Constitutional hypomorphic telomerase mutations in patients with acute myeloid leukemia

In the first cohort of 133 consecutive AML patients, we found 11 who carried a missense TERT gene variant (Table 1; Fig. 1): One was homozygous and seven heterozygous for A1062T TERT gene variant, one patient was heterozygous for a unique P65A TERT gene mutation, one patient was heterozygous for H412Y, and one was heterozygous for a unique R522K TERT mutation. None of these genetic variants were found among 198 gender-, age-, and ethnicity-matched controls (Fisher's exact test, P < 0.0001). Silent SNPs and intronic SNPs showed similar allele frequencies in patients, their matched controls, our previously reported controls (9), and the SNP database from the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/SNP/snpref.cgi?locusId = 7015), indicating a comparable genetic background among these groups [supporting information (SI) Tables S1 and S2]. No TERC mutations were observed in either patients or controls. Germ-line origin of mutations was established by analysis of nonhematopoietic tissues obtained from biopsy specimens in patients B (transbronchial lung biopsy to diagnose pneumonia), D (fine-needle liver biopsy for posttransplant hepatic complications), and H (buccal mucosa biopsy for oral lesion investigation); for patient H, the mutation also was detected in a relative from which peripheral blood cells were already available. None of the TERT-mutant patients had physical signs suggestive of DKC.

Cytogenetic analysis was available for 7 of the 11 TERT-mutant patients and was abnormal in all of them (Table 2), including trisomy 8, inversion 16, t(15;17), and complex rare karyotypes. About 55% of patients in this cohort would be predicted to have an abnormal karyotype, based on cytogenetic results in large series of AML patients (17), including those from Brazil (18); therefore, we inferred that mutation status could be associated with abnormal cytogenetics and that short and dysfunctional telomeres caused by telomerase mutations might contribute to genomic instability. In particular, trisomy 8 is usually found in 9% and inversion 16 in only 7% of AML patients (17).

To investigate the possible association between telomerase mutations and abnormal karyotype, we examined an additional 89 AML patients from a second referral institution (MD Anderson Cancer Center), who were selected based on the presence of cytogenetic abnormalities (9) (see Table 1). Again, we found a significantly higher allele frequency of missense TERT gene variants in patients, in comparison to 528 healthy controls in whom we had previously sequenced TERT (P = 0.028). Patient L was heterozygous for a unique mutation in codon V299M, patient M was homozygous for a codon 441E deletion, and patients N and O were heterozygous for an A1062T TERT gene variant (see Table 2). There was a statistically significant association between mutation status and trisomy 8 (P = 0.02) and a marginal association with inversion 16 (P = 0.08). Germ-line origin of mutations also was confirmed in 2 patients (L and O) from whom remission bone marrow slides (less than 5% blast cells) were available for genetic analysis; the other 2 patients never achieved remission and no leukemia-free sample was available for testing.

As A1062T TERT was the most common gene variant among AML patients and its allele frequency was 5 times higher in patients than in controls, we further screened 372 unselected AML patients from a third cohort diagnosed at the Leukemia/Bone Marrow Transplant Program (British Columbia Cancer Agency) specifically for the A1062T TERT gene variant. In this third cohort, the A1062T TERT allele was found in 10 patients: one was homozygous for this gene variant whereas the others were heterozygous. Among mutant patients, 6 had abnormal cytogenetic reports, including 1 with trisomy 8, 1 with inversion 16 and trisomy 22, 1 with t(8;21), and 3 (including the homozygous patient) with t(15;17). Trisomy 22 is usually found in only 3%, t(8;21) in 6%, and t(15;17) is found in 7% of AML cases (17). When the 3 cohorts were analyzed together (n = 594 patients), the A1062T TERT allele frequency was 3-fold higher in patients than in 1,110 controls (P = 0.0009) (Table 3).

Measurement of telomere length of TERT-mutant patients by Southern analysis indicated that telomeres were very short in most blast cells analyzed, independent of mutation status; mean telomere length was 3.1 kb (range, 2.4–5.9 kb) for mutant blast cells, slightly shorter that the mean observed for nonmutant blast cells (3.8; range 3.0–8.0 kb; P = 0.10; data not shown).

