Abstract

   Diabetes, a metabolic disorder characterized by hyperglycemia,
   underscores the importance of normal pancreatic β‐cell development and
   function in maintaining glucose homeostasis. Poly(A)‐specific
   ribonuclease (PARN) serves as the principal regulator of messenger RNA
   (mRNA) stability, yet its specific role in pancreatic β cells remains
   unclear. This study utilizes mice with targeted PARN deficiency in β
   cells to elucidate this role. Notably, Parn conditional knockout mice
   present unaltered β‐cell development and insulin sensitivity but
   reduced glucose‐stimulated insulin secretion (GSIS). The observed
   outcomes are corroborated in NIT‐1 cells. Furthermore, transcriptomic
   analyses reveal aberrant mRNA expression of genes crucial for insulin
   secretion in PARN‐deficient β cells. Insights from linear amplification
   of complementary DNA ends and sequencing and coimmunoprecipitation
   experiments reveal an interaction between PARN and polypyrimidine
   tract‐binding protein 1 (PTBP1), regulating the RNA stability of solute
   carrier family 30, member 8 (Slc30a8) and carbohydrate sulfotransferase
   3 (Chst3). Interference with either PARN or PTBP1 disrupts this
   stability. These data indicate that PARN deficiency hampers GSIS and
   insulin maturation by destabilizing Slc30a8 and Chst3 RNAs. These
   findings provide compelling evidence indicating that PARN is a
   potential therapeutic target for enhancing insulin maturation and
   secretion.

   Keywords: β cell, insulin maturation and secretion, RNA binding protein
     __________________________________________________________________

   This study elucidates that poly(A)‐specific ribonuclease (PARN)
   interacts with polypyrimidine tract‐binding protein 1 to directly bind
   and maintain the messenger RNA stability of solute carrier family 30,
   member 8 and carbohydrate sulfotransferase 3in β cells, thereby
   regulating insulin maturation and secretion. It is suggested that PARN
   is a potential therapeutic target for diabetes.

   graphic file with name ADVS-11-2407774-g004.jpg

1. Introduction

   Diabetes manifests as a multifaceted metabolic disorder characterized
   by chronic hyperglycemia and is typically categorized into gestational
   diabetes, CNOT^[ [46]^1 ^] type 1 diabetes mellitus,^[ [47]^2 , [48]^3
   ^] type 2 diabetes mellitus (T2DM),^[ [49]^4 ^] or other specific
   types.^[ [50]^5 ^] The etiology involves disruptions in insulin
   secretion, abnormal insulin activity, or a combination of both
   factors.^[ [51]^6 ^] Mature insulin, comprising a 30‐residue B chain
   and a 21‐residue A chain interconnected by two disulfide bonds, plays a
   pivotal role in glucose regulation.^[ [52]^7 ^] Islets constitute ≈1–2%
   of the pancreatic volume, and house β cells, which account for 60–70%
   of the islet cell population.^[ [53]^8 ^] The synthesis and secretion
   of insulin by β cells are essential for sustaining normal metabolic
   processes in organisms.

   The synthesis of insulin in response to glucose involves intricate
   regulation at multiple levels, including transcriptional,
   translational, and posttranslational processes.^[ [54]^9 ^] RNA‐binding
   proteins (RBPs), a diverse group of proteins with various structures
   and functions, play crucial roles in the posttranscriptional control of
   both mRNAs and noncoding RNAs. Their influence extends to RNA splicing,
   transport, modification, stability, and translation.^[ [55]^10 ^]
   Notably, among posttranscriptional regulators are RBPs such as ELAV
   like RNA binding protein 4 (HuD), protein disulfide isomerase (PDI),
   and polypyrimidine tract‐binding protein 1 (PTBP1), which were
   identified for their roles in insulin expression by binding to the
   untranslated region (UTR) of insulin mRNA.^[ [56]^11 ^] The cis‐acting
   sequence element in the mRNA UTR collaborates with trans‐acting
   factors, including microRNAs and RBPs, to dynamically modulate the
   translation of β cells in response to external stimuli, thereby
   contributing to glucose homeostasis. Key regulators such as CUG‐binding
   protein 1 activate phosphodiesterase 3B (PDE3B), influencing insulin
   secretion,^[ [57]^12 ^] whereas DEAD box 1 (DDX1)‐mediated inhibition
   of insulin translation exacerbates insulin resistance in individuals
   with obesity, leading to elevated blood glucose levels.^[ [58]^13 ^]
   These findings underscore the essential role of RBPs in maintaining the
   functionality of mouse β cells.

   Within eukaryotic cells, the equilibrium between mRNA synthesis and
   degradation is pivotal for mRNA homeostasis. Deadenylation, the removal
   of adenosine residues from the poly(A) tail, represents the primary and
   critical step in the mRNA degradation pathway.^[ [59]^14 ^] In mammals,
   the carbon catabolite repression 4 (CCR4)‐negative on TATA‐less (NOT)
   complex serves as the principal deadenylase, with CNOT1 being a key
   catalytic subunit.^[ [60]^15 ^] Studies have indicated that mice
   deficient in CNOT1 exhibit diminished quality white adipose tissue and
   brown adipose tissue, underscoring the role of the CCR4‐NOT complex in
   regulating gene expression and the subsequent impacts on insulin
   sensitivity, lipid metabolism, and body temperature control.^[ [61]^16
   ^] However, the influence of deadenylases on insulin synthesis and
   secretion remains unclear. Poly(A)‐specific ribonuclease (PARN), a
   prominent mammalian deadenylase, uniquely binds both the 5′ cap
   structure and 3′ poly(A) tail, increasing the mRNA degradation rate and
   processability.^[ [62]^17 ^] Consequently, our investigation sought to
   elucidate whether PARN plays a discernible role in the processes of
   insulin synthesis or secretion.

