1 Introduction
Prader–Willi syndrome (PWS) is a rare congenital disease, and PWS patients are categorized into several subtypes based upon the deletions of paternal chromosome 15q11.2–q13 [
1]. Type I and type II deletions are the major subtypes and account for 65%–75% of patients with PWS. Maternal uniparental disomy (20%–30%) and imprinting defects (2%–3%) comprise another two subtypes [
2]. The prevalence of PWS is 1 in 10 000 to 30 000, with comparable rates for males and females [
3–
5]. The clinical features of PWS vary with age and include severe hypotonia, poor sucking reflex, and feeding difficulties in early infancy. In addition, although infants with PWS manifest cognitive impairment, distinctive behavioral phenotypes, hypogonadism, short stature, and characteristic facial features, morbid obesity is the predominant symptom of children with PWS children over three years of age. For some PWS patients, a diagnosis may not be determined until morbid obesity is noted. In addition, while PWS can be diagnosed from the typical symptoms in children under three, PWS in patients without typical symptoms may not be easily diagnosed prior to three years of age [
6]. Growth hormone is currently the most widely implemented therapeutic approach for PWS patients.
Some PWS patients reportedly exhibit underestimated symptoms—including anemia in adults [
7]—and this hints possible defects in the erythrocytes of these patients. RBC anomalies have been demonstrated to be associated with arterial and venous thromboses by promoting thrombus formation and enhancing thrombus stability [
8]. PWS patients have also been shown to possess a higher incidence of thrombosis based upon a nationwide cohort study in Denmark [
9]. However, as the relationships between red blood cell defects and thrombosis in PWS patients remain speculative and controversial, the issue deserves greater study in the future. The reduced oxygen-carrying capacity of defected erythrocytes undoubtedly exacerbates nervous system-related symptoms [
10], which are the major indications in PWS patients. The defective genes in PWS thus participate in a series of biologic processes that may then affect the development or stability of erythrocytes. Thus, it is of paramount importance to examine erythrocyte-related problems in PWS patients and to ascertain the pertinent underlying mechanisms of action. We herein therefore interrogated the likely defects in RBCs from both older patients and those younger than three years of age, and embarked on an evaluation of the mechanisms responsible for the RBC phenotypes that are due to an impairment in the erythrocyte membrane skeletons and lipid bilayer. Our data collectively revealed that erythrocyte deformation may represent a novel auxiliary indicator for the early diagnosis of PWS.
2 Materials and methods
2.1 Subject enrollment and collection of clinical data
Three groups of subjects—healthy controls (n = 33), obese controls (n = 16), and PWS patients (n = 39)—were enrolled between 2019 and 2021 from the Pediatric Outpatient and Health Examination population of the Shandong Provincial Hospital Affiliated to Shandong First Medical University under the Han Chinese genetic background, with strict inclusion and exclusion criteria. Individuals who possessed other diseases such as hypothyroidism, growth hormone deficiency, infections, sexual precocity, or who received PWS-irrelevant treatments were excluded.
The controls were divided into healthy control and obese control groups based upon the China BMI-for-age SD curve that encompasses age, sex, weight, and height. The PWS patients in the current study were diagnosed using methylation-specific, multiplex ligation-dependent probe amplification (MS-MLPA). Children with a BMI more than two standard deviations above the mean (+2SD) were defined as clinically obese, while those with a BMI between two standard deviations (−2SD to +2SD) were defined as having normal weight.
All participants provided informed consent before blood sampling and clinical data collection. This study was approved by the Medical Ethics Committee of Shandong Provincial Hospital, with ethical approval number 2019-147.
2.2 Immunoblotting assay for ghost proteins and actins in erythrocytes
Erythrocytes from venous blood were separated from plasma by centrifugation at 2000 rpm at 4 °C for 15 min, followed by three washes with washing buffer (10 mmol/L HEPES-Na and 0.1 mmol/L EDTA at a pH of 7.5). We then mixed 100 μL of the collected erythrocytes with lysis buffer (2.5 mmol/L HEPES-Na and 0.1 mmol/L EDTA at a pH of 7.5) at a proportion of 1:40. Erythrocyte membranes were obtained by centrifuging cells three times at 15 000× g for 20 min at 4 °C.
