Main

The prediction that the adenomatous polyposis coli gene (APC) might have tumour suppressor activity has been supported by the fact that carcinomas developed in patients with familial adenomatous polyposis exhibited a high frequency of loss of heterozygosity in the chromosome 5q21–22 region—typical APC locus.1, 2, 3, 4 Loss of heterozygosity with respect to APC usually occurs in the normal allele. This fact suggests that both alleles of the APC gene could be inactivated by germline mutation and the somatic allele lost in colorectal carcinomas. APC protein is mutated in familial adenomatous polyposis patients and in sporadic colorectal tumours. Familial adenomatous polyposis coli is an autosomal dominant inherited disease that predisposes carriers to a high probability of colorectal carcinoma.5, 6, 7 The function of APC is still unclear. Recent evidence indicates that the APC protein functions as an intermediary between cell adhesion molecules and the cytoskeleton, that it is involved in signal transduction and in controlling apoptosis.8, 9, 10 Another study suggested that APC acts as a gatekeeper of colonic epithelial cell proliferation. Inactivation of this gatekeeper initiates neoplastic growth by creating a permanent imbalance between cell division and cell death.11 APC expression in cutaneous tumours has not been investigated thus far. The main goal of this study was to investigate APC expression, microsatellite instability in APC locus and promoter methylation status of APC in the progression of melanocytic tumours.

Materials and methods

Tumour Samples

The material investigated consisted of specimens from 152 melanocytic lesions including 36 lentigos, 88 malignant melanomas with known follow-up and 28 melanoma lymphatic metastases. Correlating non-neoplastic reference tissue was also investigated. The tumours were located as follows: head and neck 32, trunk 39, upper extremity 42, lower extremity 39. There were 79 male and 73 female patients, whose age ranged between 21 and 83 years, averaging 56 years. All primary tumours were classified according to Breslow,12 showing the following pT distribution: pT1–24, pT2–18, pT3-20 and pT4-26.

Immunohistochemical Demonstration of APC Expresssion

APC gene expression was analysed by applying rabbit polyclonal antibody C-20 raised against a peptide mapping at the carboxy terminus of APC of human origin (Santa Cruz Biotechnology Inc.).

The immunohistochemical reactions in the paraffin-embedded tumour tissue were carried out using the Stravigen Multilink kit (Biogenex Laboratories). The histological sections were mounted on uncoated slides. They were deparaffinated by xylol and then transferred to a descending alcohol series and rinsed with distilled water. Before incubation with primary antibodies, the sections were heated for 10 min on a heat plate (85°C) in citrate buffer (pH=6). Afterwards, incubation with the primary antibodies was carried overnight at 4°C at an antibody concentration of 1:50. Then, the histological specimens were rinsed with Tris buffer solution and incubated at room temperature with link (Stravigen Multilink, Biogenex Laboratories) for 45 min. After the detection reaction was performed using a label (Stravigen Multilink, Biogenex Laboratories) in combination with chromogen fast red (Biogenex Laboratories), the nuclei were counterstained with haematoxylin. In control reactions, primary antibodies were omitted. Sections were evaluated using the CAS200 image analysis system and results were expressed as percentages of immunolabelled cells (indices).

Normal control for immunohistochemical staining was positivity within the epidermis.

DNA-Extraction

Paraffin-embedded tumours and corresponding normal tissue were mounted on glass slides and stained with haematoxylin and eosin. Microdissection was performed on paraffin sections stained briefly with methylene blue. Areas of interest were scraped off the slide and subjected to digestion with proteinase K 20 mg/ml (Qiagen, Valencia CA, USA). DNA was extracted with the QIAmp DNA Mini Kit (Qiagen, Valencia, CA, USA) according to the manufacturer's recommendations. The eluted DNA was suspended in 30 μl H2O and stored at −20°C.

Microsatellite Analysis

Microsatellite analysis of locus D5S346 (APC) was performed according to the protocol of Spirio et al13 with modification. Primers applied had following sequence: forward 5′-ACTCACTCTAGTGATAAATCG-3′ (labelled with fluorescent dye HEX) and reverse 5′-AGCAGATAAGACAGTATTACTAGTT-3′. Oligonucleotides for PCR were obtained from Perkin-Elmer (Weiterstadt, Germany). Optimal PCR results were achieved by an initial 95°C denaturation step (7 min), followed by 34 cycles at 94°C (1 min)//55°C (1 min)//72°C (1.5 min), followed by a final extension step at 72°C for 5 min. The PCR mix consisted of 10 × PCR buffer (Perkin-Elmer, Weiterstadt, Germany), 100 pg primer, 10 μM of each dNTP (Perkin Elmer, Weterstadt, Germany), 2 U of Ampli Taq Gold (Perkin-Elmer, Weiterstadt, Germany). The final volume was 50 μl. Resulting PCR products were separated on 3% agarose gels and stained with ethidium bromide. Negative controls were included in each PCR run. Automatic analysis of PCR products was then performed on ABI 310 DNA-Analyzer (Perkin-Elmer, Germany) with GENESCAN software. Internal standard GENESCANTM-350, TAMRA (Perkin-Elmer, Germany) was added to check the size of the PCR fragments analysed. The peak pattern of tumour tissue was compared to that of normal reference tissue. Additional peaks in tumour tissue when compared to normal tissue indicated microsatellite instability. Loss of peaks in tumour tissue compared to normal tissue indicated loss of heterozygosity.