   In our current investigation, we initially observed a reduction in
   insulin secretion in Parn‐knockdown NIT‐1 cells subjected to
   high‐glucose conditions. We subsequently generated mice with
   β‐cell‐specific Parn knockout to systematically assess the role of PARN
   in β‐cell function. Intriguingly, the absence of PARN did not yield
   discernible alterations in islet morphology, β‐cell mass, islet size,
   or insulin sensitivity in the murine model. However, it manifested as
   glucose intolerance, insulin immaturity, and a noteworthy reduction in
   glucose‐stimulated insulin secretion (GSIS). These findings underscore
   the specific impact of PARN deficiency on aspects of glucose metabolism
   and insulin secretion without overtly affecting key structural and
   sensitivity parameters in the β cells of mice.

2. Results

2.1. PARN Inhibition Impairs Glucose‐Stimulated Insulin Secretion in NIT‐1
Cells

   We initially examined the expression of PARN in NIT‐1 cells through
   immunofluorescence (IF) to assess the role of PARN in β cells, and the
   results revealed that PARN was expressed mainly in the nucleus (Figure
   [63]1A). We subsequently confirmed a substantial reduction in Parn mRNA
   (Figure [64]S1A, Supporting Information) and protein expression levels
   upon the transfection of Parn small interfering RNAs (siRNAs) into
   NIT‐1 cells, with siRNA1 generating the most pronounced effect on gene
   expression (Figure [65]1B). Notably, there was no difference in insulin
   secretion between the PARN‐deficient NIT‐1 cells and the negative
   control (NC) cells in the presence of low glucose, but after 1 h in the
   presence of high glucose, insulin secretion was significantly reduced.
   When the background under low‐glucose conditions was subtracted from
   the high‐glucose samples, insulin secretion was decreased by ≈75%
   compared with that of the NC group (Figure [66]1C). Additionally, we
   performed insulin secretion experiments with β‐TC6 cells, and the
   results revealed that Parn knockdown in β‐TC6 cells did not affect
   basal insulin secretion, but GSIS decreased (Figure [67]1D,E), which
   was consistent with the results in NIT‐1 cells. These findings
   underscored the functional involvement of PARN in the process of GSIS.

Figure 1.

   Figure 1
   [68]Open in a new tab

   PARN inhibition impairs glucose‐stimulated insulin secretion in NIT‐1
   and β‐TC6 cells and PARN was successfully selectively deleted in β
   cell. A) Immunofluorescence double staining NIT‐1 cells with antibodies
   against insulin (red), PARN (green), nuclei were stained with
   4',6‐diamidino‐2‐phenylindole (DAPI) (blue). B) Western blotting
   analysis was performed to detect the expression level of PARN protein
   in NIT‐cells transfected with siRNA 48 h, n = 3–4. C) Insulin ELISA kit
   was used to detect the insulin secretion of NIT‐cells transfected with
   Parn siRNA1, n=3. D) Western blotting analysis was performed to detect
   the expression level of PARN protein in β‐TC6 cells transfected with
   siRNA 48 h, n = 3. E) Insulin ELISA kit was used to detect the insulin
   secretion of β‐TC6 cells transfected with Parn siRNA2, n = 3. F)
   Western blotting was used to analyze protein expression in islets and
   acinar cells, quantitative data from three independent experiments. G)
   2 months pancreatic sections were subjected to co‐IF staining using
   antibodies against insulin (red), PARN (green), nuclei were stained
   with DAPI (blue). H) Knockout scheme design in mice and polymerase
   chain reaction, the DNA templates came from mice tails. I)
   Co‐immunofluorescent images of pancreatic sectioned islets from 18
   weeks old Ctrl and cKO mice showing PARN (green), insulin (red),
   glucagon (red), and nuclei were stained with DAPI (blue), bar = 25 µm.
   ns p ≥ 0.05, *p < 0.05, ***p < 0.001, ****p < 0.0001, Student's t‐test
   and ANOVA.

2.2. PARN Was Specifically Knocked Out in Mouse β Cells

   We explored the functional role of PARN in mouse β cells by evaluating
   its expression across various mouse tissues through Western blotting.
   The results revealed increased expression of PARN in the mouse
   pancreas, with particularly elevated protein levels observed in the
   islets (Figure [69]S1B (Supporting Information) and Figure [70]1F).
   Further characterization of PARN expression patterns involved IF
   costaining of PARN and insulin, which revealed robust expression of
   PARN in β cells at different time points (Figure [71]S1C (Supporting
   Information) and Figure [72]1G). These findings suggest a potential
   role for PARN in β cells.

   We generated mice with specific Parn knockout in β cells by
   establishing insulin 2 (Ins2)‐cre; Parn^f/f mice through crosses of
   loxp‐flanked Parn ^f/f mice with Ins2‐cre mice. These mice were
   designated Parn‐specific knockout mice in β cells (cKO), whereas Parn
   ^f/f mice and other littermates served as controls (Ctrl)
   (Figure [73]1H). The negative IF costaining of PARN and insulin and the
   positive one with glucagon (Figure [74]1l). These results indicated the
   successful construction of a mouse model featuring specific Parn
   knockout in β cells.

2.3. Specific Knockout of Parn in β Cells Is Associated with Glucose
Intolerance but Not with Insulin Resistance

   To evaluate the impact of β‐cell‐specific Parn deletion in mice, we
   assessed blood glucose levels in mice fed with a normal diet (ND).
   Initially, fasting and ad libitum blood glucose levels were comparable
   between the cKO and Ctrl mice (Figure [75]2A,B). However, glucose
   tolerance was significantly impaired in the cKO mice at 12 weeks of
   age, in Ins2‐cre; Parn ^f/+ mice, glucose levels did not differ from
   those in Parn ^f/f mice (Figure [76]2C). This finding is consistent
   with previous reports indicating glucose intolerance in Ins2‐cre
   mice.^[ [77]^18 ^] Similarly, at 16 weeks of age, cKO mice presented
   significantly higher blood glucose concentrations than Ctrl mice
   (Figure [78]2E). Importantly, the loss of PARN in β cells did not
   influence insulin sensitivity, as evidenced by the insulin tolerance
   test (ITTs) conducted at 13 and 17 weeks of age (Figure [79]2D,F).
   Similar findings were observed in female Parn cKO mice (Figure [80]S2,
   Supporting Information). These results illustrated that β‐cell‐specific
   Parn knockout induced glucose intolerance without affecting insulin
   sensitivity in ND‐fed mice.

