The genetic underpinnings of ulcerative colitis relative to Crohn's disease

[Guest guest]

Dear All;

This article gives an excellent summary of the incredible progress

that has been made in understanding the genetics of inflammatory

bowel disease during the last decade, and especially during just the

last 12 months. A lot of the recent results are pointing towards Th17

cells as potential culprits in both Crohn's disease (CD) and

ulcerative colitis (UC) [the authors refer to the Th17/IL-23 pathway;

with IL-23 being involved in stimulating Th17 cell development].

Hopefully, researchers can now begin to compare and contrast UC and

CD susceptibility genes with those associated with PSC/UC and PSC/CD,

then focus on developing medications that are selective for the key

pathways involved.

Best regards,

Dave

(father of (23); PSC 07/03; UC 08/03)

________________________________________________

Gastroenterology [in Press] (2009)

Exposed: the genetic underpinnings of ulcerative colitis relative to

Crohn's disease.

Brant SR

R. Brant, M.D.

Meyerhoff Inflammatory Bowel Disease Center, Department of Medicine,

School of Medicine, Department of Epidemiology, Bloomberg School of

Public Health, s Hopkins University, Baltimore, land

Address requests for reprints to: R. Brant, M.D., Meyerhoff

IBD Center, s Hopkins University, 1501 E. Jefferson St., B136,

Baltimore, MD 21231

An Editorial to accompany publication of et al.

Investigation of Crohn's Disease Risk loci in Ulcerative Colitis

further Defines their Molecular Relationship. Gastroenterology, In

Press.

published online 05 January 2009.

Uncorrected Proof

There are many important reasons to determine why inflammatory bowel

disease (IBD) develops and what explains the distinct but related

nature of the two major IBD phenotypes, Crohn's disease (CD) and

ulcerative colitis (UC). IBD incidence continues to increase. New

biological treatments achieve remission in less than 50%, and risk

potential life-threatening complications. Intestinal resection is

frequently required for management of refractory disease and

complications. IBD is often disabling. It decreases fecundity and

successful pregnancy outcomes. Furthermore, offspring of IBD patients

are much more likely to suffer their parents' fate; there are no

methods of prevention in sight.

Defining the genetic basis of IBD will be a major step to determine

those at greatest risk of developing IBD and will uncover

pathophysiological mechanisms to apply focused interventions for

prevention and treatment. It may help predict disease course and

complications, and clarify the underlying nature of IBD type

unclassified or " indeterminate colitis " . Additionally, far too many

patients with seemingly classic UC undergo " curative " proctocolectomy

only to later develop CD.

Increased familial clustering and especially increased monzyogotic

vs. dizygotic twin concordance predicted the existence of IBD

susceptibility genes. Orholm and colleagues observed that the

increased risk of developing CD in relatives of CD patients is 10-

fold greater than that of controls, and for UC 8-fold greater,

whereas the cross disease risk (CD in a relative of a UC patient and

vice-versa) is 2-fold and 4-fold, respectively.1 This suggested that

some genes would be specific to CD, others to UC, but some would be

common to both types of IBD. Indeed, " mixed " IBD relative pairs (e.g.

CD-UC relative pairs) are relatively common, comprising one-fifth to

one-third of all IBD pairs collected for genetic linkage studies.2

Conversely, reports of CD-UC monozygotic twins are exceedingly rare

suggesting that the specific complement of IBD genes will result in

one IBD phenotype or the other.

Could discovery of IBD genes clarify the puzzle of IBD phenotype?

Early success in IBD genetics research was encouraging. The first

consistent IBD risk alleles (DNA variants associated with IBD risk)

were HLA molecules, with UC associated serotypes HLA-DR2 (DRB1*1501

and 1502 genotype) and HLA-DR1 (DRB1*0103 genotype) being the most

consistent. Interestingly, these HLA associations provide a microcosm

of the distinct as well as related genetic natures of UC and CD: HLA

DRB1*0103 was also associated with CD, but specifically with colonic

CD. By contrast, DR2 was inversely associated with CD. However, HLA

risk alleles in IBD completed only a minor part of the genetic

heritability puzzle, having at most a 3- to 4-fold risk.

