Betaine–salicylic acid cocrystal for enhanced skincare and acne treatment

Abstract

Salicylic acid (SA) is a natural lipophilic active ingredient commonly used in cosmetics and skin disease treatments, offering benefits such as exfoliation, anti-inflammation effects, antibacterial properties, oil control, and acne alleviation. However, its poor water solubility, low bioavailability, and potential side effects, such as allergies, irritation, and dryness, hinder its widespread application. In this study, we prepared a betaine–salicylic acid (BeSA) cocrystal and systematically characterized its crystal structure, biological activity, and clinical efficacy. The results showed that BeSA has significantly lower irritancy and cytotoxicity than SA, but exhibits excellent anti-inflammatory and antioxidant properties as well as high moisturizing and anti-acne efficacy, making it a potential alternative to SA. Further, quantum chemical calculations and molecular docking simulations were conducted to investigate the intrinsic mechanisms underlying the excellent bioactivity of BeSA cocrystals. This study introduces an innovative solution for safer and more effective skincare formulations based on SA and offers theoretical guidance regarding material engineering and further material optimization, which has crucial implications for both industry and academia.

We developed a betaine–salicylic acid cocrystal with enhanced bioactivity, biocompatibility, and clinical anti-acne effect. Its structure–activity relationship was investigated through single-crystal analysis and multiple simulations.

1. Introduction

Acne is a skin condition with a relatively high prevalence, reaching 80–90% during adolescence, while that during adulthood is relatively low, usually between 20% and 40%, and women are more likely to be affected.¹ Acne-induced changes in appearance may cause anxiety and low self-esteem and can even affect the quality of life and social activities of patients. Salicylic acid (SA), a bioactive compound naturally occurring in the bark of plants, such as willow and birch trees, has demonstrated efficacy in relieving acne.² Its mechanisms of action include (1) sebum secretion modulation by altering the signaling pathways of sebocytes to control sebum production, thereby improving acne symptoms;³ (2) stratum corneum exfoliation by dissolving the substances between corneocytes and removing aged corneocytes from the skin surface, thereby preventing pore blockage and reducing acne formation;⁴ (3) facilitating follicular opening clearance by penetrating the pores and dissolving dead skin cells, thereby reducing the formation of blackheads and whiteheads;⁵ and (4) anti-inflammatory and broad-spectrum antibacterial activities, alleviating skin redness, pain, and inflammation while reducing the acne symptoms associated with bacterial infections.⁶ However, the application of SA has limitations such as easy crystallization and precipitation, intense irritation, and poor tolerance.

Cocrystallization technology is crucial for the development of new pharmaceuticals as it positively influences the physical and chemical properties of drugs, their bioavailability, and therapeutic efficacy.⁷ Intermolecular solid interactions between two or more components in a cocrystal lattice can induce structural changes in drug molecules, thereby affecting their properties. This technology can be used to enhance drug solubility, improve drug stability, modulate drug release rates, and reduce local drug concentrations, thereby alleviating drug irritancy.⁸ In a previous study, we successfully prepared supramolecular crystals of an ion salt by creating a complex of salicylic acid and matrine.⁹ Compared to SA, this material exhibits lower cytotoxicity and skin irritation along with enhanced water solubility, transdermal permeability, and biological activity. Not only did it surpass or match SA in terms of anticancer, antioxidant, anti-inflammatory, and antibacterial properties, it also demonstrated satisfactory anti-acne capabilities in clinical trials. However, limited plant sources and the high price of matrine resources necessitate further modification. As an alkaloid, betaine, which is widely sourced from beets, is economically viable, non-toxic to humans, and edible, making it an attractive alternative. A notable feature of betaine is its ability to regulate the osmotic balance and accelerate cell volume recovery during cellular shrinkage, resulting in excellent moisturizing properties.¹⁰ A single betaine molecule stably accommodates 12 water molecules in its hydration shell.¹¹ Furthermore, betaine enhances skin defense against UV damage through a multi-target and multi-pathway mechanism, including reinforcement of intracellular antioxidant systems, promotion of cell proliferation and differentiation, and reduction of the upregulation of various signaling factors under UV stress.¹² These characteristics indicate the substantial potential of betaine as a key ingredient in skincare products.