Figure 2.

   Figure 2
   [81]Open in a new tab

   β‐cell specific knockout Parn induces glucose intolerance but not
   insulin sensibility. A) Blood glucose concentrations in fasting and
   free‐feeding ND male Ctrl and cKO mice, n = 3. B) Blood glucose
   concentrations in fasting and free‐feeding ND female Ctrl and cKO mice
   under normal feeding conditions, n = 3. C,E) Blood glucose excursions
   during a 2 g kg^−1 glucose tolerance test of (C) 12, (E) 16 weeks male
   mice and the area under the curve (AUC), n = 3–5. D,F) Insulin
   tolerance test of (D) 13, (F) 17 weeks male mice and the AUC, n = 3–6.
   **p < 0.01, *** p < 0.001, ****p < 0.0001, Student's t‐test and ANOVA.

2.4. PARN Is Not Required for β‐Cell Development

   We subsequently explored whether the observed impaired glucose
   tolerance in cKO mice stemmed from alterations in β‐cell mass and islet
   size. IF staining of pancreatic sections from cKO and Ctrl mice for
   insulin and glucagon revealed that Parn deletion did not influence the
   distribution of α or β cells within the islets (Figure [82]3A). At 3
   days, the β‐cell mass was not significantly different between the cKO
   and Ctrl mice (Figure [83]3B). Comparable body weights (Figure [84]3C)
   and pancreatic weights (Figure [85]3D) were compared between the two
   groups at 18 weeks. Furthermore, at 18 weeks, pancreatic sections
   subjected to insulin immunohistochemistry (IHC) staining displayed no
   significant differences in β‐cell mass (Figure [86]3E,F) or the
   distribution of islet size (Figure [87]3G) between cKO and Ctrl mice. A
   parallel phenotype was observed in female mice (Figure [88]S3A–F,
   Supporting Information). These findings suggested that Parn‐specific
   deletion in β cells had no discernible effect on the morphology or
   development of islets.

Figure 3.

   Figure 3
   [89]Open in a new tab

   Selective deletion of Parn in pancreatic β cells leads to impaired
   glucose‐stimulated insulin secretion. A) Immunofluorescent images of
   pancreatic sectioned from 18 weeks Ctrl and cKO mice showing insulin
   (red, left), glucagon (red, right), and nuclei (DAPI, blue), bar =
   10 µm. B) Representative images of IHC pancreatic sections and the
   statistical results of β‐cell mass at P3 Ctrl and cKO mice, bar =
   100 µm, n = 3. C) Body weight from 2 to 18 weeks Ctrl and cKO mice, n =
   4. D) Pancreas weight at 18 weeks Ctrl and cKO mice, n = 3. E) IHC of
   insulin pancreatic sections from 18 weeks old Ctrl and cKO mice, bar =
   100 µm. F) β‐cell mass was calculated in 18 weeks Ctrl and cKO mice, n
   = 3. G) Islet size distribution as stratified by ranges of islet area
   from 18 weeks Ctrl and cKO mice by ImageJ, n = 3. H) Serum insulin
   level at fasting, 10, 20, 30, and 60 min after glucose injection from
   18 weeks old male and female Ctrl and cKO mice, n = 3–4. I) The
   transmission electron microscopy images are from 18 weeks old Ctrl and
   cKO mice, with arrows indicating mature insulin secretion granules,
   arrow heads indicating immature insulin secretion granules, arrows
   indicating empty granules, bar = 2 µm. J) The number of total insulin
   granules calculated from the vesicle area, n = 5. K) The percentage of
   dense‐core, gray‐core, and empty insulin granules calculated from the
   vesicle area, n = 5. L) The insulin and proinsulin protein level of cKO
   and Ctrl mice was detected by Western blotting, n = 3. M) Insulin
   secretion assay on islet from 18 weeks old Ctrl and cKO male mice with
   3 and 20 mm glucose in vitro. Insulin secretion/intracellular insulin
   content ratio, bar = 20 µm, n = 3. *p < 0.05, ****p < 0.0001, Student's
   t‐test and ANOVA.

2.5. Loss of PARN in β Cells Causes GSIS Deficiency

   We investigated insulin secretion through a GSIS assay to elucidate the
   mechanism underlying impaired glucose tolerance in cKO mice with β‐cell
   Parn deficiency. Enzyme‐linked immunosorbent assays (ELISAs) revealed
   no significant difference in the serum insulin concentration between
   the Ctrl and cKO mice after fasting, but both the female and male cKO
   mice presented reduced serum insulin levels in response to glucose in
   vivo (Figure [90]3H). The intracellular islet insulin concentration was
   assessed via transmission electron microscopy (TEM) (Figure [91]3I) and
   quantified by analyzing insulin granules with ImageJ. The total number
   of insulin‐secreting granules in the cKO mice did not decrease
   (Figure [92]3J), but a decrease in the number of mature
   insulin‐secreting granules accompanied by an increase in the number of
   immature insulin‐secreting granules was observed (Figure [93]3K). The
   islets of cKO and Ctrl mice were isolated to analyze the expression
   levels of insulin and proinsulin. The level of insulin protein in the
   islets of cKO mice significantly decreased, whereas the expression of
   proinsulin protein increased significantly (Figure [94]3L). These
   results were consistent with our TEM observations, providing further
   evidence for the presence of insulin maturation disorders in cKO mice.

   Additionally, islet cells that had been isolated and cultured in vitro
   presented a decreased ratio of secreted insulin to intracellular
   insulin in the presence of 20 mm glucose (Figure [95]S3G (Supporting
   Information) and Figure [96]3M). These findings suggested that the
   absence of PARN in β cells resulted in diminished GSIS and impaired
   insulin maturation.

