Assessment of Transdermal Delivery of Topical Compounds in Skin Scarring Using a Novel Combined Approach of Raman Spectroscopy and High-Performance Liquid Chromatography

Abstract

Objective: The goal of any topical formulation is efficient transdermal delivery of its active components. However, delivery of compounds can be problematic with penetration through tough layers of fibrotic dermal scar tissue.

Approach: We propose a new approach combining high-performance liquid chromatography (HPLC) and Raman spectroscopy (RS) using a topical of unknown composition against a well-known antiscar topical (as control).

Results: Positive detection of compounds within the treatment topical using both techniques was validated with mass spectrometry. RS detected conformational structural changes; the 1,655/1,446 cm⁻¹ ratio estimating collagen content significantly decreased (p < 0.05) over weeks 4, 12, and 16 compared with day 0. The amide I band, known to represent collagen and protein in skin, shifted from 1,667 to 1,656 cm⁻¹, which may represent a change from β-sheets in elastin to α-helices in collagen. Confirmatory elastin immunohistochemistry decreased compared with day 0, conversely the collagen I/III ratio increased in the same samples by week 12 (p < 0.05, and p < 0.0001, respectively), in keeping with normal scar formation. Optical coherence tomography attenuation coefficient representing collagen deposition was significantly decreased at week 4 compared with day 0 and increased at week 16 (p < 0.05).

Innovation: This study provides a platform for further research on the simultaneous evaluation of the effects of compounds in cutaneous scarring by RS and HPLC, and identifies a role for RS in the therapeutic evaluation and theranostic management of skin scarring.

Conclusions: RS can provide noninvasive information on the effects of topicals on scar pathogenesis and structural composition, validated by other analytical techniques.

Ardeshir Bayat, MB BS, PhD

Introduction

The goal of any topical formulation should be efficient transdermal delivery of its actives in order to address the signs and symptoms of abnormal skin scarring.¹ There are three basic layers that make up skin: the outermost epidermis, the dermis (divided into an upper papillary layer and a thicker reticular layer), extending into the inner hypodermis (fatty layer).² The dermis contains all the connective tissue components which provide tensile strength and elasticity of the skin. These properties are altered through the remodeling process of scarring.³ Separating the dermis from the epidermis is a basement membrane.³ Upon wounding, there is disruption of the layers, resulting in a loss of structure, and a replacement with scar tissue.³ As there is no reconstitution back to its original state, the resultant scar may be linear, depressed (atrophic), contracted, or raised (hypertrophic or keloid) manifesting in a range of symptoms, including erythema, pruritus, and pigmentation, and as such, responsible for the significant retail market for antiscar treatments.⁴

The commonly utilized treatments include the use of inert silicone gels, oils, and sheets shown to provide hydration by occlusion.⁵ Other widely used topicals claim to provide moisturization for symptomatic relief of dryness, for example aloe vera⁶ and vitamin E,⁷ or scar reduction with improvement in erythema and elasticity, using topicals containing compounds with anti-inflammatory properties, such as epigallocatechin gallate (EGCG),⁸ and formulas such as MEBO^™ (moist exposure burn ointment) Scar ointment respectively.⁹

Delivery of active compounds however, can be problematic with penetration of the stratum corneum of normal skin let alone through abnormal dermal scars containing excessive fibrous tissue. Current widely used methods for assessment of transdermal delivery of active compounds include the use of high-performance liquid chromatography (HPLC) and mass spectrometry (MS). MS is commonly used to validate HPLC findings as it provides confirmation of the active ingredient(s),¹⁰ and LC-MS (liquid chromatography/mass spectrometry) has even been used to validate Fourier transform infrared (FTIR) spectroscopy analysis of the human skin barrier.¹¹

