Horizontal mitochondrial transfer and mitochondrial transplantation in skin: protection against UVR-induced ROS damage and enhancement of cell proliferation and wound healing

Abstract

Background

Skin homeostasis, protection against ultraviolet radiation (UVR), and wound repair depend on coordinated interactions among melanocytes, keratinocytes, and fibroblasts. Horizontal mitochondrial transfer (HMT) is a naturally occurring form of intercellular communication in which mitochondria move between cells and may contribute to stress adaptation, cellular recovery, and tissue resilience. Artificial mitochondrial transfer (AMT), performed in vitro or ex vivo, and mitochondrial transplantation (MT), involving the direct administration of isolated mitochondria in vivo, seek to therapeutically harness these biological mechanisms. However, HMT among resident skin cells and its relationship to mitochondria-based regenerative strategies remain poorly understood.

Methods

HMT among human melanocytes, keratinocytes, and fibroblasts was evaluated under basal conditions and following UVR exposure. Direct 2D coculture and transwell systems were used to assess predominantly contact-dependent and contact-independent HMT, respectively, using fluorescence microscopy. AMT was performed by delivering isolated mitochondria from human fibroblast, human Wharton’s jellymesenchymal stem/stromal cells (WJ-MSCs), or mouse bone marrow MSCs (BM-MSCs) to recipient fibroblasts, followed by assessment of mitochondrial uptake, reactive oxygen species (ROS) production, and cell proliferation. The regenerative effects of locally administered MSC-derived mitochondria were subsequently evaluated in murine and porcine primary-intention wound models: mouse BM-MSC-derived mitochondria were used in murine wounds, whereas human WJ-MSC-derived mitochondria were used in porcine wounds. Outcomes were assessed using histological analysis, the wound healing index (WHI), and, in pigs, spatial quantification of Ki67-positive cells.

Results

HMT from melanocytes to keratinocytes increased significantly after UVR exposure and occurred predominantly under direct coculture conditions, reaching approximately 39%, compared with less than 9% in transwell assays. HMT in the other donor–recipientcombinationsremainedbelow4%.AMTusing human WJ-MSC-derived mitochondria reduced UVR-induced ROS production, while mitochondria derived from both human WJ-MSCs and mouse BM-MSCs significantly increased fibroblast proliferation, although the magnitude of these effects depended on the mitochondrial dose. In murine wounds, locally administered BM-MSC-derived mitochondria enhanced early histological repair and produced effects comparable to those observed after intact BM-MSC administration. In porcine wounds, WJ-MSC-derived mitochondria increased the WHI, improved collagen-containing tissue organization, and enhanced Ki67 positivity within epidermal and dermal regions directly involved in wound repair.

Conclusions

These findings identify HMT as a cell-type-specific response in the skin, with preferential HMT from melanocytes to keratinocytes that is enhanced by UVR exposure. They also demonstrate that MSC-derived mitochondria can reduce oxidative stress, stimulate fibroblast proliferation, and promote early cutaneous repair after local administration. Together, the results establish a translational link between endogenous mitochondrial exchange and the therapeutic use of AMT and MT, supporting further development of mitochondria-based, cell-free strategies for skin injury and impaired wound healing.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12967-026-08801-y.

Highlights

Horizontal mitochondrial transfer (HMT) occurs preferentially from melanocytes to keratinocytes and is mediated mainly by direct cell–cell contact.

Ultraviolet radiation (UVR) significantly enhances melanocyte-to-keratinocyte HMT.

Fibroblasts exhibit limited spontaneous mitochondrial uptake, which may reduce their capacity to adapt to UVR-induced stress.

Artificial mitochondrial transfer (AMT) with MSC-derived mitochondria reduces UVR-induced ROS production and promotes fibroblast proliferation.

Local mitochondrial transplantation enhances early cutaneous repair in murine and porcine wound models.

AMT and mitochondrial transplantation support skin repair by improving tissue repair.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12967-026-08801-y.

Introduction

Cells can transfer mitochondria to one another through horizontal mitochondrial transfer (HMT), also referred to as intercellular mitochondrial transfer, a process in which mitochondria move from one cell to another independently of cell division [1–3]. HMT can serve diverse biological functions depending on the cell types involved, including supporting the survival of damaged cells and contributing to tissue homeostasis [4–8]. However, whether this process occurs in the skin, particularly among resident skin cells, remains unclear. It is also unknown whether skin cells can internalize mitochondria in ways that influence protective responses against ultraviolet radiation (UVR) or modulate dynamic cellular functions such as proliferation and collagen production, as has been observed in other cell types [2, 5, 8–11]. Investigating the occurrence of HMT among skin cells and demonstrating their capacity to take up exogenous mitochondria could open new avenues for therapeutic development.

In this study, we define HMT as the naturally occurring transfer of mitochondria between skin cells. In addition to investigating whether HMT occurs under physiological conditions, we explored the therapeutic potential of artificial mitochondrial transfer (AMT) and mitochondrial transplantation (MT) as strategies to counteract cellular damage and enhance wound healing. AMT involves the use of in vitro or ex vivo approaches to transfer isolated mitochondria into recipient cells [11–14], whereas MT refers to the direct administration of autologous, allogeneic, or xenogeneic mitochondria into tissues in vivo under experimental or therapeutic conditions. These approaches hold significant promise for mitigating skin damage, supporting tissue repair after injury, and treating inherited mitochondrial diseases [12, 13].

Studying the HMT process in different tissues and between cells is essential for understanding how organs and tissue homeostasis is maintained [5]. However, observing HMT between skin cells is challenging because of the diversity of cell populations within the skin and the complex interactions required to preserve tissue function and support repair after environmental damage or aging [14]. To better understand this process in the skin, we focused on melanocytes, keratinocytes, and fibroblasts, as these cell types interact closely through direct and indirect mechanisms, including the secretion of factors that support skin structure and protect against harmful stressors. Melanocytes play a key role in producing melanin within melanosomes, which are transferred to keratinocytes to protect the skin against ultraviolet radiation (UVR) [15]. Keratinocytes constitute approximately 90% of the epidermal cell population and contribute to tissue architecture, barrier protection, and extracellular vesicle (EV)-mediated crosstalk with melanocytes [15, 16]. These skin cells rely heavily on mitochondria to generate the energy required for essential functions, including melanin production, cellular proliferation, and differentiation [17, 18]. Investigating whether HMT occurs among these cells is therefore important, as it could provide new insights into how skin cells maintain tissue homeostasis under stress and support the development of mitochondria-based therapeutic strategies for skin repair and regeneration [19].

In this study, we aimed to analyze HMT between melanocytes (HEMn-HP), keratinocytes (HaCaT), and fibroblasts, which are cell types that play a crucial role in protecting the skin against UVR and promoting wound repair. We used fluorescent cell dyes to identify the cells that transfer mitochondria from donor to recipient cells in a protocol that allowed us to prevent nonspecific transfer and labeling. We investigated the mechanisms of HMT using a co-culture system to allow either direct 2D co-culture for primarily contact-dependent mitochondrial transfer (CDMT) or contact-independent mitochondrial transfer (CIMT) via transwells after 22 h of culture. Skin cells were exposed to UVR (5.4 mJ/cm² UVA/UVB) at the 18-hour mark, followed by an additional four hours to complete the 22-hour period. The results were compared with non-exposed conditions to assess changes in HMT in both 2D co-cultures and transwell assays. As we identified cells with a higher or lower capacity to receive mitochondria, we aimed to test if AMT could be a strategy to prevent UVR damage by reactive oxygen species (ROS) and to stimulate cell proliferation in those that are not receiving mitochondria. Since mesenchymal stem/stromal cells (MSCs) HMT has shown regenerative effects [20–25], we tested whether mitochondria isolated from MSCs could be used for AMT/T to decrease UVR-produced ROS and to heal skin surgical wounds in mice and swine as a therapeutic strategy that may help improve the regeneration time of wounded skin. This study provides evidence of HMT between skin cells and establishes a foundation for understanding its significance and the potential benefits of using AMT/T to repair UVR damage and enhance wound healing processes.

Results

Dynamics of HMT between skin cells

We investigated HMT among melanocytes (HEMn-HP), keratinocytes (HaCaT) and dermal fibroblasts, and whether it is modulated by ultraviolet radiation (UVR; 5.4 mJ cm⁻², UVA/UVB). Donor cells were MitoTracker Red-labelled; recipients, CellTracker Green. After 18 h of co-culture, cells received a 3-min UVR pulse and were incubated for 4 h more (22 h total). HMT and cell–cell interactions were then quantified.

To ensure our UVR dose did not compromise recipient-cell membranes, we treated fibroblasts, keratinocytes and melanocytes with either 5.4 mJ cm⁻² UVA/UVB or a positive-damage control (45 µL 3% H₂O₂ in 0.5 mL, 15 min). Trypan Blue exclusion showed no staining after UVR, whereas H₂O₂ caused dye uptake and detachment. These results show that the UVR protocol does not harm integrity and triggering of nonspecific mitochondrial transfer (Supplementary Figure 1a, 1b, 1c)

Addressing HMT between the same cell types revealed that only a very small number of fibroblasts (mean of 0.10, SD of 0.30, and mean of 0.31, SD of 0.7 under UVR exposure) transferred mitochondria to each other, with a slight, non-significant increase after UVR. Melanocytes and keratinocytes showed no evident HMT between them (Supplementary Figures 2, 3, 4).

The images in the following Results and Figures using MTR were intentionally overexposed in the red channel to enhance the detection of any potential signs of HMT in the recipient cells. This approach allows us to verify whether nonspecific labeling appears even under overexposure conditions. Notably, in some experimental conditions, CTG-labeled recipient cells show no signs of HMT, supporting the conclusion that there was no nonspecific transfer.

We evaluated the capacity of melanocytes, keratinocytes, and fibroblasts to transfer and receive mitochondria through co-culture combinations. Keratinocytes showed a greater capacity to receive mitochondria from melanocytes, with transfer rates of 26.46% (SD 7.02), increasing to 38.36% (SD 10.6) after UVR exposure, showing significant differences (***p < 0.001) (Fig. 1). Fibroblasts transferred mitochondria to a small number of keratinocytes, with a mean of 0.19% (SD 0.6), which increased significantly to 3.1% (SD 1.6) after UVR exposure (***p < 0.001) (Supplementary Figure 5). Fibroblasts transferred mitochondria to only a small fraction of melanocytes, averaging 2.18% ± 1.2 SD, and this rose slightly to 2.6% ± 1.4 SD after UVR exposure, a non-significant change (Supplementary Figure 6). Keratinocytes transferred mitochondria to fibroblasts at similarly low levels—0.30% ± 0.67 SD in controls versus 0.39% ± 0.95 SD after UVR—with no significant difference (Supplementary Figure 7). The co-cultures shown in Supplementary Figure 8 (melanocytes, MTR; fibroblasts, CTG) and Supplementary Figure 9 (keratinocytes, MTR; melanocytes, CTG) likewise displayed no detectable HMT.

Fig. 1.

Analysis of the interaction and estimation of HMT between melanocytes (mitochondria donors) and keratinocytes (mitochondria recipients) after 22 h of culture, with and without UVR exposure. a. Schematic representation of the melanocyte and keratinocyte co-culture, where donor cells were labeled with MTR and recipient cells with CTG. The arrow indicates that HMT was observed between the cells. b. Histogram generated from image and data analysis (details provided in the corresponding methods section) based on 27 data points for each condition. Normality was assessed using the D’Agostino & Pearson test, followed by non-parametric analyses, including the Mann-Whitney and Kruskal-Wallis tests, to identify statistically significant differences between conditions. A high percentage of HMT was detected, with significant differences among conditions (***p < 0.001). c. The first column presents representative images (Scale: 25 μm) used for HMT analysis. White arrows identify recipient cells that did not receive mitochondria, while red arrows with yellow borders (panels c and d) indicate cells with either internalized mitochondria or mitochondria firmly attached to the recipient cell membrane. The merged image provides a simultaneous view of both donor and recipient cells, highlighting their structure. The red fluorescence-only image offers a clearer view of potential mitochondrial uptake. The second and third columns show a 250% zoom (Scale: 10 μm), improving the identification of internalized or firmly attached mitochondria (MTR) in the recipient cells (CTG). d. Representative images of cells exposed to UVR, as described in c, are shown. A larger image (Scale: 25 μm) and a 250% zoom (Scale: 10 μm) enable the identification of cells that may have taken up mitochondria. Significant HMT, especially after UVR exposure, was observed from melanocytes to keratinocytes

Overall, our findings suggest that UVR exposure increases HMT, particularly in the transfer of mitochondria from melanocytes to keratinocytes, with possible mitochondrial internalization. The responses vary by cell type, where HMT appears to increase after UVR exposure for the other cells, though without significant differences. The minimal HMT observed between cells may result from a non-specific, random process. These results provide new insights into the communication mechanisms between skin cells and suggest a potential role for HMT in maintaining skin cell function and homeostasis.

