Expression properties, structural features and functional analysis of CALR E381A in MPN patients

Abstract

Objective: To investigate the expression properties, structural features and function of CALR E381A in myeloproliferative neoplasms (MPN) patients. Methods: In this retrospective study, 435 MPN patients admitted to the Department of Hematology, Ningbo First Hospital from July 2015 to July 2021 were selected as the study subjects. Mutations in CALR exon 9 from genomic DNA samples were identified by PCR, followed by Sanger sequencing. The physicochemical properties of the wild-type calreticulin and the p.E381A variant, and the structural information of the p.E381A variant were analyzed by using the bioinformatics databases. Growth assay of UT-7/mpl cells with CALR E381A was used for the functional analysis of CALR E381A. Results: The predominant types of CALR variants were identified as follows: p.L367fs*46 (38.1%), p.K385fs*47 (25.8%) and p.E381A (19.6%). Notably, the frequency of the p.E381A variant (c.1142A >C) in polycythemia vera or essential thrombocythemia was significantly higher than the frequency of that as a single nucleotide polymorphism (SNP) in the East Asian population. Furthermore, CALR E381A coexisted with other genetic variants, of which JAK2 V617F was more common. Bioinformatics analysis confirmed that CALR E381A did not change the physicochemical properties of the calreticulin protein, but did change the electrical charge, energy state and steric hindrance of amino acid residues at site 381. UT-7/mpl cells harboring CALR E381A overexpression did not exhibit altered cell growth, which is distinctly different from the stereotypical frameshift mutation. Conclusion: CALR E381A is not a driver mutation for the development of MPN but may be a risk SNP implying an inherited predisposition for MPN disease in East Asian populations.

Introduction

Myeloproliferative neoplasms (MPNs), characterized by overproduction of one or several blood elements, mainly include polycythemia vera (PV), primary myelofibrosis (PMF), and essential thrombocythemia (ET). Transformation can occur between different disease subtypes, mainly including ET or PV conversion into myelofibrosis (MF), and only a very small number of ET to PV. The 10-year conversion rates of PV, ET, and PMF to acute myeloid leukemia (AML) ranged from 2% to 4%, 1%, and 10% to 20%, respectively. The progress of MPNs is slow, and disease-related complications as well as transformation to AML are the important factors affecting the quality of life and survival time of patients.

The JAK2, MPL, and CALR gene mutations are the driver gene mutations of MPNs. About 95% of patients with PV carry JAK2 mutation (JAK2V617F). The gene mutation distribution of ET patients is similar to that of PMF patients. The proportions of patients carrying JAK2, CALR and MPL mutations were 50% to 60%, 30% and 5% to 8%, respectively. In addition to the above three types of driver gene mutation, some non-driver gene mutations, such as DNMT3A, TET2, ASXL2, TP53, also play a role in the clonal evolution of MPNs. However, the mechanism by which some individuals with genetic mutations develop clonal hematopoiesis and progress to MPNs on this basis is still unclear. The somatic mutations in the Janus kinase 2 gene (JAK2), the gene encoding calreticulin (CALR), and the thrombopoietin receptor gene (MPL) are considered to be the most common driver mutations in MPN and serve as the molecular basis for the diagnosis of MPN.

