Imiquimod-induced clearance of HPV is associated with normalization of immune cell counts in usual type vulvar intraepithelial neoplasia

Authors


Abstract

Recently, we reported on the efficacy of imiquimod for treatment of usual type vulvar intraepithelial neoplasia (uVIN). A histologic regression of uVIN to normal tissue was observed in 58% of patients. As success of treatment is related to clearance of high-risk human papilloma virus (HPV), the aim of our study was to assess differences in immune cell counts and in the expression of p16INK4a in VIN tissue before and after imiquimod treatment, in relation to HPV clearance and clinical response. Vulvar tissue samples taken prior to imiquimod treatment and 4 weeks after treatment were tested for the presence of HPV. Previously determined immune cell counts (CD1a, CD207, CD208, CD123/CD11c, CD94, CD4, CD8 and CD25/HLA-DR) in epidermis and dermis of 25 VIN patients and 19 healthy controls were completed with the counts for CD14 and CD68. The expression of p16INK4a was investigated by immunohistochemistry in 15 patients. Before imiquimod treatment, both HPV cleared and HPV noncleared patients showed mainly in the dermis significantly upregulated immune cell counts compared to healthy controls. However, in patients that cleared HPV and showed histologic regression already 4 weeks after imiquimod treatment, immune cell counts and p16INK4a expression were normalized. In conclusion, our data indicate that imiquimod-induced clearance of HPV results in normalization of counts for certain immune cells and is strongly correlated with histologic regression of the disease.

Vulvar intraepithelial neoplasia (VIN) is a premalignant disorder that is classified into differentiated type VIN, which is associated with lichen sclerosus, and usual type VIN (uVIN),1 which is caused by a persistent infection with a high-risk or oncogenic HPV (hrHPV, usually HPV-type 16, 18 or 33).2

Over the last decades, the incidence of uVIN has increased, most likely due to a rise in the incidence of HPV infections.3 Lifetime risk to become infected with HPV in western societies is around 80% and ∼40% of all sexually active, female adolescents are at least once infected with hrHPV.4–7 When hrHPV persists (in less than 10% of cases), premalignant disorders of the anogenital tract, such as uVIN, can develop.8–10 uVIN has invasive potential (10% of untreated cases will progress in 1–4 years) and needs to be treated proactively.11, 12

The host immune response is of critical importance in determining clearance or persistence of an HPV-infection. In natural life, during the early stages of a viral infection, CpG-rich regions in the viral DNA are recognized by toll-like receptor (TLR) 7 and 9 on the cell surface of immature dendritic cells (DCs). Upon binding, TLRs activate kinase cascades eventually activating NFκB, which will result in the production of cytokines, adhesions molecules and other effectors of the innate immune response.13 Other receptor and adhesion molecules at the DC surface can bind viral antigen and will internalize and digest the antigen, which is followed by expression on the cell membrane in major histocompatibility complex (MHC) class I and in MHC class II. Class I MHC is recognized by CD8 and class II MHC by CD4. Upon binding antigen, the DC matures and migrates to the lymph nodes where it presents antigen bound to class II MHC, to naïve T-cells. Binding to naïve T-cells will result in production of memory and effector T-cells, such as CD4+ T-helper cells, CD8+ cytotoxic T-cells, or regulatory T-cells (Treg cells).14, 15

For a long time, surgical removal of all visible lesions was the standard treatment of uVIN.16 However, the concept that the immune response in uVIN is insufficient17–19 has lead to new, less invasive treatment options in which the immunomodulator imiquimod plays a role. Imiquimod is an immune-response modulator that acts by activating the DC, thereby inducing secretion of proinflammatory cytokines and T-cell activation. This results in a profound tumor-directed cellular immune response.20–23

Recently, our group reported in a placebo-controlled, double-blinded, randomized trial on the efficacy of imiquimod treatment for uVIN.24 During this trial, patients were treated with imiquimod or placebo twice a week for a period of 16 weeks. An imiquimod-induced reduction in lesion size by ≥25% in 81% of patients was measured, while a complete response was observed in 35% of patients. Moreover, when we defined the histological diagnoses in accordance to the ISSVD guidelines1 in uVIN or normal tissue instead of VIN 1, 2 or 3, complete histologic regression of uVIN was observed in 58% of patients.

