Ann. N.Y. Acad. Sci. ISSN 0077-8923 A N N A L S O F T H E N E W Y O R K A C A D E M Y O F SC I E N C E S Special Issue: Hematopoietic Stem Cells X ORIGINAL ARTICLE Comparison of flow-FISH and MM–qPCR telomere length assessment techniques for the screening of telomeropathies Monica S. Ventura Ferreira,1,a Martin Kirschner,1,a Insa Halfmeyer,1 Natalia Estrada,2 Blanca Xicoy,2 Susanne Isfort,1 Margherita Vieri,1 Lurdes Zamora,2 Anne Abels,1 Anne-Sophie Bouillon,1 Matthias Begemann,3 Mirle Schemionek,1 Angela Maurer,1 Steffen Koschmieder,1 Stefan Wilop,1 Jens Panse,1 Tim H. Brümmendorf,1,a and Fabian Beier1,a 1 Department of Hematology, Oncology, Hemostaseology and Stem Cell Transplantation, Medical Faculty, RWTH Aachen University, University Hospital Aachen, Aachen, Germany. 2 Hematology Service, Institut Català d’Oncologia (ICO)-Hospital Germans Trias i Pujol, Institut de Recerca Contra la Leucèmia Josep Carreras, Universitat Autònoma de Barcelona, Badalona, Spain. 3 Institute of Human Genetics, Medical Faculty, RWTH Aachen University, Aachen, Germany Address for correspondence: Fabian Beier, Department of Hematology, Oncology, Hemostaseology and Stem Cell Transplantation, Medical Faculty, RWTH Aachen University, University Hospital Aachen, Pauwelsstrasse 30, 52074 Aachen, Germany. fbeier@ukaachen.de Assessment of telomere length (TL) in peripheral blood leukocytes is part of the diagnostic algorithm applied to patients with acquired bone marrow failure syndromes (BMFSs) and dyskeratosis congenita (DKC). Monochrome multiplex–quantitative polymerase chain reaction (MM–qPCR) and fluorescence in situ hybridization (flow-FISH) are methodologies available for TL screening. Dependent on TL expressed in relation to percentiles of healthy con- trols, further genetic testing for inherited mutations in telomere maintenance genes is recommended. However, the correct threshold to trigger this genetic workup is still under debate. Here, we prospectively compared MM–qPCR and flow-FISH regarding their capacity for accurate identification of DKC patients. All patients (n = 105) underwent genetic testing by next-generation sequencing and in 16 patients, mutations in DKC-relevant genes were identified. Whole leukocyte TL of patients measured by MM–qPCR was found to be moderately correlated with lymphocyte TL measured by flow-FISH (r2 = 0.34; P < 0.0001). The sensitivity of both methods was high, but the specificity of MM–qPCR (29%) was significantly lower compared with flow-FISH (58%). These results suggest that MM–qPCR of peripheral blood cells is inferior to flow-FISH for clinical routine screening for suspected DKC in adult patients with BMFS due to lower specificity and a higher rate of false-positive results. Keywords: MM–qPCR; flow-FISH; telomere length; telomeropathy; dyskeratosis congenita Introduction system is a tissue with one of the highest rates of cell turnover in the human body,2 which potentiates the Over the last decades, knowledge about telomere- importance of intact telomere maintenance mech- related pathologies has gained increasing impor- anisms for homeostasis of blood cell production. tance in understanding pathophysiology of Telomeres represent repetitive DNA sequences hematological malignancies1 in general and in and are located at the end of chromosomes. This hematopoietic stem cell disorders in particular. capping function protects chromosomes from This is mostly due to the fact that the hematopoietic fusion, degradation, and activation of DNA damage checkpoints.1,3 In most somatic cells, telomeres a These authors contributed equally to this work. shorten with each cell division due to the so-called doi: 10.1111/nyas.14248 Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences 93 published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Telomere length: flow-FISH versus MM–qPCR Ferreira et al. Figure 1. Schematic image of telomere length distribution, age, and occurrence of telomeropathies. Typical telomere attri- tion profile of patients with distinct manifestations of telomere diseases, such as Hoyeraal–Hreidarsson syndrome, dyskerato- sis congenita, aplastic anemia, and pulmonary fibrosis. The gap between short and extremely short telomere lengths (in kb) narrows with developing age (in years). Modified from Armanios, M. and E.H. Blackburn. 2012. Nat Rev Genet. 