CALR, encodes a soluble Ca²⁺-binding protein, which influences a variety of processes in calcium homeostasis, immune response, phagocytosis and signaling [1]. In 2013, mutations in exon 9 of CALR were identified in 60%-90% of patients with ET and PMF that were JAK2/MPL-mutation negative [2,3]. These results have been repeatedly confirmed in subsequent studies [4]. Previous studies showed that the type of gene mutation could affect the disease phenotype and disease progression of MPNs. For example, patients with CALR type-1 mutation (52 bp deletion) tended to develop PMF [5], while patients with CALR type-2 mutation (5 bp insertion) were more prone to ET [6]. Moreover, it was shown that MPL mutation increased the risk of ET converting to MF in contrast to JAK2V617F and CALR mutations [7]. In addition, many studies confirmed that the increased gene mutation burden of JAK2 V617F allele would increase the risk of PV or ET converting to MF [8]. It was considered that increased gene mutation burden would also promote the evolution of MPNs disease [9]. To date, more than 60 frameshift mutations in CALR have been described with all involving exon 9. The two most frequent mutations are type-1 (c.1092_1143del; p.L367fs*46) and type-2 (c.1154_1155insTTGTC; p.K385fs*47) variants [10]. It has been generally believed that mutations in CALR exon 9 are mutually exclusive to JAK2 and MPL variants as the frequency of co-occurrence is usually below 1% [11]. This exclusivity is especially noted in PV, with only one case of coexisted JAK2 V617F and CALR exon 9 mutations being reported to date [12]. This low frequency is mainly due to the fact that CALR mutations are rarely detected in PV. Accordingly, it has been suggested that screening for CALR mutations in PV is not useful [13]. Whereas, these results have been mostly found in Caucasian populations, it is unknown if there are similar results among Asian populations. Given that many investigators have shown ethnical diversity in clinical and laboratory features of MPN [14-16], it is important to determine the actual mutation profile of CALR exon 9 in Chinese patients in order to clarify its pathogenic potential in MPN.

In this context, we analyzed the CALR exon 9 of 435 Chinese patients with MPNs and defined the clinical correlates, physicochemical properties, and spatial structure of the particular type of CALR variant. The results of this study would provide some clinical reference for guiding treatment of patients with MPNs.

Materials and methods

Study population

In this retrospective study, MPN patients that presented at Ningbo First Hospital (Ningbo, Zhejiang, China) from July 2015 to July 2021 were selected as the study subjects. Diagnoses were established according to the 2008 World Health Organization (WHO) criteria. A total of 435 patients with MPN (256 males, 179 females; mean age 61.2 years old) were included in the study. The genomic DNA samples were extracted from bone marrow/peripheral blood and buccal mucosa. The clinical data from medical records were retrospectively reviewed. The study was approved by the Ethics Committee of the Ningbo City First Hospital (2020-R131), and informed written consent was obtained from all patients in the study.

Mutation detection

Mutations in CALR exon 9 were identified in the collected samples using polymerase chain reaction (PCR), followed by Sanger sequencing as reported previously [17]. The JAK2 V617F and MPL W515L/K mutations were analyzed using TaqMan quantitative PCR with mutant-specific primers [18]. From 2019, the mutations were assessed by the panel-based next-generation sequencing performed on the Illumina NextSeq 500 platform (Illumina, Inc.).

Bioinformatic analysis

The protein structure of calreticulin was retrieved from the Protein Data Bank (PDB) database (structure entry 6ENY). The online bioinformatics ProtParam tool (http://www.expasy.ch/tools/protparam.html/) was used to analyze the physicochemical properties of the wild-type calreticulin and the p.E381A variant. The secondary structure of calreticulin was analyzed using the web server PDBSum database. Finally, the structural information for the p.E381A variant was visualized and analyzed using Swiss-Pdb Viewer 4.1.0.

Transfection and cell growth assay

UT-7 cell line was obtained from the Cell Resource Center, Chinese Academy of Medical Sciences (Beijing, China). The cDNA of human c-MPL (accession number NM_005373.3) from whole gene synthesis was introduced to UT-7 cells, using NotI/NsiI restriction site of the LV5-EF1a-GFP/Puro lentivirus vector (Gene Pharma Corp., Shanghai). The stably transfected UT-7/mpl cells were validated by Western blot with anti-MPL antibody (ab172061, Abcam). These cells were suspended in IMDM containing 10% fetal bovine serum (Gibco-Life Technologies) and 10 ng/ml recombinant human thrombopoietin (rhTPO, Peprotech). CALR del52 and CALR E381A were synthesized and separately cloned into EcoRI/HindIII-digested pcDNA3.1 (+) vector (Gene Pharma Corp., Shanghai). The recombinant plasmids and the empty vector plasmid were transferred into UT-7/mpl cells, respectively, using the cell electroporation system operator H1 (Etta Biotech Co., Ltd., Suzhou). Overexpression of construct was further verified by RT-PCR. UT-7/mpl cells harboring overexpression of constructs in the absence of TPO for 5 days were assessed with a TC20 Automated Cell Counter (Bio-Rad) through Trypan Blue exclusion. UT-7/mpl cells harboring the empty vector plasmid was served as a control.

