JAK2^V617F Myeloproliferative Neoplasms Support Parallel Evolution of Independent Leukemic Clones

Abstract

Myeloproliferative neoplasms (MPNs) are hematological diseases predominantly driven by the JAK2^V617F mutation. Progression from chronic-phase MPN to secondary acute myeloid leukemia (sAML) is a severe complication that dramatically worsens disease prognosis. While progression to sAML is classically linked to MPN clones acquiring additional mutations, the absence of JAK2^V617F in some cases of post-MPN sAML cases suggests alternative mechanisms of transformation. Utilizing patient samples and in vivo modeling, we establish that leukemic clones can emerge independently of JAK2-mutant cells and undergo positive selection in the pro-inflammatory MPN environment, leading to parallel disease evolution. Genetic and pharmacological inhibition of IL-12 and TNFα mitigates this competitive advantage. Our data establish a new paradigm and show that disease progression in MPN can arise from parallel acute myeloid leukemia (pAML) clones.

Myeloproliferative neoplasms (MPNs) are a group of chronic hematological diseases driven by the acquisition of somatic mutations, predominantly JAK2^V617F, in hematopoietic stem cells (HSCs)(1–3). MPNs are characterized by the aberrant and unregulated proliferation of one or more myeloid lineages, resulting in the overproduction of mature hematopoietic cells in the bone marrow (BM) and peripheral blood (PB)(4–6). This sustained overproduction manifests as polycythemia vera (PV; excess erythrocytes), essential thrombocythemia (ET; excess platelets), or myelofibrosis (MF; BM fibrosis). Consequently, patients face increased risks of blood viscosity and clotting, organ enlargement, joint pain and swelling, BM scarring, and are at-risk for progression to more aggressive disease states. The transformation of chronic-phase MPN to secondary acute myeloid leukemia (sAML) is a severe complication traditionally linked to the acquisition of additional mutations in the MPN driver clone in genes encoding epigenetic regulators (e.g. TET2, DNMT3A, ASXL1), tumor suppressors (e.g. TP53, JARID2) and signaling molecules (e.g. N/K-RAS)(7–10). However, the JAK2^V617F driver mutation is occasionally absent in sAML that transforms from antecedent JAK2-mutant MPNs, which suggests alternative mechanisms of disease evolution(11). Proposed explanations for the absence of the JAK2^V617F mutation at the sAML stage include somatic reversion and loss of heterozygosity, but an underexplored possibility is that sAML originates from independent clones that evolve in parallel to the MPN during the natural history of clonal hematopoiesis (CH)(12, 13).

Here, we establish that pre-leukemic clones can arise independently of the MPN and outcompete the JAK2^V617F-mutant cells to manifest disease progression, challenging the traditional dogma of post-MPN sAML evolution. We leveraged a series of primary patient samples (Table S1) and pre-clinical models to investigate the growth and evolution of parallel clones in an MPN background. We demonstrate that independent clones carrying TET2 and TP53 mutations are positively selected by MPN-derived pro-inflammatory cytokines such as IL-12 and TNFα. Importantly, we demonstrate these mechanisms are amenable to targeted therapy as genetic and pharmacological inhibition of IL-12- and TNFα-signaling ameliorates the competitive advantage of TET2-mutant cells in the presence of JAK2^V617F-mutant MPN. Collectively, these data demonstrate that JAK2^V617F-mutant cells condition an environment that confers a selective advantage for the parallel growth of independent clones in the background of existing MPN. In such cases, the more accurate diagnosis is two independent diseases evolving in the same patient – the primary MPN and a parallel acute myeloid leukemia (pAML). Improved understanding of the phylogeny of MPN disease evolution may offer new clinical opportunities for these patients who currently have very limited treatment options.

