Potential regulatory role of calsequestrin in platelet Ca2+ homeostasis and its association with platelet hyperactivity in diabetes mellitus


Dali Luo, Department of Pharmacology, School of Chemical Biology & Pharmaceutical Sciences, Capital Medical University, District of Feng Tai, Street of Youanmenwai, #10 Xitoutiao, Beijing 100069, China.
Tel.: +86 10 83911517; fax: +86 10 83911517.
E-mail: luodl@ccmu.edu.cn


Summary. Background: Altered Ca2+ homeostasis contributes significantly to platelet hyperactivity in diabetes mellitus. Calsequestrin (CSQ), as a Ca2+ buffer protein in the sarcoplasmic reticulum, also regulates the Ca2+ release process in muscles. We hypothesized that CSQ may be expressed in platelets, but is altered and involved in diabetic platelet Ca2+ abnormalities and hyperaggregability. Methods: CSQ expression in platelets from streptozotocin-induced type 1 diabetes rats, type 2 diabetes volunteers and Goto-Kakizaki rats were analyzed by western blotting and RT-qPCR. Platelet Ca2+ and aggregation were evaluated with Fura2 and an aggregometer, respectively. Results: Platelets from diabetic patients and rats exhibited increased resting Ca2+ levels, and hyperactive Ca2+ and aggregation responses to agonists. This enhanced basal Ca2+ was largely dependent on intracellular Ca2+ and insensitive to inositol 1,4,5-trisphosphate receptor (IP3R) antagonism. Additionally, the expression of the skeletal CSQ isotype (CSQ-1) was detected in both rat and human platelets, but its levels were significantly lowered in diabetic platelets as compared with normal platelets. Impairment of CSQ by trifluoperazine caused concentration-dependent Ca2+ release in normal platelets and HEK293 cells. Knocking down CSQ-1 in HEK293 cells resulted in increased leakage of Ca2+, which was also insensitive to IP3R inhibition, and exaggerated Ca2+ release following carbachol treatment. Conclusions: Downregulation of CSQ-1 in diabetic platelets and impairment of CSQ-1 in normal cells leads to disturbed Ca2+ release, demonstrating a potential role for CSQ-1 in the regulation of the platelet Ca2+ release process and a possible causal contribution to diabetic platelet hyperactivity.


Intraplatelet Ca2+ ([Ca2+]i) regulates a variety of platelet functions, including shape change, secretion, aggregation, thromboxane production, and rapid surface expression of adhesion molecules. Platelets become hyperactive in Ca2+ signaling and aggregation during the course of diabetes mellitus, thereby contributing to embolization of platelet–platelet or platelet–leukocyte aggregates, atheromatous lesions, atherogenesis and thrombogenesis, and, ultimately, to visible damage to blood vessel walls [1–3].

The total profile of deregulated [Ca2+]i in the platelets of diabetes patients is enhanced Ca2+ release and Ca2+ influx in response to various agonists [2–5]. Alteration in the activities of Ca2+ effectors, including store-operated Ca2+ channels (SOCs) [6], Na+/Ca2+ exchangers (NCXs) [4], and Ca2+-ATPases in the plasma membrane and endoplasmic reticulum (ER) [5,7], have been identified and proposed to be independently responsible for the disrupted Ca2+ signaling in diabetic platelets. In addition, although the underlying mechanism is unknown, platelets from diabetic patients also show greater [Ca2+]i at baseline than healthy platelets [4–8], which may contribute, at least in part, to the spontaneous platelet aggregation [8] and resistance to antiplatelet therapy that occur in diabetic patients [3,9].

