DNA topoisomerase I and IIα expression in penile carcinomas: assessing potential tumour chemosensitivity

Authors


Daniel Berney, Barts and the London, Queen Mary’s School of Medicine and Dentistry – Department of Cellular Pathology, The Royal London Hospital 80 Newark Street, London E1 2ES, UK.
e-mail: danberney@hotmail.com

Abstract

OBJECTIVE

To examine the tissue expression of DNA topoisomerase I (Topo I) and IIα (Topo II), to pursue the possibility of future chemotherapy regimens for squamous cell carcinoma of the penis (SCCP), as high expression of Topo I might indicate sensitivity to the camptothecins, whereas high Topo II might indicate sensitivity to etoposide.

PATIENTS AND METHODS

In all, 73 patients with SCCP were reviewed and then tissue samples microarrayed. These were then stained with immunohistochemistry for Topo I, Topo II and Ki-67. Tumour stage, grade and type were available.

RESULTS

Topo II showed a strong positive correlation with the proliferation index as measured by Ki-67 (P < 0.001) but no correlation with Topo I. There were also strong correlations between tumour grade and Ki-67, and Topo II expression (both P < 0.001). Tumour type was also strongly correlated with Topo II and Ki-67 expression, with the highest expression in basaloid carcinomas and the lowest in verrucous carcinomas. However, Topo I expression was not correlated with any other tumour variable.

CONCLUSION

The expression of Topo I is grade- and type-independent, and chemotherapy using the camptothecins is unlikely to be effective. The strong positivity of Topo II in high-grade and basaloid SCCPs suggests that treatment with etoposide or other Topo II ‘poisons’ might be a better target for future clinical trials.

Abbreviations
SCC(P)

squamous cell carcinoma (of the penis)

HPV

human papilloma virus

TMA

tissue microarray

Topo

topoisomerase

H&E

haematoxylin and eosin.

INTRODUCTION

Squamous cell carcinoma of the penis (SCCP) is rare in the USA and Europe [1] but has a much higher incidence in South America and Africa. The major risk factors for the development of SCCP are environmental. Circumcision shortly after birth is protective. Poor hygiene and phimosis are well recorded as being associated with SCCP. Genital warts and human papilloma virus (HPV) [2,3], and smoking, also appear to be a risk factors. SCCPs are associated with penile intraepithelial neoplasia and squamous hyperplasia [4].

The treatment of SCCP is primarily surgical. Total penectomy has been replaced, where possible, by partial penectomy or glansectomy, with preservation of some function. Sentinel lymph-node biopsy to identify the first involved node is under evaluation [5]. Inguinal dissection is used for involved nodes. Recurrent disease is treated, with variable results, by surgery and/or radiotherapy.

SCCP is divided into several different subtypes which differ from those seen in other squamous epithelia. The vast majority are of ‘usual type’. Basaloid carcinomas are HPV-related and have an aggressive course. Verrucous carcinomas are extremely well differentiated, not associated with HPV, and in pure cases, not associated with metastasis. Papillary carcinomas are not HPV-associated and have a favourable prognosis, although metastases occur [1].

Metastases are therefore limited largely to those of the usual or basaloid subtypes. For repeated recurrences or widespread dissemination, chemotherapy remains the only viable option. There is no established chemotherapy regimen for SCCP, largely because it is rare, and there is difficulty in acquiring cases for prospective studies. The currently used regimens are of limited efficacy [6–8], mostly because of the difficulty in creating large groups of patients with metastatic SCCP that can be entered into randomized clinical trials.

However, there are now abundant reports on chemotherapy for other SCCs, and although results are variable there is evidence that chemotherapy regimens for these lesions can be developed to improve the quality of life and length of survival of patients. Such regimens should be based on multiple agents, and future treatments that have proved of benefit in other SCCs should be targeted.

The present study was developed from the wish to create a central data resource of patients with SCCP (at St George’s Hospital) that could be used to provide material for tissue microarrays (TMAs). This resource could then be used as a possible guide to future biomarkers and measurements of outcome, and as a possible guide for future therapeutic tactics in this rare disease.

The recent interest in topoisomerase I (Topo I) and topoisomerase IIα (Topo II) in SCCs led to the initial investigation of these markers as potential indicators of chemosensitivity in SCCP. The DNA topoisomerases are nuclear enzymes which assist in the coiling and uncoiling of DNA that occurs during transcription and replication. This is achieved through transient breaks in the DNA strands. Topo I breaks and rejoins a single strand, allowing the DNA to unwind along the axis of the complete strand. Topo II breaks and rejoins both strands [9,10]. The enzymes are of interest as they are targets of specific chemotherapy agents. The camptothecins, which include irinotecan and topotecan, stabilize the Topo I/DNA complexes irreversibly and thus prevent the re-ligation of the broken DNA strand. During replication, these DNA single strands are converted to double-strand breaks, which are lethal to the cell. This is an important feature of this group of drugs, as they act as ‘poisons’ rather than inhibitors, causing irreversible damage. Therefore, there is good reason for postulating that the amount of activity shown immunohistochemically by Topo I will be mirrored in the probable effectiveness of any Topo I poison used.

