External validation of models for KIR2DS1/KIR3DL1-informed selection of hematopoietic cell donors fails
External validation of models for KIR2DS1/KIR3DL1-informed selection of hematopoietic cell donors fails
复制标题
DOI：
10.1182/blood.2019002887
复制
发表时间：
2020-04-16
期刊：
影响因子：
20.3
通讯作者：
Bomhaeuser, Martin
中科院分区：
文献类型：
--
作者：
Schetelig, Johannes;Baldauf, Henning;Bomhaeuser, Martin
几项研究表明，利用杀伤细胞免疫球蛋白样受体（KIR）介导的自然杀伤（NK）细胞反应性可以降低异基因造血细胞移植后复发的风险。基于一个有前景的模型，关于KIR2DS1和KIR3DL1及其同源配体的信息可用于将供体分类为KIR有利或KIR不利。本研究的目的是在非亲缘供者造血细胞移植中对该模型进行外部验证。在校正患者年龄、改良疾病风险指数、Karnofsky体力状态、供体年龄、HLA匹配、性别匹配、巨细胞病毒匹配、预处理强度、T细胞耗竭类型和移植物类型的考克斯回归模型中，检验预测因子对总生存期（OS）和复发率的影响。分析了2222例急性髓系白血病或骨髓增生异常综合征患者的资料。通过使用基于高分辨率扩增子的下一代测序对KIR基因进行分型。在单变量分析和亚组分析中，KIR有利供体患者的OS和累积复发率与KIR不利供体患者相当。多变量考克斯回归模型校正后的OS风险比为0.99（Wald检验，P = 0.93），复发率风险比为1.04（Wald检验，P = 0.78）。我们还分别测试了激活供体KIR2DS1和KIR3DL1抑制的影响，但未发现对OS和复发风险的显著影响。因此，我们的研究表明，所提出的模型不能普遍预测NK介导的疾病控制。深入了解NK介导的同种异体反应性对于预测其对移植物抗白血病反应的贡献以及最终使用KIR基因型信息进行供体选择是必要的。
Several studies suggest that harnessing natural killer (NK) cell reactivity mediated through killer cell immunoglobulin-like receptors (KIRs) could reduce the risk of relapse after allogeneic hematopoietic cell transplantation. Based on one promising model, information on KIR2DS1 and KIR3DL1 and their cognate ligands can be used to classify donors as KIR-advantageous or KIR-disadvantageous. This study was aimed at externally validating this model in unrelated donor hematopoietic cell transplantation. The impact of the predictor on overall survival (OS) and relapse incidence was tested in a Cox regression model adjusted for patient age, a modified disease risk index, Karnofsky performance status, donor age, HLA match, sex match, cytomegalovirus match, conditioning intensity, type of T-cell depletion, and graft type. Data from 2222 patients with acute myeloid leukemia or myelodysplastic syndrome were analyzed. KIR genes were typed by using high-resolution amplicon-based next-generation sequencing. In univariable analyses and subgroup analyses, OS and the cumulative incidence of relapse of patients with a KIR-advantageous donor were comparable to patients with a KIR-disadvantageous donor. The adjusted hazard ratio from the multivariable Cox regression model was 0.99 (Wald test, P = .93) for OS and 1.04 (Wald test, P = .78) for relapse incidence. We also tested the impact of activating donor KIR2DS1 and inhibition by KIR3DL1 separately but found no significant impact on OS and the risk of relapse. Thus, our study shows that the proposed model does not universally predict NK-mediated disease control. Deeper knowledge of NK-mediated alloreactivity is necessary to predict its contribution to graft-versus-leukemia reactions and to eventually use KIR genotype information for donor selection.