Expression profile of genes involved in pathogenesis of pediatric Crohn's disease.

BACKGROUND AND AIM: Expression profiling of genes specific to pediatric Crohn's Disease (CD) patients was performed to elucidate the molecular mechanisms underlying disease cause and pathogenesis at disease onset.
METHODS: We used suppressive subtractive hybridization (SSH) and differential screening analysis to profile the mRNA expression patterns of children with CD and age- and sex-matched controls without inflammatory bowel disease (IBD).
RESULTS: Sequence analysis of 1000 clones enriched by SSH identified 75 functionally annotated human genes, represented by 430 clones. The 75 genes have potential involvement in gene networks, such as antigen presentation, inflammation, infection mechanism, connective tissue development, cell cycle and cancer. Twenty-eight genes were previously described in association with CD, while 47 were new genes not previously reported in the context of IBD. Additionally, 29 of the 75 genes have been previously implicated in bacterial and viral infections. Quantitative real-time reverse transcription polymerase chain reaction performed on ileal-derived RNA from 13 CD and nine non-IBD patients confirmed the upregulation of extracellular matrix gene MMP2 (P = 0.001), and cell proliferation gene REG1A (P = 0.063) in our pediatric CD cohort.
CONCLUSION: The retrieval of 28 genes previously reported in association with adult CD emphasizes the importance of these genes in the pediatric setting. The observed upregulation of REG1A and MMP2, and their known impact on cell proliferation and extracellular matrix remodeling, agrees with the clinical behavior of the disease. Moreover, the expressions of bacterial- and virus-related genes in our CD-patient tissues support the concept that microbial agents are important in the etiopathogenesis of CD.

Introduction

Crohn's disease (CD) is a chronic inflammatory disorder of the bowel. The cause of CD is unclear and a complex interplay between genetic, environmental and immune components has been implicated. The prevailing hypothesis for the pathogenesis of CD is that an aberrant immune response, generated against microbial agents in genetically susceptible hosts, results in chronic intestinal inflammation. Thus far, 71 genes have been implicated in CD based on genome‐wide association studies, and include genes involved in autophagy, maintenance of mucosal barrier integrity and immune regulation. , The NOD2/CARD15 on chromosome 16 was the first locus implicated, mutations of which are thought to affect bacterial recognition. Subsequently, four genes, IL10RA, IL10RB, PSMG1 and TNFRSF6B, have been linked to pediatric CD. , The polygenic nature of CD suggests that direct targeting of individual disease susceptibility genes is unlikely to be therapeutically effective. Key molecules in pathophysiology, downstream of regulatory events induced by different causative factors are more likely targets for therapeutic interventions.

Insights into key gene‐environmental interactions relevant to disease pathogenesis could help identify causative stimuli (e.g. infectious agents) based on molecular signatures of the host response. To date, microarray studies carried out on intestinal tissue of CD patients have identified several molecular biomarkers relating to inflammation, abnormal immunoregulation and cell biology, metabolism, signaling, transcription, electrolyte transport and extracellular matrix structure. , , , , , ,

The suppressive subtractive hybridization (SSH) technique provides a complementary, non‐biased approach to the identification of new genes or pathogens associated with CD. In SSH, suppression PCR normalizes the representation of rare and abundant cDNA within the target population, and the subtraction step removes common nucleic acid sequences between the target specimen and its matched control. This results in an enriched pool of sequences specific to the target population. The advantage of this approach is that no assumed knowledge of gene identity is required, as it does not rely on a defined set of gene library or conserved sequence signatures as probes for gene identification. Hence SSH complements microarray studies by identifying potentially important genes that may not be represented on the array platforms utilized by inflammatory bowel disease (IBD) microarray studies. SSH has been successfully used in the discovery of novel viruses, and the transcriptome profiling of human hepatoma and bone regeneration. , ,

In the present study, we used SSH to analyze the differential expression profile in ileal biopsies from children with CD compared with age‐ and sex‐matched non‐IBD control children. The purpose of this study was to examine the initial events occurring during CD pathogenesis.

Tissue selection.  Ileal biopsy specimens (3–6 mm3) were obtained from patients (aged 4–16) with symptoms suggestive of IBD and undergoing initial diagnostic endoscopy at the Royal Children's Hospital, Melbourne, Australia. All tissue specimens were stored in RNAlater (Ambion, Melbourne, Australia) at −70°C until nucleic acid extraction. The diagnosis of CD was established using standard clinical endoscopic and histopathological criteria according to the Montreal classification. Patients with esophagitis, mild non‐specific gastritis or no known pathological diagnosis were used as non‐IBD controls. None of the patients had received antibiotics or immunosuppressive drugs prior to endoscopy. Demographic and clinical details of patients assayed by suppressive subtractive hybridization and real‐time reverse transcription polymerase chain reaction (RT–PCR) are presented in 1, 2, respectively.

