Jonathan M. Dreyfuss

Replication of scientific findings across independent studies is essential, yet it remains challenging in high-dimensional biological datasets. A critical aspect of true replication requires detecting not only the repeated presence but also the consistent direction of an effect. Traditional methods for evaluating replication often ignore effect directionality or lack adequate statistical power when adjusting for multiple comparisons. To address this, we extend the directional MaxP test (DMT) and our screening-integrated procedure, ScreenDMT, to evaluate directional replication across an arbitrary number of independent studies. We mathematically prove ScreenDMT’s validity for an arbitrary number of studies when the studies’ sample sizes grow unbounded. Through extensive simulation studies, we demonstrate that ScreenDMT effectively controls error rates, such as the false discovery rate (FDR), in finite sample sizes and yields higher statistical power than competing methods especially for three studies. We applied the extended ScreenDMT to untargeted blood metabolomics profiles across three diverse human cohorts—ARIC (N=1,337), MESA (N=1,874), and HCHS/SOL (N=6,180)—to discover replicable biomarkers for dichotomous insomnia, long sleep, short sleep, and excessive daytime sleepiness (EDS). We found low levels of cortisol to be replicably associated with EDS, and high levels of n-formylmethionine to be associated with long sleep. Furthermore, directional pathway analysis revealed that low levels of xanthine metabolites are replicably associated with insomnia. These convergent findings validate ScreenDMT as a powerful methodology for simultaneous multi-cohort replication and offer insights into the metabolic signatures underlying sleep disorders across diverse populations.

Replenishment of pancreatic β cells is key to a cure for diabetes. β cell regeneration is achieved predominantly by self-replication, especially in rodents, but it was also shown that pancreatic duct cells can transdifferentiate into β cells. How pancreatic duct cells are transdifferentiated and whether we can manipulate transdifferentiation to replenish β cell mass are not well understood. Using a genome-wide clustered regularly interspaced short palindromic repeats (CRISPR) screen, we found that the loss of function of aldehyde dehydrogenase family 3 member B2 ( ALDH3B2 ) was sufficient to transdifferentiate cell line–based and human pancreatic duct cells into functional β-like cells. The transdifferentiated cells had substantially increased the expression of β cell marker genes, secreted insulin in response to glucose, and lowered blood glucose to near normal for 6 weeks after transplantation into streptozotocin-induced diabetic mice under the kidney capsule. Our study identifies a gene that could potentially be targeted in human pancreatic duct cells to replenish β cell mass for diabetes therapy.

Background: Refractory hypoglycemia is common in malignant insulin-secreting tumors and is associated with substantial morbidity. Ersodetug is a fully human monoclonal antibody that allosterically attenuates insulin receptor signaling with the potential to treat various forms of hyperinsulinism (HI). We retrospectively report outcomes from compassionate use of ersodetug in 8 individuals with refractory hypoglycemia due to tumor HI. Methods: Ersodetug was administered intravenously at 6 or 9 mg/kg every 1-2 weeks initially, followed by a frequency of every 2-5 weeks, as appropriate. Glycemic, functional, and safety outcomes were analyzed. Results: Eight adults (4M/4F; 24-74 years; Eastern Cooperative Oncology Group [ECOG] 1-3) received ersodetug for insulin-secreting tumors (metastatic insulinoma, n = 7; cervical neuroendocrine carcinoma, n = 1). No drug-related serious adverse events were observed. Most patients experienced improved glycemic control, including discontinuation of parenteral glucose in 6 of the 7 applicable patients (median, 4.5 days) and a 35.6% relative reduction from baseline in time in hypoglycemia (<70 mg/dL) by continuous glucose monitoring (mean 12.4-8.0%; pseudo-median paired change of -3.8 percentage points [90% CI, −8.2 to −1.1]; P = .02). These glycemic improvements permitted hospital discharge and reductions in other antihypoglycemic therapies beyond glucose infusion rate (pseudo-median change, −2.5; 90% CI, −3.5 to −1.0; P = .02), with associated improvements in ECOG status. The median treatment duration with ersodetug (11.5 months, range 5-17) tended to correspond to patient lifespan in the setting of metastatic disease. Conclusion: Ersodetug therapy resulted in reduced hypoglycemia burden, ability to discontinue parenteral glucose infusion, allowing discharge from hospital, use of fewer antihypoglycemic therapies, and improved ECOG performance status in individuals with severe, refractory tumor HI.

