About

Mark A Atkinson is the American Diabetes Association Eminent Scholar for Diabetes Research and the Jeffrey Keene Family Professor at the University of Florida. With over 40 years of investigation into type 1 diabetes, he has authored more than 700 publications and holds an h-index of 121. His research primarily focuses on understanding the causes of type 1 diabetes, developing methods to predict its onset, and finding ways to cure the disease. Throughout his career, Dr. Atkinson has contributed significantly to the field through translational research, clinical trials, and studies of pancreatic pathology, immunology, and genetics related to type 1 diabetes. He has received numerous awards for his scientific and humanitarian efforts, including awards from the Juvenile Diabetes Research Foundation, the American Diabetes Association, and international recognitions such as the Eli Lilly Award and the Jacobaeus Award from the Novo Nordisk Foundation. Dr. Atkinson has also played active roles in leadership and advisory capacities within major diabetes research organizations, including JDRF, NIH, and ADA, and has served as an editor for prominent journals. His work has advanced the understanding of disease prediction, environmental factors, stem cell and pancreatic regeneration, and immunoregulation in type 1 diabetes. He has also been involved in training and mentoring the next generation of diabetes researchers and is President of Insulin for Life USA, supporting insulin access in developing countries.

Research topics

• Medicine
• Internal medicine
• Biology
• Immunology
• Endocrinology
• Genetics
• Pathology
• Bioinformatics
• Computational biology
• Gerontology
• Environmental health
• Psychiatry
• Physiology
• Virology
• Pediatrics

Selected publications

• Spatial transcriptomics from pancreas and local draining lymph node tissue reveals a lymphotoxin-β signature in human type 1 diabetes

  Cell Reports · 2026-03-23

  articleOpen access

  This study explores the inflammatory response observed in the pancreas and pancreatic lymph nodes (pLNs) during the natural history of type 1 diabetes (T1D). Using multicell-resolution spatial transcriptomics (ST), we profile individuals without diabetes (ND), at-risk autoantibody-positive (AAb+) individuals, and T1D donors. In the T1D pancreas, we observed global upregulation of inflammation-associated transcripts, including REG family genes, C3, SOD2, and OLFM4. In the T1D pLN, LTB was significantly upregulated within the lymphoid follicles. Using an orthogonal subcellular-resolution ST platform on an independent donor set, we identified follicular B cells as the primary source of LTB in the pLN and observed increased LTB expression in lymphocytes in insulitic lesions proximal to CCL19/CCL21-expressing endothelium. Collectively, these findings highlight lymphotoxin-β and downstream chemokine signatures in the pancreatic lymphatics as well as within the insulitic lesion, which can inform future therapeutic interventions.

  PublisherDOI

• Association Between Liver Dysfunction Markers and Mortality in Heart Failure: A Systematic Review and Meta-Analysis

  Cureus · 2026-04-03

  articleOpen access

  Heart failure (HF) is a major cause of morbidity and mortality worldwide and is frequently associated with dysfunction of other organs, including the liver. Hepatic congestion and impaired perfusion from cardiac dysfunction may lead to liver injury, described as cardiohepatic syndrome. Studies suggest that liver dysfunction markers, such as the Model for End-Stage Liver Disease (MELD) and MELD excluding international normalized ratio (MELD-XI) scores, may serve as prognostic indicators in HF patients. This systematic review and meta-analysis evaluated the relationship between liver dysfunction markers and mortality outcomes in HF patients. A systematic literature search was conducted. Studies evaluating the association between liver dysfunction markers (MELD or MELD-XI scores) and mortality outcomes in adult HF patients were included. Hazard ratios (HRs) with 95% confidence intervals (CIs) were pooled using the generic inverse variance method with a random-effects model in Review Manager (RevMan) version 5.4 (The Cochrane Collaboration, Copenhagen, Denmark). Subgroup analyses were performed by liver dysfunction marker type and HF phenotype. The primary outcome was all-cause mortality. Nine studies comprising over 32,000 HF patients were included. Elevated liver dysfunction markers were significantly associated with increased mortality risk (pooled HR = 1.12, 95% CI 1.06-1.18). Substantial heterogeneity was observed (I² = 91%). Subgroup analysis showed significant associations for both MELD-XI (HR = 1.09, 95% CI 1.04-1.14) and MELD scores (HR = 1.10, 95% CI 1.06-1.14). Analysis by HF type revealed significant associations in acute HF (HR = 1.07), chronic HF (HR = 1.23), and advanced HF (HR = 1.05). Sensitivity analyses confirmed the findings, with no substantial publication bias observed. Elevated liver dysfunction markers significantly correlate with increased mortality in HF patients. Both MELD and MELD-XI scores provide prognostic information across HF populations. These findings emphasize the importance of considering hepatic dysfunction in HF assessment and suggest liver dysfunction markers may aid risk stratification. Further prospective studies are needed to determine the role of these markers in clinical decision-making.