Calsequestrin (CSQ), which has two isoforms, CSQ-1 (skeletal) and CSQ-2 (cardiac), is a ubiquitous, critically important Ca2+-binding protein found in the sarcoplasmic reticulum (SR) and ER [10]. Recent studies in muscles have indicated that, in addition to acting as a Ca2+ reservoir, CSQ serves as a luminal Ca2+ sensor, regulating ryanodine receptor (RyR)-mediated Ca2+ release via cooperation with the junctional membrane proteins triadin and junctin [10–13]. Upon impairment of CSQ expression or function, increased resting Ca2+ leak, premature spontaneous SR Ca2+ release and inducible ventricular tachycardia occur in the heart [13,14]; similarly, a moderate reduction in SR Ca2+ volume and significant increases in basal [Ca2+]i and Ca2+ release upon stimulation are found in skeletal muscle [15–17]. In some respects, this disturbed Ca2+ release resulting from CSQ deficiency resembles the hyperactive Ca2+ signaling in diabetic platelets. Additionally, recent studies have demonstrated a link between sequence variations in the CSQ gene and susceptibility to type 2 diabetes in humans [18]. Therefore, we hypothesized that, if CSQ is expressed and functions in platelets, it may be involved in the pathogenesis of altered Ca2+ homeostasis in diabetic platelets.

Materials and methods

The study protocol was approved by the Ethics Committee of the Capital Medical University (Beijing, China).

The materials, subject selection, induction of diabetes and real-time RT-PCR are described in Data S1.

Platelet preparation and aggregation

As previously described [19], after collection of platelets from the rats or human donors, the platelet pellet was adjusted with Hepes-buffered saline solution (HBSS) to 1 × 108 cells mL−1 for the Ca2+ signaling test or 3 × 108 cells mL−1 for the aggregation measurement, or stored at − 80 °C for western blot and RT-qPCR analyses. We assumed that HBSS represented 100% aggregation and that the washed platelet preparation represented 0% aggregation. Platelet aggregation after addition of a stimulator was determined as the maximum extent of aggregation.

Ca2+ measurement

As described previously [19,20], [Ca2+]i was measured in platelets and HEK293 cells after they had been incubated with 3 or 1 μm Fura2/AM for 15 min, respectively, at 37 °C. The basal [Ca2+]i was recorded in resting platelets without any stimulation, whereas Ca2+ release and Ca2+ influx in response to different agonists were measured, respectively, in Ca2+-free (0 mm Ca2+ and 0.1 mm EGTA) medium for 5 min and then in 1 mm Ca2+ (platelets) or 1.8 mm Ca2+ (HEK293 cells) containing medium for another 4 min. Changes in [Ca2+]i were expressed as 340 nm/380 nm fluorescence ratio (F340 nm/380 nm) or Ca2+ concentration calibrated according to the method of Grynkiewicz et al. [21]. In some cases, the values of delta [Ca2+]i and delta ratio represent the net increase in [Ca2+]i following treatment.

Because of the variable data for Ca2+ influx measurement in some of the platelet experiments, we also used Mn2+ quenching, measured by a decrease in fluorescence excited at 360 nm, to confirm the result of divalent cation influx across the cell membrane, as previously reported [19].

Preparation of plasmids

Four different targeting protein knockdown sequences based on the cDNA sequence of Homo sapiens CSQ-1 (GenBankTM accession number: NM001232) and a non-specific control sequence were designed with Invitrogen’s internet-based RNAi algorithm, which is available online (https://rnaidesigner.invitrogen.com/rnaiexpress/index.jsp). Sequences were verified with blast to avoid off-target gene silencing. Each of the four oligonucleotides was inserted into the pcDNA 6.2-GW/miR plasmid vector (Catalogue no. K4935-00; Invitrogen, Carlsbad, CA, USA), using the BbsI and BamHI restriction sites. The plasmid with 5′-TCAAACTCATTGGCTACTTCA-3′, showing a similar CSQ-1 silencing efficiency as that found in diabetic platelets, was selected and transfected into HEK293 cells with a previously described procedure [22].

Western blotting

Aliquots (500 μL) of platelet suspensions, the transfected subconfluent HEK293 cells and the minced skeletal and cardiac muscle tissues were lysed in RIPA buffer containing 2 mm phenylmethylsulfonyl fluoride and 2 μg mL−1 protease inhibitor cocktail (Santa Cruz, CA, USA) for 10 min on ice. Lysates (60 μg for platelets and HEK293 cells and 40 μg for muscles) were used for western blot analysis. Two different source antibodies (ABR, Thermo Fisher Scientific Inc. Rockford, IL, USA and Santa Cruz) specific for CSQ-1 and CSQ-2 were used at dilutions of 1 : 300 and 1 : 500, respectively, and antibodies specific for triadin, junctin or pan-inositol 1,4,5-trisphosphate receptor (IP3R) were used at dilutions of 1 : 300–1 : 1000, with an overnight incubation at 4 °C. The horseradish peroxidase-conjugated secondary antibodies were diluted at 1 : 3000 or 1 : 10 000, respectively.