Analysis of APC Promoter Methylation

An amount of 1 μg of DNA isolated from tumour tissue and non-neoplastic reference tissue was treated with sodium bisulphite using the CpGenome DNA Modification Kit (Intergen Company, Oxford, UK) according to the manufacturer's instructions. A volume of 2 μl of modified DNA (1/5 volume) was used for PCR amplification. Primers were APC sense 5′-GGTATATTTTCGAGGGGTACT-3′ and APC antisense 5′-TTCCCGACCCGGACTCCGC-3′ for methylated sequences, APC sense 5′-TGTGAGGGTATATTTTTGAGGGGTAT-3′ and APC antisense 5′-CTTCTCTCTCCACTTCCCAACCCA-3′ for non-methylated sequences. The PCR conditions for methylated APC-primers were as follows: 95°C hot start × 15 min, then 40 repetitive cycles of denaturation (95°C × 45 s), annealing (56°C × 60 s), extension (72°C × 60 s) followed by a final 7 min extension at 72°C. The PCR conditions for nonmethylated APC-primers were as follows: 95°C hot start × 15 min, then 40 repetitive cycles of denaturation (95°C × 45 s), annealing (56°C × 30 s), extension (72°C × 60 s) followed by a final 7 min extension at 72°C. The PCR fragments were separated on a 3% agarose gel.

Statistics

After being scanned into the spread sheet (Microsoft Excel 97), where they were made available to the statistics program via an ODBC driver, the data were analysed using the statistical analysis system (SAS, Version 7.5). on an IBM-compatibile PC under Windows NT 4.0. The median and ranges were calculated for all parameters. Mann–Whitney U-tests were used for the group comparisons between pT groups). Survival rates were calculated according to the Kaplan and Meier method.14 Cox regression was the multivariate method used for predicting the survival rate based on several parameters.15

Results

APC Expression

APC was present in all lentigos investigated and APC expression ranged between 5 and 69%. Out of 88 primary tumours, 75 were APC-positive. Index of APC-labelled cells oscillated in primary melanomas between 0 and 98%. APC expression was detected in 16 out of 28 melanoma lymphatic metastases. APC index in this group reached values between 0 and 52% (Table 1, Figure 1a–c).

Table 1 APC expression in diagnostic groups investigated
Full size table
Figure 1
figure 1

Positive cytoplasmic reaction for APC in primary melanoma and in epidermis (a, b) and in metastatic melanoma cells in lymph node (c), × 400.

Full size image

Comparison of Diagnostic Groups Investigated

Highly significant differences with regard to APC expression were found between lentigos and primary melanomas as well as between lentigos and melanoma lymphatic metastases (P<0.01). Comparison of primary melanomas and melanoma lymphatic metastases did not show any significant differences (P>0.05)

Relation between pT Groups and APC Expression

APC-positive cells were found in 21/24 pT1, 18/18 pT2, 16/20 pT3 and 20/26 pT4 tumours. APC index reached maximally 98% in pT1 and 70% in pT2 melanomas. In pT3 and pT4 tumours APC index peaked at 52% and 48% of tumour cells, respectively (Table 2). Statistically significant differences were found between pT1 and pT>1 groups for APC expression (P<0.05).

Table 2 APC expression in primary melanomas
Full size table

Microsatellite Instability Analysis in APC Locus (D5S346)

Altogether 29 cases were investigated for the presence of microsatellite instability in APC locus. Microsatellite instability was found in 5/29 cases and loss of heterozygosity only in 1/29 case (Figures 2 and 3).

Figure 2
figure 2

Example of microsatellite instability in APC locus.

Full size image
Figure 3
figure 3

Example of loss of heterozygosity in APC locus.

Full size image

APC Expression in Cases without Microsatellite Instability in APC Locus

APC-positive cells were found in 17/23 melanomas without microsatellite instability. APC expression ranged between 0 and 63% of tumour cells, reaching median of 15% (Figure 4).

Figure 4
figure 4

Boxplot demonstrating APC expression in melanomas without microsatellite instability in APC locus.

Full size image

APC Expression in Cases with Microsatellite Instability in APC Locus

Positive reaction for APC was observed only in 1/5 melanomas with instable APC locus. APC expression ranged in this tumour groups between 0 and 9% (Figure 5).