To test whether mutations decreased telomerase activity, vectors containing mutagenized TERT genes with AML-associated variants were transfected into WI38-VA13 cells along with TERC-containing vectors. The capacity of telomerase to elongate telomeres (telomerase enzymatic activity) was measured in transfected cell lysates. AML-associated TERT gene variants resulted in significantly reduced telomerase enzymatic function in comparison to wild-type TERT (Fig. 2). Cotransfection of mutant and wild-type TERT into VA13 cells at various ratios showed that mutations cause reduced telomerase activity by haploinsufficiency (see Fig. 2). In our previous study of 205 patients with apparently acquired aplastic anemia, two were found to be heterozygous for the TERT codon H412Y mutation and to have short-for-age telomeres (9). One of these patients had a family history of myelodysplasia. We also have previously demonstrated that this mutation reduces telomerase activity to ≈50% of wild-type TERT, also by haploinsufficiency (9). That TERT codon H412Y mutation is a hypomorphic variant impairing telomerase activity by haploinsufficiency has been reported by others in a family with DKC (19). In our experiments, the unique TERT codon V299M mutation did not appear to modulate telomerase's ability to elongate telomeres in vitro. However, this mutation is located in the N-terminal region, believed to be responsible for protein folding, and it is possible that telomerase interaction with its substrate or with other protein partners might be affected by this mutation. Telomerase also plays additional functions other than telomere elongation; for example, telomerase induces cell proliferation through transcriptional control of a Myc- and Wnt-related developmental program (20).

Frozen bone marrow samples were obtained at time of diagnosis from 100 consecutive patients with AML (excluding acute promyelocytic leukemia) who were diagnosed at a single institution (Division of Hematology, University of São Paulo at Ribeirão Preto Medical School, Ribeirão Preto, São Paulo, Brazil) between February 2004 and March 2006. AML was diagnosed when patients presented with ≥20% bone marrow myeloid blasts, according to World Health Organization criteria (32). AML cases were further categorized according to the FAB classification (33–35). Forty-nine patients were male and 51 female; median age was 50 years, ranging from 2 to 86 years old. Eighty-two percent of patients self-reported as White, 5% as Black, and 13% as mixed-race. Peripheral blood samples from 198 age-, sex-, geographic-, and ethnicity-matched healthy volunteers (blood donors at the Centro Regional de Hemoterapia) also were obtained as controls. Samples were collected and frozen at the time of diagnosis as part of a blood and bone marrow sample bank regulated by protocols approved by the local ethics committee; informed consent was waived by the ethics committee, and use for genetic analysis was approved by both local and Brazilian National Ethics Committee in Research (Comissão Nacional de Ética em Pesquisa, CONEP process number 25000.098750/2006–38). DNA was extracted by using the DNeasy blood kit (Qiagen). A second group of 89 patients with AML, diagnosed at MD Anderson Cancer Center and selected based on cytogenetic abnormalities were further studied. Diagnostic criteria were similar. Samples were collected after informed consent was obtained, according to the local institutional review board. Fifty patients were male and 39 were females; median age was 55 years, ranging from 18 to 84 years. Seventy percent of patients self-reported as White, 13% as Black, and 17% as Hispanic. According to cytogenetic studies, 9 patients were diploid, 10 had trisomy 8, 12 had monosomy 7, 7 had both monosomy 5 and monosomy 7, 8 had monosomy 5, 11 had t(8;21), 23 had inversion 16, and 9 had 11q⁻. Frozen cell lysates were homogenized in DNA STAT-60, and DNA extracted according to the manufacturer's instructions (IsoTex Diagnostics, Inc.). Blood or bone marrow samples from a series of 372 newly diagnosed adult AML patients from the Leukemia/Bone Marrow Transplant Program of British Columbia were obtained after informed consent was obtained with approval of the Clinical Research Ethics Board of the University of British Columbia. Diagnostic criteria and cytogenetic analysis were as described for the previous cohorts. Peripheral blood samples from 384 healthy subjects were obtained as controls. Samples consisted of anonymized leftover peripheral blood samples (age and sex available only) from individuals/patients over the entire age range for blood clotting tests at the University of British Columbia Cancer Center. Individuals had to have normal clotting results and no other illnesses for their sample to be included, and presumably represent a cross section of the normal population in Vancouver.

