A

BQ

ARRAA

KHGED

1

m(cbm(cbtep

oaaETaceet

0d

Journal of Virological Methods 179 (2012) 142– 147

Contents lists available at SciVerse ScienceDirect

Journal of Virological Methods

j ourna l ho me p ag e: www.elsev ier .com/ locate / jv i romet

hybrid-capture assay to detect HPV mRNA ratios in cervical specimens

rian Lowe ∗, Anna Fulbright, Irina NazarenkoIAGEN Gaithersburg, Inc., 1201 Clopper Road, Gaithersburg, MD 20878, United States

rticle history:eceived 31 March 2011eceived in revised form 4 October 2011ccepted 13 October 2011vailable online 20 October 2011

eywords:

a b s t r a c t

Human papillomavirus (HPV) DNA screening benefits cervical cancer diagnosis, but a few HPV infec-tions result in cancer. Assays that predict cancer are desirable. A potential biomarker is the ratio of HPVE6–7 over E2 transcripts, which may increase during early cancer progression. Modified hybrid-capturetechnology detected, in separate wells, HPV E6–7 or E2 mRNA of HPV 16 or HPV 18 in samples. Thelimit of detection was approximately 1000 copies of in vitro transcribed RNA with linear dynamic rangeapproximately four logs. No cross-reactivity between HPV 16 and HPV 18 mRNAs was detected. RNA of

uman papillomavirusene expression6–7:E2 ratioetection

SiHa cells was stable in clinical specimen pools for 67 days, as determined by RT-PCR. The ratio of HPVE6–7:E2 mRNAs was relatively high for cancer cells lines, SiHa, Caski and HeLa cells preserved in clinicalspecimen pools. A broad distribution of HPV 16 E6–7:E2 mRNA ratio was detected in a small set of clinicalspecimens with various histological diagnoses. Some specimen ratios were so high for cancer cell lines,but the significance of the results needs to be determined. This method may help determine the patternof gene expression in HPV-related disease or in other systems.

. Introduction

There are 13 well-known types of oncogenic Human papillo-avirus (HPV) that cause HPV-related cervical disease and cancer

Bosch et al., 2002). Disease stages are classified according toytological and histological methods with cervical scrapes andiopsies (Broder, 1992). Screening for oncogenic HPV DNA usingolecular tests has been useful to diagnose HPV-related disease

Sankaranarayanan et al., 2009). These current testing methodsannot precisely predict which infections may develop into cancerecause most HPV infections are transient, and clear by spon-aneous regression. Therefore, additional biomarkers are beingxplored for use in reflex assays to confirm which infections willrogress to cancer and require further treatment.

The progression of disease may be related to the expressionf certain HPV genes (Doorbar, 2007). Thus, HPV mRNA may ben additional biomarker for severe infections. Some HPV mRNAssays detect a single type of transcript species, such as the E6 or7 oncogenic sequences (Castle et al., 2007; Molden et al., 2007).hese assays may not predict only severe infections because thebundance of a single target sequence may fluctuate due to theomplex pattern of expression that occurs during the course of dis-

ase, or due to degradation of HPV from immune responses (Pettt al., 2006). In addition, mRNAs may degrade after collection, orhe number of infected cells in the collected specimen may be low,

∗ Corresponding author. Tel.: +1 301 944 7339; fax: +1 301 944 7301.E-mail address: brian.lowe@qiagen.com (B. Lowe).

166-0934/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.jviromet.2011.10.012

© 2011 Elsevier B.V. All rights reserved.

both of which may affect the assay result. As a solution, HPV assaysdesigned to detect simultaneously two target sequences of mRNAsin a ratio may be more predictive of disease than assays that detecta single mRNA species.

A ratio assay may reflect the natural pattern of HPV mRNAexpression that occurs during the progression of disease in cer-vical epithelium. After initial infection, polycistronic, pre-mRNAis transcribed from HPV DNA that exists either as circular, extra-chromosomal episomes or as DNA that is integrated into thechromosome. As pre-malignant lesions progress, the abundance ofmRNAs that encode the oncogenes, such as E6 and E7 (E6–7), mayincrease and the mRNAs encoding non-oncogenic HPV proteins,such as E2 and L1, may decrease (Doorbar, 2007). The incidence ofHPV integration may reduce further transcripts encoding E2 andother downstream genes, such as L1, because integration usuallyoccurs at the E2 loci (Vinokurova et al., 2008). The E2 gene prod-uct is an important down-regulator of oncogenic E6–7 expression(Smith et al., 2010). Thus, lower E2 levels may correlate with dis-ease progression. Therefore, it may be useful to measure the ratioof E6–7 over E2 transcripts in cervical specimens and compare thisratio with the severity of disease.