Figure 5
figure 5

Boxplot demonstrating APC expression in melanomas with microsatellite instability in APC locus.

Full size image

Comparison of APC Expression in Melanomas with and without Microsatellite Instability

Difference in APC expression between melanomas with and without microsatellite instability in APC locus was significant (P=0.0303).

Analysis of Promoter Methylation Status of APC

Methylation of APC promoter was detected in 6/10 APC-negative melanomas without metastatic events associated with them and in 9/10 APC-negative melanoma lymphatic metastases (Table 3, Figure 6).

Table 3 Frequency of APC promoter methylation in APC-negative primary and metastatic melanomas
Full size table
Figure 6
figure 6

Detection of methylated and nonmethylated APC in normal and melanoma tissues in 3% agarose gel. T1—tumour, N1—normal tissue, Non M primer—nonmethylated primer, M primer—methylated primer.

Full size image

Prognostic Significance of APC in Malignant Melanomas

Multivariate analysis

In multivariate analysis in both models, forward and backward APC proved to be a significant factor in predicting outcome (P<0.05).

Kaplan–Meier Survival Curves

Analysis of Kaplan–Meier survival curves showed that about 45% of APC-positive patients (index >0%) and none of the APC-negative patients survived 5 years. Difference between both survival curves was significant (P<0.05—log-rank test) (Figure 7).

Figure 7
figure 7

Kaplan–Meier survival curves for APC-positive and -negative melanoma patients.

Full size image

Discussion

Our study demonstrated for the first time that the altered expression of APC gene had the prognostic impact in primary melanomas. Thus far, alternations of this gene had been mostly considered to be associated with colon disease. Now that we have shown, altered APC expression in tumours outside the gastrointestinal tract allows us to postulate the general role of this gene in tumorigenesis. Similar situations are encountered in genes involved in mismatch repair. While most reports have demonstrated the role of mismatch repair genes in hereditary nonpolyposis colon cancer, it has also been shown to play a crucial role in the progression of several sporadic tumours.16, 17, 18

As mentioned in the introduction, the role of APC in tumour biology is still unclear, although there has been some speculation concerning the function of this gene.8, 9, 10, 11 The relation between the APC gene and proliferation has been investigated comprehensively. APC is known as negative regulator of proliferation (gatekeeper). Loss of expression of this gene reportedly leads to uncontrolled proliferation and dysharmony between proliferation and apoptosis and, ultimately, to accelerated tumour progression.

In this study, we demonstrated a significant difference in APC expression between melanomas with and without microsatellite instability in the APC locus. Somatic mutations and lost of heterozygosity of APC were reported for colorectal, gastric, pancreatic and cholangiocarcinomas.19, 20 To our knowledge, ours is the first investigation of loss of heterozygosity and microsatellite instability in the APC locus in malignant melanomas. In total, 80% of cases with an instable APC locus did not show any APC expression on the protein level. Aberrant APC promoter methylation has not yet been demonstrated for malignant melanomas. We concentrated our attention on APC-negative primary and metastatic melanomas to find possible correlation between loss of APC expression and promoter methylation status.

The relation between loss of APC expression and promoter methylation has been demonstrated for several sporadic tumours. Aberrant methylation of APC was associated with a loss of gene function as was demonstrated for 27% of gallbladder carcinomas.21 In prostate cancer, methylation of cyclin D2 correlated significantly with methylation of APC and with clinicopathological features of poor prognosis.22 The methylation frequency of APC correlated in prostate cancer with Gleason score.23 In breast cancer, methylation of APC was independently associated with poor outcome and was even more powerful than standard prognostic parameters.24 In colorectal cancer, the frequency of aberrant promoter methylation of APC in malignant tissues was significantly higher than in nonmalignant tissues, but methylation status of APC had no clear relationship with clinical parameters.25 Our finding of high frequency of APC methylation in malignant melanomas partly explains the mechanism responsible for the loss of APC expression and indirectly the influence of APC on the disease prognosis.

The other mechanism by which APC can influence prognosis in melanoma patients is connected with cell adhesion molecules. Colon cancer cells with mutant APC contain abnormally high levels of intracellular β-catenin. When full-length APC was added to these cells, cytoplasmic pools of ß-catenin were destabilized and eliminated.9, 26, 27 The functional implications for the colocalization of APC protein with cytoskeletal microtubules are unknown. One intriguing possibility in light of recent finding that APC associates with catenins is that APC could relay signals from catenins to the cytoskeleton. This might, in turn, interface with extracellular signals communicated through the cell adhesion molecule E-cadherin, which also binds catenins.

The main reason that melanoma therapy is unsuccessful is its metastatic spread. Strong relations between APC and cell adhesion molecules suggest that APC plays an indirect role in tumour migration. In conclusion, we submit that APC should be regarded as an important new factor in progression of malignant melanoma.