In this study, we synthesized a supramolecular cocrystal of betaine and SA (BeSA). Its structure was identified using various characterization techniques, including X-ray diffraction (XRD), nuclear magnetic resonance (NMR) spectroscopy, Fourier-transform infrared (FTIR) spectroscopy, and single-crystal analysis. Their thermal properties were studied using thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC). The cytotoxicity, irritancy, anti-inflammatory, oil-control, moisturizing, and anti-acne effects of BeSA were systematically characterized through cell experiments and human clinical trials and compared with those of SA and betaine. To better understand the structure–activity relationship, we generated single crystals and conducted quantum chemical calculations and molecular docking simulations to reveal the fundamental principles behind their bioactivity.

2. Experimental

2.1. Materials

Betaine (≥98%) and SA (≥99.5%) were acquired from Aladdin (Shanghai, China). Human keratinocytes (no. ES190826) and cell culture medium (no. KC2500) were obtained from Guangdong Biocell Biotechnology Co., Ltd. (Dongguan, China). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was purchased from Sigma-Aldrich (Guangzhou, China). Reactive oxygen species (ROS) and superoxide dismutase (SOD) assay kits were obtained from Beyotime (Shanghai, China). IL-1α, TNF-α, IL-8, and PGE2 ELISA kits were purchased from Abcam (Shanghai, China). Staphylococcus epidermidis (CICC 10294) and Propionibacterium acnes (ATCC 6919) were acquired from the China Industrial Culture Collection (Beijing, China) and American Type Culture Collection (Manassas, VA, USA), respectively.

2.2. Preparation of BeSA

First, 4 mM betaine was slowly added to 60 mL of an ethanol solution containing 4 mM SA and stirred to dissolve thoroughly. Subsequently, the solution was continuously stirred at 60 °C in a N₂ atmosphere for 24 h to ensure a thorough reaction. Afterward, BeSA crystals were collected using a rotary evaporator, washed with cold ethanol, and dried under vacuum at 40 °C for 24 h. Finally, single crystals of BeSA were prepared by slow recrystallization from its ethanol solution.

2.3. General characterization of BeSA

XRD spectra were recorded on a Rigaku D/max 2500PC X-ray powder diffractometer (Tokyo, Japan). A Hitachi SU8010 scanning electron microscope (SEM; Tokyo, Japan) was used to capture SEM images. NMR spectra were recorded on a Bruker Avance Neo 400 MHz spectrometer (Faellanden, Switzerland) using deuterated methanol. FTIR spectra were recorded using a Thermo Nicolet 380 spectrometer (Madison, WI, USA). The decomposition temperature (T_d) and melting point (T_m) were determined using a PerkinElmer TGA 8000 thermogravimetric analyzer (Waltham, MA, USA) and a Mettler Toledo DSC3 differential scanning calorimeter (Schwerzenbach, Switzerland), respectively. Single-crystal analysis of BeSA was conducted using an Agilent SuperNova, Dual, Cu at zero, AtlasS2 diffractometer (Palo Alto, CA, USA) and compared with single crystals of betaine and SA.^13,14

2.4. Cell experiments

2.4.1. Cytotoxicity

The neutral red uptake (NRU) test was performed according to OECD TG 432 with mouse NIH/3T3 fibroblasts (iCell Bioscience Inc, Shanghai, China). The cytotoxicity of betaine, SA, and BeSA on human keratinocytes was determined using the MTT assay, as described by a previous study.¹⁵

