Development of a controlled ex vivo human skin platform for quantitative evaluation of age-related functional biomarkers following application of topical treatments

Abstract

Targeting cellular senescence presents a promising approach to slow visible skin aging and promote tissue repair. However, most preclinical models fail to capture the full architecture of human skin or accommodate diverse skin types, limiting their translational relevance. To address this gap, we developed a controlled ex vivo human skin explant platform using freshly acquired tissues from donors of varying ages and Fitzpatrick skin types. This model applies standardized UVA and UVB doses to induce reproducible photodamage, enabling the assessment of both preventative and reparative effects of topical treatments. The results showed that ND-ZnO and NAC reduced levels of p16^INK4a and p53, which are key biomarkers measuring cellular senescence; ND-ZnO and exosomes lowered IL-1β expression, which is a biomarker measuring inflammation. Histological analysis confirmed these findings, with ND-ZnO-treated skins preserved epidermal structure, reduced inflammatory features, and maintained dermal collagen organization. We then conducted a four-week single-patient case study using the same ND-ZnO formulation. Visible improvements in redness, pigmentation, and texture were observed, aligned with the molecular and histological changes seen ex vivo. These findings suggested that the ex vivo platform has the potential to be used as a more inclusive, human-relavent model for evaluating and quantifying the anti-aging efficacies of topical treatments across diverse skin types and age groups.

Introduction

Skin is the largest organ in the body and is constantly exposed to the outside environment. Understanding how skin ages is important for designing treatments that help maintain skin structure and delay decline in its function. Aging of the skin occurs due to both intrinsic and extrinsic causes. Intrinsic causes include telomere shortening and changes in gene regulation over time. Extrinsic causes include ultraviolet (UV) radiation, pollution, and physical damage to the skin. These different types of stress add up over time and lead to the buildup of senescent cells in the tissue. Senescent cells release inflammatory molecules, break down the extracellular matrix, and interfere with normal repair processes^1,2.

Recent studies have shown that certain molecular markers can be useful in quantifying cellular senescence in skin^3,4. Among these, two of the most common and well-studied are p53 and p16^INK4a. These proteins reflect two different pathways of cellular stress. p53 is activated quickly when there is DNA damage or oxidative stress. It helps the cell decide whether to repair itself or undergo cell death. If p53 stays active for too long, it can lead to early signs of tissue aging^3,4. p16^INK4a builds up slowly over time in response to cell division limits and mitochondria dysfunction. It causes cells to permanently stop dividing^5–7. Both p53 and p16^INK4a are well recognized as key indicators of cellular senescence in many types of tissue.

Despite the importance of studying skin aging and the growing focus on senescence-targeting therapies, current preclinical models still face several key constraints. In vitro systems such as monolayer keratinocyte or fibroblast cultures allow efficient gene editing and high-throughput screening of candidate compounds⁸. These models are convenient and scalable, but they lack many important features of normal human skin⁹. For example, they do not capture the full range of skin cell types or the organization of the dermo-epidermal junction. They also do not model the interaction between cells and the extracellular matrix. Because of these limitations, results from in vitro systems often fail to predict outcomes in actual human skin^10–12. Animal models provide some advantages as they support systemic responses and allow the study of skin aging in a whole-organism context. But animal skin differs from human skin in many ways such as thickness, healing speed, and immune signaling¹³. Even in genetically modified mice that express senescence markers, the molecular and structural changes do not match those seen in human skin¹⁴. These differences reduce the utility of animal models for evaluating skin senescence.

Ex vivo human skin offers a promising alternative. They retain many normal human skin features, including intact epidermal and dermal layers, cell type diversity, and extracellular matrix structure. These explants come from surgical waste or donor skin and can be cultured under controlled laboratory conditions^10,15. One important constraint of this model is its limited culture lifespan. Most explants remain viable for about 5 to 10 days, depending on the media used, the size and thickness of the tissue, and intrinsic differences between donor cells. This short culture window makes it hard to be used to evaluate changes in skin quality over time. In addition, it is not feasible to obtain multiple specimens from the same donors at different time points. To use ex vivo human skin model to study aging in skin, new protocols need to be developed.

To address these limitations, we created a controlled experimental model to recapitulate aging-related stress and enables quantitative analysis of both molecular and cellular responses. In this system, skin explants were exposed to fixed doses of UVA and UVB radiation to induce acute photodamage. After UV exposure, we applied novel topical anti-aging agents either before or after UV irradiation. To assess preventive effects, the topical was applied prior to UV irradiation. To evaluate reparative effects, the topical was applied after UV irradiation. We measured key markers of cellular senescence and inflammation. These included p16^INK4a and p53, which are widely used to examine chronic and acute senescence, and interleukin-1 beta (IL-1β), a cytokine elevated in damaged and aging skin^16,17. In parallel, we performed histologic analysis to assess changes in epidermal structure, dermal collagen organization, and immune cell infiltration. By standardizing UV intensity, treatment schedule, and biomarker quantification, this platform allows consistent evaluation of aging-associated molecular events using intact human tissue.