The washed platelets were fixed with 4% formaldehyde and permeabilized with 0.1% Triton X-100. Anti-CSQ and anti-calnexin antibodies were used at a dilution of 1 : 100. The secondary antibodies, Alexa Fluor 488-labeled goat anti-rabbit and Alexa Fluor 594-labeled chicken anti-rabbit (Invitrogen), were used at a dilution of 1 : 500. Chemifluorescence detection was performed directly with a × 63 oil-immersion objective (NA1.4), using a laser-scanning microscope (Leica SP5, Mannheim, Germany).

Statistical analysis

Data are presented as the means ± standard errors of four to 14 independent measurements. Statistical comparisons between groups were carried out with a two-tailed unpaired Student’s t-test. < 0.05 was considered to be statistically significant.


CSQ expression in platelets

CSQ and other SR-associated proteins are well documented in skeletal and cardiac muscles. Thus, by referring to muscle proteins, we examined whether platelets express CSQ and other ER proteins such as triadin, junctin, calreticulin, and calnexin, which appear to affect internal Ca2+ homeostasis [10–12,23]. Four antibodies from two sources, each recognizing different specific peptides of the CSQ isotypes, were used. Actin and calnexin were measured as an indication of equal protein loading.

Consistent with previous reports [10–13], a protein band of ∼ 63 kDa and two bands of ∼ 50 kDa and ∼ 45 kDa were detected in rat skeletal muscle and cardiac muscle, respectively, with the specific antibodies against CSQ-1 and CSQ-2 (Fig. 1A). In comparison, a band of ∼ 63 kDa was detected in rat platelet preparations with antibodies against CSQ-1, and several obvious bands ranging from 50 kDa to 80 kDa with antibodies against CSQ-2. Further RT-qPCR analysis indicated that, by referring to the CSQ mRNAs of cardiac and skeletal muscles, CSQ-1 mRNA, but not CSQ-2 mRNA, was expressed in rat platelets (Fig. 1B). In accordance with this result, immunostaining with specific antibodies against the CSQ subtypes showed intense labeling of CSQ-1, but rather faint CSQ-2 staining (Fig. 1C). Therefore, these data demonstrated that rat platelets express only the CSQ-1 isotype.

Figure 1.

 Effects of diabetes on endoplasmic reticulum Ca2+ protein expression in platelets from STZ rats. (A) Immunoblots of calsequestrin (CSQ) isoforms in platelets from three normal (Con) and three diabetic (DM) rats out of eight and 12 rats, respectively. SM and CM indicate samples from skeletal and cardiac muscle, respectively. (B) CSQ-1 and CSQ-2 mRNA levels were determined, normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and compared with those in cardiac and skeletal muscles by RT-qPCR. n = 4–8 rats for each bar. (C) Immunostaining of CSQ-1, CSQ-2, and calnexin, for comparison, with specific antibodies in normal platelets. n = 3–4 independent experiments. Scale bar: 2 μm. (D, E) Immunoblots of triadin, junctin, calnexin, calreticulin and inositol 1,4,5-trisphosphate receptor IP3R in platelets with specific antibodies (D), and bar graphs summarizing the data obtained from six to twelve rats for each bar after normalization to β-actin and then to normal platelet level (E). The CSQ-1 level was also normalized to calnexin as indicated. **< 0.01 vs. normal platelets in all panels. ER, endoplasmic reticulum.

