Epigenetic regulation of INPP5D and its downstream molecular network in autoimmune pancreatitis

Abstract

Aims

Autoimmune pancreatitis (AIP) is a representative manifestation of IgG4-related disease characterized by lymphoplasmacytic inflammation and storiform fibrosis. Although epigenetic dysregulation has been implicated in autoimmune diseases, its contribution to the pathogenesis of AIP remains unclear.

Methods and results

Genome-wide DNA methylation profiling was performed using the Infinium MethylationEPIC array in tissues affected by AIP and obstructive pancreatitis (OP). Differentially methylated regions (DMRs) were identified and subjected to pathway enrichment analysis. Among the significantly enriched gene sets, GNF2_INPP5D showed the strongest enrichment and was selected for further investigation. INPP5D promoter methylation was analyzed, and its protein expression was evaluated by immunohistochemistry. Single-cell RNA sequencing (scRNA-seq) was performed to identify the cells expressing INPP5D. Pathway enrichment analysis showed significant enrichment of the INPP5D-associated immune-regulatory network in AIP. Promoter-associated CpG probes within the INPP5D locus showed predominant hypomethylation in AIP compared with the findings in OP (adjusted p < 0.05). In contrast, immunohistochemistry demonstrated reduced INPP5D expression in spindle-shaped stromal macrophage/histiocyte-lineage cells in AIP tissues. scRNA-seq analysis identified INPP5D expression predominantly in C1QC-positive macrophages and lymphocytes.

Conclusions

AIP is characterized by epigenetic remodeling of the INPP5D-associated immune-regulatory network accompanied by reduced INPP5D expression in stromal macrophage-lineage cells. These findings provide evidence for an association between altered INPP5D-related epigenetic regulation and the characteristic fibro-inflammatory microenvironment of AIP. INPP5D may represent a candidate biomarker and a potential therapeutic target in IgG4-related disease.

Introduction

INPP5D, also known as SHIP1 (Src homology 2 domain-containing inositol 5¢-phosphatase 1), is expressed in immune cells and functions as a negative regulator of the PI3K/Akt signaling pathway, making it important in regulating immune homeostasis and immune dysregulation [1]. In recent years, INPP5D has attracted increasing attention as an immunoregulatory molecule implicated in immune homeostasis and several autoimmune diseases.

Autoimmune pancreatitis (AIP) is a fibroinflammatory pancreatic disease classified as a manifestation of IgG4-related disease (Fig. 1) [2]. The pathogenesis of AIP is not fully understood, although abnormal immune cell function is thought to play an important role. To explore potential epigenetic factors underlying AIP, we compared obstructive pancreatitis (OP) associated with pancreatic ductal adenocarcinoma and AIP. Using methylation array data, we performed pathway enrichment analysis based on genes mapped to differentially methylated regions (DMRs). Among the candidate pathways identified, the INPP5D-associated immune-regulatory network showed prominent enrichment. On the basis of these findings, we selected INPP5D as a candidate molecule for further investigation of its methylation status and protein expression.

Fig. 1.

Representative histological features of AIP. Dense lymphoplasmacytic infiltration with storiform fibrosis and marked pancreatic acinar atrophy are observed. The inset shows numerous IgG4-positive plasma cells (HE; inset: IgG4 immunostaining). Scale bars: 100 μm (inset, 50 μm)

Materials and methods

Patients and tissue samples

A total of 21 patients fulfilling the diagnostic criteria for AIP and 24 patients with obstructive pancreatitis (OP) associated with pancreatic ductal adenocarcinoma (PDA) and without a history of AIP were identified. All cases were derived from surgical specimens resected between 1996 and 2022 at Shinshu University Hospital or its affiliated hospitals.

For genome-wide DNA methylation analysis, pancreatic tissue was obtained from 14 AIP and 11 OP cases. For immunohistochemical analysis, 16 AIP and 22 OP cases were included. All samples were retrieved from archived formalin-fixed, paraffin-embedded (FFPE) tissue blocks. Clinicopathological data are summarized in Table 1. All lesions were reviewed and pathologically diagnosed by two board-certified pathologists (TU and MI) using hematoxylin and eosin (HE)-stained sections. Histological variables including storiform fibrosis, lymphoplasmacytic infiltration, obliterative phlebitis, and germinal center formation were evaluated according to established histopathological features of AIP.

Table 1.