Another useful method for detection and assessment of transdermal delivery of active compounds is Raman spectroscopy (RS); a nondestructive, vibrational, spectroscopic technique that is used to probe the molecular changes within biological tissue.¹² RS provides a biomolecular fingerprint for macromolecules, which contains both chemical and physical information. As shown by Mendelsohn's group, when conformational changes occur in tissue development, disease, repair, or regeneration, they manifest in the spectra.¹³ RS is not only used for assessment of transdermal delivery but in both basic research and in the clinical setting, where analysis of normal skin composition has been demonstrated, in a recent study¹⁴ using RS to obtain quantitative, noninvasive spectral data on skin thickness. Successful discrimination between layers of the epidermis, (normally affected after formalin fixation which is also invasive), was evidenced due to the higher spatial resolution of RS (1 μm is typical) compared to other noninvasive techniques such as high-frequency ultrasound (HFUS) and optical coherence tomography (OCT).¹⁴

Clinical Problem Addressed

Topical treatments are the mainstay of management options for the prevention and symptomatic relief of scars; a high number of which are over-the-counter remedies. Despite the abundance of topicals in the market, limited scientific evidence exists for their effect. Penetration through tough fibrotic scar tissue remains problematic in the assessment of antiscar topical formulations, and the perceived lack of a robust methodology to assess transdermal delivery of their actives presents further difficulties in providing objective and quantitative analysis. We present a combined approach to the assessment of transdermal delivery of topicals, while synergistically providing information on the conformational changes taking place within the dermal scar tissue, which is corroborated by conventional histological techniques.

To demonstrate the clinical application of RS in the assessment of skin scarring, we assess transdermal delivery of a topical (MEBO Scar ointment),⁹ which has an unknown mechanism of action and no proven known actives. First, using an ex vivo human skin organ culture model, utilizing the traditional method of HPLC.¹⁵ This technique confirmed the presence of two compounds, tyramine and linoleic acid, and that their delivery could be evidenced within the different dermal layers; the confirmatory presence of these compounds was validated by MS. Effective transdermal delivery was then observed using RS. To validate these findings, the same techniques were applied on in vivo skin samples obtained from a clinical trial with the same topical and compared with a positive control, (Kelo-cote^®).¹⁶ Kelo-cote was selected, as silicone-based topicals are most commonly used in clinical practice and their effect shown to be positive in skin scarring as evidenced by a Cochrane Review.¹⁷

We therefore hypothesized that a combined approach of techniques, allows for successful validation of transdermal delivery of a topical even though the compounds within it remain unknown. Furthermore, there may be a role for RS in monitoring therapeutic effects due to the conformational changes detected within scar tissue. Figure 1 demonstrates the experimental workflow for this study; Fig. 1A represents the ex vivo, and Fig. 1B, the in vivo components of the study.

Figure 1.

Experimental workflow. (A) Describes the ex vivo work, which was initially carried out. To validate the ex vivo findings, skin samples from a clinical trial using the same topicals (B) were used to test the same techniques (HPLC, and RS). MS was then used to confirm the presence of both compounds, which were detected in HPLC. Additional conformational changes demonstrated using RS were validated with histology stains for elastin and collagen, and clinically with OCT attenuation coefficient. HPLC, high-performance liquid chromatography; MS, mass spectrometry; OCT, optical coherence tomography; RS, Raman spectroscopy. Color images are available online.

Materials and Methods

Ex vivo human skin organ culture model

We utilized the culture method previously validated and evidenced by our group.¹⁵ Discarded human skin samples were obtained from patients undergoing routine esthetic surgery, such as abdominoplasty, who were consented before their procedure with ethics approval (16/NW/0736) from the University of Manchester (Manchester, United Kingdom). All tissues were tracked and stored in a human tissue biobank following Human Tissue Act (HTA) guidelines. Skin was obtained from six patients; three samples were nonscarred, that is, normal skin, and three were scarred skin samples.

After trimming the excess hypodermal fat, circular organ culture explants were prepared with a 6 mm biopsy punch (Kai Medical). The biopsies were transferred to a 24-well plate containing transwell inserts (Corning^©), allowing them to be suspended in the culture media while keeping the epidermis exposed to the air/liquid interface, for the addition of the treatment. Each biopsy was cultured in 500 μL of Williams E culture media (Gibco^™) supplemented with 1% penicillin/streptomycin, 1% l-glutamine, 1% nonessential amino acid solution, 1% ITS +3, and 10 ng/mL hydrocortisone. The culture medium was changed daily, and the biopsies were maintained at 37°C in a humidified incubator with a 5% CO₂ atmosphere.

