INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion Appendix References

High-risk human papillomavirus (HPV) types are now considered etiological agents of cancer of the uterine cervix (5, 28). HPV infection is the most frequent sexually transmitted disease (STD) worldwide, and up to 60% of sexually active women will be infected by HPV in the genital tract (17). Irrespective of HPV infection status, fewer than 1 in 10,000 women will develop invasive cervical cancer. The fact that most HPV-infected women do not develop cytological anomalies or cancer (7) underlines the importance of factors modulating the progression of cervical disease to cancer in HPV-infected women. These factors may include the HPV genotype and molecular variant, the HPV viral load, persistence of HPV infection, coinfection with other STD agents, and the immune status of the host (7, 12, 26, 29, 30, 39).

Recent studies have demonstrated that infection with high-risk HPV types is usually transient (10, 16, 17, 34). Persistence of HPV infection substantially increases the risk of progression to clinically relevant preinvasive and invasive disease (2, 17, 18, 23, 24, 26). Persistence of HPV infection is often defined as the detection of the same HPV type in consecutive samples obtained at 3- to 6-month intervals. However, when the HPV genotype involved is prevalent in the population of women under study, reinfection with the same genotype may occur, simulating persistent infection. This detection of the same genotype may result in misclassification of consecutive transient HPV infections as persistent infections.

Genomic heterogeneity of HPV-16 isolates, the most prevalent type detected in high-grade and cancerous lesions of the uterine cervix (5, 17, 33), results from a limited number of nucleotide changes, especially in noncoding regions of the HPV genome (19). On this basis, HPV-16 can be classified into more than 40 variants (6, 20, 25, 41) which have a less than 2% variation in DNA sequence (33). Molecular variant analysis of HPV-16 isolates allows us to distinguish reinfection from persistence (11, 32). Sequence analysis of hypervariable regions of the HPV genome (11) remains the gold standard method for molecular variant analysis. However, single-strand conformation polymorphism (SSCP) analysis is an alternative to sequencing for this purpose (38). Only one group has compared the value of HPV-16 variant analysis of PCR-SSCP patterns with that of PCR-sequencing of clinical specimens (32). A simplified dot blot analysis to classify HPV-16 isolates into the major HPV-16 lineages has also been described (35).

The objective of our study was to evaluate the accuracy of PCR-SSCP analysis in classifying women who had consecutive genital samples containing HPV-16 DNA as being persistently infected or being reinfected by HPV-16. We also compared the numbers of variants identified by the SSCP and DNA-sequencing procedures. Using both methods, we investigated if sexually active women considered persistently infected with HPV-16 harbored the same variant over time.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion Appendix References

Patients' protocol. This study focused on 50 women participating in The Canadian Women's HIV Study Group (13, 14) who had at least two consecutive samples containing HPV-16 DNA sequences as determined by the L1 MY09-MY11 consensus PCR assay. Women were included from the date of the first HPV-16-positive sample to the date of loss of HPV-16, loss to follow-up, or September 1997, whichever came first. All participants provided written informed consent to participate. Ethics committees of each participating institution had approved The Canadian Women's HIV Study protocol.

Sample processing. Participants inserted and immediately withdrew a vaginal tampon before pelvic examination (8). The tampon was placed into a sterile jar containing 50 ml of transport medium composed of 10 mM Tris-HCl (pH 7.5), 50 mM EDTA, and 150 mM NaCl. Cervicovaginal lavages were obtained during pelvic examinations with 10 ml of phosphate-buffered saline (pH 7.4) (8). Specimens were refrigerated within 1 h and transported to a central laboratory on wet ice. Tampons were then squeezed to obtain a suspension of exfoliated cells. Cell suspensions from tampons and cervicovaginal lavages were pelleted after centrifugation at 2,500 rpm (IEC CENTRA-8R) for 10 min at 4°C and resuspended in 500 µl of 10 mM Tris (pH 8.3). Cell suspensions were lysed with Tween 20 at a final concentration of 0.8% (vol/vol), and digested with 250 µg of proteinase K per ml for 2 h at 45°C (8). Cell lysates were boiled for 10 min and stored at 70°C until tested. The delay between sampling and processing was never extended beyond 5 days.

