Pancreatitis-Associated PRSS1-PRSS2 Haplotype Alters T-Cell Receptor Beta (TRB) Repertoire More Strongly Than PRSS1 Expression

Abstract

Genetic variants linked to the PRSS1 and SPINK1 genes alter the risk of recurrent acute and chronic pancreatitis (RAP/CP).1Whitcomb D.C. et al.Nat Genet. 1996; 14: 141-145Crossref PubMed Scopus (1329) Google Scholar, 2Witt H. et al.Nat Genet. 2000; 25: 213-216Crossref PubMed Scopus (846) Google Scholar, 3Pfutzer R.H. et al.Gastroenterology. 2000; 119: 615-623Abstract Full Text Full Text PDF PubMed Scopus (440) Google Scholar The effects of common risk/protective haplotypes in the PRSS1-PRSS2 and SPINK1 loci in RAP/CP are poorly understood.4Whitcomb D.C. et al.Nat Genet. 2012; 44: 1349-1354Crossref PubMed Scopus (236) Google Scholar Here we define the genomics of PRSS1-PRSS2 and SPINK1 risk haplotypes at the population, organ, and single-cell level to better understand the mechanisms linking genetic variants to RAP/CP. Genotyped North American Pancreatitis Study II (NAPS2) cases of European ancestry (RAP/CP, n = 1341) and controls (n = 5691) were used for population studies (see Supplementary Material).5Whitcomb D.C. et al.Pancreatology. 2008; 8: 520-531Crossref PubMed Scopus (178) Google Scholar,6Dunbar E. et al.J Gastroenterol. 2020; 55: 1000-1009Crossref PubMed Scopus (9) Google Scholar We selected the RAP/CP PRSS1-PRSS2_rs10273639C>T risk haplotype (linked to rs6667T>C, PRSS1 p.Asn246=), spanning much of the T-cell receptor beta (TRB) locus, and the SPINK1_rs17107315C>T (p.Asn34Ser, or p.N34S) risk locus for population studies. We tested the hypothesis that the PRSS1-PRSS2 and SPINK1 risk haplotypes are linked to altered trypsin controls and their effects are synergistic. The risk of RAP/CP for SPINK1_rs17107315 TC+CC alleles was slightly higher than TC genotypes alone (TC+CC: odds ratio [OR], 3.7; 95% CI, 2.8–5.0; P = 2e–16; TC: OR, 3.4; 95% CI, 2.5–4.5; P = 3.4e–15). Homozygous SPINK1_rs17107315CC genotypes were considered pathogenic (disease causing) and excluded from SPINK1 vs PRSS1 interaction analysis. We then compared SPINK1_rs17107315TC (risk) and TT (wild-type) with 3 PRSS1-PRSS2_rs10273639 genotypes (TT [protective], TC [risk], and CC [risk]). The risk of RAP/CP with SPINK1_rs17107315TC on the PRSS1-PRSS2_rs10273639TT background was increased, but did not reach statistical significance (OR, 2.4; 95% CI, 0.91–5.87; P = .51), but was significant on PRSS1-PRSS2_rs10273639TC (OR, 3.9; 95% CI, 2.5–6.0; P = 3e–10) and CC (OR, 3.0; 95% CI, 1.8–5.0; P = 2e–5) with similar effects in alcohol and nonalcohol etiologies (not shown). However, there was no significant interaction of SPINK1 and PRSS1-PRSS2 risk haplotypes (3 × 2 χ2: P = .3) or the PRSS1-PRSS2 protective (TT) and PRSS1-PRSS1 risk (TC + CC) genotypes and the SPINK1 TT and TC genotypes (2 × 2 χ2; OR, 0.67; 95% CI, 0.27–1.42; P = .3). These associations were replicated in the UK Biobank (Supplementary Material). This suggests that although both the SPINK1 risk haplotype (linked to trypsin-dependent pathways7Hegyi E. et al.Dig Dis Sci. 2017; 62: 1692-1701Crossref PubMed Scopus (104) Google Scholar) and the PRSS1-PRSS2 risk haplotypes are associated with RAP/CP, their effects may be through different pathways. Because the PRSS1-PRSS2 risk haplotype is complex and overlaps the TRB locus.8Pu N. et al.Genes (Basel). 2021; 12Google Scholar We tested the alternative hypothesis that RAP/CP risk was associated with T-cell repertoire using expression quantitative trait loci linked to PRSS1-PRSS2_rs10273639T>C reported in Genotype-Tissue Expression (GTEx, release V8, Broad Institute; https://gtexportal.org/home/snp/chr7_142749077_T_C_b38). The strongest association was with TRBV28 (whole blood, P = 4.0e–27). For pancreas, TRBV29-1 (P = 7.3e–19) was more significant than PRSS2 (P = 4.3e–7). All of the top 173 transcripts were associated with TRB except 1—PRSS2 (PRSS1 not listed). Of note, TRBV29-1 is immediately upstream