Abstract
Glucocorticoid resistance was investigated in human leukemic CCRF-CEM cells. A mutation (L753F), which renders the human glucocorticoid receptor (hGR) gene functionally hemizygous,was identified in all CEM-derived cell lines analyzed. Allele-specific PCR identified the same mutation in lymph node biopsy material from patient CEM cells. Given the correlation between hGR concentration and glucocorticoid sensitivity, this suggests that loss of functional heterozygosity may result in resistance to glucocorticoid-based chemotherapy. The L753F mutation was probably not responsible for the ontogeny of the disease because it did not appear to be present in all leukemic cells. Thus, it is unlikely that hGR mutations would be detected in leukemic patients at presentation, but they may occur, and be selected for, during treatment. Deletions and point mutations in the hGR gene of cells selected for steroid resistance in vitro were investigated by PCR-single strand conformation polymorphism analysis. Loss of hGR mRNA expression resulted from 5′-deletion of the hGR gene and nonsense mutations in exon 6. These results provide the first evidence for somatic mutation in the hGRgene of a patient with acute lymphoblastic leukemia, offer a potential in vivo mechanism for acquisition of steroid resistance in leukemia, and suggest that screening for additional in vivo mutations will require analysis of genomic DNA.
INTRODUCTION
Corticosteroids are commonly used in the treatment of leukemia and lymphoma. However, resistance to steroid therapy is a frequent phenomenon, and early studies showed that patients who relapse after single-agent induction of an initial remission are generally refractory to further steroid therapy (1, 2, 3, 4). Subsequent studies identified a correlation between reduced GR5expression and a poor prognosis after relapse in patients with acute lymphocytic leukemia, suggesting that reduced GR expression could lead to clinical resistance (5, 6, 7, 8, 9, 10). However, the mechanism(s)by which resistance arises, as well as the mechanism by which glucocorticoids induce lymphocytolysis, remains poorly understood.
The GR is capable of direct activation or repression of gene expression through interaction with positive or negative cis-acting regulatory elements in the promoters of hormonally responsive genes(11, 12). The GR can also indirectly regulate the expression of hormonally responsive genes through protein-protein interactions with other transcription factors (13, 14, 15). Recently, we showed that the glucocorticoid receptor mutant L753F,which is defective in transactivation but which retains the ability to repress AP-1-mediated activation of the collagenase promoter, is unable to mediate a lympholytic response in the human leukemic T cell line 6TG1.1 (16, 17). In addition, using mice in which the normal GR gene had been replaced by a mutant unable to form homodimers and bind to DNA, Reichardt et al.(18) showed that although glucocorticoid treatment resulted in repression of collagenase activity, there was no steroid-induced thymolysis. Thus, it appears that the ability of glucocorticoids to induce a lympholytic response is dependent upon the ability of the GR to mediate direct activation or repression of target genes. In addition, analysis of both human and mouse cell lines has shown that the principal mechanism for in vitro acquisition of glucocorticoid resistance is somatic mutation in the GRgene (16, 19, 20, 21, 22, 23, 24). However, with the exception of alternatively spliced hGR transcripts, which have also been identified in normal individuals (25), functional or structural alterations in the GR in the cells of leukemic patients have not been identified (26). In part, this may be attributable to the fact that even a small percentage of resistant cells in an otherwise sensitive population would result in a poor clinical outcome, making it extremely difficult to identify aberrant GR structure or function in a background of normal receptor. Consequently, the value of in vitro model systems in providing insight into the mechanism of acquired in vivo drug resistance remains unclear.
CEM cells provide one of the most widely used model systems for investigating the mechanism of glucocorticoid-induced cell lymphocytolysis and the in vitro acquisition of steroid resistance in human cells. The original cell line CCRF-CEM was established from a patient with acute lymphoblastic leukemia(27). Subsequent analysis revealed that there was wide cell-to-cell variation in the degree of glucocorticoid sensitivity,leading to the establishment of clonal cell lines (28). Some of these clonal cell lines were extremely sensitive to steroid-induced lymphocytolysis, whereas others were completely resistant (28). Analysis of the GR genes in several of the glucocorticoid-sensitive clonal cell lines showed that although there is one normal GR gene, the second gene contains the mutation L753F (16, 20). Thus, the genotype of glucocorticoid-sensitive CEM cells is GR+/GR753F.