Statistical analysis

Using SPSS software version 22.0, the Z test was used to compare the frequency of mutated gene between patients and the healthy population. The probability of a gene variant detected in <5 patients was directly calculated according to the principle of binomial distribution. α=0.05 was taken as the statistical test level.

Results

Identification of CALR mutations in patients with MPN

A total of 435 patients with BCR/ABL-negative MPN at initial diagnosis were included in this study, including 148 PV, 191 ET, and 96 PMF. Mutation profiles of JAK2 V617F, CALR and MPL W515L/K in the 435 patients with MPN are shown in Table 1. In 148 patients with PV, there were 126 patients with JAK2 V617F mutations and 8 patients with CALR Exon 9 mutations, and there was no patient with MPL W515L/K mutations. In 191 patients with ET, there were 104 patients with JAK2 V617F mutations, 57 patients with CALR Exon 9 mutations, and 2 patients with MPL W515L/K mutations. In 96 patients with PMF, there were 47 patients with JAK2 V617F mutations, 32 patients with CALR Exon 9 mutations, and 3 patients with MPL W515L/K mutations. Of the 97 patients with CALR Exon 9 mutations, the proportion of p.L367fs*46 (type-1), p.K385fs*47 (type-2) and p.E381A (c.1142A >C) were 38.1%, 25.8% and 19.6%, respectively, which were the predominant mutation types, as shown in Figure 1.

Table 1.

Mutation profiles of JAK2 V617F, CALR and MPL W515L/K in 435 patients with MPN

Diagnosis	Cases (n)	JAK2 V617F [n (%)]	CALR [n (%)]	MPL W515L/K [n (%)]
PV	148	126 (85.1)	8 (5.4)	0
ET	191	104 (54.5)	57 (29.8)	2 (1.0)
PMF	96	47 (49.0)	32 (33.3)	3 (3.1)
Total	435	277 (63.7)	97 (22.3)	5 (1.1)

Abbreviations: MPN, myeloproliferative neoplasm; PV, polycythaemia vera; ET, essential thrombocytosis; PMF, primary myelofibrosis; n, number.

Figure 1.

CALR E381A in myeloproliferative neoplasm. A. Type of CALR exon 9 mutations identified in 97 patients with myeloproliferative neoplasm. B. Sequence analysis of the c.1142A >C (p.E381A) mutation in CALR exon 9.

Patients with the CALR E381A

CALR E381A was identified in 19 patients with MPN, including 8 cases of PV, 7 cases of ET, 3 cases of PMF and 1 case of CML (BCR/ABL was detected 4 years after the initial ET diagnosis). Clinical and laboratory features of patients with CALR E381A are shown in Table 2. Notably, CALR E381A could coexist with various gene alterations, of which JAK2 V617F was more common. The co-occurrence of CALR E381A and JAK2 V617F mutation was identified in 9 cases. The frequencies of CALR E381A in JAK2 V617F-positive PV, PV and ET were 3.2% (4/126), 5.4% (8/148), and 3.7% (7/191), respectively. CALR E381A is regarded as a single nucleotide polymorphism (SNP: rs143880510) with a frequency of 1.45% in the East Asian population according to the National Center for Biotechnology Information (NCBI) Allele Frequency Aggregator (ALFA) database (https://www.ncbi.nlm.nih.gov/snp/rs143880510#frequency_tab). Compared with the frequency, there was a significantly higher frequency of CALR E381A in patients with PV (Z=4.02, P<0.001), or patients with ET (Z=2.60, P<0.05). Sequencing analysis of buccal swab-isolated DNA revealed that CALR E381A was a germline variant.