2.4.2. Anti-inflammatory and anti-oxidant assays

The inflammation and oxidative stress models were established using ultraviolet B (UVB) radiation. In detail, human keratinocytes in a logarithmic growth state were incubated in six-well plates with a concentration of 2 × 10⁵ cells per well overnight at 37 °C and under 5% CO₂ and 95% relative humidity. When the plating efficiency was 40–60%, the cells in each well were treated with 2 mL of medium containing betaine, SA, or BeSA at their respective safe concentrations and incubated for 24 h. Subsequently, the cells were exposed to 300 mJ cm⁻² UVB. For ROS measurement, the cells were washed three times with phosphate-buffered saline (PBS), and ROS levels were determined using the ROS kit according to the manufacturer's instructions with the assistance of a Beckman Coulter FC500 flow cytometer (Pasadena, CA, USA). In contrast, after UVB exposure, the cells were incubated for another 24 h before SOD, IL-1α, TNF-α, IL-8, and PGE2 measurements. Finally, the SOD activity was determined using the cells, while the IL-1α, TNF-α, IL-8, and PGE2 levels were determined using the cell culture supernatant, according to their respective kit instructions.

2.5. Antibacterial assay

First, SA and BeSA solutions of various concentrations were prepared. Then, a 2.5 mL sample solution was added to 2.5 mL of the double-strength culture medium (MHB for S. epidermidis and BHI for P. acnes). A bacterial suspension containing 5 × 10⁵ to 5 × 10⁶ cells was inoculated into test tubes containing the culture medium and samples at different concentrations. The bacterial suspension was also inoculated into the culture medium without samples and called PC. Additionally, the pure culture medium without bacteria and samples was called NC. After incubation at 37 °C for 48 h, a turbid PC and a clear NC solution indicated experimental validity, and the lowest sample concentration that resulted in no bacterial growth was determined as the minimum inhibitory concentration (MIC).

2.6. Computer simulations

Density functional theory (DFT) calculations were conducted using the ORCA 5.0.4 program.¹⁶ The structures of betaine, SA, and BeSA were optimized at the B97-3c level, and their single-point energies were evaluated at the M062X 6-311++G(d,p) level using the implicit solvation model (SMD). Electrostatic potential (ESP), independent gradient model based on the Hirshfeld partition (IGMH), atoms-in-molecules (AIM), and molecular frontier orbital analyses were performed using Multiwfn and VMD software programs.¹⁷

Molecular docking and molecular dynamic (MD) simulations were performed as described previously.¹⁸ In brief, the molecular docking of SA or BeSA to the Toll/interleukin-1 receptor (TIR, PDB ID: IFYV) was performed using the AutoDock 4.2.6 program, and MD simulations were performed using GROMACS 5.0.2 program based on the optimized structures. After 10 ns, the root-mean-square deviation stabilized, indicating that the system reached equilibrium. Thus, the interactions between the ligands and TIR were analyzed based on the last 10 ns of the MD simulation trajectory.

2.7. Clinical trials

To provide the participants with a better skin feel, we prepared the testing samples by adding 3 wt% BeSA and an equal molar amount of SA to the base formula, which was an aqueous solution containing 5 wt% butanediol and 2 wt% PEG/PPG-17/6 copolymer.

An enclosed skin patch test (no. PJJC-R-2302033, approved by the Ethics Committee of Shenzhen Shinehigh Innovation Technology Co., Ltd.) was conducted according to a previous study.⁹ Briefly, 0.020–0.025 g samples were enclosed on the forearms of 31 participants (3 men and 28 women, aged 20–58 years) for 24 h. The skin condition was scored and recorded at 0.5, 24, and 48 h after the patch removal. Zero points indicated no irritation response without erythema, edema, papules, vesicles, or other skin disorders.

The anti-acne test for BeSA (no. MC-GXRT202200044) was approved by the Ethics Committee of Guangdong Weihua Testing Technology Co., Ltd. Thirty participants with acne problems (16 men and 14 women, aged 18–30 years) were asked to apply the BeSA samples to their faces after cleaning at night for 28 days without other treatments. Facial condition was recorded weekly. Subsequently, after facial cleaning, they sat still for 0.5 h in the laboratory at a constant temperature and humidity. The transepidermal water loss (TEWL) was measured using a multifunctional skin analyzer with a TEWL probe, while the porphyrin, red, and acne lesions were measured using a VISIA-CR® facial scanner. Anti-acne testing for SA (no. MC-GXRT202200045) was conducted in the same manner as was done for BeSA with a different cohort of 30 participants (13 males and 17 females, aged 20–35 years).

