Unbiased method for spectral analysis of cells with great diversity of autofluorescence spectra.

Autofluorescence is an intrinsic feature of cells, caused by the natural emission of light by photo-excitatory molecular content, which can complicate analysis of flow cytometry data. Different cell types have different autofluorescence spectra and, even within one cell type, heterogeneity of autofluorescence spectra can be present, for example, as a consequence of activation status or metabolic changes. By using full spectrum flow cytometry, the emission spectrum of a fluorochrome is captured by a set of photo detectors across a range of wavelengths, creating an unique signature for that fluorochrome. This signature is then used to identify, or unmix, that fluorochrome's unique spectrum from a multicolor sample containing different fluorescent molecules. Importantly, this means that this technology can also be used to identify intrinsic autofluorescence signal of an unstained sample, which can be used for unmixing purposes and to separate the autofluorescence signal from the fluorophore signals. However, this only works if the sample has a singular, relatively homogeneous and bright autofluorescence spectrum. To analyze samples with heterogeneous autofluorescence spectral profiles, we setup an unbiased workflow to more quickly identify differing autofluorescence spectra present in a sample to include as "autofluorescence signatures" during the unmixing of the full stained samples. First, clusters of cells with similar autofluorescence spectra are identified by unbiased dimensional reduction and clustering of unstained cells. Then, unique autofluorescence clusters are determined and are used to improve the unmixing accuracy of the full stained sample. Independent of the intensity of the autofluorescence and immunophenotyping of cell subsets, this unbiased method allows for the identification of most of the distinct autofluorescence spectra present in a sample, leading to less confounding autofluorescence spillover and spread into extrinsic phenotyping markers. Furthermore, this method is equally useful for spectral analysis of different biological samples, including tissue cell suspensions, peripheral blood mononuclear cells, and in vitro cultures of (primary) cells.

Limited impact of cancer-derived gangliosides on anti-tumor immunity in colorectal cancer.

Aberrant glycosylation is a key mechanism employed by cancer cells to evade immune surveillance, induce angiogenesis and metastasis, among other hallmarks of cancer. Sialic acids, distinctive terminal glycan structures located on glycoproteins or glycolipids, are prominently upregulated across various tumor types, including colorectal cancer (CRC). Sialylated glycans modulate anti-tumor immune responses through their interactions with Siglecs, a family of glycan-binding receptors with specificity for sialic acid-containing glycoconjugates, often resulting in immunosuppression. In this paper, we investigated the immunomodulatory function of ST3Gal5, a sialyltransferase that catalyzes the addition of α2-3 sialic acids to glycosphingolipids, since lower expression of ST3Gal5 is associated with better survival of CRC patients. We employed CRISPR/Cas9 to knock out the ST3Gal5 gene in two murine CRC cell lines MC38 and CT26. Glycomics analysis confirmed the removal of sialic acids on glycolipids, with no discernible impact on glycoprotein sialylation. Although knocking out ST3Gal5 in both cell lines did not affect in vivo tumor growth, we observed enhanced levels of regulatory T cells in CT26 tumors lacking ST3Gal5. Moreover, we demonstrate that the absence of ST3Gal5 affected size and blood vessel density only in MC38 tumors. In summary, we ascertain that sialylation of glycosphingolipids has a limited influence on the anti-tumor immune response in CRC, despite detecting alterations in the tumor microenvironment, possibly due to a shift in ganglioside abundance.

Adaptive immune changes associate with clinical progression of Alzheimer's disease.

BACKGROUND: Alzheimer's disease (AD) is the most frequent cause of dementia. Recent evidence suggests the involvement of peripheral immune cells in the disease, but the underlying mechanisms remain unclear. METHODS: We comprehensively mapped peripheral immune changes in AD patients with mild cognitive impairment (MCI) or dementia compared to controls, using cytometry by time-of-flight (CyTOF). RESULTS: We found an adaptive immune signature in AD, and specifically highlight the accumulation of PD1+ CD57+ CD8+ T effector memory cells re-expressing CD45RA in the MCI stage of AD. In addition, several innate and adaptive immune cell subsets correlated to cerebrospinal fluid (CSF) biomarkers of AD neuropathology and measures for cognitive decline. Intriguingly, subsets of memory T and B cells were negatively associated with CSF biomarkers for tau pathology, neurodegeneration and neuroinflammation in AD patients. Lastly, we established the influence of the APOE ε4 allele on peripheral immunity. CONCLUSIONS: Our findings illustrate significant peripheral immune alterations associated with both early and late clinical stages of AD, emphasizing the necessity for further investigation into how these changes influence underlying brain pathology.