Clinicopathological characteristics of patients with AIP and OP

Factors	AIP (n = 16)	OP (n = 22)	p-value
Age			0.5205
>72 years	10	11
<72 years	6	11
Sex			1
Male	13	17
Female	3	5
Serum IgG4 (mg/dL)a			0.0006
High (≥ 135 mg/dL)	9	0
Low (< 135 mg/dL)	5	14
Location (head/body–tail)b			0.707
Head	7	10
Body–tail	5	12
Complicationsc			0.0395
Present	7	9
Absent	1	13
Storiform fibrosis			< 0.0001
Present	14	0
Absent	2	22
Lymphoplasmacytic infiltration			< 0.0001
Present	15	2
Absent	1	20
Obliterative phlebitis			< 0.0001
Present	15	0
Absent	1	22
Germinal center formation			0.2029
Present	15	16
Absent	1	6

^aData on serum IgG4 levels were unavailable for 2 cases in the AIP group and 8 cases in the OP group

^bLocation data were unavailable for 3 cases in the AIP group

^cData on complications were unavailable for 8 cases in the AIP group

Bold values indicate statistically significant differences (P < 0.05)

Immunohistochemistry

Immunohistochemical staining for INPP5D and CD163 was conducted using primary antibodies against INPP5D (clone P1C1, dilution 1:200; Santa Cruz Biotechnology, Dallas, TX, USA) and CD163 (clone 10D6, ready-to-use; Leica Biosystems, Newcastle upon Tyne, UK). Immunostaining for INPP5D was performed using a Dako Omnis platform (Agilent Technologies, Santa Clara, CA, USA), whereas CD163 immunostaining was performed using a BOND-III automated immunostainer (Leica Biosystems, Wetzlar, Germany).

INPP5D and CD163 expression was independently evaluated by two board-certified pathologists (TU and MI) who were blinded to the clinical information. Areas showing the highest density of INPP5D-positive spindle-shaped stromal cells were selected as hotspots. In these hotspot regions, the numbers of INPP5D-positive cells and CD163-positive cells were counted in one high-power field.

For INPP5D immunostaining, spindle-shaped stromal cells showing distinct cytoplasmic staining were regarded as positive. For CD163 immunostaining, macrophage-lineage cells showing cytoplasmic and/or membranous staining were counted as positive. The ratio of INPP5D-positive cells to CD163-positive cells (INPP5D/CD163 ratio) was subsequently calculated for each case. Appropriate positive and negative controls were included in each staining run.

Methylation array analysis

Genome-wide DNA methylation profiling was performed on 14 AIP and 11 OP samples using the Infinium MethylationEPIC BeadChip (Illumina, San Diego, CA, USA). Beta values, ranging from 0 (unmethylated) to 1 (fully methylated), were calculated from fluorescence intensity ratios using GenomeStudio software, with higher beta values indicating greater methylation levels.

Genomic DNA was extracted from 4-µm FFPE unstained tissue sections using the QIAamp DNA Mini Kit (Qiagen, Valencia, CA, USA). Bisulfite conversion was performed using the EZ DNA Methylation Kit (Zymo Research, Orange, CA, USA), in accordance with the manufacturer’s protocol.

Pathway enrichment analysis

Differentially methylated positions (DMPs) and differentially methylated regions (DMRs) between AIP and OP were identified using the ChAMP package. Genes mapped to significant DMRs were subjected to pathway enrichment analysis using the ReactomePA package. Enrichment was assessed using a hypergeometric test, and adjusted p-values were calculated with the Benjamini–Hochberg method (false discovery rate, FDR). Candidate genes for further analysis were selected by considering both statistical significance and their reported biological relevance to autoimmune diseases and immune cell function. The Ethics Committee of Shinshu University School of Medicine approved the protocol of the present study (Approval number: 720).

Selection of candidate CpG sites for detailed analysis

CpG sites annotated to genes included in candidate immune-related gene sets identified by pathway enrichment analysis were extracted using the Illumina EPIC array annotation. To prioritize CpG sites with potential regulatory relevance to gene transcription, analyses focused on probes annotated as TSS200, TSS1500, 5′UTR, or 1stExon in the UCSC_RefGene_Group annotation. Probes with overlapping annotations (e.g., Body; TSS1500) were included when promoter-associated annotations were present.

