Use of Saliva to Identify Varicella Zoster Virus Infection of the Gut

Abstract

Background: Varicella zoster virus (VZV) establishes latency in dorsal root, cranial nerve, and enteric ganglia and can reactivate to cause zoster. Serious gastrointestinal dysfunction can result from VZV reactivation in enteric neurons (enteric zoster), but an absence of rash makes diagnosis difficult. We thus determined whether detecting VZV DNA in saliva facilitates identification of enteric zoster.

Methods: Nested and real-time polymerase chain reaction were used to validate salivary VZV DNA as a surrogate marker of VZV reactivation and then to determine the utility of that marker for the identification of those individuals within a population defined by abdominal pain that might have enteric zoster.

Results: Salivary VZV DNA was detected in 0 of 20 healthy negative controls, 11 of 16 positive controls with zoster or varicella (P < .0001), 2 of 2 patients with zoster sine herpete (P < .01), 6 of 11 patients with unexplained abdominal pain (P < .001), and 0 of 8 patients with unrelated gastrointestinal disorders. Salivary VZV DNA disappeared after recovery in 9 of 9 tested subjects with zoster, 2 of 2 with zoster sine herpete, and 5 of 5 with abdominal pain. One patient with abdominal pain and salivary VZV DNA had perforated gastric ulcers, necessitating a wedge gastrectomy. VZV DNA (vaccine type) was found in the resected stomach; immediate early (ORF63p) and late (gE) VZV proteins were immunocytochemically detected in gastric epithelium. After recovery, VZV DNA and proteins were not detected in gastric biopsies or saliva.

Conclusions: Detection of salivary VZV DNA in patients with abdominal pain helps to identify putative enteric zoster for investigation and treatment.

Keywords: abdominal pain; bleeding peptic ulcer; enteric infections; enteric nervous system; enteric neurons.

Figure 1.

Frequency of salivary varicella zoster virus (VZV) DNA in saliva of negative control (healthy) subjects, positive control (zoster/varicella) subjects, patients with zoster sine herpete, and patients with unexplained abdominal pain (without rash). A χ² analysis of the overall distribution (zoster/varicella and zoster sine herpete groups combined) was significantly different from chance (P < .0001). A, The difference in frequency between the detection of salivary VZV DNA in the positive control population is significantly greater than that in the negative control population. Fisher exact test. B, The difference in frequency between the detection of salivary VZV DNA in patients with zoster sine herpete is significantly greater than that in the negative control population. Fisher exact test. C, The difference in frequency between the detection of salivary VZV DNA in the group of patients with unexplained abdominal pain is significantly greater than that in the negative control population. Fisher exact test.

Figure 2.

Varicella zoster virus (VZV) infection was detected in saliva and resected stomach from a patient with abdominal pain, bleeding, and perforated gastric ulcer. A, DNA from saliva amplified with polymerase chain reaction. DNA encoding open reading frame (ORF) 40 and ORF67 was amplified in the patient's saliva (Pt) and in the positive control DNA prepared from a plasmid (+) but not in the negative control (−). B, The genotype of the viral DNA from resected stomach is that of vOka. DNA was digested with the Pst1 restriction endonuclease. C, Low-power micrograph showing the ulcerated mucosa of the patient's resected stomach. This is the region of the stomach from which DNA was extracted. Hematoxylin and eosin (H&E) stain; scale bar = 100 µm. D, Muscularis of the stomach. The myenteric plexus is not evident. H&E stain; scale bar = 100 µm. Abbreviations: M, size marker; Pt, DNA extracted from the resected stomach of the patient; vOka, DNA extracted from cells infected with vOka; WT, DNA extracted from cells infected with wild-type VZV (Dumas).

Figure 3.

The immunoreactivity of the varicella zoster virus proteins glycoprotein E (gE) and open reading frame (ORF) 63p are present and can be detected immunocytochemically in the mucosa of the resected stomach. A_1–6, The same field is shown in all 6 panels, illuminated to demonstrate the fluorescence of DNA (A₁), gE (A₂), and ORF63 (A₃), as well as an interference contrast image of the mucosa (A₄), the merged image of DNA + gE + ORF63 (A₅), and the merged fluorescence and interference contrast images (A₆). Note that the immunofluorescence of gE and ORF63 are coincident in the mucosa and are found in epithelial cells, underlying stromal cells of the lamina propria, and some mucosal blood vessels (linear structures). B, Control immunoreactivity. The primary antibodies to gE and ORF63p were omitted, but the sections were processed as in A. B_1–6. Again, a single field is illustrated in all 6 panels, illuminated to demonstrate the fluorescence of DNA (B₁), gE (B₂), ORF63 (B₃), an interference contrast image (B₄), the merged image of DNA + gE + ORF63 (B₅), and the merged fluorescence and interference contrast images (B₆). No green or red fluorescence can be detected. Scale bars = 50 µm.

Figure 4.

An endoscopic biopsy of the gastric mucosa obtained after the recovery of the patient following gastric surgery. The mucosa is now normal and no immunoreactivity of glycoprotein E (gE) or open reading frame (ORF) 63p can be detected. A_1–6. The same field is illustrated in all 6 panels, illuminated to demonstrate the fluorescence of DNA (A₁), gE (A₂), ORF63 (A₃), an interference contrast image (A₄), the merged image of DNA + gE + ORF63 (A₅), and the merged fluorescence and interference contrast images (A₆). The structure of the mucosa has been repaired. No ulceration, inflammation, or varicella zoster virus immunoreactivity can any longer be detected. Scale bar = 50 µm.