Advanced Environmental Surveillance and Molecular Analyses Indicate Separate Importations Rather than Endemic Circulation of Wild Type 1 Poliovirus in Gaza District in 2002

Environmental sampling and processing.

Israel initiated environmental surveillance to provide early warning of nonvaccine circulation before appearance of poliomyelitis (12). From 1989 to 1999, sewage was analyzed monthly from 20 to 36 sentinel communities in Israel and the Palestinian Authority (the West Bank and the Gaza district). After 1999, the number of sampling sites was reduced to 15 covering mainly high-risk sites. As expected, most (>98%) of the polioviruses isolated from sewage samples between 1989 and 2005 were identified as vaccine strains. Exceptions included 17 related, highly diverged VDPVs isolated between 1998 and 2006 in Israel (18-20) and wild polioviruses isolated in 1990, 1994, 1995, 1996, 1999, and 2002 almost exclusively in the Gaza district (13).

Wherever possible, untreated sewage aliquots were collected every 30 to 60 min for 24 h by in-line or external automatic composite samplers. Elsewhere, including the Gaza district, aliquots were collected manually every half hour during 3 h of peak usage in the morning. A cold temperature (approximately 4°C) was maintained during sample transport to the laboratory, storage (<1 week), and processing. The pH of a 1-liter aliquot from the approximately 1.5-liter pooled total volume was adjusted to 7.2. Stirring continued for 30 to 60 min after the addition of 80 g of polyethylene glycol 6000 (Sigma or Merck) and 17.5 g of NaCl. Samples were centrifuged (10,000 × g, 1 h, 4°C) after overnight incubation (4°C). The pellet was vigorously agitated in 15 ml of Dulbecco's calcium- and magnesium-free phosphate-buffered saline containing 0.1% Tween 80 and 15 ml of chloroform. The virus suspension was clarified by centrifugation (1,400 × g, 15 min, 4°C). Virus trapped in the pellet was recovered by centrifugation (1,400 × g, 15 min, 4°C) after vigorously shaking it in 5 ml of 3% beef extract (pH = 7.2). Both supernatants were pooled, antibiotics were added (for each 50 to 60 ml of supernatant, 0.5 ml of penicillin stock [50 mg/ml], 0.5 ml of mycostatin stock [6,250 U/ml], and 0.8 ml of PSMy stock [penicillin at 50 mg/ml, streptomycin at 50,000 U/ml, and mycostatin at 6.50 U/ml] were added) and samples were frozen at −20°C until assay. There were 2.7-fold ± 1.5-fold fewer plaques/liter recovered by this polyethylene glycol method than by the previously used Freon-based method (12) in seven samples where poliovirus was recovered by both methods and no plaques in four samples where there were two or fewer plaques recovered by the Freon method. A mean of 12% (15%, 12%, and 8.8%, respectively) was recovered from three samples spiked with Sabin 3 (100 50% tissue culture infective doses) or Sabin 1 (2 × 10⁴ 50% tissue culture infective doses).

Virus isolation.

Viruses in processed sewage were isolated under plaque-forming conditions (1-h challenge, followed by a 2% agarose overlay, 48 h of incubation at 37°C [12]). Plaque purification of environmental isolates allowed the analysis of all of the polioviruses recovered. This contrasts with the variable results obtained by methodologies using mass culture (6, 9) for replicate aliquots from individual sewage concentrates within and between laboratories (9), which can occur when different isolates dominate and outreplicate the remaining isolates.

We have evaluated different cell lines and growth temperatures for the lowest input/output ratio as follows. Polioviruses were isolated from processed sewage on monolayers of murine L20B cells expressing the human poliovirus receptor (15, 25) and on BGM cells (4). The total number of plaques that required analytic processing was 4.2-fold lower with L20B selection then with growth on BGM cells (12), followed by poliovirus identification (Table 1). OPV does not grow well at elevated temperatures (14), and we have used this feature to select for nonvaccine polioviruses as previously reported (12). Temperature selection, i.e., restricting analysis to plaques that produced a ≥50% cytopathic effect within 5 days at 40°C in 2-ml tube cultures of HEp2C cells, was systematically revalidated in 2002. Aliquots of poliovirus from each plaque were grown in 2-ml HEp2C (ATTC CCL23) tube cultures at 40°C for 5 days and at 37°C until a complete cytopathic effect was achieved (≤3 days). Uniformly high titers of isolates that grew at 40°C were prepared for further characterization by passage on 2-ml HEp2C tube cultures (37°C, 24 h).

