Manual Gating Research Articles

Analysis of the Leishmania mexicana promastigote cell cycle using imaging flow cytometry provides new insights into cell cycle flexibility and events of short duration.

Promastigote Leishmania mexicana have a complex cell division cycle characterised by the ordered replication of several single-copy organelles, a prolonged S phase and rapid G2 and cytokinesis phases, accompanied by cell cycle stage-associated morphological changes. Here we exploit these morphological changes to develop a high-throughput and semi-automated imaging flow cytometry (IFC) pipeline to analyse the cell cycle in live L. mexicana. Firstly, we demonstrate that, unlike several other DNA stains, Vybrant™ DyeCycle™ Orange (DCO) is non-toxic and enables quantitative DNA imaging in live promastigotes. Secondly, by tagging the orphan spindle kinesin, KINF, with mNeonGreen, we describe KINF's cell cycle-dependent expression and localisation. Then, by combining manual gating of DCO DNA intensity profiles with automated masking and morphological measurements of parasite images, visual determination of the number of flagella per cell, and automated masking and analysis of mNG:KINF fluorescence, we provide a newly detailed description of L. mexicana promastigote cell cycle events that, for the first time, includes the durations of individual G2, mitosis and post-mitosis phases, and identifies G1 cells within the first 12 minutes of the new cell cycle. Our custom-developed masking and gating scheme allowed us to identify elusive G2 cells and to demonstrate that the CDK-inhibitor, flavopiridol, arrests cells in G2 phase, rather than mitosis, providing proof-of-principle of the utility of IFC for drug mechanism-of-action studies. Further, the high-throughput nature of IFC allowed the close examination of promastigote cytokinesis, revealing considerable flexibility in both the timing of cytokinesis initiation and the direction of furrowing, in contrast to the related kinetoplastid parasite, Trypanosoma brucei and many other cell types. Our new pipeline offers many advantages over traditional methods of cell cycle analysis such as fluorescence microscopy and flow cytometry and paves the way for novel high-throughput analysis of Leishmania cell division.

Development of a Machine Learning Decision Support System for Clinical Flow Cytometry

Abstract High parameter flow cytometry is a highly utilized tool in both research and clinical settings for accurate immunophenotyping in immunology, oncology, and many other disciplines. With increasing dimensionality, manual analysis of data becomes more complex and time consuming. Over the past decade, several automated methods have been developed for cell population clustering and dimensionality reduction of high parameter flow cytometry data which have sped up and simplified the discovery of cell populations not observed by manual gating strategies. However, the input and output of such tools are stochastic in nature, thus making their results difficult to reuse with novel samples such as diagnostic patient samples. Several groups have recently attempted to utilize machine learning algorithms to analyze clinical flow cytometry data. These studies focused on training machine learning models with a goal of predicting a specific diagnosis for each sample. This approach can be limiting when the diagnostic features are not perfectly met to predict a specific diagnosis as well as the absence of the patient's clinical data. To our knowledge, no studies have utilized machine learning to identify and classify individual cells based on their raw fluorescence signature to various cellular populations, whether normal or abnormal, to aid a diagnostician to interpret their significance in a clinical context. In this study, we utilized a data set of ~240 million individual cells from ~9,500 flow cytometry files from 2,300 patients to train a machine learning model to predict cell population classifications as annotated by expert hematopathologists. These annotations were implemented into an end-to-end machine learning based decision support system allowing for rapid, automated analysis of high dimensional flow cytometry data with an output aiding the diagnostician in the generation of a diagnosis. A clinical validation highlighted the robustness of this methodology with 100% sensitivity and 94% specificity in the identification of abnormal B-cell populations on a cohort of 1,500 patients. This decision support system also identifies abnormal T-cell and myeloid cell populations with 100% sensitivity. Overall, the decision support system saves significant time as it removes the need for manual gating and analysis, and aids the diagnostician in identifying discrete cell populations and offering descriptive information about any abnormal populations identified.

BinaryClust2: An Integrated Computational Pipeline for Systematic Mining of Mass Cytometry Data to Assist Deep Immune Profiling in Haematological Research

Introduction: The key role of immune system in modulating treatment efficacy and clinical outcomes in the settings of haematological diseases has been well recognized through the application of flow cytometry technologies over years. Mass cytometry (MC) is a next generation cytometry platform to simultaneously profile over 40 protein markers on single-cell resolution, which facilitates clinical research on immune monitoring and biomarker discovery. Nonetheless, the high-dimensional data generated present a daunting challenge for clinical investigators. Therefore, we describe BinaryClust2 (https://github.com/JingAnyaSun/BinaryClust2), an R-based computational framework integrated with the state-of-the-art algorithms and novel cell lineage characterisation method to assist comprehensive exploration of liquid MC data. Method: Here, we present the implementation of BinaryClust2 to three example MC datasets: In-house dataset from peripheral blood mononuclear cells (PBMCs) of 9 baseline myeloproliferative neoplasm (MPN) patients (~4 million cells), previously published data from 11 MPN patients receiving influenza vaccine (~0.2million cells), and published dataset of 59 covid patients and 23 healthy donors (~2million cells). BinaryClust2 has a streamlined analytical workflow, which comprises the following steps (Figure 1A): Step 1: Quality control, batch effect evaluation and correction Data quality control is a separate step before downstream analysis which includes diagnostic plots and batch effect examination. Algorithms CytofRUV and CytoNorm are available in the pipeline to remove unwanted variations caused by batch effects. Step 2: Semi-supervised identification of main cell types BinaryClust2 adopts a knowledge-based semi-supervised approach to predict main cell types. Users are required to provide a simple marker expression matrix of pre-defined cell types along with fcs files and metadata to construct a SingleCellExperiment (SCE) object, then the embedded algorithm can automatically classify cell populations without manual annotation. Step 3: In-depth interrogation and differential testing Specific population can be further extracted from whole cells and subject to in-depth exploration using unsupervised algorithms for subpopulation discovery. Dimensionality reduction tools UMAP and TSNE, unsupervised clustering methods Phenograph and flowSOM, and various data visualization plotting functions are offered in BinaryClust2. For statistical analysis, multiple study group comparison (n>2) of cell abundance and functional marker expression is supported via Kruskal Wallis test with multiple testing correction and post hoc analysis. Results: The performance of the semi-supervised classification function was tested independently in the MPN and influenza PBMC datasets, 7 main cell lineages were identified with accuracy comparable to manual gating by human experts (Figure 1B): average F-measure reached 0.93 and 0.98 respectively. Moreover, taking manual gating as ground truth reference, BinaryClust2 outperformed the unsupervised approach flowSOM concerning accuracy (F-measure: 0.93 vs. 0.70) and speed (140s vs 339s) in MPN dataset, while remaining equivalent to the well-performing semi-supervised approach LDA in accuracy (F-measure: 0.93 vs. 0.93) but faster in runtime (140s vs 595s), as shown in Table 1. Application to covid-19 dataset by Chevrier et al. achieved reproducible results and additional discoveries. 13 main cell types were characterised, abundance of B cells, Basophils, cDCs, DN T cells, Monocytes, Neutrophils, NK cells, pDCs, CD8 T cells obtained statistical significance (all P<0.05) among study conditions (healthy, mild covid, severe covid). We also grouped markers reflecting functional status of immune cells and found Granzyme B expression was significantly increased in the majority of main immune cells of covid patients. Phenograph was further applied in neutrophils and monocytes and returned 14 and 16 subsets respectively. Conclusion: Overall, BinaryClust2 incorporates expert's prior biological knowledge in a semi-supervised fashion to accurately deconvolute well-defined main cell lineages, while also preserving the potential of unsupervised approaches to discover novel cell subsets and providing a user-friendly toolset to remove the analytical barrier for high-dimensional immune profiling.

