- Email: [email protected]
Evaluation of HPV Genotyping Assays for Archival Clinical Samples
The Journal of Molecular Diagnostics, Vol. 17, No. 3, May 2015
jmd.amjpathol.org
Evaluation of HPV Genotyping Assays for Archival Clinical Samples Gabriella Lillsunde Larsson,* Jessica Carlsson,y Mats G. Karlsson,* and Gisela Helenius* From the Departments of Laboratory Medicine* and Urology,y Faculty of Medicine and Health, Örebro University, Örebro, Sweden Accepted for publication December 19, 2014. Address correspondence to Gabriella Lillsunde Larsson, Ph.D., Department of Laboratory Medicine, Örebro University Hospital, S-701 85 Örebro, Sweden. E-mail: gabriella.lillsunde-larsson@ regionorebrolan.se.
Human papillomavirus (HPV) testing and genotyping of FFPE tissue samples is important in epidemiological investigations. Here, we compare four different HPV genotyping methods for use in FFPE clinical samples. Comparative testing was performed on 99 samples with a clinical suspicion of HPV. Specimens were analyzed with Anyplex II HPV28 detecting 28 genotypes using real-time PCR and melting curve analysis, CLART HPV2 detecting 35 genotypes using PCR and microarray detection, and MGP5þ/6þ consensus primer system together with pyrosequencing. Results were compared to a realtime PCR reference protocol detecting 14 genotypes. In total, 68% of the samples were positive for an HPV genotype using the reference protocol and MGP5þ/6þ primer system. Anyplex II HPV28 analysis and CLART HPV2 had 82% and 72% positive samples, respectively. All four methods showed good agreement when comparing the 14 genotypes included in the reference protocol. When evaluating all genotypes, the Anyplex II HPV28 assay and the CLART assay changed the status of the sample (individually or together) from negative with respect to the reference protocol to positive for either a Group 1 (n Z 4) or Group 2 (n Z 6) genotype. We conclude from this study that for an extended genotyping approach with a high sensitivity for FFPE specimens, both the Anyplex II HPV28 and CLART HPV2 assays are suitable alternatives despite minor intra-assay differences. (J Mol Diagn 2015, 17: 293e301; http://dx.doi.org/10.1016/j.jmoldx.2014.12.004)
Human papillomavirus (HPV) genotype information has become relevant in several situations. The most established setting is the possibility to separate lesions with low-risk genotypes from the lesions with high-risk genotypes in cervical dysplasia. However, further information concerning specific genotypes is, or will soon be, needed in the clinical setting, such as review of HPV infections in the vaccinated population.1 Even though cytological samples are available in many situations, formalin-fixed, paraffin-embedded (FFPE) tissue samples may be the only source for HPV analysis of the lesion on several occasions, for example to establish HPV status and genotype in genital and head-and neck tumors for prognosis2e5 and personalized treatment strategies.6 In the HPV alpha genus of high-risk types, HPV-16 and -18 are the most emergent ones that together account for about 70% of all invasive cervical cancers.7 Ten additional genotypes are included in this group carcinogenic to humans (Group 1) according to the International Agency for Research on Cancer. Additional alpha genotypes from the same clade Copyright ª 2015 American Society for Investigative Pathology and the Association for Molecular Pathology. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jmoldx.2014.12.004
are classified as Group 2 genotypes, ie, 2A, probably carcinogenic, and Group 2B, possibly carcinogenic8 (Table 1). Also, in a recent meta-analysis on cervical samples, Bzhalava et al10 presented data that suggest the addition of HPV-26, -67, -68, -69, -73, and -82 to the Group 1 classification because these genotypes were more present in invasive cervical cancer samples compared to the normal cytology. Anogenital tumors, alongside invasive cervical cancers, are also to a varying extent positive for an HPV genotype. HPV16 is frequent, but genotypes such as HPV-6, -11, -22, and -40 may also occur.7 Despite the smaller numbers of cases, these findings argue for a broader genotyping approach. Supported by the research committee of Örebro University Hospital at Örebro County Council grant OLL-405821. Portions of this work have been presented in the doctoral dissertation of G.L.L. Disclosures: G.L.L. received travel support from Seegene for participating in the 24th European Congress of Clinical Microbiology and Infectious Diseases held May 10e13, 2014, in Barcelona, Spain.
Lillsunde Larsson et al
The market for HPV testing is fast growing, and in a review, Poljak et al1 reported 54 different full genotyping tests (including all Group A genotypes), with different methodology approaches such as strip- or filter-based, microarray tests, microsphere bead tests, gel electrophoreses tests, capillary electrophoreses tests, and real-time PCR tests. Most tests, however, are only validated on
cytology samples. Tissue samples used for clinical diagnosis are mostly archived by preserving the tissue in formalin followed by paraffin embedding. Formalin fixation is excellent for preserving the morphology of the tissue, but it also creates crosslinking and degradation of the nucleic acids.11,12 Modern molecular pathology, however, has overcome those problems mainly by improved DNA extraction protocols, using shorter PCR templates and adjusted PCR protocols.13 We have routinely used a previously reported real-time PCR as the genotyping method, with primers and probes specific for the viral E6 and E7 genes.14 The protocol includes the Group 1 genotypes together with the low-risk types 6 and 11. This method shows good clinical results for both cytological and FFPE specimens. However, eight separate reactions are needed for each sample, and none of the Group 2 genotypes are included. Based on the perspective to include more genotypes, rationalize the workflow, as well as to evaluate assays using different detection technologies, a comparison of assays for FFPE samples was performed. In the Anyplex II HPV28 assay (Seegene, Seoul, Korea), detection of 28 different alpha genotypes is performed using real-time PCR followed by melting-curve analysis at 30, 40, and 50 cycles. Using TOCE technology (Seegene), variation in melting temperature values during melting curve analysis is avoided, allowing for a very sensitive multiplexed detection approach (Seegene technical manual: Anyplex II HPV28 detection). This assay has not previously been evaluated on FFPE specimens but is approved for in vitro diagnostic use in the European Union for cytology specimens. The CLART system from Genomica (Madrid, Spain) is widely used in clinical laboratories and has been approved for in vitro diagnostic use in the European Union for use in FFPE samples. It distinguishes between 35 different genotypes by a low-density array. PCR products that are tagged with biotin hybridize to the array, conjugate is added, and the product is then visualized by precipitation on spot (Genomica CLART technical manual: genotyping of human papillomavirus via genomic identification for in vitro diagnosis). Finally, a consensus approach was included in this comparative study, using the MGP5þ/6þ primers.9 The MGP5þ/6þ primers were developed from the original GP5þ/6þ primers,15 and the improved assay included five primer pairs, with extended primer length, to increase specificity for the Group 1 genotypes together with HPV-66 and -68. In a clinical evaluation by Söderlund-Strand et al,9 additional Group 2 genotypes and low-risk genotypes could also be detected. Our aim was to evaluate four different genotyping methods for use in FFPE clinical samples with a suspicion of HPV infection concerning HPV detection and specific genotype distribution. The two commercially available HPV genotyping tests, the Anyplex II HPV28 from Seegene and the CLART assay from Genomica, as well as the consensus MGP5þ/6þ primer system, were evaluated and compared to a real-time PCR reference protocol.
294
jmd.amjpathol.org
Table 1 Overview of Detectable Genotypes in the Different Assays in the Study MGP5þ/6þ IARC Reference Anyplex CLART primer Genotype classification protocol II HPV 28 HPV2 system 16 18 31 33 35 39 45 51 52 56 58 59 68 26 30 34/64 53 66 67 69 70 73 82 85 6 11 40 42 43 44 54 61 62 71 72 81 83 84 89 No.
1 1 1 1 1 1 1 1 1 1 1 1 2A 2B 2B 2B 2B 2B 2B 2B 2B 2B 2B 2B 3 3
x x x x x x x x x x x x
x x x x x x x x x x x x x x
x x x x x x x x x x x x x x
x x
x x
x x x x x x
14
x x x x x x x x
28
x x x x x x x x x x x x x x x x x x x 35
x x x x x x x x x x x x x
x
x x x x x x x
21
Genotypes detected by MGP5þ/6þ system as described by SöderlundStrand et al.9 Classification according to International Agency for Research on Cancer (IARC); 1: carcinogenic to humans, 2A: probably carcinogenic, 2B: possibly carcinogenic, 3: not classifiable as to its carcinogenity to humans. Additional low-risk genotypes are included in the Anyplex II HPV28 and CLART HPV2 assays.
-
The Journal of Molecular Diagnostics
HPV Genotyping in FFPE Tissue
Materials and Methods Samples In this study, 99 FFPE samples from patients with a clinical suspicion of HPV infection at the Department of Pathology, Örebro University Hospital, were included. The majority of the samples were cervical lesions (61%; 60 of 99) or penile carcinomas (20%; 20 of 99), and the remaining samples were a mix of samples from vulva (5%; 5 of 99), vagina (1%; 1 of 99), anus (2%; 2 of 99), head and neck (3%; 3 of 99), larynx (2%; 2 of 99), skin (4%; 4 of 99), bladder (1%; 1 of 99), and lymph node (1%; 1 of 99). All samples were anonymized before analysis.
analyzed by 7900 Fast System SDS version 2.4 software for ABI 7900 HT (Thermo Fisher Scientific). Each curve was manually surveyed, and samples with a cycle threshold <35 were considered as positives.
Genotyping with Anyplex II HPV28
DNA was extracted from FFPE samples. Tissue blocks were selected by a pathologist, and the tissue was sectioned in four 5- to 20-mm sections. Sections were placed in a clean microcentrifuge tube, deparaffinized, and incubated overnight (56 C) with proteinase K (Qiagen, Hilden, Germany). DNA was extracted using QiaAmp DNA Mini Kit (Qiagen) according to the manufacturer’s instructions. Samples were eluted in a volume of 75 mL. Quality and quantity of DNA was spectrometrically estimated (NanoDrop Technologies, Wilmington, DE). For each test, eluates were diluted to a concentration of 10 ng/mL. In every reaction and assay, 5 mL (50 ng) eluate was used. For each batch of samples, a blank water sample was included in the extraction to monitor for cross contamination. The blank sample was analyzed with the reference real-time PCR assay. For an overview of included genotypes in the different assays, see Table 1.