Topical formulations

MEBO Scar ointment (Julphar Gulf Pharmaceutical Industries, Ras al Khaimah, United Arab Emirates), comprising sesame, cactus, and beeswax, was originally formulated in China from its derivative MEBO (Julphar Gulf Pharmaceutical Industries).⁹ MEBO Scar claims to possess at least two compounds, tyramine, and linoleic acid, (unknown actives), which are thought to improve skin scarring through an enhanced remodeling process and improved microcirculation.⁹ In this study we aim to demonstrate the presence of these compounds through dermal scar tissue using the combined aforementioned techniques. Kelo-cote (Sinclair Pharma), comprising 100% silicone base was used as a positive control.

Assessment of transdermal delivery using ex vivo samples

Normal skin ex vivo organ culture samples were used to assess transdermal delivery of tyramine and linoleic acid. The biopsies were divided into three groups (n = 9). The first group had 5 μL of the treatment topical applied to the epidermis, the second group had 5 μL of the positive control applied, while the control group had neither topical applied. On days 0, 3, 7, and 10, the biopsies were washed with phosphate-buffered saline and then snap frozen and stored at −80°C ready for further analysis.

High-performance liquid chromatography

This was also based on the methodology previously demonstrated by our group.¹⁵ Three samples for each condition (n = 9) were used for the ex vivo experiments, and n = 15 for the in vivo (3 for each time point, day 0, treatment topical at weeks 8, 12, and 16, and positive control topical at week 12). An optimized Agilent 1100 system (Agilent Tech) was utilized, and data analysis performed using ChemStation (Agilent Tech) software. Pure linoleic acid and tyramine (Sigma-Aldrich, United Kingdom) were tested at different concentrations (100 mM, 10 mM, 100 μM, 10 μM), with a column (Sphereclone; Phenomenex, Inc.) ideal size of 4.6 × 250 mm, to generate standard curves for quantification. The solvent condition was acetonitrile at different concentrations (20%, 40%, 70%, 80%, and 90%) with 1% TFA (trifluoroacetic acid). The flow rate was 1ml/min at 10, 50, and 100 μL volumes. Detection of linoleic acid tyramine was at 225 nm wavelength, at 6 and 2 min, respectively.

Mass spectrometry

Skin tissue was weighed to ensure it was less than 25 mg. A total of nine samples from different conditions, such as treatment, control, and day 0 (no treatment) were used. Aliquots (1 mL of hexane and isopropranolol, at a ratio of 3:2 was added to each sample), and was homogenized in a tissue lyser (Qiagen), then centrifuged at 15,890 g for 15 min. The supernatant was collected for MS. All analyses were conducted on a QExactive Plus coupled with an Ultimate 3000 UHPLC (ultra HPLC) (Thermo). The UHPLC was equipped with an Accucore Vanquish reverse-phase column (C18—2.1 × 100 mm; 1.5 mm particle size). The solvents employed were (A) water +0.1% formic acid and (B) methanol +0.1% formic acid. The flow gradient was programmed to equilibrate at 98% A for 2 min followed by a linear gradient to 95% B over 10 min and held at 95% B for 2 min before returning to 95% A for 2 min. The flow rate was 300 mL/min. The column was maintained at 40°C and the samples chilled in the autosampler at 45°C. A sample volume of 5 μL was injected into the column. Data acquisition was conducted in full MS mode in the scan range of 90–1,350 m/z with a resolution of 70,000, an automatic gain control target of 3e⁶ and a maximum integration time of 100 ms. The samples were analyzed in positive (tyramine) and negative modes (linoleic acid) in separate acquisitions.

