Emergence of inflammatory fibroblasts with aging in Hermansky-Pudlak syndrome associated pulmonary fibrosis

Abstract

The longitudinal cellular interactions that drive pulmonary fibrosis are not well understood. To investigate the disease underpinnings associated with fibrosis onset and progression, we generated a scRNA-seq atlas of lungs from young and aged mouse models of multiple subtypes of Hermansky-Pudlak syndrome (HPS), a collection of rare autosomal recessive diseases associated with albinism, platelet dysfunction, and pulmonary fibrosis. We have identified an age-dependent increase in SAA3⁺ inflammatory lung fibroblasts in HPS mice, including in double-mutant HPS1-2 mice which develop spontaneous fibrosis. HPS1 fibroblasts show increased expression of IL-1R1, whereas alveolar type II epithelial cells from HPS2 mice induce the inflammatory gene signature in co-cultured fibroblasts. scRNA-seq of lung tissue from three HPS1 patients similarly shows the presence of inflammatory fibroblasts and increased IL1R1 expression on fibroblasts. These data posit complex interactions between dysfunctional epithelial cells, inflammatory fibroblasts, and recruited immune cells, suggesting potential opportunities for mitigation of the fibrotic cascade.

Introduction

Pulmonary fibrosis (PF) is the end stage of numerous types of interstitial lung diseases (ILD). Alveolar type II (AT2) cell dysfunction often leads to the recruitment of circulating immune cells, which expand and differentiate within the lungs^1,2,3. Together, injured epithelia and immune activation subsequently lead to fibroblast proliferation and activation, generation of fibroblast foci, collagen deposition, and pathologic remodeling of the extracellular matrix^1,4. These processes result in the thickening of alveolar septa, poor gas exchange, and eventual respiratory failure^5,6. However, the mechanisms by which cellular injury leads to fibrotic matrix deposition, and the cellular interactions responsible for driving pathologic remodeling over normal lung repair processes, remain poorly understood.

Occurring in an estimated one per million individuals⁷, Hermansky-Pudlak syndrome (HPS) is a group of 11 autosomal recessive disorders caused by mutations in ubiquitously expressed proteins⁸, including biogenesis of lysosome-related organelles complex (BLOC)-1 (HPS7, HPS8, HPS9, and HPS11), BLOC-2 (HPS3, HPS5, and HPS6), BLOC-3 (HPS1, HPS4), and adaptor protein 3 (AP-3; HPS2, HPS10)⁹. While all HPS patients exhibit oculocutaneous albinism and bleeding diathesis, HPS1, HPS2, and HPS4 patients also suffer from early onset of progressive fibrosing ILD, which is the leading cause of mortality in these patients⁵. HPS1 is the most common subtype of HPS and the genotype most associated with HPS pulmonary fibrosis (HPS-PF)^10,11.

Mouse models of HPS1 and HPS2 have increased susceptibility to pro-fibrotic stimuli, and double mutant HPS1-2 mice, which are homozygous for both HPS1 and HPS2 mutations, develop spontaneous PF at approximately one year of age^12,13. In HPS2 mice, pulmonary epithelial cell correction of AP-3 reduced fibrotic susceptibility following bleomycin injury^12,14. Exaggerated chemokine signaling from HPS2 AT2 cells was shown to increase recruitment and activation of CCR2⁺ monocytes, and inhibition of this signaling axis protected mice from fibrotic lung injury¹⁴. However, it remains unclear how the fibroblast populations become activated in HPS mice following injury, and whether the loss of AP-3 in HPS2 mice, or of BLOC-3 in HPS1 mice, directly affects fibroblast function. The objectives of this study were to characterize the cellular populations in the lungs of HPS mouse models and individuals with HPS, and to determine the cellular interactions contributing to pathologic fibroblast function and pulmonary fibrosis.

Results

Single cell RNAseq analysis of lungs from young and old HPS mouse models

We performed single cell RNA sequencing (scRNA-seq) to investigate the cellular and transcriptional factors associated with HPS-PF in clinically relevant mouse models (Fig. 1a). Samples from 8- and 60-week-old HPS mouse models were compared to age-matched C57Bl/6J wildtype (WT) controls. Double-mutant HPS1-2 mice have been shown to develop a spontaneous fibrotic phenotype with aging^13,15, and we observed increased trichrome blue staining and total lung collagen content (Fig. 1b, c). Single mutant HPS1 and HPS2 mice, which exhibit increased fibrotic susceptibility¹², were investigated to study the HPS subtypes found in patients. We included lungs from HPS2 transgenic mice (HPS2-TG) mice that express a correction of the HPS2 gene product (Ap3b1) in epithelial cells and have been shown to have reduced susceptibility to fibrotic injury compared to HPS2 mice¹². Lastly, the HPS3 mouse model served as a disease control, as pulmonary fibrosis has not been reported in individuals with HPS3, and the fibrotic response to lung injury in HPS3 mice has more closely resembled WT mice than that of HPS1 or HPS2 mice¹².

Fig. 1: Utilizing scRNA-seq to identify mechanisms of age-related fibrosis in HPS mice.

a Overview of experimental mouse models used in this study and a schematic of scRNA-seq experimental design. All HPS models are global knockouts or loss of function mutations, except HPS2-TG, which denotes transgenic HPS2 mice with lung epithelial-specific correction. b Masson’s Trichrome staining of microtome cut lung sections of 60-week-old WT and HPS1-2 mice. Scale bar = 200 µm. Imaging experiments were performed three times independently. Arrows indicate regions of increased collagen deposition. c Total lung collagen content from the lungs of naïve HPS mice, ages 8- and 60 weeks old. Data are presented as mean ± SEM and were obtained from n = 13 each for 8wk WT and 8wk HPS2 mice, n = 4 each for 8wk HPS1-2, 60wk WT, 60wk HPS1-2, 60wk HPS1, and 60wk HPS2 mice, and n = 3 each for 8wk HPS1, 8wk HPS2-TG, 8wk HPS3, 60wk HPS2-TG, and 60wk HPS3 mice. P-values were determined by two-way ANOVA followed by Dunnett’s post hoc analysis. N.S. = not significant. scRNA-seq. d UMAP of all cells recovered from scRNA-seq experiments.

Across all ages and disease subtypes, we recovered 168,037 cells from four major cell compartments (mesenchymal cells; epithelial cells; immune cells; and endothelial cells) (Fig. 1d). Cell types were identified based on the expression of known marker genes (Supplementary Fig. 1). All cell types identified were represented across all ages and HPS subtypes.

Age-dependent changes in AT2 cells from HPS mice

Within the epithelial cell compartment, we identified five major cell types: alveolar type I (AT1) cells, AT2 cells, goblet cells, club cells, and ciliated cells (Fig. 2a). Due to the crucial role that AT2 cells play in fibrotic susceptibility in HPS mouse models, we focused on the expression profile of HPS AT2 cells to determine the cell-specific factors that may predispose HPS lungs to fibrotic injury.

Fig. 2: Alveolar epithelial cell phenotypes in HPS mouse models with aging.

a UMAP of epithelial compartment following isolation and re-projection in 8- and 60-week-old HPS mice. b Split UMAP plots of epithelial cell compartments in WT, HPS1, HPS2, and HPS1-2 mice, separated by age and genotype. Arrows indicate the reduction of alveolar type II epithelial cells (AT2 cells) recovered in scRNA-seq experiments. c Immunostaining of DC-LAMP⁺ AT2 cells in 60-week-old WT, HPS1-2, HPS1, and HPS2 mice from precision cut lung sections (PCLS). Scale bar = 1 mm. aw = airways. d Quantification of DC-LAMP+ cells as a percentage of total DAPI⁺ cells. Data are presented as mean ± SEM and were obtained from n = 3 8wk WT, 8wk HPS1-2, 8wk HPS2, 60wk HPS1, and 60wk HPS2 mice, and n = 4 8wk HPS1, 60wk WT, and 60wk HPS1-2 mice. For each biological replicate, a minimum of three images were taken of lung parenchyma, and the percentage of DC-LAMP⁺ cells over total DAPI⁺ cells was calculated as the average of those images. P-values were obtained by two-way ANOVA with Dunnett’s post hoc analysis. N.S. = not significant. e Dot plot of Sftpc expression split by age and genotype. Arrow indicates reduced expression of Sftpc in 60wk HPS1-2 AT2 cells. f Dot plot of expression of transitional AT2 cell marker genes split by age and genotype. Arrows indicate increased expression of transitional AT2 cell marker genes in 8- and 60wk HPS1-2 AT2 cells. g Immunostaining of DC-LAMP⁺/KRT17⁺/KRT8⁺ aberrant transitional AT2 cells in 50 µm thick PCLS sections of 60-week-old HPS mouse models. In HPS1-2, regions of AT2 cell hyperplasia were intentionally imaged. Arrows indicate DC-LAMP⁺ cells that co-express KRT17 and KRT8. Top 20 differentially expressed genes (DEGs) between (h) 60-week-old WT and HPS1-2 AT2 cells, (i) 60-week-old WT and HPS1 AT2 cells, and (j) 60-week-old WT and HPS2 AT2 cells. k Over-representation analysis of biological processes that are upregulated in 60-week-old HPS1-2 AT2 cells compared to age-matched WT controls. l Over-representation analysis of biological processes that are downregulated in 60-week-old HPS1-2 AT2 cells compared to age-matched WT controls.

