Influence of diabetes and hypercholesterolemia on laboratory methods for hereditary spherocytosis diagnosis

Abstract Introduction Hereditary spherocytosis (HS) is characterized by decreased erythrocyte deformability resulting in hemolytic anemia. This is a heterogeneous disease regarding underlying protein deficiency, disease severity, age at diagnosis and clinical course. Although largely considered as pediatric disease, HS could be initially diagnosed also in elder patients as a result of gallstones or splenomegaly fortuitous finding. Concurrently, common adulthood metabolic disorders like diabetes or dyslipidemia are also known to impair RBC rheology and deformability. Therefore, we aimed to investigate if these diseases affect the screening and diagnostic tools used for HS diagnosis. Methods We applied our workflow for HS diagnosis on 95 pathological samples: 29 patients with diabetes, 20 with hypercholesterolemia, 17 with dyslipidemia, 6 with hypertriglyceridemia, 23 with metabolic syndrome (MS). Thus, a total of 73 samples were analyzed by automated reticulocyte analysis, 52 by cryohemolysis test, and 41 by ektacytometry osmoscan analysis as we used two out of the three tests for each individual sample. Results Applying our screening algorithm based on automated reticulocyte indices, a total of 4 samples (4.2%): one sample (5%) from the diabetes group and three samples (16.7%) from the MS group, positioned into the HS zone. However, no significant difference was found between any of the pathological groups and the controls for the cryohemolysis test or the osmoscan. Conclusion While diabetes and hypercholesterolemia are pathologic conditions known to present with decreased erythrocyte deformability and disturbed rheology, their possible concomitant presence with HS would not interfere with the screening and confirmatory laboratory methods.


