Abstract
Although glucocorticoids play an important role in the treatment of multiple myeloma, some patients do not respond or develop resistance. The glucocorticoid receptor (GR), a single gene, mediates the effects of glucocorticoids. Using a model system of a multiple myeloma cell line sensitive to glucocorticoids and its early and late resistant variants, we have analyzed mutations in the GR gene, detected the presence of different transcriptional isoforms, quantified their levels of expression, and identified the promoters that regulate their expression. Levels of GR transcripts were comparable with the expression of total GR protein. Development of resistance correlates with an overall reduction in GR mRNA levels. This decrease in GR levels is neither due to mutation of the gene nor due to methylation. GRα is the predominant isoform in the sensitive cell line decreasing in expression in the early resistant cells and virtually undetectable in late resistant cells. GR-P is expressed at equivalent levels in both sensitive and early resistant cells, whereas in the late resistant cells, GR-P is the predominant isoform. GR-A is only expressed in the early resistant cell line. GRβ is the least expressed isoform in all cell lines. Interestingly, the level of expression of exon 1-exon 2 RNA fragments remains similar in sensitive and resistant cell lines. Resistant cells became sensitive to glucocorticoids after GRα transfection. In conclusion, we show different patterns of expression of the GR isoforms and provide evidence that a decline in the expression of GRα may be associated with development of resistance. [Mol Cancer Ther 2006;5(12):3062–70]
- glucocorticoid receptor
- multiple myeloma
- mRNA isoforms
- resistance
- promoters
Introduction
Multiple myeloma is a relatively rare clonal B-cell malignancy characterized by the accumulation of terminally differentiated, antibody-producing plasma cells in the bone marrow. Glucocorticoids are lipophilic compounds derived from cholesterol that have a wide range of biological activities. In addition to their physiologic roles, glucocorticoids induce apoptosis and cell cycle arrest in lymphoid cells, playing an important role in the treatment of multiple myeloma (1). The physiologic response and sensitivity to glucocorticoids varies among species, individuals, tissues, cell types, and even during the cell cycle (2, 3). Thus, although multiple myeloma is a disease that is generally considered responsive to glucocorticoids, some patients do not respond and those that do respond eventually develop resistance to this therapy (4).
The molecular basis of glucocorticoid resistance is not completely understood. Studies addressing the mechanism of resistance have shown that the glucocorticoid receptor (GR) is the primary target of genetic alterations leading to resistance. Most of the glucocorticoid hormone effects are mediated by the GR. The vast majority of glucocorticoid-resistant cells express low levels or defective forms of the GR protein (5–7). Although abnormalities in glucocorticoid responsiveness can be attributed in many cases to specific inherited mutations, epigenetic factors are also involved because in many diseases, long-term glucocorticoid treatment can result in glucocorticoid resistance (8). Similarly, studies in patients with leukemia have suggested that low GR protein levels are associated with a poor response to glucocorticoid treatment and to a mechanism of acquiring steroid resistance (9–11).
The structural organization of the GR protein is well known (12, 13). GR contains three functional domains: the NH2-terminal part or modulating domain involved in modulation of gene transcription (14); the DNA-binding domain involved in DNA binding, receptor dimerization (15), and gene transcription enhancement (16); and the ligand-binding domain that controls the activity of the receptor as a whole through its interaction with other proteins and glucocorticoids. In addition to ligand binding, the ligand-binding domain contains a receptor dimerization function (17, 18) as well as domains for silencing of the receptor in the absence of the hormone (19, 20).
There is only one known GR gene, but several GR isoforms arise because of alternative splicing events (21). The GR gene contains at least three promoters whose utilization causes at least five separate transcripts containing different 5′-untranslated first exons (22). None of the alternate exons 1 is predicted to alter the amino acid sequence because there is an in-frame stop codon preceding the translation initiation site in exon 2 that is common to all mRNA variants. A variety of different 5′ sequences may influence post-transcriptional gene processes, such as mRNA processing, export, and stability. It has also been suggested that promoter usage may regulate the differential response of GR expression to glucocorticoids (23), the expression of specific membrane or intracellular receptor isoforms (24), or even direct GR mRNA translation (25).
