ALBERT

Description: Purpose Follicular lymphoma (FL) accounting for 20% of non-Hodgkin lymphoma, is currently considered treatable but not curable despite more effective treatment options. Idelalisib, a selective inhibitor of phosphatidylinositol 3-kinase δ (PI3Kδ) that blocks PI3Kδ-AKT signaling and promotes apoptosis, is approved for patients with relapsed/refractory (R/R) FL who have received ≥ 2 prior systemic therapies. Previous studies suggest that somatic tumor mutations identified at relapse may have an impact on the response to targeted therapies (Bartlett N et al., Blood 2018). The goal of this study was to analyze the lymphoma mutational profile to identify the prognostic value of mutations in R/R follicular lymphoma patients treated with idelalisib. Methods We performed a retrospective multicenter study of patients (pts) with relapsed follicular lymphoma and no evidence of transformation, having received at least 2 prior regimens before idelalisib treatment. Patients received idelalisib 150mg BID until progression or toxicity. Next-generation sequencing (NGS) of 51 genes was performed either at FL diagnosis and/or at relapse prior idelalisib therapy. The primary endpoint was to analyze the relationship between the mutational status and the duration of response (DOR) to idelalisib. DOR was measured from the time of initial response until documented lymphoma progression. According to DOR pts were classified as: refractory (response 〈 1 month), short-responder (1 month ≤ DOR ≤ 12 months) and long-responder (〉 12 months). Results 24 pts with R/R FL were enrolled with a median age of 62.5 years (range: 57-86). Patients had received a median of 3 prior treatments (range: 2-7). FL was refractory to rituximab in 16 pts (67%), to alkylating agents in 11 pts (46%) and to 2 or more prior treatments in 11 pts (46%). Twelve (50%) had refractory disease to the last therapy before idelalisib. Pts received idelalisib during a median of 5.5 months (range: 0.5-31). Overall response rate was 83% (n=20) including 3 (15%) complete response and 17 (85%) partial response. The median DOR was 9.5 months (range: 0-28). Eleven pts were short-responders and 9 long-responders. Four pts (17%) had refractory disease to idelalisib. Median progression-free survival and overall survival were 11.5 (range: 1-30) and 16.5 months (range: 1-56) respectively. Three pts (12.5%) are still continuing idelalisib. Twenty-one pts (87.5%) discontinued treatment mostly due to progressive disease (n=14) and adverse events (n=5); 3 pts remained progression-free after idelalisib discontinuation and observation with a median follow-up of 23 months (range: 20-24). All the pts had at least one mutation detectable for one of the targeted genes in both diagnosis (n=17) and relapse (n=20) samples. The median number of targeted genes with non-silent mutations per patient was 7 at diagnosis (range: 2-11) and 6 at relapse (range: 3-28). The most frequent genes found in 10% of patients (≥ 2) or more at diagnosis and at relapse are listed in Figure 1. The mutational profile at diagnosis predominantly included mutations in epigenetics gene family, KTM2D (n=16; 94%), EP300 (n=9; 52%), ARID1A (n=6; 35%), KTM2A (n=5; 29%) and CREBBP (n=4; 23%). The m7-FLIPI score (Pastore A et al., Lancet Oncology 2015) at diagnosis helped identifying pts with low-risk (n=12; 71%) and high-risk (n=5; 29%) of idelalisib treatment failure. The median m7-FLIPI score in the refractory group was 1.25 compared to 0.26 in the short responder group and 0.04 in the long responder group. All long responder patients had low-risk m7-FLIPI. The mutational profile at relapse was significantly enriched in mutations of TNFAIP3 (n=7; 35%) and NFKBIE (n=4; 20%), affecting the NF-kB inhibitor pathway, and mutations of transcription factors, TP53 (n=10; 50%), MEF2B (n=5; 25%), FOXO1, STAT6 and IRF4 (n=4; 20% each). Overall the mutational profile of the 3 sub-groups according to DOR was detailed in Figure 2. The genes more frequently mutated in refractory pts were: EP300 (n=3/4; 75%), B2M (n=2/4; 50%), FBXW7 (n=2/4; 50%), CARD11, CXCR4 and MYD88 (n=1/4; 25% each). Conclusion The m7FLIPI at diagnosis identifies patients with higher risk of treatment failure in patients with R/R FL treated with idelalisib. Patients with idelalisib refractory disease have more frequently mutations of EP300, B2M, FBXW7, which suggests they could be related to resistance to idelalisib. Disclosures No relevant conflicts of interest to declare.

