HPS1 Regulates the Maturation of Large Dense Core Vesicles and Lysozyme Secretion in Paneth Cells

HPS1, a BLOC-3 subunit that acts as a guanine nucleotide exchange factor of Rab32/38, may play a role in the removal of VAMP7 during the maturation of large dense core vesicles of Paneth cells. Loss of HPS1 impairs lysozyme secretion and alters the composition of intestinal microbiota, which may explain the susceptibility of HPS-associated inflammatory bowel disease. Hermansky-Pudlak syndrome (HPS) is characterized by oculocutaneous albinism, bleeding tendency, and other chronic organ lesions due to defects in tissue-specific lysosome-related organelles (LROs). For some HPS subtypes, such as HPS-1, it is common to have symptoms of HPS-associated inflammatory bowel disease (IBD). However, its underlying mechanism is largely unknown. HPS1 is a subunit of the BLOC-3 complex which functions in the biogenesis of LROs. Large dense core vesicles (LDCVs) in Paneth cells of the intestine are a type of LROs. We here first report the abnormal LDCV morphology (increased number and enlarged size) in HPS1-deficient pale ear (ep) mice. Similar to its role in melanosome maturation, HPS1 plays an important function in the removal of VAMP7 from LDCVs to promote the maturation of LDCVs. The immature LDCVs in ep mice are defective in regulated secretion of lysozyme, a key anti-microbial peptide in the intestine. We observed changes in the composition of intestinal microbiota in both HPS-1 patients and ep mice. These findings provide insights into the underlying mechanism of HPS-associated IBD development, which may be implicated in possible therapeutic intervention of this devastating condition.


INTRODUCTION
Rab2a, on LDCVs, which retains lysozyme in mature LDCVs, Loss of NOD2, LRRK2, or Rab2a from the surface of LDCVs results in lysosomal degradation of lysozyme in PCs (17). However, the regulatory mechanisms for the biogenesis and secretion of PC LDCVs are largely unknown.
HPS-1 is the predominant subtype accounting for 47.5% of reported HPS cases in the Chinese population (18). HPS1 and HPS4 proteins are found in the BLOC-3 complex. The known function of BLOC-3 is to recycle v-SNAREs on stage IV melanosomes (19) by acting as a GDP/GTP exchange factor (GEF) of Rab38 or Rab32 (20). Dysfunction of BLOC-3 leads to the retention of VAMP7 in melanosomes that affects their maturation (19) and exocytosis (21). Accumulating case reports have shown that among 11 HPS subtypes the incidence of HPS-1 is much higher than the others, and at least 10% of the reported HPS-1 cases had colitis or CD-like IBD (3). However, the role of HPS1 deficiency in the pathogenesis of CD-like IBD is still unclear. BLOC-3 is ubiquitously expressed and PC LDCV has LRO features like melanosome. However, whether HPS1 or BLOC-3 plays a role in the function of PC LDCVs, which is likely associated with the CD-like manifestations, remains elusive.
Here we use the HPS1-deficient mouse model, pale ear (ep), to explore the underlying mechanism of HPS-associated IBD. For the first time, we report compositional changes in the fecal microbiome of HPS-1 patients and ep mice and the abnormal secretion of lysozyme in ep PC LDCVs. This establishes a link between the dysfunction of PC LDCVs and the susceptibility of HPS-associated IBD.

Mice
The HPS1 deficient ep mice and C57BL/6J mice (WT) were originally obtained from Jackson Laboratory, and maintained at Dr. Richard T. Swank's lab. Mice were bred under specific pathogen-free conditions in the animal facility of the Institute of Genetics and Developmental Biology (IGDB), Chinese Academy of Sciences. Genotyping primers for the ep mutant a r e 5 ' -A C T G T G G G G T G G A C A T T T G G -3 ' a n d 5 ' -AGTGATGCGCCCTAGGCAAT-3 (348 bp amplicon), and primers for WT Hps1 are 5'-ACTGTGGGGTGGACATTTGG-3' and 5'-AGAAGCCTGCAAGCAAGACG-3' (264 bp amplicon). Age and gender matched WT and ep/+ mice were used as controls. All animal experiments were approved by the Institutional Animal Care and Use Committee of IGDB.

