Proteomic evaluation of pathways associated with dexamethasone-mediated apoptosis and resistance in multiple myeloma


F. E. Davies, Institute of Cancer Research, Brookes Lawley Building, Cotswold Road, Sutton, Surrey, UK.


We have used global protein expression analysis to characterize the pathways of dexamethasone-mediated apoptosis and resistance in myeloma. Analysis of MM.1S cells by two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) identified a series of proteins that were up- and downregulated following dexamethasone treatment. Downregulated proteins included proteins involved in cell survival and proliferation, whereas upregulated proteins were involved in post-translational modification, protein folding and trafficking. A comparison with published gene expression studies identified FK binding protein 5 (FKBP5) (also known as FKBP51), a key regulatory component of the Hsp90-steroid-receptor complex to be increased at the mRNA and protein level postdexamethasone exposure. Quantitative real time polymerase chain reaction and 2D-PAGE analysis of the dexamethasone resistant cell line MM.1R demonstrated no increase in FKBP5, consistent with its association with dexamethasone-mediated apoptosis. Western blot analysis of FKBP5 and other members of the Hsp90-receptor complex showed an increase in FKBP5 whilst FKBP4 (also known as FKBP52) and Hsp90 expression remained constant. No changes were observed in MM.1R. In conclusion, we demonstrated that following steroid receptor signalling, the cell carries out a number of adaptive responses prior to cell death. Interfering with these adaptive responses may enhance the myeloma killing effect of dexamethasone.


Glucocorticoids (GCs) are one of the most widely used treatments for multiple myeloma. In the 1970s dexamethasone was introduced as a treatment regimen as in vitro studies had demonstrated it was able to induce apoptosis in lymphocytes, and in clinical trials responses in up to 70% of myeloma patients were observed (Morgan, 2001). With the aim of improving response rates it was later used in combination with other chemotherapeutic agents, most notably in the VAD regimen, where it was combined with vincristine and adriamycin (Samson et al, 1989). More recent observations suggest that the activity of VAD is not that much greater than with single agent dexamethasone and many centres are now returning to using high-dose dexamethasone alone. A number of new agents have recently been shown to be effective, most notably thalidomide, bortezomib and lenalidomide and clinical trials demonstrate there is a synergistic response when these agents are combined with dexamethasone (Richardson et al, 2002, 2003). Despite these advances in the clinic arena, little is known about the mechanism of dexamethasone-induced apoptosis or how to enhance its activity by selecting appropriate agents that may have synergistic actions. To design effective combination chemotherapy schedules it is important to fully understand exactly how dexamethasone induces apoptosis and how resistance may occur.

Myeloma cell growth and survival is heavily dependent on the interaction of the myeloma plasma cell with the bone marrow microenvironment. The binding of the myeloma cell to the bone marrow stroma results in an upregulation of a number of important cytokines including interleukin 6 (IL6), vascular endothelial growth factor, inhibitory growth factor 1 and tumour necrosis factor α, and also protects against drug-induced apoptosis. Studies to date have used traditional protein approaches to examine the individual growth and survival pathways of myeloma plasma cells, of which the most important are mitogen-activated protein kinase (MAPK), related adhesion focal tyrosine kinase (RAFTK), phosphatidylinositol 3-kinase/Akt and nuclear factor (NF)-κB.

The key event leading to apoptosis of myeloma cell lines following dexamethasone exposure is a reduction in mitochondrial transmembrane potential, caspase 3 cleavage and poly (ADP ribose) polymerase cleavage (Chauhan et al, 1999). Overexpression of RAFTK, a tyrosine kinase, which functions as a regulator of several key signalling proteins including MAPK, Jun N-terminal kinase and Src kinases induces apoptosis which is compatible with RAFTK being an upstream mediator of caspase-induced cell death (Tokiwa et al, 1996; Chauhan et al, 1999; Pandey et al, 1999; Chauhan & Anderson, 2003). Second mitochondria-derived activator of caspases, another modulator of the caspase-dependent apoptosis pathway, is released from the mitochondria into the cytosol following treatment with dexamethasone resulting in inhibition of the inhibitor of apoptosis proteins leading to caspase activation (Chauhan et al, 2001; Ferri & Kroemer, 2001; Shi, 2001). Dexamethasone has also been shown to inhibit NF-κB, one of the central transcription factors for myeloma cells, via a number of mechanisms including the tethering of NF-κB to the glucocorticoid receptor (GR) resulting in pro-survival and anti-apoptotic gene transcriptional repression (De Bosscher et al, 2000a,b; Greenstein et al, 2003).

