Gene expression profiling of loss of TET2 and/or JAK2V617F mutant hematopoietic stem cells from mouse models of myeloproliferative neoplasms

Myeloproliferative neoplasms (MPNs) are clinically characterized by the chronic overproduction of differentiated peripheral blood cells and the gradual expansion of malignant intramedullary/extramedullary hematopoiesis. In MPNs mutations in JAK2 MPL or CALR are detected mutually exclusive in more than 90% of cases [1], [2]. Mutations in them lead to the abnormal activation of JAK/STAT signaling and the autonomous growth of differentiated cells therefore they are considered as “driver” gene mutations. In addition to the above driver gene mutations mutations in epigenetic regulators such as TET2 DNMT3A ASXL1 EZH2 or IDH1/2 are detected in about 5%–30% of cases respectively [3]. Mutations in TET2 DNMT3A EZH2 or IDH1/2 commonly confer the increased self-renewal capacity on normal hematopoietic stem cells (HSCs) but they do not lead to the autonomous growth of differentiated cells and only exhibit subtle clinical phenotypes [[4], [6], [7], [8],5]. It was unclear how mutations in such epigenetic regulators influenced abnormal HSCs with driver gene mutations how they influenced the disease phenotype or whether a single driver gene mutation was sufficient for the initiation of human MPNs. Therefore we focused on JAK2V617F and loss of TET2—the former as a representative of driver gene mutations and the latter as a representative of mutations in epigenetic regulators—and examined the influence of single or double mutations on HSCs (Lineage−Sca-1+c-Kit+ cells (LSKs)) by functional analyses and microarray whole-genome expression analyses [9]. Gene expression profiling showed that the HSC fingerprint genes [10] was statistically equally enriched in TET2-knockdown-LSKs but negatively enriched in JAK2V617F–LSKs compared to that in wild-type-LSKs. Double-mutant-LSKs showed the same tendency as JAK2V617F–LSKs in terms of their HSC fingerprint genes but the expression of individual genes differed between the two groups. Among 245 HSC fingerprint genes 100 were more highly expressed in double-mutant-LSKs than in JAK2V617F–LSKs. These altered gene expressions might partly explain the mechanisms of initiation and progression of MPNs which was observed in the functional analyses [9]. Here we describe gene expression profiles deposited at the Gene Expression Omnibus (GEO) under the accession number GSE62302 including experimental methods and quality control analyses.


Generation of mouse models of MPNs
We non-competitively transplanted 1 × 10 6 of 4 experimental groups of E14.5 FLs (WT, TET2KD, JAK2V617F, double-mutant) into lethally irradiated B6-CD45.1 mice [9]. Compared with the recipients transplanted with WT cells, the recipients of TET2KD cells showed normal blood cell count, no splenomegaly, comparable overall survival duration, and minimal extramedullary hematopoiesis of the lung and liver, indicating that TET2KD cells developed only a subtle clinical phenotype as MPNs. Recipients of JAK2V617F cells showed leukocytosis, anemia, thrombocytosis, splenomegaly, shorter survival duration, moderate extramedullary hematopoiesis, and fibrosis in bone marrow (BM) and spleen, indicating that JAK2V617F cells induced clinically primary myelofibrosis (PMF)-like MPNs. Double-mutant cells showed not only the phenotype of JAK2V617F cell recipients, but also prolonged leukocytosis, splenomegaly, and severe extramedullary hematopoiesis, with modestly shorter overall survival. These results indicated that the combination of loss of TET2 and JAK2V617F worsened the disease compared to single-mutant JAK2V617F-induced MPNs.

RNA and cDNA preparation for microarray gene expression analyses
To identify genes regulated by loss of TET2 and/or JAK2V617F, we performed microarray gene expression analyses. For the cell preparation, we sacrificed the 4 experimental groups of recipients at 10-16 weeks post-transplantation, and sorted Lineage − Sca-1 + c-Kit + cells (LSKs) from the BM of the recipients by FACSAriaII (BD Biosciences, San Jose, USA). WT-LSKs (n = 1), TET2KD-LSKs (n = 1), JAK2V617F-LSKs (n = 1), and double-mutant-LSKs (n = 1) were each pooled from 5 mice (Table 1), and preserved in TRIzol reagent. RNA samples were isolated from the each pooled 2-6 × 10 4 BM-LSKs. cDNA samples were prepared from 2-5 ng RNA samples using the Ovation Pico WTA System V2 (NuGEN, San Carlos, CA), according to the manufacturer's instructions. The NuGEN Ovation amplification methodology uses an isothermal linear amplification using DNA/RNA chimeric primers.

cDNA labeling and hybridization
Two μg of purified and amplified cDNA was used as input into the Agilent Genomic DNA Enzymatic Labeling Kit (Agilent Technologies, Palo Alto, CA), according to the manufacturer's instructions. The Cy3labeled cDNA was quantified and dye incorporation determined by Nanodrop ND-1000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA). For each hybridization, 1.65 μg of Cy3-labeled cDNA was hybridized at 65°C for 17 h to an Agilent Mouse GE 8x60K Microarray (Agilent, Design ID: 028005). After washing, microarrays were scanned using an Agilent DNA microarray scanner (Agilent).