| INTRODUC TI ON
Hereditary spherocytosis (HS) is the most frequent membranopathy and congenital hemolytic anemia in Northern Europe. 1,2 Typical HS presentation includes general signs of hemolysis combined with characteristic RBC morphology that is, spherocytes, family history, altered MCHC, and automated reticulocyte parameters that is, mean reticulocyte volume (MRV), mean sphered cell volume (MSCV), and immature reticulocyte fraction (IRF). 2,3 Novel screening algorithms were proposed, based on easily available and rapid automated hematological parameters such as delta (MCV-MSCV) 4 and MRV, 5,6 as well as IRF isolated or in a ratio to reticulocyte count. 7 These achievements were included in the latest published guidelines for RBC membrane disorders diagnosis. 3 More recent studies confirmed these findings and underlined the place of reticulocyte automated parameters in HS screening. 8,9 Specialized tests, reviewed by Bolton-Maggs et al. 2 and by King et.al. 3 are generally recommended to confirm the diagnosis of HS for example, osmotic fragility test, cryohemolysis test, acidified glycerol lysis test, and eosin 5-maleimide (EMA) binding test. In order to improve their predictive values, the combination of minimum two screening tests has been recommended. 10 Further, another diagnostic tool for HS and other membranopathies, introduced several decades ago, 11 is the osmotic gradient ektacytometry. Its importance was reassessed in the latest guidelines for RBC membrane disorders diagnosis, 3 underlining its ability to differentiate various membranopathies. Approximately at the same period, a next generation ektacytometer was introduced: Laser-assisted Optical Rotational Cell Analyzer, LoRRca, RR Mechatronics, Hoorn, the Netherlands. 12 Its clinical implementation was first reported by Da Costa et al. 13 , followed by our group. 14 Next generation ektacytometry (osmoscan) is currently a recognized method for RBC membranopathies diagnosis as an efficient intermediate step between screening tests, that is, RBC morphology, automated reticulocyte parameters, cryohemolysis, and EMA-5 binding, and confirmatory protein deficiency tests, that is, sodium dodecyl sulfate polyacrylamide gel electrophoresis or DNA-based methods. [14][15][16][17] The main underlying pathophysiological feature of HS is the deteriorated RBC mechanics and decreased cellular deformability based on structurally modified red cell membrane, decreased surface to volume ratio, and loss of elastic modulus with resultant hemolytic anemia. 18 HS is a heterogeneous disease regarding protein deficiency, disease severity, age at diagnosis, and clinical course. 2,18 Although largely considered as pediatric disease, HS could be also initially diagnosed in adults and in elder patients as a result of, for example, gallstones or splenomegaly fortuitous finding. 18 Moreover, common adulthood metabolic disorders like diabetes or dyslipidemia, are known to likewise impair RBC rheology and deformability.
Various studies, reviewed recently, 19,20 have demonstrated association of increased plasma and whole blood viscosity with abnormal RBC membrane structure and intracellular viscosity in diabetes, leading to reduced RBC cell deformability and vascular complications such as diabetic microangiopathic nephropathy and retinopathy.
Furthermore, these conditions were reported to be associated with anisocytosis, reflected by increased red distribution width, RDW. 21 Measurement of RBC deformability is performed by various techniques that could be divided in two groups: (a) individual cell measurements for example, micropipette aspiration, atomic force microscopy, optical tweezers or (b) measurements of multiple cells for example, filtration method, microfluidic filtration, laser diffractometry (ektacytometry). 19 To characterize RBC deformability, one mode is largely in use for membranopathies diagnosis: osmotic gradient ektacytometry. Shear stress (SS) is maintained constant while RBCs are mixed in the medium with gradually increasing osmolality, from 50 to 500 mOsm/kg. Measured diffraction pattern is subsequently a function of osmolality. 11 It allows evaluation of the erythrocyte response to various factors affecting their cell geometry, internal viscosity, or membrane deformability. 13 The ektacytometry has become one of the most applied methods in the field of RBC deformability study due to its accuracy, sensitivity, and convenience. 22 Using osmotic gradient ektacytometry, a slight decrease of RBC deformability of the studied diabetic subjects, together with increased membrane protein glycosylation and oxidatively damaged spectrin, was demonstrated. 23 However, in other studies, rheological behavior of the oldest and youngest 10% fractions of diabetic RBC was examined by classical rheoscope analysis, and it was found to be identical to that of normal cells. 24 Besides, no decreased erythrocyte deformability was found either by filtration or by ektacytometry. 25 Though, impaired deformability of diabetic RBC was demonstrated later by different deformability studies based on filtration and micropipette techniques and on SS diffractometer measurement by Rheodyn SSD (Myrenne GmbH, Roetgen, Germany). 26 Succeeding, a disposable microfluidic ektacytometer, RheoScan-D, (RheoMeditech) 27 combining laser diffraction system and slit rheometer has been introduced for RBC deformability measurement in terms of elongation index (EI) which assessment was in agreement with that measured by conventional ektacytometer. 27 Also, with this equipment, reduced RBC deformability was observed in diabetes, and the EI was even further decreased in diabetic patients with microvascular complications. 28 Furthermore, decreased deformability manifested by increased osmotic fragility and additional membrane protein modifications, that is, a significant quantitative reduction in ankyrin, affecting the cytoskeleton rigidity and cell fragility, and plasma protein glycation and lipid peroxidation were also reported in diabetic RBC. 29,30 Another common adult metabolic disorder, hypercholesterolemia, a cause of accelerated atherosclerosis, is characterized by increased blood cholesterol. Indirect result is disturbed blood rheology: blood viscosity, platelet activation, and RBC deformability may be affected by hypercholesterolemia. 31 Additionally, the same study demonstrated that long-term cholesterol-lowering therapy improved impaired RBC deformability, measured under a weak SS by laser diffractometric ektacytometer. Thus, it was hypothesized that improved RBC deformability may on its turn have beneficial effect on organ perfusion, including microcirculatory coronary blood flow. 31 It should be noted that analyzing erythrocyte deformability under a high SS of 23.6 dyn/cm 2 and greater gave no difference for hypercholesterolemia patients compared to control subjects. 31 In another study, EI was measured in familial hypercholesterolemia adult group without treatment, using Rheodyn SSD, at different SS ranging from 0.30 to 60 Pa. 32 Lower EI values were demonstrated for hypercholesterolemia patients at high SS >6 Pa; however, no differences were observed between patients and controls at lower SS.
In experimental in vitro conditions, the relationship between cholesterol and ATP release from RBCs was studied with a microfluidic flow-focusing device. 33 It was found that decreasing membrane cholesterol increased cell deformability and ATP release; moreover, RBC deformability and ATP release were improved by simvastatin treatment by its direct membrane action of cholesterol enrichment. 33 On the one hand, older patients are often affected by common medical conditions like diabetes or dyslipidemia, known to be characterized, inter alia, by hemorheological alterations and decreased RBC deformability. On the other hand, mild forms of HS could be suspected and diagnosed in adult or elder patient during family studies or investigation of accidentally found enlarged spleen or gallstones. As diabetes and hypercholesterolemia are pathologic conditions also known to disturb RBC deformability and rheology, we aimed to investigate if these diseases affect the screening and diagnostic tools used for membranopathies diagnosis. We focused on the screening tests already implemented in our laboratory, that is, automated reticulocyte parameters, cryohemolysis, and nextgeneration ektacytometry with osmoscan and analyzed if there were interferences on these tests in the presence of hyperglycemia or hypercholesterolemia.

