Detection of Hypertension-Induced Changes in Erythrocytes by SERS Nanosensors

Surface-enhanced Raman spectroscopy (SERS) is a promising tool that can be used in the detection of molecular changes triggered by disease development. Cardiovascular diseases (CVDs) are caused by multiple pathologies originating at the cellular level. The identification of these deteriorations can provide a better understanding of CVD mechanisms, and the monitoring of the identified molecular changes can be employed in the development of novel biosensor tools for early diagnostics. We applied plasmonic SERS nanosensors to assess changes in the properties of erythrocytes under normotensive and hypertensive conditions in the animal model. We found that spontaneous hypertension in rats leads (i) to a decrease in the erythrocyte plasma membrane fluidity and (ii) to a decrease in the mobility of the heme of the membrane-bound hemoglobin. We identified SERS parameters that can be used to detect pathological changes in the plasma membrane and submembrane region of erythrocytes.


Introduction
The development of novel methods and techniques has expanded our understanding of the molecular mechanisms underlying many pathologies and paved the way for the creation of novel diagnostic tools [1][2][3]. Cardiovascular diseases remain the leading cause of death around the world. High blood pressure is one of the most important risk factors for cardiovascular diseases, leading to organ hypoxia and consequent damage, such as heart failure, stroke, vasculopathy, and nephropathy [4][5][6]. Hypoxic conditions can develop as the result of hypertension-induced alterations in vessel structure and hemodynamics [4,7]. Another cause of tissue hypoxia is abnormal changes in hemoglobin's (Hb) affinity for oxygen (O 2 ), leading to the decreased rate of Hb saturation with O 2 in the lungs or to the decreased ability of Hb to release O 2 in peripheral tissues. Thus, the fine tuning of Hb properties ensures the optimal supply of O 2 to tissue [8,9].
Surface-enhanced Raman spectroscopy (SERS) has been successfully applied to detect different disease biomarkers in blood and other body fluids [3,10,11]. We have proposed an approach for the selective study of the Hb mb in erythrocytes with SERS using a colloidal solution of silver nanoparticles [12] and silver nanostructured surfaces (AgNSS) [13,14]. Plasmonic nanostructures allow us to achieve the enhancement of the Raman scattering from molecules near the nanostructure surface. In the case of erythrocytes, this occurs from membrane-bound Hb (Hb mb ) bound to AE1-exchanger (also known as band 3 protein). Erythrocyte plasma membrane lipids can be probed simultaneously [15]. Conventional Raman spectroscopy allows one to assess the degree of oxygenation, the conformation, and the affinity of the heme to O 2 of cytosolic Hb (Hb c ) in intact erythrocytes. However, it fails to probe Hb mb , a minor fraction of Hb, which is more sensitive to oxidative stress, changes in plasma membrane properties, and ion transport through the membrane than Hb c [16,17].
In the present work, we employed SERS with AgNSS to study the effect of spontaneous hypertension on the properties of the plasma membrane and membrane-bound hemoglobin in erythrocytes. The proposed approach is able to detect changes in erythrocyte properties in the early stages of hypertension, such as a decrease in the erythrocyte plasma membrane fluidity and a decrease in the Hb mb heme mobility.

Nanostructure Synthesis
The preparation of AgNSS was performed as described in References [13,18] with modifications. Briefly, 0.3 g of silver nitrate (Sigma, Schnelldorf, Germany, 99.999% purity) was dissolved in 40 mL of MilliQ water under constant stirring, followed by the addition of 30 mL of 20% NaOH solution. The resulting dark precipitate of silver oxide (I) was washed three times with 100 mL of MilliQ water, followed by the addition of 5 mL of 25% aqueous ammonia and 25 mL of MilliQ water, resulting in transparent precursor solution of diamine silver (I) hydroxide. Albedo ultrasonic nebulizer was used to spray the obtained solution of silver complex onto the surface of coverslips placed inside 1000 mL glass on a surface heated up to 340 • C (IKA C-MAG HS 4) for an hour with 5 min breaks every 3-4 min. The coverslips with a silver nanostructured layer were kept at 340 • C for 15 min after the completion of spraying to decompose and desorb possible residual intermediate compounds.