Clinical trial for in vivo human skin samples

Forty-five subjects were recruited and followed up over a period of 8 months, to evaluate the role of MEBO Scar (Julphar Gulf Pharmaceutical Industries) in skin scarring.¹⁷ Samples from twenty of these subjects were used for assessment of transdermal delivery using RS. All subjects signed a written consent form (UREC ref 16098), in addition to verbally consenting at each visit in keeping with the declaration of Helsinki principles. Sample size and study power was determined by an independent statistician from the University of Manchester, considering all background variables, and samples equally distributed with no bias. Recruitment was through the University volunteer intranet pages. All subjects were fit and healthy volunteers, were screened before participation, and those with any relevant medical history were excluded. The trial was registered on ISRCTN registry (ISRCTN16551998).¹⁷

After the screening appointment, all subjects received a 5 mm punch biopsy under local anesthetic, to both upper inner arms, to create a uniform scar. The subjects were then divided into 4 groups, representing a temporal sequential time point, at 4, 8, 12, and 16 weeks. The subjects attended fortnightly appointments, where noninvasive measurements (including OCT) were taken. At the first visit following the biopsy, they were given both the treatment topical, and the positive control, with instructions on how to apply them. The subjects, and the trial clinician were blinded as to which arm received treatment or control. At their final appointment (dependent on group), the scar was excised through a 6 mm punch biopsy under local anesthetic, and at this point the subject exited the study.¹⁶ All visits took place at the University Hospital of South Manchester, part of Manchester University NHS Foundation Trust (Manchester, United Kingdom). Other outcomes of this study were published in another study by our group.¹⁶

Raman spectroscopy

Biopsies were snap frozen and stored at −80°C before use. They were sliced into cross-sections of 10 μm thickness and mounted onto Raman-grade calcium fluoride slides (Crystran Ltd.). Raman analysis was undertaken using a Renishaw inVia confocal Raman microscope (Renishaw Plc.) equipped with a 785 nm laser. The instrument was calibrated using a silicon plate focused under a × 50 objective, where a static spectrum centered at 520 cm⁻¹ for 1 s at 10% power was collected.

Spectral data were collected using the Windows-based Raman environment (WiRE) 3.4 software (Galactic Industries Corp.). All spectra were acquired using the laser power adjusted on the sample to ∼28 mW. An average of six static spectral point measurements were taken in each layer. Data acquisition was optimized for our samples; using 100% power of the laser, with ten second exposure time and ten accumulations (total measurement time of 100 s), to analyze each point with the most optimal signal to noise. After data collection, any spurious spikes (cosmic rays) in the spectra were removed using the Cosmic Ray Removal function in WiRE. Raman spectra were truncated in the fingerprint region (750–1,800 cm⁻¹), baseline corrected, scaled, and smoothed (Savitzky–Golay filter) using MATLAB software version R2019b (The MathWorks, Inc.). Area under the bands peaking at 1,655 and 1,446 cm⁻¹ were calculated using in-house scripts and the ratios 1,655/1,446 were tested by one-way analysis of variance (ANOVA) and multiple compared using Tukey's test.

Histology

Formalin-fixed samples (n = 16 for each stain) were sectioned on a microtome into 5 μm slices and placed onto glass slides. The slides were deparaffinized by xylene and rehydrated through grades of alcohol. Picrosirius Red staining was performed using the Picro Sirius Red Stain Kit ab150681 (Abcam). The Picrosirius Red staining was viewed by an Olympus BX63 microscope for polarized images; collagen I was shown in red/orange and collagen III was shown in green/yellow. Elastin staining required antigen retrieval using 10% v/v citrate buffer, (Sigma-Aldrich) for 20 min. Overnight incubation with 1:600 anti-elastin antibody, ab 23747 (Abcam). Alexa Fluor 488 secondary antibody A-11034 (Thermo Fisher) was added to the sections and incubated for 40 min. DAPI, 62248 (Thermo Fisher) was used to counterstain. Sections were then dehydrated, cleaned, and mounted.

For Picrosirius Red, ImageJ was used to process the regions of interest, splitting the images into different color channels: red for collagen I quantification and green for collagen III. Analysis of elastin used Definiens Tissue Studio (Definiens) to quantify the region of interest. The frequency distribution of the collagen and elastin expression data was assessed by the Shapiro–Wilk normality test. All data passed the normality test and were evenly distributed. Subsequently, one-way ANOVA with repeated measurements was performed for both the collagen and elastin datasets. The mean collagen I/III ratio and the mean elastin level of each week were compared with the mean values of all other weeks. Differences were considered to be statistically significant when p < 0.05.