13: 693–704. https://doi.org/10.1038/nrg3246. [Correction added on November 1, 2019, after online publication: Reproduction information of Figure 1 was appended at the end of the figure legend.] end-replication problem.4 Telomere shortening 10 (NOP10), H/ACA ribonucleoprotein complex thus limits the proliferative capacity in human subunit 2 (NHP2), WD repeat–containing protein somatic cells by creating a barrier of replicative 79 (WRAP53, also known as telomerase Cajal body senescence once a critically short telomere length protein 1, TCAB1), telomerase reverse transcrip- (TL) is reached.5 tase (TERT), and telomerase RNA component Telomerase is a widely conserved enzyme that has (TERC), have been linked with the classical variant the ability of replacing lost telomeric sequences.3 of DKC.7–10 However, in approximately 40% of all High cell turnover tissues, such as the hematopoi- clinical DKC patients, no genetic aberration has yet etic system, are particularly susceptible to defects in been identified, and, consequently, the suspected telomere maintenance genes as, for example, genes diagnosis needs to be established primarily on the of the telomerase complex.6 Premature telomere basis of the patients’ clinical presentation and their loss due to defects in telomerase or other telom- family history and pedigree assesment.7–10 ere maintenance–related genes limits the prolifera- Nowadays, there is a growing body of evidence tion potential of cells, including stem cells, leading pointing to a significant incidence of late-onset to decreased tissue renewal capacity and premature hereditary telomere maintenance disorders first aging.7 manifesting themselves in adult patients. These Dyskeratosis congenita (DKC) was the first so-called “cryptic” DKC cases often present with inheritable disorder to be linked with impaired mono- or oligosymptomatic organ manifestations, telomere maintenance. Classical DKC patients are such as interstitial lung fibrosis, aplastic anemia,8,10 characterized typically by early clinical manifesta- or idiopathic liver cirrhosis. Diagnosis of cryptic tion during childhood and multiorgan involvement adult DKC is especially challenging as the pattern leading to bone marrow failure (BMF), skin and of affected organs and disease penetrance may nail affection, and idiopathic pulmonary fibrosis.8 vary extremely among individuals.11,12 In addi- Mutations in a variety of telomere maintenance tion, missed diagnosis occurs because physicians genes, such as dyskerin pseudouridine synthase typically do not expect the first manifestations of 1 (DKC1), regulator of telomere elongation an inheritable disorder in young or middle-aged helicase 1 (RTEL1), TERF1-interacting nuclear adults. Yet, the correct diagnosis is crucial for factor 2 (TINF2), conserved telomere mainte- further patient management: correctly identified nance component 1 (CTC1), nucleolar protein patients can undergo telomerase-activating therapy 94 Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences Ferreira et al. Telomere length: flow-FISH versus MM–qPCR using androgen derivatives13 or need adapted con- telomeropathy of the treating physician and/or ditioning protocols for allogeneic stem cell trans- the recommendations of the German Society of plantation to reduce the high mortality observed Hematology and Oncology (DGHO) published in DKC patients.14,15 In addition, proper genetic via Onkopedia (www.onkopedia.de). Samples were counseling of affected families, including suitable analyzed with flow-FISH and MM–qPCR, and, if donor selection in cases of available family donors deemed critically short, followed up for underlying for allogeneic stem cell transplantation, is crucial. mutations by next-generation sequencing (NGS). Accelerated shortening of TL represents the func- tional readout of directly or indirectly altered telom- DNA extraction erase function in telomeropathies. Consequently, DNA from the mononuclear cell (MNC) frac- TL measurement in peripheral blood leukocytes tion in the PB samples was extracted for MM– (PBLs) is recommended as a screening tool to iden- qPCR and NGS using the DNeasy R Blood & tify telomeropathies.6,16,17 However, due to the nat- Tissue Kit (Qiagen, Hilden, Germany) according ural age–dependent decline of TL in normal cells, to the manufacturer’s instructions. DNA was fur- the diagnostic window between shortened TL— ther cleaned with DNA Clean & Concentrator R - in relation to age-adjusted controls—and critically 5 kit (Zymo Research, Irvine, CA). For NGS, and functionally short telomeres (below the 1% 250 ng genomic DNA was used per reaction. For percentile of normal controls), which reflect an MM–qPCR, 4.98 ng genomic DNA was used per underlying telomere disease, narrows with advanc- reaction. ing age18 (Fig. 1). Targeted amplicon sequencing The actual diagnostic algorithm for germline NGS (MiSeq R , Illumina, Germany) was done as telomeropathies includes the assessment of TL previously described.20 Library preparation was from PBLs of all patients with suspected BMF done using the TruSeq R Custom Amplicon kit using fluorescence in situ hybridization (flow- (Illumina). Genetic variants/heterozygous muta- FISH) or monochrome multiplex–quantitative tions in telomere maintenance genes were screened polymerase chain reaction (MM–qPCR).10 Flow- using a self-designed panel containing the entire FISH provides the advantage of measuring TL of coding sequences for CTC1, DKC1, NHP2, NOP10, granulocytes and lymphocytes separately within RTEL1, TERC, TERT, TCAB1, USB1, and exon 