Abstract
Background: Platinum-based chemotherapy is the treatment of choice for metastatic urothelial carcinoma, which is limited by primary and secondary resistance of the tumour. The Cu2+ transporting beta polypeptide ATPase (ATP7B) is believed to play a role in this resistance. The aim of this study was to screen the ATP7B gene for mutations and loss of heterozygosity in bladder cancer and to evaluate their impact on chemotherapy resistance. Materials and Methods: DNA extracted from 17 patients with metastatic bladder cancer was analyzed by DNA sequencing, and microsatellite analysis. Results: We found 12 non-synonymous mutations and 20 synonymous mutations out of which 11 and 15, respectively, have not been previously described. Results were correlated with response to platinum-based chemotherapy: 65% of patients exhibited LOH of the ATP7B locus on chromosome 13q14.3, with a tendency to have a better response to chemotherapy. Conclusion: Although resistance is complex, LOH at the ATP7B locus might be useful in predicting chemotherapy response and needs further evaluation.
Bladder cancer is the fourth most common malignancy of males and the eighth most common of females and leads to an estimated 170,000 deaths per year worldwide (1). Bladder cancer exhibits an urothelial differentiation in 90% of cases in Western countries. Metastasized bladder cancer has an unfavourable prognosis due to rapid progression.
Limitations of platinum-based chemotherapy. Patients with metastatic disease have an average survival of 4-6 months under palliative therapy. Survival more than doubled to 12-14 months when platinum-based therapies were introduced in the late 1980s. The current standard therapy is a combination of gemcitabine and cisplatin. The median survival of patients under this regimen is 14 months, with acceptable toxicity (2). Unfortunately, for as yet unknown reasons, only 40% to 50% of all patients will respond to chemotherapy and benefit from this approach.
Mechanisms of resistance. There are different factors which potentially influence the primary efficacy of platinum on the cellular level: The transport from the extracellular compartment into the cytosol, intracellular protein binding and transport, sequestration into cellular compartments and efflux from the cell. The platinum derivatives cisplatin and carboplatin are known to be shuffled into and out of the cytoplasm by proteins involved in the metabolism of copper, a physiological heavy metal. Cisplatin is transported into the cell by the copper transporter hCTR1; if the protein is defective, cisplatin cannot be accumulated in the cell, leading to diminished efficacy of cell toxicity. Overexpression of this protein has a beneficial effect due to the resultant intracellular accumulation of cisplatin (3). Other uptake mechanisms such as endocytosis, or via the organic cation transporters (OCT) of the solute carrier family (SLC22) are further candidates for cisplatin uptake. In the cells the drug is bound to proteins such as ATOX1, metallothioneins and glutathione. These complexes can be excreted by GS-X transporters (belonging to the ATP-binding cassette proteins, previously known as multi-drug resistance protein group), possibly leading to increased inactivation and resistance (4). The copper transporters ATP7A and especially ATP7B are also candidates for excretion and inactivation.
ATP7B – more than a copper transporter. ATP7B was discovered and characterized in the context of Wilson disease, an inherited metabolic syndrome. The gene is located on chromosome 13q14.3 and encodes a copper-transporting transmembrane protein, a member of the P-type ATPases family (5). ATP7B consists of six metal binding domains, a hinge-domain, two domains acting as ATP-binding domain, eight transmembrane domains that form the copper translocation pathway and a COOH-terminal region (6). The majority of copper is metabolised in the liver, where ATP7B transports copper to the trans Golgi network which leads to the incorporation of copper into copper-dependent proteins such as ceruloplasmin. In the excretory pathway, ATP7B is inserted in intracellular vesicles, which fuse with the canalicular membrane, leading to the excretion of excess copper into the biliary system. ATP7B is thus important for copper homeostasis and for biosynthesis. Additionally ATP7B has the ability to transport not only copper but also other metals such as iron or platinum (7). Cells resistant to cisplatin have low intracellular copper levels (8). Cross-resistance for copper and cisplatin was shown in cisplatin- and copper-resistant cells respectively (7, 9). Others have supported the theory that elevated expression of ATP7B leads to cisplatin resistance (10-13). Studies of Wilson disease show that a single mutation of the ATP7B gene can lead to a non-functional protein. Around 400 mutations are known with a high geographical variation (14). Around 1% of the general population are heterozygotes for mutations inactivating ATP7B, while a larger number of the population is likely to harbour mutations of unknown significance (15). Additionally, it was shown that chromosome 13q, where the ATP7B gene is located, is often affected by chromosomal instability such as loss of heterozygosity (LOH) in high-grade urothelial carcinomas (16, 17).
Primer sequence of microsatellites flanking Cu2+ transporting beta polypeptide ATPase (ATP7B). *Fluorescence labeling.
The aim of this pilot study was to explore genetic and allelic changes of the ATP7B gene in urothelial carcinomas of the bladder and their possible use as predictive markers for chemotherapy response in the metastatic setting.