CDMT is the predominant mechanism under normal and UVR conditions mediating HMT between melanocytes and keratinocytes

HMT can occur via CDMT through TNTs or CIMT by EVs and other forms of mitochondrial transfer and uptake [3, 11, 19, 26, 27]. To identify key mechanisms of HMT in cells—whether primarily mediated by CDMT or CIMT—we focused on cocultures of skin cells that exhibited high levels of HMT, particularly under the 2D co-cultures under normal conditions and UVR exposure. The highest levels of HMT were observed between melanocytes and keratinocytes, which is why we selected them for the transwell assays. Although fibroblasts did not show significant HMT, we also included them to further investigate their transfer potential or unspecific transfer by analyzing fibroblast-to-keratinocyte and fibroblast-to-melanocyte interactions in the transwell.

In the transwell assay, melanocytes transferring mitochondria to keratinocytes resulted in 5.5% (SD 2.54) of recipient cells internalizing or firmly attaching mitochondria, increasing to 8.9% (SD 3.1) after UVR exposure (Fig. 2). For fibroblast-to-keratinocyte transfer, no mitochondrial transfer was observed in the recipient cells in the lower compartment. Similarly, when testing fibroblasts and melanocytes in the transwell to analyze other cell combinations, no mitochondrial transfer was detected via CIMT (Supplementary Figure 10).

Fig. 2.

Transwell analysis of the interaction and estimation of HMT between melanocytes (mitochondria donors in the upper compartment) and keratinocytes (mitochondria recipients in the lower compartment) after 22 h of culture, with and without UVR exposure. a. Schematic representation of MTR-labeled melanocytes in the upper compartment, separated from CTG-labeled keratinocytes in the lower compartment. Arrows indicate that CIMT was observed between the cells. b. Histogram generated from image and data analysis (details provided in the methods section), based on 27 data points for each condition. Normality was assessed using the D’Agostino & Pearson test, followed by non-parametric analyses, including the Mann-Whitney and Kruskal-Wallis tests, to identify statistically significant differences between conditions. Contact-independent HMT was detected, with significant differences among conditions (***p < 0.001). c. The first column presents representative images (Scale: 25 μm) used for CIMT analysis of keratinocytes. White arrows identify recipient cells that did not receive mitochondria, while red arrows with yellow borders (panels c and d) indicate cells with either internalized mitochondria or mitochondria firmly attached to the recipient cell membrane. The merged image provides a simultaneous view of both recipient cells and the MTR mitochondria. The red fluorescence-only image offers a clearer view of potential mitochondrial uptake. The second and third columns show a 250% zoom (Scale: 10 μm), improving the identification of internalized or firmly attached mitochondria d. Representative images of cells exposed to UVR, as described in c, are shown. A larger image (Scale: 25 μm) and a 250% zoom (Scale: 10 μm) enable the identification of cells that may have taken up mitochondria. CIMT was observed from melanocytes to keratinocytes, especially after UVR exposure

These results suggest that the predominant mechanism of HMT between melanocytes and keratinocytes is CDMT. Among the tested cells, whether in 2D co-culture or transwell, fibroblasts showed a reduced capacity to take up mitochondria, which may contribute to their functional decline with age. This finding led us to focus on fibroblasts and the potential benefits of performing AMT.

AMT to skin fibroblasts using mitochondria from MSCs and fibroblasts reduces UVR-induced ROS levels

Fibroblasts constitute an essential component of the skin and play a crucial role in the physiological repair of tissues [28]. Intriguingly, fibroblasts display heightened vulnerability to cellular injury, ROS generation, and genetic damage in both the nucleus and mitochondria when compared to other skin cells such as keratinocytes [29, 30]. In our study, we observed a reduced uptake of mitochondria by fibroblasts from themselves, as well as from melanocytes and keratinocytes, or whether these cells did not transfer mitochondria to the fibroblasts, suggesting a diminished capacity to replenish their mitochondrial pool by HMT, which may reduce their resilience to stress factors. This prompted us to investigate the effects of UVR on ROS production in fibroblasts, corroborating previous studies that showed increased ROS negatively affects cell survival and stress resilience [30–32]. We explored whether AMT of mitochondria from fibroblasts and MSCs to fibroblasts could prevent the UVR-induced increase in ROS levels.

First, we investigated whether MSCs could transfer mitochondria to fibroblasts under stress conditions other than UVR. To simulate nutritional stress, we co-cultured bone marrow-derived mouse MSCs (BM-MSCs) with fibroblasts in a low-serum, low-glucose, L-glutamine-free environment with sodium pyruvate (DMEM, 1% FBS, 1 g/L glucose) for 24 h. These conditions were compared to normal culture conditions (DMEM, 10% FBS, 4.5 g/L glucose, 2 mM L-glut, and sodium pyruvate), using mitochondrial trackers, MTG and MTR. Under normal culture conditions, no evident HMT was observed between MSCs and fibroblasts. However, under nutritional stress, there was a discernible transfer of mitochondria from MSCs to fibroblasts, potentially accompanied by mitochondrial fusion. We qualitatively assessed the assays, observing cells that either potentially transferred mitochondria or had mitochondria firmly attached (Supplementary Figure 11) [33].

After observing that MSCs were able to transfer mitochondria to fibroblasts and based on previous research that has shown that MSCs can transfer mitochondria to other cells and repair cellular and tissue damage [4]. We aimed to test the effects of this process by transferring 500 and 2,500 ng of WJ-MSC mitochondria to 100,000 fibroblasts in a P6 well and then exposing the cells to UVR, as previously described in this article.

To achieve this objective, we first evaluated the structural integrity and function of freshly isolated mitochondria from MSCs and fibroblasts using the Mitochondria Isolation Kit for Tissue (Item: 89801; Thermo Fisher Scientific, Inc., Waltham, MA, USA) as a fundamental step before performing the AMT/T assays. Representative images are shown in Supplementary Figure 12.

Using scanning electron microscopy (SEM), we visualized the complete structure of the network and intact surface of the isolated mitochondria from MSCs (Supplementary Figure 12a-b). Next, the mitochondrial membrane potential was assessed using tetramethylrhodamine (TMRM), observed through a confocal microscope (Leica Sp8) with a super-resolution module (Supplementary Figure 12c). We observed full and intense red fluorescence, corresponding to the active membrane potential of the isolated mitochondria from MSCs (Supplementary Figure 12c). Oxygen consumption was then measured using an oxygraph chamber (Hansatech Oxygraph) (Supplementary Figure 12d).

After assessing the complete structure and functionality of mitochondria isolated from MSCs (Supplementary Figure 12) and following the same isolation procedure for fibroblasts, our results indicated that increasing concentrations of MSC mitochondria (500 to 2,500 ng) labeled with MTR were rapidly internalized or firmly attached to fibroblasts within 20 min (Fig. 3a). We then exposed the fibroblasts that received mitochondria to a 3-minute pulse of UVR, measuring after the ROS levels for each condition using CellROX Green Reagent dye (Thermo Fisher Scientific, Inc., Waltham, MA, USA), by flow cytometry (Fig. 3b).

Fig. 3.

Analysis of AMT to skin fibroblasts using different concentrations of mitochondria from MSCs and fibroblasts to protect against and reduce UVR-induced ROS levels. a. Representative images showing the transfer of mitochondria from isolated MTR-labeled Wharton’s jelly MSCs (WJ-MSCs) and fibroblasts to CTG-labeled fibroblasts at doses of 500 and 2,500 ng. b. Assessment of ROS production in fibroblasts (F) that received 500 and 2,500 ng of mitochondria isolated from WJ-MSCs (MSCM) or from fibroblasts (FM) prior to UVR exposure. c. Evaluation of ROS production in fibroblasts following AMT of 500 and 2,500 ng of WJ-MSC mitochondria (MSCM). Data for b and c were obtained from five independent replicates, each with five technical repetitions by condition. Normality was assessed using the D’Agostino & Pearson test, followed by non-parametric Mann-Whitney (**p < 0.01 and *p < 0.05) and Kruskal-Wallis tests (****p < 0.0001) to identify statistically significant differences between conditions

Our assays show that UVR exposure increases ROS levels in fibroblasts. However, transferring 500 ng of mitochondria isolated from MSCs or fibroblasts reduced ROS levels, this preventing the increase, as evidenced by the mean fluorescence intensity (MFI) of cells labeled with CellROX dye (Fig. 3b). Administering a higher dose of mitochondria (2,500 ng) from both fibroblasts and WJ MSCs resulted in an increase in ROS levels in the recipient fibroblasts, likely enhancing ROS production due to UVR (Fig. 3b). To further understand how MSC mitochondria induce changes in healthy fibroblasts, we transferred 500 ng and 2,500 ng of MSC mitochondria to healthy fibroblasts. No significant change in ROS production was observed for the 500 ng condition. However, a significant increase was noted with the 2,500 ng condition, suggesting a possible enhancement of ROS production that could overlap with the effects of UVR exposure, as shown in Fig. 3b, c.

Overall, our findings suggest that fresh isolation of mitochondria using the Thermo Fisher Scientific, Inc. mitochondrial isolation kit led to functional and intact mitochondria. The transfer of mitochondria isolated from MSCs to fibroblasts can reduce ROS levels induced by UVR and may protect them from damage. However, a high mitochondrial concentration (up to 2,500 ng) may exacerbate the elevated ROS levels caused by UVR.

AMT using MSC-derived mitochondria promotes fibroblast proliferation

During the wound-healing process, fibroblasts proliferate to cover damaged areas and restore the tissue’s structural integrity [34]. This process is negatively affected by aging and UVR stress [35]. After observing a dose-dependent modulation of UVR-induced ROS by MSC-derived mitochondria, we investigated whether mitochondrial transfer could also influence fibroblast proliferation.We transferred mitochondria isolated from human WJ-MSCs to human skin fibroblasts and assessed their proliferation 72 h after the addition of 500 or 2,500 ng of mitochondria. The function and integrity of mitochondria, primarily isolated from MSCs, were evaluated prior to transfer. Using the same kit (Thermo Fisher Scientific, Inc., Waltham, MA, USA) for all cell types, we confirmed that the isolation process did not compromise the structural or functional properties of the mitochondria, supporting their suitability for the functional assays (Supplementary Figure 12).

We found that the transfer of 500 ng of human WJ MSC mitochondria significantly (**p < 0.01) increased the proliferation of 20,000 human fibroblasts compared to 2,500 ng, which also stimulated proliferation but to a lesser extent (*p < 0.05) (Fig. 4a). To further validate this finding, we measured the proliferation of mouse fibroblasts with mouse BM-MSC mitochondria under the same conditions. Similar to the results with human fibroblasts, we observed a highly significant increase in proliferation with 500 ng of mitochondria (***p < 0.001) compared to 2,500 ng (**p < 0.01). (Fig. 4b). Interestingly, AMT of 500 ng of mouse BM-MSC mitochondria in human fibroblasts significantly increased proliferation (**p < 0.01) (Fig. 4c). Our results indicate that the transfer of MSC mitochondria to fibroblasts in both human and mouse settings induces an increase in proliferation, with a dose of 500 ng being more effective for 20,000 cells in a 72 h period. Interestingly, a higher concentration of mitochondria (2,500 ng) displayed a lower capacity to induce an effect in both humans and mice fibroblasts. Finally, the transfer of mitochondria using a xenogeneic model with mouse bone marrow MSCs to human fibroblasts induces a significant effect, although to a lesser extent, with both 500 and 2,500 ng.

Fig. 4.

Stimulation of fibroblast proliferation via xenogeneic and allogeneic AMT using mitochondria isolated from MSCs. We cultured 20,000 human (H) and mouse (M) fibroblasts in a medium supplemented with 1% FBS and 1% PS. Subsequently, these fibroblasts were treated with 500 or 2,500 ng of mitochondria extracted from MSCs (both from humans and mice). After a 72-hour incubation period, cell counts were performed, and the fold-change in experimental conditions was calculated by comparing them to the control group. a. AMT of H fibroblasts with mitochondria isolated from H MSCs. b. AMT of M fibroblasts with mitochondria isolated from M MSCs. c. AMT of H fibroblasts with mitochondria isolated from M MSCs. The data presented in this figure originate from three independent replicates or assays, each consisting of at least three repetitions by condition. To assess normality, a D’Agostino & Pearson test was conducted, followed by non-parametric analyses, including Mann-Whitney, aimed at identifying statistically significant differences among the conditions (*p < 0.05, **p < 0.01, ***p < 0.001)

Localized MT enhances early cutaneous repair in murine and porcine wound models

To determine whether locally administered mitochondria could promote cutaneous repair in vivo (MT), we first used a primary-intention surgical wound model in mice. Allogeneic mitochondria isolated from mouse BM-MSCs were injected intracutaneously into the tissue adjacent to the wound margins immediately after wound closure. Intact mouse BM-MSCs were administered in parallel as a cellular reference treatment. Wound-containing tissues were collected 48 h after treatment and evaluated using hematoxylin and eosin (H&E) and Mallory’s trichrome staining. Histological findings were summarized using the wound healing index (WHI), which integrates granulation-tissue formation, inflammatory-cell infiltration, collagen organization, and the relative abundance of early and mature collagen (Fig. 5).

Fig. 5.