Analysis of clonal cell lines that were originally glucocorticoid-resistant demonstrated that resistance was the result of a defect downstream from the GR; steroid sensitivity could be restored by treatment with 5-azacytidine, and somatic hybrids between inherently steroid-resistant cells and cells lacking functional GR were steroid-sensitive (29, 30). However, Geley et al. (31) have reported that steroid-resistant CEM-C1 cells also contain one copy of the L753F mutation. This raised the possibility that the L753F mutation is present in all CCRF-CEM-derived cell lines, and perhaps even in the leukemic cells of the patient from whom CCRF-CEM was isolated. To test this possibility, we have used a variety of techniques to identify mutations in the GR gene from steroid-resistant cell lines isolated in vitro, as well as in archival biopsy material obtained from the patient from whom CCRF-CEM was isolated. Our results clearly show that, not only is it possible to readily identify mutations in the GR gene in resistant cells isolated in vitro, but that the L753F mutation present in both glucocorticoid-sensitive and -resistant cells in vitro is detectable in cells obtained from the patient from whom the original cell line was isolated. These results provide the first evidence for somatic mutation in the hGR gene in leukemic cells in vivo, and suggest that clinical resistance may, in some cases, be a consequence of such somatic mutations.
MATERIALS AND METHODS
Cells and Cell Culture.
The isolation and growth of the glucocorticoid-resistant T cell line,CEM-C1, and the glucocorticoid-sensitive CCRF-CEM-derived T cell line 6TG1.1 have been described previously (28, 32, 33). The derivation, from glucocorticoid-sensitive CEM cells, of the glucocorticoid-resistant mutant cell lines ICR27TK.3, 4R4, and BLMB1 have also been described (32, 33, 34). Cells were maintained in RPMI 1640 containing 10% fetal bovine serum and grown at 37°C in a humidified atmosphere containing 5% CO2, as described previously (33). Cell number was determined with a Coulter Counter (model ZM; Coulter Electronics, Inc., Hialeah, CA).
Paraffin Blocks.
Fixed and paraffin-embedded biopsy material from patient CEM (case number A64-307) was provided by Edmund C. Matczak (Department of Pathology, Children’s Hospital, Boston, MA). During storage, many of the blocks had become fused, and the identifying tags had become dislodged. It was therefore not possible to correlate individual blocks with specific node biopsies.
Primers.
Primer P1 (5′-CCAATTTGGAAGCCTGATC-3′), containing sequence from the middle of intron H, and primer P2 (5′-CGACTTTCTTTAAGGCAACCATT-3′),containing sequence from the 3′-UTR of exon 9, were used to amplify the coding region of exon 9 for SSCP analysis. Primer P3(5′-TTGCAGGTGGTTGAAAATCTCC-3′), containing sequence from the 3′-end of intron H, and primer P4 (5′-CCTCTACAGGACAAACTGATAG-3′), containing sequence from the 3′-UTR of exon 9, were used to amplify the region of the hGR LBD encoding codon 753 from genomic DNA. Primer P5(5′-AGGAAAAGCCATTGTCAAGAGG-3′), containing exon 8 sequence, and primer P4 were used for amplification of the region of the hGR cDNA encoding residue 753. Primers P6 (5′-CTCATACCTTTATTTCTCTT-3′) and P7(5′-GGGAAAATGACACACATACA-3′), containing sequences from introns E and F, respectively, were used to amplify exon 6 of the hGR from genomic DNA. Primers P (5′-CTTAACTATTGCTTCCAAACATT-3′), Q(5′-TCGACTTTCTTTAAGGCAACCA-3′), A (5′-ggggcgggcgCCCGAGATGTTA-3′), and B(5′-ggggcgggcgTGATGATTTCAGCA-3′) were used for allele-specific PCR (see below). Lowercase letters in primers A and B indicate the clamp sequence.
Isolation and Amplification of Genomic DNA.
