Myeloproliferative neoplasm-driving Calr frameshift promotes the development of pulmonary hypertension in mice

Abstract

Frameshifts in the Calreticulin (CALR) exon 9 provide a recurrent driver mutation of essential thrombocythemia (ET) and primary myelofibrosis among myeloproliferative neoplasms (MPNs). Here, we generated knock-in mice with murine Calr exon 9 mimicking the human CALR mutations, using the CRISPR-Cas9 method. Knock-in mice with del10 [Calr^{del10/WT (wild−type)} mice] exhibited an ET phenotype with increases of peripheral blood (PB) platelets and leukocytes, and accumulation of megakaryocytes in bone marrow (BM), while those with ins2 (Calr^ins2/WT mice) showed a slight splenic enlargement. Phosphorylated STAT3 (pSTAT3) was upregulated in BM cells of both knock-in mice. In BM transplantation (BMT) recipients from Calr^del10/WT mice, although PB cell counts were not different from those in BMT recipients from Calr^WT/WT mice, Calr^del10/WT BM-derived macrophages exhibited elevations of pSTAT3 and Endothelin-1 levels. Strikingly, BMT recipients from Calr^del10/WT mice developed more severe pulmonary hypertension (PH)—which often arises as a comorbidity in patients with MPNs—than BMT recipients from Calr^WT/WT mice, with pulmonary arterial remodeling accompanied by an accumulation of donor-derived macrophages in response to chronic hypoxia. In conclusion, our murine model with the frameshifted murine Calr presented an ET phenotype analogous to human MPNs in molecular mechanisms and cardiovascular complications such as PH.

To the editor,

CALR frameshifts provide a recurrent myeloproliferative neoplasm (MPN) driver [1]. Pulmonary hypertension (PH) is a life-threatening cardiopulmonary disease characterized by increased pulmonary arterial (PA) pressure. Bone marrow (BM)-derived cells and perivascular inflammatory infiltrates contribute to PA remodeling in PH [2, 3]. Among 5 etiological groups, the WHO group-V PH encompasses multifactorial mechanisms, including MPNs, which are often complicated by PH, with 5%-60% of the prevalence [4,5,6]. MPN-related PH is associated with crucial features, such as thromboembolism and hypermetabolic state [5]. However, the association of PH with CALR mutation remains uncertain. Here, we generated Calr^del10/WT and Calr^ins2/WT knock-in mice (Fig. 1a, Additional file 1: Fig. S1), investigated their hematopoiesis, and clarified the role of hematopoietic Calr mutation in PH using BM transplantation (BMT) and chronic hypoxia, which provokes PH [7].

Fig. 1

Hematopoietic cells with Calr mutation exacerbate the development of pulmonary hypertension in response to chronic hypoxia. a The knock-in mice with C57BL/6 J background carrying frameshifted murine Calr, del10 (Calr^del10/WT mice) and ins2 (Calr^ins2/WT mice) were generated using the CRISPR-Cas9 method. Structure of wild-type (WT) and frameshifted murine CALR proteins are shown. Both generated mutant proteins with shortened calcium-buffering sites and absent KDEL sequence, which is the signal to retain the CALR protein in the endoplasmic reticulum. b Leukocyte (white blood cell) counts (WBC), red blood cell counts (RBC), and platelet counts (PLT) in WT mice (Calr^WT/WT mice, n = 21), Calr^ins2/WT mice (n = 17), and Calr^del10/WT mice (n = 16) in the peripheral blood. ^*P < 0.05 versus the WT group. c Schematic diagram of the experimental design of bone marrow (BM) transplantation (BMT). BM cells from control Calr^WT/WT mice or Calr^del10/WT mice were injected into the lethally irradiated WT mice (C57BL/6 J mice). Four weeks after BMT, the recipient mice transplanted with the BM cells from the Calr^WT/WT mice (WT-R) or Calr^del10/WT mice (del-R) were subjected to normoxia (21% O₂) or chronic hypoxia (10% O₂) for 3 weeks. d Allele frequency of the mutant Calr in the peripheral leukocytes of recipient mice at 4 weeks after BMT (n = 15, each). e Right ventricular (RV) systolic pressure (RVSP) and RV hypertrophy determined by dividing the RV weight by the left ventricular weight including the septum (RV/LV + S) (n = 6–8). f Representative hematoxylin–eosin (HE) staining and immunohistochemistry with antibodies to anti-α smooth muscle actin (αSMA) and anti-F4/80 images in the lung of BMT recipient mice from Calr^WT/WT or Calr^del10/WT mice. Scale bars, 50 µm. g Quantitative analysis of the percentage of muscularized distal pulmonary arteries in αSMA-immunostained sections (n = 3, each). h Quantitative analysis of the pulmonary perivascular macrophages determined as F4/80-positive cells, per 30 vessels (n = 5, each). Data are presented as means ± SEM. d, e, g, h ^*P < 0.05 versus the corresponding normoxia group and ^†P < 0.05 versus the corresponding BMT recipient mice from Calr^WT/WT mice. WT-R, recipient mice transplanted with BM cells from Calr^WT/WT mice; del-R, recipient mice transplanted with BM cells from Calr^del10/WT mice. Oligonucleotides and antibodies used are listed in Additional files 8, 9

