Abstract
Objective
Pyoderma gangrenosum (PG) and erythema nodosum (EN) are the most common cutaneous manifestations of inflammatory bowel disease (IBD) but little is known regarding their etiopathogenesis.
Design
We performed a case control study comparing characteristics between IBD patients with a documented episode of PG (PG+) and/or EN (EN+) with those without PG (PG-) and EN (EN-). Data on clinical features were obtained by chart review. IBD related serology was determined using ELISA and genome-wide data generated using Illumina technology. Standard statistical tests for association were used.
Results
We identified a total 92 cases of PG and 103 cases of EN with genetic and clinical characteristics, of which 64 PG and 55 EN were available for serological analyses. Fewer male subjects were identified in the PG(+) (OR 0.6, p=0.009) and EN(+) groups (OR 0.31, p=0<0.0001). Colonic disease, previous IBD related surgery, and non-cutaneous extra-intestinal manifestations were more common among both PG(+) and EN(+) patients compared to controls. PG(+) was associated with ANCA seropositivity (p=0.03) and higher ANCA level (p=0.02) in CD. Genetic associations with PG included known IBD loci (IL8RA (p=0.00003), and PRDM1 (0.03)) as well as with USP15 (4.8×10−6) and TIMP3 (5.6 ×10−7). Genetic associations with EN included known IBD susceptibility genes (PTGER4 (p=8.8×10−4), ITGAL (0.03)) as well as SOCS5 (9.64×10−6), CD207 3.14×10−6), ITGB3 (7.56×10−6) and rs6828740 (4q26)(p <5.0 × 10−8). Multivariable models using clinical, serologic, and genetic parameters predicted PG (AUC 0.8) and EN (AUC 0.97).
Conclusion
Cutaneous manifestations in IBD are associated with distinctive genetic characteristics as well as with the similar clinical characteristics including the development of other extra-intestinal manifestations suggesting shared and distinct etiologies.
Keywords: Crohn's disease, extra-intestinal manifestations, inflammatory bowel disease, pyoderma gangrenosum, erythema nodosum, ulcerative colitis, genetics, serology
Introduction
In addition to gastrointestinal tract pathology, the inflammatory bowel diseases are associated with extra-intestinal manifestations (EIMs). including cutaneous EIMs such as pyoderma gangrenosum (PG) and erythema nodosum (EN). While troublesome in their own right, EIMs may ‘tag’ more homogenous sub-groups within IBD with unique pathogenic characteristics.
PG, observed in 0.5-5% of IBD patients, is reportedly more commonly associated with colonic disease and is a significant source of morbidity requiring intensive, in-patient multi-disciplinary approaches to management [2, 3] [4-7]. PG is a neutrophilic dermatosis and usually begins as a papule or pustule at a site of trauma with a surrounding violacious, undermined border [8]. PG can significantly affect activities of daily living, quality of life and can be cosmetically disfiguring leading to permanent scarring, significant pain, bacterial super-infection, and may require skin grafting [8, 9]. Moreover, potent immunosuppression with therapeutic agents such as intravenous cyclosporine or infliximab may be needed [10, 11]. There is limited data on genetic associations with PG, although small studies have suggested association with PSTPIP1, PTPN6, and TRAF3IP2 [12-14].
EN is the most common dermatologic manifestation in IBD and affects 3-15% of IBD patients [15-17]. Typically diagnosed clinically and rarely before the onset of IBD, lesions most commonly affect the anterior surface of the lower extremities in the form of tender, erythematous, and raised nodules with histology revealing a neutrophilic inflammatory infiltrate and panniculitis [7, 18-20] [1]. Previous studies have noted that EN is also associated with colonic disease [16] and with a more benign IBD natural history [2, 16]. Prior studies have suggested an association between the HLA and EN in IBD [14, 23].
As researchers begin to understand the molecular characteristics of IBD, there is a need to better characterize sub-groups of these complex and heterogeneous conditions. An increased understanding of the etiology of both PG and EN in IBD may highlight potential novel therapeutic pathways for the sub-set of patients who are at risk for skin manifestations. The aim of this study was to better characterize IBD disease behavior, serologic profiles, and genetic associations among patients with PG or EN.
Methods
Subjects
We reviewed two large IBD databases for cases of PG and EN. The Cedars-Sinai Medical Center IBD Research Repository contains clinical (demographics, disease phenotype, disease history, EIMs, treatment records), serologic, and genetic data on over 5000 consented IBD patients followed at our center. From this database, we identified all subjects with genome wide association (GWA) genotyping for inclusion in this study [24, 25]. Those subjects with at least one documented episode of PG [PG(+)] or EN [EN(+)] were identified and underwent a detailed confirmatory medical record review. Subjects without a history of PG [PG(−)]or EN [EN(−)] formed the control group. An episode of PG was confirmed on record review based on either dermatology consultant impression or typical features, as defined elsewhere as having three or more of the following: (i) leg or peristomal location (ii) pathergy (iii) initial pustular lesion (iv) purulent discharge (v) violacious or undermined borders (vi) crater like holes/cribiform scarring [26]. Histology was reviewed, when available, to exclude alternative diagnoses. An episode of EN was confirmed via chart review by dermatology consultation or typical features including erythematous, non-ulcerative, raised, tender bruise-like lesions usually on the legs.