-Globin and HPV DNA amplification. Five microliters from each sample was amplified for -globin DNA with the PC04 and GH20 primers to control for DNA integrity, for the presence of an adequate number of epithelial cells, and for amplification inhibitors (3, 8). -Globin-negative samples were extracted with phenol-chloroform and precipitated with ethanol (27). Five hundred nanograms of extracted nucleic acids was then reamplified for -globin. -Globin-positive samples (lysate or extracted DNA) were then tested for HPV. A volume of 5 µl of each cell lysate was amplified under standard conditions with the MY09, MY11, and HMB01 consensus HPV primers (8, 9, 13, 16). Negative, weakly positive (10 HPV-18 DNA copies), and strongly positive (HPV-6/11, -16, -31, -33, -35, -39, and -45) controls were included in each amplification run. Amplified products were spotted onto nylon membranes and were reacted as described elsewhere under stringent conditions with ³²P-labeled oligonucleotide probes for HPV-6, -11, -16, -18, -31, -33, -35, -39, -45, -51, -52, -53, -56, and -58 (3, 8, 16).

Amplification of HPV-16 LCR for molecular variant analyses. The complete long control region (LCR) of HPV-16, a hypervariable noncoding segment of the HPV-16 genome between nucleotide positions 7109 and 222, was amplified with two sets of primers (32, 37, 38): primers A and B amplified a 418-bp fragment (nucleotide positions 7109 to 7527) designated the AB fragment, while primers C and D generated a 682-bp fragment (nucleotide positions 7445 to 222) designated the CD fragment.

Molecular variant analysis by SSCP analysis. The SSCP analysis of HPV-16 LCR-amplified products was carried out as described originally by Xi et al. (37, 38). SSCP patterns of single-stranded HPV-16 PCR products from the five fragments of each HPV-16 isolate were compared with the reference patterns generated by pHPV-16. HPV-16 isolates could thus be classified into reference-like and non-reference-like variants. HPV-16 isolates were considered different variants if at least one fragment showed polymorphism. All specimens from a patient were run on the same gel. All representative isolates having SSCP patterns different from the patterns generated by pHPV-16 were compared together on one gel, which permitted comparisons of all SSCP patterns found in our population. Half of the samples were analyzed twice by SSCP analysis, including all samples from participants with isolates exhibiting a change of SSCP pattern over time. Two observers independently interpreted the gels, and both made the same interpretations of SSCP patterns, without knowledge of the results of DNA sequence analysis.

Molecular variant analysis by PCR-sequencing. Amplifications were performed as described above by replacing [-³²P]dATP by 0.25 mM dATP in the master mix. HPV-16 LCR fragments were generated with the Expand High Fidelity PCR system (Boehringer Mannheim, Laval, Quebec, Canada) which is a mixture of Taq DNA polymerase and Bwo DNA polymerase that has a low rate of misincorporation. When the intensities of the bands generated were too weak, AmpliTaq Gold DNA polymerase was used. PCR-amplified HPV-16 DNA fragments were purified by the QIAquick gel extraction kit protocol as recommended by the manufacturer (Qiagen Inc., Mississauga, Ontario, Canada).

Data analysis. HPV-16 LCR DNA sequences were compared using the BLAST sequence analysis software from Genetics Computer Group (1) to known HPV sequences contained in the GenBank database (Los Alamos National Laboratories). HPV-16 LCR DNA sequences were compared with the sequence from the HPV-16 prototype strain reported in the HPV compendium from the Los Alamos National Laboratory (25). Isolates with a DNA sequence different from that of the prototype strain were classified as non-prototype-like variants. Sequences from non-prototype-like variants were aligned and compared two by two with the BLAST sequence analysis program to further classify HPV-16 variants. Sequencing analysis was completed without knowledge of SSCP results.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion Appendix References

Consecutive detection of HPV-16 in clinical specimens. From September 1993 to September 1997, 807 women were recruited in The Canadian Women's HIV Study. Four hundred eighty-four participants were HIV seropositive, while 323 were HIV seronegative. Fifty (6.2%) of the 807 women had at least two consecutive specimens containing HPV-16 DNA and were selected for this study. Thirty-six (7.4%) of the 484 HIV-seropositive women and 14 (4.3%) of the 323 HIV-seronegative had persistent HPV-16 infection (P = 0.100). The evaluation of determinants of persistence of HPV infection is part of a global study that will be the subject of future publications. We here consider only the 50 women with persistent HPV-16 infection.