Genomic DNA was isolated from 107 cells as described (35). DNA was isolated from paraffin-embedded samples essentially as described (36, 37). Tissue scraped from paraffin blocks was incubated in 200 μl of digestion buffer A[10 mm Tris-HCl (pH 8.0), containing 100 mmNaCl, 25 mm EDTA, 0.5% SDS, and 0.1 mg/ml proteinase K]at 37°C for 5 days. After centrifugation, at 15,800 × g for 5 min, the supernatant was extracted once with phenol:chloroform:isoamyl alcohol (25:24:1), once with chloroform, and precipitated at −20°C with 100% ethanol in the presence of ammonium acetate. Nucleic acid recovered after centrifugation was washed with 70% ethanol, resuspended in 50 μl of H2O, and stored at 4°C. Alternatively, tissue sections were incubated in 100μl of digestion buffer B [10 mm Tris-HCl (pH 8.0), containing 100 mm NaCl, 25 mm EDTA, 0.1% Tween 20, and 0.1 mg/ml proteinase K] at 55°C for 3 h. After centrifugation at 15,800 × g, the supernatant was incubated at 95°C for 8 min, and 1 μl was directly used for PCR. Genomic DNA isolated from tissue culture cells was amplified in a 100 μl reaction containing 10μl of 10× PCR buffer [100 mm Tris-HCl, 500 mm KCl (pH 8.3)], 2.5 mmMgCl2, 250 nm of each deoxynucleotide triphosphate, 20 pmol of each primer, and 1 μg of DNA. After denaturation at 95°C for 2.5 min, the reaction was initiated by addition of 2.5 units of AmpliTaq DNA polymerase (PE Biosystems, Foster City, CA). Amplification of exon 9 was accomplished using 25–30 cycles of incubation at 95°C for 30 s, 48°C for 30 s, and 72°C for 1 min, followed by a final extension at 72°C for 4 min. Amplification of exon 6 was performed under the same conditions, except that primers were annealed at 58°C. Amplification of the region of the LBD containing codon 753 with primers P3 and P4 was performed under the same conditions, except that primers were annealed at 55°C. Amplified samples were extracted with phenol:chloroform:isoamyl alcohol (25:24:1) and precipitated in ethanol or purified on Ultrafree-MC spin filters (30,000 NWML; Millipore Corp.,Bedford, MA). Alternatively, PCR products were purified using the QIAquick Spin PCR Purification kit (Qiagen, Inc., Valencia, CA)according to the manufacturer’s instructions.
Allele-specific PCR.
Allele-specific PCR was performed as described by Liu et al.(38). DNA extracted from archival samples (0.9–1.5 μg)was amplified in 100 μl of 10 mm Tris-HCl buffer (pH 8.3) containing 50 mm KCl, 2.5 mm MgCl2, 0.25 mm each deoxynucleotide triphosphate, and 2.5 units of AmpliTaq DNA polymerase. Primers P, Q, wild-type forward primer A, and mutant reverse primer B were used in varying concentrations to optimize the yield of fragments PB (73 bp) and AQ(122 bp). After an initial incubation at 95°C for 3.5 min,amplification was performed by 60 cycles of incubation at 95°C for 30 s, 55°C for 1 min, and 72°C for 2 min, followed by a final extension at 72°C for 7 min. Amplified products were purified using the QIAquick Spin PCR purification kit.
RT-PCR.
Total cell RNA was isolated from 107 cells using RNAzol B (Tel-Test, Inc., Pearland, TX) according to the manufacturer’s instructions. Final pellets were suspended in H2O at a concentration of 1 mg/ml, and 1.5 μg were reverse transcribed in 30 μl of 50 mm Tris-HCl (pH 8.8) containing 50 mm KCl, 2.0 mmMgCl2, 1.33 nm random hexamers, and 50 units of Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc., Rockville, MD) at 37°C for 1 h. A single Ampliwax PCR Gem 100 wax bead (PE Biosystems) was added to each reaction mixture and melted, and the mixture was brought to room temperature. Each reaction was overlayed with 45 μl of Tris-HCl (pH 8.8) containing 50 mm KCl, 20 pmol of each primer, and 2.5 units of AmpliTaq DNA polymerase. Amplification was performed by 40 cycles of incubation at 95°C for 30 s, 55°C for 30 s, and 72°C for 1 min, followed by a final extension at 72°C for 4 min. Amplified products were extracted with phenol:chloroform:isoamyl alcohol (25:24:1), precipitated in ethanol, and fractionated by electrophoresis in 5% NuSieve agarose gels (FMC Bioproducts, Rockland,ME) either before, or after, digestion with AluI. Fragments were stained with ethidium bromide.