Microarray study design
Using the 4 experimental groups of prepared cDNA samples (each, n = 1), microarray gene expression analyses were performed. Each gene expression profile of TET2KD-LSKs, JAK2V617F-LSKs and doublemutant-LSKs was compared to that of WT-LSKs; and a gene expression profile of double-mutant-LSKs was compared to that of JAK2V617F-LSKs.
RNA and cDNA quality control analyses RNA and cDNA qualities were assessed by Nanodrop ND-1000 Spectrophotometer (Thermo Fisher Scientific) and by the 2100 Bioanalyser (Agilent). All isolated RNA samples showed appropriate A260/A280 ratios ranging between 1.56 and 1.72, but small peaks for 18 s and 28 s ribosomal RNA subunits and low RNA integrity numbers (RIN) ranging between 2.7 and 5.4 in their electropherograms, probably because of their small amounts or degradation; suggesting their insufficient qualities for the regular microarray protocol without amplification ( Fig. 1 left column). However, according to the manufacturer's instructions, the NuGEN Ovation Pico WTA System V2 enables RNA samples with RIN still around 2.0 to amplify successfully and reproducibly; therefore we performed following RNA amplification and cDNA synthesis. All amplified cDNA samples showed A260/A280 ratios ranging between 1.92 and 1.96, A260/A230 ratio between 2.30 and 2.36, and typical plots between 200 and 2000 kb; indicating their sufficient qualities for the microarray protocol with amplification ( Fig. 1 right column). Finally, all Cy3-labeled cDNA samples showed A260/A280 ratios ranging between 1.79 and 1.84, and dye incorporation rates ranging between 32.1 and 35.1.

Microarray quality control analyses
In each 4 experimental group of hybridization, the grid placements were adequate ( Fig. 2A). Obvious biases in the distributions of outlier probes were not seen, and the frequencies of non-uniform features were sufficiently low (b1%) (Fig. 2B).

Data normalization
There are total of 55,681 probes on Agilent Mouse GE 8x60K Microarray (Design ID: 028005) without control probes, and intensity values of each scanned feature were quantified using Agilent Feature Extraction software version 10.7.3.1, which performs background subtractions. Normalization was performed using Agilent GeneSpring GX version 12.6.1. (per chip: normalization to 75 percentile shift).

Data analysis (cluster analysis)
Dendrogram was constructed from unsupervised hierarchical clustering of data sets from 4 experimental groups of BM-LSKs using Pearson correlation. There were close similarities in the whole-genome expression profiles between JAK2V617F-LSKs and double-mutant-LSKs (Fig. 3).

Discussion
We present here a unique data set of mouse models of MPNs. This dataset is measured by Agilent platform and composed of gene expression profiles of normal and mutant LSKs (WT, JAK2V617F, TET2KD, double-mutant) ( Table 1). JAK2V617F-induced HSC impairments were identified in several mouse models of MPNs including ours [9,15,16]. In our model, though single-mutant JAK2V617F-cells could initiate and promote MPNs during a short-term, they showed reduced selfrenewal capacity in vitro and reduced long-term oncogenic capacity in vivo [9]. However, the double-mutant cells showed increased self- renewal capacity and could initiate and promote MPNs over the longterm [9], indicating that JAK2V617F-induced HSC impairments were restored by combined loss of TET2. Also in human, those impairments were identified in JAK2V617F-HSCs [17], and combined loss of TET2 seemed to restore the JAK2V617F-induced HSC impairments and expand the HSC compartment by altering transcriptional programs [17,18]. Therefore, we tried to uncover the restoration mechanisms by using a gene-profiling approach in mouse models. Here, we showed that many HSC fingerprint genes were down-regulated in JAK2V617F-LSKs, but the expressions of significant number of them were restored in double-mutant-LSKs (Table 2, Fig. 5), though we could not see statistically significant restoration of the profile (Fig. 4D). These expressional changes might partly explain the mechanisms of initiation and progression of MPNs. In spite of our and other studies [9,18,19], the precise mechanisms by which loss of TET2 restores the JAK2V617Finduced HSC impairments still remain poorly known, further wet and dry investigations are necessary to uncover them more precisely.