  PublisherOA PDFDOI

• Empiric azithromycin alters the upper respiratory microbiome and resistome without anti-inflammatory benefit in COVID-19

  Nature Microbiology · 2026-03-16 · 1 citations

  articleOpen access

  Azithromycin is a widely used antibiotic and was frequently used to treat hospitalized patients during the COVID-19 pandemic. The impact of empiric azithromycin use on the respiratory microbiome in patients with viral respiratory infections is unclear. Here we used longitudinal metatranscriptomics on nasal swabs from a prospective multicentre cohort of 1,164 patients hospitalized for COVID-19. We compared the upper respiratory microbiome, resistome and systemic immune response in patients treated with azithromycin (n = 366) with those who received no antibiotics (n = 474) or other antibiotics (n = 324). We found that azithromycin altered microbiome composition and increased the expression and relative proportion of macrolide/lincosamide/streptogramin (MLS) resistance genes. These changes occurred after 1 day of exposure and persisted for over a week. MLS resistance gene expression was associated with commensals and potential pathogens, while there were no differences in host inflammatory gene expression in blood and airways. This demonstrates that empiric azithromycin treatment impacts the upper respiratory microbiome and resistome without apparent anti-inflammatory benefit.

  PublisherOA PDFDOI

• Beta cell–targeted PD-1 agonist inhibits cell-mediated autoimmunity in pancreas tissue slices

  Science Advances · 2026-04-01

  articleOpen access

  This research evaluates a therapeutic approach based on tissue-targeted immunomodulation with a potential broad application to treat autoimmune diseases including type 1 diabetes (T1D). We generated a bispecific immune agonist that binds beta cells and suppresses autoreactive T cells. These bispecific molecules called immune modulating monoclonal-T cell receptor (TCR) against autoimmune disease (ImmTAAI), consist of a human-specific TCR-targeting domain fused with a programmed death-1 agonist. We used live pancreas slices to demonstrate targeting of ImmTAAI molecules to preproinsulin peptide-HLA-A2 complexes on human beta cells. ImmTAAI molecules protected beta cells from T cell killing by increasing T cell motility and inhibiting effector molecule and cytokine secretion. ImmTAAI treatment also increased the motility of islet-infiltrating T cells in slices from a donor with recent-onset T1D and preserved insulin secretion in slices cocultured with T cell avatars transduced with diabetogenic TCRs. These data demonstrate that ImmTAAI molecules have the potential to limit T cell activity locally, making this an attractive platform to elicit targeted immunoregulation in T1D.

  PublisherDOI

• Virtual multiplex staining of the pancreatic islets across type 1 diabetes progression using a Schrödinger bridge

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-04-17

  articleOpen access

  Classical hematoxylin and eosin (H&E) staining enables review of tissue morphology but lacks information regarding the molecular state of cells. Immunohistochemical (IHC) techniques label specific proteins in tissue, allowing differentiation of relevant structures that may go undetectable in H&E. However, the IHC process is complex, expensive, and time-consuming, especially for multiplex IHC (mIHC) limiting its use in large cohorts. Stain conversion of H&E to IHC using generative artificial intelligence models such as generative adversarial networks (GANs) represent one solution to this problem. However, GANs are unstable during out of distribution sampling and are prone to hallucinations or mode collapse, limiting their accuracy in challenging image conversion tasks. To address this, the field has recently turned to diffusion models. Here, we introduce Schrödinger-bridge for Multiplex ImmunoLabel Estimation (SMILE). Unlike conventional diffusion models that map from source to target through an intermediate Gaussian noise, Schrödinger-bridge diffusion models skip this step and have been shown to better preserve structures during image translation. To test the performance of SMILE, we generated a large cohort of high-fidelity H&E-mIHC image pairs from pancreatic organ donors, targeting insulin, glucagon, and CD3. Our dataset well-sampled across type-1 diabetes status, pancreas anatomical location, age, and sex. Using this cohort, we demonstrate the superiority of SMILE compared to GANs via a comprehensive evaluation framework incorporating texture, distribution, and antibody-specific metrics, as well as blinded pathologist reviews. We further confirmed the ability of SMILE to generate accurate mIHC images from H&Es generated at an external site, to perform whole slide image conversion, and to generate realistic three-dimensional maps of the pancreatic islets in non-diabetic, auto-antibody positive, and type-1 diabetic donor tissue. Finally, we performed stain conversion of paired H&E to HER2 and Ki67 images in breast cancer, confirming the superiority of SMILE in diverse stain conversion applications. Collectively, this framework provides a scalable pipeline for high-throughput proteomic inference from archival H&Es, providing transformative potential for pancreatic research and digital pathology.