In addition to the different exons 1, alternative splicing generates numerous GR protein isoforms with altered function (Fig. 1 ). Five of the main isoforms GRα, GRβ, GR-P, GRγ, and GR-A have been described previously (26–29). Although the effects of glucocorticoid are mediated through the GRα isoform, the presence of other isoforms that have altered DNA and ligand-binding domains may influence the response of a particular cell to glucocorticoids. Therefore, the complexity of glucocorticoid biology lies more in the variety of receptors themselves rather than in the ligands to which they bind.
Structure and transcriptional isoforms of the human GR gene. A, schematic representation of the different exons comprising the human GR gene. 1A1, 1A2, 1A3, 1B, and 1C represent the different alternative untranslated first exons. B, different isoforms generated by alternative splicing of the human GR gene. Each isoform can be expressed from one of the five alternative untranslated first exons. Exons 2 to 4 are common to all isoforms. When the exon 9α is present on the transcript, independently of the presence or absence of exon 9β, the GRα isoform is generated. GRβ is produced when exon 9β is joined to exon 8. When the splicing event that joins exon 7 with exon 8 fails, another GR isoform is produced, the GR-P isoform that ends at intron G. GRα, GRβ, and GR-P represent the three major 3′ ends of the GR. Another GR isoform, GRγ, is created when in the splicing junction between exons 3 and 4 three additional bases are kept, its 3′ end can be exons 9α and 9β or intron G. The elimination of exons 5 to 7 by alternative splicing cause the GR-A isoform and its COOH-terminal region can be coded by exon 9α or 9β.
Because a combination of differential promoter regulation and alternative splicing may serve as a mechanism by which a cell evades the effects of glucocorticoids, the types of isoforms expressed in cells will be important in understanding the mechanism of glucocorticoid resistance development. However, the expression pattern of all the GR isoforms in combination with the promoters that regulate their expression as well as correlating them with the protein expression levels during the development of resistance have not been yet addressed.
In the present work, using a multiple myeloma cell line model system that mimics the progression of the glucocorticoid-resistant disease, we have analyzed the presence of different GR isoforms, their levels of expression, and the promoters that regulate the expression of these isoforms at the mRNA level. Levels of GR transcripts were compared with the expression of total GR protein. Our studies show different patterns of expression of the various GR isoforms between a glucocorticoid-sensitive cell line and its early and late resistant counterparts.
Materials and Methods
Cell Lines
A myeloma cell line, MM.1, was established from the peripheral blood cells of a patient whose treatment regimen included glucocorticoids (30). From this parental heterogeneous cell population, three separate cell lines that parallel the progression of the disease and development of resistance were established: A glucocorticoid-sensitive subclone (MM.1S), a transient glucocorticoid-resistant subclone (MM.1Re), and a stable resistant subclone (MM.1RL) representative of patients in the later stages of the disease (31).
Another multiple myeloma cell line, IM-9, was obtained from the American Type Culture Collection (Manassas, VA). All cells were grown in RPMI 1640 (Life Technologies, Carlsbad, CA) supplemented with 10% heat-inactivated fetal bovine serum (Life Technologies, Invitrogen) in 5% CO2 at 37°C. All cell lines are free of mycoplasma contamination as determined by Hoechst dye staining and PCR assays.
DNA and RNA Extraction
Cells were harvested when the cultures were growing in exponential phase. DNA and total RNA were extracted using Qiagen DNA and RNeasy extraction kits (Qiagen, Valencia, CA). Total RNA was treated with DNase I to eliminate genomic DNA contamination. The integrity of total RNA was confirmed in each case using the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA).
Affymetrix GeneChip Hybridization
The sample preparation was done according to the instructions in the Affymetrix GeneChip Expression Analysis Manual (Affymetrix, Santa Clara, CA). Briefly, 10 μg of RNA were used to synthesize double-stranded cDNA following SuperScript Choice system (Invitrogen, San Diego, CA) and T7-(dT)24. cRNA was synthesized from this cDNA and biotin labeled using the Enzo BioArray RNA transcript labeling kit (Affymetrix). The cRNA was fragmented and, along with hybridization controls (Affymetrix), was hybridized to Affymetrix Human Genome U133A GeneChip array and scanned using the GeneChip Scanner 3000 System and Microarray Suite 5.0 software (Affymetrix). Two replicates of each sample were hybridized on Affymetrix oligonucleotide microarrays.