Complementing candidate gene analyses, linkage mapping approaches,

commencing in 1996, yielded several consistent familial IBD loci

(chromosomal regions that co-segregate with the IBD phenotype in

families). The first linkage locus identified, IBD1 on chromosome 16,

led to the identification of NOD2 as a major risk factor for CD, with

20-fold risk in homozygote mutation carriers and 3-fold risk in

heterozygote mutation carriers. Research soon established that NOD2

genotypes have specificity for CD as opposed to UC. Yet,

complementing also HLA-DRB1*0103, NOD2 is an independent risk factor

for ileal disease site (and complications of stricturing and internal

fisulization, Montreal classifications B2 and B3). The NOD2 discovery

also fulfilled the promise of genetics to identify mechanisms of IBD

pathophysiology: indeed, NOD2 kindled research into innate immunity

as a major factor in CD pathophysiology. NOD2 specificity for CD

highlighted the importance of gut bacteria and complemented the

association of antibodies against various gut micro-organisms and

effectiveness of antibiotic therapy for CD not UC.

NOD2, however, was only involved in 25% of CD population risk in

whites and not present in Asians. Fortunately, the advent of genome

wide association studies (GWAS) mapping has yielded a previously

unimaginable number of CD susceptibility genes. In this approach,

investigators genotype hundreds of thousands of " tagging " single

nucleotide polymorphisms (SNPs) located throughout the genome. The

first CD GWA success was the discovery in 2005 of variations of tumor

necrosis factor superfamily gene 15 (TNFSF15), using a low density

GWAS platform of only 90,000 SNPs, as the first proven gene for Asian

CD and a slightly lower risk gene for European CD.3 TNFSF15 can

induce nuclear factor kappa-B and secretion of interferon gamma in T-

cells that express death domain receptor 3 (DR3). This study has been

followed by 9 separate European ancestry GWA studies that either

examined 10,000 to 20,000 non-synonymous (amino acid altering) SNPs

(a German CD and a Wellcome Trust Case Control Consortium of the

United Kingdom [WTCCC-UK Consortium] UC studies), or 100,000 to

550,000 SNPs for association with CD (5 studies: NIDDK [National

Institutes of Diabetes Digestion and Kidney Diseases - USA and

Canadian populations], Belgian-French, WTCCC, Quebec, and German) UC

(European collaborative) and pediatric IBD (USA and European

collaborative study). Individually these studies have established 13

unique IBD loci with genome-wide evidence of association.3, 4, 5, 6, 7

The first new genes discovered, interleukin (IL) 23 receptor (IL23R)

having the next strongest risk for CD after NOD2 - 2 to 4 fold risk)

and the intestinal epithelial expressed autophagy gene, ATG16L1 (1.3-

2.0 fold risk), identified consecutively in the NIDDK and the German

non-synonymous screens, both delivered on the promise of IBD

genetics. Concurrent CD immunology and animal model research, and the

therapeutic effectiveness of anti-p40, IL23/IL12 cytokine subunit

antibodies showed that IL23 was an important cytokine for IBD chronic

inflammation. The strong association with CD and UC put IL23 related

genes and the Th17 pathway as a major focus of IBD research. The most

highly associated allele, R281Q, is also associated with psoriasis,

potentially clarifying the etiology of the marked increased disease

concordance.3 The ATG16L1 authophagy gene discovery was rapidly

followed by a second autophagy gene, IRGM, thereby opening a new

pathophysiological area for CD related research, autophagy. Unlike

IL23R, ATG16L1 and IRGM were associated only with CD and thus, like

NOD2, demonstrating that innate immune handling of bacteria

differentiates CD pathophysiology from UC. PTGER4, identified as a CD

risk gene in the Belgian-French GWAS, is a lymphocyte receptor of

prostaglandin E2, a prostaglandin responsible for a variety of

actions including gut mucosal cytoprotection. Like many of the loci

found in GWAS to be CD associated, the PTGER4 related SNPs actually

mapped in a so-called " gene-desert " region, 270 kb distally. However,

the SNPs associated were genetically linked to PTGE4 expression in

cell lines, suggesting CD risk by " long-distance " cis-regulation.