6 of the same sample.19 If a patient’s mean telomere TINF2.20,21 content is either below the first percentile of a normal control cohort or alternatively between Flow-FISH the 1st and 10th percentiles and typical clinical Vital sterile frozen MNC from PB was used symptoms or the family history is suggestive of a for the flow-FISH analysis of TL, as previously telomeropathy, further genetic testing for inher- described.22–26 Briefly, samples were prepared for ited mutations in telomere maintenance genes is cell denaturation and mixed with an FITC-labeled recommended.10 telomere-specific (CCCTAA)3-peptide nucleic acid In this study, we aimed to compare the sen- FISH probe (Eurogentec, Liège, Belgium) for DNA sitivity and specificity of MM–qPCR and flow- hybridization followed by DNA counterstaining FISH in a clinical routine screening setting for with LDS 751 (Sigma). Bovine thymocytes were telomeropathies. used as internal controls. An FC 500 flow cytometer (Becton Dickinson) was used for data acquisition. Materials and methods All measurements were carried out single-blinded Patients in triplicates. TL of bovine thymocytes was deter- Peripheral blood (PB) samples of 105 patients mined by western blot (19.515 kb) and was used from different academic centers in Germany, Aus- to convert the TL of granulocytes and lympho- tria, and Switzerland were analyzed. After written cytes into kb. Healthy controls (n = 365) were informed consent was obtained, all samples were used for age adaptation of TL for flow-FISH, as taken according to the approval by the local ethics described previously.22–26 A separate cohort of 89 committee (EK206/09). Patient inclusion criteria healthy controls was used for age adaption of TL were based on the clinical suspicion for DKC or for MM–qPCR.27,28 Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences 95 published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences Telomere length: flow-FISH versus MM–qPCR Ferreira et al. Table 1. Characteristics of the Aachen Telomeropathie Registry patients Aachen Telomeropathie Registry (ATR) Mean age ATR (in years) 45 ± 21 Gender, male:female (n) 57:48 Clinics n = 105 AA 43 Cytopenia 14 MDS 9 Family members of and known DKC 17 PNH 12 Others (e.g., IPF) 10 Blood counts Leukocytes/nL (range) 4.4 (0.6–21.7) Hemoglobin g/dL (± SD) 10.5 ± 2.3 Platelets/nL (range) 78 (1–471) Identified mutations n = 16 RTEL1 2 TERC 6 TERT 8 MM–qPCR TL analysis by MM–qPCR followed the original protocol described by Cawthon et al.29,30 Essentially, primer pairs used for telomere amplification were telg 5 –ACACTAAGGTTTGGGTTTGGGTTTGG GTTTGGGTTAGTGT–3 and telc 5 –TGTTAGG TATCCCTATCCCTATCCCTATCCCTATCCCTA ACA–3 , while signal acquisition was at 74 °C. For reference, a signal for human beta-globin was acquired at 88 °C using the primers hbgu 5 –CGGC GGCGGGCGGCCGGGGCTGGGCGGCTTCAT CCACGTTCACCTTG–3 and hbgd 5 –GCCCGG CCCGCCGCGCCCGTCCCGCCGGAGGAGAA GTCTGCCGTT–3 , as previously described.23,27,28 All measurements were carried out as single- blinded in triplicates. Leukocytes from healthy Figure 2. Correlation between MM–qPCR (T/S ratio) and lymphocyte flow-FISH (kb). (A) Linear regression plots of subjects (n = 105) were used for age adaptation telomere lengths (TLs) measured by MM–qPCR or lymphocyte of TL, which is given in T/S ratios. A T/S ratio is flow-FISH for all ATR patients. Solid line depicts data’s best fit calculated by dividing the number of copies of the (r2 = 0.34; P < 0.0001; n = 105). (B) TLs for DKC patients. telomere template (T) by the beta-globin template Solid line depicts data’s best fit (r2 = 0.29; P = 0.03; n = 16). (C) (S). For Bland–Altman analysis, T/S ratios were Bland–Altman plot for agreement analysis of flow-FISH (kb) and MM–qPCR (T/S ratio) of the TLs of all ATR patients. Bias converted into kb using four cell lines with known ± 2 SD = 0.16, LoA ranging from 0.02 to 0.31. TL (Fig. S1, online only). Statistical analysis GraphPad Prism 5.0 (GraphPad, San Diego, CA) as statistically significant. The Tel to percentile was used for statistical analysis. Statistical analysis value was calculated based on the difference (kb or was carried out using Fisher-(exact)-test and Pear- T/S ratio) of the measured TL of the patient to the son correlation. P values < 0.05 were considered respective age-adapted 1% or 5% percentile. 