In a public database [8], 138 CALR frameshifts, including the major del52 and ins5 [1], and another frameshift in 2 MPN patients (ET, myelofibrosis) exactly matching the Calr-del10, have been noted in patients with hematopoietic cancers, mostly MPNs (Additional file 7: Table S1, Additional file 1: Fig. S1f). Mouse models carrying frameshifted CALR showed ET or, rarely, myelofibrosis [9]. Likewise, Calr^del10/WT mice developed ET with phosphorylated STAT3 (pSTAT3) and cell-surface thrombopoietin-receptor (TpoR) expressions suggested to accompany mutant CALR [10], whereas Calr^ins2/WT mice showed a slight splenic enlargement (Fig. 1b, Additional files 2, 3: Fig. S2-S3).

To elucidate the roles of hematopoietic Calr mutation in PH, we performed non-competitive BMT from Calr^del10/WT mice (Fig. 1c), as we reconstituted Jak2V617F⁺ MPNs [11]. At 4 weeks after BMT, the engraftments were achieved in the BMT recipients from Calr^del10/WT mice (del-R, Fig. 1d), but their PB cell counts (Additional files 4: Fig. S4) and BM megakaryocytic distribution did not differ from BMT recipients from Calr^WT/WT mice (WT-R). We assessed right heart hemodynamics and right ventricular (RV) hypertrophy, showing that neither RV systolic pressure (RVSP) nor right ventricle/left ventricle-plus-septum weight ratio (RV/LV + S) differed between WT-R and del-R. Subsequently, del-R were exposed to chronic hypoxia (10% O₂) for 3 weeks. Strikingly, although chronic hypoxia elevated RVSP and RV/LV + S in both WT-R and del-R, these levels in del-R were significantly greater than in WT-R, suggesting that hematopoietic Calr mutation promotes PH (Fig. 1c-e).

Lung histology showed significant increases in PA medial wall thickness and muscularization, indicated by α smooth muscle actin, without thrombosis in del-R compared to WT-R under chronic hypoxia, whereas F4/80⁺ macrophages rather than TpoR⁺ cells were increased specifically in PA regions in both WT-R and del-R (Fig. 1f-h, Additional file 3: Fig. S3e). However, pSTAT3 levels were elevated in the lungs of del-R compared to WT-R after chronic hypoxia. The expression of Endothelin-1, an important vasoactive peptide involving PA remodeling in PH [4, 5], was also increased in the lungs of del-R compared to WT-R under chronic hypoxia (Fig. 2a, b, Additional file 5: Fig. S5). We visualized the Calr^del10/WT BM-derived cells using CAG-EGFP: in the lungs of BMT recipients from Calr^del10/WT/CAG-EGFP mice, donor-derived macrophages accumulated in PA regions, but donor-derived cells were not observed in vascular walls (Fig. 2c), suggesting that Calr^del10/WT BM-derived macrophages migrated into the PA regions.