In this preliminary research, HPV mRNAs that encode for E6–7,or E2, were detected using a hybrid-capture method. The mRNAtargets were isolated and hybridized to specific, short, syntheticDNA probes, and then measured by a detection system involv-

ing a RNA:DNA hybrid-capture antibody and a luminescence-basedsignal amplification (Lorincz and Anthony, 2002). The signal inten-sity corresponds linearly to the abundance of target mRNA, and tothe length of the formed RNA:DNA hybrid. A two-well assay was

ogical Methods 179 (2012) 142– 147 143

daopStiHlbu

2

2

HclbhocwwHNc

2

ncNaltwr

2

vmc6i(ncbcibh

2

cctotw

Fig. 1. Schematic diagram for the hybrid-capture assay to detect HPV mRNA. (A)General scheme for the hybrid-capture detection of HPV mRNA. HPV mRNA target(dotted line) is hybridized to capture oligodeoxyribonucleotides (short grey bars)that are coupled to a magnetic bead (circle). The RNA target is hybridized withdetection oligodeoxyribonucleotides (short black bars) to create a longer hybrid. TheRNA:DNA hybrid is bound with a hybrid-capture antibody conjugated with alkalinephosphatase (Y-shaped symbols with open circles). A chemiluminescent substrate(not shown) is added to detect the complex in a luminometer. (B) Schematic showingthe HPV genome structure with labeled genes (large grey arrows). The loci for E6–7probes (1) or E2 probes (2) are shown by black bars underneath. The arrangement

B. Lowe et al. / Journal of Virol

eveloped to measure transcripts for E6–7 in one well or E2 in second well. A high ratio indicates an increase in abundance ofncogenic E6–7 mRNA over non-oncogenic E2 mRNA. Model sam-les included in vitro transcribed mRNAs, and mRNA from culturediHa, Caski and Hela cells, which are derived from cervical cancerumors and contain HPV. HPV mRNA ratios were detected in a lim-ted number of clinical specimens. Assays were designed for twoPV types, HPV 16 and HPV 18, because they are the most preva-

ent, oncogenic HPVs, and because their pattern of expression isest known. The described assays are not available for diagnosticse.

. Materials and methods

.1. Samples and specimens

Cell lines of SiHa (HTB-35), Caski (CRL-1550), HeLa (CCL-2) andCC 1806 (CRL-2335) were obtained from ATCC (Manassas, VA) andultured by standard techniques. Residual cervical specimens iniquid-based cytology (LBC) medium (PreservCyt®, Hologics, Marl-orough, MA; 20 ml original volume) were obtained, some withistological diagnoses, after routine testing from Cytology Servicesf Maryland. Routine testing included the standard practice ofytology, with follow-up colposcopy and biopsy. Specimen poolsere composed of several of these specimens. These specimensere 5–8 months old and stored at room temperature before use.PV genotyping of some clinical specimens was done according toazarenko et al. (2008) to confirm single HPV 16 infection or toonfirm the lack of HPV DNA.

.1.1. RNA target isolationThe in vitro transcribed HPV 16 and HPV 18 RNAs for E6 (1–790

t) and E2 (2755–3852 nt) regions were prepared with standardloning techniques using HPV 16 or HPV 18 as a template (GenBankC 01526 and X05015). RNA was prepared from samples, cell linesnd specimens using either the Rneasy® Plus Mini Kit, or QIAzolysis reagent (QIAGEN Inc., Valencia, CA). For QIAzol RNA isolation,he cells preserved in LBC were isolated by centrifugation, extractedith organic solvent and the alcohol-precipitated RNA was then

esuspended in Tris-buffer (pH 7).

.1.2. Cell concentrationCells from model samples of cultured SiHa cells added to cer-

ical specimen pools that tested negative for HPV (1 ml, LBCedia) were concentrated in microfuge tubes by adsorption onto

arboxyl-modified magnetic beads (8 �l of 5% solids; catalog #5162105050350, Seradyn). This specimen-bead suspension was

ncubated at 22 ◦C for 15–30 min in a rotating microfuge block1100 rpm, Eppendorf). The cells on beads were pelleted by a mag-etic tube holder (Promega, Madison, WI). The percent of cellsollected from mixtures of known cell number was determinedy counting the cells in the leftover supernatant using a hemo-ytometer. The cells were washed with saline and then incubatedn 50–60 �l of lysis buffer for 10 min and the magnetic-carboxyl-eads were removed by magnet. The lysate was then input to theybrid-capture assay.

.2. Oligodeoxyribonucleotides

The oligodeoxyribonucleotide probes were designed to be spe-ific for HPV 16 (or 18) mRNA targets by using Blast (NCBI)omparisons. The design of capture probes (35 mers) was adjusted

o avoid cross-hybridization with other HPV types. The detectionligodeoxyribonucleotides were 33–40 mers and were complimen-ary to their targets, but not designed to avoid cross-reactivityith other HPV types. The reverse-transcription, PCR primers and

of genes and the loci for DNA probes are similar for HPV 16 and HPV 18, but theprimary sequences are unique.

TaqMan probes were designed by PrimerQuest (IDT, Coralville, IA)and Beacon Designer (Palo Alto, CA). All oligodeoxyribonucleotideswere synthesized by IDT (Coralville, IA).

2.3. Real-time, reverse-transcription and PCR (RT-PCR)

One-step RT-PCR was performed using the QuantiTect® 5× VirusMix (no ROX; QIAGEN Inc., Valencia, CA), according to vendor pro-tocol. Primer and probe sets were designed for the E6–7 region andthe E2 region using software. Real-time, RT-PCR was performedusing either a Stratagene MX3000P (Stratagene, La Jolla, CA) orBio-Rad iQTM5 (Bio-Rad, Hercules, CA) real-time PCR instrument.RT-PCR volumes were 25 �l.

