes associated with disease progression.

Reduced cardiac antioxidant defenses mediate increased susceptibility to workload-induced myocardial injury in males with genetic cardiomyopathy

Ongoing cardiomyocyte injury is a major mechanism in the progression of heart failure, particularly in dystrophic hearts. Due to the poor regenerative capacity of the adult heart, cardiomyocyte death results in the permanent loss of functional myocardium. Understanding the factors contributing to myocyte injury is essential for the development of effective heart failure therapies. As a model of persistent cardiac injury, we examined mice lacking β-sarcoglycan (β-SG), a key component of the dystrophin glycoprotein complex (DGC). The loss of the sarcoglycan complex markedly compromises sarcolemmal integrity in this β-SG-/- model. Our studies aim to characterize the mechanisms underlying dramatic sex differences in susceptibility to cardiac injury in β-SG-/- mice. Male β-SG-/- hearts display significantly greater myocardial injury and death following isoproterenol-induced cardiac stress than female β-SG-/- hearts. This protection of females was independent of ovarian hormones. Male β-SG-/- hearts displayed increased susceptibility to exogenous oxidative stress and were significantly protected by angiotensin II type 1 receptor (AT1R) antagonism. Increasing general antioxidative defenses or increasing the levels of S-nitrosylation both provided protection to the hearts of β-SG-/- male mice. Here we demonstrate that increased susceptibility to oxidative damage leads to an AT1R-mediated amplification of workload-induced myocardial injury in male β-SG-/- mice. Improving oxidative defenses, specifically by increasing S-nitrosylation, provided protection to the male β-SG-/- heart from workload-induced injury. These studies describe a unique susceptibility of the male heart to injury and may contribute to the sex differences in other forms of cardiac injury.

Fat-to-blood recirculation of partially dysfunctional PD-1+CD4 Tconv cells is associated with dysglycemia in human obesity

Obesity is characterized by the accumulation of T cells in insulin-sensitive tissues, including the visceral adipose tissue (VAT), that can interfere with the insulin signaling pathway eventually leading to insulin resistance (IR) and type 2 diabetes. Here, we found that PD-1+CD4 conventional T (Tconv) cells, endowed with a transcriptomic and functional profile of partially dysfunctional cells, are diminished in VAT of obese patients with dysglycemia (OB-Dys), without a concomitant increase in apoptosis. These cells showed enhanced capacity to recirculate into the bloodstream and had a non-restricted TCRβ repertoire divergent from that of normoglycemic obese and lean individuals. PD-1+CD4 Tconv were reduced in the circulation of OB-Dys, exhibited an altered migration potential, and were detected in the liver of patients with non-alcoholic steatohepatitis. The findings suggest a potential role for partially dysfunctional PD-1+CD4 Tconv cells as inter-organ mediators of IR in obese patients with dysglycemic.

Application of Flow Cytometry Using Advanced Chromatin Analyses for Assessing Changes in Sperm Structure and DNA Integrity in a Porcine Model

Chromatin status is critical for sperm fertility and reflects spermatogenic success. We tested a multivariate approach for studying pig sperm chromatin structure to capture its complexity with a set of quick and simple techniques, going beyond the usual assessment of DNA damage. Sperm doses from 36 boars (3 ejaculates/boar) were stored at 17 °C and analyzed on days 0 and 11. Analyses were: CASA (motility) and flow cytometry to assess sperm functionality and chromatin structure by SCSA (%DFI, DNA fragmentation; %HDS, chromatin maturity), monobromobimane (mBBr, tiol status/disulfide bridges between protamines), chromomycin A3 (CMA3, protamination), and 8-hydroxy-2'-deoxyguanosine (8-oxo-dG, DNA oxidative damage). Data were analyzed using linear models for the effects of boar and storage, correlations, and multivariate analysis as hierarchical clustering and principal component analysis (PCA). Storage reduced sperm quality parameters, mainly motility, with no critical oxidative stress increases, while chromatin status worsened slightly (%DFI and 8-oxo-dG increased while mBBr MFI-median fluorescence intensity-and disulfide bridge levels decreased). Boar significantly affected most chromatin variables except CMA3; storage also affected most variables except %HDS. At day 0, sperm chromatin variables clustered closely, except for CMA3, and %HDS and 8-oxo-dG correlated with many variables (notably, mBBr). After storage, the relation between %HDS and 8-oxo-dG remained, but correlations among other variables disappeared, and mBBr variables clustered separately. The PCA suggested a considerable influence of mBBr on sample variance, especially regarding storage, with SCSA and 8-oxo-dG affecting between-sample variability. Overall, CMA3 was the least informative, in contrast with results in other species. The combination of DNA fragmentation, DNA oxidation, chromatin compaction, and tiol status seems a good candidate for obtaining a complete picture of pig sperm nucleus status. It raises many questions for future molecular studies and deserves further research to establish its usefulness as a fertility predictor in multivariate models. The usefulness of CMA3 should be clarified.

A flow-cytometry-based pipeline for the rapid quantification of C2C12 cell differentiation

The C2C12 cell line represents a simple in vitro model for cell differentiation. Here, we present a flow-cytometry-based pipeline to quantitate C2C12 cell differentiation based on myosin heavy-chain marker expression. We describe steps for cell seeding, transfection, drug treatment, differentiation, and labeling. We then detail procedures for flow cytometry acquisition and introduce the R script FlowFate for automated analysis, including the study of dose-dependent effects of GFP-tagged genes on differentiation. For complete details on the use and execution of this protocol, please refer to Chippalkatti et al. (2023).1.

GATA transcription factors drive initial Xist upregulation after fertilization through direct activation of long-range enhancers

X-chromosome inactivation (XCI) balances gene expression between the sexes in female mammals. Shortly after fertilization, upregulation of Xist RNA from one X chromosome initiates XCI, leading to chromosome-wide gene silencing. XCI is maintained in all cell types, except the germ line and the pluripotent state where XCI is reversed. The mechanisms triggering Xist upregulation have remained elusive. Here we identify GATA transcription factors as potent activators of Xist. Through a pooled CRISPR activation screen in murine embryonic stem cells, we demonstrate that GATA1, as well as other GATA transcription factors can drive ectopic Xist expression. Moreover, we describe GATA-responsive regulatory elements in the Xist locus bound by different GATA factors. Finally, we show that GATA factors are essential for XCI induction in mouse preimplantation embryos. Deletion of GATA1/4/6 or GATA-responsive Xist enhancers in mouse zygotes effectively prevents Xist upregulation. We propose that the activity or complete absence of various GATA family members controls initial Xist upregulation, XCI maintenance in extra-embryonic lineages and XCI reversal in the epiblast.

Automated EuroFlow approach for standardized in-depth dissection of human circulating B-cells and plasma cells

Background

Multiparameter flow cytometry (FC) immunophenotyping is a key tool for detailed identification and characterization of human blood leucocytes, including B-lymphocytes and plasma cells (PC). However, currently used conventional data analysis strategies require extensive expertise, are time consuming, and show limited reproducibility.

Objective

Here, we designed, constructed and validated an automated database-guided gating and identification (AGI) approach for fast and standardized in-depth dissection of B-lymphocyte and PC populations in human blood.

Methods

For this purpose, 213 FC standard (FCS) datafiles corresponding to umbilical cord and peripheral blood samples from healthy and patient volunteers, stained with the 14-color 18-antibody EuroFlow BIgH-IMM panel, were used.