Fig. 2

STAT3 phosphorylation and Endothelin-1 expression in the lung and macrophages from Calr^del10/WT mice. a Western blot of lung homogenates of the BMT recipients from Calr^WT/WT mice (WT-R) or Calr^del10/WT mice (del-R), immunoblotted with the indicated antibodies. b Phosphorylated STAT3 (p-STAT3) to total STAT3 (t-STAT3) or Endothelin-1 to β-actin ratios are shown in the graphs. The average value for WT-R under normoxia was set to 1 (n = 5, each). ^*P < 0.05 versus the corresponding normoxia group and ^†P < 0.05 versus the corresponding WT-R. c The lethally irradiated WT C57BL/6 J mice were transplanted with the BM cells from Calr^del10/WT/CAG-EGFP mice. These recipient mice were subjected to chronic hypoxia for 3 weeks, and then the lungs were fixed and stained with the indicated antibodies. Upper images show representative immunofluorescence of the lung sections stained with anti-GFP (green) and anti-αSMA (red) antibodies and DAPI (blue). Scale bars, 50 µm. Lower images show representative immunofluorescence of the lung sections stained with anti-GFP (green) and anti-F4/80 (red) antibodies and DAPI (blue). Scale bars, 10 µm. d-g BM mononuclear cells isolated from the Calr^WT/WT or Calr^del10/WT mice were cultured in the presence of 10 ng/mL of M-CSF for 6 days. d Representative immunofluorescence images of the cells stained with anti-F4/80 (green) and DAPI (blue) are shown. More than 90% of cells were macrophages expressing F4/80. Scale bars, 25 µm. e Dot plot of flow cytometry for cultured macrophages. Red, blue, and orange dots represent cells from Calr^WT/WT mice, Calr^del10/WT mice, and negative control (mixture of WT and Calr del10 cells), respectively. Over 90% WT and del10 cells were positive for both F4/80 and CD68. SSC indicates side scatter. f The cultured macrophages were then stimulated with 0.05 µg/mL of lipopolysaccharide (LPS), a potent activator of macrophages. The mRNA expression levels of Endothelin-1 (Edn1) were analyzed at the indicated time (n = 8, each). Actb was used for normalization. The average value for the macrophages from Calr^WT/WT mice at baseline was set to 1. g Left panels show western blots on STAT3, Endothelin-1, and β-actin in the macrophages stimulated with 0.05 µg/mL of LPS. Right graphs show phosphorylated STAT3 (p-STAT3) to total STAT3 (t-STAT3) or Endothelin-1 to β-actin ratios at the indicated time. The average value for the macrophages from Calr^WT/WT mice at the baseline was set to 1 (n = 4, each). All data are presented as means ± SEM. ^*P < 0.05 versus the corresponding WT group. WT, macrophages derived from the Calr^WT/WT mice; del10, macrophages from the Calr^del10/WT mice. Oligonucleotides and antibodies used are listed in Additional files 8, 9

RNA sequencing of hematopoietic progenitors showed Calr-del10 activated JAK-STAT pathway, as well as cardiac-hypertrophy pathway that includes upregulation of Endothelin-1. Also, human CALR-del52 introduction upregulated Endothelin-1 in a macrophage cell line (Additional file 6: Fig. S6). We next obtained Calr^del10/WT macrophages by culturing BM-mononuclear cells (BM-MNCs) in the presence of M-CSF (Fig. 2d, e). The increases in the Endothelin-1 and pSTAT3 levels did not show the statistical difference between in Calr^WT/WT and Calr^del10/WT macrophages at baseline, but these levels in Calr^del10/WT macrophages were significantly more upregulated compared to Calr^WT/WT macrophages after lipopolysaccharide stimulation (Fig. 2f, g). These data suggest that BM-derived macrophages with the Calr mutation played important roles in the PA remodeling.

To date, most of studies for PH in MPN patients lack information about driver mutations, although a retrospective study indicated higher prevalence of CALR mutations in ET patients with PH than those without [6]. Besides megakaryocyte lineage with TpoR expression, a recent study indicated that transcriptional misregulation occurs with JAK-STAT activation in CALR-mutated PB-MNCs similar to JAK2-mutated PB-MNCs [12]. Our murine model revealed a hematopoietic phenotype with relevance to human MPNs with CALR mutations in terms of molecular mechanisms and PH. Further study of associations between CALR mutations and PH or macrophage activation is needed (Additional file 10).