TET2 deficiency promotes MDS-associated leukemogenesis

Dear Editor, Myelodysplastic syndrome (MDS) is a group of clonal hematopoietic disorders that frequently progress to acute myeloid leukemia (AML) [1]. However, mechanisms underlying such transformation are not yet fully understood. TET2 is one of the most frequently mutated genes in myeloid malignancies [2]. We previously demonstrated that post-translational modification of TET2 protein led to DNA hypermethylation and dysregulated gene expression in MDS hematopoietic stem and progenitor cells (HSPCs), conferring a clonal advantage [3]. TET2 down-regulation was also seen during MDS progression [4]. Herein, we retrospectively analyzed GEO datasets including AML or MDS sample cohort. We found that lower TET2 levels seen in high-risk MDS (HRMDS) were closely associated with shorter survival (Fig. S1A–C) [5]. Moreover, relative to those with wild type (WT) TET2, AML patients harboring TET2 mutations exhibited a lower survival rate and a higher likelihood of AML secondary to MDS or MPN (Fig. S1D, E) [6]. Collectively, these observations prompted us to assess TET2 function in leukemia transformation of MDS. To do so, we used a Nup98-HoxD13 (NHD13) transgenic mouse model, in which ~30% of mice develop AML. Interestingly, TET2 levels were lower in c-kit bone marrow (BM) cells of leukemiatransformed NHD13 mice relative to age-matched NHD13 mice, which developed MDS exclusively (Fig. S1F, G), suggesting transformation linked to TET2 downregulation. Thus, we crossed Tet2 conditional knockout (KO, Tet2/Mx1-Cre) or corresponding control (Tet2) mice with NHD13 mice and monitored leukemia development following poly(I:C) treatment on both genotypes (Fig. S1H, I). Notably, Tet2 deletion shortened median survival of NHD13 mice (Fig. 1A). Within 30 weeks, 5 of the 10 mice from the NHD13/Tet2-KO cohort developed AML, while none of the NHD13/ Tet2-WT mice exhibited signs of leukemia (Table S1). Specifically, leukemic NHD13/Tet2-KO mice showed increased white blood cell (WBC) counts, splenomegaly and hyper-cellularity in BM, while age-matched NHD13/Tet2-WT mice exhibited only cytopenia (Figs. 1B and S1J–L). NHD13/Tet2-KO mice showed increases in the c-kit subset and blasts in BM compared to NHD13/Tet2-WT mice (Figs. 1C, D and S1M). Moreover, secondary recipients also developed AML following transplant of leukemic NHD13/Tet2-KO BM cells (Fig. S1N, O). NHD13 transgenic mice were characterized by HoxA9 elevation [7]. Thus we evaluated TET2 function in MDS or AML patients that showed differences in HOXA9 expression. While HOXA9 or TET2 levels alone did not predict prognosis of the entire MDS population, the HOXA9/TET2 combination predicted shorter survival relative to those with HOXA9/TET2 (Fig. S1P–R). Moreover, TET2 mutation also predicted shorter overall survival in the HOXA9 AML population (Fig. S1S, T). We then transduced Tet2-WT or Tet2-KO BM cells with HoxA9 and transplanted the cells into recipient mice to monitor leukemia development. We observed that 6 of 8 HoxA9/Tet2-WT recipients survived up to 200 days, while all 8 recipients of HoxA9/Tet2-KO