Adenovirus Strains of Subgenus D Associated with Nosocomial Infection as New Etiological Agents of Epidemic Keratoconjunctivitis in Japan

At present there are 49 serotypes of adenovirus (Ad) (5, 19, 24), and they have been classified into six subgenera, A to F (22). Ad group has been shown to cause upper and lower respiratory tract infections, keratoconjunctivitis, and hemorrhagic cystitis (8). Ad3 and -4 are the most common causes of sporadic keratoconjunctivitis and pharyngoconjunctival fever (PCF), and Ad8, -19, and -37 have been responsible for sporadic cases as well as outbreaks of severe epidemic keratoconjunctivitis (EKC) in several countries, especially in east and southeast Asia, including Japan (1-4, 7). Ad8, -19, and -37 are also well known to be etiological agents for nosocomial infections (9, 12, 23).

In the autumn of 1996, we encountered a case of nosocomial infection with severe keratoconjunctivitis in the ophthalmology ward of Yokohama City University Hospital. To determine the serotype of this newly isolated Ad, two new molecular biological techniques developed by us were performed (16, 20) in addition to the standard neutralization test (NT). Conjunctival swabs were taken from eight patients with acute follicular conjunctivitis within 7 days of onset. Swabs from the conjunctiva of symptomatic eyes were collected and inoculated onto HEp-2 cells and A549 cells, which were examined for the presence of a cytopathic effect (CPE) up to 28 days. Infected cells were also identified by the fluorescent-antibody technique using mouse monoclonal antibody against Ad (Chemicon International, Inc., Temecula, Calif.), herpes simplex virus type 1 (HSV-1), HSV-2 (Micro Trak; Syva Co., Palo Alto, Calif.), and Chlamydia (Micro Trak; Syva Co.).

NT was performed by the microtiter method using HEp-2 cells and A549 cells to provide an index of CPE. All Ad prototypes and antisera against Ad1–11, -14, -19, -22, -34, -35, -37, -40, and -41 were obtained from the American Type Culture Collection (Rockville, Md.).

DNA preparation and PCR-restriction fragment length polymorphism (RFLP) analysis from the conjunctival scrapings were performed according to the method previously developed by us (16). Briefly, nested PCR was performed to amplify the 956-bp DNA fragment of the hexon conserved region. Then, differences in the restriction patterns with three restriction enzymes, EcoT14I, HaeIII, andHinfI, were combined, and the serotypes of the clinical specimens were identified by comparison with Ad prototypes.

A new method developed by us was also used to determine the serotypes (20). This method detects Ad and identifies the serotype using a combination of nested PCR and direct sequencing method targeting hypervariable regions (HVRs) that participate in serotype-specific neutralization. Direct sequencing was carried out for the nested PCR products containing seven HVRs and these sequences were determined.

Sequence data described in this article have been deposited with the GenBank/EMBL/DDBJ Data Libraries under the following accession numbers. For subgenus B2: Ad11 hexon gene, AB018424 ; Ad14, AB018425 ; Ad34, AB018426 ; Ad35, AB018427 . For subgenus D: Ad8, AB023546 ; Ad9,AB023547 ; Ad10, AB023548 ; Ad17, AB023549 ; Ad19, AB023550 ; Ad22,AB023551 ; Ad23, AB023552 ; Ad24, AB023553 ; Ad26, AB023554 ; Ad37,AB023555 ; Ad45, AB023556 ; Ad46, AB023557 ; Ad47, AB023558 . The amino acid sequences of these residues were deduced. Deduced amino acid sequences for adenoviruses (with accession numbers and references) taken from previous reports were as follows. Those for subgenus B1 were Ad3, X76549 (15), and Ad7, X76551 (14). Those for subgenus C were Ad1, X67709 (13); Ad2, J 01917 (10); Ad5, X76550 (11); and Ad6, X67710 (13). That for subgenus E was Ad4, X84646 (15). Those for subgenus F were Ad40, X51782 (22), and Ad41,X51783 (21).

Virus was isolated from three of eight patients’ clinical samples. Cultures were subpassaged several times, and it took more than 3 weeks to obtain CPE of 4+ in cultured cells. HEp-2 and A549 cell cultures harvested 2 days after showing 4+ CPE were subjected to various tests for Ad infection. Cultures exhibiting 3+ to 4+ CPE were stained with diluted group-specific anti-adenovirus hexon monoclonal antibody by direct fluorescence-antibody test.