Evaluation of regenerative potential via allogeneic MSC injection and mitochondrial infusion in mouse wounds. Various doses of isolated mitochondria (25, 50, and 100 ng) and MSCs (ranging from 2.5 to 10 × 10⁵ cells) were injected at the wound site, with two injections per side, each consisting of 25 µl. (A detailed description of the methodology is provided in the Methods section.) a. Schematic representation of the application of MSCs and mitochondria. b. Representative images showing the injection procedure in a mouse. c. Representative histological images (H&E and Trichrome staining) taken 48 h after wounding and treatment with vehicle, 25 ng mitochondria, or 250,000 MSCs. Similar results were observed with increasing concentrations of both treatments. d. The wound healing index (WHI) provides a score based on granulation tissue formation, inflammatory infiltrate, collagen organization, and the relative amounts of early and mature collagen. e. Histogram presenting WHI data from at least three independent biological replicates, each with three technical replicates. Normality was assessed using the D’Agostino & Pearson test, followed by non-parametric analyses with Mann-Whitney (****p < 0.0001) and Kruskal-Wallis (**p < 0.01) tests to identify statistically significant differences between conditions

Representative histological sections revealed distinct differences among the experimental conditions. Untreated and vehicle-treated wounds retained a more evident wound depression and discontinuity of the tissue architecture, accompanied by irregular organization of the connective tissue at the wound margins. In contrast, wounds treated with 25 ng of mitochondrial protein exhibited a shallower wound defect, a more continuous epidermal contour, and greater organization of the tissue bridging the injured area. Mallory’s trichrome staining further showed denser and more regularly distributed collagen-containing connective tissue in mitochondria-treated wounds. Administration of 2.5 × 10^5 BM-MSCs produced a similar early histological phenotype, characterized by improved epithelial continuity and prominent collagen-containing tissue surrounding the wound.

Consistent with these observations, administration of 25 ng of mitochondria significantly increased the WHI compared with untreated wounds. Increasing the mitochondrial dose to 50 or 100 ng maintained elevated WHI values but did not produce a proportional additional improvement, indicating the absence of a linear dose-response relationship within the tested range. BM-MSC-treated wounds also exhibited generally higher WHI values than the control wounds; however, increasing the cell dose from 2.5 × 10^5 to 1 × 10^6 cells did not result in a progressive enhancement of repair. Vehicle-treated wounds displayed substantial inter-animal variability and no consistent improvement relative to untreated controls. These results indicate that a relatively low mitochondrial dose was sufficient to induce a near-maximal early repair response and that isolated BM-MSC-derived mitochondria reproduced several of the histological effects observed after administration of intact BM-MSCs.

We next evaluated mitochondrial transplantation in a large-animal model with skin anatomy and repair characteristics more closely resembling those of humans. Xenogeneic mitochondria isolated from well-characterized human WJ-MSCs were administered locally around 1-cm porcine cutaneous wounds at total doses of 25, 50, or 100 ng of mitochondrial protein. Untreated and DMEM-treated wounds served as control conditions. Tissue samples were collected 48 h after treatment and evaluated by WHI scoring, H&E and Masson’s trichrome staining, and spatial quantification of Ki67-positive nuclei in predefined epidermal and dermal compartments (Fig. 6).

Fig. 6.

Local delivery of mitochondria isolated from human WJ-MSCs enhances early repair responses in porcine cutaneous wounds. Mitochondria were isolated from human WJ-MSCs and administered locally around 1 cm porcine cutaneous wounds. Each wound received vehicle or a total dose of 25, 50, or 100 ng of mitochondria, distributed among four 25 µL injections placed as close as possible to the wound margins, with two injections administered on each side of the wound, for a total injection volume of 100 µL per wound. Three assays were indepndently performed with three repetitions by condition, a total of 45 pigs were used in the assays, with one wound analyzed per animal and each animal assigned to a single experimental condition. Tissue samples were collected 48 h after wounding and treatment for histological analysis. a. Schematic representation of mitochondrial isolation from human WJ MSCs and local administration into the tissue adjacent to the wound. b. Representative photographs showing wound delineation and the local injection procedure. c. WHI in untreated control (CTL), vehicle-treated, and mitochondria-treated wounds. Individual symbols represent independent animals, with one wound analyzed per pig, and bars show group means with error bars representing (SD±/SEM). Normality was assessed before statistical analysis. Overall differences among groups were evaluated using the Kruskal–Wallis test, and the indicated pairwise comparisons were evaluated using the Mann–Whitney U test, as described in the Methods. d. Schematic representation of the anatomical regions used for spatial quantification of Ki67 positivity. On each side of the wound, the wound zone comprised the 600-µm region extending outward from the wound tip and included the wound-edge and newly formed epidermal regions together with the underlying dermis. The extreme comprised the subsequent 300-µm region located distal to the wound zone and represented tissue farther from the injury. The central wound bed was analyzed as a separate dermal compartment. Epidermal Ki67 positivity was quantified in the extreme and wound-zone regions, whereas dermal Ki67 positivity was quantified in the extreme, wound-zone, and central wound-bed regions. Areas containing hair follicles, sebaceous glands, or other skin appendages were excluded from the analysis. e. Quantification of Ki67 positivity in the epidermal and dermal compartments of CTL, vehicle-treated, and mitochondria-treated wounds. Ki67 positivity was calculated as the percentage of Ki67-positive nuclei relative to the total number of nuclei within each analyzed region. At least three non-overlapping histological areas were quantified per region and per wound, and the resulting measurements were averaged to generate a single animal-level value for statistical analysis. Box-and-whisker plots show the distribution of animal-level values across experimental conditions. For each anatomical region, the indicated pairwise comparisons among treatment groups were performed using Mann–Whitney unpaired t test, as described in the Methods. f. Representative H&E, Masson’s trichrome, and Ki67-stained tissue sections from CTL, vehicle-treated, and wounds treated with 25, 50, or 100 ng of mitochondrial protein. H&E staining illustrates overall tissue architecture and inflammatory-cell infiltration, Masson’s trichrome staining shows collagen deposition and organization, and Ki67 immunostaining identifies proliferating cells. Scale bars: d, 100 μm; f, 100 μm. ns, not significant; *P < 0.05; **P < 0.01; ****P < 0.0001

The WHI showed a dose-associated increase following mitochondrial administration. Wounds treated with 25 ng remained similar to the control conditions, whereas the 50-ng group exhibited an intermediate response. The highest WHI values were observed in wounds receiving 100 ng of mitochondrial protein, which differed significantly from untreated wounds. Vehicle-treated wounds showed the lowest mean WHI and greater variability, whereas mitochondrial treatment progressively shifted the histological scores toward a more advanced early repair phenotype.

The representative histological sections were consistent with the WHI findings. Control and vehicle-treated wounds showed less organized wound tissue, including irregular dermal architecture and heterogeneous collagen distribution. A residual separation of the wound margins remained evident in representative sections from the 25-ng group. At 50 and 100 ng, the wound tissue appeared more compact and contained denser trichrome-positive connective tissue, with collagen fibers extending through regions involved in wound repair. The 100-ng condition showed the most organized connective-tissue architecture among the mitochondrial-treatment groups. These observations support a dose-associated improvement in early collagen-containing tissue organization, although they do not demonstrate complete wound closure or mature tissue remodeling at this early endpoint.

To determine whether these histological changes were accompanied by increased cellular proliferation, Ki67 positivity was quantified separately in the epidermis and dermis according to the distance from the injury adapting the protocol of Ågren et al. (2021) [36]. No significant differences were detected in the epidermal or dermal extreme regions located distal to the wound, indicating that mitochondrial administration did not broadly increase proliferation throughout the surrounding skin. In contrast, Ki67 positivity increased within the epidermal wound zone, with the highest values observed after administration of 50 and 100 ng of mitochondrial protein. Representative immunohistochemical sections showed Ki67-positive nuclei concentrated near the wound margins and within newly forming epidermal tissue.

In the dermis, Ki67 positivity was not significantly altered in the extreme or wound-zone regions. However, a significant increase was detected within the central wound bed, where the 100-ng group exhibited the highest proportion of Ki67-positive nuclei (Fig. 6). Thus, the proliferative response was spatially restricted to epidermal and dermal compartments directly participating in wound repair rather than being diffusely induced in tissue farther from the injury.

Together, these results demonstrate that locally administered MSC-derived mitochondria retain biological activity in injured skin and promote a more advanced early repair phenotype in both small- and large-animal models. In mice, allogeneic BM-MSC-derived mitochondria produced a strong response at a relatively low dose and reproduced several histological features associated with intact BM-MSC administration. In pigs, xenogeneic mitochondria derived from human WJ-MSCs induced a dose-associated improvement in WHI, collagen-containing tissue organization, and localized cellular proliferation. Because all analyses were performed 48 h after wounding, these findings indicate an enhancement of the early repair phase rather than complete wound closure, restoration of normal skin architecture, or long-term functional regeneration.

Discussion

In this study, we focused on understanding the interactions and HMT among crucial skin constituents: melanocytes, keratinocytes, and fibroblasts. By examining which combination of these cells exchanged more mitochondria and the possible effects in counteracting UVR-induced ROS production, we observed that fibroblasts internalize fewer mitochondria compared to the other cell types. Considering the limited uptake of mitochondria by fibroblasts in co-culture assays, we aimed to determine if transferring mitochondria to fibroblasts could support their capacity to decrease ROS levels and promote proliferation. Our results showed that mitochondrial transfer to fibroblasts resulted in an interesting and positive ROS scavenging effect and increased proliferation. As wound healing is affected by ROS and UVR with effects on skin proliferation and migration [29, 37, 38], we aimed to explore a therapeutic approach by performing MT on wounded skin tissue. By using MT, we were able to improve tissue reconstitution, leading to better structuring and a reduced inflammatory response in mice and pigs. These results provide the basis for further research into HMT between skin cells and the use of MT to treat wounded tissue.

Melanocytes, keratinocytes, and fibroblasts play a pivotal role in shielding the skin from excessive UVR exposure through close interactions, providing structural protection and facilitating post-damage repair [14]. In this context, HMT may play an important role in maintaining the health of skin cells. Our findings highlight distinct capacities for HMT among these cell types. Melanocytes and HaCaT keratinocytes exhibited high rates of HMT, while fibroblasts showed minimal engagement in mitochondrial exchange, either among themselves or with other cell types. Interestingly, although melanocytes demonstrated significant transfer potential with HaCaT keratinocytes, they did not receive mitochondria when keratinocytes were the donors, and received very little from fibroblasts. Further analysis of the mechanisms involved in the transfer and uptake processes, such as the role of heparan sulfates (HS) in mitochondrial uptake, could be essential to uncover the molecular mechanisms underlying HMT in the skin [7]. Understanding the dynamics of this transfer is critical, as supporting cells that do not efficiently receive mitochondria could be key to improving their capacity to renew defective mtDNA or mitigate the damaging effects of UVR and other environmental stressors.

Skin cells are part of a complex organ where receiving mitochondria can give one cell type a competitive advantage over less fit cells. It has been observed that the differential expression of genes, such as higher levels of hemidesmosome component collagen XVII (COL17A1) in epidermal stem cells, helps these cells overcome those that are harmed or aged [39]. We hypothesize that helping cells, such as fibroblasts or other skin cells, to become healthier by AMT to them could be a strategy to maintain healthy skin tissue over time, providing them with a possible competitive advantage. Different strategies could be employed to deliver healthy mitochondria to these cells, including direct administration to the skin site, microneedles, or emulsions, facilitating their internalization [14].

Skin cells are constantly sharing intracellular material among themselves to protect against damage, which could be a reason for the observed HMT. Observing the transfer of mitochondria among skin cells in our study adds new insights to skin cell communication and the role of this process in preserving tissue function. It is well known that melanocytes transfer melanosomes to keratinocytes through vesicles, leading to the formation of a perinuclear melanin cap that protects these cells from UVR harm [40]. Moreover, previous reports suggest that keratinocytes secrete vesicles to enhance melanin synthesis and increase the activity of melanosomal proteins in melanocytes [16]. In the case of fibroblasts, they appear to uptake melanosome clusters and even apoptotic melanocytes after co-culture to support tissue function and reduce toxicity induced by environmental injuries [41]. The transfer of mitochondria and other cellular components between these cell types might play one of the many crucial roles in maintaining tissue homeostasis and adapting to environmental challenges.

Our results suggest that UVR exposure enhances HMT between melanocytes and keratinocytes with already high transfer rates. UVR exposure is a potent stimulant for stress responses and survival signals in cells [42]. This article provides early evidence of a possible new stimulant of mitochondrial transfer: UVR. UVR exposure and the induction of HMT could be potentially linked to other mechanisms related to the release of intracellular material. A key mechanism triggered by UVR is skin pigmentation. UVR exposure leads to melanosome transfer between melanocytes and keratinocytes, mediated by Myosin X-controlled melanocyte filopodia, E-cadherin, and extracellular calcium availability [43]. Additionally, UVR is a major stimulant for melanosome uptake by keratinocytes in a process that seems to be mediated by acetylcholine. Acetylcholine is released by keratinocytes, regulating the production of melanin by melanocytes and intracellular calcium mobilization induced by UVR exposure, stimulating keratinocyte phagocytosis. Thus, the cholinergic system plays a key role in melanosome uptake by keratinocytes after UVR exposure [44]. As there is a close association between mitochondria and melanosomes [45], the mentioned mechanisms in melanosome transfer may be involved in the mitochondrial transfer between these skin cells as well.