Patients
Patients recruited in this study were diagnosed with HPS-1 at Beijing Tongren Hospital from 2013 to 2018. Diagnosis of HPS was based on clinical manifestations, absence of platelet dense granules under electron microscopy, loss of platelet HPS1 or HPS4 protein by Western blotting (18). Mutations of the HPS1 gene confirmed the diagnosis of HPS-1 by next-generation sequencing (22). Patient H006 was reported previously (22). Patient H014 and H015 were reported recently (18) (Table S1).
Hospital, Capital Medical University. Fecal samples of control and HPS1 patient families were obtained after written informed consents. Fecal sample and patient data used in this study were approved by the internal review board of the bioethics committees of Beijing Tongren Hospital, Capital Medical University. The study was conducted according to the declaration of Helsinki principles.

16S rDNA Extraction and Microbiota Analysis
Sample collection, shipping, and DNA extraction were performed following the manufacturer's protocols. Fresh stool sample were collected in collection tubes with stool DNA stabilizer (1038111300, Stratec, Germany) and 16S rDNA was extracted with PSP Spin Stool DNA Plus Kit (1038110300, Stratec, Germany). All extracted DNA samples were kept at -80°C before further analysis. The variable V4 region of 16S rRNA gene was amplified from samples by IonS5 ™ XL platform using the primers 515F and 806R for bacterial diversity analysis. Cutadapt (V1.9.1, http://cutadapt.readthedocs.io/en/stable/) and Barcode were used to get the Raw Reads. The chimerical sequences were removed from the Raw Reads by comparing reads sequences to the species detection chimeric sequences (https://github.com/torognes/vsearch/) (23) to get the Clean Reads. Uparse (v7.0.1001, http://www.drive5.com/uparse/) were used for the clustering of the Clean Reads. Sequences with 97% of consistency (Identity) were clustered as the Operational Taxonomic Units (OTUs). Taxonomic analysis and community diversity were analyzed by the R software (Version 2.15.3) (24).
Alpha diversity index analysis was carried out for different groups under 97% consistency threshold; the data volume cutoff = 60481 was selected during homogenization. T-test and Wilcox rank sum test were used for parametric test and nonparametric test, respectively. T-test was conducted between groups to find inter-group differences at different levels of classification (Phylum, Class, Order, Family, Genus, Species) with significant differences (p < 0.05). By default, the results of the gate level are displayed, and if there is no significant difference in the gate level, the next level is displayed. LDA Effect Size (25) analysis was performed by the LEfSe software (http://huttenhower.sph.harvard.edu/galaxy), the default filter value LDA Score = 4.4.

Quantitative RT-PCR
Total RNA was extracted from distal ileum tissue of 5 pairs of age-matched WT and ep mice using RNeasy Mini kit (74104, Qiagen) and first-strand cDNA was generated with the iScript cDNA Synthesis Kit (1708890, Bio-Rad) according to the manufacturer's protocols. qRT-PCR analysis was performed using SYBR PrimeScript Ready Mix (Takara) in an ABI 7900 sequence detection system (Applied Biosystems). GAPDH expression was used for normalization. Total RNAs are from the dissected trunk region of zebrafish embryos. The PCR primers are: Gapdh-F: TCATCAACGGGAAGCCCATCAC, Gapdh-R: AGACTCCACGACATACTCAGCACCG, Lyz1-F: GCCAAGGTCTACAATCGTTGTGAGTTG, Lyz1-R: CAGTCAGCCAGCTTGACACCACG, Defcr1-F: TCAAGAGGCTGCAAAGGAAGAGAAC, Defcr1-R: TGGTCTCCATGTTCAGCGACAGC.

Immunoblotting
Mouse small intestine tissues were obtained from distal ileum and homogenized by RIPA lysis with protease inhibitor cocktail (Sigma-Aldrich). After 30 min incubation on ice, lysate was centrifuged at 12,000 g for 15 min to remove nuclei and cell debris. 20 mg of total protein per lane was loaded and separated within SDS-PAGE gels, and transferred to polyvinylidene difluoride membranes (PVDF, Millipore). The membrane was probed by antibodies described above and visualized with SuperSignal West Pico Plus (Thermo Fisher) by chemiluminescence imager (Sage Creation, China).