Leukemic Clones Can Arise Independently and Undergo Positive Selection in a Background of MPN

To investigate the clonal trajectory post-MPN sAML, we performed single cell genomic analysis for three paired patient samples (2 MF; 1 PV) who progressed from a JAK2^V617F-mutant MPN to a JAK2-negative sAML as defined by clinical sequencing (Fig. S1A). Single-cell sequencing data from purified CD34⁺ cells for two patients (MF UPN:950899; PV UPN:374024) demonstrated that mutations detected at the sAML stage were not detected within the JAK2-mutant clones at the MPN stage (Fig. 1A). Despite clinical genomics defining MF UPN:638574 as JAK2-negative sAML, single cell analysis of purified CD34⁺ cells clearly showed the leukemic mutations were acquired in the founding MPN clone (Fig. 1A). This was confirmed by droplet digital PCR (ddPCR) wherein the JAK2^V617F mutation was detected at both the MPN and the sAML stages (Fig. S1B). While these data show there are multiple routes to transformation of post-MPN sAML, the single cell genomics clearly show that the leukemia-initiating mutations are not always present within the JAK2-mutant clones and can arise independently and outcompete the MPN cells to drive transformation.

Fig. 1. Leukemic Clones Can Arise Independently and Undergo Positive Selection in a Background of MPN.

(A) Clonal hierarchy visualization of three paired (MPN and sAML) patient samples resolved by single-cell genomic sequencing with the x-axis expressed as time and the y-axis displaying proportionate prevalence of each clone within the population. (B) ex vivo competition assay showing VAF of engineered AAVS1, TET2, and TP53 mutations in cord blood CD34⁺ cells in co-culture with CD34⁺ cells from JAK2^V617F PV or MF patients or healthy donor (HD) BM.

Given this finding, we developed an ex vivo competition system to quantify the proliferation of independent clones in the presence of MPN cells utilizing CRISPR/Cas9 editing. 1×10⁵ CD34⁺ cells from JAK2-mutant MPN patients or healthy donor (HD) BM were co-cultured for 12-days with 2×10⁴ cord blood (CB)-derived CD34⁺ cells nucleofected with gRNAs targeting AAVS1, TET2 or TP53. CRISPR edited alleles were tracked via targeted genomic sequencing at 6-day intervals and compared to initial values 24-hours post-nucleofection. In the presence of MPN patient CD34⁺ cells, TET2- and TP53-mutant clones expanded significantly more compared to co-culture of the same clones with HD BM control CD34⁺ cells (Fig. 1B). Interestingly, PV patient cells supported the growth of both TET2- and TP53-mutant cells moreso than MF patient cells, a finding consistent with clinical reports that the JAK2-mutant MPN to JAK2 wild-type sAML trajectory is more common in PV patients(11, 13).

JAK2^V617F-Mutant MPN Accelerates Expansion of Independent TET2- and TP53-Mutant Clones