Anyplex II HPV28 detects 28 genotypes and the human gene beta-globin HBB in two multiplex reactions. Anyplex II HPV28 has previously been validated on liquid-based cytology specimens and is approved for in vitro diagnostic use in the European Union, though it has not been validated on archival FFPE tissue. Amplicon lengths are between 100 and 200 bp, and designed for the consensus gene L1. Reactions were run on a CFX96TM Real-Time PCR System (Bio-Rad Laboratories, Hercules, CA) in 20-mL reactions containing 1 HPV28 primer mix A or B, 1 Anyplex Master Mix, and approximately 50 ng of DNA. After an initial step at 50 C for 4 minutes and a denaturation step at 95 C for 15 minutes, reaction mixtures underwent 30 cycles at 95 C for 30 seconds followed by 60 C for 60 seconds and 72 C for 30 seconds. This was followed by a 55 C step for 30 seconds and a melting curve analysis at 55 to 85 C (5s/0.5 C). Reactions continued cycling at the same temperatures for 10 cycles followed by melting curve analysis, and this was repeated once more for another 10 cycles. Melting curve data were obtained at 30, 40, and 50 cycles. For each run, three positive controls, ie, plasmid constructs of all HPV types, were included as well as a nontemplate control. Results were automatically analyzed in the software Seegene Viewer version 2.0 (Seegene). Positive samples were semiquantitative, indicated by results at 30 (þþþ) cycles, 40 (þþ) cycles, or 50 (þ) cycles.
Genotyping with Reference Real-Time PCR
Genotyping with CLART HPV2
Targeting the viral gene E6 or E7, detection of 12 high-risk and two low-risk HPV, together with the human control gene beta-globin HBB, were performed with real-time PCR on the 7900 HT Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA). Reactions of 20 mL included one or two primereprobe pairs, each holding 0.9 mmol/L forward and reverse primers together with 0.2 mmol/L probe and 1 TaqMan Universal Master Mix (Life Technologies, Carlsbad, CA). Primers and probes (Thermo Fisher Scientific) were as according to Lindh et al14 with some alterations (Table 2). Primers were purified using reverse-phase cartridge purification, and probes were purified with highperformance liquid chromatography. PCR reactions underwent an initial step at 50 C for 2 minutes followed by a denaturation step at 95 C for 10 minutes before repeated cycling for 40 cycles at 95 C for 15 seconds followed by 60 C for 60 seconds. Each run of samples included positive, negative, and nontemplate controls. Approximately 50 ng of DNA was used in each reaction. Results were
CLART HPV2 (Genomica) detects 35 genotypes and the human gene CTFR in one reaction targeting the L1 region of the virus. CLART HPV2 is approved for in vitro diagnostic use in the European Union for both cytological and archival FFPE tissue. Amplicon length is 450 bp within the consensus gene L1. PCR-reactions were run on PCR equipment Eppendorf Mastercycler ep Gradient S (Eppendorf AG, Hamburg, Germany) in 50-mL reactions containing 45 mL of reaction mix (Genomica), 0.75 mmol/L MgCl (Thermo Fisher Scientific), and approximately 50 ng of DNA. After an initial denaturation step at 95 C for 5 minutes, reaction mixtures underwent 40 cycles at 94 C for 30 seconds followed by 55 C for 60 seconds, 72 C for 90 seconds, and finally a cycle of 4 C for 8 minutes. Detection of PCR product was made by a low-density microarray platform, CLART (Clinical Array Technology) according to specifications by the manufacturer. Results were automatically analyzed in CLART Human Papillomavirus 2 specific software version
DNA Extraction
The Journal of Molecular Diagnostics
-
jmd.amjpathol.org
295
Lillsunde Larsson et al Table 2 Genotype 6
11
16
18
31
33
35
39
45
51
52
56
58
59
HBB
Primers and Probes Used in the Reference Assay Sequence 50 to 30 0
Oligo 0
5 -RCGGTTYATAAAGCTAAATTGTACGT-3 50 -AGGGTAACATGTCTTCCATGCA-30 50 -VIC-AAGGGTCGCTGCCTACACTGCTGG-TAMRA-30 50 -GCTTCATAAAACTAAATAACCAGTGGAA-30 50 -GTCAGGAGGCTGCAGGTCTAGTA-30 50 -FAM-TCCAGCAGTGTAAGCAACGACCCTTCC-TAMRA-30 50 -TTGCAGATCATCAAGAACACGTAGA-30 50 -CAGTAGAGATCAGTTGTCTCTGGTTGC-30 50 -FAM-AATCATGCATGGAGATACACCTACATTGCATGA-TAMRA-30 50 -AGAGGCCAGTGCCATTCGT-30 50 -GTTTCTCTGCGTCGTTGGAGT-30 50 -VIC-TCCTGTCGTGCTCGGTTGCAGC-TAMRA-30 50 -ATTCCACAACATAGGAGGAAGGTG-30 50 -CACTTGGGTTTCAGTACGAGGTCT-30 50 -VIC-CTCCAACATGCTATGCAACGTCCTGTC-TAMRA-30 50 -ATATTTCGGGTCGTTGGGCA-30 50 -ACGTCACAGTGCAGTTTCTCTACGT-30 50 -VIC-GGACCTCCAACACGCCGCACA-TAMRA-30 50 -TCGGTGTATGTCCTGTTGGAAAC-30 50 -CATAGTCTTGCAATGTAGTTATTTCTCCA-30 50 -FAM-TGCATGATTACACCTCGGTTTCTCTACGTG-TAMRA-30 50 -GCAGGAAGCTATACAGGACAGTGTC-30 50 -CTTGGGTTTCTCTTCGTGTTAGTCT-30 50 -FAM-CCCGTTTTGTGGTCCAGCACCG-TAMRA-30 50 -GGACAGTACCGAGGGCAGTGTAA-30 50 -TCCCTACGTCTGCGAAGTCTTTC-30 50 -VIC-CATGTTGTGACCAGGCACGGCA-TAMRA-30 50 -AAAGCAAAAATTGGTGGACGA-30 50 -TGCCAGCAATTAGCGCATT-30 50 -FAM-CATGAAATAGCGGGACGTTGGACG-TAMRA-30 50 -GACATGTTAATGCAAACAAGCGAT-30 50 -CATGACGTTACACTTGGGTCACA-30 50 -VIC-TGTTCAGAGTGTTGGAGACCCCGACC-TAMRA-30 50 -TGCATTGTGACAGAAAAAGACGAT-30 50 -CTCCAGCACCCCAAACATG-30 50 -FAM-CCCGGTCCAACCATGTGCTATTAGATGA-TAMRA-30 50 -GGCATGTGGATTTAAACAAAAGGT-30 50 -TCTCATGGCGTTGTTACAGGTTAC-30 50 -FAM-CACTGCACAGCGCCCTGTCCAA-TAMRA-30 50 -TGTATGGAGAAACATTAGAGGCTGAA-30 50 -TGGACATAGAGGTTTTAGGCATCTATAA-30 50 -FAM-AGACACCGTTACATGAGCTGCTGATACGC-TAMRA-30 50 -GCTCATGGCAAGAAAGTGCTC-30 50 -GCAAAGGTGCCCTTGAGGT-30 50 -VIC-AGTGATGGCCTGGCTCACCTGGAC-TAMRA-30
F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P
Amplicon length (bp) 80
112
113
62
75
74
90
84
71
81
105
73
121
89
74
Primers are from Lindh et al14 with some alterations; VIC was used to label probes for HPV-6, -18, -33, -52, and -45. For HPV-58, HPV-11, and HPV-31, new probes were used. For HPV-18, a new reverse primer was used, as well as new primers and probe for beta-globin HBB. F, forward primer; P, probe; R, reverse primer.
8.sc.8 as well as manually surveyed using the CLART grid (Genomica).
The MGP5þ/6þ primer system detects multiple HPV genotypes by targeting the L1 region of the virus.15 Primers were
purified using desalting (Sigma-Aldrich, St. Louis, MO). Amplicon length is approximately 150 bp. PCR reactions were run on the 7500 Fast Real-Time PCR System (Thermo Fisher Scientific) in five parallel reactions. Reactions contained 1 QuantiTect SYBR Green Master Mix (Qiagen), 0.3 mmol/L forward and reverse primers (mix, five þ five) (Sigma-Aldrich), and approximately 50 ng of DNA. After a denaturation step at 95 C for 15 minutes, reaction mixtures first underwent five
296
jmd.amjpathol.org
Genotyping with MGP5þ/6þ Primers and Pyrosequencing
-
The Journal of Molecular Diagnostics
HPV Genotyping in FFPE Tissue Table 3
Concordance between Investigated Assays Compared to Reference Protocol
Reference genotypes
No. of samples
Agree- McNemar Kappa þ/þ / þ/ /þ ment P value (95% CI)
Anyplex 67 II HPV28 CLART 61 HPV2 MGP5þ/6þ 61 All genotypes Anyplex 67 II HPV28 CLART 61 HPV2 MGP5þ/6þ 63
29
0
3
97%
30
6
1
93%
31
6
1
93%
18
0
14
86%
21
6
10
84%
28
4
4
92%
Sensitivity* (95% CI)
Specificityy (95% CI)
PPVz (95% CI)
NPVx (95% CI)
0.250 0.93 (0.85e 1.00 (0.80e 0.91 (0.81e1.00) 0.96 (0.91e1.00) 1.00 (0.80e 1.00) 1.00) 1.00) 0.125 0.84 (0.73e 0.91 (0.84e 0.97 (0.91e1.00) 0.98 (0.95e1.00) 0.83 (0.71e 0.95) 0.98) 0.95) 0.125 0.84 (0.73e 0.91 (0.84e 0.97 (0.91e1.00) 0.98 (0.95e1.00) 0.84 (0.95e 0.96) 0.98) 1.00)
<0.01
0.64 (0.46e 1.00 (0.80e 0.56 (0.39e0.73) 0.83 (0.74e0.91) 1.00 (0.80e 0.81) 1.00) 1.00) 0.454 0.61 (0.43e 0.91 (0.84e 0.68 (0.51e0.84) 0.86 (0.78e0.94) 0.78 (0.62e 0.78) 0.98) 0.93) 1.000 0.81 (0.69e 0.94 (0.88e 0.88 (0.76e0.99) 0.94 (0.88e1.00) 0.88 (0.76e 0.94) 1.00) 0.99)
Calculations are made from positive and negative results only (no concordance of genotypes included). Reference method was used as the standard. For the CLART HPV2, 98 samples are used for calculations. Bold indicates a significant result. A significant difference <0.01 was found when comparing all detectable genotypes between the Anyplex II HPV 28 to the reference protocol. *Proportion of positive samples among the reference protocol positive samples. y Proportion of negative samples among the reference protocol negative samples. z Positive predictive value (PPV), proportion of reference protocol positive samples among positive samples. x Negative predictive value (NPV), proportion of reference protocol negative samples among negative samples.
cycles at 95 C for 30 seconds followed by 42 C for 30 seconds, and 72 C for 45 seconds, then 45 cycles followed with 95 C for 30 seconds followed by 64 C for 30 seconds, and 72 C for 45 seconds. A final step of 72 C for 10 minutes was included. Each run of samples included positive, negative, and nontemplate controls. Results were analyzed by 7500 Fast System SDS version 1.3.1 software (Thermo Fisher Scientific). Each curve was manually surveyed, and samples with a cycle threshold <35 were counted as positives. Sequencing of single-strand templates was performed using 1 mmol/L sequencing primer (using forward MGP5þ primers, one for each sequencing reaction) in a PSQ 96 MA system using PyroMark Gold Q96 reagents (Qiagen) with entry: cyclic_60_CTGA. The results were analyzed in the PyroMark ID/MD software version 1.0, and sequences were submitted to the National Center for Biotechnology Information Basic Local Alignment Search Tool (BLAST) using nucleotide BLAST.