2.6. Dysfunction of GSIS in High‐Fat‐Diet (HFD)‐Fed cKO Mice

   Obesity, a critical risk factor for diabetes, contributes to metabolic
   disorders.^[ [97]^19 ^] We fed both Ctrl and cKO mice a HFD from 4 to
   18 weeks of age to investigate the impact of a metabolic challenge on
   β‐cell Parn deficiency. The fed and fasting blood glucose levels were
   similar between the Ctrl and cKO mice (Figure [98]4A,B). Nevertheless,
   at 12 and 16 weeks of age, cKO mice displayed glucose intolerance
   (Figure [99]4C,E). ITTs revealed no discernible differences in blood
   glucose levels between cKO and Ctrl mice (Figure [100]4D,F), indicating
   that impaired systemic glucose homeostasis in cKO mice was unrelated to
   insulin sensitivity. We compared body weights between cKO and Ctrl mice
   to explore the potential causes of glucose intolerance and found no
   significant differences (Figure [101]4G). There were no significant
   differences in pancreatic weight, β‐cell mass, or islet size
   distribution, as assessed by insulin staining of pancreatic sections,
   between the cKO and Ctrl mice (Figure [102]4H–K). Serum insulin
   measurements further supported these findings, with cKO mice showing
   consistent fasting insulin levels but significantly lower insulin
   levels than Ctrl mice following glucose stimulation (Figure [103]4L).
   Notably, these observations were also observed in female mice (Figure
   [104]S4, Supporting Information). Collectively, these results suggested
   that HFD‐fed cKO mice maintained normal β‐cell development and that
   their impaired glucose homeostasis in response to glucose stimulation
   was attributed primarily to a substantial reduction in insulin
   secretion. Importantly, the cKO mouse phenotype was not exacerbated by
   obesity.

Figure 4.

   Figure 4
   [105]Open in a new tab

   Impairment of glucose‐induced insulin secretion in β‐cell‐specific
   Parn‐knockout mice in high fat diet. A,B) Blood glucose level in
   fasting and free‐feeding male (A) mice and female (B) HFD Ctrl and cKO
   mice, n = 3–4. C,E) Blood glucose levels during IPGTT in 12 weeks (C)
   and 16 weeks (E) Ctrl and cKO mice, with calculated AUCs, n = 4–6. D,F)
   Blood glucose levels during ITT in 13 weeks (D) and 17 weeks (F) Ctrl
   and cKO mice and the AUC, n = 4–6. G) Body weight of Ctrl and cKO mice
   during 15 weeks of HFD, n = 3–6. H) Pancreas weight at 18 weeks of age,
   n = 3. I) IHC pancreatic sections from 18 weeks Ctrl and cKO mice, bar
   = 100 µm. J) β‐cell mass in entire pancreas from 18 weeks Ctrl and cKO
   mice, n = 3. K) Islet size distribution as stratified by ranges of
   islet area from 18 weeks Ctrl and cKO mice by ImageJ, n = 3. L) Serum
   insulin was measured in 18 weeks old Ctrl and cKO mice at fasting and
   20 min after glucose injection, n = 3–4. *p < 0.05, **p < 0.01, ***p <
   0.001, ****p < 0.0001, Student's t‐test and ANOVA.

2.7. Loss of PARN in β Cells Results in Decreased Expression Levels of Genes
That Affect Insulin Secretion

   We conducted RNA‐seq profiling of islets from Ctrl and cKO mice to
   elucidate the underlying mechanisms by which PARN influences insulin
   secretion. An analysis of differentially expressed genes (DEGs)
   revealed robust repeatability across all three replicates within the
   same genotypes (Figure [106]5A). Parn mRNA expression was notably
   reduced, and the expression of 878 genes, such as synaptoporin and
   nitric oxide synthase 1, was upregulated, whereas the expression of 681
   genes, such as proprotein convertase subtilisin/kexin type 9 and
   visinin‐like protein‐1 (Vsnl1), was downregulated in cKO islets
   (Figure [107]5B,C). Hierarchical clustering of DEGs between Ctrl and
   cKO mice is visually represented in a heat map (Figure [108]5D). For
   deeper insights, we performed gene ontology (GO) and Kyoto Encyclopedia
   of Genes and Genomes (KEGG) enrichment analyses, both of which
   consistently highlighted insulin secretion (Figure [109]5E,F). A more
   granular examination of Vsnl1, 3‐hydroxy‐3‐methylglutaryl‐CoA reductase
   (Hmgcr), high‐mobility group N 3 (Hmgn3), urocortin 3 (Ucn3),
   synaptotagmin‐like 4 (Sytl4), glutamate‐ammonia ligase (Glul),
   regulatory factor X, 6 (Rfx6), and neuronatin (Nnat), which are related
   to insulin secretion genes, revealed the expression levels of these
   genes in each group (Figure [110]5G). Quantitative real‐time polymerase
   chain reaction (qRT‐PCR) confirmed the decreased expression of Vsnl1,
   Hmgcr, Hmgn3, and Nnat in the islets of cKO mice (Figure [111]5H), and
   their abundance mirrored the observed trends (Figure [112]5I). These
   findings indicated that Parn deletion disrupted the expression of
   pivotal genes implicated in insulin secretion.

Figure 5.

   Figure 5
   [113]Open in a new tab

   Loss of PARN in β cells results in decreased expression levels of genes
   that affect insulin secretion. A) The scatter plot shows the RNA‐seq
   Pearson correlation. B) Integrative Genomics Viewer (IGV) visualizes
   Parn reads values in cKO and Ctrl groups. C) Volcanic map showing the
   expression of the 878 genes up and 681 genes downregulated in cKO
   islet. D) Heat map showing differentially expressed genes in cKO islet
   at 12 weeks old. E) List of KEGG pathway enrichment analysis. F)
   Selected molecular functions enriched in different (DE) genes,
   identified by Gene Ontology analysis. G) Heat map of selected DE genes
   related to insulin secretion. H) The mRNA expression level of
   insulin‐secretion‐related genes was verified by qRT‐PCR, n = 3. I)
   Representative IGV browser tracks of gene peaks. For each panel, blue
   tracks are from Ctrl mice, red tracks are from cKO mice. *p < 0.05, **p
   < 0.01, Student's t‐test and ANOVA.