In-depth immune profiling of peripheral blood mononuclear cells in patients with pancreatic ductal adenocarcinoma reveals discriminative immune subpopulations.

Pancreatic ductal adenocarcinoma (PDAC) has a dismal prognosis with a 5-year survival of less than 10%. More knowledge of the immune response developed in patients with PDAC is pivotal to develop better combination immune therapies to improve clinical outcome. In this study, we used mass cytometry time-of-flight to undertake an in-depth characterization of PBMCs from patients with PDAC and examine the differences with healthy controls and patients with benign diseases of the biliary system or pancreas. Peripheral blood mononuclear cells from patients with PDAC or benign disease are characterized by the increase of pro-inflammatory cells, as CD86+ classical monocytes and memory T cells expressing CCR6+ and CXCR3+, associated with T helper 1 (Th1) and Th17 immune responses, respectively. However, PBMCs from patients with PDAC present also an increase of CD39+ regulatory T cells and CCR4+CCR6-CXCR3- memory T cells, suggesting Th2 and regulatory responses. Concluding, our results show PDAC develops a multifaceted immunity, where a proinflammatory component is accompanied by regulatory responses, which could inhibit potential antitumor mechanisms.

A monodispersed metal-complexing peptide-based polymer for mass cytometry enabling spectral applications.

We report the synthesis of a novel class of metal-complexing peptide-based polymers, which we name HyperMAPs (Hyper-loaded MetAl-complexed Polymers). The controlled solid-phase synthesis of HyperMAPs' scaffold peptide provides our polymer with a well-defined molecular structure that allows for an accurate on-design assembly of a wide variety of metals. The peptide-scaffold features a handle for direct conjugation to antibodies or any other biomolecules by means of a thiol-maleimide-click or aldehyde-oxime reaction, a fluorogenic moiety for biomolecule conjugation tracking, and a well-defined number of functional groups for direct incorporation of metal-chelator complexes. Since metal-chelator complexes are prepared in a separate reaction prior to incorporation to the peptide scaffold, polymers can be designed to contain specific ratios of metal isotopes, providing each polymer with a unique CyTOF spectral fingerprint. We demonstrate the complexing of 21 different metals using two different chelators and provide evidence of the application of HyperMAPs on a 13 parameter CyTOF panel and compare its performance to monoisotopic metal-conjugated antibodies.

Lymphopenia is associated with broad host response aberrations in community-acquired pneumonia.

OBJECTIVES: Lymphopenia at hospital admission occurs in over one-third of patients with community-acquired pneumonia (CAP), yet its clinical relevance and pathophysiological implications remain underexplored. We evaluated outcomes and immune features of patients with lymphopenic CAP (L-CAP), a previously described immunophenotype characterized by admission lymphocyte count <0.724 × 109 cells/L. METHODS: Observational study in 149 patients admitted to a general ward for CAP. We measured 34 plasma biomarkers reflective of inflammation, endothelial cell responses, coagulation, and immune checkpoints. We characterized lymphocyte phenotypes in 29 patients using spectral flow cytometry. RESULTS: L-CAP occurred in 45 patients (30.2%) and was associated with prolonged time-to-clinical-stability (median 5 versus 3 days), also when we accounted for competing events for reaching clinical stability and adjusted for baseline covariates (subdistribution hazard ratio 0.63; 95% confidence interval 0.45-0.88). L-CAP patients demonstrated a proportional depletion of CD4 T follicular helper cells, CD4 T effector memory cells, naïve CD8 T cells and IgG+ B cells. Plasma biomarker analyses indicated increased activation of the cytokine network and the vascular endothelium in L-CAP. CONCLUSIONS: L-CAP patients have a protracted clinical recovery course and a more broadly dysregulated host response. These findings highlight the prognostic and pathophysiological relevance of admission lymphopenia in patients with CAP.

The immunological landscape of peripheral blood in glioblastoma patients and immunological consequences of age and dexamethasone treatment.

