CALR Type 1‐Like Mutations Increase Endoplasmic Reticulum Free Ca²⁺ and Induce ERK1/2 Activation Independent of Thrombopoietin Receptor Activation

ABSTRACT

Calreticulin is a multifunctional protein found in the endoplasmic reticulum lumen that is important for calcium homeostasis and glycoprotein folding. Mutations in exon 9 of the CALR gene are the second most common genetic cause of myeloproliferative neoplasms. CALR‐mutated megakaryocyte proliferation in myeloproliferative neoplasms involves cytokine‐independent constitutive activation of JAK/STAT signaling caused by binding of mutant calreticulin to the thrombopoietin receptor. However, whether the partial or complete removal of wildtype calreticulin from the endoplasmic reticulum has additional effects on megakaryocyte biology is not clear. To explore the impact of calreticulin mutations independent of thrombopoietin receptor signaling we generated type 1‐like CALR mutations in K‐562 cells, which do not express the thrombopoietin receptor. We confirmed that the loss of endoplasmic reticulum‐retention KDEL motif causes the majority of mutant calreticulin to be secreted from cells. The CALR mutated cells have higher endoplasmic reticulum free Ca²⁺ but basal cytosolic Ca²⁺ is unchanged. Cells in which the KDEL endoplasmic reticulum retention motif was lost from all CALR alleles had increased ERp57 expression however the unfolded protein response was not induced. The calreticulin mutated cells also showed elevated basal phosphorylation of ERK1/2. Overall, these results suggest that the phenotype of type 1 CALR mutated myeloproliferative neoplasms is not solely due to cytokine independent activation of the thrombopoietin receptor by the mutant calreticulin, and that increased endoplasmic reticulum Ca²⁺ and/or basal ERK1/2 activation may contribute to the abnormal megakaryocyte proliferation characteristic of CALR mutant myeloproliferative neoplasms.

Abbreviations

Calr

calreticulin

ER

endoplasmic reticulum

ET

essential thrombocythemia

IL

interleukin

MPN

myeloproliferative neoplasm

PMF

primary myelofibrosis

PV

polycythemia vera

SOCE

store‐operated calcium entry

T1

type 1

TPO

thrombopoietin

UPR

unfolded protein response

WT

wildtype

1. Introduction

Myeloproliferative neoplasms (MPNs) are hematological malignancies characterized by increased proliferation of one or more types of myeloid cells in the bone marrow (Cacemiro et al. 2018; Grabek et al. 2020; O'Sullivan and Mead 2019). Philadelphia chromosome‐negative MPNs (classical MPNs) include polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF). These disorders have phenotype‐specific clinical characteristics with marked variability in risk of disease complication; PV involves increased red cell production, hemoglobin and hematocrit; ET involves increased megakaryocyte and platelet production with thrombocytosis; and PMF involves overproduction of megakaryocytes and granulocytes with bone marrow fibrosis (Saeidi 2016). Somatic heterozygous mutations in three driver genes, JAK2, CALR and MPL, are responsible for more than 90% of MPN cases, with phenotype‐modifying mutations characterized in several other genes (Grabek et al. 2020; Luque Paz et al. 2023; Saeidi 2016). The JAK2V617F mutation is found in approximately 95% of PV cases and more than 50% of ET and PMF cases, with MPL mutations occurring in 5‐10% of patients with ET and PMF (Kralovics et al. 2005; Spivak 2017). Mutations in the CALR gene are found in JAK2/MPL‐negative ET or PMF, accounting for 20%–30% of these cases (Araki et al. 2016; Saeidi 2016).

CALR encodes calreticulin (Calr), a 46‐kDa endoplasmic reticulum (ER)‐localized protein that belongs to the calcium‐binding chaperone family (Michalak 2023). Calr consists of three domains; the N‐domain responsible for chaperone activity, the proline‐rich lectin‐like P‐domain, and the C‐domain rich in acidic amino acids and involved in calcium binding (Nakamura et al. 2001; Wang et al. 2012). The C‐terminal KDEL motif ensures retention in the ER. Calr is involved in folding newly synthesized glycoproteins and regulation of calcium homeostasis (Lu et al. 2015; Michalak 2023). Since the first report of MPN‐associated somatic mutations in CALR (Klampfl et al. 2013; Nangalia et al. 2013) over 50 different indels have been identified in exon 9 of the CALR gene, and all lead to a +1 frameshift resulting in an altered C‐terminus with the loss of the KDEL sequence (Bilbao‐Sieyro et al. 2016; Pietra et al. 2016). Based on the extent of loss of the Ca²⁺ binding motifs in the C‐terminal domain, these mutations are classified as type 1 or type 1‐like (~65%), type 2 or type 2‐like (~32%), or other type (~3%) (Lasho et al. 2018; Tefferi, Lasho, et al. 2014; Tefferi, Wassie, et al. 2014).