2.8. PARN Interacts with PTBP1 to Regulate Genes Related to Insulin
Maturation and Secretion

   Given the status of PARN as a RBP, we investigated its mechanism of
   action in islets through linear amplification of complementary DNA ends
   and sequencing (LACE‐seq)^[ [114]^20 ^] (Figure [115]6A), focusing on
   islets from 2 months old mice. The two biological replicates within the
   PARN group exhibited robust correlations with the detected putative
   binding target RNA molecules, attaining a Pearson coefficient of up to
   0.957 (Figure [116]6B). The most highly enriched motif among these PARN
   peaks was AA (Figure [117]6C). A subsequent GO analysis revealed that
   the PARN‐binding genes were intricately linked to insulin secretion and
   T2DM (Figure [118]6D). Integration of the RNA‐seq data with
   LACE‐seq‐identified peaks revealed 30 downregulated and 40 upregulated
   transcripts in cKO mice, elucidating direct targets of PARN in islets
   (Figure [119]6E and Figure [120]S5A (Supporting Information)). Further
   scrutiny and validation of these genes revealed direct binding of PARN
   to solute carrier family 30, member 8 (Slc30a8) and carbohydrate
   sulfotransferase 3 (Chst3), both of which were downregulated in cKO
   mice (Figure [121]6F,G and Figure [122]S5B (Supporting Information)).
   The results of the Actinomycin D (ActD) experiment indicated that the
   half‐lives of Slc30a8 and Chst3 mRNAs were shortened after Parn
   knockdown. Depletion of Parn in NIT‐1 cells led to a significant
   increase in the degradation of Slc30a8 and Chst3 mRNAs
   (Figure [123]6H,I). To demonstrate the direct binding of PARN to
   Slc30a8 and Chst3, fragments of Slc30a8 and Chst3 and their
   corresponding mutant fragments were amplified and connected to the
   psi‐Check2 vector. These constructs were then cotransfected with
   pcDNA3.0‐NC and pcDNA3.0‐PARN into 293T cells. The results of the
   luciferase assay confirmed the direct binding of PARN to Slc30a8 and
   Chst3 (Figure [124]6J and Figure [125]S5B (Supporting Information)).

Figure 6.

   Figure 6
   [126]Open in a new tab

   PARN interacting with PTBP1 to regulate the expression of genes during
   insulin maturation and secretion. A) Schematic of the LACE‐seq method.
   A circled B represents biotin modification. N, random nucleotide; V can
   be A, G, or C. B) Pearson correlation between PARN LACE‐seq replicates
   in total islet for assessing the reproducibility of the data. Pearson
   correlation for the reads counts of each sample was calculated from two
   replicates. C) PARN‐binding motifs identified by LACE‐seq in mouse
   islet. D) Network showing KEGG enrichment analyses of differentially
   expressed genes. E) Venn diagram showing the overlapped genes between
   differentially expressed genes and PARN‐binding genes. F) The heat map
   showed the expression of genes involved in (E) which were related to
   insulin secretion. G) Relative mRNA expression of Chst3 and Slc30a8, n
   = 3. H) The expression of Slc30a8 in NIT‐1 cells after transfection
   with Parn siRNA and ActD treatment at different times, n  =  3. I) The
   expression of Chst3 in NIT‐1 cells after transfection with Parn siRNA
   and ActD treatment at different times, n  =  3. J) The relative signal
   intensity of Renilla was analyzed in the intracellular luciferase
   reporting system of HEK293T cells. K) Western blotting analysis was
   performed to detect the expression level of PARN and SLC30A8 (Chst3)
   protein in NIT‐1 cells transfected with siRNA and plasmid 48 h. L)
   Insulin ELISA kit was used to detect the insulin secretion of NIT‐1
   cells transfected with Parn siRNA2, n = 4. M) Pearson correlation plot
   between PARN co‐IP replicates in islet for assessing the
   reproducibility of the data. N) GO analysis of the PARN interacting
   protein corresponding genes. O) PARN–mCherry plasmid and PTBP1–eGFP
   plasmid were cotransfected in NIT‐1 cells to detect colocalization. P)
   The interaction of PARN and PTBP1 in 293T cells was detected by Western
   blotting. Q) Co‐IP detected the endogenous PARN and PTBP1 interaction
   in NIT‐1 cells. R) Localization of PTBP1 detected by
   immunofluorescence. S) The knockdown efficiency of NIT‐1 cells
   transfected with Ptbp1 siRNA was detected by qRT‐PCR, n = 3. T) The
   knockdown efficiency of NIT‐1 cells transfected with Ptbp1 siRNA was
   detected by Western blotting, n = 3. U) Expression levels of Slc30a8
   and Chst3 after transfection with Parn or Ptbp1 siRNA, n = 3, bar =
   5 µm. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, Student's
   t‐test and ANOVA.

   Furthermore, we performed additional experiments in which Parn siRNA
   was transfected alone or in combination with the pcDNA3.0‐Slc30a8 or
   pcDNA3.0‐Chst3 plasmid (Figure [127]6K). The results showed that Parn
   knockdown led to a decrease in GSIS. Compared with the control,
   overexpression of Slc30a8 or Chst3 also resulted in reduced insulin
   secretion levels. However, compared with those in the knockdown group,
   the insulin secretion levels were partially restored (Figure [128]6L).
   We knocked down Slc30a8 or Chst3 in NIT‐1 cells to further clarify the
   regulatory mechanism involved, and our cell knockdown model was
   successfully constructed (Figure [129]S5C,D,F,G, Supporting
   Information). Moreover, the GSIS results revealed that the insulin
   level in the supernatant did not differ significantly between the
   knockdown group and the NC group at low glucose; however, when Slc30a8
   or Chst3 was knocked down, insulin secretion was significantly lower in
   the high glucose group than in the control group (Figure [130]S5E,H,
   Supporting Information).