Table 1

Demographics, clinical details and genotype of pediatric patients assayed by suppressive subtractive hybridization

Patient	Sex	Age at	Montreal	Diagnosis	Endoscopic	NOD2 §	NOD2 §	NOD2 §	IL23 §	ATG16L1 §	TLR4 §
		diagnosis	Class †		Presentation ‡	(Leu1007FsinC)	(Arg702Trp)	(Gly908Arg)	(Arg381Gln)	(Thr300Ala)	(Asp299Gly)
CD‐SSH1	F	10.1	A1/L2/B1	CD	Uninflamed	DEL	CC	GG	GG	CT	GA
CD‐SSH2	M	14.5	A1/L1+4/B1	CD	Inflamed, uninflamed	DEL	CC	GG	GG	CC	AA
CD‐SSH3	F	12.8	A1/L3+4/B1	CD	Inflamed	C.DEL	CT	GG	GG	TT	AA
CD‐SSH4	M	13.3	A1/L3/B1	CD	Inflamed	DEL	CT	GG	GG	CT	AA
N‐SSH1	F	17	–	Mild gastritis	Uninflamed	DEL	CT	GG	GG	TT	AA
N‐SSH2	F	10.9	–	No pathological diagnosis	Uninflamed	DEL	CT	GG	GG	CT	AA
N‐SSH3	M	15.8	–	No pathological diagnosis	Uninflamed	DEL	CC	GG	GG	TT	AA
N‐SSH4	M	10.7	–	Esophagitis	Uninflamed	DEL	CC	GG	GG	CT	AA

A1 age ≤ 16 years old; L1 ileal location, L2 colonic, L3 ileocolonic, L4 upper gastrointestinal, B1 non‐stricturing, non‐penetrating; P perianal disease modifier.

Endoscopic presentation of ileal region where biopsy is taken. Where two biopsies taken from separate ileal locations of a patient differ in presentation, both are described here.

Genotyping of patients based on single‐nucleotide polymorphism were performed for an earlier study. Major alleles are DEL, C, G, G, C and A; for NOD2 Leu1007FsinC, Arg702Trp, Gly908Arg; IL23, ATG16L1 and TLR4 respectively.

CD, Crohn's disease; SSH, suppressive subtractive hybridization.

Table 2

Demographics and clinical details for pediatric patients assayed by real‐time reverse transcription polymerase chain reaction

Patient	Sex	Age at diagnosis (years)	Ileal biopsy endoscopic presentation †	Histology	Montreal class ‡
CD
CD1	M	14.6	Unaffected	Extensive ulceration	L3 B1
CD2	F	10.4	Affected	Complete loss of villi, extensive neutrophil infiltration	L3 B1 P
CD3	M	14.7	Affected	Occasional neutrophil infiltration	L3+4 B1 P
CD4	M	11	Unaffected	Occasional neutrophil infiltration	NA
CD5	M	11.7	Unaffected; Affected	Moderate neutrophil infiltration	L3 B1 P
CD6	F	8.3	Unaffected; Affected	Ulceration and granulation	L3+4 B1 P
CD7	M	12	Unaffected	Normal villous architecture	NA
CD8	M	9.6	Affected	Neutrophilic infiltration, crypt abscesses	L3+4 B2 P
CD9	M	13.8	Unaffected	Focal clusters of neutrophils, occasional granuloma	L3+4 B1 P
CD10	M	11.6	Unaffected	Several granulomas	L3 B1 P
CD11	F	13.4	Unaffected; Affected	Extensive ulceration	L3 B3 P
CD12	F	12.9	Affected	Moderate eosinophilic infiltration	L3+4 B1
CD13	M	12.8	Affected	Extensive leukocyte infiltration, extensive ulceration	L3+4 B1 P
Non‐IBD
N1	F	10.8	Unaffected	Normal villous architecture	–
N2	M	12	Unaffected	Normal villous architecture	–
N3	M	7.6	Unaffected	Normal villous architecture	–
N4	M	13.8	Unaffected	Normal villous architecture	–
N5	F	11.1	Unaffected	Normal villous architecture	–
N6	M	13.5	Unaffected	Normal villous architecture	–
N7	M	4.9	Unaffected	Normal villous architecture	–
N8	M	13.7	Unaffected	Normal villous architecture	–
N9	F	13.5	Unaffected	Normal villous architecture	–

Endoscopic presentation of ileal region where biopsy is taken. Where two biopsies taken from separate ileal locations of a patient differ in presentation, both are described here.