Background: Glioblastomas manipulate the immune system both locally and systemically, yet, glioblastoma-associated changes in peripheral blood immune composition are poorly studied. Age and dexamethasone administration in glioblastoma patients have been hypothesized to limit the effectiveness of immunotherapy, but their effects remain unclear. We compared peripheral blood immune composition in patients with different types of brain tumor to determine the influence of age, dexamethasone treatment, and tumor volume. Methods: High-dimensional mass cytometry was used to characterise peripheral blood mononuclear cells of 169 patients with glioblastoma, lower grade astrocytoma, metastases and meningioma. We used blood from medically-refractory epilepsy patients and healthy controls as control groups. Immune phenotyping was performed using FlowSOM and t-SNE analysis in R followed by supervised annotation of the resulting clusters. We conducted multiple linear regression analysis between intracranial pathology and cell type abundance, corrected for clinical variables. We tested correlations between cell type abundance and survival with Cox-regression analyses. Results: Glioblastoma patients had significantly fewer naive CD4+ T cells, but higher percentages of mature NK cells than controls. Decreases of naive CD8+ T cells and alternative monocytes and an increase of memory B cells in glioblastoma patients were influenced by age and dexamethasone treatment, and only memory B cells by tumor volume. Progression free survival was associated with percentages of CD4+ regulatory T cells and double negative T cells. Conclusion: High-dimensional mass cytometry of peripheral blood in patients with different types of intracranial tumor provides insight into the relation between intracranial pathology and peripheral immune status. Wide immunosuppression associated with age and pre-operative dexamethasone treatment provide further evidence for their deleterious effects on treatment with immunotherapy.

The transcriptional landscape of glycosylation-related genes in cancer.

Changes in glycosylation patterns have been associated with malignant transformation and clinical outcomes in several cancer types, prompting ongoing research into the mechanisms involved and potential clinical applications. In this study, we performed an extensive transcriptomic analysis of glycosylation-related genes and pathways, using publicly available bulk and single cell transcriptomic datasets from tumor samples and cancer cell lines. We identified genes and pathways strongly associated with different tumor types, which may represent novel diagnostic biomarkers. By using single cell RNA-seq data, we characterized the contribution of different cell types to the overall tumor glycosylation. Transcriptomic analysis of cancer cell lines revealed that they present a simplified landscape of genes compared to tissue. Lastly, we describe the association of different genes and pathways with the clinical outcome of patients. These results can serve as a resource for future research aimed to unravel the role of the glyco-code in cancer.

High-dimensional phenotyping of the peripheral immune response in community-acquired pneumonia.

Background: Community-acquired pneumonia (CAP) represents a major health burden worldwide. Dysregulation of the immune response plays an important role in adverse outcomes in patients with CAP. Methods: We analyzed peripheral blood mononuclear cells by 36-color spectral flow cytometry in adult patients hospitalized for CAP (n=40), matched control subjects (n=31), and patients hospitalized for COVID-19 (n=35). Results: We identified 86 immune cell metaclusters, 19 of which (22.1%) were differentially abundant in patients with CAP versus matched controls. The most notable differences involved classical monocyte metaclusters, which were more abundant in CAP and displayed phenotypic alterations reminiscent of immunosuppression, increased susceptibility to apoptosis, and enhanced expression of chemokine receptors. Expression profiles on classical monocytes, driven by CCR7 and CXCR5, divided patients with CAP into two clusters with a distinct inflammatory response and disease course. The peripheral immune response in patients with CAP was highly similar to that in patients with COVID-19, but increased CCR7 expression on classical monocytes was only present in CAP. Conclusion: CAP is associated with profound cellular changes in blood that mainly relate to classical monocytes and largely overlap with the immune response detected in COVID-19.

Distinct dynamics of antigen-specific induction and differentiation of different CD11c+Tbet+ B-cell subsets.

BACKGROUND: CD11c+Tbet+ B cells are enriched in autoimmunity and chronic infections and also expand on immune challenge in healthy individuals. CD11c+Tbet+ B cells remain an enigmatic B-cell population because of their intrinsic heterogeneity. OBJECTIVES: We investigated severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) antigen-specific development and differentiation properties of 3 separate CD11c+ B-cell subsets-age-associated B cells (ABCs), double-negative 2 (DN2) B cells, and activated naive B cells-and compared them to their canonical CD11c- counterparts. METHODS: Dynamics of the response of the 3 CD11c+ B-cell subsets were assessed at SARS-CoV-2 vaccination in healthy donors by spectral flow cytometry. Distinct CD11c+ B-cell subsets were functionally characterized by optimized in vitro cultures. RESULTS: In contrast to a durable expansion of antigen-specific CD11c- memory B cells over time, both ABCs and DN2 cells were strongly expanded shortly after second vaccination and subsequently contracted. Functional characterization of antibody-secreting cell differentiation dynamics revealed that CD11c+Tbet+ B cells were primed for antibody-secreting cell differentiation compared to relevant canonical CD11c- counterparts. CONCLUSION: Overall, CD11c+Tbet+ B cells encompass heterogeneous subpopulations, of which primarily ABCs as well as DN2 B cells respond early to immune challenge and display a pre-antibody-secreting cell phenotype.

