Nonisotopic Detection and Typing of Human Papillomavirus DNA in Genital Samples by the Line Blot Assay

Abstract

The line blot assay, a gene amplification method that combines PCR with nonisotopic detection of amplified DNA, was evaluated for its ability to detect human papillomavirus (HPV) DNA in genital specimens. Processed samples were amplified with biotin-labeled primers for HPV detection (primers MY09, MY11, and HMB01) and for β-globin detection (primers PC03 and PC04). Amplified DNA products were hybridized by a reverse blot method with oligonucleotide probe mixtures fixed on a strip that allowed the identification of 27 HPV genotypes. The line blot assay was compared to a standard consensus PCR test in which HPV amplicons were detected with radiolabeled probes in a dot blot assay. Two hundred fifty-five cervicovaginal lavage specimens and cervical scrapings were tested in parallel by both PCR tests. The line blot assay consistently detected 25 copies of HPV type 18 per run. The overall positivity for the DNA of HPV types detectable by both methods was 37.7% (96 of 255 samples) by the line blot assay, whereas it was 43.5% (111 of 255 samples) by the standard consensus PCR assay. The sensitivity and specificity of the line blot assay reached 84.7% (94 of 111 samples) and 98.6% (142 of 144 samples), respectively. The agreement for HPV typing between the two PCR assays reached 83.9% (214 of 255 samples). Of the 37 samples with discrepant results, 33 (89%) were resolved by avoiding coamplification of β-globin and modifying the amplification parameters. With these modifications, the line blot assay compared favorably to an assay that used radiolabeled probes. Its convenience allows the faster analysis of samples for large-scale epidemiological studies. Also, the increased probe spectrum in this single hybridization assay permits more complete type discrimination.

Human papillomavirus (HPV) is now considered a causative agent of carcinoma of the uterine cervix (36). Genital HPVs are classified into high-risk types that are associated with high-grade squamous intraepithelial lesions and cervical cancer and into low-risk types that are associated with low-grade squamous intraepithelial lesions (15, 36). HPV DNA is detected in more than 90% of patients with invasive cancer of the uterine cervix (4). Cohort studies have demonstrated that the presence of persistent infection with high-risk HPV types is predictive of cervical intraepithelial neoplasia and progression to higher grades of cervical disease (19, 25). Studies of the natural history of HPV infection, the determinants of persistent HPV infection, and the prospective impact of a diagnosis of HPV infection on cervical lesion screening and management are still needed to better define strategies aimed at preventing and treating precancerous and cancerous cervical lesions. To reach these objectives, a reliable HPV detection method that is sensitive and specific and that can be applied to large numbers of samples from cohort studies is required.

The presence of HPV in clinical specimens is established by nucleic acid hybridization tests. In order to increase the sensitivity of detection of HPV DNA, signal amplification and gene amplification methods have been developed (9). The PCR is the most sensitive method for the detection of HPV DNA sequences in clinical specimens (5, 16, 31). Since more than 30 HPV types infect the genital tract, the use of type-specific PCR assays is impractical for epidemiological studies (1, 12).

The MY09-MY11 consensus primer set targets conserved sequences in the L1 gene and can amplify a wide spectrum of genital HPV types (3, 4, 20, 24, 28, 30, 33). Amplification products are usually typed by filter-based assays with type-specific oligonucleotide probes linked to nonisotopic or isotopic labels (2). Nonisotopic detection of amplified products facilitates the use of this consensus test for large-scale testing (2, 10, 26). However, genotype determination still necessitates several hybridization reactions for typing, increasing the technical time for analysis of samples. Consequently, a one-step hybridization procedure would be desirable for the facilitation of HPV typing.

We report here on an evaluation of the line blot assay (17), a novel strip-based reverse hybridization test, for the detection of HPV DNA amplified with MY09-MY11. This nonisotopic assay was evaluated with clinical specimens obtained in two prospective studies and was compared to a standard consensus PCR test in which PCR-amplified products were detected with radiolabeled type-specific probes. Our aims were to determine the sensitivity and specificity of the line blot test for detection of the presence of HPV DNA in clinical samples as well as its reliability for the genotyping of the HPV isolates in HPV-positive samples.

MATERIALS AND METHODS

Cell lines and clinical specimens.