Using this ex vivo model, we investigated the biological effects of multiple anti-aging topical agents, including nanodiamond zinc oxide (ND-ZnO)¹⁸, N-acetylcysteine (NAC)¹⁹, and exosome-based agents^20,21. These agents represent three widely used classes of topical agents: inorganic nanomaterials, small-molecule antioxidants, and biologically derived extracellular vesicles. The results demonstrated that ND-ZnO and NAC reduced expression of p16^INK4a and p53; ND-ZnO and exosome reduced expression of IL-1β. Histologic analysis is also consistent with these molecular findings: skin samples treated with ND-ZnO showed improved epidermal integrity, more organized collagen fibers, and less inflammation compared to the untreated UV-damaged group. We conducted a four-week single subject human case study and observed visible improvements in skin tone, texture, and erythema following daily ND-ZnO application. These clinical outcomes align with the molecular and histological findings from the ex vivo model.

Results

To simulate and accelerate aging-like changes, we applied a standardized dose of UVA and UVB radiation. This standardized UV doses was selected to reproducibly induce oxidative and genotoxic stress markers while maintaining epidermal and dermal viability. We select a standardized UV doses with elevated UVB: UVA ratio to accelerate the onset of aging-associated biomarkers within the limited viable window of ex vivo skin culture^22,23. Two treatment strategies were used to evaluate both the preventative and reparative effects of the topical agents. In the preventive model, topical agents were applied 15 min before UV exposure. In the reparative model, topical agents were applied immediately after irradiation. After application of topical treatments, skin explants were incubated for 96 h at 37 °C, as shown in Fig. 1a. Following incubation, tissues were lysed and homogenized. Soluble proteins were collected from the supernatant after centrifugation, as shown in Fig. 1b. We then measured the levels of p16^INK4a and p53 using ELISA in technical triplets.

Fig. 1.

Experimental procedure and quantification of p16 and p53 expression in human ex vivo skin following UV exposure and ND-ZnO treatment. (a) Schematic diagram of the ex vivo skin aging model using UVA + UVB irradiation to induce photoaging phenotypes. (b) Following a 96-hour incubation period, skin explants were homogenized, centrifuged, and supernatants were collected for ELISA-based quantification of protein biomarkers. c–f) Both NAC and ND-ZnO were evaluated in preventive (pre-UV) and reparative (post-UV) treatment modes. (c) In the preventive arm, both ND-ZnO and NAC suppressed UV-induced p16 expression relative to untreated UV control, with ND-ZnO exhibiting a greater reduction. Data were normalized to baseline. (d) In the reparative arm, both treatments reduced p16 expression, with ND-ZnO showing significantly greater suppression compared to NAC. Data were normalized to NAC. (e) ND-ZnO and NAC pre-treatment both reduced UV-induced p53 levels. (f) Post-treatment with either compound led to p53 downregulation. All experiments were performed in technical triplicates. Bars represent mean ± SD of three replicates. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by one-way ANOVA with Tukey post hoc test.

In the preventative experiment, the UV stress caused a threefold increase in p16^INK4a expression in skin with no treatment as compared to baseline. When ND-ZnO was applied before UV exposure, this increase was reduced to 2.51-fold, suggesting that ND-ZnO can prevent some of the stress-induced senescence, as shown in Fig. 1c. Similarly, for p53, UV exposure caused a 1.25-fold increase in untreated skin relative to baseline. Pretreating the skin samples with ND-ZnO formulation lowered this to 1.11-fold, as shown in Fig. 1e.

We then tested the reparative effects of multiple skin topical treatments. Explants were first exposed to the standard UV dosage, then treated with either ND-ZnO or NAC. For NAC post-treatment, the p16 expression in the untreated skin sample was lowered by 7%. For ND-ZnO, this reduction was as dramatic as 36%, more than five-fold of the change that NAC induced, as shown in Figuer 1d. Similarly, while NAC lowered the p53 expression in untreated skin by 9%, ND-ZnO was able to lower the expression by 28%, approximately three times as effective as NAC, as shown in Fig. 1f.

To evaluate the reproducibility and donor-independence of our ex vivo skin platform, we extended the analysis to interleukin-1 beta (IL-1β). IL-1β is a UV-responsive proinflammatory cytokine that reflects upstream stress signaling and disruption of the skin barrier. We tested two different topical agents, including a ND-ZnO formulation and an exosome-containing formulation. We applied the same treatment protocols to skin explants from two donors: a 46-year-old female with Fitzpatrick skin type V and a 60-year-old female with Fitzpatrick skin type II. The skin samples were collected on different days.