Abnormal megakaryopoiesis is seen in CALR‐mutated MPN patients and mouse models (Araki et al. 2016; Chachoua et al. 2016; Marty et al. 2016; Shide et al. 2017). The primary mechanism of CALR‐driven megakaryocyte overproduction and thrombocytosis in MPN involves cytokine‐independent constitutive activation of JAK2/STAT signaling due to binding of mutant Calr to the thrombopoietin (TPO) receptor (MPL) (Araki et al. 2016; Chachoua et al. 2016; Prins et al. 2020). However, the disparity in phenotypes associated with CALR compared to JAK2 and MPL mutations suggests that there are additional molecular changes. In addition, it has been suggested that CALR mutations introduce a greater proliferative advantage compared to JAK2 mutations (Stegelmann et al. 2023). A number of groups have reported that induction of an ER stress response and/or altered calcium signaling may also contribute to the MPN phenotype (Di Buduo et al. 2020; Ibarra et al. 2022; Jaiswal et al. 2023; Jutzi et al. 2023; Pietra et al. 2016; Salati et al. 2019) although some results are contradictory. To explore the impact of type 1 CALR mutations independent of abnormal TPO receptor activation, we have generated type 1‐like CALR deletions in K‐562 cells, which do not express the TPO receptor (Methia et al. 1993). We investigate the impact on mutant Calr fate, the ER stress response, calcium homeostasis and ERK1/2 activation as well as megakaryocytic maturation.

2. Materials and Methods

2.1. Cell Culture

The human K‐562 megakaryoblastic leukemic (ATCC CRL‐243) (Tabilio et al. 1983), and U937 histiocytic lymphoma (ATCC CRL‐1593.2) cell lines were maintained in RPMI‐1640 (ThermoFisher 31800105) supplemented with 10% fetal bovine serum (FBS), sodium bicarbonate (2 g/L), penicillin (100 U/mL), and streptomycin (100 µg/mL) at 37°C in a 5% CO₂ humidified atmosphere. The identity of the cell lines was confirmed by STR analysis (DNA Diagnostics, NZ) and the cells were confirmed to be mycoplasma free using the MycoAlert™ mycoplasma detection kit (Lonza LT07‐318).

Hematopoietic stem cell‐derived megakaryocytes were obtained from CD45⁺ cells isolated and purified from peripheral blood mononuclear cells (Faiz et al. 2024; Ong et al. 2017). To induce differentiation into megakaryocytes, isolated CD45⁺ cells (1 × 10⁵ cells/mL) were cultured in StemSpan medium (StemCell Technologies, Vancouver, Canada) supplemented with 10 ng/mL thrombopoietin (TPO), interleukin (IL) 6 and IL 11 (StemCell Technologies). Cultures were maintained at 37°C in a 5% CO₂ humidified atmosphere for 14 days with media changes on days 5, 8, and 11. Platelets were obtained from whole blood was collected into ACD‐A tubes as previously described (Zsóri et al. 2013). The University of Otago Human Ethics Committee (NZ) (H20/132) approved the study and written informed consent was obtained from all participants.

2.2. CRISPR/Cas9 Gene Editing

To generate clonal K‐562 cells with type 1‐like CALR mutations, we used CRISPR/Cas9 gene editing with the introduction of double‐strand breaks (DSBs) and subsequent nonhomologous end‐joining. Briefly, sgRNAs targeting the 52‐bp exon 9 region for deletion with high on‐target efficacy scores were designed using the Benchling CRISPR Guide RNA Design Tool (Supporting Information S1: Table S1). The designed oligonucleotides with BbsI restriction enzyme‐compatible overhangs were purchased from Integrated DNA Technology. Each sgRNA was ligated into the eSp.Cas9(1.1) plasmid (Slaymaker et al. 2016) modified by the introduction of a T2A sequence and EGFP coding sequence from pSpCas9(BB)‐2A‐GFP (PX458) (Ran et al. 2013) into the Not1/Fse1 sites. eSpCas9(1.1) and pSpCas9(BB)‐2A‐GFP (PX458) were gifts from Feng Zhang (Addgene plasmid # 71814; http://n2t.net/addgene:71814; RRID:Addgene_71814; Addgene plasmid # 48138; http://n2t.net/addgene:48138; RRID:Addgene_48138). The plasmids were cotransfected into K‐562 cells using Neon™ electroporation and 48 h after transfection live cells expressing EGFP were isolated by fluorescence activated cell sorting (BD FACSAria Fusion at the Otago Micro and Nanoscale Imaging Facility, University of Otago). Clonal lines were isolated by limiting dilution and the genomic change in selected clones was determined on the Illumina MiSeq platform using the Illumina two‐step PCR amplicon approach. The first‐round PCR amplification was carried out to add 18‐bp overhang sequences to the target region using the Calr_F and Calr_R primers (Supporting Information S1: Table S1). The second PCR amplification was performed to enrich the target locus in the sample using a combination of forward and reverse barcoded Illumina Nextera XTTM DNA primers. After each amplification, the samples were purified using AmPure XP beads (Beckman Coulter). Illumina Nextera XT DNA library preparation (4 nM) and sequencing (Illumina MiSeq V2; 2 × 150 base PE instrument) were performed by the Otago Genomics Facility, University of Otago. Nine off‐target regions with the highest Benchling prediction score (> 1) exhibited no genetic modifications in the clones (data not shown).