96 Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences Ferreira et al. Telomere length: flow-FISH versus MM–qPCR Figure 3. Telomere length according to age in ATR patients measured by (A) MM–qPCR (T/S ratio), (B) lymphocyte flow-FISH in kb, and (C) granulocyte flow-FISH in kb. Percentile cutoffs to detect patients with extremely short telomere were calculated and the 1st, 10th, 50th, 90th, and 99th percentiles are represented by solid lines. The 5th and 95th percentiles are represented by a dashed line. Results or DKC family member of a patient with DKC in 16% (17/105), paroxysmal nocturnal hemoglobin- Patient cohort uria in 11% (12/105), and other diseases (for exam- The patient population studied here represents an ple, interstitial lung disease in 10% (10/105) of all adult population screened for late-onset hereditary patients). NGS was carried out in all 105 patients telomeropathies (age 45 ± 12 years; mean ± SD). and led to the identification of 16 patients with The initial diagnoses made by the treating physi- mutations in the genes TERC (n = 6), TERT (n = 7), cian were aplastic anemia in 41% (43/105), unex- DKC1 (n = 1), and RTEL1 (n = 2) (see Table 1 and plained cytopenia in 13% (14/105), myelodysplastic Table S1, online only). syndrome in 9% (9/105), clinically suspected DKC Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences 97 published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences Telomere length: flow-FISH versus MM–qPCR Ferreira et al. Correlation between flow-FISH served for granulocyte flow-FISH compared with and MM–qPCR MM–qPCR (Fig. S3A, online only). Direct comparison of flow-FISH for lymphocytes Next, we analyzed the specificity of lympho- with MM–qPCR led to a modest but highly signif- cyte flow-FISH and MM–qPCR. We observed a icant correlation between the two methodologies decrease in specificity of the lymphocyte flow-FISH (r2 = 0.34; n = 105; P < 0.0001; Fig. 2A) contrasting from 84% to 69% and 58% using either the 1st, with other studies, where weaker correlations were 5th, or 10th percentile. In comparison, MM–qPCR reported.31 When only molecularly proven DKC showed a more pronounced decline of specificity patients were considered, the correlation between from 78% to 46% to 29% using either 1st, 5th, or the methods was found to be lower (r2 = 0.29; 10th percentile, with statistically significant results n = 16; P = 0.03; Fig. 2B), in contrast with other only for the 10th percentile (P = 0.01, Fig. 4B). studies, where a stronger correlation for DKC Similar results were obtained for the comparison patients was reported.32 Overall, the range of corre- of the granulocyte flow-FISH with the MM–qPCR. lation values was low in line with previous studies For the granulocyte flow-FISH, specificity was 72%, showing similar correlations between flow-FISH 61%, and 56% for the 1st, 5th, and 10th percentiles, and MM–qPCR.31–35 respectively (Fig. S3B, online only). Specificity of Bland–Altman analysis confirmed that the over- granulocyte flow-FISH for the 10th percentile was all agreement between flow-FISH and MM–qPCR significantly higher compared with the MM–qPCR was poor. Mean ratios and limits of agreement (P = 0.02). (LoA) between flow-FISH and MM–qPCR were moderate, with an average of 0.28 kb (LoA, 3.71 to False-positive/-negative rate and −3.42 kb; Fig. 2C), as well as for DKC patients, with positive/negative predictive value of a mean ratio of –0.32 (LoA, 0.90 to –1.65 kb; Fig. 2). flow-FISH and MM-QPCR We then analyzed the rate of false-positive values. Direct comparison of flow-FISH and Lymphocyte flow-FISH displayed a rate of false MM–qPCR to detect DKC patients positivity of 13% with TL below the 1st percentile, To directly compare both techniques for their valid- increasing to 27% and 35% using the 5th and 10th ity to trigger further genetic workup, we used the percentiles, respectively. In comparison, the rate of 10th, 5th, and 1st percentiles as a trigger for genetic false-positive values achieved by MM–qPCR was testing (Fig. 3A–C), as percentiles are the estab- 19% using the 1st percentile, increasing to 46% and lished way to screen for telomeropathies. Applying 60% using the 5th or 10th percentile, respectively. NGS, 16 patients with known DKC-causing muta- Direct comparison did not reveal a statistical dif- tions were identified and used to analyze the ference for the 1st percentile, but yet a trend for the sensitivity, specificity, positive predictive value, 5th percentile (P = 0.06). The rate of false-positive and rate of false-positive values for flow-FISH and values for the 10th percentile was 60% for MM– MM–qPCR. Since flow-FISH provides results for qPCR and significantly higher compared with lym- TL of lymphocytes and granulocytes separately, phocyte flow-FISH with 35% (P = 0.03, Fig. 4C). both subpopulations—lymphocyte and granulo- A similar trend was found comparing the granu- cyte flow-FISH—were compared separately with locyte flow-FISH with MM–qPCR. MM–qPCR. The rate of false-positive values for the granulo- Sensitivity and specificity of flow-FISH cyte flow-FISH increased from 24% using the 1st and MM–qPCR percentile to 33% and 37% for the 5th and 10th Flow-FISH displayed a sensitivity of 75% with TL percentiles, respectively. In the 10th percentile, a below the 1st percentile, increasing to 88% using statistical trend (P = 0.054) favoring the granu- the 5th or 10th percentile, respectively. For MM– locyte flow-FISH compared with MM–qPCR was qPCR, the sensitivity was 69% using the 1st per- observed (Fig. S3C, online