Results

HPLC and MS

Following the methodology described above, the standard concentrations of both linoleic acid and tyramine were ascertained. Linoleic acid was present at ten times higher concentration, making it easier to detect; neither compound was present in the control topical (Fig. 2A). To test the penetration of each compound, samples were sectioned longitudinally and correlated with Hematoxylin and Eosin (H&E) staining to ensure the correct dermal layer was identified for HPLC (Fig. 2B). Whilst it was possible to detect linoleic acid in day 0 (normal skin, no treatment) and in the samples which received the positive control, it was detected at much higher levels in treatment topical samples, although notably this peak was smaller with deeper penetration, where it can be presumed that linoleic acid was present in much lower quantities. Tyramine was much more difficult to detect and was present in low quantities in all samples. In Fig. 2C the average peak area of linoleic acid in the treatment samples is much greater, implying that it was easier to detect in in vivo samples. As with the ex vivo samples, the quantities detected were lower with deeper penetration, and the average peak area for tyramine was much lower, although still detectable in the day 0 and control samples. To validate the presence of both compounds, MS was used to detect linoleic acid and tyramine in treatment samples, the average concentration of both compounds is shown in Fig. 2D.

Figure 2.

Comparison of the detected concentrations of tyramine and linoleic acid in the topicals, and treated samples through HPLC and MS. (A) Demonstrates the presence of both compounds in the treatment topical, and neither in the positive control. (B) Shows the results for HPLC on ex vivo samples (n = 9): an average of the area under each peak (for linoleic acid and tyramine) was taken to quantify the detection of both compounds in each layer. (C) Shows the results for HPLC on in vivo samples (n = 15) using the average area under each peak for both compounds in each layer; as with (B), tyramine is present in much lower quantities. In (D) the average concentration of each compound detected in the skin samples with the treatment topical applied (n = 9) is shown. Key: D0, day 0, normal skin, no treatment; Control, samples which received the positive control topical; Treatment, samples which received the treatment topical. E, epidermis; P, papillary dermis; R, reticular dermis. Color images are available online.

Raman spectroscopy

Measuring the spectra of the treatment topical, distinctive peaks were noted at 1,042–1,120, 1,266–1,303, 1,439, and 1,656 cm⁻¹ (Fig. 3A) in the Raman spectra of samples treated with the topical; although all these peaks may account for lipid, protein, and nucleic acid vibrations in “normal” skin (listed in Fig. 3B),¹⁸ they were of higher attenuation. Figure 3C and D demonstrates the spectra for ex vivo samples for day 0, the positive control, and the treatment topical in the epidermis and dermis, respectively. The spectra are offset to facilitate visualization, and subtle changes are noted in the spectra of dermis and epidermis collected from tissue subjected to treatment and positive control compared with day 0.

Figure 3.

RS of the treatment topical compared with the positive control and no topical. (A) Demonstrates the spectrum of the treatment topical with the key bands of high attenuation highlighted. (B) A table with the band assignments for the peaks present within the treatment topical, which may also account for constituents of normal skin. (C) Shows spectra for the treatment topical against day 0 and the positive control in the epidermis. (D) Shows spectra for the treatment topical against day 0 and the positive control in the dermis. The spectra in (C, D) are offset to facilitate visualization. There are changes within the bands at 1,040–1,120 and 1,200–1,320 cm⁻¹, which are regions highlighted in the treatment topical in (A). Color images are available online.

In Fig. 4, the treatment topical skin spectra are represented at different time points, in the epidermis and dermis. The time points represent the length of application of the topical, that is, for 4, 8, 12, or 16 weeks compared with baseline (day 0). As previously described,¹⁹ skin structures most relevant to scarring are visible at the amide I, amide III, and proline bands. The amide I band represents collagen I in the dermis at 1,665 cm⁻¹, and at 1,654 cm⁻¹ in the epidermis representing the α-helix component of keratin.²⁰ The amide III band is broad with a peak at 1,246 cm⁻¹ associated with the proline-rich (nonpolar) and at 1,271 cm⁻¹ proline-poor (polar) regions within collagen.²⁰ Another band of note is at 1,446 cm⁻¹, which is attributed to CH₂ scissoring.²¹ Previous studies²¹ have demonstrated that the 1,655/1,446 cm⁻¹ band area ratio is correlated to the collagen content. When applying this to the spectra collected at each time point for the dermis in 4B, there is a significant decrease between the collagen content at weeks 4, 12, and 16 (p < 0.05), compared with day 0 (normal skin collagen content); despite the rise in collagen content by week 16, it does not return to normal levels (Fig. 4C).