For mRNA stability analyzed by RT-PCR, fresh SiHa cells werepreserved in pooled LBC cervical specimens that previously con-tained no HPV as indicated by consensus HPV PCR (Nazarenkoet al., 2008). These samples were incubated at room temperaturefor up to 67 days. Two 1-ml aliquots were removed periodically(3, 13, 26, 42 and 67 days) and the RNA was isolated by QIAzolreagent. The HPV mRNA level was determined using a real-time,RT-PCR for E6–7 (5′–3′; forward primer GCACCAAAAGAGAACTG-CAATGT, reverse primer CATATACCTCACGTCGCAGTAACT, TaqManprobe FAM-CAGGACCCACAGGAGCGACCCAGA-BHQ1). Each reac-tion contained the mRNA from approximately 125,000 SiHa cells(SiHa cells have two copies of integrated, HPV 16 genome with noHPV E2 region).

Consensus HPV PCR (Nazarenko et al., 2008) for important HPVtypes was used to indicate whether a cervical specimen pool con-tained HPV.

2.4. Hybrid-capture assay

A schematic diagram for the hybrid-capture assay for mRNA isshown in Fig. 1. The assay is loosely based on the digene HC2 HPVDNA Test® (QIAGEN Inc.); except mRNA (not DNA) is the target andthe probes consist of synthetic DNA (not RNA), the alkali denatu-ration step is not included, and the formed RNA:DNA hybrids arecaptured on magnetic beads instead of an ELISA plate.

The cell pellet (prepared by cell concentration using un-modified, magnetic beads or by the Qiazol procedure) or in vitro

transcribed RNA, was lysed (50–60 �l, RLT Plus lysis buffer, QIA-GEN Inc.) to release and unwind the RNA. The RNA was capturedby adding 10 �l of oligodeoxyribonucleotide-modified magneticbeads (0.01% solids, 1 �m, Seradyn). The target RNA was captured

1 ogical Methods 179 (2012) 142– 147

ob

(bsc1csmu

aD(bPttotHibttF

uwbmpwbaml

3

3

Lcr1ab

3d

tE1ilod1b

Table 1DNA probe sequences for HPV 16 and HPV 18. Capture oligodeoxyribonucleotideswere modified with a 5′ biotin.

Target 5′–3′ capture sequences

HPV 16 E6–7 GTTTGCAGCTCTGTGCATAACTGTGGTAACTTTCTHPV 16 E6–7 CAGTAACTGTTGCTTGCAGTACACACATTCTAATAHPV 16 E6–7 ACATATATTCATGCAATGTAGGTGTATCTCCATGCHPV 16 E6–7 AAGGTTACAATATTGTAATGGGCTCTGTCCGGTTCHPV 16 E6–7 ATTAACAGGTCTTCCAAAGTACGAATGTCTACGTGHPV 16 E2 CAATAGTCTATATGGTCACGTAGGTCTGTACTATCHPV 16 E2 CAAGGCTAACGTCTTGTAATGTCCACTTTTCATTAHPV 16 E2 TATAAACCATAATAGTCAACTTGACCCTCTACCACHPV 16 E2 TTGGTCACGTTGCCATTCACTATCATATGTAAGTGHPV 16 E2 CTGATCTTGGTCGCTGGATAGTCGTCTGTGTTTCTHPV 18 E6–7 TCATAGTGGTCTATGATTTTGTCCTGCACGCAACTHPV 18 E6–7 TCCAATCCTCGGTTTTGTATCGACTTTGTGCAAGGHPV 18 E6–7 TGTGACTTACACAGGTAGCGGTTTTGTCCCATGTTHPV 18 E6–7 TGGGTTGACAGGTCCACAATGCTGCTTCTCCGCGAHPV 18 E6–7 CCACCAATATTTGTACACTATCTGGAATTGCAACAHPV 18 E2 ATACACAGGTTATTTCTATGTCTTGCAGTGAAGTGHPV 18 E2 GCACTGGCCTCTATAGTGCCCAGCTATGTTGTGAAHPV 18 E2 CATAGAAGGTCAACCGGAATTTCATTTTGGGGCTCHPV 18 E2 CGGGCTGGTAAATGTTGATGATTAACTCCATCTATHPV 18 E2 CAGGGTGTTCAGAAACAGCTGCTGGAATGCTCGAA