A1 age ≤ 16 years old; L1 ileal location, L2 colonic, L3 ileocolonic, L4 upper gastrointestinal, B1 non‐stricturing, non‐penetrating B2 stricturing, B3 penetrating; P perianal disease modifier.

CD, Crohn's disease; IBD, inflammatory bowel disease.

Sample preparation and RNA extraction.  Each biopsy was mechanically homogenized, the supernatant harvested, and RNA extracted using the AllPrep DNA/RNA Mini Kit (Qiagen, Melbourne, Australia) according to the manufacturer's protocol. All extractions were conducted in a biological safety cabinet class II.

Suppressive subtractive hybridization.  The CD‐specific subtractive library was constructed using the PCR‐Select cDNA Subtraction Kit according to the user manual provided (Clontech, Palo Alto, CA, USA). An overview of the SSH technique is described in Figure S1. Ileal RNA were obtained separately from four CD and four non‐IBD patients, then pooled into CD and non‐IBD groups for the SSH assay. The patient groups were matched based on sex, mean age and common genotypes associated with CD, to minimize heterogeneity.

Differential screening.  The library of differentially expressed cDNA specific to the CD population was constructed using the TOPO TA cloning kit (Invitrogen, Melbourne, Australia). Five thousand randomly selected clones from the CD‐specific subtractive library were spotted onto Hybond nylon membrane (Amersham Biosciences, Sydney, Australia) in 384 × 3 by 2 arrays by the Australian Genome Research Facility (AGRF), Melbourne. CD‐specific sequences were detected by reverse hybridization with digoxigenin (DIG)‐labeled probes (Roche, Sydney, Australia) synthesized directly from cDNA of the CD and non‐IBD subtractive library, according to manufacturer's protocol (DIG applications manual for filter hybridization, Roche). Clones with greater than three times hybridization affinity to the CD‐library‐specific probes as compared to non‐IBD‐library‐specific probes were selected for sequencing.

Sequence identification and data analysis.  Sanger sequencing of differentially expressed clones was performed using an ABI 3730 DNA analyzer (Applied Biosystems, Melbourne, Australia) at the AGRF, Melbourne. The ChromasPro software (Technelysium, Brisbane, Australia) was used to remove adaptor and vector sequences, and sequences were blasted against GenBank (http://www.ncbi.nlm.nih.gov/BLAST). Annotated sequences were submitted to SOURCE (http://smd.stanford.edu/cgi-bin/source/sourceBatchSearch), where all gene symbols and chromosome locations were obtained. Functional assignment was determined using the University of California, Santa Cruz (UCSC) genome browser and National Center for Biotechnology Information (NCBI) Entrez Gene database. Ingenuity Systems' IPA software (Ingenuity Systems Inc., http://www.ingenuity.com) was used to group the differentially expressed genes into biologically relevant networks.

Quantitative real‐time RT–PCR.  Expression of selected human genes (REG1A, MMP2 and ANPEP) in ileal biopsy was analyzed using quantitative real‐time RT–PCR. Commercially available clones from OriGene (RPL32: SC119501, MMP2: SC321560, ANPEP: SC119422, REG1A: SC122637) were used for real‐time RT–PCR method establishment. First strand cDNA was synthesized using the Superscript III RT kit (Invitrogen) according to manufacturer's instructions. Oligonucleotide primers spanning two different exons of each target gene were selected based on published sequences or designed using Primer3 Output software to avoid amplifying genomic DNA. The primers used are detailed in Table S1.

Quantification of cDNA by real‐time PCR was performed using the SYBR GreenER qPCR Super mix for ABI PRISM (Invitrogen), in accordance with manufacturer's instructions. Analysis of real‐time RT–PCR reactions and quantification of RNA was determined using the 7300 System Sequence Detection Software Version 1.4 (Applied Biosystems). Each sample was analyzed in triplicate. Gene expression levels for individual patient samples were normalized relative to the expression of ribosomal protein L32 (RPL32) housekeeping gene. Calculations were based on the Pfaffl method, a mathematical method based on the real‐time PCR efficiencies. The OriGene clone cDNA (125 fg) of each gene was used as the calibrator in every assay to allow for direct comparison of gene expression for all samples analyzed across multiple assays.

Statistical analysis.  The Mann–Whitney U‐test was used to compare the difference in median values between gene expression in CD and non‐IBD patient samples. A P‐value of less than 0.05 was considered statistically significant. All statistical tests were performed using SigmaStat, version 3.5 (SyStat Software Inc., San Jose, CA, USA).

