Mutational analysis of bone marrow mesenchymal stromal cells in myeloid malignancies

Microenvironment is actively involved in the pathophysiology of hematopoietic malignancies. Mesenchymal stromal cells (MSC) support hematopoiesis through the production and secretion of cytokines, cell–cell interactions and immunomodulating properties. These cells are defined, plastic-adherent, growing cells, are known to express CD105, CD73, and CD90, lack common hematopoietic antigens, and differentiate into osteoblasts, adipocytes, and chondroblasts in vitro [1]. Anomalies of multiple MSC features have been described in myelodysplastic syndromes (MDS) and acute myeloid leukemia (AML). These include significantly reduced growth, proliferative and differentiating capacity, premature replicative senescence, abnormal expression of surface molecules and chemokines, and reduced ability to support hematopoietic stem and progenitor cell growth in long-term culture assays [2]. The molecular basis of MSC dysfunction in MDS and AML is still under investigation. Previous studies have reported the occurrence of non-clonal chromosomal aberrations in bone marrow MSC isolated from patients with MDS and AML, which only very rarely correspond to the cytogenetic markers observed in the hematopoietic leukemic clone of the same individual [3]. Somatic mutations of multiple genes have been described in myeloid malignancies and often concur to identify distinct leukemia subtypes or prognostic subgroups. Recently, deep sequencing approaches have identified new recurrent mutations of genes involved in epigenetic and spliceosome machineries in AML and MDS samples 4, 5 and 6. We investigated the frequency of recurrent mutations of epigenetic and spliceosomal genes, and of FLT3 and NPM1 genes, in matched bone marrow hematopoietic cells and MSC isolated from 41 patients with myeloid malignancies. The study population included 9 patients with de novo AML, 9 with MDS, 7 with chronic myeloproliferative neoplasms (MPN), 3 with secondary AML (sAML, 2 evolved from MDS, 1 from MPN), and 13 with therapy-related myeloid neoplasms (7 t-AML and 6 t-MDS). Bone marrow mononuclear cells (BM-MNC) were isolated at the time of diagnosis by Ficoll gradient centrifugation. MSC were expanded using Mesencult medium (Stem Cell Technologies, Voden Medical Instruments, Milan, Italy) in plastic-adherent cultures up to the second passage. Flow cytometry analysis confirmed the standard MSC phenotype (CD45 negative, CD73 positive, CD90 positive, CD105 positive) in more than 99% of the MSC population. DNA was extracted from BM-MNC and MSC using the QIAamp DNA Mini Kit (Qiagen, Milan, Italy). The following hot-spot mutations were studied on genomic DNA by Sanger sequencing (ABI PRISM 3100; Applied Biosystems/Life Technologies, Milan, Italy): IDH1 R132, IDH2 R140 and R172, DNMT3A R882, U2AF1 S34 and R35, SF3B1 exons 13–14 and 15–16, and SRSF2 exon 1 [7]. In addition, FLT3 and NPM1 mutations were analyzed in patients with de novo or therapy-related AML. FLT3 internal tandem duplication and tyrosine kinase domain mutations were studied by reverse transcription polymerase chain reaction (RT-PCR) and restriction fragment length polymorphism RT-PCR, respectively, whereas NPM1 exon 12 mutations were detected by RT-PCR with high-resolution melting curve analysis, followed by Sanger sequencing of positive cases. In BM-MNC, FLT3 internal tandem duplication and FLT3 tyrosine kinase domain mutations were found in two patients (one patient with de novo AML and one patient with a t-AML), whereas NPM1 exon 12 mutations were present in six AML patients (4 de novo, 3 COSM158604 and 1 COSM1319219; 1 sAML, COSM158604; 1 t-AML, COSM28937). No FLT3 and NPM1 mutations were found in MSC compartment in all studied patients. Mutations of IDH1 R132 were found in BM-MNC isolated from one patient with de novo AML (R132C) and in two patients with t-AML (one R132H and one R132L). One t-AML patient had an IDH2 R140Q mutation in BM-MNC, whereas no mutations were detected at the codon R172. Neither IDH1 nor IDH2 mutations occurred in matched MSC. Mutations of DNMT3A R882 were found in BM-MNC from one patient with de novo AML (R882H), one with t-AML (R882H) and one with sAML evolved from polycythemia vera (R882C), but not in corresponding MSC. One U2AF1 S34Y mutation was found in BM-MNC isolated from one patient with intermediate-risk MDS, and a SF3B1 K666N mutation was detected in BM-MNC from one patient with AML following a primary MDS. SRFS2 P95 mutations were found in BM-MNC isolated from two patients with de novo AML (P95L and P95R), three patients with MDS (P95L, P95H and P95 frameshift mutation), and one patient with MPN (P95H). Again, U2AF1, SF3B1, and SRFS2 mutations were absent in MSC. All mutations described were heterozygous. Results of mutational analysis in hematopoietic BM-MNC are reported in Figure 1. Figure 1. Distribution of IDH1, IDH2, DNMT3A, U2AF1, SF3B1, SRSF2, NPM1 and FLT3 mutations in the hematopoietic compartment of the study population. No mutations were found in the mesenchymal stem cell compartment. Each column represents a single patient. Black boxes represent mutations. FLT3 and NPM1 mutational analysis was not performed in patients with MDS, MPN and t-MDS (white boxes). AML = acute myeloid leukemia, MDS = myelodysplastic syndromes, MPN = myeloproliferative neoplasms, sAML = secondary AML, t-AML = therapy-related AML, t-MDS = therapy-related MDS. Figure optionsDownload full-size imageDownload high-quality image (563 K)Download as PowerPoint slide We found no mutations for any of the studied genes in the MSC compartment, both in carriers of mutations in the hematopoietic compartment and in wild-type patients. The absence of NPM1 and FLT3 mutations in MSC from AML patients has been previously reported [3]. So far, the prevalence of somatic mutations of epigenetic and spliceosomal genes in the MSC compartment in patients with myeloid malignancies is not known. Mutations of DNMT3A and IDH1/2 have been described in 22% and 16% of AML patients and in 8% and 12% of MDS patients, respectively 4, 8 and 9. These mutations have been associated with deregulated DNA methylation and poor prognosis. Abnormal DNA methylation has been advocated as responsible for the premature senescence and reduced growth of MSC in MDS. Specific patterns of DNA methylation in MSC have been reported in MDS subtypes compared with healthy donors [2]. According to our results, abnormal methylation profiles described in MSC are unlikely to be related to mutations of epigenetic regulatory genes, even though a wider mutational analysis, including also ASXL1 and TET2 genes, could be more informative. Abnormal splicing patterns of multiple genes have been described in myeloid neoplasms because of mutations of spliceosome machinery genes [10]. This phenomenon is considered to play a significant role in myeloid leukemogenesis because of selective missplicing of tumor-associated genes. The contribution of missplicing to MSC dysfunction in myeloid neoplasms is still a matter of investigation. The absence of recurrent spliceosome gene mutations in MSC contrasts with the hypothesis that these mutations may play a significant role. We conclude that common mutations of genes involved in epigenetic regulation and spliceosome machinery are absent in the mesenchymal compartment of leukemic bone marrow and are restricted only to the malignant myeloid clone, in agreement with the distinct origin in hematopoietic cells.