Introduction

Rett Syndrome (RTT; MIM312750) is the leading cause of a pervasive and complex neurodevelopmental disease in females. More than 95% of individuals with classic RTT carry sporadic de novo loss-of-function mutations in the X-linked methyl-CpG-binding protein 2 (MECP2) gene [1], which encodes a nuclear protein that binds methylated CpGs and regulates gene expression [2]. Other genes have been linked to additional rare variant RTT, including cyclin-dependent kinase-like 5 (CDKL5), forkhead box protein G1 (FOXG1), and the Netrin G1 genes [3, 4]. Although a wide variability in phenotypic severity is observed, in its classic form, RTT is characterized by a unique disease progression, beginning with an uneventful early infancy (6–18 months), followed by stagnation and regression of growth, motor, language, and social skills later in development [5, 6]. Notably, breathing disorders (i.e., chronic mild hypoxia with impaired lung gas exchange) are considered as a hallmark feature of the disease being among the key symptoms of RTT [7,8,9,10], and contributing to the developmental abnormalities in the brain and to a high rate of sudden and unexpected death [11]. It is widely recognized that the systemic oxidative stress (OS), triggered by associated hypoxia, acts as a primary driving force in the pathogenesis mechanism of classic RTT, this being supported by the evidence that OS biomarkers are related to the severity of neurological symptoms, mutation type, and clinical presentation [12, 13, 20]. The occurrence of a redox imbalance in RTT has been reported both in patients [12,13,14,15,16] and in experimental mouse models [17]. Notably, brain oxidative damage occurs already in the pre-symptomatic stage, as inferred by studies carried out in Mecp2-mutant murine models of Rett synrome [18,19,20,21].

Red blood cell (RBC) shape is abnormal in RTT and it appears to be mainly modulated by OS. In particular, leptocytes were found to be the predominantly altered erythrocyte shape in untreated typical RTT [22]. The percentages of abnormal RBCs shape were found to be related to an increased intraerythrocytic and plasmatic non-protein-bound iron (NPBI; i.e., free iron), and to the evidence of a membrane oxidative damage, as demonstrated by the increased formation of esterified F2-IsoPs and 4-HNE Pas [22, 23]. In typical RTT patients, the evidences of OS damage in RBCs are strongly related to hypoxic conditions, as indicated by decreased SpO2 and PaO2 levels, coupled with increased methemoglobin (Met-Hb) and carboxyhemoglobin (CO-Hb) concentrations in RTT with respect to controls [20, 22]. Supplementation with ω-3 PUFAs has been shown to be able to partially rescue RBCs shape and to reduce clinical severity, improving pulmonary oxygen exchange and cardiopulmonary physiology [22]. Altogether, these findings suggest that monitoring of physiological properties of RBC as a function of OS damage can represent an important diagnostic and prognostic tool in this postnatal neurological disorder, in which the lung seems to represent an unexpectedly key organ for the disease pathogenesis and progression.

We have focused the present study on RBCs in order to evaluate the impact of oxidation and redox imbalance on several functional properties of this important biosensor. In particular, we have investigated for the first time in RTT patients the kinetics of oxygen release by RBCs and the intrinsic affinity of the hemoglobin (Hb) to possibly detect differences in the functional properties of RBCs related to oxygen exchanges and Hb affinity. Moreover, in order to identify alterations in antioxidant defense and metabolic process in RBCs from RTT, the energy metabolism of erythrocytes was analyzed by ion-pairing HPLC method and related to the levels of malondialdehyde (MDA), a marker of membrane lipid peroxidation [24], and to the reactive oxygen species scavenging capacity (ORAC assay).

Materials and methods

Subjects

A total of 15 female patients with clinical diagnosis of typical RTT (mean age: 11–15 years, range 2–21), as well as ten healthy female controls of comparable age free of any medication (mean age: 11–21 years, range 2–22), were enrolled in the study. RTT diagnosis and inclusion/exclusion criteria were based on the recently revised RTT nomenclature consensus. Blood samplings in the control group were carried out during routine health checks, always after informed consent while blood samplings in Rett patients were obtained during the periodic clinical checks. Institutional review board approval and informed consent from either the parents or the adult control subjects as well as the approval by the local ethical committee were obtained.