Next, the relative abundance of CSQ-1 and other ER Ca2+ regulatory proteins in streptozotocin (STZ) rat platelets was determined by normalization with β-actin or calnexin. The results showed that, as compared with non-diabetic platelets, CSQ-1 expression in diabetic platelets was reduced by 24.0% ± 2.5% and 32.1% ± 2.2% at the protein level (= 0.0012 and = 0.0009, normalized by β-actin and calnexin, respectively; Fig. 1A,D,E), and by 72.8% ± 3.7% at the mRNA level (= 0.00033; Fig. 1B), whereas the levels of triadin, junctin, calreticulin (∼ 46 kDa in cardiac muscle), calnexin and IP3R did not change.

CSQ expression was also evaluated in platelets from type 2 diabetes volunteers and GK rats, a genetic rat model of type 2 diabetes [24]. As in rat platelets, CSQ-1 but not CSQ-2 (data not shown) was detected in human platelets (Fig. 2A). Again, a significant reduction in the level of this isoform was observed in diabetic volunteers as compared with gender-matched and age-matched healthy subjects (32.2% ± 4.1%, = 0.0019; Fig. 2B). Although residual mRNAs of glyceraldehyde-3-phosphate dehydrogenase and calnexin were observed, CSQ-1 mRNA was undetectable in human platelets, probably because of an exceptionally low level of mRNA in human platelets (data not shown). GK rat platelets also showed a lower level of CSQ-1 than matched control rat platelets (35.5% ± 3.7%, = 0.0035; Fig. 2C,D).

Figure 2.

 Effects of diabetes on calsequestrin (CSQ) expression in platelets from type 2 diabetes donors and GK rats. (A, B) Immunoblots of CSQ-1 in platelets from three healthy (Con) and three type 2 diabetes (DM) individuals (A), and the levels of CSQ-1 after normalization to β-actin and summarized data (in the panel) from 14 donors for each group (B). (C, D) Immunoblots of CSQ-1 in platelets from three normal and three GK rats (C), and bar graphs summarizing the data obtained from six rats for each group (D). **< 0.01 vs. normal platelets in all panels.

All of these data demonstrate that muscle-like CSQ, in particular CSQ-1, is expressed in both human and rat platelets; however, this Ca2+ storage protein is downregulated in diabetes, suggesting the possibility of a Ca2+ signaling dysfunction in diabetic platelets.

Platelet Ca2+ homeostasis and aggregation

Functional evaluation of platelets was performed in healthy subjects, diabetic patients and rats by measuring basal [Ca2+]i, Ca2+ signaling and aggregation responses to G-protein-coupled receptor agonists, including thrombin (0.05 and 0.1 U mL−1) and ADP (20 μm). To further determine the contributors to platelet Ca2+ release, we also utilized the ER Ca2+ pump inhibitors thapsigargin (TG) and 2,5-di-(t-butyl)-1,4-hydroquinone (BHQ), an inositol 1,4,5-trisphosphate (IP3) analog (IP3/BM, a membrane-permeable IP3), the IP3R antagonist 2-aminoethoxydiphenylborate (2-APB) [20], and nifedipine, an inhibitor of nicotinic acid adenine dinucleotide phosphate receptor (NAADPR). These reagents have been shown to act on different internal Ca2+ stores in platelets; at a low concentration, TG inhibits Ca2+-ATPases of IP3R-operated Ca2+ pools, whereas BHQ affects the pump of NAADPR-mediated Ca2+ stores [25,26].

To identify the contribution of internal Ca2+ and extracellular Ca2+ to basal [Ca2+]i, basal [Ca2+]i was monitored in resting platelets that were incubated in Ca2+-free and then in 1 mm Ca2+-containing medium (Fig. 3A). Diabetic platelets showed higher resting [Ca2+]i than non-diabetic platelets in both conditions. Noticeably, following normalization, the amount of elevated basal [Ca2+]i in diabetic platelets depended more on the internal Ca2+ source (∼80%) than on the external one (∼20%; Fig. 3B), suggesting greater Ca2+ leakage from internal pools at rest in diabetic platelets than in normal platelets.

Figure 3.

