Multiscale topology classifies cells in subcellular spatial transcriptomics.

Spatial transcriptomics measures in situ gene expression at millions of locations within a tissue1, hitherto with some trade-off between transcriptome depth, spatial resolution and sample size2. Although integration of image-based segmentation has enabled impactful work in this context, it is limited by imaging quality and tissue heterogeneity. By contrast, recent array-based technologies offer the ability to measure the entire transcriptome at subcellular resolution across large samples3-6. Presently, there exist no approaches for cell type identification that directly leverage this information to annotate individual cells. Here we propose a multiscale approach to automatically classify cell types at this subcellular level, using both transcriptomic information and spatial context. We showcase this on both targeted and whole-transcriptome spatial platforms, improving cell classification and morphology for human kidney tissue and pinpointing individual sparsely distributed renal mouse immune cells without reliance on image data. By integrating these predictions into a topological pipeline based on multiparameter persistent homology7-9, we identify cell spatial relationships characteristic of a mouse model of lupus nephritis, which we validate experimentally by immunofluorescence. The proposed framework readily generalizes to new platforms, providing a comprehensive pipeline bridging different levels of biological organization from genes through to tissues.

Prolidase Deficiency Causes Spontaneous T Cell Activation and Lupus-like Autoimmunity.

Prolidase deficiency (PD) is a multisystem disorder caused by mutations in the PEPD gene, which encodes a ubiquitously expressed metallopeptidase essential for the hydrolysis of dipeptides containing C-terminal proline or hydroxyproline. PD typically presents in childhood with developmental delay, skin ulcers, recurrent infections, and, in some patients, autoimmune features that can mimic systemic lupus erythematosus. The basis for the autoimmune association is uncertain, but might be due to self-antigen exposure with tissue damage, or indirectly driven by chronic infection and microbial burden. In this study, we address the question of causation and show that Pepd-null mice have increased antinuclear autoantibodies and raised serum IgA, accompanied by kidney immune complex deposition, consistent with a systemic lupus erythematosus-like disease. These features are associated with an accumulation of CD4 and CD8 effector T cells in the spleen and liver. Pepd deficiency leads to spontaneous T cell activation and proliferation into the effector subset, which is cell intrinsic and independent of Ag receptor specificity or antigenic stimulation. However, an increase in KLRG1+ effector CD8 cells is not observed in mixed chimeras, in which the autoimmune phenotype is also absent. Our findings link autoimmune susceptibility in PD to spontaneous T cell dysfunction, likely to be acting in combination with immune activators that lie outside the hemopoietic system but result from the abnormal metabolism or loss of nonenzymatic prolidase function. This knowledge provides insight into the role of prolidase in the maintenance of self-tolerance and highlights the importance of treatment to control T cell activation.

Xenotransplantation of tannic acid-encapsulated neonatal porcine islets decreases proinflammatory innate immune responses.

BACKGROUND: Islet transplantation with neonatal porcine islets (NPIs) is a promising treatment for type 1 diabetes (T1D), but immune rejection poses a major hurdle for clinical use. Innate immune-derived reactive oxygen species (ROS) synthesis can facilitate islet xenograft destruction and enhance adaptive immune responses. METHODS: To suppress ROS-mediated xenograft destruction, we utilized nanothin encapsulation materials composed of multilayers of tannic acid (TA), an antioxidant, and a neutral polymer, poly(N-vinylpyrrolidone) (PVPON). We hypothesized that (PVPON/TA)-encapsulated NPIs will maintain euglycemia and dampen proinflammatory innate immune responses following xenotransplantation. RESULTS: (PVPON/TA)-encapsulated NPIs were viable and glucose-responsive similar to non-encapsulated NPIs. Transplantation of (PVPON/TA)-encapsulated NPIs into hyperglycemic C57BL/6.Rag or NOD.Rag mice restored euglycemia, exhibited glucose tolerance, and maintained islet-specific transcription factor levels similar to non-encapsulated NPIs. Gene expression analysis of (PVPON/TA)-encapsulated grafts post-transplantation displayed reduced proinflammatory Ccl5, Cxcl10, Tnf, and Stat1 while enhancing alternatively activated macrophage Retnla, Arg1, and Stat6 mRNA accumulation compared with controls. Flow cytometry analysis demonstrated significantly reduced innate immune infiltration, MHC-II, co-stimulatory molecule, and TNF expression with concomitant increases in arginase-1+ macrophages and dendritic cells. Similar alterations in immune responses were observed following xenotransplantation into immunocompetent NOD mice. CONCLUSION: Our data suggest that (PVPON/TA) encapsulation of NPIs is an effective strategy to decrease inflammatory innate immune signals involved in NPI xenograft responses through STAT1/6 modulation without compromising islet function.

