Hermansky-Pudlak syndrome type 1 causes impaired anti-microbial immunity and inflammation due to dysregulated immunometabolism

Abstract

Hermansky-Pudlak syndrome (HPS) types 1 and 4 are caused by defective vesicle trafficking. The mechanism for Crohn's disease-like inflammation, lung fibrosis, and macrophage lipid accumulation in these patients remains enigmatic. The aim of this study is to understand the cellular basis of inflammation in HPS-1. We performed mass cytometry, proteomic and transcriptomic analyses to investigate peripheral blood cells and serum of HPS-1 patients. Using spatial transcriptomics, granuloma-associated signatures in the tissue of an HPS-1 patient with granulomatous colitis were dissected. In vitro studies were conducted to investigate anti-microbial responses of HPS-1 patient macrophages and cell lines. Monocytes of HPS-1 patients exhibit an inflammatory phenotype associated with dysregulated TNF, IL-1α, OSM in serum, and monocyte-derived macrophages. Inflammatory macrophages accumulate in the intestine and granuloma-associated macrophages in HPS-1 show transcriptional signatures suggestive of a lipid storage and metabolic defect. We show that HPS1 deficiency leads to an altered metabolic program and Rab32-dependent amplified mTOR signaling, facilitated by the accumulation of mTOR on lysosomes. This pathogenic mechanism translates into aberrant bacterial clearance, which can be rescued with mTORC1 inhibition. Rab32-mediated mTOR signaling acts as an immuno-metabolic checkpoint, adding to the evidence that defective bioenergetics can drive hampered anti-microbial activity and contribute to inflammation.

Fig. 1. Mass cytometry of HPS-1 PBMCs identifies distinct inflammatory monocyte populations.

a FlowSOM analysis of CyTOF data defines 100 cell clusters (individual nodes) organized into 13 metaclusters (node number and color) for 12 controls and 8 HPS-1 patients. We denoted clusters with surface markers that resembled particular cell types. b Specific metacluster 11 and metacluster 5 cell clusters are significantly associated with HPS-1 patients. Metacluster 11.1 was defined by CXCR3, CD64++, CD62L++, metacluster 11.2 by Beta7++, CD64++, CD62L++ and 11.3 by CD16. c Principal component analysis of FlowSOM clusters separates HPS-1 patients and healthy controls. d viSNE analysis of CD14⁺ monocytes defines distinct populations of monocytes in patients with HPS-1 highly expressing CD62L and CD64. e Histogram of marker expression in patients with HPS-1 or Healthy Controls. f Heatmap with hierarchical clustering of inflammatory proteins of O-link inflammation panel for 50 controls, 5 HPS-1, 2 HPS-1 PF, and 4 HPS-1 IBD patients. g Dot plots of IL-1α and TNF protein expression. h Principal component analysis of the O-link inflammatory panel. For CyTOF analysis, populations are compared with an unpaired t-test with Bonferroni-Dunn correction for multiple comparisons. Data are mean+/− SEM. *P < 0.05 **P < 0.01 ***<0.001. All P values are adjusted. For O-link analysis, the Benjamini, Krieger, and Yekutieli multiple nonparametric t-test was performed, ***adjusted p < 0.001, ****adjusted p < 0.0001.

Fig. 2. Spatial transcriptomics uncovers granuloma-associated signatures in the intestine of an HPS-1 patient.

a HPS-1 gastrointestinal tissue used for spatial transcriptomics and immunofluorescently labelled for DNA, CD3, and CD68 staining. b Principal component analysis reveals cellular populations and anatomical regions as main drivers of variation. c Cell deconvolution of spatial transcriptomic profiles, mapped against a selection of cellular profiles provided by human adult gut scRNAseq atlas from the Human Cell Atlas project. d Heatmap of differentially expressed genes (p < 0.001) across the granuloma core, granuloma periphery and mucosa for CD68+CD3− CD45+cells. e HPS1 and HPS4 expression in different immune cell types (n = 4–16). Immune cells from peripheral blood were FACS sorted. Monocytes and monocyte-derived macrophages were isolated using CD14 beads and macrophage differentiation involved M-CSF treatment of monocytes for 5 days. Dot plot, mean and SEM are provided.**p < 0.01, ****p < 0.0001; Statistical significance was determined using a Tukey multiple comparisons test on log10-transformed data.

Fig. 3. HPS1 deficiency leads to a dysregulated metabolic program.

a Differential expression of HPS-1 and control macrophages. Known IBD-related cytokines are bolded. B Weighted Gene Correlation Network Analysis reveals a module highly associated with the HPS-1 disease trait; ClueGO was performed on the top 200 hub genes of this module. c Heatmap of the genes contributing to the LDL particle receptor catabolic signature. d Gene set enrichment analysis using hallmark pathways reveals a reduction in fatty acid metabolism. e Gene set enrichment analysis highlights decreased oxidative phosphorylation. f Sanger sequencing of HPS1 knockout HAP1 cells and control as a comparison. g Oxygen consumption rate of control and HPS1 knockout cells. Error bars in SEM. 12 replicates from one experiment, representative of two independent experiments. h MitoSOX staining on control and HPS1 knockout cells (n = 6). i Serum-starved HAP1 cells were treated with 2.5 μg/mL LDL-BODIPY for 3 h (n = 7). j HAP1 cells were stimulated with cholesterol for 2 h and stained with BODIPY 493/503 (n = 5). *p < 0.05, **p < 0.01, ****p < 0.0001; Ratio paired t-test on mean fluorescence intensity. Unpaired t-test on OCR log-transformed data. Padj = adjusted p-value; NES normalized enrichment score, R & A rotenone and antimycin A.