2.8. Statistical analysis

Statistical analyses were conducted using GraphPad Prism 9.5 software. The cell experiments were performed in triplicate, and the results were expressed as mean ± standard deviation (SD) and compared using one-way analysis of variance tests. The clinical trials were conducted with no less than 30 participants, with results expressed as mean ± SD and compared using the paired t-test for the same group and unpaired t-test between the two groups. Differences were considered statistically significant at P < 0.05.

3. Results and discussion

3.1. Preparation and characterization of BeSA

The green cocrystal BeSA was prepared from naturally derived betaine and SA at a molar ratio of 1 : 1 (Fig. 1a). The XRD pattern of BeSA is distinctly different from those of betaine and SA, indicating the formation of a novel crystal structure (Fig. 1b). ¹H and ¹³C NMR spectra indicated that the BeSA was of high purity (Fig. 1c and S1†). Compared with betaine, the significant shift in H between the amino and the carboxyl groups in BeSA suggests strong hydrogen bonding, which led to a decrease in magnetic shielding around the carboxyl group of betaine. In the FTIR spectrum of SA, a broad absorbtance zone at 2530–3240 cm⁻¹ is observed for O–H stretching, which was attributed to the strong intramolecular hydrogen bonding of SA (Fig. 1d). Meanwhile, the C O stretching peak at 1660 cm⁻¹ corresponds to carboxylic acid dimers, consistent with the single-crystal analysis of SA (Fig. S2†). The broad FTIR peak of betaine around 3300 cm⁻¹ may be due to moisture absorption, and the peak at 3020 cm⁻¹ corresponds to the CH₃ group. The C O stretching peak at 1630 cm⁻¹ indicated a red shift, possibly due to solid electrostatic interactions between betaine molecules (Fig. S3†). Compared to SA, a blue shift was observed in the O–H stretching peak of BeSA, whereas a red shift occurred in the C O peak compared to betaine, indicating that supramolecular interactions between SA and betaine weaken the intramolecular hydrogen bonding of SA and the electrostatic interactions of betaine. Consequently, BeSA exhibits thermal properties that are distinct from those of SA and betaine (Fig. 1e). The T_m and T_d values of SA and betaine are very similar, whereas BeSA remains liquid over a wider temperature range, making it easier to process. The T_m of BeSA is approximately 112 °C, significantly lower than that of the precursors, indicating a low eutectic solvent. Its T_d is 220 °C, significantly higher than that of SA, indicating enhanced thermal stability.

Fig. 1. Synthesis and characterization of betaine–salicylic acid cocrystal (BeSA). (a) Formation scheme and morphology of BeSA. (b–e) X-ray diffraction (XRD) patterns, ¹H nuclear magnetic resonance spectroscopy (NMR), Fourier-transform infrared (FTIR) spectra, and thermal properties of BeSA and its precursor compounds.