SSCP.
SSCP was performed essentially as described by Orita et al.(39). A single primer was 32P-end-labeled with[γ-32P]ATP (Amersham Pharmacia Biotech, Inc.,Piscataway, NJ) using T4 polynucleotide kinase. One labeled primer (P2 for exon 9 and P6 for exon 6) and one unlabeled primer (P1 for exon 9 and P7 for exon 6) were then used to amplify genomic DNA as described above. Amplified DNA (10 μl) was added to 90 μl of 10 mm EDTA containing 0.1% SDS, and 2 μl of each sample were mixed with an equal volume of sample buffer (95% formamide containing 20 mm EDTA, 0.05% xylene cyanol, and 0.05% bromphenol blue), heated at 80°C, and loaded onto 16 × 16 × 0.1-cm nondenaturing 6% polyacrylamide gels (acrylamide:bisacrylamide, 29:1) containing 10% glycerol. Electrophoresis was performed at constant power (30 W) at 20°C (exon 6) or 30°C (exon 9). After electrophoresis, gels were transferred to 3MM paper and dried under vacuum at 80°C; 32P-labeled DNA was visualized by autoradiography.
Analysis of BclI and AluI Polymorphisms.
Genomic DNA was digested with BclI and analyzed by Southern blotting using a probe specific for the 5′-portion of the hGR gene as described previously (40). Alternatively, genomic DNA amplified using primers P3 and P4 was digested with AluI, fractionated by agarose gel electrophoresis, and stained with ethidium bromide. In some cases, PCR of exon 9 was performed using 32P-end-labeled primer P3. In these cases, DNA was visualized by autoradiography after agarose gel electrophoresis.
Cloning and Sequencing.
Amplified DNA was cloned into pCRII (Invitrogen, Carlsbad, CA) as described by the manufacturer. DNA was sequenced manually as described previously (16). Alternatively, allele-specific PCR products were cloned into PCR2.1 (Invitrogen), and 300–500 ng of purified plasmid were cycle sequenced using T7 forward or M13 reverse sequencing primers, the Big Dye Cycle Sequencing System (PE Biosystems), and an ABI Prism 377 DNA sequencer (PE Biosystems).
RESULTS
Identification and Expression of the L753F Mutation in the hGR Gene in Glucocorticoid-resistant Cells.
We have shown previously that the glucocorticoid-sensitive CEM cell line 6TG1.1 contains one normal hGR gene and one mutant gene(L753F) that encodes a protein unable to induce transcription from a GRE-containing promoter and that cannot mediate an apoptotic response (16, 17, 41). The presence of this mutation in cells grown in culture can be readily detected by PCR-SSCP analysis of hGR exon 9. DNA isolated from IM-9 cells,homozygous for the wild-type hGR gene, yields a single electrophoretic species, whereas DNA isolated from 6TG1.1 cells,containing both mutant and wild-type genes, yields two electrophoretic species (Fig. 1,B). The more rapidly migrating of these species comigrates with DNA amplified from ICR27TK.3 cells, which contain only the mutant L753F hGR gene (16, 40), demonstrating that the more slowly migrating band is derived from the wild-type gene and that the more rapidly migrating band is derived from the mutant L753F gene. PCR-SSCP also identified the presence of a mutant hGR gene in CEM-C1 cells, confirming the results of Geley et al. (31), who showed that, like glucocorticoid-sensitive 6TG1.1 cells, these cells also contain one normal and one mutant hGR gene. Surprisingly, PCR-SSCP analysis also indicated the presence of both normal and mutant hGR exon 9 sequences in the glucocorticoid-resistant cell line 4R4 (Fig. 1 B, Lane 5). Analysis in this cell line of a BclI RFLP in the 5′ region of the hGR gene indicated that the 5′-portion of the gene is deleted (Ref.20; data not shown). Thus, steroid resistance in this cell line appears to be the result of a partial gene deletion.