Primers and Probes
Primers for sequencing of promoter A and for sequencing of exons 3 to 8 have been described before (22, 32). The remaining primers for regular PCR and sequencing were designed with Primer3 software (33) available online.3 Primers for real-time PCR quantification of promoters 1A1, 1A2, 1A3, 1B, and 1C were identical to previous study (34). The other primers were designed with Primer Express software (Applied Biosystems, Foster City, CA). The sequences of all the primers and probes used in the present study are listed in Supplementary Table S1.4
Conventional PCR
Detection of the different transcripts expressed from the five promoters was done with SuperScript One-Step reverse transcription-PCR (RT-PCR) system for long templates (Invitrogen) according to the manufacturer's instructions. Amplification of the different exons from DNA samples were done using Platinum PCR SuperMix (Invitrogen). Conventional PCRs were carried out on a MBS Satellite Thermal Cycler (Thermo Hybaid, Milford, CA).
Quantitative Real-time RT-PCR
The levels of GR mRNA expression were determined by one-step real-time PCR with primers specific for GRα, GRβ, GRγ, and GR-P localized on exons 8 to 9α, 8 to 9β, 3 to 4, and exon 7-intron G, respectively, using Taqman Chemistry (Applied Biosystems). The assays to quantify the different exon 1-containing transcripts were designed to amplify a RNA fragment spanning each specific exons 1 and 2. The 18S rRNA from the same samples served as an internal standard. All reactions were carried out in an ABI Prism 7900HT Sequence Detection System (Applied Biosystems) using standard cycling conditions for real-time assays. The efficiency of all real-time PCR systems was between 95% (slope of −3.45) and 107% (slope of −3.1).
Cloning of GR Isoforms
cDNA coding for GRα, GRβ, and GR-P containing promoters 1B and 1C were isolated from MM.1S and cloned into the TA site of the pCR-XL-TOPO vector (Invitrogen) and transformed into bacteria. Positive clones were identified by amplification of exons 5 and 6 of the GR cDNA. The correct sequence of the inserts was confirmed by DNA sequencing. These plasmids were used to generate distinct 10-fold dilution series of standards starting from concentrations of 105 molecules/μL for relative quantification of GR transcripts.
Immunoblot
Cells were lysed under nondenaturing conditions with cell lysis buffer (Cell Signaling Technology, Inc., Beverly, MA) supplemented with 1 mmol/L phenylmethylsulfonyl fluoride and complete protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN). Protein samples were separated on a NuPAGE Novex Tris-acetate 4% to 8% gel (Invitrogen) and transferred to Immobilon-P transfer membranes (Millipore, Billerica, MA). Mouse monoclonal anti-human GR (BD Biosciences PharMingen, San Jose, CA), mouse monoclonal anti-human actin (Sigma-Aldrich, St. Louis, MO), and anti-mouse IgG horseradish peroxidase–linked (Amersham, Piscataway, NJ) antibodies were used according to the manufacturer's recommendations. Blots were developed by means of addition of peroxidase substrate, with enhancement by using SuperSignal West Pico chemiluminescent substrate (Pierce Biotechnology, Rockford, IL) solution, and exposed to hyperfilm enhanced chemiluminescence films (Amersham). Serial dilutions ranging from 1,140 × 1010 to 0.13·× 1010 molecules of purified human GR protein (Sigma-Aldrich) were used to generate a standard curve to calculate the amount of GR protein in each cell line.
Flow Cytometry
To measure the expression of intracellular GR and active caspase-3, cells were fixed with 4% paraformaldehyde, permeabilized with saponin, and stained with mouse anti-GR monoclonal and anti-IgG1 conjugated with FITC and rabbit anti-caspase-3, active form conjugated with phycoerythrin (BD Biosciences, PharMingen). The same FITC-labeled anti-IgG1 antibody was used as negative control. An identical protocol without saponin was used to measure the membrane expression of GR proteins. Antibody reaction was measured using a FACScan (Becton Dickinson, Mansfield, MA) and analyzed by CellQuest software (Becton Dickinson). The expression of GR antigen was calculated as the mean fluorescence index [MFI; MFI = mean fluorescence of anti-GR antibody / mean fluorescence of anti-IgG1 antibody] in the total population.