Blood sampling

Blood was collected in heparinized tubes and split into two aliquots; one was processed according to the HPLC analysis, the other was centrifuged at 3500 rpm for 5 min at 4 °C and the supernatant, containing plasma and buffy coat, was removed. The erythrocyte pellet was washed three times and resuspended in isotonic buffer solution (58 mM NaP, 77.5 mM NaCl, 10 μM CaCl2), with a final pH value of 7.4. Red blood cell suspension was then split and differently processed for kinetic analysis and oxygen-binding measurements.

Kinetics of oxygen release by RBCs

The kinetics of oxygen release by RBCs was carried out, within 1 h from withdrawal, by rapid mixing experiments using the SX18.MV stopped-flow apparatus (Applied Photophysics, Salisbury, UK) equipped with a diode array for spectra acquisition over 1 ms time range [25, 26]. Suspension of completely oxygenated red cells (in isotonic buffer solution, pH 7.4, 15 μM total heme concentration after mixing) was mixed with a deoxygenated isotonic solution containing 20 mmol L−1 sodium dithionite (Fluka Biochemika) at 37 °C.

Kinetic traces have been analyzed according to the following equation:

$$ OD_{\text{obs}} = OD_{0} \pm \sum\limits_{i = 1}^{i = r} {DOD_{i} \cdot \exp ( - {}^{i}k \cdot t)^{n} } , $$
(1)

where OD obs is the observed optical density at a selected wavelength and at a given time interval, OD 0 is the optical density at t = 0, r is the number of exponentials, DOD i is the optical density change associated to the exponential i, i k is the rate constant of the exponential i, t is the time, and n is the Hill coefficient, as a phenomenological expression of the cooperativity of the kinetic process, accounting for the markedly autocatalytic O2 release.

Equilibrium oxygen-binding measurements

Oxygen-binding measurements were obtained spectrophotometrically, at 37 °C, by tonometric method [27]. To determine its internal volume (i.e., V tot) the tonometer was weighed before and after filling with distilled H2O. Red blood cell suspensions (corresponding to about 15 μM total heme) were deoxygenated until deoxy-Hb was formed, as from the absorption spectra. Aliquots of air (ranging between 1 and 10 mL) were then injected into the tonometer by syringes; the oxygen pressure after each addition of air could be calculated according to the following relationship

$$ pO_{2} ({\text{mmHg}}) = 152 \cdot \frac{{V_{\text{ad}} }}{{V_{\text{res}} }}, $$
(2)

where V ad is the total volume of air added, V res (=V tot−V sample) is the residual tonometer gas volume after RBC sample introduction, and 152 mmHg corresponds to the O2 pressure at 1 atmosphere of air. After each addition, red cell suspension was incubated for 10 min at 37 °C, rotating the horizontally lying tonometer at 120 rpm around its axis; the absorbance change upon oxygenation was monitored by a Jasco-710 spectrophotometer (Tokyo, Japan) at 560 versus 577 nm. Oxygenation of the intraerythrocytic Hb was brought about in a step-wise manner so that each addition of O2 would change the absorbance of the suspension by a given percentage (corresponding to the saturation degree Y) of the total absorbance change observed for complete oxygenation. Fractional saturation of Hb was determined from the absorbance changes, and the free oxygen concentration was calculated by subtracting the amount of bound oxygen from total oxygen present after each addition of air, as from Eq. (2). The cooperative nature of oxygen binding to Hb in RTT and controls was examined by Hill plot analysis, according to the following equation

$$ Y = \frac{{\left[ {HbO_{2} } \right]}}{{\left[ {Hb} \right] + \left[ {HbO_{2} } \right]}} = \frac{{K_{\text{ov}} \cdot pO_{2}^{n} }}{{1 + K_{\text{ov}} \cdot pO_{2}^{n} }}, $$
(3a)

where Y is percentage of oxygenated HbO 2, K ov (mmHg−1) is the overall oxygen-binding affinity, pO 2 is the oxygen pressure, and n is the Hill coefficient, expressing the cooperativity degree. A more appropriate and common way to represent O2-binding isotherms is by the logarithmic plot, according to the following equation

$$ \log \left( {\frac{Y}{1 - Y}} \right) = \log (K_{\text{ov}} ) + n \cdot \log (pO_{2} ), $$
(3b)

