Fate of systemically and locally administered adipose‐derived mesenchymal stromal cells and their effect on wound healing

Abstract There is increasing interest in the use of adipose‐derived mesenchymal stromal cells (ASCs) for wound repair. As the fate of administered cells is still poorly defined, we aimed to establish the location, survival, and effect of ASCs when administered either systemically or locally during wound repair under physiological conditions. To determine the behavior of ASCs, a rat model with wounds on the dorsal aspect of the hind paws was used and two treatment modes were assessed: ASCs administered systemically into the tail vein or locally around the wound. ASCs were transduced to express both firefly luciferase (Fluc) and green fluorescent protein to enable tracking by bioluminescence imaging and immunohistological analysis. Systemically administered ASCs were detected in the lungs 3 hours after injection with a decrease in luminescent signal at 48 hours and signal disappearance from 72 hours. No ASCs were detected in the wound. Locally administered ASCs remained strongly detectable for 7 days at the injection site and became distributed within the wound bed as early as 24 hours post injection with a significant increase observed at 72 hours. Systemically administered ASCs were filtered out in the lungs, whereas ASCs administered locally remained and survived not only at the injection site but were also detected within the wound bed. Both treatments led to enhanced wound closure. It appears that systemically administered ASCs have the potential to enhance wound repair distally from their site of entrapment in the lungs whereas locally administered ASCs enhanced wound repair as they became redistributed within the wound bed.


| INTRODUCTION
In the field of regenerative medicine, cell-based therapies utilizing human mesenchymal stromal cells (MSCs) have generated a great deal of interest for wound healing. [1][2][3] The exact mechanism underlying the therapeutic benefits of MSCs has yet to be fully elucidated. Their role in tissue maintenance, repair, and regeneration is suggested to stem from their multipotent differentiation capacity, immunomodulatory properties, antimicrobial effect, paracrine signaling, and their ability to migrate to sites of injury. 2,[4][5][6] Although MSCs were originally isolated from bone marrow (BM), 7 a source used with increasing frequency is subcutaneous adipose tissue. 4 Both BM-MSCs and adipose-derived MSCs (ASCs) are candidate cell types for treating wounds; laboratory and clinical studies, however, favor BM-MSCs. 3,8 ASCs are considered to be a potentially useful cell population because of their availability, ease of harvest, high cell yield, and ability to be expanded in culture for clinical use. 4,[9][10][11] MSCs are believed to migrate to damaged tissue in a process called homing. 12 In vitro, MSCs are attracted by cytokines and chemokines that are known to be upregulated under conditions of inflammation. [13][14][15] However, in vivo, homing of MSCs has not been unequivocally shown. Some studies have reported homing of systemically administered MSCs in mouse models of type 2 diabetes, 16 traumatic brain injury, 17 and burn injury, 18 whereas others reported intrapulmonary cell trapping 19 and minimal evidence for homing. 20 Where recirculation of MSCs was shown to occur after becoming trapped in the lungs, only a low percentage of MSCs could be found at the injury site or in non-target organs. [20][21][22][23] Most of these studies evaluated the homing of BM-MSCs. Although BM-MSCs and ASCs share biological characteristics, differences in their phenotype, differentiation potential, gene expression profile, and immunomodulatory activity have been noted. 11,24,25 Thus, differences in their homing capacity may also exist, establishing a need to investigate the homing of ASCs and compare this to available BM-MSCs homing data.
Commonly suggested routes of MSC transplantation include systemic and local injection. 19,[26][27][28][29] Local injection relies on the activity of locally transplanted MSCs at the target site. Systemic injection requires circulating MSC to exit the blood vessels and to migrate to the target site. However, whether administered systemically or locally, where the cells go to and whether they can bind, engraft and even survive in vivo, is still not well understood, particularly for ASCs.
Insights into the basic biology of administered ASCs, their biodistribution and their engraftment into damaged tissue, need to be understood to achieve clinical efficiency.
In this study, we evaluate the distribution and survival of ASCs in vivo when administered either systemically or locally using a standardized rat model of cutaneous wound repair under physiological conditions. Defining the location of administered ASCs in vivo is complicated by the lack of a single identifying marker. In vitro, ASCs are characterized by their co-expression of a panel of surface markers; however, this is not possible to achieve in vivo. 30 Molecular imaging offers the opportunity to track cells in intact organisms. Fluorescent reporters, such as green fluorescent protein (GFP), do not give a sufficiently strong signal to allow for efficient detection in vivo. 31 A useful method for in vivo cell tracking in small animal models is bioluminescence imaging (BLI). [32][33][34] BLI allows for non-invasive imaging over a longer period and provides information on the viability 35 of cells after transplantation. In this study, we used both BLI and GFP to enable in vivo tracking and post-mortem identification, respectively.
The therapeutic potential of MSCs for cutaneous wound repair is being investigated in clinical trials for burn wounds, 36,37 diabetic foot ulcers, [37][38][39] and critical limb ischemia. 40,41 Although preclinical and clinical studies using MSCs to enhance wound repair have shown positive outcomes, the benefit of MSC administration for wound healing has not yet been clearly established. 39,[42][43][44][45] As a secondary aim, we evaluated the effect of different modes of ASC administration on wound closure (WC) time. Information is provided on the biodistribution of ASCs and their tissue-specific location post administration, as well as their effect on WC time. By understanding the fate of administered ASCs in response to injury, this study contributes to establishing the optimal mode of administration required in the clinical setting.