Recently, further information regarding the different functional pathways involved in the mechanisms of drug-induced apoptosis and resistance has become available because of the use of a more global approach with oligonucleotide arrays. However, the level of RNA expression does not necessarily correlate with the changes at the functional protein level (Anderson & Seilhamer, 1997). Advances in the use of proteomic technologies now provide a robust approach to study multiple signalling pathways simultaneously. We have, therefore, used a global proteomic-based approach to look at GC-induced apoptosis in an in vitro myeloma cell line model.


Cell culture and reagents

The dexamethasone sensitive and resistant subline MM.1S and MM.1R (Greenstein et al, 2003; Moalli et al, 1992; Goldman-Leikin et al, 1989) were cultured in RPMI-1640 medium with glutamine (Gibco Invitrogen, Paisley, UK), 10% fetal bovine serum (FBS; Gibco Invitrogen) and 1% penicillin/streptomycin (Gibco Invitrogen) in 5% CO2 at 37°C. Cells at a density of 0·5 × 106/ml were treated with a range of 0·01–10 μmol/l dexamethasone (Sigma-Aldrich, Gillingham, UK) for 24 h. Response to dexamethasone was determined using the 4,5-dimethylthiazol-2-yl-2,5-diphenyltetrazolium bromide (MTT) colorimetric assay according to manufacturers instructions (Chemicon, Chandlers Ford, UK).

Protein extraction

Serum starved cells at a density of 0·5 × 106/ml were incubated in 10% FBS/RPMI medium in 5% CO2 at 37°C with and without 1 μmol/l dexamethasone for 24 h. Cells were washed three times with ice-cold phosphate buffered saline and finally with ice-cold isotonic (0·25 mol/l) sucrose, followed by incubation with lysis buffer (7 mol/l urea, 2 mol/l thiourea, 4% w/v CHAPS (3,3-cholamidopropyl-dimethylammonio-1-propanesulfonate), 1% w/v dithiothreitol (DTT) 0·8% v/v Pharmalyte pH3-10, 1 mg/ml of Pefabloc) containing Complete™ mini protease inhibitor cocktail (Roche, Lewes, UK; 1 tablet/2·5 m) at room temperature for 30 min with intermittent vortexing. Following ultracentrifugation at 42 000 g for 60 min at 15°C, samples were aliquoted and stored at −80°C. Protein concentration was determined using the Bio-Rad protein assay, based on a modified Bradford method (Bio-Rad, Hemel Hemstead, UK). All experiments were performed in triplicate.

Two dimensional polyacrylamide gel electrophoresis

Isoelectric focussing.  Protein (100 μg for analytical gels and 1 mg for preparative gels) was run on 18 cm pH 4–7 precast IPG strips (Amersham Biosciences, Little Chalfont, UK). Protein was diluted in reswelling buffer (7 mol/l urea, 2 mol/l thiourea, 4% w/v CHAPS 0·46% w/v DTT, 0·2% v/v Pharmalyte 3–10, bromophenol blue) and loaded onto immobilized pH gradient (IPG) strips by in gel rehydration (30 V, 13 h) using the IPGphor system (Amersham Biosciences, Little Chalfont, UK). Focusing was carried out at 150 V (1 h), 300 V (1 h) 600 V (1 h), 600 V–3500 V (2 h) and 3500 V to end (total of 65000 Vh) using the Multiphor system (Amersham Biosciences, Little Chalfont, UK). Strips were stored at −80°C (Craven et al, 2002).