There are two classes of chemotherapeutic drugs that target Topo II; Topo II poisons and Topo II catalytic inhibitors. Topo II poisons act by stabilizing enzyme-DNA complexes leading to DNA breaks, which eventually results in apoptosis. Etoposide, an epipodophyllotoxin, and doxorubicin, an anthracycline, are examples of Topo II poisons.

Catalytic inhibitors act by inhibiting Topo II but at a specific stage of the Topo II catalytic cycle, where both DNA strands are intact. As a result they do not cause DNA breaks. Etoposide and doxorubicin have been used in the treatment of several diseases, e.g. malignant lymphoma, leukaemia, testicular cancer, breast and small cell lung cancer.

In the present study we report an experiment to examine potential markers of chemosensitivity of SCCPs, concentrating on Topo I and II. Ki-67 was also used as a standard method to examine the proliferation rate. As DNA replication is required, a high proliferative fraction is thought to indicate those tumours most likely to respond to therapy.

PATIENTS AND METHODS

In all, 73 tumour samples were retrieved from the Pathology Department of St George’s Hospital between 2001 and 2005. The tumours were obtained from excision biopsy, glansectomy or penectomy, the surgery being conducted by one consultant urologist (N.W.). Sections stained with haematoxylin and eosin (H&E) were reviewed by a consultant histopathologist (C.C.) and areas suitable for TMA were chosen. Data for each sample, including patient age, tumour grade and stage, were recorded. The tumours were graded into three groups, as mild (grade 1), moderate (grade 2) and severe (grade 3). The TNM classification of penile cancer was used [1].

A TMA of five blocks of penile samples was created using a manual microarrayer; 1-mm cores were taken and, where possible, three cores were taken from each tissue sample. Then 4 µm sections were cut from each block, and one section from each block was stained with H&E, to check that the areas taken were representative. Immunochemistry was then performed using the avidin-biotin technique with the following antibodies: Topo IIα (mouse monoclonal antibody, NCL-TOPOIIA, Novocastra, Newcastle-upon-Tyne, UK), Ki67 (rabbit antibody, No: A0047, Dako, Carpinteria, CA, USA) and Topo I (mouse monoclonal antibody, NCL-TOPOI, Novocastra). A dilution of 1:100 was used for Topo I and II, and 1:1000 for Ki67. Examples of the immunostaining are shown in Fig. 1a–c. The sections were then counterstained in haematoxylin for 5 min.

Figure 1.

Expression in SCCP of: a, Topo 1, showing nuclear staining; b, Topo II, with strong nuclear staining; and c, Ki-67. All originals ×40.

Human tonsillar tissue was used as a positive control for all three antibodies, to show positivity in germinal centres and basal epithelium. Negative controls were incubated with BSA buffer in place of the primary antibody. The staining pattern in all three proteins was nuclear. Mitoses were noted to be invariably positive as a good internal positive control for all antibodies.

Sections were scored by a consultant genitourinary pathologist (D.B.) unaware of origin, and semiquantitatively. Each spot on the TMA was given scored as 0–100% in 10% increments, representing the percentage of positively stained nuclei. Sections that could not be evaluated were not scored; this included sections that had fallen out during the immunostaining or cutting process, and cores that did not have adequate tissue, i.e. high levels of fibrous tissue.

The correlations between Topo II, Ki-67 and p53 was evaluated using Spearman’s test. The Shapiro-Wilks test was used to assess the data for normality. The association between Topo I, Topo II and Ki-67 protein expression, and tumour grade and stage, was evaluated using the Kruskal–Wallis test. In all analyses P < 0.05 was considered to indicate statistical significance.

In the analyses the verrucous tumours were excluded from grade and stage analyses as they are by definition low-grade and low-stage. The basaloid and usual type were included in these analyses as they are all capable of metastasis and progression.

RESULTS

Because of the variable loss of cores from the TMA blocks, data were available on a variable number of tumours, depending on the group used for analysis. However, loss of cores was generally low (<5%); 73 tumours were available for analysis, including 56 usual, seven basaloid, one warty carcinoma and nine verrucous carcinomas. Verrucous carcinomas were excluded from the grade and stage testing, as they are by definition grade 1 and low-stage and have a benign natural history, and might have biased the results. Of the remainder, 13 were of grade 1, 29 were grade 2 and 21 were grade 3.

Shapiro-Wilks testing for all the results showed that they were unlikely to be from a normal distribution, and therefore comparisons were made using the Kruskal–Wallis test.

Ki-67 expression varied from 10% cellular positivity to 90%. Expression was strongly correlated with the grade of tumour examined, being higher in those of high grade (Fig. 2a). There was a significant difference between all three grades of tumour (grade 1 and 2, P < 0.001; grade 1 and 3, P < 0.001; grade 2 and 3, P = 0.002) However, Ki-67 showed no association with stage of tumour. There was a striking association with tumour type, with basaloid having a high proliferative index and verrucous, as expected, a low proliferative index, with the usual type being intermediate (difference usual vs basaloid, P < 0.001; usual vs verrucous, P = 0.020; basaloid vs verrucous, P < 0.001).

Figure 2.