Single crystals of BeSA were prepared and analyzed to visualize its molecular structure. The BeSA cocrystal belongs to the orthorhombic space group of Pbca; each unit cell has eight BeSA molecules with the following lattice parameters: a = 10.211, b = 10.775, and c = 22.531 Å (Tables S1–S8†). The molecular structure of BeSA is shown in Fig. 2a. The carboxyl groups of SA and betaine are closely positioned, forming hydrogen bonds. When viewed along the b-axis, BeSA molecules are arranged in two wavy layers, each exhibiting a sandwich structure with parallel SA molecules spaced 5.28 Å apart and sandwiching betaine molecules in between (Fig. 2b). Within the same layer, the dihedral angle between adjacent phenyl rings is 80.13°, whereas that between adjacent phenyl rings from different layers is 12.31°. SA within BeSA cocrystals still includes an intramolecular hydrogen bond with a length of 1.83 Å, longer than that in the SA crystal (Fig. S2†), indicating weaker intramolecular hydrogen bonding, consistent with the FTIR results (Fig. 1d). Along the b-axis, the betaine and SA molecules are alternately connected through O1–H1⋯O4 and O3–H4⋯O4 hydrogen bonds, forming a linear chain. The length of O3–H4⋯O4 is 1.53 Å, shorter than the intermolecular hydrogen bonds within the SA crystal, suggesting a stronger hydrogen bonding interaction between betaine and SA, and resulting in higher thermal stability (Fig. 1e). Viewed along the a-axis, the BeSA molecules are also arranged in two layers, with neighboring BeSA molecules within the same layer exhibiting axial symmetry, while those in different layers exhibit central symmetry (Fig. 2c). Viewed along the c-axis, the BeSA molecules also displayed wavy stacking with hydrogen bonding within layers but not between layers (Fig. 2d). The wavy arrangement of the BeSA molecules resulted in a more loosely stacked structure than that of the betaine and SA crystals, leading to a lower melting point (Fig. 1e). Hirshfeld surface analysis revealed that compared to pure SA, the SA molecules in BeSA cocrystals exhibit comparable O⋯H and C⋯H contacts and more H⋯H contacts but no C⋯C and C⋯O contacts with surrounding atoms (Fig. 2e). This indicates a lack of π–π stacking interactions between BeSA molecules, which is further demonstrated by the presence of more red concavities in the shape-index map and more blue borders in the curvedness map (Fig. 2e). The reduction in π–π stacking can be attributed to the insertion of betaine molecules, which increases the angle or spacing between adjacent SA molecules (Fig. 2b and S2†).

Fig. 2. Single-crystal analysis of betaine–salicylic acid cocrystal (BeSA). (a) Molecular structure of BeSA. (b–d) Molecular arrangement of BeSA along different axes, with hydrogen bonds represented by the dotted green lines and parallel benzene rings highlighted in the same color. (e) Hirshfeld surface analysis of salicylic acid (SA) and BeSA, mapped with d_norm, shape index, curvedness, and 2D fingerprint plots.

We also conducted DFT simulations to elucidate the structure-activity relationship of BeSA. The ESP maps revealed that betaine is a polar molecule with a high proportion of polar surface area, consistent with its hydrophilicity, while SA has an ESP distribution more concentrated around zero, consistent with its lipophilicity (Fig. 3a). IGMH analysis indicated that in BeSA, the most positive site of SA forms strong hydrogen bonds with the most negative site of betaine (Fig. 3b), resulting in a reduced negative potential compared with betaine, consistent with ¹H NMR shifts (Fig. 1c). Molecular frontier orbital theory was used to infer the reactivity of BeSA (Fig. 3c). The highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) of BeSA are both located on SA, with distributions that are almost identical to those of a single SA molecule, indicating that SA is the primary active site of BeSA and that they have similar antioxidant mechanisms. Compared to SA, BeSA exhibited higher HOMO, LUMO, and bandgap energies (E_g), suggesting that it is more prone to losing electrons, less prone to gaining electrons, and has higher stability (Fig. 3c). The phenolic hydroxyl group of SA plays a significant role in radical scavenging activity.¹⁹ The phenolic hydroxyl bond length of BeSA was longer than that of SA, with a greater charge difference at the ends, indicating easier dissociation (Fig. 3d). Current radical scavenging mechanisms mainly include (1) hydrogen atom transfer (HAT), with the key thermochemical parameter of bond dissociation enthalpy (BDE); (2) sequential proton loss electron transfer (SPLET) involving thermochemical parameters of adiabatic ionization potential (IP) and proton dissociation enthalpy (PDE); and (3) single-electron transfer–proton transfer (SET-PT), involving thermochemical parameters of proton affinity (PA) and electron transfer enthalpy (ETE).^20,21 The comparable BDE, IP + PDE, and PA + ETE all demonstrated similar antioxidant abilities (Fig. 3e),^22,23 making BeSA a potential alternative candidate to SA.