Single-cell RNA sequencing analysis of INPP5D expression in PDA

Single-cell RNA sequencing (scRNA-seq) data from PDA were analyzed to determine the cells expressing INPP5D in the tumor microenvironment. Publicly available datasets were processed using the Seurat R package (v4.1.0). After normalization and clustering, major cell populations—including epithelial cells, fibroblasts, endothelial cells, and immune cell subsets (T cells, B cells, NK cells, and myeloid cells)—were annotated based on canonical marker expression. The expression of INPP5D was visualized using uniform manifold approximation and projection (UMAP) and violin plots.

Statistical analysis

Continuous variables were expressed as median values with ranges and compared using the Mann–Whitney U test. Categorical variables were analyzed using Pearson’s chi-squared test or Fisher’s exact test, as appropriate. Comparisons of INPP5D-positive cell counts, CD163-positive cell counts, and the INPP5D/CD163 ratio between the AIP and OP groups were performed using the Mann–Whitney U test. Clinicopathological analyses and immunohistochemical data analyses were conducted using JMP version 13 (SAS Institute, Cary, NC, USA).

For single-cell RNA sequencing analysis, statistical analyses were performed using EZR (Easy R) version 1.66, a graphical user interface for R developed by The R Foundation for Statistical Computing (Vienna, Austria), following the analytical framework used in our previous studies. A two-sided p-value of < 0.05 was considered statistically significant.

Results

Selection of INPP5D as a candidate gene

Pathway enrichment analysis of DMRs revealed multiple gene sets significantly enriched in AIP compared with the findings in OP. Among these, the GNF2_INPP5D gene set showed particularly strong enrichment (n = 9, OR = 55.1, p = 9.62 × 10^− 13, adjusted p = 1.34 × 10⁻⁹). Consistent with this, dot plot visualization demonstrated that GNF2_INPP5D ranked among the top enriched gene sets, characterized by a high gene ratio and a low adjusted p-value (Fig. 2). On the basis of this strong statistical enrichment and the well-established immunoregulatory functions of INPP5D in multiple immune cell types, we selected it for further analysis of its methylation status and protein expression. The overlapping genes within the enriched GNF2_INPP5D gene set included SASH3, TAPBP, TAP1, INPP5D, ARHGDIB, HLA-A, HLA-B, HLA-E, and PSMB8.

Fig. 2.

Enrichment analysis of DMR-associated gene sets in autoimmune pancreatitis. Dot plot showing gene set enrichment results based on genes mapped to DMRs identified between AIP and OP. The x-axis represents the gene ratio, calculated as the proportion of DMR-associated genes overlapping each gene set relative to the total number of genes in that gene set, while the y-axis indicates individual gene sets. The size of each dot reflects the number of overlapping genes, and the color scale corresponds to the −log10-transformed adjusted p-value. Among the enriched pathways, the GNF2_INPP5D gene set demonstrated particularly strong statistical enrichment, highlighting INPP5D-associated epigenetic alterations characteristic of AIP

Methylation patterns of INPP5D-related immune genes

To further characterize methylation abnormalities associated with the enriched GNF2_INPP5D gene set, CpG methylation patterns of related immune-regulatory genes were evaluated. The analyzed genes included multiple antigen-presentation and immune-regulatory genes.

Among the analyzed CpG sites, the majority demonstrated lower methylation levels in AIP than in OP. Specifically, CpG sites showing relative hypomethylation in AIP versus OP were identified as follows: SASH3, 10 versus 0 sites; TAPBP, 7 versus 2 sites; TAP1, 9 versus 0 sites; INPP5D, 13 versus 4 sites; ARHGDIB, 22 versus 0 sites; HLA-A, 5 versus 3 sites; HLA-B, 15 versus 1 site; HLA-E, 16 versus 4 sites; and PSMB8, 46 versus 7 sites.

Overall, 143 of 164 CpG sites (87.2%) demonstrated relative hypomethylation in AIP. Although a minority of CpG sites showed the opposite pattern, predominant hypomethylation in AIP was consistently observed across all nine genes.

Methylation status of INPP5D

Probe-level analysis identified 18 CpG probes mapped to the INPP5D locus. Various probes were located within promoter-associated regions (TSS200, TSS1500, or 5′UTR) and/or annotated as enhancer elements, with some probes meeting both criteria (Fig. 3). Visualization of methylation differences (Δβ values) demonstrated that multiple promoter-associated probes exhibited lower methylation levels in AIP than in OP, and several of these probes showed statistically significant hypomethylation after correction for multiple tests (adjusted p < 0.05).

Fig. 3.

