Serum mirna signature diagnoses and discriminates murine colitis subtypes and predicts ulcerative colitis in humans

Transcription

1 Serum mirna signature diagnoses and discriminates murine colitis subtypes and predicts ulcerative colitis in humans Emilie Viennois 1*, Yuan Zhao 1, 2, Moon Kwon Han 1, Bo Xiao 1, 3, Mingzhen Zhang 1, Meena Prasad 4, 5, Lixin Wang 1, 4, Didier Merlin 1, 4. 1 Institute for Biomedical Sciences, Center for Diagnostics and Therapeutics, Georgia State University, Atlanta, GA 333, USA. 2 Department of Gastroenterology, Zhongshan Hospital, Fudan University, Shanghai, China. 3 Institute for Clean Energy and Advanced Materials, Faculty for Materials and Energy, Southwest University, Chongqing, 4715, China. 4 Veterans Affairs Medical Center, Decatur, GA, USA. 5 Emory University, Department of Medicine, Atlanta, GA, USA *Author for correspondence Tel.: +1 (44) Fax: +1 (44) eviennois@gsu.edu

2 Fig. S1: Spontaneous development of colitis in IL1 -/- mice. A, WT and IL-1 -/- 135-day-old mice were subjected to H & E staining of colon samples. Arrowheads indicate lymphocyte infiltration. B, Lcn-2, which is used as a marker of intestinal inflammation, was monitored from day 4 to day 127 by ELISA (n=6-7). Fig. S2: Heat map and unsupervised hierarchical clustering on the 5 mirnas with highest standard deviations. Total RNAs from serum collected from 3 animals was used for mirna profiling. Clustering was performed on all 3 samples, using the top 5 mirnas with highest standard deviations. The normalized (dcp) values were analyzed. Fig. S3: Number of detected samples and quality control of the high throughput approach. A, Number of mirnas detected per sample (count) and their average amplification threshold (Cp) values in each sample. B, Raw Cp values obtained from the two controls, UniSp6 and UniSp3, which were used to test the assay performance for each sample. C, Hemolysis was assessed using the ratio of mir-451a to mir-23a-3p. Possible erythrocyte microrna contaminations were indicated by a Delta Cq (mir-23a-3p - mir-451a) 5 (orange area). Delta Cq 7 8 or more indicated a high risk of hemolysis (red area). Fig. S4: Assessment of inflammation in various mouse models. The degree of inflammation was assessed in arthritic mice (A), DSS-treated mice (B) and TLR5 -/- mice (C). A, The levels of Lcn-2 were measured in the sera obtained from CAIA mice at D12 (after induction of arthritis) and in the same mice at D2 (before induction of arthritis). Histological scores were displayed (A). B, The levels of Lcn-2 were measured in the feces of WT mice after 7 days (D7) of

3 exposure to 3% DSS in the drinking water, and compared to the levels obtained from the same mice at D. C, The levels of Lcn-2 were measured in the sera of TLR5 -/- mice and corresponding WT controls. Significance was determined using either an unpaired t-test or Mann Whitney test if the concentration values did not follow a normal distribution (B) (*, p<.5; and **, p<.1) and compared to the corresponding non-inflamed control. D, Hemolysis was assessed using the ratio of mir-451a to mir-23a-3p expression levels in sera arthritic and its controls (D12 and D-2 respectively), DSS-treated and water-control (WT-DSS-D7 and WT-D respectively) and TLR5- /- and its corresponding WT controls mice. Possible erythrocyte microrna contaminations were indicated by a Delta Cq (mir-23a-3p - mir-451a) 5 (orange area). Fig. S5: Lcn-2 levels in the sera of various mouse model of inflammation. Lcn-2 levels were measured in the sera of WT, D7 DSS, TLR5 -/-, D28 IL1 -/-, D98 IL1 -/- and arthritic mice. Data were presented as the means +/- S.E.M. (n=5-18). Significance was determined using Kruskal- Wallis followed by a Dunn s multiple comparison post-test (*, p<.5; and **, p<.1). Fig. S6: mirna signature in the context of cage effect and genetic background. PCoA of the euclidean distance matrix of the expression of the nine signature mirnas in sera of WT mice hosted in four different cages (A) or representing two different genetic backgrounds, Bab/C or C57/Bl5 (cage A or cage B) (B). Fig. S7: Effect of Anti-TNFα treatment on IL1-/- mice. A, Colon section were stained by H&E, scale bar=1µm. B, Hemolysis was assessed using the ratio of mir-451a to mir-23a-3p expression levels in sera of D28 IL1-/-, D98 IL1 -/- PBS-treated and day 98 IL1 -/- anti-tnfα-