Fig. 3. Density functional theory (DFT) simulations of betaine–salicylic acid cocrystal (BeSA). (a) Electrostatic potential (ESP) analysis of betaine, salicylic acid (SA), and BeSA. (b) Independent gradient model based on the Hirshfeld partition (IGMH) analysis of BeSA. (c–e) Frontier molecular orbital analysis, phenolic hydroxyl group analysis, and antioxidant ability comparison of SA and BeSA.

3.2. Bioactivities of BeSA

Safety and efficacy are key factors in evaluating biomaterials. In this study, we assessed the safety of BeSA using cytotoxicity and NRU assays. The highest concentration at which cell viability remained above 90% was considered the safe concentration. SA exhibited high cytotoxicity towards human keratinocytes, with a safe concentration of only 0.004 mg mL⁻¹ (0.03 mM) (Fig. 4a). In contrast, betaine displayed extremely low cytotoxicity, with a safe concentration of up to 5 mg mL⁻¹ (Fig. 4b). This indicated that their cocrystallization would not exacerbate biocompatibility issues of SA. The safe concentration of BeSA was 0.125 mg mL⁻¹ (0.49 mM) (Fig. 4c), significantly higher than that of SA, suggesting that supramolecular interactions in BeSA improved the biocompatibility of SA. This assumption was further confirmed by NRU tests, which showed that BeSA was not phototoxic, and its half-inhibitory concentration (IC₅₀) was more than three times that of SA, indicating a much lower irritancy (Fig. 4d). We also investigated the anti-inflammatory effects of betaine, SA, and BeSA at safe concentrations using a UV irradiation-induced inflammatory keratinocyte model. Post-UV irradiation, the inflammatory cytokines IL-1α, IL-8, PGE2, and TNF-α were overexpressed (NC), but their levels were effectively reduced by the drugs (PC), confirming the successful establishment of the model (Fig. 4e–h). These four cytokines are key players in skin infections, inflammation, and disorders, including acne, psoriasis, and contact dermatitis. BeSA significantly reduced IL-1α levels, while betaine and SA were ineffective (Fig. 4e). Both SA and BeSA lowered IL-8 levels, with BeSA showing a much stronger effect (Fig. 4f). Betaine decreased PGE2 upregulation, whereas SA was ineffective; however, the impact of BeSA was not significantly different from that of betaine (Fig. 4g). Betaine, SA, and BeSA significantly and equivalently inhibited TNF-α overexpression (Fig. 4h). These results indicate that BeSA possesses excellent anti-inflammatory effects and inhibited the overexpression of these four inflammatory cytokines, whereas betaine and SA were only partially effective. BeSA also demonstrated strong antioxidant effects, consistent with the DFT calculations, and scavenged ROS and enhanced SOD activity. Furthermore, the antibacterial effect of BeSA was superior to that of SA, with a similar MIC against P. acnes and a lower MIC against S. epidermidis. Therefore, BeSA appears to be a promising alternative to SA, offering improved safety and efficacy.

Fig. 4. Bioactivities of betaine–salicylic acid cocrystal (BeSA). (a–c) Cytotoxicity of betaine, salicylic acid (SA) and BeSA. (d) NRU tests of SA and BeSA in NIH/3T3 cells and their corresponding IC₅₀ values. (e–h) Anti-inflammatory effects of betaine, SA, and BeSA. (i and j) Anti-oxidation performance of BeSA. (k) Antibacterial activity of SA and BeSA. Results are presented as mean ± SD for n = 3; ns = no significant difference; *p < 0.05; ⁺p < 0.05, ⁺⁺p < 0.01 vs. NC.

