Protein arginine methyltransferase 5 orchestrates the homeostasis and reconstitution of intestinal stem cell niche via balancing urea cycling

Abstract

Background

Protein arginine methyltransferase 5 (PRMT5) is highly expressed in many cancers and is a potential therapeutic target. It is also expressed in the small intestine, suggesting a role in intestinal health. This study explores PRMT5’s function in both normal physiology and radiation-induced intestinal injury (RIII), focusing on its effects on intestinal stem cells (ISCs) and their niche.

Methods

We examined PRMT5 expression in healthy and radiation-damaged intestines and treated mice and organoids with AMI-1, a PRMT5 inhibitor. Epithelial lineage composition, ISC proliferation, inducible nitric oxide synthase (iNOS) levels, and organoid activity were assessed. The impact of PRMT5 deficiency on ISC function was studied in vitro, and RNA-Seq and qRT-PCR were used to explore its effects on the urea cycle.

Results

PRMT5 was highly expressed in intestinal crypts. AMI-1 treatment reduced small intestine length, altered epithelial morphology, and increased secretory cells. In healthy intestines, PRMT5 inhibition enhanced Olfm4 + ISCs and induced iNOS expression. After radiation, PRMT5 deficiency inhibited ISC proliferation and caused Paneth cell acidification in the ISC niche. Organoids showed reduced vitality. PRMT5 deficiency disrupted the urea cycle, upregulated iNOS, increased NO production, and elevated lipid and ROS levels, impairing ISC homeostasis.

Conclusions

PRMT5 is critical for maintaining intestinal homeostasis and regeneration. Its deficiency disrupts ISC niche function, highlighting PRMT5 as a potential target for treating intestinal disorders.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12964-026-02971-4.

Background

Gastrointestinal (GI) tract is responsible for the digestion and absorption of ingested food and fluids and is of critical importance for human health [1, 2]. The epithelium of GI tract is also one of the most radiosensitive organs, and it is susceptible to injury and acute syndrome after high doses exposure to ionizing radiation [3]. Radiotherapy is one of the most frequently used treatment for pelvic cancers, thus it inevitably leads to different degrees of intestinal epithelial injury [4]. Our previous studies have found that inducible nitric oxide synthase (iNOS) is physiologically expressed in Paneth cells (PCs) and plays a pivotal role during the homeostatic maintenance for the niche intestinal stem cells (ISCs) [5]. However, the specific molecular mechanisms are still unclarified and still deserve further exploration.

Protein methylation on arginine residues was first reported in the late 1960s and early 1970s [6]. This post transcriptional modification (PTM) depends on protein arginine methyltransferases (PRMT), and PRMT1 is the first member of PRMT family discovered in 1996 [7]. Protein arginine methyltransferase family comprises nine members, classified into three types based on their catalytic activity [8]. Type I enzymes (PRMT1, 2, 3, 4, 6, and 8) mediate asymmetric dimethylation of arginine residues, type II enzymes (PRMT5 and 9) catalyze symmetric dimethylation, and type III (PRMT7) enzymes are monomethyltransferases (MMT) [9]. Among all these PRMTs, previous studies have shown that PRMT5 is highly expressed in cancer cells and tissues, and that its overexpression is directly correlated with the progression of a variety of solid malignancies [10]. PRMT5 is also the only PRMT that requires the obligatory cofactor methylosome protein to exert its function [11]. PRMT5 together with methylated histone 50 (MEP50) (also known as WDR77) forms an octameric heterodimer that catalyzes the methylation of arginine residues in histone tails, leading to repression of tumor suppressor genes and activation of cell proliferation and invasion genes [12]. In GI tract, PRMT5 is highly expressed in gastric and colorectal cancers, and early results indicate that PRMT5 promotes cancer cell growth through transcriptional activation of oncogenes and/or inhibition of tumor suppressor proteins [13, 14].

However, in recent years, PRMT5 has been noticed due to its protective functions in tissues. For example, some studies have shown that PRMT5 acts as a tumor suppressor protein in the early development stage of gastric cancer. And PRMT5 has also been shown to protect the intestine by modulating the function of goblet cells [15, 16]. While PRMT5 comes into the limelight as a therapeutic target, the corresponding PRMT5 inhibitors are emerging. Of all the PRMT5 inhibitors, Arginine N-methyltransferase inhibitor 1 (AMI-1) is a potent, cell-permeable, reversible inhibitor of protein arginine N-methyltransferases [17–19]. It inhibits PRMTs by blocking peptide-substrate binding, and specifically inhibits the arginine methyltransferase activity but not the activity of lysine methyltransferase. In colorectal cancer, AMI-1 has been reported to inhibit PRMT5 activity by reducing arginine methylation of eIF4E and FGFR3 [20]. Unfortunately, the roles of PRMT5 in the small intestine are still lack of careful investigation and exploration.

In this study, we aimed to investigate the viscounty of PRMTs in the niche of ISCs. Using RNA-Seq of isolated crypts from small intestine, we unexpectedly found that PRMT5 was highly expressed in the crypts, and its expression was further upregulated following radiation-induced intestinal injury (RIII). These findings suggest that PRMT5 may play an essential role in maintaining small intestinal function. To further validate its function, we used AMI-1 to inhibit PRMT5 expression in both healthy and radiation damaged small intestines. And it was found that PRMT5 deficiency could lead to the disruption of intestinal urea cycle and dysregulation of arginine metabolism, which resulted in an increase of iNOS expression and NO production. This, in turn, caused abnormal stem cell behavior and acidification of Paneth cells, ultimately impairing intestinal health.

Methods

Animals and ethical statement

Six- to eight-week-old C57BL/6 mice were ordered from HFK Bioscience Co., LTD (Beijing, China). Lgr5-EGFP-IRES-CreERT2 (Lgr5^CreERT2) mice were obtained from Jackson Laboratory (Bar Harbor, ME, USA). All the mice were maintained in a specific pathogen-free (SPF) facility with a 12 h light/dark cycle, and they were allowed to take food and water ad libitum. All the mice were acclimated for a week before the experimental use. All the experimental procedures were complied with the guidelines for the Care and Use of Laboratory Animals of National Institutes of Health (NIH) and approved by the Ethics Committee of Sichuan Cancer Hospital & Institute (SCCHEC-04-2024-036).