We also included IBD cases with GWA data available from the National Institute for Diabetes and Digestive and Kidney Diseases IBD Genetics Consortium (NIDDK-IBDGC) (Cedars-Sinai Medical Center, Johns Hopkins University, University of Montreal, University of Pittsburgh, University of Toronto, Yale University, and University of Chicago (Data Coordinating Center). The validity and reliability of the clinical features collected in this database has been previously documented [28]. The clinical and genetic data on IBD patients (care was taken to exclude any CSMC patients in the NIDDK cohort in order to avoid double ‘reporting’) were combined with the samples from the CSMC database (Supplementary Figures 1a and 1b). The serologic analysis was limited to the subjects identified in the CSMC IBD Research Repository. The CSMC IBD Research Repository contained 4137 IBD subjects with GWA data. Of these, 64 were PG(+), 55EN(+), and 4073 PG/EN(−). The NIDDK IBDGC database contained 1619 IBD subjects with GWA data (28 PG(+), 48(EN+), and 1591 PG/EN(−)). The genetic and clinical analyses were performed on the combined datasets. Therefore, this analysis consisted of 5756 subjects (92 PG(+), 103 EN(+), and 5664 PG/EN(−)). The serologic analysis was limited to the subjects in the CSMC IBD Research Repository (total 4137 subjects: 64 PG(+), 55 EN(+); and 4073 PG/EN(−)). A total of 14 patients with both EN and PG were identified and were included in both EN and PG analyses.
Clinical and Serologic Phenotyping
Clinical data were collected at all contributing centers by chart review and included patient demographics (age, gender, ethnicity, race), family history, disease duration, disease phenotype (IBD subtype and disease location/behavior according to the Montreal Classification), surgical history, treatment record, smoking status, and presence or absence of other extra-intestinal manifestations [29]. IBD-associated serologies (anti-saccharomyces cerevisiae antibodies (ASCA IgG and IgA), perinuclear anti-nuclear cytoplasmic antibody (pANCA), anti-flagellin (anti-CBir1), anti-outer membrane porin C (anti-OmpC) and anti-Pseudomonas fluorescens-associated sequence I2 (anti-I2) were measured by ELISA, as previously described [30]. Antibody levels were determined and results expressed as ELISA units (EU/mL) that are relative to CSMC laboratory (immunoglobulin [Ig] A-I2, IgA-OmpC) or a Prometheus Laboratory standard (SanDiego, CA; IgA and IgG ASCA) having been derived from a pool of patient sera with well-characterized disease found to have reactivity to these antigens. Quantitation of IgG anti-CBir1 reactivity was expressed in ELISA units derived based on a proportion of reactivity relative to a standardized positive control. Quantitative serologic results were converted to binary variables (i.e. positive or negative). Qualitative positivity to any antibody was defined as being greater than cut-off values 2 standard deviations above mean control titers for each assay. All assays were performed blind to any knowledge of patient clinical characteristics.
Genotyping and genotype quality control
Genotyping was performed at Cedars Sinai Medical Center using Illumina Human610 platforms for CD and HumanCNV370 platforms for UC [23, 24]. The NIDDK IBDGC samples were genotyped using the HumanHap300 and HumanHap550 platforms as previously described [31, 32]. In the genotyped datasets individual and genotype missingness, allele frequencies, and deviations from Hardy-Weinberg Equilibrium were calculated using the PLINK software package (http://pngu.mgh.harvard.edu/∼purcell/plink). Of the 5756 IBD cases, 21 failed quality control (QC) metrics, in which we required a genotyping call rate of > 95%, inbreeding coefficient of < 0.05, and lack of cryptic relatedness which was defined as unexpected relatedness not reflected by known family structure and can sometimes be caused by sample contamination. Analysis of population substructure with Eigenstrat identified 134 subjects that were outside of the main Caucasian cluster on principle components analysis and those subjects were subsequently removed. SNPs with a call rate of < 0.95, with minor allele frequency (MAF) of < 0.01, and that strongly deviated from Hardy-Weinberg equilibrium (P < 1 × 10−7) were removed. After the cleaning of the genotype datasets, for comparison purpose, we imputed SNPs separately for each genotyping platform using IMPUTE2 (mathgen.stats.ox.ac.uk/impute/impute_v2.html) with the HapMap Phase 2 (www.hapmap.org, all population groups combined) serving as the reference. After the imputation, a further stringent cleaning procedure was applied thereby removing any imputed SNP with a certainty score below 0.8, quality score below 0.5 and MAF < 0.01 in any of the cohorts, leaving 2,017,629 SNPs available for analysis.
Statistical Analysis
Univariate analysis
Univariate analyses were conducted to test for differences between PG(+) and PG(−) subjects and between EN(+) and EN(−) subjects. Fisher Exact Test was used to test for associations between PG/EN status and clinical characteristics. Serology was assessed both qualitatively (i.e. percent seropositive) and quantitatively (i.e. median antibody level) using the Fisher's Exact Test and the Wilcoxon Rank Sum Test, respectively. A p-value <0.05 was considered statistically significant.
Genetic analysis
In the imputed datasets, the association between PG and EN status and each single SNP were evaluated using logistic regression with SNPTEST(http://mathgen.stats.ox.ac.uk/genetics_software/snptest/snptest.html) in the Cedars and NIDDK cohorts separately. Top 4 principle components calculated using Eigenstrat (http://genepath.med.harvard.edu/∼reich/Software.htm)were included in the analysis to control for potential confounding effects due to population structure. We then performed a meta-analysis combining the Cedars and NIDDK cohorts to obtain combined estimates using an inverse variance weighting of cohort-specific estimates. Genetic associations with PG and EN were assessed using both a genome wide approach using, given that the skin EIMs are rare conditions, a nominal level of significance of p<5×10−5, and at 99 previously identified IBD susceptibility loci using a level of significance of p<0.05 [33, 34].
Multivariable analysis
Models to predict PG/EN development were created using step-wise multiple logistic regression, combining IBD clinical features, serology, significant IBD susceptibility SNP's, and top GWA hits. The models built for PG analysis included: (1) clinical features alone, (2) serology alone, (3) genetics alone, (4) clinical features + serology, and (5) clinical features + serology + genetic associations. The models build for EN analysis included (1) genetics alone, (2) clinical features alone, and (3) clinical + genetics. The area under the Receiver Operating Characteristic (ROC) curve (AUC [95% CI]) was used to measure the predictive performance of these models for PG and EN.