Sodium dodecyl sulphate polyacrylamide gel electrophoresis.  A 10%T/3·3% large format gels were poured using the ISO-DALT multicasting system (Amersham Biosciences, Little Chalfont, UK). 4%T stacking gels were added to the top of each gel prior to use. Strips were thawed and equilibrated in equilibration buffer [6 mol/l urea, 30% v/v glycerol 2% w/v Sodium dodecyl sulphate (SDS) in 50 mmol/l Tris–HCL and pH 6·8] containing 1% w/v DTT for 15 min followed by 10 min in equilibration buffer containing 4% w/v iodoacetamide. Strips were rinsed in running buffer (24 mmol/l Tris, 200 mmol/l glycine and 0·1% w/v SDS) and placed on top of the gel. Molecular weight markers (Novex Marker 12; Novex Toronto, Canada) were added and the strip and markers were sealed in place using 1% LMP ultrapure agarose in running buffer (Gibco Invitrogen, Paisley, UK). Gels were electrophoresed overnight 12·5°C in an ISO-DALT gel tank (Amersham Biosciences, Little Chalfont, UK) at 17 mA/gel, stained using the OWL silver staining kit (OWL Separation Systems, Portsmouth, NH, USA) and scanned using a personal densitometer SI (Molecular Dynamics Amersham Biosciences, Little Chalfont, UK). All gels were run in triplicate and analysed using Melanie 3 software (Gene-Bio, Geneva, Switzerland). Protein spots with a percentage volume of ≥0·04 and a change of twofold or more (P < 0·05; Students t-test) were selected. The entire experiment was also repeated for verification of the results. For preparative gels, gels were stained using a modified protocol from the Plus One silver staining kit, (Amersham Biosciences, Little Chalfont, UK) (Unwin et al, 2003).

Mass spectrometry

Protein spots of interest were identified and excised from the gel. Spots were destained using 50 mmol/l sodium thiosulphate, 15 mmol/l potassium ferricyanide, washed with H20 for 20 min then equilibrated with 25 mmol/l ammonium bicarbonate for 20 min. Spots were dehydrated with acetonitrile (ACN) for 20 min, rehydrated with 25 mmol/l ammonium bicarbonate, dehydrated with ACN and dried in a Speed Vac. For tryptic digestion 200 μl 0·1 mg/ml trypsin (16 000 U/mg) (Promega, Southampton, UK) was prepared in ice-cold resuspension buffer (supplied with the enzyme) and made up to 500 μl 25 mmol/l ammonium bicarbonate to give a working stock. Gel pieces were rehydratated in 5 μl for 45 min on ice. A further 30 μl 25 mmol/l ammonium bicarbonate was added and the samples were incubated at 37°C for 4 h after which the reaction was stopped by placing on dry ice. After defrosting, the supernatant was placed in a clean tube and two further extractions were performed from the gel piece with 5% formic acid, incubating in a sonicating water bath for 15 min. Peptides were pooled and dried in a Speed Vac.

Peptides were analysed using a 4700 Proteomics Analyzer (Amersham Biosciences, Little Chalfont, UK) in MS reflector positive mode. The collated MS and/or MS/MS data were used to search the NCBI database using Protein Prospector with a mass accuracy of 20 ppm.

Western blotting

Protein separation and transfer was carried out using the Bio-Rad Protean 3 system (Bio-Rad, Hemel Hempstead, UK) Proteins were separated by SDS- polyacrylamide gel electrophoresis (PAGE) on 10% gels. The separated proteins were transferred onto nitrocellulose (Schleicher and Scheull, London, UK) in 25 mmol/l Tris, 192 mmol/l glycine, 20% methanol for 75 min at 250 mA with cooling. Membranes were washed with 1 × TBS-T (20 mmol/l Tris, 137 mmol/l NaCl, pH 7·6 with 1% Tween 20) for 5 min and blocked (1 × TBS, 0·1% Tween-20 with 5% w/v non-fat dry milk) for 2 h at room temperature. Membranes were washed three times for 5 min in TBS-T and incubated with primary antibody in primary antibody dilution buffer [1 × TBS, 0·1% Tween-20 with 5% bovine serum albumin (BSA)] as follows; FKBP5 (also known as FKBP51) (F-13), FKBP4 (also known as FKBP52) [HSP-56 (P-14)], and actin (C-11) (Santa Cruz Biotechology, Heidleberg, Germany), Hsp90 (Stressgen Bioreagents, Victoria, Canada) and GR (E-20) (BD biosciences, Cambridge UK). Membranes were then washed three further times and incubated in horseradish peroxidase-conjugated secondary antibody, anti-Goat IgG or anti-Rabbit IgG (Sigma-Aldrich, Gillingham, UK). Following three final washes, detection was performed using an enhanced chemiluminescence method (Amersham Biosciences, Little Chalfont, UK). Triplicate experiments were performed for verification of data.