Transfection of MM.1R Cells with GR cDNA
MM.1R cells (2 × 106) were resuspended in solution from nucleofector kit V, mixed with 2 μg plasmid (pCMX-PL2-GRα, Dr. Evans' Lab, Salk Institute, La Jolla, CA) and transfected with an Amaxa Nucleofector I apparatus (Amaxa, Inc., Gaithersburg, MD) using the O-20 program. After transfection, cells were cultured for 24 h before treatment with 10 μmol/L dexamethasone. GR and active caspase-3 expressions were analyzed by flow cytometry after 24 h of treatment.
Results
Gene Expression Profile in MM.1S and MM.1RL
To identify changes at the expression level that may contribute to the glucocorticoid-resistant phenotype, the gene expression profile in the glucocorticoid-resistant cell line MM.1RL was compared with the sensitive parent cell line, MM.1S. These gene expression profiles were examined using oligonucleotide microarrays containing ∼22,500 oligonucleotide probes corresponding to 14,500 well-characterized human genes.
Comparative analysis of gene profiles of glucocorticoid-sensitive versus glucocorticoid-resistant multiple myeloma cells showed significantly lower levels of GR expression in resistant cells. The intensity signal (log) of the fold change is listed in Fig. 2 . Expression of GR is reduced more than five times in glucocorticoid-resistant cells. To obtain independent verification of the steady-state RNA levels changes seen in array analysis, quantitative PCR analysis was done focusing on the GR gene, which shows the highest difference of expression between the cell lines.
Representation of the intensity signal (log) and the fold changes of the genes whose expression is repressed in MM.1RL cell line compared with that in MM.1S cell line based on the criteria of at least a 1.5-fold difference.
Real-time RT-PCR results confirmed the differences in expression levels of GR observed in the microarray analysis (Table 1 ). It should be noted that RNAs used for real-time RT-PCR were obtained from independent experiments, further verifying the changes. Our data are consistent with a previous report showing low levels of GRα in MM.1R cells by Northern blot and oligonucleotide arrays (35, 36).
Transcript copy number per microgram of total RNA of the different GR transcriptional isoforms and promoters expressed in MM.1S, MM.1Re, and MM.1RL cell lines
Expression of GR Protein Levels
To correlate the mRNA levels with the protein levels, the concentrations of the GR protein were measured by densitometric analysis of GR-immunoreactive bands on immunoblots. The number of GR molecules was extrapolated from a standard curve done with purified human GR protein (Fig. 3A ) and normalized to the number of cells used in the study. Because the antibody used in the study was not able to differentiate between the different isoforms, the immunoblots reflect the levels of total GR protein in MM.1S, MM.1Re, and MM.1RL cell lines (Fig. 3B). Quantitative analysis of these bands indicates that the GR protein is expressed at 100-fold lower levels in the MM.1Re cell line than in the sensitive cell line, and the protein is below the limit of detection in MM.1RL cell line.
Expression levels of total GR protein in MM.1S, MM.1Re, and MM.1RL cell lines. Total GR protein was detected by using a monoclonal antibody common to all isoforms and measured by densitometric analysis of GR-immunoreactive bands or flow cytometry. A, immunoblot of a standard curve obtained with serial dilutions of purified GR protein. The numbers under each lane represent the number of GR molecules loaded on the gel. B, Western blot for total GR protein in MM.1S, MM.1Re, and MM.1RL cell lines. C, expression levels of cytoplasmic and membrane-bound GR protein measured by flow cytometry with and without the use of a cell-permeabilizing agent, respectively. MFI of cells expressing GR protein. Positive values are higher than one.
Although the GR protein is primarily expressed intracellularly, a cell membrane–associated receptor has been described (37). This membrane isoform has been associated with the expression of exon 1A-containing transcripts (24). To test if MM.1S, MM.1Re, and MM.1RL cells express this type of GR and to differentiate the intracellular and extracellular GR expression, we did a flow cytometry assay. When studying the level of expression of intracellular GR receptor by determining the MFI, we observed results similar to those obtained by immunoblots. That is, MM.1S cell line expresses GR at higher levels than the MM.1Re, and the expression of GR protein was not detected in MM.1RL (Fig. 3C). These differences in intracellular expression of GR are statistically significant (P = 0.0067) when the three cell lines were compared and analyzed with a Kruskal-Wallis statistical test. In contrast, the membrane-bound GR expression is not significantly different (P = 0.1795) among cell lines. In agreement with our RNA expression results with exon 1A-containing transcripts, we did not detect the expression of a membrane receptor in these cell lines. This result was in contrast to the data in IM-9 cell line (a positive control), where we were able to detect a membrane form of the GR (data not shown).