Ethical considerations.  This study received ethics approval from the Human Ethics Committee of the Royal Children's Hospital (EHRC no. 23003). Written and informed consent was obtained from each individual, parent or guardian prior to enrolment in the study.

Functional classification of differentially expressed genes specific to Crohn's disease ileum.  Sequence analysis of 1000 differentially expressed clones from the CD subtraction library identified 863 clones with high homology to GenBank sequences. These included 430 clones, which had matches to human mRNA sequences representing 75 annotated genes. The remaining clones had sequence similarity to mitochondrial and ribosomal genes, hypothetical proteins, expressed sequence tag (EST), human chromosomes, bacterial and animal genes.

The 75 annotated genes were assigned to eight functional clusters based on information obtained from the UCSC genome browser and NCBI Entrez Gene database. The map location, gene function and frequency of SSH clone representation for each gene is listed in Table S2. We noted an enrichment of immune function genes and inflammatory mediators (Cluster I and II); extracellular matrix, remodeling, and ion transport coding genes (Cluster III); metabolic enzymes and signal transducers (Cluster IV); genes involved in cell‐cycle regulation (Cluster V); cancer‐related genes (Cluster VI); transcription factors and post‐transcription modifiers (Cluster VI) and genes with unknown function (Clusters VIII).

Real‐time RT–PCR confirmation of SSH results.  To assess the quality of the SSH data, genes representing different clone abundance levels were selected for real‐time RT–PCR quantification on ileal biopsies. Three genes were selected based on their representation of the SSH detection frequency range (high: > 50; moderate: 10–50; low: < 10), and also on potential functional interest with respect to CD pathogenesis. REG1A (55 clones) was selected based on its cell proliferative function and earlier reports of upregulation in colonic tissue of adult CD patients. , MMP2 (12 clones) is involved in wound healing and has been proposed to have a protective role in colitis by regulating barrier function and vascularisation. ANPEP (2 clones) has previously been reported to be a receptor for coronavirus.

Real‐time RT–PCR analysis of the three genes was conducted on ileum‐derived RNA from 13 CD and nine non‐IBD patients, in triplicate. For CD patients CD5, CD6 and CD11, biopsies taken from both endoscopically affected and unaffected ileal locations were used in the analysis. Individual gene expression levels for each sample were represented as fold change ratios relative to the expression of positive controls (OriGene clones for MMP2, ANPEP and REG1A). The individual expression levels (fold change value) of each gene for the biopsy samples of the 13 CD and nine non‐IBD patients are depicted in Figure 1.

Figure 1

The relative expression levels of REG1A, MMP2 and ANPEP in ileal biopsies from 13 Crohn's disease (CD) and nine non‐inflammatory bowel disease (IBD) patients. The relative expression ratio of each gene was calculated based on real‐time reverse transcription polymerase chain reaction (RT–PCR) efficiency and the crossing point deviation of the target patient sample versus the internal RPL32 control, according to Pfaffl.

Using the Mann–Whitney statistical test for non‐parametric and unpaired populations, the transcript expression levels of MMP2 were found to be significantly higher in CD ileal biopsies as compared to non‐IBD ileal biopsies (P = 0.001). The CD population had a trend towards a higher level of REG1A transcript expression, although the difference was not statistically significant (P = 0.063). There was no significant difference in ANPEP transcript expression between CD and non‐IBD patient samples (P = 0.305). The real‐time RT–PCR results validated that genes represented by > 10 clones enriched by subtractive hybridization were expressed in higher abundance in CD as compared with non‐IBD ileal biopsies.

Analysis of REG1A, MMP2 and ANPEP gene expression across the CD patient samples revealed interesting patterns of expression. Using a fold change ratio of 1 as reference, four CD ileum samples (CD1, CD2, CD8, CD11un) with high levels of MMP2 expression, had low or negligible REG1A and ANPEP expression (Fig. 1). This inverse pattern of expression was also observed in the CD ileum samples where MMP2 gene expression was high.

Comparison of the CD expression profile represented in the SSH library with published microarray data.  To contextualize our SSH findings, we compared our results with the data tables from seven microarray studies published previously, that had reported differential expression of genes between inflamed biopsies of CD and non‐inflamed biopsies of non‐IBD controls. , , , , , , Of the 75 annotated genes, 28 genes have been previously analyzed by microarray (Table 3). The genes were either reported to be upregulated (n = 16), downregulated (n = 10) or variable (n = 2) depending on biopsy site assayed. There were 47 genes identified in this study that have not been previously described in the context of IBD investigations.