AMI-1 and radiation treatment

After the acclimatization for a week, mice were randomly divided into four groups, including the control group, AMI-1 group, radiation group, and radiation + AMI-1 group. In order to establish the model of radiation induced intestinal injury, mice were anesthetized by intraperitoneal (i.p) injection of 1% sodium barbiturates before irradiation. The whole abdomens of the mice were irradiated using an X-Rad 320 irradiator (PXI, Connecticut, USA) for a single dose of 8 Gy, and the dose rate was 75 cGy/min. During the radiation, all the other body parts of mice were protected by lead pates. In the AMI-1 group, mice were intraperitoneally injected with AMI-1 (HY-18962, MCE, China) at a dosage of 200 mg/kg/day for 7 consecutive days. In the radiation + AMI-1 group, mice were given AMI-1 in the same way for 3 consecutive days after irradiation. Mice in the control group and radiation group were only injected with the same volume of PBS.

Tissue collection and histological staining

Mice were injected with BrdU (100 mg/kg) at 90 min before tissue collection. Tissue samples were harvested at the indicated time points. Intestinal tissues were fixed with 4% cold paraformaldehyde (PFA, BL539A, Biosharp, China) for 72 h, and then dehydrated and paraffin embedded using the standard histological protocol of our laboratory. Sections with 4 μm thickness were subjected to hematoxylin-eosin (H&E) staining. Different intestinal segments, including duodenum, jejunum, ileum, and colon, were stained. Goblet cells were stained with an alcian blue kit (#E670107, BBI) following the manufacturer’s protocol. Briefly, alcian blue was used for acid mucus polysaccharide detection, and PAS was used to stain the neutral mucus substance. Nuclei were stained with fast red or hematoxylin. Histological images were captured with an Olympus BX53 microscope. The morphological parameters, including villous height, crypt depth, and crypt density, were measured using Image J (NIH, USA).

Immunological staining

Slides were deparaffinized in xylene and rehydrated. After that, antigen retrieval was performed for 20 min in boiling Tris-EDTA antigen retrieval solution (#BL618A, Biosharp). When slides were cooled down, antigen blocking was conducted using 1% bovine serum albumin (BSA, #A7906, Sigma−Aldrich, USA) containing 0.5% Triton X-100. Primary antibodies were diluted and incubated with tissues overnight at 4℃. The specific details of primary antibodies were listed in the supplementary Table S1. On the second day, HRP-linked secondary antibody solution (ZSBio, Beijing, China) was incubated, and a DAB kit (ZSBio, Beijing, China) was used for immunohistochemistry (IHC)-based visualization. For immunofluorescent (IF) staining, Alexa Fluor™ 488 or 594 labeled highly cross-adsorbed donkey anti-rabbit or mouse IgG (H + L) antibodies (Thermo Fisher, USA) were used, and nuclei were stained by DAPI (Vector Laboratories, Burlingame, CA). Slides were carefully washed and mounted with anti-fade mounting medium (Beyotime, China). IHC images were captured using an Olympus BX53 microscope. IF staining images were taken by a Nikon confocal microscope (Nikon, Japan) or an Axio Observer with Apotome 3 (ZEISS, Germany).

Organoid culture and treatment

Isolation of small intestinal crypts and organoid culture was performed following our previous protocol [21]. Briefly, fresh small intestinal segments were quickly isolated from wild-type or Lgr5-CreERT2 transgenic mice for organoid culture. Intestinal samples were flushed with ice-cold PBS for three times to remove the contents. Intestines were cut into 3 to 5 mm pieces, and then incubated in PBS supplemented with 5 mM EDTA (25,300,096, Invitrogen, USA) and 1x penicillin/streptomycin solution for 30 min on ice. Crypts and villi were detached by vigorous shaking and examined under a microscope. The supernatant was filtered through a 70 μm nylon strainer to enrich the crypts. Individual crypts were collected by centrifugation at 4°C and 800 ×g for 3 min. The density of crypts was determined, and the crypts were centrifuged and mixed with Matrigel (#354230, Corning, USA). Organoids were cultured using IntestiCult™ Organoid Growth Medium (OGM) (Stemcell Technologies, Canada). Culture medium was replaced every three days. AMI-1 was supplemented into the medium to mimic the deficiency of PRMT5. Images of organoids were taken by Cytation 5 (Agilent Technologies, USA).

Cell culture

Intestinal epithelial cell line IEC-6 was introduced from American Type Culture Collection (ATCC, USA), and Caco-2 was obtained from Cell Bank of Chinese Academy of Sciences (Shanghai, China). All the cell lines were maintained in our laboratory. Cells were cultured at 37℃ in an incubator (Thermo Fisher, USA) with 5% CO₂. Culture medium was used following the provider’s instruction and supplemented with 10% FBS and 1% penicillin/streptomycin. Cells were treated by different concentrations of AMI-1 ranging from 0.25 mM to 1.25 mM to mimic PRMT5 deficiency. IEC-6 or Caco-2 cells were incubated with AMI-1 for 24 h. For staining, cells were fixed with 4% PFA for 30 min, washed with cold PBS. Cells are subsequently used for the detection of various indicators.

siRNA mediated Prmt5 knockdown.

IEC-6 cells were seeded and transfected at approximately 60–70% confluence. Scratch and Prmt5 siRNAs were purchased from Sangon Biotech (Shanghai, China). Transfection complexes (250 µL Opti-MEM, 10 µL Lipofectamine 2000, and 25 µM siRNA) were prepared and incubated for 20 min at room temperature to allow complex formation. Cells were then treated with the transfection mixture containing siRNA for 24 h, after which the medium was replaced with fresh complete medium. Total RNA was extracted at 48 h post-transfection to assess the knocking-down efficiency by qRT-PCR.

ROS detection

The intracellular level of reactive oxygen species (ROS) was detected using a dihydroethidium (DHE) probe (S0063, Beyotime, China) following the manufacture’s protocol. Briefly, DHE was dissolved in DMSO, and loaded into the culture medium at a final concentration of 5 µM. Cells were incubated with DHE probe at 37℃ for 30 min before observation.