only). centile, increasing to 94% using the 5th or 10th In the next step, we analyzed the positive predic- percentile, respectively. Sensitivity between lym- tive value. We observed a decrease in lymphocyte phocyte flow-FISH and MM–qPCR did not differ flow-FISH from 46% to 33% and 27% using the 1st, significantly (Fig. 4A). Similar results were ob- 5th, or 10th percentile, respectively. MM–qPCR 98 Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences Ferreira et al. Telomere length: flow-FISH versus MM–qPCR Figure 4. Patient detection rates of MM–qPCR and lymphocyte flow-FISH according to the 1st, 5th, and 10th percentile cutoffs. (A) Sensitivity. (B) Specificity. (C) False-positive values. (D) Positive predictive value. Black bars represent the lymphocyte flow- FISH and light gray bars represent the MM–qPCR. showed a more pronounced decline from 35% We combined (1) the threshold of the 1% per- to 24% to 19% using 1st, 5th, or 10th percentile, centile of lymphocyte flow-FISH with the 1% respectively (Fig. 4D). Comparison of the positive percentile of the MM–qPCR, (2) the threshold predictive values did not differ significantly. of 5% of lymphocyte flow-FISH with the 1% per- Similar results were observed for the granulocyte centile of the MM–qPCR, and (3) 5% threshold of flow-FISH. Here, we observed a positive predictive lymphocyte flow-FISH with the 1% percentile of value of 35%, 24%, and 19% for the 1st, 5th, or 10th the granulocyte flow-FISH. Of note, all combina- percentile, respectively (Fig. S3D, online only). tions resulted in a nonsignificant lower sensitivity Finally, analysis of the negative predictive value as lymphocyte flow-FISH alone (Fig. 5A and and false negativity did not reveal any significant B). By contrast, all combinations showed a non- differences for the MM–qPCR group and granulo- significant trend (all P values ≤ 0.1) to higher cyte flow-FISH (Fig. S4A and B and Table S2, online specificity compared with lymphocyte flow-FISH only). alone (Fig. 5B). Analyzing the rate of false-positive values, the Combination of lymphocyte flow-FISH with combination of the 5% percentile cutoff for the MM–qPCR or granulocyte flow-FISH lymphocyte flow-FISH combined with the 1% Based on our previous results, lymphocyte flow- percentile of the MM–qPCR rates of false-positive FISH resulted in higher specificity and lower values (27% versus 12%, P = 0.01, Fig. 5C) was rates of false-positive values compared with MM– significantly lower compared with lymphocyte qPCR or granulocyte flow-FISH. Here, we analyzed flow-FISH alone. A similar trend was found com- whether a combination of lymphocyte flow-FISH bining lymphocyte and granulocyte flow-FISH with MM–qPCR or granulocyte flow-FISH can fur- (27% versus 16%, P = 0.14, Fig. 5C). ther improve the sensitivity, specificity, rate of false- Finally, focusing on the positive predictive val- positive values, or positive predictive value, espe- ues, we observed a nonsignificant tendency to cially when applied for the 5th or 10th percentile. higher values in all combination groups compared Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences 99 published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences Telomere length: flow-FISH versus MM–qPCR Ferreira et al. Figure 5. Different cutoffs of lymphocyte flow-FISH combined with MM–qPCR or granulocyte low-FISH. (A) Graphical depic- tion of different combinations. (B) Sensitivity. (C) Specificity. (D) False-positive values. (E) Positive predictive values. The cutoff of the 1% percentile of lymphocyte flow-FISH with the 1% percentile of the MM–qPCR, 5% of lymphocyte flow-FISH combined with the 1% percentile of the MM–qPCR, and the 5% threshold of lymphocyte flow-FISH with the 1% percentile of the granulocyte flow-FISH was used. with lymphocyte flow-FISH (Fig. 5D and Table S3, patients with newly diagnosed BMF syndromes. online only). Few studies have thoroughly compared and vali- dated the different available methods for telomere Discussion assessment within clinical routine.31,32 Previously Valid screening tools to reliably identify published studies were either limited by the analysis patients with a high likelihood of an underlying of cohorts of family members of DKC patients or telomeropathy are urgently needed, particularly compared both techniques without testing the for the diagnostic workup of adult and pediatric application for TL screening of these techniques 100 Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences Ferreira et al. Telomere length: flow-FISH versus MM–qPCR within a prespecified diagnostic algorithm.31,32 showed similar sensitivity and specificity. However, Hence, our study introduces a prospective by increasing the threshold to 5th or the recom- comparison of these two potentially applicable mended 10% percentile, flow-FISH of the lym- methods to the routine TL assessment on the phocyte subpopulation demonstrated significantly largest cohort of