Quantitative real time-polymerase chain reaction

RNA was extracted from cell lines and CD138+ bone marrow plasma cells with and without 24-h dexamethasone exposure using the RNeasy® Mini kit (Qiagen Ltd., Crawley, UK) following the manufacturers instructions. Total RNA was reverse transcribed using random hexamers, pdN6 (Roche Diagnostics Mannheim, Germany), dNTPs (Bioline, Ltd. London UK) and Superscript reverse transcriptase (Invitrogen, Paisley, UK). An ABL PCR was performed to determine the cDNA state and to detect any DNA contamination. SYBR Green RQ-PCR (quantitative real time-polymerase chain reaction) (Applied Biosystems, Warrington, UK) was performed in triplicate to determine the relative expression of FKBP5 and FKBP4 relative to the control gene GUSB. Primers sequences were designed for intron/exon boundaries for FKBP5; F 5′-TGTCAAAGAGAAGGGAACCGTATAC, R 5′-AACCAGGACACTATCTTCCCATACTG, FKBP4; F-5′AATTCAGAAGAGAAGCTGGAACAGA, R 5′-CTTGCTTGTATTTACCTTCCTTGAAGT and GUSB F 5′-GAAAATATGTGGTTGGAGAGCTCATT, R 5′-CCGAGTGAAGATCCCCTTTTTA (Invitrogen, Paisley, UK). Thermal cycling conditions were 2 min at 50°C, 10 min at 95°C, and 40 cycles at 95°C for 15 s and 60°C for 1 min.


Treatment of MM.1S with dexamethasone

Protein from untreated and treated (1 μmol/l dexamethasone for 24 h) MM.1S cells was separated by 2D PAGE using a pH range of 4–7. An average number of 951 spots were identified on these gels. Following treatment with dexamethasone significant changes in expression pattern of twofold or more (P < 0·05) were identified; 24 spots were downregulated and three spots were upregulated (Fig 1A and B). 18 of the 27 spots were identified by mass spectrometry (Table I).

Figure 1.

 Two-dimentional polyacrylamide gel electrophoresis (2D-PAGE) analysis of the protein expression profile of MM.1S with and without exposure to dexamethasone. (A) 2D gel of MM.1S after 2D-PAGE, using a pH 4–7 non-linear gradient. 24-downregulated spots (full circles) and three upregulated spots (broken circles) were identified with labelling corresponding to the identifications given in Table I. (B) Representation of the average percentage volume of each spot in both MM.1S (white bar) and MM.1S following dexamethasone exposure (black bar). Error bars represent ± 1 SD. All gels were run in triplicate and the whole experiment repeated twice.

Table I.   Identities of the protein spots with an altered expression pattern upon treatment of the MM.1S cell line with dexamethasone.
Spot*Protein name (synonym where applicable)Database accession numberFunctional clusteringTheoretical mass (kDa)/isoelectric point Ratio†
  1. D1 (−2·8), D2 (−4·2), D3 (−2), D6 (−2·8), D12 (−6·3), D18 (−7·1), D19 (−2·2), D20 (−7·9) and D23 (−2·4) were not identified.

  2. *D designates a downregulated spot after exposure to dexamethasone whilst U designates upregulation.

  3. †−ve value indicates a decrease in fold change after exposure to dexamethasone whilst a +ve value designates an increase.

  4. ‡TrEMBL accession number (Computer annotated supplement to Swiss-Prot).

  5. §GenBank accession number (NIH genetic sequence database).

  6. ¶Swiss-Prot accession number (Protein knowledge database).