Materials and Methods
Materials. Samples from 17 patients with metastatic high-grade urothelial carcinoma (pT2 or higher, G3) with measurable disease were analyzed. The formalin-fixed paraffin-embedded tissue was obtained after transurethral resection or cystectomy according to standard guidelines (18). In these patients, chemotherapy consisted of a minimum of two cycles of a palliative cisplatin-based regimen. Re-staging was performed with computed tomography scans of thorax, abdomen and pelvis after two cycles of chemotherapy. Re-staging was re-evaluated following the RECIST 1.0 criteria (19). Response was defined as partial remission or complete remission in the imaging. Out of these 17 cases, tumour tissue, morphological normal urothelium and muscle tissue was obtained by microdissection. These tissues were identified by a senior pathologist and microdissected. Following the microdissection the DNA was extracted as described elsewhere (20).
Amplification of DNA. DNA was amplified by target-specific primers in standard polymerase chain reactions together with a negative control. Only highly specific results were used for further analysis.
Microsatellite analysis. The microsatellite locations and primer sequences (see Table I) were taken from the literature (21, 22). In each case, DNA derived from the three tissue types was amplified separately and analyzed by capillary electrophoresis. A LOH was assumed if there was a relative reduction of the specific DNA of more than 50% in the tumour compared to normal tissue (smooth muscle).
Mutation analysis. Primers for amplification of the exons of ATP7B were taken from the literature and are shown in Table II (23, 24). ATP7B exons 6, 8, 13, 14, 15, 16, 17 and 20 were chosen for analysis because they are sites of the most prevalent polymorphisms in northwest Europe (14).
Denaturing high-performance liquid chromatography (DHPLC): For DHPLC, the amplified DNA was taken and used either in pure form to screen for heterozygotic changes or mixed with amplified DNA of a healthy Caucasian donor to screen for homozygotic changes. Melting times and temperatures were taken from preliminary work of our group (24). All heteroduplex cases indicating a sequence alteration were further analyzed by direct sequencing.
Microsatellite D13S297 in tumour (left), normal urothelium (middle) and normal smooth muscle of the bladder wall (right) of case 9. Intensity diagram below the histomorphologic image: The decrease of 67% of one of the two alleles in the tumour (left) in comparison to normal urothelium and smooth muscle reveals the loss of heterozygosity in the tumour. Above the image is the nucleotide sequence with the mutation M3, which is lost in the tumour compared to the morphological normal urothelium.
Sequencing: After purification, direct cycle sequencing was carried out. Both forward and reverse strands were sequenced. The sequences were compared with the published wild-type sequence using the Basic Local Alignment Search Tool BLAST provided by the National Center for Biotechnology Information, USA (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Non-synonymous mutations were defined as changes in nucleic acid sequence predicted to induce a change of the amino acid sequence. Synonymous mutations were defined as changes of the nucleic acid sequence without impact on the amino acid sequence during translation.
Statistical methods. Binary variables were analysed using cross-tables and Fisher's exact test. Numerical data were described with median, mean, standard deviation and quartiles, separated into groups for responders and non-responders, and displayed as box plots. All statistical tests were conducted two-sided and a p-value <0.05 was considered to indicate statistical significance. No correction of p-values for multiple testing was performed.
Results
Clinical response. The median patient age was 66 years (range=44-79, SD=10.8 years). Overall, 70.6% of patients were male, 29.4% were female. The evaluation of response after two cycles of chemotherapy revealed progressive disease in 11/17 (65%) patients. 6/17 (35%) patients had a partial remission. No complete remissions or stable disease occurred in this group.
Primer sequences for the amplification of the selected exons.
LOH. The microsatellite analysis detected LOH in 11/17 (65%) of the cases. Figure 1 exemplarily shows the reduction of the signal of one allele in the tumour in comparison with normal tissue in case 9. Table III gives the detailed results of the analysis.
Results of microsatellite analysis. Cases with more than 50% decrease were considered loss of heterozygosity (LOH).
Non-synonymous mutations including localisation in the ATP7B-gene with subsequent amino acid change and localisation in tertiary structure (25).
Mutations and polymorphisms. We found 12 non-synonymous mutations (NMs) and 20 synonymous mutations (SMs), most of them not previously described. None of the mutations appeared in more than one patient except for mutation #11 (see Tables IV, V and VI) (25).
Comparison with Wilson disease: The results were compared with the Wilson Disease Database (http://www.wilsondisease.med.ualberta.ca/database.asp) and the ATP7B Mutation Database (http://www.umd.be/ATP7B/). The majority of the discovered mutations have not been described before. NM 11 at position 3419 was described in the aforementioned databases and is considered as a non-pathogenic polymorphism. A pathogenic mutation at position 3122 is known in Wilson disease with a change from arginine to proline. In our case, in NM 6, the arginine is substituted by glutamine. SM 1, 3, 4, 6, 7 are listed in the databases as polymorphisms.