ATM is required for SOD2 expression and homeostasis within the mammary gland.

PURPOSE: ATM activates the NF-κB transcriptional complex in response to genotoxic and oxidative stress. The purpose of this study was to examine if the NF-κB target gene and critical antioxidant SOD2 (MnSOD) in cultured mammary epithelium is also ATM-dependent, and what phenotypes arise from deletion of ATM and SOD2 within the mammary gland. METHODS: SOD2 expression was studied in human mammary epithelial cells and MCF10A using RNAi to knockdown ATM or the NF-κB subunit RelA. To study ATM and SOD2 function in mammary glands, mouse lines containing Atm or Sod2 genes containing LoxP sites were mated with mice harboring Cre recombinase under the control of the whey acidic protein promoter. Quantitative PCR was used to measure gene expression, and mammary gland structure was studied using histology. RESULTS: SOD2 expression is ATM- and RelA-dependent, ATM knockdown renders cells sensitive to pro-oxidant exposure, and SOD mimetics partially rescue this sensitivity. Mice with germline deletion of Atm fail to develop mature mammary glands, but using a conditional knockout approach, we determined that Atm deletion significantly diminished the expression of Sod2. We also observed that these mice (termed AtmΔ/Δ) displayed a progressive lactation defect as judged by reduced pup growth rate, aberrant lobulo-alveolar structure, diminished milk protein gene expression, and increased apoptosis within lactating glands. This phenotype appears to be linked to dysregulated Sod2 expression as mammary gland-specific deletion of Sod2 phenocopies defects observed in AtmΔ/Δ dams. CONCLUSIONS: We conclude that ATM is required to promote expression of SOD2 within the mammary epithelium, and that both ATM and SOD2 play a crucial role in mammary gland homeostasis.

The SSBP3 co-regulator is required for glucose homeostasis, pancreatic islet architecture, and beta-cell identity.

OBJECTIVE: Transcriptional complex activity drives the development and function of pancreatic islet cells to allow for proper glucose regulation. Prior studies from our lab and others highlighted that the LIM-homeodomain transcription factor (TF), Islet-1 (Isl1), and its interacting co-regulator, Ldb1, are vital effectors of developing and adult ß-cells. We further found that a member of the Single Stranded DNA-Binding Protein (SSBP) co-regulator family, SSBP3, interacts with Isl1 and Ldb1 in ß-cells and primary islets (mouse and human) to impact ß-cell target genes MafA and Glp1R in vitro. Members of the SSBP family stabilize TF complexes by binding directly to Ldb1 and protecting the complex from ubiquitin-mediated turnover. In this study, we hypothesized that SSBP3 has critical roles in pancreatic islet cell function in vivo, similar to the Isl1::Ldb1 complex. METHODS: We first developed a novel SSBP3 LoxP allele mouse line, where Cre-mediated recombination imparts a predicted early protein termination. We bred this mouse with constitutive Cre lines (Pdx1- and Pax6-driven) to recombine SSBP3 in the developing pancreas and islet (SSBP3ΔPanc and SSBP3ΔIslet), respectively. We assessed glucose tolerance and used immunofluorescence to detect changes in islet cell abundance and markers of ß-cell identity and function. Using an inducible Cre system, we also deleted SSBP3 in the adult ß-cell, a model termed SSBP3Δß-cell. We measured glucose tolerance as well as glucose-stimulated insulin secretion (GSIS), both in vivo and in isolated islets in vitro. Using islets from control and SSBP3Δß-cell we conducted RNA-Seq and compared our results to published datasets for similar ß-cell specific Ldb1 and Isl1 knockouts to identify commonly regulated target genes. RESULTS: SSBP3ΔPanc and SSBP3ΔIslet neonates present with hyperglycemia. SSBP3ΔIslet mice are glucose intolerant by P21 and exhibit a reduction of ß-cell maturity markers MafA, Pdx1, and UCN3. We observe disruptions in islet cell architecture with an increase in glucagon+ α-cells and ghrelin+ ε-cells at P10. Inducible loss of ß-cell SSBP3 in SSBP3Δß-cell causes hyperglycemia, glucose intolerance, and reduced GSIS. Transcriptomic analysis of 14-week-old SSBP3Δß-cell islets revealed a decrease in ß-cell function gene expression (Ins, MafA, Ucn3), increased stress and dedifferentiation markers (Neurogenin-3, Aldh1a3, Gastrin), and shared differentially expressed genes between SSBP3, Ldb1, and Isl1 in adult ß-cells. CONCLUSIONS: SSBP3 drives proper islet identity and function, where its loss causes altered islet-cell abundance and glucose homeostasis. ß-Cell SSBP3 is required for GSIS and glucose homeostasis, at least partially through shared regulation of Ldb1 and Isl1 target genes.