Target 5′–3′ detection sequences

HPV 16 E6–7 ATACTATGCATAAATCCCGAAAAGCAAAGTCATATACHPV 16 E6–7 ATTTATCACATACAGCATATGGATTCCCATCTCTATHPV 16 E6–7 GTCTATACTCACTAATTTTAGAATAAAACTTTAAACHPV 16 E6–7 GTTCTAATGTTGTTCCATACAAACTATAACAATAATHPV 16 E6–7 CTAATTAACAAATCACACAACGGTTTGTTGTATTGCTHPV 16 E6–7 CCTGTGGGTCCTGAAACATTGCAGTTCTCTTTTGGTGCATHPV 16 E6–7 TGTGCTTTGTACGCACAACCGAAGCGTAGAGTCACACTTGHPV 16 E6–7 TTATGGTTTCTGAGAACAGATGGGGCACACAATTCCTAGTHPV 16 E6–7 TTTTCTTCAGGACACAGTGGCTTTTGACAGTTAATACACHPV 16 E6–7 ATATTATGGAATCTTTGCTTTTTGTCCAGATGTCTTTGCHPV 16 E6–7 CTGCAACAAGACATACATCGACCGGTCCACCGACCCCTTHPV 16 E6–7 ATGATTACAGCTGGGTTTCTCTACGTGTTCTTGATGATHPV 16 E6–7 CTCCTCTGAGCTGTCATTTAATTGCTCATAACHPV 16 E6–7 AGTAGAGATCAGTTGTCTCTGGTTGCAAATCTAHPV 16 E6–7 TGCTTGTCCAGCTGGACCATCTATTTCATCCTCHPV 16 E2 TTTTATACATCCTGTTGGTGCAGTTAAATACACTTHPV 16 E2 CCATCAAACTGCACTTCCACTGTATATCCATGTTTHPV 16 E2 TCCAGTTTGTATAATGCATTGTATTGCATATGTCTHPV 16 E2 AGTTACTGATGCTTCTTCACAAATATATATATGTGHPV 16 E2 CTTTATTTTTACTATATTTTTCTGCATCATCTTTAAAHPV 16 E2 CATAATATTACCTGACCACCCGCATGAACTTCCCATAHPV 16 E2 AGAGGATACTTCGTTGCTGCTAAACACAGATGTAGGAHPV 16 E2 CGGGGTGGTTGGCCAAGTGCTGCCTAATAATTTCAGGHPV 16 E2 CTGCACAAAATATGTTCGTATTCCTTCATGAACATAAHPV 16 E2 TCGGTGCCCAAGGCGACGGCTTTGGTATGGGTCGCGGHPV 16 E2 CACACATTTAAACGTTGGCAAAGAGTCTCCATHPV 16 E2 ATTTTCATAATGTGTTAGTATTTTGTCCTGAHPV 16 E2 TAGTTTTTGGTATTTTAACTTGAGACAAAAAHPV 16 E2 TCATATAGACATAAATCCAGTAGACACTGTAAHPV 16 E2 TAATAAATAGCACATTCTAGGCGCATGTGTTTCHPV 16 E2 TTAATATGTTTAAATCCCATTTCTCTGGCCTTGHPV 16 E2 TTTGATACAGCCAGTGTTGGCACCACTTGGTGGHPV 16 E2 AGTTGCAGTTCAATTGCTTGTAATGCTTTATTCHPV 16 E2 CTATATTGTGAGTTATATATTGTTTCTAACGTTHPV 16 E2 TAGTGGTGTGGCAGGGGTTTCCGGTGTCTGGCTHPV 16 E2 TAACAATTGCACTTTTATGTTTTACATTATGTCHPV 16 E2 GGAGCACTGTCCACTGAGTCTCTGTGCAACAACTHPV 16 E2 TCCTTTGTGTGAGCTGTTAAATGCAGTGAGGATTHPV 16 E2 CTATGGGTGTAGTGTTACTATTACAGTTAATCCGHPV 16 E2 CATTTTAAAGTATTAGCATCACCTTTTAAATGTAHPV 16 E2 CAATGTACAATGCTTTTTAAATCTATATCTTAAAHPV 16 E2 CTGTCCAATGCCATGTAGACGACACTGCAGTATAHPV 18 E6–7 GGGTCGCCGTGTTGGATCCTCAAAGCGCGCCATHPV 18 E6–7 TTCAGTTCCGTGCACAGATCAGGTAGCTTGTAHPV 18 E6–7 TCTGTAAGTTCCAATACTGTCTTGCAATHPV 18 E6–7 CACCACAAATAAATCTTTAAATGCAAATTCAAATACCHPV 18 E6–7 ATTTATGGCATGCAGCATGGGGTATACTGTCTCTATAHPV 18 E6–7 GTCTTAATTCTCTAATTCTAGAATAAAAATCTATAC

44 B. Lowe et al. / Journal of Virol

nto these oligodeoxyribonucleotide-modified beads during incu-ation at 60 ◦C for 30 min with 1100 rpm rotation.

The streptavidin-modified, magnetic beads were pre-coupledstandard techniques) with specific, biotinylated, oligodeoxyri-onucleotides to enable capture of specific mRNA. There were five,equence-specific capture probes per bead per target. Four, hybrid-apture assays for HPV mRNAs were designed to be specific to HPV6 E6–7, HPV 16 E2, HPV 18 E6–7 or HPV 18 E2, by using spe-ific capture probes (Table 1 ). The Blast program confirmed thepecificity of the capture probes, so that no less than about fiveismatches occurred when a specific probe was aligned with any

nintended HPV target.The lysate was diluted 1:3 with water and split into two wells of

96-well microtiter plate for detection of either E6–7 or E2 mRNA.etection DNA probes (4.2 mM each, 33–45 nt) were hybridized

60 ◦C for 30 min with 1100 rpm rotation) to the target RNA in auffer composed of a 5:8 mixture of HC2 Denaturation Reagent:robe Diluent (QIAGEN Inc). There were 15 detection probes forhe E6–7 target and 27 detection probes for the E2 target. Detec-ion probes complemented the entire length of either the E6–7r E2 coding region for each captured mRNA target. These detec-ion probes were not designed to avoid cross-reactivity with otherPV types. Their function was to allow signal amplification via

ncreased hybrid length and binding of the hybrid-capture anti-ody with alkaline phosphatase. The lengths of the formed hybridargets were approximately 740 bp for the E6–7 and 1500 bp forhe E2. The probe loci for hybrid-capture probes are indicated inig. 1B.