Characterization of Nanostructures
The fabricated AgNSS were visualized by a scanning electron microscope (SEM) NVision 40 (Carl Zeiss). The near-field optical characterization was performed by scatteringtype scanning near-field optical microscope s-SNOM "NeaSNOM" (Neaspec, Munich, Germany) based on an atomic force microscope that uses sharp silicon tips covered with a Pt-Ir coating as a near-field probe. The scanning was carried out in a tapping mode with an oscillation frequency of~250 kHz. The tip-sample interface was illuminated under ∼50 • relative to the sample surface by a linearly P-polarized Avesta TiC, Ti:Sapphire continuous tunable laser (700-1000 nm). In this s-SNOM setup, the laser beam is focused on the nanostructures by an upper parabolic mirror. A tip-scattered near-field signal is collected by the same parabolic mirror and goes to a high-sensitive detector, then analyzed after.

Animals
Male 15-or 16-week-old rats from two strains (3 rats in each group) were used in the experiments: spontaneously hypertensive rats (SHR), and Wistar-Kyoto rats (WKY) (BIBCh, Pushchino, Russia Arterial blood pressure (ABP) was measured noninvasively using the photoelectric plethysmography technique [19] at least 3 times for each rat after adaptation and handling. For WKY rats, the ABP was 139 ± 4 mm Hg; for SHR rats, the ABP was 203 ± 4 mm Hg.

Preparation of Erythrocyte Ghosts
Blood was collected during the decapitation of rats in 15 mL glass tubes containing heparin (10 units/mL). Erythrocyte ghosts (enclosed vesicles consisting of erythrocyte plasma membrane with membrane-bound hemoglobin (Hb mb ) and submembrane cytoskeleton without cytosolic Hb) were obtained as described in Reference [12]. Briefly, red blood cell mass was hemolyzed in 20 volumes of phosphate buffer (4.7 mM Na 2 HPO 4 , 0.3 mM NaH 2 PO 4 (pH 7.4), 4 • C), followed by three washings and centrifugations (3500 g, 40 min). The erythrocyte ghosts were concentrated by centrifugation (13000 g, 40 min).

SERS Measurements
SERS measurements were performed using an InVia Raman microscope (Renishaw, New Mills, Wotton-under-Edge, Gloucestershire, United Kingdom) equipped with a 20 mW 514 nm argon laser and power neutral density filter (50%). All the spectra were collected using a 20× NA 0.4 objective and 20 s acquisition time. Laser power was 1-5 mM per registration spot. A silicon wafer was used for calibration. The SERS spectra of erythrocytes and erythrocyte ghosts were registered as described previously [12,14]. Before the SERS measurements, suspensions of erythrocyte ghosts were diluted 5 times with the Alen saline (145 mM NaCl, 5 mM KCl, 4 mM Na 2 HPO 4 , 1 mM NaH 2 PO 4 , 1 mM MgSO 4 , 1 mM CaCl 2 , pH 7.4). To perform SERS measurements, 300 µL of the suspension of interest were dropped on the AgNSS placed into a glass-bottom Petri dish, then SERS spectra were collected from several places.

SERS Spectra Analysis and Statistics
SERS spectra were processed using open source software Pyraman available at https:// github.com/abrazhe/pyraman accessed on 22 November 2021. The baseline was subtracted in each spectrum and ratios of peaks intensities I 1638 /I 1375 , I 1175 /I 1375 and I 2872 /I 2927 were calculated after the baseline substraction.
Statistical data were analyzed using GraphPad Prism 8.4 (https://www.graphpad. com accessed on 22 November 2021). The Mann-Whitney U test was used for group comparisons.