of PRSS1 and is highly expressed in the pancreas and salivary gland. These data indicate that the PRSS1-PRSS2 risk haplotype is more likely associated with altered immune responses to pancreatic injury. To evaluate PRSS1 × SPINK1 at tissue and single-cell levels, we studied human tissue. Pancreatic samples were prospectively collected for total RNA sequencing (n = 15) and single-cell RNA sequencing (n = 4) from the GREAT1 (Genomic Resources for Enhancing Available Therapies, ClinicalTrials.gov number NCT04306939) (n = 14) and the Prospective Autos (Autologous Islet Transplantation for Treatment of Pancreatic Disease, Institutional Review Board approval number 0609M91887) (n = 5) studies (see Supplementary Material). Expression of SPINK1 and PRSS1 RNA transcripts was evaluated for SPINK1_rs17107315_C (p.Ser34) and PRSS1-PRSS2 risk haplotype linked to rs667 (PRSS1 p.Asn246=). Among heterozygous SPINK1 p.Asn34Ser (rs17107315 TC) individuals (n = 3), SPINK1_rs17107315_T (p.Asn34) represented 78.5% ± 4.3% of mRNA reads, and the risk allele, SPINK1_rs17107315_C (p.Ser34), represented 21.5% ± 4.3% of mRNA reads (C is 57% lower than T; Welch 2-sample t test, P = 8.3e–05), the first demonstration of altered RNA expression in human tissue linked to SPINK1 p.Asn34Ser.8Pu N. et al.Genes (Basel). 2021; 12Google Scholar Among heterozygous PRSS1 p.Asn246= (rs6667 TC) individuals (n = 12), the low-risk PRSS1_rs6667_T allele represented 47.5% ± 4.0% of mRNA reads and the high-risk allele PRSS1_rs6667_C represented 52.5% ± 4.0% of mRNA reads (C is 5% higher than T; Wilcoxon signed ranked test, P = .021). High variability in expression suggests that independent variants also affect PRSS1 expression. Thus, low expression of the SPINK1_rs17107315_C (p.Ser34) explains increased RAP/CP risk,8Pu N. et al.Genes (Basel). 2021; 12Google Scholar but the minimal increase in PRSS1 expression from PRSS1_rs6667_C (risk) cannot explain the higher risk of the PRSS1-PRSS2 risk haplotype. The relative expression of PRSS1 and SPINK1 among different pancreatic cell subtypes was evaluated using single-cell RNA sequencing. A heatmap identified 10 major cell types using 10 cell marker genes to classify each cell type (Supplementary Figure 1). A large fraction of cells were partially undifferentiated acinar-type cells,9Blobner B.M. et al.Am J Physiol Gastrointest Liver Physiol. 2021; 321: G449-G460Crossref PubMed Scopus (4) Google Scholar with some genes from other cell types expressed at low levels (Supplementary Figure 1, column 2 [undifferentiated cells] and column 4 [acinar cells]). Figure 1A is a UMAP (Uniform Manifold Approximation and Projection) plot of all cell types. Expression of PRSS1 and SPINK1 in cell types is shown using feature plots (Figure 1B and C). The PRSS1/SPINK1 ratio ranged from 5.92 to 7.38 (mean of 6.47), which is slightly lower than the PRSS1/SPINK1 ratio of 18.5 from GTEx (https://gtexportal.org/home/eqtls/tissue?tissueName=Pancreas). The slightly higher SPINK1 levels in our samples may be due to underlying inflammation.10Khalid A. et al.Gut. 2006; 55: 728-731Crossref PubMed Scopus (42) Google Scholar The PRSS1/SPINK ratio was high in subsets of undifferentiated acinar-type cells (circled in Figure 1A and expanded in Figure 1D into 9 subgroups [0–8]). Violin plot showed the PRSS1 and SPINK1 expression level in subclusters (Figure 1E). The PRSS1/SPINK1 expression ratio was highest in subclusters 5 and 6 (Figure 1F). KEGG pathway enrichment in subcluster 6 indicated that cells with high PRSS1/SPINK1 ratios also had marked inflammation and immune gene expression responses, consistent with unregulated trypsin activity and cell injury vs acinar cells (Supplementary Table 1). This indicates that cells with low SPINK1 relative to PRSS1 expression are undergoing intracellular damage, likely related to uninhibited trypsin activity.7Hegyi E. et al.Dig Dis Sci. 2017; 62: 1692-1701Crossref PubMed Scopus (104) Google Scholar