Aims: Considering the sparse information about the clinical utility of the novel immunohistochemical marker ProEx C in histological sections, a decision was taken to study the pattern of ProEx C expression in normal/benign cervical epithelium (N/B), low-grade squamous intraepithelial lesion (LGSIL) and high-grade squamous intraepithelial lesion (HGSIL), as well as the association of ProEx C expression with human papillomavirus (HPV) genotypes.
Methods: 100 cervical samples, including 21 N/B cervices, 16 LGSILs, 61 HGSILs and two cervical invasive carcinomas, were obtained from conisation and hysterectomy. Surgical specimens were arranged in three tissue microarrays and stained for ProEx C. Ninety-three samples were HPV genotyped. Genotyping was performed by DNA amplification and hybridisation with genotype-specific probes on a low-density DNA array.
Results: ProEx C-positive expression in more than the lower third of the epithelium was observed in 14.3% of N/B, 62.5% of LGSIL and 90.2% of HGSIL. Seventy percent of HPV positivity was found in cases with expression in more than the lower third of the epithelium. Of 31 cases that were positive for HPV16, 16.1% showed ProEx C expression restricted to one or two basal layers, and 83.9% showed ProEx C expression in more than the lower third of the epithelium.
Conclusions: ProEx C is significantly associated with HPV16 infection and is a useful adjunct in the identification of LGSIL and HGSIL in histological sections when expressed in more than the lower third of the epithelium.
Statistics from Altmetric.com
ProEx C is a useful marker in the identification of squamous intraepithelial lesions in cervical histological sections.
ProEx C expression in more than the lower third of the epithelium and diffuse pattern of expression are significantly associated with higher-grade lesions.
ProEx C positive expression is significantly associated with HPV16 infection.
Human papillomavirus (HPV) is the leading cause of cervical cancer. Although HPV infections are among the more common sexually transmitted diseases, only a small subset of HPV infections will cause this type of cancer.1 The search for prognostic markers has been a continuing quest. Because of the oncogenic mechanism of HPV, cell-cycle-related proteins have been extensively investigated for their prognostic value in both cytological and histological specimens.
Recently, transcriptional profiling and cDNA microarray studies have identified two cell-cycle-related proteins, minichromosome maintenance protein-2 (MCM2) and topoisomerase II-alpha (TOP2A), the genes of which are overexpressed in cervical cancer2 3 and seem to act in early S phase, allowing recognition of replication origin and DNA unwinding.4 ProEx C is a recently developed immunohistochemical marker for detection of MCM2 and TOP2A proteins. Recent studies have shown the utility of this marker in cytological specimens as an ancillary test to distinguish normal and benign lesions such as metaplasia from cervical cancer precursor lesions.5 6 However, there is little reported experience of the use of this marker in tissue sections.7–9 To the best of our knowledge, no studies have so far attempted to relate ProEx C expression to specific HPV genotypes. The aim of this work was to evaluate ProEx C immunostaining in cervical tissue specimens using a tissue microarray platform and to compare it with HPV genotype detected by PCR amplification and hybridisation on low-density DNA arrays.
MATERIALS AND METHODS
A total of 100 specimens obtained from conisations and hysterectomies were arranged in three tissue microarrays. These specimens comprised 21 normal/benign cervices (N/B) (14 normal, two metaplasias and five cervicitis), 16 low-grade squamous intraepithelial lesions (LGSILs), 61 high-grade squamous intraepithelial lesions (HGSILs; 17 cervical intraepithelial neoplasia (CIN) II and 44 CIN III/in situ carcinoma), and two cervical invasive carcinomas (one squamous cell carcinoma and one adenocarcinoma). Normal/benign cervices were obtained from female patients who had undergone hysterectomies because of myomas. All identifiers were deleted to protect patient confidentiality, and the study was approved by the local ethics board. Each case was represented by two cores, each 1 mm in diameter. Two pathologists (SOR and JAO) performed a double-blind study for histological classification of the cases.
Tissue sections were deparaffinised, and antigen retrieval was carried out by treating the slides with Target Retrieval Solution pH 6.0 (S1699; Dako, Copenhagen, Denmark) using PTLink (Dako). The tissues were then washed three times with Tris buffer, pH 7.6, containing Tween 20 (S3006; Dako). Sections were incubated with ProEx C primary antibody (clone MCM2 26H6.19, MCM2 27C5.6, TOP2A SWT3D1; Tripath-Imaging, Burlington, North Carolina, USA) in compliance with the manufacturer’s instructions. Endogenous peroxidase was blocked by the use of 3% hydrogen peroxide in water for 5 min. Envision (Dako) polymer-based system was used for antigen detection. Two positive controls for ProEx C were used: a section of a Bowen disease sample, as described by Freeman et al10 for MCM2 expression, and a section of a HGSIL sample, as described by Shi et al.7 As a ProEx C negative control, a section of normal cervix was used. A reagent-negative control was performed by omitting the primary antibody and replacing it with Dako Real antibody diluent (S2022).
Slides were counterstained with H&E, washed and mounted with DPX mountant for microscopy (360294H; VWR International Ltd, Poole, Dorset, UK).