Statistics
Sensitivity=Specificity=K=PPV=NPV 1:96 SE
ð1Þ
where PPV is the positive predictive value and NPV is the negative predictive value. For sensitivity or specificity of 100%, we used the Poisson distribution: Sensitivity=specificity 1:96
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sensitivity=specificity n
ð2Þ
SPSS Statistics version 22 (IBM, New York, NY) was used for calculation of the McNemar’s test.
Results
The data were separated using two approaches: comparing results from the 14 genotypes specified in the reference protocol and comparing results from all genotypes detected by each assay. The results from Anyplex II HPV28, CLART HPV2, and MGP5þ/6þ were compared against the reference protocol. Comparisons between tests were made using the McNemar’s test and by calculating Cohen’s kappa values. Kappa values between 0 and 0.2 were classified as slight agreement, between 0.2 and 0.4 as fair agreement, 0.4 to 0.6 as moderate agreement, 0.6 to 0.8 as substantial agreement, and above 0.8 as almost perfect agreement.16 No kappa values at the actual borders were
The Journal of Molecular Diagnostics
reported. McNemar’s test was chosen because genotyping data for all methods are from the same samples. Sensitivity, specificity, and positive and negative predictive values were calculated using the reference protocol as standard. Ninetyfive percent confidence intervals were calculated with the formula,
-
jmd.amjpathol.org
Here, 99 clinical FFPE samples were analyzed with Anyplex II HPV28, CLART HPV2, and the MGP5þ/6þ primer system, which were compared to a reference PCR protocol. Valid genotyping results were obtained from all samples and methods except for one sample that was invalid for the CLART HPV2 assay. In total, counting all results, 68% (67 of 99) of the samples were positive for an HPV genotype using the reference protocol and the MGP5þ/6þ primer system. When analyzed with the Anyplex II HPV28, 82% (81 of 99) of the samples were positive, and with the CLART HPV2
297
Lillsunde Larsson et al Table 4
Samples with the Same Genotyping Result in Three of Four Assays
No.
Consensus sample result (3 of 4 assays)
Deviating result
Assay
Detection level*
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
HPV-6 HPV-6 HPV-16 HPV-16 HPV-16 HPV-16 HPV-16 HPV-18 HPV-18 HPV-31 HPV-45 HPV-51 Negative Negative Negative Negative Negative Negative HPV-16 HPV-18 HPV-45 Negative Negative HPV-16/HPV-39 HPV-16
HPV-6/HPV-70 HPV-6/HPV-33 HPV-16/HPV-42/HPV-53 HPV-16/HPV-40 HPV-16/HPV-70 HPV-16/HPV-53 HPV-16/HPV-56 HPV-18/HPV-44 HPV-18/HPV-70/HPV-6 HPV-31/HPV-53 HPV-45/HPV-43 HPV-51/HPV-6 HPV-18 HPV-68/HPV-70 HPV-31 HPV-33/HPV-42 HPV-42 HPV-42 HPV-16/HPV-70 Negative Negative HPV-66 Invalid HPV-16 HPV-16/HPV-51
Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 CLART HPV2 CLART HPV2 CLART HPV2 CLART HPV2 CLART HPV2 MGP5þ/6þ primer system Reference protocol
50 cycles 50 cycles 50 cycles 50 cycles 50 cycles 50 cycles 50 cycles 40 cycles 50 cycles 50 cycles 50 cycles 50 cycles 50 cycles 40 cycles/50 cycles 50 cycles 50 cycles 40 cycles 50 cycles NA NA NA NA NA NA NA
Detection level is only shown for Anyplex II HPV28. *Detection level of deviating result. Genotypes detected after melting curve analysis at indicated cycle number. NA, not applicable.
assay, 72% (71 of 98) of the samples were positive. Raw data on all genotyping results are provided in Supplemental Table S1.
Detection of Genotypes Specified by Reference Protocol When only comparing the 14 genotypes from the reference protocol, kappa values ranged from 0.84 to 0.93, and the highest kappa value was seen between the reference protocol and Anyplex II HPV28 (95% CI, 0.85e1.00). No significant differences between methods (number of positive samples) were found. Agreement rates varied between 93% (CLART HPV2 and the MGP5þ/6þ primer system) and 97% (Anyplex II HPV 28). Anyplex II HPV28 had, relative to the reference protocol, the highest sensitivity of the assays that were compared (100%; 95% CI, 80%e100%), whereas CLART HPV2 and the MGP5þ/6þ primer system had the highest specificity (97%; 95% CI, 91%e100%). Positive predictive values varied between 96% and 98%, and the negative predictive value between 83% and 100% (Table 3).
Anyplex II HPV28 to the reference protocol. Agreement rates, as well as kappa values were lower, compared to only including the 14 genotypes, ranging between 84% to 92% and 0.61 to 0.81, respectively. Again, Anyplex II HPV28 had a sensitivity of 100% relative to the reference protocol (95% CI, 80%e 100%). Specificity was lower when including all genotypes in the calculation, between 56% and 88%, reflecting the differences in number of detectable genotypes. Highest specificity was noted for MGP5þ/6þ primer system (88%; 95% CI, 76%e99%). Positive predictive values varied between 83% and 94%, and the negative predictive values between 78% and 100% (Table 3).
Genotyping Differences between Assays
When comparing all detectable genotypes, a significant difference (McNemar test, P < 0.01) was found when comparing
When comparing all four assays and taking into consideration all genotypes possible, 55 of all samples (n Z 99) had the same genotyping result in all four assays. Results that were identical in all four assays were negative samples (n Z 15) and single infections of HPV-16 (n Z 22), HPV-6 (n Z 7), HPV-18 (n Z 4), HPV-31 (n Z 1), HPV-33 (n Z 1), HPV-35 (n Z 1), HPV-39 (n Z 1), HPV-45 (n Z 1), HPV-58 (n Z 1), and HPV-59 (n Z 1). Samples with deviating results between assays were further evaluated. Twenty-five of the 99 samples had the same genotyping result in three of four assays, but differed in
298
jmd.amjpathol.org
Detection of All Genotypes
-
The Journal of Molecular Diagnostics
HPV Genotyping in FFPE Tissue Table 5
Samples with the Same Genotyping Result in Two of Four Assays
Consensus sample result No. (2 of 4 assays)
Deviating result 1
Assay
Detection level*
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
HPV-16/HPV-70 HPV-45/HPV-42 HPV-51/HPV-43 HPV-18/HPV-58/HPV-53 HPV-31/HPV-33/HPV-16 HPV-68/HPV-70 HPV-70 HPV-42 HPV-66/HPV-73 HPV-66/HPV-42 HPV-66/HPV-43/HPV-70 HPV-16 HPV-39/HPV-66 Negative HPV-16 HPV-35
Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 CLART HPV2 CLART HPV2 Reference protocol
40 cycles HPV-16/HPV-70 40 cycles Negative 30 cycles HPV-43 40 cycles HPV-18 50 cycles HPV-33 30 cycles/50 cycles Negative 50 cycles HPV-70 40 cycles HPV-42 40 cycles HPV-66 40 cycles/50 cycles HPV-66 50 cycles HPV-70 50 cycles HPV-18 40 cycles/50 cycles HPV-39 NA Negative NA HPV-16 NA Negative
HPV-16 HPV-45 HPV-51 HPV-18/HPV-58 HPV-31/HPV-33 HPV-68 Negative Negative Negative Negative Negative Negative Negative HPV 56 Negative HPV 58
Deviating result 2 Assay CLART HPV2 CLART HPV2 MGP 5þ/6þ primer MGP 5þ/6þ primer MGP 5þ/6þ primer Reference protocol CLART HPV2 CLART HPV2 CLART HPV2 CLART HPV2 CLART HPV2 Reference protocol Reference protocol MGP 5þ/6þ primer MGP 5þ/6þ primer MGP 5þ/6þ primer
system system system
system system system
Detection level is only shown for Anyplex II HPV28. *Detection level of deviating result for Anyplex II HPV 28 assay. Genotypes detected after melting curve analysis at indicated cycle number. NA, not applicable.
one assay from the other. Among the 25 deviating samples, 18 had a different Anyplex II HPV28 result, 5 had a different result from the CLART HPV2 assay, and 1 sample each from the reference protocol and the MGP5þ/6 þ primer system showed deviating results (Table 4). Of the 18 diverging Anyplex II HPV28 assay results, 12 samples had additional genotypes found in addition to what the other assays detected, and in six samples Anyplex II HPV28 assay detected genotypes in samples that the other assays had classified as negatives. Genotypes were detected at 40 or at 50 cycles and the genotypes found were from International Agency for Research on Cancer Group 1 (n Z 5), Group 2A (n Z 1), Group 2B (n Z 7), or of the low-risk genotypes (n Z 9). For four samples, the Anyplex II HPV28 results changed the sample status from being negative to positive for a high-risk-type Group 1 genotype or a Group 2A genotype. Five samples differed by the CLART HPV2 assay only. In one of the five samples, result for the genomic DNA control failed, leaving the sample result invalid. Another one of the five samples found an additional genotype compared to the consensus result, and for one sample with the consensus result negative, CLART HPV2 genotyped the sample as HPV-66 (Group 2B). For two of the samples, CLART HPV2 failed to repeat the result reported from the other three assays (that were Table 6
concordant). The MGP5þ/6þ primer system failed to detect the same genotypes as the three other assays in one sample only. The reference protocol also had one sample diverging from the three other assays, in which both HPV-16 and HPV-51 were initially detected, but when re-analyzed with Anyplex II HPV28, CLART HPV2, and the MGP5þ/6þ primer system, the sample was positive only for HPV-16. In subdividing the four assay results further, 16 of the samples had the same genotyping result in two of four assays (Table 5). In brief, the Anyplex II HPV28 assay detected additional genotypes in 13 of 16 samples, whereas the CLART HPV2 added HPV information in 7 of the cases, whereas 2 cases were considered negative. Three samples had different genotyping results in all assays tested (Table 6).