The resulting hybrids affixed to magnetic beads were pelletedsing a magnetic plate holder (Ambion). The hybrid-bead complexas washed on the magnetic plate with a saline, detergent-based

uffer (pH 7.5). The complex was incubated (45 ◦C, 30 min) withonoclonal Hybrid Capture® antibodies conjugated to alkaline

hosphatase (DR1; QIAGEN Inc.). This complex was then washedith HC2 wash buffer (QIAGEN Inc.). The complex was then incu-

ated (22 ◦C, 15 min, rotation 300 rpm) with a chemiluminescent,lkaline phosphatase substrate (DR2, QIAGEN Inc.). The signal waseasured in relative luminescence units (RLU) using a DML 2000

uminometer (QIAGEN Inc.).

. Results

.1. Stability of HPV mRNA

The stability of the HPV mRNA in SiHa cells that were fixed inBC medium was determined using a real-time, RT-PCR assay. Theycle threshold, a measure of mRNA abundance, of the RT-PCRs waselatively stable up to 42 days and then shifted by approximately–2 cycles for the 42- and 67-day aliquots (Fig. 2). This shift mayccount for a reduction in target mRNA of approximately three-foldased on theoretical PCR kinetics.

.2. Analytical performance of the hybrid-capture mRNAetection assay

The hybrid-capture assay was first performed for pure, in vitroranscribed HPV RNA targets for HPV 16 E6–7, HPV 16 E2, HPV 186–7 and HPV 18 E2, all in separate wells. The results for the HPV6 E6–7 assay are shown in Fig. 3A. The assay detected approx-

mately 1000 copies, or more, of RNA per reaction. There was ainear dependence of signal on target input with a dynamic range

f 3–4 logs. A similar dependence of signal on target input wasetected for the other three RNA targets including HPV 16 E2, HPV8 E6–7, and HPV 18 E2 transcripts (not shown). No signal aboveackground resulted when an HPV 18 RNA target was assayed with

HPV 18 E6–7 TTTTCCAATGTGTCTCCATACACAGAGTCTGAATAATHPV 18 E6–7 CCTTATTAATAAATTGTATAACCCAGTGTTAGTTAGTHPV 18 E6–7 TGCTGGATTCAACGGTTTCTGGCACCGCAGGCAHPV 18 E6–7 ATCGTCGTTTTTCATTAAGGTGTCTAAGTTTTTC

B. Lowe et al. / Journal of Virological Methods 179 (2012) 142– 147 145

Table 1 (Continued )

Target 5′–3′ detection sequences

HPV 18 E6–7 TTGGAGTCGTTCCTGTCGTGCTCGGTTGCAGCACGAATGHPV 18 E6–7 ATGCATACTTAATATTATACTTGTGTTTCTCTGCGTCGHPV 18 E6–7 TAAATGCAATACAATGTCTTGCAATGTTGCCTTAGGTCCHPV 18 E6–7 TTCATCGTTTTCTTCCTCTGAGTCGCTTAATTGCTCGTGAHPV 18 E6–7 CAACATTGTGTGACGTTGTGGTTCGGCTCGTHPV 18 E6–7 AATTCTGGCTTCACACTTACAACACATACAHPV 18 E6–7 GGTCGTCTGCTGAGCTTTCTACTACTAGCTCHPV 18 E6–7 TTACTGCTGGGATGCACACCACGGACACACAAAGGAHPV 18 E2 TAAACGTTCCGAAAGGGTTTCCTTCGGTGTCTGCATHPV 18 E2 ATACTGTATTTGGCTGTCTATGTCTTTACTGTCATTTHPV 18 E2 AAAGAATATTGCATTTTCCCAACGTATTAGTTGCCAHPV 18 E2 GGTGGTTTAATGTCTGTATGCCATGTTCCCTTGCTGCHPV 18 E2 CTTTACTTTTTGAAATGTTATAGGCTGGCACCACCTHPV 18 E2 CCTTGTAGGGCCATTTGCAGTTCAATAGCTTTATGTGHPV 18 E2 GTTCTGTATTCCATAGTTCCTCGCATGTGTCTTGCAGTGHPV 18 E2 TTGTACTGTTTGGCCACCTTTTTTAAAGCAGTGAGTAGHPV 18 E2 TAGGTCATACAATTGTCTTTGTTGCCATCAAAATATACHPV 18 E2 CCTGCATCAGTCATATAATACACACTGTCCCATGCTACAHPV 18 E2 ACGTGTTGTACCCTTCCTTTACATAATACAATCCCCHPV 18 E2 ATATTTTTCACATTCACTTTTAAATTCTATATAAAHPV 18 E2 TTCCCAAAATGTACTTCCCACGTACCTGTGTTCCCHPV 18 E2 TACTGCACATAGAGTCATTACAATCAATTACATTAHPV 18 E2 AACAAGCTGAGTAGCGGATACCGTGTCGTCACTGGHPV 18 E2 CTGGAATACGGTGAGGGGGTGTGCTGTAGCTGTTTHPV 18 E2 GGCCGTAGGTCTTTGCGGTGCCCACGGACACGGTGHPV 18 E2 GTCCACAGTGTCCAGGTCGTGTAGCAGCCGACGTCTHPV 18 E2 TTTGTTGTTGCCTGTAGGTGTAGCTGCACCGAGAAGHPV 18 E2 TATAGGCGTAGTGTTACCACTACAGAGTTTCCGTCTHPV 18 E2 ACATTTTAAACTGTTTCTGTCACCTTTTAAATGTATHPV 18 E2 TAGTGGTCGCTATGTTTTCGCAATCTGTACCGTAAHPV 18 E2 GCACCTGTCCAATGCCAGGTGGATGATATATCTCTA