Expression of Different Transcriptional Isoforms
The analyses of the different transcripts expressed in MM.1S, MM.1Re, and MM.1RL cell lines were done by long-range PCR using forward primers specific for each first exon and reverse primers representing the main 3′ structures of GR mRNAs that are exon 9α, exon 9β, and intron G. These PCRs would result in amplification of GRα, GRβ, and GR-P, respectively. All the amplified segments obtained were of the expected sizes excluding major deletions or rearrangements within the GR cDNA. Because the GRγ isoform differs from GRα, GRβ, and GR-P by the insertion of only one amino acid, it cannot be separated due to the lack of resolution of amplification products on agarose gels.
Transcripts containing exons 1A1, 1A2, or 1A3 in combination with exon 9α were not detected in any of the cell lines (Fig. 4A, left ). In contrast, the expression of GRα isoform driven by promoters B and C was detected in the sensitive and early resistant cell lines but only by promoter C in the late resistant cell line. In addition to GRα, when GR RNA with a reverse primer located in exon 9α was amplified, we were able to detect a transcript corresponding to the GR-A isoform (described as in ref. 31) only in the early resistant cell line (Fig. 4A, left, lanes 5 and 6). When the GR-A transcript is present, another fragment, shorter than the wild-type transcript for GRα, was also detected on the electrophoresis (Fig. 4A, left, *, lanes 5 and 6). This fragment is also visible when primers located in exons 4 and 8 are used to prime the amplification reaction (Fig. 4B). Neither the GR-A isoform nor the new fragment is detected in the MM.1S or MM.1RL cell lines. Sequencing of this fragment reveals a double sequence identical at both ends of the sequence but differing in the central part, suggesting the presence of a heteroduplex (data not shown). To further confirm the presence of a heteroduplex, wild-type GR and GR-A bands were isolated from gel, mixed together, heat denaturated, allowed to cool at room temperature, and resolved in an agarose gel. The appearance of the medium-sized fragment confirms the generation of a heteroduplex between the amplification products corresponding to wild-type GRα and GR-A isoforms (Fig. 4C, *, lane 5). As expected, a similar size fragment corresponding to GR-A was again visible when the GRβ isoform is amplified and resolved on a gel (Fig. 4A, middle, lane 5).
Expression of GR transcript isoforms in MM.1S, MM.1Re, and MM.1RL cell lines. The detection of the different GR transcripts was done by one-step RT-PCR suitable for long templates. A, promoters regulating the α, β, P, and A isoforms. Lanes 2 to 6, GR transcripts containing the different untranslated first exons of the GR gene (A1, A2, A3, B, and C). Left, middle, and right, GRα, GRβ, and GR-P isoforms, respectively. First lane in gel corresponds to a molecular size marker showing two bands of 2,000 and 3,000 bp. B, expression of the GR-A isoform in three cell lines. The detection of the GR-A isoform was achieved by RT-PCR amplification of a mRNA fragment between exons 4 and 8 of the GR gene. Lanes 2 to 4, expression of wild-type (WT) GR and GR-A transcriptional isoforms in MM.1S, MM.1Re, and MM.1RL cell lines, respectively. *, unidentified PCR product. C, identification of a heteroduplex. Amplification products of wild-type GR (lane 1) and GR-A (lane 2) were mixed together and resolved on a gel with or without previous incubation at 94°C. M, molecular size marker with bands every 100 bp; the smallest band corresponds to a size of 100 bp; *, heteroduplex. D, PCR assay of the GRβ isoform. In addition to the β transcript, the gel shows amplification of GRα transcripts with a 3′-untranslated region that includes exon 9β (♦). Lanes 2 and 3, MM.1S and IM-9 cell lines.