Table 3

Genes identified by SSH in this study that have previously been associations with CD

Genes symbol	Study reference	Tissue site	Gene expression in CD	SSH clone abundance
REG1A	10, 11, 12	Colon	Upregulated	55
	13	TI	Downregulated
CEACAM5	13	TI	Upregulated	17
CD74	8	Colon	Upregulated	12
MMP2	9	Sigmoid colon	Upregulated	12
IGHG1	9	Sigmoid colon	Upregulated	11
PSME2	14	Colon	Upregulated	9
IGL@	11	Colon	Upregulated	9
REG1B	12	Colon	Upregulated	7
LGALS4	11	Colon	Upregulated	5
OLFM4	13	All intestinal sites	Upregulated	5
SERPINA1	8	Colon	Upregulated	4
GBP1	13	All intestinal sites	Upregulated	4
APOB	13	All intestinal sites	Upregulated	4
		TI	Downregulated
CEACAM6	13	TI	Upregulated	3
CANX	11	Colon	Upregulated	2
ALDOB	10	Intestinal mucosa	Upregulated	2
HLA‐DRA	8	Colon	Upregulated	1
DMBT1	10, 13	All intestinal sites	Upregulated	1
SRGN	12	Colon	Upregulated	1
EIF4EBP2	13	TI	Downregulated	5
SLC5A1	13	TI	Downregulated	4
TGOLN2	13	TI	Downregulated	3
ANPEP	8, 12	Colon	Downregulated	2
UGT2B17	13	TI	Downregulated	2
HIST1H1B	13	All intestinal sites	Downregulated	1
TTRAP	13	TI	Downregulated	1
LAPTM5	13	TI	Downregulated	1
GDA	13	TI	Downregulated	1

CD, Crohn's disease; SSH, suppressive subtractive hybridization; TI, terminal ileum.

Gene networks.  To identify biological and functional networks based on potential gene interactions among the 75 SSH enriched genes, we utilized the “Core” program of the Ingenuity Pathway Analysis Software. The majority of the 75 genes were classified into six networks comprising the following functions: (i) antigen presentation, inflammatory response, cancer; (ii) cancer, cell cycle, cellular compromise; (iii) connective tissue development and function, tissue morphology, developmental disorder; (iv) infection mechanism, genetic disorder, nutritional disease; (v) cell signaling, cellular assembly and organization, cellular function and maintenance; and (vi) amino acid metabolism, molecular transport, small molecule biochemistry (Table 4).

Table 4

Gene networks represented by suppressive‐subtractive‐hybridization‐enriched genes

Gene network	Top functions	Genes involved	Number of genes
1	Antigen presentation, inflammatory response, cancer	ANPEP, APOB, CANX, CD74, CEACAM5, CEACAM6, CTSS, DDC, DMBT1, GBP1, HLA‐DRA, IGHG1, ITGB1, LGALS4, MMP2, OLFM4, PARP9, PSME2, PTMA, SERPINA1, SLC5A1, TGOLN2, UGT2B17	23
2	Cancer, cell cycle, cellular compromise	C12ORF35, DOCK9, EVL, GBP3, HIST1H1B, LAPTM5, LPHN1, MACF1, MAN1A1, MARK3, OTUD4, PABPC1, PRKCSH, SRGN, TUBA1B	15
3	Connective tissue development and function, tissue morphology, developmental disorder	ALDOB, APH1A, CAP1, EIF4EBP2, GUF1, HNF4G, RBM17, REG1A, REG1B, SF3B1, STOM, TTRAP, TMEM66, XRN1	14
4	Infection mechanism, genetic disorder, nutritional disease	APPBP2, CLCA1, CRIM1, HSD11B2, IGL@, NACA, PLS1, SCP2, SLC26A3	9
5	Cell signaling, cellular assembly and organization, cellular function and maintenance	DNAJC5, EEF1A1, GDA, NRF1, PLCB3, PRKAA1, PSAP, VAV2	8
6	Amino acid metabolism, molecular transport, small molecule biochemistry	SLC17A7	1

Network 1 contained the highest number of SSH genes. Interestingly, 18/23 genes in this network have been previously reported in microarray studies. The five newly identified genes within this network are cathepsin (CTSS), DOPA decarboxylase (DDC), integrin beta 1 (ITGB1), poly ADP‐ribose polymerase (PARP9) and prothymosin alpha (PTMA). Figure 2 depicts a schematic representation of this gene network. CTSS and ITGB1 appear to be involved in multiple pathways, including several direct and indirect associations with the previously reported genes.