The cervical carcinoma cell line HeLa (which contains 40 copies of HPV type 18 [HPV-18] DNA per cell) was obtained from the American Type Culture Collection (Rockville, Md.) and was maintained in Eagle’s minimum essential medium supplemented with 10% fetal calf serum. Two hundred fifty-five genital specimens were collected from 255 women enrolled in two different cohort studies investigating the determinants of persistent HPV infection. One hundred sixty-two cervicovaginal lavage specimens were from The Canadian Women’s HIV study (8, 18). This study evaluates the relationship between genital HPV infection and cervical disease progression in relation to human immunodeficiency virus (HIV)-induced immune deficiency. The cervicovaginal lavage specimens were selected on the basis of initial results obtained with MY09-MY11 amplification reactions and detection with isotopic type-specific probes. This selection ensured the inclusion of all HPV types detected in the standard consensus PCR test and ensured the inclusion of specimens containing multiple HPV types. Ninety-three consecutive cervical brushings were from an ongoing study on the determinants of HPV persistence in young adult women attending McGill University (29). All samples were tested by the line blot assay and the standard consensus PCR test without knowledge of previous results or the patients’ clinical status. Consent was obtained from each participant. Both projects had the approval of the ethics committees of the institutions involved.

Processing of clinical samples.

Cervicovaginal lavage specimens were obtained with 10 ml of phosphate-buffered saline (pH 7.4) by standard procedures (35). The cells were then centrifuged in an IEC Centra-8R centrifuge at 2,500 rpm for 10 min at 4°C, resuspended in 500 μl of 10 mM Tris-HCl (pH 8.2), and stored frozen at −70°C until they were processed (7). The cell suspensions were thawed, lysed by the addition of Tween 20 and Nonidet P-40 (each at a final concentration of 0.4% [vol/vol]), and digested with 250 μg of proteinase K per ml for 2 h at 45°C. Cell lysates were boiled for 10 min and stored at −70°C until they were tested.

Exfoliated endo- and ectocervical cells from the uterine cervix were obtained with the Accelon combi cervical biosampler and resuspended in 2 ml of 10 mM Tris-HCl (pH 7.4) with 0.1 mM EDTA (TE) buffer. Two hundred microliters of the cell suspension was lysed with Tween 20 (final concentration, 0.8% [vol/vol]) and digested with 250 μg of proteinase K per ml at 45°C for 2 h. The lysates were purified with GlassMAX resin (Gibco-BRL, Burlington, Ontario, Canada) according to the recommendations of the manufacturer and were resuspended in 50 μl of TE buffer. The cell lysates were boiled for 10 min and were stored at −70°C until they were tested. Five microliters of processed sample was tested in each PCR assay. All samples had tested positive for β-globin with primers PC04 and GH20 (3, 8).

Standard consensus PCR test.

HPV DNA was amplified under standard conditions with the MY09, MY11, and HMB01 consensus HPV primers as described previously (8, 11, 20). Amplifications of HPV DNA and β-globin DNA were done in separate reactions. The amplification mixture contained 6.5 mM MgCl₂, 50 mM KCl, 2.5 U of Taq DNA polymerase (AmpliTaq; Roche Molecular Diagnostics, Mississauga, Ontario, Canada), 200 μM each dATP, dCTP, dGTP, and dTTP, and 50 pmol of each primer. Negative and weakly positive (25 HPV-18 DNA copies) controls were included to monitor contamination and the overall endpoint sensitivity of each PCR run. Amplifications were performed in a 480 Thermocycler (Perkin-Elmer Cetus, Montréal, Quebec, Canada) for 40 cycles with the following cycling parameters: 95°C for 1 min, 55°C for 1 min, and 72°C for 1 min. The amplified products were spotted onto nylon membranes and were reacted under stringent conditions with ³²P-labeled oligonucleotide probes for types 6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 53, 56, and 58 (2, 3, 20). The measures used to avoid false-positive reactions due to contamination have been described elsewhere (8).

Line blot assay.

Lysates were amplified by the consensus L1 PCR protocol described above, with the following modifications made to the amplification mixture (17): the use of 6 mM MgCl₂, 7.5 U of AmpliTaq Gold DNA polymerase, 600 μM dUTP, and 200 μM each dATP, dCTP, and dGTP, and biotin-labeled primers (MY09, MY11, HMB01, PC04, and GH20). A rapid amplification profile was used in a TC 9600 thermal cycler (Perkin-Elmer Cetus): activation of AmpliTaq Gold at 95°C for 9 min; denaturation at 95°C for 20 s, annealing at 55°C for 30 s, and extension at 72°C for 30 s for 40 cycles; and terminal extension at 72°C for 5 min (17).