  PublisherDOI

• Machine learning models predict long COVID outcomes based on baseline clinical and immunologic factors

  Communications Medicine · 2026-01-03 · 1 citations

  articleOpen access

  The post-acute sequelae of SARS-CoV-2 (PASC), also known as long COVID, remain a significant health issue that is incompletely understood. Predicting which acutely infected individuals will develop long COVID is challenging due to the absence of established biomarkers, clear disease mechanisms, or well-defined sub-phenotypes. Machine learning (ML) models may address this gap by leveraging clinical data to enhance diagnostic precision. Clinical data, including antibody titers and viral load measurements collected at the time of hospital admission, are used to predict the likelihood of acute COVID-19 progressing to long COVID. Machine learning models are trained and evaluated for predictive performance. Feature importance analysis is performed to identify the most influential predictors. The machine learning models achieve median AUROC values ranging from 0.64 to 0.66 and AUPRC values between 0.51 and 0.54, demonstrating predictive capabilities. Low antibody titers and high viral loads at hospital admission emerge as the strongest predictors of long COVID outcomes. Comorbidities—such as chronic respiratory, cardiac, and neurologic diseases—and female sex are also identified as significant risk factors. Machine learning models identify patients at risk for developing long COVID based on baseline clinical characteristics. These models guide early interventions, improve patient outcomes, and mitigate the long-term public health impacts of SARS-CoV-2. Long COVID, or post-acute sequelae of SARS-CoV-2, is a prolonged health condition that can occur after acute COVID-19 infection. However, the ability to predict who will develop long COVID remains limited due to the absence of clear tests or biomarkers. We looked at patients’ medical information, including the amount of virus in their body at hospital admission, and how strong their immune response was. Using computer programs that can find hidden patterns in large sets of data, we discovered that people with a weaker immune response, higher amounts of virus, certain long term health problems and women are more likely to develop long COVID. This study highlights that computer-based tools could help doctors identify high-risk patients early and provide care that may prevent long-term complications. Jayavelu, Samaha et al., apply machine learning models on hospital admission data, including antibody titers and viral load, to identify patients at high risk for Long COVID. Low antibody levels, high viral loads, chronic diseases, and female sex are key predictors, supporting early, targeted interventions.

  PublisherOA PDFDOI

• Author Correction: Redox regulation of m6A methyltransferase METTL3 in β-cells controls the innate immune response in type 1 diabetes

  Nature Cell Biology · 2025-11-17

  articleOpen access
  PublisherOA PDFDOI

• Evaluating COVID-19 severity prediction and immune dynamics with NULISAseq: Insights from the IMPACC study.

  PubMed · 2025-12-01

  articleOpen access

  The National Institutes of Health-funded IMPACC (IMmunoPhenotyping Assessment in a COVID-19 Cohort) evaluated longitudinal clinical and immunological features of human patients hospitalized for COVID-19. This study focuses on comparing the novel NULISAseq assay with the Olink platform using a subset of participants to assess their efficacy in predicting COVID-19 severity and understanding immune response dynamics. Our findings reveal that NULISAseq could provide superior detectability and dynamic range across various targets. Elastic net analysis demonstrated that specific proteins, including amphiregulin, effectively predict COVID-19 severity from sera at admission (samples drawn within 96 h of admission), with a test area under the curve of 0.84. Longitudinal analysis identified significant differences in multiple targets, including IL-5 and interferons, between low- and high-severity groups over time. Additionally, association rule mining suggested potential early markers predictive of later immune cell changes. These findings emphasize the potential of NULISAseq for comprehensive profiling, early prediction, and identification of targeted therapeutic interventions in COVID-19.