D4Phosphotyrosine independent ligand for the Lck SH2 domain p62 (sequestosome 1)Q13501Cell growth and maintenance77·78/5·29−1·9
D5Tubulin alpha (alpha tubulin)5174477§Structural50·15/4·94−2·2
D7Ionisine-5′- monophosphate dehydrogenase 2 (IMP dehydrogenase 2)P12268Cell growth and maintenance55·80/6·44No expression postdex
D8Tubulin beta-5 chain (tubulin 5 beta)P04350Structural49·67/4·78−5·3
D9Cytosolic non-specific dipeptidase (GNDP dipeptidase 2)Q96KP4Cell growth and maintenance52·78/5·66−1·9
D10Phosphoribosylglycinamideformyltransferase isoform 2 24660366§Cell growth and maintenance46·03/6·34−2·0
D11Gamma enolase (enolase 2)P09104Metabolism47·26/4·9−3·2
D13Eukaryotic translation elongation factor 1 gammaQ96CU2Protein synthesis50·12/6·25−4·8
D14Ubiquitin-like 1 activating enzyme E1A (SUMO-1 activating enzyme subunit 1)Q9UBE0Cell growth and maintenance38·35/5·11−2·1
D15Proliferating cell nuclear antigen (PCNA)P12004Cell growth and maintenance28·76/4·57−3·2
D16Tyrosine 3/tryptophan 5 monooxygenase activation protein, epsilon polypeptide 5803225§Signal transduction29·17/4·63−2·3
D17Inositol-(myo)1(or 4)-monophosphate 15031789§Cell growth and maintenance30·19/5·16−4·6
D21Tumour protein, translationally-controlled 1 (fortilin)4507669§Cell survival19·59/4·84−3·0
D22Deoxyuridine 5′-triphosphate nucleotidohydrolase 3041664§Cell growth and maintenance26·71/9·65−2·2
D24Tubulin alpha (alpha-tubulin) (part sequence)5174477§Structural22·21/4·94−5·1
U1Prolyl 4-hydroxylase alpha-1 subunit precursor (4-PH alpha-1)P13674Protein synthesis61·05/5·7+2·9
U2FK506 binding protein 5 (FKBP5)Q13451Signal transduction51·21/5·71+3·5
U3FK506 binding protein 5Q13451Signal transduction51·21/5·71+2·4

Interestingly, of the 15 spots identified to be downregulated, 10 proteins are involved in processes necessary for successful cell proliferation or survival. These include PCNA (D15), a cofactor for DNA polymerase which is involved in DNA repair; eukaryotic translation elongation factor 1 gamma (D13), a subunit of the elongation factor 1 complex which is responsible for the enzymatic delivery of aminoacyl tRNA to the ribosome; inosine-5′-monophosphate dehydrogenase 2 (D7) which is involved in the de novo purine synthesis pathway and, when inhibited, results in abrupt cessation of DNA synthesis; SUMO-1 activating enzyme subunit 1 (D14), an enzyme involved in protein degradation and the control of the levels of critical proteins within the cell; tyrosine 3/tryptophan 5 monooxygenase activation protein, epsilon polypeptide (D16), a member of the 14-3-3 family of proteins which mediates signal transduction by binding to phosphoserine-containing proteins, and interacts with RAF1; and fortilin (D21), a protein involved in cell survival and apoptosis regulation. Three of the 15 downregulated proteins are involved in maintaining the structure of the cell; two isoforms of tubulin α and tubulin β5 (Spots D5, D8 and D24).

Three proteins were identified as being upregulated following dexamethasone treatment. Two of these were identified as FKBP5, (U2 and U3), a protein known to be important in protein folding and trafficking (Vermeer et al, 2003). The presence of two spots suggests either the potential presence of two isoforms or evidence of post-translational modification such as glycosylation or phosphorylation. The third protein prolyl 4-hydroxylase alpha-1 subunit precursor (4-PH alpha-1) (U1) is a protein important in post-translational modification of proteins.