cells developed lethal AML, starting at day 62 (Fig. S1U, V), suggesting that Tet2 deletion promotes AML transformation and in agreement with outcomes seen in NHD13/Tet2-KO mice. To define mechanisms underlying MDS progression, we evaluated the Tet2-KO vs. Tet2-WT NHD13 BM compartment at a pre-leukemic stage (20-weeks-old). At that time point, neither genotype showed signs of leukemia (Fig. S2A), but Tet2 deletion increased the number of monocytes in BM of NHD13 mice (Fig. S2B–D). Importantly, BM cells from NHD13/Tet2-KO mice showed an increase in the Linc-kitSca-1 (LK) population relative to those of Tet2-WT NHD13 mice, whereas the Linc-kitSca-1 (LSK) population was unchanged by Tet2 deletion (Figs. 1E and S2E). Increases in the LK subset are likely due to decreased apoptosis following Tet2-KO (Fig. 1F). Within the LK subset, we observed increases in common myeloid progenitors (CMPs) and granulocyte-monocyte progenitors (GMPs) in NHD13/Tet2-KO mice (Fig. S2F, G). Moreover, Tet2 loss did not alter the cell cycle of c-kit cells (Fig. S2H) or that of the LK subset (not shown). Colonyforming cell (CFC) assays revealed that NHD13/Tet2-KO BM cells formed colonies in the absence of cytokines (Fig. S2I). In the presence of cytokines, Tet2-KO cells exhibited a slightly higher CFC number than did Tet2-WT cells, the difference was further amplified in the context of NHD13 (Fig. S2J, K). Tet2-KO cells also exhibited higher replating capacity than did Tet2-WT cells (Fig. S2L). We next transplanted LK cells (CD45.2) from pre-leukemic NHD13/Tet2-KO or corresponding control NHD13 mice into lethally-irradiated secondary recipients to assess leukemogenicity. As expected, NHD13/Tet2-KO cell transplantation increased the percentage of CD45.2 cells and WBCs in peripheral blood (PB) relative to NHD13/Tet2-WT cells (Figs. 1G and S2M, N). By 16 weeks post-transplantation, 14 of 18 NHD13/Tet2-KO recipients developed AML and exhibited increased numbers of c-kit cells and blasts in PB (Fig. S2O, P). Notably, NHD13/Tet2-KO transplants showed shorter survival than NHD13/Tet2-WT transplants (Fig. S2Q). Moreover, we also analyzed the transplants using donor c-kit cells (CD45.2) from WT or Tet2-KO mice (Fig. S2R, S) and observed that recipients from both genotypes survived up to 24 weeks following transplantation, with no signs of leukemic transformation (data not shown). Collectively, these results indicate that Tet2-KO-mediated leukemogenesis is associated with expansion of the MDS HSPC (LK subset) pool. Tet2 loss in HSPCs can lead to hypermutagenicity [8]. To evaluate these outcomes, we performed whole-exome sequencing of c-kit cells from pre-leukemic NHD13/Tet2-KO vs. matched NHD13/Tet2-WT mice. Relative to NHD13/Tet2-WT mice, we observed 271 newly acquired alterations and 199 alterations with increased variant allele frequency (VAF, fold-change >1.5) in NHD13/Tet2-KO mice (Fig. S3A and Table S2). KEGG analysis of these alterations (271+ 199) in NHD13/Tet2-KO mice revealed significant enrichment of genes related to cancer and signaling pathways (Fig. S3B). We next focused on the top 70 altered genes