patients with clinical suspicion of higher specificity and lower rates of false-positive telomere disease so far. results compared with MM–qPCR. Hence, flow- We confirm previous observations showing FISH, when used to screen for suspected telom- that the two mostly used high-throughput TL eropathy might be a fast, reliable tool, even sparing assays—flow-FISH and MM–qPCR—do not cor- resources (e.g., less genetic testing) compared with relate constantly in their output,10,32 as several MM–qPCR. factors contribute to the difficulty in the reliable In addition, while previous data led to the recom- reproducibility of telomere measurements.36–38 mendation of a 10% threshold for further genetic Briefly, automated flow-FISH is considered the workup, our data illustrate that a 5% threshold leads method of choice for the clinical routine setting to equal performance.10,41 of DKC screening as the main advantages of the The question whether an absolute TL threshold flow-FISH technique are its high precision, repro- of, for example, 6.5 kb might potentially be more ducibility, and sensitivity.19 In addition, flow-FISH suitable to reliably identify DKC patients, partic- provides telomere distribution across different ularly at an older age, warrants further validation. leukocyte subsets. Main disadvantages are the high Interestingly, a combination of lymphocyte flow- effort and quality control requirements for trans- FISH and MM–qPCR or granulocyte flow-FISH did port, processing, and analyses of the samples,39 and not result in an increased sensitivity or specificity, prerequisite of a viable cell population. but in a slightly reduced rate of false-positive values. By contrast, MM–qPCR requires only small One limitation of our study is that the rate of amounts of DNA (i.e., it is not dependent on viable false-positive results might be lower since our NGS cells) and moderate method-related effort, exper- screening is not covering rare or unknown genetic tise, and diligence.29,30 Main disadvantages of the DKC–causing mutations. However, a previous study technique are the fact that high-quality DNA is using whole-exome sequencing showed a relevant needed and MM–qPCR does not provide telom- cluster of mutated genes in BMF patients covered ere distribution within different cellular subsets. by our NGS panel. Instead, results typically correspond to the TL aver- One explanation for our observation of the sig- age in all analyzed cells. This can be misleading nificant decrease of specificity and higher false- as, for example, specific hematopoietic subpopula- positive rates for MM–qPCR is the fact that MM– tions can be affected to different degrees by underly- qPCR TL analysis was performed on the whole ing acquired disease entities. As such, lymphocytes MNC fraction, including both the lymphoid and are mostly clonal in patients with chronic lympho- myeloid populations. Other factors, like DNA qual- cytic leukemia, whereas granulocytes are nonclonal. ity or intrinsic characteristics of MM–qPCR, might In noninheritable disorders, the latter can lead to play an additional role explaining the observed dif- contamination by nondisease-affected cells, namely, ferences with flow-FISH. nonclonal versus clonal cells, and misinterpretation Clinically, a correct TL assessment is of utmost of results. importance for affected patients and their fami- Therefore, no optimal screening method for lies and has immediate clinical consequences.14,42 inherited telomere maintenance disorders, such as First, prematurely shortened telomeres repre- DKC, has been clearly identified until now. In our sent the functional readout of impaired telomere current analysis, however, we found that flow-FISH maintenance.2 Second, NGS screening for relevant of lymphocytes so far is the most accurate method DKC mutations reveals various polymorphisms.43 to screen for DKC. Our findings were in line with Identification of true mutations is frequently dif- previous studies showing the advantage of lympho- ficult since the actual capacity of bioinformatics cyte flow-FISH over other techniques.18,40 As long tools to predict the functional role of a muta- as the 1% percentile was used to trigger further tion/polymorphism is limited.44 Our data suggest genetic workup, both flow-FISH and MM–qPCR that MM–qPCR-based TL between 1% and 10% Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences 101 published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences Telomere length: flow-FISH versus MM–qPCR Ferreira et al. percentile should not be used as a functional Table S1. Detected DKC-associated mutations in support for a possible disease-causing mutation. RTEL1, DKC1, TERC, and TERT. Even for lymphocyte (and granulocyte) flow-FISH, Table S2. Relevant statistical findings, flow-FISH careful interpretation is of importance, especially compared with MM–qPCR. in patients in whom TL ranges between the 5% and 10% percentiles. Table S3. Relevant