There was a reduction in the proportion of AT2 cells recovered in HPS1-2 mice by scRNA-seq, consistent with previous reports¹⁶ (Fig. 2b). Immunostaining and quantification showed a significant reduction in AT2 cells in both 8- and 60-week-old HPS1-2 mice compared to age-matched controls, consistent with previous reports¹⁶ (Fig. 2c, d). Within the AT2 cells that were recovered by scRNA-seq, there was a reduction in Sftpc gene expression in 60-week-old HPS1-2 AT2 cells compared to age-matched WT controls (Fig. 2e). HPS1-2 AT2 cells also showed increased expression of genes associated with transitional AT2 cells^17,18, such as Krt8, Krt18, Krt19, Lgals3, Cldn4, and Sfn (Fig. 2f). Immunostaining showed increased co-expression of DC-LAMP, KRT8 and KRT17, markers of pathologic transitional AT2s¹⁹, in HPS AT2 cells (Fig. 2g). Together, these results show a significant reduction in AT2 cell numbers over time with emergence of increased numbers of pathological transitional KRT8⁺/KRT17⁺ AT2 cells in the lungs of HPS mice.

Among the top differentially expressed genes, Ccnd1 was upregulated in HPS1-2 AT2 cells compared to age-matched WT controls, as well as in both HPS1 and HPS2 AT2 cells (Fig. 2h–j). The Ccnd1 protein product, cyclin D1, is crucial for mitotic progression during proliferation, but over-expression has been associated with cellular senescence^20,21,22. Over-representation analysis (ORA) of differentially expressed genes showed upregulation of biological processes associated with leukocyte activation and inflammation in 60-week-old HPS1-2 mice, as well as both HPS1 and HPS2 single mutant models, suggesting a common function across all aged HPS AT2 cells in pro-fibrotic subtypes (Fig. 2k, l, and Supplementary Fig. 2a, b). AT2 cells from HPS1-2 mice also showed increased expression of genes associated with transforming growth factor beta (TGF-β) activation, and decreased expression of genes involved in lipid biosynthetic and metabolic processes. Markers of AT2 cell differentiation, transitional AT2 cell, and p53 signaling were consistent with previous results from bulk RNAseq of AT2 cells¹⁶ (Supplementary Fig. 2c).

To investigate the factors which may drive fibrotic lung injury in aged double mutant HPS1-2 mice compared to single mutant mice, we also evaluated DEGs and performed ORA comparing 60-week-old HPS1-2 AT2 cells with expression from 60-week-old single mutant HPS mice. HPS1-2 AT2 cells showed an upregulation in genes associated with cell migration compared to AT2 cells from HPS1 mice (Supplementary Fig. 3a, b). Similarly, HPS1-2 AT2 cells showed an upregulation in genes associated with p53-mediated apoptotic signaling compared to HPS2 AT2 cells (Supplementary Fig. 3c, d). Collectively, these results suggest an additive impact in the HPS1-2 model.

Increased total immune cells and interstitial macrophages in HPS mice

From the myeloid-lineage immune cells, we focused our analyses on five major cell types: alveolar macrophages, interstitial macrophages, dendritic cells, inflammatory monocytes, and classical monocytes (Fig. 3a). ORA of differentially expressed genes between the myeloid cell compartments of 60-week-old WT and HPS1-2 mice showed upregulation of genes involved in collagen catabolic processes (Fig. 3b), and downregulation of genes associated with Fc-receptor signaling (Fig. 3c). A number of genes associated with p53 intrinsic signaling were downregulated in HPS2 myeloid cells yet upregulated in HPS1-2 myeloid cells, similar to our observations in AT2 cells (Supplementary Figs. 4, 5). Previous studies have noted changes in lymphocytes recovered from BAL fluid in HPS mouse models. Defects in select lymphoid cells have been proposed to impair cytotoxic function^13,23, and a recent publication has noted the potential for type 2 innate lymphoid cells to drive fibrosis in HPS1 mouse models²⁴. From these scRNA-seq data, we did not observe significant shifts in any lymphoid cell proportions.

Fig. 3: Increased myeloid immune cells and interstitial macrophage populations in aged HPS mice.

a UMAP of myeloid lineage immune compartment following isolation and re-projection in 8- and 60-week-old HPS mice, focused on macrophages, dendritic cells, and monocytes. b Over representation analysis of biological processes that are upregulated in myeloid cells from 60-week-old HPS1-2 mice compared to age matched WT controls. c Over representation analysis of biological processes that are downregulated in myeloid cells from 60-week-old HPS1-2 compared to age matched WT controls. d Representative immunostaining of CD68⁺ cells in HPS mice, by age and genotype. Imaging experiments were performed on 50 µm thick PCLS from at least three independent biological replicates per age and genotype. Arrows indicate regions of increased CD68⁺ cell clustering. Scale bar = 100 µm. e Percentage of CD68⁺ cells per DAPI⁺ cells. Data are presented as mean ± SEM and were obtained from n = 3 8wk HPS2, 8wk HPS2-TG, 60wk WT, 60wk HPS1, 60wk HPS2, and 60wk HPS2-TG mice, and n = 4 8wk WT, 8wk HPS1-2, 8wk HPS1, and 60wk HPS1-2 mice. For each biological replicate, a minimum of three images were taken of lung parenchyma, and the percentage of CD68⁺ cells was calculated as the average of those images. P values were obtained by two-way ANOVA with Dunnett’s post hoc analysis. f Proportion of interstitial macrophages (MΦ) (CD45⁺CD11b⁺MHCII⁺CD64⁺CD24^-) in 8-week-old mice by flow cytometry, as a percentage of total CD45⁺ cells. Data are presented as mean ± SEM and were obtained from n = 3 each for HPS1 and HPS2 mice, n = 6 HPS1-2 mice, and n = 7 WT mice. g Proportion of interstitial macrophages (MΦ) in 60-week-old mice by flow cytometry, as a percentage of total CD45⁺ cells. Data are presented as mean ± SEM and were obtained from n = 3 each for WT, HPS1, and HPS2 mice, and n = 4 HPS1-2 mice. P values were obtained by one-way ANOVA with Dunnett’s post hoc analysis. N.S. = not significant.

By immunofluorescence imaging, we observed a significant increase in the percentage of CD68⁺ cells in the lungs of 60-week-old HPS1-2 and HPS2 mice compared to age-matched WT controls (Fig. 3d, e). Notably, there was no significant difference in the percentage of CD68⁺ cells between 60-week HPS2-TG and WT mice, suggesting that correction of HPS2 in the lung epithelia is sufficient to prevent the increase in CD68⁺ cells observed in the lungs of HPS2 mice.

We next investigated whether the increase in the CD68⁺ cell population in aged HPS1-2 lungs was associated with changes in the abundance of different immune cell types. Flow cytometry analysis showed a significant increase in the proportions of interstitial macrophages from the lungs of 60-week-old HPS1-2 and HPS1 mice compared to age-matched WT controls (Fig. 3f-g, Supplementary Fig. 6a – c). Immunostaining for CD11b/CX3CR1 double-positive cells in HPS1-2 mice confirmed the presence of interstitial macrophages (Supplementary Fig. 6d). Although HPS2 mice had an increase in CD68⁺ cells, the relative composition of myeloid cell types was similar to WT.

Previous studies have identified two distinct interstitial macrophage subclusters which can be identified by specific markers, including major histocompatibility complex II (MHCII)^25,26,27. Further analysis of our interstitial macrophage population revealed no statistically significant differences in the proportion of MHCII^hi vs. MHCII^lo macrophages in any HPS genotype at any age (Supplementary Fig. 6e, f). Taken together, these data show significant increases in the number of myeloid cells in HPS1-2 and HPS2 mice, with a shift towards an increased proportion of interstitial macrophages in aged HPS1-2 and HPS1 mice.

Age-dependent inflammatory fibroblast emergence in HPS

Within the mesenchymal cell compartment, we identified seven major cell types: airway smooth muscle cells (ASM), ductal myofibroblasts, mesothelium, adventitial fibroblasts (MFAP5-PDGFRα), pericytes, vascular smooth muscle cells (VSM), and alveolar fibroblasts (WNT2-PDGFRα) (Fig. 4a). Despite the increased collagen content in the lungs of 60-week-old HPS1-2 mice, we did not observe increased expression of genes associated with fibroblast activation within alveolar fibroblast populations from HPS1-2 mice (Supplementary Fig. 7a). Immunostaining of activated fibroblast markers, such as α-SMA, SM22α, or CTHRC1²⁸, was similar in 60-week-old HPS mice compared to age matched WT controls (Supplementary Fig. 7b). We performed DEG and ORA on alveolar fibroblasts from HPS1-2, HPS1, and HPS2, and still did not observe any gene ontology (GO) biological processes that were associated with collagen deposition or extracellular matrix remodeling (Supplementary Fig. 8). HPS1-2 and HPS1 fibroblasts showed an upregulation of genes associated with inflammatory processes (Supplementary Fig. 8a–d). HPS2 fibroblasts showed a downregulation of genes associated with collagen-signaling and upregulation of collagen metabolism (Supplementary Fig. 8e, f). Comparing HPS1-2 with HPS1 and HPS2 alveolar fibroblasts, we observed an upregulation of genes associated extracellular matrix catabolism and immune cell migration in HPS1-2 compared to HPS1, and an increase in collagen-associated signaling in HPS1-2 fibroblasts compared to HPS2 (Supplementary Fig. 9).