The cell membranal skeleton of erythrocytes is a pseudohexagonal meshwork of ghost proteins that include spectrin, actin, protein 4.1R, ankyrin, and actin-associated proteins. These ghost proteins [
11] laminate the inner membrane surface and attach to the overlying lipid bilayer of erythrocytes. For ghost-protein immunoblotting assay, we incubated the membranes with cold lysis buffer for 20 min, and then the extracted ghost proteins were subjected to SDS-PAGE and stained with colloidal Coomassie Blue or blotted with primary antibodies. The primary antibodies we employed were anti-spectrin α and anti-band 3 from Abcam (England); anti-spectrin β from Santa Cruz (USA); anti-protein 4.1, anti-dematin, and anti-stomatin from Proteintech (USA); and anti-EPB42 from Novus (USA).
We added four volumes of Triton-lysis buffer containing 2.5% Triton X-100 to the immunoblotting assay for F-actin and G-actin to extract erythrocyte protein. Supernatants (soluble fractions) were then separated from the pellets (membrane skeleton fractions) by centrifugation, followed by SDS-PAGE and blotting using primary antibodies generated against F-actin and G-actin (Proteintech, USA).
2.3 Lipidomic analyses of erythrocyte membrane lipids
The membrane lipid profiles of erythrocytes were obtained by performing a high-coverage targeted lipidomics analysis using high-performance liquid chromatography (HPLC) coupled with multiple reaction monitoring (MRM). Lipids were extracted from membranes using a modified Bligh and Dyer extraction procedure and dried in a SpeedVac in OH mode. All lipidomic analyses were performed on an Exion LC-system coupled with a QTRAP 6500 PLUS system according to the standard protocols of the manufacturer (SCIEX, USA). The steps mentioned above were completed by LipidALL Technologies Co., Ltd.
Individual lipids from different lipid classes were quantified relative to their respective internal standards. We conducted multivariate statistical analysis using orthogonal partial least squares-discriminant analysis (OPLS-DA). Differential lipids were identified with a variable importance on projection (VIP) value ≥ 1 and P ≤ 0.05.
2.4 Statistical analysis
We applied unpaired two-tailed Student’s t tests to examine the differences in two groups using the GraphPad Prism v9 software, and we defined statistical significance at a P value < 0.05. The distribution of parameters was analyzed with the Shapiro–Wilk’s test using SPSS, version 25. For data that conformed to a normal distribution, a parameter test was used directly, and for data that did not conform to a normal distribution, t-test was used after log-transformation.
3 Results and discussion
3.1 Deformed erythrocytes and mild anemia noted in children over three years of age with PWS
As shown in Tab.1 and Fig.1, 16 PWS patients (11 males and 5 females) and 16 obese control (OC) individuals (11 males and 5 females) were enlisted. Considering that obesity could quantitatively and qualitatively influence erythrocytes [
12–
14], 16 healthy control (HC) individuals (8 males and 8 females) were enlisted as normal controls (Fig.1 and Tab.1). Of note, no significant difference was found between two genders (Fig.1). Importantly, the BMI did not differ significantly between obese controls and PWS patients while BMI in healthy control group was lower than another two groups. Moreover, no significant difference was found in ages, liver functions, the levels of plasma lipids (triglycerides (TG), cholesterol (CHOL), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C)), and blood glucose among three groups (Fig.1 and Tab.1).
Since some PWS patients with hypothyroidism received symptomatic treatment, their thyroid functions were basically recovered despite a lower level of thyroid-stimulating hormone (TSH) relative to that in the HC and OC groups (Tab.1). In addition, growth hormone treatment in this study was administered to PWS patients, leading to the recovery of insulin-like growth factor-1 (IGF-1) levels (Tab.1). Thus, the impact of diminished IGF-1 on RBCs might be limited and was thus not considered.