tonsgs

dtscHesT

FT(ad

Fig. 3. Analytical performance of the hybrid-capture assay for HPV 16 E6–7 RNA.(A) Dependence of luminescence signal output (average RLU, n = 3 samples, errorbars show standard deviation) on target input (RNA copies per reaction) for ahybrid-capture assay. Signal:noise ratio is given above bars. (B) Dependence ofluminescence signal output (average RLU, n = 4) on the number of complemen-tary detection probes per assay for the same target input (1 × 105 copies, HPV 16E6–7 in vitro transcribed RNA). In this experiment, the hybrid length increased in

HPV 18 E2 TATGTTACAGTCAGTATTCCTGTTTTTTCATTGCCTHPV 18 E2 GTATTTAAAAATTTTGTTCTTTGTGTTTCACTATGG

he HPV 16 specific capture and detection probes, or visa versa. Nother cross-reactivity experiments were conducted. Although theumber of capture probes used was five, three capture probes wereufficient for equivalent detection of 1000 copies. The use of a sin-le capture probe resulted in attenuated signal for 1000 copies (nothown).

In addition to the abundance of RNA molecules, the assay signalepended on length of formed RNA:DNA hybrid. To demonstratehis, equivalent amounts of HPV 16 E6–7 RNA were captured ineparate wells using magnetic beads coupled with the HPV 16 E6–7apture oligodeoxyribonucleotides. An increasing number of the

PV 16 detection probes were added to each separate well. Thus,ach well had RNA:DNA hybrids of successively longer length, untilome wells contained completely hybridized RNA targets (Fig. 3B).he signals for the wells increased in correspondence to the number

ig. 2. Stability of HPV RNA in SiHa cells preserved in a LBC clinical specimen pool.he RT-PCR plots show the assay signal (y-axis) plotted against PCR cycle numberx-axis) for samples of SiHa cells incubated over the course of 67 days. Symbolsre star, 3 days; square, 13 days; triangle, 26 days; filled diamond, 42 days; openiamond, 67 days. Values are an average of two replicates for each day.

wells with the addition of 5, 10 and 15 probes. The sample labeled as 19 + 5 (x-axis) contained 5 non-complementary probes in addition to the 15 complementaryprobes.

of detection probes until a plateau was reached at the well in whichthe target was completely hybridized (15 detection probes added).The further addition of five, non-complementary probes did notincrease the assay signal.

3.3. Detection of E6–7:E2 ratios in cancer cell lines

HPV 16 mRNA was detected for SiHa cells preserved in a poolof LBC clinical specimens that previously did not contain HPV. Thecell concentration procedure using magnetic, carboxyl-beads wasapplied to pellet the cells. Ninety-five percent of the cells werepelleted in 30 min using this procedure, as determined by cellcounting. The resulting cell pellets were lysed and the lysate vol-ume was divided equally into two wells. HPV 16 E6–7 transcriptswere assayed in one well and HPV 16 E2 were assayed in a secondwell. The SiHa mixture expressed abundant E6–7 transcripts, butnot E2 (Fig. 4A). These assays for HPV 16 detected only a negligiblesignal when HPV 18 mRNA of HeLa cells (1 × 106 cells) was used asa target (signal:noise < 2; not shown). Cross-reactivity with otherHPV types was not tested. The ratio of HPV 16 E6–7 and E2 in SiHacells was calculated from this data. The maximum signal for the E2hybrid may be proportionally greater than for E6–7 hybrid due toits increased length. For this reason, the E2 signal was multipliedby a compensating factor of 0.51 when calculating E6–7:E2 ratiosfor cells and specimens. The HPV 16 E6–7:E2 ratio was 8.2 or higherfor SiHa cells depending upon the number of cells in the assay. Thismethod was used to calculate the HPV E6–7:E2 ratio for other can-

cer cell lines that express HPV transcripts. These include Caski andHeLa, which express HPV 16 and HPV 18, respectively (Fig. 4B). Theratio for the SiHa and HeLa cell lines was relatively higher than forCaski cells.

146 B. Lowe et al. / Journal of Virological Methods 179 (2012) 142– 147

Fig. 4. Detection of HPV E6–7 and E2 mRNA in cancer cells lines. (A) The dependenceof signal: noise (average, n = 3) on number of SiHa cells per assay was plotted for theHPV 16 E6–7 (black bars) and E2 (grey bars); for the two assays in separate wells. Forthese assays, the background noise was obtained from the signal of a control assaywmb

rceti1wtlatwta

Fww1

ith no target added, approximately 50 RLU. (B) The values for HPV E6–7 and E2RNA assays were plotted as a ratio for the cancer cell lines, SiHa, Caski and HeLa;

ars represent the average ratios of three replicate experiments.

