Smc3 Haploinsufficiency and Smc3 Deletion Alter Hematopoiesis In Vivo

Recurrent mutations in SMC3, encoding a cohesin subunit, have been identified in acute myeloid leukemia (AML) and other myeloid malignancies by our group and others. SMC3 mutations are heterozygous in AML patients. Missense, nonsense, and splice site mutations have been observed across all domains of SMC3. Given the breath of mutations, it is important to determine whether these represent recurrent loss-of-function mechanisms, or if some might have dominant negative effects. To determine the impact of Smc3 deletion on hematopoiesis, we studied both Smc3 haploinsufficient and Smc3 deficient mice as models of loss-of function and dominant negative phenotypes respectively.

The Smc3 haploinsufficient mouse model has a lacZneo gene trap inserted in intron 3-4 of Smc3, which leads to a premature transcription stop and therefore produces a truncated and dysfunctional protein. The homozygous Smc3^trap allele is embryonic lethal. The Smc3^trap/+mice have an early growth defect, although their body weight catches up to wild type mice after 6 weeks of age. We found no difference in spleen weights, peripheral blood counts, and bone marrow (BM) compositions between Smc3^trap/+ and wild type mice. The Smc3^trap/+ BM cells formed similar number of colonies as wild type cells when plated in methylcellulose in vitro and lost self-renewal capabilities after replating for two weeks. Competitive repopulation assay in vivo showed neither advantage nor disadvantage for the Smc3^trap/+BM cells (n=10). Thus, Smc3^trap/+BM cells have normal colony forming capacity in vitro and normal homeostatic feedback in vivo.

Further, we generated Smc3 conditionally deficient mice by removing the gene-trap cassette, which retains the loxP sites flanking exon 4 (Smc3^fl), and crossing these mice with either Vav1-Cre^+/- or ERT2-Cre^+/- to delete the allele (Smc3^fl/+/Vav1-Cre^+/- is constitutively haploinsufficient in hematopoietic cells, whereas Smc3^fl/+/ERT2-Cre^+/-is only haploinsufficient when induced with tamoxifen). We characterized both models by serial replating assays, flow cytometry assays for hematopoietic stem/progenitor cells (HSPCs), and BM lineage in vitro and found no difference in these mice compared to the Smc3^fl/+control. In contrast to the Smc3^fl/+/Mx1-Cre^+/- mice (Viny et al. JEM 212 (11): 1819-1832), we observed a significant competitive disadvantage for the Smc3^fl/+/ERT2-Cre^+/-BM cells (p<0.0001, n=10), most pronounced in Gr1+ myeloid cells in vivo (p<0.0001), implying Smc3 haploinsufficiency alters hematopoiesis in those mice in vivo.

We characterized the effects of homozygous Smc3 loss on hematopoiesis in the inducible Smc3^fl/fl/ERT2-Cre^+/- mice by treating mice with tamoxifen at 6 weeks of age (Smc3^fl/fl/Vav1-Cre^+/- is embryonic lethal). Deletion of Smc3 led to rapid bone marrow failure and 100% lethality with a median survival of 8 days (n=4, 2 independent experiments). At the time of death, we observed severe reduction in the sizes of spleen (Sp) and thymus (Thy), in total number of BM, Sp, and Thy cells, and in white blood counts, lymphocytes, monocytes, and platelets. The Smc3 deficient BM cells had decreased levels of Smc3 by Western blot. The impact of Smc3 deletion on HSPC functions in vivo was assessed by a competitive repopulation assay of Smc3^fl/fl/ERT2-Cre^+/-BM cells (p<0.0001, n=10). Recipient mice were treated with tamoxifen after 6-week engraftment. After tamoxifen-mediated deletion, Smc3 deficient cells were rapidly outcompeted in vivo, indicating complete loss of HSPC functions.

Collectively, these results suggest that Smc3 is necessary for normal hematopoiesis and for HSPC functions. The AML-associated SMC3 mutations are therefore unlikely to be dominant negative because the complete loss of Smc3 is incompatible with hematopoiesis.