In NT tests, these isolates were not significantly neutralized with rabbit immune serum to the Ad prototype used in this study. However, weak reactions between these isolates and Ad8 or Ad9 were observed. The serotype of these isolates was tentatively named Ad8/9. Antiserum against Ad9 gave titers 128-fold lower against these strains than against Ad9 prototype antigen. On the other hand, antiserum against Ad8 gave titers 16-fold lower against these strains than against Ad8 prototype antigen.

The PCR-RFLP cleavage patterns of the eight clinical samples were identical and showed the same patterns as those of Ad8 prototype (Fig.1). The restriction enzyme patterns of clinical samples were different from those of Ad9 prototype with two of three enzymes, EcoT14I and HaeIII.

• Open in new tab
• Download powerpoint

Fig. 1.

Agarose gel electrophoresis shows the cleavage patterns of the 956-bp amplified products digested by EcoT14I (a),HaeIII (b), and HinfI (c). Numbers above the lanes correspond to serotypes of Ad and lane M contains molecular weight standards (φX174 HincII digest). Among subgenus D, we could distinguish seven serotypes of the Ad prototype employed in this study, Ad8, -9, -10, -17, -19, -22, and -37, with two enzymes,EcoT14I and HaeIII.

The amino acid sequences of the hexon HVR genes derived from the nucleotide sequences of the eight clinical samples were identical. The clinical isolates, Ad8/9, had a higher average homology in the hexon HVRs with subgenus D serotypes than with other serotypes. This indicates that these strains belong to the subgenus D. Among the serotypes of subgenus D, Ad8/9 had the highest mean maximum homology with Ad9 prototype, 75.0%. In five out of seven HVRs, HVR1, -2, -3, -5, and -7, the highest DNA homology was seen between Ad8/9 and Ad9, 65.7, 94.1, 90.9, 76.2, and 81.8%, respectively (Table1).

In this study, we isolated three strains of an uncommon Ad which differed from established Ads as a causative agent of conjunctivitis from eight patients diagnosed clinically as having EKC. The fact that our present strain showed a weak reaction with rabbit hyperimmune serum to both Ad8 and Ad9 means this strain may be an intermediate strain between Ad8 and Ad9, called intermediate-type virus, Ad8/9. It may have originated by recombination of Ad8 and Ad9 within the neutralization related antigen. PCR-RFLP analysis using the conserved region of the hexon identified this Ad8/9 strain as Ad8. These results suggest that this strain is most similar to Ad8 in whole hexon region and has partial replacement in the area of the neutralization epitopes, that is, HVR1 to HVR7, leading to a sequence similar to that of Ad9. From the viewpoint of the evolution of Ads, it is interesting to speculate on the origin of this strain. The recombination in the hexon region seen in this Ad8/9 has not been previously reported, and simultaneous outbreaks caused by the two Ad serotypes have been reported (18).

In the PCR-sequencing method using seven HVRs, Ad8/9 had the highest mean maximum homology with Ad9 prototype, 75.0%. However, this rate is not so high. These results support the finding that Ad8/9 was not neutralized completely by antiserum against Ad9. In our previous study, the predicted amino acid homology of these HVRs suggested three regions, HVR4, -5, and -7, to be good candidates for neutralization epitopes, because of their low homology among human Ads (20). In fact, the average homology rate of HVR4, -5, and -7, 76.5%, indicated that Ad8/9 was most closely related to Ad9. The highest average homology in HVR4, -5, and -7, between Ad8/9 and Ad9, strongly supported the weak reaction in NT with rabbit immune serum to Ad9.

As a result of the sudden emergence of nosocomial infection in association with conjunctivitis, it would be of interest to investigate whether the properties of Ad8/9 vary from those of Ad8 and Ad9 prototypes and whether it is serologically related to the other Ad associated with conjunctivitis. It remains to be determined whether this Ad8/9 is a new Ad serotype, candidate Ad50.

Some patients showed corneal involvement in the form of subepithelial punctate keratitis and enlarged preauricular lymph nodes, which is commonly seen in Ad8 infection (1, 7). The conjunctival symptoms of acute follicular conjunctivitis were not similar to those of PCF, but were similar to those of EKC. EKC caused by subgenus D, Ad8, Ad19, and Ad37 has been reported to be associated with many cases of nosocomial infection (1, 2, 4, 6). In the case of infection in our hospital, it was shown that the degree of contagiousness or pathogenicity for this serotype, Ad8/9, is the same as that for Ad8, Ad19, and Ad37, belonging to subgenus D.

In our previous survey, we did not detect this new serotype in Japan and east Asia (1-3, 7, 17). Thus, the seroprevalence of this serotype in the general population is estimated to be low and outbreaks of adenoviral conjunctivitis caused by this serotype in Japan should be taken into consideration.

• Copyright © 1999 American Society for Microbiology