statistical findings, flow-FISH Special attention is warranted in cases of clinical compared with various combinations of techniques suspicion of typical DKC or overt DKC symptoms and cutoffs, see manuscript. without detectable mutations, representing up to Figure S1. Comparison of MM–qPCR (T/S ratio) 30–40% of all pediatric/adolescent patients.8 Here, and lymphocyte flow-FISH (kb). Bland–Altman the assessment of TL is clearly recommended to plot for agreement analysis of flow-FISH (kb) and support the diagnosis of DKC, representing a func- MM–qPCR (T/S ratio) of the telomere lengths of tional disease-defining marker. DKC patients. Bias ± 2 SD = 0.13, LoA ranging In summary, we conclude that MM–qPCR is well from 0.05 to 0.22. suitable and satisfactory for TL analysis in research projects and retrospective analysis of large cohorts Figure S2. Patient detection rates of MM–qPCR of DNA samples. In clinical routine settings, flow- and granulocyte flow-FISH according to the 1st, FISH is superior over MM–qPCR and should be 5th, and 10th percentile cutoffs. (A) Sensitivity. (B) recommended for prospective screening of DKC Specificity. (C) False-positive values. (D) Positive patients based on TL assessment. predictive value. Dark gray bars represent the gran- ulocyte flow-FISH and light gray bars represent the Author contributions MM–qPCR. M.S.V.F. performed the MM–qPCR experiments, Figure S3. Patient detection rates of MM–qPCR and analyzed and interpreted the data, and wrote the flow-FISH according to the 1st, 5th, and 10th per- manuscript. M.K. performed parts of the NGS centile cutoffs. (A) Negative predictive value. (B) experiments, analyzed the NGS data, analyzed and False-negative value. Black bars represent the lym- interpreted the data, and revised the manuscript. phocyte flow-FISH, dark gray bars represent the I.H. collected the clinical data, performed flow- granulocyte flow-FISH, and light gray bars repre- FISH, assisted with NGS analysis, and interpreted sent the MM–qPCR. the data. N.E. performed parts of the MM–qPCR experiments and analyzed the data. B.X., M.V., L.Z., Competing interests and S.W. analyzed and interpreted the data. A.S.B. T.H.B. and F.B. receive scientific support from assisted with data collection and analyzed the data. RepeatDx, Vancouver, Canada. Lucia Vankann is M.B. and A.M. analyzed NGS results and inter- receiving honoraria (unrelated to this study) from preted the data. S.I. and J.P. interpreted the data RepeatDx. and revised the manuscript. T.H.B. analyzed and interpreted the data and revised the manuscript. F.B. References conceived and planned the study design, interpreted 1. Drummond, M.W. et al. 2007. Concise review: telomere the data, and wrote the manuscript. biology in normal and leukemic hematopoietic stem cells. Stem Cells 25: 1853–1861. Acknowledgments 2. Blasco, M.A. 2007. Telomere length, stem cells and aging. We thank Lucia Vankann for excellent technical Nat. Chem. Biol. 3: 640–649. 3. Martinez, P. & M.A. Blasco. 2011. Telomeric and extra- assistance. This work was partially supported by a telomeric roles for telomerase and the telomere-binding pro- START Grant (N° 691743, RWTH Aachen Univer- teins. Nat. Rev. Cancer 11: 161–176. sity) to M.S.V.F. and supported by a grant of the 4. de Lange, T. 2009. How telomeres solve the end-protection “Stiftung Lichterzellen.” problem. Science 326: 948–952. 5. Blasco, M.A. 2005. Telomeres and human disease: ageing, Supporting information cancer and beyond. Nat. Rev. Genet. 6: 611–622. 6. Young, N.S. 2010. Telomere biology and telomere diseases: Additional supporting information may be found in implications for practice and research. Hematology Am. Soc. the online version of this article. Hematol. Educ. Program 2010: 30–35. 102 Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences Ferreira et al. Telomere length: flow-FISH versus MM–qPCR 7. Calado, R. & N. Young. 2012. Telomeres in disease. F1000 patients with acute myeloid leukemia. Ann. Hematol. 96: Med. Rep. 4: 8. 1457–1461. 8. Calado, R.T. & N.S. Young. 2009. Telomere diseases. N. Engl. 28. Wenn, K. et al. 2015. Telomere length at diagnosis of chronic J. Med. 361: 2353–2365. phase chronic myeloid leukemia (CML-CP) identifies a sub- 9. Calado, R.T. & N.S. Young. 2008. Telomere maintenance and group with favourable prognostic parameters and molecular human bone marrow failure. Blood 111: 4446–4455. response according to the ELN criteria after 12 months of 10. Townsley, D.M., B. Dumitriu & N.S. Young. 2014. Bone mar- treatment with nilotinib. Leukemia 29: 2402–2404. row failure and the telomeropathies. Blood 124: 2775–2783. 29. Cawthon, R.M. 2009. Telomere length measurement by a 11. Dokal, I. 2011. Dyskeratosis congenita. Hematology Am. Soc. novel monochrome multiplex quantitative PCR method. Hematol. Educ. Program 2011: 480–486. Nucleic Acids Res. 37: e21. 12. Kirwan, M. & I. Dokal. 