Furthermore, our study put forward that reduction in lesion size was correlated to normalization of the numbers of immune cells.25 In the present study, this concept was investigated further. Therefore, we retrospectively assessed correlations between numbers of immature, mature and plasmacytoid DCs, NK-cells, monocytes, macrophages and T-cells in vulvar skin of 25 patients with uVIN before and 4 weeks after imiquimod treatment with the degree of histological response and clearance of HPV. Data from 19 healthy women were used as reference values. We also studied the presence of p16INK4a in relation to HPV status before and after imiquimod treatment. P16INK4a is a member of the INK4a family of cyclin dependent kinase inhibitors and during a persistent hrHPV infection, the viral oncoproteins E6 and E7 can interfere with the cell cycle regulatory pathway, which results in overexpression of p16INK4a. Therefore, immunohistochemistry with p16INK4a is used as a diagnostic aid and positivity of p16INK4a is a surrogate marker for hrHPV infection.26–29

In addition to our previous article in which it was observed that VIN patients have disturbed immune cell counts,30 we observed in the present study that already 4 weeks after treatment, HPV clearance and subsequent normalization of p16INK4a expression indicates a full or almost complete clinical response and is characterized by normalization of counts for certain immune cells in the epidermis and dermis.

Abbreviations

CIN: cervical intraepithelial neoplasia; DCs: dendritic cells; HPV: human papilloma virus; hrHPV: high-risk human papilloma virus; mDCs: mature dendritic cells; MHC: major histocompatibility complex: NK: natural killer; pDCs: plasmacytoid dendritic cells; PDT: photodynamic therapy; TLR: toll-like receptor; Treg cells: regulatory T-cells; (u)VIN: (usual type) vulvar intraepithelial neoplasia

Material and Methods

Patients

For the immunohistochemical staining, frozen tissue sections of 4-mm punch biopsies were used. Biopsies from 25 women with multifocal uVIN who participated in a placebo-controlled RCT25 were taken before treatment with imiquimod (at time = 0 weeks) and 4 weeks after treatment with imiquimod (at time is 20 weeks). During this trial, patients were treated with imiquimod or placebo twice a week for a period of 16 weeks. All biopsies were immediately frozen in liquid nitrogen and stored at −80°C until further analysis. Furthermore, 19 women who underwent elective vulvar surgery for cosmetic reasons served as healthy controls.

For staining with p16INK4a, formalin-fixed paraffin-embedded biopsies were used,25 taken before and 4 weeks after imiquimod treatment.

Histological diagnosis of all biopsy specimens was performed by 2 experienced gynaecologic pathologists. In the previous study by van Seters et al.,25 biopsy specimens were classified into VIN 1, 2 and 3. In the present study, biopsy specimens were classified uVIN according to the new ISSVD guidelines1 (previously VIN 2 or 3). In this classification, VIN 1 no longer exists and is considered normal.

All frozen tissue samples were analyzed for the presence of the 14 most prevalent hrHPV-types by using a standard GP5+/6+ PCR enzyme immunoassay followed by reverse line blot analysis, as described previously.31 For reasons explained in the results section, 2 samples were also analyzed with the SPF10 DEIA/LiPA version 1 test, which is a more sensitive HPV test.32 The HPV detection limit of the SPF10 DEIA/LiPa is 65 bp, and it detects 25 different HPV genotypes.

Retrospectively, patients were divided into 2 groups, namely patients who cleared HPV (HPV cleared) and patient who did not clear HPV (HPV noncleared) after imiquimod treatment.

Medical Ethical Committees approved our study design and all women provided voluntarily written informed consent.

Immunohistochemical staining on frozen tissue sections

Immunohistochemical staining was performed on 6-μm-thick frozen tissue sections for the following markers: CD1a, classical marker for immature DCs/ Langerhans cells (Orthobiotech, Bridgewater, NJ); CD207, marker for immature DCs expressing Langerin (DCGM4; Beckman Coulter, Fullerton, CA); CD208, marker for mature DCs (104.G4; Beckman Coulter); CD94, marker for NK cells (HP.3b1; Beckman Coulter); CD14, marker for monocytes (M0825; Dako Denmark, Glostrup, Denmark); CD68, marker for macrophages (M0718; Dako); CD4, marker for T-helper cells (MT.310; Dako); CD8, marker for cytotoxic T-cells (DK25; Dako); CD25/HLA-DR, marker for Treg cells (AKT-1; Dako/1E5; Sanquin); for plasmacytoid DCs (CD123+/CD11c), a double staining with both markers was used [anti-CD123 (9F5; Becton Dickinson, Alphen aan den Rijn, The Netherlands) and anti-CD11c (SHCL-3; Becton Dickinson)]. Staining and light microscopic evaluation of the obtained data were performed as described earlier.30