To explore competition dynamics between MPN cells and independent clones, we leveraged our MPN patient derived xenograft (PDX) system(14). TET2 and TP53 mutations were engineered into CB-derived CD34⁺ cells using CRISPR/Cas9 homology-directed repair (HDR) to introduce TET2^1216* and TP53^R248Q knock-in (KI) mutations into CB-derived CD34⁺ cells (Fig. S2). These mutations were chosen from the limited literature describing TET2 and TP53 mutations as the most common variants in JAK2-negative post-MPN sAML(15, 16). A single-stranded oligo donor nucleotide (ssODN) was designed to introduce a silent mutation in the inert AAVS1 locus to serve as a trackable negative control genetic barcode. As most TET2 and TP53 mutations in myeloid malignancies lead to loss of function(17–22), indels resulting from CRISPR/Cas9 cutting but failed ssODN directed repair were additionally tracked. These edits should also provide the cells with a competitive advantage and increase the clonal complexity within a given experiment. PDX models were established by co-transplanting 2.0×10⁴ cells from each of the AAVS1, TET2 and TP53 nucleofected populations into NSGS immunodeficient mice with 1.0×10⁵ CD34⁺ cells derived from either HD BM confirmed to be JAK2^V617F-negative by ddPCR (control; n=4), JAK2-mutant MF (n=4), or JAK2-mutant PV (n=4) patients (Table S1). Flow cytometric analysis was performed to confirm PB lineage reconstitution (Fig. S3), and PDX models supported robust engraftment of human cells (Fig. 2A) in the BM (Fig. 2B). Notably, mice engrafted with PV patient cells supported significant expansion of human HSCs (mCD45− hCD45+ LINEAGE− CD34+ CD38− CD45RA− CD90+) in the BM (Fig. 2C). PDXs recapitulated classical phenotypes of the patients from which they were derived (WBC, HCT, splenomegaly; Fig. 2D–F) and served as a relevant platform to observe how different MPN subtypes support the expansion of independent TET2- and TP53-mutant clones. Targeted genomic sequencing was performed on hCD45+ cells isolated from the BM of recipient mice at the conclusion of each independent transplant (16-weeks) to quantify gRNA-mediated variant allele fraction (VAF). These data were compared to pre-transplant values to analyze the growth dynamics of parallel clones isolated from a JAK2^V617F-mutant MPN-derived PDX host in comparison to the same pool of edited cells in a HD control host. MF and PV patient cells supported the growth of TET2- and TP53-mutant clones significantly more than HD control BM across four separate experiments established with independent starting MPN patient material (Fig. 2G). Notably, cohorts established from PV patient cells exhibited the highest expansion of independent clones, particularly for TET2-mutant cells, mirroring results from the ex vivo experiment. These data highlight the robust expansion capacity of TET2- and TP53-mutant clones within JAK2^V617F-mutant MPN environments in vivo.

Fig. 2. JAK2^V617F-Mutant MPN Accelerates Expansion of Independent TET2- and TP53-Mutant Clones.

(A) Representative flow cytometry plot depicting human cell engraftment in BM of NSGS mice. (B) Quantification of overall hCD45+ BM engraftment, (C) human HSC number, (D) white blood cell count, (E) hematocrit and (F) spleen weights compiled across PDX experiments. (G) Relative fold change of CRISPR modified reads in hCD45+ BM cells 16-weeks post-transplant compared to day 0 values of tracked AAVS1, TET2, and TP53 engineered mutations in cord-blood CD34⁺ cells (n= 12 recipient mice sourced from 4 separate donors for HD, PV, and MF). **p ≤0.01, ****p ≤0.0001

Low Burden of Jak2^V617F-Mutant Cells Specifically Supports Parallel Expansion of Tet2-Mutant Clones

With the finding that independent TET2 and TP53 clones display a significant growth advantage in a JAK2^V617F-mutant environment, we established murine chimera models to provide a platform for mechanistic studies. These systems allow for precise control of the burden of each mutant cell population at the start of each transplant, facilitating a more detailed study of clone behavior and interaction than afforded by PDX models. We utilized mice with inducible expression of Jak2^V617F in the hematopoietic system (Vav-Cre;Jak2^V617F/+ = “Jak2^V617F”), which generate a PV-like phenotype, or wild-type (WT) control mice as a host background and Tet2 heterozygous loss-of-function (Vav-Cre;Tet2^fl/+ = “Tet2^Δ/+”) or Tp53 heterozygous knock-in (Tp53^R172H/+) mice as the competing test populations. 2.5×10⁶ BM cells from either Jak2^V617F or WT control mice (CD45.2) were mixed with 5×10⁵ BM cells from either Tet2^Δ/+, Tp53^R172H/+, or WT test BM (CD45.1/2) and transplanted into lethally irradiated recipients (CD45.1) to produce a starting fraction of “test” BM of approximately 15%.