Fig. 4: Age-dependent enrichment of inflammatory lung fibroblasts in HPS models associated with fibrosis.

a UMAP of mesenchymal compartment following isolation and re-projection in 8- and 60-week-old HPS mice. b UMAPs of select inflammatory fibroblast marker genes within alveolar fibroblasts used to identify inflammatory fibroblasts. c Proportion of alveolar fibroblasts that were identified as inflammatory fibroblasts, separated by age and genotype. d Dot plot of expression of inflammatory fibroblast marker genes in alveolar fibroblasts split by age and genotype. Arrows indicate genotypes with increased expression of inflammatory fibroblast-associated genes. e Split UMAP of inflammatory fibroblast subpopulations in alveolar fibroblasts across age and genotype. f Representative immunofluorescence images from 50 µm thick PCLS showing the enrichment of SAA3⁺ cells in lungs of HPS1, HPS2, and HPS1-2 mice compared to age-matched WT controls, HPS3 mice, and HPS2-TG mice. Imaging experiments were performed on three independent biological replicates per age and genotype. Arrows indicate regions of increased SAA3⁺ expression. Scale bar = 100 µm.

Within the scRNA-seq dataset, we noted a subset of alveolar fibroblasts that were enriched in HPS mice. Differential gene expression analysis of these cells compared to other alveolar fibroblasts revealed increased expression of inflammatory genes, including Saa3, Hp, Lcn2, as well as other genes that are downstream of inflammatory signaling, such as Il6, Tnfaip3, and Ccl2 (Fig. 4b). This gene signature is similar to the transitional/inflammatory fibroblasts recently described by others^29,30,31. We observed an increased proportion of alveolar fibroblasts that exhibited this gene signature in HPS1-2 mice at both 8- and 60 weeks of age compared to age-matched WT controls, as well as in alveolar fibroblasts from 60-week-old HPS1 and HPS2 mice (Fig. 4c). Fibroblasts from 60-week-old HPS1-2 mice showed the strongest increase in expression of these inflammatory genes, while a subset of these genes, specifically Saa3, Hp, Tnfaip3, and Il6, were detectable in fibroblasts from the lungs of 8-week-old HPS1-2 mice (Fig. 4d, e). There was minimal or no expression of these inflammatory markers in alveolar fibroblasts from HPS2-TG and HPS3 mice at either 8- or 60 weeks of age.

Consistent with transcriptomic data, SAA3⁺ staining was prominent in both 8- and 60-week-old HPS1-2 mice compared to age-matched WT controls (Fig. 4f). In examining lung sections from single mutant HPS1 and HPS2 mice, minimal SAA3⁺ expression was present in the samples from younger mice, with marked increases in SAA3⁺ cells evident in lungs from 60-week-old mice. SAA3 expression was minimal in the lungs of HPS2-TG and HPS3 mice at all ages. Immunostaining with PDGFRα confirmed that the majority of SAA3⁺ staining co-localized with PDGFRα expressing fibroblasts (Supplementary Fig. 10a).

In vivo, SAA3⁺ cells appear to be distributed throughout the lungs. In aged HPS1-2 mice, immunostaining shows these cells are located in regions of both AT2 cell depletion and AT2 cell retention, as well as regions both with and without CD68⁺ cell accumulation (Supplementary Fig. 10b). While these cells can be located around fibrotic regions in HPS1-2 mice, they can also be observed in regions absent of collagen deposition (Supplementary Fig. 10c). Importantly, within regions of collagen accumulation, SAA3⁺ fibroblasts appear distinct from collagen type I expressing cells in the lung, consistent with previous reports²⁹ (Supplementary Fig. 10d). Taken together, the genotype-specific presence of inflammatory fibroblasts under naïve conditions prior to onset of fibrosis supports the conclusion that this fibroblast cell state is relevant for pulmonary fibrosis in HPS mouse models.

Increased expression of IL-1R1 in HPS1 fibroblasts

Based on the increased prevalence of inflammatory SAA3⁺ fibroblasts in aged HPS lungs, we evaluated inflammatory cytokine gene expression in other cell types to investigate the source of inflammation. We observed an increase in Il1b expression in interstitial macrophages from 60-week-old HPS1-2 and HPS1 mice compared to age-matched WT controls (Supplementary Fig. 11a). We did not detect any significant increases in inflammatory cytokine gene expression in HPS AT2 cells (Supplementary Fig. 11b). Given this increase in Il1b gene expression in interstitial macrophages, and the role of interleukin-1 signaling in inflammatory cell activation, we evaluated both IL-1β and IL-1α concentrations in whole lung homogenate. Both IL-1β and IL-1α were significantly increased in the lung tissue homogenate of 60-week HPS1-2 mice, but not in 8-week-old HPS1-2 mice or in other HPS mice regardless of age (Supplementary Fig. 11c, d). We did not detect significant differences in IL-1β concentrations in bronchoalveolar lavage fluid (BALF) from any of the HPS models, nor did we detect significant differences in IL-1β from BALF-derived HPS macrophages cultured in vitro (Supplementary Fig. 11e, f).

To further evaluate potential macrophage-fibroblast interactions, we isolated and cultured PDGFRα⁺ fibroblasts and macrophages from 60-week-old WT and HPS1-2 mice. After 24 h in culture, cell culture supernatant was collected from WT and HPS1-2 macrophages and used to stimulate the cultured fibroblasts for an additional 24 h (Supplementary Fig. 11g). Consistent with our scRNA-seq data, HPS1-2 fibroblasts displayed increased expression of Saa3 and Tnfaip3 under baseline conditions and exhibited an increase in both Saa3 and Tnfaip3 gene expression following stimulation from either WT or HPS1-2 macrophage supernatant (Supplementary Fig. 11h, i). In contrast, there was no significant difference in Saa3 or Tnfaip3 expression in WT fibroblasts exposed to cell culture supernatant from either WT or HPS1-2 macrophages, suggesting that under these conditions, secreted factors from HPS1-2 macrophages were insufficient to induce the inflammatory fibroblast cell state in WT fibroblasts.

We then examined the expression of inflammatory cytokine receptors, such as IL-1-, TNF-, and IL-6 receptors in HPS fibroblasts. We noted an increase in Il1r1 gene expression in 60-week-old HPS1-2 fibroblasts (Fig. 5a). There was also an increase in Il1r1 expression in 60-week-old HPS1, but not HPS2, fibroblasts. Immunocytochemistry imaging of 60-week-old primary murine lung fibroblasts showed greater IL-1R1 expression in HPS1-2 and HPS1 fibroblasts (Fig. 5b), and flow cytometry analysis of 60-week-old PDGFRα⁺ fibroblasts indicated a significant increase in cell surface expression of IL-1R1 on HPS1-2 fibroblasts (Fig. 5c, Supplementary Fig. 12a, b). In vivo, IL-1R1⁺ cells appeared to be distributed throughout the lung in both WT and HPS mice, although there was greater IL-1R1 expression in 60-week-old HPS1-2 and HPS1 mice compared to WT (Fig. 5d). While we did observe IL-1R1⁺ fibroblasts located in close proximity to AT2 cells in HPS1-2 mice, we also observed IL-1R1 expression in AT2 cell depleted regions of the lung (Supplementary Fig. 12c). Therefore, it is unclear whether proximity to AT2 cells is important for IL-1R1 expression in this context.

Fig. 5: Genotype-specific mechanisms associated with inflammatory fibroblast cell state in HPS mice.

a Dot plot of inflammatory receptor gene expression in 60-week-old alveolar fibroblasts split by genotype. Arrow indicates Il1r1, which shows increased expression in 60-week-old HPS1-2 and HPS1 fibroblasts relative to age-matched WT controls. b Representative immunocytochemistry images of IL-1R1 and phalloidin in primary PDGFRα⁺ fibroblasts from 60-week-old mice. Imaging experiments were performed on three independent biological replicates per genotype. Regions within the white square are magnified in the image inset. Scale bars = 50 µm (wide) or 10 µm (inset). c Cell surface expression of IL-1R1 on PDGFRα⁺ fibroblasts from 60-week-old mice determined by flow cytometry. Data are presented as the cumulative cell surface expression of IL-1R1 from four WT mice, five HPS1 mice, three HPS2 mice, and five HPS1-2 mice. Mean ± SEM are presented in the upper right corner of each plot. d Representative immunostaining from 50 µm PCLS showing increased expression of IL-1R1 in 60-week-old HPS1-2 and HPS1 mice compared to age-matched WT controls. Scale bar = 100 µm. e Relative gene expression of IL1R1 in MRC5 fibroblasts following 48 h knockdown experiment. Data are presented as mean ± SEM and were obtained from n = 6 replicates for Scramble control and HPS1 knockdown (denoted siHPS1), and n = 4 replicates for AP3B1 knockdown (denoted siHPS2). P-values were determined by one-way ANOVA followed by Sidak’s multiple comparisons post-hoc analysis. f Experimental outline for (g–i). Relative gene expression of (g) Saa3, (h) Ccl2, and (i) Il6 in WT fibroblasts cultured for 21 days with WT, HPS1-2, HPS1, HPS2, or HPS2-TG AT2 cells in distal lung organoid model. Data are presented as mean ± SEM and were obtained from n = 15 organoids containing WT AT2 cells, n = 3 organoids containing HPS1-2 and HPS1 AT2 cells, and n = 4 organoids containing HPS2 and HPS2-TG AT2 cells. For Ccl2 gene expression, n = 3 organoids containing HPS2-TG AT2 cells. For Il6 gene expression, n = 14 organoids containing WT AT2 cells. P values were determined by one-way ANOVA followed by Sidak's post-hoc analysis. N.S. = not significant. Figure (f) created with Biorender.com.