Thus, SPINK1 protects the pancreas by inhibiting active PRSS1 (trypsin). Low SPINK1 expression is a risk for cell injury and RAP/CP. The PRSS1-PRSS2 risk haplotype has little effect on PRSS1 expression or interaction with SPINK1. The PRSS1-PRSS2 risk haplotype is strongly associated with altered expression of TRBV29-1 and other TRB transcripts. These data suggest that the strong risk of the PRSS1-PRSS2 haplotype for RAP/CP may be through variant TRB repertoires that alter immune phenotypes. Co-investigators of the GREAT1 and University of Pittsburgh Pancreas Study Group: Randall E. Brand MD, Celeste Shelton Ohlsen, PhD, CGC, Jami L. Saloman, PhD, and David C. Whitcomb, MD, PhD, Division of Gastroenterology, Hepatology and Nutrition, Department of Medicine, University of Pittsburgh, Pittsburgh, PA; H. J. Park, PhD, Department of Human Genetics, University of Pittsburgh, Pittsburgh, PA; and Kenneth K. Lee, MD, Alessandro Paniccia, MD, and Amer Zureikat, MD, Division of Gastrointestinal Surgical Oncology, Department of Surgery, University of Pittsburgh, Pittsburgh, PA. Co-investigators of the Prospective Autos and the University of Minnesota Group: Melena D. Bellin, MD, Department of Pediatrics, University of Minnesota, Minneapolis, MN and Greg Beilman, MD, Division of Critical Care/Acute Care Surgery, Department of Surgery, University of Minnesota, Minneapolis, MN. Co-investigators of the North American Pancreatitis Study II (NAPS2) Genetics Consortium: Samer Alkaade, MD, Stephen Amann, MD, Michelle A. Anderson, MD, MSc, Peter Banks, MD, Randall E. Brand, MD, Darwin L. Conwell, MD, Gregory A. Cote, MD, MS, Christopher E. Forsmark, MD, Timothy B. Gardner, MD, Andres Gelrud, MD, Nalini M. Guda, MD, Michele D. Lewis, MD, Thiruvengadam Muniraj, MD, PhD, Georgios I. Papachristou, MD, PhD, Joseph Romagnuolo, MD, Bimaljit S. Sandhu, MD, Stuart Sherman, MD, Vikesh K. Singh, MD, MSc, Adam Slivka, MD, PhD, Charles Melbern Wilcox, MD, and Dhiraj Yadav, MD. Additional NAPS2 control samples from the Pittsburgh regions were contributed by Seth M. Weinberg, PhD, Department of Oral and Craniofacial Sciences, School of Dental Medicine, University of Pittsburgh, Pittsburgh, PA and Mary L. Marazita, PhD, Center for Craniofacial and Dental Genetics, Department of Oral and Craniofacial Science, School of Dental Medicine; Department of Human Genetics, School of Public Health, Clinical and Translational Sciences, and Department of Psychiatry, School of Medicine, University of Pittsburgh, Pittsburgh, PA. The authors acknowledge the contribution of the following collaborators: Kaare Christensen, Andrew Czeizel, Frederic W. Deleyiannis, Eleanor Feingold, Jacqueline C. Hecht, Myoungkeun Lee, Lina Moreno-Uribe, Jeff C. Murray, Katherine Neiswanger, John R. Shaffer, George Wehby, and the staff of the Center for Craniofacial and Dental Genetics, University of Pittsburgh, Pittsburgh, PA. The authors also acknowledge the contributions of the following individuals to the NAPS2 studies: Nalini Guda, MD, Aurora St. Luke’s Medical Center, Milwaukee, WI; Peter Banks, MD, Brigham & Women’s Hospital, Boston, MA; Darwin Conwell, MD, Brigham & Women’s Hospital, Boston, MA (current affiliation is University of Kentucky, Lexington, KY); Simon K. Lo, MD, Cedars-Sinai Medical Center, Los Angeles, CA; Timothy Gardner, MD, Dartmouth-Hitchcock Medical Center, Hanover, NH; the late John Baillie, MD, Duke University Medical Center, Durham, NC; Christopher E. Forsmark, MD, University of Florida, Gainesville, FL; Thiruvengadam Muniraj, MD, PhD, Griffin Hospital, CT (current affiliation is Yale University, New Haven, CT); Stuart Sherman, MD, Indiana University, Indianapolis, IN; Vikesh Singh, MD, Johns Hopkins University, Baltimore, MD; Mary Money, MD, Washington County Hospital, Hagerstown, MD; Michele Lewis, MD, Mayo Clinic, Jacksonville, FL; Joseph Romagnuolo, MD, Medical University of South Carolina, Charleston, SC (current affiliation is also Ralph H. Johnson