Transmission Electron Microscope Assay
Mouse distal ileum sections were fixed over night by 0.1 M sodium phosphate buffer (pH 7.2) containing 2.5% glutaraldehyde and 2% paraformaldehyde at 4°C and washed by double distilled water for 4 times to remove glutaraldehyde. Tissue sections were then fixed by 1% osmic acid for 1 h at 4°C followed by dehydration with an acetone gradient (30-50-70-80-90-100-100-100%). Infiltration and embedment with an Embed 812 kit (EMS #14120, Electron Microscopy Sciences) were operated as the protocol described. 70 nm slices were prepared with Leica EM UC7 Ultramicrotome and images were obtained with a JEM 1400 electron microscope (JEOL, Japan).

Immunofluorescence Confocal Imaging
Mouse distal ileum tissue were fixed overnight with 4% paraformaldehyde in PBS buffer at 4°C and embedded with paraffin. 5 mm slices were prepared with a rotary paraffin microtome (Leica RM2255) and mounted on positively charged slides, followed by antigen retrieval of 20 min steaming with 0.01 M sodium citrate buffer (pH 6.0). Tissue slices were then blocked with 1% normal goat serum (AR0009, Wuhan booster, China) in PBS buffer for 30 min at room temperature, incubated in primary antibody solutions (lysozyme 1:200, procryptdin 1:50) at 4°C overnight, washed by PBS buffer for 3 times and incubated in secondary antibody solutions (1:1,000) and fluorescence conjunct lectin. Nucleus were stained by DAPI while mounting. Confocal images were acquired with a 100xoil objective with NA 1.40 on an ECLIPSE Ti-C2 confocal microscope (Nikon, Japan). Images were obtained using the NIS-Elements AR 3.2 software provided by Nikon, and analyzed with NIH Image J.

Image Analysis
Quantification of LDCVs and Paneth cell phenotype was done by two blinded investigators independently. Five randomly selected images acquired by a 100× oil objective with NA 1.40 per animal were quantified. For degranulation analysis of hematoxylin-eosin (H-E staining), 2 discontinuous slices per animal were quantified; all crypts observed on the slices were included. For degranulation analysis of immunofluorescence staining, 20 crypts of each genotype were quantified.

In Vivo LPS Treatment
Ep and littermate control mice received LPS (Cat. code: tlrl-eblps, InvivoGen, diluted to 5 mg/ml in endotoxin-free water, 200 mg LPS/30 g body weight) or 200 ml endotoxin-free water (mock treatment) by mouth. Mice were sacrificed 1.5 h after treatment, and intestinal tissues were processed as described above. Serial sections were used for HE and immunofluorescence staining of lysozyme and procryptdin.

In Vitro LPS Treatment
Intestinal crypts were isolated as described (17). Pelleted crypts were suspended with 2 ml iPIPES buffer (10 mM PIPES, 137 mM NaCl) and then treated with LPS (2 mg/ml) or endotoxin-free water (mock). At the designated time points after treatment, 5, 10, or 15 min, crypts and supernatant were collected by centrifugation at 4°C. For immunoblotting, crypt lysates were prepared as described above. Supernatant was mixed with 1% protease inhibitor cocktail before subjected to immunoblotting.

OptiPrep Gradient Assay
Mouse small intestinal crypts were homogenized in 500 ml HB buffer (250 mM sucrose, 20 mM Tris-HCl, 1 mM EDTA, pH 7.4) with protease inhibitor cocktail followed by a centrifugation at 1,000 g for 10 min at 4°C to obtain post-nuclear supernatant. 50 ml of the supernatant was saved as input control, 400 ml of the supernatant was loaded onto 20 ml linear 5-50% linear Optiprep gradient (Axis-Shield, Norway) in HB buffer with a Gradient Master (Biocomp, USA) and centrifuged at 30,000 rpm for 16 h in an SW41 rotor (Beckman, USA) at 4°C. Thirteen equal fractions were collected from top to bottom and protein was precipitated by TCA. The precipitated protein and input control were analyzed by SDS-PAGE and immunoblotting.

Statistics
All values in the text are shown as mean ± s.e.m. The two-way ANOVA test and Tukey test were used in the microbiota analysis. Student's t-test was used in other experiments to compare means of two groups and Tukey test was used to compare means of more than two groups. The differences of p <0.05 were considered as statistically significant.