The GRβ isoform was detected in MM.1S and MM.1Re cell lines only (Fig. 4A, middle), and as with GRα isoform, transcripts containing exon 1A1, 1A2, or 1A3 in combination with exon 9β were not detected in any of the cell lines. The expression of the β transcriptional isoform is regulated once more by promoters B and C in the sensitive cell line and only by promoter C in the early resistant cell line. GRβ was not detected in MM.1RL cell line, but because the level of expression of this isoform in most cell lines studied is very low and the sensitivity of long-range PCRs is not very high, we cannot discard the possibility for the presence of GRβ in MM.1RL cell line. In addition to the presence of GRβ, on the electrophoresis scan, a slow migrating fragment was also seen (Fig. 4D, ♦). Because the PCR assay designed to detect the GRβ isoform has the reverse primer located on exon 9β, we speculate that the slow migrating fragment corresponds to a GRα transcript with a longer 3′-untranslated region that includes exon 9β at the very end of the transcript (see Fig. 1). The identity of this transcript was confirmed by extracting and sequencing the slow migrating fragment. For the rest of the study, the extension time on the long-range PCR was adjusted to only detect the GRβ isoform.
To identify the expression pattern of the GR-P transcriptional isoform, primers located on the three promoters (forward primers) and on intron G (reverse primer) were used. As observed for the GRα isoform (Fig. 4A, compare right and left), GR-P is detected in the three cell lines, being regulated by promoters B and C in MM.1S and MM.1Re cell lines and by promoter C in MM.1RL cell line. Once again, there is no amplification product in any of exon 1A-containing transcripts of GR-P. In agreement with a deletion of exon 7 in the GR-A isoform, this isoform is not detected when a reverse primer located on intron G is used in the PCR to detect the GR-P isoform.
Expression Levels of Each Isoform
The above results described types of GR isoforms expressed in each cell line. However, because the function of GR isoforms is dependent on the number of transcripts, expression levels of the different isoforms and promoters (Table 1) were quantified by real-time PCR assays using plasmid clones of each isoform as standards. An overall comparison of GR transcript levels in MM.1S, MM.1Re, and MM.1RL cell lines showed that GR transcripts are greatly reduced in the resistant cell lines with an almost complete diminution in the late resistant cell line (Table 1).
Analysis of the relative expression of GRα isoform, the mediator of glucocorticoid responsiveness, in MM.1S, MM.1Re, and MM.1RL revealed a gradual diminution of this isoform with the progression of the disease. As noted in Table 1, the highest expression levels of GRα is in the MM.1S cell line, with 100- and 100,000-fold decrease in the expression of this isoform in early and late resistant cell lines, respectively. Compared with the expression levels of the other GR isoforms, GRα is the predominantly expressed isoform in the sensitive cell line.
GRβ is the least abundant of all isoforms in the three cell lines. However, as with the GRα isoform, the GRβ RNA expression undergoes a decline with the progression of the disease, and it is almost below the lower detection limit for the MM.1RL cell line. GRβ is expressed 10,000- and 1,000-fold less than GRα in MM.1S and MM.1Re cell lines, respectively.
Interestingly, there is not much difference in the expression levels of GR-P isoform between the sensitive and the early resistant cell line (Table 1). In these cell lines, the expression of GR-P is very high but 100-fold less than GRα in MM.1S and at the same level in MM.1Re. Again, the late resistant cell line is the one that expresses lowest level of this isoform.
Using real-time PCR, we were able to detect the GRγ isoform that only differs in three bp from the other isoforms. All three cell lines express the GRγ isoform, although at different levels. Whereas the quantity of this message is high in both sensitive and early resistant cell lines, the expression of GRγ in MM.1RL cell line is extremely low and close to the limit of detection.
As shown in Table 1, MM.1Re expresses all isoforms at lower level than the sensitive cell line. Moreover, in the late resistant cell line, there is a dramatic decrease in the transcript levels of all GR isoforms; whereas we were able to detect very low levels of GRα and GR-P, the GRβ and GRγ isoforms were almost undetectable. Surprisingly, when analyzing the expression levels of the first and second exons, the three cell lines express fragments of promoter-exon 2 transcripts at high levels. In agreement with the results of transcriptional isoforms detected by long-range PCR, exon 1A-containing transcripts are not detectable by real-time PCR. In addition, the levels of exon 1B/C-exon 2 expression were similar in all cell lines with exon 1C-containing transcripts 7 to 10 times more abundant than exon 1B-containing transcripts.