Ethical Considerations
The Institutional Review Board at Cedars-Sinai Medical Center and the GRCs of the NIDDK IBDGC approved this study and enrolled subjects had provided written informed consent.
Results
Clinical Associations
Pyoderma Gangrenosum
Overall, 92 PG(+) cases were identified among 5756 subjects for the genetic and clinical analysis while the serologic analysis was limited to 64 PG(+) among 4137 samples (see Methods section above and supplementary figures 1a and 1b).
Table 1a shows the demographic and disease features stratified according to PG status. PG was more common in women and patients with PG were diagnosed with IBD at an older age. Among those with CD, fewer subjects had small bowel disease in the PG(+) group but perianal disease and colonic disease were more common compared to the PG(−) group. Eye EIMs (i.e. iritis, uveitis), IBD associated arthritis and ankylosing spondylitis were more prevalent among those with PG.
Table 1. Clinical and Demographic Characteristics of PG Cohort.
| Clinical Feature | PG(+), n = 92 (%) | PG(−), n = 5664 (%) | P | OR (95% CI) |
|---|---|---|---|---|
| Male | 34 (37) | 2825 (51) | 0.009 | 0.6 (0.4–0.9) |
| Crohn's disease | 57 (62) | 3580 (63) | NS | — |
| Ulcerative colitis/IBDU | 35 (38) | 1851 (37) | NS | — |
| Mean age of IBD onset (yr) | 28.5 | 24.8 | 0.026 | — |
| Active smoking | 30 (33) | 869 (26) | NS | — |
| Family history of IBD | 23 (27) | 439 (20) | NS | — |
| Small bowel involvement | 43 (77) | 1403 (88) | 0.021 | 0.5 (0.2–0.9) |
| Perianal CD | 33 (54) | 470 (29) | <0.0001 | 2.8 (1.6–4.7) |
| Colonic disease | 55 (89) | 1148 (71) | 0.001 | 3.2 (1.5–7.1) |
| CD disease behavior | ||||
| B1 | 19 (34) | 623 (39) | NS | — |
| B2 | 17 (30) | 659 (41) | ||
| B3 | 25 (45) | 514 (32) | ||
| Previous IBD-related surgery | 61 (66) | 1629 (48) | <0.0001 | 2.2 (1.4–3.4) |
| IBD-associated arthritis | 26 (28) | 248 (9) | <0.0001 | 3.8 (2.4–6.1) |
| Ankylosing spondylitis | 4 (4) | 31 (1) | 0.029 | 3.8 (1.3–11) |
| Ophthalmologic (iritis, uveitis) | 18 (20) | 93 (4) | <0.0001 | 6.7 (3.8–11.5) |
| Primary sclerosing cholangitis | 2 (2) | 51 (2) | NS | — |
OR, odds ratio; IBDU, inflammatory bowel disease unclassified.
Erythema Nodosum
Overall, 103 EN(+) cases were identified among 5756 subjects for the genetic and clinical analysis while the serologic analysis was limited to 55 EN(+) among 4137 samples (see Methods section and supplementary figures 1a and 1b).
Demographic and clinical characteristics associated with EN are shown in table 1b. Similar to the findings seen in PG, EN co-occured with other EIMs and was more common in women and patients with colonic disease.
Serologic Associations
CD patients with PG were more likely to be ASCA negative, CBir1 negative and ANCA positive as well as having lower median ASCA and higher median ANCA levels in keeping with colonic disease location (supplementary figure 2a and 2b). No significant associations were seen between IBD-associated serologies and EN apart from a borderline association with median anti-CBir1 level (supplementary figures 2c and 2d) in CD cases. No significant serological differences were seen in UC between the two groups (data not shown).
Genetic Associations
Pyoderma Gangrenosum
A number of known IBD susceptibility loci showed significant association with PG including the loci containing IL8RA (p=0.00003), MUC17 (0.01), MMP24 (0.01), WNK2 (0.01), DOCK9 (0.02), PRDM1 (p=0.03) and NDIFIP1 (0.04) (table 3a). While no genetic associations with PG reached a genome wide level of significance, several SNP's met the nominal level of significance of p<5×10−5 (table 2a and supplementary figure 2a).
Table 3. Genetic Associations with PG Assessed at Known IBD Susceptibility Loci.
| SNP | “Risk” Allele | Chromosome | P | OR (95% CI) | Gene |
|---|---|---|---|---|---|
| rs2382817 | C | 2q35 | 0.00003 | 0.4 (0.3–0.6) | IL8RA, TMBIM1, TNS1, GPBAR1, ILSRB |
| rs1734907 | A | 7q22 | 0.01 | 1.8 (1.2–2.7) | MUC17, AP1S1, SERPINE1, VGF, ARS2 |
| rs6088765 | G | 20q11 | 0.01 | 1.6 (1.1–2.3) | MMP24, PROCR, EIF6, EDEM2 |
| rs4743820 | T | 9q22 | 0.01 | 0.6 (0.4–0.9) | WNK2, NINJ1, ZNF484, FGD3, SUSD3 |
| rs95S7195 | T | 13q32 | 0.02 | 0.6 (0.3–0.9) | DOCK9, GPRI8, UBAC2, EBI2, TM9SF2 |
| rs6568421 | G | 6q21 | 0.03 | 1.5 (1.1–2.2) | PRDM1 |
| rs11230563 | T | 11q12 | 0.03 | 1.5 (1.0–2.1) | — |
| rs6863411 | T | 5q31 | 0.04 | 0.7 (0.5–0.9) | SPRY4, PCDHI2, GNPDA1, NDFIP1 |
OR, odds ratio.