2.3. Western Blotting Analysis

Whole cell lysates were prepared in RIPA buffer (50 mM Tris‐HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% NP‐40, 0.25% sodium deoxycholate) containing protease and phosphatase inhibitor cocktail (MS‐SAFE, Sigma‐Aldrich;). The cell suspension was vortexed briefly, incubated on ice for 10 min, and centrifuged at 13,000 rpm for 10 min at 4°C and the supernatant was stored at −80°C. The concentration of the whole cell lysate was determined using the BCA assay.

Secretome samples were prepared by plating cells in Opti‐MEM I Reduced Serum Medium (ThermoFisher 11058021) in a 6‐well plate. After 24 h at 37°C in a 5% CO₂ incubator the supernatant was collected by centrifugation (1000 g, 5 min, RT and then 30,000 g, 30 min, RT) and concentrated by ultrafiltration centrifugation using a 10 kDa cut‐off membrane (Amicon, Millipore).

Whole cell lysates or secretome samples were resolved by SDS‐PAGE under reducing conditions and transferred to nitrocellulose membrane. The membrane was stained with REVERT 700 total protein stain (LI‐COR Biosciences), washed (6.7% (v/v) glacial acetic acid, 30% (v/v) methanol), and probed with the indicated primary and secondary antibodies (Supporting Information S1: Table S2). The blots were visualized with the Odyssey Fc Imaging System (LI‐COR Biosciences), and the band intensities were quantified with Image Studio Lite version 5.2 (LI‐COR Biosciences). For quantification of ERp57, calnexin and PDI, signal intensity was normalized to total REVERT intensity and expression relative to WT was calculated. For quantification of ERK1/2 phosphorylation the signal intensities of pERK1/2 and ERK1/2 were normalized to tubulin and the ratio of pERK1/2 to ERK1/2 was calculated.

2.4. Mass Spectrometry

Secretome samples prepared as described above were further processed by the filter‐aided sample preparation protocol (Wisniewski 2017). In brief samples were reduced and alkylated on filter in 5 mM tris(2‐carboxyethyl)phosphine and 10 mM iodoacetamide respectively, washed in 100 mM triethylammonium bicarbonate buffer and digested on filter with trypsin (proteomics grade, Promega) in a 1/20 ratio of trypsin/total protein.

Peptide samples were subjected to targeted parallel reaction monitoring mass spectrometry using a nano‐flow liquid chromatography coupled 5600+ TripleTOF mass spectrometer (AB Sciex). We were unable to target mutant‐specific peptides because the altered C‐terminal region of mutant Calr is composed of a stretch of positively charged amino acids with several lysine or arginine residues resulting in too small or too large peptides using common proteases such as trypsin, chymotrypsin, Lys‐C, Glu‐C, or Arg‐C. Hence, we selected four peptides in the C‐ or P‐domain of Calr that are common for both mutant and WT to detect the secretion of mutant Calr. The following Calr target peptides were monitored in an unscheduled analysis: EQFLDGDGWTSR ([M+2H]²⁺ at m/z 705.82), GLQTSQDAR ([M+2H]²⁺ at m/z 488.25), KPEDWDEEMDGEWEPPVIQNPEYK ([M+3H]³⁺ at m/z 987.44), IKDPDASKPEDWDER ([M+2H]²⁺ at m/z 900.92). In addition, two serotransferrin peptides were targeted as reference peptides for intensity normalization: MYLGYEYVTAIR ([M+2H]²⁺ at m/z 739.87), EDPQTFYYAVAVVK ([M+2H]²⁺ at m/z 815.41). Each sample was measured in four technical replicates resulting in an intra‐assay coefficient of variation in percent (CV%) of < 6% for all measurements and target peptides. Raw data were analyzed using the Skyline software (https://skyline.ms/project/home/begin.view) (Pino et al. 2020).

2.5. Monitoring Intra‐ER Free Ca²⁺ and Cytosolic Ca²⁺

Intra‐ER free Ca²⁺ was measured using Mag‐Fluo‐4, AM (Rossi and Taylor 2020) (ThermoFisher M14206). Briefly, cells were plated into a black clear‐bottom 96‐well plate with 200 μL of 1 mg/mL BSA/KRH (1 mg/mL BSA, 135 mM NaCl, 5 mM KCL, 1 mM MgSO₄, 0.4 mM K₂HPO₄, 5.5 mM glucose, 20 mM HEPES, pH 7.4) solution and incubated at 37°C for 60 min. After incubation, cells were washed and resuspended in 190 μL of KRH solution, followed by 30 min incubation at 37°C. Finally, calcium flux was monitored for up to 10 min with the injection of 10 μM thapsigargin at 1 min at λ _EX 483‐14, λ _EM 530‐30 using a CLARIOstar microplate reader (BMG Lab Technologies, USA). To quantify intra‐ER Ca²⁺, Mag‐Fluo‐4, AM signal intensity was normalized with Hoechst staining in each well. Briefly, Hoechst stain was added at a final concentration of 1 mg/mL to the wells, the plate was incubated for 30 min, and the fluorescence intensity was measured at λ _EX 355, λ _EM 460 using a CLARIOstar microplate reader (BMG Lab Technologies, USA).