Quantitative analysis of malondialdehyde and energy state

Peripheral venous blood samples were collected in heparinized tubes. After 10 min centrifugation at 1853×g and 4 °C, carried out within 2 min from withdrawal, erythrocytes were washed twice with a large volume of PBS. After the second wash, packed erythrocytes were gently resuspended with PBS to obtain a 5% hematocrit. 500 μL of this suspension was deproteinized by adding ice-cold 70% HClO4 (20 μL). The suspension was centrifuged at 20,690×g for 10 min at 4 °C, neutralized by adding 5 μM K2CO3 (20 μL) in ice, filtered through a 0.45 μM Millipore-HV filter and then analyzed by HPLC (100 μL) for the simultaneous direct determination of MDA and adenine nucleotides [24, 28]. The employed HPLC assay for the energy state analysis offers the possibility to determine high-energy phosphate content without any chemical manipulation of samples, except for perchloric acid deproteinization in order (i) to minimize the risk of modifications in the pattern of metabolites by proteins and (ii) to avoid any possible HPLC analytical column obstruction caused by proteins [29, 30]. Concentrations of high-energy phosphates were determined on 100 μL of perchloric acid extract by an ion-pairing HPLC method [29,30,31] using a Kromasil 250 × 4.6 mm, 5 μm particle size column, with its own guard column (Eka Chemicals AB, Bohus, Sweden), and using tetrabutylammonium hydroxide as the pairing reagent. Briefly, the separation of different metabolites was obtained by forming a step gradient (adapted to the column length increase compared to the original method) [30] with two buffers of the following composition: buffer A, 10 mM tetrabutylammonium hydroxide, 10 mM KH2PO4, 0.25% methanol pH 7.00; buffer B, 2.8 mM tetrabutylammonium hydroxide, 100 mM KH2PO4, 30% methanol pH 5.50. The gradient was: 10 min 100% buffer A; 3 min 90% buffer A; 10 min 70% buffer A; 12 min 55% buffer A; 15 min 45% buffer A; 10 min 25% buffer A; 5 min 0% buffer A. The flow rate throughout the chromatographic runs was 1.2 mL/min and the column temperature was kept at a constant 23 °C using water-jacketed glassware. The HPLC apparatus consisted of a Surveyor LC Pump (ThermoFinnigan Italia, Rodano, Milan, Italy) connected to a Surveyor PDA Detector (ThermoFinnigan Italia) at 200–300 nm. Acquisition and analysis of data were performed using the ChromQuest program (ThermoQuest Italy). Comparison of areas, retention times, and absorbance spectra of the peaks of sample chromatograms with those of freshly prepared ultrapure standards made it possible to identify and calculate the concentration of the different metabolites. Both Hb and percentage haemolysis were calculated by standard hematological techniques [30], using a Jasco-685 double beam spectrophotometer. Same chromatograms also allowed us to identify and calculate MDA concentrations. ATP/ADP and ATP/AMP ratios, as well as NAD+ content (μM) were used to identify intracellular energy content.

Oxygen radical absorbance capacity (ORAC) assay

The ORAC assay is based on the dose- and time-dependent decrease in the fluorescence intensity of β-phycoerythrin (β-PE) (Sigma-Aldrich Co, St. Louis, MO) when oxidized by oxygen radicals [32]. It measures the antioxidant capacity of a substance in terms of its ability to inhibit or delay β-PE peroxidation. AAPH [2.2′-Azobis(2-aminopropane) dihydrochloride], purchased from Polyscience (Warrington, PA), was used as the free-radical generator. Briefly, 20 μL of fresh whole blood was diluted in 10 mL of 75 mM phosphate buffer pH 7.0. After centrifugation to remove RBC ghosts, 100 mL of supernatant was used to evaluate redox scavenging activity. The final reaction mixture (2 mL) contained 1750 mL of 75 mM phosphate buffer (pH 7.0) and 100 μL of β-PE (1 mg dissolved in 88.2 mL of 75 mM phosphate buffer pH 7). A total of 0.1 mL of 20 μM Trolox (6-hydroxy-2,5,7,8-tetramethyl-2-carboxylic acid) in 75 mM phosphate buffer pH 7, buffer alone (blank), or hemolysed blood were added. After 15 min at 37 °C, 50 μL of 160 mM AAPH in a phosphate buffer was added. Beta-PE fluorescence was measured with a Varian Cary Eclipse Fluorescence Spectrofotometer (Varian Ltd., Madrid, Spain) at λ = 546 nm (λ excitation) and λ = 573 nm (λ emission). Measurements were made every 2.5 min at 37 °C for 1 h or until the fluorescence variation dropped below 2%. The ORAC of the sample was expressed as Micromol Trolox Equivalents/g and calculated as [(As−Ab)/(At−Ab)]ka, where As is the area under the curve (AUC) of β-PE in the sample, calculated with the Origin 2.8 Integration Program (MicroCal Software, LLC, Northhampton, MA), At is the AUC of the Trolox, Ab is the AUC of the control, k is the dilution factor (1:500 for the blood), and a is the concentration of the Trolox in mmol L−1.