Experimental Design and AgNSS Characterization
Hypertension-related effects were studied in two types of samples obtained from the blood of hypertensive and normotensive rats: on diluted suspensions of erythrocytes and diluted suspensions of erythrocyte ghosts (Figure 1a).
Erythrocyte ghosts serve as a simplified experimental model of the submembrane region of erythrocytes, since they represent enclosed vesicles of the erythrocyte plasma membrane with Hb mb , which maintains its interaction with AE1-exchanger [20,21] (Figure 1b). The submembrane region of intact erythrocytes is more complicated and consists of the plasma membrane, Hb mb interacting with the AE1-exchanger, and the cytoskeleton with some amount of Hb c in close vicinity to the inner membrane surface (Figure 1b). To record SERS spectra from molecules in the submembrane region of the studied preparations a small volume of the diluted erythrocyte or erythrocyte ghost suspension was placed on plasmonic nanostructured surfaces (AgNSS in a glass Petri dish). The AgNSS, which were formed by the ultrasonic spraying method [13], have a complex morphology of intersecting circles consisting of silver nanoparticles of different sizes from 5-20 nm up to 100 nm ( Figure 2a). Erythrocyte ghosts serve as a simplified experimental model of the submembrane region of erythrocytes, since they represent enclosed vesicles of the erythrocyte plasma membrane with Hbmb, which maintains its interaction with AE1-exchanger [20,21] ( Figure 1b). The submembrane region of intact erythrocytes is more complicated and consists of the plasma membrane, Hbmb interacting with the AE1-exchanger, and the cytoskeleton with some amount of Hbc in close vicinity to the inner membrane surface (Figure 1b). To record SERS spectra from molecules in the submembrane region of the studied preparations a small volume of the diluted erythrocyte or erythrocyte ghost suspension was placed on plasmonic nanostructured surfaces (AgNSS in a glass Petri dish). The AgNSS, which were formed by the ultrasonic spraying method [13], have a complex morphology of intersecting circles consisting of silver nanoparticles of different sizes from 5-20 nm up to 100 nm ( Figure 2a). These nanostructures provide a stable and reproducible SERS signal from purified biomolecules, erythrocytes, and mitochondria, as was demonstrated in our previous works [13,18]. Thus, we have already demonstrated the reproducible SERS spectra from Hbmb in erythrocytes and cytochrome c in mitochondria recorded from various places on the same AgNSS or from different AgNSS. We also showed the stability of SERS spectra of erythrocytes and mitochondria registered from the same region of AgNSS in time proving the absence of AgNSS degradation in the sample and the absence of the erythrocyte or mitochondria damage by nanostructures [13,18].
Before SERS measurements, the AgNSS were characterized by s-SNOM to visualize  Erythrocyte ghosts serve as a simplified experimental model of the submembrane region of erythrocytes, since they represent enclosed vesicles of the erythrocyte plasma membrane with Hbmb, which maintains its interaction with AE1-exchanger [20,21] ( Figure 1b). The submembrane region of intact erythrocytes is more complicated and consists of the plasma membrane, Hbmb interacting with the AE1-exchanger, and the cytoskeleton with some amount of Hbc in close vicinity to the inner membrane surface (Figure 1b). To record SERS spectra from molecules in the submembrane region of the studied preparations a small volume of the diluted erythrocyte or erythrocyte ghost suspension was placed on plasmonic nanostructured surfaces (AgNSS in a glass Petri dish). The AgNSS, which were formed by the ultrasonic spraying method [13], have a complex morphology of intersecting circles consisting of silver nanoparticles of different sizes from 5-20 nm up to 100 nm (Figure 2a). These nanostructures provide a stable and reproducible SERS signal from purified biomolecules, erythrocytes, and mitochondria, as was demonstrated in our previous works [13,18]. Thus, we have already demonstrated the reproducible SERS spectra from Hbmb in erythrocytes and cytochrome c in mitochondria recorded from various places on the same AgNSS or from different AgNSS. We also showed the stability of SERS spectra of erythrocytes and mitochondria registered from the same region of AgNSS in time proving the absence of AgNSS degradation in the sample and the absence of the erythrocyte or mitochondria damage by nanostructures [13,18].
Before SERS measurements, the AgNSS were characterized by s-SNOM to visualize the light-sample interaction. The s-SNOM image (Figure 2c) revealed the existence of randomly distributed and strongly localized electromagnetic excitations-hot spots [22,23] that originate mainly due to the gaps between conglomerations of Ag nanoparticles. The comparative analysis of the s-SNOM image (recorded at λ = 720 nm) was performed by directly superimposing one on the corresponding topographical These nanostructures provide a stable and reproducible SERS signal from purified biomolecules, erythrocytes, and mitochondria, as was demonstrated in our previous works [13,18]. Thus, we have already demonstrated the reproducible SERS spectra from Hb mb in erythrocytes and cytochrome c in mitochondria recorded from various places on the same AgNSS or from different AgNSS. We also showed the stability of SERS spectra of erythrocytes and mitochondria registered from the same region of AgNSS in time proving the absence of AgNSS degradation in the sample and the absence of the erythrocyte or mitochondria damage by nanostructures [13,18].
Before SERS measurements, the AgNSS were characterized by s-SNOM to visualize the light-sample interaction. The s-SNOM image (Figure 2c) revealed the existence of randomly distributed and strongly localized electromagnetic excitations-hot spots [22,23] that originate mainly due to the gaps between conglomerations of Ag nanoparticles. The comparative analysis of the s-SNOM image (recorded at λ = 720 nm) was performed by directly superimposing one on the corresponding topographical image obtained for a typical conglomeration of Ag nanoparticles (Figure 2b,c). The localization of several hot spots can be clearly seen inside the conglomeration; their appearance should be directly attributed to the gaps between individual Ag nanoparticles. The wavelength used for s-SNOM characterization (λ = 720 nm) is far from the 514 nm argon laser that was used in the SERS experiments. However, this is justified since gap-induced field enhancements are associated with the boundary conditions for the electric field [24][25][26] and therefore weakly depend on the excitation wavelength. Such places on AgNSS with high local electric fields ensure the enhancement of Raman scattering from molecules that are localized not only on the AgNSS, but also on some distance from nanoparticles (e.g., all Hb molecules in the submembrane region of erythrocytes and Hb mb in erythrocyte ghosts).