Innate immunity predictors of HIV-1 risk and pathogenesis vary with reproductive hormones, pregnancy, and lactation, yet the underlying mechanisms remain unclear. We hypothesized that pregnancy-associated physiological adaptations alter systemic microRNA (miRNA) expression, thereby regulating immunity, pathogenesis and susceptibility to infection. We analyzed 174 serum samples from 88 participants in a longitudinal cohort from Uganda and Zimbabwe across pre-pregnancy (PP), pregnancy (P), and postpartum breastfeeding (BF). Cell-free peripheral blood miRNAs (n = 2083) were profiled using HTG EdgeSeq. Pregnancy-specific miRNAs were identified by intersecting differentially expressed (DE) miRNAs from P vs. PP and P vs. BF comparisons. miRNA targets and pathways were analyzed using miRWalk, Cytoscape/ClueGO, and cytoHubba. Pregnancy was associated with DE miRNAs (29 upregulated and 131 downregulated) targeting 2733 validated genes. Enriched pathways (FDR < 0.05) included adaptive immune response, Hippo Signaling, Cellular Senescence, HSV-1 infection, and two cancer-related pathways. Pregnancy-enriched targets within each pathway overlapped with the HIV–host interactome by 37–88%. Network analysis identified 47 hub genes interacting with 18 HIV-1 proteins, with Tat and gp120 being most connected viral and HLA-A being the most connected host protein. These findings indicate that pregnancy-driven systemic miRNAs target the HIV–host interactome and specifically identify pregnancy-enriched central hub genes involved in cell cycle control, viral immune evasion and replication to be further investigated for their predictive value in HIV acquisition and pathogenesis in longitudinal cohorts and experimental settings.

Biological aging greatly impacts the body’s ability to handle glucose, and represents a major risk factor the development and progression of type 2 diabetes (T2D). Nonetheless, despite advances in cellular senescence research and the development of new senolytic therapies, the heterogeneity of cellular senescence in the human endocrine pancreas, as well as its roles in normal aging, remains to be elucidated at the single-cell level. Here, we performed single-cell–resolved spatial proteomics and transcriptomics on intact pancreas from 26 donors (ages 20–80) and multiplexed single-cell RNA sequencing and functional assays on dispersed islets from 14 donors (ages 34–69). We identify two discrete SnC subpopulations distinguished by relative expression of CDKN1A and CDKN2A. CDKN1A ⁺ senescent cells (SnCs) exhibit loss of β-cell identity, impaired insulin secretion, and a proinflammatory SASP associated with increased islet immune infiltration. In contrast, CDKN2A ⁺ SnCs retain transcriptional identity and functional competence, with lower inflammatory signaling. Together, these findings identify heterogeneous and functionally divergent senotypes in the human pancreas, distinguishing an adaptive ( CDKN2A ⁺) from a maladaptive ( CDKN1A ⁺) senescence program, thus providing a mechanism-guided framework for senescence-targeted therapies in T2D.

Adipocytes exist along a functional spectrum: white adipocytes are energy storing, and brown adipocytes have thermogenic capacity such that activation may counteract obesity-related disease. In between are UCP1-expressing beige adipocytes, which can transition between these two energetic states. We previously showed that bone morphogenetic protein 7 (BMP7), a member of the TGF-β superfamily, enables differentiation of brown preadipocytes to mature thermogenic cells. To see whether immortalized, clonal human white and brown preadipocytes (hWAs and hBAs, respectively) would become more thermogenic in response to BMP exposure, we treated them with BMP7 or BMP4 for the first 7 days of a 30-day differentiation protocol. In hBAs, absence of either BMP7 or BMP4 led to lower expression of brown-specific markers and oxygen consumption relative to 7 days with either BMP. hWAs treated for 7 days with either BMP did not increase expression of thermogenic protein UCP1 nor induce a brown-like transcription profile. However, BMP-treated hWAs produced adipocytes that had higher basal and drug-induced maximal oxygen consumption, which was UCP1-independent and due substantially to the futile creatine cycle. Our results demonstrate that energetically quiescent hWAs can be pushed into an energy-expending phenotype without transdifferentiation into beige adipocytes, providing a new approach to treat obesity-related metabolic disease.

Somatically acquired mitochondrial DNA (mtDNA) mutations accumulate with age, but the mechanisms and consequences of this accumulation are poorly understood. Here we show that transient injuries induce a burst of persistent mtDNA mutations that impair resilience to future injuries. mtDNA mutations suppressed energy-intensive nucleotide metabolism. Repletion of adenosine, but not other nucleotides, restored adenosine triphosphate generation, which required a nuclear-encoded purine biosynthetic enzyme, adenylate kinase 4 (AK4). Analysis of 369,912 UK Biobank participants revealed a graded association between mutation burden and chronic kidney disease severity as well as an independent increase in the risk of future acute kidney injury events ( P < 10 ⁻⁷ ). Heteroplasmic mtDNA mutations may therefore reflect the cumulative effect of acute injuries to metabolically active cells, impairing major functions in a fashion amenable to nuclear-controlled purine biosynthesis.