  PublisherOA PDFDOI

• Single-Islet Proteomics Maps Pseudo-Temporal Islet Immune Responses and Dysfunction in Stage 1 Type 1 Diabetes

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-11-12

  preprintOpen access

  Progressive β-cell dysfunction precedes the onset of type 1 diabetes (T1D), yet the molecular mechanisms driving early T1D development remain poorly understood. Although single-cell RNA-sequencing has uncovered transcript-level changes in human islet cells, it offers limited insight into the heterogeneity of distinct islet microenvironments. Here, we applied a single-islet proteomics workflow to profile intra-donor islet heterogeneity in three stage 1 T1D cases with matched non-diabetic controls and define in situ protein signatures of pseudo-temporal islet dysfunction. Intra-donor analyses of ~100 individual islets per donor revealed highly consistent proteomic patterns reflecting pseudo-time progression of islet immune responses and β-cell dysfunction. Several pathways, including extracellular matrix remodeling and mRNA processing, were identified as closely associated with progressive islet immune activation and loss of β-cell function. These findings provide robust proteome-wide evidence of the progression of islet dysfunction, offer a valuable resource for investigating early mechanisms of T1D pathogenesis-including novel candidates for functional studies-and underscore the utility of single-islet spatial proteomics for examining islet heterogeneity in T1D.

  PublisherDOI

• 3D imaging of human pancreas suggests islet size and endocrine composition influence their loss in type 1 diabetes

  Nature Communications · 2025-12-11 · 3 citations

  articleOpen accessSenior authorCorresponding

  A high-definition description of pancreatic islets would prove beneficial for understanding the pathophysiology of type 1 diabetes (T1D), yet significant knowledge gaps exist in terms of their size, endocrine cell composition, and number in both health and disease. Here, 3-dimensional (3D) analyses of pancreata from control persons without diabetes (ND) demonstrate approximately 50% of islets are insulin-positive (INS + ) glucagon-negative (GCG-). Non-diabetic individuals positive for a single Glutamic acid decarboxylase autoantibody (GADA + ) yet at increased risk for disease consistently demonstrate endocrine features, including islet volume and cell composition, closely resembling the age-matched ND controls. In contrast, pancreata from individuals with short-duration T1D demonstrate significantly reduced islet density and a dramatic loss of INS + GCG- islets with preservation of large INS + GCG+ islets. The size and cellular composition of pancreatic islets may, therefore, represent influential factors that impact β-cell loss during T1D disease progression. LSFM imaging revealed 3D islet maps: GADA+ cases mirrored controls, while short-duration T1D showed loss of small INS + GCG– islets but preserved larger mixed INS + GCG+ ones, highlighting islet size- and composition-dependent vulnerability

  PublisherDOI

Recent grants

• NIH Grant R21DK077580

  NIH · $191k · 2009

• Regional and lobular heterogeneity of human pancreas morphology and function in type 1 diabetes pathogenesis

  NIH · $1.7M · 2020–2025

• Co-registration of Cell Organization, Phenotype and Function in the Human Pancreas During Type 1 Diabetes

  NIH · $2.5M · 2021–2027

• NIH Grant UC4DK108132

  NIH · $3.6M · 2020

• NIH Grant M01RR000082

  NIH · $30.1M · 2011

Frequent coauthors

• Clive Wasserfall

  University of Florida

  323 shared

• Desmond Schatz

  University of Florida

  260 shared

• Todd M. Brusko

  161 shared

• Martha Campbell‐Thompson

  University of Florida

  119 shared

• Michael J. Haller

  117 shared

• María J. Redondo

  Texas Children's Hospital

  99 shared

• Jay M. Sosenko

  88 shared

• Carmella Evans‐Molina

  80 shared

Education

• PhD, Pathology & Laboratory Medicine

  University of Florida

  1987

Awards & honors

• Gerold and Gayla Grodsky award (2001)
• Mary Tyler Moore & S. Robert Levine M.D. award (2004, 2008,…
• JDRF’s David Rumbough award (2005)
• Eli Lilly Award for Outstanding Scientific Achievement from…
• Barbara Davis Award for contributions to the field seeking t…