   We conducted immunoprecipitation mass spectrometry to identify proteins
   influencing insulin secretion that potentially interacted with PARN.
   The scatterplot revealed strong repeatability between the two sets of
   experimental data (Figure [131]6M). GO enrichment analysis of potential
   interacting proteins highlighted functions related to RNA processing,
   RNA stability, and RNA splicing (Figure [132]6N). PTBP1, a crucial
   RNA‐binding protein that governs mRNA stability and processing, emerged
   as a noteworthy candidate. The colocalization of mCherry–PARN and green
   fluorescent protein (GFP)–PTBP1 in NIT‐1 cells was evident upon
   transfection (Figure [133]6O). Further experimentation through
   co‐immunoprecipitation (co‐IP) with FLAG‐PARN and hemagglutinin
   (HA)–PTBP1 in HEK293T cells confirmed their interaction
   (Figure [134]6P). The IP results demonstrated that endogenous PARN
   interacted with PTBP1 in NIT‐1 cells (Figure [135]6Q).

   We investigated the impact of Ptbp1 deletion on the mRNA levels of the
   PARN‐bound targets Slc30a8 and Chst3 in NIT‐1 cells to assess the
   necessity of the PARN–PTBP1 interaction for maintaining RNA stability.
   IF staining verified PTBP1 expression in NIT‐1 cells (Figure [136]6R).
   The successful knockdown of Ptbp1 with siRNAs, notably with siRNA2, was
   confirmed (Figure [137]6S,T). qRT‐PCR analysis revealed aberrant
   Slc30a8 and Chst3 mRNA levels in NIT‐1 cells lacking Parn or Ptbp1,
   which was consistent with findings from β cells with a specific Parn
   deletion (Figure [138]6U). These results underscored the collaborative
   role of PARN and PTBP1 in regulating the stability of the Slc30a8 and
   Chst3 mRNAs.

3. Discussion

   Elevated insulin secretion in response to glucose represents a critical
   facet in assessing β‐cell function, with compromised insulin secretion
   identified as a key factor in diabetes pathogenesis.^[ [139]^21 ^]
   Therefore, we investigated the role of PARN in pancreatic β cells
   through the specific knockout of Parn in these cells in mice. Our data
   indicated that PARN played a vital role in maintaining blood glucose
   homeostasis in mice by modulating insulin secretion under
   glucose‐stimulated physiological conditions following Parn depletion in
   β cells. Consequently, we propose that PARN is a pivotal factor
   involved in sustaining β‐cell function.

   The normal development and function of β cells are pivotal determinants
   of insulin secretion.^[ [140]^22 ^] Studies have implicated the
   RNA‐binding protein fox‐1 homolog in mediating alternative splicing to
   modulate insulin secretion.^[ [141]^23 ^] In our investigation, we
   observed that PARN loss did not impede β‐cell development but
   significantly inhibited GSIS, which was attributed to concurrent
   reductions in insulin maturation and secretion. This assertion is
   supported by both in vitro and in vivo data.

   RBPs pervasively regulate transcripts throughout their life cycle.^[
   [142]^24 ^] Our results indicated that reduced mRNA stability
   contributed the diminished levels of gene transcripts associated with
   insulin synthesis and secretion upon Parn deletion. RNA‐seq data
   revealed significant upregulation and downregulation of a multitude of
   genes related to insulin secretion, such as Vsnl1,^[ [143]^25 ^]
   Hmgcr,^[ [144]^26 ^] Hmgn3,^[ [145]^27 ^] and Nnat,^[ [146]^28 ^] in
   cKO mice compared with control mice.

   Moreover, our LACE‐seq analysis revealed direct binding of PARN to
   Slc30a8 and Chst3. Slc30a8 encodes zinc transporter 8 (ZnT8), which is
   expressed predominantly on the secretory granule membrane of pancreatic
   β cells, where it governs the synthesis, storage, and secretion of
   insulin.^[ [147]^29 ^] Loss of ZnT8 in mouse models is associated with
   abnormal dense core formation within insulin granules, leading to
   impaired GSIS.^[ [148]^30 ^] Human studies have linked the Slc30a8
   polymorphism, particularly the rs13266634C allele, to a significant
   reduction in insulin secretion.^[ [149]^31 ^] Notably, conflicting
   evidence exists, as certain investigations indicate that the loss of
   Slc30a8 function may increase insulin secretion.^[ [150]^32 ^] A recent
   study has shown that the diminished activity of ZnT8 in β cells
   exacerbates human islet amyloid polypeptide (hiAPP)‐induced toxicity,
   resulting in impaired β‐cell function and disrupted glucose homeostasis
   under metabolic stress.^[ [151]^33 ^] Our study further revealed the
   binding of PARN to Chst3, suggesting its potential role in regulating
   insulin secretion. Consistent with this discovery, carriers of the C
   allele at rs4148941 within Chst3 were shown to have a 32% reduction in
   glucagon like peptide 1 (GLP‐1)‐stimulated insulin.^[ [152]^34 ^]