Statistical analysis

All samples were run on HPLC twice, and ORAC was performed in triplicate. Results are given as an average among patients ±SD. Data were entered into the GraphPad Prism statistical analysis program (GraphPad, San Diego). Comparison of results was performed by one-way analysis of variance (ANOVA). Bonferroni’s test was used for multiple comparisons and Tukey’s test was used to analyze media and standard deviations in different experiments. Differences between groups were considered statistically significant when p < 0.05.

Results and discussion

It is widely recognized that RTT may arise from systemic physiological abnormalities including OS [12,13,14, 20], mitochondrial abnormalities [33, 34], and immune dysregulation/inflammation [35, 36], rather than being a purely central nervous system disorder. In particular, the oxidative hypothesis is able to explain several features of this syndrome (e.g., its genotype–phenotype correlation and clinical heterogeneity), being supported by the evidence that OS biomarkers are related to neurological symptoms severity, mutation type, and clinical presentation [37, 38]. The erythrocyte seems to be one of the most important tools of antioxidant defenses in the whole body [39]. If the oxidative insult of the microenvironment overcomes the RBCs’ defenses, this cell undergoes oxidative alterations, such as change of rheologic/functional properties [22]. It is conceivable that, in the presence of an intense and chronic OS, it is not possible to fully repair the damage [39]. In RBC of RTT patients oxidants induce an altered RBC shapes (i.e., leptocytes) and increased iron release in a free redox active form (NPBI, i.e., non-protein-bound iron), together with an increased level of erythrocyte membrane esterification of F2-IsoPs and 4-HNE Pas [22, 23, 40]. Interestingly, despite the oxidative damage to the RBC membrane, no pits and holes, usually associated with lipid peroxidation, were observed in the erythrocytes from RTT patients [22, 39] and, correspondingly, on the clinical side, no hemolytic anemia is commonly observed in RTT patients [22].

Moreover, no significant difference was observed in the RBC antioxidant enzyme activities (i.e., glutathione peroxidase, glutathione reductase, and catalase activities) between RTT patients and the control group, whereas erythrocyte superoxide dismutase activities were significantly decreased in Rett subjects as compared with controls [41].

In typical RTT patients, a hypoxic condition is evidenced [42,43,44], and it is associated with the abnormal erythrocyte shapes and with an increase in Met-Hb and CO-Hb concentrations in comparison to controls, coupled with a decrease in peripheral and arterial O2 levels [22, 23]. These alterations were more prominent when associated to an enhancement of iron release and lipid peroxidation, preventable by ω-3 PUFAs [22]. Overall, these findings suggest that, in this particular form of postnatal neurological syndrome, in which breathing disorders and the related levels of oxidative damage in erythrocytes and plasma significantly contribute to the progression of the disease, red blood cell can represent a valuable cell target model for monitoring the degree of phenotype severity. Our study was aimed to investigate, in RTT patients, the impact of oxidation and redox imbalance on several functional properties of RBCs, never deeply investigated before. The kinetics of oxygen release by RBCs and the intrinsic affinity of the intraerythrocytic Hb have been investigated for the first time in RTT subjects to possibly detect differences in the functional properties of RBCs, related to oxygen exchange and Hb affinity.