Table 2. Clinical and Demographic Characteristics of EN Cohort.
| Clinical Feature | EN(+), n = 103 (%) | EN(−), n = 5664 (%) | P | OR (95% CI) |
|---|---|---|---|---|
| Male | 25 (24) | 2823 (51) | <0.0001 | 0.3 (0.2–0.5) |
| Crohn's disease | 69 (67) | 3578 (66) | NS | — |
| Ulcerative colitis/IBDU | 34 (33) | 1851 (34) | NS | — |
| Mean age of IBD onset (yr) | 23.5 | 24.8 | NS | — |
| Active smoking | 70 (69) | 868 (26) | <0.0001 | 6.2 (4.0–9.5) |
| Family history of JBD | 31 (30) | 439 (20) | 0.012 | 1,8 (1.2–2.7) |
| Small bowel involvement | 36 (49) | 1405 (87) | <0.0001 | 0.1 (0.1–0.2) |
| Perianal CD | 26 (30) | 470 (29) | NS | 1.0 (0.6–1.6) |
| Colonic disease | 77 (88) | 1148 (71) | 0.001 | 2.9 (1.5–5.5) |
| CD disease behavior | ||||
| B1 | 35 (48) | 623 (39) | NS | — |
| B2 | 21 (30) | 659 (41) | ||
| B3 | 29 (41) | 514 (32) | ||
| Previous IBD-related surgery | 68 (66) | 1628 (48) | <0.0001 | 2.1 (1.4–3.2) |
| IBD-associated arthritis | 49 (48) | 248 (9) | <0.0001 | 8.7 (5.8–13.1) |
| Ankylosing spondylitis | 2 (2) | 31 (1) | NS | 3.8 (1.3–11) |
| Ophthalmologic (iritis, uveitis) | 26 (25) | 93 (4) | <0.0001 | 9.2 (5.6–15) |
| Primary sclerosing cholangitis | 2 (2) | 34 (3) | NS | — |
OR, odds ratio; TBDU, inflammatory bowel disease unclassified.
Erythema Nodosum
A number of known IBD susceptibility loci showed significant association with EN including PTGER4 (p=8.7 × 10−4), ITGAL (0.03), and IKZF1 (0.03) (table 3b). There were several SNP's that met the nominal level of genome-wide significance of p<5×10−5 (table 2b, supplementary figure 2b). A SNP on chromosome 4q26 reached genome wide level (p = 4.94×10−8).
Multivariable Logistic Regression Analysis
Pyoderma Gangrenosum
Multiple logistic multiple regression was performed to create predictive models for PG development (figure 1a). Model 1 examined the importance of genetic associations in predicting PG. The SNP's rs2382817 (IL8RA) and rs130555 (TIMP3) remained significant in the model, with an AUC of 0.73 [95% CI 0.7-0.76]. Model II examined clinical features in predicting PG and a history of perianal disease and eye manifestations remained significant. Model III examined the significance of serology on PG, and p-ANCA status remained significant in this model while Model IV involved the combination of clinical features and serology. Finally, Model V examined the combination of clinical, serologic, and genetic data for predicting PG, and performed the best with an AUC of 0.80 [0.73-0.87].
Figure 1.
A, Multivariable model for predicting PG development. B, Multivariable model for predicting EN development.
Erythema Nodosum
Similar models were created for EN (figure 1b). Model 1 examined the predictive value of genetic associations in EN (AUC, 0.76 [95% CI 0.71-0.80]) and it appeared that using genetics alone was modestly predictive of EN development. Model II employed clinical features including gender, small bowel involvement, IBD related arthritis, and eye involvement with significantly greater success in predicting EN (AUC, 0.95 [95% CI 0.94-0.97]). Model III, a combination of clinical and genetic parameters, showed only a modest improvement than model II (AUC 0.97).
Discussion
Extra-intestinal manifestations (EIMs) can often be more debilitating and difficult to treat than the underlying IBD and the ability to predict the development of these lesions could have implications on management. For example, patients at high-risk of PG may be educated regarding the importance of early detection and reporting of skin lesions. An increased understanding of the pathogenesis of EIMs may define sub-sets of disease and potentially identify novel therapeutic targets. The current study is among the first and largest in the IBD literature to characterize genetic, clinical and serologic associations of both PG and EN.
In keeping with previous findings, skin EIMs occur more commonly in women[23,24]. We observed that patients with PG were diagnosed with IBD at an older age and developed PG much later into IBD disease course, with a mean onset of 15 years after IBD diagnosis (data not shown), which is longer than the 2-7 years previously reported in the literature [36, 37]. Earlier studies have suggested that PG was more common in UC, however more recent studies have shown stronger associations with CD [6, 36, 38]. In a population-based study, Bernstein et al. showed a prevalence of PG of 1.3% among CD patients compared to 0.8% among those with UC [6]. However, we found no difference in PG episodes among IBD disease type i.e. CD or UC in our cohort of patients largely collected from tertiary centers. While there are conflicting data, colonic disease has been more consistently demonstrated to be associated with EIM's [17, 36, 39] and our findings support this in both EN and PG. It has been proposed that the predominance of colonic disease in a subject with EIM's may be related to interactions between colonic disease and the microbiome with subsequent cross-reacting of skin, joint, eye, etc. antigens with gut antigens [40]. A number of studies have demonstrated the co-occurrence of EIMs and this phenomenon is present in the cohorts from this study [17, 39]. In one study from Switzerland, patients with IBD-associated arthritis had an OR of 2.9 for subsequent PG development [17]. Moreover, certain EIM's tended to cluster, with EN and musculoskeletal manifestations commonly occurring together and PG occurring together with inflammatory eye manifestations. In this study, subjects with PG had a higher prevalence of iritis, uveitis, IBD associated arthritis, and ankylosing spondylitis (AS) and EN was associated with oral ulcers, IBD related arthritis, and ophthalmologic involvement providing additional evidence of common underlying pathophysiology to the IBD EIMs.
