n = 68 - Advances in Inflammatory Bowel Diseases

Using the Pediatric RISK and PROTECT
Inception Cohorts to Define Important Clinical
and Biologic Phenotypes
Ted Denson, MD
Cincinnati Children’s Hospital Medical Center and
the University of Cincinnati College of Medicine
Disclosures & Objectives
• Grant support: NIH & CCFA
• Describe aims & composition of RISK & PROTECT cohort
• Share early analyses from RISK cohort
• Summarize future goals for using cohorts to refine patient
clinical and biologic phenotypes
• Lee Denson
• Yael Haberman
•CCHMC Bioinformatics
• Bruce J Aronow
• Phillip Dexheimer
•Curtis Huttenhower
•Timothy L Tickle
• Enrolling sites
• Thomas D. Walters,
SickKids, Toronto,
• Subra Kugathasan,
Emory-Children’s Center,
Atlanta, GA
•Ramnik J Xavier
•Dirk Gevers
Current IBD Clinical Phenotyping
• Suspect the diagnosis
–History, exam, lab tests
• Determine disease location
–Upper endoscopy, colonoscopy, CT/MR enterography
• Classify as CD, UC, or IBD-U
–Montreal or Paris classification for disease location,
extent, severity, and complications
Importance of IBD Clinical Phenotypes
Louis et al Best Pract & Res Clin Gastro 2011
Crohn’s Disease Progresses on “Conventional
Therapy” in Children: 1988-2002
•34% at 5 yrs
•Vernier-Massouille et al.
Combined Genetic:Paneth Cell Phenotype
Stappenbeck et al Gastro 2013
High Risk CD Patient for Complications:
• NOD2 gene mutation
• Ileal location
• High titer anti-microbial serologies
CD1: High Risk Anti-TNF Responsive
Validate this model in a
prospective inception cohort
Siegel et al IBD 2011
Clinical use of Gene Expression Panels to
Improve Diagnostic or Prognostic Accuracy
Several gene expression diagnostics for oncology
Afirma Thyroid Cancer test
56,540 thyroid cancer cases per year
Indeterminate pathology in 30%
Expression of 142 genes in thyroid biopsy
49 site validation in 3789 patients: 92% accuracy
Prevent 25,000 thyroid resections per year
Charge: $4200, covered by Medicare and third party
Alexander et al NEJM 2012
CCFA Sponsored Pediatric Clinical Research
Network: PRO-KIIDS RISK Study
•1112 children with CD at
diagnosis between 20082012
• Follow-up to 2017
Paris Classification
Environmental Exposures
Microbial Community/Gene Expression
Define patients with
complication / surgery
• Develop and validate a composite risk score for
complicating CD based upon age, cytokine and microbial
antigen sero-reactivity, and early anti-TNF therapy
• Compare the effectiveness of early versus late
introduction of anti-TNF upon rates of complicating
• Profile host gene expression & the microbial community
structure and function in tissue and stool samples
• Identify and validate intestinal gene expression and
microbial community profiles associated with
complicated behavior which will improve the accuracy of
a model based upon clinical, genetic, and serologic
RISK Cohort
Original Specific Aims Funded in 2009
Progress to date in 2013
1a: Recruit 1100 children (age at diagnosis < 16 years) with
newly diagnosed inflammatory CD using standardized
diagnostic criteria.
Recruited 1112 CD. Additionally, we have recruited 160 UC, 129 IBDU, and 393
non-IBD as disease / healthy controls.
1b: Collect clinical and demographic information from well-characterized
newly diagnosed CD patients.
Collected comprehensive data at enrollment and every 6 months on every
subject with the diagnosis of CD, UC, and IBDU.
1c: Collect serum samples at diagnosis and annually for
both current serological immune markers including ANCA, ASCA, OMPC, I2 and
CBir1 and newly identified markers including GM-CSF auto-antibodies.
Serum was collected and all the stated antibodies were measured at diagnosis
for all CD, UC, IBDU, and non-IBD control. Serum has been collected annually on
CD, UC, and IBDU.
1d: Collect and genotype samples for known candidate genes and newly
discovered variants which may influence the stricturing/penetrating outcome,
and/or CD
Immunochip genotyping (over 200,000 SNPs) has been completed in all CD, UC,
and IBDU and non-IBD control subjects. Preliminary analysis shows that we have
found novel early onset susceptibility variants.
1e: Collect mucosal biopsies in a subset inception cohort stored for subsequent
aggregative gene expression analysis to define minimal genomic signatures
will distinguish CD patients who experience complicated disease behavior from
those who do not.