Flow cytometric analysis was performed to evaluate PB lineage reconstitution (Fig. S4). By 16-weeks post-transplant, Jak2^V617F-mutant host cells significantly supported expansion of both Tet2^Δ/+ and Tp53^R172H/+ test populations in the PB compared to a WT control host background (Fig. 3A–B). However, the WT test population also exhibited an engraftment increase in the presence of Jak2^V617F-mutant cells (Fig. 3B). This engraftment advantage was largely restricted to peripheral lymphoid cells, consistent with prior reports that JAK2^V617F mutations confer lymphoid deficiency. Within the PB myeloid compartment in a Jak2^V617F-mutant background, the WT test population exhibited significantly reduced chimerism compared to Tet2^Δ/+ and Tp53^R172H/+ test populations (Fig. 3C–D). BM analysis 18-weeks post-transplant revealed a similar competitive advantage of Tet2^Δ/+ and Tp53^R172H/+ cells in a Jak2^V617F-mutant background (Fig. 3E). Interestingly, the effect was most pronounced in the hematopoietic stem/progenitor cell (HSPC; c-Kit+ Sca-1+ Lineage-“KSL”) compartment rather than the most primitive long-term hematopoietic stem cells (HSCs; KSL CD48− CD150+)(Fig. 3F–G).

Fig. 3. Low Burden of Jak2^V617F-Mutant Cells Specifically Supports Parallel Expansion of Tet2-Mutant Clones.

(A, B) Peripheral blood engraftment of WT control, Tet2^Δ/+ and Tp53^R172H/+ test cells in (A) WT or (B) a Jak2^V617F host BM. (C, D) PB myeloid (Gr-1⁺ Cd11b⁺) engraftment of WT control, Tet2^Δ/+ and Tp53^R172H/+ test cells in (C) WT or (D) Jak2^V617F host BM. (E-G) Quantification of WT, Tet2^Δ/+, Tp53^R172H/+ test cell (E) BM engraftment, (F) HSPC (Lineage⁻ Sca-1⁺ c-Kit⁺) number and (G) HSC (Lineage⁻ Sca-1⁺ c-Kit⁺ CD48⁻ CD150⁺) number in the BM of recipient mice 18-weeks post-transplant (n =10 / group). (H-J) Quantification of WT control and Tet2^Δ/+ cell (H) PB engraftment, (I) BM engraftment and (J) HSPC numbers in recipient mice transplanted with indicated fractions of Jak2^V617F BM input cells (n = 5 / group). (K-N) Quantification of WT control, Tet2^Δ/+, Tp53^R172H/+ test cell (K) PB engraftment, (L) BM engraftment, (M) HSPC numbers and (N) HSC numbers in recipient mice transplanted with 50% Jak2^V617F BM input (n= 5–9 / group). *p ≤0.05, **p ≤0.01, ***p ≤0.001, ****p ≤0.0001

As Tet2-mutant clones demonstrated an enhanced growth advantage in a Jak2^V617F-environment compared to Tp53-mutant clones in both PDX and murine chimeras, we aimed to determine the minimum Jak2^V617F-mutant cell burden required to support parallel Tet2-mutant clone expansion. A titration experiment was performed wherein BM cells from Jak2^V617F mice were diluted at predetermined ratios with WT BM while maintaining a fixed Tet2^Δ/+ cell input (15%). There was a threshold effect between starting Jak2^V617F cell input and the conferred competitive advantage of Tet2^Δ/+ cells with a ~35% Jak2^V617F mutant cell burden being the threshold needed to support robust expansion of Tet2^Δ/+ cells (Fig. 3H–J).

Based on these data, we established chimeras with 50% Jak2^V617F-mutant BM, 35% WT support BM, and 15% Tet2^Δ/+, Tp53^R172H/+ or WT test BM to determine if this was also sufficient to induce outgrowth of parallel Tp53-mutant cells. As anticipated, a 50% starting Jak2-input supported positive selection of Tet2^Δ/+ cells in the PB, BM, and HSPC compartments. However, this Jak2^V617F-mutant burden did not support engraftment of Tp53^R172H/+ or WT cells unlike the higher (85%) Jak2^V617F-mutant cell input (Fig. 3K–M). These findings highlight a specific interaction between Jak2^V617F-mutant MPN cells and independent Tet2-mutant clones.