Poliovirus plaques isolated from 51 (49.5%) of 103 consecutive sewage samples were characterized (Tables 2 and 3). Two were wild type. All of the remaining isolates were vaccine, and 27.5% of them grew at 40°C. Most (92%) of the vaccine isolates that grew at 40°C were of the same serotype as virus from the same plaque cultured at 37°C. Occasionally, two or more virus particles remained physically associated during passage through the sewer system, subsequent treatment, and challenge. Mixed-serotype microclumping was confirmed by selective outgrowth of each serotype after neutralization of the other type(s). The microclumping frequency was underestimated since only microclumps containing roughly equal amounts of different serotypes would have been detected. Temperature selection reduced the number of mixed-serotype plaques because of the low probability that more than one isolate was capable of growth at 40°C. Microclumping was further reduced by vigorous agitation during processing; however, this increased the number of isolates with identical or nearly identical substitutions.

At 37°C, all three poliovirus serotypes were recovered in nearly equal numbers (38% type 1, 30% type 2, and 31% type 3). In contrast, type 2 isolates predominated (62%) among isolates recovered after growth at 40°C, as found previously (18).

Quality for sewage collection and processing was assured by using quantitative plaque isolation of enterovirus from processed sewage on either BGM or HEp2C cell monolayers.

Typic differentiation (TD) and intratypic differentiation (ITD) of poliovirus isolates.

Poliovirus serotypes (TD) were determined by microneutralization (2) or immunofluorescence by using type-specific monoclonal antibodies (catalog no. 3331, 3332, and 3335; Chemicon International, Inc., Temecula, CA). Isolates were characterized as Sabin-like or non-Sabin-like (ITD) by microneutralization with monoclonal and/or cross-adsorbed polyclonal anti-Sabin vaccine antisera (12).

TD and ITD were also inferred by dot blot hybridization to Sabin vaccine-specific RNA probes (5) and/or by sequence analysis (17). For sequence analysis, viral RNA was amplified (Ready-to-Go reverse transcription-PCR beads; Amersham-Pharmacia Biotech Ltd., Buckinghamshire, England), purified (QIAquick gel extraction kits [QIAGEN GmbH, Hilden, Germany] or HighPure PCR product purification kits [Roche Diagnostics, Indianapolis, IN]) with Y7 sense and Q8b antisense primers and sequenced automatically with the Q8b antisense primer (17). Molecular TD was provided with the Sequencher sequence analysis program (Gencodes Corporation, Ann Arbor, MI). When parameters were set for ≥85% homology, the antisense strand VP1/2A isolate sequence (≥500 nucleotides) only formed a contig with the actual sequences (both strands) of its corresponding Sabin serotype. A contig is a continuous sequence of bases derived from a set of aligned sequence fragments. Furthermore, the percent nucleotide sequence difference between the query and the Sabin reference within this contig provided ITD (e.g., ≤1% for vaccine, >1% for VDPVs). Both strands were sequenced whenever the difference exceeded 0.5%. Isolate sequences failed to form contigs when they were of poor quality, represented poliovirus mixtures, were from non-polioviruses, or were from wild poliovirus (divergence, >15%). Inspection of the graphic output distinguished among the possibilities. TD of wild viruses was obtained by reiteratively decreasing the percent homology until a contig was formed.

Wild polioviruses PV1/5763-1T/PAL02 and PV1/5816-2T/PAL02 (accession numbers AM056055 and AM056056, respectively) were isolated from sewage samples 5763 and 5816 collected from the Gaza district on 16 February and 4 August 2002 (Tables 2 and 3). They were revealed to be type 1 by iteratively increasing the allowed maximum divergence in Sequencher by 2% steps until a contig formed with Sabin 1. Both strands were sequenced. These isolates grew at 40°C, were non-OPV by dot blot hybridization, and were not neutralized by anti-Sabin monoclonal or polyclonal anti-poliovaccine antisera.

Epidemiological characterization of the two wild-type viruses.

