SMC3 encodes a subunit of cohesin associated with recurrent heterozygous mutations in acute myeloid leukemia (AML) and other myeloid malignancies. About one-third of these mutations are nonsense and splice-site mutations and the rest are missense. To understand whether these missense mutations might have dominant-negative effects or phenocopy loss-of-function effects, we compared the consequences of Smc3 deficient and haploinsufficient mouse models.

We found that homozygous deletion of Smc3 during embryogenesis (Smc3fl/fl/Vav1-Cre+/-) is embryonic lethal. Grossly, the E13.5 Smc3fl/fl/Vav1-Cre+/- embryos were indistinguishable in size and appearance from other genotypes, except the lack of obvious fetal livers and near complete absence of CD45+Gr1+CD11b+ fetal liver cells, consistent with myeloid biased hematopoietic failure.

Likewise, somatic deletion of homozygous Smc3 alleles in adult mice led to complete hematopoietic failure. In Smc3fl/fl/ERT2-Cre+/- mice, tamoxifen-induced Smc3 deletion led to rapid bone marrow (BM) failure and death. In competitive transplantation, the Smc3fl/fl/ERT2-Cre+/- BM cells were quickly outcompeted after tamoxifen treatment, indicating complete loss of hematopoietic stem/progenitor cell (HSPC) functions. Again, the effect was most pronounced within the myeloid compartment. Thus, Smc3 is necessary for both embryonic and adult hematopoiesis.

We examined the consequences of Smc3 haploinsufficiency using three mouse models: Smc3fl/+/CMV-Cre+/-, Smc3fl/+/Vav1-Cre+/-, and Smc3fl/+/ERT2-Cre+/- . We found that Smc3 haploinsufficiency led to a normal number of female pups in X-linked CMV-Cre intercrosses, and the female pups had no obvious defects in growth or blood counts, implying tolerance of embryonic heterozygous deletion of Smc3 . In ex vivo methylcellulose assays, Smc3 haploinsufficiency did not result in increased number of colonies, and the Smc3 haploinsufficient BM cells did not replate beyond two weeks. In competitive transplantations, we observed a significant competitive disadvantage in the Smc3fl/+/ERT2-Cre+/- BM cells, most pronounced in the Gr1+ myeloid cells.

The patterns of missense and nonsense mutations observed in AML patients suggested that SMC3 likely acts as a tumor suppressor, andthat haploinsufficiency would augment pre-leukemic self-renewal. However, we observed a competitive disadvantage across multiple experiments. This contradicts the competitive advantage observed when Smc3 haploinsufficiency was induced using Mx1-Cre (Viny et al . JEM 212 (11): 1819-1832). Therefore, we examined whether Smc3 haploinsufficiency might augment self-renewal under specific conditions of hematopoietic stress, including pIpC. We again observed a marked competitive disadvantage in the Smc3fl/+/ERT2-Cre+/- BM cells following tamoxifen induction, although the myeloid competitive disadvantage was partially ameliorated in the pIpC-treated group, whereas 5-FU treatment augmented the myeloid competitive disadvantage.

In AML patients, SMC3 mutations frequently co-occurred with DNMT3A mutations. We, therefore, asked whether Smc3 haploinsufficiency might lead to a competitive advantage if it occurred in the background of Dnmt3a haploinsufficiency. With the addition of Dnmt3a haploinsufficiency, the severe myeloid competitive disadvantage was partially abrogated, but the significant competitive disadvantage in other lineages remained intact; even with constitutive Dnmt3a haploinsufficiency, Smc3 haploinsufficiency did not result in competitive growth of BM cells.

In summary, our data demonstrate that homozygous Smc3 deletion was incompatible with embryonic or adult hematopoiesis, with the greatest defects noted within myeloid compartment, indicating that AML-associated SMC3 mutations are unlikely to have dominant-negative effects. Smc3 haploinsufficiency resulted in a myeloid competitive disadvantage, which was partially ameliorated by pIpC treatment, and accelerated by the stem cell stress of 5-FU. In the presence of concurrent Dnmt3a haploinsufficiency, the myeloid disadvantage was partially abrogated, while the disadvantage in other lineages remained unperturbed, suggesting interactions between the mutations.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution