Abstract
Introduction
Somatic mutations in Hematopoietic Stem Cells and Progenitors (HSC/HCP) are very frequent in patients with acquired Immune-mediated Aplastic Anemia (IAA) being found in about a third of patients at diagnosis and up to 65-75% after Immuno-Suppressive Treatment (IST). In our systematic study investigating 170 patients enrolled in the EBMT RACE randomized, prospective clinical trial (NCT02099747) the CUB and Sushi Multiple Domains 1 (CSMD1) gene for the first time emerged as one of the most frequently mutated genes. Here we present a deep analysis of CSMD1 mutations in IAA patients enrolled in the RACE trial, looking for the frequency of mutations, its association with clinical variables and other somatic mutations, the dynamics of the mutated clones across the 2-year study period and a in silico structural analysis predicting functional consequences.
Methods
The RACE trial enrolled 197 IAA patients randomized to receive standard IST (anti-thymocyte globulin [ATG] and cyclosporine A) with or without eltrombopag. Bone marrow aspirate samples were collected at baseline, 6 months, and 24 months and analyzed at King's College London using custom targeted gene panels incorporating unique molecular identifiers. The HaloHS 31 gene panel was performed on 170, 150 and 103 patient samples collected at baseline, 6 and 24 months post randomization (TP6 and TP24), respectively. Appropriate variant caller protocols were utilized to identify candidate variants. An arbitrary Variant Allele Frequency (VAF) threshold of 0.2% was exploited to define mutations.
Results
At baseline (n=170), no CSMD1 mutation was found in any IAA patient. At later time points, mutations in the CSMD1 gene were found in 4/150 (2.7%) patients at TP6 and in 16/103 (15.5%) patients at TP24. Many patients harbored multiple mutations, with a total of 11 mutations detected in 4 patients at TP6, and 48 mutations in 16 patients at TP24. This resulted in a global mean number of CSMD1 mutations per patients increasing from 0 at baseline, to 0.07 and 0.47 at TP6 and TP24, respectively. The proportion of patients carrying CSMD1 mutations and the mean number of CSMD1 mutations per patients were not associated with age (≤40 vs >40 years), disease severity at baseline (severe vs very severe), presence of somatic mutations in myeloid-cancer genes at baseline, treatment arm (eltrombopag yes vs no) and hematological response.
The VAF of CSMD1-mutated clones was usually very low, with 9/11 (81.8%) and 42/48 (87.5%) clones remaining below a 5% VAF at TP6 and TP24 (VAF range was 0.9-34.4% at TP6 and 0.8-48.5% at TP24), with unpredictable behavior. However, the general rule is that CSMD1-mutated clones become detectable only after IST, with the highest frequency at TP24, with very low VAF and often being multiple in the same patient.
Patients with CSMD1 mutated clone often showed mutations in other myeloid cancer genes, especially ETV6, PHF6, PPM1D, PTPN11, RAD21 and STAG2.
The CSMD1 protein consists of 14 CUB domains alternated with 14 Sushi domains, followed by further 15 Sushi domains, a transmembrane domain and an intracytoplasmic tail. The Sushi domains are known as Regulator of Complement Activation (RCA) which preferentially disable the classical complement pathway (CCP). Looking structurally at the specific mutations, no hotspot mutation was found; at TP6 frameshift and missense mutations were 8/11 (72.7%) and 3/11 (27.3), whereas at TP24 they were 42/48 (87.5%) and 6/48 (12.5%), respectively. These mutations result in major structural changes of CSMD1, which in most cases are truncated proteins lacking the transmembrane domain.
ConclusionsCSMD1 mutations are commonly found after IST in IAA patients. We hypothesize that the secretion of truncated CSMD1 proteins retaining RCA domains may result in a paracrine effect which may spare mutated HSC/HCP from some CCP-mediated damage. The observation that all these clones emerge after IST makes ATG as the most likely trigger for such CCP activation via immune complexes. Although CSMD1 is a known tumor suppressor gene, in the context of IAA clones carrying specific CSMD1 mutation might have a transient survival advantage in the context of hematopoietic toxicity during ATG treatment. Thus, CSMD1-mutated HSC/HCP seem to be an example of pruning selection, where some intrinsic advantage transiently emerging in specific micro-environmental setting may eventually become evident at the time of hematological recovery.