Cytosolic Ca²⁺ was measured using Fura‐2‐AM (Roe et al. 1990) (ThermoFisher F1221). Briefly, cells were incubated with 3 µM Fura‐2‐AM and plated in a black clear‐bottom 96‐well plate in 200 μL of 1 mg/mL BSA/KRH solution and incubated at 37°C for 30 min. After incubation cells were washed and resuspended in 190 μL of KRH solution, incubated at 37°C for 10 min, and calcium flux was monitored for up to 10 min at λ _EX 335‐12 and 380‐12, λ _EM 510‐30 using a CLARIOstar® microplate reader (BMG Lab Technologies, USA).

2.6. Quantifying ER Volume and Surface Area

Cells were transfected with pcDNA‐D1ER (Palmer et al. 2004) (a gift from Prof Peter Jones, University of Otago) using Lipofectamine 3000 (ThermoFisher L300001) and incubated at 37°C for 24 h before plating in a 35 mm culture dish in Ca²⁺‐free KRH solution. D1ER expressing cells were kept at 37°C using a Stage Top incubator (Tokai‐Hit) and live‐cell images were acquired with a 60 × oil objective at optimal pixel size and interval using a confocal microscope (FV3000, Olympus) at λ _EX 436‐20 and λ _EM 535‐40. A total of 28 Z‐slices per cell were collected using NIS‐Elements, Advanced Research 4.50 software. Quantification of ER volume, surface area, and volume to surface ratio in > 80 cells per biological replicate was performed using the Fiji 3D Objects Counter plugin (Bolte and cordelières 2006; Schindelin et al. 2012).

2.7. RNA Isolation and RT‐qPCR

Total RNA was isolated from K‐562 cells, mature megakaryocytes and platelets using TRIzol (ThermoFisher 15596026). First‐strand cDNA synthesis was performed with 1 µg of total RNA in 20 µL reaction volume using SuperScript IV VILO master mix (ThermoFisher 11766050). RT‐qPCR reactions were run on LightCycler^TM 480 Instrument II System (Roche Life Science) using PowerUp SYBR Green Master Mix (ThermoFisher 4367659). All the procedures were followed as recommended by the manufacturers. The qPCR primer sequences are in Supporting Information S1: Table S1. RT‐qPCR reactions were performed in triplicate, and the basket normalization method with two reference genes (HPRT and GAPDH) was performed to quantify the relative mRNA expression (Hellemans et al. 2007; Vandesompele et al. 2002; VanGuilder et al. 2008). The following equation was used to quantify the relative fold change of target mRNA expression normalized to two reference genes.

$$\text{Relative}\text{fold}\text{change}=\frac{{\left({\mathrm{E}}_{\text{GOI}}\right)}^{\Delta \text{CtGOI}}}{\text{GeoMean}\left[{\left({\mathrm{E}}_{\text{REF}}\right)}^{\Delta \text{CtREF}}\right]}.$$

ΔCt = Calibrator Ct‐Sample Ct. GOI: gene of interest, REF: reference gene and E: efficiency.

2.7.1. Assessment of Megakaryocytic Maturation

K‐562 cells were resuspended in Dulbecco's PBS and stained for 30 min in the dark at room temperature with anti‐CD61 PE (1:20; BD Biosciences 1075384), anti‐CD41a V450 (1:20; BD Biosciences 1011309), and Zombie Green (1:100; BioLegend B334686). Cells were washed twice with Dulbecco's PBS, resuspended in FACS buffer (0.01% sodium azide, 0.1% BSA in Dulbecco's PBS) and analyzed using the Guava easyCyte 5 HPL benchtop flow cytometer (Merck Millipore) or the BD LSRFortessa cell analyzer (BD Biosciences). Data analysis was performed with FlowJo version 10 (FlowJo, Ashland, USA) and data are presented as geometric mean fluorescence intensity (MFI) of the live cell (Zombie negative) population.