In addition to generating a conformation detectable by SSCP analysis,the L753F mutation results in the loss of an AluI site in exon 9, creating a fortuitous RFLP that can also be used to identify the presence and expression of the mutant hGR gene. To confirm the results of the PCR-SSCP analysis, DNA isolated from glucocorticoid-sensitive and -resistant cell lines was therefore examined for the presence of this polymorphism. AluI digestion of DNA amplified from IM-9 cells generated the three fragments (37, 85, and 119 bp) characteristic of the wild-type gene,whereas digestion of DNA isolated from ICR27TK.3 cells yielded only two fragments of 37 and 204 bp (Fig. 2,B). As expected, DNA isolated from glucocorticoid-sensitive 6TG1.1 cells and glucocorticoid-resistant CEM-C1 cells yielded a combination of the two patterns, indicating the presence of both the wild-type and mutant genes and confirming the results of the SSCP analysis. A combination of both patterns was also obtained after AluI digestion of DNA isolated from 4R4 cells, providing additional evidence that the 3′ region of the wild-type hGRgene is present in these cells. This is more clearly seen in Fig. 2 C, where digestion of DNA amplified from 4R4 cells using 32P-labeled primer P3 yielded both the labeled 85-bp fragment characteristic of the normal gene and a labeled 204-bp fragment characteristic of the L753F gene. It therefore appears that steroid resistance in 4R4 cells is the result of a partial gene deletion.
Deletion of the 5′ portion of the wild-type gene could result in loss of hGR expression. Alternatively, it could result in fusion of the remaining portion of the hGR gene to an upstream gene and synthesis of a chimeric protein containing the COOH-terminal portion of the hGR under the control of a heterologous promoter. To distinguish these possibilities, RNA isolated from 4R4 cells was subjected to RT-PCR, and the amplified product was digested with AluI to identify the presence of transcripts containing wild-type and mutant hGR exon 9 sequences. In these experiments, the 5′ primer for amplification of the cDNA corresponds to sequence in exon 8 of the hGR gene(42, 43), thereby eliminating the possibility of amplification of contaminating genomic DNA. AluI digestion of RT-PCR products obtained from 4R4 cell mRNA yielded, exclusively,the 289- and 37-bp products predicted for the transcripts containing the mutant sequence (Fig. 3 B). The same pattern was seen after digestion of RT-PCR product prepared from ICR27TK.3 cells in which the entire wild-type hGR gene is deleted as the result of an unbalanced translocation involving chromosomes 5 and 15 (40). Fragments of 119 and 170 bp, characteristic of digestion of material amplified from wild-type mRNA, were not detected after digestion of material amplified from 4R4 cell mRNA, indicating that transcripts containing wild-type sequence are not expressed in 4R4 cells. These fragments were easily detected after digestion of material prepared from IM-9 cells, which contain two copies of the wild-type gene. In addition, consistent with the presence of both normal and mutant genes in 6TG1.1 and CEM-C1 cells, digestion of material obtained after RT-PCR of mRNA from these cells yielded all four fragments. Thus, deletion of the 5′ portion of the hGR in 4R4 cells does not result in the synthesis of a chimeric transcript containing sequence encoding the LBD of the hGR.
The L753F Mutation Was Present in the Patient from Whom CEM Cells Were Established.