LDB1-mediated transcriptional complexes are sensitive to islet stress.

Excess nutrients and proinflammatory cytokines impart stresses on pancreatic islet ß-cells that, if unchecked, can lead to cellular dysfunction and/or death. Among these stress-induced effects is loss of key ß-cell transcriptional regulator mRNA and protein levels required for ß-cell function. Previously, our lab and others reported that LIM-domain complexes comprised the LDB1 transcriptional co-regulator and Islet-1 (ISL1) transcription factor are required for islet ß-cell development, maturation, and function. The LDB1:ISL1 complex directly occupies and regulates key ß-cell genes, including MafA, Pdx1, and Slc2a2, to maintain ß-cell identity and function. Given the importance of LDB1:ISL1 complexes, we hypothesized that LDB1 and/or ISL1 levels, like other transcriptional regulators, are sensitive to ß-cell nutrient and cytokine stresses, likely contributing to ß-cell (dys)function under various stimuli. We tested this by treating ß-cell lines or primary mouse islets with elevating glucose concentrations, palmitate, or a cytokine cocktail of IL-1ß, TNFα, and IFNγ. We indeed observed that LDB1 mRNA and/or protein levels were reduced upon palmitate and cytokine (cocktail or singly) incubation. Conversely, acute high glucose treatment of ß-cells did not impair LDB1 or ISL1 levels, but increased LDB1:ISL1 interactions. These observations suggest that LDB1:ISL1 complex formation is sensitive to ß-cell stresses and that targeting and/or stabilizing this complex may rescue lost ß-cell gene expression to preserve cellular function.

Islet transplantation into brown adipose tissue can delay immune rejection.

Type 1 diabetes is an autoimmune disease characterized by insulin-producing ß cell destruction. Although islet transplantation restores euglycemia and improves patient outcomes, an ideal transplant site remains elusive. Brown adipose tissue (BAT) has a highly vascularized and antiinflammatory microenvironment. Because these tissue features can promote islet graft survival, we hypothesized that islets transplanted into BAT will maintain islet graft and BAT function while delaying immune-mediated rejection. We transplanted syngeneic and allogeneic islets into BAT or under the kidney capsule of streptozotocin-induced diabetic NOD.Rag and NOD mice to investigate islet graft function, BAT function, metabolism, and immune-mediated rejection. Islet grafts within BAT restored euglycemia similarly to kidney capsule controls. Islets transplanted in BAT maintained expression of islet hormones and transcription factors and were vascularized. Compared with those in kidney capsule and euglycemic mock-surgery controls, no differences in glucose or insulin tolerance, thermogenic regulation, or energy expenditure were observed with islet grafts in BAT. Immune profiling of BAT revealed enriched antiinflammatory macrophages and T cells. Compared with the kidney capsule control, there were significant delays in autoimmune and allograft rejection of islets transplanted in BAT, possibly due to increased antiinflammatory immune populations. Our data support BAT as an alternative islet transplant site that may improve graft survival.

NDRG1 is induced by antigen-receptor signaling but dispensable for B and T cell self-tolerance.