In this article, a dose of UVR at 5.4 mJ/cm² (UVA/UVB) induced the transfer of mitochondria. However, it has been reported that a combined dose of UVA at 10 J/cm² and UVB at 0.5 J/cm² induces high cytotoxicity in HaCaT keratinocytes and malignant melanoma cells [46]. This could imply that the selection of 5.4 mJ/cm² (UVA/UVB) is sufficient to study HMT between skin cells without causing significant cytotoxicity. Understanding better how different doses of UVA/UVB stimulate mitochondrial transfer or induce varying levels of cellular stress, potentially leading to cell death, could help determine how to optimize this process to maintain healthy tissue function and HMT.

UVR induces major alterations in cells depending on the wavelength and dose, ranging from growth inhibition and proliferation to damaging membrane phospholipids, proteins, and nucleic acids [47]. To determine if exposure to UVR at 5.4 mJ/cm² was inducing cell permeabilization, which could lead to an unspecific estimation of HMT, we performed an assay exposing the cells to two major stressors: our UVR exposure protocol and H₂O₂. The cells were labeled with Trypan Blue to assess membrane integrity and survival. Interestingly, we observed that cells exposed to UVR did not change significantly in morphology and were not labeled in blue. In contrast, some cells exposed to H₂O₂ turned blue and were positive to the Trypan Blue test. This result supports the specificity of the HMT estimation between melanocytes, keratinocytes, and fibroblasts, indicating that the cell membranes maintained their integrity. H₂O₂ is known to induce HMT in somatic cells, where the transfer of damaged mitochondria stimulates other cells, such as MSCs, to produce more mitochondria and share them with the harmed counterparts [48].

Even though our results show that HMT between cells could have positive effects in reducing ROS and increasing the proliferation of skin cells, it is very important to note that long-term assessment of the transfer and possible changes in the shared mitochondria could be associated with tissue decay. Prolonged UVR exposure has been linked to cellular senescence in skin cells, affecting both keratinocytes and fibroblasts [49, 50]. Notably, research has indicated that senescent fibroblasts secrete more EVs, which impact neighboring cells such as keratinocytes and elevate interleukin-6 (IL-6) levels [51]. The possibility of increased EV production by skin cells following UVR-induced senescence might be connected to the release of mitochondria with effects that still need to be further understood. One possibility is that HMT could be essential for maintaining cellular function. However, it could also serve as a vehicle to induce tissue senescence. Further research is needed to elucidate whether the net effect of HMT under different conditions is beneficial or detrimental to overall skin health.

Cells with an incapacity to receive new mitochondria from other cells could be more susceptible to damage [3, 6, 52]. This is why we focused on fibroblasts. Fibroblasts seem to not exchange mitochondria among themselves; however, in further analysis we were able to observe that MSCs successfully transfer mitochondria to these cells. Based on this evidence, we performed AMT with isolated mitochondria of MSCs to fibroblasts and observed if the production of ROS decreases by this treatment. We observed that the transfer of 500 ng MSC mitochondria, compared to 2,500 ng and UVR exposed cells, effectively reduced ROS in comparison to isolated fibroblast mitochondria. A higher concentration of mitochondria (2,500 ng) could impair cellular function. Mitochondria that were not uptaken by the recipient cell or assimilated by the cells could be malfunctioning and perceived by the cells as a damage signal, sometimes inducing proliferation [53]. The exposure of recipient cells to high concentrations of mitochondria, as seen with 2,500 ng, may induce a shift in metabolism and possibly detrimental phenotypic changes. Previous studies have shown a similar response in MDA-MB-231 cells, where a higher amount of mitochondria induced a decrease in cell proliferation [54]. However, it remains puzzling and unclear how mitochondria inside cells will behave, as it seems that different concentrations induce contrasting effects [55, 56].

We verified that our isolation procedure, performed using the same Thermo Fisher kit across all cell types, preserves mitochondrial integrity regardless of the source. Pre-transfer assessments showed intact double membranes, sustained membrane potential, and ADP-stimulated respiration in MSC-derived mitochondria (Supplementary Figure 12); similar results were obtained with MSC mitochondria from other partner laboratories, consistent with previous findings from our group [57]. Consistently, Fig. 3 demonstrates that fibroblast-derived mitochondria, once transferred, significantly attenuate UVR-induced ROS in recipient fibroblasts, exhibiting bioactivity comparable to that of MSC-derived mitochondria.

Different mechanisms have been associated with the intake of extracellular mitochondria (ex-mito) by the cell and their effects inside. It has been observed that mitochondrial uptake may depend on HS on recipient cells such as macrophages, which are able to be recipients of neighboring adipocytes mitochondria [7]. The uptake of ex-mito from stressed somatic cells acts as a danger signal, inducing an anti-apoptotic phenotype in MSCs, promoting mitochondrial biogenesis, and increasing the capacity of MSCs to donate their mitochondria to injured cells and combat ROS [48]. A better understanding of how skin cells accept mitochondria, whether the process is mediated by HS, and if the internalized mitochondria induce mitophagy or mitochondrial biogenesis are key aspects that need to be addressed.

In our study, we utilized HaCaT cells grown in EpiLife medium supplemented with HMGS2 instead of HKGS. The use of HMGS2, which contains basic FGF instead of EGF, raised the question of whether this substitution could affect the integrity and behavior of keratinocytes. Literature indicates that both FGF and EGF share common intracellular signaling pathways, and FGF can stimulate similar responses in keratinocytes [58]. Given the role of keratinocytes within the melanocyte skin ecosystem, where factors like FGF, NGF, and endothelin-1 are naturally present and support melanocyte growth, exposing keratinocytes to HMGS2 is a reasonable option [59]. The FGFR signal is determined by the combination of the HS co-receptor and the concentration of the FGF ligand, which is relevant since HMT is mediated by HS [60]. The exposure of HaCaT keratinocytes to FGF could influence their HMT capacity and the internalization of mitochondria by recipient cells when co-cultured. The stimulation of keratinocytes or other skin cells to uptake higher concentrations of mitochondria through the use of growth factors or other agents could stimulate cells to undergo this process more effectively. This artificial enhancement could potentially improve or restore the lost functions of skin tissue over time.

In light of the different rates at which melanocytes, keratinocytes, and fibroblasts engage in HMT, our investigation centered on understanding the nature of this transport, whether by CDMT or CIMT. CDMT can be addressed by 2D co-cultures, and CIMT by transwells. Determining the mechanism of transfer by CDMT and CIMT is challenging, especially in identifying the precise mode of transport of mitochondria, such as TNTs for CDMT or EVs by CIMT, and later quantifying how much mitochondria is being transferred. As observed, the amount of internalized or firmly attached mitochondria varies among the recipient cells. We assessed this transfer by fluorescent microscopy and performed several washes to ensure that mitochondria were being transferred or at least attached to the recipient cell membranes. In some cases, we observed that the transferred mitochondria take space in the recipient cell cytoplasm, leaving the nuclei space free, which reasonably suggests internalization (Figs. 1, 2, 3 and Supplementary Figures 2–11) Further quantification of the amount of transferred mitochondria could be assessed with the use of confocal microscopes, fluorescence-activated cell sorting, single-cell analysis, or other methods available to understand and define the transfer [27, 61].

Identification of the transfer mechanism during co-culture assays, whether through TNTs or EVs, may require the use of several techniques, including potent inhibitors of cell-to-cell interactions and microvesicle release. Among the different methodologies involved in studying their formation, options include using actin inhibitors, monitoring TNT-mediated vesicle transfer, and observing the reception of the cargo by recipient cells [62]. To determine the transfer by EVs, it is necessary to process the cell supernatant through centrifugation, ultracentrifugation, nanoparticle tracking, western blot, flow cytometry, and confocal imaging studies, among others [63]. It has been observed that mitochondria can physiologically exist as ex-mito (mitochondria without wrapping membranes). This mechanism of CIMT still represents a challenge for further study as there are many mechanisms of liberation and characterization of ex-mito [3, 26]. It is key to perform analysis of mitochondrial structure using fluorescence-activated cell sorting, fluorescence microscopy, transmission electron microscopy, oxygen consumption analysis, among others [64]. In this article, we provide the first evidence of the transfer of mitochondria between skin cells and more specifically from melanocytes to keratinocytes. Based on our quantification of the transferred mitochondria, we suggest that the transfer mechanism is mainly due to CDMT. Without a doubt, further characterization of the transfer mechanism will be important to address the pathways of transport and to study how to improve the transfer of mitochondria, especially if clear evidence supports the role of HMT as a key mechanism to maintain tissue health.

The transwell assay is a widely used methodology to study CIMT. By creating a physical separation between donor cells in the upper compartment and recipient cells in the lower compartment, this method facilitates the identification of EVs or other transfer mechanisms that pass through a 3 μm mesh membrane filter. In our transwell assay, we focused on cell combinations with a high transfer rate under both normal and UVR-exposed conditions to compare them to direct co-culture. Specifically, we observed that HMT from melanocytes to keratinocytes decreased in the transwell setup, suggesting that this transfer primarily occurs through CDMT. This emphasizes the role of physical interactions in facilitating mitochondrial exchange and suggests that UVR exposure impacts the mechanisms of transfer more profoundly when cells are not in direct contact.

In other examples of the validity of the transwell, upon co-culturing MSCs and corneal epithelial cells, an augmentation in HMT was observed subsequent to oxidative stress induced by rotenone. This increase coincided with a rise in cell-to-cell connections, particularly through TNTs, which appeared to confer protection against oxidative damage [23]. Conversely, when these interconnected cells were physically separated using the transwell setup, the HMT exhibited a reduction, and consequently, no observable safeguarding effect on the affected cells was noted. This disparity underscores the significant role of direct cell-to-cell interactions, possibly mediated by TNTs, in facilitating efficient HMT and conferring cellular protection in the context of oxidative stress [23].

The differences observed in the pattern of HMT between melanocytes, keratinocytes, and fibroblasts, the use of 2D co-culture and transwells, the observation of mitochondria near the nuclei, and the exposure of cells to UVR provide reasonable evidence of a release of mitochondria through CDMT and CIMT and its uptake by recipient cells. Methodological concerns published in the literature have raised awareness of the potential for nonspecific staining and observation of transfer [65]. These concerns and proposed solutions have improved the protocols for using MTR, particularly in this study. It is crucial to meticulously wash MTR-labeled donor cells four to five times to prevent dye leakage [65]. A key insight into nonspecific MTR coloring is the observation of high transfer rates with no differences among conditions, which is not the case in our article [65]. For example, co-culturing cells in a transwell seems to inhibit the nonspecific passage of mitochondria [65]. Even if our results reasonably show that HMT occurs between skin cells, further use of other methods to label mitochondria is recommended, such as using stable transgene systems that encode for mitochondria-localized tags or fluorescent proteins [27]. However, this technique also requires setup and analysis to ensure it does not perturb mitochondrial function or cause immunogenicity [27]. More recently, genetically encoded mitochondrial tracing systems have enabled the stable labeling of donor mitochondria and their visualization within the intracellular compartment of recipient cells. However, these approaches typically require transfection and antibiotic selection, making them technically demanding and potentially capable of altering cellular phenotype, mitochondrial function, or cell viability. Such modifications may, in turn, affect the intrinsic capacity of donor cells to release mitochondria or of recipient cells to internalize them. Therefore, labeling conditions must be carefully optimized and validated to ensure that the tracing strategy itself does not influence the frequency, mechanism, or biological properties of HMT [66].

Wound healing and maintaining cell functionality are key areas that need to be fully addressed when studying the potential effects of AMT/T, from in vitro experiments to in vivo studies and eventually clinical trials. Finding the best in vivo model is essential for accurately assessing the efficacy and safety of AMT/T before moving on to clinical trials [34]. After observing that fibroblasts with MSC-derived mitochondria showed reduced ROS production following UVR exposure, we aimed to determine whether AMT/T could induce fibroblast proliferation. Our findings indicate that the AMT from both human and mouse MSCs to human or mouse fibroblasts successfully promoted the proliferation of these cells.