Figure 2

Gene network containing antigen presentation, inflammatory response and cancer as top functions. Genes enriched by suppressive subtractive hybridization (SSH) are highlighted in red, blue and green, while the other molecules serve as intermediates in the gene interactome. Genes previously reported as upregulated (red) in Crohn's disease (CD) population are closely associated with inflammatory proteins, such as major histocompatibility complex (MHC) class II complex, NF‐κB complex and 26S proteasome. Genes previously reported as downregulated (blue) in CD population are either indirectly activated by interferon‐α and interleukin (IL)‐1, or indirectly associated with ITGB1. Of the new genes identified (highlighted green), CTSS and ITGB1 appear to be involved in multiple pathways within the inflammatory network. The different interactions include direct (solid lines) or indirect (dashed lines) interactions; binding (straight line), activation (arrow), inhibition (truncated line) or either activation or inhibition (truncated line with arrow).

Genes associated with microbial pathogenesis.  To elucidate evidence for microbial pathogenesis, the 75 functionally annotated genes were individually searched against the NCBI Entrez Gene database for reported functional associations with viral or bacterial infections. A total of 29 genes associated with microbial pathogenesis were identified (Table 5).

Table 5

Differentially expressed genes associated with microbial pathogenesis

Gene symbol	Gene name	Function
ANPEP	Alanyl (membrane) aminopeptidase	Receptor for human coronavirus 229E 27
CEACAM6	Carcinoembryonic antigen‐related cell adhesion molecule 6 (non‐specific cross reacting antigen)	Receptor for adherent invasive Escherichia coli, abnormally expressed by ileal epithelial cells in Crohn's disease patients 28
ITGB1	Integrin beta 1	Receptor for Kaposi sarcoma herpesvirus KSHV.HHV8, and Helicobacter pylori, promotes infection by human metapneumovirus 29
CD74	CD74 molecule, major histocompatibility complex, class II invariant chain	CD74 receptor facilitates the adhesion of H. pylori to gastric epithelial cells 30
TTRAP	TRAF and TNF receptor associated protein	Facilitates lentiviral integration 31
DMBT1	Deleted in malignant brain tumors 1	Facilitator of HIV‐1 transcytosis, broad bacterial‐binding specificity (LRR) inhibits LPS‐induced TLR4‐mediated NF‐kappaB activation 32
TUBA1B	Tubulin, alpha 1b	HIV‐1 binding to CD4 permissible cells induce acetylation of tubulin, facilitating HIV cell fusion, involved in EPEC and EHEC infection 33
CTSS	Cathepsin S	Mammalian reoviruses utilize CTSS for disassembly of the virus outer capsid and activation of the membrane penetration machinery 34
NRF1	Nuclear respiratory factor 1	Human T lymphotropic virus type 1 transactivates the promoter for T cell tropic HIV‐1 through association with NRF 35
MAN1A1	Mannosidase, alpha, class 1A, member 1	Processing of gp160 of HIV 36
EEF1A1	Eukaryotic translation elongation factor 1 alpha 1	Interacts with hepatitis deltavirus RNA and HIV gag protein, possibly permitting packaging of viral RNA into virion 37
TGOLN2	Trans‐golgi network protein 2	Involved in the final envelopment of herpesviruses 38
CANX	Calnexin	Interacts with measles virus protein F and hemagglutinin 39
MMP2	Matrix metallopeptidase 2 (gelatinase A, 72 kDa gelatinase, 72 kDa type IV collagenase)	HIV‐1 induces MMP2 expression in astrocytes 40
SERPINA1	Serpin peptidase inhibitor, clade A (alpha‐1 antiproteinase, antitrypsin), member 1	Specifically induced in Helicobacter pylori infection, inhibitor of HIV replication 41
OTUD4	OTU domain containing 4	Expressed only in HIV‐1 infected cell 42
MACF1	Microtubule‐actin cross‐linking factor 1	Parvovirus infection induces the upregulation of MACF1 43
PLS1	Plastin 1 (I isoform)	PLS1 is upregulated in HIV‐1‐infected human monocyte‐derived macrophages 44
MUC17	Mucin 17	MUC17 is upregulated upon infection by atypical enteropathogenic Escherichia coli 45
CLCA1	Chloride channel accessory 1	CLCA1 plays a role in bacterial‐induced mucus hypersecretion 46
EIF4EBP2	Eukaryotic translation initiation factor 4E binding protein 2	Adenovirus infection inactivates translational inhibitors 4E‐BP1 and 4E‐BP2 47
SLC5A1	Solute carrier family 5 (sodium/glucose cotransporter), member 1	HIV Tat induces SGLT1 mis‐sorting and impairs intestinal glucose absorption 48
PABPC1	Poly(A) binding protein, cytoplasmic 1	Rotavirus nsp3 expression directs PABC1 from cytoplasm to nucleus, in poliovirus, cleavage of PABP contributes to viral translation shutoff that is required for the switch from translation to RNA replication 49
SF3B1	Splicing factor 3b, subunit 1, 155 kDa	Vpr, the viral protein R of HIV‐1, induces G(2) cell cycle arrest and apoptosis in mammalian cells via binding to a subunit of multimeric SF3B 50
PSME2	Proteasome (prosome, macropain) activator subunit 2 (PA28 beta)	Upregulates presentation of viral MHC 1 51
PTMA	Prothymosin, alpha	Inhibitor of HIV‐1 expression 52
HLA‐DRA	Major histocompatibility complex, class II, DR alpha	Particular HLA class II region haplotypes affect the probability that an HBV infection will become persistent 53
LRRC25	Leucine rich repeat containing 25	Contains motifs involved in bacterial LPS recognition
XRN1	5′‐3′ exoribonuclease 1	XRN1 possess strong anti‐RNA virus activity by degrading uncapped RNA 54

