Severe aplastic anemia (SAA) can be treated with either immunosuppressive therapy (IST) or hematopoietic stem cell transplantation (HSCT). Response to IST depends on control of the immune response and on a hematopoietic stem cell (HSC) compartment capable of repopulating the bone marrow after therapy. Telomeres are protective repetitive DNA sequences at the ends of chromosomes; telomeres shorten with each cell division due to DNA polymerase deficiency to fully duplicate telomeric ends. Cells with critically shortened telomeres undergo proliferative senescence, apoptosis, and genomic instability. Excessive shortening of telomeres has been proposed as a potential biomarker for HSC exhaustion, and loss-of-function mutations in telomerase genes cause excessive telomere erosion and associate with marrow failure. To test the hypothesis that telomere length is a predictor of successful outcome after IST in SAA, telomere length of peripheral blood leukocytes were measured by quantitative PCR and expressed as the relative ratio of telomere repeat copy number to single gene copy number (T/S ratio) in 168 consecutive patients prior to IST therapy. The cohort consisted of patients with SAA (mean age, 34 years; range, 4–82) enrolled in three sequential protocols at the National Institutes of Health from 2003 to 2008 (ClinicalTrials.gov identifiers, NCT00001964, NCT00260689, and NCT00061360); all patients received IST based on horse anti-thymoglobulin (h-ATG) plus cyclosporine (CsA). Additional treatments depended on specific protocol arm: mycophenolate mofetil (32 patients), rabbit-ATG in place of h-ATG (31 patients), or rapamycin (35 patients); 70 patients received h-ATG and CsA only. None of the patients had clinical findings suggestive of dyskeratosis congenita. All adult patients or legal guardians signed informed consent according to NHLBI Institutional Review Board. Response, relapse, and clonal evolution rates were similar across all regimens. Using both a univariate and multivariate Cox regression model, telomere length was not associated with response (partial or complete) to IST at 6 months. However, for patients who initially responded to therapy, telomere length as a continuous variable inversely correlated with relapse rate (P=0.01). When telomere length was treated as a categorical variable, patients with shorter telomeres (below the 50th percentile of telomere distribution) had 2.5 times higher probability to relapse in five years then did patients with longer telomeres (P=0.002). Since telomere length physiologically shortens with aging, patients were stratified as being either greater than or equal/less than fifty years old to address whether shortened telomeres were a surrogate marker of age. In patients over 50 only (n=54), short telomere length but not age predicted relapse (P =0.04) in a multivariate model including telomere length, age, and pre-treatment blood counts, indicating that telomere length associated with relapse independently of age. Approximately 10% of patients undergoing IST eventually develop secondary clonal evolution. In this cohort 11/168 patients developed cytogenetic abnormalities, myelodysplastic syndrome, or leukemia in 5 years, and telomere length inversely correlated with risk for clonal evolution (P=0.01). These results indicate that leukocytes’ telomere length predicts sustained response to IST, possibly serving as a biological marker for HSC reserve. While initial response to treatment results from reducing cytotoxic T-cells, patients with short telomeres may quickly reach a critical length threshold where further stem cell division is not possible. Patients in this cohort with the shortest telomeres also were at risk for developing clonal evolution.

Disclosures: Cooper:NIH-Pfizer: Research Funding, This research year was made possible through the Clinical Research Training Program, a public-private partnership supported jointly by the NIH and Pfizer Inc (via a grant to the Foundation for NIH from Pfizer Inc).

Our findings support the theory that short telomeres promote genomic instability in human cells and human disease.

Author notes

Corresponding author

Sign in via your Institution