Innate immunity predictors of HIV risk are influenced by reproductive hormone levels, pregnancy, and lactation status. However, the molecular mechanisms underlying these associations remain unclear. MicroRNAs (miRNAs), as post-transcriptional regulators, play key roles in immune regulation and host-virus interactions. We hypothesized that physiological adaptations to pregnancy include the modulation of systemic miRNA expression, which in turn regulates host genes involved in HIV interaction, potentially influencing susceptibility during pregnancy. To test this, we leveraged a large longitudinal cohort from Uganda and Zimbabwe and analyzed 174 serum samples from 88 participants in pre-pregnancy (PP), pregnancy (P), and breastfeeding (BF) states using the HTG EdgeSeq platform. Differentially expressed (DE) miRNAs in pregnancy were identified with false discovery rate (FDR) < 0.1 (29 upregulated and 131 downregulated) by intersecting pairwise comparisons (P vs. PP and P vs. BF). Validated gene targets (2,733) were identified via miRWalk and modular enrichment analysis performed using Cytoscape/ClueGO. Enriched pathways (FDR<0.05) included Adaptive Immune Response, Hippo Signaling, Cellular Senescence, HSV-1 Infection, and two cancer-related pathways. Genes from pregnancy-enriched pathways overlapped with the known HIV-host interactome at the range of 37 to 88%. From this overlap, the Maximal Clique Centrally (MCC) score (cytoHubba) identified 47 unique hub genes acting as essential regulatory nodes of protein-protein interaction (PPI) network. These hub genes were further examined by their expression patterns predicted by DE miRNAs, and explored for their interactions with the HIV interactome, revealing connectivity with 18 HIV proteins - highest with Tat and gp120 - and impact on the HIV replication process. HLA-A was the most connected host gene, followed by BCL2L1, EIF2AK2, PTEN, and CYCS. These findings support the hypothesis that pregnancy-driven systemic miRNAs may shape HIV susceptibility by regulating hub genes with central role in viral evasion of host immunity. GRAPHIC

Background: Physiological changes during pregnancy result in altered maternal drug metabolism that impacts efficacy and safety of therapeutics in pregnant women. Pregnancy induced hormonal, immunologic, or metabolic changes may also influence and alter drug disposition. Despite research efforts focused on pharmacokinetics of medications used in pregnant women in the past decade, knowledge gaps exist in understanding how pregnancy influences drug disposition and placental drug transporters. Moreover, there is a scarcity of research in understanding the safety and effectiveness of therapeutics in lactating women. This study aimed to determine the effect of pregnancy on levels of miRNAs regulating drug metabolizing enzymes and transporters (DMET). Methods: We utilized longitudinal serum specimens collected in 3-month intervals from 88 women who became pregnant during follow-up in a large prospective study of hormonal contraception and HIV acquisition in Uganda and Zimbabwe. We used the HTG EdgeSeq platform coupled with Illumina sequencing to obtain the global miRNA transcriptome in paired specimens collected before, during and after pregnancy. To identify differentially expressed (DE) miRNAs that distinguish pregnancy from pre-conception or breastfeeding we used mixed effect model accounting for multiple samples from the pregnancy event and controlling for fixed effects of batch, country, Nugent score category and sexually transmitted infections. To identify hormonally regulated miRNAs independently associated with Box-Cox-transformed levels of progesterone (P4), β-estradiol (E2), and sex-hormone binding protein (SHBG) we controlled in addition for age, pregnancy and breastfeeding P-values were corrected using the Benjamini-Hochberg false discovery rate (FDR). DMET-targeting miRNAs were identified using miRTarBase focusing on interactions verified by 3’-UTR luciferase reporter assay and overlapped with DE miRNAs with FDR < 0.05. Results: Of 140 DMET-targeting miRNAs among the 2079 miRNAs in the global peripheral blood transcriptome, 41 unique DMET-targeting miRNAs were found to be DE during pregnancy – 38 differentiating pregnancy from preconception and 9 differentiating pregnancy from breastfeeding. The 56 DMETs confirmed as targets of the DE miRNAs included 8 members of the ABC (ATP-binding cassette) transporter family, all abundantly expressed in the placenta, and 4 members of the cytochrome P450 Phase 1 enzyme family with major role in xenobiotics detoxification. The study also revealed a strong (FDR<0.05), predominantly positive association between specific DMET-targeting miRNAs and sex hormone-binding globulin (SHBG) levels, suggesting a miRNA-mediated downregulation of DMETs as SHBG levels rise during pregnancy. Conclusion: This research provides crucial insights into the molecular mechanisms underlying altered drug disposition in pregnant and lactating women, paving the way for improved therapeutic management and personalized medicine in these populations.