We collected over 5300 individual biopsies from 950 subjects. DNA and RNA
extraction has been completed for all biopsies, with sufficient quality and
quantity to support 16S microbial community profiling and RNASeq gene
expression analysis in > 95% of samples.
1f: Establish a Centralized Data Coordinating Center based at Toronto Sick Kids
utilizing a Clinipace web-based Data Management Platform and a Biospecimen
Procurement Center and Repository at Emory University.
A DCC and Bio-repository were established and have functioned very well to
support the RISK study. Details regarding quality control processes for both data
and sample management are provided in the Research Plan and the Appendix.
RISK Enrollment Birth & Environmental History
Comparative Effectiveness of Early IM vs
anti-TNF for Week 52 SFR in RISK
552 with complete data and
1 yr f/u
n = 68
Early IM only
No early
n = 236
• Propensity Score Matching
n = 68
IM only
n = 68
No early
n = 68
Hyams et al Gastro 2013
12 Month Outcomes For The
Three Early Therapy Approaches:
•CS-free, Surgery free
Yes (n=136)
No (n=68)
Early anti-TNFα
only (n=68)
58 (85%)
10 (15%)
Early IM only
41 (60%)
27 (40%)
No early
37 (54%)
31 (46%)
•No difference between early IM and no early IM
Hyams et al Gastro 2013
What Does This Mean?
• In clinically similar populations of children with Crohn’s
disease, early (<3 mon) therapy with anti-TNFα was
superior to early IM or no early immunotherapy despite
later addition of those agents: PCDAI remission, CRP,
• There was no particular clinical or laboratory
characteristic that helped predict response or nonresponse to an initial therapeutic decision
• It doesn’t mean that everyone should get anti-TNFα
therapy, rather that we need to better define further
characteristics of patients, such as genetics, serology,
microbiome. Confirmatory studies will be required.
Hyams et al Gastro 2013
Using Next Gen Sequencing to Classify the Intestinal
Microbiome and Genome at Diagnosis
• Processed ~ 5300 intestinal biopsies
from 950 CD, UC, IBDU patients and nonIBD controls
• DNA yield: 10,500 (8,468,12,670) ng
• RNA yield: 11,490 (9,351, 13,640) ng
• RNA quality sufficient for PCR or
RNASeq in >95%
• Microbiome: 1000 ng DNA
• RNASeq: 1000 ng RNA
• Completed ileal & rectal RNASeq for 317
CD, 79 UC, 9 IBDU & 52 non-IBD control
• Completed ileal, rectal, and fecal 16Smicrobial DNA profile for 477 CD & 221
non-IBD control
Ileal Gene Expression Panel Associated with
Complicated Behavior in High Risk Ser2+ Patients
Multivariate Analysis by Linear Models
Genes from the APOA1
module (APOA1, CXCL9)
Genes from DUOX2 module
Clinical phenotype (Ctl, UC,
endoscopic severity (ileal
deep ulcers)
Clinical severity (PCDAI)
• ileal microbial
•Controlling for: age, gender, body mass index (BMI), and NOD2,
FUT2, and ATG16L1 risk allele carriage.
Haberman et al 2013
Host:Microbe Profiles Present at Diagnosis May
Define IBD Biologic Phenotypes
CD2 greater clinical
severity at
Lower rate of SFR
with conventional
Two-fold higher
anti-TNF exposure
Haberman et al 2013
Treatment Response Phenotype:
“5-ASA Responsive UC”
Disease Severity
Enteral Nutrition
1U01 DK 095745-01
• Define rates of week 52 steroid-free remission with 5ASA as sole maintenance therapy in an inception cohort
of 430 pediatric UC patients receiving standardized
induction therapy with 5-ASA/corticosteroids
• Test clinical and biologic predictors of week 52 SFR
• Define the host rectal global pattern and of gene
expression and rectal and fecal microbial community
structure and function
• Identify mucosal host and microbial biologic pathways
associated with initial treatment response and week 52
PROTECT Recruitment
Future IBD Phenotyping
RISK & PROTECT cohorts: 2000 CD & UC patients
Prospective clinical course & treatment responses
Genetics & serology in all
Tissue gene expression and microbial community profile in fifty percent
Optimizing technique to isolate RNA & DNA from clinical path specimens in
order to perform gene expression and microbial community analysis on all
Key to maintain accurate long-term follow-up to link these host and
microbial profiles to clinical outcomes
Utilize to define combined clinical/biologic/treatment response phenotypes

similar documents