Veterans Affairs Medical Center, Charleston, SC); Robert Hawes, MD, Indiana University, Indianapolis, IN (then at Medical University of South Carolina, Charleston, SC and current affiliation is Orlando Health, Orlando, FL); Gregory A. Cote, MD, Christopher Lawrence, MD, Medical University of South Carolina, Charleston, SC (current affiliation is Oregon Health & Science University, Portland, OR); C. Mel Wilcox, MD, University of Alabama, Birmingham, AL; Michelle A. Anderson, MD, University of Michigan, Ann Arbor, MI (current affiliation is Mayo Clinic, Scottsdale, AZ); Stephen T. Amann, MD, North Mississippi Medical Center, Tupelo, MS; Babak Etemad, MD, Ochsner Medical Center, New Orleans, LA (current affiliation is Main Line HealthCare Interventional Gastroenterology, Wynnewood, PA); Mark DeMeo, MD, Rush University Medical Center, Chicago, IL; Samer Al Kaade, MD, St Louis University, St Louis, MO (current affiliation is Mercy Clinic Gastroenterology, St Louis, MO); Michael Kochman, MD, University of Pennsylvania, Philadelphia, PA; the late M. Michael Barmada, PhD, Randall E. Brand, MD, Bernie Devilin PhD, Adam Slivka, MD, PhD, David C. Whitcomb, MD, PhD, Dhiraj Yadav, MD, MPH, Jessica LaRusch, PhD, Judah N. Abberbock, PhD, Gong Tang, PhD, Michael O’Connell, PhD, Kimberly Stello, Emil Bauer, Elizabeth Kennard, PhD, Stephen R. Wisniewski, PhD, University of Pittsburgh, Pittsburgh, PA; the late Frank Burton, MD, St Louis University, St Louis, MO; James DiSario, MD, University of Utah Health Science Center, Salt Lake City, UT (current affiliation is Monterey Bay GI Consultants, Monterey, CA); Bimaljit S. Sandhu, MD, Virginia Commonwealth University, Richmond, VA (current affiliation is St Mary's Hospital, Richmond, VA); William Steinberg, MD, Washington Medical Center, Washington, DC (currently retired). The authors thank Kim Stello, Lori Kelly, Esther Lee, and Matthew Henkel for technical assistance. An abstract including some of these data was presented in Digestive Disease Week in 2020. Dongni Fu’s current affiliation is the Division of Gastroenterology, Xiangya Hospital, Central South University, Changsha, Hunan, China. Brandon M. Blobner’s current affiliation is Bluesphere Bio, Pittsburgh, Pennsylvania. Phil J. Greer’s current affiliation is Ariel Precision Medicine, Pittsburgh, Pennsylvania. David C. Whitcomb, MD, PhD (Conceptualization: Lead; Formal analysis: Supporting; Funding acquisition: Lead; Investigation: Equal; Methodology: Equal; Project administration: Lead; Resources: Lead; Supervision: Lead; Writing – original draft: Equal; Writing – review & editing: Equal). Dongni Fu, BS (Conceptualization: Supporting; Data curation: Supporting; Formal analysis: Lead; Funding acquisition: Supporting; Investigation: Equal; Methodology: Equal; Visualization: Lead; Writing – original draft: Equal; Writing – review & editing: Equal). Brandon M. Blobner, PhD (Data curation: Equal; Formal analysis: Equal; Investigation: Supporting; Methodology: Equal; Visualization: Supporting; Writing – original draft: Supporting; Writing – review & editing: Supporting). Phil J. Greer, MS (Data curation: Supporting; Formal analysis: Supporting; Investigation: Supporting; Methodology: Supporting; Supervision: Supporting; Writing – review & editing: Equal). Celeste Shelton Ohlsen, PhD, CGC (Data curation: Supporting; Formal analysis: Supporting; Writing – review & editing: Equal). Jami Saloman, PhD (Data curation: Supporting; Investigation: Supporting; Writing – review & editing: Equal). Hyun Jung Park, PhD (Methodology: Supporting; Supervision: Supporting; Writing – review & editing: Supporting). Melena Bellin, MD (Data curation: Equal; Funding acquisition: Supporting; Investigation: Supporting; Resources: Equal; Writing – review & editing: Equal). Gregory Beilman, MD (Data curation: Equal; Writing – review & editing: Equal). Kenneth K. Lee, MD (Data curation: Supporting; Writing – review & editing: Equal). Alessandro