Figure 4.

RS of in vivo treated skin scar samples at different time points. (A) Day 0 (no treatment, normal skin), compared with skin scar samples with the treatment topical, which had been applied for 4, 8, 12, and 16 weeks in the epidermis. (B) Day 0 (no treatment, normal skin), compared with skin scar samples with the treatment topical, which had been applied for 4, 8, 12, and 16 weeks in the dermis. In (A, B) the spectra are offset to facilitate visualization. Tukey's test was used to evaluate the significance of the band area ratio (1,655/1,446 cm⁻¹) relating to collagen content in (C); distinct letters indicate statistically significant differences between the groups (p < 0.05), that is, weeks 4, 12, and 16 were statistically significant compared with day 0. Color images are available online.

Interestingly, at week 16 in the treated group, RS showed a shift of the amide I peak from 1,667 to 1,656 cm⁻¹ in the dermis (Fig. 5). This shift in the peak may represent a change in protein conformation, from β-pleated sheets in elastin²² in the region of 1,665–1,675 cm⁻¹ to a region of 1,650–1,656 cm⁻¹, which represents α-helix components of collagen.²² Changes in the amide I peak representing secondary structural changes in protein have been recognized in skin previously: comparing normal skin to cancerous skin lesions²³ although never in the context of human skin scars. To corroborate these findings, traditional histology stains for collagen and elastin were carried out on skin scar samples, which had the treatment topical applied for 4, 8, and 12 weeks. The images in the column on the left of Fig. 6A represent Picro Sirius Red staining, imaged under polarized light. On day 0, there is an even distribution of collagen I and collagen III fibers. At weeks 4, and 8, there are more of the yellowish fibers of collagen III. At week 12, the distribution to collagen fibers is beginning to even out; Fig. 6B represents the ratio of collagen I/III fibers: at weeks 4 and 8, this is significantly less than day 0, and at week 12 this ratio was significantly higher than day 0 (p < 0.0001). The column on the right in Fig. 6A is anti-elastin antibody immunofluorescence staining. Even at week 12, there are very few elastin (green) fibers visible in the dermis. Figure 6C demonstrates a significant decrease (p < 0.05) in elastin levels at week 4 compared with day 0, and despite a rise at week 8, there is a significant decrease (p < 0.05) also at week 12 compared with day 0. While the RS collagen ratio does not demonstrate an increase to day 0 levels, the decrease in elastin and increase collagen over time fits with the process of fibrosis and corroborates with the shift in the amide I peak at week 16. Furthermore, clinical correlation using the OCT attenuation coefficient (Fig. 6D) when volunteers had noninvasive measurements, demonstrated a significant decrease in both treated and positive control arms compared with day 0 by week 4, which increased and was significantly greater in the treated arm (compared with the control), at week 16 (p < 0.05) (Fig. 6E), although still not at baseline levels—as demonstrated using RS.

Figure 5.

Protein conformational changes identified using RS at week 16 in treated samples. A shift in the amide I peak representing β-sheet contribution of elastin to α-helix component of collagen. RS therefore detected changes in wound healing and remodeling. Color images are available online.

Figure 6.

Assessment of collagen and elastin using histology and clinical correlation with OCT. (A) Shows the polarized light images the Picrosirius stain for collagen on the left, and immunofluorescence stain for elastin on the right at day 0, weeks 4, 8, and 12. On the Picrosirius stain, collagen I fibers stain red/orange, and collagen III yellow/green. Elastin fibers stain green with nuclei in blue. (B) Demonstrates the collagen I/III ratio, which is significantly decreased at weeks 4 and 8 and increased at week 12 compared with day 0, ***p < 0.0001. (C) Demonstrates the decrease in elastin at weeks 4 and 12, which was significant compared with day 0, *p < 0.05. (D) Demonstrates use of the OCT device and the image from which the attenuation compensation is taken. The attenuation coefficient represents collagen deposition, which is significantly less in both arms compared with day 0 by week 4 (p = 0.031), then increases in both arms, with a significant increase in the treatment arm compared with the control at week 16 (p = 0.047), although it has not quite returned to baseline (day 0) levels (E). Star symbol indicates significant p-values. Color images are available online.