3. Results

3.1. K‐562 Cells With Type 1 CALR Mutations Secrete Mutant Calreticulin

To study the impact of type 1 CALR mutations independent of constitutive activation of the TPO receptor we made use of K‐562 cells, which do not express this receptor (Methia et al. 1993) (Supporting Information S1: Figure S1). We targeted deletion of the 52‐bp region of CALR exon 9 (the most common type 1 mutation) using a dual guide Crispr/Cas9 approach. K‐562 cells have three copies of the CALR allele (Naumann et al. 2001), and four clonal lines with different mutation profiles were studied (Table 1, Supporting Information S1: Tables S3 and S4). The C‐terminus region of Calr consists of three stretches of negatively charged residues, and the categorization of CALR exon 9 mutants that have been reported in ET and PMF patients usually depends on the number of stretches deleted (Gángó et al. 2018; Pietra et al. 2016). Tefferi et al used a statistical model, AGADIR, which calculates helix propensity for 31 unique residues altered by CALR mutations to subclassify CALR mutations (Tefferi, Lasho, et al. 2014). Accordingly, we categorized our CALR mutations based on their predicted effect on three negatively charged amino acid stretches. All CALR mutations that we generated exhibited similar characteristics to type 1 mutations with the loss of two negative amino acid stretches (Supporting Information S1: Table S4). To confirm our categorization of CALR mutations, we calculated the pI value of each mutant Calr starting from codon A352. We found that the pI values differed significantly between the CALR mutations but were similar within each group, confirming our categorization of CALR mutants (Supporting Information S1: Table S4). Genetic alterations resulting in truncated mutant CALR were referred to as other type (OT) because of the lack of a common 36‐basepair consensus sequence of the mutated C‐terminal. Based on the CALR mutant subclassification, we termed our clones as C20^T1*/T1*/OT, C21^WT*/T1*/T1*, C22^WT*/T1*/OT, and C24^T1*/T1*/OT (* denotes ‐like).

Table 1.

Mutant Calr classification.

Clone	Amino acid change	Mutant Calr classificationa	Predicted Calr molecular weight (Da)b
C20	Leu367Argfs*61	T1*	48025
	Leu367Glnfs*59	T1*	47713
	Glu369Glyfs*20	OT	42346
C21	Glu369Glyfs*10	WT*	45065
	Glu364Thrfs*57	T1*	47043
	Lys368Argfs*63	T1*	48295
C22	Gln365_Glu369del	WT*	45793
	Asp362Glufs*51	T1*	46928
	Lys368Argfs*9	OT	41134
C24	Arg366Thrfs*46	T1*	45821
	Lys368Argfs*63	T1*	48295
	Glu370Thrfs*13	OT	41735

^a Mutation type is as defined (Gángó et al. 2018; Tefferi, Lasho, et al. 2014). WT* and T1* denote WT‐like and type 1‐like, respectively, OT indicates other type.

^b Predicted molecular weight excluding the signal peptide. Predicted molecular weight of WT Calr excluding the signal peptide is 46,449 Da, noting that WT Calr resolves at 55,000–60,000 Da by SDS PAGE (Varricchio et al. 2017).

We compared the expression of Calr in whole cell lysates and the secretome in the four clonal lines and the parental cells. Using a polyclonal anti‐Calr antibody, Calr was only detected in whole cell lysates from the parental cells and clones C21 and C22, which each contain a WT‐like CALR allele (Figure 1A). The size of the Calr was lower than WT in C21 and C22, consistent with their predicted molecular weight (Table 1). Expression also appeared lower, consistent with the presence of only one WT‐like allele in these clones. Using a monoclonal anti‐Calr antibody, which recognizes both WT and type 1 mutant Calr (Han et al. 2016), a faint band of presumably mutant Calr was detected in clone C20 and an additional band was detected in clone C21 (Figure 1B). Several studies have reported that mutant Calr acts as a chaperone for the TPO receptor and is cotrafficked to the cell surface via the secretory pathway with partially mature receptor (Masubuchi et al. 2020; Pecquet et al. 2019), and that intracellular mutant Calr levels are not increased by an autophagy inhibitor (Han et al. 2016). In contrast, Mansier et al. reported that proteasome‐mediated degradation is responsible for the lower expression of mutant Calr (Mansier et al. 2019). Therefore, we examined whether the mutant Calr was being secreted or degraded. Secretome samples were analyzed by western blotting analysis and mass spectrometry. There was very little secretion of WT Calr. However, all the clonal lines showed increased secretion with highest levels of Calr secreted from clones C20 and C24 (Figure 1C), consistent with all Calr in these clones lacking the KDEL ER retention signal. We next quantified Calr secretion by mass spectrometry, finding that the clones with loss of the KDEL motif on all CALR alleles (C20 and C24) had the highest secretion of mutant Calr, followed by the clones with retention of KDEL motif in one allele (C21 and C22) (Figure 1D). To determine if mutant Calr undergoes proteasomal degradation, we analyzed the expression of Calr after a 24 incubation with MG132, a proteasomal inhibitor (Supporting Information S1: Figure S2). There was no increase in the expression of the mutant Calr in any of the clonal cell lines. In summary, our data demonstrate that the presence of the altered C‐terminal domain lacking the KDEL retention motif leads to secretion of the mutant Calr rather than proteasomal degradation, and that secretion does not require the presence of the TPO receptor.

Figure 1.

Characterization of CALR Type 1‐like clonal cell lines. (A) Calr expression in whole cell lysates from WT cells and clonal cell lines using a polyclonal anti‐Calr antibody. Green is Calr, red is Revert total protein stain. (B) Calr expression in whole cell lysates from WT cells and clonal cell lines using a monoclonal anti‐Calr antibody. Tubulin is the loading control. (C) Calr in secretome samples from WT cells and clonal cell lines using a monoclonal anti‐Calr antibody. (D) Mass spectrometry targeted parallel reaction monitoring to quantify the secretion of Calr. Bar graph represents the average signal intensities of selected Calr peptides normalized to serotransferrin.

3.2. K‐562 Cells Expressing Type 1‐Like CALR Mutant Proteins Exhibit Elevated ER Ca²⁺ and ERK1/2 Phosphorylation

Previous studies showed that type 1 CALR mutants influence ER Ca²⁺ levels (Pietra et al. 2016) and store‐operated calcium entry (SOCE) activity (Di Buduo et al. 2020). To gain further insights into Ca²⁺ handling in the Calr‐mutant expressing K‐562 clonal lines, we first determined basal cytosolic and ER free Ca²⁺ levels using Fura 2, a ratiometric fluorescent dye, and Mag‐Fluo‐4 AM, a low‐affinity Ca²⁺ indicator, respectively (Figure 2). While there was no difference in basal cytosolic Ca²⁺ between WT cells and the clonal lines (Figure 2A), basal ER free Ca²⁺ was significantly increased in clones C21 and C24, with clones C20 and C22 also having an indication of increased basal ER Ca²⁺ (Figure 2B). To determine whether the elevated ER free Ca²⁺ was due to an expanded ER, we determined ER volume and surface area using D1ER. There was no difference in either of these parameters between WT and the Calr mutant clonal lines (Figure S3), confirming that the increased ER free Ca²⁺ is most likely due to loss of ER Ca²⁺ binding capacity due to secretion of mutant Calr from the cells.

Figure 2.

CALR mutations increase basal ER Ca²⁺ and ERK1/2 phosphorylation. (A, B) Relative resting cytoplasmic (A) and ER (B) Ca²⁺ compared to WT. N = 3 ± SD. *p < 0.05 compared to WT (one sample t‐test). (C) Time course of intra luminal ER Ca²⁺ flows in CALR mutant clonal cell lines. Cells were loaded with Mag‐Fluo‐4AM and intra ER Ca²⁺ pools were depleted, in the absence of extracellular Ca²⁺, by addition of 10 μM thapsigargin (Tg). N = 3 ± SEM. (D) Quantification of ERK1/2 phosphorylation.

We next determined whether having either less (C21 and C22) or no (C20 and C24) WT‐like Calr in the ER altered Ca²⁺ release from the ER in response to the sarco/endoplasmic reticulum Ca²⁺ ATPase inhibitor thapsigargin. Thapsigargin caused an immediate emptying of ER Ca²⁺ stores in WT and all clonal cell lines (Figure 2C) demonstrating that the ability to release Ca²⁺ via SOCE is not altered by the expression of mutant Calr. There was a greater decrease in the fluorescent signal in the clonal lines compared to WT cells, consistent with the higher basal ER free Ca²⁺.

Given the role of MAPK signaling both downstream of constitutive JAK/STAT pathway activation in Calr mutant megakaryocytes (Chachoua et al. 2016; Kollmann et al. 2017) and in human megakaryocyte maturation (Di Buduo et al. 2016), we next determined whether the presence of mutant Calr leads to basal activation of the Raf‐MEK‐ERK MAP kinase pathway by assessing ERK1/2 phosphorylation. The four clonal lines had similar ERK1/2 phosphorylation, and this was significantly higher than WT cells for three of the four clonal lines (C20, C21, C22) (Figure 2D). Thus, MAPK signaling is activated independent of the mutant Calr/TPO receptor interaction in type 1 CALR mutant K‐562 cells.

3.3. ERp57 Expression Is Increased in Type 1‐Like Calr Mutant Cells Independent of UPR Induction

Calr and its membrane‐bound homolog calnexin assist glycoprotein folding in the ER by recruitment of other chaperones such as ERp57 (PDIA3) (Michalak 2023). We therefore determined whether expression of calnexin and ERp57, as well as PDI (PDIA1, a non‐Calr‐associated chaperone) changed in response to the decreased level of intracellular Calr. There was a significant increase in ERp57 expression in clones C20 and C24, in which all alleles lack the ER retention KDEL motif, and a moderate increase in clones C21 and C22 clones which retain one WT‐like allele (Figure 3A). Calnexin and PDI levels were comparable between WT and all clonal lines (Figure 3B,C).

Figure 3.

ERp57 is increased but UPR is not induced in type 1 mutant CALR cells. (A–C) Expression of ERp57 (A), calnexin (B) and PDI (C) in whole cell lysates of WT cells and clonal cell lines. N = 3 ± SD, *p < 0.05, **p < 0.01 compared to WT (one sample t‐test). (D–G) Relative expression of CHOP (D), BiP/GRP78 (E), HERPUD1 (F) and Xbp1s/t (G). N = 3 ± SD. *p < 0.05 compared to WT (one‐way ANOVA).

It has been reported that activation of the unfolded protein response (UPR), particularly via the IRE1α/XBP1 splicing pathway, contributes to the abnormal megakaryocyte proliferation seen in MPN (Jutzi et al. 2023; Nam et al. 2019). Given the partial loss of both Ca²⁺ ‐binding and glycoprotein folding capacity in the ER expected with the reduction in Calr levels, we hypothesized that we would see UPR activation in the clonal cell lines. We assessed expression of CHOP, BiP/GRP78P and HERPUD1, and Xbp1 splicing, as indicators of UPR activation. There was no difference between the WT cells and the clonal cell lines with type 1 CALR mutations apart from increased CHOP expression in one cell line (C24) (Figure 3D–G). It has also been reported that overexpression of type 1 mutant Calr in K‐562 cells inhibits UPR induction by exogenous stressors (Salati et al. 2019). However we found no difference in the induction of BiP or CHOP expression, or XBP1 splicing, in response to thapsigargin in any of the clonal cell lines compared to the WT cells (Supporting Information S1: Figure S4).

3.4. Type 1‐Like CALR Mutant Proteins Do Not Induce Megakaryocytic Maturation of K‐562 Cells

Finally, we determined whether the expression of type 1‐like Calr was sufficient to drive megakaryocytic maturation of K‐562 cells in the absence of an exogenous stimulus such as PMA. by analyzing surface expression of CD61 and CD41a and the levels of the megakaryopoiesis transcription factors GATA1, RUNX1, and NFE2. There was no increase in basal expression of CD61 or CD41a (Figure 4A,B) or the levels of the megakaryocyte ‐related transcription factors GATA1, RUNX1, and NFE2 (Figure 4C–E), in any of the clonal lines.

Figure 4.

CALR mutations do not induce megakaryocytic maturation of K‐562 cells. Expression of CD61 (A), CD41a (B), GATA1 (C), RUNX1 (D) and NEF2 (E) in WT and CALR mutant clonal cell lines. N = 3 ± SD. *p < 0.05 compared to WT (Kruskal‐Wallis test).

4. Discussion

The discovery of CALR mutations in MPN has prompted many researchers to investigate the molecular mechanism of pathogenic mutant CALR with a view to developing effective treatments (Faiz et al. 2025). It is well established that binding of mutant Calr to the TPO receptor, leading to the constitutive activation of JAK/STAT signaling, is the primary cause of CALR mutated MPN pathogenesis (Araki et al. 2016; Chachoua et al. 2016; Elf et al. 2016, 2018; Marty et al. 2016; Shide et al. 2017). However, other pathways are thought to contribute to mutant CALR‐driven MPN oncogenesis (Lau et al. 2015). Notably, only 16.7% of CALR‐mutated patients respond to Ruxolitinib treatment (Guglielmelli et al. 2016) suggesting that JAK/STAT signaling downstream of the TPO receptor is not the only aberrant signaling responsible for MPN pathogenesis.

In this study, we generated type 1‐like CALR mutant expressing K‐562 cells using CRISPR/Cas9 technology. A gene editing approach has been used to model CALR mutations in human cell lines (Shide et al. 2017), murine cell lines (Abdelfattah and Mullally 2018), and in a recent study to correct the patient‐derived iPSCs carrying homozygous or heterozygous CALRdel52 or CALRins5 mutations (Olschok et al. 2021). We used K‐562 cells in this study for two reasons: 1) they have the ability to undergo in vitro megakaryocytic differentiation (Herrera et al. 1998) and 2) they lack the TPO receptor, so can serve as a model system to investigate the role of TPO‐independent events. The goal of this study was to examine if type 1 CALR mutations affected Ca²⁺ homeostasis and/or ER protein homeostasis and related signaling pathways, which could aid MPN oncogenesis. We studied four clonal lines, two that expressed only mutant Calr and two that expressed WT and mutant Calr, noting that CALR mutations in MPN patients are almost always heterozygous.

Consistent with previous findings of type 1 mutant Calr (Han et al. 2016), the mutant Calr was secreted from the cells. We did not find any evidence of the proteasomal degradation of mutant Calr previously reported (Mansier et al. 2019). The proteasomal degradation observed by Mansier et al (Mansier et al. 2019) may be a consequence of the use of an exogenous overexpression system and seems unlikely to occur in vivo based on our and others' results.

