Genetic studies reveal an unexpected negative regulatory role for Jak2 in thrombopoiesis

Myeloproliferative neoplasms (MPN) are characterized by clonal proliferation of mature myeloid cells. Somatic mutations in the JAK2 pathway, including JAK2V617F,^1-4 JAK2 exon 12 mutations,⁵  and in the thrombopoietin (TPO) receptor MPL⁶  are observed in MPN patients. JAK kinase inhibitors attenuate splenomegaly and constitutional symptoms in MPN,^7,8  but cannot induce molecular responses. JAK inhibitor therapy results in the variable development of anemia and thrombocytopenia preventing dose intensification and maximal target inhibition.⁹  Cytopenias have been considered on-target effects of JAK2 inhibitors inseparable from therapeutic efficacy. However, the risk of thrombocytopenia varies among JAK2 inhibitors,¹⁰  suggesting thrombocytopenia may not be a direct consequence of JAK2 inhibition.

JAK2 mediates cytokine signaling from erythropoietin, TPO, and granulocyte-macrophage (GM)–colony-stimulating factor receptors.¹¹  Germline deletion of Jak2 causes embryonic lethality by deficient definitive erythropoiesis,^12,13  which was confirmed by a conditional Jak2 allele and germline Cre-recombinase.¹⁴  Conditional Jak2 loss in hematopoietic stem cells (HSCs)/progenitor cells induces anemia and thrombocytopenia,^15,16  whereas deletion of Mpl in platelets (PLTs) leads to thrombocytosis.^17-19  Although these studies highlight the significance of Jak-Stat signaling for steady state hematopoiesis, the specific requirement for Jak2 in different terminally differentiated myeloid lineages, including PLTs, has not been clarified. We, therefore, investigated the effects of PLT-specific Jak2 loss in vivo.

Jak2^f/f mice¹⁴  were crossed with Pf4-Cre mice.²⁰  Animal care was performed according to institutional and international guidelines.

PLTs were purified on a sepharose CL-2B column, and stimulated with TPO and lysates probed for Jak2, Stat3/5, Tyk2, β-actin (Cell Signaling), and Mpl (Wei Tong, Philadelphia).

Bone marrow was stained for lineage markers, Sca-1, c-Kit, CD41, CD150, CD48, CD16/32, and CD105²¹  (eBioscience), and PLTs for CD42, CD61, and Mpl, and analyzed/sorted on FACSAria III (BD Biosciences). Megakaryocytes (MKs) were purified by CD41-phycoerythrin-magnetic beads (Stemcell Technologies).¹⁸ 

RNA was isolated in TRIzol (Life Technologies) and reverse transcribed by the Verso cDNA Kit (Thermo) for quantitative polymerase chain reaction with SYBR Green (Applied Biosystems) and β-actin for normalization.

Acetylcholinesterase-positive MK colonies were grown from 0.5 × 10⁵ bone marrow cells or 3000 megakaryocyte-erythroid progenitors (Pre-MegE), MK progenitors (MKPs), or Lin⁻Sca1⁺Kit⁺ (LSK) cells in MegaCult-C (Stemcell Technologies) with TPO, IL-6, and IL-3 (R&D). Myeloid colonies from 4000 Pre-MegE, MKP, or 200 LSK cells were cultured in MethoCult (Stemcell Technologies) with stem cell factor, IL-3, IL-6, and erythropoietin.

TPO was determined by Quantikine enzyme-linked immunosorbent assay (R&D) in serum, or in media²²  incubated with purified PLTs.