Paniccia, MD (Data curation: Supporting; Writing – review & editing: Equal). Amer Zureikat, MD (Data curation: Supporting; Writing – review & editing: Equal). Robert Lafyatis, MD (Methodology: Supporting; Project administration: Supporting; Resources: Supporting; Writing – review & editing: Equal). Randall E. Brand, MD (Data curation: Supporting; Methodology: Supporting; Resources: Supporting; Writing – review & editing: Equal). Samer Alkaade, MD (Data curation: Supporting; Writing – review & editing: Equal). Stephen Amann, MD (Data curation: Supporting; Writing – review & editing: Equal). Michelle A. Anderson, MD, MSc (Data curation: Supporting; Writing – review & editing: Equal). Peter Banks, MD (Data curation: Supporting; Writing – review & editing: Equal). Darwin Conwell, MD, MSc (Data curation: Supporting; Writing – review & editing: Equal). Gregory Coté, MD, MS (Data curation: Supporting; Writing – review & editing: Equal). Christopher E. Forsmark, MD (Data curation: Supporting; Writing – review & editing: Supporting). Timothy B. Gardner, MD (Data curation: Supporting; Writing – review & editing: Equal). Andres Gelrud, MD, MMSc (Data curation: Supporting; Writing – review & editing: Equal). Nalini M. Guda, MD (Data curation: Supporting; Writing – review & editing: Equal). Michele Lewis, MD (Data curation: Supporting; Writing – review & editing: Equal). Thiruvengadam Muniraj, MD, PhD (Data curation: Supporting; Writing – review & editing: Equal). Georgios I. Papachristou, MD, PhD (Data curation: Supporting; Writing – review & editing: Equal). Joseph Romagnuolo, MD, MSc (Data curation: Supporting; Writing – review & editing: Equal). Bimaljit S. Sandhu, MD (Data curation: Supporting; Writing – review & editing: Equal). Stuart Sherman, MD (Data curation: Supporting; Writing – review & editing: Equal). Vikesh K. Singh, MD, MSc (Data curation: Supporting; Writing – review & editing: Equal). Adam Slivka, MD, PhD (Data curation: Supporting; Writing – review & editing: Equal). Charles Melbern Wilcox, MD (Data curation: Supporting; Writing – review & editing: Equal). Dhiraj Yadav, MD, MPH (Data curation: Supporting; Formal analysis: Supporting; Writing – review & editing: Equal). Mary Marazita, PhD (Data curation: Supporting; Resources: Supporting; Writing – review & editing: Supporting). Seth Michael Weinberg, PhD (Data curation: Supporting; Resources: Supporting; Writing – review & editing: Supporting). The North American Pancreatitis Study II (NAPS2) is a multicenter, cross-sectional cohort study of RAP and CP in 3 phases. The first phase established the primary cohort,5Whitcomb D.C. et al.Pancreatology. 2008; 8: 520-531Crossref PubMed Scopus (178) Google Scholar followed by a replication cohorte1Conwell D.L. et al.Dig Dis Sci. 2017; 62: 2133-2140Crossref PubMed Scopus (50) Google Scholar for genome-wide association studies.4Whitcomb D.C. et al.Nat Genet. 2012; 44: 1349-1354Crossref PubMed Scopus (236) Google Scholar The current study included patients of European ancestry, as these were the initial cases and controls that were originally genotyped.4Whitcomb D.C. et al.Nat Genet. 2012; 44: 1349-1354Crossref PubMed Scopus (236) Google Scholar Additional well-phenotyped controls from the Pittsburgh, PA region were included with the original NAPS2 controls after comparing genotyping methods and minor allele frequencies of important variants. The additional controls were from the Center for Craniofacial and Dental Genetics, Department of Oral Biology, and Department of Oral and Craniofacial Sciences School of Dental Medicine, University of Pittsburgh, Pittsburgh, PA (see Acknowledgments). The final cohort of cases (RAP: n = 475; CP: n = 866, total: n = 1247) and controls (n = 5691) were similar in ancestry, but the cases were older (controls: 29 ± 19 years; cases: 49 ± 16 years; P < 1e–10), male (controls: 41.3%; cases: 50.2%; P = 8e–10), and had a higher body mass index (controls: 20 ± 13 kg/m2; cases: 25 ± 7.3 kg/m2; P = 1e–10). The NAPS2 cohort (ClinicalTrials.gov number NCT01545167) and NAPS2 genome-wide association study consortium data were used for population-based studies. All blood samples for genomic DNA were obtained from consented patients under Institutional Review Board–approved protocols. DNA was genotyped for a genome-wide association study using the Illumina HumanOmniExpress BeadChip,4Whitcomb D.C. et al.Nat Genet. 