EHEC, enterohaemorrhagic Escherichia coli; EPEC, enteropathogenic Esherichia coli; HBV, hepatitis B virus; LPS, lipopolysaccharide; LRR, leucine rich region; MHC, major histocompatibility complex; NRF, nuclear respiratory factors; OUT, operational taxonomic unit; PABP, poly A binding protein; TNF, tumor necrosis factor; TRAF, tumour necrosis factor receptor—associated factor 1.

Discussion

The pathogenesis of CD is thought to involve a complex interplay between the microbiome, the environment and multiple genetic factors. To gain further insights into the gene regulation processes involved, several gene array analyses have been performed using surgical resections or endoscopic biopsies of the colon obtained during treatment of adults with known IBD. , , , , , However, the chronicity of the disease process and variability of treatments used are likely to have influenced gene expression profiles in these patients. Our study used tissue obtained at initial diagnosis in treatment‐naive children with early onset disease. To date, there have been very few studies of events at the genetic level during early disease onset in children. A recent study examining the genome‐wide expression profile of pediatric IBD patients was conducted using colonic tissue.

Our study extends these initial gene expression profile studies by comparing ileal biopsies from a pediatric cohort of CD and non‐IBD patients. SSH analysis led to the identification of 75 functionally annotated genes, specific to the CD cohort. Comparison of our SSH data with existing microarray studies revealed that 47 of these genes are novel and 28 genes have been previously identified by microarray to be either upregulated or downregulated in the CD population.

Gene networks.  The antigen presentation, inflammatory response and cancer gene network (Network 1) comprise one‐third of the genes identified by SSH, with a high proportion of genes previously identified to be differentially expressed in CD. This is partially attributable to acute inflammation of the biopsies of CD patients as compared with the non‐inflamed biopsies of non‐IBD controls. Differences in gene expression profiles between inflamed and non‐inflamed CD terminal ileum have been recently described. Relative to non‐IBD controls, the gene expressions of IL‐8 and SAA1 were reportedly much higher in inflamed CD terminal ileum as compared to non‐inflamed CD terminal ileum.

New genes identified within this network include CTSS, DDC, ITGB1, PARP9 and PTMA. Based on the molecular interactions depicted in this network, CTSS and ITGB1 appear to be involved in multiple pathways associated with inflammatory complexes (major histocompatibility complex [MHC] class II complex and NF‐κB complex), and with other genes previously reported as upregulated in CD population. CTSS is mainly expressed in antigen‐presenting cells and is required for the degradation of MHC‐class‐II‐associated invariant chains, necessary for proper MHC class II antigen presentation. , Integrins, which include ITGB1, are membrane receptors involved in cell adhesion and several processes, including immune response. ITGB1 is expressed during hypoxic conditions, and can serve as an indicator of intestinal wound repair, which occurs only in a hypoxic environment.

REG1A and MMP2 expression.  The REG1A gene is involved in regulation of cell proliferation, and has been proposed to function as a mitogenic and/or an anti‐apoptotic factor in ulcerative colitis (UC)‐colitic cancer progression. Its high expression levels have been correlated with the severity of intestinal inflammation in patients with UC, and microarray studies have reported its upregulation in the colon of adult IBD patients. , , Similarly, we identified an upregulation of REG1A in the terminal ileum of pediatric CD patients. This was however contrary to a recent study comparing the expression of REG1A in the terminal ileum of adult CD and non‐IBD controls, which reported a downregulation in REG1A expression. The difference in REG1A expression could indicate a distinction between the pathogenesis of early onset CD and adult‐onset CD. Based on the knowledge that REG1A gene expression is associated with cancer development, the high level of REG1A expression in the terminal ileum of some CD pediatric patients could indicate an increased risk for colorectal cancer development. Individuals with early onset CD have been previously described to have an increased risk of developing colorectal cancer.