The line blot assay was completed as described previously (17). Twenty-five microliters of AMPLICOR denaturation solution was added to 50 μl of the PCR-amplified products. Seventy microliters of denatured PCR products was added to each well of an AMPLICOR typing tray that contained 3 ml of hybridization solution (4× SSPE [1× SSPE is 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA [pH 7.7], 0.1% sodium dodecyl sulfate) that had been prewarmed to 53°C and a strip of HPV oligonucleotide and β-globin probes. The probe mixtures for the following 27 HPV genotypes were fixed on distinct lines on each strip: types 6, 11, 16, 18, 26, 31, 33, 35, 39, 40, 42, 45, 51, 52, 53, 54, 55, 56, 57, 58, 59, 66, 68, MM4, MM7, MM8, and MM9. The tray was incubated in a shaking water bath at 53°C for 30 min. The hybridization solution was aspirated from each well, and 3 ml of washing solution containing 1× SSPE and 0.1% sodium dodecyl sulfate was added at room temperature and was aspirated. Again, 3 ml of washing solution was added to each well and the plate was incubated at 53°C for 15 min. The washing buffer was aspirated, and 3 ml of AMPLICOR streptavidin-horseradish peroxidase conjugate was added. The tray was shaken gently for 30 min at room temperature. The conjugate was aspirated and 3 ml of washing buffer was added. The trays were shaken for 10 min on a platform shaker. This step was repeated once. After aspiration of the washing buffer, 3 ml of citrate buffer was added to each well and was aspirated. The substrate was prepared by mixing 0.01% H₂O₂ and 0.1% ProClin in a 0.1 M citrate solution with 0.1% tetramethylbenzidine in dimethylformamide. Three milliliters of substrate was added to each well. The trays were shaken at 70 rpm for 5 min at room temperature. The substrate was removed, the strips were rinsed with distilled water and stored in citrate buffer, and the results were read within 30 min.

When discordant results between the standard consensus PCR test and the line blot assay were encountered, the samples were retested by both assays. PCR products from the line blot assay were also spotted onto a nylon filter and were hybridized with type-specific radiolabeled oligonucleotide probes, as described above for the standard consensus PCR assay. Samples for which results remained discordant were then retested by the line blot assay without the addition of β-globin primers to the amplification reaction mixture. These lysates were also tested by the line blot assay by using the ultrasensitive amplification cycling profile (17) (activation of AmpliTaq Gold at 95°C for 9 min; denaturation at 95°C for 60 s, annealing at 55°C for 60 s, and extension at 72°C for 60 s for 40 cycles; and terminal extension at 72°C for 5 min).

Statistical method.

The crude percent agreement between both detection methods was the percentage of paired samples with identical results. The agreements for overall positivity (HPV DNA positive), for positivity for high-risk HPV types as a group (types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, and 58), and for positivity for each type were calculated. The unweighted kappa statistic was calculated to adjust for chance agreement between HPV detection methods (14). In general, a kappa value above 0.75 represents excellent agreement beyond chance. The sensitivity and specificity of the line blot assay were calculated by considering the standard consensus PCR test as the “gold standard.” The mean number of types detected per sample by each PCR test was compared by the Mann-Whitney rank sum test, since the distribution of types per sample was not normally distributed. Chi-square analysis was performed to compare proportions.

RESULTS

HPV DNA detection.

HPV DNA was detected in 47.5% (121 of 255) and 43.5% (111 of 255) of samples by using the line blot assay and the standard consensus PCR test, respectively. However, only 14 genotypes (types 6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 53, 56, and 58) were identified by both assays. The following analyses were restricted to these 14 types detectable by the standard consensus and the line blot assays. Specimens that tested positive by the line blot assay for an HPV type not included in the latter 14 types were considered negative in the comparisons.

First, the line blot and the standard consensus PCR assays were compared for their abilities to detect the presence of HPV DNA of the 14 detectable types enumerated above in 255 genital specimens. HPV DNA was detected in 96 (37.7%) and 111 (43.5%) of the 255 samples by the line blot assay and the standard consensus PCR test, respectively. Identical results were obtained by both tests for 236 (92.6%) of 255 samples (94 HPV-positive samples and 142 HPV-negative samples) for a very good agreement for the detection of the presence of HPV DNA in genital samples (kappa statistic of 0.82).