Toll-like receptors (TLRs) are a family of pattern recognition receptors that play crucial roles in skin immune responses.²⁴ They recognize various pathogen- and damage-associated molecules, triggering downstream inflammation via signaling pathways, such as MyD88/NF-κB, leading to the release of inflammatory mediators and cytokines.^25,26 SA reduces skin inflammation through multiple mechanisms, and its inhibition of the NF-κB pathway suggests that TLRs may be potential targets of its action. TLR2 and TLR4 are the primary TLRs involved in skin inflammation, sharing a conserved cytoplasmic domain (TIR) that is essential for initiating downstream signaling.^27,28 To explore why BeSA has lower irritation and superior anti-inflammatory performance, we performed molecular docking of SA and BeSA with TIR.

The optimized docking results showed that BeSA formed more hydrogen bonds with TIR than with SA, implying stronger binding energy, which may explain its enhanced anti-inflammatory efficacy (Fig. 5a). In the TIR–BeSA system, the hydrogen bonds between SA and TIR were weaker and longer than those in the TIR–SA system, explaining the low irritation by BeSA. MD simulations confirmed these results (Fig. 5b). In these systems, TIR acts as a hydrogen-bond donor, whereas SA and BeSA are acceptors. All the hydrogen bonds between SA and TIR were N–H⋯O, whereas those between BeSA and TIR contained both N–H⋯O and stronger O–H⋯O bonds. The latter typically has higher binding energy, which may further enhance the anti-inflammatory effects of BeSA. In addition, betaine in the TIR–BeSA complex partially occupied the SA-binding sites in TIR, which may have reduced irritation (Tables S9 and S10†). Notably, BeSA only altered the local structure of the TIR without protein denaturation, suggesting its high biosafety potential (Fig. 5c). Energy analysis showed that both the TIR–SA and the TIR–BeSA systems possess similar solvation free energies (Fig. 5d), including ΔG_pb, ΔG_np, and ΔG_sol, likely attributed to the excellent solubility of betaine. The positive ΔG_pb and ΔG_sol indicate that polar interactions (e.g., hydrogen bonding) are less stable in water than in vacuum, leading to a detrimental impact of solvation on drug–protein binding. However, the TIR–BeSA system exhibited a more negative ΔG_ele than the TIR–SA system, making ΔG_polar change from positive to negative and resulting in a lower ΔG_total. The enhanced Coulomb electrostatic interactions in the TIR–BeSA system may be related to the stronger electronegativity of BeSA (Fig. 3a), which favors the formation of hydrogen bonds with TIR as an acceptor. As a result, BeSA produced a stronger attraction with residues in the TIR (e.g., Phe637, Lys677, and Thr685), resulting in an increased number and stability of hydrogen bonds compared to the TIR–SA system (Fig. 5b).

Fig. 5. Molecular docking analysis of salicylic acid–Toll/interleukin-1 receptor (SA–TIR) and betaine–salicylic acid cocrystal–Toll/interleukin-1 receptor (BeSA–TIR) complexes. (a) Optimized docking results. (b) Number of hydrogen bonds in SA–TIR and BeSA–TIR complexes. (c) Structure superposition diagram after molecular dynamic (MD) simulations. (d) Binding free energies (ΔG_ele, Coulomb electrostatic interaction energy; ΔG_vmd, van der Waals interaction energy; ΔG_MM = ΔG_ele + ΔG_vmd, molecular mechanics energy in vacuum; ΔG_pb, polar solvation energy; ΔG_np, non-polar solvation energy; ΔG_sol = ΔG_pb + ΔG_np, total solvation energy; ΔG_polar = ΔG_ele + ΔG_pb, total polar binding energy; ΔG_nonpolar = ΔG_vmd + ΔG_np, total non-polar binding energy; ΔG_total = ΔG_polar + ΔG_nonpolar, total binding energy). (e) Residue contributions.

