On January 15, 2012, a 13-month-old girl enrolled in a childcare center in Rockville, Maryland, experienced 3 episodes of vomiting within 1 hour, after which she had diarrhea or loose stools for ≈1 week. Within 24 hours after this child’s onset of symptoms, 2 family members reported multiple episodes of vomiting and diarrhea that lasted >3 days. Because several children and teachers at the childcare center reported similar symptoms, parents of children enrolled at the childcare center were alerted to the possibility of a gastroenteritis outbreak.

The patient reported here was subsequently enrolled in National Institutes of Health clinical study NCT01306084, after receipt of informed consent from the mother. Fecal samples were collected from the child and examined for norovirus RNA by reverse transcription PCR. Viral RNA was detected for 4 weeks after the onset of symptoms, and viral RNA quantification reached up to 1.7 × 10⁸ genome copies/g feces. Sequence and phylogenetic analyses of VP1 from the virus (designated norovirus Hu/GII.4/RockvilleD1/2012/U.S.) showed that it grouped within the newly emerging virus GII.4 Sydney_2012 cluster (Figure 1, panel A) (11).

On December 10, 2012, a gastroenteritis outbreak occurred in a different childcare center in Bethesda, Maryland, at which the same child (now 24 months of age) was enrolled. The child experienced vomiting, diarrhea, and fatigue that lasted ≈2 days, and a similar disease pattern developed in a family member 24 hours after the onset of the child’s symptoms. Fecal samples were again positive for norovirus with 5.3 × 10⁹ genome copies/g of feces, and viral RNA was detected up to 3 weeks after disease onset. Phylogenetic and sequence analyses revealed a GII.6 norovirus (designated norovirus Hu/GII.6/BethesdaD1/2012/U.S.), most closely related to GII.6 noroviruses detected in Miami (Florida, USA) and Texas (USA) in 1994 and 1997, respectively (Figure 1, panel A).

Norovirus strains GII.4 and GII.6 differed by ≈38% in VP1 sequences; most amino acid sequence variation occurred in the capsid protruding domain (29%; 163/556). Alignment of the VP1 sequences from the 2 strains in this study showed several gaps; each strain bore discrete regions of amino acid insertions or deletions that differed from those in the other strain. The GII.6 VP1 (547 aa long) contained 3 insertions at positions 296 (11 residues), 310 (2 residues), and 344 (3 residues); GII.4 VP1 (540 aa long) did not contain these insertions. The same alignment showed GII.4 VP1 insertions at positions 190 (1 residue), 373 (1 residue), and 390 (7 residues) (Figure 2, Appendix). Of note, most gaps in the alignment of the VP1 sequences were present in or near recently described GII.4 epitopes (7,12), thereby suggesting that these residues might play a role in defining the antigenic specificity of the 2 genotypes.

To understand the absence of immunity to the second norovirus and to gain insight into the specificity of the mucosal immune response after sequential norovirus infections, we developed virus-like particles from both noroviruses. We used the Bac-to-Bac Baculovirus Expression System (Invitrogen, Carlsbad, CA, USA) and baculovirus–infected Sf-9 cells for virus-like particle production as described elsewhere (7). The corresponding virus-like particles were used to test for IgA in feces with an ELISA that used polyclonal goat anti-human IgA conjugated with horseradish peroxidase (1:2,000 dilution; KPL, Gaithersburg, MD, USA) as the detector antibody and 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS; KPL) as the substrate. During the child’s first infection with norovirus GII.4, virus-specific IgA was not detectable until 3 weeks after infection (Figure 1, panel B). At the onset of symptoms during the second infection with norovirus GII.6, a rapid anamnestic response specific for GII.4 virus-like particles developed. The GII.6-specific IgA response was detected at day 6 after infection, and IgA was detectable in feces until the last daily collection at 4 weeks (Figure 1, panel B). Examination of feces for norovirus-specific IgG after the second infection showed a similar anamnestic response that endured throughout the period examined (data not shown).

The child’s clinical history, young age, and absence of early detectable norovirus-specific fecal IgA suggest that she experienced a primary infection with GII.4 norovirus in the first childcare center outbreak. During the second infection with GII.6 norovirus, a rapid anamnestic GII.4-specific IgA response developed and persisted up to 4 weeks as the slower primary GII.6-specific IgA response developed ≈1 week after symptom onset. Follow-up fecal samples collected at 14 weeks after the second infection contained little or no detectable levels of either GII.6- or GII.4-specific IgA, indicating an eventual decline in mucosal IgA titers against each genotype (data not shown).

Previous data from human volunteer studies suggest that homologous immunity to Norwalk virus (genotype GI.1) lasts from 2 months to 2 years (13) and that a rapid and specific mucosal IgA response (likely anamnestic) was a correlate of protection (14). The data presented here are consistent with the development of a genotype-specific, short-lived mucosal IgA response to norovirus infection that, when stimulated anamnestically, might provide little or no protection against other norovirus genotypes.

This study shows that a young child can experience 2 episodes of acute gastroenteritis caused by distinct norovirus genotypes (GII.4 and GII.6) within 1 year. Reinfection with distinct genotypes can commonly occur in younger persons, as recently demonstrated in a longitudinal study of norovirus infection in infants and young children in Peru (15). Genotypes within the 2 major genogroups of human noroviruses might represent distinct serotypes. The mechanisms of enduring norovirus immunity in the development of cross-protective and effective vaccines need to be elucidated.