To test whether loss of HPS1 gene expression directly impacts IL1R1 expression, we performed siRNA knockdown experiments in the MRC5 human lung fibroblast cell line with siRNAs to target both HPS1 and AP3B1 (Supplementary Fig. 12d). Transfection with HPS1 or AP3B1 siRNA led to a ~ 70% decrease in gene expression compared to scramble controls, respectively (Supplementary Fig. 12e, f). In line with previous studies³², knockdown of HPS1 led to an increase in MYH10 expression (Supplementary Fig. 12g). MRC5 fibroblasts transfected with HPS1 siRNA, but not AP3B1 siRNA, showed a significant increase in IL1R1 gene expression compared to scramble control, suggesting that the reduction of HPS1 is sufficient to drive increased IL1R1 expression under these conditions (Fig. 5e). To evaluate if these changes could also be observed in extrapulmonary fibroblasts, we repeated this experiment with primary human dermal fibroblasts (Supplementary Fig. 12h). While we were able to achieve a significant knockdown of both HPS1 and AP3B1, we did not observe any significant increases in IL1R1 expression, suggesting cell type specificity of these findings (Supplementary Fig. 12i–k).

Taken together, these results suggest that an increase in IL-1R1 expression on HPS1-2 and HPS1 fibroblasts is linked to HPS1 expression and is associated with the altered inflammatory fibroblast phenotype in HPS. However, IL-1R1 expression was normal in HPS2 fibroblasts, suggesting an alternative mechanism may be responsible for emergence of inflammatory fibroblasts in HPS2.

AT2 cells from HPS2 mice induce inflammatory fibroblast gene expression in vitro

AT2 cells have been shown to be a key initiator of fibrotic susceptibility in HPS mice¹². To investigate AT2 cell-fibroblast interactions in HPS, we isolated these two cell types and co-cultured them in distal alveolar organoid models (Supplementary Fig. 13a). Organoid co-cultures containing HPS1-2 AT2 cells and HPS1-2 fibroblasts showed significant reductions in both colony forming efficiency (CFE) and size compared to WT organoids (Supplementary Fig. 13b–d). Furthermore, WT AT2 cells co-cultured with HPS1-2 fibroblasts also showed a significant reduction in CFE, suggesting that HPS1-2 fibroblasts fail to support AT2 cell survival and organoid formation.

We next evaluated the AT2-fibroblast interactions in the single-mutant HPS models. Organoid formation was unaltered in the presence of HPS1 AT2 cells or fibroblasts. In contrast, HPS2 fibroblasts failed to support WT AT2 cell CFE and size in the organoid system, similar to results with HPS1-2 fibroblasts. Colony-forming efficiency and average organoid size were restored when WT AT2 cells were cultured with HPS2-TG fibroblasts.

Transcriptomic analyses and immunostaining results suggested that correction of AP-3 in HPS2-TG epithelial cells is sufficient to reduce the expression of SAA3⁺ fibroblasts, as demonstrated by the lack of inflammatory fibroblasts in the lungs of HPS2-TG mice (Fig. 4d). Therefore, we proceeded to directly test the impact of HPS AT2 cells on fibroblast gene expression profiles in the organoid system. WT fibroblasts cultured with HPS1-2 and HPS2 AT2 cells exhibited increases in Saa3, Ccl2, and Il6 expression in comparison to WT fibroblasts cultured with WT AT2 cells (Fig. 5f–i). AT2 cells from HPS2-TG mice did not induce increased inflammatory gene expression in WT fibroblasts. In testing AT2 cells from HPS1 mice in the organoid system, WT fibroblasts inflammatory gene expression was not altered. Il1r1 expression was not significantly increased in WT fibroblasts following co-culture with HPS AT2 cells, suggesting that HPS AT2 cells do not induce expression of IL-1R1 in fibroblasts in this model system (Supplementary Fig. 13e).

Taken together, these results demonstrate the epithelial dependence of inflammatory fibroblast phenotypes in HPS2 mice, with an HPS2 fibroblast phenotype that also fails to support organoid growth. However, our data suggest differences in HPS1 fibroblast phenotypes without demonstration of altered HPS1 epithelial-fibroblast interactions in the organoid system.

HPS1 patient lung tissue shows a reduction in alveolar epithelial cells with increased presence of both interstitial macrophages and inflammatory fibroblasts

To understand the relevance of the observations made in HPS mouse models, we performed scRNA-seq on lung tissues obtained at the time of lung transplant from three different individuals with HPS1. Previously published scRNA-seq datasets from control and interstitial lung disease (ILD) patients were integrated with these HPS1 patient datasets to evaluate common gene expression patterns in both HPS1 disease and other ILDs^28,33,34 (Fig. 6a, b).

Fig. 6: Human HPS1 fibrotic lung tissue demonstrates loss of AT2 cells and increased prevalence of inflammatory fibroblasts.

a UMAP of all cells, grouped by compartment, from HPS1 patient scRNA-seq dataset merged with publicly available control and ILD patient scRNA-seq datasets^28,33,34. b UMAP of cells recovered and number of cells recovered from three separate HPS1 patient lung tissues (blue) overlaid on control tissues and ILD atlas. c Cell type proportions within the epithelial compartment of human scRNA-seq atlas, separated by disease classification. NSIP = nonspecific interstitial pneumonia, Chronic HP = chronic hypersensitivity pneumonitis. d UMAP of select inflammatory fibroblast marker genes (top panel) and fibrotic fibroblast marker genes (bottom panel) within PDGFRα⁺ fibroblasts used to identify each fibroblast subtype, respectively. e Proportion of PDGFRα⁺ fibroblasts that were classified as either normal, inflammatory, fibrotic, or both inflammatory and fibrotic in human scRNA-seq atlas by disease type. f Split UMAPs of control, HPS1, and IPF PDGFRα⁺ fibroblasts by identified fibroblast subtype. g Hematoxylin & eosin and immunofluorescent staining of tissue from control and HPS1 patient tissue. Sequential sections were used to localize SFRP2⁺ and CTHRC1⁺ fibroblasts in control and HPS1 patient tissue. Imaging experiment was performed on one HPS1 patient tissue and one control patient tissue. Arrows indicate similar regions across sequential tissue sections, which highlight the overlap of SFRP2⁺ cells with CTHRC1⁺ cells in fibrotic regions, as determined by H&E stain. Scale bars = 100 µm. h Dot plot of top 20 DEGs between control tissue and HPS1 patient PDGFRα⁺ fibroblasts. Gene expression of IL1R1 is bolded and highlighted with an arrow.

Within the epithelial cell compartment, the HPS1 patient tissue showed the lowest proportion of AT2 cells (Fig. 6c). Immunofluorescence imaging of HPS1 patient tissue confirmed sparse clusters of HTII-280⁺ cells scattered throughout the lung parenchyma (Supplementary Fig. 14a). DEG analysis of HPS1 patient AT2 cells showed upregulation of both KRT17 and KRT8 compared to control tissue, suggesting that these cells share similar characteristics to pathological transitional AT2 cells (Supplementary Fig. 14b), which we confirmed by immunostaining for KRT8 and KRT17 in HTII-280⁺ AT2 cells (Supplementary Fig. 14c). Similar to HPS1-2 mouse models, ORA of HPS1 patient AT2 cells showed increased expression of genes associated with the production of TGF-β, and downregulation of processes involved in mitochondrial function and aerobic cellular respiration (Supplementary Fig. 14d). Differing from findings in the murine models, CCND1 expression in HPS1 patient AT2 cells was similar to levels in lung samples from healthy controls (Supplementary Fig. 14e).

Within the myeloid cell compartment, HPS1 patient samples had an increased proportion of interstitial macrophages compared to healthy controls, which was also observed in other ILD etiologies (Supplementary Fig. 15a, b). SPP1, a pathologic macrophage marker described in idiopathic pulmonary fibrosis (IPF)³⁵, was upregulated in HPS1 patient myeloid cells (Supplementary Fig. 15c). ORA of differentially expressed genes between control and HPS1 patient myeloid lineage immune cells showed inhibition of Fc-receptor signaling and downregulation of biological processes involving mitochondrial respiration (Supplementary Fig. 15d). We observed an increased expression of IL1B in HPS1 patient-derived interstitial macrophages, as in other ILD etiologies (Supplementary Fig. 15e).

Unlike the aged HPS mouse models, PDGFRα⁺ fibroblasts from HPS1 patients showed an upregulation of hallmark fibrotic fibroblast genes, including CTHRC1, POSTN, SPARC, and numerous collagen- and extracellular matrix-associated genes (Supplementary Fig. 16a). Because SAA3P is a pseudogene in humans³⁶, we utilized gene markers previously used to describe inflammatory fibroblasts in human tissue samples²⁹. Using this gene set, we observed a large proportion of inflammatory fibroblasts in the HPS1 patient samples, as well as a subset of cells that expressed genes characteristic of both of inflammatory and fibrotic subtypes (Fig. 6d – f, Supplementary 16b). To evaluate the localization of both inflammatory and fibrotic fibroblasts within HPS1 patient tissue samples, we performed immunostaining on sequential sections. In contrast to previous studies which appeared to show inflammatory fibroblasts near but distinct from fibrotic fibroblasts²⁹, our data show SFRP2⁺ inflammatory fibroblasts alongside CTHRC1⁺ pathologic fibroblasts within diseased regions (Fig. 6g).