IL-12 and TNFα Drive Expansion of Tet2-Mutant Clones in a Jak2^V617F-Mutant Environment

As our data show a particularly strong selection of independent Tet2-mutant clones in an MPN background, we focused mechanistic studies to define this relationship. As the selective advantage for Tet2^Δ/+ cells in and MPN background was most evident at the level of HSPCs, gene expression profiling was performed by RNA-sequencing analysis of Tet2^Δ/+ HSPCs (CD45.1/2+ c-Kit+ Sca-1+ Lineage-) isolated from either WT or Jak2^V617F-mutant hosts 18-weeks post-transplant. Analysis of differentially expressed genes (DEGs; Fig. 4A) demonstrated that Tet2^Δ/+ HSPCs from a Jak2^V617F-mutant environment exhibited a pronounced proliferation bias, exemplified by upregulation of Mki67, and increased expression of pro-myeloid differentiation genes such as Elane, Mpo, and Cebpe. The inflammatory response was also heightened with elevated levels of S100a8 and S100a9 which function as key promoters of myeloid inflammation and are routinely upregulated in HSPCs in the contexts of myeloid skewing and chronic inflammatory states(23). This myeloid differentiation bias occurred at the expense of lymphoid priming marked by decreased expression of Dntt and Lck, essential genes for lymphoid lineage commitment. Genes classically associated with HSC identity, such as Hoxa10, Fgd5, Vldlr, Hlf, and Mecom, were also downregulated in Tet2^Δ/+ HSPCs from a Jak2^V617F-mutant environment (Fig. 4B), suggesting these cells were being pushed towards proliferation at the expense of self-renewal. Collectively, these findings indicate that a Jak2^V617F-mutant environment drives Tet2-mutant HSPCs toward a proliferative, pro-myeloid, inflammatory phenotype. We next interrogated whether the gene expression changes observed in Tet2^Δ/+ HSPCs from a Jak2^V617F-mutant environment were due to specific effects on Tet2^Δ/+ cells and not a general effect for any HSPCs exposed to MPN cells by RNA-seq comparison of Tet2^Δ/+ and WT HSPCs isolated from a Jak2^V617F-mutant environment. Differential expression analysis confirmed that the transcriptomic alterations previously identified in Tet2^Δ/+ HSPCs remained dysregulated in Tet2^Δ/+ HSPCs compared to WT HSPCs from a Jak2^V617F-mutant environment (Fig. S5). These findings demonstrate that the identified gene dysregulation is unique to Tet2-mutant cells within a Jak2^V617F-mutant milieu and not a generalized feature of HSPCs exposed to MPN cells.

Fig. 4. IL-12 and TNFα Drive Expansion of Tet2-Mutant Clones in a Jak2^V617F-Mutant Environment.

(A) Volcano plot showing differentially expressed genes between Tet2^Δ/+ HSPCs from a Jak2^V617F environment compared to Tet2^Δ/+ HSPCs from a WT environment. (B) LogFC values for Tet2^Δ/+ HSPCs DEGs from indicated pathways. (C) Volcano plot of gene set over-representation analysis (ORA) identifying IL-12 signaling as the most significantly enriched pathway in Tet2^Δ/+ HSPCs from a Jak2^V617F environment. (D-E) Serum cytokine levels (normalized to WT control chimeras) of (D) IL-12 and (E) TNFα from murine mixed chimeras (n= 9–15 / group). (F) Serum cytokine fold change (normalized to HD control mice) for PV and MF PDXs plotted against relative increase in VAF of individually tracked TET2 mutations engineered into cord blood CD34⁺ cells (n= 12 recipient mice sourced from 4 separate donors for each HD, PV, and MF). (G) Representative flow cytometry plots showing intracellular pSTAT4 staining in Tet2^Δ/+ BM cells in WT or Jak2^V617F backgrounds. (H-I) Median fluorescent intensity (MFI) of intracellular pSTAT4 levels in (H) whole BM and (I) HSPCs (n = 5 / group). *p ≤0.05, **p ≤0.01, ***p ≤0.001, ****p ≤0.0001