 Ca2+ hyperreactivity and hyperaggregability in response to stimulation in platelets from diabetic patients and rats. (A, B) Typical traces illustrate basal intraplatelet Ca2+ ([Ca2+]i) in resting platelets incubated in Ca2+-free medium, and following addition of 1 mm Ca2+ to the medium (A), and the summarized data of normal (Con) and diabetic subjects or rats (DM) (B). Note that (1)/(2) = 79% of the total elevated basal [Ca2+]i in diabetic platelets, representing an internal Ca2+ source-dependent component. (C) Typical traces illustrate the Ca2+ response to thrombin (0.1 U mL−1) in Ca2+-free and Ca2+-containing medium in normal and diabetic platelets. (D, E) Bar graphs summarizing the data of the Ca2+ responses to thrombin, ADP (20 μm) (D), thapsigargin (TG) (0.5 μm), 2,5-di-(t-butyl)-1,4-hydroquinone (BHQ) (20 μm), inositol 1,4,5-trisphosphate (IP3)/BM (20 μm), and ionomycin (5 μm) (E). (F) Thrombin (0.05 and 0.1 U mL−1)-induced and ADP (20 μm)-induced maximal aggregation were assessed in washed platelets from normal and diabetic donors or rats. All of the data were obtained from seven to 14 subjects in each group; *< 0.05 and **< 0.01 vs. normal subjects or control rats in all of the panels.

Next, we evaluated the Ca2+ release response in Ca2+-free medium and Ca2+ influx following the addition of 1 mm Ca2+ induced by thrombin or ADP. As previously reported [4–7], both of the Ca2+ signaling processes were significantly augmented in diabetic platelets (Fig. 3C,D). Additionally, higher Ca2+ release was observed in response to IP3/BM (20 μm), BHQ (20 μm) and TG (0.5 and 1 μm) in diabetic platelets, whereas the total releasable Ca2+, assessed by the Ca2+ response to the Ca2+ ionophore ionomycin (5 μm), remained unaltered (Fig. 3E). This suggests that the Ca2+ storage volume is unaffected, but that Ca2+ release is dramatically upregulated in diabetic platelets, regardless of IP3R stimulation or Ca2+-ATPase inhibition.

In accordance with the alterations in Ca2+ signaling, greater platelet aggregation in response to thrombin and ADP was found with diabetic platelets than with the parallel controls (Fig. 3F). Thus, these results demonstrate that hyperaggregability of diabetic platelets correlates with abnormal Ca2+ activity, manifested by enhanced basal [Ca2+]i, and exaggerated Ca2+ release and Ca2+ entry following agonist treatment.

CSQ function

To determine whether the Ca2+ signaling disorders and the CSQ expression change found in diabetic platelets correlated with each other, we used trifluoperazine (TFP), a tricyclic antidepressant. This drug has been found to prevent Ca2+ binding to CSQ-1 and CSQ-2 with Kd values in the micromolar range, leading to increased Ca2+ efflux from the SR, and muscle-related side effects similar to those observed in CSQ-knockout mice [14–17,27,28]. However, TFP also has some non-specific effects, including inhibition of calmodulin [29], ER Ca2+-ATPase [30] and K+ channels [31] in many cell types, all of which may affect intracellular Ca2+ activity. To distinguish between these non-specific TFP effects and the CSQ inhibition effect, we used N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide (W7), a calmodulin antagonist, and TG, a specific inhibitor of ER pumps but not of plasma Ca2+ pumps. Treatment of normal rat platelets with TFP (10–50 μm) for 10 min in Ca2+-free medium caused a gradual [Ca2+]i increase in a concentration-dependent fashion, whereas W7 (50 or 100 μm) failed to induce any noticeable Ca2+ response (Fig. 4A). Additionally, like W7 (100 μm), 30 μm TFP significantly attenuated the ADP-mediated Ca2+ release (Fig. 4B). This effect is most likely attributable to its inhibition of calmodulin, because TFP-related calmodulin inhibition has been shown to attenuate Ca2+ signaling in many other studies [28,29,31]. Furthermore, depletion of the ER by TFP-induced Ca2+ leakage may further enhance this effect. Similar effects of TFP and W7 were also observed in HEK293 cells, in which TFP, but not W7, elevated basal [Ca2+]i, but both of them inhibited the Ca2+ release induced by carbachol (a muscarinic receptor agonist) in a dose-dependent manner (Fig. 4B).

Figure 4.