Experiments were performed to determine the HPV E6–7:E2atios in heterogeneous mixtures of cancer cells and non-cancerells that both express HPV E6–7 and E2 transcripts in un-qual ratios. The SiHa cells, which have a relatively high E6–7:E2ranscript ratio, were added and preserved in a pool of clinical spec-mens (LBC medium) that was positive for only HPV 16. The HPV6 E6–7:E2 ratio of the pooled specimens was approximately one,ith no added SiHa cells. Serial dilutions of SiHa cells were added

o 2 ml aliquots of this specimen pool. The sample RNA was iso-ated by QIAzol extraction. The results of the HPV E6–7 and E2ssays were expressed as a ratio (Fig. 5). The addition of SiHa cells

o the HPV-positive pool resulted in an increased E6–7:E2 ratio,ith a substantial increase of approximately 4.2-fold upon addi-

ion of 100,000 SiHa cells. In comparison, the ratio for SiHa cellslone (33,000 cells) was about nine (Fig. 4B).

ig. 5. Detection of the HPV 16 E6–7:E2 transcript ratio in a mixture of SiHa cellsith the cells from a pool of HPV-positive specimens. Cultured SiHa cells were mixedith a pool (2 ml) of HPV-positive, liquid-based cytology specimens (approximately

00,000 total cells in 2 ml).

Fig. 6. The HPV 16 E6–7:E2 ratio in cervical specimens. The E6–7:E2 ratio wasplotted from the hybrid-capture assay results (see text).

3.4. Ratios in clinical specimens

The HPV 16 E6–7:E2 mRNA ratio was determined also in alimited number (n = 13) of selected cervical specimens using thehybrid-capture mRNA assay for HPV 16 E6–7 and E2. These speci-mens were pre-selected from those which had known histologicaldiagnoses, and were confirmed by PCR to include only HPV 16. Addi-tional, fresh, HPV 16 infected specimens and specimens with onlyHPV 18 infection were not available. The specimen RNA was iso-lated by QIAzol extraction. There was a broad distribution of ratiosfor all histological grades, but two specimens each for CINI andCINIII had a relatively high ratio above two, with the maximumratio of 3.8 for one CINIII specimen (Fig. 6). The mean ratios were1.7, 1.1 and 1.7 for CINI, CINII, and CINIII, respectively.

4. Discussion

This hybrid-capture assay detected in vitro transcribed RNA withgood linearity and dynamic range of approximately 3–4 logs. Thisanalytical performance is similar to that of hybrid-capture detec-tion of DNA (Cullen et al., 1997). There was no cross-reactivitybetween the HPV 16 and HPV 18 mRNA due to the specificity of thecapture oligodeoxyribonucleotides. The cross-reactivity of all thevarious HPV types was not tested. This is a limitation of the assaythat requires further research. Detection of E6–7 or E2 mRNA fromeither HPV 16 or HPV 18 was demonstrated by assays in separatewells. HPV E6–7:E2 ratios may be calculated from these separateassays. The use of short DNA probes for target capture and detectionallow flexibility for assay design with various targets. For example,the assays may be designed to detect any single HPV type (typing)in a single well, or to detect simultaneously multiple, specific HPVsequences of various types (screening), or to detect an mRNA ratioof any number of type such that there are two wells per type.

The usefulness of the hybrid capture to assess to signal thelength of an mRNA sequence was particularly interesting (Fig. 3B).The relative length of a transcript or the existence of a trunca-tion or deletion may be measured in a linear fashion. For example,HPV transcript structures in various HPV types may be character-ized to determine whether a specific transcript, such as E2, hasbeen spliced or truncated. Capture probes designed across splicesor within introns may be able to detect the presence or absence ofmRNA splice forms.

A relatively high HPV E6–7:E2 ratio of 8, or greater, was deter-mined for SiHa and HeLa cervical cancer cells cancer cells lines

with a slightly lower ratio for Caski cells. This is consistent withthe known expression pattern for these cells (Baker et al., 1987). Inthese cell lines, the E6–7 gene products are oncogenic and the E2down-regulates E6–7 transcription. Upon HPV integration into the

ogical

ces

tctaascF

apEmkifuldc

fihbotmawnfsd

A

M

B. Lowe et al. / Journal of Virol

hromosome, the E2 gene is disrupted and is not expressed (Boshartt al., 1984). These cellular events may increase the E6–7:E2 tran-cript ratio, which may be associated with disease progression.

These cellular processes may cause ratios in clinical specimenso be elevated. Cell lines, however, may not precisely reflect a pre-ancer situation, so a limited number of clinical specimens wereested. Some cervical specimens had relatively high E6–7:E2 ratios,bove two, but these were not dependent on the grade of lesions might be expected. Reasons for this result include that cervicalcrape samples may comprise heterogeneous mixtures of superfi-ial cells with either high or low E6–7:E2 ratios, as addressed inig. 5.

Another reason for lack of strong correlation between high rationd severe lesions may be that some assay interference occurs fromolycistronic, pre-mRNA sequences that are complimentary to both6–7 and E2 capture probes. HPV has an assortment of spliced andulti-gene transcripts for which the pattern of expression is not

nown completely (Zheng and Baker, 2006). Finally, the possibil-ty remains in which the specimens with a high ratio were thoseor which disease was progressing. The relatively few specimenssed in this study limit the discussion of results and a prospective,

ongitudinal study that tracks the patient outcomes is required toetermine whether the E6–7:E2 ratio is an important diagnostic forancer in cervical scrapes.