Reduced Immune-Cell Infiltration in MHC Class I Negative DLBCL

Background Diffuse large B-cell lymphoma (DLBCL) is an aggressive malignancy and the most common type of non-Hodgkin B-cell lymphoma. It is a genetically, biologically and clinically heterogeneous disease and current research focuses on identifying stratification biomarkers to identify high-risk patients and guide selection of therapy. The genetic landscape of DLBCL is well characterized, but how these mutations affect the lymphoma microenvironment is not fully understood. We aimed to categorize DLBCL cases based on tumor microenvironment (TME) features and to identify how tumor cell features including loss of MHC expression shape the TME. Methods Diagnostic biopsies from 58 high-risk de novo DLBCL NOS from two Nordic phase II clinical trials (PMID: 32380536, PMID: 23247661) were imaged by Hyperion imaging mass cytometry. Single-cell segmentation was followed by clustering of 14 lineage markers and manual annotation to identify eight cell types. Using CytoMAP (PMID: 32320656), the cells were divided into 20-µm-radius neighborhoods and clustered to define six types of tissue regions (Figure 1). Unsupervised hierarchical clustering of the region prevalence in each image was used to classify different types of tumor microenvironments. To identify tumor cell features, tumor-specific expression of markers related to immune evasion (CD54, PD-L1) and prognostic markers (FOXP1, Ki67) were quantified by manual gating of positive cells. Cases were classified as MHC class I loss when less than 50% of the tumor cells expressed HLA A or HLA B, and Ki67-positive when more than 10% of the tumor cells were gated as positive. The LymphGen algorithm was used to classify tumor genotypes. Results Clustering of the region prevalence in each image revealed six types of tumor microenvironments; immune-cell depleted (47%), tumor/immune cell mixed (29%) and four TMEs with low tumor-cell fractions that were dominated by CD4 T cells (7%), CD8 T cells (7%), M1/M2 macrophages (5%) or collagen-positive stromal cells (5%). In cases classified as immune-cell depleted, more than 70% of the cells were tumor cells and less than 15% were infiltrating immune cells. The four groups with low tumor-cell fractions were merged for statistical analyses. TME subclasses were not correlated with tumor genotypes, but DLBCL cases classified as immune-cell depleted were overrepresented among the Ki67-positive cases. Loss of MHC class I was seen in 34% of the cases and was associated with the immune-cell depleted TME subclass. Furthermore, MHC class I loss correlated with lower infiltration of CD4 and CD8 T cells. Tumor-cell specific expression of MHC class II, PD-L1, CD54, or FOXP1 did not significantly correlate with the TME subclasses. Conclusion In conclusion, our study contributes to improved understanding of the DLBCL microenvironment through mapping of the diversity of immune-cell composition. We found that loss of MHC class I expression was associated with an immune cell-depleted TME. The prominent loss of MHC has potential impact for the selection of patients for CAR T-cell therapy vs. bispecific antibodies/T-cell engagers. Together, refined classification tools of DLBCL based on genetic aberrations and microenvironment architecture may improve strategies of personalized medicine.

Immunomodulatory Effects of Chemo-Immunotherapy ± Idelalisib in Chronic Lymphocytic Leukaemia (CLL)

