Article Text


The impact of epithelial biomarkers, local immune response and human papillomavirus genotype in the regression of cervical intraepithelial neoplasia grades 2–3
  1. Irene Tveiterås Øvestad1,5,
  2. Einar Gudlaugsson1,5,
  3. Ivar Skaland1,
  4. Anais Malpica3,4,
  5. Ane Cecilie Munk1,2,5,
  6. Emiel A M Janssen1,
  7. Jan P Baak1,5
  1. 1Department of Pathology, Stavanger University Hospital, Stavanger, Norway
  2. 2Department of Gynecology, Stavanger University Hospital, Stavanger, Norway
  3. 3Department of Pathology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas, USA
  4. 4Department of Gynecologic Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas, USA
  5. 5The Gade Institute, Medical-Odontologic Faculty, University of Bergen, Bergen, Norway
  1. Correspondence to Professor Dr J P A Baak, Department of Pathology, Stavanger University Hospital, Box 8100, Stavanger 4068, Norway; jpabaak{at}


Background and aims 15–30% of cases of cervical intraepithelial neoplasia grades 2–3 (CIN2–3) detected in punch biopsies regress spontaneously (ie, show CIN1 or less in the follow-up cone). Epithelial retinoblastoma protein (pRb), tumour suppressor protein (p53), HPV genotype and immunoreactive cells have been reported to be helpful in predicting regression but their interaction in regression prediction is unknown.

Material and methods 55 cases of CIN2–3 in cervical biopsies with subsequent cervical cones were studied retrospectively to assess how epithelial biomarkers, immunoreactive cells (with immunohistochemistry) and high-risk (hr) HPV genotypes (by the AMPLICOR and linear array tests) prognostically interact with epithelial pRb and p53.

Results 18% of CIN2–3s regressed (median biopsy–cone interval of 12.0 weeks, range 5.0–34.1 weeks). CIN2–3s that regressed had higher epithelial pRb and p53, lower stromal CD25+ and CD138+, and higher CD8 cells than persistent lesions. They also had higher ratios of CD4+/CD25+ and CD8+/CD25 in stroma and epithelium. HPV16 correlated with low pRb and low CD8+. With multivariate analysis a combined high ratio of CD8+/CD25+ in the stroma, high epithelial pRb and p53 expression had independent value to predict the regression.

Conclusions CIN2–3 lesions with a non-hrHPV16 infection, high ratios of stromal CD8+/CD25+ and high epithelial expression of pRb or p53 are associated with spontaneous regression.

  • Cervix
  • HPV
  • immunohistochemistry
  • immunopathology
  • molecular pathology

Statistics from


High grade cervical intraepithelial neoplasia (CIN2–3) is a common disease which is usually treated surgically (ie, either with a loop electrosurgical excision procedure (LEEP) or with a cold knife cone). If untreated the majority of CIN3 lesions will persist,1 2 and with prolonged follow-up, 31% of the cases develop invasive squamous carcinoma, compared to <1% in CIN3 cases with conventional surgical treatment.3 On the other hand, 11–40% of CIN2–3 cases can regress spontaneously (ie, maximally CIN1 or less detected in follow-up cones).4 Cervical cone excision has been associated with an increased risk for preterm birth in later pregnancies.5 6 Therefore, identifying those CIN2–3 lesions that will regress spontaneously is of clinical value because it will prevent the overtreatment of young patients with CIN2–3. It should be noted that CIN lesions microscopically show a wide variation of changes, which are classified as grade 1, 2 or 3. However, due to the subtle and gradual changes during carcinogenesis, unambiguous classification of CIN1 to CIN2 to CIN3 can be quite difficult. As a result, subjective CIN grading may not always be reproducible,7 and it is therefore important to include objective, independent laboratory tests. A quantitative interactive image analysis system (QPRODIT system, Leica, Cambridge, UK) used for Ki67 immunoquantitative features has been shown to have important prognostic value, exceeding that of CIN grading.8

The difference in the natural history of a high grade CIN (ie, regression, persistence or progression) relates to the balance between viral and host/immunological factors.9 10 Altered transcriptional regulation and immortalisation of human papillomavirus (HPV) transformed cells occurs on integration of the HPV genome into the host genome. This results in up-regulation of the oncogenic HPV proteins E6 and E7, and consequently binding and degradation of retinoblastoma protein (pRb) by E7 and degradation of tumour suppressor protein (p53) by E6. This in turn leads to uncontrolled cellular proliferation, DNA damage and chromosomal instability.11