Discussion Based on the perspective to evaluate genotyping assays with a high sensitivity for FFPE specimens, four different genotyping methods were used on 99 clinical samples with a suspicion of HPV infection. Generally, the Anyplex II HPV28 and CLART assays detected more positives in this material compared to the reference method, not surprisingly, given the broader
Samples with Different Genotyping Result in All Four Assays
No.
Reference protocol
Anyplex II HPV28
CLART HPV2
MGP5þ/6þ primer system
1 2 3
Negative HPV-33 Negative
HPV-70 HPV-33/HPV-66/HPV-73 HPV-66/HPV-73
HPV-82 HPV-33/HPV-66 HPV-66
HPV-91 HPV-90 HPV-90
The Journal of Molecular Diagnostics
-
jmd.amjpathol.org
299
Lillsunde Larsson et al genotyping possibilities with these tests. The MGP5þ/6þ assay found the same number of positives as the reference method, which has been used in our laboratory with good reproducibility in both internal and external control programs. Primers were originally validated to linear array (Roche, Basel, Switzerland) and have also exhibited a 100% proficiency in the World Health Organization LabNet proficiency panel study. All four methods showed good agreement when comparing the Group 1 genotypes together with the low-risk genotypes 6 and 11. Anyplex II HPV28 had the highest agreement rate when compared to the reference protocol; furthermore, it detected positive samples among the samples classified as negatives by the reference protocol. It is notable that the MGP5þ/6þ system had the same performance as the CLART assay. When comparing all genotypes, we found a significant difference with respect to positives between the reference protocol and Anyplex II HPV28. This finding was not surprising because of the wider genotype detection spectrum of the Anyplex II HPV28. We further evaluated the genotyping results that differed between methods. Anyplex II HPV28 detected more genotypes compared to the other assays. This was seen both as coinfecting genotypes and as single infections in samples that were classified as negatives by the other three assays. The CLART assay detected additional genotypes compared to the other assays, but also failed to repeat some results. The additional genotypes detected by Anyplex II HPV28 were also included in the CLART assay repertoire, and the failure of detection might be due to sensitivity differences between the assays. Despite the fact that the CLART assay is validated for FFPE material, the 450-bp amplicon may sometimes be difficult to amplify because of degradation of DNA. Anyplex II HPV28 also uses 10 more cycles in the PCR amplification, compared to CLART. The MGP5þ/6þ primer system was surprisingly concordant with the reference protocol. However, it did not manage to distinguish between coinfecting genotypes. This method could be improved by a type-specific sequencing, which might also validate the uncommon genotypes that were found. However, considering all genotyping results, the more extensive genotyping assays (Anyplex II HPV28 and CLART) changed the status of the sample (individually or together) from negative with respect to the reference protocol to a positive for Group 1 (n Z 4) and Group 2 (n Z 6) genotypes. The Group 1 genotypes should ideally have been detected by the other assays as well, but may have been unidentified because of a lower sensitivity compared to Anyplex II HPV28 and the CLART. Anyplex II HPV28 found many genotypes at 50 cycles, but how to clinically interpret such a result (ie, clinical sensitivity) is, however, not clear. One limitation of this study is also that samples were analyzed one time only with each assay, because of material limitations. Repeated results could strengthen the relevance of these results. The observed differences in the spectrum of detected genotypes within individual samples
300
could possibly also be due to the fact that assays were not run in parallel. Initially, samples were extracted for DNA, and the reference protocol was performed. Sample DNA was kept at 20 C until reanalyzed with the other assays. Study time spanned over a 1-yr period and could perhaps explain the loss of genotypes as a result of the freezeethaw process. The value of the more uncommon results of the Group 2 and the low-risk genotypes is not easily interpreted. Thus, from a research perspective, it is naturally valuable to increase our knowledge on coinfecting and uncommon genotypes. In the longer perspective, an enhanced genotyping will also provide more knowledge on the genotypes present and viral characteristics can be further validated in terms of variant, load, integration, and methylation. These data may add further knowledge to why certain infections persist, and lead to prevention strategies. Also, detailed genotyping information is important when evaluating HPV-related disease in populations where other concomitant infections exist. HIV has been associated with an increased risk of developing HPV-induced cancers,17 and the distribution and amount of HPV genotypes are also different in these areas.18,19 In conclusion, for an extended genotyping approach with a high sensitivity for FFPE specimens, Anyplex II HPV28, as well as CLART HPV2, is a suitable alternative despite minor intra-assay differences. In the clinical context, both methods find more Group 1 and 2 HPV-positive cases and could thereby contribute to personalized treatment.
Supplemental Data Supplemental material for this article can be found at http://dx.doi.org/10.1016/j.jmoldx.2014.12.004.
References 1. Poljak M, Cuzick J, Kocjan BJ, Iftner T, Dillner J, Arbyn M: Nucleic acid tests for the detection of alpha human papillomaviruses. Vaccine 2012, 30(Suppl 5):F100eF106 2. Larsson GL, Helenius G, Andersson S, Elgh F, Sorbe B, Karlsson MG: Human papillomavirus (HPV) and HPV 16-variant distribution in vulvar squamous cell carcinoma in Sweden. Int J Gynecol Cancer 2012, 22:1413e1419 3. Larsson GL, Helenius G, Andersson S, Sorbe B, Karlsson MG: Prognostic impact of human papilloma virus (HPV) genotyping and HPV-16 subtyping in vaginal carcinoma. Gynecol Oncol 2013, 129: 406e411 4. Hall JS, Iype R, Armenoult LS, Taylor J, Miller CJ, Davidson S, de Sanjose S, Bosch X, Stern PL, West CM: Poor prognosis associated with human papillomavirus alpha7 genotypes in cervical carcinoma cannot be explained by intrinsic radiosensitivity. Int J Radiat Oncol Biol Phys 2013, 85:e223ee229 5. Wang CC, Lai CH, Huang YT, Chao A, Chou HH, Hong JH: HPV genotypes predict survival benefits from concurrent chemotherapy and radiation therapy in advanced squamous cell carcinoma of the cervix. Int J Radiat Oncol Biol Phys 2012, 84:e499ee506 6. Kofler B, Laban S, Busch CJ, Lorincz B, Knecht R: New treatment strategies for HPV-positive head and neck cancer. Eur Arch Otorhinolaryngol 2014, 271:1861e1867
jmd.amjpathol.org
-
The Journal of Molecular Diagnostics
HPV Genotyping in FFPE Tissue 7. Bruni L, Barrionuevo-Rosas L, Albero G, Aldea M, Serrano B, Valencia S, Brotons M, Mena M, Cosano R, Muñoz J, Bosch FX, de Sanjosé S, Castellsagué X: Human Papillomavirus and Related Diseases in the World. Barcelona, Spain, ICO Information Centre on HPV and Cancer (HPV Information Centre), 2014 8. Schiffman M, Clifford G, Buonaguro FM: Classification of weakly carcinogenic human papillomavirus types: addressing the limits of epidemiology at the borderline. Infect Agent Cancer 2009, 4:8 9. Söderlund-Strand A, Carlson J, Dillner J: Modified general primer PCR system for sensitive detection of multiple types of oncogenic human papillomavirus. J Clin Microbiol 2009, 47:541e546 10. Bzhalava D, Guan P, Franceschi S, Dillner J, Clifford G: A systematic review of the prevalence of mucosal and cutaneous human papillomavirus types. Virology 2013, 445:224e231 11. Ludyga N, Grunwald B, Azimzadeh O, Englert S, Hofler H, Tapio S, Aubele M: Nucleic acids from long-term preserved FFPE tissues are suitable for downstream analyses. Virchows Arch 2012, 460:131e140 12. Srinivasan M, Sedmak D, Jewell S: Effect of fixatives and tissue processing on the content and integrity of nucleic acids. Am J Pathol 2002, 161:1961e1971 13. Fairley JA, Gilmour K, Walsh K: Making the most of pathological specimens: molecular diagnosis in formalin-fixed, paraffin embedded tissue. Curr Drug Targets 2012, 13:1475e1487
The Journal of Molecular Diagnostics
-
jmd.amjpathol.org
14. Lindh M, Gorander S, Andersson E, Horal P, Mattsby-Balzer I, Ryd W: Real-time Taqman PCR targeting 14 human papilloma virus types. J Clin Virol 2007, 40:321e324 15. de Roda Husman AM, Walboomers JM, van den Brule AJ, Meijer CJ, Snijders PJ: The use of general primers GP5 and GP6 elongated at their 3’ ends with adjacent highly conserved sequences improves human papillomavirus detection by PCR. J Gen Virol 1995, 76(Pt 4): 1057e1062 16. McGinn T, Wyer PC, Newman TB, Keitz S, Leipzig R, For GG: Tips for learners of evidence-based medicine: 3. Measures of observer variability (kappa statistic). CMAJ 2004, 171:1369e1373 17. Denny LA, Franceschi S, de Sanjose S, Heard I, Moscicki AB, Palefsky J: Human papillomavirus, human immunodeficiency virus and immunosuppression. Vaccine 2012, 30(Suppl 5):F168eF174 18. Denny L, Adewole I, Anorlu R, Dreyer G, Moodley M, Smith T, Snyman L, Wiredu E, Molijn A, Quint W, Ramakrishnan G, Schmidt J: Human papillomavirus prevalence and type distribution in invasive cervical cancer in sub-Saharan Africa. Int J Cancer 2014, 134:1389e1398 19. Ndiaye C, Alemany L, Ndiaye N, Kamate B, Diop Y, Odida M, Banjo K, Tous S, Klaustermeier JE, Clavero O, Castellsague X, Bosch FX, Trottier H, de Sanjose S: Human papillomavirus distribution in invasive cervical carcinoma in sub-Saharan Africa: could HIV explain the differences? Trop Med Int Health 2012, 17:1432e1440
301
jmd.amjpathol.org
Evaluation of HPV Genotyping Assays for Archival Clinical Samples Gabriella Lillsunde Larsson,* Jessica Carlsson,y Mats G. Karlsson,* and Gisela Helenius* From the Departments of Laboratory Medicine* and Urology,y Faculty of Medicine and Health, Örebro University, Örebro, Sweden Accepted for publication December 19, 2014. Address correspondence to Gabriella Lillsunde Larsson, Ph.D., Department of Laboratory Medicine, Örebro University Hospital, S-701 85 Örebro, Sweden. E-mail: gabriella.lillsunde-larsson@ regionorebrolan.se.