Genomic distribution and methylation status of CpG probes mapped to the INPP5D locus. Schematic representation of the INPP5D genomic region showing RefSeq gene annotations, promoter-associated regions (TSS200, TSS1500, and 5′UTR), and annotated enhancer elements. Vertical bars indicate the positions of CpG probes included in the methylation array analysis. Differential methylation between AIP and OP is visualized as Δβ values for individual probes, highlighting the distribution of methylation changes across promoter and adjacent regulatory regions

Immunohistochemistry of INPP5D and CD163

Immunohistochemical analysis of 16 AIP and 22 OP cases revealed that INPP5D was predominantly expressed in spindle-shaped stromal macrophage/histiocyte-lineage cells, in both AIP and OP tissues. Representative images are shown in Fig. 4. In INPP5D-positive cases, the expression pattern was diffuse without apparent regional heterogeneity. Likewise, in cases showing reduced or lost INPP5D expression, the reduction was diffusely distributed and appeared unrelated to the degree of lymphoplasmacytic infiltration. Lymphoid follicles and lymphocytes also exhibited INPP5D positivity in both groups.

Fig. 4.

Comparison of CD163-positive macrophages and INPP5D expression between OP and AIP. Representative images of OP (A–C) and AIP (D–F). HE is shown in panels A and D, CD163 in panels B and E, and INPP5D immunostaining in panels C and F. All images were obtained from the corresponding area of each lesion. Numerous CD163-positive macrophage/histiocyte-lineage cells are observed in both lesions, whereas INPP5D expression in these cells is more prominent in OP than in AIP. Scale bars = 50 μm

CD163-positive macrophage/histiocyte density was significantly higher in AIP than in OP (median 121.0 [93.3–153.3] vs. 72.5 [49.3–104.3], p = 0.0043). In contrast, INPP5D-positive stromal cell density was significantly higher in OP than in AIP (17.5 [8.5–49.3] vs. 7.5 [3.0–13.3], p = 0.0297). Consequently, the INPP5D/CD163 ratio was significantly lower in AIP (0.058 [0.022–0.157] vs. 0.303 [0.118–0.564], p = 0.0016)(Fig. 5).

Fig. 5.

Quantitative comparison of CD163-positive cells, INPP5D-positive cells, and the INPP5D/CD163 ratio between AIP and OP. Box plots showing the densities of CD163-positive macrophage/histiocyte-lineage cells (A), INPP5D-positive stromal cells (B), and the INPP5D/CD163 ratio (C) in AIP and OP. Statistical comparisons were performed using the Mann–Whitney U test

Single-cell RNA sequencing reveals INPP5D expression in C1QC-positive macrophages

scRNA-seq analysis demonstrated that INPP5D was predominantly expressed in lymphocytes and myeloid-lineage cells, with minimal expression in epithelial or fibroblast clusters (Fig. 6). Subclustering of the myeloid compartment revealed a distinct population of C1QC-positive macrophages showing strong INPP5D expression (Fig. 6).

Fig. 6.

Distribution of INPP5D expression revealed by scRNA-seq analysis in PDA. A UMAP visualization of major cell clusters (epithelial, fibroblast, endothelial, T, B, NK, and myeloid cells). B INPP5D expression is predominantly observed in lymphocytes and myeloid cells. C The myeloid domain was classified into multiple subclusters. D Myeloid subclustering identifies a C1QC-positive macrophage subset with high INPP5D expression

Discussion

Although INPP5D is a well-established immunoregulatory molecule [3, 4], its epigenetic alterations and potential involvement in autoimmune pancreatitis have not previously been characterized.

DMR-based pathway enrichment analysis showed significant enrichment of the GNF2_INPP5D gene set in AIP. This indicates that genes transcriptionally linked to INPP5D are epigenetically altered in AIP. This enrichment suggests the presence of broader epigenetic alterations involving immune-regulatory networks in AIP. However, enrichment analysis does not specify the direction of gene regulation. Therefore, our findings suggest that the INPP5D-related immune-regulatory network is selectively affected by disease-specific epigenetic changes in AIP, rather than being uniformly suppressed.

A particularly important finding of the present study is the predominant hypomethylation across multiple genes within the GNF2_INPP5D-associated network, including multiple antigen-presentation and immune-regulatory genes, in AIP compared with the findings in OP. These genes are closely related to antigen presentation, interferon-associated immune responses, and leukocyte immune regulation. TAP1, TAPBP, PSMB8, and HLA class I molecules are key components of the antigen-processing and presentation pathway, whereas SASH3 and ARHGDIB are involved in lymphocyte signaling and immune-cell regulatory pathways [5–7]. These findings suggest broad epigenetic remodeling of an immune-regulatory network in AIP [8, 9].

