P1633 - Disease Activity in Ulcerative Colitis Is Strongly Associated With the Injury-Repair Response

Jeffery M. Venner, MD, MSc¹, Katelynn S. Madill-Thomsen, PhD², Shane Pon, BSc², Karen Kroeker, MD, MSc, FRCPC², Farhad Peerani, MD², Levinus Dieleman, MD, PhD², Karen Wong, MD², Daniel Baumgart, MD, PhD, MBA², Philip F. Halloran, MD, PhD², Brendan P. Halloran, MD^2
1University of Manitoba, Winnipeg, MB, Canada; ²University of Alberta, Edmonton, AB, Canada

Introduction: Ulcerative colitis (UC) is a chronic inflammatory condition affecting the colonic epithelium. We used an established microarray-based system to analyze a set of 128 UC biopsies (115 patients), assessing gene expression associated with the colon’s response to injury in UC.

Methods: 128 UC colon biopsies were collected at the University of Alberta Hospital (Edmonton, AB) and Cedars-Sinai Hospital (Los Angeles, CA) during standard of care colonoscopy and processed using Affymetrix microarrays. Principal component analysis (PCA) and archetypal analysis (AA) visualized relationships between biopsies and previously annotated injury-associated transcript sets. AA assigned each biopsy to one of three groups, and scores to each biopsy relating it to all three groups.

Results: Spearman correlations (Table 1A) were highest between the endoscopic Mayo score and the injury-repair-associated transcripts (IRRAT, 0.64, P=4.7x10^-16), immunoglobulin transcripts highly associated with chronic injury and fibrosis (IGT, 0.63, P=3.0x10^‑15), endothelial transcripts (ENDAT, 0.61, P=1.8x10^-14), and parenchymal dedifferentiation i.e. epithelial solute carrier loss (CT2, -0.60, P=6.5x10^-14).

PCA separated injury from no injury in PC1 (Figure 1A). T cell transcripts (QCATs), interferon-gamma inducible transcripts (GRITs) and targets of biologics (IL12, TNFA, ITGA4/B7) separated from injury transcripts in PC2.

We assigned three AA groups and visualized biopsies in PCA (Figure 1B, colored by AA membership). Group 1 (grey, N=44) biopsies had little parenchymal dedifferentiation and low expression of injury-associated transcripts. Groups 2 (red, N=44) and 3 (blue, N=40) had increased expression of injury-associated transcript sets and dedifferentiation compared to Group 1 (Table 1B). Although Group 3 was endoscopically similar to Group 1 (P >0.05), Group 3 showed elevated injury-associated transcript set expression (e.g. IRRAT) and increased parenchymal dedifferentiation (CT2).

Discussion: Assessment of UC biopsies using AA and previously annotated injury-associated gene sets reveals two groups of biopsies that are endoscopically similar though one group has increased molecular abnormalities, thus revealing heterogeneity unrelated to the Mayo score. A molecular system based around PCA and AA could enhance and refine UC disease assessment by allowing for quantitation and qualification of injury in biopsies obtained at endoscopy i.e. a level of resolution beyond conventional endoscopic scoring.


Figure: Figure 1. (A) PCA factor analysis of injury and inflammation transcripts in UC biopsies, (B) Three AA groups of 128 UC biopsies visualized in PCA space.

Table 1A. Transcript set correlation with Endoscopic Mayo Score (0-3) in ulcerative colitis biopsies (N=128)
Transcript Set		Spearman correlation coefficient	P-value
Colon-associated – solute carriers (CT2, parenchymal function)		-0.60	6.5x10-14
Injury-repair-associated (IRRAT)		0.64	4.7x10-16
Endothelial-associated (ENDAT)		0.61	1.8x10-14
Danger-associated molecular patterns (DAMPs)		0.57	2.5x10-12
Alternative Macrophage Activation-associated (AMAT)		0.50	2.7x10-9
Macrophage-associated (QCMAT)		0.58	8.9x10-13
Immunoglobulins (IGT)		0.63	3.0x10-15
T cell-associated (QCAT)		0.40	3.2x10-6
IFNG-inducible (GRIT)		0.48	8.4x10-9
Table 1B. Mean transcript set scores (expression) and mean clinical scores/counts in each of the three Archetype Groups.
	Variable	Group 1Grey – Normal (N=44 biopsies)	Group 2 Red – Injury 1 (N=44 biopsies)	Group 3Blue – Injury 2 (N=40 biopsies)
Clinical Features (#)	Mayo 0 biopsies (#)	15	0	9
	Mayo 1 biopsies (#)	13	3	14
	Mayo 2 biopsies (#)	14	18	13
	Mayo 3 biopsies (#)	2	23	4
Mean Clinical Features (median)	Mean endoscopic Mayo Score (median)	1.07 (1.0)	2.45 (3.0)	1.3 (1.0)
	Mean total Mayo Score (median)	3.2 (2.0)	7.2 (7.0)	4.4 (3.5)
Molecular Features (Transcript Set Score Expression Means)	Colon-associated – solute carriers (CT2, parenchymal function)	0.04	-0.53	-0.15
	Injury-repair-associated (IRRAT)	-0.14	0.80	0.34
	Endothelial-associated (ENDAT)	-0.03	0.49	0.17
	Danger-associated molecular patterns (DAMPs)	0.07	0.71	0.38
	Alternative Macrophage Activation-associated (AMAT)	-0.12	0.60	0.22
	Macrophage-associated (QCMAT)	-0.05	0.64	0.21
	Immunoglobulins (IGT)	0.05	1.21	0.31
	T cell-associated (QCAT)	0.02	0.67	0.37
	IFNG-inducible (GRIT)	0.15	0.65	0.34
*Note. Groups 2 vs 1 and 2 vs 3 differed in total Mayo score (P=7.5E-6 and 0.001, respectively), and in endoscopic Mayo Score (P=4.1E-12, and 9.2E-9). Groups 3 vs 1 were not significantly different in Total Mayo Score (P=0.2) or Endoscopic Mayo Score (P=0.25).

Table: Table 1.


Disclosures:
Jeffery Venner indicated no relevant financial relationships.
Katelynn Madill-Thomsen indicated no relevant financial relationships.
Shane Pon indicated no relevant financial relationships.
Karen Kroeker indicated no relevant financial relationships.
Farhad Peerani indicated no relevant financial relationships.
Levinus Dieleman indicated no relevant financial relationships.
Karen Wong indicated no relevant financial relationships.
Daniel Baumgart indicated no relevant financial relationships.
Philip Halloran indicated no relevant financial relationships.
Brendan Halloran indicated no relevant financial relationships.


Jeffery M. Venner, MD, MSc¹, Katelynn S. Madill-Thomsen, PhD², Shane Pon, BSc², Karen Kroeker, MD, MSc, FRCPC², Farhad Peerani, MD², Levinus Dieleman, MD, PhD², Karen Wong, MD², Daniel Baumgart, MD, PhD, MBA², Philip F. Halloran, MD, PhD², Brendan P. Halloran, MD². P1633 - Disease Activity in Ulcerative Colitis Is Strongly Associated With the Injury-Repair Response, ACG 2021 Annual Scientific Meeting Abstracts. Las Vegas, Nevada: American College of Gastroenterology.