Of the 113 samples positive for HPV DNA by at least one PCR method, HPV DNA was detected by one method only in 19 samples. In two of these samples, infections with single HPV types (once each with HPV-39 and HPV-53) were detected only by the line blot assay. The testing of biotin-labeled amplified products from these two samples with isotopic probes confirmed the presence of these HPV types. Repeat testing of the two samples by the line blot assay also confirmed the presence of HPV DNA, but the standard consensus PCR test remained negative. In the other 17 samples, the presence of HPV DNA from 16 infections caused by single HPV types and from 1 infection caused by several HPV types was detected only by the standard consensus PCR test. Biotin-labeled PCR products from line blot assays for these samples were tested by a dot blot assay with radiolabeled probes and scored negative for HPV. This last experiment confirmed that these samples were falsely negative by the line blot assay because of differences in the amplification process between the two PCR assays and not because of a lack of sensitivity of the strip hybridization assay. When the standard consensus PCR was considered to be the gold standard, the sensitivity and the specificity of the line blot assay for detection of the presence of HPV DNA reached 84.7% (94 of 111) and 98.6% (142 of 144), respectively.

HPV DNA type identification.

The distributions of the 27 types detected by the line blot assay and the 14 types detected by the standard consensus PCR assay are presented in Table 1. When only the 14 HPV types detectable by both assays were considered, 46 samples (40.7% of the 113 HPV-positive samples) contained multiple HPV types by at least one PCR assay, while 32 contained multiple types by both assays.

TABLE 1.

Comparison of the line blot assay and a standard consensus PCR test for detection and typing of HPV DNA in 130 HPV-positive genital samples^a

HPV type	No. of samples positive by the following assay:
	Standard consensus PCR	Line blot	M-Line blot-1	M-Line blot-2
6	14	12	14	14
11	1	1	1	1
16	7	6	7	7
18	10	9	9	9
26	NTb	0	0	0
31	13	11	13	13
33	9	5	7	8
35	6	2	3	5
39	17	15c	16c	18c
40	NT	0	0	0
42	NT	0	0	0
45	7	7	7	7
51	13	10	11	13
52	16	11	13	15
53	19	18c	19c	20c
54	NT	13	14	14
55	NT	2	4	4
56	24	13c	19c	24c
57	NT	0	0	0
58	12	7	11	12
59	NT	3	3	3
66	NT	12	12	12
68	NT	5	5	5
MM4	NT	0	0	0
MM7	NT	16	16	16
MM8	NT	11	11	11
MM9	NT	3	3	3

^aLysates from 255 samples were amplified in parallel by the line blot assay and a standard consensus PCR assay. One hundred thirty samples contained HPV DNA by at least one method. For the column labeled M-Line blot-1, the 37 samples that had initially provided false-negative line blot assay results were retested by the line blot assay without coamplification of β-globin. For the column labeled M-Line blot-2, the 37 samples that had initially provided false-negative line blot assay results were retested by the line blot assay without coamplification of β-globin and with the ultrasensitive profile (see Results). For both of these columns, the results obtained for these 37 samples were combined with the results obtained for the HPV-positive samples for which concordant results between the line blot and the standard consensus PCR tests were obtained. 

^bNT, not tested. 

^cResults for samples that tested positive for one type by the line blot assay but negative by the standard consensus PCR test for that type are included. 

When the typing results are considered, there was a good agreement of 83.9% (214 of 255 samples) between the two PCR assays: 142 specimens tested negative by both tests and 72 scored positive for the same type(s) (kappa statistic, 0.67). However, 41 samples yielded discordant results for HPV typing: in 4 samples, the line blot assay detected types undetected by the standard consensus PCR test, and in 37 samples, the standard consensus PCR detected types undetected by the line blot assay. The line blot assay identified one sample that was positive for HPV-39 and one sample that was positive for HPV-53, but both of the samples were negative for these types by the standard consensus PCR assay. These samples were also found to be positive by testing biotinylated PCR products by a dot blot assay with radiolabeled probes, confirming the presence of these two types in the samples. HPV-56 was detected by the line blot assay but not by the standard consensus PCR test in two other samples that contained more than one HPV type. In 37 samples, the standard consensus PCR assay detected more HPV types than the line blot test. The types not detected in the latter samples included HPV-56 (13 samples), HPV-52 and -58 (5 samples each), HPV-33 and -35 (4 samples each), HPV-51 and -39 (3 samples each), HPV-6, -53, and -31 (2 samples each), and HPV-16 and -18 (1 sample each). In eight specimens, more than one HPV type was not detected by the line blot assay.