Results

The BIgH-IMM antibody panel allowed identification of 117 different B-lymphocyte and PC subsets. Samples from 36 healthy donors were stained and 14 of the datafiles that fulfilled strict inclusion criteria were analysed by an expert flow cytometrist to build the EuroFlow BIgH-IMM database. Data contained in the datafiles was then merged into a reference database that was uploaded in the Infinicyt software (Cytognos, Salamanca, Spain). Subsequently, we compared the results of manual gating (MG) with the performance of two classification algorithms -hierarchical algorithm vs two-step algorithm- for AGI of the cell populations present in 5 randomly selected FCS datafiles. The hierarchical AGI algorithm showed higher correlation values vs conventional MG (r2 of 0.94 vs. 0.88 for the two-step AGI algorithm) and was further validated in a set of 177 FCS datafiles against conventional expert-based MG. For virtually all identifiable cell populations a highly significant correlation was observed between the two approaches (r2>0.81 for 79% of all B-cell populations identified), with a significantly lower median time of analysis per sample (6 vs. 40 min, p=0.001) for the AGI tool vs. MG, respectively and both intra-sample (median CV of 1.7% vs. 10.4% by MG, p<0.001) and inter-expert (median CV of 3.9% vs. 17.3% by MG by 2 experts, p<0.001) variability.

Conclusion

Our results show that compared to conventional FC data analysis strategies, the here proposed AGI tool is a faster, more robust, reproducible, and standardized approach for in-depth analysis of B-lymphocyte and PC subsets circulating in human blood.

Mass cytometry as a tool in target validation and drug discovery

Mass cytometry provides highly multiparametric data at a single cell level, coupling the specificity and sensitivity of time-of-flight mass spectrometry with the single-cell throughput of flow cytometry. It offers great value in interrogating the potentially heterogenous impact that a drug may have on a biological system, allowing an investigator to capture not just changes in cell behavior, but how these changes may differ between cell subtypes. In this chapter, we review the technical details of the platform as well as its limitations, before describing our approach to planning and running a mass cytometry experiment. A series of method modules, spanning the staining process through to data cleaning, are described that are then combined to create three separate experiments. The first experiment illustrates a core process in mass cytometry: the validation and titration of a metal-conjugated antibody reporter. The second experiment explores the impact of a kinase inhibitor on cell cycle and apoptosis pathways of a human myeloma cell line. And the third experiment exploits the multiparametric capability of mass cytometry, by exploring the differential expression changes in a transcription factor upon drug treatment across the cellular compartments of a peripheral blood mononuclear cell sample.

A cell-level discriminative neural network model for diagnosis of blood cancers

MOTIVATION

Precise identification of cancer cells in patient samples is essential for accurate diagnosis and clinical monitoring but has been a significant challenge in machine learning approaches for cancer precision medicine. In most scenarios, training data are only available with disease annotation at the subject or sample level. Traditional approaches separate the classification process into multiple steps that are optimized independently. Recent methods either focus on predicting sample-level diagnosis without identifying individual pathologic cells or are less effective for identifying heterogeneous cancer cell phenotypes.

RESULTS

We developed a generalized end-to-end differentiable model, the Cell Scoring Neural Network (CSNN), which takes sample-level training data and predicts the diagnosis of the testing samples and the identity of the diagnostic cells in the sample, simultaneously. The cell-level density differences between samples are linked to the sample diagnosis, which allows the probabilities of individual cells being diagnostic to be calculated using backpropagation. We applied CSNN to two independent clinical flow cytometry datasets for leukemia diagnosis. In both qualitative and quantitative assessments, CSNN outperformed preexisting neural network modeling approaches for both cancer diagnosis and cell-level classification. Post hoc decision trees and 2D dot plots were generated for interpretation of the identified cancer cells, showing that the identified cell phenotypes match the cancer endotypes observed clinically in patient cohorts. Independent data clustering analysis confirmed the identified cancer cell populations.

AVAILABILITY AND IMPLEMENTATION

The source code of CSNN and datasets used in the experiments are publicly available on GitHub (http://github.com/erobl/csnn). Raw FCS files can be downloaded from FlowRepository (ID: FR-FCM-Z6YK).

Streamlining N-terminally anchored yeast surface display via structural insights into S. cerevisiae Pir proteins

Surface display co-opts yeast's innate ability to embellish its cell wall with mannoproteins, thus converting the yeast's outer surface into a growing and self-sustaining catalyst. However, the efficient toolbox for converting the enzyme of interest into its surface-displayed isoform is currently lacking, especially if the isoform needs to be anchored to the cell wall near the isoform's N-terminus, e.g., through a short GPI-independent protein anchor. Aiming to advance such N-terminally anchored surface display, we employed in silico and machine-learning strategies to study the 3D structure, function, genomic organisation, and evolution of the Pir protein family, whose members evolved to covalently attach themselves near their N-terminus to the β-1,3-glucan of the cell wall. Through the newly-gained insights, we rationally engineered 14 S. cerevisiae Hsp150 (Pir2)-based fusion proteins. We quantified their performance, uncovering guidelines for efficient yeast surface display while developing a construct that promoted a 2.5-fold more efficient display of a reporter protein than the full-length Hsp150. Moreover, we developed a Pir-tag, i.e., a peptide spanning only 4.5 kDa but promoting as efficient surface display of a reporter protein as the full-length Hsp150. These constructs fortify the existing surface display toolbox, allowing for a prompt and routine refitting of intracellular proteins into their N-terminally anchored isoforms.

A topic modeling approach reveals the dynamic T cell composition of peripheral blood during cancer immunotherapy

We present TopicFlow, a computational framework for flow cytometry data analysis of patient blood samples for the identification of functional and dynamic topics in circulating T cell population. This framework applies a Latent Dirichlet Allocation (LDA) model, adapting the concept of topic modeling in text mining to flow cytometry. To demonstrate the utility of our method, we conducted an analysis of ∼17 million T cells collected from 138 peripheral blood samples in 51 patients with melanoma undergoing treatment with immune checkpoint inhibitors (ICIs). Our study highlights three latent dynamic topics identified by LDA: a T cell exhaustion topic that independently recapitulates the previously identified LAG-3+ immunotype associated with ICI resistance, a naive topic and its association with immune-related toxicity, and a T cell activation topic that emerges upon ICI treatment. Our approach can be broadly applied to mine high-parameter flow cytometry data for insights into mechanisms of treatment response and toxicity.

Computational Methods for Single-Cell Proteomics

Advances in single-cell proteomics technologies have resulted in high-dimensional datasets comprising millions of cells that are capable of answering key questions about biology and disease. The advent of these technologies has prompted the development of computational tools to process and visualize the complex data. In this review, we outline the steps of single-cell and spatial proteomics analysis pipelines. In addition to describing available methods, we highlight benchmarking studies that have identified advantages and pitfalls of the currently available computational toolkits. As these technologies continue to advance, robust analysis tools should be developed in tandem to take full advantage of the potential biological insights provided by these data.

AliGater: a framework for the development of bioinformatic pipelines for large-scale, high-dimensional cytometry data

Motivation

AliGater is an open-source framework to accelerate the development of bioinformatic pipelines for the analysis of large-scale, high-dimensional flow cytometry data. AliGater provides a Python package for automatic feature extraction workflows, as well as building blocks to construct analysis pipelines.

Results

We illustrate the use of AliGater in a high-resolution flow cytometry-based genome-wide association study on 46 immune cell populations in 14 288 individuals.

Availability and implementation

Source code and documentation at https://github.com/LudvigEk/aligater and https://aligater.readthedocs.io.