To discern the most dysregulated signaling pathways that may be driving the observed transcriptional changes, over-representation analysis (ORA) was performed on the identified DEGs. ORA identified IL-12-mediated signaling as the most significantly dysregulated cancer-related pathway between Tet2-mutant HSPCs isolated from a Jak2^V617F-mutant background compared to from a WT background (Fig. 4C). To further classify over-represented pathways, DEGs were quantitatively scored based on how frequently they appear in a cancer-related pathway relative to their frequency across all cancer-related pathways and used to weight gene set terms by their relevance across the dataset. Uniform manifold approximation and projection (UMAP) was then applied for dimensionality reduction to visualize relationships between enriched pathways. Again, among the identified clusters, IL-12-mediated signaling emerged as the most highly enriched pathway. Adjacent points in the UMAP space, inferring biological relevance of over-represented gene sets, included TNF receptor signaling and IL-12 signaling mediated by STAT4, highlighting these related signaling pathways as potential selection factors for Tet2^Δ/+ HSPC expansion in a Jak2^V617F-mutant environment (Fig. S6).

It is well-documented that MPN cells secrete high levels of many pro-inflammatory cytokines(24–27). Moreover, we and others have shown that many common CH mutations impart growth advantages to the mutant clones under different conditions of inflammation(28–32). To determine if the observed gene expression changes in Tet2^Δ/+ HSPCs in a Jak2^V617F-mutant environment were associated with altered cytokine levels due to the MPN cells, global cytokine profiling of PB serum from murine chimera and PDX experiments was performed. The overlap of cytokines increased in MPN models compared to relevant control comparators revealed four candidates - IL-1, IL-12, IL-27, and TNFα (Fig. S7A). We sought to determine if any of these cytokines might create a competitive advantage for TET2-mutant clones. CB-derived CD34⁺ cells were CRISPR/Cas9-engineered with TET2 mutations and cultured with each candidate cytokine at concentrations ranging from 1–100 ng/mL for 6-days. TNFα and IL-12 were the most effective in accelerating TET2-mutant cell proliferation in vitro (Fig. S7B). Moreover, from in vivo studies, increasing IL-12 and TNFα levels correlated with TET2-mutant clonal expansion in murine (Fig. 4D,E) and PDX transplants (Fig. 4F). This effect was most pronounced in PDX mice established from PV patient cells (Fig. 4F). To determine the source of the aberrant signaling, IL-12 and TNFα levels were measured in serum from transplant donor mice of each genotype. IL-12 levels were significantly elevated in Jak2^V617F-mutant mice compared to WT and Tet2^Δ/+ donor mice. Conversely, TNFα was significantly elevated in Tet2^Δ/+ donor mice (Fig. S8). Notably, TNFα has been reported to correlate with TET2-mutant CH and IL-12 is a known TNFα stimulator that is elevated in MPN patients(33–38). This suggests a model whereby IL-12 secreted by MPN cells induces TNFα over-production by Tet2-mutant cells to condition an environment that fosters their development.