Detection of lipid droplets

Intracellular lipid droplets were detected using the Lipid Droplet Staining Kit (C2050S, Beyotime, China) according to the manufacturer’s instructions. Briefly, cells were washed with PBS and fixed with fixative solution for 10–15 min at room temperature. After fixation, cells were incubated with the working dye solution in the dark for 10–20 min. Subsequently, cells were washed with PBS to remove excess dye and counterstained with DAPI for nuclear visualization. Stained cells were imaged using a fluorescence microscope.

Determination of nitric oxide (NO)

Cells were seeded into 96-well plates and treated with AMI-1 (0.75 mM) for 24 h. For the radiation group, cells were irradiated using X-Rad 320 for a single dose 15 Gy. At the end of the treatment, NO levels were determined using the Nitric Oxide Assay Kit (S0021S, Beyotime, China) following the manufacture’s protocol. Briefly, cell supernatants were collected and centrifuged. 50 µL of standard and experimental samples were added into each well of a 96-well plate. After that, 50 µL of Griess Reagent I was loaded, then followed by 50 µL of Griess Reagent II. Cells were incubated at room temperature in avoid of light. The absorbance was measured by Cytation 5 at 540 nm.

RNA extraction and qRT-PCR

Total RNA was isolated from mouse intestinal tissues or cultured cells using RNAiso (9109, TaKaRa, Japan) according to the manufacturer’s protocol. The quality and concentration of RNA samples were examined using a NanoDrop2000 spectrophotometer. Reverse transcription was performed using Hifair II 1st Strand cDNA Synthesis SuperMix (11137ES60, YEASEN, China). qPCR was conducted using Hieff qPCR SYBR Green Master Mix (11203ES08, YEASEN, China) using a C1000 machine (Bio-Rad, USA). Primer sequences used in this study were listed in the supplementary Table S2. Gene expression results were normalized to that of β-actin, and the relative expression was determined by the 2^−ΔΔCt method.

RNA-Seq assay

Fresh small intestinal crypts were isolated from mice in healthy condition or after different treatments as indicated in the relative panels. Intestinal crypts were gently washed and centrifuged in could DEPC-treated molecular-grade water (10601ES76, YEASEN), then quickly frozen by liquid nitrogen. Crypts were preserved in -80°C freezer for the next analysis. RNA-Seq was performed in accordance with our previous description (Refs: Journal of Pathology). Briefly, total RNA extraction, RNA integrity evaluation, library construction, and sequencing were performed by OE Biotech Co., Ltd. (Shanghai, China). RNA-seq data analysis was also performed by OE Biotech Co., Ltd.

Transmission electron microscopy (TEM)

IEC-6 cells were cultured in T75 flasks until cells grew into the confluency of about 80%. Culture medium was replaced by fresh medium containing AMI-1 at the final concentration of 0.75 mM. After 24 h treatment, IEC-6 cells were detached by 0.25% Trypsin solution containing 0.02% EDTA. Cells were collected, washed, and quickly fixed for 1 h in cold 0.1 M sodium Cacodylate-HCl buffer (pH 7.4) containing 4% PFA and 1% glutaraldehyde, then washed in 0.1 M cacodylate buffer containing 0.1 M glycine. Cells were further postfixed in 1% osmium tetroxide in 0.1 M cacodylate buffer, dehydrated in a methanol series to propylene oxide, and embedded in epoxy resin. Ultrathin sections were placed on nickel grids and then examined and photographed by transmission electron microscopy (Tecnai-10 electron microscope; Philips, Amsterdam, The Netherlands) at an accelerating voltage of 100 kV.

Statistical analysis

All the values were presented as the mean ± SD. Data analysis was performed using GraphPad Prism 9 (GraphPad Software, USA). Comparisons between two different groups were conducted by two-tailed unpaired Student’s t-test. Differences with P values less than 0.05 were considered statistically significant (*:P < 0.05, **:P < 0.01, ***: P < 0.001, ****: P < 0.0001). P > 0.05 was considered to be nonsignificant (n.s).

Results

PRMT5 is important for maintaining the gross structure and histology of intestine under normal physiological condition

Of all the nine members in the PRMT family, PRMT5, PRMT1 and CARM1 (PRMT4) have the highest expression in cancer [11]. We found that the gene expressions of PRMT1, PRMT5 and PRMT4 within healthy small intestinal crypts also ranked the top three (Fig. 1A). This phenomenon made us to infer that PRMT5 might be necessary to maintain the physiology of intestine. Therefore, we treated mice with AMI-1 (the specific inhibitor of PRMT5) by intraperitoneal injection for six times and observed the body weight of the mice (Fig. 1B). And there were no significant differences in body weight (Fig. 1C). However, the gross morphology of GI tract was significantly changed compared with that of the control group. It was found that the length of the small intestines in AMI-1 treated mice was significantly decreased than that of the control group, and there was no significant change in the length of the colons (Fig. 1D, E). We further analyzed the histological parameters of intestines using H&E staining and quantified the length of the villi and the depth of the crypts, and additionally measured the mucosal thickness of the colon (Fig. 1F-I). The results demonstrated that AMI-1 predominantly affected the histological characteristics of jejunum and ileum, while it also impacted mucosal thickness of the colon. Therefore, these results prove PRMT5 is highly expressed in small intestine and PRMT5 inhibition could disrupt the histology of intestine in healthy mice.

Fig. 1.