To assess the requirement for Jak2 in PLTs, we deleted Jak2 in PLTs and MKs using Cre-recombinase expressed under the Pf4 promoter.^14,19,20,23  Pf4-Cre–positive Jak2^f/f mice showed loss of Jak2 messenger RNA (mRNA) expression in PLTs and MKs demonstrating complete excision, and we observed a modest reduction in Jak2 mRNA expression in MKPs (P = .029) (Figure 1A). Jak2 mRNA expression was unaffected in Pre-MegE or LSK cells (n = 3 to 6 mice/group). Western blot analysis revealed loss of Jak2 protein expression in Pf4-Cre–positive Jak2^f/f PLTs, despite continued Mpl expression (Figure 1B-C). Jak2 deletion in PLTs abrogated Jak2, Stat3, Stat5, and Tyk2 phosphorylation upon TPO stimulation (Figure 1D). The absence of residual Tyk2 phosphorylation was consistent with cell line studies, suggesting that TPO-mediated activation of Tyk2 requires Jak2.²⁴ 

We next investigated the impact of PLT-specific Jak2 deletion in vivo. Pf4-Cre–positive Jak2^f/f mice developed thrombocytosis with average PLTs of 4720 G/l (range 3436 to 6460 G/l, P < .0001) at 2 months and 5966 G/l (3204 to 7892 G/l, P < .0001) at 5 months of age (n = 5 to 17 mice/group;Figure 1E). PLT survival assessed by in vivo biotinylation and PLT volume were normal in Pf4-Cre–positive Jak2^f/f mice compared with littermate controls (Figure 1E). PLT-specific Jak2 loss resulted in leukocytosis with a mean white blood cell (WBC) count of 23.7 G/l (vs 11.9 G/l in controls, P = .005) with proportionate expansion of all leukocyte subsets (Figure 1F-G). Hematocrit levels were not significantly affected. PLT-specific Jak2 haploinsufficiency had no impact on blood counts compared with controls. These data demonstrate that Jak2 expression and activity in PLTs are not required for terminal PLT production.

The thrombocytosis observed in Pf4-Cre–positive Jak2^f/f mice suggested PLT/MK-specific Jak2 deletion led to non-autonomous expansion of Jak2-expressing stem/progenitor cells. Pf4-Cre–positive Jak2^f/f mice had expansion of Pre-MegE (209% of controls, P < .0001), MKP (932% of controls, P = .0001), and MKs (606% of controls, P < .0001) (n = 3 to 12 mice/group;Figure 2A,D). Histopathologic analysis confirmed bone marrow megakaryocytosis (Figure 2B). MK colony formation from Pf4-Cre–positive Jak2^f/f bone marrow and sorted Pre-MegE or MKP was enhanced (Figure 2C). We observed an increase in colony-forming unit (CFU)-GEMM and CFU-GM colonies but not BFU-E colonies from Pf4-Cre–positive Jak2^f/f mice (see supplemental Figure 1A-B, available on the Blood Web site). These data demonstrate that PLT/MK-specific Jak2 loss induces expansion of megakaryopoiesis in vivo.

We next assessed the impact of PLT-specific Jak2 loss on HSC/progenitor cells in vivo. We observed a significant expansion in LSK number (567% of controls, P < .0001), (n = 5 to 11 mice/group; Figure 2E). LSK cells from Pf4-Cre–positive Jak2^f/f mice showed enhanced CFU-GEMM and CFU-GM colony formation (supplemental Figure 1C) and marked expansion in megakaryocytic colony formation on a per-cell basis (Figure 2F). This suggested the possibility that Pf4-Cre–positive Jak2^f/f mice had selective expansion of MK-“primed” stem cells.²⁵  Consistent with recent studies showing MK bias of c-Kit^hi HSC,²⁶  Pf4-Cre–positive Jak2^f/f mice had significant expansion of c-Kit^hiLin⁻Sca1⁺Kit⁺CD150⁺CD48⁻ HSC (Figure 2G-H). These data demonstrate PLT-specific Jak2 loss induces expansion of MK-biased stem/progenitors.