The intracellular mechanisms behind the reduction of ROS production induced by UVR after fibroblasts receive mitochondria from MSCs are of significant interest. Similar work has been performed in vivo, but it was applied to improving cardiac function after infarction. It has been shown that isolated MSC mitochondria, rather than fibroblast mitochondria, produce better results in promoting cardiac function recovery. Interestingly, another study found that the transfer of isolated MSC mitochondria inhibited H₂O₂-induced senescence by decreasing the fluorescent signal in a cytometry assay via p-ERK activity and reducing the expression of senescence markers [67]. The transfer of mitochondria isolated from platelets to human dermal fibroblasts enhanced wound healing in vitro and reduced ROS, even in cells exposed to different stressors such as hydrogen peroxide, cisplatin, and TGF-beta [68]. MSCs were able to transfer mitochondria to mice and human fibroblasts with a defect in the nuclear DNA-encoded NDUFS4 subunit of complex I. The transfer of mitochondria by MSCs was visualized in 13.2% of human fibroblasts and 6% of mouse fibroblasts, resulting in lowered cellular ROS. However, the expression and activity of the CI protein were not rescued [69]. In contrast to the evidence related to a decrease in ROS levels, it has been observed that damaged somatic mitochondria transferred to MSCs lead to an increase in ROS levels, which activates mitophagy and mitochondrial biogenesis and supports the MSCs’ capacity to protect cardiac tissue [48]. Additionally, studying the tumor microenvironment has revealed that the transfer of dysfunctional mitochondria from macrophages induces ROS accumulation in recipient cancer cells, activating ERK signaling and promoting proliferation in vivo [70]. It seems that the effects of HMT, AMT/T and MT depend on the donor and recipient cell types, as well as the context, the type of stress, its complications, and the roles of multiple elements within the complex ecosystem of wounded or diseased tissue. The transfer of MSC mitochondria to damaged cells and tissues appears to lower ROS levels but still induces proliferation. Verifying mitochondrial viability is crucial; however, it is challenging to ensure that all cells uptake the same quality and number of mitochondria. Fibroblasts are not similar to cancer cells, but many aspects of their biology could affect how they respond to mitochondrial transfer by decreasing ROS and inducing proliferation. Further research into the specific conditions and mechanisms governing these interactions will be crucial for developing effective mitochondrial transfer therapies for various pathological conditions.

UVR induces profound cellular changes, including alterations in proliferation, metabolic rewiring, and even carcinogenesis. For instance, UVB exposure in keratinocytes results in a biphasic activation of nicotinamide adenine dinucleotide phosphate oxidase 1 (NOX1), impairing nucleotide excision repair and contributing to UVB-induced skin cancer [71]. Metabolic reprogramming also precedes primary skin tumor formation, as evidenced in UVB-induced skin cancer models, where glycolysis, the tricarboxylic acid (TCA) cycle, and fatty acid β-oxidation are decreased in the early stages of photocarcinogenesis. In contrast, the distal part of the electron transport chain (ETC) is upregulated, and reductive glutamine metabolism, alongside dihydroorotate dehydrogenase (DHODH) activity, plays a crucial role in sustaining high ETC activity. Notably, mice with impaired DHODH activity or ETC functionality fail to develop pre-malignant and malignant lesions [72]. Furthermore, chronic inhibition of DHODH using leflunomide (LFN) blocks UVB-induced tumor initiation, highlighting the transcriptional regulation of DHODH by STAT3 in this context [73]. These findings raise intriguing questions about the role of HMT in skin UVB-induced carcinogenesis. Specifically, investigating the effects of HMT between melanocytes and keratinocytes on tumor formation under such conditions is essential. Additionally, AMT may influence this process, necessitating studies to evaluate both the therapeutic potential and safety of applying isolated mitochondria to the skin. A deeper understanding of the molecular mechanisms underlying HMT is essential to elucidate its implications for skin biology and its potential interactions with UVR-induced carcinogenic processes.

Considering that an excess of ROS and a lack of proliferation from fibroblasts can affect wound regenerative capacity, we conducted a study to investigate whether the transfer of MSCs or mitochondria isolated from MSCs could help to regenerate tissue faster using a wound model in mice [34]. Our results showed that transferring 25 ng of MSC mitochondria or 250,000 MSCs had similar effects. Interestingly, we did not observe any immunological response that led to a decrease in wound repair. On the contrary, the wounds seemed to heal faster, even in a model organism that already has a fast repair capacity. The source of the cells used to isolate mitochondria was the bone marrow of C57BL/6 mice, and we applied these mitochondria to wild-type Swiss white mice, representing an allogeneic MT. It has been previously shown that both autologous and allogeneic MT do not induce an immunological response, suggesting no inflammatory effects associated with MT [74–78]. The alloresponse and allorejection of MT were investigated after the intraperitoneal injection of increasing concentrations of mitochondria and dosages over different days. The experiments were performed using BALB/cJ mice receiving mitochondria donated from C57BL/6J mice. No evidence of an immune reaction was observed after seven days [78]. Furthermore, we observed that the AMT by MitoCeption from human MSCs to human in vivo polarized CD4 + Th17 cells significantly increased the number of Treg cells, with an increase in FOXP3-producing cells and a decrease of IL-17. This assay demonstrated that allogeneic AMT seems to lower a pro-inflammatory phenotype, transforming Th17 cells into a regulatory type [22]. Achieving a 100% absence of inflammatory response from isolated mitochondria that are allogenically transferred to different donors may be difficult to guarantee. Mitochondria are delicate organelles that may suffer from various environmental stressors, which can damage their structure and induce a pro-inflammatory response similar to damage-associated molecular patterns (DAMPs). Protecting the mitochondrial structure during transfer and transplantation is crucial for a successful transfer in certain scenarios and to not activate an immune response [2, 3, 6]. However, understanding how different mitochondrial gene types could induce different responses in recipient cells is challenging. The incorporation of transferred mitochondria and their fusion with the endogenous mitochondrial network are not always assured. Nonetheless, the transfer of specific resistance traits coded by the mtDNA of the transferred mitochondria is possible in mammalian cells. This could mean that the acquisition of a single nucleotide polymorphism (SNP) coding for an advantage could be integrated into the recipient cell and change its genotype and phenotype [79]. In a recent study by Jacoby and colleagues, mitochondrial augmentation therapy (MAT) was applied to individuals with single large-scale mitochondrial DNA deletion syndromes (SLSMDs) as part of a compassionate use program. CD34 + cells were collected from patients, augmented ex vivo with maternally derived healthy mitochondria, and then reintroduced to the patients. MAT resulted in decreased heteroplasmy, increased full-length mitochondrial DNA, and improved ATP content in peripheral blood mononuclear cells [80]. Taking all this together, there is reasonable evidence to suggest that intact mitochondria can be transplanted allogenically, transferring some traits of their mtDNA without causing an evident immune response.

The induction of cell proliferation, a decrease in ROS, and the promotion of an immunoregulatory phenotype by immune cells, along with findings on the therapeutic effects of transplanted mitochondria in wounded tissue (such as ischemic heart tissue), provide compelling evidence to further analyze the role of mitochondria in skin wound repair [22, 68, 81, 82]. In this study, we used minimally invasive and clinically relevant primary-intention wound-healing models in mice and pigs to evaluate the early effects of MT and MSC administration [38]. These models were selected because of their reproducibility, practicality, and reduced postoperative care requirements compared with secondary-intention wound models. In particular, they minimize the need for prolonged administration of antibiotics, analgesics, and anti-inflammatory drugs, thereby reducing potential pharmacological confounders. The 48-hour endpoint allowed us to assess early tissue responses to treatment. In mice, which exhibit relatively rapid wound-healing kinetics, this time point permitted the detection of initial changes in inflammation, granulation tissue formation, and collagen organization. In pigs, Ki67 immunostaining provided evidence of early proliferative activity within the treated wounds. Studies incorporating longer follow-up periods will therefore be necessary to determine whether the early effects observed after mitochondrial or MSC administration translate into durable improvements in wound closure and tissue quality. However, extended experiments may also require more intensive postoperative care and prolonged exposure to analgesics or other medications, which could independently influence wound healing and potentially modify the therapeutic effects of transplanted mitochondria or MSCs [34, 83, 84].

In addition, the decision to use a linear incision in our skin wound model was deliberate and based on ethical and methodological considerations that offer several key advantages over the more commonly used circular incision [85]. Circular wounds, pose challenges such as increased stress, pain, and risk of infection for the animal models, as well as susceptibility to wound contraction, which could confound the effects of therapeutic interventions [34]. Linear incisions, in contrast, provide a uniform wound edge, simplifying the measurement of key parameters such as re-epithelialization and wound closure along a single axis, thereby reducing variability and enhancing data reliability [34]. Additionally, linear wounds are more clinically common, as they closely mimic surgical incisions encountered in human medical practice, increasing the translational applicability of the findings.

Importantly, although our histological analyses provided valuable insights into the structural aspects of wound healing, further cellular-level assessments are essential for a more detailed understanding of the underlying biological processes. Investigating cellular responses, including immune cell phenotypes, fibroblast and keratinocyte behavior, and the molecular signaling pathways involved in the repair process, would provide a more holistic view of how transplanted mitochondria and MSCs contribute to wound regeneration. For example, examining immune cell recruitment and polarization, mitochondrial dynamics in recipient cells, and intercellular communication within the wound environment could elucidate critical mechanisms that drive the observed therapeutic effects.

Such additional cellular analyses could not only validate and expand upon the findings of this study but also help identify potential biomarkers or molecular targets for enhancing wound healing therapies. By combining histological outcomes with cellular and molecular-level assessments, future studies could provide a more comprehensive framework for understanding and optimizing the use of mitochondria and MSCs in clinical wound healing applications.

The development of Advanced Therapeutic Medicinal Products (ATMPs) is a highly diverse field, with potential avenues emerging from observations of factors mediating intercellular processes that demonstrate regenerative or reparative effects. MSCs, one of the most prominent cell-based agents in regenerative medicine, have been shown to produce factors that induce tissue repair [19, 86–88]. These factors have led to the development and clinical validation of various products, including microvesicles, conditioned media, and cocktails of immunomodulatory molecules. Interestingly, mitochondria actively transferred from MSCs to other cells have emerged as a promising therapeutic agent for skin and other tissues [22, 24, 89, 90]. This study provides the first evidence that HMT occurs between skin cells and demonstrates that exposing fibroblasts to isolated mitochondria enhances their proliferation and resilience to ROS induced by UVR exposure. Together with other findings in the field, this data suggests that isolated mitochondria and their targeted transfer to cells or tissues could represent a novel class of ATMPs. Furthermore, AMT and MT have shown the potential to regulate immune responses without eliciting significant adverse reactions from the recipient’s immune system. However, understanding the allogeneic effects of MT is crucial for elucidating its long-term safety and efficacy [2, 14, 19, 52, 91]. Despite these promising implications, it is important to address the potential risks and limitations associated with mitochondrial manipulation. Key challenges include ensuring the quality and functionality of isolated mitochondria, preventing potential immune reactions in certain contexts, and developing robust preservation methods to maintain mitochondrial integrity during storage and transport. Advancing our understanding of these issues will be critical for translating mitochondrial therapies into clinical practice and realizing their potential as transformative ATMPs in regenerative medicine.

In conclusion, this study reveals a cell-type-specific pattern of HMT in the skin, characterized by preferential HMT from melanocytes to keratinocytes, particularly following UVR exposure, whereas fibroblasts exhibited limited participation in spontaneous mitochondrial exchange. These findings suggest that HMT may represent a previously underrecognized component of the cellular response of the skin to environmental stress. Importantly, the targeted delivery of MSC-derived mitochondria to fibroblasts reduced UVR-induced ROS production and promoted cell proliferation in vitro, while local MT enhanced early histological repair and localized proliferative responses in murine and porcine wounds. Together, these results establish a conceptual link between endogenous mitochondrial exchange and the therapeutic potential of AMT and MT in cutaneous repair. Nevertheless, further studies are required to define the molecular mechanisms governing mitochondrial release and uptake, determine the intracellular fate and functional integration of transferred mitochondria, establish optimal dosing and delivery strategies, and assess long-term safety and regenerative efficacy. Addressing these questions will be essential for translating mitochondria-based interventions into reproducible and clinically relevant therapies for skin injury, UVR-induced damage, and impaired wound healing.

Methodology

Cell culture

Human Epidermal Melanocytes (HEM, neonatal) Highly Pigmented (HP) (HEMn-HP, Thermo Fisher Scientific, Inc., Waltham, MA, USA) were cultured in M254 medium (Item: M-254-500; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with human melanocyte growth supplement (HMGS2) (Item: S-016-5; Gibco, Cascade Biologics, Portland, OR, USA). Human immortalized keratinocytes (HaCaT) were acquired from Thermo Fisher Scientific, Inc., and grown in EpiLife medium (Item: MEPI500CA; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with HMGS2. Keratinocytes (HaCaT) were cultured with FGF, Insulin, Transferrin, Bovine Pituitary Extract, Endothelin-1, FBS, Heparin, and Hydrocortisone (HMGS2), and more particularly FGF instead of EGF. It has been shown in the literature that both FGF and EGF have common intracellular signaling pathways, and FGF could be used to stimulate similar responses [58]. Given that keratinocytes are already part of the melanocyte skin ecosystem and that FGF, NGF, and endothelin-1 are keratinocyte-derived factors supporting melanocyte growth, exposing keratinocytes to HMGS2 could help us to replicate the skin environment [92]. Normal human dermal fibroblasts (primary) were donated in collaboration with Maroun Khoury at the IMPACT Center. One fibroblast line was purchased from Lonza Biologics (Basel, Switzerland; Item: CC-2511), and the other two were isolated from foreskin collected from healthy donors with informed consent from the parents. Mouse Bone Marrow (BM) MSCs were characterized and donated in collaboration with Patricia Luz Crawford from the Universidad de los Andes, Santiago, Chile. These cells were isolated from the bone marrow of wild-type C57BL/6 mice and characterized as previously described [93–95]. Human WJ MSCs, isolated from the umbilical cord of healthy and full-term infants and characterized as previously described [96], were donated by Maria Ines Mitrani and AssureImmune, Inc. in Miami. Mice BM MSCs and fibroblasts were cultured in high glucose Dulbecco’s modified Eagle’s medium (DMEM, Item: 12430-054) supplemented with 1% penicillin/streptomycin solution (PS; 10,000 U/mL and 10,000 µg/mL, respectively; Item: 17-0602E), 1% L-glutamine (200 mM), 25 mM HEPES (all from Gibco, Cascade Biologics, Portland, OR, USA), and 10% Foetal Bovine Serum, Qualified (FBS) (Item: 10437-028, Gibco, Cascade Biologics, Portland, OR, USA). Human MSCs were grown in a medium composed of 20% FBS and 1% penicillin/streptomycin in α-MEM from Gibco or 5% PLT max (Millcreek), heparin, and 1% penicillin/streptomycin. All cells were cultured in a 37 °C incubator with 5% CO₂ and split with 1X Trypsin from Lonza (Item: BE02-007E) after a wash with Phosphate-Buffered Saline (PBS) without calcium or magnesium.