In the preventative model, IL-1β expression surged to 124% in donor 1 compared to baseline. The ND-ZnO formulation was able to downregulate the IL-1β expression to 65% of the baseline after treatment, as shown in Fig. 2a. The trend was consistent in donor 2, IL-1β expression surged to 149% of baseline without treatment, and ND-ZnO downregulated the IL-1β expression to only 42% of the baseline. Figure 2b. This means that, in both donors, ND-ZnO can prevent the UV stress, and at the same time downregulate the IL-1β expression to lower than the existing level in the baseline, as shown in Fig. 2b.

Fig. 2.

IL-1β expression in ex vivo human skin following ND-ZnO treatment in multiple donors. (a–d) IL-1β levels were quantified by ELISA to assess inflammatory response following UVA + UVB exposure and topical treatment. Both ND-ZnO and exosome formulations were evaluated under pre-treatment and post-treatment conditions across two donor samples. (a, c) Results from Donor 1 (46-year-old female, Fitzpatrick skin type IV–V) showing IL-1β suppression by both treatments, with ND-ZnO exhibiting stronger suppression compared to the exosome formulation. (b, d) Results from Donor 2 (60-year-old female, Fitzpatrick skin type I–II) confirming similar trends, with both treatments lowering IL-1β expression relative to untreated UV control, and ND-ZnO showing a more pronounced effect. Data were normalized to the exosome-treated group for cross-comparison. All experiments were conducted in technical triplicates. Bars represent mean ± SD of triplicate samples. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by one-way ANOVA with Tukey post hoc test.

In the reparative model, the results were normalized to exosome-treated level. The exosome-containing regime lowers the IL-1β expression by 5% as compared to the untreated skin, while the ND-ZnO formulation is able to suppress the IL-1β expression by 30% as compared to the untreated, which is 5-fold more effective than the exosome formulation as shown in Fig. 2c. Similarly, while the exosome-containing regime downregulated IL-1β expression by 27% in donor 2, ND-ZnO downregulated IL-1β expression by 75% as compared to the untreated sample, which is almost three-fold of the exosome-containing sample, as shown in Fig. 2d. The results obtained were consistent using explants derived from the two donors.

Unlike in vitro cell models, ex vivo human skin explants retain intact tissue structure. This includes the epidermis, dermis, and adnexal units, which allow evaluation of treatment effects in a more physiologically relevant context. One key advantage of our model is the ability to compare responses across different Fitzpatrick skin types. These comparisons help capture variations in pigmentation, inflammation, and matrix remodeling responses in different skin types that are not represented in simplified monolayer cultures.

We used this model to study histological changes following ND-ZnO treatment. The preventative and reparative experimental paradigms were applied, similar as describe above. Explants were irradiated with UVA and UVB to induce visible tissue injury. We analyzed four treatment conditions: untreated baseline, UV-exposed untreated control, ND-ZnO pre-treatment, and ND-ZnO post-treatment. Following treatment, samples were formalin-fixed, paraffin-embedded, sectioned, and stained with hematoxylin and eosin (H&E). A board-certified dermatopathologist, blinded to treatment groups, evaluated five histological features: epidermal integrity, collagen remodeling, adnexal preservation, inflammation, and re-epithelialization.

The first cohort of samples was from a 31-year-old female donor with Fitzpatrick skin type II (Donor lot 2). In untreated baseline samples, the skin showed normal histology with well-stratified epidermis, intact rete ridges, organized collagen, and no inflammation, as shown in Fig. 3a. After UV exposure, the skin exhibited photodamage. This included epidermal thinning, dermo-epidermal separation, loss of adnexal structures, and perivascular leukocyte infiltration, as shown in Fig. 3b. When ND-ZnO was applied before UV exposure, the epidermal structure and rete ridges were preserved. Dermal collagen remained intact, and minimal cytoplasmic vacuolization was observed, as shown in Fig. 3c. In the post-treatment group, ND-ZnO partially restored epidermal stratification and collagen alignment. Inflammatory infiltrate was reduced compared to untreated UV-damaged samples, as shown in Fig. 3d. These findings were reflected in the histological scores listed in Table 1.

Fig. 3.