In previous studies, low pRb and/or p53 levels in the deep layers of the epithelium correlated with persistent HPV lesions and could act as potential biomarkers for high-risk HPV (hrHPV) E6 and E7 function.9 12 In line with these findings a low expression of pRb and p53 has been found to be strongly predictive for progression of CIN1 lesions.13 Moreover, immunoreactive cells like CD8+, CD138+, CD25+ and CD4+ are important factors for predicting regression or not.10 However, the interaction and independent prognostic value of epithelial biomarkers, immunoreactive cells and HPV genotype is unknown.

The current study analyses whether epithelial biomarkers (pRb, p53, CK13, CK14, Ki67, PPH3), different types of immunoreactive cells (CD4, CD8, CD25, CD138, FoxP3) and hrHPV genotype have independent prognostic value for regression or not in CIN2–3 lesions.

Materials and methods

Approvals from the Regional Medical Ethics Committee of Helse Vest, Norway (#205/06), the Norwegian Social Science Data Service (#5909/06) and the Norwegian Data Inspectorate (#11512) were obtained prior to the initiation of this study.


Sixty-two patients who underwent conisation at Stavanger University Hospital were included. All patients had an abnormal cytology specimen collected as part of routine cervical screening procedures. They were all positive for the AMPLICOR HPV test and high risk HPV genotypes were detected by the Linear Array test. All patients had cervical punch biopsy with CIN2–3 before the conisation. The histological diagnoses of the cervical punch biopsy and cervical cones were confirmed by a second pathologist blinded to the original diagnosis/clinical data and aided by additional p16 and Ki67 immunoquantitation.8 14 On this second review, the cervical biopsies were diagnosed as CIN1 (1 case), CIN2 (7 cases) and CIN3 (54 cases), and the cone biopsies as normal (9 cases), CIN1 (3 cases), CIN2 (8 cases) and CIN3 (42 cases). The single CIN1 punch biopsy case was excluded from this study. Moreover, in six of the punch biopsies the CIN2–3 lesions in the paraffin blocks had been cut through and therefore were inadequate for immunohistochemical assessment, leaving 55 cases (10 regression, 45 non-regression) for the study. In addition, immunohistochemical staining was missing for CD138 in 3 cases (1 regression and 2 persistent), Foxp3 in 1 regression case, and CD4, pRb, p53 and p16 in 1 persistent case. A total of 53 cases were available for the multivariate analysis (44 persistent and 9 regressions).

AMPLICOR HPV and linear array HPV genotyping test

The AMPLICOR HPV test (Roche Molecular Systems, Roche Diagnostics, Mannheim, Germany) was used for the HPV testing. The test permits simultaneous PCR amplification of HPV target DNA from 13 hrHPV genotypes (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59 and 68), and β-globin DNA as a cellular control. The Roche Linear Array HPV Genotyping test (Roche Molecular Systems) was used for genotyping. This test detects 37 HPV genotypes (6, 11, 16, 18, 26, 31, 33, 35, 39, 40, 42, 45, 51, 52, 53, 54, 55, 56, 58, 59, 61, 62, 64, 66, 67, 68, 69, 70, 71, 72, 73, 81, 82, 83, 84, 89 and IS39 (subtype of 82)). The Linear Array HPV genotyping strips were manually interpreted by two observers using the Linear Array HPV Genotyping Test Reference Guide provided by the manufacturer.

The AMPLICOR HPV test and Linear Array HPV Genotyping tests were performed as described previously in detail.10


The protocol for immunohistochemistry, type/manufacturer and dilution of the antibodies p16, Ki67, CD4, CD8, CD25, CD138 and FOXp3 have been described previously.10 For the epithelial biomarkers the following antibodies and dilutions were used: CK-13 (clone KS-1A3, 1/200, Novocastra, Newcastle upon Tyne, UK); CK-14 (clone LL002, 1/40, Novocastra); pRb (clone 13A10, 1/25, Novocastra); p53 (clone DO-7, 1/200, DAKO, Glostrup, Denmark, S3002); PhosphoHistone (PPH3; polyclonal, 1/1500, Polyclonal Cell Signalling Solutions, Lake Placid, New York, USA); and p63 (clone 4A4, 1/50, DAKO, S3002).