Mice
The Nup98-HoxD13 (NHD13) transgenic mice were from Jackson Laboratory. The Tet2 flox/flox (Tet2 fl/fl ) and Mx1-Cre mice were mated on a C57BL/6J genetic background. The NHD13 mice were crossed with Tet2 fl/fl /Mx1-Cre or Tet2 fl/fl mice to generate NHD13/Tet2 fl/fl /Mx1-Cre and control NHD13/Tet2 fl/fl mice. Both genotypes were injected with poly(I:C) at 6-weeks-old. All mice were drug or test naive and not involved in previous procedures. Mice of the same gender and age were randomly divided into groups. Investigators were blinded to mouse genotype while performing treatment or monitoring engraftment or survival. Mice were bred at the City of Hope (COH) animal facility and received autoclaved water and clean food. All mice were subjected to 12-hour light/dark cycles and kept in controlled ambient room temperature and air humidity conditions. All animal procedures were conducted in accordance with established institutional guidance and approved protocols of the Institutional Animal Care and Use Committee at COH.

Cell line and primary cell culture
The human MDS cell line MDS-L was provided and developed by Dr. Kaoru Tohyama. According to the inventor, cells were cultured in RPMI 1640 medium with 10% FBS and 1% penicillinstreptomycin in the presence of 30 ng/mL human recombinant human IL-3. For MDS-L, no known reference STR profiling is available. To avoid misidentification, cross-contamination, or genetic drifting during the experiments, we replaced the cultured cells with original frozen stock (100 vials totals) which was originally directly provided from inventor Dr. Kaoru Tohyama. Human CB CD34+ and primary MDS CD34+ cells were maintained in StemSpan Serum-Free Expansion Media (StemCell Technologies, Inc.) supplemented with recombinant human SCF (50 ng/mL), Flt3 ligand (100 ng/mL), TPO (100 ng/mL), IL-3 (25 ng/mL), and IL-6 (10 ng/mL).

BM c-kit + cell selection
BM c-kit + cells were isolated using mouse CD117 microBeads (Miltenyi Biotec) according to the kit protocol. Briefly, total BM cells were counted and resuspended in MACS buffer. After staining with CD117 microbeads (20µL per 10 7 total cells, 15 min at 4℃), cells were washed and applied onto MS columns, which were rinsed and placed in a magnetic field. Columns were washed three times with 500 µL buffer and then removed from the separator. Magnetically-labeled cells were flushed out with 1mL buffer. c-kit + cells were counted and used for further experiments.
Chimerism of donor-derived hematopoietic cells was monitored by flow cytometry.

RT-qPCR analysis
For RT-qPCR, RNA was isolated using Trizol reagent (Invitrogen) or an RNeasy micro kit (QIAGEN), following standard manufacturers' protocols. cDNA was amplified using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Quantitative real time PCR was performed using TaqMan or SYBR Green master mix (Life Technologies) with 0.2 mM Taqman probe (Life Technologies) or gene-specific primers. Signals were detected with a QuantStudio 7 Flex Real-Time PCR system (Life Biotechnology). Relative expression levels were determined after normalization to GAPDH levels.

Lentivirus transduction and cell transfection
First, 293T cells were transiently transfected with pCDH plasmids and pMD2G/pSAPX2 packaging plasmids. Then, 24-36 hours later, supernatants containing replication-incompetent lentiviruses were collected and concentrated using PEG-it (System Biosciences). After titration, target cells were exposed to virus (MOI=10), and 48 hours later, cells were sorted by flow cytometry based on GFP/RFP expression.

In vitro colony forming assays
Total BM or c-kit + cells were obtained and suspended in methylcellulose medium (M3434, StemCell Technologies; colonyGEL 1201, ReachBio Research Labs) and then seeded into plates, according to the manufacturer's protocol. Colonies were counted 7 days later. For replating assays, cells from each plate were harvested and replated at 5 × 10 3 per well.

Cell viability analysis
Cell viability was determined using the Cell Titer-Glo Luminescent Cell Viability Assay Kit (G7570, Promega). Briefly, cells were cultured under different treatment conditions and transferred to 96-well plates for 30 min. Kit substrate and buffer reagents were then added to each well and mixed to allow cell lysis. Plates were read on a Microplate reader (Beckman culture DTX880). Results were expressed as a percentage of untreated cells, for three replicates.

Whole-exome sequencing (WES)
Initial WES was carried out to identify mutations in gene exomes. Three NHD13/Tet2 fl/fl and five NHD13/Tet2 fl/fl /Mx1-Cre mice from the same pair of breeders (NHD13/Tet2 fl/fl and Tet2 fl/fl /Mx1-Cre) were used to exclude the confounding results of germline mutations. Tet2 was deleted by injection of poly(I:C) at 6-weeks-old and c-kit + cells were isolated at 20-weeks-old. Genomic DNA was captured with the Agilent SureSelect all mouse exon probes according to the manufacturer's protocol. 100-bp paired-end sequencing was performed using an Illumina HiSeq 2500 system. Raw sequencing reads were mapped to the whole mouse genome (mm10) using PEMapper/PECaller (https://github.com/wingolab-org/pecaller) with default settings, and variant bases were annotated with SeqAnt (http://seqant.genetics.emory.edu/).