Peripheral tolerance prevents the initiation of damaging immune responses by autoreactive lymphocytes. While tolerogenic mechanisms are tightly regulated by antigen-dependent and independent signals, downstream pathways are incompletely understood. N-myc downstream-regulated gene 1 (NDRG1), an anti-cancer therapeutic target, has previously been implicated as a CD4+ T cell clonal anergy factor. By RNA-sequencing, we identified Ndrg1 as the third most upregulated gene in anergic, compared to naïve follicular, B cells. Ndrg1 is upregulated by B cell receptor activation (signal one) and suppressed by co-stimulation (signal two), suggesting that NDRG1 may be important in B cell tolerance. However, though Ndrg1-/- mice have a neurological defect mimicking NDRG1-associated Charcot-Marie-Tooth (CMT4d) disease, primary and secondary immune responses were normal. We find that B cell tolerance is maintained, and NDRG1 does not play a role in downstream responses during re-stimulation of in vivo antigen-experienced CD4+ T cells, demonstrating that NDGR1 is functionally redundant for lymphocyte anergy.

The transcriptional co-regulator LDB1 is required for brown adipose function.

OBJECTIVE: Brown adipose tissue (BAT) is critical for thermogenesis and glucose/lipid homeostasis. Exploiting the energy uncoupling capacity of BAT may reveal targets for obesity therapies. This exploitation requires a greater understanding of the transcriptional mechanisms underlying BAT function. One potential regulator of BAT is the transcriptional co-regulator LIM domain-binding protein 1 (LDB1), which acts as a dimerized scaffold, allowing for the assembly of transcriptional complexes. Utilizing a global LDB1 heterozygous mouse model, we recently reported that LDB1 might have novel roles in regulating BAT function. However, direct evidence for the LDB1 regulation of BAT thermogenesis and substrate utilization has not been elucidated. We hypothesize that brown adipocyte-expressed LDB1 is required for BAT function. METHODS: LDB1-deficient primary cells and brown adipocyte cell lines were assessed via qRT-PCR and western blotting for altered mRNA and protein levels to define the brown adipose-specific roles. We conducted chromatin immunoprecipitation with primary BAT tissue and immortalized cell lines. Potential transcriptional partners of LDB1 were revealed by conducting LIM factor surveys via qRT-PCR in mouse and human brown adipocytes. We developed a Ucp1-Cre-driven LDB1-deficiency mouse model, termed Ldb1ΔBAT, to test LDB1 function in vivo. Glucose tolerance and uptake were assessed at thermoneutrality via intraperitoneal glucose challenge and glucose tracer studies. Insulin tolerance was measured at thermoneutrality and after stimulation with cold or the administration of the ß3-adrenergic receptor (ß3-AR) agonist CL316,243. Additionally, we analyzed plasma insulin via ELISA and insulin signaling via western blotting. Lipid metabolism was evaluated via BAT weight, histology, lipid droplet morphometry, and the examination of lipid-associated mRNA. Finally, energy expenditure and cold tolerance were evaluated via indirect calorimetry and cold challenges. RESULTS: Reducing Ldb1 in vitro and in vivo resulted in altered BAT-selective mRNA, including Ucp1, Elovl3, and Dio2. In addition, there was reduced Ucp1 induction in vitro. Impacts on gene expression may be due, in part, to LDB1 occupying Ucp1 upstream regulatory domains. We also identified BAT-expressed LIM-domain factors Lmo2, Lmo4, and Lhx8, which may partner with LDB1 to mediate activity in brown adipocytes. Additionally, we observed LDB1 enrichment in human brown adipose. In vivo analysis revealed LDB1 is required for whole-body glucose and insulin tolerance, in part through reduced glucose uptake into BAT. In Ldb1ΔBAT tissue, we found significant alterations in insulin-signaling effectors. An assessment of brown adipocyte morphology and lipid droplet size revealed larger and more unilocular brown adipocytes in Ldb1ΔBAT mice, particularly after a cold challenge. Alterations in lipid handling were further supported by reductions in mRNA associated with fatty acid oxidation and mitochondrial respiration. Finally, LDB1 is required for energy expenditure and cold tolerance in both male and female mice. CONCLUSIONS: Our findings support LDB1 as a regulator of BAT function. Furthermore, given LDB1 enrichment in human brown adipose, this co-regulator may have conserved roles in human BAT.