Identifying cancer-associated leukocyte profiles using high-resolution flow cytometry screening and machine learning

Background

Machine learning (ML) is a valuable tool with the potential to aid clinical decision making. Adoption of ML to this end requires data that reliably correlates with the clinical outcome of interest; the advantage of ML is that it can model these correlations from complex multiparameter data sets that can be difficult to interpret conventionally. While currently available clinical data can be used in ML for this purpose, there exists the potential to discover new "biomarkers" that will enhance the effectiveness of ML in clinical decision making. Since the interaction of the immune system and cancer is a hallmark of tumor establishment and progression, one potential area for cancer biomarker discovery is through the investigation of cancer-related immune cell signatures. Hence, we hypothesize that blood immune cell signatures can act as a biomarker for cancer progression.

Methods

To probe this, we have developed and tested a multiparameter cell-surface marker screening pipeline, using flow cytometry to obtain high-resolution systemic leukocyte population profiles that correlate with detection and characterization of several cancers in murine syngeneic tumor models.

Results

We discovered a signature of several blood leukocyte subsets, the most notable of which were monocyte subsets, that could be used to train CATboost ML models to predict the presence and type of cancer present in the animals.

Conclusions

Our findings highlight the potential utility of a screening approach to identify robust leukocyte biomarkers for cancer detection and characterization. This pipeline can easily be adapted to screen for cancer specific leukocyte markers from the blood of cancer patient.

Proteome Mapping of the Human Pancreatic Islet Microenvironment Reveals Endocrine-Exocrine Signaling Sphere of Influence

The need for a clinically accessible method with the ability to match protein activity within heterogeneous tissues is currently unmet by existing technologies. Our proteomics sample preparation platform, named microPOTS (Microdroplet Processing in One pot for Trace Samples), can be used to measure relative protein abundance in micron-scale samples alongside the spatial location of each measurement, thereby tying biologically interesting proteins and pathways to distinct regions. However, given the smaller pixel/voxel number and amount of tissue measured, standard mass spectrometric analysis pipelines have proven inadequate. Here we describe how existing computational approaches can be adapted to focus on the specific biological questions asked in spatial proteomics experiments. We apply this approach to present an unbiased characterization of the human islet microenvironment comprising the entire complex array of cell types involved while maintaining spatial information and the degree of the islet's sphere of influence. We identify specific functional activity unique to the pancreatic islet cells and demonstrate how far their signature can be detected in the adjacent tissue. Our results show that we can distinguish pancreatic islet cells from the neighboring exocrine tissue environment, recapitulate known biological functions of islet cells, and identify a spatial gradient in the expression of RNA processing proteins within the islet microenvironment.

IFN-γ independent markers of Mycobacterium tuberculosis exposure among male South African gold miners

BACKGROUND

The prevalence of tuberculosis among men who work in the gold mines of South Africa is among the highest in the world, but a fraction of miners demonstrate consistently negative results upon tuberculin skin test (TST) and IFN-γ release assay (IGRA). We hypothesized that these "resisters" (RSTRs) may display unconventional immune signatures of exposure to M. tuberculosis (M.tb).

METHODS

In a cohort of RSTRs and matched controls with latent TB infection (LTBI), we profiled the functional breadth of M.tb antigen-specific T cell and antibody responses using multi-parameter flow cytometry and systems serology, respectively.

FINDINGS

RSTRs and LTBI controls both exhibited IFN-γ independent T-cell and IgG antibody responses to M.tb-specific antigens ESAT-6 and CFP-10. Antigen-specific antibody Fc galactosylation and sialylation were higher among RSTRs. In a combined T-cell and antibody analysis, M.tb lysate-stimulated TNF secretion by T cells correlated positively with levels of purified protein derivative-specific IgG. A multivariate model of the combined data was able to differentiate RSTR and LTBI subjects.

INTERPRETATION

IFN-γ independent immune signatures of exposure to M.tb, which are not detected by approved clinical diagnostics, are readily detectable in an occupational cohort uniquely characterized by intense and long-term infection pressure. Further, TNF may mediate a coordinated response between M.tb-specific T-cells and B-cells.

FUNDING

This work was supported by the US National Institutes of Health (R01-AI124348 to Boom, Stein, and Hawn; R01-AI125189 and R01-AI146072 to Seshadri; and 75N93019C00071 to Fortune, Alter, Seshadri, and Boom), the Doris Duke Charitable Foundation (Davies), the Bill & Melinda Gates Foundation (OPP1151836 and OPP1109001 to Hawn; and OPP1151840 to Alter), Mass Life Science Foundation (Fortune), and Good Ventures Fund (Fortune).

Cytotoxic T Cells Targeting Spike Glycoprotein Are Associated with Hybrid Immunity to SARS-CoV-2

mRNA vaccination of individuals with prior SARS-CoV-2 infection provides superior protection against breakthrough infections with variants of concern compared with vaccination in the absence of prior infection. However, the immune mechanisms by which this hybrid immunity is generated and maintained are unknown. Whereas genetic variation in spike glycoprotein effectively subverts neutralizing Abs, spike-specific T cells are generally maintained against SARS-CoV-2 variants. Thus, we comprehensively profiled human T cell responses against the S1 and S2 domains of spike glycoprotein in a cohort of SARS-CoV-2-naive (n = 13) or -convalescent (n = 17) individuals who received two-dose mRNA vaccine series and were matched by age, sex, and vaccine type. Using flow cytometry, we observed that the overall functional breadth of CD4 T cells and polyfunctional Th1 responses was similar between the two groups. However, polyfunctional cytotoxic CD4 T cell responses against both S1 and S2 domains trended higher among convalescent subjects. Multimodal single-cell RNA sequencing revealed diverse functional programs in spike-specific CD4 and CD8 T cells in both groups. However, convalescent individuals displayed enhanced cytotoxic and antiviral CD8 T cell responses to both S1 and S2 in the absence of cytokine production. Taken together, our data suggest that cytotoxic CD4 and CD8 T cells targeting spike glycoprotein may partially account for hybrid immunity and protection against breakthrough infections with SARS-CoV-2.

Uncovering the hidden structure of dynamic T cell composition in peripheral blood during cancer immunotherapy: a topic modeling approach

Immune checkpoint inhibitors (ICIs), now mainstays in the treatment of cancer treatment, show great potential but only benefit a subset of patients. A more complete understanding of the immunological mechanisms and pharmacodynamics of ICI in cancer patients will help identify the patients most likely to benefit and will generate knowledge for the development of next-generation ICI regimens. We set out to interrogate the early temporal evolution of T cell populations from longitudinal single-cell flow cytometry data. We developed an innovative statistical and computational approach using a Latent Dirichlet Allocation (LDA) model that extends the concept of topic modeling used in text mining. This powerful unsupervised learning tool allows us to discover compositional topics within immune cell populations that have distinct functional and differentiation states and are biologically and clinically relevant. To illustrate the model's utility, we analyzed ∼17 million T cells obtained from 138 pre- and on-treatment peripheral blood samples from a cohort of melanoma patients treated with ICIs. We identified three latent dynamic topics: a T-cell exhaustion topic that recapitulates a LAG3+ predominant patient subgroup with poor clinical outcome; a naive topic that shows association with immune-related toxicity; and an immune activation topic that emerges upon ICI treatment. We identified that a patient subgroup with a high baseline of the naïve topic has a higher toxicity grade. While the current application is demonstrated using flow cytometry data, our approach has broader utility and creates a new direction for translating single-cell data into biological and clinical insights.