Characterization of myelodysplastic syndromes hematopoietic stem and progenitor cells using mass cytometry.

BACKGROUND: Myelodysplastic syndromes (MDS) at risk of transformation to acute myeloid leukemia (AML) are difficult to identify. The bone marrows of MDS patients harbor specific hematopoietic stem and progenitor cell (HSPC) abnormalities that may be associated with sub-types and risk-groups. Leukemia-associated characteristics of such cells may identify MDS patients at risk of progression to AML and provide insight in the pathobiology of MDS. METHODS: Bone marrow samples from healthy donors (n = 10), low risk (n = 12) and high risk (n = 13) MDS patients were collected, in addition, AML samples for 5 out of 6 MDS patients that progressed. Mass cytometry was applied to assess expression of stem cell subset and leukemia-associated immunophenotype markers. RESULTS: We analyzed the data using FlowSOM to cluster cells with similar expression of 10 commonly used stem cell markers. Metaclusters (n = 20) of these clusters represented populations of cells with a related phenotype, largely resembling known stem cell subsets. Within specific subsets, intra-cellular expression levels of pCREB, IkBα, or pS6 differed significantly between healthy bone marrow (HBM) and MDS or consecutive secondary AML samples. CD34, CD44, and CD49f expression was significantly increased in high risk MDS and AML-associated metaclusters. We identified MDS/sAML cells with aberrant phenotypes when compared to HBM. Such cells were observed in clusters of both primary MDS and secondary AML samples. CONCLUSIONS: High-dimensional mass cytometry and computational data analyses enabled characterization of HSPC subsets in MDS and identification of leukemia stem cell populations based on their immunophenotype. Stem cells in MDS that display leukemia-associated features may predict the risk of developing AML.

INFLECT: an R-package for cytometry cluster evaluation using marker modality.

BACKGROUND: Current methods of high-dimensional unsupervised clustering of mass cytometry data lack means to monitor and evaluate clustering results. Whether unsupervised clustering is correct is typically evaluated by agreement with dimensionality reduction techniques or based on benchmarking with manually classified cells. The ambiguity and lack of reproducibility of sequential gating has been replaced with ambiguity in interpretation of clustering results. On the other hand, spurious overclustering of data leads to loss of statistical power. We have developed INFLECT, an R-package designed to give insight in clustering results and provide an optimal number of clusters. In our approach, a mass cytometry dataset is overclustered intentionally to ensure the smallest phenotypically different subsets are captured using FlowSOM. A range of metacluster number endpoints are generated and evaluated using marker interquartile range and distribution unimodality checks. The fraction of marker distributions that pass these checks is taken as a measure of clustering success. The fraction of unimodal distributions within metaclusters is plotted against the number of generated metaclusters and reaches a plateau of diminishing returns. The inflection point at which this occurs gives an optimal point of capturing cellular heterogeneity versus statistical power. RESULTS: We applied INFLECT to four publically available mass cytometry datasets of different size and number of markers. The unimodality score consistently reached a plateau, with an inflection point dependent on dataset size and number of dimensions. We tested both ConsenusClusterPlus metaclustering and hierarchical clustering. While hierarchical clustering is less computationally expensive and thus faster, it achieved similar results to ConsensusClusterPlus. The four datasets consisted of labeled data and we compared INFLECT metaclustering to published results. INFLECT identified a higher optimal number of metaclusters for all datasets. We illustrated the underlying heterogeneity within labels, showing that these labels encompass distinct types of cells. CONCLUSION: INFLECT addresses a knowledge gap in high-dimensional cytometry analysis, namely assessing clustering results. This is done through monitoring marker distributions for interquartile range and unimodality across a range of metacluster numbers. The inflection point is the optimal trade-off between cellular heterogeneity and statistical power, applied in this work for FlowSOM clustering on mass cytometry datasets.

Generation of immunocompetent syngeneic allograft mouse models for pediatric diffuse midline glioma.