PRMT5 is necessary to maintain the gross structure and histology of intestine under physiological condition. A Transcriptome heatmap for the expression of PRMT family members under normal physiological condition of small intestine (n = 3). B Schematic diagram of AMI-1 treatment in physiological conditions. C Statistical graph of the body weight changes of mice. D Representative gross images of the whole gastrointestinal tract of mice. E Statistical analysis for the lengths of small intestines and colons between control group and AMI-1 group. F Representative H&E images of different intestinal segments from the control group and the AMI-1 group (Bar = 100 μm). G Statistical analysis for the length of villi in different small intestinal segments. H Statistical analysis for the cryptal depth in small intestinal segments. I Comparison for mucosal thickness of colon between two groups. There were at least 3 mice included in each group for every comparison. **: P < 0.01, ****: P < 0.0001, ns: P > 0.05

PRMT5 deficiency leads to abnormal secretory epithelial lineage differentiation

Since AMI-1 caused significant changes of intestinal morphology, we further examined the epithelial differentiation using IHC staining. First, we confirmed that AMI-1 could effectively inhibit the expression of PRMT 5 in the crypts (Fig. 2A, B), and the mRNA level of Prmt5 likewise proved the point (Fig. S1A). To explore how the inhibition of PRMT5 would impact the constitution of intestinal epithelium, we performed IHC staining against MUC2 to show goblet cells and found an apparent increase in goblet cells on villi (Fig. 2A, C). Mucins stained by Alcian blue with different pH values also showed the same results (Fig. 2F, G). Meanwhile, we found that inhibition of PRMT5 also resulted in a significant increase of DCLK1⁺ Tuft cells and ChrgA⁺ enteroendocrine cells (Fig. 2A, D, E). Paneth cells are the key players of mucosal defense and ISC niche, and the antimicrobial peptide secreted by Paneth cells and the mucus layer of the goblet cells together build up the physical and chemical barrier of intestinal epithelium. Interestingly, when we examined the cryptal cells, it was observed that Paneth cells located at the base of small intestinal crypts showed an expansion in both the quantity and cellular size (Fig. 2F, H,I, J). Thus, these results support that the inhibition of PRMT5 leads to an obvious abnormality in secretory epithelial lineage differentiation, which might change intestinal mucosal immunity and ISC niche.

Fig. 2.

PRMT5 deficiency causes abnormal differentiation towards secretory epithelial lineages. A Representative IHC images stained for PRMT5, goblet cells (MUC2), Tuft cells (DCLK1), and enteroendocrine cells (EECs, ChrgA) between control group and AMI-1 group (Bar = 100 μm). B Quantification for average optical density (AOD) values of PRMT5. C Statistical analysis for the quantity of MUC2⁺ goblet cells per villus. D Comparison for the number of DCLK1⁺ Tuft cells per villus-crypt axis. E Statistical analysis of the number of ChrgA⁺ EECs in the villus-crypt axis. F Representative staining images of Alcian blue-positive goblet cells and Lysozyme⁺ Paneth cells between two groups (Bar = 100 μm). G Quantitative analysis for goblet cells stained by Alcian blue at different pH values. H Quantitative analysis for Paneth cells (PCs) per crypt. I Statistical analysis for the area of PCs. J 3D demonstration for Lysozyme fluorescent intensity in PCs based on the above pictures. There were 3 mice used in each group for every comparison. *: P < 0.05, **: P < 0.01, ***: P < 0.001

Inhibition of PRMT5 enhances iNOS expression in small intestinal stem cells

Paneth cells are the niche cells of ISCs, and due to the changes of Paneth cells after AMI-1 treatment, we further explored the effects of PRMT5 inhibition on the behavior of ISCs. Firstly, we detected the proliferative cells by IHC staining against Ki67. It was shown that the inhibition of PRMT5 by the administration of AMI-1 significantly decreased the number of Ki67⁺ proliferating epithelial cells within small intestinal crypts (Fig. 3A, B). However, there was an increased number of Olfm4 + ISCs (Fig. 3A, C). Meanwhile, we also examined the mRNA expression of Lgr5, which is the marker of active ISCs. It was confirmed that Ami-1 treatment definitely induced an upregulation of Lgr5 gene expression at the transcriptional level in AMI-1-treated samples (Fig. S1B). Because PRMT5 is a methyltransferase of arginine that shares the same substrate as iNOS, we postulated that whether the expression of iNOS would be affected by in the inhibited activity of PRMT5. Previously, we have reported that Paneth cells could express iNOS in physiology. Surprisingly, here we observed that AMI-1 treatment caused a large number of ISCs to express iNOS (Fig. 3A, D). In order to confirm this result, we also performed IF co-staining of Lysozyme and iNOS, and the results were consistent with the previous observation that iNOS expression was increased in ISCs when the expression of PRMT5 was inhibited by AMI-1 (Fig. 3E, F). To further confirm the AMI-1 induced co-localization of iNOS with ISCs, we performed immunofluorescence analysis using Lgr5 reporter mice and observed a clear overlap between iNOS signals and EGFP-positive Lgr5⁺ ISCs (Fig. S2). Our data demonstrated that the inhibition of PRMT5 would reduce the proliferation of cryptal epithelial cells, and AMI-1 unexpectedly increased the number of ISCs with ectopic iNOS expression, which means an iNOS shifting from Paneth cells to ISCs, indicating a potential compensatory mechanism in ISC niche.

Fig. 3.

PRMT5 deficiency increases iNOS expression in small intestinal stem cells. A Staining of Ki67, Olfm4, and iNOS between the control group and AMI-1 group (Bar = 50 μm). B Comparison of Ki67-positive cells within intestinal crypts (n = 3 each group). C Statistical analysis of Olfm4⁺ ISCs in each crypt (n = 3 each group). D Quantification of iNOS expression in Paneth cells and ISCs (n = 3 each group). E Distribution of iNOS within crypts co-stained by Lysozyme (Paneth cell) in control and AMI-1 groups (Bar = 50 μm). F 3D constructs for the signaling distribution of iNOS in the crypts. **: P < 0.01, ***: P < 0.001, ****: P < 0.0001

PRMT5 deficiency exacerbates histological injuries of small intestine after radiation