The utilization of cells and cell lines provided from established institutions in our study was reviewed by the Universidad San Francisco de Quito USFQ Internal Review Board (USFQ-IRB), which determined that it did not require further approval (Document number: N°. 065IN-2021-CEISH-USFQ dated on May 7th, 2021).

Cell labeling

First, donor cells were labeled with 285 nM MitoTracker Red (MTR, Item: M7512; Invitrogen, Thermo Fisher Scientific, Inc., Waltham, MA, USA), while recipient cells were labeled with 5 µM Cell Tracker Green (CTG, Item: C7025; Life Technologies, Thermo Fisher Scientific, Inc., Waltham, MA, USA) for 20 min in T75 flasks, following previously established protocols [54, 97, 98]. After staining, the cells were thoroughly washed with 10 mL PBS, and fresh media was added to the flasks. After 2 h, an additional wash step was performed. On the following day, donor cells were first plated. After they attached, which took overnight, they were washed with PBS, and then the recipient cells were added to the culture dish using a detachment reagent without EDTA. Subsequent assays were co-culture assays, transwell experiments and others performed in this article.

Culture and co-culture conditions and determination of cells with internalized or firmly attached mitochondria

After labeling, 50,000 cells of each cell type (donor MTR and recipient CTG) were plated in a P6 well plate and incubated for 22 h at 37 °C with 5% CO2. For UVR-exposed cells, at 18 h of culture and co-culture, the cells were exposed to a 3-minute UVR pulse and then incubated for an additional 4 h, totaling 22 h of culture. UVR exposure details will be provided later in this section. The medium from the donor cells was used when co-culturing with a different recipient cell type. For example, complete DMEM medium was used for co-culturing fibroblasts donating mitochondria to HEMn-HP recipient cells. Prior to live imaging, cells were verified to have 50–60% confluence, washed twice, and fresh medium was added to remove any unattached or non-internalized mitochondria. Images were captured using a Nikon Eclipse Ts2R Microscope. Three biological replicates (assays) were conducted, with three technical repetitions for each condition (control or UVR-exposed cells). For each technical repetition, three random fields were selected, and 10 representative images were captured per field to identify and count the total number of cells that had received mitochondria or had mitochondria firmly attached to their membranes. Verification of cell counts and mitochondria presence was performed at 20X or 40X magnification. For data analysis, the mean number of mitochondria-receiving cells from all captured images was calculated, resulting in three data points representing each repetition of each condition. In total, 27 data points were collected across the three biological replicates. The analysis was performed using GraphPad Prism version 10.3.1 for Windows, GraphPad Software, Boston, Massachusetts US. To determine the potential for internalization, we adjusted the focus of the cells and observed whether the green-labeled cells co-localized with red dots, resulting in a yellowish color, which would indicate internalized mitochondria surrounding the nuclei of the recipient cells. The data are presented as the percentage of positive cells that demonstrated firm attachment or internalization of mitochondria.

Verification of no permeabilization of cell membrane after UVR exposure

50,000 cells were cultured per well in a 12-well plate for 24 h with the specific media for each cell type, establishing three experimental conditions: Control, UVR Lamp exposure, and H₂O₂ treatment. After the initial culture period, the UVR group corresponding to an energy density of 5.4 mJ/cm² was exposed to ultraviolet radiation for three minutes with the plate caps on. Four hours following UVR exposure, the wells in the H₂O₂ group were treated with 45 µL of 3% hydrogen peroxide in 0.5 ml for 15 min to induce an increase in ROS levels associated with oxidative stress. After treatment, all cells were washed with 250 µL of Phosphate-Buffered Saline (PBS) and stained with 500 µL of 1X Trypan Blue. The cells were then incubated for 10 min at 37 °C. To assess membrane permeability, imaging was performed randomly to the wells immediately before and after staining using a Nikon Eclipse Ts2R Microscope and equipped with an OMAX Digital Camera A35180U3 made in China. For the qualitative assessment, three biological replicates (three different assays) were performed, with three technical repetitions for each condition. Images shown were randomly selected (Supplementary Figure 1).

UVR exposure

For the UVR-exposed cell conditions, after staining and 18 h in culture at 50–60% confluence (to ensure enough space for cells to differentiate between those that transferred and received mitochondria), the cells were exposed to a UVR lamp in a three minute pulse. The medium used was specific to the donor cell type; for example, if HaCaT cells (donor cells) were labeled with MTR and co-cultured with fibroblasts, the medium was EpiLife medium supplemented with HMGS2, in which they received the UVR. The UVR lamp primarily emitted a 254 nm wavelength UVC inside a BS-02 UVR system (Purifier Logic Class II, KS, USA) for three minutes with the plate cover on. The plastic cover has the capacity to filter most of the UVC light, a procedure studied and accepted as a standard in photobiology [99]. The power of the UVR lamp that received the cells with the plastic cover on was measured by an UVA/UVB Light Meter 85,009 (Sper Scientific, Scottsdale, AZ 85260, USA). The power of the UVR lamp that reached the cells with the plastic cover on was 0.03 mW/cm², which, for a three-minute exposure, results in an energy density of 5.4 mJ/cm². If cells were left without the plastic cover, they received 0.06 mW/cm², which for three minutes results in an energy density of 10.8 mJ/cm². This higher exposure resulted in cellular stress and the formation of vesicles in their cytoplasm. We estimated the dosage of UVR in setup assays using additional time points and the power of 0.03 mW/cm² and 0.06 mW/cm² (data not shown).

Changes in HMT were assessed four hours later using the Nikon Eclipse Ts2R Microscope, completing 22 h. The four-hour time point was chosen to estimate an increase of HMT. Literature indicates that four hours is an early yet sufficient time to observe mitochondrial uptake [22, 100–102].

The experiments were conducted by separating the cells into different plates to independently maintain the conditions for cells with and without UVR exposure.

Transwells

Transwell assays were set up by plating donor cells (MTR-labeled) in the upper chamber and recipient cells (CTG-labeled) in the lower chamber, separated by a 3 μm pore membrane (Transwell, Item: 3415; Corning Incorporated, Kennebunk, ME, USA). A p24-sized transwell was used to test mitochondrial transfer via CIMT under both non-UVR and UVR conditions, as previously described for the 2D co-cultures. A total of 22 h of interaction was allowed, with UVR exposure performed at the 18-hour mark for three minutes, followed by four additional hours to assess any changes post-UVR exposure. Three biological replicates were conducted, with three technical repetitions for each condition (control or UVR-exposed cells). For each technical repetition, three random fields were selected, and 10 representative images were captured per field to identify and count the total number of cells that had received mitochondria or had mitochondria firmly attached to their membranes. Verification of cell counts and mitochondria presence was performed at 20X or 40X magnification. For data analysis, the mean number of mitochondria-receiving cells from all captured images was calculated, resulting in three data points representing each repetition of each condition. In total, 27 data points were collected across the three biological replicates. The analysis was performed using GraphPad Prism version 10.3.1 for Windows, GraphPad Software, Boston, Massachusetts US.

Mitochondria isolation

Mitochondria isolation was performed according to Cabrera et al., 2019 [52]. The Mitochondria Isolation Kit for Tissue (Item: 89801; Thermo Fisher Scientific, Inc., Waltham, MA, USA) was used, according to the manufacturer’s guidelines, starting with 10–20 × 10⁶ labeled or not labeled donor cells, as some assays were performed without the need of observing the transferred fluorescent mitochondria. To ensure the extraction of highly purified mitochondrial isolates, a final recovery step involved centrifugation at 3,000 x g for 15 min. For additional purification, a second wash was conducted with the same centrifugation parameters to eliminate any residual Reactive C from the kit. The mitochondrial pellet was then resuspended in 1 mL of donor cell medium without serum. To quantify mitochondrial concentration, a Pierce™ Coomassie Plus (Bradford) Assay Kit (Item: 1856210, Thermo Fisher Scientific, Inc., Waltham, MA, USA) was employed. Typically, the yield of mitochondria ranged from 30 to 50 µg/mL. After isolation, the mitochondrial suspension was stored at 4 °C until subsequent use in AMT or MT experiments.

AMT/T by modified MitoCeption

AMT/T was conducted using an adaptation of the MitoCeption protocol developed by Caicedo et al. (2015) [54]. Specifically, the final centrifugation step between the co-incubated cells and donor mitochondria was omitted to better replicate the co-culture conditions [54]. Throughout the procedure, the thermal shock was consistently maintained. The quantity of mitochondria employed was determined based on the protein concentration derived from the Bradford estimation protocol. To achieve the desired concentration, mitochondria were appropriately diluted, aiming for a final volume of 50 to 100 µl. This diluted mitochondrial solution was then distributed in close proximity to the target cells. The choice of using 500 ng (0.5 µg) and 2,500 ng (2.5 µg) of mitochondria for transfer was based on our previous publications and assessments of the optimal quantity of mitochondria to induce cell proliferation in our laboratory [22, 52, 54, 97]. Further methodological details of the MT assays to mice and swine skin wounds are provided later in this section.

Scanning electron microscopy (SEM)

Once isolated, the mitochondria were fixed in 2.5% glutaraldehyde and subsequently washed with PBS. The samples were then dehydrated through a graded ethanol series, increasing from 50% to 100% concentration. Critical point drying was performed using the EM CDP300 (Leica) system, involving 12 exchange cycles until reaching 72.8 bar and 32 °C. Afterward, the samples were sputter-coated with an 80/20% gold-palladium mixture using the ACE200 (Leica) coater. The visualizations were carried out using a VEGA 3 scanning electron microscope (Tescan) at 30 kV.

TMRM-based mitochondrial membrane potential assay

Isolated mitochondria were be labeled with tetramethylrhodamine (TMRM, 100 nM) for 30 min at 37 °C. Following staining, the mitochondria were centrifuged at 16,000 g for 10 min at 4 °C, washed with 1X PBS, and centrifuged again under the same conditions. The staining were visualized using a Leica Sp8 confocal microscope equipped with a super-resolution module, maintained at 37 °C and 5% CO2.

Mitochondrial functional respirometry characterization

Mitochondria were isolated from 2 million murine MSCs using the Cell Mitochondria Isolation Kit (ThermoFisher Scientific, USA), following the manufacturer’s instructions. Oxygen consumption of the isolated mitochondria was assessed using an oxygraph chamber (Hansatech Oxigraph). Briefly, the mitochondrial pellet was re-suspended in Mitobuffer (pH 7.5), consisting of 50 mM sucrose, 5 mM HEPES, and 2 mM EGTA. Oxygen concentration and flow rate were recorded in the Mitobuffer solution in an open Oxigraph chamber for 10 min until a stable signal was obtained. A mitochondrial suspension with substrates (final concentrations: 5 mM glutamate, 5 mM pyruvate, 2 mM malate) was then added to the chamber, and oxygen flux was recorded for 5–7 min until a stable signal was achieved. This was followed by measurements of coupled respiration (1 mM ADP) and uncoupled cellular respiration (1 µM oligomycin), each recorded for 5–7 min.

ROS labeling

After AMT and UVR exposure, cells were labeled with a final concentration of 5 µM CellRox Green Reagent (Item: C10444; Invitrogen, Thermo Fisher Scientific, Inc., Waltham, MA, USA) for the detection of ROS associated with oxidative stress. The labeling process was conducted for 20 min, following the manufacturer’s instructions. Subsequently, cells were washed with PBS, and trypsin was added to detach all cells from the mitocepted wells. Cells were then centrifuged at 1,000 g for 5 min, and the resulting pellet was resuspended in 250 µl of PBS for flow cytometry analysis using the BD Accuri™ CSampler. Size and granularity parameters were utilized to identify the fibroblast population, and the Mean Fluorescent Intensity (MFI) emitted by the CellRox Green Reagent at 475 to 650 nm in the FL1 channel was measured within this population and converted to the fold for each data point. Five biological replicates, each with five technical repetitions per condition, were performed. The analysis was performed using GraphPad Prism version 10.3.1 for Windows, GraphPad Software, Boston, Massachusetts US.