 Effects of interfering with calsequestrin (CSQ) expression in normal rat platelets and HEK293 cells. (A) Typical traces show that trifluoperazine (TFP) gradually raised intraplatelet Ca2+ ([Ca2+]i) in resting platelets, and inhibited the subsequent Ca2+ release response to ADP (20 μm) in Ca2+-free medium, whereas N-(6-aminohexyl)-5-chloro-1-naphthalenesulfonamide (W7) (100 μm) only abrogated the ADP-induced Ca2+ release. (B) Summarized data from five to eight samples for each point indicate the dose-dependent effects of TFP and W7 on Ca2+ release in normal rat platelets and HEK293 cells, respectively. Carb in the lower panel stands for carbachol (50 μm), used to induce the Ca2+ release response. (C) Bar graphs show the inhibitory effect of 2-aminoethoxydiphenylborate (2-APB) (75 μm for 5 min) on TFP-induced Ca2+ release but not on that induced by 0.5 μm thapsigargin (TG). (D, E) CSQ-1 expression in HEK293 cells was interfered with by different concentrations of plasmids (1, 1.5 and 2 μg mL−1) carrying small interfering RNA (siRNA) against human CSQ-1. The immunoblots (D) and bar graphs (E) depict the knockdown efficiency analyzed from four independent experiments for each bar. (F–H) Typical traces illustrate the alterations in basal [Ca2+]i (1), and release (2) and influx (3) of Ca2+, in response to 50 μm carbachol in Ca2+-free and 1.8 mm Ca2+-containing medium in HEK293 cells (F). The statistical data obtained from six to eight independent experiments for each point (G) reveal a linear correlation, indicated with R2-values, between the levels of CSQ-1 and elevations in basal [Ca2+]i and the Ca2+ release response to carbachol in CSQ-1-KD cells (H). *< 0.05 and **< 0.01 vs. cells prior to TFP or W7 treatment, or control (Con)-siRNA, and ##< 0.01 in panel C vs. TFP treatment with dimethylsulfoxide (DMSO) as a vehicle.

Therefore, the TFP-induced Ca2+ increase above baseline was probably attributable to internal Ca2+ release and/or inhibition of ER Ca2+ pumps, rather than inhibition of calmodulin or K+ channels, because K+ channel blockade presumably results in secondary Ca2+ entry rather than Ca2+ release [31]. A previous report has demonstrated that TFP inhibits Ca2+-ATPases in the SR of platelets and myocytes at a concentration five times higher than that used for promoting Ca2+ leakage [30]. In our study, TFP at 40 μm, a concentration that does not affect Ca2+-ATPase activity in human platelets [30], elicited similar levels of Ca2+ release as those following platelet exposure to TG (1 μm, data not shown). Additionally, the patterns of Ca2+ release were different: TG induced an obvious Ca2+ release peak that lasted for 2–4 min (data not shown), whereas TFP gradually elevated [Ca2+]i during the 10-min recording (Fig. 4A). Furthermore, the TFP-associated [Ca2+]i increase was significantly sensitive to inhibition of IP3R by 2-APB, whereas Ca2+ release caused by TG was only slightly affected (Fig. 4C). Thus, in contrast to TG, it is likely that TFP mediates Ca2+ leak through its inhibition of CSQ, which is in line with recent findings in muscle studies [17,27,28].