| Patients
Automated reticulocyte count, cryohemolysis test, and ektacytometry osmoscan analysis were performed during a three-month period over a total of 95 samples from out-clinic patients sent to the laboratory as part of their monitoring. Chronic diabetes patients were followed by our endocrinologist department. Chronic dyslipidemia patients were followed in our department of gastroenterologyhepatology or in the department of vascular diseases. Samples were selected on daily basis with regard to the following criteria in order to avoid any bias related to vitamin or nutritional deficiency or con-

| Cryohemolysis test
Cryohemolysis test was performed following Streichman et al. 34 with minor adjustments as previously described. 5

| LoRRca MaxSis ektacytometer and osmoscan curve
Osmotic gradient is generated by mixing low-osmolar (40 mOsm/ kg) and high osmolar (600 mOsm/kg) solutions, to which the blood sample is automatically and continuously mixed, as previously described. 14

| Statistical analysis
Statistical analysis was realized by Analyse-it Software, Ltd and GraphPad Prism software. Statistical significance for parameters with normal distribution was evaluated by unpaired t test; for parameters with non-Gaussian distribution, Mann-Whitney t test and one-way ANOVA Kruskal-Wallis with one-way analysis of variance (Dunn's multiple comparison test) were used; a level of p < 0.05 was considered as statistically significant.

| RE SULTS
A total of 95 individual pathological samples were included in the study. Generally, each sample was tested with 2 out of the three tests: reticulocyte count and cryohemolysis or osmoscan. A presentation of the total number of samples per group and per test is shown on Table 1. Because of the restricted number of patients in the group of hypertriglyceridemia, this group was merged in all following analysis to the dyslipidemia group.
When studying data obtained from the automated hematological analyzer, several statistical differences were found (Table 2 and Figure 1). There were no differences for hemoglobin level as patients presented no anemia following the selection criteria. However, MCV was significantly higher in the dyslipidemia group, MSCV was significantly lower in diabetic and MS groups, which explains why delta (MCV-MSCV) was higher for all pathological groups. Regarding reticulocyte number and their indices, only in MS group we observed slight elevation in the reticulocyte count, still remaining in the reference range. MRV was significantly lower in diabetes and MS groups and IRF showed statistical differences for several groups. However, Ret/IRF ratio showed no difference for any of the pathological groups compared to the control group.
Applying our screening algorithm with defined cut-offs for HS screening, 5

| DISCUSS ION
With the present study, we aimed to analyze if common pathologic conditions such as diabetes or hypercholesterolemia interfere with variations in cholesterol-phospholipid ratio, fatty acid saturation and membrane proteins influence the membrane dynamic properties. 26 With regard to hypercholesterolemia, RBC deformability was found decreased, conceivably as a consequence of modifications in lipid membrane composition and in ATP release intensity. 33 However, evaluation of RBC deformability is challenging, due to the absence of standardized techniques. Furthermore, the observed discrepancies between various studies may be explained by differences in the studied populations or in the methodologies used to measure RBC deformability. 26 Cryohemolysis test 34 is based on the measurement of hemolysis of washed RBC diluted in hypertonic solution at 37°C and then subjected to lower temperature (0°C) provoking lipid phase transition and massive hemolysis. The range of cryohemolysis in normal subjects is 3%-15%; in HS, it is >20%. 38 It was also demonstrated that artificially provoked spherocytes, which are known to be very fragile osmotically, presented with normal cryohemolysis. 34 This suggests that surface-volume ratio reduction alone does not affect the cryohemolysis level. It was postulated that the straining inflicted on the cytoskeleton during the temperature changes exposed de- pathology groups and the control group ( Figure 3). We used in our study the osmoscan modus of LoRRca as this is the one actually largely used in hereditary membranopathies diagnosis. [14][15][16][17] To our knowledge, this study is the first one analyzing diabetic and dyslipidemia samples by osmotic gradient ektacytometry performed by LoRRca.
Decreased deformability was previously observed in 66% of the studied diabetic population by conventional osmotic gradient ektacytometry, analyzed by osmotic gradient ektacytometry (0-500 mOsm) under constant SS (120 dyn/cm 2 or 12 Pa) and by iso-osmotic osmoscan at 290 mOsm under SS gradient (0-212 dyn/ cm 2 ). 23 However, these studies were performed at lower SS than All studies demonstrating decreased deformability of RBC in diabetes 26,28,[40][41][42] or in hypercholesterolemia, 31,32 using RheoScan-D 28,42 or Rheodyn SSD 26,32,40,41 or a similar laser diffractometric method, 31 were based on measuring EI at lower SS than LoRRca in its osmoscan modus does. The discrepancy observed between our study and the previous studies could be explained by the distinctive instruments and specifically by the differences in the applied SS. However, the aim of our study was to investigate if there were interferences by diabetes or hypercholesterolemia in osmoscan used for HS diagnosis. Obviously, these common adulthood pathologies would not potentially interfere with a concomitantly present membranopathy under LoRRca osmoscan mode.
In conclusion, even if diabetes and hypercholesterolemia are pathologic conditions known to present with reduced RBC deformability and rheology, their potential concomitant occurrence with HS would not interfere with the laboratory diagnostic methods.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.