   PTBP1, a member of the heterogeneous ribonucleoprotein family, exhibits
   dynamic shuttling between the nucleus and cytoplasm and plays pivotal
   roles in diverse cellular processes.^[ [153]^35 ^] Its functions
   include pre‐mRNA alternative splicing and roles in processes such as
   mRNA polyadenylation, nuclear pore export, mRNA stability, subcellular
   localization, replication, and internal‐ribosome‐entry‐site‐mediated
   translation.^[ [154]^36 ^] Our investigation revealed an interaction
   between PARN and PTBP1, which is crucial for maintaining the stability
   of the Slc30a8 and Chst3 mRNAs. Upon glucose stimulation of β cells,
   PTBP1 undergoes nucleocytoplasmic translocation, where cytoplasmic
   PTBP1 binds to and stabilizes mRNAs encoding proteins associated with
   secretory granules (SGs), consequently increasing translation.^[
   [155]^11 , [156]^37 ^] Notably, nuclear PTBP1 levels decrease in
   stimulated nondiabetic islets but remain unaltered in T2DM islets.
   Moreover, PTBP1 has emerged as a novel risk gene for T2DM linked to
   diminished insulin secretion.^[ [157]^38 ^] These findings underscore
   PTBP1 as a crucial posttranscriptional regulator of insulin‐related SG
   components. As Ptbp1 deletion suppresses GSIS, a plausible hypothesis
   is that Parn deletion may compromise PTBP1 function, contributing to
   diminished GSIS.

   In conclusion, our study elucidated the positive impact of PARN on
   pancreatic β cells. The diminished RNA stability of Slc30a8 and Chst3,
   coupled with defects in proinsulin processing and reduced insulin
   secretion, collectively contributed to impaired GSIS in cKO mice
   (Figure [158]7 ). This insight provides a theoretical foundation for
   current gene therapy approaches for the treatment of diabetes. However,
   owing to the substantial RNA requirements of poly(A)‐tail‐primed
   sequencing (2P‐seq) and the challenges in obtaining islets, this
   technology was not employed to investigate the 3′ UTR of RNA.

Figure 7.

   Figure 7
   [159]Open in a new tab

   Model of PARN functional mechanisms in mouse islet. PARN and PTBP1
   interact to regulate Slc30a8 and Chst3 mRNA stability. In
   PARN‐deficient β cells, reduced Slc30a8 and Chst3 mRNA levels lead to
   fewer mature insulin secretary granules and reduced insulin secretion.

4. Experimental Section

Animals

   Ins2‐Cre mice were procured from the Jackson Laboratory (JAX:003573),
   while Parn transgenic mice were created via clustered regularly
   interspaced short palindromic repeats (CRISPR)/CRISPR‐associated
   protein 9 (Cas9) technology. The single guide RNAs (sgRNAs) were
   synthesized with a MEGAshortscript T7 Transcription Kit (Ambion) in
   accordance with the manufacturer's guidelines. For Parn^f/f mice, DNA
   fragments encompassing exon 7 of the Parn gene flanked by two loxP
   sites and two homology arms served as donor templates. The Cas9 protein
   (New England Biolabs (NEB), M0646) and sgRNA were coincubated, and the
   resulting Cas9–sgRNA complex, along with the donor templates, was
   injected into C57BL/6 zygotes. These manipulated zygotes were then
   transferred into pseudopregnant CD1 female mice. The resulting founder
   mice were genotyped and subsequently bred with C57BL/6 mice. HFD‐fed
   mice were subjected to a 60% high‐fat diet (Hfkbio, 10060, China) from
   4 to 18 weeks of age. Parn floxed and Ins2‐Cre mice were generated via
   PCR using DNA extracted from their tails. The sgRNA and primer
   sequences utilized for genotyping are provided in Table [160]S1
   (Supporting Information).

   All the mice were housed in a controlled environment with standard
   lighting conditions (12 h light/dark cycle) under the supervision of
   the Animal Care and Use Committee of China Agricultural University
   (AW92113202‐3‐3).

IF and IHC

   The pancreas was meticulously isolated and subsequently fixed with a 4%
   paraformaldehyde solution overnight at 4 °C. Paraffin‐embedded sections
   were subjected to both IF and IHC analyses. These sections were
   subjected to antigen retrieval via Tris‐HCl or
   ethylenediaminetetraacetic acid buffer, followed by a 1 h incubation in
   10% normal goat serum at room temperature for blocking. The sections
   were subsequently incubated overnight at 4 °C with primary antibodies.
   Gentle washing with phosphate‐buffered saline (PBS) preceded a 1 h
   incubation with secondary antibodies, both fluorescent and
   histochemical, at room temperature. IF images were captured using a
   laser confocal microscope (Nikon, A1), while analysis of the IHC
   sections was conducted through digital pathology section scanning
   (Roche, Ventana DP 200). Details of the antibodies utilized are
   provided in Table [161]S2 (Supporting Information).

Western Blotting

   Tissue lysates were prepared using radio immunoprecipitation assay
   (RIPA) buffer (Beyotime, P0013B) and subjected to analysis via sodium
   dodecyl sulfate– polyacrylamide gel electrophoresis (SDS‐PAGE). The
   proteins were resolved via SDS‐PAGE and subsequently transferred onto
   polyvinylidene fluoride membranes for 2 h at 4 °C. Next, the membranes
   were transferred to 5% skim milk, blocked at room temperature for
   30 min, and incubated with primary antibodies overnight at 4 °C. After
   incubation, the membrane was thoroughly washed with Tris‐buffered
   saline containing Tween and then incubated with secondary antibodies
   for 1 h at room temperature. To ensure equal loading, the housekeeping
   protein (glyceraldehyde‐3‐phosphate dehydrogenase, GAPDH) was
   quantified in each gel. Visualization of the results was accomplished
   using a chemiluminescent imaging system (Tanon 5200). The specific
   details of the antibodies used can be found in Table [162]S2
   (Supporting Information).

Intraperitoneal Glucose Tolerance Test (IPGTT), ITT, and GSIS

   For the IPGTT, 12 and 16 weeks old mice were fasted for 16 h, followed
   by intraperitoneal injection of glucose (2 g kg^−1). Blood glucose
   levels were assessed in tail blood samples at 0, 20, 40, 60, 80, and
   120 min postinjection. In the ITT, mice aged 13 and 17 weeks were
   subjected to a 4 h fast, after which they received an intraperitoneal
   injection of insulin (0.75 U kg^−1). Blood glucose levels were measured
   in tail blood samples at 0, 20, 40, 60, 80, and 120 min following
   insulin administration. Serum was collected at fasting and after
   glucose injection for 10, 20, 30, and 60 min to detect GSIS. Serum
   insulin levels were quantified using a mouse insulin ELISA kit (Crystal
   Chem, 90080).

