Background/Aims—Loss of function of the retinoblastoma (Rb) tumour suppressor gene, located on chromosome 13, is common in many inherited and sporadic forms of cancer. Inactivation of its gene product by oncogenic human papillomaviruses (HPV) plays a key role in the genesis of cervical cancer. It has been shown previously that non-melanoma skin cancers of renal transplant recipients and immunocompetent patients with skin cancer also frequently harbour potentially oncogenic HPV types. This study aimed to examine the integrity of the Rb gene in histologically confirmed squamous cell carcinomas (SCCs) from renal transplant recipients and immunocompetent patients with skin cancer.
Methods—Loss of heterozygosity (LOH) at the Rb locus was examined in 13 histologically confirmed SCCs using the D13S153 microsatellite marker, which is located in exon 2 of the Rb gene. Loss of a second marker, D13S118, distal telomerically to the Rb gene at 13q14.3 was also analysed.
Results—Of the 13 HPV associated SCCs examined 11 were informative (two SCCs were homozygous for both microsatellite markers). LOH at the D13S153 locus was found in four of the 10 informative SCCs and LOH at the D13S118 locus was found in five of the 11 informative cases. Overall, seven of the 11 informative cases showed LOH at one or other locus. This represents a high degree of chromosomal instability in these tumours. The expression of the Rb gene product in the 11 informative cases was analysed immunohistochemically. Expression of Rb was detected in 10 of the 11 SCCs examined. No correlation between the HPV status of the tumours and the expression of Rb was found. Although the only SCC not to express Rb also demonstrated LOH at the D13S153 locus, the remaining SCCs that had LOH at 13q14 were able to express Rb.
Conclusion—Another tumour suppressor gene located at 13q14 might be responsible for the genesis of these tumours.
Statistics from Altmetric.com
The rare childhood cancer retinoblastoma is partially caused by the inactivation of the Rb tumour suppressor gene, located on the long arm of chromosome 13 at 13q14.2. Its gene product is a nuclear phosphoprotein that plays a crucial role in the regulation of cell proliferation, differentiation, and signal transduction.1 Loss of Rb function leads to the development of a wide range of inherited and sporadic forms of cancer.2, 3 The inactivation of Rb can occur through mutation of the gene or by interaction with the oncoproteins of DNA tumour viruses.
Loss of heterozygosity (LOH) is regarded as an important genetic mechanism in the development of malignant neoplasia and is responsible for inherited retinoblastoma. LOH at 3p and 17p has been described previously in cutaneous squamous cell carcinomas (SCCs).4, 5 In SCCs of the head and neck, LOH is frequently detected at 13q14.6 Deletions of portions of chromosome arm 13q or mutations of genes located on this arm are found in a large number of different human cancers. Most notably, the Rb gene and BRCA2 (a tumour suppressor gene associated with a predisposition to breast cancer), which is located at 13q12.q3.7 In cervical carcinomas, the E7 oncoproteins of human papillomavirus type 16 (HPV-16) and HPV-18 bind competitively to Rb, thus inhibiting the binding of Rb to its normal physiological partners.8
There is no evidence to suggest that the E7 protein of the Epidermodyplasia verruciforms (EV) HPV types prevalent in cutaneous SCCs acts in a similar manner to the E7 oncoprotein of HPV-16 and HPV-18. Therefore, another mechanism of Rb inactivation may occur. Here, we describe a study to examine the integrity of the Rb gene in cutaneous SCCs. Allelic loss at the D13S153 locus (located in exon 2 of the Rb gene) and at the D13S118 locus (distal telomerically to Rb at 13q14.3) was examined in 13 histologically confirmed cutaneous SCCs. The effect of LOH on Rb expression in these tumours was also examined immunohistochemically.
Materials and methods
SAMPLES FOR LOH ANALYSIS
Thirteen histologically confirmed SCCs and matched peripheral blood (normal controls) were chosen for LOH analysis. SCCs were sectioned and microdissected to remove any contaminating normal tissue. A fresh sterile scalpel blade was used each time a new case of SCC was dissected. The surface of the stereomicroscope was cleaned with ethanol between each case. Total genomic DNA was isolated from microdissected tumour tissue by proteinase K digestion (0.5 mg/ml in 100 μl of lysis buffer)9, 10 at 55°C for 72 hours. After incubation, the proteinase K was inactivated by heating to 95°C for 10 minutes. The tubes were then centrifuged at 12 000 ×g for 30 seconds. The remaining supernatant was used for the LOH analysis. Total genomic DNA was extracted from matched peripheral blood using the Isolate II™ kit (Cruachem, Glasgow, Scotland).