Histologic assessment and photograph of human ex vivo skin following aging stress and ND-ZnO treatment. (a-d) Fitzpatrick skin type II donor. (a) Left: Baseline untreated skin exhibits intact epidermis, normal rete ridge architecture, and well-organized collagen in the dermis. Right: photograph of explant showing even tone and fine skin texture. (b) Left: UV-exposed, untreated skin displays pronounced photodamage including epidermal thinning, epidermal-dermal junction separation, keratinocyte dropout, and disrupted collagen. Right: surface image reveals wrinkling and hyperpigmentation. (c) Left: ND-ZnO pre-treatment protected the explants with no interface separation, except mild basal vacuolization and mild inflammatory cell infiltration. Right: skin surface remains even-toned with minimal disruption. (d) Left: ND-ZnO post-treatment shows partial repair of UV-induced damage with no dermal-epidermal separation and reduced structural disruption. Right: surface appearance is smoother with decreased pigment mottling. (e–h) Fitzpatrick skin type IV-V donor. (e) Left: Baseline explant demonstrates stratified epidermis and intact dermis. Right: skin surface shows homogeneous pigmentation with no visible signs of damage. (f) Left: UV-damaged untreated sample presents with full thickness epidermal destruction, epidermal-dermal split, and collagen disarray. Right: increased wrinkling and darkened surface features are evident. (g) Left: ND-ZnO pre-treatment maintains epidermal morphology and dermal collagen integrity, with minimal inflammation. Right: reduced surface coarseness and pigmentation changes compared to untreated. (h) Left: ND-ZnO post-treatment yields near-complete structural preservation, with reappearance of elongated rete ridges and resolution of skin inflammation. Right: skin surface shows smoother texture and reduced UV-induced pigmentation.

Table 1.

Assessment of skin injury across five parameters donor lot 2 and donor lot 4.

Sample	Treatment	Epidermal integrity loss	Collagen remodeling	Adnexal changes	Inflammation	Re-epithelialization
a	Baseline	1	0	1	0	0
e		1	0	1	0	0
b	UV Negative	4	0	3	2	0
f		3	1	3	2	0
c	ND-ZnO Pre-Treatment	1	0	1	0	0
g		1	0	1	0	0
d	ND-ZnO Post-Treatment	2	0	1	0	0
h		2	0	1	1	0

0 = no significant change, 1 = minimal, 2 = mild, 3 = moderate, 4 = severe.

The second cohort of samples was from a 46-year-old female donor with Fitzpatrick skin type V (Donor lot 4). In this cohort, untreated baseline samples showed a preserved epidermal layer and dense dermal collagen, as shown in Fig. 3e. After UV exposure, histological damage was observed. This included disruption of epidermal architecture and fragmentation of dermal collagen fibers, as shown in Fig. 3f. When ND-ZnO was applied before UV exposure, tissue morphology was better maintained. Rete ridges were visible, and there was less cytoplasmic and structural damage, as shown in Fig. 3g. ND-ZnO post-treatment samples showed partial recovery of tissue architecture and reduced inflammation when compared to UV-exposed controls, as shown in Fig. 3h. Histological injury scores for this donor are also reported in Table 1.

Evaluating treatment responses across different skin types is important for improving the translational relevance of dermatologic product development. In real-world settings, patients have varying levels of skin pigmentation, barrier properties, and inflammatory sensitivity. These differences can affect how skin responds to external stress and to topical agents. For example, darker skin types often have stronger barrier function but may also show different patterns of post-inflammatory hyperpigmentation. Lighter skin types may be more susceptible to UV-induced damage and erythema. In this study, we used ex vivo skin explants from donors representing both low and high Fitzpatrick skin types. This allowed us to assess how ND-ZnO performs in different physiologic contexts. The explant model preserves key features of native skin, including cellular composition and matrix structure, which makes it suitable for testing interventions across varied skin types. The consistent results observed across donors with different pigmentation and age support the model’s robustness. By expanding experimentation across the Fitzpatrick scale, this approach helps reveal population-level differences in response to oxidative stress and anti-aging agents. It also provides early insight into the dosing and safety profiles needed for broad application. These findings can potentially contribute to the design of more inclusive and responsive dermatologic treatments.

To evaluate whether the molecular and histological improvements observed in our ex vivo model can be translated to visible clinical improvement in human subjects, we carried out a 4-week pilot case study with a male participant with Fitzpatrick Skin type II using a topical formulation containing ND-ZnO. This participant applied the product daily over the course of four weeks. After treatment, the subject showed clear improvements in facial skin appearance. These included a reduction in erythema, less diffuse redness, and more even pigmentation. The changes were especially noticeable in the malar and perioral areas, as shown in Fig. 4.

Fig. 4.

Clinical improvement following 4-week topical ND-ZnO treatment. Representative images of a male subject at baseline and after 4 weeks of once daily application of an ND-ZnO-containing formulation. Improvements in facial erythema and skin tone uniformity were observed, particularly in the malar and perioral regions.