Box-and-whisker plots of: a, Ki-67 expression in different grades of tumour; b, Topo I expression in different grades of tumour; c, Topo II expression in different grades of tumour; and d, Topo II expression in different types of penile carcinoma. B, basaloid; U, usual type; and V, verrucous.

Topo I expression varied from 0% to 90% of positive cells. There was no association between Topo I expression and grade (Fig. 2b), stage or type of tumour in any of the analyses, and nor was there any correlation between Topo I expression and the other two markers in the analyses. The expression appeared to be entirely unpredictable and independent of any tumour variable that could be analysed.

Topo II expression in each case varied from 0% to 60% positivity of cells examined. Generally Topo II expression mirrored the Ki-67 expression, and was strongly associated with grade and type of tumour, but not with stage. However, the association with grade was not as strong as that of Ki-67 (Fig. 2c), with significant differences between grade 1 and grades 2/3 (P < 0.001) but not between grades 2 and 3 (P = 0.300). However, there were marked differences between the different types of tumour (Fig. 2d; usual vs basaloid, P = 0.004; usual vs verrucous, P < 0.001; basaloid vs verrucous, P < 0.001).

Relationships between the different markers were examined; there was a very strong positive correlation between the Ki-67 fraction and Topo II expression, with a Spearman’s rank correlation coefficient of 0.83 (P < 0.001).

DISCUSSION

The rarity of SCCP makes it difficult to conduct clinical trials into the effectiveness of chemotherapy in this malignancy. It is therefore important that biomarker analyses are used to indicate the likelihood of a given regimen being successful, so that prospective clinical trials can guide potential future therapies. We have generated, to our knowledge, one of the largest molecular studies of SCCP, to assist in long-term analyses of this disease and assist in future treatment regimens.

Chemosensitivity assays can have two roles; possibly the most important is to report the lack of a specific marker, indicating that a particular therapy is unlikely to be useful in the clinical trial setting. The second role is to report high expression in specific tumours, and to attempt to correlate expression with response in a particular clinical trial. Studies of this are in their infancy, although the use of herceptin treatment in tumours which overexpress the cERB2 gene have shown the way forward for this type of investigation.

The tissue evaluation of Topo I for the effect of Topo poisons appears theoretically to be ideal in this type of study, because expression in a given tissue could be directly correlated with response. The heterogeneity between tumours in the expression of Topo I has been recognized previously in a study where many genes were examined in a wide variety of tumour types [11]. In practice, these other genes might affect the response to Topo I, and include metabolism enzymes (CES2, UGT1A1, CYP3A4, CYP3A5), and cellular transporters of the topoisomerase poison (ABCB1, ABCC1, ABCC2, ABCC3, ABCC5, ABCG2). However, as an initial assessment of a candidate chemotherapeutic regimen, an analysis of the protein that is being poisoned or inhibited seems a logical analysis, as if this is deficient, then assessment of secondary proteins will be purposeless. Topo I inhibitors are likely to only work well in high cycling cells [10,12], therefore Ki-67 and Topo II provide ancillary information on the likelihood of response. Only if Ki67 and/or Topo II are high is there a likelihood of a Topo I poison response.

It is also valuable to examine the findings in similar tumours that are more common than SCCP, and compare the data found in those tumours. SCCs occur in many organs, including the cervix, the bronchi, oesophagus, skin and head and neck carcinomas. The most well studied group in this latter category are the head and neck tumours arising from oral squamous mucosa. One study showed that Topo I showed increased expression in oral squamous tumours of high grade [12]. These results are different from the present, highlighting how findings in one tumour cannot necessarily translate to another of similar type. Topo I poisons have been used in head and neck tumours, with mixed results, although similar to most other trials, there are no comparisons with pathological marker expressions [13].

There are more extensive reports on the other two markers used in the present study. Topo II is consistently positively associated with tumour grade in a wide variety of squamous tumours [14–16]. Ki-67 is the only marker used in this study that has previously been examined in SCCP [17]; those results were similar to the present, but numbers were low in the only study examining all tumours. Intense basal expression of Ki-67 was reported in verrucous SCCP [18]; it has been shown to be strongly associated with grade and prognosis in most other squamous tumours [13–15,19].

SCC of the cervix is probably the tumour which best resembles SCCP. Both are HPV-associated genital cancers, although the prevalence of cervical cancer is much higher than SCCP. In cervical SCC, topotecan combined with cisplatin was recently shown to give a survival advantage over cisplatin alone [20], but did not reduce the quality of life [21]. There are no data on the expression of Topo I in cervical carcinoma.

We conclude that SCCP does not appear to show the same pattern of Topo I expression as other squamous tumours that have been investigated. By contrast, Topo II appears to be closely associated with Ki-67, reflecting the findings in other studies. These data might be important in the design of prospective clinical trials, and help to explain any possible variable response to Topo I poisons. Therefore we suggest that in future trials, Topo I poisons are likely to be effective only in selected high-grade tumours.

ACKNOWLEDGEMENTS

This project is supported by The Jean Shanks Foundation. DB and SK are supported by The Orchid Appeal.

CONFLICT OF INTEREST

None declared.

Ancillary