Proliferation

20,000 cells were cultured in their standard medium and plated into P6 wells. On the following day, cells were washed with PBS, and culture media was replaced with medium containing 1% FBS and 1% PS. On day three, recipient cells, either mouse or human fibroblasts, were incubated with either 500 or 2,500 ng mitochondria isolated from 2 to 3 million donor cells, specifically either WJ- MSCs or mice BM MSCs. Following a 72 h incubation period, the cells were counted, and the fold-change of the experimental conditions was determined by comparing them to the control group. This fold-change analysis was utilized to assess the assays among the different cell types and sources of mitochondria. The 72 h time point was selected based on the time required for fibroblasts to grow and divide without inhibiting proliferation due to close cell contact and spatial limitations, as determined by our setup assays. Three biological replicates or assays were performed, with three technical repetitions for each condition. The analysis was performed using GraphPad Prism version 10.3.1 for Windows, GraphPad Software, Boston, Massachusetts US.

In vivo studies

Ethical statement

The methods employed in this research, including those used in mice and swine, were approved by the Bioethics Committee for the Use of Animals in Research and Teaching at the School of Veterinary Medicine of Universidad San Francisco de Quito (USFQ). Approval was granted in accordance with animal welfare standards and the principles of Replacement, Reduction, and Refinement (3Rs) in animal research, as proposed by Russell and Burch in 1959 [103]. The mouse assays were conducted under approval number 2020-001, whereas the swine assays were approved under number 2025-006. The methods used in this study were based on those previously described by Villagómez et al. in 2021 [85] and were adapted to the swine model with minor modifications.

The methods are reported based on the ARRIVE 2.0 guidelines [104]:

Study design

This study was designed to evaluate the effect of MT on cutaneous wound healing using minimally invasive surgical models intended to reduce animal harm, infection risk, and pain. Both small- and large-animal models were included to assess the regenerative potential of MT and its translational relevance.

In the murine model, allogeneic mitochondria isolated from BM-MSCs were administered at three doses: 25 ng, 50 ng, and 100 ng. Additional murine experimental groups received whole BM-MSCs at doses of 250,000, 500,000, and 1 × 10⁶ cells, allowing comparison between isolated mitochondrial treatment and intact cell administration. Healthy 6-week-old male Swiss white mice, weighing 23–25 g, were used for these assays. Control groups included untreated wounds and wounds treated with DMEM alone. The experimental unit for the murine assays was a single mouse. The wounds were analyzed at 48 h.

To further assess the translational potential of MT in a large-animal model, a xenogeneic swine wound-healing assay was performed using weaned piglets (Sus scrofa domestica) aged 21–31 days, with an approximate body weight of 6–9 kg. Animals were selected after veterinary examination to confirm good general health and the absence of pre-existing skin lesions. In this model, only isolated mitochondria derived from 2 × 10^6–2.5 × 10^6 human WJ-MSCs were applied to the wounds at doses of 25 ng, 50 ng, and 100 ng. No intact MSCs were administered in the swine model.

The rationale for this experimental design was to first determine, in a controlled murine setting, whether mitochondrial application to skin wounds could promote repair without inducing overt inflammatory or adverse local effects, while also comparing its performance with BM-MSC-based treatment. Although rodent models provide valuable preliminary evidence of regenerative activity, a large-animal model was incorporated because swine skin more closely resembles human skin in anatomical and physiological properties. Therefore, the swine model enabled evaluation of the wound-healing effects of isolated xenogeneic mitochondria in a system with greater translational relevance for human applications, while also allowing assessment of potential local immune or adverse responses following application of mitochondria derived from another species.

Overall, this design allowed evaluation of the dose-dependent effects of isolated mitochondria on wound-healing outcomes in both murine and swine models, while comparison with whole BM-MSC administration was performed specifically in the murine setting.

Sample size

Each experimental condition included three independent experimental replicates, with three animals per replicate, resulting in a total of nine animals per condition. For the murine assays, a total of 72 mice were used and allocated across eight experimental conditions: an untreated control group, a DMEM-treated control group, three groups treated with isolated mitochondria at doses of 25 ng, 50 ng, and 100 ng, and three groups treated with BM-MSCs at doses of 250,000, 500,000, and 1 × 10⁶ cells.

For the swine assays, a total of 45 weaned piglets were used, with nine animals allocated to each experimental condition. The swine groups included an untreated control group, a DMEM-treated control group, and three groups treated with isolated mitochondria at doses of 25 ng, 50 ng, and 100 ng.

The sample size was determined based on a power analysis informed by previous studies, including Villagómez et al. [85], which demonstrated that three animals per group per replicate were sufficient to detect biologically meaningful differences in wound-healing outcomes with 80% statistical power. The analysis considered the expected effect size, variability in the wound-healing response, and a significance level of α = 0.05.

The use of three independent replicates per condition, resulting in a minimum of nine animals per group, was intended to improve the robustness and reproducibility of the findings across experimental conditions. At the same time, the study design was aligned with the 3R principles by minimizing the number of animals used while preserving sufficient statistical power to obtain reliable and interpretable results.

Inclusion and exclusion criteria

The inclusion criteria were established a priori. The murine study included healthy, wild-type male Swiss white mice aged 6 weeks and weighing 23–25 g. The porcine study included healthy weaned piglets that underwent a veterinary examination before enrollment to confirm their general health and the absence of pre-existing injuries or skin lesions.

Animals presenting signs of illness, injury, abnormal behavior, or any other condition that could compromise their welfare or interfere with the experimental outcomes were excluded before the study. Once enrolled and assigned to an experimental group, no additional exclusion criteria were applied.

For the murine experiments, each experimental condition included nine mice, distributed across three independent experimental replicates, with three mice per replicate. This sample size was applied consistently to the untreated control, vehicle control, mitochondrial-treatment, and BM-MSC-treatment groups.

For the porcine experiments, each experimental condition included three pigs. This sample size was applied consistently across the untreated control, vehicle control, and mitochondrial-treatment groups.

Randomisation

Randomization was used to allocate mice and pigs to the control and treatment groups. Individual animals were selected without a predetermined pattern and assigned to the corresponding experimental conditions, including untreated control, vehicle control, mitochondrial transplantation, and, in the murine study, MSC administration. This approach was intended to minimize selection bias and promote a balanced distribution of animals across the experimental groups.

The order in which treatments were administered and outcome measurements were performed was also varied across animals to reduce the potential influence of procedural order or other systematic effects. All animals were maintained under the same environmental and husbandry conditions throughout the study to minimize external sources of variability.

Blinding

Blinding was not implemented during treatment administration or outcome assessment in the murine wound-healing experiments. In the porcine study, however, histological evaluation and the quantification of Ki67-positive cells were performed by investigators blinded to the experimental group assignments.

Outcome measures

The primary outcome measure was the early wound-healing response, assessed 48 h after injury using the WHI in both the murine and porcine models. The WHI was determined from histological sections of tissue collected from the wound sites and provided an integrated assessment of granulation tissue formation, inflammatory-cell infiltration, collagen deposition and organization, and the relative abundance of early and mature collagen, following the scoring approach described by Villagómez et al. (2021) [85]. Murine tissue sections were stained with hematoxylin and eosin (H&E) and Mallory’s trichrome, whereas porcine tissue sections were evaluated using H&E and Masson’s trichrome staining. Because wounds were analyzed at a single 48-h endpoint, the WHI was interpreted as a measure of the advancement of the early repair phase rather than complete wound closure or long-term tissue regeneration.

Secondary outcome measures included qualitative and quantitative histological assessment of tissue architecture, inflammatory-cell infiltration, granulation tissue formation, and collagen-containing connective-tissue organization. Representative stained sections were also examined to identify treatment-associated changes in wound structure and tissue integrity.

In the porcine model, cellular proliferation was additionally evaluated by Ki67 immunohistochemistry. Ki67 positivity was quantified separately within predefined epidermal and dermal compartments. On each side of the wound, the wound zone comprised the 600-µm region extending outward from the wound tip, including the wound-edge and newly formed epidermal regions together with the underlying dermis. The extreme region comprised the subsequent 300-µm segment located distal to the wound zone and represented tissue farther from the injury. The central wound bed was evaluated as a separate dermal compartment. Epidermal Ki67 positivity was quantified in the wound-zone and extreme regions, whereas dermal Ki67 positivity was assessed in the wound zone, extreme region, and central wound bed. This compartment-based strategy was adapted from quantitative immunohistochemical approaches that evaluate epidermal and dermal wound regions separately and exclude appendage-containing areas to avoid confounding the proliferative-cell counts.

Areas containing hair follicles, sebaceous glands, or other skin appendages were excluded because these structures contain physiologically proliferative cell populations that could bias Ki67 measurements. For each anatomical region, Ki67 positivity was calculated as the percentage of Ki67-positive nuclei relative to the total number of nuclei. At least three non-overlapping histological areas were quantified per region and per wound. These measurements were averaged to obtain a single value for each anatomical region in each animal, with the individual pig, rather than each microscopic field, considered the experimental unit. Histological evaluation and Ki67 quantification in the porcine study were performed by investigators blinded to treatment allocation.

Statistical methods

Data analysis was performed using GraphPad Prism version 10.3.1 for Windows (GraphPad Software, Boston, Massachusetts, USA). The WHI was used to quantify tissue repair, and data were presented as mean ± standard deviation (SD). Normality was assessed using the D’Agostino & Pearson test. For non-normally distributed data, non-parametric tests were applied, including the Mann-Whitney U test (****p < 0.0001) for pairwise comparisons and the Kruskal-Wallis test (**p < 0.01) for comparing multiple groups.

Experimental procedures

Murine wound model

For each experimental condition, including the control groups, a 1-cm linear incision was surgically created on the dorsal skin of each mouse under anesthesia. Animals were anesthetized with xylazine (0.5 mg/kg) and ketamine (80 mg/kg), following the procedure described by Villagómez et al. (2021) [85]. The incision was subsequently closed using Monocryl^® sutures. Treatment preparations were formulated in a final volume of 100 µL of DMEM. The mitochondrial-treatment groups received 25, 50, or 100 ng of mitochondrial protein isolated from mouse BM-MSCs. For the cellular-treatment groups, 2.5 × 10^5, 5 × 10^5, or 1 × 10^6 BM-MSCs were resuspended in the same final volume. The vehicle-control group received 100 µL of DMEM without mitochondria or cells, whereas the untreated control group received no local injection. Each preparation was administered once, immediately after wound closure, through four 25-µL intracutaneous injections placed adjacent to the wound margins, with two injections administered on each side of the incision. Injections were performed using a 0.3-mL insulin syringe fitted with an 8-mm, 31-gauge needle. The mice were monitored for 48 h after treatment and were then euthanized for collection of the wound-containing tissue and subsequent histological analysis.

Porcine wound model

Before surgery, the pigs were sedated and anesthetized using a 2% xylazine formulation in combination with ketamine administered at 5 mg/kg. The same anesthetic protocol was applied in both experimental phases to reduce stress and provide adequate sedation, muscle relaxation, and analgesia during the procedure.

Following aseptic preparation of the dorsolumbar region, a 1-cm linear incision was created in the dorsal skin of each animal and closed using Monocryl^® sutures. Freshly isolated mitochondria derived from well-characterized human WJ-MSCs were resuspended in DMEM at total mitochondrial protein doses of 25, 50, or 100 ng in a final volume of 100 µL. The vehicle-control group received 100 µL of DMEM without mitochondria, whereas the untreated control group received no local injection. Intact MSCs were not administered in the porcine model. As in the murine model, each preparation was administered once, immediately after wound closure, through four 25-µL intracutaneous injections placed as close as possible to the wound margins. Two injections were administered on each side of the incision, resulting in a total injection volume of 100 µL per wound. Each animal was assigned to a single experimental condition, and one wound was analyzed per pig. The animals were monitored for 48 h after treatment under veterinary supervision. At the experimental endpoint, wound-containing tissue samples were collected for WHI determination, histological evaluation using hematoxylin and eosin and Masson’s trichrome staining, and spatial quantification of Ki67-positive cells in predefined epidermal and dermal compartments.

Acclimatization periods

Mice were housed in the animal facility of Universidad San Francisco de Quito (USFQ) and acclimatized for at least 7 days before the experimental procedure. During this period, the animals were monitored to confirm their general health and adaptation to the housing environment. Mice were maintained under standard laboratory conditions, including controlled temperature and humidity, a 12-h light/12-h dark cycle, and access to food and water according to institutional animal-care practices. All murine surgical procedures were performed in the animal surgery facility at USFQ. The porcine experiments were conducted in three independent facilities located south of Quito: a farm managed by Universidad de las Fuerzas Armadas (ESPE), a farm in Chaupi, and a farm in Panzaleo. At each location, the piglets were housed under conditions appropriate for young pigs. Pens included a heated area maintained at approximately 30–34 °C and provided approximately 3.5–4 m² of usable space per litter or group, allowing the animals to move and rest comfortably. The pens had concrete floors designed to facilitate liquid drainage and routine cleaning. Dry bedding composed of straw and wood shavings was provided and replaced or cleaned daily to maintain appropriate hygiene and comfort. Adequate ventilation and air exchange were ensured through well-distributed, covered windows that protected the animals from direct exposure while allowing continuous airflow. Animals were monitored throughout the housing and experimental periods for signs of illness, discomfort, or abnormal behavior.