4 treated mice. Possible erythrocyte microrna contaminations were indicated by a Delta Cq (mir-23a-3p - mir-451a) 5 (orange area). The serum expression levels of let-7d-3p (C) and mmu-mir-21a-5p (D) were compared between D28 IL1 -/- mice, D98 IL1 -/- PBS-treated and day 98 IL1 -/- anti-tnfα-treated mice and expressed as Fold change over Day 28 control. Significance was tested using Kruskal-Wallis followed by a Dunn s multiple comparison posttest. Fig. S8: mirna signature in IBD patients. The expression levels of the 9 deregulated mirnas pertaining to the signature identified in mice were quantified by qpcr in sera of patients with ulcerative colitis (UC) and controls. A, Overall misclassification error rate using from 1 (right) to 9 (left) mirnas of the signature panel. B-C, CRP (B) and Lcn-2 (C) levels were measured by ELISA in the sera of control and ulcerative colitis patients. Data were presented as the means +/- S.E.M. (n=11). Significance was evaluated using either an unpaired t-test (B) or Mann Whitney test when the concentration values did not follow a normal distribution (C).

5 A WT IL1 -/- B Lcn-2 [ng/g feces] IL1-/- WT Age (days) Fig. S1

6 Fig. S2

7 A Number of mirna detected per sample (count) Count Average (detected in all, Cp) average amplification threshold (Cp) B 24 UniSp3 UniSp Raw Cp C dcq (mir-23a - mir-451) Fig. S3

8 A Lcn-2 (ng/ml) ** Clinical score D-2 D12 D-2 D12 B Lcn-2 [ng/g feces] ** Lcn-2 (ng/ml) C * 1 D D7 WT TLR5-/- D 6 dcq (mir-23a - mir-451) Fig. S4

9 Lcn-2 (ng/ml) *** NS ** Fig. S5

10 A PC3 (16%) Cage 1 Cage 3 Cage 2 Cage 4 PC3 (4.7%) PC1 (75%) B PC2 (9.6%) C57/Bl6 (cage A) Balb/C C57/Bl6 (Cage B) PC3 (2%) PC1 (87%) Fig. S6

11 A Anti-TNFα PBS B dcq (mir-23a - mir-451) C Fold change over Day D28 control let-7d-3p D98-PBS D98- Anti-TNFα Fold change over Day 28 D D28 control mir-21a-5p D98-PBS D98- Anti-TNFα Fig. S7

12 A Number of mirna Misclassification error Value of threshold B C CRP (ng/ml) Lcn-2 (ng/ml) Control Ulcerative colitis Control Ulcerative colitis Fig. S8

13 Table S1: Differentially expressed mirnas mirna name p-value Day 3 Day 58 Day 86 Day 144 Day 128 mmu-mir-29b-3p E mmu-mir-122-5p E mmu-mir-335-5p E mmu-mir-148a-3p E mmu-mir-15-5p 1.541E mmu-mir-192-5p E mmu-mir-152-3p E mmu-mir-14-3p E mmu-mir-194-5p E mmu-mir-331-3p E mmu-mir-195a-5p E mmu-mir-14-3p E mmu-mir-146a-5p E mmu-mir-375-3p E mmu-mir-199a-3p E mmu-mir-142-5p E mmu-mir-29a-3p E mmu-mir-125b-5p E mmu-mir-154-5p E mmu-mir-22-5p E mmu-mir-378a-5p E mmu-mir-29c-3p E mmu-mir-99a-5p 8.944E mmu-mir-342-3p E mmu-mir-3d-5p mmu-mir-29a-5p mmu-mir-132-3p mmu-mir-143-3p mmu-mir-423-3p mmu-mir-99b-5p mmu-mir-365-3p mmu-mir-328-3p Table S1

14 Table S2: mirna sequences mirna name mirna sequence 5-3 mmu-mir-29b-3p mmu-mir-122-5p mmu-mir-335-5p mmu-mir-148a-3p mmu-mir-15-5p mmu-mir-192-5p mmu-mir-14-5p mmu-mir-194-5p mmu-mir-195a-5p mmu-mir-14-3p mmu-mir-146a-5p mmu-mir-375-3p mmu-mir-199a-3p mmu-let-7d-3p mmu-mir-21a-5p mmu-mir-23a-3p mmu-mir-451a TAGCACCATTTGAAATCAGTGTT TGGAGTGTGACAATGGTGTTTG TCAAGAGCAATAACGAAAAATGT TCAGTGCACTACAGAACTTTGT TCTCCCAACCCTTGTACCAGTG CTGACCTATGAATTGACAGCC CAGTGGTTTTACCCTATGGTAG TGTAACAGCAACTCCATGTGGA TAGCAGCACAGAAATATTGGC TACCACAGGGTAGAACCACGG TGAGAACTGAATTCCATGGGTT TTTGTTCGTTCGGCTCGCGTGA ACAGTAGTCTGCACATTGGTTA CTATACGACCTGCTGCCTTTCT TAGCTTATCAGACTGATGTTGA ATCACATTGCCAGGGATTTCC AAACCGTTACCATTACTGAGTT Table S2