PCR DETECTION OF MICROSATELLITE MARKERS D13S153 AND D13S118
Two CY5 labelled D13S153 and D13S118 microsatellite markers (fig 1) were amplified using the primers described previously by Hudson and Weissenbach et al.11, 12 The PCR reaction mixture contained 1.5 mM MgCl2, 200 μM dNTPs, 1 μM of the 1312R/TG primer pair, and 0.25 μM of the AFM058xd6a/m primer pair. The reaction was performed in an automated DNA thermal cycler (Hybaid, Ashford, Middlesex, UK) programmed for 30 cycles of denaturation at 94°C for one minute, annealing at 50°C for one minute, and polymerisation at 72°C for one minute. Approximately 100 ng of total genomic DNA extracted from microdissected tumours and peripheral blood was used as template for all PCR protocols. All PCR protocols were carried out using Taq polymerase (Promega, Madison, Wisconsin, USA) (1.5 U/50 μl reaction). The size of the PCR product generated by means of the 1312R/TG primer pair is 187–201 bp and that generated by means of the AFM058xd6a/m primer pair is 212–236 bp.
ANALYSIS OF LOSS OF HETEROZYGOSITY
Analysis of LOH at the D13S153 and D13S118 loci was performed on an ALFexpress™ automated DNA sequencer (Pharmacia, Uppsaala, Sweden) using the Allele links™ software package (Pharmacia). CY5 labelled PCR products were run on 6% (wt/vol) polyacrylamide gels and measured by laser excitation of the CY5 label. PCR product size was determined by comparison with CY5 labelled size markers (50–500 bp). The intensity of CY5 labelled PCR bands representing each allele was calculated as a peak area for each band. Using the peak areas determined, allele ratios for each tumour sample and its corresponding normal sample were calculated and used to determine LOH at the D13S153 and D13S118 loci.
CY5 labelled PCR products were diluted between 1/2 and 1/10 in loading dye, depending on the concentration of each product. Before loading, the samples and the CY5 labelled size marker were heated to 95°C for five minutes and immediately quenched on ice. A DNA negative control was included in each gel. Two sets of PCR products from each case were analysed on separate ALF gels and a mean peak ratio calculated and used to determine allelic loss. The Allele links™ software package automatically calibrates the peak height between the normal and the tumour sample and the determination of the allele loss is based on the ratio of the tumour sample divided by overall ratio of the normal sample. The cut off point for determining LOH on the ALFexpress™ automated DNA sequencer used in our study was calculated previously to be 0.74 (that is, 99.5% of paired normal cases would give an allele ratio of 0.74 or greater).13 Therefore, tumour samples with an allele ratio of 0.74 or less when compared with the normal allele ratio were deemed to show allelic loss.
The 13 cutaneous SCCs analysed for LOH at the D13S153 and D13S118 loci were analysed immunohistochemically for Rb expression.
Immunohistochemical analysis of Rb expression was performed using the RB1 antibody (Dako, Glostrup, Denmark). All tissues were fixed in 10% unbuffered formalin and embedded in paraffin wax. Serial sections (4 μm thick) were cut and stained with the RB1 antibody, using the avidin–biotin method. Briefly, sections were subjected to microwave antigen retrieval and were placed in 0.01 M sodium citrate (pH 6.0) and microwaved at 850 W for 22 minutes. All sections were then incubated with primary antibody for 30 minutes, followed by two washes in Tris buffered saline (TBS) (four minutes each) and visualisation of the antibody–antigen reaction with the Duet system (Dako). The peroxidase reaction was developed using diaminobenzidine for three minutes and sections were counterstained with Harris's haematoxylin. Positive controls were included for each test. Negative controls (a serial section with the primary antibody omitted) were also included. Results were recorded independently, assessing the proportion of positive cells and the intensity of staining relative to the positive controls.
LOSS OF HETEROZYGOSITY AT THE D13S153 AND D13S118 LOCI IN CUTANEOUS SSC
The rate of informativity was 77% (10 of 13) at the D13S153 locus and 85% (11 of 13) at the D13S118 locus. LOH at D13S153 and D13S118 was determined by applying a previously established LOH cut off point of 0.74.13
At the D13S153 locus, four of the 10 informative cases showed allelic loss (fig 1). In all cases the degree of loss was approximately 50% or greater. At the D13S118 locus, five of the 11 informative cases demonstrated allelic loss (for example, fig 2). However, the degree of loss was less than 50% in four of the five cases. Only one SCC showed a large degree of allelic loss (89%) at D13S118. Overall, seven of the 11 informative cases demonstrated allelic loss at either D13S153 or D13S118. Only two SCCs showed loss of both markers (fig 1).
IMMUNOHISTOCHEMICAL DETECTION OF Rb EXPRESSION IN CUTANEOUS SCCS ANALYSED FOR LOH
Table 1 shows the relation between allelic loss and Rb expression. Of the HPV associated SCCs analysed for LOH, 12 of 13 showed Rb expression in the invasive tumour. Of the four samples that demonstrated LOH at the D13S153 locus, three retained the ability to express Rb (fig 3). Overall, six of the seven SCCs that demonstrated LOH at either locus were found to express Rb (table 1).