Introduction: Chronic lymphocytic leukaemia (CLL) is characterised by a dysfunctional immune profile. However, the precise impact of therapy on this altered immune state, both short and long term, remains unclear. The collection of pre-treatment and longitudinal blood samples in the NCRI phase III RIAltO trial (NCT01678430) offered a unique opportunity to address this question in the context of frontline chemo-immunotherapy (CIT) containing bendamustine or chlorambucil. The factorial randomisation to idelalisib versus placebo also provided a framework for investigating the immunomodulatory effects of PI3Kδ inhibition in this setting. Methods: We conducted mass cytometry analysis on samples obtained from CLL patients enrolled in the RIAltO trial. Patients were randomly assigned to receive ofatumumab plus either bendamustine or chlorambucil, with or without idelalisib. Samples were collected before treatment and at early (median 5.9 [IQR 5.7 - 6.4] months) and late (18.5 [16.1 - 22.6] months) post-treatment timepoints. Three healthy controls (HC) were included for comparison. CD3 + T cells were analysed using a customised CyTOF panel comprising 36 markers. Uniform manifold approximation and projection (UMAP) and manual gating analysis were utilised to evaluate global and specific T-cell changes. Results: In all, 177 samples from 79 CLL patients, obtained before or after CIT, were analysed. Baseline characteristics were similar between the different treatment arms. Compared to HC, treatment-naïve CLL patients exhibited a distinct T-cell profile with increased regulatory T cells, decreased naïve and central memory (CM) cells, elevated effector memory (EM) cells, reduced non-senescent cells, and increased expression of HLA-DR, PD-1, TOX, T-bet, and Eomes. Following CIT, the T-cell profile was significantly altered but remained distinct from that of HCs. CIT-induced changes persisted at the late post-treatment timepoint, indicating a sustained effect. Analysis of specific T-cell populations showed that, although some CLL-associated features became less prominent following CIT (expression of PD-1, TOX and TIGIT), most were enhanced (shift from naïve and CM cells towards a terminal effector phenotype; shift from a non-senescent to a terminally senescent phenotype; increased HLA-DR and T-bet expression). The post-treatment immune profiles were similar in patients receiving bendamustine or chlorambucil. Despite its immunomodulatory effects when used as a single agent, idelalisib also induced minimal additional changes. Furthermore, the post-treatment immune profile was similar in patients achieving minimal residual disease (MRD) negativity versus those who remained MRD positive, suggesting a direct pharmacodynamic effect. Finally, non-metric multidimensional scaling (NMDS) analysis demonstrated that CIT shifted the overall CLL-associated immune profile further away from that of HC, and that the shift was independent of chemotherapy allocation, idelalisib, or the level of CLL cytoreduction achieved. Conclusion: CIT results in profound immune changes that persists up to 18.5 [16.1 - 22.6] months following treatment. Furthermore, the overall effect of these changes further differentiates the CLL-associated immune profile from that of HC, deepening the existing T-cell dysfunction irrespective of chemotherapy choice, or the addition of idelalisib. Our study has important potential implications for patients receiving CIT in a range of clinical settings and supports the move towards more targeted agents that are less likely to cause indiscriminate immune perturbation.

Cryopreservation Changes the Immune Effector Cell Composition of Peripheral Blood Stem Cell Grafts: An Analysis from the DKMS and NMDP Graft Composition Study

The impact of the cellular composition of PBSC grafts on the immune reconstitution as well as graft-versus-host and graft-versus-leukemia effects after allogeneic hematopoietic cell transplantation (alloHCT) is only poorly understood. DKMS and NMDP joined forces to comprehensively characterize the cellular composition of 2,000 PBSC products using a 34-color flow cytometry panel to define immune effector cell subsets of interest and hematopoietic stem cells. In parallel, we have created a biobank of unstimulated donor PBMC and PBSC graft samples for future analyses. As part of this larger effort, we investigated the effect of cryopreservation on the cellular composition of PBSC products. Cryopreservation of PBSC grafts was used frequently during the COVID19-pandemic and facilitated more flexible timelines for transplantation. Still, one third of the US patients and 15% of the German patients continue to receive cryopreserved PBSC grafts. Ongoing studies are investigating the impact of cryopreservation on immune-effector cell function and the outcome of hematopoietic stem cell transplantation. Several observational studies have reported delayed neutrophil and platelet engraftment with cryopreserved products. More recently, less chronic GVHD was reported for patients who had received cryopreserved PBSC products. Cryopreservation affects the cellular composition and hence has the potential to change immune effector cell functions of the graft, but available data on the alterations of cell population frequencies are scarce. We analyzed the cellular composition of 20 fresh grafts and compared them to their frozen and thawed counterparts. Cryopreservation of product samples followed routine clinical procedures using a 10% DMSO containing cryo-medium and a controlled rate freezing. After thawing at 37°C, cells were washed, stained with a 34-parameter phenotyping panel, fixed and analyzed immediately on a full spectrum flow cytometer. Following spectral unmixing and manual spillover correction, proportions of immune cells were assessed using manual gating. First, we investigated the relative frequencies of 10 non-overlapping individual cell populations relative to viable CD45+ cells, including several lymphocyte populations, monocytes, granulocytes, hematopoietic stem cells (HSC) and dendritic cells. Granulocytes showed the largest decrease after cryopreservation and thawing (18.2 % vs 3.5 %; p<0.001). The loss of granulocytes results in an increase of relative frequencies of many other immune cell populations including monocytes (35.4 % vs 44.7 %; p=0.018), B cells (5.8 % vs 9.6 %; p=0.006), and HSC (1.0 % vs 1.5 %; p<0.001). We therefore also calculated changes relative to the viable non-granulocyte cells. By doing so, a significant relative increase of cells was observed for HSC (1.2 % vs 1.6 %; p<0.001), whereas monocyte and B cell frequencies changed only marginally. This suggests that HSC are more resilient to freezing than other cell types. When comparing the proportions of 40 other immune cell (sub-) populations relative to their parent population in the gating hierarchy, we found a profound and significant decrease of the relative frequencies for CD16+ myeloid dendritic cells (mDCs, 68 % vs 30 % of all mDCs; p<0.001) and granulocytic myeloid derived suppressor cells (gMDSCs, 2 % vs 1 % of all MDSCs; p<0.001). In an independent set of analyses performed at Roswell Park Comprehensive Cancer Center, Buffalo, comparable results were obtained with a significant increase of HSC, a significant decrease of CD16+ mDCs, and a trend of a decrease of gMDSCs. As CD16+ mDCs may have a role in the development of chronic GVHD after alloHCT, this decrease may be meaningful. In summary, we found that cryopreservation of PBSC allografts resulted in a decreased relative frequency of granulocytes, and CD16+ mDCs, as well as an increased relative frequency of HSCs. The results provide some insight into the impact of cryopreservation on PBSC graft composition and warrant further efforts to explore the potential mechanisms for the differential outcomes observed using fresh vs cryopreserved PBSCs.