Ca²⁺ homeostasis plays a vital role in megakaryocyte maturation function with ER Ca²⁺ release required for normal megakaryocyte adhesion and proplatelet formation (Andrea Di Buduo et al. 2014). However, the impact of CALR mutations on cellular Ca²⁺ homeostasis is not fully understood. The replacement of the Ca²⁺‐binding, negatively charged amino acids with a stretch of positively charged ones, as well as secretion of mutant Calr from cells may disrupt the Ca²⁺ buffering activity within the ER lumen and therefore alter Ca²⁺‐dependent signaling pathways. Analysis of ex vivo differentiated megakaryocytes from subjects with type 1 or 2 mutant Calr demonstrated increased cytosolic Ca²⁺ spiking compared to control cells, and decreased interaction of mutant Calr with STIM1 and ERp57, but the basal levels of cytosolic or ER Ca²⁺ were not reported (Di Buduo et al. 2020). In another study of ex vivo differentiated WT or type 1 mutant CALR megakaryocytes, basal cytosolic Ca²⁺ was unchanged, and Ca²⁺ spiking wasn't observed, but cytosolic Ca²⁺ increased to a higher level in the type 1 mutant cells following depletion of ER Ca²⁺ stores with cyclopiazonic acid (Pietra et al. 2016). Our results directly confirm that loss of type 1 Calr from the ER results in increased ER free Ca²⁺. Determining whether increased ER Ca²⁺ contributes to the abnormal megakaryocyte proliferation in MPN requires further investigation. It has been proposed that tumorigenic cell lines rely on persistent IP3R‐mediated Ca²⁺ transfer from the ER to mitochondria to maintain mitochondrial function and cell survival (Cárdenas et al. 2016). The functional consequences and detailed mechanism of ER‐mitochondria Ca²⁺ interplay in response to type 1 CALR‐mediated Ca²⁺ alterations remains unexplored.

CALR mutations mediate constitutive activation of the TPO receptor due to mutant Calr binding (Araki et al. 2016; Chachoua et al. 2016; Elf et al. 2016, 2018; Marty et al. 2016; Shide et al. 2017), with aberrant MAPK signaling downstream of the JAK/STAT cascade helping to drive oncogenic transformation (Chachoua et al. 2016; Kollmann et al. 2017). We found that K‐562 clones expressing type 1‐like Calr had significantly increased basal phosphorylation of ERK1/2. This suggests that in type 1 Calr cells, abnormal ERK1/2 activation could be due to both TPO receptor‐dependent and independent signaling. The receptor‐independent mechanism is unclear. Calcium activation of ERK1/2 is well described (Wiegert and Bading 2011) with multiple potential partners including CaMKII, CaMKI, PYK2/RAFTK and Ras‐GRP. Further studies are warranted to determine whether elevated ER Ca²⁺ activates MAPK signaling and if so, which specific pathway is involved.

The loss of Calr from the ER appears to be compensated by increased expression of ERp57, though not an obvious UPR. Mass spectrometry analysis suggested reduced ERp57 expression after overexpression of MBP‐tagged type 1 Calr although this was not confirmed by western blot analysis (Pronier et al. 2018). In contrast, increased ERp57 was observed in ex vivo cultured mouse megakaryocytes expressing type 1 mutant Calr (Jutzi et al. 2023). The absence of an obvious UPR in the mutant cells was unexpected, especially for the two clonal lines with no WT Calr expression. While UPR induction in response to mutant Calr has been implied from RNA‐Seq and microarray data (Jutzi et al. 2023; Nam et al. 2019; Salati et al. 2019), there have been few direct investigations. Salati et al (Salati et al. 2019) overexpressed del52 and ins5 mutant Calr in K‐562 cells, and reported impairment of UPR induction, but did not compare basal expression levels of UPR targets. In contrast to Salati et al we observed robust induction of the UPR in the mutant cell lines, indicating that loss of Calr does not interfere with UPR signaling. We note that the UPR responses reported by Salati et al were relatively modest, even in WT cells (Salati et al. 2019). The impact of IRE1α inhibition on Calr^Δ52/+ mice has also been assessed, but no impact was seen on platelet counts, or the frequency of either long‐term hematopoietic stem cells or megakaryocyte‐erythroid progenitors (Jutzi et al. 2023).

In conclusion, our study describes an in vitro system for evaluating the TPO‐receptor independent effects of type 1 mutant Calr, which is a critical consideration for drug development. Specifically, we show type 1‐like Calr expressing K‐562 cells have elevated ER Ca²⁺ levels and increased baseline ERK1/2 phosphorylation. Either or both of these changes may contribute to the abnormal megakaryocyte proliferation seen in CALR mutant MPN.

Author Contributions

Conceptualization, visualization, writing – original draft, writing – reviewing and editing: Mifra Faiz and Elizabeth C. Ledgerwood. Investigation, methodology and formal analysis: Mifra Faiz, Caitlin Dunstan‐Harrison and Elizabeth C. Ledgerwood. Funding acquisition, project administration, resources, supervision: Elizabeth C. Ledgerwood.

Supporting information

Faiz and Ledgerwood Supplementary Data.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.