Of the 37 samples with discordant results, 20 specimens contained multiple types, while 17 contained only one type as determined by the standard consensus PCR assay. Samples for which discordant results were found between the line blot assay and the standard consensus PCR test had a greater mean number of types per sample than HPV-positive samples for which the results were concordant (2.0 ± 1.2 versus 1.5 ± 0.9; P = 0.008). A significantly greater number of samples in the group of HPV-positive samples with discordant results between the two PCR tests than in the group of HPV-positive samples with concordant results also contained multiple HPV types (22 of 41 samples versus 24 of 72 samples; P = 0.047).

When samples were classified as positive or negative for high-risk HPV types (HPV-16, -18, -31, -33, -35, -39, -45, -51, -52, -56, and -58), the rate of agreement between the two PCR tests rose to 93.3% (238 of 255 samples), for a kappa value of 0.84. Of the 99 samples classified as positive for high-risk HPV types by the standard consensus PCR test, 84 were positive by the line blot assay (sensitivity, 84.9%), while 154 of the 156 samples negative by the standard consensus PCR test tested negative by the line blot assay (specificity, 98.7%).

Since 255 samples were tested for the presence of 14 types, 3,570 results for HPV typing could be compared between these two PCR methods. By considering each type individually, 3,524 of 3,570 typing results (98.7%) were concordant between these two assays. When concordance was calculated after exclusion of samples that were HPV negative by both assays, the results for 1,564 (97.1%) of 1,610 positive typing results were identical by both assays.

HPV typing results by line blot assay by avoiding coamplification.

HPV and β-globin sequences were coamplified only by the line blot assay. Since in our experience coamplification reduces the level of sensitivity for HPV detection with consensus L1 primers (6), we avoided β-globin coamplification in the standard consensus PCR assay. When coamplification of β-globin and HPV DNA was done by the line blot assay, 10 copies of HPV-18 scored negative in two of five PCR runs (data not shown). The same positive control scored positive in five of five PCR runs when primers for β-globin were not added to the master mixture (data not shown). In 37 samples, the line blot assay detected fewer types than the standard consensus PCR test. Retesting of these samples by the line blot and the standard consensus PCR assays yielded the same results as those obtained in the first amplification run (data not shown). The latter 37 samples were again amplified by the line blot assay but the β-globin primers were not added to the amplification mixture.

First, the impact of avoiding coamplification was evaluated with the rapid amplification cycling profile (Table 1, M-Line blot-1). A set of 50 random samples from the pool of positive samples for which concordant results had been obtained was retested by the line blot assay without coamplification of β-globin: the results remained identical to those obtained in the first amplification run (data not shown). When only the presence or absence of HPV DNA is considered, 12 of the 17 specimens initially negative by the line blot assay but positive by the standard consensus PCR test were positive by the line blot assay, for a sensitivity of the line blot assay of 95.5% (106 of 111 samples). When typing results are considered, 21 of the 37 discrepancies (56.8%) were resolved. Of the 21 samples with resolved discrepancies, 9 contained more than one HPV type by the standard consensus PCR assay (the types resolved included type 56 for three women, type 6 for one woman, type 31 for one woman, type 33 for one woman, type 51 for one woman, type 52 for one woman, and type 58 for one woman) and 12 contained one type by the standard consensus PCR assay (type 56 for three women, type 58 for three women, type 6 for one woman, type 16 for one woman, type 33 for one woman, type 35 for one woman, type 52 for one woman, and type 53 for one woman). Results from two additional samples that contained multiple types were partly resolved. In the latter samples, types 31 and 39 were detected by the line blot assay without coamplification but still remained falsely negative for types 51 and 56. The results for 14 samples remained unresolved, including 9 samples with more than one type (types undetected by the modified line blot assay included HPV-56 for 3 samples and HPV-51, -33, -39, -35, -58, -52, and -53 for one sample each) and 5 samples with one type by the standard consensus PCR assay (the types undetected by the modified line blot assay included HPV-52 for 2 samples, HPV-35 for 2 samples, and HPV-18 for 1 sample). The mean number of types per sample containing multiple types that remained unresolved was similar to that for samples containing multiple types that were resolved (2.8 ± 1.4 types versus 2.6 ± 0.6 types; P = 0.715).