Altered Composition in Fecal Microbiota of HPS-1 Patients and ep Mice
Fecal samples of 3 HPS-1 patients (age 6-7 years, Table S1) from 3 unrelated families and 4 healthy children (age 6-12 years) from 4 unrelated families along with samples from their parents were collected for 16S rDNA sequencing. Taxonomy-based analysis showed no significant gross alteration in the phylum composition of HPS-1 fecal microbiome compared to the control group ( Figure 1A). Shannon diversity revealed no differences between HPS-1 patients and controls ( Figure 1B). LDA effect size (LEfSe) is an analytical tool used to find and interpret the biomarkers (genes, pathways, classification units, etc.) between groups of high-dimensional data. It most likely explains differences between classes by coupling standard tests for statistical significance with additional tests encoding biological consistency and effect relevance, and emphasizes statistical significance and biological correlations (25). To identify the characteristics of different abundances and the associated categories, LEfSe was used to interpret the comparison of HPS-1 patients and controls. Prevotellaceae in the phylum Bateroidetes is the trait type in HPS-1 patients at the family level while healthy controls show predominantly Ruminoccacceae, a family in phylum Firmicutes ( Figure 1C). Furthermore, no dominant species of the HPS-1 parent group were found in comparison to the healthy parent group ( Figure  S1). Similarly, a previous study reported that Prevotella is more abundant in fecal samples from patients with diarrheapredominant irritable bowel syndrome compared to healthy controls (26). Thus, HPS-1 patients, but not their parents, had altered composition of fecal microbiota.
The HPS-1 mouse mutant, ep (or ep/ep), carries a spontaneous IAP element insertion in the Hps1 gene resulting in the absence of HPS1 protein (27). To investigate whether ep mice have the risk of IBD as shown in HPS-1 patients, we sequenced fecal 16S rDNA from age and gender paired adult wild-type (WT) and ep mice. Taxonomy-based analysis suggested that the relative abundance of Firmicutes decreased in ep mice ( Figure 1D). Shannon diversity analysis revealed no significant differences between ep mice and WT controls, but a higher individual variation within the ep mice group was observed ( Figure 1E). LEfSe revealed that Prevotellaceae in phylum Bacteroidetes was also the dominant type in ep mice ( Figure 1F). Taken together, these results suggest that HPS1 deficiency affects the intestinal microbiota both in humans and mice, which may confer a risk for IBD development.

HPS1 Deficiency Leads to Altered LDCV Morphology in PCs
To determine whether HPS1 is involved in regulating the biological function of PCs, we first detected HPS1 expression in murine small intestine crypts containing PCs, transient amplifying cells and stem cells. Western blotting confirmed that HPS1 protein was expressed in WT crypts, and was absent in ep crypts (Figure 2A). Unfortunately, we did not test the intracellular localization of HPS1 because our HPS1 antibody did not work well for immunostaining. We then analyzed the phenotypes of LDCVs and PCs in ep mice to investigate the role of HPS1 in PCs. ep PC LDCVs displayed a typical dense core structure compared to WT PC LDCVs ( Figure 2B). We visualized PC LDCVs with FITC-UEA-1 which specifically binds to glycoprotein or fucose residue of lipopolysaccharide on the granule membrane ( Figure 2C). PC morphology analysis was performed as described (6) ( Figure  2F). For both ep and control mice, PCs with normal morphology were predominant (an average of 84% in WT mice vs. 83% in ep mice). The proportion of abnormal PC subtypes in ep mice had no significant changes compared to the WT mice ( Figure 2G). Furthermore, quantification of the immunofluorescence staining images showed that there was a significantly enlarged average diameter of LDCVs in ep mice compared to WT ( Figure 2D) and a significantly increased number of LDCVs per PC in ep mutants ( Figures 2E, S2). Thus, HPS1 deficiency is associated with an abnormal morphology in PC LDCVs.