Human papillomavirus (HPV) testing and genotyping of FFPE tissue samples is important in epidemiological investigations. Here, we compare four different HPV genotyping methods for use in FFPE clinical samples. Comparative testing was performed on 99 samples with a clinical suspicion of HPV. Specimens were analyzed with Anyplex II HPV28 detecting 28 genotypes using real-time PCR and melting curve analysis, CLART HPV2 detecting 35 genotypes using PCR and microarray detection, and MGP5þ/6þ consensus primer system together with pyrosequencing. Results were compared to a realtime PCR reference protocol detecting 14 genotypes. In total, 68% of the samples were positive for an HPV genotype using the reference protocol and MGP5þ/6þ primer system. Anyplex II HPV28 analysis and CLART HPV2 had 82% and 72% positive samples, respectively. All four methods showed good agreement when comparing the 14 genotypes included in the reference protocol. When evaluating all genotypes, the Anyplex II HPV28 assay and the CLART assay changed the status of the sample (individually or together) from negative with respect to the reference protocol to positive for either a Group 1 (n Z 4) or Group 2 (n Z 6) genotype. We conclude from this study that for an extended genotyping approach with a high sensitivity for FFPE specimens, both the Anyplex II HPV28 and CLART HPV2 assays are suitable alternatives despite minor intra-assay differences. (J Mol Diagn 2015, 17: 293e301; http://dx.doi.org/10.1016/j.jmoldx.2014.12.004)
Human papillomavirus (HPV) genotype information has become relevant in several situations. The most established setting is the possibility to separate lesions with low-risk genotypes from the lesions with high-risk genotypes in cervical dysplasia. However, further information concerning specific genotypes is, or will soon be, needed in the clinical setting, such as review of HPV infections in the vaccinated population.1 Even though cytological samples are available in many situations, formalin-fixed, paraffin-embedded (FFPE) tissue samples may be the only source for HPV analysis of the lesion on several occasions, for example to establish HPV status and genotype in genital and head-and neck tumors for prognosis2e5 and personalized treatment strategies.6 In the HPV alpha genus of high-risk types, HPV-16 and -18 are the most emergent ones that together account for about 70% of all invasive cervical cancers.7 Ten additional genotypes are included in this group carcinogenic to humans (Group 1) according to the International Agency for Research on Cancer. Additional alpha genotypes from the same clade Copyright ª 2015 American Society for Investigative Pathology and the Association for Molecular Pathology. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jmoldx.2014.12.004
are classified as Group 2 genotypes, ie, 2A, probably carcinogenic, and Group 2B, possibly carcinogenic8 (Table 1). Also, in a recent meta-analysis on cervical samples, Bzhalava et al10 presented data that suggest the addition of HPV-26, -67, -68, -69, -73, and -82 to the Group 1 classification because these genotypes were more present in invasive cervical cancer samples compared to the normal cytology. Anogenital tumors, alongside invasive cervical cancers, are also to a varying extent positive for an HPV genotype. HPV16 is frequent, but genotypes such as HPV-6, -11, -22, and -40 may also occur.7 Despite the smaller numbers of cases, these findings argue for a broader genotyping approach. Supported by the research committee of Örebro University Hospital at Örebro County Council grant OLL-405821. Portions of this work have been presented in the doctoral dissertation of G.L.L. Disclosures: G.L.L. received travel support from Seegene for participating in the 24th European Congress of Clinical Microbiology and Infectious Diseases held May 10e13, 2014, in Barcelona, Spain.
Lillsunde Larsson et al
The market for HPV testing is fast growing, and in a review, Poljak et al1 reported 54 different full genotyping tests (including all Group A genotypes), with different methodology approaches such as strip- or filter-based, microarray tests, microsphere bead tests, gel electrophoreses tests, capillary electrophoreses tests, and real-time PCR tests. Most tests, however, are only validated on
cytology samples. Tissue samples used for clinical diagnosis are mostly archived by preserving the tissue in formalin followed by paraffin embedding. Formalin fixation is excellent for preserving the morphology of the tissue, but it also creates crosslinking and degradation of the nucleic acids.11,12 Modern molecular pathology, however, has overcome those problems mainly by improved DNA extraction protocols, using shorter PCR templates and adjusted PCR protocols.13 We have routinely used a previously reported real-time PCR as the genotyping method, with primers and probes specific for the viral E6 and E7 genes.14 The protocol includes the Group 1 genotypes together with the low-risk types 6 and 11. This method shows good clinical results for both cytological and FFPE specimens. However, eight separate reactions are needed for each sample, and none of the Group 2 genotypes are included. Based on the perspective to include more genotypes, rationalize the workflow, as well as to evaluate assays using different detection technologies, a comparison of assays for FFPE samples was performed. In the Anyplex II HPV28 assay (Seegene, Seoul, Korea), detection of 28 different alpha genotypes is performed using real-time PCR followed by melting-curve analysis at 30, 40, and 50 cycles. Using TOCE technology (Seegene), variation in melting temperature values during melting curve analysis is avoided, allowing for a very sensitive multiplexed detection approach (Seegene technical manual: Anyplex II HPV28 detection). This assay has not previously been evaluated on FFPE specimens but is approved for in vitro diagnostic use in the European Union for cytology specimens. The CLART system from Genomica (Madrid, Spain) is widely used in clinical laboratories and has been approved for in vitro diagnostic use in the European Union for use in FFPE samples. It distinguishes between 35 different genotypes by a low-density array. PCR products that are tagged with biotin hybridize to the array, conjugate is added, and the product is then visualized by precipitation on spot (Genomica CLART technical manual: genotyping of human papillomavirus via genomic identification for in vitro diagnosis). Finally, a consensus approach was included in this comparative study, using the MGP5þ/6þ primers.9 The MGP5þ/6þ primers were developed from the original GP5þ/6þ primers,15 and the improved assay included five primer pairs, with extended primer length, to increase specificity for the Group 1 genotypes together with HPV-66 and -68. In a clinical evaluation by Söderlund-Strand et al,9 additional Group 2 genotypes and low-risk genotypes could also be detected. Our aim was to evaluate four different genotyping methods for use in FFPE clinical samples with a suspicion of HPV infection concerning HPV detection and specific genotype distribution. The two commercially available HPV genotyping tests, the Anyplex II HPV28 from Seegene and the CLART assay from Genomica, as well as the consensus MGP5þ/6þ primer system, were evaluated and compared to a real-time PCR reference protocol.
294
jmd.amjpathol.org
Table 1 Overview of Detectable Genotypes in the Different Assays in the Study MGP5þ/6þ IARC Reference Anyplex CLART primer Genotype classification protocol II HPV 28 HPV2 system 16 18 31 33 35 39 45 51 52 56 58 59 68 26 30 34/64 53 66 67 69 70 73 82 85 6 11 40 42 43 44 54 61 62 71 72 81 83 84 89 No.
1 1 1 1 1 1 1 1 1 1 1 1 2A 2B 2B 2B 2B 2B 2B 2B 2B 2B 2B 2B 3 3
x x x x x x x x x x x x
x x x x x x x x x x x x x x
x x x x x x x x x x x x x x
x x
x x
x x x x x x
14
x x x x x x x x
28
x x x x x x x x x x x x x x x x x x x 35
x x x x x x x x x x x x x
x
x x x x x x x
21
Genotypes detected by MGP5þ/6þ system as described by SöderlundStrand et al.9 Classification according to International Agency for Research on Cancer (IARC); 1: carcinogenic to humans, 2A: probably carcinogenic, 2B: possibly carcinogenic, 3: not classifiable as to its carcinogenity to humans. Additional low-risk genotypes are included in the Anyplex II HPV28 and CLART HPV2 assays.
-
The Journal of Molecular Diagnostics
HPV Genotyping in FFPE Tissue
Materials and Methods Samples In this study, 99 FFPE samples from patients with a clinical suspicion of HPV infection at the Department of Pathology, Örebro University Hospital, were included. The majority of the samples were cervical lesions (61%; 60 of 99) or penile carcinomas (20%; 20 of 99), and the remaining samples were a mix of samples from vulva (5%; 5 of 99), vagina (1%; 1 of 99), anus (2%; 2 of 99), head and neck (3%; 3 of 99), larynx (2%; 2 of 99), skin (4%; 4 of 99), bladder (1%; 1 of 99), and lymph node (1%; 1 of 99). All samples were anonymized before analysis.
analyzed by 7900 Fast System SDS version 2.4 software for ABI 7900 HT (Thermo Fisher Scientific). Each curve was manually surveyed, and samples with a cycle threshold <35 were considered as positives.
Genotyping with Anyplex II HPV28
DNA was extracted from FFPE samples. Tissue blocks were selected by a pathologist, and the tissue was sectioned in four 5- to 20-mm sections. Sections were placed in a clean microcentrifuge tube, deparaffinized, and incubated overnight (56 C) with proteinase K (Qiagen, Hilden, Germany). DNA was extracted using QiaAmp DNA Mini Kit (Qiagen) according to the manufacturer’s instructions. Samples were eluted in a volume of 75 mL. Quality and quantity of DNA was spectrometrically estimated (NanoDrop Technologies, Wilmington, DE). For each test, eluates were diluted to a concentration of 10 ng/mL. In every reaction and assay, 5 mL (50 ng) eluate was used. For each batch of samples, a blank water sample was included in the extraction to monitor for cross contamination. The blank sample was analyzed with the reference real-time PCR assay. For an overview of included genotypes in the different assays, see Table 1.