Description: The polycomb group protein Bmi-1 is a well known determinant of hematopoietic stem cell function. Bmi-1-/- mice display severe hematopoietic defects, including progressive loss of hematopoietic cells from the bone marrow. Bmi-1 is dispensable for hematopoietic stem cell specification, but essential for their maintenance, an effect attributable to its ability to promote HSC self-renewal. The mechanism by which Bmi-1 regulates this process is not completely understood. Bmi-1 has been shown to repress the INK4A/ARF locus encoding the cell cycle inhibitors p16ink4a and p19arf , to interact with the E4F1 protein and to regulate the DNA damage response pathway, however experimental manipulation of these proteins/pathways only partially rescues the hematopoietic defects of the Bmi-1-/-mice. It thus appears that the mechanism by which Bmi-1 regulates HSC self-renewal remains to be determined. Towards this goal, we purified Bmi-1 containing protein complexes from cellular extracts and identified Bmi-1 interaction partners by mass spectrometry. We observed that the protein Ubap2l, which has never been shown to associate with Bmi-1 and for which no link with polycomb group protein function has been described, was consistently found in Bmi-1-containing protein complexes. Immunoprecipitation experiments revealed that Ubap2l indirectly associates with Bmi-1 via an interaction with the polycomb group protein Rnf2. We then evaluated the possibility that Ubap2l might be involved in the regulation of HSC activity. We observed that Ubap2l transcripts are more abundant in primitive HSC populations compared to total BM. CFC assays performed with BM cells infected with Ubap2l shRNAs revealed that Ubap2l knockdown causes a modest and progressive loss of progenitor activity when cells are kept in culture, with multipotent and bipotent progenitors being substantially more affected than unipotent progenitors. We transplanted these cells in mice and observed a gradual decrease in the percentage of donor derived cells expressing Ubap2l shRNAs in the peripheral blood of the recipient mice, with the most striking effect observed 16 weeks post-transplantation in the BM. Bmi-1 has been shown to regulate the proliferative capacity of both progenitor and stem cells, and its deletion in BM cells is known to dramatically reduce the reconstitution activity of these cells at early time points following transplantation. In contrast, Ubap2l appears to preferentially regulate LTR-HSC activity. We tested the effects of Ubap2l silencing on leukemic cells in vivo and observed that a reduction of Ubap2l levels in these cells had an important impact on their ability to reconstitute recipient mice, suggesting that Ubap2l also plays a role in leukemic stem cell activity. We determined if the mechanism by which Ubap2l regulates HSC activity is related to Bmi-1 function by simultaneously introducing Bmi-1 cDNA and Ubap2l shRNAs in BM cells and found that Bmi-1 is able to rescue the long-term reconstitution defect caused by Ubap2l downregulation in these cells. We observed that Ubap2l silencing does not significantly affect the expression of the known Bmi-1 targets p16ink4a and p19arf, implying that Ubap2l regulates HSC activity via a Bmi-1-dependent mechanism that does not involve repression of the INK4A/ARF locus. One explanation for the two Bmi-1 dependent mechanisms at play in the regulation of HSC activity could be that Bmi-1 is part of two separate protein complexes, each regulating different aspects of hematopoietic cell function. To test this hypothesis, we fractionated cellular extracts and were indeed able to resolve two distinct Bmi-1 containing protein complexes, distinguishable by the presence of Ubap2l. Based on the results we obtained, we propose a model in which two different Bmi-1 containing protein complexes regulate hematopoietic stem cell function. An Ubap2l-independent complex, which is most likely involved in the repression of the INK4A/ARF locus, and could be responsible for the effects of Bmi-1 on multipotent progenitors and STR-HSCs, and an Ubap2l-dependent complex, which operates via a yet to be defined mechanism unrelated to p16Ink4a and p19Arf, and would account for the effects of Bmi-1 on LTR-HSC activity. These results position Ubap2l as a key regulator of LTR-HSC activity and unveil a novel protein complex mediating the effects of Bmi-1 on LTR-HSCs. Disclosures: No relevant conflicts of interest to declare.

Description: Key Points UBAP2L interacts with BMI1 as part of a novel Polycomb subcomplex. UBAP2L regulates HSC activity via a mechanism unrelated to the repression of the Ink4a/Arf locus.