Analysis of persistent HPV-16 infection by the SSCP method. SSCP analysis was used to evaluate the genomic diversity of HPV-16 LCR DNA amplified from 144 clinical specimens. Overall, 141 of the 144 specimens (104 cervicovaginal lavages, 37 tampons) from 50 women with persistent HPV-16 infection gave a positive result for HPV-16 by SSCP analysis when the primer pair A-B or C-D was used. We obtained adequate signals for SSCP analysis for 141 (97.9%) of 144 samples using primers A and B and for 118 (81.9%) of 144 samples using primers C and D. Examples of SSCP patterns are shown in Fig. 1. The three samples for which an interpretable SSCP pattern could not be obtained with the AB fragment did not generate an SSCP pattern with the CD fragment and generated weak signals for HPV-16 DNA by the L1 consensus PCR assay (data not shown).

View larger version (40K): [in this window] [in a new window]   FIG. 1.   SSCP patterns of fragments AB and CD. Samples were amplified with primers located in the LCR and subjected to nondenaturing gel electrophoresis after digestion with restriction enzymes, as described in Materials and Methods. ds is for double-stranded DNA. A' is for digested PCR products without denaturation, while A is for denatured digested PCR products.

Comparison of results of SSCP analysis and DNA sequencing for analysis of HPV-16 variants. Sequencing of the AB fragment (PCR-sequencing) generated by amplification of the HPV-16 LCR served as the gold standard test to evaluate the ability of PCR-SSCP analysis to identify HPV-16 variants. Mutations defining each variant are displayed in Fig. 2 and compared with results obtained by SSCP analysis. The nucleotide positions that were frequently involved in mutations identifying variants included nucleotide position 7192 (change of a G to a T), found in 16 variants, and nucleotide positions 7232 (change of an A to a C) and 7488 (change of a G to an A), found in 4 variants each (Fig. 2). Overall, PCR-sequencing identified 22 different variants in comparison with only 15 variants identified by SSCP analysis. SSCP analysis of the AB fragment defined as the prototype strain six HPV-16 strains identified by DNA sequence analysis as variants. Other isolates generated SSCP pattern 12 but had a different DNA sequence (sequences V, XXI, and XXII). Variants 10 and 15 generated different SSCP patterns of the CD fragment but had identical AB fragments by SSCP analysis and PCR-sequencing (Fig. 2). Ten of the 28 women for whom SSCP analysis identified the presence of a reference-like variant were infected by nonprototypic variants as determined by DNA sequencing (data not shown).

View larger version (17K): [in this window] [in a new window]   FIG. 2.   DNA sequence analysis from variants of HPV-16. The sequence variant was determined arbitrarily from the combinations of mutations identified in the amplified AB fragment. The reference sequence was that of the prototype strain reported in the HPV compendium as described in Materials and Methods. The SSCP pattern was determined according to the patterns of bands obtained by SSCP analysis of the AB and CD fragments. Two SSCP patterns were described for sequence variant III on the basis of differences in the migrations of the CD fragment between these two variants, but only the AB fragment was sequenced, illustrating the advantage of testing all the LCR instead of only one section of the LCR.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion Appendix References

We demonstrate here the usefulness of molecular variant analysis in identifying women with persistent HPV-16 infection. This is the largest published study to date refining the definition of HPV persistence in HIV-infected women using DNA sequencing, and it is also the largest evaluation comparing in parallel SSCP analysis to sequencing for HPV-16 variant analysis of women at risk for STD. Variant analysis has been used to investigate the epidemiology of HPV-16, the phylogenetic analysis of evolution of HPV-16, and determinants of cervical disease progression (4, 6, 11, 19-21, 32, 38-41). HPV-16 nonprototype variants may confer a higher risk than that of the prototype variant for the development of high-grade cervical or anal lesions (36, 39).

Targeting hypervariable regions of the HPV genome increases the probability of finding molecular variants (38). SSCP analysis of the HPV-16 LCR was first described by Xi et al. (38). Using their primers, only 2.1% of HPV-16-positive samples could not be amplified. In comparison to the number of variants identified by PCR-sequencing, the primer pair C-D did not perform as well as the primer pair A-B in terms of the number of variants identified (14 of 22 [63.6%] versus 6 of 22 [27.3%]; P = 0.034, z statistic test) and the number of isolates that generated visible bands (141 of 144 [97.9%] versus 118 of 144 [87.5%]; P = 0.001, McNemar's test). Of the nine women showing a change of SSCP pattern over time or the addition of weak signals suggestive of a minor variant, seven changes were revealed with the AB fragment and two were revealed with the CD fragment only. Another group has also confirmed the superiority of primer pair A-B for SSCP and sequence analyses (32).