Anyplex II HPV28 detects 28 genotypes and the human gene beta-globin HBB in two multiplex reactions. Anyplex II HPV28 has previously been validated on liquid-based cytology specimens and is approved for in vitro diagnostic use in the European Union, though it has not been validated on archival FFPE tissue. Amplicon lengths are between 100 and 200 bp, and designed for the consensus gene L1. Reactions were run on a CFX96TM Real-Time PCR System (Bio-Rad Laboratories, Hercules, CA) in 20-mL reactions containing 1 HPV28 primer mix A or B, 1 Anyplex Master Mix, and approximately 50 ng of DNA. After an initial step at 50 C for 4 minutes and a denaturation step at 95 C for 15 minutes, reaction mixtures underwent 30 cycles at 95 C for 30 seconds followed by 60 C for 60 seconds and 72 C for 30 seconds. This was followed by a 55 C step for 30 seconds and a melting curve analysis at 55 to 85 C (5s/0.5 C). Reactions continued cycling at the same temperatures for 10 cycles followed by melting curve analysis, and this was repeated once more for another 10 cycles. Melting curve data were obtained at 30, 40, and 50 cycles. For each run, three positive controls, ie, plasmid constructs of all HPV types, were included as well as a nontemplate control. Results were automatically analyzed in the software Seegene Viewer version 2.0 (Seegene). Positive samples were semiquantitative, indicated by results at 30 (þþþ) cycles, 40 (þþ) cycles, or 50 (þ) cycles.
Genotyping with Reference Real-Time PCR
Genotyping with CLART HPV2
Targeting the viral gene E6 or E7, detection of 12 high-risk and two low-risk HPV, together with the human control gene beta-globin HBB, were performed with real-time PCR on the 7900 HT Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA). Reactions of 20 mL included one or two primereprobe pairs, each holding 0.9 mmol/L forward and reverse primers together with 0.2 mmol/L probe and 1 TaqMan Universal Master Mix (Life Technologies, Carlsbad, CA). Primers and probes (Thermo Fisher Scientific) were as according to Lindh et al14 with some alterations (Table 2). Primers were purified using reverse-phase cartridge purification, and probes were purified with highperformance liquid chromatography. PCR reactions underwent an initial step at 50 C for 2 minutes followed by a denaturation step at 95 C for 10 minutes before repeated cycling for 40 cycles at 95 C for 15 seconds followed by 60 C for 60 seconds. Each run of samples included positive, negative, and nontemplate controls. Approximately 50 ng of DNA was used in each reaction. Results were
CLART HPV2 (Genomica) detects 35 genotypes and the human gene CTFR in one reaction targeting the L1 region of the virus. CLART HPV2 is approved for in vitro diagnostic use in the European Union for both cytological and archival FFPE tissue. Amplicon length is 450 bp within the consensus gene L1. PCR-reactions were run on PCR equipment Eppendorf Mastercycler ep Gradient S (Eppendorf AG, Hamburg, Germany) in 50-mL reactions containing 45 mL of reaction mix (Genomica), 0.75 mmol/L MgCl (Thermo Fisher Scientific), and approximately 50 ng of DNA. After an initial denaturation step at 95 C for 5 minutes, reaction mixtures underwent 40 cycles at 94 C for 30 seconds followed by 55 C for 60 seconds, 72 C for 90 seconds, and finally a cycle of 4 C for 8 minutes. Detection of PCR product was made by a low-density microarray platform, CLART (Clinical Array Technology) according to specifications by the manufacturer. Results were automatically analyzed in CLART Human Papillomavirus 2 specific software version
DNA Extraction
The Journal of Molecular Diagnostics
-
jmd.amjpathol.org
295
Lillsunde Larsson et al Table 2 Genotype 6
11
16
18
31
33
35
39
45
51
52
56
58
59
HBB
Primers and Probes Used in the Reference Assay Sequence 50 to 30 0
Oligo 0
5 -RCGGTTYATAAAGCTAAATTGTACGT-3 50 -AGGGTAACATGTCTTCCATGCA-30 50 -VIC-AAGGGTCGCTGCCTACACTGCTGG-TAMRA-30 50 -GCTTCATAAAACTAAATAACCAGTGGAA-30 50 -GTCAGGAGGCTGCAGGTCTAGTA-30 50 -FAM-TCCAGCAGTGTAAGCAACGACCCTTCC-TAMRA-30 50 -TTGCAGATCATCAAGAACACGTAGA-30 50 -CAGTAGAGATCAGTTGTCTCTGGTTGC-30 50 -FAM-AATCATGCATGGAGATACACCTACATTGCATGA-TAMRA-30 50 -AGAGGCCAGTGCCATTCGT-30 50 -GTTTCTCTGCGTCGTTGGAGT-30 50 -VIC-TCCTGTCGTGCTCGGTTGCAGC-TAMRA-30 50 -ATTCCACAACATAGGAGGAAGGTG-30 50 -CACTTGGGTTTCAGTACGAGGTCT-30 50 -VIC-CTCCAACATGCTATGCAACGTCCTGTC-TAMRA-30 50 -ATATTTCGGGTCGTTGGGCA-30 50 -ACGTCACAGTGCAGTTTCTCTACGT-30 50 -VIC-GGACCTCCAACACGCCGCACA-TAMRA-30 50 -TCGGTGTATGTCCTGTTGGAAAC-30 50 -CATAGTCTTGCAATGTAGTTATTTCTCCA-30 50 -FAM-TGCATGATTACACCTCGGTTTCTCTACGTG-TAMRA-30 50 -GCAGGAAGCTATACAGGACAGTGTC-30 50 -CTTGGGTTTCTCTTCGTGTTAGTCT-30 50 -FAM-CCCGTTTTGTGGTCCAGCACCG-TAMRA-30 50 -GGACAGTACCGAGGGCAGTGTAA-30 50 -TCCCTACGTCTGCGAAGTCTTTC-30 50 -VIC-CATGTTGTGACCAGGCACGGCA-TAMRA-30 50 -AAAGCAAAAATTGGTGGACGA-30 50 -TGCCAGCAATTAGCGCATT-30 50 -FAM-CATGAAATAGCGGGACGTTGGACG-TAMRA-30 50 -GACATGTTAATGCAAACAAGCGAT-30 50 -CATGACGTTACACTTGGGTCACA-30 50 -VIC-TGTTCAGAGTGTTGGAGACCCCGACC-TAMRA-30 50 -TGCATTGTGACAGAAAAAGACGAT-30 50 -CTCCAGCACCCCAAACATG-30 50 -FAM-CCCGGTCCAACCATGTGCTATTAGATGA-TAMRA-30 50 -GGCATGTGGATTTAAACAAAAGGT-30 50 -TCTCATGGCGTTGTTACAGGTTAC-30 50 -FAM-CACTGCACAGCGCCCTGTCCAA-TAMRA-30 50 -TGTATGGAGAAACATTAGAGGCTGAA-30 50 -TGGACATAGAGGTTTTAGGCATCTATAA-30 50 -FAM-AGACACCGTTACATGAGCTGCTGATACGC-TAMRA-30 50 -GCTCATGGCAAGAAAGTGCTC-30 50 -GCAAAGGTGCCCTTGAGGT-30 50 -VIC-AGTGATGGCCTGGCTCACCTGGAC-TAMRA-30
F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P F R P
Amplicon length (bp) 80
112
113
62
75
74
90
84
71
81
105
73
121
89
74
Primers are from Lindh et al14 with some alterations; VIC was used to label probes for HPV-6, -18, -33, -52, and -45. For HPV-58, HPV-11, and HPV-31, new probes were used. For HPV-18, a new reverse primer was used, as well as new primers and probe for beta-globin HBB. F, forward primer; P, probe; R, reverse primer.
8.sc.8 as well as manually surveyed using the CLART grid (Genomica).
The MGP5þ/6þ primer system detects multiple HPV genotypes by targeting the L1 region of the virus.15 Primers were
purified using desalting (Sigma-Aldrich, St. Louis, MO). Amplicon length is approximately 150 bp. PCR reactions were run on the 7500 Fast Real-Time PCR System (Thermo Fisher Scientific) in five parallel reactions. Reactions contained 1 QuantiTect SYBR Green Master Mix (Qiagen), 0.3 mmol/L forward and reverse primers (mix, five þ five) (Sigma-Aldrich), and approximately 50 ng of DNA. After a denaturation step at 95 C for 15 minutes, reaction mixtures first underwent five
296
jmd.amjpathol.org
Genotyping with MGP5þ/6þ Primers and Pyrosequencing
-
The Journal of Molecular Diagnostics
HPV Genotyping in FFPE Tissue Table 3
Concordance between Investigated Assays Compared to Reference Protocol
Reference genotypes
No. of samples
Agree- McNemar Kappa þ/þ / þ/ /þ ment P value (95% CI)
Anyplex 67 II HPV28 CLART 61 HPV2 MGP5þ/6þ 61 All genotypes Anyplex 67 II HPV28 CLART 61 HPV2 MGP5þ/6þ 63
29
0
3
97%
30
6
1
93%
31
6
1
93%
18
0
14
86%
21
6
10
84%
28
4
4
92%
Sensitivity* (95% CI)
Specificityy (95% CI)
PPVz (95% CI)
NPVx (95% CI)
0.250 0.93 (0.85e 1.00 (0.80e 0.91 (0.81e1.00) 0.96 (0.91e1.00) 1.00 (0.80e 1.00) 1.00) 1.00) 0.125 0.84 (0.73e 0.91 (0.84e 0.97 (0.91e1.00) 0.98 (0.95e1.00) 0.83 (0.71e 0.95) 0.98) 0.95) 0.125 0.84 (0.73e 0.91 (0.84e 0.97 (0.91e1.00) 0.98 (0.95e1.00) 0.84 (0.95e 0.96) 0.98) 1.00)
<0.01
0.64 (0.46e 1.00 (0.80e 0.56 (0.39e0.73) 0.83 (0.74e0.91) 1.00 (0.80e 0.81) 1.00) 1.00) 0.454 0.61 (0.43e 0.91 (0.84e 0.68 (0.51e0.84) 0.86 (0.78e0.94) 0.78 (0.62e 0.78) 0.98) 0.93) 1.000 0.81 (0.69e 0.94 (0.88e 0.88 (0.76e0.99) 0.94 (0.88e1.00) 0.88 (0.76e 0.94) 1.00) 0.99)
Calculations are made from positive and negative results only (no concordance of genotypes included). Reference method was used as the standard. For the CLART HPV2, 98 samples are used for calculations. Bold indicates a significant result. A significant difference <0.01 was found when comparing all detectable genotypes between the Anyplex II HPV 28 to the reference protocol. *Proportion of positive samples among the reference protocol positive samples. y Proportion of negative samples among the reference protocol negative samples. z Positive predictive value (PPV), proportion of reference protocol positive samples among positive samples. x Negative predictive value (NPV), proportion of reference protocol negative samples among negative samples.
cycles at 95 C for 30 seconds followed by 42 C for 30 seconds, and 72 C for 45 seconds, then 45 cycles followed with 95 C for 30 seconds followed by 64 C for 30 seconds, and 72 C for 45 seconds. A final step of 72 C for 10 minutes was included. Each run of samples included positive, negative, and nontemplate controls. Results were analyzed by 7500 Fast System SDS version 1.3.1 software (Thermo Fisher Scientific). Each curve was manually surveyed, and samples with a cycle threshold <35 were counted as positives. Sequencing of single-strand templates was performed using 1 mmol/L sequencing primer (using forward MGP5þ primers, one for each sequencing reaction) in a PSQ 96 MA system using PyroMark Gold Q96 reagents (Qiagen) with entry: cyclic_60_CTGA. The results were analyzed in the PyroMark ID/MD software version 1.0, and sequences were submitted to the National Center for Biotechnology Information Basic Local Alignment Search Tool (BLAST) using nucleotide BLAST.