Reversion of GR Resistance
To determine if we can revert the glucocorticoid-resistant phenotype of MM.1R cells, we transfected these cells with a plasmid containing the coding sequence of the GRα gene. After 48 h of transfection, we only detect a 4.9-fold induction on the expression of GRα in the transfected cells with respect to nontransfected cells. However, after 24 h of dexamethasone treatment, we observed an 8.5-fold induction (P ≤ 0.05) in the expression of GRα in transfected cells versus a 1.9-fold induction in nontransfected cells (Fig. 5A ). After 24 h of dexamethasone treatment, the induction of apoptosis in GRα-expressing cells was measured by flow cytometry using an antiactive caspase-3 antibody. As shown in Fig. 5B, after 24 h of dexamethasone treatment, there is only a 1.5-fold induction in apoptosis in nontransfected cells, in contrast with a 3-fold induction in GRα-transfected cells. These data suggest that presence of GRα isoform is associated with sensitivity to glucocorticoid.
Detection of GRα protein and apoptosis in transfected (black columns) and nontransfected (white columns) MM.1RL cells before and after 10 μmol/L dexamethasone treatment for 24 h. Cells were transfected with a plasmid containing the GRα coding region. A, changes in intracellular GRα expression. Expression was measured by flow cytometry 48 h after transfection. Values are expressed as the fold change compared with nontransfected and untreated MM.1RL cells. B, induction of apoptosis detected using an antibody antiactive caspase-3. Values are expressed as the fold change with respect to nontransfected and untreated MM.1RL cells.
Discussion
Previous studies have shown that although MM.1S is sensitive to glucocorticoids, early and late resistant variants, developed by constant exposure to glucocorticoids, do not respond to dexamethasone-induced apoptosis. Although, the primary mechanism of glucocorticoid resistance has been directed toward the expression of the GR, a detailed analysis at expression levels is not available. The aim of our study was to identify GR variables that are associated with the resistant variants of the primary cell line.
The only known GR gene results into multiple GR isoforms generated by alternative splicing. To determine if the prevalence of various GR isoforms is associated with resistance, we analyzed the expression of these isoforms. Because our study quantified the number of transcripts of each isoform, it allows us not only to evaluate the amount of each isoform in different cell lines but also to compare the number of transcripts of all isoforms in one cell line.
The effects of glucocorticoids are mediated through the GRα isoform (27) that resides in the cytoplasm in absence of hormone. Several investigations have shown that reductions in the number of GRα molecules/cell result in a decrease in glucocorticoid sensitivity (9–11, 38, 39). Hence, GRα seems to be the most important isoform about sensitivity (40). This observation agrees with our results that show a gradual reduction in the number of GRα transcripts during the development of glucocorticoid resistance. For example, the GRα isoform is 40-fold more abundant in MM.1S than in MM.1Re, which in turn is 1,000-fold greater than in MM.1RL (Table 1). Moreover, the expression of total GR protein is proportional with GRα mRNA levels (Fig. 3B; Table 1).
Although after transfection we did not get a high number of cells expressing GRα, we were able to detect a 3-fold apoptotic induction. This phenomenon can be explained as the need of a certain threshold level of intracellular GR for glucocorticoid-mediated cell death in MM.1 cells, suggesting that glucocorticoid-mediated up-regulation of the GR is required for hormone-induced apoptosis, similar to the glucocorticoid effects described previously in T cells (41). Our current study did not focus on glucocorticoid-mediated transactivation of promoters; previous studies have shown such effects after transfection of GRα in COS cells (42). Taken together, these data show that low expression of GRα in MM.1R cells may be the main factor implicated in the resistance.
An alternative isoform is GRβ that differs from GRα on the COOH-terminal end of the protein. GRβ fails to bind hormone or activate gene transcription and is located in the nucleus in the absence of ligand (43). This splice variant has been proposed to have a dominant-negative effect on the actions of GRα; however, there is still much controversy about the functional significance of GRβ, especially with respect to its putative inhibiting activities on GRα (44). Our data suggest that in multiple myeloma cells relative to GRα, the level of GRβ isoform is very low. Furthermore, as with GRα, there was a gradual decrease in its expression in resistant cells. Hence, it seems that in the present model system, the GRβ isoform may not be implicated in the development of resistance to glucocorticoids.