P16INK4a staining

Staining for p16INK4a was performed on 4-μm-thick paraffin-embedded tissue sections. The tissue sections were first deparaffinized overnight and rehydrated through graded series of xylene and alcohol. Endogenous peroxidase was blocked in H2O2 for 5 min, followed by washing the sections in dH2O. The slides were placed in a citrate buffer (pH 6) and heated in a 900 W microwave for 1 × 6.5 min, 1 × 3 min and—after adding dH2O—again for 1 × 3 min. After cooling for 1 hr, tissues were washed in PBS for 3 × 5 min. Unspecific binding was blocked with 200 μl PBS with 0.3% bovine serum albumine (BSA) for 30 min. Sections were then incubated overnight at 4°C in a dark humid chamber with the mouse monoclonal antibody p16INK4a (clone JC8, sc-56330, Santa Cruz biotechnology, Santa Cruz, CA) in a dilution of 1:100. After incubation with the primary antibody, sections were incubated with 200 μl biotin-conjugated goat antimouse immunoglobulin (Dako) diluted 1:400 in PBS with 0.3% BSA for 30 min. This was followed by incubation with 200 μl streptavidin-biotinylated-peroxidase complex (ABComplex) with horse radish peroxidase (HRP) conjugate (Dako) for 30 min in dark. Between all incubation steps, sections were washed with PBS for 3 × 5 min. Staining was developed using 0.133 g diaminobenzidine (DAB, Fluka, Sigma Aldrich, Buchs SG, Switzerland) in 200 ml PBS with 600 μl H2O2 and the reaction was stopped after 10 min by washing in dH2O. Finally, sections were counterstained with haematoxylin (Klinipath, Duiven, The Netherlands) for 3 sec, washed with flushing water and dehydrated through graded series of alcohol and xylene.

Light microscopic evaluation was performed blinded. Nuclear and cytoplasmic staining was considered as a positive reaction. Sections were scored semiquantitative for the percentage of p16INK4a positive cells: 0–25% positive (i), 25–50% positive (ii), 50–75% positive (iii), >75% positive (iv). Sections were also scored for the intensity of the staining: absent (i), light (ii), moderate (iii) and strong (iv).33 The total score per sample was obtained by multiplying percentage with intensity.

Statistical analysis

Data analysis was performed with the use of the SPSS 15.0 software package for Windows. The Kolmogorov-Smirnov test was used to identify the distribution for the cell types. Because a non-normal distribution was observed for some cell types, the nonparametric Mann-Whitney test was used for evaluation of differences in cell counts between the following groups: HPV cleared vs. healthy controls, and HPV noncleared vs. healthy controls, both before and after imiquimod treatment. Differences in p16INK4a expression before and after treatment within the HPV cleared and HPV noncleared VIN patients were assessed with the use of the Wilcoxon signed rank test.

The Mann-Whitney test was used to assess differences between the HPV cleared and HPV noncleared patients. A two-tailed p-value of p < 0.05 was chosen to represent statistical significance.

Results

Patients

The median age of the 25 VIN patients was 39 years (range, 22–56 years). In the healthy control group, the median age was 40 years (range 19–56 years). Cryosections were obtained from al these patients, but paraffin-embedded samples were limited and only obtained from 15 of 25 VIN patients before as well as after treatment.

Twenty-three patients were smokers, 2 patients had a history of immunosuppressant use, and 14 patients had a history of CIN. Patients had not received any treatment for at least 3 months before treatment with imiquimod. Of all controls, 6 were smokers and 6 had a history of smoking, none of the controls had a history of immunosuppressant use and none of them had a history of CIN.