Ex vivo skin models provide a useful platform for studying UV-induced acute skin injury. However, not all senescence-associated biomarkers perform equally well in this system. To evaluate biomarker performance and determine selection criteria, we tested two proteins linked to cellular aging: β-galactosidase (β-gal) and tissue factor pathway inhibitor-2 (TFPI-2)^24,25. These proteins were chosen based on three factors. First, they are biologically relevant to aging. Second, they respond to UV-induced damage. Third, they can be measured using ELISA at the protein level. Although both markers have been linked to aging in human tissue, their compatibility with ex vivo systems had not been tested systematically.

β-galactosidase is often used as a marker of senescence, especially in cultured cells. It reflects lysosomal activity that increases when cells enter growth arrest. In most experiments, SA-β-gal is measured by histochemical staining under acidic conditions. This approach works well in monolayer culture. However, β-gal is not unique to senescent cells. It can also be produced by metabolically active, non-senescent cells, which increases background signal in biochemical assays. In our ex vivo system, we found that β-gal expression changed only slightly after UV exposure. When measured by ELISA, the signal showed a narrow dynamic range and poor contrast between control and stressed samples, as shown in Fig. 5a. These results suggest that β-gal may not be well-suited for ex vivo skin models.

Fig. 5.

UV-induced changes of beta-gal and TFPI-2 in ex vivo human skin models. (a) Human β-galactosidase (β-gal) levels were modestly reduced in UV-exposed untreated skin relative to baseline, indicating impaired lysosomal senescence-associated enzyme expression. (b) TFPI-2 expression increased nearly threefold in UV-exposed untreated skin compared to baseline, consistent with stress-induced matrix remodeling. Data normalized to baseline. Bars represent mean ± SD in the technical duplicates; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by one-way ANOVA with Tukey post hoc test.

TFPI-2 is a secreted protease inhibitor that regulates extracellular matrix remodeling. It is mainly expressed in the dermis and is involved in fibroblast senescence and tissue structure maintenance. Previous studies have shown that TFPI-2 is upregulated after UV exposure and in response to inflammatory signals. In our system, TFPI-2 levels increased nearly threefold in UV-treated explants compared to baseline, as shown in Fig. 5b. The strong and consistent change in expression indicates that TFPI-2 responds well to photodamage. Its dermal origin and protein-level detectability make it a promising biomarker for tracking structural changes associated with aging in skin explants.

Discussion

In this study, we utilized a controlled ex vivo human skin explant platform to quantify molecular and structural changes related to senescence after UV-induced damage. The model uses calibrated doses of UVA and UVB to create consistent stress conditions. Topical treatments were applied before or after UV exposure to measure preventative or reparative effects. We measured IL-1β levels using ELISA on donor skin tissues from different age groups, Fitzpatrick skin types, and harvest times. Results were reproducible across two independent donor lots. We then validated these findings through double-blinded histological analysis by a board-certified pathologist. Reduced IL-1β levels corresponded with less inflammation and better tissue preservation. These results show that the platform can detect biomarker changes consistently across diverse donor backgrounds. With both molecular and histological readouts, this model offers a practical and inclusive tool to evaluate topical anti-aging treatments on skin from varied ages and skin types.

Using these protocols, we tested multiple topical formulations including nanodiamond zinc oxide (ND-ZnO), N-acetylcysteine (NAC), and exosome-based formulations. ND-ZnO containing formulation outperformed NAC and exosome-based formulations in skin aging-related biomarkers analyzed. Histologic analysis from the same ex vivo human extracts confirmed the protective effects of ND-ZnO. We further use histologic analysis from the same ex vivo human extracts to validate the ELISA results, and the results further proved the protective and reparative effects of ND-ZnO.

The ND-ZnO has a core-shell structure with porous zinc oxide spheroids cluster uniformly around central nanodiamond (ND) particles. Prior studies have shown that exposure of this ultramicronized nanodiamond zinc oxide preparation to solar-simulated radiation generated approximately 80% less ROS compared to irradiation of conventional zinc oxide preparations²⁶. Moreover, in previous in vitro models, ND-ZnO increased collagen synthesis in human dermal fibroblasts following UV irradiation¹⁸. These two mechanisms of action: reduction of ROS and increased collagen synthesis, may explain the superior performance of ND-ZnO in our ex vivo study. While we did not directly measure ROS or DNA damage endpoints in this study, the reduction in senescence and inflammation markers, along with more preserved epidermal and dermal architecture, supports the hypothesis that ND-ZnO achieves these effects by reducing ROS and preserving collagen structure and, by extension, the extracellular matrix.

To assess the translational relevance of this platform, we conducted a follow-up four-week clinical pilot study evaluating the same ND-ZnO formulation in a single patient. The observed visible improvement in pigmentation and texture mirrored the molecular and histologic improvements documented in the ex vivo assays. Although the current clinical case study remains exploratory and limited in scale, the results aligned with our ex vivo model, demonstrating the potential of using this platform to predict human skin responses.