TERT sequencing was performed as previously described (9). Dot blot hybridization was performed as described previously (36). Briefly, PCR products were generated across exon 15 of hTERT from genomic DNA of AML patients and normal donors, and 3 to 100 ng per kilobase of each amplicon was spotted onto Hybond-N⁺ membranes (Amersham) to produce 96-well format dot blots. Allele specific oligonucleotides were end-labeled with ³²P and hybridized sequentially to the dot blots in tetramethylammonium chloride (Sigma) hybridization solution at 53 °C. Following posthybridization washes, the dot blots were exposed to Kodak Blue-XB1 X-ray film.

Telomere length of blast cells was determined by analysis of terminal restriction fragments obtained by digestion of genomic DNA with HinfI and RsaI restriction enzymes by using the TeloTAGG Telomere Length Assay (Roche Applied Science) following the manufacture's instructions. Briefly, one microgram of genomic DNA was digested at 37 °C overnight; DNA fragments were separated by 0.8% agarose gel electrophoresis and transferred to a nylon membrane by Southern blotting overnight. The blotted DNA was hybridized with a digoxigenin-labeled probe specific for telomeric sequences and incubated with anti-digoxigenin-specific antibody coupled with alkaline phosphatase and incubated with alkaline phosphatase highly sensitive chemiluminescent substrate. Average terminal restriction fragment lengths were determined by comparing to the molecular weight standard by using the ImageQuantTL digital analyzer (Amersham).

The mutations to encode TERT V299M, R522K, K570N, and A1062T were introduced into the pcDNA3-FLAG-TERT plasmid by using the QuikChange Site-Directed Mutagenesis Kit (Stratagene) and the Advantage-GC2 PCR kit (Clontech). Primers were designed by using the manufacturer's guidelines (see SI Appendix). Each 50-μl reaction contained 10-μl 5X GC2 PCR Buffer, 10-μl GC-Melt, 1-μl 50X GC2 Polymerase Mix, 5 nmol of each dNTP (Invitrogen), 1.25 ng of each primer, and 100-ng plasmid DNA. After amplification, DNA from each reaction was purified by using a PCR Purification Kit (Qiagen) following the manufacturer's instructions, except that DNA was eluted with 50-μl 1X New England Digestion Buffer 4 (New England Biolabs). To each eluate, 40 units DpnI (New England Biolabs) was added, and the reaction mixture was incubated at 37 °C for at least an 1 h. The plasmids were amplified in TOP-10 cells (Invitrogen) according to the manufacturer's protocol. Colonies were harvested and grown in liquid culture media (Luria-Bertani Broth) from which DNA was isolated (QIAprep Spin Miniprep Kit, Qiagen). To sequence the insert, the DNA was amplified by using the Advantage-GC2 PCR Kit as described above, except 10 pmol of a forward and a reverse primer were used (see SI Appendix); the amplicon was purified by using ExcelaPure 96-Well UF PCR Purification plates (Edge BioSystems) and direct sequencing was performed with BigDye Terminator version 3.1 (Applied Biosystems). Sequencing products were purified in sephadex G-50 (Sigma-Aldrich) columns (0.45-μm, 96-filter plate, Pall Corporation) for excess dye extraction. Specific primers for sequencing were designed (see SI Appendix), and sequencing products were analyzed in an automated genetic-sequence analyzer (ABI Prism 3100, Applied Biosystems). Plasmids with only the desired nonsynonymous mutation were purified (HiSpeed Plasmid Maxi Kit, Qiagen), and the clone sequence was confirmed. All PCRs were performed in an MJ Research PTC-200 Thermocycler (Global Medical Instrumentation). For TERT codon P65A and 441E deletion, wild-type vector was mutagenized by Mutagenex, sequence was confirmed by direct sequencing of the whole insert, and plasmids were purified by using the HiSpeed Plasmid Maxi Kit.

pcDNA3-Flag-TERC was kindly provided by Dr. Hinh Ly (Emory University). The sequence was confirmed, the plasmid was transformed into TOP-10 cells, and purified by using the HiSpeed Plasmid Maxi Kit.