Importantly, several of these genes are known to be induced under interferon-associated inflammatory conditions [5, 10]. Previous studies have demonstrated activation of IFN-related immune pathways in IgG4-related disease and AIP, particularly in association with plasmacytoid dendritic cells and innate immune activation [11, 12]. Therefore, the coordinated hypomethylation observed in the present study may be associated with disease-specific activation of interferon-responsive immune programs within the fibro-inflammatory microenvironment of AIP.

Recent studies have highlighted INPP5D as a key immunoregulatory molecule. For example, Chou et al. reported that INPP5D negatively regulates inflammatory responses in microglia. Loss of INPP5D results in activation of the NLRP3 inflammasome and excessive secretion of proinflammatory cytokines [13]. Because histiocyte-lineage cells share functional characteristics with microglia [14], a similar regulatory mechanism may operate in stromal macrophage-lineage cells in AIP.

An additional conceptual framework that may help explain the present findings is the emerging concept of exhaustion-like remodeling in chronically stimulated macrophages. Recent studies have demonstrated that persistent inflammatory stimulation, aging-associated immune dysregulation, and fibrotic tissue remodeling can induce exhaustion-like states characterized by impaired homeostatic regulation, defective phagocytosis, altered lysosomal activity, and sustained inflammatory signaling [15–17]. Because INPP5D plays important roles in PI3K-associated immune signaling, phagocytic activity, and intracellular immune homeostasis [3, 4], reduced INPP5D protein expression in spindle-shaped stromal macrophage-lineage cells in AIP may be associated with maladaptive or exhaustion-like immune remodeling under chronic inflammatory and fibrotic stress.

DMR-based pathway analysis demonstrated enrichment of the GNF2_INPP5D gene set in AIP. Although co-expression does not necessarily imply identical function, several mechanisms may explain the coordinated methylation changes observed in the present study. INPP5D and related genes may be regulated by common transcription factors such as STATs or NF-κB [18–20], share overlapping PI3K-associated regulatory pathways [21], and be coordinately expressed within histiocyte-lineage transcriptional programs [22]. Collectively, these mechanisms may explain the observed coordinated changes of methylation in AIP.

In AIP, a substantial subset of CD163-positive M2-polarized macrophages is likely to express C1QC and to participate in immune regulation and stromal fibrosis [23]. Recent studies have suggested that C1QC-positive macrophages are involved in tissue remodeling, complement regulation, lipid metabolism, and fibrosis-associated immune regulation [24, 25]. Given their abundant infiltration within fibrotic stroma, it appears more plausible that INPP5D expression is selectively reduced within C1QC-expressing macrophages rather than that the macrophage population itself is diminished. Such dysfunction or exhaustion-like remodeling may be associated with impaired immune homeostasis and persistent fibro-inflammatory responses in AIP.

AIP is further characterized by abundant infiltration of CD163-positive M2-polarized macrophages [26, 27]and plasmacytoid dendritic cells (pDCs) [12, 28–30], which promote IgG4-related immune responses through the production of IL-10, TGF-β, and type I interferons [31–33]. These immune-cell populations may contribute to the enrichment of INPP5D-associated signatures observed in AIP. By contrast, OP predominantly shows macrophage infiltration associated with secondary inflammatory responses induced by ductal obstruction [34]. Collectively, these findings suggest an association between epigenetic remodeling of the INPP5D-associated immune-regulatory network and the characteristic fibro-inflammatory microenvironment of AIP, rather than merely reflecting inflammatory cell infiltration. However, further functional studies are required to determine whether these alterations play a direct biological role in disease pathogenesis.

Importantly, despite promoter hypomethylation of the INPP5D locus in AIP, immunohistochemistry demonstrated reduced INPP5D protein expression in stromal macrophage/histiocyte-lineage cells. This apparent discrepancy suggests that bulk methylation profiles do not necessarily reflect functional protein expression within specific immune-cell subsets. Such discordance may reflect cell-type–specific regulation, compensatory immune feedback, or exhaustion-like remodeling of macrophage-lineage cells within chronically inflamed fibrotic lesions [15, 17]. Collectively, these findings suggest that epigenetic remodeling of the INPP5D-associated immune-regulatory network is associated with the characteristic fibro-inflammatory microenvironment of AIP rather than merely reflecting inflammatory cell infiltration.