The ex vivo model has its constraints. First, the viable window for skin explants is relatively short (typically under 7–10 days), which limits the study of long-term remodeling or chronic treatment regimens. Our study empirically maximized the UV dosage to induce a sufficient UV burden without destruction of the skin structure during the incubation period, but it is difficult to translate such UV stress directly to chronic aging burden.

Second, not all biomarkers are suitable for ex vivo analysis. We tested β-galactosidase and TFPI-2, and found that effective biomarkers in this system must both respond to UV-induced damage and be detectable at the protein level using ELISA. Moreover, while TFPI-2 showed a strong response to UV, its baseline levels vary significantly with donor age, making interpretation more complex²⁵. Our current donor cohort did not include individuals under 30, limiting our ability to evaluate age-related changes. In future work, we aim to include younger donors to better assess the therapeutic effects of anti-aging agents on TFPI-2 expression.

Third, biological replication is often limited by tissue availability and cost, which restricts the number of donors per treatment condition. As this study remains exploratory at this point, some assays were performed on single-donor lots, and we used technical replicates to ensure assay accuracy. While such effort may be helpful in identifying suitable biomarkers for ex vivo models and comparing treatment efficacy between donors with different ages and Fitzpatrick skin types, the small sample number is not sufficient to generate robust statistical conclusions. Hence, in future studies, multiple biological replicates should be allocated to each assay. Future studies may benefit from larger sample sizes and integrating histochemical detection methods or single-cell profiling to improve spatial and cell-type-specific resolution.

Materials and methods

Ex vivo human skin culture

Full-thickness human skin (Genoskin, MA) was obtained one day post-abdominoplasty from healthy adult female donors with informed consent, in accordance with the Declaration of Helsinki and all applicable regulations. The skin tissue was obtained from two donor lots of de-identified, post-surgical samples provided with informed consent for research use.

Skin was cleaned of subcutaneous fat and residual blood before 8-mm biopsy punches were prepared. All biopsies were cultured in a 12-well plate at an air-liquid interface with fresh culture medium provided by the manufacturer and maintained at 37 °C in a tissue incubator with 5% CO₂ and 95% humidity. Explants were equilibrated for 12 h prior to experimental procedures.

Topical formulations and application doses

All topical agents were applied to the surface of each skin punch at a uniform dose of 10 mg/cm² unless otherwise specified. This formulation and concentration were used consistently across all experiments in this study.

ND-ZnO formulation: The ND-ZnO topical formulation consisted of 15% (w/w) nanodiamond zinc oxide dispersed in an inert cream base. The vehicle contained water, glycerin, dimethicone, and emulsifying agents to ensure stable dispersion and uniform topical delivery. At the applied dose of 10 mg/cm², the effective ND-ZnO exposure was 1.5 mg/cm².

N-acetylcysteine (NAC) formulation: N-acetylcysteine was prepared as a 1% (w/v) aqueous solution by dissolving NAC at 10 mg/mL in phosphate-buffered saline. The solution was applied topically at a dose of 10 mg/cm², corresponding to a NAC exposure of 0.1 mg/cm².

Exosome formulation: The exosome-containing topical preparation was applied at 10 mg/cm². Each applied dose contained approximately 5 × 10⁷ human platelet-derived exosomes. The formulation vehicle consisted of deionized water, glycerin, and polyacrylate crosspolymer-6 to support viscosity and topical spreadability.

UV irradiation of ex vivo skin explants

Ex vivo skin explants were subjected to a standardized cumulative aging stimulus using a custom-fabricated UV irradiation unit housed within a biological safety cabinet to maintain sterility. The irradiation source consisted of an array of four UVA and two UVB emitters. Prior to experimentation, the irradiance was measured using a calibrated UV meter. Measurements taken at the sample working distance of 5 cm from the source confirmed a radiometric irradiance of approximately 2.78 mW/cm² for UVA and 0.625 mW/cm² for UVB.

For the irradiation procedure, tissue culture plates containing the skin explants were placed 5 cm from the emitters. The samples were exposed for a duration of 1 h, and this exposure protocol was repeated on the following day to generate a total cumulative dose of 20 J/cm² UVA and 4.5 J/cm² UVB over two consecutive days. The “baseline” control samples were maintained under identical conditions in the biosafety cabinet for the same duration but in the absence of UV exposure.

Aging model and treatment paradigms

Donor Lot 1 (37-year-old female donor; Fitzpatrick skin type II): this experimental cohort was used to assess molecular senescence markers (p53 and p16^INK4a) via ELISA. Two treatment paradigms were employed:

Aging-Preventive Model: ND-ZnO formulation is applied topically to each skin punch 15 min prior to UV exposure. Explants were then exposed to UVA/UVB irradiation as described above. Between exposures, samples were maintained at 37 °C in fresh culture medium. Following irradiation, samples were incubated for 24 h at 37 °C in fresh culture medium.