Mezigdomide Treatment in Relapsed-Refractory Myeloma Patients Shifts Bone Marrow NK and T Cell Populations from Exhaustion to Activation

Background The effect of immunomodulatory drugs (IMiDs) and novel cereblon E3 ligase modulators (CELMoDs) on the tumour microenviroment (TME) of multiple myeloma (MM) patients remains poorly understood. A recent multi-center phase 1/2 trial (CC-92480-MM-001) of CELMoD Mezigdomide (MEZI) demonstrated clinical response in triple-class refractory, heavily pre-treated MM patients (Richardson et al, EHA abstracts 2020). Here, we report results from high dimensional mass cytometry immune profiling of baseline and treatment samples from relapsed-refractory MM (RRMM) patients to understand the effect of MEZI on the bone marrow (BM) TME. In addition, we analyze baseline samples to evaluate BM immune phenotypes that may correlate with response to MEZI. Methods We identified 45 RRMM patients with paired bone marrow aspirate samples pre- (screening) and post- (cycle 2 onwards) MEZI treatment, and 32 additional patients with isolated screening samples. Of the 45 patients, 39 received a dose of 0.8-1.0mg daily for 14 or 21 days of a 28-day cycle, 6 received 0.2-0.6mg on varying schedules as part of the trial design. CD138-negative BM mononuclear cells were isolated from these patients and profiled at single-cell resolution using a previously described 39-marker TME CyTOF-panel (van Oekelen, ASH 2021). Both manual hierarchical gating and unsupervised computational workflows were used for CyTOF data analysis. Around 5 million total cells were analyzed and over 100 immune populations. Samples with <2500 total events were excluded for quality-control. Population frequencies used for comparison are reported as median percentage of the parent population. Statistical comparisons between groups were assessed using the Wilcoxon test. Results Upon MEZI treatment, both adaptive and innate BM immune compartments demonstrated a significant increase in activated populations and marked reduction in exhausted/senescent populations (Figure 1A). Total NK and NKT populations expanded on MEZI (p<0.01 for both). Both CD4 and CD8 T cell populations shifted towards an effector phenotype with significant increases in effector memory (CCR7-CD45RA-) populations (p<0.001 for both) and concurrent reductions of and naïve (p<0.001), central memory (p<0.01 for CD4, NS for CD8) and TEMRA populations (p<0.001). HLADR-expressing CD4, CD8 and NKT populations significantly increased on MEZI (p<0.001 for all) as did ICOS-expressing CD4 (p<0.01), NK (p<0.001) and NKT cells (p<0.05). Senescent NK, NKT, CD4 and CD8 populations expressing KLRG1 decreased on MEZI treatment (p<0.01 for all) and CD57 positive CD8 and NKT cell populations also reduced (p<0.05). CD8, NK and NKT cells expressing the inhibitory checkpoint molecule TIGIT, showed a near 3-fold reduction of on treatment (p<0.001 for all). Although PD1-positive populations did not change on MEZI, PD1/TIGIT double-positive CD4 and CD8 populations showed significant reduction (p<0.05 for both). A dose-dependent reduction of TIGIT-expressing CD8 and NKT populations was seen at higher doses of MEZI (p<0.05 for both), suggesting a direct effect of MEZI on the BM TME. We further correlated depth-of-response with baseline immune composition in 77 patients (Figure 1B). Patients with progression of disease (n=9) on MEZI, when compared to all other patients (n=68), were found to have higher frequencies of PD1 (p<0.05) and TIGIT (p<0.001) positive, as well as dual PD1 and TIGIT positive (p<0.001), CD8 and NKT populations. Number of prior lines (median = 6) and type of MM therapy prior to MEZI did not significantly affect baseline PD1 or TIGIT positive populations. Conclusion MEZI treatment leads to significant activation of adaptive and innate BM immune populations with concurrent reduction of exhausted and senescent populations. These findings are concordant with prior observations of IMiD-mediated immune-stimulation. Patients progressing through MEZI had higher levels of immune checkpoint expression at baseline, which may be helpful for patient selection for this novel CELMoD. These results also suggest that MEZI may be useful in combination with other agents for increasing immune activation and reducing immune exhaustion. Further analyses comparing the effect of different CELMoDs on the BM TME are ongoing.

Multi-Omics Analysis Reveals That TET2 Loss Epigenetically Primes Monocytes for Inflammatory Responses Via AP-1 Signaling