Fig. 4. Rab32-mediated enhanced mTORC1 signaling in HPS1 deficiency.

a Cholesterol stimulation was performed for 2 h prior to staining with BODIPY 493/503 in patient monocyte-derived macrophages (n = 5). Representative flow cytometry histogram plots. b Schematic of lysosomal immunoprecipitation and confocal microscopy of TMEM192-3xHA and LysoTracker (LT) in HAP1 cells. Lysosomal immunoprecipitation in cells approach was assessed through LAMP2 and CTSC Western blot. c LysoIP proteomics (n = 3) and Western blotting of lysosomal immunoprecipitation samples for mTOR, LAMP2 and TMEM192-3xHA tag (n = 4). d HAP1 cells were serum-starved overnight and treated with 50 μM rapamycin for 3 h prior to pS6 staining (n = 4). e Quantification of Western blot of control and HPS-1 patient macrophage samples for pS6 and GAPDH (n = 3 for healthy controls and HPS-1 patients). Quantification was performed using ImageJ. Western blot image used for quantification. f HAP1 cells were treated with 25 μg/mL LDL for 1 h and then stained for pS6 (n = 3). g LDL uptake in HAP1 cells using LDL-BODIPY for 1 h at 25 μg/mL were pre-treated with 50 μM rapamycin for 3 h (n = 4). h HAP1 cells were transfected with siRNA against Rab32 or a control pool and probed for S6 phosphorylation (n = 5). Statistics: Unpaired parametric test on log10-transformed mean fluorescence intensity data for a. Paired t-test on ratio data for c. Ratio paired t-test on mean fluorescence intensity (d, f, g). An unpaired t-test on ratio data was performed for the pS6/GAPDH quantification.

Fig. 5. mTORC1 inhibition and fatty acids rescue HPS1 bacterial clearance defects.

a HAP1 cells were treated with 10 nM bafilomycin A1 for 2 h and the FlowCellect autophagy kit was used (n = 5). b The gentamicin protection assay was performed in HAP1 cells using GFP-Salmonella Typhimurium (n = 8), Adherent Invasive E. coli (n = 3) or Staphylococcus aureus (n = 4). c Gentamicin protection assay of healthy donor and HPS-1 patient-derived macrophages (n = 10). Representative agar plate of gentamicin protection assay. d Quantification of number of bacteria in each cell using confocal microscopy in control and HPS-1 macrophages (7 controls, 6 HPS-1 patients). The average bacteria per cell in each donor was used for statistical tests, where we performed an unpaired t-test. Representative confocal images; scale bar 5 μm. White asterisk denotes DAPI + GFP- bacteria; yellow asterisk DAPI + GFPbright bacteria; purple asterisk DAPI + GFPdim bacteria. e Differential expression of HPS1 patient and control macrophages following Salmonella Typhimurium infection (n = 4). f HAP1 cells were stimulated with 50 μM cholesterol for 2 h prior to the gentamicin protection assay (n = 5). g HAP1 cells were treated with free fatty acids (FFA) for 2 h prior to a gentamicin protection assay (n = 5). h HAP1 cells were treated with 50 μM rapamycin for 3 h followed by the gentamicin protection assay (n = 4, Salmonella Typhimurium). i Macrophages were treated with 50 μM rapamycin for 2 h prior to Salmonella Typhimurium infection and the gentamicin protection assay (n = 6); each dot represents one patient. Significance was determined using a paired t-test on the autophagy flux values. For gentamicin protection assays, we used a ratio paired t-test for HAP1 cells and an unpaired t-test on log-transformed data for patient macrophages. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Fig. 6. Summary of intestinal inflammation and molecular mechanisms in HPS1 deficiency.

In health, monocytes from the blood differentiate into macrophages, which interact with regulatory T cells and fibroblasts, promoting homeostasis. On a molecular level, lipid droplet metabolism and lower mTORC1 levels facilitate the degradation of microbes, where Rab32 plays a key role in phagosome maturation. In HPS1 deficiency, activated monocytes are recruited from the blood, differentiating into macrophages that secrete inflammatory mediators such as TNF, IL-1, OSM and PTGS2. On an intracellular level, high levels of mTORC1 activity mediated through Rab32 and deregulated immune metabolism dampen anti-microbial activity. The suppressed degradation of bacteria and increased cytokine production in HPS-1 patients can lead to tissue inflammation involving crypt abscesses, epithelial barrier damage, and fistulizing disease. Figure elements are derived from Servier Medical Art.