Notably, HPS1 fibroblasts showed increased IL1R1 expression (Fig. 6h). While IL1R1 receptor expression was also increased in IPF, nonspecific interstitial pneumonia (NSIP), and unclassifiable ILD, expression was most pronounced in HPS1 patients (Supplementary Fig. 16c). Immunostaining of HPS patient tissue showed a notable increase in expression of IL-1R1 compared to control (Supplementary Fig. 16d). Together, this data suggests that HPS mouse models recapitulate human HPS-PF in several key features, as both models and disease show a reduction in AT2 cells, expansion of interstitial macrophages, and a high proportion of inflammatory fibroblasts that could be driven through increased expression of IL-1R1 within the fibroblast population.

Discussion

Hermansky-Pudlak syndrome is among the most penetrant genetic etiologies of pulmonary fibrosis, and the recognizable clinical features provide opportunity for early detection and treatment⁵. The scRNA-seq dataset generated in this study allows for in-depth analysis of the genotype-specific, cell-type-specific, and age-dependent transcriptomic changes, with comparisons between HPS murine models and human HPS disease (Fig. 7).

Fig. 7: Models of inflammatory fibroblast emergence in HPS mouse models.

Based on the results from this study, we have identified an enrichment of inflammatory fibroblasts in 60-week-old HPS1-2 mice, that appears to coincide with numerous other observations, including increased expression of IL-1R1, increased AT2-signaling, and increases in myeloid cell accumulation, including in the proportion of interstitial macrophages. Using single mutant HPS mouse models to deconvolute the causes of these phenotypes, our results suggest that the mechanisms responsible for the emergence of these inflammatory fibroblasts are specific to the different HPS subtypes. In HPS1 mice, in addition to the observed increase in interstitial macrophages, distal lung fibroblasts have a significant increase in cell surface expression of IL-1R1, which could lead to an exaggerated response to IL-1β stimulation and a resultant hyperresponsive inflammatory response. HPS2 mice have increased total myeloid cell accumulation within the lungs, and an unknown AT2-derived factor appears to induce the inflammatory fibroblast phenotype, with a reduction of inflammatory fibroblasts occurring with epithelial-cell selective correction of Ap3b1 in the HPS2-TG model. HPS1-2 mice appear to possess features of both HPS1 (i.e., increased cell surface expression of IL-1R1) and HPS2 (i.e., AT2-derived inflammatory stimulation) mouse models and likely explains the increased severity of lung fibrosis observed in this model. Created with Biorender.com.

Based on integration with data from previously published ILD scRNA-seq atlases³³, our findings demonstrate that HPS-PF shares many key features with other forms of progressive fibrosing ILDs including IPF. Historically, therapies tested and approved for use in IPF have not been made available to individuals with HPS-PF, attributable in part to uncertainty about common pathobiology. Because of bleeding risk and use of genetic testing, lung biopsy is typically not performed for diagnosis of HPS-PF, and thus limited human lung tissue has been available for study. Only 13 lung transplants for HPS were performed in the past five years (personal communication, D. Appell, HPS Network, April 2024), and therefore the access to three HPS lung explants provided a unique scientific opportunity. Within our HPS1 patient scRNA-seq dataset, we found that many of the distinct cell types and markers of cell activation that have been identified through prior fibrotic ILD atlases, including SPP1⁺ macrophages³⁵, CTHRC1⁺ fibrotic fibroblasts²⁸, and KRT8⁺/KRT17⁺ aberrant transitional cells^17,19,37,38. HPS-PF tissue shared other key similarities with fibrotic ILD, including an increased proportion of interstitial macrophages and increased expression of fibrotic and inflammatory fibroblast cell subsets. These results highlight the shared cellular features of HPS-PF and other ILD etiologies, providing further rationale to evaluate therapies developed for IPF for use in individuals with HPS.

Our studies in HPS murine models enabled analysis of age-dependent cellular and molecular pathology. A striking finding by both scRNA-seq and immunofluorescence imaging was the lack of fibrotic fibroblasts in HPS mouse models, particularly in 60-week-old HPS1-2 mice, despite evidence of increased collagen deposition as observed by Trichrome staining, collagen I immunofluorescence, and measured by Sircol assay total lung collagen content. Instead, we observed a robust increase in an inflammatory fibroblast population, marked by expression of Saa3, Lcn2, Il6, and Ccl2. This fibroblast population is similar to what has been previously described in both mouse models of pulmonary fibrosis and human IPF tissue^29,30,31. Furthermore, the prevalence of these inflammatory fibroblasts correlated with HPS model disease severity, with the largest proportion of inflammatory fibroblasts observed in aged HPS1-2 mice. This fibroblast population was not observed in HPS2-TG and HPS3 mice, which show minimal pulmonary defects¹², suggesting that these inflammatory fibroblasts are closely associated with the occurrence of fibrosis. HPS1 patient tissue also showed marked enrichment of inflammatory fibroblasts, further supporting disease relevance.

Serum amyloid A3 is a pleiotropic protein which is rapidly induced as part of the local immune response and has been shown to induce cellular migration and activation of both innate and adaptive immune cells, as well as endothelia^39,40,41,42. In our scRNA-seq dataset, Saa3 specifically marked the inflammatory fibroblast subpopulations, which allowed us to use this marker to study this unique cell type. However, as SAA3P is a pseudogene in humans, we needed to identify an alternative marker gene for human tissue. While other groups have used soluble frizzled-related peptide 1 (SFRP1) as a marker of these inflammatory fibroblasts^31,43, we observed very low levels of SFRP1 expression in our dataset. We observed that SFRP2, which has also previously been used to identify inflammatory fibroblasts²⁹, was upregulated specifically within our inflammatory fibroblast subsets, and therefore was used to identify our cell type of interest in human tissue.

Previous studies have proposed that inflammatory fibroblast subpopulations may act as “transitional” states between naïve conditions and CTHRC1⁺ fibrotic fibroblast states^29,30. In HPS1-2 mice, we observed the emergence and persistence of inflammatory fibroblasts starting as early as 8 weeks of age without detectable progression to fibrotic fibroblasts. Human HPS1 patient tissue showed presence of both inflammatory fibroblasts and fibrotic CTHRC1⁺ fibroblasts, with these two fibroblast populations appearing to co-exist within diseased lung tissue regions. In comparison to the advanced fibrotic state observed in the HPS1 patient tissue and in other experimental mouse models, the apparent lack of CTHRC1⁺ fibroblasts in the naïve HPS1-2 mouse model may relate to the relatively mild extent of fibrotic disease. In HPS models, the early emergence and subsequent persistence of inflammatory fibroblasts raises the possibility that these cells may have functions beyond simply serving as an intermediary cell state. Importantly, as our studies were performed in unchallenged murine models, these results do not exclude the possibility that inflammatory fibroblasts may be primed to transition into fibrotic fibroblasts following lung injury. We note that the presence of this inflammatory fibroblast phenotype alone is insufficient to induce fibrotic injury, as young HPS1-2 mice, as well as aged single mutant HPS1 and HPS2 models, had regions of prominent SAA3⁺ staining yet no significant increases in collagen deposition or other hallmarks of fibrosis. This finding also raises the possibility that other cell types, such as epithelia, may contribute to collagen deposition in HPS1-2 mouse models⁴⁴.

In examining the epithelial cell compartment in the scRNA-seq dataset, we found an overall reduction in the number of AT2 cells in HPS murine models associated with pulmonary fibrosis. These findings are consistent with flow cytometry analysis of whole lung tissue and immunostaining quantification recently reported by our group¹⁶. This reduction in overall numbers of AT2 cells occurs despite the presence of regions of “AT2 cell hyperplasia”, as has been reported by other groups^{10,15,45,46,47}, suggesting the potential for differing methodology to impact the overall quantitative data from what may be regional tissue differences. Further, due to the relatively low yield of epithelial cells across all HPS samples and apparent batch-specific effects, future studies with in situ methodologies are needed to draw definitive conclusions regarding various epithelial cell proportions.

When we explored biological processes that were altered in AT2 cells in our dataset, ORA found increased expression of genes associated with the production of TGF-β in AT2 cells from both 60-week-old HPS1-2 mice and HPS1 human lungs. A recent publication has described an AT2 cell-specific autocrine TGF-β signaling loop, wherein bleomycin-treated AT2 cells activate p53 signaling, and thus trigger release of TGF-β⁴⁸. This is intriguing, as a study by our group has also identified signs that HPS1-2 AT2 cells undergo p53-dependent senescence¹⁶. Ccnd1, a gene commonly associated with senescence, was the only gene that consistently upregulated across HPS1-2, HPS1, and HPS2 mouse models. As senescent epithelial cells have been shown to trigger fibroblast activation¹⁶, it is plausible that senescence-associated pathologies in AT2 cells contribute to the onset and progression of HPS-PF. One of the major features observed in AT2 cells from both HPS mouse models and HPS1 patient tissue was the presence of KRT8⁺/KRT17⁺ pathological transitional AT2 cells. These pathologic AT2 cell states have been readily observed to activate profibrotic processes in local lung environments^19,49,50. Moreover, this cell state has been associated with AT2 cell senescence^19,51, supporting a connection between Ccnd1 expression, KRT8⁺/KRT17⁺ AT2 cell expansion, and the progression of fibrotic lung injury. It is important to note, however, that we did not observe any changes in CCND1 expression or changes in p53-mediated signaling by ORA in human HPS1 patient tissue. This raises the possibility that such findings are no longer present or detectable by scRNA-seq in end-stage lung disease samples, or that other factors may be driving AT2 cell dysfunction in human disease that we do not observe in these murine models.