2012; 44: 1349-1354Crossref PubMed Scopus (236) Google Scholar and HumanCoreExome BeadChip V1 (Illumina), as described in detail previously.4Whitcomb D.C. et al.Nat Genet. 2012; 44: 1349-1354Crossref PubMed Scopus (236) Google Scholar,6Dunbar E. et al.J Gastroenterol. 2020; 55: 1000-1009Crossref PubMed Scopus (9) Google Scholar Detailed demographic information, additional genotyping, imputation, and analysis methods have also been described previously.4Whitcomb D.C. et al.Nat Genet. 2012; 44: 1349-1354Crossref PubMed Scopus (236) Google Scholar,6Dunbar E. et al.J Gastroenterol. 2020; 55: 1000-1009Crossref PubMed Scopus (9) Google Scholar The initial single nucleotide polymorphism (SNP) array was limited to European ancestries, so we chose patients of European ancestries for population genetics studies. Pancreas tissue samples from consented and well-phenotyped patients undergoing pancreatic surgery were obtained at University of Pittsburgh Medical Center through the Genomic Resources for Enhancing Available Therapies (GREAT1; ClinicalTrials.gov number NCT04306939) study, the Prospective Autos (Prospective Study of Outcomes Following Pancreatectomy and Autologous Islet Transplantation for Treatment of Pancreatic Disease; Institutional Review Board approval number 0609M91887) at the University of Minnesota or the POST (Prospective Observational Study of TPIAT) study (ClinicalTrials.gov number NCT03260387) at the University of Pittsburgh and the University of Minnesota. The tissue from the University of Pittsburgh was used for both total RNA sequencing (RNA-seq) and single-cell RNA-seq (scRNA-seq) and included “normal” pancreas as surgical waste during primary resection of tumors of the distal bile duct (n = 2, Whipple) and mucinous cystic neoplasm of the tail (n = 1). The remaining samples were from total pancreatectomy with islet autotransplantation (TPIAT). TPIAT is performed for RAP or early-established CP while pancreatic function remains (especially islet function) so that significant pancreatic tissue is present after resection. TPIAT samples from Pittsburgh for RAP and duct obstruction with early CP (n = 1), hereditary pancreatitis (n = 1), idiopathic CP (n = 1), and alcohol-associated CP (n = 2). Tissues from Minnesota were from patients undergoing TPIAT for pancreatitis etiologies of hereditary (n = 4), idiopathic (n = 4), and alcohol-associated (n = 1). Pieces of fresh pancreatic tissue samples were divided and immediately processed for histology, RNA-seq using RNAlater or single-cell isolation for scRNA-seq, as described previously.9Blobner B.M. et al.Am J Physiol Gastrointest Liver Physiol. 2021; 321: G449-G460Crossref PubMed Scopus (4) Google Scholar Total RNA extraction and library preparation were performed at the University of Pittsburgh Genomics Research Core using the manufacturer’s recommendations, as described previously.e2Blobner B.M. et al.Pancreas. 2020; 49: 1037-1043Crossref PubMed Scopus (5) Google Scholar Barcoded RNA-seq libraries were sequenced using an Illumina NextSeq 500 (Illumina) with 75-cycle, high-output, paired-end sequencing. Reads were aligned to GRCh38. The specific method of alignment and quality control was the same as described previously.e2Blobner B.M. et al.Pancreas. 2020; 49: 1037-1043Crossref PubMed Scopus (5) Google Scholar Samples for scRNA-seq were minced with sterile scissors and then incubated in 10 mL Dulbecco’s modified Eagle medium/F12 with 1 mg/mL collagenase+protease inhibitor (Sigma, catalog number C6079) at 37°C for 20 minutes. A vortex mixer was used to re-suspend cells every 5 minutes to created single-cell suspensions, as described in detail previously.9Blobner B.M. et al.Am J Physiol Gastrointest Liver Physiol. 