The results presented above and those from other laboratories suggest that the L753F mutation is present in all cell lines derived from CCRF-CEM cells. This raises the question of whether this mutation was also present in the leukemic cells of the individual from whom this cell line was isolated. To address this question, DNA was isolated from paraffin blocks containing biopsy material obtained from patient CEM during the course of her therapy (27). Unfortunately, DNA isolated from paraffin blocks was badly degraded and proved a poor substrate for PCR amplification of fragments of the size used for SSCP or AluI RFLP analysis. Allele-specific PCR of shorter fragments, using one forward primer (primer A) containing wild-type sequence and one reverse primer (primer B) terminating in the mutant sequence, was therefore used to examine archival DNA for the presence of the L753F mutation (Fig. 4,A). As expected, allele-specific PCR of DNA isolated from IM-9 cells yielded wild-type fragment AQ, whereas allele-specific PCR of DNA isolated from ICR27TK.3 cells yielded only mutant fragment PB(Fig. 4,B). In contrast, DNA isolated from the GR+/GR753F cell line 6TG1.1 yielded both fragments. When archival DNA was examined using conditions optimized for amplification of mutant fragment PB, the 73-bp fragment containing the L753F mutation was identified in four of seven samples examined (Fig. 4,C, Lanes 6, 8, and 10 and data not shown). Archival material also contained the wild-type hGR gene, because fragment AQ was detected after amplification of the same samples using conditions optimized for amplification of this fragment (Fig. 4 C, Lanes 7 and 9 and data not shown). It is highly improbable that these results reflect contamination with exogenous DNA, because reactions from which template was omitted yielded no identifiable fragments, and no mutant band was ever detected after amplification of IM-9 cell DNA.
To confirm the presence of the L753F mutation in archival material, DNA isolated from two different samples was amplified,cloned, and sequenced. Table 1 shows the frequency with which the L753F mutation was detected. Four of 44 clones examined contained the L753Fmutation, establishing the presence of this mutation in at least some of the leukemic cells of patient CEM. However, the frequency was substantially lower than the 0.5 expected if every cell contained one copy of the mutant gene, suggesting that not every leukemic cell contained this mutation or that the biopsy material was contaminated with normal cells. In either case, however, these results establish,for the first time, the presence of an hGR mutation in a patient with acute lymphoblastic leukemia.
Nonsense Mutations Result in Loss of hGR mRNA Expression.
Although allele-specific PCR is well suited to the identification of known mutations, it cannot be used to identify unknown mutations. To identify other mutations in the hGR gene responsible for steroid resistance, PCR-SSCP was used to scan for mutations in exons 3–9 of the hGR gene in glucocorticoid-resistant cell lines derived from glucocorticoid-sensitive 6TG1.1 cells after mutagenesis with bleomycin (40). The results for PCR-SSCP of exon 6 of the hGR genes in glucocorticoid-resistant BLMB1 cells are presented in Fig. 5,A. In addition to the normal electrophoretic species that comigrates with material amplified from IM-9 and various CEM-derived cell lines, an additional band with slightly greater mobility was detected. Sequencing of genomic clones derived after amplification of exon 6 revealed the presence of clones with normal exon 6 sequence,derived from amplification of the gene containing the L753Fmutation, and clones containing a single C→T mutation in the first nucleotide of codon 615, resulting in the creation of a premature stop codon in the hGR mRNA (Fig. 5 B). The genotype of BLMB1 cells is therefore GRQ615Stop/GR753F.
In some genes, the occurrence of a nonsense mutation in an exon upstream from the terminal coding exon results in loss of expression of mature mRNA (44, 45). To determine whether the hGR belongs to this class or whether upstream nonsense mutations could potentially result in the synthesis of truncated receptor proteins lacking portions of the LBD, RT-PCR of exon 9 was used to assess mRNA expression from the mutant GRQ615Stop gene, as well as other bleomycin-induced mutant hGR genes. AluI digestion of material amplified from all cell lines yielded the 289-and 37-bp fragments characteristic of mRNA synthesized from the mutant L753F gene. However, the 119- and 170-bp bands,characteristic of mRNA containing normal exon 9 sequence, were absent after digestion of material amplified from BLMB1 cell mRNA (Fig. 5 C, Lane 6), indicating that these cells contain little if any mRNA from the mutant Q615Stop gene. A comparable result was seen for mRNA isolated from BLMA4 cells in which a different nonsense mutation (GRL587Stop) is present in exon 6,6but not for mRNA isolated from glucocorticoid-resistant BLMA2 cells, in which there is an R611I missense mutation in exon 6, or from mRNA isolated from hGR mutants BLMA2 and BLMA5, which contain nonsense mutations in codon 772 of exon 9.6 Thus, nonsense mutations in upstream exons of the hGR gene result in loss of mRNA expression.