LOH has been shown to play a causal role in the development of basal cell carcinoma, malignant melanoma, and various other types of skin cancer.14, 15 Previously, LOH in cutaneous squamous cell carcinomas has primarily been detected on chromosome 9, although a minimal amount of LOH has been seen at 3p, 13q, and 17p.16 Non-random allelic loss at 3p, 11p, and 13q has been associated with HPV mediated immortalisation of keratinocytes and the subsequent loss of terminal differentiation.17 Actinic keratoses, a UV related precancerous lesion of the skin, has also been shown to exhibit a high degree of LOH at these same chromosomal regions.16
The cut off point for determining LOH varies widely in other studies. Some researchers have used a cut off point of 0.50 (that is, a 50% reduction or greater).18, 19 This cut off point was established on non-microdissected tissue and assumes that tumours with no contaminating normal tissue will give an allele ratio of 0.00. However, when the clonality of tumours is questionable and accumulated genetic damage is accrued as the tumour progresses this may not always be the case. Other researchers have applied a cut off point of 0.70 or 0.80; however, no explanation for choosing either was offered.20, 21 Previous studies have also determined LOH by densitometric analysis of PCR bands, and this technique is not as sensitive as fluorescent based measurements. For the purposes of LOH determination, a cut off point of 0.74 was used, which had been determined previously on the same ALFexpress DNA sequencer used in our study.13 This cut off point was determined by analysing the variation among paired normal samples, and it was determined that 99.5% of normal cases would give an allele ratio of 0.74 or greater.
Our results show that LOH at 13q14.2–13q14.3 is frequent in cutaneous SCCs of renal transplant recipients. Overall, seven of the 11 informative cases showed loss of one or other microsatellite marker. This degree of allelic loss is far greater than that reported in any other study of LOH in cutaneous SCCs. Previous studies have estimated the degree of LOH in cutaneous SCCs to range from 7% to 28%, predominantly on chromosome 9.4, 16
Despite the high degree of allelic loss in and around the Rb tumour suppressor gene demonstrated in these tumours, loss of Rb protein expression was seen in only one of the 13 SCCs examined for LOH. This suggests that another tumour suppressor gene in the region 13q14.2–13q14.3 could be inactivated by the allelic loss shown in these tumours. In contrast, laryngeal tumours, oesophageal cancer, and Merkel cell carcinomas with allelic loss at 13q14.2 generally show abnormal staining patterns for Rb when examined immunohistochemically.22–24
Frequent allelic loss at 13q14.2–13q14.3 has been demonstrated in a wide variety of other cancers, in particular SCC of the head and neck, lung cancer, breast cancer, and oral SCC.6, 25–27 In many cases, Rb does not appear to be inactivated, suggesting that another tumour suppressor gene near the D13S153 and D13S118 loci may play a role in the development of these tumours (if Rb is not mutated or the loss of one allele does not affect the cell through haploinsufficiency).6, 27, 28 In B cell chronic lymphocytic leukaemia (B-CLL), allelic loss at 13q14.3 is frequent, although normal Rb expression is retained.29 Two putative tumour suppressor genes, leu1 and leu2, were thought to be crucially lost in all cases of B-CLL; however, it has since been shown that this is not the case so that leu1 and leu2 may not be tumour suppressor genes.30 A further 46 expressed sequence tags have been assigned to the region 13q14.3, and any one of these could be a candidate tumour suppressor gene, the loss of which might contribute to a wide variety of human cancers.31
Amplification of 11q13 in SCC of the head and neck is well known and may target the cyclin D1 gene.28 In oesophageal cancer, it has been suggested that the Rb pathway could be circumvented by either Rb inactivation or cyclin D1 overexpression.32 LOH at chromosome 9p21, where the p16INK4 locus is located, has been detected in 71% of sporadic melanomas.33 The increase in proliferative potential associated with Rb inactivation can be seen in cells with p16INK4 inactivation through mutation or altered methylation, and in cells with overexpression of cyclin dependent kinase 4. Thus, Rb itself may be expressed normally but its function in regulating the cell cycle might be diminished by other means.34
In summary, the results of our study show that a high degree of chromosomal instability at 13q14.2–13q14.3 exists in cutaneous SCCs of renal transplant recipients. The expression of Rb in the tumours exhibiting allelic loss in and around the Rb locus appears to be normal, suggesting that another as yet unknown tumour suppressor gene could be inactivated by the allelic loss observed.
We would like to express our sincere thanks to the medical staff of the dermatology clinic at Beaumont Hospital for providing samples from patients with skin cancer; we also thank C O'Flatharta for technical assistance. We would also like to acknowledge the Irish Cancer Society and the Royal College of Surgeons in Ireland for their financial support.
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.