Statistics
Sensitivity=Specificity=K=PPV=NPV 1:96 SE
ð1Þ
where PPV is the positive predictive value and NPV is the negative predictive value. For sensitivity or specificity of 100%, we used the Poisson distribution: Sensitivity=specificity 1:96
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi sensitivity=specificity n
ð2Þ
SPSS Statistics version 22 (IBM, New York, NY) was used for calculation of the McNemar’s test.
Results
The data were separated using two approaches: comparing results from the 14 genotypes specified in the reference protocol and comparing results from all genotypes detected by each assay. The results from Anyplex II HPV28, CLART HPV2, and MGP5þ/6þ were compared against the reference protocol. Comparisons between tests were made using the McNemar’s test and by calculating Cohen’s kappa values. Kappa values between 0 and 0.2 were classified as slight agreement, between 0.2 and 0.4 as fair agreement, 0.4 to 0.6 as moderate agreement, 0.6 to 0.8 as substantial agreement, and above 0.8 as almost perfect agreement.16 No kappa values at the actual borders were
The Journal of Molecular Diagnostics
reported. McNemar’s test was chosen because genotyping data for all methods are from the same samples. Sensitivity, specificity, and positive and negative predictive values were calculated using the reference protocol as standard. Ninetyfive percent confidence intervals were calculated with the formula,
-
jmd.amjpathol.org
Here, 99 clinical FFPE samples were analyzed with Anyplex II HPV28, CLART HPV2, and the MGP5þ/6þ primer system, which were compared to a reference PCR protocol. Valid genotyping results were obtained from all samples and methods except for one sample that was invalid for the CLART HPV2 assay. In total, counting all results, 68% (67 of 99) of the samples were positive for an HPV genotype using the reference protocol and the MGP5þ/6þ primer system. When analyzed with the Anyplex II HPV28, 82% (81 of 99) of the samples were positive, and with the CLART HPV2
297
Lillsunde Larsson et al Table 4
Samples with the Same Genotyping Result in Three of Four Assays
No.
Consensus sample result (3 of 4 assays)
Deviating result
Assay
Detection level*
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
HPV-6 HPV-6 HPV-16 HPV-16 HPV-16 HPV-16 HPV-16 HPV-18 HPV-18 HPV-31 HPV-45 HPV-51 Negative Negative Negative Negative Negative Negative HPV-16 HPV-18 HPV-45 Negative Negative HPV-16/HPV-39 HPV-16
HPV-6/HPV-70 HPV-6/HPV-33 HPV-16/HPV-42/HPV-53 HPV-16/HPV-40 HPV-16/HPV-70 HPV-16/HPV-53 HPV-16/HPV-56 HPV-18/HPV-44 HPV-18/HPV-70/HPV-6 HPV-31/HPV-53 HPV-45/HPV-43 HPV-51/HPV-6 HPV-18 HPV-68/HPV-70 HPV-31 HPV-33/HPV-42 HPV-42 HPV-42 HPV-16/HPV-70 Negative Negative HPV-66 Invalid HPV-16 HPV-16/HPV-51
Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 CLART HPV2 CLART HPV2 CLART HPV2 CLART HPV2 CLART HPV2 MGP5þ/6þ primer system Reference protocol
50 cycles 50 cycles 50 cycles 50 cycles 50 cycles 50 cycles 50 cycles 40 cycles 50 cycles 50 cycles 50 cycles 50 cycles 50 cycles 40 cycles/50 cycles 50 cycles 50 cycles 40 cycles 50 cycles NA NA NA NA NA NA NA
Detection level is only shown for Anyplex II HPV28. *Detection level of deviating result. Genotypes detected after melting curve analysis at indicated cycle number. NA, not applicable.
assay, 72% (71 of 98) of the samples were positive. Raw data on all genotyping results are provided in Supplemental Table S1.
Detection of Genotypes Specified by Reference Protocol When only comparing the 14 genotypes from the reference protocol, kappa values ranged from 0.84 to 0.93, and the highest kappa value was seen between the reference protocol and Anyplex II HPV28 (95% CI, 0.85e1.00). No significant differences between methods (number of positive samples) were found. Agreement rates varied between 93% (CLART HPV2 and the MGP5þ/6þ primer system) and 97% (Anyplex II HPV 28). Anyplex II HPV28 had, relative to the reference protocol, the highest sensitivity of the assays that were compared (100%; 95% CI, 80%e100%), whereas CLART HPV2 and the MGP5þ/6þ primer system had the highest specificity (97%; 95% CI, 91%e100%). Positive predictive values varied between 96% and 98%, and the negative predictive value between 83% and 100% (Table 3).
Anyplex II HPV28 to the reference protocol. Agreement rates, as well as kappa values were lower, compared to only including the 14 genotypes, ranging between 84% to 92% and 0.61 to 0.81, respectively. Again, Anyplex II HPV28 had a sensitivity of 100% relative to the reference protocol (95% CI, 80%e 100%). Specificity was lower when including all genotypes in the calculation, between 56% and 88%, reflecting the differences in number of detectable genotypes. Highest specificity was noted for MGP5þ/6þ primer system (88%; 95% CI, 76%e99%). Positive predictive values varied between 83% and 94%, and the negative predictive values between 78% and 100% (Table 3).
Genotyping Differences between Assays
When comparing all detectable genotypes, a significant difference (McNemar test, P < 0.01) was found when comparing
When comparing all four assays and taking into consideration all genotypes possible, 55 of all samples (n Z 99) had the same genotyping result in all four assays. Results that were identical in all four assays were negative samples (n Z 15) and single infections of HPV-16 (n Z 22), HPV-6 (n Z 7), HPV-18 (n Z 4), HPV-31 (n Z 1), HPV-33 (n Z 1), HPV-35 (n Z 1), HPV-39 (n Z 1), HPV-45 (n Z 1), HPV-58 (n Z 1), and HPV-59 (n Z 1). Samples with deviating results between assays were further evaluated. Twenty-five of the 99 samples had the same genotyping result in three of four assays, but differed in
298
jmd.amjpathol.org
Detection of All Genotypes
-
The Journal of Molecular Diagnostics
HPV Genotyping in FFPE Tissue Table 5
Samples with the Same Genotyping Result in Two of Four Assays
Consensus sample result No. (2 of 4 assays)
Deviating result 1
Assay
Detection level*
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
HPV-16/HPV-70 HPV-45/HPV-42 HPV-51/HPV-43 HPV-18/HPV-58/HPV-53 HPV-31/HPV-33/HPV-16 HPV-68/HPV-70 HPV-70 HPV-42 HPV-66/HPV-73 HPV-66/HPV-42 HPV-66/HPV-43/HPV-70 HPV-16 HPV-39/HPV-66 Negative HPV-16 HPV-35
Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 Anyplex II HPV28 CLART HPV2 CLART HPV2 Reference protocol
40 cycles HPV-16/HPV-70 40 cycles Negative 30 cycles HPV-43 40 cycles HPV-18 50 cycles HPV-33 30 cycles/50 cycles Negative 50 cycles HPV-70 40 cycles HPV-42 40 cycles HPV-66 40 cycles/50 cycles HPV-66 50 cycles HPV-70 50 cycles HPV-18 40 cycles/50 cycles HPV-39 NA Negative NA HPV-16 NA Negative
HPV-16 HPV-45 HPV-51 HPV-18/HPV-58 HPV-31/HPV-33 HPV-68 Negative Negative Negative Negative Negative Negative Negative HPV 56 Negative HPV 58
Deviating result 2 Assay CLART HPV2 CLART HPV2 MGP 5þ/6þ primer MGP 5þ/6þ primer MGP 5þ/6þ primer Reference protocol CLART HPV2 CLART HPV2 CLART HPV2 CLART HPV2 CLART HPV2 Reference protocol Reference protocol MGP 5þ/6þ primer MGP 5þ/6þ primer MGP 5þ/6þ primer
system system system
system system system
Detection level is only shown for Anyplex II HPV28. *Detection level of deviating result for Anyplex II HPV 28 assay. Genotypes detected after melting curve analysis at indicated cycle number. NA, not applicable.
one assay from the other. Among the 25 deviating samples, 18 had a different Anyplex II HPV28 result, 5 had a different result from the CLART HPV2 assay, and 1 sample each from the reference protocol and the MGP5þ/6 þ primer system showed deviating results (Table 4). Of the 18 diverging Anyplex II HPV28 assay results, 12 samples had additional genotypes found in addition to what the other assays detected, and in six samples Anyplex II HPV28 assay detected genotypes in samples that the other assays had classified as negatives. Genotypes were detected at 40 or at 50 cycles and the genotypes found were from International Agency for Research on Cancer Group 1 (n Z 5), Group 2A (n Z 1), Group 2B (n Z 7), or of the low-risk genotypes (n Z 9). For four samples, the Anyplex II HPV28 results changed the sample status from being negative to positive for a high-risk-type Group 1 genotype or a Group 2A genotype. Five samples differed by the CLART HPV2 assay only. In one of the five samples, result for the genomic DNA control failed, leaving the sample result invalid. Another one of the five samples found an additional genotype compared to the consensus result, and for one sample with the consensus result negative, CLART HPV2 genotyped the sample as HPV-66 (Group 2B). For two of the samples, CLART HPV2 failed to repeat the result reported from the other three assays (that were Table 6
concordant). The MGP5þ/6þ primer system failed to detect the same genotypes as the three other assays in one sample only. The reference protocol also had one sample diverging from the three other assays, in which both HPV-16 and HPV-51 were initially detected, but when re-analyzed with Anyplex II HPV28, CLART HPV2, and the MGP5þ/6þ primer system, the sample was positive only for HPV-16. In subdividing the four assay results further, 16 of the samples had the same genotyping result in two of four assays (Table 5). In brief, the Anyplex II HPV28 assay detected additional genotypes in 13 of 16 samples, whereas the CLART HPV2 added HPV information in 7 of the cases, whereas 2 cases were considered negative. Three samples had different genotyping results in all assays tested (Table 6).