The distribution of ProEx C staining was scored by a blinded pathologist (MPG) as negative (no staining), bpb+ (basal 1–2 layers), 1+ (lower third), 2+ (lower two-thirds), or 3+ (more than the lower two-thirds and up to full thickness). “Diffuse pattern” was defined as >50% dysplastic squamous epithelium stained, and “focal pattern” as <50% of dysplastic squamous epithelium stained, as proposed by Shi et al.7
Slices 4 μm thick were cut from the original lesions, and genomic DNA extraction was performed using the GeneAll kit (General Biosystems, Seoul, Korea) according to the instruction manual. Viral genotypes were determined using the Hybribio HPV genoarray kit (Hybribio, Hong Kong, China) in accordance with the manufacturer’s instructions. Briefly, amplification of the L1 conserved region with biotin-labelled primers was performed, and amplicons were denaturalised and subjected to flow-through hybridisation on a low-density array consisting of a nylon 6.6 biodyne membrane coated with 21 HPV type-specific probes corresponding to high-risk genotypes 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68, low-risk genotypes 6, 11, 42, 43, 44, and genotypes with undetermined risk 53 and CP8340. Membrane was incubated with streptavidin–alkaline phosphatase conjugate and treated with nitroblue tetrazolium substrate. Colour development indicates the presence of the corresponding HPV genotype. A hybridisation-positive control containing biotin, a PCR internal control, an HPV-negative control and an HPV-positive (HPV18) control were included.
Statistical analysis was performed using SPSS V14.0. The χ2 test was used to evaluate statistical significance. p<0.05 was considered significant. The McNemar index was used to evaluate sample variability between the two cores (related samples), and the κ coefficient was applied for concordance analysis between the two observers. Logistic regression was used for sensitivity and specificity calculations.
Tissue microarray validation
For concordance of pathological diagnoses between the two observers, the κ coefficient was 85.7% (95% CI 72.0% to 99.3%). The McNemar index revealed that sample variability between the two cores representing the same specimen was ⩽ 10% for both histological classification and ProEx C expression.
ProEx C immunostaining
ProEx C immunostaining was restricted to cell nuclei in all samples. No negative staining for ProEx C was observed in any case. Some nuclei of stromal cells, lymphocytes and germinal centre cells also stained positive (data not shown).
ProEx C expression in more than the first third of the epithelium (⩾1+) was observed in 14.3% of N/B, 62.5% of LGSIL and 90.2% of HGSIL (p<0.001) (table 1 and figs 1 and 2). The two cases of carcinoma showed 3+ ProEx C expression.
Fifty-eight out of 71 cases (81.7%) with a diffuse pattern showed ProEx C expression ⩾1+ (p<0.05) (table 2).
When assessing the utility of ProEx C expression above the basal 1–2 epithelial layer (bpb+) in distinguishing N/B epithelium from LGSIL and HGSIL, sensitivity was 84.4% (95% CI 74.4% to 91.7%) and specificity was 85.7% (95% CI 63.7% to 97.0%), and in distinguishing N/B and LGSIL from HGSIL, sensitivity was 90.2% (95% CI 79.8% to 96.3%) and specificity 64.9% (95% CI 47.5% to 80.2%).
Association of ProEx C immunostaining with HPV genotypes
The HPV genotype test was performed in 93 samples. HPV genome was detected in 87 cases. As shown in table 2, cases with bpb+ ProEx C expression showed 29.9% of HPV positivity, and in cases with expression ⩾1+, HPV presence was 70.1% (p>0.05).
The most common HPV genotype was HPV16 (35.6% of total). Of HPV16-positive cases, 16.1% showed bpb+ staining and 83.9% ProEx C expression ⩾1+ (p = 0.027) (table 2).
This association was not seen for the rest of the HPV genotypes (p>0.05). No significant relationship was observed between ProEx C expression and HPV co-infection. Association analysis between grade of lesion and HPV16 infection did not reach statistical significance (p = 0.115).
Sensitivity and specificity calculations and confidence intervals for significant associations are provided online in a supplementary table (table S1).
The use of tissue microarray has been shown to be feasible in the study of precursor lesions of cervical cancer.11–17 Our concordance and sample variability analyses support these previous studies. In the present study, we observed that ProEx C immunostaining in more than the basal 1–2 layers was significantly associated with higher-grade lesions and a diffuse pattern of expression.
Available information about ProEx C expression in tissue sections is sparse, but our results agree with those reported by Shi et al,7 Badr et al8 and Pinto et al,9 which showed a strong association between HGSIL and ProEx C positivity in the lower and upper fractions of the epithelium.7–9
As reported by others, we found that ProExC is expressed in basal layers in normal epithelium, some stromal cell nuclei and germinal centres.7 8 10 This feature may serve as a useful positive internal control.
Given the different oncogenic potential of each high-risk HPV genotype,1 we found it interesting to HPV genotype all our cases and relate this to ProEx C expression. Our findings indicate that ProEx C expression tends to be associated with HPV infection. Badr et al,8 using a cocktail of high-risk HPV probes for in situ hybridisation on 56 biopsy samples, found this association to be significant. In our study, this association becomes significant for cases positive for the most oncogenic HPV genotype, HPV16.
Our results show that ProEx C could be a useful marker, especially for distinguishing N/B cervical epithelium from LGSIL and HGSIL. To summarise, although it seems that ProEx C is not a specific marker for the proliferation induced by HPV infection, its expression is significantly associated with HPV16 infection and has clinical utility in identifying precursor lesions of the uterine cervix when expressed in more than the lower third of the epithelium.
We are grateful to Ana Díez from the Centro Nacional de Investigaciones Oncológicas, Madrid, for her help with tissue microarray construction, and Luis Polo from the Hospital Universitario Virgen de la Arrixaca, Murcia, for his help with the early stages of the project. Thanks go to P Mason for reviewing the English version of this paper.
web only appendix 62/2/159
▸ A supplementary table is published online only at http://jcp.bmj.com/content/vol62/issue2
Funding: This work was supported by a grant from Fundación para la Formación e Investigación Sanitarias, Consejería de Sanidad de la Región de Murcia, Spain and Fundación Cajamurcia.
Competing interests: None.
Ethics approval: Obtained.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.