No Apparent Changes of Lysozyme and Procryptdin Expression in ep PCs
To explore whether the abnormal LDCV morphology leads to altered expression of the contents in ep PCs, we carried out immunofluorescence staining of mouse small intestine slices. Lysozyme and procryptdins, major proteins of LDCVs which are main functional AMPs associated with mucosal anti-microbial activity and gastrointestinal innate immune response, were located within granule structures and were partially colocalized with Rhod-UEA1, which concentrates on the surface of mature LDCVs (28,29), in both control heterozygotes (ep/+) and ep mice. Lysozyme and procryptdin colocalized with UEA1 at most LDCVs while a small amount of LDCVs were only positive for the antimicrobial peptides (white arrow) or only positive for UEA1 (empty arrow) ( Figures 3G, H), but the distribution of each AMPs showed no difference between ep and heterozygotes ( Figures 3I, J). We next tested whether the transcription and expression of the two proteins in ep PCs was changed. Lyz1, the gene encoding lysozyme, and Defa1, a representative gene of procryptdins, were quantified by real time RT-PCR. The expression levels of the two genes in the small intestine of ep mice were not significantly changed compared with WT ( Figures 3A, B). Furthermore, no significant changes of lysozyme and procryptdins in the small intestine of ep mice were detected by Western blotting although the high variations were observed among samples ( Figures 3C-F). These results suggest that HPS1 deficiency does not affect the transcription or expression of lysozyme and procryptdins.

Lysozyme Secretion Is Impaired in ep PCs
To investigate whether the abnormal LDCV morphology in ep mice has an effect on PC function, age and gender matched ep/+ and ep mice were orally treated with lipopolysaccharide (LPS), which is a main component of gram negative bacterium cell wall and a TLR2/4/9 agonist triggering PC degranulation and secretion of AMPs (30). Hematoxylin-eosin (HE) staining analysis of small intestine sections harvested after 1.5 h revealed PC degranulation in control ep/+ mice while the ep PCs with mature morphology had no obvious degranulation ( Figure 4A). We classified intestinal crypts according to the morphology and distribution of LDCVs in PCs (31,32). Type I and type II crypts contain empty PCs or secreted LDCVs in crypts lumen, type III crypts contain non-degranulated PCs with abnormal vacuolar structures ( Figure 4B). Statistical analysis showed that after 1.5 h of LPS treatment, the proportion of type I and type II crypts in ep/+ small intestine sections was significantly increased compared to equally treated ep mice and untreated ep/+ mice (ctrl), while the proportion of type III crypts that were not supposed to be affected by LPS treatment showed no significant differences between the two genotypes ( Figures  4C, D). We observed that PCs with type I and II crypts in treated ep/+ mice had fewer lysozyme positive granules and UEA1 labeled LDCVs, but the number of procryptdin positive granules between the two genotypes showed no significant difference ( Figures 4E-G). Lysozyme secretion after LPS treatment was analyzed with isolated small intestinal crypts in vitro. We treated isolated crypts with LPS and followed lysozyme secretion into the supernatants. The results showed that lysozyme in supernatant fraction from isolated WT crypts was significantly increased 5 min after LPS treatment and reached a peak at 10 min, while un-secreted lysozyme in lysate fraction remained unchanged (Figures 5A,  B). In control ep/+ mice, supernatant lysozyme had similar changes to WT, while intracellular lysozyme content increased slightly at 10 min of LPS treatment compared to the mock treatment but decreased at 15 min ( Figure 5C). However, lysozyme in either supernatant or lysate fractions from ep crypts did not show such changes after LPS treatment compared to mock treatment ( Figure 5D). Taken together, HPS1 deficient PCs were unable to release lysozyme in response to LPS stimulation, and heterozygous ep/+ mice showed a slight defect in lysozyme secretion possibly due to dose effects of HPS1 protein.