Histological response and presence of HPV-DNA and p16INK4a

Before imiquimod treatment, 19 of the 25 VIN patients were positive for HPV 16, 5 patients were positive for HPV 33, and 1 patient was HPV DNA-negative.25 The HPV DNA-negative patient was excluded from further analysis. Four weeks after imiquimod treatment, 7 patients were positive for HPV 16, 3 were positive for HPV 33 and 14 patients were HPV DNA-negative. All healthy controls were HPV DNA-negative.

Immunohistochemical analysis for p16INK4a expression was performed on tissue sections from 15 patients before and after imiquimod treatment. Before treatment, 12 of 15 patients were HPV 16 positive and 3 patients were HPV 33 positive. After treatment, 4 patients were HPV 16 positive, 2 were HPV 33 positive and 9 patients were HPV DNA-negative.

P16INK4a scores and staining before and after treatment in relation to histologic regression and viral clearance are shown in Table 1 and in Figure 1. Median score of p16INK4a expression (percentage × intensity) before imiquimod treatment was 12 for the HPV cleared and 12 for the HPV noncleared patients. After treatment, p16INK4a expression was significantly decreased in HPV cleared patients (p = 0.008) but not in HPV noncleared patients (p = 0.084). In addition, p16INK4a score after treatment was significantly lower in HPV cleared patients compared to HPV noncleared patients (p = 0.006). Interestingly, two patients who cleared HPV were still diagnosed with uVIN after treatment and additional observations and experiments were performed to clarify this situation.

Figure 1.

P16INK4a expression in control tissue and in VIN lesions before and after imiquimod treatment. (a) hrHPV negative control tissue does not show expression of p16INK4a. (b, c) p16INK4a expression before (b) and after (c) imiquimod treatment in a patient that did show clearance of HPV. (d, e) p16INK4a expression before (d) and after (e) imiquimod treatment in a patient that did not clear HPV (photographs ae, magnification: ×10).

Table 1. Histologic diagnosis after imiquimod treatment and p16INK4A scores in HPV-cleared and HPV-noncleared patients before and after imiquimod treatment
inline image

Patient 1 had a clinical response to imiquimod treatment of 76–99% and was negative for HPV and for p16INK4a expression as measured in a biopsy obtained 4 weeks after treatment. The patient was followed for an additional period after which it was observed that at that point all lesions had disappeared. To date (>5 years), this patient remains disease-free.

Patient 2 had a clinical response to imiquimod treatment of less than 25%, and remained positive for p16INK4a, although the HPV test indicated clearance of the virus. Follow-up of this patient showed several recurrences after imiquimod treatment. To rule out the possibility of having missed an HPV infection, a different more sensitive HPV test was performed (SPF10 DEIA/LiPA version 132). Using this sensitive HPV test for both patients we could not detect the presence of HPV viral DNA, signifying the absence of hrHPV infection.

Evaluation of immune cell numbers in VIN lesions before and after imiquimod treatment and in normal vulvar skin

Epidermis

immunohistochemical analysis of the epidermis showed significant differences in immune cell counts between HPV cleared (n = 14) or noncleared (n = 10) VIN patients and controls (n = 19) (Table 2).

Table 2. Immunocompetent cell counts in epidermis in controls, in patients who cleared HPV and patients who did not clear HPV upon imiquimod treatment
  1. CD1a+, immature dendritic cells (DCs)/Langerhans cells; CD207+, immature DCs expressing Langerin; CD208+, mature DCs; CD8+, cytotoxic T-cells; CD4+, T-helper cells; CD94+, natural killer cells; CD25+/HLA-DR+, Treg cells; CD123+/CD11c, plasmacytoid DCs; CD14+, monocytes; CD68+, macrophages.

inline image

Before treatment, CD8+ cytotoxic T-cells were significantly lower in HPV cleared patients as compared to controls (Table 2, p-value 1 vs. 2), while CD14+ monocytes were higher in HPV cleared patients and in HPV noncleared patients compared to healthy controls (Table 2, p-value 1 vs. 2 and p-value 1 vs. 4). After treatment, counts for CD8+ and CD14+ cells returned to control levels in HPV cleared patients (Table 2, p-value 1 vs. 3, Figs. 2a, 2b, and 3), while in HPV noncleared patients the elevated CD14+ cell numbers did not return to control levels (Table 2, p-value 1 vs. 5). In addition, CD123+/CD11c pDCs, CD208+ mDCs and CD68+ macrophages were also significantly higher in HPV noncleared patients after treatment compared to controls (Table 2, p-value 1 vs. 5).