Table 2.

Clinical and laboratory features of patients with CALR E381A

Case	Gender/Age	Diagnosis	Coexistence	WBC (109/L)	HB (g/L)	PLT (109/L)	Thrombosis related events	Splenomegaly
1	M/53	PV	JAK2 V617F	10.3	236.	751	N	Y
2	F/55	PV	JAK2 V617F	9.4	191	306	Cerebral thrombosis	N
3	M/59	PV	JAK2 V617F	14.9	212	273	N	Y
4	F/69	PV	JAK2 V617F	8.5	180	403	Myocardial infarction	N
5	M/53	PV	N	4.02	191	119	Calf vein thrombosis	N
6	M/57	PV	N	6.41	196	173	N	N
7	M/65	PV	N	5.92	187	178	N	Y
8	M/50	PV	N	4.04	198	212	N	N
9	M/71	ET	JAK2 V617F	9.73	152	861	N	N
10	F/30	ET	JAK2 V617F	9.5	132	740	Habitual abortion	Y
11	F/47	ET	JAK2 V617F	9.3	125	613	N	N
12	F/32	ET	JAK2 V617F	10.5	153	1544	N	Y
13	F/22	ET	p.K385fs*47 (type-2)	10.3	191	751	N	Y
14	M/73	ET	MPL W515L	4.42	129	1082	Intestinal bleeding	Y
15	M/41	ET	N	11.2	141	1474	N	Y
16	F/69	PMF	N	15.1	76	406	Pulmonary embolism	Y
17	F/54	PMF	JAK2 V617F	16.5	116	711	N	Y
18	M/62	PMF	NRAS (p.Gly12Ser)	29.9	107	116	N	Y
19	M/50	CML (ET)	BCR/ABL	13.82	103	424	N	Y

Structural features of CALR E381A based on bioinformatic analysis

The c.1142A >C variant is a missense mutation occurring in the C-terminus of calreticulin and leads to the Glu381Ala amino acid substitution (Figure 2A). ProtParam tool showed that this point mutation did not significantly change the physicochemical properties of the protein compared with that of the wild-type one, as seen in Table 3. Both proteins were acidic, negatively charged, hydrophilic and unstable.

Figure 2.

Structural features of CALR E381A. A. The E381A mutation occurred in the C-terminal domain of calreticulin. B. Presentation of residues with forces vectors before and after the E381A substitution. C. Glu and Ala at amino acid 381 of calreticulin depicted as a Van der Waals surface (carbon atoms, grey; oxygen atoms, red; nitrogen atoms, blue).

Table 3.

Physicochemical properties of the wild-type calreticulin and the p.E381A variant

Parameters		Wild type	p.E381A
Number of amino acids		417
Molecular weight		48141.56	48083.82
Theoretical pI		4.29	4.3
Number of negatively charged residues		109	108
Number of positively charged residues		50	50
Total number of atoms		6598	6592
Estimated half-life	(Mammalian reticulocytes, in vitro)	30 hours
	(Yeast, in vivo)	>20 hours
	(Escherichia coli, in vivo)	>10 hours
Instability index		47.56	46.78
Aliphatic index		59.14	59.38
Grand average of hydropathicity		-1.104	-1.092

Swiss-Pdb Viewer showed that the formation of intermolecular hydrogen bonds did not change, but the space occupied by the amino acid side chain became smaller with the mutation, resulting in reduced steric hindrance (Figure 2B). Concomitantly, both the Van der Waals force and the energy decreased at amino acid 381 of calreticulin (Figure 2C; Table 4). Glu is a negatively charged, polar, hydrophilic amino acid. The substitution of Ala for the Glu residue resulted in a neutral, nonpolar, hydrophobic position with a lower energy state. This was predicted to be possibly deleterious based on its combined annotation-dependent depletion (CADD) score of 20.4.