Previous studies have shown that PLTs and MKs regulate serum TPO through internalization attenuating TPO levels in thrombocytosis.^27,28  We hypothesized that Jak2 loss in PLTs and MKs would compromise TPO internalization and maintain high steady state TPO despite thrombocytosis. We assessed TPO levels in Pf4-Cre–positive Jak2^f/f mice, controls, and mice expressing MPLW515L,²⁹  which develop a similar extent of thrombocytosis and MK expansion as Pf4-Cre–positive Jak2^f/f mice. Serum TPO was markedly reduced in MPLW515L mice (TPO 230^+/− 37 pg/mL in MPLW515L vs 2787^+/− 204 pg/mL in controls, P < .0001), but not in Pf4-Cre–positive Jak2^f/f mice (TPO 2524^+/− 230 pg/mL, P < .0001) (n = 8 to 13 mice/group;Figure 2I). Jak2-deficient PLTs internalized less TPO than PLTs from Pf4-Cre–negative Jak2^f/f controls (ΔTPO 884 pg/mL vs 1051 pg/mL, P < .01; Figure 2K). Mpl mRNA expression was unchanged in MKs of Pf4-Cre–positive Jak2^f/f mice (Figure 1A), while Mpl total and surface protein was modestly reduced in Jak2-deficient PLTs (Figures 1C and 2J). These data suggest that TPO levels are maintained at an inappropriately high rheostat in Pf4-Cre–positive Jak2^f/f mice due to reduced TPO internalization by Jak2-deficient PLTs.

In conclusion, our data demonstrate that Jak2 is dispensable for the production of MKs and PLTs. This finding suggests that thrombocytopenia observed with certain JAK2 inhibitors is due to JAK2 inhibition in more proximal stem/progenitor cells consistent with conditional Jak2 knockout studies in hematopoiesis,^15,16  or potentially due to inhibition of other targets. Analogous to genetic Jak2 loss, Jak2 inhibition in PLTs/MKs may actually enhance PLT production,³⁰  but may be overruled by the impact of Jak2 inhibition on stem/progenitors. Complementary to previous studies on the role of Mpl in PLTs,^17-19  our results demonstrate an important role for PLT-specific Jak2 signaling in regulating TPO turnover, and show that PLT Jak2 signaling negatively regulates stem/progenitor cells, including MK-primed HSCs. Notably, HSCs biased toward megakaryopoiesis have been shown to express higher levels of MK-specific genes, including Mpl.^25,26  Increased Mpl in MK-primed HSCs might enhance susceptibility to TPO in Pf4-Cre–positive Jak2^f/f mice and contribute to thrombocytosis. Recent studies further suggest a role for MKs in regulating HSCs and have shown that PLT secretory factors can impact HSC numbers.³¹  Taken together, these data underscore the dynamic interaction between PLTs, MKPs, and stem/progenitors, and suggest that signaling in mature myeloid lineages can alter stem/progenitor function in vivo.

The authors thank the members of the Levine Laboratory, Omar Abdel-Wahab, Christopher Park, Barry Coller, and Lorena Buitrago for helpful comments and discussion. The authors also thank P. Jan Hendrikx, Tom Baumgartner, Christi O’Donnell, and Paul Byrne from the Flow Cytometry Core Facility of Memorial Sloan Kettering Cancer Center for their support in sorting megakaryocyte progenitors, as well as Wei Tong (The Children’s Hospital of Philadelphia) for the kind gift of anti-Mpl antibody, and Kay-Uwe Wagner for Jak2^f/f mice.

This study was supported by grants from the National Institutes of Health, National Cancer Institute (R01 CA151949-05) (R.L.L.), the Swiss National Science Foundation, the Swiss Cancer League, and the Huggenberger-Bischoff Foundation for Cancer Research (S.C.M.). R.L.L. is a Leukemia and Lymphoma Society Scholar.

Contribution: S.C.M., M.D.K., B.A.W., L.M.L., L.B., M.K., N.B., and S.M. performed experiments; and S.C.M. and R.L.L. designed research, analyzed data, and wrote and edited the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Ross L. Levine, Human Oncology and Pathogenesis Program, Leukemia Service, Department of Medicine, Memorial Sloan Kettering Cancer Center, 1275 York Ave, Box 20, New York, NY 10065; e-mail: leviner@mskcc.org.