Rationale for procedures

The experimental procedures were designed to evaluate the regenerative potential of MT in cutaneous wound healing and, in the murine model, to compare its effects with those of intact BM-MSCs. The linear, primary-intention wound model was selected to reproduce a standardized surgical injury while minimizing tissue loss, animal discomfort, and the risk of infection. Local intracutaneous administration adjacent to the wound margins was used to deliver the treatments directly to the tissue undergoing repair. The WHI, as described by Villagómez et al. (2021) [85], provided a standardized histological assessment of early tissue repair, including inflammatory-cell infiltration, granulation tissue formation, and collagen deposition and organization.

Following surgery, mice received analgesic treatment for pain management and were monitored regularly for signs of distress, abnormal behavior, impaired mobility, wound complications, or other adverse events. At 48 h after treatment, the mice were euthanized by cervical dislocation in accordance with the approved institutional protocol and applicable animal-welfare guidelines. Euthanasia was performed by trained personnel under veterinary supervision, in compliance with the requirements of the USFQ Animal Research Committee and the principles of Replacement, Reduction, and Refinement.

In the porcine experiments, the animals were maintained under anesthesia during wound creation and treatment administration. Analgesia was provided after the procedure, and the wounds were disinfected and monitored for signs of infection, inflammation, dehiscence, pain, or other complications. At the 48-h endpoint, the pigs were anesthetized again to allow collection of the wound-containing tissue using a biopsy punch without causing avoidable pain or distress. The biopsy sites were subsequently disinfected and treated according to veterinary recommendations. The pigs were not euthanized as part of the study and, after postoperative recovery and veterinary evaluation, remained under the care of the farm owners and returned to routine farm management activities.

Histological analysis and Ki67 immunohistochemistry

Wound-containing tissue samples collected from murine and porcine models, fixed, processed, embedded in paraffin, and sectioned for histological and immunohistochemical analysis. Sections were mounted on electrostatically charged microscope slides (Leica Biosystems) to improve tissue adherence. Before staining, the slides were heated at 60 °C for 30 min to promote adhesion of the tissue sections and reduce the risk of detachment during subsequent processing.

Routine histological evaluation was performed using hematoxylin and eosin (H&E) staining to assess overall tissue architecture, inflammatory-cell infiltration, granulation tissue formation, and re-epithelialization. Masson’s trichrome staining was used to evaluate collagen-containing connective tissue, collagen deposition, and tissue organization within the wound.

Ki67 immunohistochemistry was performed using a BOND-MAX automated staining platform (Leica Biosystems). Commercial reagents supplied by the manufacturer were maintained under controlled refrigerated conditions at approximately 5 °C and included BOND Dewax Solution for deparaffinization, BOND Epitope Retrieval Solution 1 at pH 6.0 for antigen retrieval, BOND Wash Solution, and the BOND Polymer Refine Detection system.

Ki67 was detected using a mouse monoclonal anti-Ki67 antibody, clone K2, at a concentration of 1 mg/L. The automated staining protocol consisted of heat-induced epitope retrieval with BOND Epitope Retrieval Solution 1 at 95 °C for 20 min, followed by incubation with the primary antibody for 40 min. Sections were then sequentially incubated with the BOND Primer amplification reagent for 15 min and BOND Polymer for 15 min. Chromogenic development was performed using the BOND Polymer Refine 3,3′-diaminobenzidine (DAB) detection system for 16 min. Nuclear counterstaining was carried out with hematoxylin for 18 min, after which the slides were washed and permanently mounted.

Image acquisition and spatial quantification of Ki67

Histological and immunohistochemical sections were examined using a ZEISS Primostar 3 microscope equipped with an integrated digital camera. Images were acquired under standardized magnification and illumination conditions. Ki67-positive and total nuclei were quantified using ZEISS Labscope software.

Ki67 expression was analyzed separately in predefined epidermal and dermal compartments. On each side of the wound, the wound zone was defined as the 600-µm region extending outward from the wound tip and included the wound-edge and newly formed epidermal tissue together with the underlying dermis. The extreme region comprised the subsequent 300-µm segment located distal to the wound zone and represented tissue farther from the injury. The central wound bed was evaluated as an independent dermal compartment.

Epidermal Ki67 positivity was quantified in the wound-zone and extreme regions, whereas dermal Ki67 positivity was evaluated in the wound-zone, extreme, and central wound-bed regions. Areas containing hair follicles, sebaceous glands, or other skin appendages were excluded because these structures contain physiologically proliferative cells that could confound the analysis. For each compartment, Ki67 positivity was calculated using the following formula:

At least three non-overlapping microscopic fields were quantified for each anatomical region in each wound. Values obtained from the different fields were averaged to generate a single animal-level value for each compartment. Histological evaluation and Ki67 quantification were performed by investigators blinded to the experimental treatment groups.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

AC conceived the study, designed the overall experimental strategy, supervised the project, and coordinated the integration of the in vitro and in vivo studies. AC and ABA performed the core coculture, transwell, artificial mitochondrial transfer, and mitochondrial transplantation experiments. MBA, ABA, AAPM, AHV, PR, PA, DT, DB, GZ, and AC conducted and analyzed the 2D coculture and transwell experiments examining mitochondrial-transfer dynamics among skin cells under basal conditions and following UVR-induced oxidative stress. SP contributed to the coculture experiments and the assays presented in Supplementary Figure 1. KZ, SCM, and CVC independently repeated selected experiments to assess the reproducibility of the observed mitochondrial-transfer dynamics. AC, LFS, and TMG performed and analyzed the fibroblast-proliferation assays following AMT/T. AC, GZ, PLC, IMM, MIM, VAB, and MK contributed to cell isolation and provided expertise in the culture and coculture of MSCs, fibroblasts, and other relevant cell populations. AC, BA, and ENL optimized and validated the procedures used to assess the structural integrity, viability, and functionality of isolated mitochondria before transfer. SC, LF, BA, PLC, ADC, and ENL performed cell-culture experiments using human and mouse MSCs and fibroblasts, contributed to the optimization of AMT/T protocols, and supported the evaluation of mitochondrial structure and function. LFS, PA, DT, BA, GZ, DB, LF, ADC, KZ, SC, DV, AAPM, DFCH, PMA, FC, PLC, IMM, MIM, and MK provided methodological support, data interpretation, and assistance with the integration of the in vitro and in vivo findings and the preparation of the final figures. AC, ABA, AV, ML, ASU, SDPS, MS, TB, MBA, AHV, GS, PP, FT, GD, DS, PR, PA, DV, DFCH, PMA, FC, IMM, and RFD contributed to the design, optimization, and execution of the murine and porcine wound-healing studies. Their contributions included animal handling and health monitoring, surgical procedures, mitochondrial and MSC administration, postoperative care, tissue collection, histological processing, image acquisition, quantitative image analysis, data analysis, and interpretation of the therapeutic effects of mitochondrial transplantation and MSC administration. AC wrote the original draft of the manuscript. TB, GS, PP, GD, PMA, FC, PLC, IMM, MIM, MK, ENL, VAB, RFD, and TMG made substantial contributions to the critical review, scientific interpretation, and revision of the manuscript. All authors reviewed, edited, and approved the final version of the manuscript.

Funding

This work was supported by the Escuela de Medicina at the Colegio de Ciencias de la Salud COCSA, Universidad San Francisco de Quito (USFQ). Funding was provided by the Air Force Office of Scientific Research (AFOSR) under award/project number FA9550-20-1-0407 and by the Corporación Ecuatoriana para el Desarrollo de la Investigación y Académica, CEDIA, through the CEPRA XIV-2020-04, MITOCHONDRIAS project.

Data availability

All relevant data supporting the findings of this study are available from the corresponding author upon request to acaicedo@usfq.edu.ec.

Declarations

Ethics approval and consent to participate

This study involved the use of established and primary human-derived cells and cell lines, together with in vivo murine and porcine models. The ethical review and approvals applicable to each component are detailed below.

Consent for publication

Not applicable.

Human-derived cells and cell lines

No human biological material was prospectively collected directly from patients or donors by the investigators for the purposes of this study. The human-derived cells and cell lines were obtained from recognized commercial suppliers or collaborating academic and biomedical institutions, as described in the Methods section. The provenance and supporting documentation associated with these materials—including, where applicable, their original collection, informed-consent procedures, authorization for research use, institutional transfer, storage, and permitted secondary use—were submitted to and evaluated by the Research Ethics Committee on Human Subjects of Universidad San Francisco de Quito (CEISH-USFQ). For the primary human fibroblasts originally isolated from neonatal foreskin, the documentation confirmed that the tissue had been collected from healthy donors after informed consent was obtained from their parents or legal guardians. Following its review, the CEISH-USFQ determined that no additional ethical approval was required for the use of these materials in the present study (Document No. 065IN-2021-CEISH-USFQ, dated May 7, 2021). The ethical assessment was conducted in accordance with the applicable Ecuadorian regulatory framework and the international ethical principles governing research involving human-derived biological materials. These include the principles of the World Medical Association’s Declaration of Helsinki, the WMA Declaration of Taipei on Ethical Considerations regarding Health Databases and Biobanks, and the Council for International Organizations of Medical Sciences’ International Ethical Guidelines for Health-related Research Involving Humans. These frameworks require appropriate documentation of the origin and authorized use of human biological materials, protection of donor dignity, autonomy, privacy, and confidentiality, and independent ethical evaluation of their storage, transfer, and secondary research use. Based on the absence of direct participant recruitment, intervention, or prospective sample collection by the study investigators, together with the satisfactory review of the materials’ provenance and supporting documentation, the CEISH-USFQ identified no ethical or regulatory impediment to their use in this research.

Animal studies

All experimental procedures involving mice and swine were reviewed and approved before study initiation by the Bioethics Committee for the Use of Animals in Research and Teaching of the School of Veterinary Medicine at Universidad San Francisco de Quito (USFQ). The murine experiments were conducted under approval No. 2020-001, and the porcine experiments were conducted under approval No. 2025-006. The studies were conducted in accordance with the applicable Ecuadorian regulatory framework governing the use of animals in research, including the Ley Orgánica de Sanidad Agropecuaria, its General Regulations, and the applicable regulations issued by the Agencia de Regulación y Control Fito y Zoosanitario (AGROCALIDAD). This framework requires prior ethical approval of research involving live animals and establishes standards for their rational, ethical, and humane use.

The protocols were also evaluated in accordance with the International Guiding Principles for Biomedical Research Involving Animals, developed by the Council for International Organizations of Medical Sciences and the International Council for Laboratory Animal Science (CIOMS–ICLAS). These principles constitute an international ethical framework for animal research, recognize animals as sentient beings, and require the principles of Replacement, Reduction, and Refinement to be incorporated into the design and conduct of scientific studies.

The studies additionally followed the animal-welfare recommendations of the World Organisation for Animal Health (OIE/WOAH), particularly those contained in the Terrestrial Animal Health Code—Use of Animals in Research and Education. The AGROCALIDAD framework expressly incorporates the CIOMS principles and OIE/WOAH standards as ethical references for research involving animals.

Directive 2010/63/EU of the European Parliament and of the Council on the protection of animals used for scientific purposes was also used by the USFQ Bioethics Committee as a complementary international benchmark for ethical assessment. Although this Directive is not the governing legislation in Ecuador, it provides internationally recognized standards concerning the mandatory application of the 3Rs, animal housing and care, anesthesia and analgesia, classification of procedural severity, humane endpoints, and the prevention or minimization of pain, suffering, distress, and lasting harm.

Before approval, the protocols underwent a prospective harm–benefit assessment in which the anticipated scientific and potential medical value of the research was weighed against the possible harm to the animals. The ethical review considered the scientific justification for using animals, the appropriateness of the selected species, the number of animals required, the statistical design, the availability of non-animal alternatives, the severity and duration of the procedures, and the measures established to protect animal welfare. The Ecuadorian framework specifically requires an evaluation of risks and benefits and the justification of the species, age, sex, weight, and number of animals included in a study.

The principles of Replacement, Reduction, and Refinement were applied throughout the study. Alternative methods were considered whenever appropriate; the number of animals was limited to the minimum necessary to obtain scientifically valid and interpretable results; and the procedures were refined to prevent or minimize pain, distress, discomfort, and lasting harm.

All experimental procedures were performed by appropriately trained personnel or under veterinary supervision. Suitable anesthesia, analgesia, postoperative care, monitoring procedures, and predefined experimental or humane endpoints were implemented according to the approved protocols. Animals were treated as sentient beings, and all reasonable measures were taken to preserve their health and welfare and to avoid unnecessary suffering. The study and its animal procedures are reported in accordance with the ARRIVE 2.0 guidelines.

Competing interests

AC is the scientific founder and advisor of Dragon Biomed, entrepreneurial initiative at Universidad San Francisco de Quito (USFQ) institute of Biomedical Research iBioMed. He also serves as a scientific advisor to the Research and Development Department at Luvigix and provides scientific advisory support to the Super Space Foods Research Institute. In these roles, AC provides scientific guidance and expertise but does not participate in operational management or decision-making processes. AHV is a research trainee at Luvigix and has no decision-making authority within the company. MK is the Chief Scientific Officer of Cells for Cells and REGENERO. All other authors declare no competing interests.