Loss of function mutations in TET2 are common in clonal hematopoiesis (CH) and myelodysplastic syndromes (MDS). Increased inflammation and stem cell capacity have been reported in Tet2 KO mouse models, but molecular mechanisms and cell type-specific contributions to explain this phenotype are lacking. To create models of CH-like TET2 deficiency, we transplanted Mx1-CRE Tet2 fl/fl (Tet2 KO) or Tet2 fl/fl (both CD45.1 +) in CD45.2 + recipientsfollowing radiation-free conditioning with ACK2 (anti-cKit) antibody injection. Six months post-transplant, we flow sorted CD45.1 +Tet2 KO and Tet2 fl/fl Lin -Sca-1 +cKit + (LSK) cells, granulocyte-monocyte progenitors (GMP), and monocytes for RNA and assay for transposase-accessible chromatin (ATAC) sequencing. Differential gene expression analysis identified 902 (468 up, 434 down) differentially expressed genes (DEGs) in LSK, 134 (95 up, 39 down) in GMP, and 474 (125 up, 349 down) in monocytes (p adj<0.05). Gene set enrichment analysis (GSEA) identified that the top Hallmark gene sets in Tet2 KOLSK cells were related to MYC targets, oxidative phosphorylation, and E2F targets. In contrast, GMP and monocytes were most enriched for IFN-alpha and gamma gene sets along with IL6/JAK/STAT3 signaling. ATACseq revealed increased accessibility of myeloid-related transcription factor (TF) binding motifs in Tet2 KOLSK cells (CEBPA, PU.1-IRF(ETS:IRF), and AP-1 TF FOSL2(bZIP)), while prediction of cis-regulating regions via GREAT identified downregulated peaks associated with stem cell maintenance and upregulated peaks associated with myeloid mediated immunity. Peaks upregulated in Tet2 KOGMP cells were enriched for the motifs recognized by CEBPE, ISRE(IRF)) and the AP-1 TF JUND, while upregulated peaks in monocytes were enriched for RUNX1 and CEBPA motifs, as well as AP-1 TF FOS::JUNB. GREAT analysis in GMPs identified cis-regulatory regions related to myeloid mediated immunity, while monocytes possessed those associated with biotic stimuli. To better dissect heterogeneity across primitive and mature cell types driven by TET2 loss, we performed scRNAseq on Tet2 KO and Tet2 WT wholebone marrow (WBM) and sorted LSK cells. We identified 18 transcriptionally distinct cell subsets in a combined analysis of WBM. Tet2 KOprogenitor and myeloid subsets were all enriched for gene sets related to IFN-alpha and gamma and immune signaling. Interestingly, primitive and cycling HSPCs were also enriched for MYC target and oxidative phosphorylation gene sets, while these were negatively enriched for GMP, monocyte progenitor, and monocyte Tet2 KO subsets. Further, the IL-6/JAK/STAT3 signaling gene set was enriched in later myeloid cell types, but not in early or cycling HSPCs. These observations led us to hypothesize that Tet2 KO BM cell populations have differentially enhanced phosphoprotein signaling responses to key cytokines. We used a 36-antibody mass cytometry panel to perform deep immunophenotyping and signaling response studies in Tet2 KO BM. FlowSOM analysis identified clusters corresponding to cKIT +CD34 + myeloid progenitors, a Ly6C +CD34 +cKIT + progenitor, and Ly6Chi monocyte clusters were enriched in Tet2 KO BM, while manual gating revealed expansion of LSK, MPP3, and monocytes. We found that activation of pSTAT3 and pSTAT1, in response to either IL-6 or IFNγ, respectively, were significantly higher in Tet2 KO Ly6Chi monocytes, while the more primitive Ly6C +CD34 +cKIT + progenitors demonstrated only an increased response to IFNγ at pSTAT1. Further, prolonged exposure to IL-6 in immortalized Tet2 KO Hoxb8-ER GMP-like cells led to significantly higher levels of pSTAT3 over 4 hours. Since AP-1 TFs are drivers of inflammation and were enriched in our ATACseq dataset, we tested the effects of an AP-1 inhibitor on the Tet2 KO Hoxb8-ER cells. We found that the AP-1 inhibitor T5224 dose dependently reduced the expansion of Tet2 KO Hoxb8 progenitor cells in base media and inhibited the IL-6 mediated expansion of Tet2 KO Hoxb8 cells. Our data show that TET2 loss rewires BM HSPCs towards an inflammatory, myeloid phenotype. Starting at the GMP stage, we see increases in phosphoprotein responses to IFNγ, and then the addition of hypersensitivity to IL-6 in monocytes. TET2 loss leads to an increase in accessible AP-1 motifs in Tet2 KO cells, while inhibition of AP-1 reduces fitness of cells, suggesting a possible therapeutic strategy for TET2-mutated myeloid malignancy.

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

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. 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. 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. 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. 40min, 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. 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.

CRB1-Associated Retinal Dystrophy Patients Have Expanded Lewis Glycoantigen-Positive T Cells.

Eye inflammation may occur in patients with inherited retinal dystrophies (IRDs) and is seen frequently in IRDs associated with mutations in the CRB1 gene. The purpose of this study was to determine the types of inflammatory cells involved in IRDs, by deep profiling the composition of peripheral blood mononuclear cells of patients with a CRB1-associated IRD. This study included 33 patients with an IRD with confirmed CRB1 mutations and 32 healthy controls. A 43-parameter flow cytometry analysis was performed on peripheral blood mononuclear cells isolated from venous blood. FlowSOM and manual Boolean combination gating were used to identify and quantify immune cell subsets. Comparing patients with controls revealed a significant increase in patients in the abundance of circulating CD4+ T cells and CD8+ T cells that express sialyl Lewis X antigen. Furthermore, we detected a decrease in plasmacytoid dendritic cells and an IgA+CD24+CD38+ transitional B-cell subset in patients with an IRD. Patients with a CRB1-associated IRD show marked changes in blood leukocyte composition, affecting lymphocyte and dendritic cell populations. These results implicate inflammatory pathways in the disease manifestations of IRDs.

Using kV Triggered Imaging and Liver Dome Position to Reduce the Dosimetric Error Caused by Breath Hold Variability for Liver Stereotactic Body Radiation Therapy

In a previous study, we demonstrated that manual gating using kV triggered imaging and liver dome position can reduce targeting errors caused by breath hold variability for liver stereotactic body radiation therapy (SBRT). In this study, we quantified the dosimetric error caused by breath hold variability and investigated the effect of liver dome gating on reducing dosimetric error. Twenty-five liver SBRT patients treated with deep inspiration breath-hold were included in this IRB approved study. Volumetric modulated arc therapy was used to deliver 30-60 Gy in 1-5 fractions. To verify the breath-hold reproducibility during treatment, a KV triggered image was acquired at the beginning of each breath-hold. The liver dome position was visually compared with the expected upper/lower liver boundaries created by expanding/contracting the liver contour 5mm in the superior-inferior direction. If the liver dome position was within the boundaries, delivery continued; otherwise, beam was held manually and the patient was instructed to take another breath hold until the liver dome position was within boundaries. To calculate delivered dose, for each fraction, the treatment plan was divided into sub-beams, each corresponding to one breath hold using delivery log files. The triggered images were registered to the planning CT to determine the liver position during each breath hold. Dose delivered during each breath hold was calculated by shifting the isocenter of the sub-beam according to the liver position. Breath holds discarded by gating were excluded since no dose was delivered during these breath holds. Delivered fractional doses were compared with planned fractional doses using GTV D99 and liver Dmean. To estimate delivered dose without gating, the first "corrective" breath hold taken after the discarded breath holds was replaced with the prior discarded breath hold and dose calculation was repeated. Seven hundred eleven triggered images from 91 treatment fractions were analyzed. Without gating, in 11 of the 91 fractions from 7 of the 25 patients, delivered GTV D99 reduced > 0.50 Gy from planned value (range 0.51-1.68 Gy, 3-10% of planned fractional GTV D99). Liver dome gating was able to detect/exclude irreproducible breath holds in 8 of the 11 fractions, increasing the delivered GTV D99 by 0.70 Gy per fraction on average (range 0.21-1.63 Gy). With liver dome gating, delivered fractional GTV D99 was comparable to planned value for all fractions (12.96 +/- 5.19 Gy vs 13.04 +/- 5.18 Gy, p > 0.05). Liver mean dose was not affected by breath hold variability or gating. Fractional liver Dmean was 2.26 +/- 1.19 Gy from plan, 2.27 +/- 1.21 Gy for delivery with gating and 2.27 +/- 1.20 Gy for delivery without gating. Breath hold variability may cause tumor underdose. Liver dome gating using kV triggered imaging reduces dosimetric error and ensures tumor coverage for liver SBRT.