Correlation of global mRNA expression studies with global protein studies

To further determine the potential role of the proteins identified in this study, we compared the list of proteins with data from previously published gene expression profiling studies. Two previous studies have characterized the gene expression profile of MM.1S following dexamethasone exposure (Chauhan et al, 2002; Greenstein et al, 2003). At 6 h, 36 genes were identified as having a >2·5-fold change in expression and, at 24 h, 84 genes were noted to change by at least twofold. Greenstein et al (2003) compared the two independent gene expression data sets and identified 12 genes in common between the two studies; 11 were upregulated and one was downregulated. The common genes included those involved in GC signalling, growth factor receptor genes, enzymes and apoptosis. Two of the genes noted to have an altered RNA expression level in these studies (FKBP5 and P4HA1) were also identified in the current global protein analysis; FKBP5 and 4-PH alpha-1.

Treatment of MM.1R with dexamethasone

A comparative 2D analysis of MM.1R, the resistant cell line with a mutated GC receptor was also performed following treatment with 1 μmol/l dexamethasone for 24 h. Over the pH range 4–7, 19 spots were identified to be downregulated and seven upregulated following treatment with dexamethasone. The majority of these changes were seen in the acidic pH range, and within regions not previously studied in detail in MM.1S. However, five of the downregulated spots did correspond with those identified in MM.1S following dexamethasone exposure. Four of these spots had positive identifications; PCNA, tyrosine 3/tryptophan 5 monooxygenase activation protein, epsilon polypeptide and fortilin (Table I). Their downregulation in the MM.1R cell line suggests the presence of either residual signalling through the GR or off target signalling via an alternative pathway. The remainder of the changes seen when MM.1S was treated with dexamethasone were not seen when MM.1R was treated, suggesting these changes are specifically linked to GC signalling and drug-induced apoptosis. For example the spots corresponding to FKBP5, which were upregulated following treatment with dexamethasone, did not change the following drug exposure in MM.1R (Fig 2A).

Figure 2.

 Changes in the levels of FK binding protein 5 (FKBP5) protein and mRNA following dexamethasone treatment. (A) The upper panels demonstrate the induction of FKBP5 in MM.1S following dexamethasone exposure. The lower panels show the two spots corresponding to FKBP5 were not upregulated following exposure to dexamethasone in MM.1R. (B) quantitative real time polymerase chain reaction confirms the increase in FKBP5 levels in the dexamethasone sensitive cell line MM.1S, with little change in the resistant cell line MM.1R.

The steroid receptor complex following dexamethasone exposure

As FKBP5 was identified to be upregulated in MM.1S but not MM.1R in the mRNA expression studies (Chauhan et al, 2002; Greenstein et al, 2003) and in our global protein analysis, we chose to further characterize the association of FKBP5 protein with dexamethasone-mediated apoptosis. Initially, we confirmed the changes in mRNA expression using real time PCR. The sensitive cell lines (MM.1S) showed induction of FKBP5 on exposure to dexamethasone of 20·9-fold whilst the resistant cell line showed minimal induction of 1·3, confirming the global gene expression analysis and suggesting the inability to induce FKBP5 may correspond with dexamethasone resistance (Fig 2B). FKBP5 is a member of the Hsp90 steroid receptor complex that has been shown to be important in the maintenance of the conformation of the GR in a state capable of binding steroid (Pratt & Toft, 1997). A second immunophilin, FKBP4 is also associated with dexamethasone signalling through the steroid receptor (Davies et al, 2002) (Fig 3A). Thus the expression of the GR, the molecular chaperone Hsp90 and FKBP4 were also investigated. Protein levels of FKBP4 did not change following treatment with dexamethasone whereas the protein levels of FKBP5, were elevated from 8 to 24 h after exposure to dexamethasone confirming the 2D-PAGE findings (Fig 3B). The expression profile for the molecular chaperone Hsp90 remained constant whilst the profile for the GR decreased over time. To further confirm these findings were related to dexamethasone-induced signalling we investigated the expression level of these proteins in the dexamethasone resistant MM.1R cell line (Fig 3C). No changes in protein expression of FKBP5, FKBP4 or GC were observed.

Figure 3.