Finally, in AT2 cells from HPS mice, we also observed upregulation of biological processes involved in immune cell activation and migration. This finding may be relevant for the significant increase in CD68⁺ cells in the lungs of HPS mice⁵², as prior studies demonstrated that increased chemokine production and secretion from AT2 cells are crucial for excess macrophage recruitment and activation in HPS mouse models^14,52. Notably, although increased CCL2 protein levels have been reported in several studies^14,52, we did not detect enhanced Ccl2 gene expression from HPS mouse AT2 cells, suggesting potential contribution of a non-transcriptional regulatory mechanism.

We identified increases in the number of myeloid-lineage immune cells in an age-dependent manner in HPS mice. Notably, the extent of CD68⁺ immune cell accumulation also had a relationship to severity of HPS subtype. In addition to the increase in total immune cells, we observed a significant increase in the proportion of interstitial macrophages in the lungs of HPS1-2 and HPS1 mice, but not HPS2 mice. Studies have shown that the expansion of interstitial macrophage populations may be an early sign of fibrotic injury, occurring as early as five days post bleomycin administration^26,53,54. In HPS1 mice, the expansion of interstitial macrophages following bleomycin injury has been associated with upregulation of Tgfb1¹⁴, suggesting that these macrophages may be important in activating fibrotic programs in resident fibroblast populations. Further, the consistent increase in interstitial macrophages observed across human ILD scRNA-seq datasets supports the notion that this cell type plays an important role in fibrosis^26,53,54.

Utilizing the scRNA-seq dataset, we attempted to determine the cellular interactions responsible for the increase in inflammatory fibroblasts in HPS, including possible transcriptomic evidence of AT2 cell- or macrophage-mediated signaling to fibroblasts. However, the emergence of inflammatory fibroblasts appears to occur  in the absence of significant changes in inflammatory cytokine gene expression in either AT2 cells or myeloid cells. Furthermore, studies with supernatant from cultured HPS macrophages did not induce the inflammatory fibroblast gene expression profile in WT fibroblasts. Interleukin-1β has been implicated as a key cytokine in the onset and progression of fibrosis^55,56, including in the context of transitional AT2 cells⁵⁷ and activation of inflammatory fibroblasts²⁹. Inhibition of IL-1 signaling can mitigate fibrotic injury^{56,58,59,60,61}, and in bleomycin injury models, treatment with pirfenidone or nintedanib has resulted in significant reductions in IL-1β levels in lung tissue^62,63. We identified significant increases in IL-1α and IL-1β concentrations in lung tissue in 60-week-old HPS1-2 mice, but not in 8-week-old HPS1-2 mice, the timepoint when SAA3⁺ inflammatory fibroblasts initially appeared. Therefore, our data suggest that increased concentrations of IL-1 ligand are not directly responsible for the fibroblast phenotype in HPS mice, and additional studies are required to understand the signals responsible for the expansion of SAA3⁺ fibroblasts in HPS models.

Notably, IL-1R1 gene and protein expression were increased in fibroblasts from HPS1-2 and HPS1 mice. Knockdown experiments in MRC5 fibroblast cell lines demonstrated that the reduction of HPS1 alone is capable of increasing IL1R1 expression in vitro. Further, increased expression of IL1R1 was also observed in fibroblasts in HPS1 patient scRNA-seq data, suggesting that the loss of HPS1 may be linked to the upregulation of IL-1R1 in both mouse models and HPS1 patients. With these results, we hypothesize that the increase in receptor expression may cause exaggerated signaling in the absence of increased IL-1 ligand concentrations due to increased sensitivity to the ligand. However, additional studies would be required to confirm this hypothesis. Moreover, these results raise further questions. In mouse models, Il1r1 expression is significantly increased in 60-week-old HPS1 and HPS1-2 but is not observed in 8-week-old mice, suggesting there is an age-related component to the changes in expression. We did not observe significant changes in IL-1R1 gene expression in non-mesenchymal cell types in murine models or human tissue, raising the question of why this phenomenon is only observed in fibroblasts. Taken together, our findings suggest a novel mechanism by which exaggerated IL-1R1 expression in HPS1 fibroblasts may contribute to HPS-PF, thus providing a novel targetable strategy to test the effects of reducing IL-1 signaling. Future studies are required to understand how the loss of the HPS1 protein product BLOC-3 leads to increased IL-1R1 receptor expression, and what role age or other factors may play in augmenting this response.

In contrast to the findings in HPS1, we did not observe significant changes in IL-1R1 gene expression or cell surface expression in HPS2 fibroblasts, which led us to question how the loss of Ap3b1 in HPS2 causes the expansion of inflammatory fibroblasts in vivo. Previous studies by our lab have demonstrated that AT2 cell dysfunction underlies fibrotic susceptibility in HPS2 mice, as epithelial cell-specific correction of the Ap3b1 mutation is sufficient to reduce susceptibility to fibrotic injury^12,14. Here, we provide further evidence that AT2 cell dysfunction in HPS2 mice is also sufficient to induce the inflammatory fibroblast cell state. In the HPS2 model, we observed significant increases in inflammatory fibroblasts by both gene expression and immunostaining. Notably, inflammatory fibroblast gene expression and SAA3⁺ fibroblasts by immunostaining were restored to levels similar to WT mice in HPS2-TG mice, which have an epithelial-specific Ap3b1 correction. Moreover, in organoid models, we observed that HPS2 AT2 cells were sufficient to induce the inflammatory phenotype in WT fibroblasts, as observed by increased expression of Saa3, Ccl2, and Il6, while these results did not occur in organoids using AT2 cells from HPS2-TG mice. Interestingly, HPS1 AT2 cells failed to induce the inflammatory phenotype in WT fibroblasts in the organoid system, suggesting different AT2 cell-to-fibroblast interactions in HPS1 and HPS2 disease. However, it is important to note that our organoid experiments utilized 8-week-old AT2 cells and thus the results do not preclude the possibility that aged HPS1 AT2 cells may play a role in driving the inflammatory fibroblast expansion, or the increase in Il1r1 expression. Transcriptomic analysis of HPS2 AT2 cells suggested potential factors responsible for driving the inflammatory fibroblast cell state, including increased expression of Serpine2, which has previously been implicated in development of IPF⁶⁴, and increased expression of Il33, which has been shown to induce inflammation in cells neighboring the airway epithelium in SARS-COV-2 infection⁶⁵ and play a role in bleomycin-induced fibrosis in HPS1 mouse models²⁴. As lung tissue from patients with HPS2 is not available, additional studies are required to evaluate whether these potential factors induce SAA3⁺ fibroblast emergence in HPS2 mice.

As noted above, there were numerous disease characteristics that were observed in HPS1 mouse models, but not HPS2, and vice versa. There are differences in age of onset and disease progression between HPS1 and HPS2 patients, though historical paradigms have considered HPS1 and HPS2 as varying severity of a similar disease due to association with ILD^5,6,66,67. However, because HPS1 and HPS2 are due to loss of distinct proteins with distinct trafficking defects^46,68,69, it is not surprising that differences continue to emerge in the cell biology and model system phenotypes.

In summary, we have identified that an inflammatory fibroblast population is strongly enriched in the lungs of HPS mice in an age-dependent manner. This inflammatory cell state appears to be driven by increased IL-1 signaling due to increased expression of IL-1R1 in HPS1 and through a yet unknown mechanism involving AT2 cells in HPS2 mice. Inflammatory fibroblasts from young HPS2 and HPS1-2 mouse models have a reduced capacity to support AT2 cell survival and growth in organoid models, suggesting that these cells may in turn alter AT2 cell growth and proliferation. Further studies are required to understand the cell-intrinsic and cell-extrinsic factors which are responsible for these fibroblast cell states, and test whether pharmacotherapies directed at inflammatory fibroblasts may have a role in mitigating HPS pulmonary fibrosis progression.

Methods

Mice

All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC), and mice were housed in American Association for Accreditation of Laboratory Animal Care accredited facilities at the Children’s Hospital of Philadelphia (CHOP) where they received ad libitum access to chow and water. HPS mouse models were maintained as homozygous mutants on the C57Bl/6J (WT) background as previously described¹². C57Bl/6 mice were bred and maintained to serve as controls. Six- to nine-week-old mice were used as the young cohort (“8-week-old”), and 60+ -week-old mice were used as the old cohort (“60-week-old”). Data from male and female mice were pooled. We have complied with all relevant ethical regulations for animal use.

