|Home | About | Journals | Submit | Contact Us | Français|
Noroviruses (formerly Norwalk-like viruses) are a major cause of acute gastroenteritis worldwide and are associated with a significant number of nosocomial and food-borne outbreaks. In this study we show that the human secretor FUT2 gene, which codes for an α(1,2)-fucosyltransferase synthesizing the H-type 1 antigen in saliva and mucosa, is associated with susceptibility to norovirus infections. Allelic polymorphism characterization at nucleotide 428 for symptomatic (n = 53) and asymptomatic (n = 62) individuals associated with nosocomial and sporadic norovirus outbreaks revealed that homozygous nonsense mutation (428G→A) in FUT2 segregated with complete resistance for the disease. Of all symptomatic individuals, 49% were homozygous (SeSe) and 51% heterozygous (Sese428) secretors, and none were secretor negative (se428se428), in contrast to 20% nonsecretors (se428se428) among Swedish blood donors (n = 104) (P < 0.0002) and 29% for asymptomatic individuals associated with nosocomial outbreaks (P < 0.00001). Furthermore, saliva from secretor-positive and symptomatic patients but not from secretor-negative and asymptomatic individuals bound the norovirus strain responsible for that particular outbreak. This is the first report showing that the FUT2 nonsecretor (se428se428) genotype is associated with resistance to nosocomial and sporadic outbreaks with norovirus.
Noroviruses are the major cause of acute gastroenteritis worldwide among adults and are associated with the illness “winter vomiting disease,” characterized by a short incubation period (24 to 48 h) and significant vomiting and diarrhea. While the viruses are highly contagious, with attack rates up to 70% (12), volunteer challenge studies have shown that a subset of individuals remain uninfected even after repeated challenges (2, 20, 26). At present, it is not clear why a fraction of individuals remain uninfected in norovirus outbreaks and why some volunteers are repeatedly resistant to experimental Norwalk virus inoculations. Recently, it was suggested that histo-blood group antigens and the secretor status might be associated with experimental Norwalk virus infections (4, 10, 11, 16, 17).
The FUT2 gene, which is responsible for the secretor phenotype, encodes an α(1,2)-fucosyltransferase that regulates the expression of the ABH antigens in saliva and mucosal tissues and secretions. The FUT2 gene has a significant polymorphism with typical ethnic specificity (13). The nonsense mutation 428G→A (Trp143→stop) is characteristic for the dominating nonsecretor allele (se428) in Europeans and appears in about 20% of the Caucasian population (13).
The facts that about 20% of Europeans are nonsecretors and norovirus attack rates seldom exceeds 80% in a given outbreak led us to investigate, throughout a series of prospective studies, if the FUT2 secretor gene was associated with resistance to nosocomial and sporadic outbreaks caused by genogroup II (GGII) noroviruses, which dominate in all parts of the world (14).
The study included 115 saliva samples, of which 53 were collected from norovirus-infected and symptomatic individuals and 62 from asymptomatic individuals.
A patient with gastroenteritis was defined as a patient with vomiting (one or more times in 24 h) and/or diarrhea (more than two watery stools in 24 h).
All fecal samples were screened for viruses by electron microscopy (EM) as described previously (6).
RNA was extracted from 100 μl stool suspension by use of the guanidinium thiocyanate-silica extraction method (1).
Specimens were investigated for the presence of Norwalk-like human calicivirus by a reverse transcription-PCR (RT-PCR) as previously described by Vinjé et al. (23, 24). Briefly, RNA was extracted from 50 μl of a 10% stool suspension using the guadinium thiocyanate-silica extraction method (1). RT was performed as follows. Five microliters of extracted RNA was annealed with 50 pmol JV 13 (5′-TCA TCA TCA CCA TAG AAA GAG) in a total volume of 9 μl and then added to a reaction mix containing 10 mM Tris, pH 8.3, 50 mM KCl, 3 mM MgCl2, 1 mM deoxynucleoside triphosphate, and 100 units of Moloney murine leukemia virus RT (Superscript; Life Technologies). Reaction mixtures were incubated for 1 h at 42°C. Five microliters of the RT reaction was added to a PCR mix composed of 10 mM Tris, pH 9.2, 75 mM KCl, 1.5 mM MgCl2, 0.2 mM deoxynucleoside triphosphate, 15 pmol JV12 (5′-ATA CCA CTA TGA TGC AGA TTA), and 2.5 U Taq polymerase. Forty reaction cycles were carried out with 1 min at 94°C, 1.5 min at 37°C, and 1 min at 74°C followed by a final incubation at 74°C for 7 min. One-fifth of the reaction volume was analyzed on agarose gels. A positive and specific reaction resulted in a 326-bp product, located in the gene for the RNA-dependent RNA polymerase.
