DFT Simulation of the Vibrational Spectrum of Cholesteryl Esters: A New Physical Therapy Proposal for Targeted Clearance of Atherosclerotic Lipid Plaques

1. Introduction

Atherosclerosis (AS), the primary pathological basis of cardiovascular diseases, starts with endothelial dysfunction, accompanied by abnormal lipid deposition and amplified chronic inflammation [1,2,3]. The dynamic deposition of low-density lipoprotein (LDL) in the vascular wall is a key pathogenic factor. Cholesteryl ester, the esterified forms of cholesterol stored in LDL, constitute a major component of the LDL lipid core and directly drive the AS lesion progression. After endothelial damage, LDL lipids infiltrate the intima, oxidizing into pro-inflammatory oxLDL in oxidative stress and acidic environments. Monocytes migrate to the arterial intima, differentiating into macrophages that engulf oxLDL through scavenger receptors [4]. Macrophages upregulate scavenger receptors and downregulate cholesterol transporters, causing cholesterol and cholesteryl ester deposition in pathological states, inflammation and foam cell formation, which are the early signs of AS [5,6]. Consequently, foam cell apoptosis or necrosis creates a growing necrotic core of cholesteryl esters, cholesterol crystals, and debris, thereby increasing vascular lesion and rupture risks [7].

AS results from various factors, both hereditary and environmental. Unhealthy lifestyles, including high-fat diets, smoking, chronic stress, and irregular sleep patterns, are widely linked to a higher AS incidence [8]. Current treatments mainly involve statins and PCSK9 inhibitors, combined with anti-inflammatory therapy [9,10,11]. However, these traditional methods lack targeting specificity. Also, some patients are resistant to or have adverse reactions to statins [9,12]. Nanocarriers, which can passively accumulate at the lesion by Enhanced Permeation and Retention (EPR) effect or actively target specific receptor by precise surface modification, have gained significant attention [13,14,15,16]. Gao’s team developed a macrophage membrane coated reactive oxygen species (ROS)-responsive nanoparticles (NPs)-derived drug delivery system [16]. Kong et al. comprehensively explained the cascade activation mechanism of the NLRP3 inflammasome signaling pathway and proposed treating AS by targeting key inflammatory pathways [17]. New photodynamic therapy technologies like nanomotor driven methods, can better target endothelial cells, achieving precision drug delivery and treatment [18,19,20,21]. Sun’s team recently identified the endogenous receptors CYSLTR2 and P2RY6 for ceramide. They revealed that ceramide exacerbates atherosclerosis and chronic kidney disease by activating these receptors and inflammasomes, offering new targets for managing lipid metabolism disorders and residual lipid risk [22]. In addition, laser technology shows great potential in AS diagnosis and treatment. Intravascular optical coherence tomography (IVOCT) combines time-domain (TD) and Fourier-domain (FD) techniques, while intravascular photoacoustic imaging (IVAP) integrates optical excitation and acoustic signal detection to significantly enhance micrometer-level resolution for identifying plaque component [23]. The excimer laser uses ultraviolet light to break molecular bonds in AS lesion plaques, reducing them to small particles [24,25]. However, traditional laser therapy lacks molecular targeting and poses a risk of thermal damage to surrounding tissues. Moreover, stent and catheter implantation during the procedure may also lead to new issues such as vascular intimal hyperplasia and thrombosis [23,25].

This study proposes a new method for laser-targeted therapy based on photon-phonon resonant absorption (PPRA). We select four typical cholesterol esters, which are important components of lipid plaques: cholesterol linoleate (CLA), cholesterol oleate (COA), cholesterol palmitate (CPA), and cholesterol stearate (CSA). The infrared (IR) and Raman spectra of these esters were calculated by using first-principles methods, and the vibrational modes were assigned accordingly. This work offers a theoretical basis and a new direction for the precise laser intervention therapy of AS.