Discussion

Our study has demonstrated how HPLC and RS can be used to assess transdermal delivery of compounds; penetrating scar tissue even when the precise composition and mechanism of action of the topical remains unknown. Identifying penetration within each dermal layer is also possible using HPLC, but relies upon correlation of each section with H&E staining or careful tissue dissection. HPLC also relies on ensuring appropriate set up and usage of additional substances (solvents), although it has a higher signal sensitivity than RS (where surface enhanced Raman scattering [SERS] may have to be employed to improve signal).^24,25 HPLC also detected the overall concentration of the target compound in the whole sample, as it is based on the analysis of extracts; while Raman is complementary as it allows for the detection of levels of the compounds nondestructively in a spatial manner with a pixel resolution of ∼1 μm. The use of cross-sectional samples whereby, under the high magnification of the confocal Raman microscope, dermal layers can be identified aids this process.

The novel finding of conformational changes within the scar tissue over sequential time points, could prove to be a major clinical benefit in the assessment of scarring, and provide additional objective and quantitative therapeutic assessment of topical formulations. RS requires minimal sample preparation or fixation and is nondestructive (unlike scanning electron microscopy, immunohistochemistry [IHC], or quantitative polymerase chain reaction [qPCR]), and requires no labeling, unlike some of these techniques. There is no loss of spatial information (as in MS or qPCR), or sample processing (unlike western blots, qPCR, DNA microarrays), which is time consuming. RS is also cost effective compared with more traditional methods (however, we note that the instrument is costly, but after purchase no additional materials/chemicals are required). A combined approach of HPLC with RS to evaluate clinical effects of the topical should therefore prove beneficial. Validation of this model using other techniques, such as FTIR microscopy, OCT, HFUS, and traditional histology, may also prove useful.

There is however, an ever-increasing role of RS in the clinical field; in terms of the skin, it has been used to assess functional effects of delivery of sunscreens²⁶ for example. RS has the capability of identifying the unique fingerprint of disease pathology, and changes within biomarkers in the skin.²³ Indeed, despite the huge clinical need for treatment for skin scarring, although there are a number of subjective scales for scar assessment³ there is no current objective, and quantitative method for scarring classification and corresponding therapeutics.

There are limitations to this study: first, the reliance of the ability to section longitudinally through each dermal layer and accurately correlate with corresponding H&E sections in HPLC. It is also important to acknowledge that skin explants do not have the same permeability, elasticity, movement, and pharmacokinetics as intact skin, however, our group has previously published work¹⁵ (and more recently other groups)^27,28 demonstrating that it is possible to assess topical penetration on skin explants. The inability to use a true placebo compared with the treatment (since the treatment topical has no evidence to prove known actives) is another limitation, although in this case we were able to demonstrate how compounds could be identified in the absence of known proven activity. Finally, we also acknowledge that the study has been conducted on a limited number of individuals and this to some degree is due to the invasive sampling required to obtain tissue for HPLC and RS.

We have attempted to validate our RS results with histology and clinical correlation, and although no direct comparisons can be made, we have demonstrated trends within the different data sets. Extending the histology time points, and indeed the RS analysis beyond weeks 12 and 16, respectively in order to evaluate wound healing over a longer time period, could provide more information. RS, to our knowledge, has not been able to differentiate between Collagen I and Collagen III, although previous studies have demonstrated a differentiation between types I and IV.¹⁹ Collagen differentiation of Collagen types I and III has however been successfully demonstrated using FTIR.²⁹ Although in both of these studies, pure collagen was used (not from biological tissue).

Finally, although we have evidenced the treatment topical within the scarred skin samples using RS, many of the compounds naturally exist within the skin. The use of SERS,³⁰ may help identify peaks not within the normal skin spectra, which would aid in removing this problem, but this is invasive as metallic nanoparticles need to be added into the skin.