The section adjacent to the sections used for immunostaining was cut and stained with H&E to ensure the presence of the same CIN lesion in all test sections (sandwich technique).

Semi-quantitative scoring of immunohistochemical staining of biopsies from cervix

Table 1 summarises the localisation of expression and the method for quantitation of the different biomarkers. The extent and degree of immunopositivity was assessed by consensus scoring by three observers (IO, JB, EG) (field of vision for ×40 objective=0.52 mm, numerical aperture=0.65). For all the immunostained sections, the most severely dysplastic area with the most intense p16 staining was interpreted. For p16, Ki67, CD4, CD8, CD25, CD138 and FOXp3, the interpretation was done according to a previous described protocol.10 The epithelium was subjectively divided into two layers, an upper and a lower half, and all biomarkers were assessed in each half separately as described previously.15 The stroma adjacent to the basal membrane was evaluated in a strip measuring 500 μm deep below the epithelial basal membrane. This was performed by using a ×40 objective with a field diameter of 520 μm and by viewing the membrane just visible in the outermost part of the field of vision.

Table 1

Overview of the methods to interpret the different biomarkers used in this study

Statistical analysis

SPSS V.15 was used. The continuous variables were divided into two different subgroups, using a threshold value assessed by receiver operating curve (ROC) analysis (MedCalc Software, Mariakerke, Belgium).16 Some variables had to be log Normal transformed to obtain normal distribution. The Fisher exact test was used to compare the two groups for the categorical immunoquantitative variables; p values ≤0.05 were considered significant. Logistic regression (Wald test) was performed to predict the assessed biomarkers from hrHPV status. A multivariate Cox proportional hazard analysis was used to compute HRs, which reflect the RR of regression of high-grade CIN associated with the expression status of the biomarkers.


The median cervical punch biopsy to cone time interval was 12.0 weeks (range 5.0–34.1 weeks). There were no significant differences in time interval for the regression versus non-regression cases (p=0.78). HPV16 or combined HPV16/other high risk genotypes, and non-type 16 HPV high risk genotypes were found in the cytology specimens of all cases. HPV genotype (as HPV16, either alone or combined with other high risk genotypes, vs all others) correlated with a low percentage (<40%) of pRb in the deep layer (p=0.05) and a low number (<113) of CD8 positive cells in stroma (p=0.02) (figure 1A).

Figure 1

(A) A low percentage of retinoblastoma protein (pRb) positive cells and low numbers of cytotoxic CD8 positive T-cells correlate with high-risk human papillomavirus (hrHPV)-16. (B) High ratios of CD8/CD25 positive cells correlate with high percentages of pRb for cases with regression. HrHPV E7 is responsible for down-regulating both pRb and MHC I and thereby disrupting presentation of antigens to CD8+ T-lymphocytes and formation of antigen specific CD8+ T-cell clones. (C) Cases with high percentages of pRb and tumour suppressor protein (p53) together with high ratios of CD8/CD25 positive cells are most likely to regress.

The p16 staining intensity in the upper epithelium layer was stronger in the cases with no regression (p=0.02) (table 2). The regression cases had lower numbers of CD25+ in stroma and epithelium, and lower numbers of CD138+ and higher numbers of CD8+ in stroma than the persistent cases. They also had higher ratios of stromal and intraepithelial CD4+/CD25+ and CD8+/CD25+. For CK13+, CK14+ or p63+ the differences between regression and persistent cases were small and not significant. PPH3+ and Ki67+ also showed no significant differences in the two groups. CIN2–3s that regressed had higher percentages of pRb and p53 positive cells in both deep and upper layers of the epithelium than the persistent cases.

Table 2

Correlation between the different variables in the CIN2–3 lesions with and without regression

The sensitivity/specificity for predicting regression/non-regression for pRb and p53 epithelium deep layer is 100%/84% and 80%/73%, respectively; for the immunomarkers CD138+ and CD8+ in stroma, 67%/77% and 60%/70%, respectively; and for the ratios LNCD4+/CD25+ and LNCD8+/CD25+ in stroma, 70%/89% and 70%/76%, respectively. (table 2) The results of multivariate (Cox model) analysis confirmed the hypothesis that a high ratio of CD8+/CD25+, high pRb and high p53 are independent predictors for regression. Variables listed in the second part of table 3 were not selected by the Cox model analysis (figure 1B–C, table 3).