2. Methods

As solid cholesteryl esters are amorphous in structural growth, simulating their intermolecular interactions requires a huge supercell. Besides, their IR and Raman spectra have no characteristic peaks in the hydrogen-bonding regions. Thus, four single-molecule models were used to study their intramolecular vibrational modes. The $\mathrm{D}\mathrm{M}\mathrm{o}{\mathrm{L}}^3$code of the first-principles density functional theory (DFT) was employed to perform geometric optimization and vibrational spectrum simulation [26]. It is found that the Perdew–Burke–Ernzerhof exchange-correlation functional (PBE) is more accurate in calculating the vibrational spectrums of cholesteryl esters [27]. The convergence tolerance was set to $1\times {10}^{-5}$ Ha and the self-consistent field convergence criterion was set to $1\times {10}^{-9}$ eV/atom. We selected DNP with the DSPP pseudopotential as the basis set, and the energy cutoff was 3.7 $\mathrm{Å}$. Based on the harmonic approximation, the dynamic processes of all vibrational modes for four cholesteryl esters were analyzed. This enabled us to assign the reported experimental peaks and identify the vibrational modes with strong photon–phonon coupling.

3. Results and Discussion

The cholesteryl ester is structurally composed of the cyclopentaphenanthrene ring system, which is attached to the aliphatic hydrocarbon by an ester bond at the third position of the steroid [3]. Specifically, the fatty acid of CLA (${\mathrm{C}}_{45}{\mathrm{H}}_{76}{\mathrm{O}}_2$) is linoleic acid with 18 carbon atoms and two cis C=C bonds, showing a di-unsaturated feature. That of COA (${\mathrm{C}}_{45}{\mathrm{H}}_{78}{\mathrm{O}}_2$) is also an 18-carbon unsaturated fatty acid with a single cis C=C bond. CPA (${\mathrm{C}}_{43}{\mathrm{H}}_{76}{\mathrm{O}}_2$) contained a 16- carbon saturated palmitic acid with no C=C bonds and CSA (${\mathrm{C}}_{45}{\mathrm{H}}_{80}{\mathrm{O}}_2$) is combined with an 18-carbon saturated stearic acid. Common specific structural formulas of the four cholesteryl esters are shown in Figure 1. It could be found that their differences mainly lie in the length of the fatty acid chain, the degree of unsaturation, the number and position of C=C bonds. A single molecule of CLA contains 123 atoms, which means there are $123\times 3-3=366$ vibrational modes that can be detected by IR and Raman spectra. Similarly, COA, CPA, and CSA have 372, 360, and 378 vibrational modes, respectively. However, the intensity of each peak in IR and Raman spectra may vary depending on the selection rules. Therefore, we focused on the characteristic IR peaks with high PPRA and used the Raman peaks for validation.

Figure 2 presents the simulated IR and Raman spectra of four cholesteryl esters (CLA, COA, CPA, CSA). Through dynamic analysis of the vibrational modes, we classified them into four energy bands, which were skeletal vibrations, C–H bending vibrations, C=C/C=O stretching vibrations, and C–H stretching vibrations. Vibrations with wavenumbers below $500 \mathrm{c}\mathrm{m}⁻¹$, which involve complex motions of carbon chains and cooperative skeletal deformations, were assigned to skeletal vibrations. A sharp peak observed at around 430 cm⁻¹ in the Raman spectra in pervious experiments was attributed to the CH₂ deformation in the ring, which was considered as one of the cholesterol backbone marker bands [28,29]. Here we observed similar vibrations at $450 {\mathrm{c}\mathrm{m}}^{-1}$ (CLA), $439 {\mathrm{c}\mathrm{m}}^{-1}$ (COA), $461 {\mathrm{c}\mathrm{m}}^{-1}$ $\left(\mathrm{C}\mathrm{P}\mathrm{A}\right)$ and $435 {\mathrm{c}\mathrm{m}}^{-1}$ (CSA). Dynamic analysis indicated these modes correspond to cooperative deformations of ${\mathrm{CH}}_2/{\mathrm{CH}}_3$ groups near the cholesterol benzene ring and the benzene ring skeleton as shown in Figure 3.