HPS1 Deficiency Affects the Subcellular Localization of VAMP7
To further study the underlying mechanism of impaired lysozyme secretion in ep PC LDCVs, inferred from the function of HPS1 in melanosomes (19), we reasoned that HPS1 may participate in the recycling of v-SNARE proteins from PC LDCVs, which may be important for the maturation of LDCVs and concomitantly the release of LDCV cargo. HPS1 has been implicated in the recycling of VAMP7 from melanosomes to REs through the activation of Rab38/Rab32 (19). To address whether this pathway functions in PC LDCVs, we measured the total protein level and subcellular location of VAMP7 in ileum. Western blotting was performed to determine the expression of Rab38, Rab32, VAMP7, and other SNARE or regulatory proteins in small intestine tissue ( Figures S3 A-C). The protein levels of VAMP7, Rab38, and Rab32 ( Figure S3 D-F) together with VAMP4, VAMP8, synaptotagmin, and snapin ( Figures S3G,  H) were not affected by either the loss of HPS1 or upon LPS stimuli.
To explore whether the subcellular distribution of these proteins is altered, lysates of small intestinal crypts from ep mice and littermate ep/+ controls were fractionated by gradient centrifugation ( Figure 6A). Judging by the organelle markers, Golgi (GM130) and TGN (TGN38) were enriched in fraction 2, lysosomes (LAMP1) in fractions 9-11, endosomal structures (33) in fraction 11, and LDCVs (Lyz) in fraction 12. Among these proteins, VAMP7 in ep/+ crypts concentrated at the high density fractions labeled by LDCV cargo lysozyme (the 11th fraction 18%) and endosomal component TfR (the 12th fraction, 16%) and at the low density fractions enriched with Golgi and TGN (the 2nd fraction, 17%). However, in ep crypts, VAMP7 partially shifted from the 11th fraction (5%) to the 2nd fraction (28%) ( Figure 6D). This distributional shift suggests that in the absence of HPS1, VAMP7 failed to be removed from LDCVs and mostly remained in the Golgi and TGN (Figure 7). We also observed a slight shift of Rab38 to the middle density fractions (the 3rd to 5th fractions) in ep crypts ( Figure 6B), and Rab32 had similar distributional shift in ep crypts ( Figure 6C). These results suggest that deficiency in HPS1 GEF activity alters the membrane association of Rab32/38 which may affect the removal of VAMP7 from LDCVs, and that loss of HPS1 does not affect the cargo sorting-in process but rather compromises the pathway for cargo sorting-out, thus impairing the maturation of LDCVs.