Description: Abstract 2370 TIF1gamma (or TRIM33) is an ubiquitous nuclear protein that belongs to the transcriptional intermediary factor 1 family. Human and mouse TIF1gamma are closely related to zebrafish moonshine (mon), a gene whose mutations disrupt embryonic and adult hematopoiesis with severe red blood cell aplasia. Targeted deletion of Tif1gamma is embryonic lethal in mice. In zebrafish and human CD34+ cells, TIF1gamma functionally links positive elongation factors such as p-TEFb and FACT to blood specific transcription complexes (e.g. the SCL/TAL1 complex) to regulate elongation of genes by antagonizing Pol II pausing. TIF1gamma also affects the human hematopoietic progenitor cell response to the cytokines of the transforming growth factor-beta superfamily through various mechanisms. Recently, we showed that the loss of Tif1gamma in mouse hematopoietic stem cells (cFES-Cre-Tif1gamma) favors the expansion of the granulo-monocytic progenitor compartment. The gene deletion induces the age-dependent appearance of a cell-autonomous myeloproliferative disorder with myelodysplastic features, monocytosis, and hepatosplenomegaly that recapitulates essential features of human chronic myelomonocytic leukemia (CMML). Interestingly, TIF1gamma is almost undetectable in leukemic cells of 35% of patients with CMML. This down-regulation is related to the hyper-methylation of CpG sequences in the gene promoter. Our results demonstrated that TIF1gamma is an epigenetically regulated tumor suppressor gene in hematopoietic cells. In addition, an altered production of peritoneal macrophages was observed in our mouse model. These macrophages did not adhere to the plastic and were morphologically abnormal in vitro. In bone marrow and in Lin- progenitor cells, Tif1gamma deletion leads to a significant decrease of cfms (Csf-1r) expression, required for the differentiation, proliferation, and survival of monocytic phagocytes. We also identified in CMML patients the association between low levels of TIF1gamma and cFMS (Aucagne et al., J. Clin. Invest., 121, 2361–2370, 2011). To gain insight into the possible mechanism accounting for diminished accumulation of macrophages, we examined the expression of c-Fms. We show that level of its expression is reduced significantly in blood monocytes isolated from Tif1gamma-deleted mice. When Tif1gamma-deleted sorted myeloid cells were induced to differentiate into macrophages in presence of CSF-1, a delayed production of few abnormal large macrophages was observed. Apoptosis was associated with this alteration of differentiation. This phenomenon was also characterized in young mice not developing the disease yet. The morphological abnormalities were correlated with very important alterations of specific macrophage differentiation markers. Expression of specific transcription factors involved in macrophage differentiation was deeply deregulated. Moreover, macrophage function such as migration, cytokine or chemokine secretion in response to LPS was altered. Likewise, in vitro differentiation of monocytes into dendritic cells was also abnormal. Altogether, our results suggest that monocyte plasticity is at least partially orchestrated by Tif1gamma. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 2366 Colony-stimulating factor-1 (CSF-1 or M-CSF) triggers the differentiation of human peripheral blood monocytes into macrophages through and integrated cytokine/transcription factors circuitry. Using microarray profiling to explore the role of microRNAs (miRNAs) in this molecular circuitry, we identified the down-regulation of miR-142-3p in human macrophages obtained from CSF-1-treated monocytes. We show that miR-142-3p is a repressor of the transcription factor EGR2 (Early Growth Response 2) through direct 3'UTR interactions. Interestingly, EGR2 binds the promoter of the pre-miR-142-3p gene to negatively regulate its expression, identifying a self-regulatory feedback loop. Enforced expression of miR-142-3p in primary human monocytes as well as decreased expression of miR-142-3p observed in monocytes from patients with a chronic myelomonocytic leukemia further assess the link between miR-142-3p and EGR2 expression in these cells. A chemical inhibition of the Src kinase family prevents the regulation loop induced by CSF-1. Thus, our study uncovers an EGR2/miR-142-3p circuitry which regulates CSF-1 driven differentiation of human monocytes into macrophages. Disclosures: No relevant conflicts of interest to declare.

Description: MOZ (MOnocytic leukaemia Zinc finger protein) (also called MYST3 or KAT6A) is a member of the MYST family of HATs which likely acetylate H4K16. The MLL (MixedLineageLeukemia) gene is a frequent target for recurrent chromosomal translocations found in AML and ALL. MLL (KMT2A) is a methyl-transferase targeting H3K4. It was shown that MOZ/CBP leukemia, as observed in MLL-rearranged leukemias, harbors abnormal levels of homeobox (HOX) genes expression. HOX transcription factors have a crucial function in hematopoiesis regulation. In addition, HOXA5, HOXA7, and HOXA9 are often considered to be pivotal HOX genes for MLL transformation, constituting downstream targets of MLL. In our study, we characterize an in vivo cooperation between MOZ and MLL. MOZ co-localizes and interacts with MLL in vivo. We also demonstrate that WDR5, an adaptor protein essential for trimethylation of H3K4, belonging to the MLL complex, co-localizes and interacts with MOZ. Moreover, by in vitro pull-down assays, we show that MOZ can interact with the lysine 4 of histone H3 methylated by MLL, as WDR5. Hence, MLL, WDR5 and MOZ cooperate together on the same targets corresponding to the trimethylated H3K4. MOZ may couple histone acetylation to histone methylation throughout its recruitment on lysine methylated by MLL. Furthermore, ChIP experiments in HSCs indicate that MLL and MOZ are both recruited to HOXA5, HOXA7, and HOXA9 genes promoters in correlation with the presence of H3K9K14ac, H4K16ac, H3K4me2 and H3K4me3, specific epigenetic marks of transcription activation. Remarkably, knockdown of MLL or MOZ inhibits MOZ or MLL recruitment and their specific epigenetic marks from HOX genes promoters, respectively. These alterations are associated with abnormalities in hematopoietic differentiation of HSCs. Our data suggest that the histone methyl-transferase MLL and the MYST family histone acetyl-transferase MOZ functionally cooperate to modulate HOX genes expression through specific epigenetic modifications in HSCs. In conclusion, we provide an example of a mechanism involving a direct cross-talk between two histone modifying enzymes, facilitating the rapid remodeling of chromatin in order to favor the commitment of human HSCs via HOX genes regulation.