Discussion Based on the perspective to evaluate genotyping assays with a high sensitivity for FFPE specimens, four different genotyping methods were used on 99 clinical samples with a suspicion of HPV infection. Generally, the Anyplex II HPV28 and CLART assays detected more positives in this material compared to the reference method, not surprisingly, given the broader
Samples with Different Genotyping Result in All Four Assays
No.
Reference protocol
Anyplex II HPV28
CLART HPV2
MGP5þ/6þ primer system
1 2 3
Negative HPV-33 Negative
HPV-70 HPV-33/HPV-66/HPV-73 HPV-66/HPV-73
HPV-82 HPV-33/HPV-66 HPV-66
HPV-91 HPV-90 HPV-90
The Journal of Molecular Diagnostics
-
jmd.amjpathol.org
299
Lillsunde Larsson et al genotyping possibilities with these tests. The MGP5þ/6þ assay found the same number of positives as the reference method, which has been used in our laboratory with good reproducibility in both internal and external control programs. Primers were originally validated to linear array (Roche, Basel, Switzerland) and have also exhibited a 100% proficiency in the World Health Organization LabNet proficiency panel study. All four methods showed good agreement when comparing the Group 1 genotypes together with the low-risk genotypes 6 and 11. Anyplex II HPV28 had the highest agreement rate when compared to the reference protocol; furthermore, it detected positive samples among the samples classified as negatives by the reference protocol. It is notable that the MGP5þ/6þ system had the same performance as the CLART assay. When comparing all genotypes, we found a significant difference with respect to positives between the reference protocol and Anyplex II HPV28. This finding was not surprising because of the wider genotype detection spectrum of the Anyplex II HPV28. We further evaluated the genotyping results that differed between methods. Anyplex II HPV28 detected more genotypes compared to the other assays. This was seen both as coinfecting genotypes and as single infections in samples that were classified as negatives by the other three assays. The CLART assay detected additional genotypes compared to the other assays, but also failed to repeat some results. The additional genotypes detected by Anyplex II HPV28 were also included in the CLART assay repertoire, and the failure of detection might be due to sensitivity differences between the assays. Despite the fact that the CLART assay is validated for FFPE material, the 450-bp amplicon may sometimes be difficult to amplify because of degradation of DNA. Anyplex II HPV28 also uses 10 more cycles in the PCR amplification, compared to CLART. The MGP5þ/6þ primer system was surprisingly concordant with the reference protocol. However, it did not manage to distinguish between coinfecting genotypes. This method could be improved by a type-specific sequencing, which might also validate the uncommon genotypes that were found. However, considering all genotyping results, the more extensive genotyping assays (Anyplex II HPV28 and CLART) changed the status of the sample (individually or together) from negative with respect to the reference protocol to a positive for Group 1 (n Z 4) and Group 2 (n Z 6) genotypes. The Group 1 genotypes should ideally have been detected by the other assays as well, but may have been unidentified because of a lower sensitivity compared to Anyplex II HPV28 and the CLART. Anyplex II HPV28 found many genotypes at 50 cycles, but how to clinically interpret such a result (ie, clinical sensitivity) is, however, not clear. One limitation of this study is also that samples were analyzed one time only with each assay, because of material limitations. Repeated results could strengthen the relevance of these results. The observed differences in the spectrum of detected genotypes within individual samples
300
could possibly also be due to the fact that assays were not run in parallel. Initially, samples were extracted for DNA, and the reference protocol was performed. Sample DNA was kept at 20 C until reanalyzed with the other assays. Study time spanned over a 1-yr period and could perhaps explain the loss of genotypes as a result of the freezeethaw process. The value of the more uncommon results of the Group 2 and the low-risk genotypes is not easily interpreted. Thus, from a research perspective, it is naturally valuable to increase our knowledge on coinfecting and uncommon genotypes. In the longer perspective, an enhanced genotyping will also provide more knowledge on the genotypes present and viral characteristics can be further validated in terms of variant, load, integration, and methylation. These data may add further knowledge to why certain infections persist, and lead to prevention strategies. Also, detailed genotyping information is important when evaluating HPV-related disease in populations where other concomitant infections exist. HIV has been associated with an increased risk of developing HPV-induced cancers,17 and the distribution and amount of HPV genotypes are also different in these areas.18,19 In conclusion, for an extended genotyping approach with a high sensitivity for FFPE specimens, Anyplex II HPV28, as well as CLART HPV2, is a suitable alternative despite minor intra-assay differences. In the clinical context, both methods find more Group 1 and 2 HPV-positive cases and could thereby contribute to personalized treatment.
Supplemental Data Supplemental material for this article can be found at http://dx.doi.org/10.1016/j.jmoldx.2014.12.004.
References 1. Poljak M, Cuzick J, Kocjan BJ, Iftner T, Dillner J, Arbyn M: Nucleic acid tests for the detection of alpha human papillomaviruses. Vaccine 2012, 30(Suppl 5):F100eF106 2. Larsson GL, Helenius G, Andersson S, Elgh F, Sorbe B, Karlsson MG: Human papillomavirus (HPV) and HPV 16-variant distribution in vulvar squamous cell carcinoma in Sweden. Int J Gynecol Cancer 2012, 22:1413e1419 3. Larsson GL, Helenius G, Andersson S, Sorbe B, Karlsson MG: Prognostic impact of human papilloma virus (HPV) genotyping and HPV-16 subtyping in vaginal carcinoma. Gynecol Oncol 2013, 129: 406e411 4. Hall JS, Iype R, Armenoult LS, Taylor J, Miller CJ, Davidson S, de Sanjose S, Bosch X, Stern PL, West CM: Poor prognosis associated with human papillomavirus alpha7 genotypes in cervical carcinoma cannot be explained by intrinsic radiosensitivity. Int J Radiat Oncol Biol Phys 2013, 85:e223ee229 5. Wang CC, Lai CH, Huang YT, Chao A, Chou HH, Hong JH: HPV genotypes predict survival benefits from concurrent chemotherapy and radiation therapy in advanced squamous cell carcinoma of the cervix. Int J Radiat Oncol Biol Phys 2012, 84:e499ee506 6. Kofler B, Laban S, Busch CJ, Lorincz B, Knecht R: New treatment strategies for HPV-positive head and neck cancer. Eur Arch Otorhinolaryngol 2014, 271:1861e1867
jmd.amjpathol.org
-
The Journal of Molecular Diagnostics
HPV Genotyping in FFPE Tissue 7. Bruni L, Barrionuevo-Rosas L, Albero G, Aldea M, Serrano B, Valencia S, Brotons M, Mena M, Cosano R, Muñoz J, Bosch FX, de Sanjosé S, Castellsagué X: Human Papillomavirus and Related Diseases in the World. Barcelona, Spain, ICO Information Centre on HPV and Cancer (HPV Information Centre), 2014 8. Schiffman M, Clifford G, Buonaguro FM: Classification of weakly carcinogenic human papillomavirus types: addressing the limits of epidemiology at the borderline. Infect Agent Cancer 2009, 4:8 9. Söderlund-Strand A, Carlson J, Dillner J: Modified general primer PCR system for sensitive detection of multiple types of oncogenic human papillomavirus. J Clin Microbiol 2009, 47:541e546 10. Bzhalava D, Guan P, Franceschi S, Dillner J, Clifford G: A systematic review of the prevalence of mucosal and cutaneous human papillomavirus types. Virology 2013, 445:224e231 11. Ludyga N, Grunwald B, Azimzadeh O, Englert S, Hofler H, Tapio S, Aubele M: Nucleic acids from long-term preserved FFPE tissues are suitable for downstream analyses. Virchows Arch 2012, 460:131e140 12. Srinivasan M, Sedmak D, Jewell S: Effect of fixatives and tissue processing on the content and integrity of nucleic acids. Am J Pathol 2002, 161:1961e1971 13. Fairley JA, Gilmour K, Walsh K: Making the most of pathological specimens: molecular diagnosis in formalin-fixed, paraffin embedded tissue. Curr Drug Targets 2012, 13:1475e1487
The Journal of Molecular Diagnostics
-
jmd.amjpathol.org
14. Lindh M, Gorander S, Andersson E, Horal P, Mattsby-Balzer I, Ryd W: Real-time Taqman PCR targeting 14 human papilloma virus types. J Clin Virol 2007, 40:321e324 15. de Roda Husman AM, Walboomers JM, van den Brule AJ, Meijer CJ, Snijders PJ: The use of general primers GP5 and GP6 elongated at their 3’ ends with adjacent highly conserved sequences improves human papillomavirus detection by PCR. J Gen Virol 1995, 76(Pt 4): 1057e1062 16. McGinn T, Wyer PC, Newman TB, Keitz S, Leipzig R, For GG: Tips for learners of evidence-based medicine: 3. Measures of observer variability (kappa statistic). CMAJ 2004, 171:1369e1373 17. Denny LA, Franceschi S, de Sanjose S, Heard I, Moscicki AB, Palefsky J: Human papillomavirus, human immunodeficiency virus and immunosuppression. Vaccine 2012, 30(Suppl 5):F168eF174 18. Denny L, Adewole I, Anorlu R, Dreyer G, Moodley M, Smith T, Snyman L, Wiredu E, Molijn A, Quint W, Ramakrishnan G, Schmidt J: Human papillomavirus prevalence and type distribution in invasive cervical cancer in sub-Saharan Africa. Int J Cancer 2014, 134:1389e1398 19. Ndiaye C, Alemany L, Ndiaye N, Kamate B, Diop Y, Odida M, Banjo K, Tous S, Klaustermeier JE, Clavero O, Castellsague X, Bosch FX, Trottier H, de Sanjose S: Human papillomavirus distribution in invasive cervical carcinoma in sub-Saharan Africa: could HIV explain the differences? Trop Med Int Health 2012, 17:1432e1440
301