Isolated Islets

   The mice were anesthetized with 1% pentobarbital sodium, and pancreas
   isolation was achieved by injecting collagenase P (Sigma, 11213857001)
   into the pancreas through the common bile duct. Following isolation,
   the digestion process involved shaking at 37 °C and 170 rpm for 8 min,
   and it was terminated using Hank's buffer (8 g NaCl, 0.4 g KCl,
   0.0489 g MgSO[4], 0.0479 g Na[2]HPO[4], 0.06 g KH[2]PO[4], 0.0468 g
   MgCl[2], 0.185 g CaCl[2]·2H[2]O, 0.3 g d−(+)‐glucose, 0.35 g NaHCO[3],
   and ddH[2]O to 1 L). Islets were subsequently manually selected from
   the digested tissue. Postisolation, the islets were cultured in Roswell
   Park Memorial Institute (RPMI) 1640 supplemented with 10% fetal bovine
   serum (FBS) (VITECH, SE200‐ES) and a penicillin–streptomycin solution
   (Beyotime, C0222). Groups of ten islets were transferred to cell
   culture dishes containing 3 or 20 mm glucose Krebs‐Ringer bicarbonate
   HEPES buffer (KRBH) buffer for 1 h. The intracellular insulin content
   was determined via acid–alcohol, and the supernatant and intracellular
   insulin levels were detected by ELISA. Insulin and proinsulin
   expression levels were analyzed in islets isolated from Ctrl and cKO
   mice by Western blotting.

Measurement of β‐Cell Mass and Islet Size

   The pancreas was weighed and subsequently fixed. 5 µm sections were
   obtained at 200 µm intervals for IHC. The β‐cell fractional area was
   determined by quantifying the percentage of the insulin‐positive
   pancreatic area in relation to the total pancreatic area for each
   section. To calculate the β‐cell mass, the β‐cell fractional area was
   multiplied by the pancreatic weight. The islet distribution ratio was
   computed as the ratio of the number of islets in different regions to
   the total number of islets.

TEM

   Isolated mouse islets were initially fixed overnight with 3%
   glutaraldehyde. The islets were subsequently rinsed with PBS. A second
   fixation was subsequently performed using 1% osmic acid, followed by
   additional rinsing with PBS. The islets then underwent a gradient
   dehydration process involving 50% ethanol, 70% ethanol, 90% ethanol,
   90% acetone, and 100% acetone. Sequential soaking in acetone/embedding
   agent solutions with varying ratios (1:1, 1:2) and acetone/embedding
   agent ratios (1:3) was subsequently performed. Finally, the tissue was
   immersed in pure embedding agent for embedding. The resulting embedded
   blocks were extracted, and ultrathin sections were generated. These
   sections were subjected to staining with uranyl acetate for 30 min,
   followed by three rinses with double‐distilled water. Subsequent
   staining with lead citrate for 10 min was carried out, followed by an
   additional three rinses with distilled water. Following drying, the
   sections were observed via transmission electron microscopy (Hitachi,
   HT7800).

Cell Culture

   The NIT‐1 mouse pancreatic β‐cell line was cultured in RPMI 1640 medium
   (Gibco, C11875500BT) supplemented with 10% FBS (VITECH, SE200‐ES) and a
   penicillin–streptomycin solution (Beyotime, C0222). β‐TC6 cells were
   cultured in high‐glucose Dulbecco's modified Eagle's medium (DMEM;
   Gibco, C11995500BT). To detect insulin secretion, the cells were
   pretreated with KRBH containing 3 mm glucose for 1 h, which was then
   replaced with KRBH containing 20 mm glucose for 1 h. Additionally, the
   human embryonic kidney cell line HEK293T was cultured in DMEM (Gibco,
   C11995500BT).

RNA Sequencing and qRT‐PCR

   The total RNA was extracted from isolated islets via the TRIzol method
   (TaKaRa, RNAiso plus 9109). The RNA integrity and quantity were
   subsequently and meticulously assessed using an Agilent 2100
   bioanalyzer. For sequencing, the fluorescence signal captured by the
   sequencer was converted into a sequencing peak through computer
   software, facilitating the acquisition of sequence information for the
   targeted fragments. Differential expression analysis of the two
   combinations of DEGs was conducted using DESeq2 software (version
   1.20.0). Complementary DNA (cDNA) synthesis was accomplished using a
   RNA Mini Kit (Tiangen, KR107‐01). qRT‐PCR analysis was performed on a
   real‐time PCR system (Light Cycle 96 Instrument) with SYBR Green Master
   Mix (Yeasen, 11198ES03). Detailed information about the primers used
   can be found in Table [163]S1 (Supporting Information).

LACE‐Seq

   LACE‐seq was executed with islet samples obtained from 12 weeks old
   mice following a previously outlined protocol.^[ [164]^20 ^] In brief,
   each sample was subjected to UV‐C light irradiation on ice and was
   administered 3 times for 1 min each to induce cross‐linking. Subsequent
   to cross‐linking, RNA immunoprecipitation was conducted, and the
   immunoprecipitated RNA underwent fragmentation and dephosphorylation.
   The subsequent steps included reverse transcription, strand affinity
   bead capture of the first‐strand cDNA, poly(A) tailing, pre‐PCR, in
   vitro transcription, RNA purification, reverse transcription polymerase
   chain reaction, and deep sequencing.

   Upon the acquisition of raw data, an initial filtering process was
   applied to obtain high‐quality clean data. The clean data were
   subsequently aligned with the reference genome of the project species
   (mm10), and whole‐genome peak calling was subsequently executed on the
   resulting alignment results. The binding preferences of the proteins