We hypothesized that IL-12 secreted by MPN cells directly acts upon TET2-mutant cells to phosphorylate STAT4, a key mediator of IL-12 signaling(39–41). Intracellular flow cytometry analysis (Fig. 4G) revealed that pSTAT4 levels were elevated in Tet2^Δ/+ BM cells isolated from a Jak2^V617F-mutant background 10-weeks post-transplant compared to those isolated from a WT background or from WT cells from either background (Fig. 4H). Strikingly, the pSTAT4 elevation was most significant in the Tet2^Δ/+ HSPC population isolated from a Jak2^V617F-mutant environment (Fig. 4I). This indicates a potential mechanistic link between IL-12 signaling and Tet2^Δ/+ HSPC proliferation in a Jak2^V617F-mutant context, suggesting that targeting this pathway could mitigate the competitive advantage in this genetic setting.

Inhibition of Inflammatory Cytokines Mitigates the Competitive Advantage of TET2-Mutant Cells in a Jak2-Mutant Environment

With the identification of IL-12 and TNFα as potential selective pressures for independent Tet2-mutant clone expansion in parallel to MPN, we aimed to evaluate the functional consequences of disrupting these pathways. To enhance the clinical relevance, we replicated mutational burdens that more closely mirror those observed in MPN patients with input thresholds determined by the Jak2^V617F titration experiment. As such, chimeras utilized in functional studies were established by transplanting 1.5×10⁶ Jak2^V617F BM cells (CD45.2) with 4.5×10⁵ test BM (CD45.1/2) and 1.05×10⁶ WT support cells (CD45.1) into lethally irradiated recipients (CD45.1) to produce a starting fraction of test BM of approximately 15% in an environment composed of approximately 50% Jak2^V617F-mutant host BM. First, to assess the implications of TNFα on the observed competitive advantage of Tet2-mutant clones, Vav-Cre;Tet2^fl/+ mice were crossed with a TNFα-receptor genetic deletion model (p55/p75 germline knockout = “KO”) to establish mice wherein Tet2-mutant hematopoietic cells lack receptors for TNFα. To determine the effect of ameliorating TNFα signaling in Tet2-mutant cells, BM cells from either Jak2^V617F or WT mice were mixed with either Tet2^Δ/+, TNFαrp55^−/−p75^−/− (= “TNFr^KO”), Tet2^Δ/+/TNFr^KO, or WT test BM and co-transplanted with WT support. Blood (Fig. 5A–C) and BM analysis (Fig. 5D) revealed the competitive advantage of Tet2-mutant cells in a Jak2^V617F-mutant environment was significantly blunted by genetic deletion of TNFα receptors on the Tet2-mutant cells.

Fig. 5. Inhibition of Inflammatory Cytokines Mitigates the Competitive Advantage of TET2-Mutant Cells in a Jak2-Mutant Environment.

(A, B) Peripheral blood engraftment of WT control, Tet2^Δ/+, TNFr^KO and Tet2^Δ/+;TNFr^KO, test cells in (A) WT or (B) a Jak2^V617F host BM (n= 5 / group). (C) 16-week peripheral blood engraftment of Tet2^Δ/+ and Tet2^Δ/+, TNFr^KO test cells in WT or Jak2^V617F host BM (n= 5 / group). (D) 18-week BM engraftment of Tet2^Δ/+ and Tet2^Δ/+, TNFr^KO test cells in WT or Jak2^V617F host BM (n= 5 / group). (E-H) Tet2^Δ/+ test cell (E) 16-week PB engraftment, (F) 18-week BM engraftment, (G) HSPC number and (H) HSC number chimeric mice receiving neutralizing antibodies against IL-12 and/or TNFα (n= 4–5 / group). (I) qPCR gene expression analysis of Tet2^Δ/+ HSPCs from indicated treatments of genetic models depicting levels of indicated genes (normalized to non-treated controls; n = 2–3 / group). (J-L) Analysis of (J) spleen weight, (K) hematocrit and (L) modified read fold change for CRISPR engineered TET2 mutations in cord blood CD34⁺ cells in treatment-control and IL-12-neutralizer treated PDX mice co-transplanted with PV UPN:702759 CD34⁺ cells (n= 2–3 / group). (M) Schematic depicting mechanism of PB cells providing a selective advantage to independent TET2-mutant clones through cytokine support. *p ≤0.05, **p ≤0.01, ***p ≤0.001, ****p ≤0.0001