The most frequent variant identified by SSCP analysis in our group of persistently infected women was the reference-like strain. Variant analysis is limited by the extent to which HPV variants tend to occur in tight geographical clusters. If a variant occurs frequently in a population, then it becomes an insensitive marker of persistence because of the high probability of reinfection by the same variant. Fortunately, 14 non-reference-like variants identified by PCR-sequencing accounted for nearly 50% of persistent HPV-16 infections in our population. The ethnic diversity of women recruited and the fact that participants in The Canadian Women's HIV Study were sexually active and continuously exposed to STD agents, as well as the national basis for recruitment, reduced the possibility of finding a limited number of variants. PCR-sequencing permitted us to identify 22 variants from 141 HPV-16 isolates detected in 50 sexually active women. A greater number of HPV-16 variants can be found in populations at risk for STD (32).

Although others found that PCR-sequencing and PCR-SSCP analysis gave comparable results for HPV-16 variant analysis, these comparisons were done only on selected specimens exhibiting different SSCP patterns (37-39). Since SSCP analysis and sequencing were conducted only on a subset of samples, these experiments were restricted mainly to the evaluation of the specificity of SSCP analysis by demonstrating that changes in SSCP patterns corresponded to a base substitution(s). Besides our work, only one study compared in parallel sequencing to SSCP analysis of HPV-16 PCR products (32). Those researchers also reached the conclusion that the resolution of the SSCP method for variant analysis was less than that of PCR-sequencing.

We chose to sequence fragment AB, since it exhibits more variability than fragment CD (32) and since nearly all HPV-16 isolates could be amplified with primer pair A-B. In our population of women persistently infected with HPV-16, differences in SSCP patterns were reproducible and could be explained by base substitutions. Although our results demonstrate that SSCP analysis is an adequate alternative to sequencing in identifying HPV-16 variants in consecutive specimens, we found, as reported by others (20, 31, 32, 38), that a more complete analysis of the variable region can be obtained by sequence analysis of PCR-amplified products. We found seven variants defined by sequencing that did not generate different SSCP patterns in our cohort.

In contrast to others (32, 37, 38) who found no change of predominant variant in 344 consecutive specimens from 70 patients, we observed a change of HPV-16 variant for 6% of women with consecutive specimens containing HPV-16 DNA. Three of these four women would have been misclassified as having persistent infection without molecular variant analysis. These differences with our study could be explained by several factors. First, 70% of our study population was HIV seropositive. Second, women in our cohort were exposed to HPV, as demonstrated by the high rate of prevalence of HPV DNA detection (13) and the high risk for STD in participating women, as presented elsewhere (13, 14). Finally, the initial study that used SSCP analysis to study persistence (37) may have underestimated the frequency of reinfection since two variants accounted for up to 70% of all HPV-16 variants identified. It is of interest that another team in Brazil has identified, as we have, a change of variant in consecutive specimens (11). Participant WQ 505 illustrates the complexity of variant analysis. A new variant replaced the first variant but was then replaced in the last sample by the first variant. Since we did not sequence more than five clones of HPV-16 PCR products per sample, we could not exclude the possibility that the first variant was still present in the second specimen at a low copy number. This woman may thus be persistently infected by one variant as well as having been transiently infected by a new variant.

Establishment of infection with multiple variants has been demonstrated (6, 15, 19, 21, 22, 37, 38). We demonstrate here that even in a high-risk population, detection of a new variant is a rare event but does occur. SSCP patterns for five women suggested the presence of more than one variant. Since we did not systematically sequence multiple clones for each sample, we could rely only on PCR-SSCP analysis to estimate the prevalence (10%) of coinfection with more than one variant.

Our experiments confirmed that PCR-SSCP analysis is an alternative in identifying women persistently infected by HPV-16. However, SSCP analysis has a lower resolution in identifying molecular variants than sequencing. The additional information on the nature of persistent infection provided by molecular variant analysis was useful for 6% of women who would have been misclassified as persistently infected by HPV-16. Molecular variant analysis may become more interesting if specific variants are associated with increased risk of disease progression or persistence of HPV infection.