Figure 2.

Numbers of immuncompetent cells in control and VIN tissues before and after imiquimod treatment in patients who cleared or did not clear HPV from the site of the lesion. (a, b) CD8+ and CD14+ cells in epidermis of VIN patients before and after imiquimod treatment. (cf) CD208+, CD123+/11c, CD4+ and CD8+ cells in dermis of VIN patients before and after imiquimod treatment. Median ± SD. *p < 0.05, **p < 0.005, ***p < 0.001, significant different from controls (Mann-Whitney test).

Figure 3.

CD14, CD4 and CD8 positive cells in control and VIN tissues before and after imiquimod treatment in patients who cleared or did not clear HPV from the site of the lesion (magnification: ×10).

CD1a+ Langerhans cells in HPV cleared and noncleared patients were not different from healthy controls before treatment, but were significantly higher in HPV cleared patients after treatment (Table 2, p-value 1 vs. 3).

No significant differences before treatment were seen when we compared the HPV cleared patients with HPV noncleared patients (Table 2, p-value 2 vs. 4). However, when we performed the same analysis after treatment, the number of CD14+ and CD68+ cells were significantly higher in the HPV noncleared group (Table 2, p-value 3 vs. 5).

Dermis

In the dermis, differences in cell numbers were much more pronounced (Table 3). Before imiquimod treatment, there were significantly higher numbers of CD4+ T-helper cells, CD8+ cytotoxic T-cells, CD25/HLA-DR+ Treg cells, CD208+ mDCs and CD123+/CD11c pDCs in HPV cleared patients vs. controls (Table 3, p-value 1 vs. 2). Also in the HPV noncleared group, the numbers of these cells were higher, although this was not significant for CD123+/CD11c pDCs and CD25/HLA-DR+ Treg cells (p = 0.079, p = 0.082 respectively, Table 3, p-value 1 vs. 4). In the dermis, similar to the epidermis, in HPV cleared patients cell numbers returned to normal after treatment (Table 3, p-value 1 vs. 3), this in contrast to the cell numbers in HPV noncleared patients that did not return to normal (Table 3, p-value 1 vs. 5, Figs. 2c–2f and 3).

Table 3. Immunocompetent cell counts in dermis in healthy controls, patients who cleared HPV and patients who did not clear HPV upon imiquimod treatment
  1. CD1a+, immature dendritic cells (DCs)/Langerhans cells; CD207+, immature DCs expressing Langerin; CD208+, mature DCs; CD8+, cytotoxic T-cells; CD4+, T-helper cells; CD94+, natural killer cells; CD25+/HLA-DR+, Treg cells; CD123+/CD11c, plasmacytoid DCs; CD14+, monocytes; CD68+, macrophages.

inline image

When we compared HPV cleared and HPV noncleared patients, no significant differences in cell counts were seen before treatment (Table 3, p-value 2 vs. 4), while after treatment, CD208+, CD8+, CD4+, CD94+, CD14+ and CD68+ cells were significantly lower in the HPV cleared patients in comparison to the HPV noncleared patients (Table 3, p-value 3 vs. 5).

Individual values for immune cell counts per patient, combined with the data on viral clearance and p16INK4a expression, are shown in the Supporting Information Table 1.

Discussion

The observations in our study demonstrate that clearance of HPV infection already 4 weeks after treatment with imiquimod, results in normalization of immune cell counts and normalization of p16INK4a expression at the site of the lesion. Winters et al.34 also studied immune cell counts and measured the HPV status of VIN patients before and after treatment with imiquimod combined with photodynamic therapy (PDT). Interestingly, Winters observed that besides all complete responders, 2 of 4 partial responders and 1 patient with stable disease had become HPV-DNA negative. In our study, 12 of 14 patients that cleared HPV 4 weeks after imiquimod treatment showed complete histologic regression of uVIN and an additional patient showed a somewhat delayed complete regression. An explanation for the delayed regression could be that clearance of HPV precedes regression of uVIN. This has been shown previously in patients with abnormal cervical smears, in which hrHPV clearance preceded regression of cervical lesions by an average of 3 months.35

