Two Novel Knock-in and Knock-out of Setd2 Alleles Cooperate with Mll-Af9 Knock-in Allele to Accelerate Leukemia Development

SETD2 protein is a histone methyltransferase that is specific to lysine-36 of histone H3. SETD2 also contains a novel transcriptional activation domain (Set2 Rpb1 interacting domain, SRI domain), which has been found to be associated with hyperphosphorylated RNA polymerase II. Functioning as a tumor suppressor gene, SETD2 is enriched with missense, nonsense and frameshift mutations in many types of cancers. Previously we had identified 6% SETD2 mutations in acute myeloid leukemia (AML) and acute lymphoid leukemia (ALL), with 22% enriched in MLL leukemia. Recent studies also reported that SETD2 mutations were enriched at relapse in pediatric B-cell ALL, suggesting that SETD2 mutations may contribute to relapse and chemotherapy resistance.

There are equal amounts of missense mutations and nonsense/frame-shift mutations of SETD2 observed in acute leukemia patients. To understand the function of SETD2 mutations in leukemia development and chemotherapy resistance, we generated two Setd2 mutant alleles using CRISPR-Cas9 method: 1) Setd2^F2478L/WT allele, which is equivalent to the SETD2-F2505L mutation found in an AML patient. Setd2-F2478L mutation is located in the SRI domain, which completely loses the interaction with the C-terminal domain (CTD) of RNA pol II. 2) Setd2-exon6^KO/WT allele, resulting in a frame shift and nonsense mediated decay of Setd2 mRNA. Setd2 expression in c-Kit⁺ BM cells of Setd2^F2478L/WT mice is similar to WT mice at both mRNA and protein levels. However, Setd2 expression in c-Kit⁺ BM cells of Setd2-exon6^KO/WT decreased 50% at both mRNA and protein levels, which indicated the nonsense mediate decay of the exon6 KO allele. Both Setd2^F2478L/WT and Setd2-exon6^KO/WT mice have shown a similar decrease in the H3K36me3 mark and other related epigenetic changes in purified c-Kit⁺ BM cells. The H3K36me3 ratio of WT versus Setd2^F2478L/WT versus Setd2-exon6^KO/WT is 1: (0.415 ± 0.21): (0.432 ± 0.18), *P < 0.01.

These two mouse models showed a series of similar phenotypes and pathological changes. Firstly, homozygous Setd2 mutation showed early embryonic lethality, which is consistent with previous reports that homozygous deletion of exon4 and exon5 of Setd2 resulted in embryonic lethality at E10.5-E11.5. Setd2^F2478L/WT mice were bred with Setd2^F2478L/WT mice,5 WT, 31 Setd2^F2478L/WT, and 0 Setd2^{F2478L/ F2478L} pups were found from a total 6 litters (n=36). Similarly, Setd2-exon6^KO/WT mice were bred with Setd2-exon6^KO/WT mice, 11 WT, 17 Setd2-exon6^KO/WT and 0 Setd2-exon6^KO/KO were found from a total 4 litters (n=28). This data indicates that homozygous Setd2 mutation is early embryonic lethal. At 10 weeks, the body weights of Setd2^F2478L/WT and Setd2-exon6^KO/WT were 15.28% ± 2.31% (*P < 0.05) and 12.29% ± 3.18% (*P < 0.05) lower than the WT littermates, and remained low during their normal life span. Secondly, Setd2 mutations changed HSPCs profile, increased numbers of early progenitors in in vivo CFU re-plating assays of BM cells. The 2nd re-plating CFU colonies derived from Setd2^F2478L/WT bone marrow increased by 41% ± 6.34% compared to the WT bone marrow, while those CFU colonies derived from Setd2-exon6^KO/WT increased by 36% ± 5.12%. Consistent with these findings, the mutant mice increased CMP population compared to the WT littermates. However, the mutant cells stopped growing in the 3rd re-plating, which indicated that the Setd2 mutant cells are not fully transformed. Thirdly, we observed cooperative function of SETD2 mutation in the development of myeloid leukemia (eg. MLL translocation). Setd2^F2478L/WT was bred with MLL-Af9 knock-in mice, and colony-forming re-plating assays demonstrated cooperative effects of Setd2^F2478L/WT with MA9. MLL-Af9/Setd2^F2478L/WT mice resulted in a significantly higher (90.99% ± 10.27%) yield of total colonies and growth advantage in the multiple rounds of plating comparing to MLL-Af9 cells. MA9 knock-in mice cohort died of AML at 181days, but MLL-Af9/Setd2^F2478L/WT mice died at 128 days (*P < 0.01).

In conclusion, our two novel LOF Setd2 mutation alleles (Setd2^F2478L and Setd2-exon6^KO) showed similar epigenetic, cellular, and growth retardation phenotypes, and these Setd2 mutation alleles cooperate with MLL-AF9 to accelerate leukemia development. Thus, these novel Setd2 mutation alleles will be useful to help us understand transcriptional, epigenetic, and tumor suppression roles of SETD2 alone and in concert with other oncoproteins.