Real-time, reverse-transcriptase and PCR and other RNA ampli-cation assays are widely used for mRNA detection because of theirigh sensitivity. The hybrid-capture assay is less sensitive, but maye less prone to false negatives due to target degradation by virtuef having multiple probes along an extended target; which may bet-er tolerate mutations and breaks in the mRNA target. Compared to

ultiplex PCR, the hybrid-capture assay may more easily capturend detect multiple types of sequence-specific mRNAs in a singleell. In addition, hybrid-capture assays are not prone to contami-ation with amplicons. For these reasons, the hybrid-capture assay

or detection of HPV mRNAs may help determine the role of expres-ion in HPV-related disease, which may have ultimate use for canceriagnostics. These assays are not available for diagnostic use.

cknowledgements

The authors wish to thank former QIAGEN staff members Nicolealandro and Angela Brand for their technical assistance.

Methods 179 (2012) 142– 147 147

References

Baker, C.C., Phelps, W.C., Lindgren, V., Braun, M.J., Gonda, M.A., Howley,P.M., 1987. Structural and transcriptional analysis of human papillo-mavirus type 16 sequences in cervical carcinoma cell lines. J. Virol. 61,962–971.

Bosch, F.X., Lorincz, A., Munoz, N., Meijer, C.J.L.M., Shah, K.V., 2002. The causal rela-tion between human papillomavirus and cervical cancer. J. Clin. Pathol. 55,244–265.

Boshart, M., Gissmann, L., Ikenberg, H., Kleinheinz, A., Scheurlen, W., zur Hausen,H., 1984. A new type of papillomavirus DNA, its presence in genital can-cer biopsies and in cell lines derived from cervical cancer. EMBO J. 3,1151–1157.

Broder, S., 1992. The Bethesda System for reporting cervical/vaginal cytologic diag-noses – report of the 1991 Bethesda Workshop. JAMA 267, 1892.

Castle, P.E., Dockter, J., Giachetti, C., Garcia, F.A.R., McCormick, M.K., Mitchell, A.L.,Holladay, E.B., Kolk, D.P., 2007. A cross-sectional study of a prototype car-cinogenic human papillomavirus E6/E7 messenger RNA assay for detection ofcervical precancer and cancer. Clin. Cancer Res. 13, 2599–2605.

Cullen, A.P., Long, C.D., Lorincz, A.T., 1997. Rapid detection and typing of herpes sim-plex virus DNA in clinical specimens by the hybrid capture II signal amplificationprobe test. J. Clin. Microbiol. 35, 2275–2278.

Doorbar, J., 2007. Papillomavirus life cycle organization and biomarker selection.Dis. Markers 23, 297–313.

Lorincz, A., Anthony, J., 2002. A system for nucleic acid detection by signalamplification technology. In: Van Dyke, K., Van Dyke, C., Woodfork, K. (Eds.),Luminescence Biotechnology: Instruments and Applications. CRC Press, BocaRaton, FL, pp. 149–158.

Molden, T., Kraus, I., Skomedal, H., Nordstrom, T., Karlsen, F., 2007. PreTectTM

HPV-Proofer: real-time detection and typing of E6/E7 mRNA from carcinogenichuman papillomaviruses. J. Virol. Methods 142, 204–212.

Nazarenko, I., Kobayashi, L., Giles, J., Fishman, C., Chen, G., Lorincz, A., 2008. A novelmethod of HPV genotyping using hybrid capture sample preparation methodcombined with GP5+/6+ PCR and multiplex detection on Luminex XMAP. J. Virol.Methods 154, 76–81.

Pett, M.R., Herdman, M.T., Palmer, R.D., Yeo, G.S.H., Shivji, M.K., Stanley, M.A.,Coleman, N., 2006. Selection of cervical keratinocytes containing integratedHPV16 associates with episome loss and an endogenous antiviral response. Proc.Natl. Acad. Sci. 103, 3822–3827.

Sankaranarayanan, R., Nene, B.M., Shastri, S.S., Jayant, K., Muwonge, R., Budukh,A.M., Hingmire, S., Malvi, S.G., Thorat, R., Kothari, A., Chinoy, R., Kelkar, R.,Kane, S., Desai, S., Keskar, V.R., Rajeshwarkar, R., Panse, N., Dinshaw, K.A.,2009. HPV screening for cervical cancer in rural India. N. Engl. J. Med. 360,1385–1394.

Smith, J.A., White, E.A., Sowa, M.E., Powell, M.L., Ottinger, M., Harper, J.W., Howley,P.M., 2010. Genome-wide siRNA screen identifies SMCX, EP400, and Brd4 as E2-dependent regulators of human papillomavirus oncogene expression. Proc. Natl.Acad. Sci. 107, 3752–3757.

Vinokurova, S., Wentzensen, N., Kraus, I., Klaes, R., Driesch, C., Melsheimer, P.,

Kisseljov, F., Durst, M., Schneider, A., Knebel Doeberitz, M., 2008. Type-dependent integration frequency of human papillomavirus genomes in cervicallesions. Cancer Res. 68, 307–313.

Zheng, Z.-M., Baker, C.C., 2006. Papillomavirus genome structure, expression, andpost-transcriptional regulation. Front. Biosci. 11, 2286–2302.