Investigators from the WTCCC-UK, NIDDK and Belgian-French groups

noted that the observed number of independent SNPs associated in CD

GWAS below genome-wide levels of significance was far greater than

that expected by chance, suggesting that many would tag true, albeit

low-risk, CD genes. Therefore, we performed a meta-analysis of three

CD GWASs with a massive screen of 3200 cases and 4800 controls and

replication using 3600 and 3100 case control equivalents.8 We

identified an additional 29 CD loci, 21 with combined association

evidence at p-values < 2 × 10-8. The findings established further the

contribution of genetics in elucidating mechanisms underlying CD

pathophysiology. Numerous IL23/TH17 pathway genes were associated in

addition to IL23R. ITLN1, a lactoferrin and glactofurranose binding

protein expressed in brush border membranes in the small intestine

and like ATG16L1 and NOD2, is believed to have a role in innate

immunity. PTPN2 and PTPN22, are associated with T1D (and PTPN22 with

rheumatoid arthritis), but the PTPN22 T1D associated allele is

protective in CD. MST1 and NKX2-3, two genes also previously not on

the radar of CD pathophysiology, regulate cell types critical to CD -

macrophage and Peyer patch development, respectively. SNPs were

associated with CD at the HLA locus at BTLN2, in tight linkage-

disequilibrium with SNPs of the HLA-DRB1 gene.

With the numerous CD genes now established, a large part of the

genetic foundation of CD pathophysiology has become transparent. The

next step to further clarify IBD's genetic foundation was to

determine which loci are also relevant to UC or are unique to CD. In

this month's Gastroenterology, the WTCCC-UK group completed this next

step and tested each of the new meta-analysis SNPs associated.9

Consistent with the new paradigm of IBD genetics research, the study

was highly powered with 2500 UC cases and 4100 population controls

being genotyped. The study rounds out the WTCCC-UK findings from a

recent report that examined 1800 UC cases and 1300 controls for UC

relevance of the then confirmed 13 GWAS identified loci (excluding

the NOD2 and IBD5 loci), and provided evidence that MST1 and NKX2-3

are genes for both UC and CD, with effects for UC on a par with that

observed for CD.4 In the present study, et al. provide

further evidence that the Th17/IL23 pathway genes are associated with

both IBD phenotypes: the IL23R ligand IL12B, JAK2, STAT3 and CDKAL1

were all significantly associated with UC at Odds Ratios very similar

to that observed for CD. IL18 receptor accessory protein (IL18RAP), a

gene with " just perceptible " risk for CD in the meta-analysis (OR

1.07) showed somewhat higher association with UC (OR 1.17). Perhaps,

not surprisingly given the longstanding evidence for HLA-DRB1

associations with both CD and UC, the rs3763313 SNP that tagged SNPs

from BTLN2 to HLA-DRQA was also significantly associated with UC. The

findings from the et al. study complemented a recent German

study that also examined several novel CD genes in UC and likewise

found evidence that STAT3, NKX2-3 and the cyclin cell cycle

transcription regulator CCNY were genes for all IBD.6 Two recent GWAS

also showed the potential clinical relevance of IBD molecular

genetics. A high density UC study identified IL10 as a UC and CD

gene, suggesting that recombinant IL10 treatment be reconsidered in

UC.10 The first pediatric IBD GWA study identified two genes equally

strong for UC and CD, TNFRSF6B and PSMG1. These were not observed in

the high-powered meta-analysis suggesting specific genes for early

onset IBD and perhaps explaining pediatric IBD often presenting as UC

but developing into CD with maturation.

et al. also genotyped, in 1560 CD cases and 3000 controls, a

SNP for the extracellular matrix protein 1 (ECM1), discovered in the

WTCCC UC non-synonymous screen. ECM1 had been also found to have

modest association with ankylosing spondylitis, a disease found in 3%

of patients with UC and 8% of patients with CD. Interestingly, ECM1

was not associated with CD with 80% power to identify a 1.1 fold

effect. Why ECM1 would be specific to UC and not CD may help uncover

UC pathophysiology, as ECM1 is known to be expressed in small and

large bowel and to activate NF-kappaB. It will be interesting to see

if ECM1 can induce epithelial auto-antibodies specific to UC, as ECM1

auto-antibodies have been found in lichen sclerosis, a mucocutaneous

disease with high rates of other autoimmune disorders, although not

IBD. Indeed, an important feature of the et al. report, and

a model for IBD genetics research, was recognizing the importance of

negative " absence of association " findings: the investigators

reported their power as well as the sample size required to achieve

80% power to replicate genetic associations in UC (P-value 0.0017) at

the observed meta-analysis odds ratios of the SNP for CD. Of 19 CD

loci with negative UC results, there was adequate (74% or greater)

power to exclude only 5 loci including Inducible T-Cell Costimulator

Ligand (ICOSLG) gene, a gene expressed in colon and involved in IL10

and IL2 regulation. The T1D associated PTPN2 and PTPN22 genes, appear

to be CD specific by their lack of replication in either the Fisher

et al. or et al. studies, although the German UC study

observed borderline association for PTPN2 (P = 0.002).