To evaluate this finding in a more translational system, we administered murine biosimilars of TNFα and IL-12 monoclonal neutralizing antibodies (adalimumab / ustekinumab respectively) into mouse chimeric models between weeks 4–10 post-transplant. Inhibition of TNFα and IL-12 dramatically reduced Tet2^Δ/+ cell engraftment in a Jak2^V617F-mutant background in the PB (Fig. 5E), BM (Fig. 5F) and HSPC populations (Fig. 5G). Notably, neutralization of IL-12 and TNFα suppressed Tet2^Δ/+ cells down to the HSC level – an effect most significantly achieved by IL-12 neutralization (Fig 5H). Gene expression analysis of Tet2^Δ/+ HSPCs isolated from a Jak2^V617F-mutant background at the end of the 18-week study from control and treated cohorts showed targeting the IL-12/TNFα axis normalized dysregulation in proliferation (Mki67) and self-renewal (Hoxa10, Hlf, Fgd5, and Mecom) pathways. Similar normalization of these abnormal gene expression programs was also observed in Tet2^Δ/+ HSPCs genetically deficient for TNFα receptors (Fig. 5I). Thus, genetic and pharmacological inhibition of aberrant MPN cytokine signaling abrogates the competitive advantage of Tet2^Δ/+ HSPCs in the presence of Jak2^V617F-mutant cells. To evaluate the effect of IL-12 neutralization in a human system, we utilized a JAK2^V617F-mutant PV patient sample that previously demonstrated support of parallel TET2-mutant clones (UPN:702759). 1×10⁵ CD34⁺ PV patient cells were co-transplanted with 2.0×10⁴ CB-derived CD34⁺ cells harboring gRNA-mediated TET2 mutations. Between weeks 10–16 post-transplant, a human IL-12 neutralizing agent was administered to the treatment group. Control mice exhibited signs of disease progression characterized by increased spleen weight and decreased hematocrit, while mice receiving the IL-12 neutralizing agent retained PV-like pathologies (Fig. 5J–K). Strikingly, IL-12 neutralization was able to mitigate the competitive advantage of TET2-mutant clones in the presence of PV patient cells (Fig. 5L). These findings reveal that targeting IL-12 could be a potential therapeutic approach for restricting TET2-mutant CH in the background of an existing JAK2^V617F-mutant driven MPN to minimize future risk of disease progression to pAML (Fig. 5M).

Conclusion

These data establish that JAK2^V617F-mutant cells potentiate parallel expansion of independent clones as a non-classical trajectory of disease progression in MPN. We show that MPN cells drive this parallel evolution through an IL-12/TNFα cytokine axis, which biases Tet2-mutant progenitors towards increased proliferation and myeloid differentiation. Genetic and pharmacological inhibition of IL-12 and TNFα resulted in both functional and molecular rescue of these phenotypes. Our findings provide crucial insights into how sAML which lacks the JAK2^V617F driver mutation can evolve from an antecedent JAK2^V617F-mutant MPN. These results highlight a specific example of clonal dynamics and cell competition within a JAK2^V617F-mutant context; however, it is important to acknowledge that clones with other CH mutations (e.g. DNMT3A) as well as MPNs driven by other mutations (CALR, MPL) may confer different patterns of clonal evolution. We aim to leverage these findings to enhance disease surveillance in MPN populations. With improved genomic monitoring, targeted interventional therapies to prevent the expansion of emerging pre-leukemic clones could reduce risk of parallel disease evolution in MPN, a significant therapeutic advantage for a substantial subset of patients.

Data and materials availability:

Data not available in the main text or the supplementary materials are available through request to the corresponding author. All raw read data (FASTQ files) for RNA sequencing are publicly available at the Gene Expression Omnibus database accession number GSE308233.