Given the above results, we think that PRMT5 is required for the intestinal homeostasis. We wonder how the inhibition of PRMT5 would impair the epithelial reconstitution after RIII. Firstly, we found an enhanced expression of Prmt5 in small intestinal crypts after 8 Gy irradiation in RNA-Seq dataset (Fig. 4A, Fig. S3). Consistent with these data, qPCR results also confirmed that radiation exposure induced an upregulation of Prmt5 gene expression (Fig. S1C). In order to block the activity of PRMT5 after RIII, we intraperitoneally injected mice with AMI-1 after 8 Gy irradiation for three consecutive days (Fig. 4B). Notably, AMI-1 treatment increased the sensitivity to radiation injury, and mice couldn’t survive beyond day 3–4 post-irradiation, precluding analysis for later time points. These findings indicate that PRMT5 inhibition exacerbates radiation-induced intestinal injury and impairs survival. We didn’t observe any significant difference in the body weights of mice (Fig. 4C). From the gross GI images, the length of the small intestine in both the 8 Gy group and the 8 Gy + AMI-1 group was significantly shortened compared with that of the control group, but the length of colon did not change significantly. Although the length of small intestine in 8 Gy + AMI-1 group was not shorter than the 8 Gy group, we found that its small intestinal morphology deteriorated more seriously, especially the cecum was swollen and full of fluids (Fig. 4D, E). Histological staining further showed that the injection of AMI-1 after RIII further aggravated the damage in intestinal epithelium (Fig. 4F). Interestingly, we noticed that PRMT5 deficiency after RIII resulted in the appearance of many vacuolar structures in the crypt portion of small intestine. At the same time, the length of intestinal villi was significantly lower in both the 8 Gy and 8 Gy + AMI-1 groups than the control group, and the length of intestinal villi was even shorter in the 8 Gy + AMI-1 group than in the 8 Gy group (Fig. 4G), and the crypt depth was significantly higher in both the 8 Gy and 8 Gy + AMI-1 groups than in the control group (Fig. 4H). All these results suggest that the suppression of PRMT5 could aggravate the epithelial damage of small intestine after RIII.

Fig. 4.

Inhibition of PRMT5 by AMI-1 exacerbates intestinal epithelial injury after RIII. A Heatmap for the transcriptome of PRMT family members in intestinal crypts after radiation injury (n = 3 each group). B Schematic diagram of whole abdominal irradiation (WAI) and AMI-1 treatment. C Statistical analysis of body weight changes after WAI and AMI-1 injections. D Representative gross image of the whole GI tract of mice in the control group, 8 Gy group and 8 Gy + AMI-1 group (n = 2 per group). E Statistical analysis for the length of small intestine and colon in different groups. F Representative H&E staining images of small intestines in the three groups (Up: Bar = 500 μm; Down: Bar = 100 μm). G Quantitative analysis for the villous length in different experimental groups. H Comparison for the depth of crypts in each group. At least 3 mice were used for analysis in each group for every comparison. ***: P < 0.001, ****: P < 0.0001, ns: P > 0.05

PRMT5 deficiency alters the number and function of secretory epithelial cells after RIII

We further investigated the functions of PRMT5 in RIII and its differential regulation compared to undamaged intestine. Similarly, we confirmed the inhibition of AMI-1 by IHC staining against PRMT5 and mRNA expression of its gene (Fig. 5A, B, Fig. S1C, Fig. S4). IHC against MUC2 and Alcian Blue staining (at varying pH levels) revealed that 8 Gy radiation exposure induced severe depletion of goblet cells, reducing their numbers to nearly undetectable levels. However, AMI-1 treatment significantly increased the number of goblet cells on the villi under the same radiation conditions (Fig. 5A, C, F, G). Following this, we similarly evaluated Tuft cells and enteroendocrine cells. While PRMT5 inhibition showed no significant effect on the abundance of Tuft cells, it induced a substantial increase of enteroendocrine cells (Fig. 5A, D, E). IF staining revealed that Paneth cells subjected to both irradiation and AMI-1 treatment exhibited more pronounced cellular enlargement compared to non-irradiated model, along with a significant increase in cell numbers (Fig. 5F, H, I, J). These results are consistent with our previous findings in the non-irradiated injury model, indicating comparable effects of PRMT5 inhibition on the morphology of Paneth cells. We conclude that PRMT5 would trigger multiple defensive responses and its inhibition disrupts normal secretory and protective functions of small intestine after RIII.

Fig. 5.

PRMT5 deficiency alters the number and function of secretory epithelial cells after RIII. A Representative IHC staining images of PRMT5, MUC2, DCLK1, and ChrgA between the 8 Gy group and 8 Gy + AMI-1group (Bar = 100 μm). B Statistical analysis for the AOD values of PRMT5. C Quantitative analysis for the number of MUC2⁺ goblet cells in villus. D Comparison for the amount of DCLK1⁺ Tuft cells per villus-crypt axis. E Statistical analysis of ChrgA⁺ EECs per villus-crypt axis between two groups. F Representative staining images of Alcian blue at two pH values and Lysozyme in the 8 Gy group and 8 Gy + AMI-1 group. (Bar = 100 μm and 50 μm as indicated). G Statistics for the number of goblet cells at different acidity between two groups. H Quantitative analysis for the number of PCs per crypt. I Statistical analysis for the size of PCs in the two groups. J Three-dimensional constructs of Lysozyme fluorescence intensity in PCs. There were at least 3 mice used in each group for every comparison. **: P < 0.01, ***: P < 0.001, ****: P < 0.0001, ns: P > 0.05

PRMT5 suppression aggravates loss of ISCs and enhances Paneth cell acidification after RIII

Subsequently, we assessed how PRMT5 suppression affected the cellular proliferation and stem cell activity in the circumstance of radiation. It was observed that the inhibition of PRMT5 by AMI-1 following radiation injury significantly reduced the number of Ki67⁺ and BrdU⁺ proliferating cells (Fig. 6A-C, Fig. S5). RIII usually caused a decreased quantity of Olfm4⁺ ISCs at day 3 after radiation exposure. Unexpectedly, PRMT5 inhibition after RIII led to a marked loss of Olfm4⁺ ISCs, approaching complete loss of Olfm4⁺ ISCs in certain crypts (Fig. 6A, D). Interestingly, the expression pattern of iNOS within crypts closely mirrored the IHC staining of Olfm4, indicating a potential association between iNOS level and stem cell maintenance. Specifically, iNOS expression was partially reduced after RIII. In combination with AMI-1 treatment, an obvious reduction of iNOS expression was observed in the crypts, while an increased iNOS expression was noted in the villi (Fig. 6A, E). As described above, H&E staining revealed the presence of numerous vacuole-like structures at the base of crypts after radiation and AMI-1 treatment (Fig. 4F). To further characterize these structures, we stained tissues with Alcian blue of different pH levels. Notably, more vacuoles were stained by Alcian blue at pH 2.5 compared to pH 1.0. Concurrently, IF staining against Lysozyme identified that these vacuolated cells were actually Paneth cells. Collectively, these findings demonstrate that suppressing the activity of PRMT5 after RIII leads to the enlargement and acidification of Paneth cells (Fig. 6F). PRMT5 inhibition alone could not induce any acidification in Paneth cells in physiology (Fig. 2F). However, some Paneth cells became Alcian blue-positive after RIII, particularly when stained at pH 2.5, suggesting an increased accumulation of acidic mucins or other acidic components. Importantly, PRMT5 inhibition by AMI-1 further enhanced Alcian blue staining and was accompanied by an increase in Paneth cell size, indicating alterations in their secretory profile. These changes may impair the ability of Paneth cells to support the reconstitution of the ISC niche. Collectively, these findings suggest that PRMT5 is required for proper epithelial regeneration after RIII.

