Acute Myeloid Leukemia (AML) is a disease
characterized by the aberrant expansion of white blood cells leading to death
from infection, hemorrhage or organ infiltration. AML is genetically
heterogeneous, however, activating mutations in the RasGTPases (typically K- or
N-Ras) are found in 15-44% of AML. Likewise, the sustained activation or over-expression
of signal transducer and activator of transcription 3 (STAT3) is a common event
in AML (up to 50% of cases) and correlates with poor therapeutic responses. The
current dogma suggests that STAT3 is constitutively tyrosine phosphorylated
leading to sustained transcription of STAT3 target genes which provide a proliferative
and survival advantage to transformed cells. This paradigm is important, but
needs to be amended to address the recent discovery of non-canonical,
mitochondrial activity of STAT3. We found that STAT3 is phosphorylated on S727
which is necessary for mitochondrial activity, but not on Y705 which is
required for transcriptional activity in a potent mouse model of K-Ras driven myeloproliferative
disease (MPD). Moreover, mutation of S727 in STAT3 decreases KRasG12D driven
MPD onset, severity and significantly extends survival in vivo. To differentiate between the nuclear and mitochondrial
activities of STAT3 we engineered KRasG12D, STAT3-/- bone marrow
progenitor cells to express wild-type STAT3, STAT3Y705F or a mitochondrially-restricted
STAT3 mutant. The expression of each of these STAT3 variants were similarly
capable of rescuing growth factor independent myeloid colony formation in
colony formation unit assays. Together these data confirm that the mitochondrial
pool of STAT3 is important for K-RasG12D driven AML in vivo.