2021; 321: G449-G460Crossref PubMed Scopus (4) Google Scholar Cell suspensions were transferred to the University of Pittsburgh Health Sciences Core Research Facilities Single Cell RNA Sequencing Laboratory (Robert Lafyatis, MD, Director) for single-cell complementary DNA library preparation using droplet-based technology from 10X Genomics. Libraries were generated according to the 10X Genomics Chromium Single-Cell 3′ v2 protocol (https://support.10xgenomics.com/single-cell-gene-expression/library-prep/doc/user-guide-chromium-single-cell-3-reagent-kits-user-guide-v2-chemistry). The library was loaded onto an Illumina NextSeq500 with 2 × 75-cycle, paired-end sequencing by the University of Pittsburgh Genomics Research Core. Single-cell libraries were sequenced at the University of Pittsburgh Medical Center Genome Center (Pittsburgh, PA). Transcriptome data are publicly available at: https://github.com/Whitcomb-Lab/scRNA-Seq_analysis. We previously demonstrated the risk for RAP/CP associated with the PRSS1-PRSS2 rs10273639_C haplotype.4Whitcomb D.C. et al.Nat Genet. 2012; 44: 1349-1354Crossref PubMed Scopus (236) Google Scholar Since rs10273639 is in a noncoding region of the locus (5′ to PRSS1 in GRCh37, intronic in GRCh38), we identified a linked SNP within the PRSS1 coding region as a surrogate for the risk/protective haplotype in transcription analysis. PRSS1 SNP rs6667_C (p.N246=) is in linkage disequilibrium with rs10273639 (r2 = <0.8–1.0 depending on cohorts) (Figure 1A). The Integrative Genomics Viewer e3Robinson J.T. et al.Nat Biotechnol. 2011; 29: 24-26Crossref PubMed Scopus (7714) Google Scholar was used for counting SPINK1 rs17107315 C and T allele transcripts and PRSS1-PRSS2 rs6667 C and T transcripts in RNA-seq bam-files. TOPHATe4Trapnell C. et al.Bioinformatics. 2009; 25: 1105-1111Crossref PubMed Scopus (9032) Google Scholar was used for alignment and expression analysis and Seurate5Stuart T. et al.Cell. 2019; 177: 1888-1902.e21Abstract Full Text Full Text PDF PubMed Scopus (4541) Google Scholar was used for single-cell RNA-seq data. Cells expressing <200 and >3000 genes or >10% mitochondrial RNA were excluded. For single-cell analysis, a Jackstraw test was performed on the remaining cells to identify the significant principal components. Cell clusters were identified by feature plots first, then by identifying the cluster markers and submitting the cluster markers to CellMarkere6Zhang X. et al.Nucleic Acids Res. 2019; 47: D721-D728Crossref PubMed Scopus (420) Google Scholar (http://biocc.hrbmu.edu.cn/CellMarker/) and PanglaoDBe7Franzen O. et al.Database (Oxford). 2019; (2019)PubMed Google Scholar (https://panglaodb.se/). UMAP (Uniform Manifold Approximation and Projection)e8McInnes L. et al.arXiv:1802.03426. 2018; Google Scholar was performed for dimensional reduction (https://arxiv.org/abs/1802.03426). STRINGe9von Mering C. et al.Nucleic Acids Res. 2005; 33: D433-D437Crossref PubMed Scopus (1155) Google Scholar was used to conduct pathway analysis (https://string-db.org/). In RNA-seq, relative expression of SPINK1 rs17107315 T and C transcripts were analyzed by Welch 2-sample t test (2-sided) and PRSS1-PRSS2 rs6667 T and C transcripts were analyzed by Wilcoxon signed ranked test because the distribution of the data was non-normal. The frequency of SPINK1 rs17107315 TC, TT, and PRSS1-PRSS2 rs10273639 TT, TC, and CC in genome-wide association studies was determine in the case–control association using the χ2 test. A P value of <.05 was considered statistically significant. Analysis was performed using R, version 3.6.2. A replication study of the SPINK1 prevalence and SPINK1, PRSS1-PRSS2 risk haplotype interaction was performed on tag SNPs in the UK Biobank with pancreatitis patients phenotyped, as reported previously by Ariel Precision Medicine.e10Spagnolo D.M. et al.Clin Transl Gastroenterol. 