Artificial intelligence in clinical multiparameter flow cytometry and mass cytometry-key tools and progress

There are many research studies and emerging tools using artificial intelligence (AI) and machine learning to augment flow and mass cytometry workflows. Emerging AI tools can quickly identify common cell populations with continuous improvement of accuracy, uncover patterns in high-dimensional cytometric data that are undetectable by human analysis, facilitate the discovery of cell subpopulations, perform semi-automated immune cell profiling, and demonstrate potential to automate aspects of clinical multiparameter flow cytometric (MFC) diagnostic workflow. Utilizing AI in the analysis of cytometry samples can reduce subjective variability and assist in breakthroughs in understanding diseases. Here we review the diverse types of AI that are being applied to clinical cytometry data and how AI is driving advances in data analysis to improve diagnostic sensitivity and accuracy. We review supervised and unsupervised clustering algorithms for cell population identification, various dimensionality reduction techniques, and their utilities in visualization and machine learning pipelines, and supervised learning approaches for classifying entire cytometry samples.Understanding the AI landscape will enable pathologists to better utilize open source and commercially available tools, plan exploratory research projects to characterize diseases, and work with machine learning and data scientists to implement clinical data analysis pipelines.

A cell-level discriminative neural network model for diagnosis of blood cancers

Motivation

Precise identification of cancer cells in patient samples is essential for accurate diagnosis and clinical monitoring but has been a significant challenge in machine learning approaches for cancer precision medicine. In most scenarios, training data are only available with disease annotation at the subject or sample level. Traditional approaches separate the classification process into multiple steps that are optimized independently. Recent methods either focus on predicting sample-level diagnosis without identifying individual pathologic cells or are less effective for identifying heterogeneous cancer cell phenotypes.

Results

We developed a generalized end-to-end differentiable model, the Cell Scoring Neural Network (CSNN), which takes the available sample-level training data and predicts both the diagnosis of the testing samples and the identity of the diagnostic cells in the sample, simultaneously. The cell-level density differences between samples are linked to the sample diagnosis, which allows the probabilities of individual cells being diagnostic to be calculated using backpropagation. We applied CSNN to two independent clinical flow cytometry datasets for leukemia diagnosis. In both qualitative and quantitative assessments, CSNN outperformed preexisting neural network modeling approaches for both cancer diagnosis and cell-level classification. Post hoc decision trees and 2D dot plots were generated for interpretation of the identified cancer cells, showing that the identified cell phenotypes match the cancer endotypes observed clinically in patient cohorts. Independent data clustering analysis confirmed the identified cancer cell populations.

Availability

The source code of CSNN and datasets used in the experiments are publicly available on GitHub and FlowRepository.

Contact

Edgar E. Robles: roblesee@uci.edu and Yu Qian: mqian@jcvi.org.

Supplementary information

Supplementary data are available on GitHub and at Bioinformatics online.

HIV-1 Protease as DNA Immunogen against Drug Resistance in HIV-1 Infection: DNA Immunization with Drug Resistant HIV-1 Protease Protects Mice from Challenge with Protease-Expressing Cells

DNA immunization with HIV-1 protease (PR) is advanced for immunotherapy of HIV-1 infection to reduce the number of infected cells producing drug-resistant virus. A consensus PR of the HIV-1 FSU_A strain was designed, expression-optimized, inactivated (D25N), and supplemented with drug resistance (DR) mutations M46I, I54V, and V82A common for FSU_A. PR variants with D25N/M46I/I54V (PR_Ai2mut) and with D25N/M46I/I54V/V82A (PR_Ai3mut) were cloned into the DNA vaccine vector pVAX1, and PR_Ai3mut, into a lentiviral vector for the transduction of murine mammary adenocarcinoma cells expressing luciferase 4T1luc2. BALB/c mice were DNA-immunized by intradermal injections of PR_Ai, PR_Ai2mut, PR_Ai3mut, vector pVAX1, or PBS with electroporation. All PR variants induced specific CD8+ T-cell responses revealed after splenocyte stimulation with PR-derived peptides. Splenocytes of mice DNA-immunized with PR_Ai and PR_Ai2mut were not activated by peptides carrying V82A, whereas splenocytes of PR_Ai3mut-immunized mice recognized both peptides with and without V82A mutation. Mutations M46I and I54V were immunologically silent. In the challenge study, DNA immunization with PR_Ai3mut protected mice from the outgrowth of subcutaneously implanted adenocarcinoma 4T1luc2 cells expressing PR_Ai3mut; a tumor was formed only in 1/10 implantation sites and no metastases were detected. Immunizations with other PR variants were not protective; all mice formed tumors and multiple metastasis in the lungs, liver, and spleen. CD8+ cells of PR_Ai3mut DNA-immunized mice exhibited strong IFN-γ/IL-2 responses against PR peptides, while the splenocytes of mice in other groups were nonresponsive. Thus, immunization with a DNA plasmid encoding inactive HIV-1 protease with DR mutations suppressed the growth and metastatic activity of tumor cells expressing PR identical to the one encoded by the immunogen. This demonstrates the capacity of T-cell response induced by DNA immunization to recognize single DR mutations, and supports the concept of the development of immunotherapies against drug resistance in HIV-1 infection. It also suggests that HIV-1-infected patients developing drug resistance may have a reduced natural immune response against DR HIV-1 mutations causing an immune escape.

Postnatal depletion of maternal cells biases T lymphocytes and natural killer cells' profiles toward early activation in the spleen

The maternal cells transferred into the fetus during gestation persist long after birth in the progeny. These maternal cells have been hypothesized to promote the maturation of the fetal immune system in utero but there are still significant gaps in our knowledge of their potential roles after birth. To provide insights into these maternal cells' postnatal functional roles, we set up a transgenic mouse model to specifically eliminate maternal cells in the neonates by diphtheria toxin injection and confirmed significant depletion in the spleens. We then performed immunophenotyping of the spleens of two-week-old pups by mass cytometry to pinpoint the immune profile differences driven by the depletion of maternal cells in early postnatal life. We observed a heightened expression of markers related to activation and maturation in some natural killer and T cell populations. We hypothesize these results to indicate a potential postnatal regulation of lymphocytic responses by maternal cells. Together, our findings highlight an immunological influence of maternal microchimeric cells postnatally, possibly protecting against adverse hypersensitivity reactions of the neonate at a crucial time of new encounters with self and environmental antigens.

Combined unsupervised and semi-automated supervised analysis of flow cytometry data reveals cellular fingerprint associated with newly diagnosed pediatric type 1 diabetes

An unbiased and replicable profiling of type 1 diabetes (T1D)-specific circulating immunome at disease onset has yet to be identified due to experimental and patient selection limitations. Multicolor flow cytometry was performed on whole blood from a pediatric cohort of 107 patients with new-onset T1D, 85 relatives of T1D patients with 0-1 islet autoantibodies (pre-T1D_LR), 58 patients with celiac disease or autoimmune thyroiditis (CD_THY) and 76 healthy controls (HC). Unsupervised clustering of flow cytometry data, validated by a semi-automated gating strategy, confirmed previous findings showing selective increase of naïve CD4 T cells and plasmacytoid DCs, and revealed a decrease in CD56brightNK cells in T1D. Furthermore, a non-selective decrease of CD3+CD56+ regulatory T cells was observed in T1D. The frequency of naïve CD4 T cells at disease onset was associated with partial remission, while it was found unaltered in the pre-symptomatic stages of the disease. Thanks to a broad cohort of pediatric individuals and the implementation of unbiased approaches for the analysis of flow cytometry data, here we determined the circulating immune fingerprint of newly diagnosed pediatric T1D and provide a reference dataset to be exploited for validation or discovery purposes to unravel the pathogenesis of T1D.