Synonymous mutations including localisation in the Cu2+ transporting beta polypeptide ATPase (ATP7B)-gene.
Distribution of LOH and non-synonymous mutations to responders and non-responders.
Statistical analysis: The data were correlated with the clinical parameter ‘response’, which included partial remission and complete remission. LOH was present in bladder cancer from 5/11 (45%) responders while tumours in non-responders did not have LOH.
Point-mutations were found in tumours of 2/8 (25%) responders and in 4/9 (44%) non-responders. The mutation NM11, which was the only common NM in our study group, was found in 5/11 (45%) of the non-responder group and in 1/6 (17%) of the responder group (see Figure 2).
The number of NMs was higher in the non-responder group than in the responder group. The number of affected microsatellite loci was lower in the non-responder group than in the responder group (see Figure 3). Although statistical analysis showed the described tendencies, none of these results were statistically significant.
Box plot of the absolute number of non-synonymous mutations and microsatellite losses in the responder vs. non-responder group. The black band shows the median. The boxes mark the first and third quartile, while the whiskers indicate one standard deviation above and below the mean. The small circles show outliers.
Distribution of mutations and Loss of heterozygosity in tumour (LOH) in tissue and by cases. SM, Synonymous mutation; NM, non-synonymous mutation, *no SM/NM.
Discussion
In this study we tried to analyze the role of genetic and allelic alterations of the ATP7B gene in primary resistance to platinum-based treatment. Due to the exploratory design of this study, the number of cases analyzed was low, restricting statistical analysis. We found 12 NMs mutations and 20 SMs. The comparison of NMs and SMs in tumour, morphologically normal urothelium and muscle tissue showed that most genetic alterations were already present in normal tissue. In all cases with LOH in the tumour tissue, we observed that only the mutated allele was deleted in favour of the wild-type allele of ATP7B (data not shown). The significance of this finding is unclear, but other investigations have shown that allele-specific loss of mutations might be advantageous for tumors (26).
Out of the examined mutations, only mutation NM11 occurred in more than one case (~35%). The described mutation is located in exon 16 and lies in the cytoplasmic domain of the protein. Non-responders exhibited the mutation NM11 more often than responders (Figure 3). In the context of Wilson disease, this mutation is mostly considered non-pathogenic (not inactivating), which makes it likely that it is either a silent polymorphism or an activating mutation. Furthermore, we found that NMs were present more often in the non-responder group. Additionally there were a higher absolute number of mutations in the non-responder group. Assuming that activating mutations lead to increased resistance while inactivating mutations or LOH lead to chemosensitivity, it might be that these mutations are activating. This speculation could also explain why the cases with LOH and a higher number of microsatellite losses were more often found in the responder group. Although our findings were not statistically significant, the analysis of LOH at the ATP7B locus might be fruitful for response prediction, even though other genes close to this locus might also be involved in chemotherapy response (e.g. the retinoblastoma gene on 13q14.1-q14.2).
Resistance is complex. Although the exact mechanism of resistance is unclear, there is more evidence supporting the role of ATP7B in platinum resistance. Down-regulation of ATP7B via liposomal siRNA led to better response to cisplatin and shrinking of the tumours in a mouse model (27). Another study analyzed the ATP7B transcription in colorectal carcinoma using quantitative PCR. Patients with down-regulated ATP7B had a longer response to platinum-based chemotherapy (28).
While these are interesting observations, it is clear that a monodimensional approach will not be sufficient to explain such a complex process as chemotherapy resistance. A study by Leonhardt et al. showed that two liver cell lines with a comparable expression of ATP7B had different cisplatin sensitivities. Cells derived from Atp7b −/− mice showed an increased resistance to cisplatin (29). This leads to the question if protective mechanisms do exist in the cell, which are up-regulated in reaction to ATP7B deficiency? A possible explanation might be that cisplatin not only uses ATP7B as a transporter but also interacts with it, leading to activation of the protein. This activation could lead to intracellular copper deficiency, which in turn could lead to the up regulation of protective mechanims such as the glutathione system (30). These interactions might be more important mechanisms in resistance than drug efflux itself.
There are other possible players in platinum-resistance such as CTR1, OCTs of the SLC22A-family, glutathione and GS-X of the ABC protein family. A better understanding of the interaction of these mechanisms at the level of genetic alterations, allelic changes and changes in protein expression might lead to the development of predictive marker(s) and individualised therapy.
Acknowledgements
We would like to thank Birgit Geist, Daniela Angermeier and Professor Karl-Friedrich Becker and his Group (Institute for Pathology) as well as Dr. med. Markus Groffmann (Department of Urology) for their help and advice during this project.
- Received July 6, 2013.
- Revision received July 24, 2013.
- Accepted July 25, 2013.
- Copyright© 2013 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.