HPRT mutation analysis
The HPRT mutation assay was conducted to evaluate spontaneous forward mutation frequency.
Briefly, K562 cells were selected in Hypoxanthine-aminopterin-thymidine (HAT) medium for 3 days to remove pre-existing HPRT mutants. Next, the cells were washed and seeded in normal culture medium for 13 days for mutant expression. After the phenotypic expression period, the cells were added to 96-well microplates (4 × 10 5 cells/well) in culture medium with selection using the toxic analogue 6-TG (from Sigma UK) at 0.6 μg/ml. Plating efficiency was determined by culturing 200 cells/well in the absence of 6-TG. After 14 days of culturing, colonies of 20+ cells in diameter were counted as viable expansion. The mutation frequency was determined by dividing the number of 6-TG-resistant colonies by the total number of cells plated, normalized by the plating efficiency.

5-hydroxymethylcytosine DNA-IP (hMeDIP) sequencing and analysis
hMeDIP was performed according to a protocol provided in the SimpleDIP Hydroxymethylated DNA IP kit (Cell Signaling Technology Inc.). First, illumina barcode adapters were ligated to sonicated genomic DNA prior to hMeDIP, and DNA was then denatured and incubated with anti-5hmC antibody at 4 ℃ overnight. The DNA/antibody complex was then captured on protein G magnetic beads. Enriched DNA was purified and sequenced followed by standard Illumina protocols on the Illumina HiSeq 2500 platform (Illumina). Sequencing reads were aligned to the mouse genome build mm10 using the Burrows-Wheeler Aligner (BWA, v0.7.5a).

hMeDIP-qPCR analysis
hMeDIP assays were performed based on the manufacturer's instructions using the SimpleDIP kit.
DNA was denatured and IP'd with anti-5hmC or IgG control antibodies and protein G magnetic beads. After three washes, 5hmC-modified DNA was eluted from beads, purified and used as template for qPCR.

Statistical analysis
Data obtained from independent experiments were expressed as means ± SEM, and n represented the number of samples. Unless otherwise specified, two-tailed Student's t-tests were used to compare differences between groups. GraphPad Prism 8 software (La Jolla, CA) was used for statistical analyses. P<0.05 represented a statistically significant difference.

Sequences of shRNAs
shArih2 #1: shArih2 #2:       Figure S5. Tet2 deficiency induces a state of hypermutagenicity.   Table S1. Tet2 loss accelerates leukemogenesis in the NHD13 mouse model. Included shows all mice from the survival analyses ( Figure 1D). CBC, flow cytometry and morphologic analyses were included in the determination of cause of death. In stained panels, leukemia blasts are indicated by green arrows and dysplastic cells by red. Tarsl2 Newly acquired mutations Higher VAF mutations (VAF FC>1.5) Table S2. KEGG pathway analysis of mutations with increased VAF in NHD13 /Tet2 -KO mice. 470 genetic alterations with significantly increased VAF (fold change > 1.5) or newly acquired mutations in NHD13 /Tet2 -KO mice were analyzed for enrichment in KEGG gene sets using DAVID Functional Annotation Tools. Pathways were ranked based on the number of enriched genes.  Table S3. Top altered genes with AML survival data in PRECOG. 470 genes (newly acquired mutations or higher VAF mutations) in NHD13 /Tet2 -KO mice were selected and the top 70 altered genes (VAF FC>2, P<0.05) were imported into PRECOG. Shown are 37 genes with meta-z-scores indicating their AML prognostic significance.