The deficiency in Th2-like Tfh cells affects the maturation and quality of HIV-specific B cell response in viremic infection

Optimal T follicular helper (Tfh) cells function is important to promote the development of germinal centers and maturation of high affinity antigen-specific B cells. We have found that the expression of CXCR3 defines distinct Tfh subsets: CXCR3⁺ Th1-like Tfh cells mainly producing single IFN-γ and dual IL-21/IFN-γ and CXCR3^- Th2-like Tfh cells mainly producing single IL-4 and dual IL-21/IL-4 cytokines. CXCR3^- Th2-like Tfhs are significantly reduced during ongoing HIV replication. While the percentage of Th2-like Tfh cells correlates with that of total and cycling HIV-specific B cells, the percentage of CXCR3⁺ Th1-like Tfhs correlates with HIV-specific B cells expressing T-bet and CXCR3. Of note, only IL-4 and IL-21 cytokines boosted efficient maturation of HIV-specific B cells while IFN-γ induced expression of T-bet and CXCR3 in B cells. Interestingly, total and HIV-specific CXCR3⁺ B cells showed lower rate of somatic hypermutation, as compared to CXCR3^- B cells. Therefore, the imbalance in Th2/Th1-like Tfhs affects B cell responses in viremic HIV infection.

An R-based reproducible and user-friendly preprocessing pipeline for CyTOF data

Mass cytometry (CyTOF) has become a method of choice for in-depth characterization of tissue heterogeneity in health and disease, and is currently implemented in multiple clinical trials, where higher quality standards must be met. Currently, preprocessing of raw files is commonly performed in independent standalone tools, which makes it difficult to reproduce. Here, we present an R pipeline based on an updated version of CATALYST that covers all preprocessing steps required for downstream mass cytometry analysis in a fully reproducible way. This new version of CATALYST is based on Bioconductor’s SingleCellExperiment class and fully unit tested. The R-based pipeline includes file concatenation, bead-based normalization, single-cell deconvolution, spillover compensation and live cell gating after debris and doublet removal. Importantly, this pipeline also includes different quality checks to assess machine sensitivity and staining performance while allowing also for batch correction. This pipeline is based on open source R packages and can be easily be adapted to different study designs. It therefore has the potential to significantly facilitate the work of CyTOF users while increasing the quality and reproducibility of data generated with this technology.

BCyto: A shiny app for flow cytometry data analysis

Flow cytometry (FCM) is a widely used technique to simultaneously measure various characteristics of a cell or particle. Most of reliable software for FCM data analysis are commercially available at high costs. A few R packages are available; however, programming skills are required. In addition, currently, no package offers an interactive user interface (UI) with efficient tools for conventional FCM data analysis. BCyto is an open-source R package/shiny app that allows cell biologists with no coding skills to reliably analyse FCM data in R. BCyto provides intuitive axis transformation for better data visualization, easy visualization of compensation plots and modification of compensation matrices, fast generation of backgating and overlay plots, as well as built-in proliferation and dimensionality reduction tools. BCyto will not only improve accessibility to FCM data analysis but might also serve as an environment for the development of further R-based computational FCM algorithms.

Techniques for Profiling the Cellular Immune Response and Their Implications for Interventional Oncology

In recent years there has been increased interest in using the immune contexture of the primary tumors to predict the patient's prognosis. The tumor microenvironment of patients with cancers consists of different types of lymphocytes, tumor-infiltrating leukocytes, dendritic cells, and others. Different technologies can be used for the evaluation of the tumor microenvironment, all of which require a tissue or cell sample. Image-guided tissue sampling is a cornerstone in the diagnosis, stratification, and longitudinal evaluation of therapeutic efficacy for cancer patients receiving immunotherapies. Therefore, interventional radiologists (IRs) play an essential role in the evaluation of patients treated with systemically administered immunotherapies. This review provides a detailed description of different technologies used for immune assessment and analysis of the data collected from the use of these technologies. The detailed approach provided herein is intended to provide the reader with the knowledge necessary to not only interpret studies containing such data but also design and apply these tools for clinical practice and future research studies.

Cas9-induced large deletions and small indels are controlled in a convergent fashion

Repair of Cas9-induced double-stranded breaks results primarily in formation of small insertions and deletions (indels), but can also cause potentially harmful large deletions. While mechanisms leading to the creation of small indels are relatively well understood, very little is known about the origins of large deletions. Using a library of clonal NGS-validated mouse embryonic stem cells deficient for 32 DNA repair genes, we have shown that large deletion frequency increases in cells impaired for non-homologous end joining and decreases in cells deficient for the central resection gene Nbn and the microhomology-mediated end joining gene Polq. Across deficient clones, increase in large deletion frequency was closely correlated with the increase in the extent of microhomology and the size of small indels, implying a continuity of repair processes across different genomic scales. Furthermore, by targeting diverse genomic sites, we identified examples of repair processes that were highly locus-specific, discovering a role for exonuclease Trex1. Finally, we present evidence that indel sizes increase with the overall efficiency of Cas9 mutagenesis. These findings may have impact on both basic research and clinical use of CRISPR-Cas9, in particular in conjunction with repair pathway modulation.

CRISPR/Cas9 system has revolutionized science and therapy, but DNA damage it causes often goes beyond the desired ’precision editing’. Here, the authors identify general and target specific DNA repair pathways responsible for unwanted mutagenesis.

Tofacitinib Inhibits Leukocyte Trafficking Across the Intestinal Endothelial Barrier in a Specific Cohort of Ulcerative Colitis Patients

The JAK/STAT inhibitor tofacitinib, recently approved for the treatment of ulcerative colitis, is found to modulate the intestinal endothelial barrier functions in directing the leukocyte adhesion and transmigration in ulcerative colitis patients displaying high levels of endothelial STAT3/STAT6 phosphorylation.

Breaking the Immune Complexity of the Tumor Microenvironment Using Single-Cell Technologies

Tumors are not a simple aggregate of transformed cells but rather a complicated ecosystem containing various components, including infiltrating immune cells, tumor-related stromal cells, endothelial cells, soluble factors, and extracellular matrix proteins. Profiling the immune contexture of this intricate framework is now mandatory to develop more effective cancer therapies and precise immunotherapeutic approaches by identifying exact targets or predictive biomarkers, respectively. Conventional technologies are limited in reaching this goal because they lack high resolution. Recent developments in single-cell technologies, such as single-cell RNA transcriptomics, mass cytometry, and multiparameter immunofluorescence, have revolutionized the cancer immunology field, capturing the heterogeneity of tumor-infiltrating immune cells and the dynamic complexity of tenets that regulate cell networks in the tumor microenvironment. In this review, we describe some of the current single-cell technologies and computational techniques applied for immune-profiling the cancer landscape and discuss future directions of how integrating multi-omics data can guide a new "precision oncology" advancement.

Using High-Resolution Differential Cell Counts (HRDCCs) in Bovine Milk and Blood to Monitor the Immune Status over the Entire Lactation Period

Differential cell counts in milk offer a deeper insight into the immunology of the mammary gland and might even provide information about the systemic health status of a dairy cow. Consequently, their potential as a diagnostic method to identify biomarkers has been a subject of research for quite some time. The objective of our study was to closely monitor the immune status of eight healthy dairy cows throughout their whole lactation. For this, high-resolution differential cell counts in milk and blood were determined by means of flow cytometry, which included 10 subpopulations of the 3 main populations of immune cells and their viability. Milk and blood samples were taken twice a week in the first 100 days after calving and once a week during the remaining lactation period: in total, 55 (52–57) blood and 55 (52–57) milk samples per animal. In addition, six well-established routine laboratory biomarkers, i.e., haptoglobin, calcium, and different metabolic parameters (non-esterified fatty acids, β-hydroxybutyric acid, bilirubin, and glutamate dehydrogenase), were analyzed in all blood samples. Furthermore, a standard differential blood cell count was performed on all blood samples. We found substantial differences between cell count progressions in the blood and milk. The distribution of cell populations in the blood remained mostly stable throughout the lactation, albeit at different individual levels. Several cell populations in the milk showed a noticeable dynamic over time, which caused a separation of different lactation stages in clustering analyses. Gamma delta T cells and CD4+ T cells in the milk stood out as they showed characteristic fluctuations during the course of lactation as well as minor changes in the case of inflammation. The determination of a differential cell count has the potential to be a sensitive diagnostic and prognostic tool in bovine milk. Further studies need to show to what extent the method is suitable for routine diagnostics and how to deal with animal-specific differences.