The kinetics of oxygen release has been followed by the stopped-flow and analyzed according to Eq. (1) [26, 27]. Data are reported in Table 1 together with those of Hb oxygen affinity (K ov values) and cooperativity (Hill’s coefficient, n), which were analyzed according to Eq. (3b). The reported values of n refer to those obtained by equilibrium oxygen affinity, since they are more readily compared with literature data. According to this experimental approach, it appears evident that RBCs from RTT patients and those from healthy controls display closely similar parameters both for Hb O2 affinity and rate of O2 release at 37 °C. This becomes evident also by comparing two typical O2 release kinetic processes by RTT patients and controls (Fig. 1a) and two typical O2-binding isotherms (Fig. 1b).

Table 1 Kinetics of oxygen release by RBCs and Hb oxygen affinity determination
Full size table
Fig. 1
figure 1

a Time courses for the release of O2 from red blood cells. Kinetic traces, at pH 7.4 and 37 °C at λ = 414 nm, for RTT erythrocytes (asterisk symbol) and healthy controls (open circle). Continuous lines correspond to the non-linear least squares fitting of data according to Eq. (1). b Hill plots of oxygen-binding curves of Hb. Oxygen-binding isotherms are shown, at pH 7.4 and 37 °C, for RTT erythrocytes (asterisk symbol) and healthy controls (open circle). Continuous lines correspond to the non-linear least squares fitting of data according to Eq. (3b)

Full size image

RBCs show a characteristic biconcave shape, which is of key importance for all their functions (i.e., deformability and O2 exchange), but it also very susceptible to morphological changes with consequent lack of functionality. Changes in osmolarity, pH conditions, and presence of oxidants are reported among the factors inducing morphological modifications and changes in rheological properties of RBCs [23, 45,46,47]. Thus, although the oxygen release depends markedly on the surface area/volume ratio and internal environment of the cell, as a result of our data, the in vitro dynamic of oxygen release by RBCs doesn’t seem to be compromised by the severe alteration of morphology observed in RTT erythrocytes [22]. Furthermore, our results show that the oxygen affinity and the cooperative nature of oxygen binding, determined spectrophotometrically for Hb of patients, were in every respect similar to those obtained for controls (Table 1; Fig. 1b). Even though this information concerns only the intrinsic functional properties of Hb and eventual alterations in the intraerythrocytic metabolites and allosteric modulators which affect the affinity of the Hb (i.e., pH, carbon dioxide, adenosine triphosphate or 2,3-bisphosphoglycerate) [48, 49], we can roughly postulate that the highly pro-oxidant environment of the erythrocyte in RTT (i.e., increased NPBI, alteration in Met-Hb and CO-Hb concentrations) doesn’t seem to have a role in affecting the Hb oxygenation properties. Anyhow, it must be stressed that this experimental approach is able to detect only functional alterations related to the role of Hb as O2 carrier. Therefore, functional modifications due to shape abnormalities or altered rheological properties of the RTT erythrocytes “in vivo” might not be detected, since the experimental conditions do not reproduce what actually occurs in the lung and/or in the system capillaries, where RBCs are squeezed and O2 release takes place under grossly shape deforming conditions, which eventually magnify membrane elasticity alterations.

Therefore, the close functional similarity between RTT erythrocytes and those from healthy controls simply rules out that the gas exchange alterations, previously reported [22], could be referred to a modified intrinsic O2 transport function of RTT RBCs. On the other hand, it also comes out in a clearcut fashion that respiratory alterations do not stem from a reduced capability of O2 transport for RBCs from RTT patients, as indicated by the closely similar O2 affinity and cooperativity (see Table 1).