We conclude that using RS, overcomes the difficulties in detecting topical formulation compounds at different depths using HPLC (i.e., the depth of penetration can be quantified to ensure uniformity across all sections); and can be optimized to assess the transdermal delivery of topicals through human skin scars. We have also demonstrated the potential of RS for simultaneous evaluation of the effect of these compounds; with further evaluation of specific bands which may describe scar composition.

Innovation

RS overcomes the difficulty in identifying compounds and positive penetration through scar tissue, validated by HPLC and MS; we have proved this is possible even when the composition of the topical is unknown. There is a role for RS in the clinical setting, where there is a lack of objective, and quantifiable means to assess scarring, and additionally, monitor therapeutic effects by evaluation of structural changes within scar tissue. These findings can be corroborated by traditional techniques such as IHC, however, RS is advantageous as it is nondestructive, does not require fixation or labeling, and easily replicable.

Key Findings

• A combined approach of HPLC and RS can be used to assess transdermal delivery, even when the precise composition of a topical is unknown.

• RS overcomes the difficulty of discriminating between dermal layers using HPLC, although HPLC has a higher signal sensitivity.

• The novel detection of conformational changes using RS provides a role for its use in the clinical setting in the assessment of scarring and monitoring therapeutic effects of antiscar topicals.

• This combined approach can be validated using more traditional techniques; MS to confirm the presence of compounds, IHC to confirm structural changes.

Abbreviations and Acronyms

FTIR

Fourier transform infrared

H&E

Hematoxylin and Eosin

HFUS

high-frequency ultrasound

HPLC

high-performance liquid chromatography

IHC

immunohistochemistry

MEBO

moist exposure burn ointment

MS

mass spectrometry

OCT

optical coherence tomography

qPCR

quantitative polymerase chain reaction

RS

Raman spectroscopy

SERS

surface-enhanced Raman scattering

WiRE

Windows-based Raman environment

Acknowledgments and Funding Sources

The authors specially thank Martin Isabelle, Renishaw Plc. for RS input and for test measurements carried out on RA816 Biological Analyzer at Renishaw Plc.

R.B. thanks Rehanna Sung, Manchester Institute of Biotechnology, for HPLC analytical support. H.M. thanks the University of Liverpool (Liverpool, United Kingdom) for funding and support.

This project was partially funded by Julphar Gulf Pharmaceutical Industries and also by NIHR Manchester Biomedical Research Center.

Author Disclosure and Ghostwriting

R.A.K. is an employee of Julphar Gulf Pharmaceutical Industries. No competing financial interests exist for all other authors with the exception of R.A.K. The content of this article was expressly written by the authors listed. No ghostwriters were used to write this article.

About the Authors

Rubinder Basson, MB BS, is a clinical doctor and surgical trainee in Manchester, currently undertaking a PhD under Dr. Ardeshir Bayat. Cassio Lima, PhD, is a Postdoctoral Researcher in the Department of Biochemistry at the University of Liverpool; he has a specialist interest in FTIR in skin. Howbeer Muhamadali, PhD, is a Research Associate in the Department of Biochemistry at the University of Liverpool, with research in metabolomics using a variety of analytical techniques. Weiping Li, MSc, is a Research Technician at the University of Manchester. Katherine Hollywood, PhD, is an analytical chemist and Senior Experimental Officer for Synbiochem at the University of Manchester. Ludanni Li, MSc, is a Masters student at the University of Manchester. Mohamed Baguneid, MD, is the Chair of Surgery, Al Ain Hospital, UAE. Rawya Al Kredy, BSc, works in R&D at Julphar Gulf Pharmaceutical Industries. Royston Goodacre, PhD, is Professor in Biological Chemistry at the University of Liverpool. He is a leading expert in metabolomics and RS and editor-in-chief of the journal Metabolomics. Ardeshir Bayat, MB BS, PhD, is a Clinician Scientist and Reader at the University of Manchester, an active member of the Wound Healing Society and on the Board of Directors for the Wound Healing Foundation. He is Principal Investigator and corresponding author for this article.