Aging-Reparative Model: Explants were first irradiated with the same two-dose UV protocol. Immediately following irradiation, either ND-ZnO formulation or N-acetylcysteine was applied topically. Samples were then incubated for 24 h under standard culture conditions before tissue was snap-frozen for protein extraction and subsequent ELISA analysis.

Donor Lot 2 (31-year-old female donor; Fitzpatrick skin type II): this experimental cohort was used to evaluate structural changes via hematoxylin and eosin (H&E) staining and histopathological scoring.

Aging-Preventive Model: ND-ZnO or an exosome-containing topical skin preparation was applied 15 min before irradiation. Explants were then exposed to UVA/UVB irradiation as described above. Between exposures, samples were maintained at 37 °C in fresh culture medium. Explants were then incubated for 96 h at 37 °C, fixed in 10% formalin, and processed for paraffin embedding, sectioning, and H&E staining for histological analysis.

Aging-Reparative Model: Explants were first irradiated with the same UV protocol. Following the second exposure, samples were immediately treated with either ND-ZnO or the exosome-containing skin preparation. Explants were then incubated for 96 h at 37 °C, fixed in 10% formalin, and processed for paraffin embedding, sectioning, and H&E staining for histological analysis.

Donor Lot 3 (38-year-old female donor; Fitzpatrick skin type III): this experimental cohort was used to evaluate the change of biomarkers tissue factor pathway inhibitor-2 (TFPI2) and β-galactosidase.

Explants were first irradiated with the same UV protocol. Samples were then incubated for 96 h under standard culture conditions. Following treatment, tissue was harvested, snap-frozen, and processed for ELISA quantification of TFPI2 expression and β-galactosidase expression.

Donor Lot 4 (46-year-old female donor; Fitzpatrick skin type V) and Donor Lot 5 (60-year-old female with Fitzpatrick skin type II): these two cohorts were used to assess both inflammatory signaling and tissue integrity following defined aging stress and topical interventions via hematoxylin and eosin (H&E) staining and histopathological scoring.

Two treatment paradigms were implemented:

Preventive Model: To evaluate early inflammatory responses and tissue preservation, ND-ZnO or exosome-containing skin preparation was applied 15 min prior to UV irradiation. Explants were then exposed to a UVA/UVB as described above. Between exposures, samples were maintained at 37 °C in fresh culture medium. Explants were then incubated for 96 h at 37 °C. After incubation, tissue samples were harvested and either snap-frozen for IL-1β quantification via enzyme-linked immunosorbent assay (ELISA) or fixed in 10% neutral buffered formalin (NBF) for histopathological analysis.

Inflammatory Marker and Histology Evaluation. Reparative Model: Explants were first irradiated with the same two-dose UV protocol, then treated with ND-ZnO and exosome formulation. Tissues were incubated for 96 h at 37 °C prior to harvest. Parallel samples were processed for IL-1β ELISA and fixed in NBF for paraffin embedding, sectioning, and hematoxylin and eosin (H&E) staining. Histological scoring was performed to assess epidermal and dermal response under blinded conditions.

Protein extraction and quantification

Following incubation, skin samples from doner lot 1 were snap-frozen in liquid nitrogen and stored at -80 °C. For protein extraction, tissues were homogenized on ice using a handheld tissue homogenizer (Omni THQ, Omni International) in ice-cold 1× Cell Extraction Buffer (Abcam, ab270057) supplemented with protease inhibitors. Lysates were incubated overnight at 4 °C on a rotator, followed by centrifugation at 14,000 rpm for 20 min at 4 °C. The clarified supernatant was collected and stored at -80 °C until analysis.

Total protein concentration was determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) per the manufacturer’s protocol. Absorbance at 562 nm was measured using a Varioskan LUX multimode plate reader (Thermo Fisher Scientific), and protein concentrations were interpolated from a BSA standard curve.

Quantification of senescence markers

Expression levels of p53, p16 ^INK4a, TFPI2, β-galactosidase, and IL-1β were quantified in tissue lysates using the Human p53 ELISA Kit (Abcam, ab171571), Human p16 ELISA Kit (Abcam, ab229402), Human GLB1/Beta-galactosidase ELISA Kit (Abcam, ab309424), Human TFPI2 ELISA Kit (Abcam, ab213836), and Human IL-1 beta ELISA Kit (Abcam, ab214025) respectively. All assays were performed according to the manufacturers’ protocols. Samples were loaded in duplicate, with 500 ng of normalized total protein per well. Plates were developed and read at 450 nm using the Varioskan LUX plate reader. Concentrations were fitted to the kit standards.