In the present study, the single-cell RNA-seq analysis was primarily used to complement the methylation findings by identifying the cellular distribution of INPP5D. Future studies using disease-specific single-cell RNA-seq or spatial transcriptomic datasets should further characterize INPP5D-associated macrophage subpopulations, their transcriptional programs, and downstream signaling pathways.

This study has several limitations. First, the functional significance of the observed epigenetic alterations remains unclear, and the immunophenotypic characteristics of INPP5D-expressing stromal cells were not fully characterized. In addition, because methylation analysis was performed using bulk tissue specimens, cell-type-specific epigenetic alterations could not be fully resolved. Future studies integrating spatial transcriptomics and functional analyses will help clarify the cellular and biological significance of INPP5D-associated immune remodeling in AIP. Furthermore, because the present study was observational, causal relationships between epigenetic alterations of INPP5D and autoimmune pancreatitis cannot be established. Although coordinated DNA methylation changes and reduced INPP5D protein expression were observed, it remains unclear whether these alterations directly contribute to disease pathogenesis or merely represent secondary changes associated with chronic inflammation and fibrosis. Recent studies have demonstrated the utility of Mendelian randomization for distinguishing causal from non-causal associations identified by omics analyses. Therefore, future studies integrating genetic epidemiological approaches, such as Mendelian randomization, together with functional validation may help determine whether genetically regulated INPP5D expression plays a causal role in the development of autoimmune pancreatitis and IgG4-related disease [35, 36].

Conclusion

In AIP, coordinated hypomethylation was observed in the INPP5D locus and related immune-regulatory genes, whereas INPP5D protein expression was reduced in stromal macrophage/histiocyte-lineage cells. These findings suggest that AIP is characterized by coordinated epigenetic alterations involving the INPP5D-associated immune-regulatory network together with reduced INPP5D protein expression in stromal macrophage-lineage cells. The present findings provide evidence that the INPP5D-associated immune-regulatory network is associated with the characteristic fibro-inflammatory microenvironment of AIP. Further functional studies are required to determine whether INPP5D plays a direct biological role in disease pathogenesis. If validated, INPP5D may represent a potential biomarker and therapeutic target for IgG4-RD.

Abbreviations

AIP

Autoimmune pancreatitis

DMP

Differentially methylated position

DMR

Differentially methylated region

DNA

Deoxyribonucleic acid

EZR

Easy R

FDR

False discovery rate

FFPE

Formalin-fixed paraffin-embedded

HE

Hematoxylin and eosin

HLA

Human leukocyte antigen

HPF

High-power field

IFN

Interferon

IgG4-RD

Immunoglobulin G4-related disease

IL

Interleukin

INPP5D

Inositol polyphosphate-5-phosphatase D

MHC

Major histocompatibility complex

NK

Natural killer

NLRP3

NLR family pyrin domain containing 3

OP

Obstructive pancreatitis

OR

Odds ratio

PDA

Pancreatic ductal adenocarcinoma

PI3K

Phosphoinositide 3-kinase

pDC

Plasmacytoid dendritic cell

scRNA-seq

Single-cell RNA sequencing

SHIP1

Src homology 2 domain-containing inositol 5′-phosphatase 1

TGF-β

Transforming growth factor-beta

TMA

Tissue microarray

Treg

Regulatory T cell

UMAP

Uniform manifold approximation and projection

Authors' contributions

Takeshi Uehara: Conceptualization, Methodology, Formal analysis, Supervision, Project administration, Funding acquisition, Writing – original draft, Writing – review & editing. Mai Iwaya: Investigation, Validation. Shiho Asaka: Investigation. Tomoyuki Nakajima: Investigation, Resources. Shotaro Komamura: Investigation. Hiroshi Sawaguchi: Investigation. Shunsuke Imamura: Investigation. Akira Nakamura: Investigation. Shinsuke Sugenoya: Investigation. Akira Shimizu: Investigation. Hiroyoshi Ota: Supervision, Writing – review & editing.

Funding

This study was supported by the Hokuto Foundation for Bioscience (grant award to TU). This work was supported by JSPS KAKENHI Grant Numbers JP20K07405 and JP24K10121.

Data availability

The data generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Shinshu University School of Medicine (Approval No. 720). Because this was a retrospective study using archived human tissue specimens, the requirement for written informed consent to participate was waived by the Ethics Committee of Shinshu University School of Medicine. An opt-out approach was adopted. The study was conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing Interest

The authors declare no competing interests.