DNA was extracted from saliva samples, and the FUT2 (secretor) genotype was determined using sequence-specific primers and PCR (PCR-SSP) as described previously (3, 21). The same antisense primer (5′-GGCTGCCTCTGGCTTAAAG) was used in all PCR amplifications but was paired with the sense primers for FUT2 wild-type 385A (5′-AGGAGGAATACCGCCACAT), mutated 385T (5′-GAGGAGGAATACCGCCACT), wild-type 428G (5′-GCTACCCCTGCTCCTGG), mutated 428A (5′-CGGCTACCCCTGCTCCTA), wild-type 571C (5′-TAGGGGTCCATGTTCGCC), and mutated 571T (5′-GTAGGGGTCCATGTTCGCT), giving fragments of 574-, 575-, 530-, 532-, 388-, and 390-bp lengths. The human growth hormone gene was used as an internal PCR control, giving a 428-bp product.
Saliva samples were boiled, centrifuged at 10,000 × g for 5 min, and diluted 1:500 in enzyme-linked immunosorbent assay (ELISA) coating buffer (0.1 M carbonate-bicarbonate buffer, pH 9.6) essentially as described previously (4, 17). After 2 h of incubation at 37°C followed by overnight incubation at 4°C, the plates were washed three times with washing buffer and blocked (bovine serum albumin [3%]-phosphate-buffered saline) for 60 min at 37°C followed by three more washes. Stool sample suspensions (10%) were diluted 1:2 in phosphate-buffered saline-Tween 20 (0.05%)-bovine serum albumin (0.5%) and incubated for 2 h at 37°C followed by three washes and incubation with peroxidase-labeled genogroup I and II-specific norovirus polyclonal antibody (DAKO, Denmark) for 1 h at 37°C. The reaction was developed using 3′,3′,5′,5-tetramethylbenzidine (ICN Biochemicale).
Norovirus-specific PCR products were sequenced using the BigDye Terminator cycle sequencing kit (Perkin-Elmer) on an automated ABI PRISM model 3100 machine. Sequence analysis, alignments, and phylogenetic comparisons were done using the Lasergene software package (DNASTAR, Inc., Madison, Wis.).
Fisher′s exact test (two-sided) was used to test significance of differences in distribution of secretor-positive (SeSe or Sese428) and secretor-negative (se428se428) individuals among symptomatic, asymptomatic, and control individuals.
The study was approved by local ethical committees at the involved universities and included informed consent for genetic testing of saliva samples.
Information about the secretor status in the Swedish population was obtained from 104 unselected plasma donors genotyped by PCR-restriction fragment length polymorphism for the FUT2 gene 428G→A nonsense mutation (8, 15). A majority, 54%, were heterozygous secretors (Sese428), 26% homozygous secretors (SeSe), and 20% homozygous nonsecretors (se428se428), consistent with previous reports that about 20% of Europeans are genetically nonsecretors.
On 20 September 2002, a patient with gastroenteritis was transferred to an internal medicine ward (ward A). Two days later, an assistant nurse in the ward fell ill with acute vomiting and diarrhea. On the following 5 days, 14 additional medical staff and 11 patients on the ward got sick with gastroenteritis. Thus, all together, 15 medical staff and 12 patients had gastroenteritis within 8 days (Fig. (Fig.1,1, Ward A). The symptoms and the outbreak were in agreement with a norovirus infection (i.e., acute onset of gastroenteritis that lasts for a few days) (12). A stool sample taken from one symptomatic patient contained norovirus as determined by electron microscopy and norovirus-specific PCR (12). Genotyping revealed that the virus belonged to the genogroup II cluster and was Lordsdale-like (LDV), a genotype most commonly found in nosocomial outbreaks in Europe (14).