2008. Dyskeratosis congenita: a 30. Cawthon, R.M. 2002. Telomere measurement by quantita- genetic disorder of many faces. Clin. Genet. 73: 103–112. tive PCR. Nucleic Acids Res. 30: e47. 13. Ziegler, P. et al. 2012. Telomere elongation and clinical 31. Gutierrez-Rodrigues, F. et al. 2014. Direct comparison of response to androgen treatment in a patient with aplastic flow-FISH and qPCR as diagnostic tests for telomere length anemia and a heterozygous hTERT gene mutation. Ann. measurement in humans. PLoS One 9: e113747. Hematol. 91: 1115–1120. 32. Khincha, P.P. et al. 2017. Correlation of leukocyte telomere 14. Alter, B.P. 2017. Inherited bone marrow failure syndromes: length measurement methods in patients with dyskeratosis considerations pre- and posttransplant. Blood 130: 2257– congenita and in their unaffected relatives. Int. J. Mol. Sci. 2264. 18: 1765. 15. Townsley, D.M. et al. 2016. Danazol treatment for telomere 33. Behrens, Y.L. et al. 2017. Comparison of different methods diseases. N. Engl. J. Med. 374: 1922–1931. for telomere length measurement in whole blood and blood 16. Brummendorf, T.H. et al. 2001. Limited telomere shortening cell subsets: recommendations for telomere length measure- in hematopoietic stem cells after transplantation. Ann. N.Y. ment in hematological diseases. Genes Chromosomes Cancer Acad. Sci. 938: 1–7; discussion 7–8. 56: 700–708. 17. Brummendorf, T.H. et al. 2001. Telomere length dynamics in 34. Wand, T. et al. 2016. Telomere content measurement in normal individuals and in patients with hematopoietic stem human hematopoietic cells: comparative analysis of qPCR cell-associated disorders. Ann. N.Y. Acad. Sci. 938: 293–303; and flow-FISH techniques. Cytometry A 89: 914–921. discussion 303–294. 35. Gadalla, S.M. et al. 2016. The limitations of qPCR telomere 18. Alter, B.P. et al. 2015. Telomere length in inherited bone mar- length measurement in diagnosing dyskeratosis congenita. row failure syndromes. Haematologica 100: 49–54. Mol. Genet. Genomic Med. 4: 475–479. 19. Baerlocher, G.M. et al. 2006. Flow cytometry and FISH to 36. Lai, T.P., W.E. Wright & J.W. Shay. 2018. Comparison of measure the average length of telomeres (flow FISH). Nat. telomere length measurement methods. Philos. Trans. R. Soc. Protoc. 1: 2365–2376. Lond. B Biol. Sci. 373: 20160451. 20. Kirschner, M. et al. 2018. Recurrent somatic mutations 37. Montpetit, A.J. et al. 2014. Telomere length: a review of are rare in patients with cryptic dyskeratosis congenita. methods for measurement. Nurs. Res. 63: 289–299. Leukemia 32: 1762–1767. 38. Aubert, G., M. Hills & P.M. Lansdorp. 2012. Telomere length 21. Weidner, C.I. et al. 2016. DNA methylation in PRDM8 is measurement-caveats and a critical assessment of the avail- indicative for dyskeratosis congenita. Oncotarget 7: 10765– able technologies and tools. Mutat. Res. 730: 59–67. 10772. 39. Wang, F. et al. 2013. Robust measurement of telomere length 22. Beier, F. et al. 2005. Accelerated telomere shortening in in single cells. Proc. Natl. Acad. Sci. USA 110: E1906– glycosylphosphatidylinositol (GPI)-negative compared with E1912. GPI-positive granulocytes from patients with paroxysmal 40. Alter, B.P. et al. 2007. Very short telomere length by flow nocturnal hemoglobinuria (PNH) detected by proaerolysin fluorescence in situ hybridization identifies patients with flow-FISH. Blood 106: 531–533. dyskeratosis congenita. Blood 110: 1439–1447. 23. Schrader, A. et al. 2018. Actionable perturbations of dam- 41. Alter, B.P. et al. 2012. Telomere length is associated with dis- age responses by TCL1/ATM and epigenetic lesions form the ease severity and declines with age in dyskeratosis congenita. basis of T-PLL. Nat. Commun. 9: 697. Haematologica 97: 353–359. 24. Werner, B. et al. 2015. Reconstructing the in vivo dynamics 42. Alter, B.P. 2007. Diagnosis, genetics, and management of of hematopoietic stem cells from telomere length distribu- inherited bone marrow failure syndromes. Hematology Am. tions. eLife 4: e08687. Soc. Hematol. Educ. Program 2007: 29–39. 25. Beier, F. et al. 2015. Telomere dynamics in patients with del 43. Dokal, I. et al. 2015. Clinical utility gene card for: dysker- (5q) MDS before and under treatment with lenalidomide. atosis congenita—update 2015. Eur. J. Hum. Genet. 23: 558– Leuk. Res. 39: 1292–1298. 558. 26. Brummendorf, T.H. et al. 2001. Telomere length in leukocyte 44. Roy, S. et al. 2018. Standards and guidelines for validating subpopulations of patients with aplastic anemia. Blood 97: next-generation sequencing bioinformatics pipelines: a Joint 895–900. Recommendation of the Association for Molecular Pathol- 27. Ventura Ferreira, M.S. et al. 2017. Evidence for a pre-existing ogy and the College of American Pathologists. J. Mol. Diagn. telomere deficit in non-clonal hematopoietic stem cells in 20: 4–27. Ann. N.Y. Acad. Sci. 1466 (2020) 93–103 © 2019 The Authors. Annals of the New York Academy of Sciences 103 published by Wiley Periodicals, Inc. on behalf of New York Academy of Sciences