To further confirm the CSQ-associated regulation of Ca2+ release in non-muscle cells, we tested several human cell lines for knockdown evaluation. We selected HEK293 cells, because of the obvious expression of endogenous CSQ-1 in these cells. These cells were transfected with 1, 1.5 or 2 μg mL−1 of plasmids carrying small interfering RNA (siRNA) against CSQ-1 for 48 h (denoted as CSQ-1-KD cells), and this was followed by immunoblotting analysis to determine CSQ-1 levels. The basal [Ca2+]i and Ca2+ responses to 50 μm carbachol were monitored with the same protocol as used in the platelet experiments. As compared with the cells treated with the control siRNA, CSQ-1 expression in CSQ-1-KD cells was interfered with by CSQ-1 siRNA in a dose-dependent manner (Fig. 4D–E), and the decreases were within the range observed in diabetic platelets (20–40%; Figs 1E and 2C–E). As expected, these cells showed significant increases in basal [Ca2+]i and carbachol-induced Ca2+ release (Fig. 4F,G), but no changes in the internal Ca2+ volume (assessed by ionomycin treatment; data not shown) and the Ca2+ influx response, implying that the augmented Ca2+ entry in diabetic platelets may not be attributable to the downregulated CSQ. The enhanced basal [Ca2+]i and carbachol-induced Ca2+ release correlated well with the decrease in CSQ-1 level (see R2-values in Fig. 4H), and the TFP-induced [Ca2+]i increases correlated well with the TFP concentrations used in platelets and HEK293 cells (Fig. 4B). Therefore, all of the data suggest a role for CSQ in the maintenance of internal Ca2+ pool stabilization in both resting and stimulated states in these non-muscle cells.

Target effector(s) of CSQ

Finally, we used pharmacologic methods to investigate possible target(s) for CSQ modulation of basal [Ca2+]i and Ca2+ release in platelets. Because Ca2+ mobilization is predominantly regulated by IP3R and NAADPR in platelets [25,26], we used 2-APB to block IP3R and SOCs [20], and nifedipine to inhibit NAADPRs. As shown in Fig. 5A,C, 2-APB (75 and 100 μm) did not affect basal [Ca2+]i in Ca2+-free medium, but did inhibit the rise in basal [Ca2+]i when Ca2+ was added to the medium in both control and diabetic platelets from humans or rats. Neither nifedipine (3 μm) nor the combination of nifedipine and 2-APB affected the internal Ca2+-dependent basal [Ca2+]i (data not shown). A similar result was also found in CSQ-1-KD HEK293 cells treated with 2-APB (data not shown).

Figure 5.

 Effect of inositol 1,4,5-trisphosphate receptor inhibition on Ca2+ hyperreactivity in diabetic platelets from patients and rats. (A, B) Typical traces illustrate the effects of 2-aminoethoxydiphenylborate (2-APB) (75 μm for 5 min) on basal intraplatelet Ca2+ ([Ca2+]i) (A), and the Ca2+ response to 0.1 U mL−1 thrombin, in normal (Con) and diabetic (DM) platelets (B) in Ca2+-free and then Ca2+-containing medium, respectively. The arrows in (A) indicate the inhibitory effect of 2-APB on the extracellular Ca2+-dependent basal [Ca2+]i. (C, D) Statistical data for the effects of 2-APB on basal [Ca2+]i (C), and the thrombin-induced Ca2+ release and Ca2+ influx (D). n = 5–6 human platelet experiments; n = 10–12 rat platelet tests. Note that 2-APB inhibition efficiency, indicated by IC50 in (D), is greater for Ca2+ entry than for Ca2+ release response, and also greater for normal platelets than for diabetic platelets from rats. **< 0.01 vs. normal platelets, ##< 0.01 vs. dimethylsulfoxide (DMSO).

As expected, 2-APB caused dose-dependent inhibition of thrombin-induced Ca2+ release and Ca2+ entry in platelets from normal and diabetic rats (Fig. 5B,D). Although with less potency for Ca2+ release than Ca2+ influx, and for diabetic platelets than normal platelets (see IC50 values in Fig. 5D), 2-APB at the highest concentrations almost completely blocked the Ca2+ release and Ca2+ entry responses (the equal remaining Ca2+ release responses in normal and diabetic platelets were further abolished by nifedipine; data not shown). Therefore, in agreement with previous reports [2,3,6], the hyperreactive Ca2+ release and Ca2+ influx following stimulation in diabetic platelets were IP3R-mediated and SOC-mediated, respectively.


In this study, platelet Ca2+ signaling and aggregation dysfunction was confirmed in type 1 and type 2 diabetes rats and in type 2 diabetes patients. Importantly, in addition to the observation that the increased basal [Ca2+]i came mostly from internal Ca2+ pools, the major finding of this study was that CSQ, in particular CSQ-1, is highly expressed, and may function as a regulator of ER Ca2+ cycling in platelets. As a result, the downregulation of CSQ-1 expression found in diabetic platelets may cause enhanced Ca2+ release at rest and following stimulation.