A Machine Learning Approach to the Classification of Acute Leukemias and Distinction From Nonneoplastic Cytopenias Using Flow Cytometry Data

OBJECTIVES

Flow cytometry (FC) is critical for the diagnosis and monitoring of hematologic malignancies. Machine learning (ML) methods rapidly classify multidimensional data and should dramatically improve the efficiency of FC data analysis. We aimed to build a model to classify acute leukemias, including acute promyelocytic leukemia (APL), and distinguish them from nonneoplastic cytopenias. We also sought to illustrate a method to identify key FC parameters that contribute to the model's performance.

METHODS

Using data from 531 patients who underwent evaluation for cytopenias and/or acute leukemia, we developed an ML model to rapidly distinguish among APL, acute myeloid leukemia/not APL, acute lymphoblastic leukemia, and nonneoplastic cytopenias. Unsupervised learning using gaussian mixture model and Fisher kernel methods were applied to FC listmode data, followed by supervised support vector machine classification.

RESULTS

High accuracy (ACC, 94.2%; area under the curve [AUC], 99.5%) was achieved based on the 37-parameter FC panel. Using only 3 parameters, however, yielded similar performance (ACC, 91.7%; AUC, 98.3%) and highlighted the significant contribution of light scatter properties.

CONCLUSIONS

Our findings underscore the potential for ML to automatically identify and prioritize FC specimens that have critical results, including APL and other acute leukemias.

Durable Expansion of TCR-δ Meta-Clonotypes After BCG Revaccination in Humans

Mycobacterium bovis bacille Calmette-Guérin (BCG) has been used for 100 years and prevents disseminated tuberculosis and death in young children. However, it shows only partial efficacy against pulmonary tuberculosis (TB) in adults, so new vaccines are urgently needed. The protective efficacy of BCG depends on T cells, which are typically activated by pathogen-derived protein antigens that bind to highly polymorphic major histocompatibility complex (MHC) molecules. Some T cells recognize non-protein antigens via antigen presenting systems that are independent of genetic background, leading to their designation as donor-unrestricted T (DURT) cells. Whether live whole cell vaccines, like BCG, can induce durable expansions of DURT cells in humans is not known. We used combinatorial tetramer staining, multi-parameter flow cytometry, and immunosequencing to comprehensively characterize the effect of BCG on activation and expansion of DURT cell subsets. We examined peripheral blood mononuclear cells (PBMC) derived from a Phase I study of South African adults in which samples were archived at baseline, 3 weeks, and 52 weeks post-BCG revaccination. We did not observe a change in the frequency of total mucosal-associated invariant T (MAIT) cells, invariant natural killer T (iNKT) cells, germline encoded mycolyl-reactive (GEM) T cells, or γδ T cells at 52 weeks post-BCG. However, immunosequencing revealed a set of TCR-δ clonotypes that were expanded at 52 weeks post-BCG revaccination. These expanded clones expressed the Vδ2 gene segment and could be further defined on the basis of biochemical similarity into several 'meta-clonotypes' that likely recognize similar epitopes. Our data reveal that BCG vaccination leads to durable expansion of DURT cell clonotypes despite a limited effect on total circulating frequencies in the blood and have implications for defining the immunogenicity of candidate whole cell TB vaccines.

Advances in automated real-time flow cytometry for monitoring of bioreactor processes

Flow cytometry and its technological possibilities have greatly advanced in the past decade as analysis tool for single cell properties and population distributions of different cell types in bioreactors. Along the way, some solutions for automated real-time flow cytometry (ART-FCM) were developed for monitoring of bioreactor processes without operator interference over extended periods with variable sampling frequency. However, there is still great potential for ART-FCM to evolve and possibly become a standard application in bioprocess monitoring and process control. This review first addresses different components of an ART-FCM, including the sampling device, the sample-processing unit, the unit for sample delivery to the flow cytometer and the settings for measurement of pre-processed samples. Also, available algorithms are presented for automated data analysis of multi-parameter fluorescence datasets derived from ART-FCM experiments. Furthermore, challenges are discussed for integration of fluorescence-activated cell sorting into an ART-FCM setup for isolation and separation of interesting subpopulations that can be further characterized by for instance omics-methods. As the application of ART-FCM is especially of interest for bioreactor process monitoring, including investigation of population heterogeneity and automated process control, a summary of already existing setups for these purposes is given. Additionally, the general future potential of ART-FCM is addressed.

Reversible Shielding and Immobilization of Liposomes and Viral Vectors by Tailored Antibody-Ligand Interactions

Controlling the time and dose of nanoparticulate drug delivery by administration of small molecule drugs holds promise for efficient and safer therapies. This study describes a versatile approach of exploiting antibody-ligand interactions for the design of small molecule-responsive nanocarrier and nanocomposite systems. For this purpose, antibody fragments (scFvs) specific for two distinct small molecule ligands are designed. Subsequently, the surface of nanoparticles (liposomes or adeno-associated viral vectors, AAVs) is modified with these ligands, serving as anchor points for scFv binding. By modifying the scFvs with polymer tails, they can act as a non-covalently bound shielding layer, which is recruited to the anchor points on the nanoparticle surface and prevents interactions with cultured mammalian cells. Administration of an excess of the respective ligand triggers competitive displacement of the shielding layer from the nanoparticle surface and restores nanoparticle-cell interactions. The same principle is applied for developing hydrogel depots that can release integrated AAVs or liposomes in response to small molecule ligands. The liberated nanoparticles subsequently deliver their cargoes to cells. In summary, the utilization of different antibody-ligand interactions, different nanoparticles, and different release systems validates the versatility of the design concept described herein.

CD4 and CD8 co-receptors modulate functional avidity of CD1b-restricted T cells

T cells recognize mycobacterial glycolipid (mycolipid) antigens presented by CD1b molecules, but the role of CD4 and CD8 co-receptors in mycolipid recognition is unknown. Here we show CD1b-mycolipid tetramers reveal a hierarchy in which circulating T cells expressing CD4 or CD8 co-receptor stain with a higher tetramer mean fluorescence intensity than CD4-CD8- T cells. CD4+ primary T cells transduced with mycolipid-specific T cell receptors bind CD1b-mycolipid tetramer with a higher fluorescence intensity than CD8+ primary T cells. The presence of either CD4 or CD8 also decreases the threshold for interferon-γ secretion. Co-receptor expression increases surface expression of CD3ε, suggesting a mechanism for increased tetramer binding and activation. Targeted transcriptional profiling of mycolipid-specific T cells from individuals with active tuberculosis reveals canonical markers associated with cytotoxicity among CD8+ compared to CD4+ T cells. Thus, expression of co-receptors modulates T cell receptor avidity for mycobacterial lipids, leading to in vivo functional diversity during tuberculosis disease.