Fig. 6.

PRMT5 suppression exacerbates loss of ISCs and promotes acidification in Paneth cells after irradiation injury. A IHC staining of Ki67, BrdU, Olfm4, and iNOS between the 8 Gy group and 8 Gy + AMI-1 group (Bar = 100 μm). B Statistical analysis for the number of Ki67⁺ cells within small intestinal crypts. C Quantification for BrdU⁺ proliferating cells within crypts. D Comparison for the number of Olfm4⁺ ISCs per group in two groups. E Statistical analysis of the percentage area of iNOS expression in the small intestinal crypt and villus regions in the 8 Gy and 8 Gy + AMI-1 groups. F Representative staining images of H&E, Alcian blue, and Lysozyme between the 8 Gy group and 8 Gy + AMI-1group (Bar = 50 μm). There were at least 3 mice used in each group for every comparison. **: P < 0.01, ***: P < 0.001, ****: P < 0.0001, ns: P > 0.05

PRMT5 deficiency limits the growth and regeneration of intestinal organoids

Intestinal organoid serves as an excellent ex vivo model to investigate the stem cell function, since it provides a specific three-dimensional architecture to simulate the structure of intestinal crypt and villus in vivo. ISCs and Paneth cells are the fundamental units responsible for initiating and sustaining the growth of intestinal organoids [22–24]. We treated intestinal organoids with different concentrations of AMI-1 at the beginning of growth. And it was found that PRMT5 deficiency caused a dose-dependent inhibition on the organoid-forming efficiency at Day 1 (Fig. 7A, Fig. S6). Subsequently, we quantified the number of buds per organoid on Day 3 and also observed AMI-1 treatment suppressed the budding capacity of intestinal organoids (Fig. 7A, B). On Day 5, the surface areas of intestinal organoids gradually decreased compared with the control group accompanied with the increasing of AMI-1 doses (Fig. 7A, C). Organoid growth was almost completely arrested at higher concentrations. In the 0.5 mM and 0.75 mM AMI-1 groups, only a few bud structures remained viable, whereas organoids treated with 1 mM AMI-1 were entirely dead at Day 5. To further investigate the importance of PRMT5 in mature organoids, we also treated mature organoids with the same pattern of AMI-1 and evaluated both the budding number and organoid area following treatment. It was shown that Ami-1 treatment was able to induce the breakdown of intestinal organoids with the increasing of doses and incubation time (Fig. 7D, E, F). Consistently, organoid-based analysis further revealed that PRMT5 inhibition could lead to a reduction in the population of EGFP-positive Lgr5⁺ ISCs, accompanied by an increase in UEA-1-positive secretory lineage cells, including Paneth and goblet cells (Fig. S7). These findings recapitulate the in vivo phenotypes and support a shift in epithelial lineage composition upon PRMT5 inhibition. In addition, we also irradiated the organoids with a single dose of 5 Gy. We found the organoids without PRMT5 inhibition gradually decreased their sizes after irradiation. Upon Ami-1 treatment, irradiated organoids exhibited more rapid progressive degeneration. Notably, AMI-1 treatment induced organoid disintegration within 24 h (Fig. 7G). Therefore, these results demonstrate that PRMT5 is indispensable for both the growth and regeneration of intestinal organoids.

Fig. 7.

PRMT5 inhibition limits the growth and regeneration of intestinal organoids. A Representative brightfield images of organoids derived from healthy crypts and treated with different doses of AMI-1 from Day 0 (Bar = 200 μm). B Quantification for the buds per organoids on Day 3. C Statistical analysis for the surface area of intestinal organoids on Day 5. D Representative images of mature intestinal organoids treated with AMI-1 after a 3-day growth (Bar = 200 μm). E Quantitative analysis for the number of buds per organoid on Day 1 after AMI-1 treatment. F Statistical results for the area of mature organoids on Day 2 after the administration of AMI-1. G Representative images of mature organoids after a 5 Gy radiation and treated with AMI-1 (Bar = 200 μm). *: P < 0.05, **: P < 0.01, ***: P < 0.001, ****: P < 0.0001, ns: P > 0.05

PRMT5 inhibition disrupts urea cycling and induces ammonia-driven death in IECs

We also used cell lines of intestinal epithelial cells (IECs) to explore the molecular mechanisms of PRMT5. Briefly, we treated IEC-6 cells with 0.75 mM AMI-1 and found the formation of vacuoles in their cytoplasm at 24 h and 48 h. The same phenomenon was also observed after irradiation-induced injury (Fig. 8A). We also confirmed these results in Caco-2 cells. Moreover, cell viability of both IEC-6 and Caco-2 was significantly decreased with the increasing doses of AMI-1 (Fig. S8). Using transmission electron microscopy (TEM), we found that AMI-1 induced PRMT5 inhibition led to an apparent damaged morphology of mitochondria, such as swelling, dissolution, fragmentation, vacuolization (red arrow), and even disappearance of cristae structure. Additionally, mild expansion of the endoplasmic reticulum and widening of the lumen were observed (Fig. 8B). Vacuole formation and mitochondrial swelling are characteristics of ammonium-induced cell death, which is a unique form of cell death that differs from apoptosis, ferroptosis, and necroptosis [25].

Fig. 8.