The vibrational modes in the range from 500 to 1600 cm⁻¹ involving various dihedral angle deformations of C–H bonds were assigned to C–H bending vibrations. The calculated IR spectra of the four cholesteryl esters showed many peaks in this region. We observed distinct vibrational peaks at around $720 {\mathrm{c}\mathrm{m}}^{-1}$ in IR spectra of CLA and COA, with similar modes detected in the other two cholesterol esters. We presented four vibrational modes at $719 {\mathrm{c}\mathrm{m}}^{-1}$ (CLA), $740 \mathrm{c}\mathrm{m}⁻¹$ (COA), $736 \mathrm{c}\mathrm{m}⁻¹$ (CPA), and $727 \mathrm{c}\mathrm{m}⁻¹$ (CSA) as shown in Figure 3. Iyengar et al. classified the peak near $720 {\mathrm{cm}}^{-1}$ as the rocking vibration of the CH₂ groups in the lipid chain [30,31]. In this work, for unsaturated CLA and COA, they were attributed to H–C=C–H rocking vibrations, whereas those for saturated CSA and CPA, they originated from CH₂ rocking vibrations. We observed two strong vibrational peaks of CLA at 719 cm⁻¹ and 729 cm⁻¹, which correspond to the out-of-plane bending vibration of the H–C=C–C–H groups. For unsaturated cholesteryl esters CLA and COA, the presence of the H-C=C-H groups in their fatty acid chains leads to a greater degree of freedom, resulting in stronger corresponding vibrational peaks in this region. Smith identified strong ester peaks at around $1100 \mathrm{c}\mathrm{m}⁻¹$, and $1200 \mathrm{c}\mathrm{m}⁻¹$, corresponding to C–C–O and O–C–C stretching vibration, respectively [32]. Wrobel et al. also noted that the absorption peaks in the $1160-1175 \mathrm{c}\mathrm{m}⁻¹$ region were originated from the C–O–C stretching vibrations, which were only found in esters and not in fatty acids or cholesterol, could be a standard for distinguishing esters from alcohols and acids [31]. In contrast, our analysis of four strong IR peaks at $1188 \mathrm{c}\mathrm{m}⁻¹$ (CLA), $1195 \mathrm{c}\mathrm{m}⁻¹$ (COA), $1145 \mathrm{c}\mathrm{m}⁻¹$ (CPA), and $1150 \mathrm{c}\mathrm{m}⁻¹$ (CSA) as shown in Figure 5, indicated that these vibrations are primarily attributed to bending vibrations of $\mathrm{C}\mathrm{H}₂$ groups near the ester group, accompanied by C–O stretching vibrations. Ignored these details, our theoretical analysis classified as CH₂ bending modes in this band. Broad peaks in the $1400-1500 {\mathrm{c}\mathrm{m}}^{-1}$ region are observed in IR spectra of the four esters. These modes are identified as C–H bending vibrations of ${\mathrm{CH}}_2/{\mathrm{CH}}_3$ groups too. Experimental studies have reported Raman peaks near$1440 {\mathrm{c}\mathrm{m}}^{-1}$ cholesterol esters and IR peaks near $1470 \mathrm{c}\mathrm{m}⁻¹$, attributed to ${\mathrm{CH}}_2/{\mathrm{CH}}_3$ scissoring and bending vibrations [28,29,30,31].

CLA has a special peak at around $1690 {\mathrm{c}\mathrm{m}}^{-1}$ in the Raman spectrum, which is distinct from other cholesteryl esters. Analysis of this vibrational mode revealed that there were two modes at $1682 {\mathrm{c}\mathrm{m}}^{-1}$and $1687 {\mathrm{c}\mathrm{m}}^{-1}$, which originate from the asymmetric and symmetric stretching vibrations of two C=C bonds in the FA chain. And there was a vibrational mode at $1707 {\mathrm{c}\mathrm{m}}^{-1}$ corresponding to the stretching vibration of C=C bond in the benzene ring. The three vibrational modes are depicted in Figure 6. The coupling effect of these C=C vibrational modes results in a strong, narrow-bandwidth Raman peak of CLA. For COA, which is also unsaturated, the C=C stretching vibrations were observed at $1706 {\mathrm{c}\mathrm{m}}^{-1}$ in the FA chain and at $1692 {\mathrm{c}\mathrm{m}}^{-1}$ in the benzene ring. The saturated CPA and CSA has only one C=C in the benzene ring, corresponding to the vibrations at $1680 {\mathrm{c}\mathrm{m}}^{-1}$ and $1681 {\mathrm{c}\mathrm{m}}^{-1}$, respectively. The reported experimental C=C vibrations occured at $1660-1680 {\mathrm{c}\mathrm{m}}^{-1}$ [28,30,33], and the theoretical results here show slightly blue shifts.