DISCUSSION
In this study, we reported the abnormal morphology (increased number and enlarged size) in HPS1 deficient LDCVs, a type of LROs in Paneth cells. This alteration may impair the secretion of lysozyme, a key anti-microbial peptide associated with mucosal anti-microbial activity and gastrointestinal innate immune response. The abnormal secretion of lysozyme thereafter leads to the changes in the composition of intestinal microbiota, which may confer a risk for IBD development. Mechanistically, HPS1 is likely associated with the recycling of VAMP7 from LDCVs to promote the maturation of LDCVs. The immature LDCVs are defective in lysozyme secretion. This mechanism is conserved in melanosome maturation (19). Exploring the role of HPS1 in the biogenesis of LDCVs in Paneth cells provides new insights into the susceptibility of HPS-associated IBD and the role of HPS1 in innate immunity.
HPS1 functions as a GEF of Rab32/38 that mediates the conversion of Rab32/38 between the inactive GDP binding form and the active GTP binding form (20). In the melanocyte cell line MNT-1, Rab38 locates at melanosomes and indirectly interacts with VAMP7 via VARP (34). In this study we reported that in normal Paneth cells, Rab38, Rab32, and VAMP7 partially colocalized at LDCVs. VAMP7 is likely to be routed back to RE after recycling from LDCVs. In HPS1-deficient Paneth cells, VAMP7 shifted from LDCVs and endosome-associated fractions to the Golgi fractions, suggesting that the HPS1-Rab32/38-VAMP7 axis is likely to be a general mechanism in the biogenesis of LROs. This raises the possibility that LDCVs and REs interact for VAMP7 recycling by the action of HPS1. However, the machinery of how Rab32/38 mediates VAMP7 recycling awaits further investigation.
The NOD2-LRRK2-Rab2a pathway is an important molecular switch in the maturation of PC LDCVs that specifically regulates the sorting of lysozyme (17). LRRK2 protein contains a leucine rich repeat domain, an armadillo repeat, a kinase domain, an anchor protein domain, a RAS (GTPase) domain and a WD40 repeat (35). LRRK2 is also associated with Parkinson's disease (36) and is involved in vesicle transport, autophagy, cytoskeleton formation and mitochondrial function (37). The transportation and kinase activity of cytoplasmic LRRK2 are regulated by Rab32/38. LRRK2 binds to Rab32/38 in a GTP-dependent way via the armadillo repeat domain, and the complex is mainly colocalized at the endosomes and trafficking vesicles (38). Overexpression of constitutively activated mutant Rab32 led to increased LRRK2 localization to late endosomes and multi-vesicular bodies (39). We observed a slight shift of Rab2a in our gradient fractionation assay (data not shown). Whether HPS1 is the GEF of Rab2a, or regulates the NOD2-LRRK2-Rab2a pathway via its interaction with Rab32/38 needs to be investigated.
Case reports have shown that among eleven HPS subtypes the frequency of HPS-1 is much higher than the others (18), and about 10% of the reported HPS-1 cases had colitis or CD-like IBD (3). Development of HPS1-IBD, although it is at low frequency, may depend on complex genetic and environmental factors. There were no indications to perform endoscopy for our HPS-1 patients due to their young ages and no IBD symptoms. CD-like IBD have also been reported in some cases with HPS-4 or HPS-6 (3, 4). HPS1 and HPS4 form a tight BLOC-3 complex which functions as a GEF for Rab32/38 (20). BLOC-3 and Rab32/38 are involved in cargo removal (e.g. VAMP7) during LRO maturation (19). HPS6, HPS5, and HPS3 form the BLOC-2 complex which is known to be involved in cargo sorting during LRO biogenesis. BLOC-1 and AP-3 cooperate in this cargo sorting pathway (1,40). During LRO biogenesis, the HOPS complex plays an important role in autophagy which is required for LRO maturation probably by degradation of excessive membrane or proteins (41). Thus, deficiency in any of these HPACs may have defects in LRO biogenesis, which In ep crypt, the increase of Rab38 in the middle density fragments is significant (the 3rd~5th fractions, ep/+ crypts 9%, ep crypts 16%, *p < 0.05). (C) Statistical analysis of Rab32 distribution in (D). Rab32 in ep crypts significantly decreases at the low density fragment (the 1st fraction: ep/+ crypts 25%, ep crypts 16%, *p < 0.05) and at the 9th fraction (ctrl 4%, ep 2%, *p < 0.05). (D) Statistical analysis of VAMP7 distribution in (A). VAMP7 in ep crypts significantly increases in TGN38 and GM130 enriched fraction (the 2nd fraction, ctrl crypts 17%, ep crypts 28%, *p < 0.05) and the 4th fraction (ep/+ crypts 2%, ep crypts 6%, **p < 0.01), and decreases in the RE enriched 11th fraction (ep/+ crypts 18%, ep crypts 5%, *p < 0.05). 4 mice of each genotype were analyzed, values are means ± s.e.m. of three replicates of immunoblotting data. thereafter impairs LRO secretion. Regarding this general mechanism of LRO dysfunction in HPS patients, it is plausible to have multi-organelle defects of tissue-specific LROs such as melanosomes in melanocytes and LDCVs in PCs. However, symptoms or severity of HPS-associated IBD may vary in different HPS subtypes or in different susceptibility conditions of HPS patients. The secretion of lysozyme and the release of LDCV are essential for the maintenance of enteral homeostasis. We here report the altered fecal microbiota composition in HPS1deficient ep mice similar to the CD mouse models (42) and CD patients (43), reflecting a vulnerable enteral homeostasis. The interaction of genetics and the environment contributes to CD pathogenesis. Because the mutant ep mice were bred under specific pathogen-free facilities, ep did not have typical symptoms of IBD. HPS-1 patients in this study also had no CD symptoms at their young ages, but the fecal microbiota dominant phenotype Prevotellaceae including Prevotella predisposes a higher risk of diarrhea-predominant irritable bowel syndrome (26). However, whether fecal microbiota transplantation can alleviate the development or symptoms of HPS-associated IBD requires further investigation.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the internal review board of the bioethics committees of Beijing Tongren Hospital, Capital Medical University. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin. The animal study was reviewed and approved by Institutional Animal Care and Use Committee of Institute of Genetics and Developmental Biology, Chinese Academy of Sciences.

AUTHOR CONTRIBUTIONS
WL, ZL, and JY designed the research and wrote the manuscript. WL and ZL supervised the study. XH and CH analyzed the data and provided funding support. JY led the experimental investigation. QZ, YP, ZH, LY, YY, ZZ, and CZ performed some assays and analyzed the data. AW and TL diagnosed and collected the HPS patient samples. All authors contributed to the article and approved the submitted version.