Distal and proximal cis-regulatory elements sense X chromosome dosage and developmental state at the Xist locus

Developmental genes such as Xist, which initiates X chromosome inactivation, are controlled by complex cis-regulatory landscapes, which decode multiple signals to establish specific spatiotemporal expression patterns. Xist integrates information on X chromosome dosage and developmental stage to trigger X inactivation in the epiblast specifically in female embryos. Through a pooled CRISPR screen in differentiating mouse embryonic stem cells, we identify functional enhancer elements of Xist at the onset of random X inactivation. Chromatin profiling reveals that X-dosage controls the promoter-proximal region, while differentiation cues activate several distal enhancers. The strongest distal element lies in an enhancer cluster associated with a previously unannotated Xist-enhancing regulatory transcript, which we named Xert. Developmental cues and X-dosage are thus decoded by distinct regulatory regions, which cooperate to ensure female-specific Xist upregulation at the correct developmental time. With this study, we start to disentangle how multiple, functionally distinct regulatory elements interact to generate complex expression patterns in mammals.

New interpretable machine-learning method for single-cell data reveals correlates of clinical response to cancer immunotherapy

We introduce a new method for single-cell cytometry studies, FAUST, which performs unbiased cell population discovery and annotation. FAUST processes experimental data on a per-sample basis and returns biologically interpretable cell phenotypes, making it well suited for the analysis of complex datasets. We provide simulation studies that compare FAUST with existing methodology, exemplifying its strength. We apply FAUST to data from a Merkel cell carcinoma anti-PD-1 trial and discover pre-treatment effector memory T cell correlates of outcome co-expressing PD-1, HLA-DR, and CD28. Using FAUST, we then validate these correlates in cryopreserved peripheral blood mononuclear cell samples from the same study, as well as an independent CyTOF dataset from a published metastatic melanoma trial. Finally, we show how FAUST's phenotypes can be used to perform cross-study data integration in the presence of diverse staining panels. Together, these results establish FAUST as a powerful new approach for unbiased discovery in single-cell cytometry.

•

An interpretable machine-learning method for cytometry data analysis is developed

•

Using this, candidate biomarkers of response to therapy are identified and visualized

•

The method is used to validate our findings on two additional cytometry datasets

•

It is shown how to integrate findings across datasets with heterogeneous marker panels

Our article introduces a new method, FAUST, which combines novel algorithms for clustering, cluster matching, variable selection, and feature selection. While these algorithms were developed for application to high-dimensional single-cell data—and our article validates this application area with multiple case studies—they are general purpose and can be applied to any collection of related real-valued matrices one wishes to partition.

Some useful features of these algorithms to the broader data science community include the following: they estimate the number of clusters across a dataset, they can be applied independently to each matrix in the set of matrices one wishes to cluster, they match clusters across matrices on the basis of data-driven annotations, and the annotations are interpretable in relation to the initial measurement variables. We provide an open-source implementation of our method, https://github.com/RGLab/FAUST, targeting data structures optimized for use in cytometry data analysis.

Development of an advanced flow cytometry based high-resolution immunophenotyping method to benchmark early immune response in dairy cows

The determination of the somatic cell count of a milk sample is one of the most common methods to monitor udder health of a dairy cow. However, this procedure does not take into account the fact that cells in milk present a great variety of different cell types. The objective of our study was to establish a high-resolution differential cell count (HRDCC) by means of flow cytometry in blood and milk. We were able to detect ten subpopulations among the three main populations of immune cells and to determine their viability. Additionally, blood samples were analyzed for common laboratory biomarkers, i.e. differential blood counts, haptoglobin levels and several metabolic parameters. In this first feasibility study, we used three different vaccines to stimulate the immune system of five healthy cows each. Samples were collected shortly before, in between and after the vaccinations. Using multivariate statistical methods we saw a diagnostic benefit when HRDCCs were included compared to only the standard laboratory parameters. The impacts of all three vaccinations on the immune system were visible in blood HRDCCs as well as in milk HRDCCs. Cluster of Differentiation 8⁺ (CD8⁺) T cells, B cells and monocyte/macrophage subpopulations were among the most important and statistically relevant parameters for all treatments in both biofluids. Moreover, in one of the treatment groups intermediate monocytes showed a significant increase after both vaccinations. Although the use of HRDCC in blood or milk was shown to be highly relevant for early systemic diagnostic, to confirm these subpopulations further investigations in cows of different breed, lactation stage or health status are required.

FlowCT for the analysis of large immunophenotypic datasets and biomarker discovery in cancer immunology

Large-scale immune monitoring is becoming routinely used in clinical trials to identify determinants of treatment responsiveness, particularly to immunotherapies. Flow cytometry remains one of the most versatile and high throughput approaches for single-cell analysis; however, manual interpretation of multidimensional data poses a challenge to capture full cellular diversity and provide reproducible results. We present FlowCT, a semi-automated workspace empowered to analyze large datasets that includes pre-processing, normalization, multiple dimensionality reduction techniques, automated clustering and predictive modeling tools. As a proof of concept, we used FlowCT to compare the T cell compartment in bone marrow (BM) vs peripheral blood (PB) of patients with smoldering multiple myeloma (MM); identify minimally-invasive immune biomarkers of progression from smoldering to active MM; define prognostic T cell subsets in the BM of patients with active MM after treatment intensification; and assess the longitudinal effect of maintenance therapy in BM T cells. A total of 354 samples were analyzed and immune signatures predictive of malignant transformation in 150 smoldering MM patients (hazard ratio [HR]: 1.7; P <.001), and of progression-free (HR: 4.09; P <.0001) and overall survival (HR: 3.12; P =.047) in 100 active MM patients, were identified. New data also emerged about stem cell memory T cells, the concordance between immune profiles in BM vs PB and the immunomodulatory effect of maintenance therapy. FlowCT is a new open-source computational approach that can be readily implemented by research laboratories to perform quality-control, analyze high-dimensional data, unveil cellular diversity and objectively identify biomarkers in large immune monitoring studies.

High-dimensional immune phenotyping of blood cells by mass cytometry in patients infected with hepatitis C virus

OBJECTIVES

Chronic hepatitis C virus (HCV) infection affects the immune system. Whether elimination of HCV with direct-acting antivirals (DAA) restores immunity is unclear. We used mass cytometry to get a broad and in-depth assessment of blood cell populations of patients with chronic HCV before and after DAA therapy.

METHODS

Before and 12 weeks after sustained virological response (SVR12) to DAA therapy, 22 cell populations were analysed by mass cytometry in blood collected from ten healthy control individuals and 20 HCV-infected patients with (ten patients) or without (ten patients) human immunodeficiency virus (HIV) infection.

RESULTS

HCV infection altered the frequency of 14/22 (64%) blood cell populations. At baseline, the frequencies (median, interquartile range (IQR); control, HCV, HCV/HIV) of intermediate monocytes (1.2, IQR 0.47-1.46; 1.76, IQR 0.83-2.66; 0.78, IQR 0.28-1.77), non-classical monocytes (1.11, IQR 0.49-1.26; 0.9, IQR 0.18-0.99; 0.54, IQR 0.28-1.77), conventional dendritic cells type 2 (0.55, IQR 0.35-0.59; 0.31, IQR 0.16-0.38; 0.19, IQR 0.11-0.36) and CD56^dim natural killer cells (8.08, IQR 5.34-9.79; 4.72, IQR 2.59-6.05) 3.61, IQR 2.98-5.07) were reduced by 35% to 65%, particularly in HCV/HIV co-infected patients. In contrast, activated double-negative T cells (0.07, IQR 0.06-0.10; 0.10, IQR 0.09-0.19; 0.19, IQR 0.12-0.25), activated CD4 T cells (0.28, IQR 0.21-0.36; 0.56, IQR 0.33-0.77; 0.40, IQR 0.22-0.53) and activated CD8 T cells (0.23, IQR 0.14-0.42; 0.74, IQR 0.30-1.65; 0.80, IQR 0.58-1.16) were increased 1.4 to 3.5 times. Upon stimulation with Toll-like receptor ligands, the expression of cytokines was up-regulated in 7/9 (78%) and 17/19 (89%) of the conditions in HCV- and HCV/HIV-infected patients, respectively. Most alterations persisted at SVR12.