| MATERIALS AND METHODS
All animal experiments were approved by the local veterinary authority (authorization number GE/70/18).

Significance statement
The exogenous administration of adipose-derived mesenchymal stromal cells (ASCs) holds promise as a treatment strategy for wound healing by promoting tissue repair and regeneration. However, the best route of administration is still not well defined. Herein, this article describes the biodistribution and survival of systemically vs locally administered ASCs. Interestingly, both routes of administration led to enhanced wound repair as seen by earlier wound closure.
Systemically administered ASCs have the potential to enhance wound repair distally from their site of entrapment in the lungs, whereas locally administered ASCs migrate into the wound bed.

| Transduction of ASCs
ASCs were transduced with a dual lentivector expressing GFP and firefly luciferase (Fluc), pCWX-UBI-Fluc-PGK-GFP. To determine the amount of lentivector needed to transduce greater than 70% of the cells, a multiplicity of infection (MOI) of 0, 2, 5, and 10 was tested (n = 4). A MOI of 10 was used for all further experiments. ASCs at P1/P2 were plated at 5 × 10 3 cells/cm 2 and allowed to adhere for 24 hours. Lentivectors were added and the cultures left for 72 hours before replacing the medium with fresh CGM. At 80% confluence, ASCs were trypsinized and an aliquot prepared for flow cytometric analysis (n = 6) as described below to determine their immunophenotype and the percentage of ASCs expressing GFP. To determine whether ASCs also expressed Fluc,

| In vitro differentiation capacity of ASCs
To confirm the differentiation capacity of non-transduced and transduced ASCs, they were induced to differentiate along adipogenic and osteogenic lineages using specific induction media. Cell cultures were trypsinized, counted, phenotypically characterized, and replated into