Using this premise we sought to investigate RBC phenotypes inspired by reported anemia in adults with PWS. Although no significant difference was observed in RBC number; hemoglobin (HGB) concentration, hematocrit (HCT), and mean corpuscular volume (MCV) were reduced by 4%, 7%, 7%; and 6%, 10%, 6% in the PWS group compared with the HC and OC groups, respectively (Fig.1–1D, P < 0.05), indicating slight anemia in these PWS patients. In support of this finding, the serum levels of erythropoietin (EPO), an indicator of anemia and hypoxia, were increased by 8% and 6% in the PWS group compared with the HC and OC groups, respectively (Fig.1, P < 0.05). The serum iron level was also significantly decreased by 34% and 26% in the PWS group compared with the HC and OC groups, respectively (Fig.1, P < 0.05), implying a limited iron supply for erythropoiesis. These results therefore indicated a mild iron-deficiency-associated microcytic anemia phenotype in RBCs from the PWS patients over three years old that was not found in the HC and OC groups.
Intriguingly, we observed a significant rise in RBC distribution width (RDW) by roughly 15% in the PWS patients relative to controls (Fig.1,
P < 0.05). RDW reflects range variation in RBC sizes and depicts morphologic changes and deformation [
15]. With these encouraging observations, we obtained blood smears that showed marked morphologic abnormalities in RBCs from PWS patients, including deformed RBCs with a bull’s-eye appearance and RBC debris (Fig.1). These results implied an elevated fragility of RBCs from PWS patients. We then deployed the osmotic fragility assay with fresh blood as described previously [
16], and observed a considerably augmented osmotic fragility in RBCs taken from patients with PWS and immersed in 0.30%–0.45% NaCl; however, this fragility was not found in RBCs from individuals in the HC and OC groups (Fig.1,
P < 0.05). These results thus revealed deformed RBCs and increased susceptibility to hemolysis upon osmotic stress in PWS patients over the age of three years.
When we addressed the clinical value of erythrocyte defects in the diagnosis of PWS, we observed that the ratio of RDW to HCT (RDW/HCT) was markedly increased in PWS patients by 24% and 28% in comparison to the HC and OC controls, respectively (Fig.1, P < 0.05). This ratio clearly discriminated the PWS patients from the control individuals without any overlap (Fig.1). Furthermore, the RDW/HCT ratio was found to be positively correlated with the results of MS-MLPA (Fig.1), showing a much more robust correlation than with the other parameters. Given that obesity and hormonal disorders are the typical manifestations in PWS, we further examined the receiver operating characteristic (ROC) curves for the indices of RDW/HCT ratio, BMI, FT4, and IGF-1, where the area under the ROC curve (AUCROC) represented the predictive power in recognizing PWS (Fig.1). As a result, the AUCROC value for RDW/HCT was much greater than that for the other indices, suggesting that this ratio possessed a superior ability in the diagnosis of PWS (Fig.1).
Our collective results revealed mild anemia, defective erythrocytes, and slight hemolysis in children over three years of age with PWS; these phenotypes have not been previously reported for PWS children. More importantly, RBC deformation—in particular the RDW/HCT ratio—may be employed clinically as an auxiliary index in the early diagnosis of PWS. However, it should be aware that the sample size of PWS subjects was relatively small and not matched with the sizes of HC and OC groups due to the rare incidence of PWS and the difficulty of blood sampling from the PWS children. More PWS subjects should be enrolled to strengthen these findings in the future.