A further aspect reviewed in our study concerned the antioxidant defense capability and the energy metabolism in RBCs from RTT. Detection of the main substances of energy metabolism (i.e., ATP, AMP and NAD+) on cell extracts has been undertaken by employing the HPLC ion-pairing method and it has been related to the MDA levels and ORAC values. As expected, data clearly envisage a severe OS condition in erythrocytes from RTT patients (see Table 2; Fig. 2). In particular, an increased level of MDA (up to tenfold) was observed in RTT patients together with a significant reduced antioxidant defense capability (ORAC −39.2%) (Table 2; Fig. 2a, b respectively). As reported in Table 2, the levels of the high-energy phosphorylated compounds were significantly altered in RTT red cells. In particular, a relevant depletion of ATP was observed, as indicated by a concentration of ≈1823 µmol L−1 in controls and ≈1266 µmol L−1 in RTT erythrocytes. A concomitant increase of ADP and AMP (up 99 and 33%, respectively) was detected with a resulting decrease of ATP/ADP and ATP/AMP ratios (−64 and −47%, respectively). In addition, an increase (up to 59%) of the NAD values, considered as the total nicotinic coenzyme pools (NAD+ + NADH), were observed in RTT red cells. Most of the energy is used in erythrocytes to protect iron and other cytoplasmic and membrane molecules from oxidation through either scavenging of redox reactive species and/or reduction of the already oxidized molecules. In RBCs, where neither mitochondria nor nuclei are present, the level of high-energy compounds is strictly modulated and directly linked to the redox state of erythrocytes, such that an increase in redox-protective response depletes the RBCs from high-energy compounds, like NADPH (GSH/GSSG regulator), NADH (cofactor in the Met-Hb reductase reaction), and ATP [29, 30, 50]. Therefore, determination of high-energy phosphate nucleotides (i.e., ATP/ADP, ATP/AMP ratios), reflecting changes in cell energy state, can provide information about the intracellular redox state of RBCs. On the other hand, it worth outlining that the extent of ATP decrease does not affect significantly the O2 affinity of intraerythrocytic Hb [51].

Table 2 Erythrocyte energy metabolism, lipid peroxidation (MDA), and ORAC assay results
Full size table
Fig. 2
figure 2

Levels of lipid peroxidation (MDA) (a) and oxygen radical absorbance capacity (ORAC assay) (b). The ORAC values are expressed as mM Trolox equivalent (TEs). Lines indicate mean with SD and symbols correspond to individual data points from patients with Rett syndrome (n = 13) and controls (n = 10)

Full size image

In Fig. 3 we sketch the complexity of factors influencing the respiratory mechanism, underlining alterations exhibited by RTT patients. It is quite obvious that the whole regulatory process is very complex and certainly the gas exchange abnormality observed in RTT patients cannot be the result of a single alleged mechanism, but it is rather the result of several contributing factors, involving OS and chronic subclinical inflammation in which terminal bronchioles and alveoli are likely a major inflammatory target of the disease [44].

Fig. 3
figure 3

Hypothetical diagram of respiratory pathophysiology in typical Rett Syndrome. ↑ increase; ↓ decrease; ↔ unchanged; RB-ILD respiratory bronchiolitis-interstitial lung disease; Δ(A-a)PO 2 oxygen alveolar–arterial gradient; F 2 -IsoPs F2- Isoprostanes; IE-NPBI intra-erythrocyte non-protein-bound iron; P-NPBI plasma non-protein-bound iron; GSSG glutathione oxidized form; GSH glutathione reduced form; 4-HNE Pas 4-hydroxynonenal protein adducts; MDA malondialdehyde; Fe 2+/3+ redox free iron; GAPDH glyceraldehyde-3-phosphate dehydrogenase; expr. expression; B3AT band 3 anion transport protein; ox. oxidation; ATP adenosine triphosphate; ADP adenosine diphosphate; AMP adenosine monophosphate; PAO 2 partial alveolar pressure of oxygen; PaO 2 partial arterial pressure of oxygen; PcO 2 (PaO 2 ) partial capillary pressure of oxygen; K rate of oxygen release; K ov oxygen affinities; ORAC reactive oxygen species scavenging capacity; Θ oxygen diffusion coefficient; *Present work

Full size image

Therefore, on the basis of the present study, we can hypothesize that the dynamics of the oxygen exchanges of erythrocytes with the environment and the intrinsic O2 transport function of RTT RBCs is normal and that the observed pulmonary gas exchange impairment in RTT patients should be rather attributed to some inflammatory alteration which affects O2 diffusion in the lungs, possibly at the level of terminal bronchioles and alveoli. In this perspective, such hypothesis could underline a connection between the redox imbalance, indeed detected in RBCs, and the inflammation-linked histological and functional alterations of tissues in the pulmonary system, bringing about the respiratory syndrome.

In conclusion, from this investigation on RBCs from RTT the “OS hypothesis” emerges as a systemic metabolic alteration, likely spread over to all body districts of RTT patients, and the respiratory alteration appears only as a consequence of this dismetabolism in pulmonary tissues without any consequence on the blood gas transport function.

In any case, our findings strengthen the importance of RBCs as a suitable biological tool for detecting markers of oxidative damage in RTT and underline the link between the biochemistry of OS and the pathogenesis of the disease.