Specific Norovirus Interaction with Lewis x and Lewis a on Human Intestinal Inflammatory Mucosa during Refractory Inflammatory Bowel Disease

Abstract

Inflammatory bowel disease (IBD), which includes Crohn's disease (CD) and ulcerative colitis (UC), is related to immunological and microbial factors, with the possible implication of enteric viruses. We characterized the interaction between human noroviruses (HuNoVs) and blood group antigens in refractory CD and UC using HuNoV virus-like particles (VLPs) and histological tissues. Immunohistochemistry was conducted on inflammatory tissue samples from the small intestine, colon, and rectum in 15 CD and 9 UC patients. Analysis of the regenerative mucosa of the colon and rectum revealed strong expression of sialylated Lewis a (sLe^a) and Lewis x (sLe^x) antigens and HuNoV VLP binding in the absence of ABO antigen expression in both UC and CD. Competition experiments using sialidase, lectins, and monoclonal antibodies demonstrated that HuNoV attachment mostly involved Le^a and, to a lesser extent, Le^x moieties on regenerative mucosa in both UC and CD. Further studies will be required to understand the implications of specific HuNoV binding to regenerative mucosa in refractory IBD.IMPORTANCE Inflammatory bowel diseases (IBD), including Crohn's disease (CD) and ulcerative colitis (UC), are progressive diseases affecting millions of people each year. Flare-ups during IBD result in severe mucosal alterations of the small intestine (in CD) and in the colon and rectum (in CD and UC). Immunohistochemical analysis of CD and UC samples showed strong expression of known tumoral markers sialyl Lewis a (CA19.9) and sialyl Lewis x (CD15s) antigens on colonic and rectal regenerative mucosa, concurrent with strong human norovirus (HuNov) VLP GII.4 affinity. Sialidase treatment and competition experiments using histo-blood group antigen (HBGA)-specific monoclonal antibodies and lectins clearly demonstrated the implication of the Lewis a moiety and, to a lesser extent, the Lewis x moiety in HuNov recognition in regenerative mucosa of CD and UC tissues. Further studies are required to explore the possible implications of enteric viruses in the impairment of epithelial repair and dysregulation of inflammatory pathways during severe IBD.

Keywords: Crohn’s disease; HBGA; gut inflammation; inflammation; norovirus; ulcerative colitis.

FIG 1

VLP binding specificity demonstrated in a healthy duodenal sample from a blood group A patient who underwent duodenopancreatectomy. Slides were pretreated with either 50 mM sodium periodate (NaIO₄) or α1,2 fucosidase (boiled or cold) before incubation with HuNoV GII.4 VLP. Mutated ΔD373 GII.4 VLPs were used as negative controls. For this experiment and the following, VLP binding was detected with GII.4 VP1-specific MAb labeled with peroxidase. Peroxidase activity was detected by colorimetry using H₂O₂ and 3,3′-diaminobenzidine (DAB), giving a brown staining. Mock and pretreatments are indicated above each image.

FIG 2

HBGA detection and HuNoV binding in CD ileum (patient 5), proximal colon (patient 9), transverse colon (patient 12), and distal colon (patient 15) showing quiescent mucosa. For this figure and the next two figures, blood group A, Le^a, Le^x, sLe^a, and sLe^x antigens were detected with specific Mabs in a colorimetric assay using DAB. The α1,2 fucose moiety characterizing the H antigen was detected using fluorescein isothiocyanate (FITC)-conjugated UEA-I lectin. Detected antigens are indicated above each column. Patient numbers and tissue sample types are indicated at the left of each row.

FIG 3

HBGA detection and HuNoV binding in CD ileum (patient 6), transverse colon (patient 13), and sigmoid (patient 16) showing regenerative mucosa. Detected antigens are indicated above each column. Patient numbers and tissue sample types are indicated at the left of each row.

FIG 4

HBGA detection and HuNoV binding in UC colon (patient 19) and rectum (patient 23) with quiescent mucosa (QM), and sigmoid (patient 22) and rectum (patient 24) with severe inflammation and regenerative mucosa (RM). The detected antigen is indicated above the column panel. QM, RM, patient numbers, and tissue sample types are indicated at the left of each row.

FIG 5

Role of the sialic acid moiety into HuNoV attachment. Attached VLP to mucosa and presence of sLe^a and sLe^x were detected in CD (patient 13) and UC (patient 22) macrodissected tissue samples following sialidase treatment. The efficacy of the sialidase treatment for the removal of the sialic acid was controlled by the absence of immunostaining following incubation with sLe^a- and sLe^x-specific Mabs. Mock and sialidase treatments are indicated on the left of each row. Detected antigens are indicated above each column panel. Specific staining is indicated with an arrowhead.

FIG 6

Role of the Lewis antigens in HuNoV specific attachment to pathological CD (patient 13) and UC (patient 22) mucosae. Tissues were first preincubated with a combination of sialidase and lectins (UEA-I and HPA). Presence or absence of treatment is indicated above the column panel by positive and minus signs, respectively. The objective was then to specifically inhibit VLP with Lewis-specific Mabs. Slides were first pretreated with specific MAbs against Le^a (αLe^a), Le^x (αLe^x), or both combined before HuNoV GII.4 VLP incubation. The Mabs used for the experiments are indicated in bracket at the left side of each row. For the control, VLPs were directly incubated on tissue without any pretreatment with sialidase, lectins, or Mabs. VLP attachment is indicated by arrowheads and magnified areas correspond to dashed boxes.