3.2 Similar RBC disorders in PWS patients under three years of age.
We next tested the applicability of RBC deformation in PWS children under three years of age, and enlisted 10 PWS patients below three years of age (6 males and 4 females) and also healthy controls of comparable age (8 males and 8 females) (Fig.2). The weights of the PWS patients and controls were within the normal range based on the China BMI-for-age SD curve. Analogous to the observations for PWS children over three years of age, we also uncovered mild microcytic anemia in the PWS children under three years of age as evidenced by a reduction in HGB by 3%, HCT by 4%, MCV by 6%, and MCH by 3% compared with the HC group (Fig.2–2E, P < 0.05). We also noted consistent increases in RDW and RBC deformation in the younger PWS patients (Fig.2 and 2G, P < 0.05), indicating the occurrence of RBC disorders at early ages even below three years. The RDW/HCT ratio was increased by 18% in PWS patients under three years of age, consistent with our observation of an elevated RDW/HCT ratio in PWS patients over three years of age. Furthermore, we could apply the RDW/HCT ratio to discriminate the PWS patients under three years of age from the age-matched control individuals (Fig.2, P < 0.05). Additional correlation analysis corroborated the significance of the RCW/HCT ratio in recognizing PWS at an early stage (Fig.2).
We herein ascertained that the mild anemia and erythrocyte deformation in PWS patients under three years of age were in agreement with the same indices observed in patients older than three years of age. The augmented RDW/HCT ratio was also applicable to the early stages of PWS prior to developing obesity. These findings provide an alternative strategy beyond obesity and genome sequencing for the early diagnosis of PWS, particularly for those PWS patients lacking classical complications.
3.3 Compromised membrane skeletal assembly in RBCs from children with PWS
The membrane skeleton of erythrocytes essentially dictates the structure of the lipid bilayer and endows the membrane with the durability and flexibility to survive in the circulation [
11,
17,
18]. Erythrocyte membrane skeleton is principally composed of spectrin, actin, and their associated proteins (tropomyosin, tropomodulin, adducin, and dematin); protein 4.1R family members; and ankyrin. The spectrin–actin-based membrane skeleton is linked to the plasma membrane through a variety of protein–protein interactions, principally the ankyrin and actin-protein 4.1 junctional complexes. The dynamic balance in the ratio of G-actin to F-actin maintains the stability of membranal structures while factors that disturb the assembly of F-actin lead to the loss of connectivity between ghost proteins [
11].
We then examined the RBC ghost proteins through SDS-PAGE analysis, and as shown in Fig.3, we noted a visible diminution in ghost proteins of approximately 15–40 kDa and 55 kDa in RBCs from PWS patients compared with those from the HCs. To verify these observations, we conducted Western blotting analysis on the ghost proteins spectrin α and β, band 3, proteins 4.1 and 4.2, dematin, and stomatin (Fig.3). Congruent with the SDS-PAGE staining results, most ghost proteins were reduced in PWS RBCs relative to the HC RBCs—particularly with respect to protein 4.1, dematin, and stomatin (Fig.3). In support of our findings, dematin deficiency was previously reported to cause RBC ghost instability and hemolysis [
19]; and the loss of protein 4.1 [
20] and stomatin [
21] also led to severe erythrocyte deformation in mouse models.
As summarized in Fig.3,
MAGEL2—a gene completely deleted in our PWS patients—has been documented to be involved in F-actin assembly [
22]. As a partner of ubiquitin E3 ligases, MAGEL2 regulates the aggregation of F-actin in endosomes, and
MAGEL2 deletion leads to the disruption of endosome-mediated communication between the plasma membrane and cellular organelles [
22,
23]. Furthermore, loss of
MAGEL2 also inhibits the fundamental physiologic processes intrinsic to neurons, such as migration and axonal outgrowth [
23,
24]. Thus, it could be assumed that genetic deletion of
MAGEL2 in PWS patients would result in disrupted assembly of the RBC membrane skeleton by inhibiting F-actin formation. In support of our hypothesis, we observed a reduction in F-actin and an increase in G-actin in PWS erythrocytes compared with control erythrocytes (Fig.3). These findings demonstrated that the defects in the ghost proteins and the compromised membranal skeleton assembly were due to inhibited F-actin formation in PWS patients.