PRMT5 deficiency impairs urea cycling in intestinal epithelial cells and induces ammonia-driven cell death. A Representative morphology of normal and irradiated IEC-6 cells with or without 0.75 mM AMI-1 treatment (Bar = 50 μm). B Representative TEM images of IEC-6 cell in control and in the presence of 0.75 mM AMI-1 for 24 h. Red arrows show mitochondria (Bar = 500 nm). C Statistical charts show the values of Fragments Per Kilobase of transcript per Million (FPKM) mapped reads for enzymes involved in urea cycling (n = 3 in each group). D qPCR validation of Cps1 and Arg1 in crypts from different groups, including Ctrl, AMI-1, 8 Gy, and 8 Gy + AMI-1. E qPCR analysis of iNOS mRNA levels in the crypts of four groups: Ctrl, AMI-1, 8 Gy, and 8 Gy + AMI-1. F Statistical chart showing the ratio of NO production in four groups of IEC-6 cells. G Representative images of LD540-stained lipid droplets and ROS detected by DHE in normal and irradiated (15 Gy) IEC-6 cells treated by AMI-1 (0.75mM) for 24 h (Bar=: 100 μm–50 μm as shown). *: P < 0.05, **: P < 0.01, ***: P < 0.001, ****: P < 0.0001, ns: P > 0.05

RNA-Seq results revealed that enzymes related to urea cycling were expressed by intestinal crypts, including carbamoyl-phosphate synthetase 1 (Cps1), ornithine transcarbamylase (Otc), argininosuccinate synthetase 1 (Ass1), argininosuccinate lyase (Asl) and arginase 1 (Arg1). Although Arginase 2 is predominantly expressed in healthy intestine, Arginase 1 was increased while Arginase was down after irradiation injury, indicating radiation triggers the ammonia detoxification function of the urea cycle (Fig. 8C). Cps1 is the rate-limiting enzyme in the urea cycle, and Arg1 catalyzes the conversion of ammonia into urea. We found that the levels of Cps1 and Arg1 were reduced upon AMI-1-induced PRMT5 inhibition, suggesting a disruption of urea cycle homeostasis. Given that PRMT5 and urea cycle enzymes share arginine as a common substrate, these changes are likely secondary to altered arginine metabolism rather than direct regulation by PRMT5 (Fig. 8D). These results suggested that urea cycling was impaired after PRMT5 inhibition, which led to further accumulation of ammonia and arginine. Importantly, upon inhibition of PRMT5 expression, the mRNA levels of NOS2 were increased and accompanied by an elevation of NO content (Fig. 8E, F). Moreover, genetic silencing of PRMT5 by siRNA recapitulated these effects, resulting in similar alterations in Arg1 and iNOS expression, thereby supporting the specificity of PRMT5 inhibition (Fig. S9). Overproduction of NO can lead to alterations in lipid metabolism, resulting in accumulation of lipid droplets. Additionally, the increased NO may react with superoxide (O₂⁻) to form peroxynitrite (ONOO⁻), a potent oxidant that can induce cellular oxidative damage, leading to an increase in ROS production (Fig. 8G). Consistently, antioxidant treatment with NAC partially reversed the AMI-1-induced increase in ROS levels and alleviated the associated cellular phenotypes, supporting a functional role of ROS downstream of PRMT5 inhibition (Fig. S10). In summary, PRMT5 inhibition disrupts urea cycle in IECs and induced toxic ammonia accumulation and cell death. This process is accompanied by less ARG1, more iNOS, and increased NO production, which triggers lipid peroxidation and cellular damage (Fig. 9).

Fig. 9.

Mechanistic overview of PRMT5 regulation of intestinal stem cell niche in homeostasis and after RIII

Discussion

PRMT5 is an important member of PRMT family, which is highly expressed in various types of cancers. Although it has long been considered a potential therapeutic target, often studied as an oncogenic factor [10, 14, 26–29], there are emerging reports suggesting PRMT5 plays multiple roles in tumorigenesis and other diseases, including tumor-suppressive effects [15, 16]. In the current study, we focused our attention on the role of PRMT5 in GI tract, especially the small intestine. Using transcriptome sequencing analysis of intestinal crypts, we discover that PRMT5 is also highly expressed in physiological intestinal crypts. Furthermore, its expression would be increased after RIII. Given these observations, we hypothesize that PRMT5 plays an indispensable role in both homeostasis and reconstitution of the ISC niche.

In this study, we treated mice with AMI-1, which is the specific inhibitor of PRMT5. Intriguingly, we noticed that PRMT5 deficiency made small intestine become shorter with an obvious abnormality in secretory lineage differentiation of IECs. It also significantly increased the quantity of Olfm4⁺ ISCs and expanded both the number and size of their surrounding Paneth cells. Notably, with the inhibition of PRMT5, iNOS expression was shifted from Paneth cells to ISCs (Fig. 3A, D, E, Fig S2). PRMT5 inhibition exacerbated intestinal damage after RIII by a dramatic loss of ISC and further expansion and acidification of Paneth cells (Figs. 4, 5 and 6). We further validated the role of PRMT5 using intestinal organoids and also found that PRMT5 is indispensable to the growth and maturation of healthy organoids and the repairing after radiation injury (Fig. 7). These results show novel experimental evidence that PRMT5 is required for the homeostasis and reconstitution of ISC niche.

Urea cycling is an important physiological function in mammals, which converts toxic ammonia into non-toxic urea [30, 31]. The previous studies show that urea cycling primarily takes place in the liver [32]. However, the major source of ammonia in human body is small intestine, because ammonia could be originating from diet, luminal bacterial activity, and deamination of glutamine in IECs [33, 34]. According to the reports, in the liver, Arginase 1 plays the dominant role in urea cycling, while other organs predominantly express Arginase 2 [35]. Our transcriptomic analysis of small intestinal crypts revealed that Arginase 1 expression was lower than that of Arginase 2 under normal physiological condition. However, after radiation injury, the expression of Arginase 1 became surged, and Arginase 2 was dramatically decreased. Additionally, all the urea cycling related enzymes exhibited a basal expression (Fig. 8A-C). We proposed that when ammonia levels increase following radiation-induced damage, intestinal epithelial cells may initiate their own urea cycling to detoxify the overproduction of ammonia. However, when the activity of PRMT5 was blocked by AMI-1, the rate-limiting enzyme of the urea cycle (CPS1) and the key transaminase (ARG1) were both downregulated, indicating the disruption of urea cycling. These changes resulted in more accumulation of ammonia and arginine. Interestingly, AMI-1 treatment in physiology and after RIII also enhanced the expression of iNOS and led to an increased content of NO in IECs, thereby inducing the accumulation of lipid droplets and the production of ROS (Fig. 8D-G). All these molecular mechanisms ultimately in turn affects the function of ISCs and Paneth cells and ultimately exacerbate intestinal epithelial damage.