The genetic contributions to the EIMs remain unclear, however, we have identified several putative genetic associations (from the GWA approach) between IBD and PG including GPBAR1 which encodes a G-protein involved in bile salt absorption, highly expressed in the ileum and colon, and interestingly, associated with PSC [44, 45] providing additional evidence of shared etiology for the EIMs. PG was also associated with TIMP3, an important inhibitor of metalloproteinase involved in the degradation of the extracellular matrix [46]. Given the large contribution of the extracellular matrix to skin morphology, this may be a particularly pertinent finding [47]. We also identified associations between known IBD susceptibility genes and PG including IL8RA, a known mediator of neutrophil migration [33, 34, 48] which may be highly relevant given the role of neutrophils in PG pathogenesis. We identified several SNP's that met our nominal level of for genome wide significance (p<5×10−5) and one SNP (rs6828740) reached the highly stringent level for genome wide level of significance (p<5×10−8) for association with EN. This SNP, rs6828740, is located on chromosome 4q26 in a gene desert highly conserved across species that contains a binding site for PRDM1, a known IBD and autoimmune disease ‘gene’ also associated with PG development in this study (table 1b). Other putatively associated genes of interest include: SOCS5 (2p21), a negative regulator of cytokine signaling [59]; CD207 (2p13), a non-classical antigen-processing pathway gene that facilitates uptake of antigens for presentation to T-cells [60]; and ITGB3 (17q21), an integrin involved with cell adhesion and cells-surface mediated signaling that is associated with asthma and diabetes susceptibility [61]. Many of the genes that we have identified play a role in antigen presentation and cytokine signaling/production that contribute both, to the inflammation of the intestinal tract during IBD flares, as well as to the lympho-histiocytic infiltrate of the dermis and panniculitis seen with EN lesions [1].Our multivariable model for PG showed that the combination of clinical features, serology, and genetic associations performed better in predicting PG development than any variable alone. Our results add to the growing number of published examples of predictive multiplatform models. [49, 50].
We recognize that this study, despite being the largest of its type performed to date, has a number of limitations including the sample size. Furthermore, there is no definitive definition or definitive diagnostic test for PG or EN and all studies to date have been based on clinical impression from typical dermatologic features [8]. The diagnoses of PG and EN, in this study, were made by experienced IBD clinicians or dermatologists based at large tertiary referral IBD centers in North America. Reassuringly, our clinical associations with skin manifestations replicate results from other studies as well as identifying novel demographic, clinical, serological, and genetic associations. These findings should be replicated both independently and through combination with this existing cohort. The multivariable model represent an interesting and potentially powerful example of multiplatform predictive tools.
In summary, we have better characterized the clinical, serologic, and genetic associations in IBD patients who develop PG or EN. While they are clearly distinct entities from one another, we demonstrate that there are shared associations suggesting common underlying mechanisms and pathways. These findings may also suggest potential novel therapeutic targets for managing these extra-intestinal manifestations including TIMP inhibitors and anti-cell adhesion strategies.
Supplementary Material
Table 4. Genetic Associations with PG Using a Genome-Wide Approach.
| SNP | “Risk” Allele | Chromosome | P | OR (95% CI) | Genes |
|---|---|---|---|---|---|
| rs10184704 | C | 2q14 | 6.3 × 10−6 | 2.3 (1.6–3.3) | GYPC |
| rs13392177 | C | 2q35 | 4.8 × 10−6 | 2.3 (1.6–3.4) | APRC2/GPBAR1/HMG1L9 |
| rs161857 | T | 5q23 | 7.3 × 10−6 | 3.0 (1.9–4.9) | TNFAIP8/HSD17B4/FABP5L6 |
| rs11174436 | A | 12q14 | 4.6 × 10−6 | 2.3 (1.6–3.4) | FAM19A, USP15, MON2 |
| rs12607055 | C | 18q12 | 1.7 × 10−6 | 3.5 (2.1–5.9) | RIT2/SYT4 |
| rs17770334 | C | 18q21 | 1.8 × 10−6 | 3.1 (1.9–4.8) | SERP1NB3/4/5J7J11J13, BCL2, VPS4B, KDSR |
| rs130555 | C | 22q12 | 5.6 × 10−7 | 2.5 (1.7–3.5) | TIMP3 |
OR, odds ratio.
Table 5. Genetic Associations with EN Assessed at Known IBD Susceptibility Loci.
| SNP | “Risk” Allele | Chromosome | P | OR (95% CI) | Gene |
|---|---|---|---|---|---|
| rs11742570 | C | 5p13 | 8.77 × 10−4 | 0.56 (0.4–0.8) | PTGER4 |
| rs11150589 | T | 16p11 | 0.027 | 1.41 (1.0–1.9) | 1TGAL |
| rs1456896 | C | 7p12 | 0.03 | 1.6 (1.1–2.3) | IKZF1 |
OR, odds ratio.
Table 6. Genetic Associations with EN Using a Genome-Wide Approach.
| SNP | “Risk” Allele | Chromosome | P | OR (95% CI) | Gene |
|---|---|---|---|---|---|
| rs7586612 | A | 2p21 | 9.64 × 10−6 | 2.01 (1.5–2.7) | SOCS5, MCFD2, TTC7A, CALM2 |
| rs390966 | G | 2p13 | 3.14 × 10−6 | 2.17 (1.6–3.0) | CD207, NAGK, CLEC4F, MCEE |
| rs6828740 | C | 4q26 | 4.94 × 10−8 | 4.14 (2.5–6.9) | — |
| rs9405444 | A | 6p25 | 8.07 × 10−6 | 2.21 (1.6–3.1) | FOXQ1 |
| rs10138297 | G | 14q21 | 7.29 10−6 | 2.63 (1.7–4.0) | ARF6, METTL21D, SOS2 |
| rs11079757 | A | 17 | 7.6 × 10−6 | 2.2 (1.6–3.1) | JTGB3, MYL4, CCD27 |
OR, odds ratio.