The GR-P isoform first described in tumor cells from a glucocorticoid-resistant myeloma patient (28, 45) is encoded by exons 2 to 7 and part of intron G as a unique COOH-terminal tail. This truncated isoform lacks a large part of the ligand-binding domain, including the domains for silencing of the receptor in the absence of hormone and transcriptional activation. The GR-P isoform seems to be present in several hematologic tumor cells as well as in normal lymphocytes. There is controversy about its role in glucocorticoid sensitivity with reports suggesting that it might contribute to the resistant phenotype (45) and others providing evidence that GR-P can enhance the activity of GRα in some cells (46). In our study, compared with other variants, GR-P seems to be the predominant isoform in resistant cell lines. This isoform is expressed at similar levels in sensitive and early resistant cell lines. Moreover, when compared with GRα, we observed a change in the GR-P/GRα ratio, where GR-P is expressed 100-fold less than GRα in MM.1S, at similar levels in MM.1Re and 10 times more in MM.1RL cell lines. In the late resistant cell line, this is the most abundant isoform, indicating that the GR-P isoform may be implicated in the development of resistance.
Another isoform named GRγ is the result of the retention of three bp between exons 3 and 4 due to the use of an alternative splice donor site. Consequently, a mRNA coding for an additional arginine on the DNA-binding domain is produced. GRγ is a ligand-dependent transcription factor with reduced transactivating activity (29). GRγ seems rather ubiquitously expressed, although its function is presently unknown. In our model system, GRγ is present at lower levels than GRα isoform in all cell lines and lowest in the late resistant cell line. Moreover, there is not a difference in the pattern of expression of this isoform with respect to the others between the three cell lines. Taken together, these data suggest that GRγ may not be implicated in glucocorticoid resistance in this model.
Moalli et al. (28) described the GR-A isoform, which presents an internal deletion in the first portion of the ligand-binding domain, including a region important in the phenomenon of hormone down-regulation, nuclear localization, and transactivation of target genes. In GR-A mRNA, exons 5 to 7 are precisely excised due to an alternative splicing event. The pattern of expression of this isoform between the sensitive and resistant cell lines is unique. In the sensitive cell line, GR-A is not expressed, but its expression is induced as an early step in the development of resistance. Once the resistant phenotype is achieved in a stable manner, the expression of GR-A is inhibited.
When the quantity of each isoform (except GR-A) is compared in these three cell lines, a clear pattern appears showing an overall gradual decline in the GR isoforms as cells become resistant to glucocorticoids. Although our study did not focus on the mechanisms for the gradual decrease in global GR transcripts in resistant cell lines, a few points emerged. First, the differential level in the expression of the GR isoforms may be due to the use of different promoters in these cell types. However, the same promoters, promoters B and C, regulate the different expression patterns of all these GR isoforms in the three cell lines. Second, in all three cell types, the same promoter, promoter C, is used the most. Thus, it seems that alternative splicing rather than promoter usage may result in the differential expression of the GR isoforms in the sensitive and resistant cell lines. Third, although a gradual decrease in full-length GR transcripts is seen during the development of resistance, the level of expression of exon 1-exon 2 RNA fragments remains the same in sensitive and resistant cell lines (Table 1). This observation suggests that an inhibition in the initiation of transcription may not be the cause for the overall reduction of GR mRNAs in resistant cell lines. These observations and the fact that GR mRNA has a half-life of 1 to 6 h and that we are not able to detect GR transcripts in the late resistant cell line at any point but consistently obtain high levels of expression of exon 1-exon 2 fragments rule out the possibility of a faster RNA degradation in the resistant cell line. In conclusion, these four observations point to a block on the transcriptional elongation as a mechanism in the development of resistance. Currently, using isolated nuclei, we are pursuing this postulate.
Footnotes
↵3 http://www-genome.wi.mit.edu/genome_software/other/primer3.html.
↵4 Supplementary material for this article is available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).
Grant support: National Cancer Institute grant CA85915.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Accepted October 26, 2006.
- Received June 12, 2006.
- Revision received September 24, 2006.
- American Association for Cancer Research