Empirical observations (Table 4, center column) of three parameters, the maximum rate of substitutions, the distribution of substitutions along the poliovirus genome and within codons, and the transition/transversion ratio for isolates from a 1-year (1987 to 1988) poliovirus outbreak in Israel (17), were used to develop a profile for outbreaks evolving from a single polioviral progenitor (POESP). The POESP (Table 4) was formulated from an outbreak of short duration so that it would most closely represent evolutionary changes expected to occur during a short interval between two or more environmental and/or clinical isolations. Similar values (not shown) were extracted from an outbreak of longer duration (9). Of the three parameters, the rate of substitution carries the most weight for determining whether isolates were directly related through silent endemic circulation or introduced separately from one or more external reservoirs. The confidence for ruling out a direct chain of transmission increases the higher the substitution rate is above POESP expectations, especially when the interval between isolations is short. Thus, it is very unlikely that PV1/5763-1T/PAL02 and PV1/5816-2T/PAL02, collected 6 months apart in the Gaza district, were directly related by endemic circulation since the substitution rate was 17-fold higher than expected (Table 4). Poliovirus genomic RNA frequently recombines with RNA from other polioviruses or non-poliovirus enteroviruses (1, 8). Identical recombination patterns, like those described for Israeli VDPVs (19), are strong indications of a direct relationship to a common ancestor (genomic recombination, Table 4).

Relative similarity (RS), the percent nucleotide similarity between local isolates divided by the percent similarity between either isolate and the most similar nonlocal isolate, provides another guideline for distinguishing separate introductions from endemic circulation. Sequences from databases of the Global Specialized Poliovirus Reference Laboratories at the CDC and the Pasteur Institute and from the WHO-supported (TSA/18/181/526) project on environmental surveillance in Egypt at the KTL laboratory were screened for similarity to the Gaza district sequences. Egyptian environmental isolates from Minya (accession no. AF53118 and AY923837) shared the highest homology with PV1/5763-1T/PAL02, while isolates from Asyut (accession no. AF545125 and AF545126) and Abu Qurqas (accession no. AF545121) shared the highest homology with PV1/5816-2T/PAL02. An RS of <1 for both 2002 Gaza district isolates compared with these Egyptian isolates (Table 4) indicated that both local isolates were more closely related to isolates from external reservoirs than they were related to each other. This strengthened confidence for separate introductions as indicated by POESP, as did a consensus neighbor-joining phylogenetic tree constructed with the ClustalX program (22) with sequence data bootstrapped 1,000 times (Fig. 1).

• Open in new tab
• Download powerpoint

FIG. 1.

Phylogenetic relationship between wild poliovirus environmental isolates from Gaza and southern governates in Egypt. Consensus neighbor-joining phylogenetic trees were constructed by using the ClustalX program (22) with sequence data bootstrapped 1,000 times. Trees were visualized by using njplot (M. Gouy, Laboratoire de Biometrie et Biologie Evolutive, Université Lyon, CNRS, Lyon, France [http://pbil.univ-lyon1.fr/software/njplot.html ]).

In conclusion, we have documented the effectiveness of our improved sewage surveillance algorithm (isolation, treatment, and analysis) for identifying viral circulation before the appearance of clinical cases and have provided a paradigm for identifying the location of remote reservoirs from which the local isolates were most likely imported. The sewage treatment and viral isolation modifications described here significantly reduced laboratory costs and hands-on time. The surveillance algorithm worked well for sewage with a high OPV background in the environment when the Israeli and Palestinian immunization programs included OPV (21, 23) and in the absence of OPV when Israel switched to the exclusive use of inactivated poliovaccine (19). Finally, our results illustrate the significant international contribution this algorithmic approach can provide to the Global Polio Eradication Initiative. Specifically, a poor fit with the POESP and a strong RS linkage between local isolates and recent external isolates should trigger surveillance and possibly active intervention at the remote reservoir, where the virus lineage may also still circulate silently.

Nucleotide sequence accession numbers.

The accession numbers for wild type 1 environmental isolate sequences PV1/5763-1T/PAL02 and PV1/5816-2T/PAL02 are AM056055 and AM056056, respectively. The Egyptian environmental isolates (KTL database) are derived from a WHO-supported (TSA/18/181/526) project on environmental surveillance in Egypt and have accession numbers AF545121, AF545125, AF545126, AY531188, and AY923837.