Two micrograms of pcDNA3-Flag-TERC and 2 μg of either pcDNA3-Flag, wild-type pcDNA3-Flag-TERT, or mutant (P65A, V299M, R522K, K570N, A1062T) pcDNA3-Flag-TERT DNA were transfected into telomerase-deficient WI-38 VA13 cells (ATCC) at 60% confluence in each well of 6-well polystyrene dishes by using Superfect Transfection Reagent (Qiagen), according to manufacturer's instructions. VA13 cells were plated the day before transfection and grown in RPMI 1640 media (Mediatech) containing 10% FBS and antibiotics. This cell line, derived from human-lung fibroblasts, is telomerase-negative and maintains telomeres by a recombination-based pathway called ALT (alternative lengthening of telomeres). In cotransfection studies, wild-type and mutant pcDNA3-Flag-TERT were used in ratios of 3:1, 1:1, and 1:3 with a total of 2-μg pcDNA-Flag-TERT (wild-type and mutant) for each well. Subsequent manipulations of the cell extracts were carried out as described previously (16). Briefly, cells were scraped off the dishes 48 h after transfection, washed 3 times in PBS, and lyzed with CHAPS buffer. Total cellular RNA was extracted with TRIzol reagent (Invitrogen), treated with DNaseI (Roche Molecular Biochemicals), and purified with TRIzol reagent. Transfections were carried out at least twice and in triplicates for each mutation.

RNA concentrations were estimated by absorbance at 260 nm (Nanodrop). Expression of TERT, TERC, and β2-microglobulin (β2M) were assayed by RT PCR (TaqMan Gene Expression Assays, Applied Biosystems) according to the manufacturer's instructions with the exceptions listed below. A standard curve of 10-fold serial dilutions from 2.5 × 10⁸ to 250 copies pGRN145-TERT (ATCC) or TERC cDNA was run in parallel with 50-ng RNA from each treatment group. For TERT, a primer and probe premix were used (TaqMan, Applied Biosystems). For TERC, each reaction contained 900 nM of each primer and 250 nM probe (see SI Appendix). To more accurately quantify the amount of amplification-capable RNA for each sample, expression of B2M was assessed. A standard curve of 10-fold serial dilutions from 10⁶ to 1 copy pCMV6-XL5-B2M (Origene) and 50 ng of sample RNA (by A260) were run in parallel by using a probe and primer premix (Applied Biosystems). All standards and unknowns were run in triplicate, and each sample was run in the same position of the plate for measuring expression of TERT, TERC, and β2M. All plates were run in duplicate. A single PTC-200 Thermal Cycler (MJ Research) with a Chomo4 Fluorescence Plate Reader (MJ Research) was used for all of the aforementioned plates; the results were analyzed with MJ Opticon Monitor Analysis Software (Bio-Rad). All standard curves were linear (R²>0.99); all nontemplate controls were negative; the pcDNA3-FLAG transfected control was negative for TERT and TERC expression, confirming that the VA13 cells were not expressing telomerase; and all experimental values fell within the linear range. TERT:β2M ratio for each well was calculated, and the normalized, mean ratio for each transfection group was used to correct telomerase activity for variation in transfection efficiency.

Protein concentrations were assayed in triplicate twice for each sample according to the DC Protein Assay (Bio-Rad) instructions on a Victor3 (PerkinElmer) except that absorbance was taken at 650 nm instead of 750 nm. A standard curve of 2, 1, 0.5, 0.25, and 0 mg/ml BSA in CHAPS was run in parallel with samples to be run on the same telomerase activity plate. All standard curves were linear (R² > 0.99), and all experimental values fell within the linear range.

The fluorescent telomerase repeat amplification protocol (TRAPeze XL, Chemicon) was used to measure telomerase activity as described by the manufacturer. We used 250 ng of protein per unknown reaction; PCR was performed in a GeneAmp PCR System 9700 (PerkinElmer); and fluorescence was measured in a Victor3. Each standard, control, and unknown was run in triplicate. This procedure was performed twice for each transfection group each time the transfection was performed. All standard curves were linear (R² > 0.95) and all experimental values fell within the linear range. No negative controls, including CHAPS and RNA from the pcDNA3-FLAG-transfected group, showed more than 0.03% wild-type telomerase activity. Telomerase activity was corrected for the level of TERT expression in each transfection group.

Differences in the frequencies of coding-sequence variations between samples from patients and those from controls were evaluated by means of Fisher's exact test. For A1062T TERT gene variant, because of the large number of individuals analyzed in both groups, differences in minor allele frequency between patients and controls were determined by χ-square test. Differences in Telomere lengths between mutant and nonmutant patients were compared by using the nonparametric Mann–Whitney U test and considering a P value <0.05 as statistically significant.