Next, the impact of avoiding coamplification was investigated by using the ultrasensitive amplification profile for these 37 samples (17). In this profile, each step of the 40 cycles of amplification is extended to 1 min instead of 20 or 30 s. Discrepancies were resolved for 33 (89.2%) samples and were partly resolved for 1 sample (Table 1). By combining these new results obtained by the line blot assay without β-globin primers with initial results obtained by the line blot assay for samples with concordant results, the sensitivity of the line blot assay without coamplification for the detection of HPV types reached 96.3% (105 of 109 samples; 95% confidence interval, 92.6 to 100.0%).

DISCUSSION

This report describes the evaluation of a practical nonisotopic method for the typing of HPV L1 DNA amplified by PCR from exfoliated cells obtained by cervicovaginal lavage or from cervical scraping. The method involves a reverse hybridization reaction of biotin-labeled amplified DNA with oligonucleotide probes fixed on a strip. We have compared this novel assay to a standardized test that served as a gold standard. Both assays used the same primer pair that has demonstrated good levels of sensitivity and reproducibility in other studies (27). Several differences are found between the line blot assay and the standard consensus PCR test. Coamplification of HPV and β-globin DNA was accomplished only by the line blot assay. The line blot assay uses a novel thermostable DNA polymerase that is activated by high temperatures and that does not require hot start. In the line blot assay, biotin-labeled PCR products are reacted to probes fixed onto lines on a strip. The standard consensus PCR test detected with radiolabeled probes amplicons fixed on nylon membranes. The line blot assay can be completed in a short time because amplified DNA is hybridized with all relevant probes in one reaction instead of several reactions.

There was a good agreement between these two assays. The agreement exceeded 90% for HPV DNA detection, detection of high-risk HPV types as a group, and typing of HPV when coamplification was not done. The line blot assay was not as sensitive as the standard consensus PCR assay, especially with samples with more than one HPV type. Samples with discordant results between the two PCR tests were more likely to be infected with multiple types. An important proportion of these discrepancies were found with HPV-56 or -58 infection and rarely affected frequently encountered high-risk types such as HPV-16. Avoidance of coamplification and the use of the ultrasensitive amplification profile (17) resolved most (89%) of the discrepancies between the two PCR tests. Although faster, the rapid amplification profile should not be used for the sensitive detection of HPV DNA. In the first evaluation of the line blot assay (17), the rapid amplification profile was also less sensitive that the ultrasensitive profile. Previous work with the line blot assay did not report a loss of sensitivity due to coamplification (17). This variability in the performance of the line blot assay could be in part related to the strategy of synthesis of biotin-labeled primers: batch synthesis of biotin-labeled primers was not controlled to ensure an equivalent representation and synthesis of all degenerate primers. Underrepresentation of some degenerate primers could lead to inefficient amplification for some types. The use of a rapid amplification profile could also explain the loss in sensitivity with coamplification. We did not evaluate coamplification of HPV and β-globin DNA using the ultrasensitive profile because limited amounts of sample remained. The line blot assay had an excellent specificity, considering that biotin-labeled PCR products from line blot assay-positive and standard consensus PCR-negative samples reacted with radiolabeled probes in a dot blot assay.

The performance of nonisotopic hybridization assays is often comparable to that of methods that use radiolabeled probes for detection of HPV DNA amplified products (2, 10, 21–23, 32). The use of nonisotopic probes in filter-based hybridization assays has been described with consensus PCR tests, but successive tests with individual probes for each type were still required (2, 27, 37). Manipulation of strips in the line blot assay is more convenient than the handling of filters. Some nonisotopic tests (10, 21, 23, 32) have the advantage of using 96-well microtiter plates which are easier to manipulate and which are more convenient than filter-based assays. However, most require the testing of samples with multiple individual probes and could not detect as many types as the number described here for the line blot assay with MY09-MY11 PCR products.

A reverse hybridization assay has also been described for HPV detection (34). Compared to the line blot assay, complete HPV genomic probes were fixed on membranes instead of type-specific oligonucleotide probes. Because a limited number of types were evaluated and because of the results of previous work demonstrating cross-hybridization when long probes are used to type MY09-MY11 amplicons (10, 13), cross-hybridization could still be encountered in that system.

A substantial number of samples contained HPV types that have been recently described, such as types 66, MM7, and MM8. This could result from the selection of HPV-positive samples. However, these results underscore the importance of keeping assays current as new types are discovered.

In conclusion, the line blot assay compared favorably to a standard isotopic method for detection of PCR products with clinical specimens. It was nearly as sensitive as the standard radioisotopic assay even with samples infected with multiple HPV types, especially if coamplification was avoided and if the ultrasensitive amplification profile was used. This convenient format allows one to test one sample for 27 types in one reaction.