Acknowledgments
Funding: This study was supported by The Helmsley Charitable Trust. IBD Research at Cedars-Sinai is supported by USPHS grant PO1DK046763 and the Cedars-Sinai F. Widjaja Foundation Inflammatory Bowel andImmunobiology Research Institute Research Funds. Genotyping at CSMC is supported in part by the National Center for Research Resources (NCRR) grant M01-RR00425, UCLA/Cedars-Sinai/Harbor/Drew Clinical and Translational Science Institute (CTSI) Grant (UL1 TR000124-01), the Southern California Diabetes and Endocrinology Research Grant (DERC) (DK063491). Project investigators are supported by The Helmsley Charitable Trust (D.P.B.M.), The European Union (D.P.B.M.), The Crohn's and Colitis Foundation of America (CCFA) (D.P.B.M.), The Feintech Family Chair in IBD (S.R.T.), The Joshua L. and Lisa Z. Greer Chair in IBD Genetics (D.P.B.M.), The Abe and Claire Levine Chair in Pediatric IBD (MD), and grants, DK062413, DK046763-19, AI067068, HS021747 (D.P.B.M.).
Footnotes
AW and BH performed the chart review and wrote the first draft of the manuscript. DB performed the non-genetic and modeling associated statistical analyses. SRT recruited patients and obtained funding for the study. DPM, MD, PF, AI, MK, SB, IO, RD, JR, MS, and EV recruited and phenotyped subjects. DP phenotyped study subjects. JIR and TH performed and supervised genotyping and QCing of genotype data. DS, GYM, and DPM conceived and supervised the study.
Disclosures: Marla Dubinsky and Mark Silverberg, consultant Prometheus Labs
References
- 1.Huang BL, Chandra S, Shih DQ. Skin manifestations of inflammatory bowel disease. Front Physiol. 2012;3:13. doi: 10.3389/fphys.2012.00013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Tromm A, et al. Cutaneous manifestations in inflammatory bowel disease. Z Gastroenterol. 2001;39(2):137–44. doi: 10.1055/s-2001-11153. [DOI] [PubMed] [Google Scholar]
- 3.McCallum DI, Kinmont PD. Dermatological manifestations of crohn's disease. Br J Dermatol. 1968;80(1):1–8. doi: 10.1111/j.1365-2133.1968.tb11898.x. [DOI] [PubMed] [Google Scholar]
- 4.Lindgren A, Wallerstedt S, Olsson R. Prevalence of Crohn's disease and simultaneous occurrence of extraintestinal complications and cancer. An epidemiologic study in adults. Scand J Gastroenterol. 1996;31(1):74–8. doi: 10.3109/00365529609031630. [DOI] [PubMed] [Google Scholar]
- 5.Greenstein AJ, Janowitz HD, Sachar DB. The extra-intestinal complications of Crohn's disease and ulcerative colitis: a study of 700 patients. Medicine (Baltimore) 1976;55(5):401–12. doi: 10.1097/00005792-197609000-00004. [DOI] [PubMed] [Google Scholar]
- 6.Bernstein CN, et al. The prevalence of extraintestinal diseases in inflammatory bowel disease: a population-based study. Am J Gastroenterol. 2001;96(4):1116–22. doi: 10.1111/j.1572-0241.2001.03756.x. [DOI] [PubMed] [Google Scholar]
- 7.Mir-Madjlessi SH, Taylor JS, Farmer RG. Clinical course and evolution of erythema nodosum and pyoderma gangrenosum in chronic ulcerative colitis: a study of 42 patients. Am J Gastroenterol. 1985;80(8):615–20. [PubMed] [Google Scholar]
- 8.Powell FC, Su WP, Perry HO. Pyoderma gangrenosum: classification and management. J Am Acad Dermatol. 1996;34(3):395–409. doi: 10.1016/s0190-9622(96)90428-4. quiz 410-2. [DOI] [PubMed] [Google Scholar]
- 9.Thornton JR, et al. Pyoderma gangrenosum and ulcerative colitis. Gut. 1980;21(3):247–8. doi: 10.1136/gut.21.3.247. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Friedman S, et al. Intravenous cyclosporine in refractory pyoderma gangrenosum complicating inflammatory bowel disease. Inflamm Bowel Dis. 2001;7(1):1–7. doi: 10.1097/00054725-200102000-00001. [DOI] [PubMed] [Google Scholar]
- 11.Regueiro M, et al. Infliximab for treatment of pyoderma gangrenosum associated with inflammatory bowel disease. Am J Gastroenterol. 2003;98(8):1821–6. doi: 10.1111/j.1572-0241.2003.07581.x. [DOI] [PubMed] [Google Scholar]
- 12.Lin Z, et al. Failure of anakinra treatment of pyoderma gangrenosum in an IBD patient and relevance to the PSTPIP1 gene. Inflamm Bowel Dis. 2011;17(6):E41–2. doi: 10.1002/ibd.21684. [DOI] [PubMed] [Google Scholar]
- 13.Nesterovitch AB, et al. Alteration in the gene encoding protein tyrosine phosphatase nonreceptor type 6 (PTPN6/SHP1) may contribute to neutrophilic dermatoses. Am J Pathol. 2011;178(4):1434–41. doi: 10.1016/j.ajpath.2010.12.035. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ciccacci C, et al. TRAF3IP2 gene is associated with cutaneous extraintestinal manifestations in Inflammatory Bowel Disease. J Crohns Colitis. 2012 doi: 10.1016/j.crohns.2012.02.020. [DOI] [PubMed] [Google Scholar]
- 15.Lebwohl M, Lebwohl O. Cutaneous manifestations of inflammatory bowel disease. Inflamm Bowel Dis. 1998;4(2):142–8. doi: 10.1002/ibd.3780040209. [DOI] [PubMed] [Google Scholar]
