In the present study, we demonstrate that the increase in MDR-1 mRNA levels induced by iHDACs inhibitors in pancreatic adenocarcinoma cell lines does not parallel an increase in Pgp protein or in Pgp activity. This observation is important since we and others have reported that the histone deacetylase inhibitors TSA and SAHA induce two major effects in several drug-resistant cell lines: down-regulation of Pgp and induction of apoptosis. In that sense, we have demonstrated that TSA markedly reduced Pgp expression in the L1210R drug-resistant murine cell line [20
] and “sensitized” these cells to daunomycin [26
], as shown in Figure C, where TSA treatment increases DNM accumulation in L1210R cells, suggesting that iHDACs might have a therapeutic potential against chemoresistant tumours. However, other authors had reported an increase in the steady-state level of MDR1 mRNA upon treatment with TSA [17
], suggesting that iHDACs could not be utilized in combination with other cytotoxic agents that are substrates of Pgp. Our results in pancreatic carcinoma cell lines, which are in agreement with our previously published results in colon carcinoma cell lines [19
], strongly suggest that expression of MDR1 mRNA is necessary but not sufficient for Pgp protein expression.
Next, we tried to identify the putative mechanisms involved in this phenomenon and we concluded that a translational blockade of Pgp expression takes place in pancreatic carcinoma cell lines, in agreement with our previous studies in colon carcinoma cell lines after TSA treatment [19
] and in the human erytroleukaemia K-562 cell line [28
Our data suggest that the translational blockade of MDR1 mRNA in the colon and pancreatic carcinoma cell lines and in K-562 cells could be overcome by alterations in the 5′
end of the MDR1 mRNA in the resistant variants of these cell lines. These results are especially relevant since we have previously demonstrated the relationship between the expression of the long 5′
UTR MDR1 mRNA and the final expression of an active Pgp protein [19
]. The origin and nature of these MDR1 mRNA isoforms became clear when Raguz et al. [25
] reported the presence of an ABCB1 gene upstream promoter in breast carcinoma samples. Both promoters would translate the same protein because they use the same ATG codon, but the mRNA transcribed from the upstream promoter is approximately 285
bp longer in its 5′
end than the MDR1 mRNA transcribed from the downstream promoter. Our data, together with results published by other groups, strongly suggest that expression of MDR1 mRNA is necessary but not sufficient for Pgp protein expression, indicating that MDR1 mRNA is subjected to a negative translational control. During the acquisition of chemoresistance there is a switch from the downstream to the upstream ABCB1 gene promoter, and this promoter transcribes a MDR1 mRNA that is translated more efficiently. To sustain this hypothesis, we have demonstrated that the expression of an active Pgp protein correlates with the activation of the upstream promoter of the ABCB1 gene in several K-562 cellular sublines obtained by selective pressure with increasing concentrations of daunomycin [19
] and Figure .
In addition, the results shown herein demonstrate that short and long 5′UTR MDR1 mRNAs are differentially regulated by the histone deacetylase inhibitor TSA, indicating that both promoters are differentially regulated by iHDACs. This is a compelling observation because TSA is able not only to downregulate the promoter responsible for active Pgp protein expression but also to induce apoptosis in colon and pancreatic carcinoma cells, sensitizing them to other chemotherapeutic agents that are substrates of Pgp. In addition, we observed that TSA increased MDR1 mRNA in the parental K-562 cells, whereas TSA decreased MDR1 mRNA levels in the daunomycin-resistant K-562 sublines which express Pgp protein and employ the USP promoter. This suggests again a differential regulation by TSA of both ABCB1 promoters.
We have also study the effect of TSA on the regulation of the RUNDC3B gene, since this is a nested gene transcribed from the ABCB1complementary DNA strand. These studies were designed to test the hypothesis that the regulation of the ABCB1 nested gene RUNDC3B could interfere with the alternative expression of both 5′
UTR MDR1 mRNAs, since some of the exons of the RUNDC3B mRNA lie on the complementary strand of the ABCB1 gene, as shown in Figure A. Our results demonstrate that TSA is able to increase RUNDC3B mRNA levels, independently of the ABCB1 promoters that are active in these cell lines. RUNDC3B has been related to a more metastatic phenotype in breast cancer patients [29
]. According to this, we could speculate a selection of RUNDC3B expression in colon and pancreatic carcinoma cell lines, instead of a functional Pgp protein expression, since expression of RUNDC3B could represent an additional advantage in their evolution to a more aggressive phenotype. However, taking into consideration our results, we cannot conclude that the expression of RUNDC3B could be in some way incompatible with the ABCB1 USP promoter expression forcing the cells to use the DSP promoter.
Although our data fit quite well with the existence of two ABCB1 promoters, some authors [30
] have found that the additional 5′
-UTR that is important for P-glycoprotein expression is due not to the use of an alternative promoter, but mostly to epigenetic changes in the region where this additional exon lies (112
kb upstream of the short MDR1 mRNA transcription initiation site). In fact, and based on chromatin immunoprecipitation assays and transient transfection of reporter genes under the control of the putative region where USP promoter lies, they concluded that the reason for the alternative 5′
-UTR MDR1mRNA expression is due to epigenetic changes, mostly related to histone H3 acetylation. The situation in the putative upstream promoter cannot be explained only by histone acetylation because TSA generally produces a state of hyperacetylation due to histone deacetylases inhibition. Global general hyperacetylation is related to an increase in transcription and however upstream promoter transcription is inhibited. Epigenetic control of Pgp expression has been previously suggested by different authors, which found that demethylating agents were able to induce Pgp- mediated chemoresistance in different cellular models [31
], We have some preliminary data with primary cultures derived from tumours of patients with colon cancer. In some of these cultures, we found that Pgp expression was lost. In some clonal populations obtained by extreme dilution of these cultures the long MDR1mRNA isoform and Pgp expression after treatment with DNA demethylating agents was recovered, suggesting that DNA methylation is involved in the expression of the MDR1mRNA isoforms (data not shown). However, it is unclear whether Pgp expression is regulated by two different promoters or by epigenetic mechanisms that modifies the activity of a single promoter. Since TSA is an epigenetic drug, from our results it is difficult to discriminate whether TSA downregulates the USP promoter or, alternatively modifies the epigenetic status of a specific region of the ABCB1 gene. In any case, we can affirm that TSA inhibits the expression of an active P-glycoprotein. The authors that argument the epigenetic changes as the main reason for the long 5′
-UTR MDR1mRNA production suggest that the nested gene RUNDC3B is only expressed when the long 5′
-UTR MDR1mRNA is expressed, or in other words that there is a correlation between the expression of both mRNAs. Our data do not support this hypothesis, because as seen in Figure B, RUNDC3B mRNA is expressed in cell lines that express the short 5′
-UTR MDR1 mRNA, suggesting that the expression of both genes is not regulated simultaneously by the same epigenetic changes in a specific genomic region.