As the field of PRMT5 research moves from understanding its role in oncogenesis to gradually uncovering its tumor-suppressive effects, it has been identified PRMT5 as a tumor suppressor in gastric cancer, proposing that the heterogeneity in cancer onset and progression arises from the fact that PRMT5, as a major type II arginine methyltransferase, targets different substrates in different circumstances [11, 15, 36]. The findings in this study show global knockout of PRMT5 could lead to mouse lethality, suggesting PRMT5 might regulate the function of ISCs and their niche, which inspired our research. Subsequently, Hernandez and colleagues investigated the role of PRMT5 in colonic tissue, and they proposed that when the integrity of mucosal barrier was disrupted, PRMT5 might be upregulated as a protective mechanism to promote the production of antimicrobial peptides and facilitate mucosal barrier repairing [16]. This finding aligns with our results showing a role for goblet cells in mucosal defense in the small intestine. However, it is important to note the structural differences between small intestine and colon, and especially there is the absence of Paneth cells in the colon. Our study firstly expands on this by examining the acidification of Paneth cells, revealing an additional role for PRMT5 in the ISC niche of small intestine. Secretory cell acidification not only supports antimicrobial defense but also reflects secretory cell differentiation, vesicle maturation, and overall cellular metabolic homeostasis. Disruption of acidification can thus impact lysosomal function, mitochondrial stress, and cell viability beyond antimicrobial activity [37, 38].

In addition, we explored the underlying molecular mechanisms of PRMT5 deficiency-induced changes. It was demonstrated that PRMT5 deficiency disrupted the urea cycling in intestinal epithelial cells, which led to the accumulation of toxic ammonia and disrupted arginine metabolism in IECs, consistent with previous findings that PRMT5 is essential for maintaining intestinal epithelial homeostasis [39]. We found that AMI-1 induced PRMT5 inhibition could result in an increasing of iNOS expression, which resulted in an elevated NO production, ultimately impairing ISCs intestinal health. Now we still believe that the majority of the urea cycle occurs in the liver, which remains the key organ for ammonia detoxification. However, our study firstly proves that PRMT5 mediated urea cycling is also of great importance for the homeostasis and reconstitution of ISCs and their niche. Additionally, liver diseases caused by the dysfunction of GI tract are of significant practical relevance, so urea cycling in intestine is definitely worthy of further investigation. Moreover, it is also notable that ammonia is a neurotoxic substance [34], and the increased ammonia levels in blood can lead to an elevated ammonia concentration in the brain, causing astrocyte swelling, increased blood-brain barrier permeability, altered brain metabolism and neurotransmission, and brain edema [40–42]. Ammonia also impairs the potassium buffering function of astrocytes, ultimately leading to seizures [43]. This indicates that the function of PRMT5 and urea cycling in intestines, besides liver, may provide a new therapeutic target for treating epilepsy or other neurodegenerative diseases through the regulation of blood ammonia levels during the interaction between gut and brain, which is an exciting avenue worth of further exploration.

In conclusion, here we demonstrate that PRMT5 is capable to orchestrate ISCs and their niche in both physiology and radiation-induced injury of small intestine for the first time, and this study also elucidates the potential underlying molecular mechanisms involved during PRMT5 deficiency. These findings highlight the potential risks associated with systemic PRMT5 inhibition in cancer treatment. Since radiotherapy is still the conventional strategy used for abdominal and pelvic cancers, PRMT5 and urea cycling should be paid more attention during the management of radiation induced enteritis. It also inspires us to reconsider the crosstalk between gut and other organs due to urea cycling. Overall, these findings highlight the multifaceted roles of PRMT5 and suggest new therapeutic strategies for intestinal, liver, and brain-related disorders.

Abbreviations

AMI

1 Arginine N-methyltransferase inhibitor 1

AOD

Average optical density

Arg1

Arginase 1

Asl

Argininosuccinate lyase

Ass1

Argininosuccinate synthetase 1

ChrgA

Chromogranin A

Cps1

Carbamoyl-phosphate synthetase 1

DCLK1

Doublecortin like kinase 1

DHE

Dihydroethidium

EECs

Enteroendocrine cells

FPKM

Fragments Per Kilobase of transcript per Million

ER

Endoplasmic reticulum

GI

Gastrointestinal

Gy

Gray

H&E

Hematoxylin-eosin

IECs

Intestinal epithelial cells

IF

Immunofluorescent

IHC

Immunohistochemistry

iNOS

Inducible nitric oxide synthase

i.p

Intraperitoneal

ISCs

Intestinal stem cells

MEP50

Methylated histone 50

MUC2

Mucin2

NIH

National Institutes of Health

NO

Nitric Oxide

Otc

Ornithine transcarbamylase

PCs

Paneth cells

PR

Protein arginine methyltransferase 5

PTM

Post transcriptional modification

RIII

Radiation-induced intestinal injury

RNA-Seq

RNA sequencing

ROS

Reactive oxygen species

SPF

Specific pathogen-free

TEM

Transmission electron microscopy

WAI

Whole abdominal irradiation

Authors' contributions

D.L and N.W conceived this study. N.W, X.W and Q.L performed the experiments and data collection. Y.L, H.T and B.L helped with tissue preparation and histological analysis. Y.H, Z.X, X.L and L.H helped with data analysis and interpretation. The manuscript was written and revised by N.W and D.L. All the authors have read and approved this manuscript.

Funding support

This work is supported by National Natural Science Foundation of China (82570643, 81874255, 81673089), Science and Technology Department of Sichuan Province (2025NSFJQ0058).

Data availability

The data are available depending on reasonable requests to the corresponding author.

Declarations

Ethical statement

This study has been approved by the Ethics Committee of Sichuan Cancer Hospital (SCCHEC-04-2024-036).

Competing interests

The authors declare no competing interests.