- 16.Timani S, Mutasim DF. Skin manifestations of inflammatory bowel disease. Clin Dermatol. 2008;26(3):265–73. doi: 10.1016/j.clindermatol.2007.10.018. [DOI] [PubMed] [Google Scholar]
- 17.Vavricka SR, et al. Frequency and risk factors for extraintestinal manifestations in the Swiss inflammatory bowel disease cohort. Am J Gastroenterol. 2011;106(1):110–9. doi: 10.1038/ajg.2010.343. [DOI] [PubMed] [Google Scholar]
- 18.Evans PE, Pardi DS. Extraintestinal manifestations of inflammatory bowel disease: focus on the musculoskeletal, dermatologic, and ocular manifestations. MedGenMed. 2007;9(1):55. [PMC free article] [PubMed] [Google Scholar]
- 19.Trost LB, McDonnell JK. Important cutaneous manifestations of inflammatory bowel disease. Postgrad Med J. 2005;81(959):580–5. doi: 10.1136/pgmj.2004.031633. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Weinstein M, Turner D, Avitzur Y. Erythema nodosum as a presentation of inflammatory bowel disease. CMAJ. 2005;173(2):145–6. doi: 10.1503/cmaj.050201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Orchard T. Extraintestinal complications of inflammatory bowel disease. Curr Gastroenterol Rep. 2003;5(6):512–7. doi: 10.1007/s11894-003-0042-6. [DOI] [PubMed] [Google Scholar]
- 22.Kaufman I, et al. The effect of infliximab on extraintestinal manifestations of Crohn's disease. Rheumatol Int. 2005;25(6):406–10. doi: 10.1007/s00296-004-0467-8. [DOI] [PubMed] [Google Scholar]
- 23.Orchard TR, et al. Uveitis and erythema nodosum in inflammatory bowel disease: clinical features and the role of HLA genes. Gastroenterology. 2002;123(3):714–8. doi: 10.1053/gast.2002.35396. [DOI] [PubMed] [Google Scholar]
- 24.McGovern DP, et al. Genome-wide association identifies multiple ulcerative colitis susceptibility loci. Nat Genet. 2010;42(4):332–7. doi: 10.1038/ng.549. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.McGovern DP, et al. Fucosyltransferase 2 (FUT2) non-secretor status is associated with Crohn's disease. Hum Mol Genet. 2010;19(17):3468–76. doi: 10.1093/hmg/ddq248. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Hadi A, Lebwohl M. Clinical features of pyoderma gangrenosum and current diagnostic trends. J Am Acad Dermatol. 2011;64(5):950–4. doi: 10.1016/j.jaad.2010.01.049. [DOI] [PubMed] [Google Scholar]
- 27.Diseases, N.I.o.D.a.D.a.K. Inflammatory Bowel Disease Genetics Consortium. 2012 [cited 2012 2012]; Available from: http://medicine.yale.edu/intmed/ibdgc.
- 28.Dassopoulos T, et al. Assessment of reliability and validity of IBD phenotyping within the National Institutes of Diabetes and Digestive and Kidney Diseases (NIDDK) IBD Genetics Consortium (IBDGC) Inflamm Bowel Dis. 2007;13(8):975–83. doi: 10.1002/ibd.20144. [DOI] [PubMed] [Google Scholar]
- 29.Silverberg MS, et al. Toward an integrated clinical, molecular and serological classification of inflammatory bowel disease: Report of a Working Party of the 2005 Montreal World Congress of Gastroenterology. Can J Gastroenterol. 2005;19(Suppl A):5–36. doi: 10.1155/2005/269076. [DOI] [PubMed] [Google Scholar]
- 30.Dubinsky MC, et al. Increased immune reactivity predicts aggressive complicating Crohn's disease in children. Clin Gastroenterol Hepatol. 2008;6(10):1105–11. doi: 10.1016/j.cgh.2008.04.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Duerr RH, et al. A genome-wide association study identifies IL23R as an inflammatory bowel disease gene. Science. 2006;314(5804):1461–3. doi: 10.1126/science.1135245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Silverberg MS, et al. Ulcerative colitis-risk loci on chromosomes 1p36 and 12q15 found by genome-wide association study. Nat Genet. 2009;41(2):216–20. doi: 10.1038/ng.275. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Anderson CA, et al. Meta-analysis identifies 29 additional ulcerative colitis risk loci, increasing the number of confirmed associations to 47. Nat Genet. 2011;43(3):246–52. doi: 10.1038/ng.764. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Franke A, et al. Genome-wide meta-analysis increases to 71 the number of confirmed Crohn's disease susceptibility loci. Nat Genet. 2010;42(12):1118–25. doi: 10.1038/ng.717. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Kattan MW, Scardino PT. Prediction of progression: nomograms of clinical utility. Clin Prostate Cancer. 2002;1(2):90–6. doi: 10.3816/cgc.2002.n.010. [DOI] [PubMed] [Google Scholar]
- 36.Polcz M, Gu J, Florin T. Pyoderma gangrenosum in inflammatory bowel disease: the experience at Mater Health Services' Adult Hospital 1998-2009. J Crohns Colitis. 2011;5(2):148–51. doi: 10.1016/j.crohns.2010.10.006. [DOI] [PubMed] [Google Scholar]
- 37.Veloso FT. Extraintestinal manifestations of inflammatory bowel disease: do they influence treatment and outcome? World J Gastroenterol. 2011;17(22):2702–7. doi: 10.3748/wjg.v17.i22.2702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Powell FC, et al. Pyoderma gangrenosum: a review of 86 patients. Q J Med. 1985;55(217):173–86. [PubMed] [Google Scholar]