B-144 A Comparative Analysis of Unsupervised Machine Learning Algorithms for Polyploidy Analysis Using Flow Cytometry

Abstract Background Recently, there has been a growing interest in using cell differentiation inducers to produce platelets in vitro by inducing megakaryocytic differentiation in cell lines. During this process, cells undergo polyploidization through endomitosis. To measure the degree of polyploidy, cell cycle analysis by flow cytometry is commonly used. However, manual gating can be time-consuming and prone to human error, especially when analyzing large datasets. To address this issue, this study aimed to compare the performance of unsupervised machine learning algorithms for polyploidy analysis of flow cytometry data. Methods The K562 cell line (Korea Cell Line Bank, Korea) was cultured under four conditions, including 1 nM and 2.5 nM of PMA (phorbol 12-myristate 13-acetate; Sigma Aldrich, USA) and 5 μg/mL of phytosphingosine (TCI, Japan), for four days. Propidium iodide (Sigma Aldrich, USA) was used to stain the cells for ploidy analysis, and flow cytometry (Navios, Beckman Coulter Inc, USA) was used to obtain the intensity values. Twelve flow cytometry measurements were obtained for K562 cells cultured under four different conditions, with each condition tested in triplicate. Unsupervised machine learning algorithms, including K-Means, Bisecting K-Means (BKM), Mini Batch K-Means (MBKM), Agglomerative Clustering (AC), Gaussian Mixture Models (GMM), K-Medoids, and Partitioning Around Medoids (PAM), were used to generate boundaries among three clusters (2n, 4n, and ≥8n). The boundary values generated by each algorithm were comparatively analyzed by calculating the standard derivation index (SDI) on each boundary. The concordance of each algorithm with manual gating of flow cytometry data was evaluated by calculating residual errors. Results A total of 168 SDI values were obtained from seven different unsupervised machine learning algorithms for two boundaries among 2n, 4n, and ≥8n clusters of four culture conditions in triplicate. When comparing the average (±95% confidence interval) of SDI for each algorithm for the entire dataset, K-Means showed the most acceptable average SDI value of −0.07 (±0.17) with the narrowest confidence interval followed by −0.13 (±0.54) of AC, −0.26 (±0.18) of MBKM, −0.53 (±0.2) of K-Medoids, −0.58 (±0.2) of PAM, 0.74 (±0.3) of GMM, and 0.84 (±0.35) of BKM. In addition, the average (±95% confidence interval) of the absolute value of the residual errors of each algorithm was calculated, with K-Means showing the lowest value of 6.67 (±2.3), indicating higher concordance with manual gating. AC, MBKM, K-Medoids, PAM, BKM, and GMM had higher average residual errors of 10.04 (±2.5), 10.46 (±4.07), 11.13 (±4.65), 12.04 (±4.83), 12.54 (±3.0), and 12.75 (±3.5), respectively. Conclusion K-means demonstrated the best performance for clustering each ploidy among the seven unsupervised machine learning algorithms tested for the flow cytometry data. In addition, the algorithms achieved acceptable concordance with manual gating in most cases in this study. Therefore, unsupervised machine learning algorithms show promise in automating the gating of flow cytometry data for polyploidy analysis and measuring megakaryocytic differentiation in cell lines.

Abstract 21 Machine Learning Approach for Automatic Enumeration of Cord Blood Stem Cells by Flow Cytometry

Abstract Introduction Stem cell laboratories measure the number of viable CD34+ and CD45+ cells in their products by flow cytometry following a standard protocol (ISHAGE). Although the ISHAGE protocol is well documented, its correct application can be challenging. Automated gating algorithms have been created to address this issue, but they have limitations and can only be used in certain conditions. Recent developments in artificial intelligence (AI) and machine learning (ML) have made these approaches more accessible and practical to use. For flow cytometry data, the Cytobank platform now offers the Automatic Gating Algorithm (AGA), an AI-ML-based feature allowing users to easily implement their own gating strategies. Objectives The performance of the AGA based on the ISHAGE protocol gating was evaluated using cord blood post-thaw flow cytometry data. Methods In this study, the AGA was trained using 29 flow cytometry cord blood data files analyzed according to the ISHAGE gating strategy. Using the model, an inference analysis was conducted on two series of 12 data samples acquired from two independent operators. The enumeration of viable CD34+ cells was manually performed and compared to the AI analyses. Results The mean differences and standard deviation in vCD34+ counts between the manual and the automatic gating are 6.0% ± 2.7 for Operator #1 and 7.7% ± 2.1 for Operator #2. The correlation coefficients between vCD34+ results are 0.9981 for Operator #1 vs AI, and 0.9699 for Operator #2 vs AI. These results show that the AI inference provides results comparable to human manual gating, and are within the accepted difference of 10% between duplicates as defined by ISHAGE convention. Discussion This study provides a testing ground for using AI-ML-based automatic gating and supports its use in stem cell laboratories with acceptable performance and accuracy. The application of AI and ML for vCD34+ enumeration can potentially lead to the development of more widely applicable automatic gating algorithms. This could improve standardization by providing a common tool for analyzing flow cytometry data.

Classifying flow cytometry data using Bayesian analysis helps to distinguish ALS patients from healthy controls.