Human tissue samples

After Institutional Review Board (IRB) approvals and after informed consent, Hermansky-Pudlak Syndrome type I (HPS1) patient tissue was obtained from lung explants at the time of lung transplant at one of three lung transplant centers: University of Pennsylvania (IRB protocol #813685), University of Florida (IRB protocol #00035974), and Duke University (IRB protocol #00035974). Samples from the University of Florida and Duke University were processed at Vanderbilt University (IRB protocol #171657). Age-matched control tissue from deceased donors was obtained through the University of Pennsylvania Lung Biology Institute Human Tissue Lung Bank via The Prospective Registry of Outcomes in Patients Electing Lung Transplantation (PROPEL) study (IRB protocol #813685), which is a single-center prospective cohort study of matched donor-recipient pairs transplanted at the University of Pennsylvania (2011-2024). All ethical regulations relevant to human research participants were followed.

scRNA-seq library preparation and next-generation sequencing

Following euthanasia, mice were intubated, and 2.5 ml of Dispase I (BD Biosciences, BD354235) was instilled into the airways. The mice were extubated, lungs were removed en bloc and incubated in 1 ml Dispase at 37 °C for 6 min. Lung lobes were then separated and placed in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco 11965-092) containing 1 M HEPES (Gibco 15630-080), 1% penicillin/streptomycin, 5% FBS and 120 U/ml DNase I (Sigma Aldrich, D4527), followed by homogenization using the gentleMACS dissociator (Miltenyi Biotec). Suspensions were centrifuged at 400 × g for 5 min followed by incubation with ACK lysis buffer (Quality Biological 118-156-101) for five minutes on ice. The lysis buffer was neutralized with complete DMEM without DNase, and the resultant cell suspensions were filtered through a 40 µm filter. Live cell counts were determined by Acridine Orange/Propidium Iodine staining using the Nexcelom cell counter, and cells were loaded into the 10X Chromium Controller according to manufacturer’s instructions. After generation of libraries using the 10X Single Cell 3’ v3.1 protocol according to manufacturer’s instructions, sequencing was performed on an Illumina HiSeq 2500 instrument at a depth of approximately 250 million reads per sample.

Due to the low recovery of mesenchymal cells and epithelial cells, a second iteration of scRNA-seq was performed. Lungs were harvested and digested by a collagenase, dispase, DNase (CDD) mixture as previously described⁷⁰. Cells were stained for CD45 or EPCAM, and the cell populations were isolated by fluorescence-activated cell sorting (FACS). Cells were pelleted, live cell counts were determined by Trypan Blue and cell populations were re-combined at a ratio of 1 CD45⁺:4 CD45⁻ cells, or 1 EPCAM⁺:1 EPCAM⁻ cell. Subsequent loading, library generation, and sequencing were performed as described above.

Single-cell RNA sequencing read alignment and gene expression analysis

Reads in FASTQ format were aligned to the NCBI RefSeq GRCm39 (GCF_000001635.27) genome and reads with unique molecular identifiers (UMIs) were quantified at the gene level using STAR-solo⁷¹, with soloUMIdedup = ”1MM_CR”, soloUMIfiltering = ”MultiGeneUMI_CR”, and soloCellFilter = ”EmptyDrops_CR”. Ambient RNA was removed from the generated count matrices using the scvi-tools⁷² implementation of scAR⁷³. Ambient-corrected counts matrices were subsequently processed using scanpy⁷⁴ and anndata⁷⁵. Cells were filtered to include only those cells with greater than 300 unique genes detected/cell, fewer than 5317 unique genes detected/cell (the 95^th percentile), and with fewer than 15% of unique reads being attributed to mitochondrial genes. Genes were filtered from the combined dataset if detected in fewer than 20 cells. Putative cell doublets were identified and removed using the scvi-tools implementation of the SOLO doublet identification model⁷⁶.

Gene expression was calculated from scAR-corrected UMI-counts by first normalizing to CPM using the sc.pp.normalize_total function with target_sum = 1e6 and then calculating ln(CPM + 1) using the sc.pp.log1p function. The top 2000 highly-variable genes in the dataset were identified using the sc.pp.highly_variable_genes function, with flavor = ”cell_ranger” and n_top_genes = 2000. An initial principle component analysis was performed using the sc.tl.pca function with svd_solver = ”arpack”. This PCA was subsequently batch-corrected using the pytorch implementation of the harmony package⁷⁷. An additional low-dimensional embedding of the data was generated using the scvi-tools SCVI model, with categorical_covariate_keys = [“sex”, “age”] and continuous_covariate_keys =[“log1p_total_counts”]. A cell-cell neighborhood was calculated using the sc.pp.neighbors function, with n_neighbors=int(0.25*np.sqrt(adata.n_obs)) and use_rep = ”X_scVI”, i.e., using the SCVI low-dimensional embedding. This cell-cell neighborhood was used to generate a 2-dimensional uniform manifold approximation and projection (UMAP)⁷⁸ using the sc.tl.umap function with init_pos = ”spectral”. Finally, cell clustering was performed on the cell-cell network using the Leiden algorithm⁷⁹ via the sc.tl.leiden function with resolution = 1.

Putative cell compartment labels, and subsequently cell type labels, were assigned to cells using the scvi-tools implementation of the SCANVI model⁸⁰. Briefly, a small set of cells was used as ‘seeds’ for each of the expected cell compartment/type categories, identified by their maximal expression of established marker genes for each cell compartment/type. The gene expression profiles of these ‘seed’ cells were then used to build the SCANVI model and predict putative cell compartment/type labels of unlabeled cells in the dataset. Cell compartment/type labels were then manually collapsed on a per-cell-cluster (Leiden cluster) basis to refine the final cell compartment/type labels using the additional information provided by cell-cell neighborhood network clustering.

In human scRNA-seq datasets, publicly available datasets were processed from cell-by-gene UMI counts matrices and merged with data generated in this study. All cells within this merged dataset were filtered to include cells with greater than 300 unique genes detected/cell, fewer than 3100 unique genes detected/cell, and with fewer than 20% of unique reads being attributed to mitochondrial genes. Expression normalization, batch-corrected PCA, cell-cell neighborhood graph construction, clustering, and cell-type labeling were all performed as described above.

For all over-representation analysis, DEGs with log fold changes > 0.5 and false discovery rates < 0.05 were used for analysis. The top fifteen pathways based on term fold enrichment were presented.

Antibodies

Primary antibodies used were guinea pig-anti-mouse DC-LAMP (1:5000, Synaptic Systems 391-005), rat-anti-mouse CD68 (1:500, Invitrogen MA5-16674), rabbit-anti-mouse CD68 (1:500, Invitrogen 98876), rat-anti-mouse SAA3 (1:300, Abcam ab231680), rabbit-anti-mouse IL-1R1 (1:100, Invitrogen PA5-97866), rabbit-anti-mouse SM22 (1:1000, Abcam ab14106), rabbit-anti-mouse αSMA (1:250, Sigma Aldrich A2547), rabbit-anti-mouse CTHRC1 (1:500, Proteintech 16534-1-AP; 1:500, Abcam ab85739), rabbit-anti-mouse PDGFRα (1:1:500, Cell Signaling D1E1E), rat-anti-mouse CD11b (1:100, Invitrogen 14-0112-82), rabbit-anti-mouse CX3CR1 (1:100, Invitrogen 14-6093-81), goat-anti-mouse collagen I (1:500, Southern Biotech 1310-01), rabbit-anti-human SFRP2 (1:50, Abcam ab137560), rat-anti-mouse KRT8 (1:500, DHSB TROMA-I), rabbit-anti-human KRT17 (1:1000, Abcam ab53707), and mouse-anti-human HTII-280 (1:50, Terrace Biotech TB-27AHT2-280). Secondary antibodies used were donkey-anti-rabbit Alexa Fluor 488 (1:250, Invitrogen A21206), donkey-anti-mouse Alexa Fluor 488 (1:250, Invitrogen A21202), donkey-anti-rat Alexa Fluor 555 (1:250, Invitrogen A48270), donkey-anti-goat Alexa Fluor 647 (1:250, Invitrogen A21447), and donkey-anti-guinea pig 647 (1:400, Jackson Immunoresearch 706-605-148).

Precision cut lung slice immunofluorescence

Lungs were inflation/fixed with 4% PFA at a 30 cm constant pressure for 10 min, followed by overnight fixation in 4% PFA. Fixed lungs were washed in 1x DPBS and then embedded in 4% low melting point agarose (IBI Scientific, Dubuque, IA). Tissues were sectioned using a vibratome (VT1000S, Leica Biosystems) in 200 µm thick sections.

For staining, sections were incubated with a permeabilization buffer (1% Triton-X in 1x DPBS) for one hour at room temperature followed by incubation in staining buffer (1% bovine serum albumin, 0.3% Triton-X in 1x DPBS) with primary antibodies for 72 h at 4 °C. Samples were washed and incubated with secondary antibodies for 24 h at 4 °C. Samples were cleared using Scale A2 buffer (4 M urea, 10% glycerol, 0.1% Triton-X 100 in water) at 4 °C for 4–7 days, DAPI stained for 24 h at room temperature in Scale B4 buffer (8 M urea, 0.1% Triton X in water), and then washed in Scale B4 buffer without DAPI for 24 h at room temperature. Samples were re-incubated in Scale A2 for two to ten days at 4 °C until imaging was performed. Samples were imaged using the Leica DMi8 Confocal microscope and analyzed with Leica Application Suite X software.

Microtome cut lung section immunofluorescence

Lung tissue samples were paraffin-embedded and sectioned 7 µm thick using a Leica HistoCore BioCut microtome. Following heat-induced antigen retrieval, samples were incubated with primary antibody overnight at 4 °C followed by incubation with secondary antibody for one hour at room temperature the following day. DAPI was added, followed by coverslip mounting. Slides were imaged using the Leica DMi8 confocal microscope and analyzed with Leica Application Suite X software.