Spontaneous Hypertension Affects Conformation of Heme in Hb Bound to AE1-Exchanger
The SERS spectra of erythrocytes and erythrocyte ghosts recorded in low-frequency region (600-1800 cm −1 ) originate mainly from the heme molecules of oxyhemoglobin (oHb) (Figure 3a) [12]. The oxygenation of all hemoglobin molecules in studied samples occurred during the high dilution of blood in erythrocyte suspension and during the procedure of ghost preparation. The complete oxygenation of Hb is attributed to SERS peaks that are sensitive to the Hb oxygenation state: the combination of intensive peaks at 1375, 1585, and 1638 cm −1 is the signature of oHb, whereas peaks assigned to deoxyhemoglobin (1355, 1555, and 1606 cm −1 ) and methemoglobin (1360-1365, 1565, 1603 cm −1 ) are absent [16,[27][28][29]. The peaks at 1638 and 1585 cm −1 correspond to vibrations of methine bridges; the peaks at 1375 and 1175 cm −1 correspond to symmetric and asymmetric vibrations of pyrrole half rings, respectively; the peak at 1123 cm −1 reflects methyl group vibrations. The SERS spectra of erythrocytes and erythrocyte ghosts were stable over time and reproducible for different AgNSSs or different spots on the same region. Figure S1a (Supporting Information) demonstrates SERS spectra of erythrocyte ghosts recorded from three different regions of AgNSS. It can be seen that the spectrum structure and the input of main peaks into the overall spectrum are similar for all the SERS spectra. The same reproducibility was demonstrated with Hb mb in erythrocytes in our previous work [13], showing the absence of the negative effect of the biological sample on AgNSS and vice versa. An important parameter that is usually used to characterize SERS structures is the enhancement factor (EF). EF is calculated from concentrations of the molecule, giving Raman and SERS spectra and the peak intensities of a chosen peak in Raman and SERS spectra. Figure S1b (Supporting Information) demonstrates the Raman and SERS spectra of erythrocytes with a description of the EF calculation for the peak at 1638 cm −1 estimated to be ≈ 2.5 × 10 4 .
OR PEER REVIEW 6 of 12  To investigate the conformational properties of Hb mb heme, we calculated the ratios of the SERS peak intensities: I 1638 /I 1375 and I 1175 /I 1375 (Figure 3b,c). The intensity of the peak at 1375 cm −1 is used as a normalizing factor, since its intensity does not depend on the heme conformation and can be regarded as a constant value in our experimental conditions. As was demonstrated earlier [30][31][32], the heme b in Hb predominantly exists in two conformations: planar and domed [30][31][32]. The first ratio, I 1638 /I 1375 , shows the probability of the planar heme conformation. The higher this ratio is, the higher the probability of a planar conformation of the heme is [33,34]. Moreover, the use of ratios of peak intensities causes the spectral analysis to be independent of the local SERS signal intensity. Our results clearly indicate the absence of the effect of hypertension on the probability of the planar heme conformation (Figure 3b). The ratio of I 1175 /I 1375 peak intensities represents the probability of the asymmetric pyrrole half-ring vibrations and can be used as a marker of heme mobility: the higher this ratio is, the higher the in-plane mobility of the heme is [33,34]. We found that this ratio is more dispersed in erythrocytes than in erythrocyte ghosts (Figure 3c), which may be explained by the more heterogeneous Hb population in the submembrane region of intact erythrocytes than in erythrocyte ghosts (Figure 1b). We observed no differences in the I 1175 /I 1375 ratio in the SERS spectra of erythrocytes from normotensive and hypertensive rats (Figure 3c). In contrast, we observed a significant decrease in this ratio for erythrocyte ghosts from SHR rats, indicating the decreased in-plane mobility of the heme in Hb mb (Figure 3c). Moreover, this marker negatively correlates with ABP ( Figure 4).