- 39.Monsen U, et al. Extracolonic diagnoses in ulcerative colitis: an epidemiological study. Am J Gastroenterol. 1990;85(6):711–6. [PubMed] [Google Scholar]
- 40.Bhagat S, Das KM. A shared and unique peptide in the human colon, eye, and joint detected by a monoclonal antibody. Gastroenterology. 1994;107(1):103–8. doi: 10.1016/0016-5085(94)90066-3. [DOI] [PubMed] [Google Scholar]
- 41.Ruemmele FM, et al. Diagnostic accuracy of serological assays in pediatric inflammatory bowel disease. Gastroenterology. 1998;115(4):822–9. doi: 10.1016/s0016-5085(98)70252-5. [DOI] [PubMed] [Google Scholar]
- 42.Quinton JF, et al. Anti-Saccharomyces cerevisiae mannan antibodies combined with antineutrophil cytoplasmic autoantibodies in inflammatory bowel disease: prevalence and diagnostic role. Gut. 1998;42(6):788–91. doi: 10.1136/gut.42.6.788. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Vasiliauskas EA, et al. Perinuclear antineutrophil cytoplasmic antibodies in patients with Crohn's disease define a clinical subgroup. Gastroenterology. 1996;110(6):1810–9. doi: 10.1053/gast.1996.v110.pm8964407. [DOI] [PubMed] [Google Scholar]
- 44.Hov JR, et al. TGR5 sequence variation in primary sclerosing cholangitis. Dig Dis. 2011;29(1):78–84. doi: 10.1159/000324138. [DOI] [PubMed] [Google Scholar]
- 45.Information, T.N.C.f.B. GPBAR1 G protein-coupled bile acid receptor 1 [Homo sapiens] 2012 [cited 2012 2012]; Available from: http://www.ncbi.nlm.nih.gov/gene/151306.
- 46.Information, T.N.C.f.B. TIMP3 TIMP metallopeptidase inhibitor 3 [Homo sapiens] 2012 [cited 2012 2012]; Available from: http://www.ncbi.nlm.nih.gov/gene/7078.
- 47.Alberts B, et al. Cell junctions, cell adhesion, and the extracellular matrix. Molecular Biology of the Cell. 2002:1065–125. [Google Scholar]
- 48.Wu Y, et al. A chemokine receptor CXCR2 macromolecular complex regulates neutrophil functions in inflammatory diseases. J Biol Chem. 2012;287(8):5744–55. doi: 10.1074/jbc.M111.315762. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Lichtenstein GR, et al. Combination of genetic and quantitative serological immune markers are associated with complicated Crohn's disease behavior. Inflamm Bowel Dis. 2011;17(12):2488–96. doi: 10.1002/ibd.21661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Dubinsky MC, et al. Genome wide association (GWA) predictors of anti-TNFalpha therapeutic responsiveness in pediatric inflammatory bowel disease. Inflamm Bowel Dis. 2010;16(8):1357–66. doi: 10.1002/ibd.21174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Bevilacqua JL, et al. Doctor, what are my chances of having a positive sentinel node? A validated nomogram for risk estimation. J Clin Oncol. 2007;25(24):3670–9. doi: 10.1200/JCO.2006.08.8013. [DOI] [PubMed] [Google Scholar]
- 52.Danese S, et al. Extraintestinal manifestations in inflammatory bowel disease. World J Gastroenterol. 2005;11(46):7227–36. doi: 10.3748/wjg.v11.i46.7227. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Goudet P, et al. Characteristics and evolution of extraintestinal manifestations associated with ulcerative colitis after proctocolectomy. Dig Surg. 2001;18(1):51–5. doi: 10.1159/000050097. [DOI] [PubMed] [Google Scholar]
- 54.Fleisher M, et al. Infliximab in the treatment of steroid refractory erythema nodosum of IBD. Gastroenterology. 2002;122(Suppl 1):A618. [Google Scholar]
- 55.Mendoza JL, et al. [Extraintestinal manifestations in inflammatory bowel disease: differences between Crohn's disease and ulcerative colitis] Med Clin (Barc) 2005;125(8):297–300. doi: 10.1157/13078423. [DOI] [PubMed] [Google Scholar]
- 56.Georgiou S, et al. Cutaneous Manifestations of Inflammatory Bowel Disease. Hospital Chronicles. 2006;1(3):158–168. [Google Scholar]
- 57.Veloso FT, Carvalho J, Magro F. Immune-related systemic manifestations of inflammatory bowel disease. A prospective study of 792 patients. J Clin Gastroenterol. 1996;23(1):29–34. doi: 10.1097/00004836-199607000-00009. [DOI] [PubMed] [Google Scholar]
- 58.Papp M, et al. Utility of serological markers in inflammatory bowel diseases: gadget or magic? World J Gastroenterol. 2007;13(14):2028–36. doi: 10.3748/wjg.v13.i14.2028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Information, T.N.C.f.B. SOCS5 suppressor of cytokine signaling 5 [Homo sapiens] 2012 [cited 2012 2012]; Available from: http://www.ncbi.nlm.nih.gov/gene/9655#reference-sequences.
- 60.Information, T.N.C.f.B. CD207 molecule, langerin [Homo sapiens] 2012 [cited 2012 2012]; Available from: http://www.ncbi.nlm.nih.gov/gene/50489.
- 61.Information, T.N.C.f.B. ITGB3 integrin, beta 3 (platelet glycoprotein IIIa, antigen CD61) [Homo sapiens] 2012 [cited 2012 2012]; Available from: http://www.ncbi.nlm.nih.gov/gene/3690.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.