Background: Diffuse midline gliomas (DMG) are highly malignant incurable pediatric brain tumors. A lack of effective treatment options highlights the need to investigate novel therapeutic strategies. This includes the use of immunotherapy, which has shown promise in other hard-to-treat tumors. To facilitate preclinical immunotherapeutic research, immunocompetent mouse models that accurately reflect the unique genetic, anatomical, and histological features of DMG patients are warranted. Methods: We established cell cultures from primary DMG mouse models (C57BL/6) that were generated by brainstem targeted intra-uterine electroporation (IUE). We subsequently created allograft DMG mouse models by orthotopically implanting these tumor cells into syngeneic mice. Immunohistochemistry and -fluorescence, mass cytometry, and cell-viability assays were then used to verify that these murine tumors recapitulated human DMG. Results: We generated three genetically distinct allograft models representing histone 3 wildtype (H3WT) and K27M-mutant DMG (H3.3K27M and H3.1K27M). These allograft models recapitulated the histopathologic phenotype of their human counterparts, including their diffuse infiltrative growth and expression of DMG-associated antigens. These murine pontine tumors also exhibited an immune microenvironment similar to human DMG, characterized by considerable myeloid cell infiltration and a paucity of T-lymphocytes and NK cells. Finally, we show that these murine DMG cells display similar sensitivity to histone deacetylase (HDAC) inhibition as patient-derived DMG cells. Conclusions: We created and validated an accessible method to generate immunocompetent allograft models reflecting different subtypes of DMG. These models adequately recapitulated the histopathology, immune microenvironment, and therapeutic response of human DMG, providing useful tools for future preclinical studies.

Epigenetic Signatures Discriminate Patients With Primary Sclerosing Cholangitis and Ulcerative Colitis From Patients With Ulcerative Colitis.

Background: Primary sclerosing cholangitis (PSC) is a chronic inflammatory liver disease affecting the intra- and extrahepatic bile ducts, and is strongly associated with ulcerative colitis (UC). In this study, we explored the peripheral blood DNA methylome and its immune cell composition in patients with PSC-UC, UC, and healthy controls (HC) with the aim to develop a predictive assay in distinguishing patients with PSC-UC from those with UC alone. Methods: The peripheral blood DNA methylome of male patients with PSC and concomitant UC, UC and HCs was profiled using the Illumina HumanMethylation Infinium EPIC BeadChip (850K) array. Differentially methylated CpG position (DMP) and region (DMR) analyses were performed alongside gradient boosting classification analyses to discern PSC-UC from UC patients. As observed differences in the DNA methylome could be the result of differences in cellular populations, we additionally employed mass cytometry (CyTOF) to characterize the immune cell compositions. Results: Genome wide methylation analysis did not reveal large differences between PSC-UC and UC patients nor HCs. Nonetheless, using gradient boosting we were capable of discerning PSC-UC from UC with an area under the receiver operator curve (AUROC) of 0.80. Four CpG sites annotated to the NINJ2 gene were found to strongly contribute to the predictive performance. While CyTOF analyses corroborated the largely similar blood cell composition among patients with PSC-UC, UC and HC, a higher abundance of myeloid cells was observed in UC compared to PSC-UC patients. Conclusion: DNA methylation enables discerning PSC-UC from UC patients, with a potential for biomarker development.

Discovery and quantification of plastic particle pollution in human blood.

Plastic particles are ubiquitous pollutants in the living environment and food chain but no study to date has reported on the internal exposure of plastic particles in human blood. This study's goal was to develop a robust and sensitive sampling and analytical method with double shot pyrolysis - gas chromatography/mass spectrometry and apply it to measure plastic particles ≥700 nm in human whole blood from 22 healthy volunteers. Four high production volume polymers applied in plastic were identified and quantified for the first time in blood. Polyethylene terephthalate, polyethylene and polymers of styrene (a sum parameter of polystyrene, expanded polystyrene, acetonitrile butadiene styrene etc.) were the most widely encountered, followed by poly(methyl methacrylate). Polypropylene was analysed but values were under the limits of quantification. In this study of a small set of donors, the mean of the sum quantifiable concentration of plastic particles in blood was 1.6 µg/ml, showing a first measurement of the mass concentration of the polymeric component of plastic in human blood. This pioneering human biomonitoring study demonstrated that plastic particles are bioavailable for uptake into the human bloodstream. An understanding of the exposure of these substances in humans and the associated hazard of such exposure is needed to determine whether or not plastic particle exposure is a public health risk.