Table 4.

Energy computations for Glu and Ala at amino acid 381 of calreticulin determined using the GROMOS96 simulation software of the Swiss-Pdb Viewer

Residue	Bond (KJ/mol)	Angle (KJ/mol)	Torsion (KJ/mol)	Improper torsion (KJ/mol)	Total (KJ/mol)
Glu	0.969	2.991	3.332	1.62	8.912
Ala	0.518	1.196	0.942	1.617	4.273

Note: In this simulation, all computations were done in vacuo without reaction field.

Functional analysis of CALR E381A

We generated the UT-7/mpl cell line by lentivirus transduction, which stably expressed TPO receptor (MPL) and could maintain proliferation in a TPO-dependent manner. When UT-7/mpl cells with overexpression of constructs were cultured for 5 days in TPO-free medium, the number of viable cells harboring CALR E381A or control decreased gradually, only cells harboring CALR del52 showed the proliferation of live cells (Figure 3). These results indicate that CALR E381A was not sufficient to alter growth characteristics of UT-7/mpl as CALR del52 did.

Figure 3.

Growth assay of UT-7/mpl cells with CALR E381A. Live cells with CALR E381A were counted everyday using an automated cell counter in the absence of TPO compared with UT-7/mpl cells harboring CALR del52 or empty plasmid.

Discussion

Due to the mutation of genes in the human body, the active proliferation of one or more lineages of myeloid cells mainly occurs in patients with myeloproliferative tumors. The findings of clonal molecular markers, such as JAK2, MPL and CALR, are of great significance for understanding the molecular pathogenesis, diagnosis and treatment of MPNs. Detection of genetic mutations has become the main diagnostic method for MPNs. The prevalence of JAK2V617F mutation is as high as 95% in PV patients and about 50%-60% in ET and PMF patients. The mutation of JAK2V617F gene occurs at 1849^th nucleotide in No. 14 exon of the JAK2 gene, changing from guanine G to thymine T, resulting in the 617^th valine codon of the pseudokinase domain (JH2) being replaced by phenylalanine acid, which eliminates the negative regulation of tyrosine kinase activity. At present, JAK2V617F gene mutation has become a diagnostic marker of MPN. Moreover, the MPL gene, located at 1p34, encodes a TPO receptor through JAK-STAT signaling and plays an important role in the proliferation and differentiation of megakaryocyte, and platelet maturation. It was found in patients with ET and PMF that the mutation of this gene occurred at the 515^th amino acid changing from tryptophan (W) to leucine (L) or lysine (K), which was presented as MPL-W515L/K. Rare mutations include MPLW515A/R and MPLS505N. These mutations appear as gain-of-function mutations. At present, many studies at home and abroad pointed out that patients with different driver gene mutations had significantly different disease characteristics [19]. For example, patients with positive CALR mutation had a younger age of onset, a lower risk of thrombosis, and a relatively better prognosis in contrast to patients with JAK2V617F mutations. It is characterized by a relatively better prognosis [20]. Previous studies mostly believed that MPNs patients only showed mutations in one of the above three driver genes, and the driver genes were mutually exclusive [21]. However, in recent years, MPNs cases with co-expression of JAK2V617F/CALR and JAK2V617F/MPL have been reported [22]. Double mutation of driver genes suggests that the driver gene may not alone cause the pathogenesis of MPN, and the mutations at different sites may promote each other and cause the occurrence of disease together.