How far do we have to go to complete the genetic portrait of IBD?

Although the genetic basis for CD has been outlined and relevance to

UC is now being elucidated, the genetic basis for IBD requires

further investigation. The unique features of UC are only now coming

to light with the first high density UC GWAS studies at or near

completion, and then the necessary meta-analyses studies required for

power to find the weak UC genes. Although we know the CD and now the

common UC loci, the SNPs associated are in most cases simply tagging

SNPs that mark the territory. Thus, deep sequencing studies and

functional analyses are needed to identify the specific disease

causing SNPs that will let us know the true colors to paint for these

loci.

To complete the IBD genetic portrait will require studies with

analyses for copy number variant (CNV) polymorphisms and epigenetic

effects, determination of the specific phenotype subgroup tendencies

when present, and massive sample sizes to fill in the shading from

gene-gene and gene-environment interactions. Lastly, perhaps we will

develop separate portraits for the familial IBD gallery, a gallery

that will include the genes for those several remaining IBD linkage

loci. These loci will likely contain high penetrance disease alleles,

too rare to be identified by GWAS common tagging SNPs, but possible

to find by careful linkage disequilibrium mapping approaches and deep

sequencing the IBD pedigrees that link to the loci. Figure 1.

Figure 1 (not attached). Genes significantly associated with CD

(triangles), UC (circles) or both (diamonds) as reported in recent

genome wide association, CD meta-analysis and follow up studies

plotted by Odds Ratios of risk allele for UC vs. controls as compared

to same/equivalent risk allele for CD vs. controls. Genes listed are

those with allele frequencies reported for CD and UC. Data

preferentially from largest studies. Gene symbols outside axes lines

have significant evidence for UC only or CD only.

References

1. Orholm M, Iselius L, Sorensen TI, et al. Investigation of

inheritance of chronic inflammatory bowel diseases by complex

segregation analysis. BMJ. 1993;306:20-24.

2. Brant SR, Shugart YY. Inflammatory bowel disease gene hunting by

linkage analysis: rationale, methodology, and present status of the

field. Inflamm Bowel Dis. 2004;10:300–311.

3. Cho JH, Weaver CT. The genetics of inflammatory bowel disease.

Gastroenterology. 2007;133:1327-1339.

4. Fisher SA, Tremelling M, CA, et al. Genetic determinants

of ulcerative colitis include the ECM1 locus and five loci implicated

in Crohn's disease. Nat Genet. 2008;40:710-712.

5. Raelson JV, Little RD, Ruether A, et al. Genome-wide association

study for Crohn's disease in the Quebec Founder Population identifies

multiple validated disease loci. Proc Natl Acad Sci U S A.

2007;104:14747-14752.

6. e A, Balschun T, Karlsen TH, et al. Sequence variants in

IL10, ARPC2 and multiple other loci contribute to ulcerative colitis

susceptibility. Nat Genet. 2008;40:1319-1323.

7. Kugathasan S, Baldassano RN, Bradfield JP, et al. Loci on 20q13

and 21q22 are associated with pediatric-onset inflammatory bowel

disease. Nat Genet. 2008;40:1211-1215.

8. Barrett JC, Hansoul S, Nicolae DL, et al. Genome-wide association

defines more than thirty distinct susceptibility loci for Crohn's

disease. Nat Genet 2008.

9. C, A., Massey D, C. O., Barrett J, C., et al.

Investigation of Crohn's disease risk loci in Ulcerative Colitis

further defines their molecular relationship. Gastroenterology: W.B.

Saunders, 2008. In Press.

10. e A, Balschun T, Karlsen TH, et al. Replication of signals

from recent studies of Crohn's disease identifies previously unknown

disease loci for ulcerative colitis. Nat Genet. 2008;40:713-715.