Saliva samples were obtained from 41 of the medical staff and from 19 of the patients, including 10 of the 12 patients and 13 of the 15 medical staff with gastroenteritis. DNA was obtained from 50 of the 60 saliva samples and examined for mutations partially or completely inactivating FUT2 (385A→T, 571C→T, and 428G→A) by PCR-SSP, a technique previously used for Lewis secretor and ABH genotyping (3, 21) (Fig. (Fig.2).2). Of the symptomatic individuals from this outbreak, 47% were homozygous secretors (SeSe), 53% heterozygous secretors (Sese428), and most interestingly, none were secretor negative (Fig. (Fig.3),3), in contrast to 19% (se428se428) who either were not exposed or were resistant to infection. Three of the six nonsecretors were medical staff (ages 39, 45, and 59), and three were patients (ages 88, 89, and 89). Interestingly, two of the three secretor-negative patients were nursed together in the same rooms as symptomatic patients with gastroenteritis, of which one was the index patient. In spite of the potential exposure from a symptomatic roommate, none of these two secretor-negative patients had any clinical symptoms of gastroenteritis. To investigate if the outbreak strain could bind to saliva from infected individuals, a saliva-based ELISA was established. As shown in Fig. Fig.4,4, saliva from secretor-positive and ill individuals but not from secretor-negative individuals bound the LDV associated with the outbreak (LDV, Ward A).
The second nosocomial outbreak with norovirus occurred in a pediatric ward (Fig. (Fig.1,1, Ward B) on 22 November 2002 with an index patient with acute gastroenteritis. Within the following 6 days, 12 patients and 19 medical staff fell ill with gastroenteritis (Fig. (Fig.1,1, Ward B). In fecal samples collected from ill patients, norovirus were identified by electron microscopy and PCR for three patients. Genotyping by reversed line blotting (25) revealed that all viruses again belonged to genogroup II and were Lordsdale-like. Saliva samples were collected from 38 of the medical staff, of which 12 had symptoms. Unfortunately, saliva could not be collected from patients. Of the 38 saliva samples investigated, secretor status could be established in 28. None of the 7 symptomatic medical staff was secretor negative, in contrast to 43% (9/21) secretor-negatives among the medical staff that remained without clinical symptoms (Fig. (Fig.33).
A third nosocomial outbreak affected orthopedic patients (Fig. (Fig.1,1, Ward C) and started on 8 December 2002 with the introduction of a patient with gastroenteritis. One day later, another patient in the ward had similar symptoms, and thereafter, patients and medical staff fell ill during the following 10 days. In all, 14 patients and 28 medical staff experienced gastroenteritis. The epidemic and symptoms were in agreement with norovirus infection. Norovirus was also confirmed by PCR in stool samples from four patients. Once again the norovirus identified belonged to genogroup II and the Lordsdale-like cluster. Saliva samples were obtained from 26 of the medical staff, of which 16 had gastroenteritis. Saliva samples could not be collected from patients. Secretor status was established from 18/26 salivas, and none of the 12 symptomatic individuals was a nonsecretor, in contrast to 50% (3/6) among those who remained asymptomatic. To better understand the phenotype correlation between secretor status and resistance to norovirus infection, saliva samples from secretor and nonsecretor individuals were tested in a saliva ELISA. As shown in Fig. Fig.4,4, saliva from ill secretor individuals but not from nonsecretors recognized the virus strain responsible for the outbreak, strongly suggesting that receptors for norovirus GGII are present in saliva from secretors but not from nonsecretors.