Decreased Fluidity of Plasma Membrane of Erythrocytes under Spontaneous Hypertension
To study the properties of the erythrocyte plasma membrane, we recorded SERS

Decreased Fluidity of Plasma Membrane of Erythrocytes under Spontaneous Hypertension
To study the properties of the erythrocyte plasma membrane, we recorded SERS spectra in the high-frequency region (2700-3100 cm −1 ) from erythrocyte ghosts. The SERS spectra of erythrocyte ghosts contain three main peaks at 2872, 2927, and 2962 cm −1 (Table 1, Figure 3d), corresponding to =CH 2 asymmetric, =CH 2 symmetric, and -CH 3 group vibrations, respectively [35]. These peaks are known to originate from lipids and proteins. In intact erythrocytes, the submembrane region contains different proteins; therefore, we did not record the SERS spectra of intact erythrocytes. In erythrocyte ghosts, lipids are the main component of the membrane region with the only subpopulation of Hb molecules bound to the AE1 protein. Thus, the SERS peaks at 2872, 2927, and 2962 cm −1 are assumed to originate mainly from lipids [15]. The ratio I 2872 /I 2927 may indicate the ordering of the lipid phase. An increase in this ratio indicates an increase in the number of lipids in trans conformation, which make membrane less fluid [36]. We propose the use of the ratio I 2872 /I 2927 to evaluate the fluidity of the erythrocyte ghost membrane (Figure 3e). The observed increase in this ratio indicates a decrease in the membrane fluidity. This ratio is statistically significantly higher in SHR rats (Figure 3e), or rats with elevated ABP. This means that spontaneous hypertension leads to the changes in lipid ordering and fluidity in the plasma membrane of erythrocytes. We also found that the ratio I 2872 /I 2927 (lipid ordering marker) positively correlates with arterial blood pressure (ABP) in rats (Figure 4a), whereas the ratio I 1175 /I 1375 (reflecting the in-plane mobility of pyrrole half rings in a Hb mb heme) negatively correlates with ABP ( Figure 4b). There is no correlation between the ratio I 1638 /I 1375 (the probability of the planar conformation of a Hb mb heme) and ABP (Figure 4c). We notice that there is a strong negative correlation between the ratios I 1175 /I 1375 and I 2872 /I 2927 , showing that the in-plane heme mobility in Hb mb decreases with the decrease in the plasma membrane fluidity.

