Gfi136N As a Novel Marker and Therapeutic Target of MDS and AML

Growth Factor Independence 1 (GFI1) is a hematopoietic transcription factor that plays a crucial role in the development of myeloid precursor cells. Its variant single nucleotide polymorphism called GFI136N (whereby Serine at position 36 is replaced by Asparagine) has been described to play a role in acute myeloid leukemia (AML) development. The prevalence of this variant is about 5-7% in the healthy Caucasian populations.

Patients with myelodysplastic syndrome (MDS) show disturbed bone marrow function and are at risk to develop AML. Identifying prognostic markers and potential targets is essential for improved MDS therapy.

We explored how GFI136N influences onset and progression of MDS and investigated the biological mechanisms behind our findings.

To examine the role of GFI136N with regard to progression of MDS to AML, we characterized the status of GFI1 in 201 German-, 350 US- and 86 Dutch MDS patients. Our results revealed that 11-13% of patients were heterozygous for GFI136N compared to 5-7% in the healthy population. Patients carrying GFI136N also showed a 2-3-fold higher risk of AML-development and inferior leukemia-free survival than patients carrying the common variant GFI136S. The onset of MDS and AML tended to be at younger ages for GFI136N carriers and they did not respond to demethylating therapy to the same extent as GFI136S homozygous MDS patients. Whole exome sequencing revealed that patients carrying GFI136N had mutations in epigenetic modifiers such as EZH2 and expressed higher levels of different oncogenes such as HOXA9.

To dissect the mechanisms behind these findings and to study the role of GFI136N in AML development, we used GFI136N and GFI136S knock-in mice that expressed either of the two human variants instead of murine Gfi1. We transduced hematopoietic progenitor cells from GFI136S or GFI136N expressing mice with retroviral vectors overexpressing oncofusion proteins AML1-ETO9a or MLL-AF9, which can be recurrently found in AML patient cohorts. GFI136N expressing cells that had been transduced with either AML1-ETO9a or MLL-AF9 generated significantly more cells and colonies than GFI136S controls.

In order to confirm our findings in vivo, we used three different AML and MDS mouse models crossed to mice carrying either GFI136N or the common GFI136S.

As a first in vivoapproach, we used a mouse strain that expresses the human oncofusion protein CBFB-MYH11 (inv(16)) frequently found in AML patients. Mice carrying GFI136N and expressing CBFB-MYH11 showed a significantly accelerated onset of AML compared to mice carrying GFI136S (p=0.004). We found similar results in a second AML mouse model expressing MLL-AF9. Using the NUP98-HOXD13 MDS mouse model a marginally significant decrease in leukemia-free survival was observed in mice carrying GFI136N.

To investigate molecular mechanisms that could cause the observed effects of GFI136N, we used ChipSeq and RNASeq. Our results revealed that GFI136N lineage negative cells show a genome-wide higher degree of H3K4 methylation and H3K9 acetylation than cells carrying GFI136S. GFI1 recruits among others LSD1 and HDAC to demethylate H3K4 and deacetylate H3K9 and thus GFI136N fails to initiate epigenetic changes previously shown for GFI1, leading to higher expression of various oncogenes such as HoxA9 and Kras. A number of MDS and AML patients are treated with HDAC and LSD1 inhibitors, but based on the above data we would predict that this therapy would likely fail in GFI136N patients and that therapy with histone acetylase (HAT) inhibitors or histone methyltransferase inhibitors would be more beneficial for these patients.

To test this hypothesis in vitro, we treated the above-described GFI136N and GFI136S expressing AML1-ETO9a or MLL-AF9 leukemic cells with HAT or HDAC inhibitors. HAT inhibitors inhibited growth of preleukemic GFI136N expressing cells more efficiently than HDAC inhibitors.

Similarly in vivo, we transplanted irradiated mice with GFI136N or GFI136S expressing MLL-AF9 leukemic cells. We found that HAT inhibitors tended to prolong the survival of mice transplanted with GFI136N leukemic cells more than mice transplanted with GFI136S leukemic cells.

In summary, GFI136N is a novel prognostic marker for MDS patients that predisposes to AML likely via epigenetic changes at different oncogenes. We also suggest that HAT inhibitors could be more beneficial to GFI136N carrier patients thus enabling a more individualized therapy.