2022; 19e00455Google Scholar The UK Biobank cohort includes primarily subjects older than 40 years and there are fewer patients with early-onset and idiopathic pancreatitis. Patients with pancreatitis from the UK Biobank (n = 1249) were classified using ICD-10 codes and compared with matched controls (n = 23,256), as described previously.e10Spagnolo D.M. et al.Clin Transl Gastroenterol. 2022; 19e00455Google Scholar The minor allele frequencies of the SPINK1 p.N34S risk allele is 1.3% in controls and 2.5% in CP+RAP (P = 4.095e–06), confirming an association with moderate effect size and replicating the NAPS2 cohort findings. The effect of the PRSS1-PRSS2_rs10273639 genotypes on SPINK1_rs17107315TC (risk) was calculated. As with NAPS2, the risk of RAP/CP with SPINK1_rs17107315TC on the PRSS1-PRSS2_rs10273639TT (protective) background was increased, but did not reach statistical significance (OR, 1.37; 95% CI, 0.51–3.00; P = .37). Unlike NAPS2, the risk of SPINK1_rs17107315TC on the rs10273639TC in the UK Biobank did not reach statistical significance (OR, 1.39; 95% CI, 0.86–2.15; P = .13) but, like NAPS2, the SPINK1_rs17107315TC on the PRSS1-PRSS2_rs10273639CC (risk) background was significant (OR, 2.57; 95% CI, 1.60–3.00; P = 9.16e–05), replicating the NAPS2 findings. The test of interaction between the SPINK1 and PRSS1-PRSS2 risk haplotypes in the UK Biobank data was non-significant, as seen in NAPS2 for either the SPINK1 and PRSS1-PRSS2 risk haplotypes (3 × 2 χ2: P = .11) or the PRSS1-PRSS2 protective (TT) and PRSS1-PRSS1 risk (TC + CC) genotypes and the SPINK1 TT and TC genotypes (2 × 2 χ2; OR, 0.67; 95% CI, 0.25–1.53; P = .47).Supplementary Table 1The Enriched KEGG Pathways in Cluster 6 vs Acinar Cells Based on Gene Expression Signatures Highly Expressed in Cluster of Cells With High PRSS1/SPINK1 RatiosTop 25 pathwaysAcinar cellsCluster 61hsa04971: gastric acid secretionhsa05144: malaria2hsa04976: bile secretionhsa05150: Staphylococcus aureus infection3hsa04657: IL-17 signaling pathwayhsa05152: tuberculosis4hsa04964: proximal tubule bicarbonate reclamationhsa04145: phagosome5hsa05417: lipid and atherosclerosishsa04612: antigen processing and presentation6hsa05208: chemical carcinogenesis - reactive oxygen speciehsa05140: leishmaniasis7hsa04972: pancreatic secretionhsa04610: complement and coagulation cascades8hsa05020: prion diseasehsa05323: rheumatoid arthritis9hsa04915: estrogen signaling pathwayhsa04620: Toll-like receptor signaling pathway10hsa05134: Legionellosishsa04621: NOD-like receptor signaling pathway11hsa04260: cardiac muscle contractionhsa04668: TNF signaling pathway12hsa05167: Kaposi sarcoma–associated herpesvirus infectionhsa05142: Chagas disease (American trypanosomiasis)13hsa04714: thermogenesishsa05164: influenza A14hsa00190: oxidative phosphorylationhsa05321: inflammatory bowel disease15hsa04918: thyroid hormone synthesishsa05310: asthma16hsa04141: protein processing in endoplasmic reticulumhsa05133: pertussis17hsa04612: antigen processing and presentationhsa05143: African trypanosomiasis18hsa04668: TNF signaling pathwayhsa05330: allograft rejection19hsa05012: Parkinson diseasehsa04657: IL-17 signaling pathway20hsa05415: diabetic cardiomyopathyhsa05322: systemic lupus erythematosus21hsa04530: tight junctionhsa04060: cytokine–cytokine receptor interaction22hsa04010: MAPK signaling pathwayhsa04672: intestinal immune network for IgA production23hsa05418: fluid shear stress and atherosclerosishsa04062: chemokine signaling pathway24hsa04024: cAMP signaling pathwayhsa04979: cholesterol metabolism25hsa05166: human T-cell leukemia virus 1 infectionhsa05145: toxoplasmosiscAMP, cyclic adenosine monophosphate; MAPK, mitogen-activated protein kinase; TNF, tumor necrosis factor. Open table in a new tab cAMP, cyclic adenosine monophosphate; MAPK, mitogen-activated protein kinase; TNF, tumor necrosis factor.