Another patient, who cleared HPV after treatment, however, showed a clinical response of less than 25% and continuing elevated expression of p16INK4a. Our first thought was that the method used to detect viral DNA was not sensitive enough. However, even a more sensitive method on paraffin sections adjacent to those used for p16INK4a detection could not detect HPV. In other words, despite imiquimod-induced clearance of HPV the disease had only marginally improved and p16INK4a level remained elevated. These findings seem to indicate that in this particular case the disease may have progressed beyond HPV control. Elevation of p16INK4a, normally a hallmark of hrHPV,26–29 in this case can be considered a compensation for abnormal cell proliferation as is observed in many HPV-negative carcinomas.28, 36, 37

HPV clearance results in normalization of immune cell counts

As persistent HPV infection is the underlying cause of VIN and the host immune response is primarily directed against this HPV infection, we focused our investigations on differences in local immunity between patients who were able to clear HPV and patients who were not able to clear HPV upon local treatment with imiquimod. Our most important finding is that, already 4 weeks after the end of treatment, imiquimod-induced clearance of HPV results in normalization of the numbers of immune cells at the site of the lesion (Table 2 and 3). Although imiquimod treatment is generally well tolerated and for a number of patients a good alternative for otherwise mutilating treatments like vulvectomy or local excision, it is of interest to know beforehand which patients will respond to therapy and which patients will not. Winters et al.34 reported that Treg cells in nonresponders before treatment were significantly increased compared to numbers in responders, a finding which we could not confirm. However, we did compare the numbers of immune cells before treatment between those patients that cleared HPV upon treatment vs. those patients that did not clear HPV to find other possible predictors of HPV clearance upon imiquimod treatment. Our results demonstrate lower CD8+ cell numbers in the epidermis of patients that cleared HPV, but not in patients who did not clear HPV when we compared to healthy control levels (Table 2). Furthermore, CD123+ pDCs in the epidermis and CD68+ cells in the dermis of HPV noncleared patients were significantly higher compared to controls (Tables 2 and 3), which was not the case in the HPV cleared group. In theory, the above indicated CD-markers can be used to predict success or failure of imiquimod treatment to clear HPV. However, when we compared the HPV cleared group directly with the HPV noncleared group (before treatment, Tables 2 and 3, p-value 2 vs. 4), we did not measure any significant differences between the 2 groups. On the basis of our current findings, we consider it too early to do any recommendations on the above markers to predict HPV clearance upon imiquimod treatment.

Previously, our study group reported on the immunosuppressive state of the epidermis of VIN patients, demonstrated by a decrease in CD1a+ and CD8+ cells.30 Interestingly, the present study shows a higher number of CD14+ monocytes in the epidermis of VIN patients compared to controls. It is known that CD14+ cells in an environment rich of TGF-β1 (which is produced by most immune cells), and in response to CCL20 (for which CD14+ cells express the receptor molecule CCR6) differentiate into immature Langerhans cells (CD207+). Under conditions of cutaneous inflammation, when the proinflammatory cytokine GM-CSF is released along with TGF-β1, CD14+ cells can differentiate into mature Langerhans cells (CD1a+).38–41 This is in accordance with our findings. After treatment, we observed a decreased number of CD14+ monocytes (p = 0.005, Wilcoxon signed rank test) in the HPV cleared epidermis, while the number of CD1a+ Langerhans cells in this group seemed to be increased (p = 0.074, Wilcoxon signed ranked test). During successful imiquimod treatment, CD14+ monocytes in the epidermis might have differentiated into Langerhans cells.

An important limitation of the present study is the absence of biopsy specimens during treatment to assess the specific changes in immune cells induced by imiquimod that are associated with the clearance of lesions. Such data would be important in identifying the specific immune cells that must be targeted for successful treatment.

In conclusion, our data indicate that already 4 weeks after the end of treatment, imiquimod-induced clearance of HPV (confirmed by normalization of p16INK4a expression) results in normalization of immune cell counts and is strongly correlated with histologic regression of the disease. Interestingly, van Seters et al.24 could show that histologic regression is maintained till at least 1 year after the start of treatment and preliminary long-term follow-up data (>5 years) further confirm the success of imiquimod treatment.

Acknowledgements

We thank Dr. W. Quint for performing the SPF10 DEIA/LiPA version 1 test.

Ancillary