CALR gene, located at 19p13, is a highly conserved endoplasmic reticulum molecular chaperone, mainly involved in the regulation of calcium ion balance, protein folding and processing. It was found CALR frameshift mutations existed in JAK2/MPL-negative ET and PMF patients (50%-60% ET and 75% PMF) [23]. This mutation is an insertion/deletion of exon 9 and results in a specific C-carboxy terminus lacking the endoplasmic reticulum retention sequence KDEL, leading to altered biological function. It was also found that the specific C-carboxy terminus blocked the P-domain, which prompted the N-domain in mutant CALR to preferentially interact with MPL, further inducing JAK2 activation and cytokine-independent growth [24]. Constitutive activation of the receptor by this mutant molecular chaperone is considered as a novel molecular mechanism of cellular transformation in MPNs [25]. There are more than 50 CALR frameshift mutations reported in previous literature. The most common mutation types are the type-1 variant (p.L367fs*46) caused by a 52-base deletion (C.1099_1150del52bp) and type-2 variant (p.K385fs*47) caused by a 5-base TTGTC insertion (c.1154_1155insTTGTC).

We unexpectedly found that CALR E381A was not rare in MPN. As a SNP, CALR E381A has been previously identified in East Asian patients with ET [26]. In this study, the frequency of CALR E381A in patients with PV or ET was significantly higher than that as a SNP, especially in the PV patients, with a frequency as high as 5.4%.

It is noteworthy that CALR E381A not only coexists with JAK2 V617F in PV, but also appears to be a “collaborator” that could coexist with BCR/ABL, MPL W515L, NRAS and even a frameshift mutation of its own gene, according to the genetic assay results for 19 patients. Based on previous research [27], it seems reasonable to believe that CALR E381A as a germline mutation, is an early event, while other mutations including JAK2 V617F, BCR/ABL and frameshift mutations are later events in MPN pathogenesis.

Generally, it is considered that frameshift mutations in CALR are pathogenic for leading to a change in the C-terminus peptide with the loss of most of the acidic domain and the KDEL motif [28]. CALR del52, as a representative of stereotypical frameshift mutations in CALR, is known to abnormally bind with and activate MPL which continues to act on JAK2/STAT5 signaling as well as on the MAPK pathway, justifying the cell hyper-proliferation [29]. Structurally, CALR E381A is a point mutation that maintains C-terminus and KDEL motif in CALR, which is distinctly different from the stereotypical frameshift pattern. Cell proliferation assay also confirmed that CALR E381A showed a failure of cell growth without TPO supplementation, whereas the situation from frameshift CALR 52del was rather different. We suppose that CALR E381A may not be sufficient to be a driver mutation, but contribute to clonal evolution events, like a predisposing factor that offers “fertile ground” for pathogenic somatic mutations in the early pathogenesis of MPN [30]. Given that the effects of CALR E381A was investigated by transient overexpression in vitro, further studies are needed to determine the long-term role of CALR E381A in vivo.

Bioinformatics analysis showed that CALR E381A changed the electrical charge, steric hindrance and energy state at position 381 in the C-domain of calreticulin, which is an important amino acid site and has been reported of anomalies repeatedly, even though its function has not yet been determined [31-33]. The property of the acidic C-domain of calreticulin binding multiple calcium ions with low affinity has been corroborated [34]. Thus, it makes sense to assume that CALR E381A may alter the affinity of the C-domain binding with Ca²⁺, thereby affecting a range of biological processes. The detailed mechanism underlying this process is worth of exploration.

In conclusion, CALR E381A not only has a significantly higher frequency in patients with PV or ET than that as a SNP, but also is able to coexist with various genetic variants including JAK2 V617F. It may serve as a marker that would be beneficial in the diagnosis of MPN and a risk SNP implying an inherited predisposition for the development of MPN in East Asian populations. However, our study had some limitations. First, a small number of patients were included in the study. Second, the effects of other concomitant somatic mutations were not considered. Third, the animal experiments of MPN were not conducted. Finally, a comparison with JAK2-mutated MPN was not performed. More experiments are needed in future for further confirmation.

Disclosure of conflict of interest

None.