DISCUSSION
The presence of the L753F mutation, routinely found in one of the hGR genes of glucocorticoid-sensitive and glucocorticoid-resistant CCRF-CEM cell lines (16, 20, 23, 31), was identified in biopsy material from the patient (CEM)from whom the cell line CCRF-CEM was derived (27). Polymorphisms have been identified in the coding and noncoding regions of the hGR gene (46, 47, 48, 49), some of which have been associated with accumulation of abdominal fat and obesity(50, 51), hyperinsulinemia (52), or increased corticosteroid sensitivity (53). Inherited germ-line mutations have been identified in patients with generalized inherited glucocorticoid resistance (43, 54, 55), and a de novo dominant-negative germ line mutation has been identified in a patient with sporadic generalized glucocorticoid resistance(56). In addition, a somatic mutation in exon 2 has been identified in a patient with Nelson’s syndrome (57). However, despite the widespread use of glucocorticoids in the treatment of leukemias, lymphomas, and multiple myeloma, somatic hGRmutations responsible for the acquisition of glucocorticoid resistance in malignant human cells has only been demonstrated in cultured cells(16, 20, 21, 22, 23, 24, 31, 58).
The frequency of genomic clones containing the L753Fmutation was substantially <50%, indicating that only a fraction of the leukemic cells contained the L753F mutation, and suggesting that this mutation occurred during progression of the disease. Although it is not possible to state with certainty at what point in the progression of the disease the biopsy material was obtained, it appears that several rounds of corticosteroid therapy preceded the collection of material (27). It is therefore unlikely that this mutation contributed to the development of the disease. However, numerous in vitro studies have established a direct correlation between receptor concentration and the ability of corticosteroids to induce a lytic response in sensitive T cells(59, 60, 61). Indeed, we showed recently that a 2-fold increase in hGR concentration is sufficient to render otherwise steroid-resistant CEM cells sensitive to the growth-inhibitory and cytotoxic effects of dexamethasone (62). In addition,several studies of leukemic blasts have found a direct correlation between hGR concentration and clinical prognosis (5, 6, 7, 63, 64, 65), and loss of heterozygosity at the hGR locus has been identified in a substantial proportion of patients with adrenocorticotropic hormone-secreting pituitary adenomas in which negative feedback of adrenocorticotropic hormone production is attenuated (66). Thus, loss of one functional copy of the hGR gene may result in decreased glucocorticoid sensitivity and/or partial glucocorticoid resistance.
The L753F mutation could have arisen spontaneously and been selected during multiple rounds of chemotherapy. Alternatively,one of the other drugs in the combination regimens used for treatment could have induced this mutation in the hGR gene. Indeed, we have shown previously that several common chemotherapeutic agents are capable of in vitro induction of mutations in the hGR gene that result in steroid resistance. Using the fact that CEM cells are heterozygous for the BclI RFLP in the 5′portion of the hGR gene, it was shown previously that the hGR gene is deleted in two cell lines derived after chemical mutagenesis or treatment with the radiomimetic agent bleomycin(40). Using this same polymorphism, Ashraf and Thompson showed that the 5′-end of the hGR gene was deleted in the spontaneously derived glucocorticoid-resistant cell line 4R4(20). To assess the integrity of the 3′-end of the hGR gene, the L753F polymorphism in exon 9 was examined. Surprisingly, the 3′-end of the wild-type gene was intact in 4R4 cells, demonstrating that, at least in vitro,glucocorticoid resistance can result from partial deletion of the hGR gene. The presence of a microdeletion at the boundary of exon 6 and intron F of the hGR responsible for glucocorticoid resistance in a family with familial glucocorticoid resistance demonstrates that in vivo deletions are also possible (67). Thus, although the L753Fmutation is unlikely to be present in other leukemic patients, a sufficient number of naturally occurring polymorphisms in the 5′- and 3′-ends of the hGR gene have been identified to allow screening for full or partial deletion of the hGR gene in clinical samples. Identification of additional polymorphisms would facilitate the identification and mapping of even smaller deletions.