Given its wide availability and cost-effectiveness, multidimensional flow cytometry (mFC) became a core method in the field of immunology allowing for the analysis of a broad range of individual cells providing insights into cell subset composition, cellular behavior, and cell-to-cell interactions. Formerly, the analysis of mFC data solely relied on manual gating strategies. With the advent of novel computational approaches, (semi-)automated gating strategies and analysis tools complemented manual approaches. Using Bayesian network analysis, we developed a mathematical model for the dependencies of different obtained mFC markers. The algorithm creates a Bayesian network that is a HC tree when including raw, ungated mFC data of a randomly selected healthy control cohort (HC). The HC tree is used to classify whether the observed marker distribution (either patients with amyotrophic lateral sclerosis (ALS) or HC) is predicted. The relative number of cells where the probability q is equal to zero is calculated reflecting the similarity in the marker distribution between a randomly chosen mFC file (ALS or HC) and the HC tree. Including peripheral blood mFC data from 68 ALS and 35 HC, the algorithm could correctly identify 64/68 ALS cases. Tuning of parameters revealed that the combination of 7 markers, 200 bins, and 20 patients achieved the highest AUC on a significance level of p < 0.0001. The markers CD4 and CD38 showed the highest zero probability. We successfully validated our approach by including a second, independent ALS and HC cohort (55 ALS and 30 HC). In this case, all ALS were correctly identified and side scatter and CD20 yielded the highest zero probability. Finally, both datasets were analyzed by the commercially available algorithm 'Citrus', which indicated superior ability of Bayesian network analysis when including raw, ungated mFC data. Bayesian network analysis might present a novel approach for classifying mFC data, which does not rely on reduction techniques, thus, allowing to retain information on the entire dataset. Future studies will have to assess the performance when discriminating clinically relevant differential diagnoses to evaluate the complementary diagnostic benefit of Bayesian network analysis to the clinical routine workup.

A biology-driven deep generative model for cell-type annotation in cytometry.

Cytometry enables precise single-cell phenotyping within heterogeneous populations. These cell types are traditionally annotated via manual gating, but this method lacks reproducibility and sensitivity to batch effect. Also, the most recent cytometers-spectral flow or mass cytometers-create rich and high-dimensional data whose analysis via manual gating becomes challenging and time-consuming. To tackle these limitations, we introduce Scyan https://github.com/MICS-Lab/scyan, a Single-cell Cytometry Annotation Network that automatically annotates cell types using only prior expert knowledge about the cytometry panel. For this, it uses a normalizing flow-a type of deep generative model-that maps protein expressions into a biologically relevant latent space. We demonstrate that Scyan significantly outperforms the related state-of-the-art models on multiple public datasets while being faster and interpretable. In addition, Scyan overcomes several complementary tasks, such as batch-effect correction, debarcoding and population discovery. Overall, this model accelerates and eases cell population characterization, quantification and discovery in cytometry.

Adipose derived stem cell extracellular vesicles modulate primary human macrophages to an anti‐inflammatory phenotype in vitro

AbstractEVs released by adipose derived stem cells (ADSCs) have shown promise as a therapeutic for tissue repair because of their purported immune‐regulatory properties. Extracellular vesicles (EVs) from ADSCs could be beneficial in improving graft retention rates for autologous fat grafting (AFG) post‐mastectomy as, currently, grafted tissue rates are variable. Enriching grafted tissue with ADSC‐EVs may improve retention rates by modulating macrophages resident within both the breast and lipoaspirate. We aimed to identify key macrophage phenotypes that are modulated by ADSC‐EVs in vitro. ADSCs were isolated from lipoaspirates of women undergoing AFG and characterised by flow cytometry and differentiation potential. ADSC‐EVs were isolated from culture media and characterised by tuneable resistive pulse sensing, transmission electron microscopy and Western blot. Primary monocyte‐derived macrophages were polarized to an M1‐like (GM‐CSF, IFNγ), M2‐like phenotype (M‐CSF, IL‐4) or maintained (M0‐like; M‐CSF) and ADSC‐EVs were co‐cultured with macrophages for 48 h. Flow cytometry and high‐dimensional analysis clustered macrophages post co‐culture. A manual gating strategy was generated to recapitulate these clusters and was applied to a repeat experimental run. Both runs were analysed to examine the prevalence of each cluster, representing a unique macrophage phenotype, with and without ADSC‐EVs. Following the addition of ADSC‐EVs, M0‐like macrophages demonstrated a reciprocal shift of cell distribution from a cluster with a ‘high inflammatory profile’ (CD36+++CD206+++CD86+++; 16.5 ± 7.0%; p &lt; 0.0001) to a cluster with a ‘lower inflammatory profile’ (CD36+CD206+CD86+; 35 ± 21.5%; p &lt; 0.05). M1‐like macrophages shifted from a cluster with a ‘high inflammatory profile’ (CD206++CD11b++CD36++CD163++; 26.1 ± 9.4%; p = 0.0024) to a ‘lower inflammatory profile’ (CD206+CD11b+CD36+CD163+; 72.8 ± 8.7%; p = 0.0007). There was no shift in M2‐like clusters following ADSC‐EV treatment. ADSC‐EVs are complex regulators of macrophage phenotype that can shift macrophages away from a heightened pro‐inflammatory state.

Merging and imputation of flow cytometry data: a critical assessment.

Although most modern techniques and analysis methods in multiparameter flow cytometry (MFC) allow for increased dimensionality for the characterization and quantification of cell populations, most MFC applications depend on flow cytometers measuring relatively small (<16) numbers of parameters. When more markers than the available parameters need to be acquired, these are commonly distributed over multiple independent measurements that include a backbone of common markers. Several methods have been proposed to impute values for combinations of markers that were not measured simultaneously. These imputation methods are frequently used without proper validation and knowledge of their effects on data analysis. We evaluated the performance of existing imputation software (Infinicyt, CyTOFmerge, CytoBackBone, and cyCombine) in approximating known measured expression data in terms of similarity in visual appearance, cell expression, and gating in different datasets by splitting MFC samples into separate measurements with partially overlapping markers and re-calculating missing marker expression. Out of the assessed packages, CyTOFmerge showed the most accurate approximation of the known expression in terms of similar expression values and concordance with manual gating, with a mean F-score between 0.53-0.87 when retrieving cell populations in different datasets. Performance remained inadequate for all methods, with only limited similarity at the cell level. In conclusion, the use of imputed MFC data should take such limitations into account and include independent validation of results to justify conclusions. This article is protected by copyright. All rights reserved.