 The Hsp90 steroid receptor complex and dexamethasone. (A) In the absence of dexamethasone FK binding protein 5 (FKBP5) is the primary immunophilin. Upon dexamethasone binding FKBP5 is displaced by FKBP4, another immunophilin that allows the entry of the complex into the nucleus. The complex then dissociates allowing binding of glucocorticoid receptor (GR) to DNA binding motifs. (B) Western blotting of MM.1S protein after exposure to dexamethasone identified no change in expression of Hsp90 or FKBP4 with time, an increase in the expression of FKBP5 from 8 to 24 h, and a decrease in the expression of the GR. An actin antibody was used as a loading control. Images are representative of a triplicate experiment. (C) Western blotting analysis of MM.1S and MM.1R treated versus control protein after exposure to dexamethasone for 24 h identified no change in expression of Hsp90, FKBP5, FKBP4 or GR in MM.1R. Images are representative of a triplicate experiment.

The role of FKBP5 in dexamethasone resistance

Studies looking at GC resistance in other disease models have suggested that cells resistant to steroids do not increase FKBP5 on exposure to dexamethasone and laboratory measurements of FKBP5 may be useful as a screening test to predict clinical response to therapy (Vermeer et al, 2003). To investigate this in myeloma, we performed RQ-PCR for FKBP5 and FKBP4 expression on a series of cell lines and patients cells before and after treatment with dexamethasone (Fig 4A–C). Sensitive cell lines (Jim1 and MM.1S) and patients cells showed an induction of FKBP5 of between 5·4- and 20·9-fold upon exposure to dexamethasone. Cell lines known to be resistant to dexamethasone because of either downregulation (Greenstein et al, 2003) or mutation of the GR (Karkera et al, 1997) (MM.1R and U266B respectively) showed minimal induction of 1·3- to 2·6-fold respectively. A number of cell lines resistant to dexamethasone but with a functional GR also showed induction of FKBP5 (Jim3, JJN3 and H929) of between 4·2- and 10·8-fold. This suggests that upregulation of FKBP5 is an early event related to activation of the GC receptor and that other mechanisms of resistance downstream to the Hsp90 steroid complex are important in conferring steroid resistance in these cases. Minimal changes in FKBP4 mRNA levels were seen.

Figure 4.

 Investigation of the role of FK binding protein 5 (FKBP5) in resistance to dexamethasone. (A) 4,5-dimethylthiazol-2-yl-2,5-diphenyltetrazolium bromide (MTT) assay for cell viability of seven myeloma cell lines after exposure to 0·01–10 μmol/l dexamethasone for 24 h MM.1S (♦), MM.1R (bsl00001) cell lines Jim1 (bsl00066), Jim3 (△), U266B (bsl00041), JJN3 (bsl00043) and H929 (bsl00000). (B) Western blot analysis of the basal levels of GR expression in the cell line panel, demonstrating that all cell lines except MM.IR expression the glucocorticoid receptor. (C) Upper panel. Quantitative real time polymerase chain reaction identified an increase in FKBP5 expression in the dexamethasone sensitive cell lines MM.1S and Jim1 and patient cells after treatment, compared with the two cell lines with mutated GR, MM.1R and U266B. The resistant cell lines Jim3, JJN3 and H929 upregulated FKBP5 suggesting that further downstream mechanisms may mediate effects and confer resistance. Lower panel. Minimal change in FKBP4 was demonstrated following treatment with dexamethasone.


To date, studies in myeloma investigating dexamethasone-induced apoptosis have used pathway-specific approaches, however advances in technology now enable a more global approach to the analysis of protein expression changes to be undertaken. Because of the requirement for a relatively large amount of protein material to run 2D gels combined with subsequent mass spectrometry for protein identification we have used a myeloma cell line model to identify protein changes. Using the MM.1S cell line 27 spots were found to be significantly altered following dexamethasone treatment and 18 of these were identified. Importantly, the proteins identified were involved in cellular growth and maintenance, a number of which have been previously seen in global gene expression analyses. Our study of the resistant cell line MM.1R demonstrated five spots that were also upregulated in the MM.1S cell line, leading us to conclude that, despite the cell line being resistant to dexamethasone, there was either residual signalling through the GR or activation of an alternative signalling pathway. Interestingly, none of the proteins previously identified to be involved in dexamethasone-induced apoptosis using Western blot techniques were identified using this technique (Chauhan et al, 1999; De Bosscher et al, 2000a,b). This is probably because our methodology primarily looked for changes in protein expression rather than changes in activation status.