In addition to its well-recognized role as a Ca2+ buffer, CSQ also actively participates in the modulation of the Ca2+ release process. Functional studies have shown that, when triadin and junctin are present, CSQ maximally inhibits RyR at physiologic Ca2+ concentrations in the SR lumen, but that a further increase in the Ca2+ concentration to 4 mm results in dissociation of CSQ from triadin and junction, and relief of the inhibitory influence of the quaternary complex (RyR, CSQ, triadin, and junctin) on Ca2+ release [10–13]. When CSQ expression is interfered with, a disturbance in SR function (see Introduction) leads to cardiac arrhythmia and muscle-related adverse effects, such as uncontrollable muscle movements, severe muscle stiffness, and spasms [13,14,17,27,28]. Therefore, a functional role for CSQ in the regulation of Ca2+ release has been well characterized in muscles.

Although CSQ is also expressed in non-muscle tissues and cells, such as neurons [32], liver [33], pancreas [34], and neutrophils and HeLa cells [34], and colocalizes with IP3R and Ca2+-ATPase in the ER, its biological features and function in the regulation of Ca2+ release in these cells are presently unknown. This study identified a protein with a similar molecular mass to skeletal CSQ in platelets from rats and humans, and in HEK293 cells, providing new insights into the regulation of Ca2+ homeostasis by CSQ-1 in non-muscle cells. Indeed, as compared with normal platelets, profound Ca2+ signaling alterations, including increased basal [Ca2+]i and enhanced mobilization and influx of Ca2+ in response to agonists, and reduced CSQ-1 levels, were detected in both type 1 and type 2 diabetes platelets. Knockdown of CSQ-1 in HEK293 cells caused an increase in basal [Ca2+]i and exaggerated Ca2+ release following agonist stimulation. Additionally, platelets and HEK293 cells with TFP-impaired CSQ also exhibited elevated Ca2+ leakage. However, in contrast to the findings in diabetic platelets and CSQ-1-KD cells, TFP further abrogated the agonist-induced Ca2+ release in both platelet and HEK293 cells, owing to its inhibitory effect on calmodulin, an action that has been well characterized in other studies [29–31].

In cardiac myocytes, upregulation of RyR function has been identified as the molecular mechanism for the increased Ca2+ release induced by deficient CSQ [10–13]. This may not be the case in platelets, which lack RyR. The dysfunctional NCX in diabetic platelet [4] is also not responsible, because, except for prolonging the recovery time of agonist-induced Ca2+ release to baseline, inhibition of NCX with KB-R7943 (a specific NCX antagonist) did not change basal [Ca2+]i in diabetic platelets (data not shown). IP3R is probably one of the targets for CSQ-modulated Ca2+ release, because 2-APB not only inhibited the enhanced Ca2+ release following stimulation in diabetic platelets, but also partially blocked TFP-induced Ca2+ release in normal platelets. However, other unknown mechanism(s), which is not affected by either 2-APB or nifedipine, is also involved in the CSQ-related Ca2+ leakage, and requires further investigation.

In summary, the study reported herein demonstrates that CSQ in platelets is involved in the maintenance of internal Ca2+ pool stabilization via negative regulation of Ca2+ release in the resting state and upon stimulation. However, this regulatory effect of CSQ is somehow disrupted by partial CSQ downregulation in diabetes, thereby causing an altered Ca2+ release signal in diabetic platelet. A recent finding that human subjects with variations in the CSQ gene are susceptible to diabetes [18] may provide evidence at the gene level in this regard. Therefore, CSQ is probably a potential contributor to the increased Ca2+ leakage in the resting state and the exaggerated Ca2+ release following stimulation in diabetic platelets. Given the persistent and uncontrollable CSQ-deregulated Ca2+ signaling in diabetic platelets, our findings may also provide a potential mechanism for the spontaneous platelet aggregation and the resistance to antiplatelet therapy that are often observed in diabetic patients.


This study was supported by grants from the National Natural Science Foundation (30973537, 30772574), and the Beijing Natural Science Foundation (7082018).

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.