The increased levels of MMP2 observed in CD ileum are consistent with previous studies conducted on colonic tissue where MMP2 is highly expressed in the intestinal epithelia during IBD. , Other studies have suggested the involvement of MMP2 in the regulation of epithelial barrier function. Since epithelial barrier dysfunction plays a central role in the pathogenesis of intestinal inflammation, the increased expression of MMP2 may serve as a response to counteract tissue damage, hence protecting against colitis.

The fluctuation in REG1A and MMP2 gene expression between ileal biopsies of different patients and also between biopsies taken at different ileal locations of the same patient, suggest a spatial‐temporal nature of gene regulation during early CD pathogenesis. This finding is consistent with the clinical nature of CD, with its patchy distribution.

Microbial associations.  Twenty‐nine of the 75 genes identified in this study have functional roles in the processes of bacterial or viral infection. Evidence of host response in facilitating viral infection is demonstrated by the enrichment of gene products involved in viral attachment (ANPEP, ITGB1), viral entry, vesicular trafficking and transcytosis of viral proteins (TUBA1B, DMBT1, CTSS); lentiviral integration (TTRAP); viral translation (PABPC1, EIF4EBP2) and replication (NRF1); virion glycoprotein processing (MAN1A1); packaging (TGOLN2, EEF1A1) and possibly release (CANX). Evidence of response to bacterial infection is reflected by the enrichment of receptors for adherent invasive Escherichia coli and Helicobacter pylori (CEACAM6, CD74). ,

The enrichment of MMP2, SERPINA1, OTUD4, MACF1, PLS1, MUC17 and CLCA1 transcripts suggests the presence of infectious agent(s) early in disease pathway as these genes have previously been reported to be upregulated during bacterial or viral infections. , , , , , , The involvement of SLC5A1 and SF3B1 gene products in the impairment of intestinal glucose absorption and apoptosis due to HIV‐1‐induced glucose channel mis‐sorting and cell cycle arrest suggest the occurrence of viral activities in early CD pathogenesis. , The PSME2, PTMA, HLA‐DRA, LRRC25 and XRN1 genes or gene products have been previously reported to be associated with defense against viral and bacterial infections. , , , It is possible that these genes are differentially expressed in CD patients in response to infectious triggers.

Our study recognizes the limitation of the SSH technique whereby the CD subtraction library contained clones that are not differentially expressed, as shown by the ANPEP expression data. This limitation was also observed in previous studies. Preliminary SSH data presented in this study were verified either by real‐time PCR quantification or comparison to microarray data from studies performed on individuals with and without IBD. Several of the genes anecdotally identified in the context of CD by our study have roles in microbial pathogenesis, promoting inflammation, epithelial remodeling, vesicular transport or cell differentiation and proliferation. These processes are relevant to CD pathogenesis, hence future investigations into the association between these novel gene candidates and CD could contribute to the understanding of the disease.

Figure S1 Suppressive subtractive hybridization method. Restriction endonuclease‐digested tester DNA was split into two pools and ligated with Adaptor 1 or Adaptor 2R. Two successive rounds of hybridization with excess restriction endonuclease‐digested driver DNA followed. Thereafter, single‐stranded components of the adaptors were filled in. Exponential amplification of tester‐specific sequences is used to enrich for potential differentially expressed genes. Type a molecules are significantly enriched, differentially expressed sequences, while cDNA that are not differentially expressed form type c molecules with the driver. The concentration of high‐ and low‐abundance sequences is equalized, whereby highly abundant molecules re‐anneal to form type b and d molecules. During the second hybridization, remaining equalized and subtracted single‐stranded tester cDNA reassociate to form type e hybrids, with different ends corresponding to sequences of Adaptor 1 and Adaptor 2R (adapted from Clontech PCR‐Select cDNA subtraction kit user manual [BD Biosciences]).

Table S1 Primers used for real‐time reverse transcription polymerase chain reaction quantification of ANPEP, REG1A, MMP2 and RPL32

Table S2 Differentially expressed genes specific to Crohn’s Disease (CD) ileum. Genes within each functional category are listed in order of clone abundance

Supporting info item

Click here for additional data file.

Supporting info item

Click here for additional data file.