Expression of Livin, an antiapoptotic protein, has been associated with a poor outcome in patients with several different solid tumors. By contrast, Choi and colleagues in this issue of Blood show that its expression has independent favorable prognostic significance in childhood acute lymphoblastic leukemia.
Livin (aka, ML-IAP or KIAP) is a member of the inhibitor of apoptosis protein (IAP) family, and has 2 splice variants (α- and β-isoforms).1 It antagonizes both the extrinsic death receptor and intrinsic mitochondria-based apoptotic pathways by inhibiting caspases 3, 7, and 9, and by degrading the pivotal apoptotic regulator Smac/DIABLO through the ubiquitin-proteasome pathway.2 Not surprisingly, Livin expression has been associated with a poor outcome in patients with neuroblastoma, melanoma, or superficial bladder cancer. The majority of tissue samples from patients with non–small-cell lung cancer also express Livin; silencing of its expression by RNA interference sensitizes the cancer cells in these samples to various proapoptotic stimuli.3
In this issue of Blood, Choi and colleagues studied Livin expression in the leukemic cells of 222 children with newly diagnosed acute lymphoblastic leukemia (ALL). They report associations of this protein with several recognized favorable features of ALL patients: female sex, age 1 to 9 years, leukocyte count less than 50 × 109/L, and the presence of t(12;21). Livin expression was found in only 1 of 32 cases with unfavorable chromosomal abnormalities (t(9;22) or 11q23 rearrangement). Surprisingly, it was also rare in hyperdiploid (≥ 50) ALL (1 of 26 cases), a genetic subtype generally associated with a favorable prognosis. Finally, it was associated with an increased apoptotic response to methylprednisolone treatment ex vivo, a rapid early treatment response in vivo, and more importantly, a very favorable treatment outcome (5-year relapse-free survival rate of 97.9% ± 4.0%, compared with 64.9% ± 11.8% for patients without this feature). Livin expression retained independent prognostic significance in a multivariate analysis.
How does one account for the seemingly paradoxical prognostic impact of Livin expression in childhood ALL? The type of treatment administered can be discounted, as the patients were not assigned to risk groups based on Livin expression, even though they were managed on 6 different clinical protocols (29 patients underwent hematopoietic stem cell transplantation). The authors provide one possible explanation: the cleaved form of Livin may serve as a proapoptotic regulator.4 A recent study showed that silencing of the Livin β-isoform, but not the α-isoform, by RNA interference blocked the growth of HeLa cells in clonogenic survival assays, and sensitized the cells to various proapoptotic stimuli.1 A reasonable corollary of this idea is that leukemic lymphoblasts preferentially express the α-isoform of Livin, rendering them more susceptible to apoptosis. Although the authors did not analyze their cases on the basis of Livin isoform expression, one case studied with a cytotoxicity assay expressed more of the α-isoform than the β-isoform.
Why should hyperdiploid (≥ 50) ALL cases have a very low rate of Livin expression in contrast to much higher rates found in other prognostically favorable groups? Although such cases had a relatively poor treatment outcome in this study (5-year relapse-free survival of 65.4% ± 29%), a finding consistent with low Livin expression, this cytogenetic feature generally confers a favorable prognosis (5-year event-free survival rate of ∼ 90% in many reported series).5 Moreover, the low frequency of hyperdiploidy (≥ 50) in the series of Choi et al (12.8%) contrasts sharply with the 25% rate in other series.5 Thus, either the cohort of patients on which this analysis was based does not reflect the general ALL population, or the cytogenetic studies failed to detect some hyperdiploid (≥ 50) cases, a common result with this genetic subtype owing to the propensity of the leukemic cells to undergo spontaneous apoptosis in vitro.
Should Livin expression be used as a prognostic factor for treatment stratification of ALL patients, as suggested by the authors? The plethora of highly predictive risk factors in childhood ALL argues against this application.5 However, if Livin expression is independently confirmed as an exceptionally robust favorable prognostic indicator, it might be useful as a means to identify potentially curable patients within otherwise high-risk groups—with one caveat. Any reduction of treatment based on Livin expression should be done judiciously, as therapy exerts independent prognostic strength in childhood ALL and its inappropriate reduction could nullify the favorable influence of Livin and other factors on clinical outcome.
Conflict-of-interest disclosure: The author declares no competing financial interests. ▪