Importantly, this study demonstrated that using global protein analysis in combination with gene expression studies can reduce the complexity of changes identified at the RNA level and has identified a number of proteins important in mediating dexamethasone apoptosis. We have confirmed the relevance of this approach through the in depth study of FKBP5. Two protein spots corresponding to FKBP5, were noted to be increased following dexamethasone treatment in MM.1S but did not change in intensity in MM.1R suggesting that the upregulation of FKBP5 is involved in dexamethasone-mediated apoptosis through the GR-induced signalling pathway.

The FKBP5 is a 51 kDa immunophilin protein family member that plays a role in immunoregulation and basic cellular processes involving protein folding and trafficking. It is also a member of the Hsp90 steroid receptor complex (Pratt & Toft, 1997), which is one of the most abundant cellular chaperone proteins functioning to prevent protein aggregation, aiding in protein stabilization together with the facilitation of the activation of many of its bound proteins. The Hsp90 steroid receptor complex is important in the maintenance of the GC receptor in a conformation capable of binding steroid. Thus the induction of expression of one of the members of the Hsp90 steroid receptor complex by dexamethasone is indicative of a positive feedback loop through which signalling via the receptor can be maintained.

Based on in vitro data showing the induction of myeloma cell apoptosis and synergy with conventional chemotherapy and proteasome inhibitors, the Hsp90 inhibitors such as 17AAG and 17DMAG have recently been introduced into the clinic for the treatment of multiple myeloma (Mitsiades et al, 2006). Although HSP90 inhibitors can be combined with conventional chemotherapeutic agents, the results of the present study would suggest that the combination of a HSP90 inhibitor with dexamethasone should be used with care, as interfering with the chaperone function of the HSP90 steroid receptor complex may result in an antagonistic effect, blocking the delivery of the GC receptor apoptotic signal to the nucleus (Tago et al, 2004).

Previous studies in endocrine diseases have suggested the upregulation of FKBP5 may be used as a clinical surrogate marker for dexamethasone sensitivity. Our study suggests the upregulation of FKBP5 may not be a clinically useful marker in myeloma. We demonstrate that FKBP5 levels are upregulated in cell lines that are sensitive to dexamethasone, and levels remain stable in cells with resistance mediated by downregulation or mutation of the GR receptor. However FKBP5 was also upregulated in dexamethasone resistant cell lines with a functional GR, suggesting that FKBP5 upregulation is an early event and that other downstream events may mediate dexamethasone resistance in myeloma.

A number of mechanisms for dexamethasone resistance have been noted in myeloma and other diseases. In endocrine diseases one mechanism explaining drug resistance is the constitutive upregulation of FKBP5 reducing the affinity of the GR for its ligand (Reynolds et al, 1999; Denny et al, 2000; Vermeer et al, 2003). However this does not appear to be the case in myeloma patients as examination of the expression level of FKBP5 in 30 myeloma patients with newly presenting disease and relapsed refractory disease using the Affymetrix Chip U95AV2 oligonucleotide array showed no statistical difference in FKBP5 expression between cases sensitive and resistant to dexamethasone (data not shown). Other mechanisms include the binding of the myeloma cell to the bone marrow stroma, resulting in a protective effect because of increased secretion of pro-survival cytokines from the bone marrow microenvironment (Tu et al, 2000; Hideshima et al, 2001) and the overexpression of the oncogenes FGFR3 and MYC (Medh et al, 2001; Pollet et al, 2002). The results from this study would suggest that although downregulation and mutation of the GC receptor, with the associated lack of induction of FKBP5 and the HSP90 steroid receptor complex, are important in some myeloma cases, mechanisms downstream to the HSP90 steroid receptor complex contribute frequently to dexamethasone resistance in myeloma.

The further characterization of the upstream signalling proteins as well as the downstream effector genes may provide the framework for the improved use of dexamethasone in the treatment of multiple myeloma.


We would also like to acknowledge the help and expertise of Dr Alison Ashcroft and Dr Richard Unwin. This work is supported by funding from the Department of Health UK (KRU, ED and FED), the Leukaemia Research Fund UK (FED and GJM) and Cancer Research UK (RAC, SH, NFT and REB).