The L753F mutation causes an increased rate of ligand dissociation from both the unactivated and activated hGR(34). However, this is accompanied by an increased rate of association of ligand with the unactivated receptor, resulting in an apparently normal equilibrium dissociation constant (34). Consequently, assays that measure the concentration of unactivated hGR would fail to detect any abnormality in receptor number or affinity. This is probably true of other mutations in the LBD, as well as in other regions of the receptor. In addition, as the data presented here suggest, not every cell in the leukemic population must necessarily contain a receptor mutant for therapy to fail, thereby limiting the sensitivity of assays that measure receptor concentration in a heterogeneous population of cells. Thus, simple measurement of receptor concentration, or affinity, is almost certainly inadequate to identify subtle receptor defects contributing to glucocorticoid resistance.
SSCP-PCR identified a point mutation creating a premature stop codon in exon 6 of the hGR. Interestingly, no mRNA from this mutant gene was detected. This is comparable to the lack of mRNA expression seen from the Δ4 allele in individuals with familial glucocorticoid resistance (43) and suggests that the hGRbelongs to that family of genes the transcripts of which are degraded if mutation introduces a premature stop codon (44, 45). Indeed, a nonsense mutation in any region of the hGR from the distal portion of exon 2 through exon 8 appears to result in loss of mRNA expression.6 Thus, although spontaneous or induced nonsense mutations could potentially occur with high frequency in the hGR gene, it is unlikely that they would result in the synthesis of a truncated protein with aberrant or constitutive function. The hGR, therefore, appears similar to the human androgen and vitamin D receptors, in which nonsense mutations in upstream exons result in substantial loss of mRNA and protein expression (68, 69). These results also suggest that the protein truncation test, widely used to detect nonsense mutations, or analysis of cDNA clones prepared from leukemic cell mRNA would fail to detect a significant percentage of hGRmutations resulting in functional loss of heterozygosity. Similarly,analysis of cDNAs by SSCP or other methods for detection of hGR coding region mutations would fail to detect nonsense mutations in exons 2–8, perhaps accounting for the failure of Soufi et al. (26) to find mutations in the leukemic cells of patients with chronic lymphocytic leukemia.
The presence of the L753F mutation was demonstrated in patient CEM, from whom the cell line CCRF-CEM was derived. The results presented here suggest that this mutation was not responsible for the ontogeny of the leukemia. However, it may have contributed to emergence of a glucocorticoid-resistant cell population. The determination of whether somatic mutation in the hGR gene is a significant cause of glucocorticoid resistance will require detailed analysis of the hGR gene in a variety of leukemias, lymphomas, and multiple myeloma.
ACKNOWLEDGMENTS
The opinions or assertions contained herein are the private ones of the authors and are not to be construed as official or reflecting the view of the Department of Defense or the Uniformed Services University of the Health Sciences.
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.
This work was supported by USPHS Grant CA32226 from the National Cancer Institute and Uniformed Services University of the Health Sciences Grant R075CW (to J. M. H.).
The abbreviations used are: GR, glucocorticoid receptor; hGR, human glucocorticoid receptor; LBD, ligand binding domain; SSCP, single-strand conformational polymorphism; UTR,untranslated region; RT-PCR, reverse transcription-PCR.
A. G. Hillmann, K. Multanen, J. Ramdas, and J. M. Harmon, manuscript in preparation.
. | No. of mutant clones . | No. of wild-type clones . | Frequency . |
---|---|---|---|
Sample 1 | 3 | 19 | 0.18 |
Sample 2 | 1 | 25 | 0.04 |
. | No. of mutant clones . | No. of wild-type clones . | Frequency . |
---|---|---|---|
Sample 1 | 3 | 19 | 0.18 |
Sample 2 | 1 | 25 | 0.04 |
Fragment PQ was amplified from two different archival samples, cloned into PCR2.1, and sequenced as described in “Materials and Methods.” The numbers of genomic clones containing mutant(L753F) and wild-type sequences are indicated.