CONCLUSIONS

Chronic HCV and HCV/HIV infections induce profound and durable perturbations of innate and adaptive immune homeostasis.

Longitudinal immune dynamics of mild COVID-19 define signatures of recovery and persistence

SARS-CoV-2 has infected over 200 million and caused more than 4 million deaths to date. Most individuals (>80%) have mild symptoms and recover in the outpatient setting, but detailed studies of immune responses have focused primarily on moderate to severe COVID-19. We deeply profiled the longitudinal immune response in individuals with mild COVID-19 beginning with early time points post-infection (1-15 days) and proceeding through convalescence to >100 days after symptom onset. We correlated data from single cell analyses of peripheral blood cells, serum proteomics, virus-specific cellular and humoral immune responses, and clinical metadata. Acute infection was characterized by vigorous coordinated innate and adaptive immune activation that differed in character by age (young vs. old). We then characterized signals associated with recovery and convalescence to define and validate a new signature of inflammatory cytokines, gene expression, and chromatin accessibility that persists in individuals with post-acute sequelae of SARS-CoV-2 infection (PASC).

Unsupervised flow cytometry analysis in hematological malignancies: A new paradigm

Ever since hematopoietic cells became "events" enumerated and characterized in suspension by cell counters or flow cytometers, researchers and engineers have strived to refine the acquisition and display of the electronic signals generated. A large array of solutions was then developed to identify at best the numerous cell subsets that can be delineated, notably among hematopoietic cells. As instruments became more and more stable and robust, the focus moved to analytic software. Almost concomitantly, the capacity increased to use large panels (both with mass and classical cytometry) and to apply artificial intelligence/machine learning for their analysis. The combination of these concepts raised new analytical possibilities, opening an unprecedented field of subtle exploration for many conditions, including hematopoiesis and hematological disorders. In this review, the general concepts and progress achieved in the development of new analytical approaches for exploring high-dimensional data sets at the single-cell level will be described as they appeared over the past few years. A larger and more practical part will detail the various steps that need to be mastered, both in data acquisition and in the preanalytical check of data files. Finally, a step-by-step explanation of the solution in development to combine the Bioconductor clustering algorithm FlowSOM and the popular and widely used software Kaluza® (Beckman Coulter) will be presented. The aim of this review was to point out that the day when these progresses will reach routine hematology laboratories does not seem so far away.

Analyzing high-dimensional cytometry data using FlowSOM

The dimensionality of cytometry data has strongly increased in the last decade, and in many situations the traditional manual downstream analysis becomes insufficient. The field is therefore slowly moving toward more automated approaches, and in this paper we describe the protocol for analyzing high-dimensional cytometry data using FlowSOM, a clustering and visualization algorithm based on a self-organizing map. FlowSOM is used to distinguish cell populations from cytometry data in an unsupervised way and can help to gain deeper insights in fields such as immunology and oncology. Since the original FlowSOM publication (2015), we have validated the tool on a wide variety of datasets, and to write this protocol, we made use of this experience to improve the user-friendliness of the package (e.g., comprehensive functions replacing commonly required scripts). Where the original paper focused mainly on the algorithm description, this protocol offers user guidelines on how to implement the procedure, detailed parameter descriptions and troubleshooting recommendations. The protocol provides clearly annotated R code, and is therefore relevant for all scientists interested in computational high-dimensional analyses without requiring a strong bioinformatics background. We demonstrate the complete workflow, starting from data preparation (such as compensation, transformation and quality control), including detailed discussion of the different FlowSOM parameters and visualization options, and concluding with how the results can be further used to answer biological questions, such as statistical comparison between groups of interest. An average FlowSOM analysis takes 1-3 h to complete, though quality issues can increase this time considerably.

Multi-omics reveals clinically relevant proliferative drive associated with mTOR-MYC-OXPHOS activity in chronic lymphocytic leukemia

Chronic Lymphocytic Leukemia (CLL) has a complex pattern of driver mutations and much of its clinical diversity remains unexplained. We devised a method for simultaneous subgroup discovery across multiple data types and applied it to genomic, transcriptomic, DNA methylation and ex-vivo drug response data from 217 Chronic Lymphocytic Leukemia (CLL) cases. We uncovered a biological axis of heterogeneity strongly associated with clinical behavior and orthogonal to the known biomarkers. We validated its presence and clinical relevance in four independent cohorts (n=547 patients). We find that this axis captures the proliferative drive (PD) of CLL cells, as it associates with lymphocyte doubling rate, global hypomethylation, accumulation of driver aberrations and response to pro-proliferative stimuli. CLL-PD was linked to the activation of mTOR-MYC-oxidative phosphorylation (OXPHOS) through transcriptomic, proteomic and single cell resolution analysis. CLL-PD is a key determinant of disease outcome in CLL. Our multi-table integration approach may be applicable to other tumors whose inter-individual differences are currently unexplained.

Acquisition of High-Quality Spectral Flow Cytometry Data

Flow cytometry allows the visualization of physical, functional, and/or biological properties of cells including antigens, cytokines, size, and complexity. With increasingly large flow cytometry panels able to analyze up to 50 parameters, there is a need to standardize flow cytometry protocols to achieve high-quality data that can be input into analysis algorithms. Without this clean data, algorithms may incorrectly categorize the cell populations present in the samples. In this protocol, we outline a comprehensive methodology to prepare samples for polychromatic flow cytometry. The use of multiple washing steps and rigorous controls creates high-quality data with good separation between cell populations. Experimental data acquired using this protocol can be analyzed via computational algorithms that perform end-to-end analysis. © 2020 by John Wiley & Sons, Inc. Basic Protocol 1: Preparation of single-cell suspension for flow cytometry Support Protocol 1: Lung preparation Support Protocol 2: Counting cells on a flow cytometer Basic Protocol 2: Surface and intracellular flow cytometry staining Support Protocol 3: Single-color bead controls.

Agile workflow for interactive analysis of mass cytometry data

MOTIVATION

Single-cell proteomics technologies, such as mass cytometry, have enabled characterization of cell-to-cell variation and cell populations at a single-cell resolution. These large amounts of data, require dedicated, interactive tools for translating the data into knowledge.

RESULTS

We present a comprehensive, interactive method called Cyto to streamline analysis of large-scale cytometry data. Cyto is a workflow-based open-source solution that automates the use of state-of-the-art single-cell analysis methods with interactive visualization. We show the utility of Cyto by applying it to mass cytometry data from peripheral blood and high-grade serous ovarian cancer (HGSOC) samples. Our results show that Cyto is able to reliably capture the immune cell sub-populations from peripheral blood and cellular compositions of unique immune- and cancer cell subpopulations in HGSOC tumor and ascites samples.

AVAILABILITYAND IMPLEMENTATION

The method is available as a Docker container at https://hub.docker.com/r/anduril/cyto and the user guide and source code are available at https://bitbucket.org/anduril-dev/cyto.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.