Analysis of the glyco-code in pancreatic ductal adenocarcinoma identifies glycan-mediated immune regulatory circuits.

Pancreatic ductal adenocarcinoma (PDAC) remains one of the most aggressive malignancies with a 5-year survival rate of only 9%. Despite the fact that changes in glycosylation patterns during tumour progression have been reported, no systematic approach has been conducted to evaluate its potential for patient stratification. By analysing publicly available transcriptomic data of patient samples and cell lines, we identified here two specific glycan profiles in PDAC that correlated with progression, clinical outcome and epithelial to mesenchymal transition (EMT) status. These different glycan profiles, confirmed by glycomics, can be distinguished by the expression of O-glycan fucosylated structures, present only in epithelial cells and regulated by the expression of GALNT3. Moreover, these fucosylated glycans can serve as ligands for DC-SIGN positive tumour-associated macrophages, modulating their activation and inducing the production of IL-10. Our results show mechanisms by which the glyco-code contributes to the tolerogenic microenvironment in PDAC.

Carboxylesterase-1 Assisted Targeting of HDAC Inhibitors to Mononuclear Myeloid Cells in Inflammatory Bowel Disease.

BACKGROUND AND AIMS: Histone deacetylase inhibitors [HDACi] exert potent anti-inflammatory effects. Because of the ubiquitous expression of HDACs, clinical utility of HDACi is limited by off-target effects. Esterase-sensitive motif [ESM] technology aims to deliver ESM-conjugated compounds to human mononuclear myeloid cells, based on their expression of carboxylesterase 1 [CES1]. This study aims to investigate utility of an ESM-tagged HDACi in inflammatory bowel disease [IBD]. METHODS: CES1 expression was assessed in human blood, in vitro differentiated macrophage and dendritic cells, and Crohn's disease [CD] colon mucosa, by mass cytometry, quantitative polymerase chain reaction [PCR], and immunofluorescence staining, respectively. ESM-HDAC528 intracellular retention was evaluated by mass spectrometry. Clinical efficacy of ESM-HDAC528 was tested in dextran sulphate sodium [DSS]-induced colitis and T cell transfer colitis models using transgenic mice expressing human CES1 under the CD68 promoter. RESULTS: CES1 mRNA was highly expressed in human blood CD14+ monocytes, in vitro differentiated and lipopolysaccharide [LPS]-stimulated macrophages, and dendritic cells. Specific hydrolysis and intracellular retention of ESM-HDAC528 in CES1+ cells was demonstrated. ESM-HDAC528 inhibited LPS-stimulated IL-6 and TNF-α production 1000 times more potently than its control, HDAC800, in CES1high monocytes. In healthy donor peripheral blood, CES1 expression was significantly higher in CD14++CD16- monocytes compared with CD14+CD16++ monocytes. In CD-inflamed colon, a higher number of mucosal CD68+ macrophages expressed CES1 compared with non-inflamed mucosa. In vivo, ESM-HDAC528 reduced monocyte differentiation in the colon and significantly improved colitis in a T cell transfer model, while having limited potential in ameliorating DSS-induced colitis. CONCLUSIONS: We demonstrate that monocytes and inflammatory macrophages specifically express CES1, and can be preferentially targeted by ESM-HDAC528 to achieve therapeutic benefit in IBD.

Bayesian log-normal deconvolution for enhanced in silico microdissection of bulk gene expression data.

Deconvolution of bulk gene expression profiles into the cellular components is pivotal to portraying tissue's complex cellular make-up, such as the tumor microenvironment. However, the inherently variable nature of gene expression requires a comprehensive statistical model and reliable prior knowledge of individual cell types that can be obtained from single-cell RNA sequencing. We introduce BLADE (Bayesian Log-normAl Deconvolution), a unified Bayesian framework to estimate both cellular composition and gene expression profiles for each cell type. Unlike previous comprehensive statistical approaches, BLADE can handle > 20 types of cells due to the efficient variational inference. Throughout an intensive evaluation with > 700 simulated and real datasets, BLADE demonstrated enhanced robustness against gene expression variability and better completeness than conventional methods, in particular, to reconstruct gene expression profiles of each cell type. In summary, BLADE is a powerful tool to unravel heterogeneous cellular activity in complex biological systems from standard bulk gene expression data.