Table 3

Results of multivariate survival analysis (Cox model) for the different variables of the cases with and without regression (variables not included are listed in the second part of the table)


High epithelial expression of pRb and p53 together with high ratios of stromal CD8+/CD25+ are predictors for regression in CIN2–3 cases. HPV16 positivity correlated with low percentages of pRb in epithelial deep layer and low numbers of CD8+. In addition, p16 staining intensity in the upper epithelial layers was lower in the regression cases. These results show that there are significant differences in the micro environment of regressive and persistent CIN2–3 lesions, and that the different features are correlated with each other yet contribute independently to regression or persistence. These findings support the hypothesis and confirm previous reports that the first signs of a regressing CIN2–3 lesion are increased pRb, p539 and CD8 cells,10 reflecting the down-regulation of E6 and E7 due to responses by the host. On the other hand, a low ratio of CD8+/CD25 cells together with low pRb and p53 is associated with a persistent non-regressive lesion and immune tolerance by suppression of effector T-cells against E6 and E7 (figure 1C). Interestingly, pRb in the lower half of the epithelium and CD8 in the stroma is also associated with HPV genotype, with low values in HPV16 infected CIN2–3 lesions (figure 1A). HPV16 has been associated with increased severity of diagnosis.17 18 The mean duration of an HPV16/18 infection is estimated to be 1.2 years,19 and during this period, various cytokines and chemokines, presenting antigen and expressing adhesion and inhibitory molecules lead to either tolerance or active immune responses. This reflects how the host can remain ignorant of HPV for long periods as viral growth accompanies the maturation of the keratinocyte as it progress up the epithelium and may result in a persistent chronic infection.

To allow virus replication, the HPV E7 delays differentiation and upgrades cell proliferation of the keratinocytes by degrading pRb. The induction of cellular proliferation through degradation of pRb, activation of cyclin-dependent kinases (CDKs)20 and disruption of the activity of CDK inhibitors p21 and p27 by E721 is a necessary step to keep the cells in the replicating phase. In order to overcome the cellular response to cell stress, the E6 protein binds to the tumour suppressor protein p5322 and is also able to block its transcriptional activity.23 The differences in expression of pRb and p53 between the regression and persistent cases in this study illustrate that the features necessary for cellular immortalisation and transformation are different.

A biopsy from a CIN lesion represents a snapshot of the morphological changes of an ongoing disease and therefore provides a picture of cellular interactions in situ. During tumour progression, the microenvironment undergoes alterations; the nature and composition of the inflammatory infiltrates change, and the expression of biomarkers such as pRb, p53 and p16 is altered. Together with the HPV genotype all these factors interact, resulting in the abovementioned snapshot. Therefore, each one of these factors could possibly serve as a predictor for CIN2–3 regression or persistence.

Enrolment of patients with precancerous lesions or early stage cervical cancer for future clinical trials regarding therapeutic vaccines24 should include examination of the patient's immune system with immunomarkers like CD138, CD8 and the ratios of CD8+/CD25+ and CD4+/CD25+ in stroma before and after vaccination. Adding the combined biomarkers pRb and p53 in deep layer of the epithelium would show cases where punch biopsies predict regression. These patients may not require conisation. Moreover, the combined biomarkers could perhaps also serve as surrogate success predictors of non-surgical treatment of CIN or therapeutic vaccines and thereby avoid adverse side effects of conisation.5 6

While histological CIN grade 2 is the typical clinical threshold for treatment, there is an increasing awareness that this is the least reproducible of CIN diagnosis. In order to minimise the risk of overcalling biopsies in the current study, the histological diagnoses of the punch biopsies were confirmed independently by a second pathologist and an additional verification of the CIN grade by Ki67 immunoquantitation.8

The authors are aware that the low number of regression cases is a weak point of the study and although the results appear to be clinically relevant, they need to be confirmed with the study of additional cases.

Take-home message

In cervical intraepithelial neoplasia grades 2–3 lesions, human papillomavirus genotype, immunoreactive cells, retinoblastoma protein (pRb) and tumour suppressor protein (p53) interact to induce regression or lead to persistent disease.


The authors would like to thank Anne Elin Varhaugvik and Britt Fjæran for their technical support.


View Abstract


  • Funding Helse Vest, project #507030 and the Stavanger University Hospital #2009/632.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the Regional Medical Ethics Committee of Helse Vest, Norway (#205/06), the Norwegian Social Science Data Service (#5909/06) and the Norwegian Data Inspectorate (#11512).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

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.