Increased levels of circulating granulocytic myeloid‑derived suppressor cells in lumbar disc herniation.

Myeloid-derived suppressor cells (MDSCs) expand when the body undergoes inflammatory diseases and chronic diseases. However, its role in intervertebral disc degeneration remains unclear. The present study aimed to characterize specific subsets of MDSCs as potential indicators of disease progression in patients with lumbar disc herniation (LDH). The Gene Expression Omnibus (GEO) database was used to analyze the changes in granulocyte MDSCs (G-MDSCs). Peripheral blood samples were collected from 40 patients with LDH and 15 healthy controls, and flow cytometry was used to characterize different subsets of MDSCs. All subjects underwent lumbar spine magnetic resonance imaging. Then, t-distributed stochastic neighborhood embedding and FlowSOM were used to analyze the data obtained by CytoFlex. The correlation between circulating MDSCs and the clinicopathological stage of LDH was then further analyzed. The GEO database predicted that G-MDSCs were highly expressed in patients with LDH. The frequency of circulating G-MDSCs increased with Pfirrmann stage III and IV, while the percentage of mononuclear MDSCs (M-MDSCs) only increased. Patient age and sex did not correlate with the frequency of circulating G-MDSCs and M-MDSCs. The results of the computer algorithm analysis were consistent with those of our manual gating. The present study showed that the occurrence of LDH led to changes in the MDSC subpopulation in the circulating peripheral blood of patients, and the frequency of circulating G-MDSCs in patients with clinical stage III and IV LDH increased with the degree of degeneration. The determination of G-MDSCs can be used as an auxiliary examination item for LDH.

#5664 PERITONEAL DIFFUSE PODOPLANIN EXPRESSION IS ASSOCIATED WITH SEVERE ARTERIOLOPATHY IN CHILDREN ON PERITONEAL DIALYSIS

Abstract Background and Aims Podoplanin is a transmembrane glycoprotein binding chemokines and modulating inflammatory reactions. In healthy peritoneum, it is expressed in mesothelial cells and in lymphatic endothelium, expression beyond these structures (diffuse podoplanin staining, DPS) has been observed in adult patients with encapsulating peritoneal sclerosis (EPS). Method Parietal peritoneal tissues from 71 children on PD fluids with low glucose degradation product (GDP) content underwent digital histomorphometry. DPS and pathomechanisms of peritoneal membrane transformation and vasculopathy were hierarchically clustered. A panel of 13 peritoneal cell markers was selected and tissue sections were stained with metal conjugated antibodies and acquired using imaging mass cytometry (IMC; Hyperion Imaging System®). After single-cell segmentation, spillover correction, and normalization, cell types were identified using manual gating to train a random forest classifier for final cell type identifications using published workflows. Single-cell analysis was used to characterize peritoneal cells with positive podoplanin signals. Results DPS was present in 22% of the PD patients, all devoid of any clinical, radiological or histological feature of EPS. DPS positive patients were younger, had higher dialytic glucose exposure (median 103 vs. 154 g/m²/day, p = 0.002), and lower peritoneal arteriolar lumen to vessel (L/V) ratio, i.e. exhibited more arteriolar lumen narrowing. PD duration [median 20 (14.5, 48) months] and peritonitis incidence were similar as in non-DPS patients. In multivariable analysis, glucose exposure, PD duration, epithelial-to-mesenchymal transformed (EMT) cells and lower patient body surface area (BSA) independently predicted the presence of DPS. DPS and lower mesothelial surface coverage predicted lower arteriolar L/V ratio. In subgroups matched for BSA, PD duration and dialytic glucose exposure, DPS patients had a lower L/V ratio, (0.44 ± 0.13 vs. 0.31 ± 0.18, p = 0.003) and higher submesothelial CD45 leucocyte, CD68 macrophage and EMT cell count. Submesothelial lipopolysaccharide and hyaluronic acid receptor CD44 were increased in DPS positive patients, hyaluronan synthase 1 and 2 were similar. Among DPS positive patients, DPS intensity grade was higher in those with history of bacterial or fungal peritonitis (7 out of 15). Two patients had three and four peritonitis episodes respectively, their DPS pattern did not differ from patients with history of only one peritonitis episode. Hierarchical clustering demonstrated highest similarity of DPS positive areas with CD68 positive areas. In IMC analysis of DPS positive areas, highest podoplanin signals were derived from M2 classified macrophages (CD68+CD163+) and fibroblastic cells [αSMA positive/non-endothelial/non-macrophage (αSMA+PROX1-CD31-CD68-CD163-)], signals derived from M1 classified macrophages (CD68+CD163-) were lowest. Conclusion Peritoneal DPS in pediatric patients on low-GDP PD is associated with lower BSA, higher dialytic glucose exposure, history of infectious peritonitis and peritoneal inflammatory and EMT cell invasion, and predicts obliterating peritoneal vasculopathy. We identified M2 macrophages as major source of extra-lymphatic peritoneal podoplanin accumulation, suggesting their role in severe peritoneal transformation.

Perencanaan Control Valve Pada Head Tank PLTA Tulungagung Menggunakan PLC

Head Tank of the Tulungagung Hydro Electric Power Plan of PT PJBUP Brantas still has a manual control valve which consists of a gate valve and a pair of float valves. There are some problems when the pipe of the float valve is blocked by cediment and debris, it could affect the cooling water system and make damage to the power plant. In this research, there would build a plan of the automatic control valve of the Tulungagung HEPP. PLC Siemens Simatic Step 7 was used as a control base on a SCADA software Wonderware Intouch version 10.0.0. In this modification plan, a pair of float valve changed over solenoid valve and manual gate valve change over motorized valve. For the water level that would be ultrasonic sensor SICK UM 30 as a input for the power meter Omron Procces Meter K3MA-J. The reliability of this modification are excellence and complete for the auto/manual control even for the long distance one. Moreover this modification also has excellence safety.