We also compared the anti-acne efficacy of BeSA and SA in clinical trials (Fig. 6) after conducting patch tests to ensure the skin safety (Table S11†). TEWL is a critical parameter for assessing the barrier function of the skin. Higher TEWL values indicate greater water dissipation through the skin and poorer barrier function of the stratum corneum.²⁹ Porphyrins are bacterial metabolites that parasitize hair follicles and fluoresce under UV radiation. They reflect sebum secretion from the skin and are often associated with the development of acne.³⁰ Red lesions reflect skin problems related to blood vessels, including acne and inflammation.³¹ As shown in Fig. 6a, treatment with SA or BeSA significantly reduced TEWL, porphyrin, red lesions, and acne levels on the subjects' faces, with BeSA showing a more pronounced effect over time. At 14 days, BeSA treatment reduced TEWL, porphyrin, red lesions, and acne by 13.12%, 18.33%, 21.29%, and 19.62%, respectively, outperforming SA treatment (12.44%, 14.61%, 13.01%, and 12.24%, respectively); however, the difference between the two groups was not significant (Fig. 6b). At 28 days, BeSA treatment led to a 43.45%, 40.80%, 44.37%, and 37.51% reduction in TEWL, porphyrin, red lesions, and acne, significantly higher than those after SA treatment (21.90%, 26.16%, 24.52%, and 24.01%, respectively), indicating that BeSA had superior repair, antibacterial, anti-inflammatory, and anti-acne effects, possibly due to its low irritation and high efficacy (Fig. 4). Furthermore, BeSA also scored higher in self-assessment of oil control, acne reduction, and absorption, leading to a greater likelihood of repurchase, highlighting its considerable market potential. In summary, the excellent repair, antibacterial, anti-inflammatory, and anti-acne efficacies of BeSA (Fig. 6c) make it a promising alternative to SA, offering a better choice for acne treatment and skin care.

Fig. 6. Clinical trials of 3 wt% betaine–salicylic acid cocrystal (BeSA) and equimolar salicylic acid (SA). (a) Changes in skin condition of subjects' faces following treatment with BeSA and SA. (b) Percentage change in transepidermal water loss (TEWL), porphyrin, red lesions, and acne levels on subjects' faces after treatment with BeSA and SA. (c) Representative photographs of subjects after treatment with BeSA. (d) Satisfaction survey results from subjects following treatment with BeSA and SA. Results are presented as mean ± standard deviation for n = 30, *p < 0.05; **p < 0.01.

4 Conclusions

In this study, we synthesized a novel cocrystal, BeSA, from betaine and SA and systematically compared its crystal structure, biological activity, and clinical efficacy with those of SA. Single-crystal analysis and DFT calculations revealed that strong hydrogen bonding from betaine altered the electron distribution of SA, enhancing thermal stability and reactivity in BeSA. Molecular docking and MD simulations illustrated that BeSA binds more strongly to the target protein, suggesting higher bioactivity. Consequently, BeSA exhibited lower cytotoxicity, better anti-inflammatory and antibacterial properties in vitro, and superior repair, oil control, anti-inflammatory, and anti-acne effects in clinical trials. In summary, the cocrystallization of SA with betaine improved its bioactivity and clinical efficacy, making BeSA an excellent alternative to SA for acne treatment and skin care. This study provides a new perspective on the application of SA cocrystals in topical transdermal delivery and supports the development of novel anti-acne treatments. However, this work still has some limitations, including the simplified solution sample not reflecting the real formula in practice, the lack of fundamental research on transdermal delivery, and the relatively small number of subjects in clinical trials. We will continue to work on these issues in the future to promote the clinical application of BeSA.

Data availability

The data supporting this article have been included as part of the ESI.† Crystallographic data for this paper can be obtained free of charge from The Cambridge Crystallographic Data Centre (CCDC 2101906).

Author contributions

Conceptualization, methodology, and data curation were carried out by Z. W. and M. W. Z. W. wrote the original draft of the manuscript. Q. T., Y. L., H. W., M. Z., and X. L. were responsible for investigation, validation, and visualization. J. Z. provided supervision and contributed to the review and editing of the manuscript.

Conflicts of interest

There are no conflicts to declare.