Hematoxylin & Eosin staining

Lung sections were deparaffinized and rehydrated, washed, and sequentially stained with Harris Modified Hematoxylin, differentiation solution, and eosin. Slides were then washed, mounted, imaged using the Leica DMi8 Thunder brightfield microscope and analyzed with Leica Application Suite X software.

Masson’s trichrome staining

Lung sections were deparaffinized and rehydrated and then placed in Bouin’s solution overnight at room temperature. Sections were washed, then sequentially stained with Weigert’s solution, Biebrich solution, phosphomolybdic-phosphotungstic solution, analine blue solution, and 1% acetic acid. Slides were mounted and imaged using the Leica DMi8 Thunder brightfield microscope and analyzed with Leica Application Suite X software.

Bronchoalveolar lavage and lung homogenate

Following euthanasia, mice were intubated with a 22G endotracheal tube, and bronchoalveolar lavage fluid (BALF) was collected by three serial lavages using 1 ml of 1x DPBS. Lavage fluid was centrifuged at 1000 × g for 10 min at 4 °C. BALF supernatant was collected, supplemented with 0.1% NP-40 and 1X protease inhibitor, and used to measure IL-1β concentrations by ELISA according to manufacturer’s instructions (R&D Systems MLB00C-1). The BALF-derived cell pellet was resuspended in Hanks Balanced Salt Solution (HBSS) and plated in a 96-well plate at a density of 5 × 10⁴ cells/well. Cells were cultured as previously described⁸¹, and cell supernatant IL-1β concentration was measured by ELISA according to manufacturer’s instructions.

Post-lavage lungs were perfused, resected, and frozen on dry ice. Lungs were homogenized in 2 ml of 1x DPBS supplemented with 0.1% NP-40 and 1X protease inhibitor using the Qiagen Tissue Lyser II system. Samples were then centrifuged at 10,000 × g for 10 min at 4 °C. IL-1α and IL-1β concentrations in homogenates were measured by ELISA according to manufacturer’s instructions.

Sircol collagen assay

Total lung collagen content was measured using the Sircol assay according to manufacturer instructions¹².

Primary murine fibroblast immunocytochemistry

Fibroblasts were isolated and plated in 8-well culture slides (Corning 354118) pre-coated with rat collagen type I. After fibroblasts reached ~80% confluence, cells were washed once fixed with 2% PFA for 20 min at room temperature. Cells were permeabilized with 0.25% Triton-X for 10 min and then incubated with primary antibody overnight at 4 °C, followed by incubation with secondary antibody for one hour at room temperature the following day. DAPI was added, followed by coverslip mounting. Slides were imaged using the Leica DMi8 confocal microscope and analyzed with Leica Application Suite X software.

Flow cytometry

For immune cell flow cytometry, mice were euthanized by sodium pentobarbital overdose and lung tissue was digested with CDD mixture as described above. Cells were incubated with anti-MHCII (clone M5/114.15.2), anti-CD24 (clone M1/69), anti-CD45 (clone 30-F11), anti-CD64 (clone S18017D), anti-CD11b (clone M1/70), anti-CD11c (clone N418; all antibodies diluted 1:200, Biolegend) and anti-SiglecF (clone E50-2440, 1:200, BD Biosciences) for 30 minutes at 4 °C. Cells were washed and resuspended in FACS buffer (HBSS, 2% FBS, 2 mM EDTA, 25 mM HEPES, 1% Penn/Strep). Data was acquired on the CytoFlex S and analyzed by FlowJo software (version 10.9.0; BD Biosciences). Cell types were defined according to previous definitions⁵³.

For cell surface analysis of IL-1R1 expression by flow cytometry, mice were euthanized, and lung tissue was digested with CDD mixture as described above. Cells were incubated with anti-CD45 (clone 30-F11), anti-CD31 (clone W18222B), anti-EPCAM (clone G8.8), anti-CD140a (clone APA5), and anti-IL-1R1 (Invitrogen; clone 191) for 30 minutes at 4 °C. Cells were washed and resuspended in FACS buffer (HBSS, 2% FBS, 2 mM EDTA, 25 mM HEPES, 1% Penn/Strep). Data was acquired on the CytoFlex S and analyzed by FlowJo software.

Fibroblast stimulation with macrophage supernatant

To isolate cells, mice were euthanized, lungs were perfused with ice-cold 1x Dubecco’s phosphate buffered saline (DPBS), and lung tissue was homogenized using a CDD mixture as described above. Whole lung cell suspension was divided, and cells were incubated with either anti-CD45, anti-CD31, anti-EPCAM, and anti-CD140a for PDGFRα⁺ fibroblast harvest, or anti-CD45 and anti-CD64 (clone S18017D) for macrophage harvest. After washing, cells were resuspended in FACS buffer containing either 7-AAD (for fibroblasts) or DRAQ-7 (for macrophages) live/dead stain and then sorted and collected on the FACS Jazz (BD Biosciences). Collected fibroblasts were resuspended in complete media (DMEM, 10% FBS, 1% penicillin/streptomycin), and macrophages were resuspended in macrophage-specific media (RPMI1640, 10% FBS, 1% penicillin/streptomycin, 20 ng/ml M-CSF (VWR, 10780-694)) and plated on a 24-well plate at a density of 5 × 10⁴ cells/well. Twenty-four hours after plating, fibroblasts were washed and placed in stimulation media (DMEM, 2% FBS, 1% penicillin/streptomycin). Macrophage cell culture supernatant was collected and added to fibroblast culture at a concentration of 1:10. Twenty-four hours after stimulation, cell culture media was removed, and fibroblasts were resuspended in TriZol for downstream RNA and cDNA processing as described below.

Primary organoid models

Lung tissue was digested using the CDD mixture as described above. Cells were stained as described above with anti-CD45, anti-CD31, and anti-EPCAM, as well as either anti-CD140a for isolation of PDGFRα⁺ fibroblasts, or anti-CD104 (clone346-11A) for AT2 cells. AT2 cells (CD45^-EPCAM⁺CD31^-CD104^-) and PDGFRα+ fibroblasts (CD45^-EPCAM^-CD31^-CD140a⁺) were sorted and collected on the FACS Jazz. Cells were co-cultured at a ratio of 5 000 AT2s:50 000 fibroblasts in MTEC Plus media as previously described for 21 days⁸². After culture, organoids were imaged using the Leica DMi8 Thunder microscope to determine colony forming efficiency (CFE) and average organoid size. Organoids with an area ≥ 1500 µm² were counted, consistent with previous studies⁸³. N = 15 biological replicates in triplicate or quadruplicate for organoids containing WT AT2 cells and WT fibroblasts; 4 biological replicates in triplicate or quadruplicate with organoids generated from HPS2 or HPS2-TG mice, and 3 biological replicates in quadruplicate in organoids generated from HPS1 or HPS1-2 mice.

After imaging, Matrigel was dissolved in Dispase in DMEM (StemCell 07923), and cells were dissociated with 0.05% trypsin. The cell suspension was incubated with anti-Epcam for 15 min, followed by in incubation with biotin-dynabeads. Magnetic sorting was used to enrich for EPCAM⁺ (“AT2s”) and EPCAM^- (“Fibroblast”) subpopulations. Cell subpopulations were centrifuged and resuspended in TriZol for downstream RNA and cDNA processing.

RNA interference

MRC5 fibroblasts (ATCC) or primary human dermal fibroblasts (gift from Dr. Michael Marks) were transfected with either non-targeting control siRNA (IDT 51-01-14-03), HPS1 specific siRNA (previously described³²), or AP3B1 specific siRNA (Invitrogen AM16708). Transfections were performed using a total of 20 nM siRNA with Lipofectamine RNAiMAX according to manufacturer’s instructions. Forty-eight hours after transfections, fibroblasts were washed twice with 1x DPBS and resuspended in TriZol for downstream RNA and cDNA processing.

Quantitative polymerase chain reaction (qPCR)

Total RNA was isolated using TriZol reagent (Invitrogen) and purified by the RNeasy Mini kit (Qiagen) following the manufacturer’s instructions. cDNA was prepared using High-Capacity RNA-to-cDNA kit according to manufacturer’s instructions (Invitrogen). qPCR was performed on QuantStudio 5 (Applied Biosciences) using the TaqMan Fast Advanced Master Mix (ThermoFisher). The following TaqMan Primer-probes were used: Gapdh (Mm99999915_m1), Saa3 (Mm00441203_m1), Il6 (Mm00446190_m1), Ccl2 (Mm00441242_m1), Tnfaip3 (Mm00437121_m1), Il1r1 (Mm00437237_m1), HPS1 (Hs00945781_g1), AP3B1 (Hs00186383_m1), IL1R1 (Hs00991010_m1), MYH10, (Hs00992055_m1). qPCR reactions were performed in triplicate and changes in gene expression were determined based on the delta-delta Ct method⁸⁴, with each sample gene expression normalized to Gapdh.

Statistics and reproducibility

At least three biological replicates were used for all experiments. All statistical analysis was performed using GraphPad Prism (v 10.2.3). All statistical analysis was performed as two-tailed analysis. All error Tests used for statistical analysis are noted in each figure legend. A p-value < 0.05 was considered statistically significant.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.