3.4 Dysregulated lipid homeostasis contributes to an unstable plasma membrane in PWS
We subsequently determined the changes in other cardinal components of the plasma membrane (i.e., lipids) in the coordination of ghost proteins. Membrane lipids are indispensable for the function, stability, and fluidity of the erythrocyte membrane [
25]. Actin-based superstructures intimately interact with the plasma membrane in a variety of cellular locations that include highly curved regions of the plasma membrane, providing directionality to actin polymerization [
26]. To explore the likely alterations in lipid profiles, we executed targeted lipidomics on the erythrocytes from PWS patients and from the HC and OC individuals. As shown in Fig.4 and 4B, the membrane lipid compositions of erythrocytes in PWS patients were distinct from those in the control groups, revealing altered membrane lipid profiles in PWS erythrocytes. Compared with the HC group, free fatty acids (FFAs) constituted the most significantly changed membrane lipid class in erythrocytes from PWS patients, followed by diacylglycerol (DAG), sphingomyelin (SM), lysophosphatidylcholine (LPC), and lysophosphatidylethanolamine (LPE) (Fig.4). Phosphatidylethanolamine (PE), FFAs, and DAG comprised the top three classes of significantly changed membrane lipids in the erythrocytes from PWS patients compared with the OC group (Fig.4). Although DAG regulates the calcium pump on the erythrocyte membrane, our data indicated no involvement in the maintenance of erythrocyte deformability or stability [
27]. As one of the major membrane phospholipids, SM (together with cholesterol) constitutes the predominant constituent of lipid rafts [
28]. However, there was no significant change in cholesterol in PWS erythrocytes. PE is predominantly present on the cytoplasmic side of the membrane bilayer, and PE externalization induces the phagocytosis of erythrocytes. However, as enzymatic methylation of PE increases erythrocyte membrane fluidity [
28], we next focused on the effects of modifications in FFAs, LPE, and LPC on the erythrocyte membrane from PWS patients.
As shown in Fig.4 and 4F, the content of unsaturated FAs—but not of saturated FAs—in erythrocytes was significantly attenuated in PWS patients compared with that in the controls (
P < 0.05). As a crucial component of the phospholipids in plasma membranes, unsaturated FAs are necessary to maintain membrane fluidity; and high levels of ω-6 polyunsaturated FAs were previously reported in the membranes of erythrocytes from obese children [
29]. Thus, reduced unsaturated FAs in erythrocytes from PWS patients cannot account for the observed obesity. Similarly, the LPE and LPC contents were significantly decreased in PWS erythrocytes compared with those in the controls (Fig.4 and 4H,
P < 0.05). Phospholipids exhibit a cylindrical shape due to the comparable sizes of hydrophilic head groups and fatty acid chains, while LPE and LPC exhibit cone shapes due to their large polar head and thin hydrophobic tail [
30]. Thus, the migration of cone-shaped LPE and LPC to the outer leaflet of the bilayer increases the outer surface area and enhances the intrinsic curvature of the plasma membrane (Fig.5), and the decrease in LPE and LPC thereby greatly compromises the membrane fluidity of erythrocytes. Membrane skeleton interacts intimately and communicates bidirectionally with membrane lipids, and phosphoinositol phospholipids provide directionality to actin polymerization. However, protein changes in the membrane skeleton would also be expected to cause alterations in lipid compositions. Our results collectively revealed simultaneous membrane lipid disorders in addition to a disrupted membrane skeleton in erythrocytes from patients with PWS (Fig.5). We postulate that disordered membrane lipid profiles contribute to the reduction in membrane fluidity and ultimately generate the defect characteristic of RBC morphology and the lowered anti-stress capacities observed in patients with PWS.
4 Conclusions
We uncovered slight anemia in children with PWS, significant RBC deformation as reflected by hemolysis, increased RDW and conversely diminished HCT, and a greatly elevated RDW/HCT ratio in PWS patients relative to healthy controls or individuals with conventional obesity. Mechanistically, we discerned a compromised membrane skeletal assembly and membrane lipid composition that partially accounted for the observed RBC disorders in these PWS patients; this was likely due to genetic alterations or abnormal DNA methylation. To summarize, our findings resolved the molecular bases underlying RBC deformation and anemia and provided a proof-of-concept indicator in screening for PWS during early childhood.
PDF
(3498KB)