To obtain further information about the role of FUT2 in norovirus infections, fecal and saliva samples were collected from community outbreaks. Of salivas collected from 19 individuals associated with 3 community outbreaks with noroviruses belonging to genotypes GGII/Melksam, GGI/Sindlesham, and another GGII strain, 4 individuals were asymptomatic and 15 symptomatic with vomiting and diarrhea. Of the asymptomatic individuals, two were Sese428 and two SeSe, and of the 15 symptomatic individuals, 47% (7/15) were SeSe and 53% (8/15) Sese428. Thus, none of the 15 symptomatic individuals was a nonsecretor.
Disease studies have given different results on whether individual ABO (H) blood groups are related to an increased risk of norovirus infection or not. Hutson et al. (10) first reported that individuals with the O phenotype were more likely to be infected with Norwalk virus belonging to genogroup I viruses, whereas individuals with an A or B histo-blood group antigen had a decreased risk of infection. A similar observation was made by Hennessy et al. (7) in investigating a norovirus outbreak at a British military field hospital, finding a reduced susceptibility of blood group B individuals to symptomatic norovirus infections. These observations are in contrast to those of Meyer et al. (18), who reported that persons with the O phenotype were significant less affected than expected from the normal distribution. Furthermore, Rockx et al. (22) recently reported that individuals secreting type B antigens in saliva were significantly protected against infection with genogroup I virus (2 of 22 individuals) and were also less likely to acquire norovirus-specific immunoglobulin G.
Possible explanations for these different observations may be that the secretor status was not determined in all cases and, secondly, whether it was determined by phenotyping rather than by genotyping (3, 19, 21). Norwalk virus have been found to bind to gastrointestinal cells in vitro independently of ABO (H) blood groups but with dependence on secretor status (17) and found to correlate to resistance against genogroup I viruses (16), suggesting that the FUT2 gene and secretor status are susceptibility markers for Norwalk virus infection. The fact that nonsecretors do not express blood group A or B in saliva or other body fluids could suggest a coincidental correlation between the ABO phenotype in blood and infection of the intestinal mucosa. Indeed, individuals can have the A or B phenotype in the blood, but if secretor negative, they do not express the A or B antigen in saliva or mucosa. However, it cannot be ruled out that certain noroviruses are secretor independent, since multiple receptor specificities have been reported from in vitro studies (5, 9).
Our studies, the first to include a host genetic approach and to examine secretor status for nosocomial and sporadic outbreaks, revealed several novel observations. We report for the first time that a nonsense mutation (428G→A) in the FUT2 gene is strongly associated (P < 0.00001) with resistance to winter vomiting disease. In fact, all (100%) individuals with a confirmed norovirus infection studied have so far been secretor positive (Table (Table11).
Furthermore, we also show that only homozygous nonsecretors (se428se428) are protected. The inactivating mutation examined in this study is by far the most common in Europe and was found to occur homozygously in 20% among Swedish plasma donors (Table (Table1).1). However, it should been mentioned that protective immunity might very well play a role, although correlates of protection remains to be identified.
In vitro studies have shown that binding of Norwalk-like viruses to ABO, Lewis, and secretor histo-blood group antigens are strain specific and that different attachment mechanisms exist (4, 9). Since cells transfected with the FUT2 gene enhance norovirus binding to nonpermissive cells (17), it is reasonable to believe that H type 1 or related blood group antigens act as receptors for norovirus. Our study supports this hypothesis, and we observed that the saliva ELISA by Marionneau et al. (17) was successfully applied to show that the virus causing disease also could be recognized by the saliva of the patient. This new observation, together with our secretor polymorphism data, provide biochemical and genetic explanations for resistance to noroviruses belonging to the common GGII Lordsdale-like cluster.
This study was supported by the Swedish Research Council (grant 8266), the European Union (grants QLRT-1999-00634 and QLRT-1999-00594), and the Health Research Council of Southeast Sweden.
We thank Gustaf Rydell for statistical calculations, Ann-Christin Hammarlund at the Hospital Infection Control Unit, and the medical staffand patients at the departments of Medicine, Orthopaedics, and Pediatrics, Lund University Hospital, Lund, who participated in the studies.