Histological analysis on H&E staining slides

Following the incubation, skin samples were fixed in 10% neutral-buffered formalin (NBF) for 24 h. Fixed samples were then stored in 10% NBF for 24 h. Each explant was paraffin-embedded, sectioned, and stained with hematoxylin and eosin (H&E). Two sections per sample were generated for downstream analysis.

Digital whole-slide images (WSI) were acquired using high-resolution scanning and reviewed using SlideViewer software. Four experimental groups were evaluated for each donor: untreated baseline, UV-only negative control, ND-ZnO pre-treatment, and ND-ZnO post-treatment.

Histological evaluation was conducted by a board-certified dermatopathologist who was blinded to all experimental parameters, including group identity, treatment assignment, donor background, and associated clinical or procedural metadata. Blinding was rigorously maintained throughout scoring and narrative assessment.

Tissue sections were evaluated according to criteria outlined by the International Harmonization of Nomenclature and Diagnostic Criteria (INHAND)^27,28. A semi-quantitative grading scale (0 to 4) was used to assess the following parameters: epidermal integrity, dermal collagen remodeling or fragmentation, adnexal structure preservation or loss, inflammatory infiltration, and evidence of re-epithelialization. Scores were defined as follows: 0 = no significant change, 1 = minimal, 2 = mild, 3 = moderate, 4 = severe.

Statistical analysis

All statistical analyses were performed using GraphPad Prism (version 8.4.0). Data are presented as mean ± standard deviation (SD). Normality and homogeneity of variance were assessed prior to hypothesis testing. Comparisons between two groups were performed using the unpaired Student’s t-test. For comparisons involving more than two groups, one-way or two-way ANOVA was used, followed by post hoc testing as appropriate. A p-value < 0.05 was considered statistically significant.

Single-subject topical application study

A prospective, single-subject case study was conducted to evaluate the clinical effects of a topical nanodispersed zinc oxide (ND-ZnO) formulation on facial skin. Subject selection was performed by a board-certified dermatologist based on the presence of mild facial erythema, uneven skin tone, and early clinical signs of photodamage and aging, including telangiectasia and rough texture in the malar and perioral regions.

One healthy adult male participant was enrolled after providing written informed consent for participation and publication of clinical photographs. The study was conducted by Validated Claim Support. The participant applied the ND-ZnO formulation once daily to the entire face for 28 consecutive days without concurrent use of other skincare products. Standardized facial photographs were taken at baseline (Day 0) and post-treatment (Day 28) using cross-polarized lighting under fixed imaging conditions. Clinical assessment was performed by independent, blinded reviewers.

Ethics statement

The clinical portion of this study (Clinical Study Number: CS241011) was conducted by Validated Claim Support (VCS) under their Quality Management System. The study involved the topical application of a cosmetic serum to intact, non-lesional facial skin in adult participants. No drug or device claims were made, and no medical procedures or sensitive personal data were involved. The study excluded at-risk populations and used ingredients with an established safety profile. Based on a formal internal review, VCS determined that the study posed minimal risk and did not require Institutional Review Board (IRB) submission under applicable U.S. federal regulations (45 CFR 46). All participants provided written informed consent prior to enrollment, including consent for photography and publication of de-identified images

Human skin explants used in the ex vivo assays were obtained from Genoskin, Inc. (Cambridge, MA, USA) through their IRB-approved Human Tissue Donation Program (IRB Study Number 21-08-536; IRB Committee A-IRB0000080). All tissue specimens were remnant surgical materials collected under written informed consent and fully de-identified in compliance with HIPAA and GDPR prior to transfer. In accordance with U.S. federal regulations, the use of these tissues qualifies as exempt human specimen research and does not require additional IRB approval by the receiving institution. Procurement and use of all tissues followed the Declaration of Helsinki and all applicable U.S. and international ethical guidelines. No personally identifiable information is included in this manuscript.

Author contributions

X.D.Z., X.B., and C.T. conceived and designed the study. X.D.Z., N.A., and E.R. performed the experiments. X.D.Z. analyzed the data, and J.T. contributed to histological data interpretation. All figures were prepared by X.D.Z. The manuscript was drafted by X.D.Z., N.A., E.R., C.T., and X.B., and reviewed and edited by X.B., C.T., J.T., and X.D.Z.

Funding

This work was supported by internal research funding from B.A.I. Biosciences.

Data availability

All data supporting the findings of this study are available within the article. Additional raw data and analysis scripts are available from the corresponding author upon reasonable request.

Declarations

Competing interests

X.D.Z, C.T., and X.B. are employees of B.A.I. Biosciences. X.D.Z. and X.B. are named inventors on a patent application related to the nanodiamond zinc oxide formulation described in this study, which is solely owned by B.A.I. Biosciences. The remaining authors declare no competing interests.