Discussion
We applied plasmonic SERS nanostructures to study the properties of the membranebound hemoglobin and plasma membrane of erythrocytes and erythrocyte ghosts received from rats with normal and elevated arterial blood pressure. Our results revealed hypertension-induced changes in erythrocyte properties at the molecular level.
It should be noted that high requirements are imposed on SERS nanosensors for biological application-first of all, the reproducibility and stability of the Raman enhancement [37]. The SERS nanosensors used in this study proved themselves to be resistant to the biological preparations and have shown the ability to provide reproducible SERS spectra for a variety of biological objects [13,15,18]. Calculating the ratios of specified SERS peak intensities rather than analyzing absolute peak intensities helps to avoid the influence of variation in signal intensity on the interpretation of the results.
We used the ratio of SERS peak intensities I 2872 /I 2927 to study the microfluidity of erythrocyte plasma membranes. We revealed decreased membrane fluidity in erythrocyte ghost membranes from hypertensive rats (SHR) compared to healthy normotensive rats (WKY) (Figure 3e), which fits well with clinically relevant data [38].
One of the explanations of this phenomenon is the increased level of cholesterol. An increased level of cholesterol in erythrocyte membranes from people with hypertension and other CVD was shown by Raman spectroscopy and EPR [38,39] and commercial enzymatic assays [40,41]. Increasing cholesterol leads to a decrease in membrane fluidity [42,43] and changes in the erythrocyte function [44]. The SERS ratio I 2872 /I 2927 was proposed as a marker of cholesterol amount by Faried et al. [45]. Thus, an increase in this ratio, indicating an increase in the membrane stiffness, may be explained by an increased amount of cholesterol in the erythrocyte plasma membrane.
For the first time, we observed a decrease in the in-plane mobility of heme in Hb bound to AE1-exchanger in hypertensive rats (Figure 3c), which is impossible to detect by other methods. Other powerful techniques such as photoacustics and NIRS [46,47] allow determining only Hb c oxygenation, but not other properties. Raman spectroscopy is able to evaluate oxygenation, the affinity of Hb to oxygen, and the conformational properties of Hb c , but not Hb mb .
We suggest that the detected changes in Hb mb can be explained by the increasing stiffness of the membrane (Figure 4d), which leads to a reduction in the mobility of AE1exchanger [48], which in turn directly interacts with Hb mb and may affect the conformation and mobility of Hb mb globin. This may reduce the ability of a heme to adapt to changing oxygen concentrations and, thus, to disturb the normal functioning of Hb mb and impair oxygen supply to tissues in hypertensive conditions.

Conclusions
The use of SERS plasmonic nanosensors allowed us to detect hypertension-induced changes in the plasma membrane and in the conformation of heme in membrane-bound Hb in erythrocytes and erythrocyte ghosts. We revealed a decrease in erythrocyte plasma membrane fluidity and a decrease in the in-plane mobility of heme pyrrole rings in membranebound Hb under hypertension, which may affect the affinity of hemoglobin to oxygen. The proposed SERS-based approach may be used to develop novel diagnostic tools to detect early pathologies and to assess treatment outcomes in CVD and beyond.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/bios12010032/s1, Figure S1: (a) SERS spectra of erythrocyte ghosts recorded from different places in a randomly chosen AgNSS. X-axis, Raman shift, cm −1 ; Y-axis, SERS intensity, vertical scale bar corresponds to 1000 arbitr.u. For a clearer representation, the spectra are shifted vertically. Numbers above a peak show maximum positions. The similarity of the spectral structure demonstrates the high reproducibility of the SERS spectra of erythrocyte ghosts. (b) Raman and SERS spectra of erythrocyte suspensions. X-axis, Raman shift, cm −1 ; Y-axis, Raman or SERS intensity. Raman spectrum (black) was recorded from the erythrocyte suspension obtained by a 50-time blood dilution; SERS spectrum (red) was recorded from an erythrocyte sample obtained by a 2000-time blood dilution. Laser wavelength was 514 nm with a power of 1 mW per registration spot, with 60 s accumulation time; objective 20x, NA 0.4.  Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.