There is a dominant peak at $1744$, $1722$, $1734$, and $1738 {\mathrm{c}\mathrm{m}}^{-1}$ respectively in the simulated IR spectra of these four cholesterol esters corresponding to C=O stretching vibration as shown in Figure 7, which matches the reported experimental result [30,31,34,35,36]. It can be concluded that the C=O stretching vibrations of cholesteryl esters are in the range from $1720$ to $1750 {\mathrm{c}\mathrm{m}}^{-1}$, independent of the carbon chain length [37]. For the dynamic process, please see the supplementary file. The other strong peaks near$3000 \mathrm{c}\mathrm{m}⁻¹$ exhibit notable bandwidth features, mainly arising from C–H stretching vibrations of ${\mathrm{CH}}_2/{\mathrm{CH}}_3$ groups, aligning with previous lipid characteristic reports [28,29,30,31,36]. Some experiments have assigned the vibrations near $2850-2900 {\mathrm{c}\mathrm{m}}^{-1}$ and $2900-2950 {\mathrm{c}\mathrm{m}}^{-1}$ to the symmetric and asymmetric stretching vibrations of ${\mathrm{CH}}_2$ and ${\mathrm{CH}}_3$, respectively [30,31]. However, the calculated vibrational modes were limited in the range of $2900-3200 {\mathrm{c}\mathrm{m}}^{-1}$ . Analysis revealed that the vibrational modes at $2950-3000 {\mathrm{c}\mathrm{m}}^{-1}$ are primarily due to symmetric stretching of C–H bonds in ${\mathrm{CH}}_2$/${\mathrm{CH}}_3$ groups, while peaks at $3000-3200 {\mathrm{c}\mathrm{m}}^{-1}$ mainly originate from asymmetric stretching of C–H bonds in ${\mathrm{CH}}_2$/${\mathrm{CH}}_3$ groups of the FA chain, as shown in Figure 8, which shows specific differences from previous experiments. Additionally, in the IR spectra of CSA, the intensity of C–H vibrations is slightly stronger than that of C=O vibrations. This may be attributed to its longer carbon chain with more CH₂ units and a saturated FA chain without C=C restrictions.

The calculated IR and Raman spectra in this study show certain differences in the characteristic peaks of the fingerprint region compared to the previously reported experimental spectra, which is mainly attributed to the simplification of theoretical models. These calculations were based on monomolecular model, ignoring the molecular environment and intermolecular non-covalent interactions. Thus, the positions or intensities of peaks in the fingerprint region, which are sensitive to intermolecular interactions, may be quite different. However, the strong IR peaks originated from C=O vibrations presented good agreement with the experimental reports. It is located in the range of $1720-1750 {\mathrm{c}\mathrm{m}}^{-1}$, i.e. 52 THz.

In this paper, we mainly focus on the C=O vibration peaks which exhibit high PPRA efficiency. In principle, the IR signals are due to photon-phonon resonance absorption while the Raman scattering signals are different frequency photon-phonon coupling. Since the light intensity of Raman scattering is ${10}^{-7}$ lower than that of IR, it is obviously that the best photothermal effect rely on the matchable frequencies between photon and phonon. Provided a laser of which the frequency is at around 52 THz focuses on the plaque, the C=O vibrations will obtain energy level transitions quickly. On the other way, the high electronegativity of C=O enables this group to participate in the stable formation of intermolecular hydrogen bond networks. The high vibrational energy of C=O bond may facilitate the hydrogen bonds breaking, and then destabilize the hydrophobic core of the lesion plaque and induce the phase transition of ester components from solid or liquid crystalline states to a liquid state. This disruption also affects unsaturated cholesteryl ester lipid droplets by weakening intermolecular forces and reducing viscosity. Cholesterol ester molecules experience reduced volume and increased mobility after phase transition, making them more susceptible to hydrolysis by esterases and subsequent excretion from the cell. This process consequently accelerates reverse cholesterol transport and diminishes lipid deposition within the vessel wall, thereby alleviating atherosclerosis.