| In vivo rat model and ASC administration
Female Wistar rats (≥10 weeks old, ≥250 g) were fed a standard diet and given water ad libitum. For all procedures, animals were anesthetized by inhalation of 3%-5% isoflurane (Baxter AG, Opfikon, Switzerland). Wounds on the dorsal aspect of the hind paws were created bilaterally by removing a full-thickness skin area of 1.2 cm × 0.8 cm.
One day after wounding, 27 animals received a single injection of 2 × 10 6 Fluc-GFP positive ASCs systemically into the tail vein and 21 animals locally with 2 × 10 5 Fluc-GFP positive ASCs in total, split between two sides of each wound (proximally at the ankle and distally at the toes). One day post wound creation was chosen for administration so that the inflammatory phase would already be initiated to attract the ASCs. Moreover, by postponing the injection of the ASCs, we avoided the probability of the cells being washed away due to bleeding immediately after wound creation. For each experiment, ASCs isolated from two rats were cultured and transduced separately before being pooled for either systemic or local administration. Nine control animals received a single injection of sodium chloride (NaCl) systemically into the tail vein. Both feet were evaluated separately for each animal. All animals received a semi-occlusive wound dressing that was changed at every evaluated time-point ( Figure 1).

| In vivo tracking of administered ASCs by BLI
Animals treated systemically and locally were imaged by BLI at 3 hours, 24 hours, 48 hours, 72 hours, d7, d15, and at WC. Prior to imaging, animals received a single IP injection of firefly D-luciferin potassium salt (150 mg/kg body weight) prepared in PBS. After 20 minutes, imaging was performed using the Xenogen IVIS spectrum in vivo imaging system with an XGI-8 Gas anesthesia system. A photographic image of the animal followed by a luminescent image was recorded by the camera. To exclude background luminescence emitted from the ASCs themselves, a single animal that received non-transduced ASCs at the same concentration and mode of administration was included as a control for every BLI experiment. For quantification, three regions of interest (ROI) were manually selected to quantify the luminescent signal in the lung (ROI 1) and the wound area (right foot, ROI 2; left foot ROI 3). The ROI size was kept constant between animals and the intensity of the luminescent signal was recorded as total flux (average photons per second, p/s). 50 All images were analyzed using the Living Image 4.3.1 software.

| Histological assessment of GFP-positive ASCs
For systemically and locally treated groups, the lungs and the entire foot with the wound and the surrounding uninjured skin were harvested at 3 hours, 24 hours, 72 hours, day 7 (d7), d15, and at WC and fixed in 4% formaldehyde (Merck Millipore, Massachusetts). After fixation, the wound area along with the uninjured surrounding skin was removed from the foot and processed for paraffin embedding ( Figure 2A). The lungs were kept whole and processed for paraffin embedding.

| Assessment of wound repair
The time taken to complete WC and the contraction/re-epithelialization ratio were assessed as previously described. 51 Wound size was documented immediately after wounding and until WC by digital photography at a constant distance with a ruler next to the wound for scaling.
The photos were analyzed using ImageJ 1.48v software 52 to determine the wound area. At complete WC (ie, full epithelialization), the surface of the scar with hairless skin was measured and considered to correspond to the area of the wound that had healed by re-epithelialization.
The surface of the wound that healed by contraction was estimated by subtraction of the epithelialized surface from the original wound area measured at day 0 ( Figure 7C).

| Statistical analysis
Data are expressed as means ± SD. Statistical analyses were carried out using Prism (GraphPad, California). A Mann-Whitney test was used to assess differences between data means. Grouped data were assessed by performing multiple t tests using the Holm-Sidak method.
A P value of <−.05 between data means was considered significant.

| Transduced ASCs maintained their immunophenotype and differentiation capacity
Isolated ASCs were plastic adherent in culture and displayed a typical fibroblast morphology ( Figure 3A). Evaluation of GFP expression established that the lowest MOI needed to transduce at least 70% of the ASCs was 10 ( Figure 3B). As the Fluc gene is under the control of a different promoter than the GFP gene, we determined whether our newly transduced cells also expressed Fluc. When D-luciferin, the molecular substrate for firefly luciferase, was added to the transduced cells in vitro, emitted luminescence was detected. An increase in luminescence, quantified as total flux, was observed with increasing MOIs F I G U R E 2 Schematic showing how the locally treated feet samples were processed, sectioned, stained for IHC, and analyzed. A, The feet were collected, fixed, and then processed for embedding in paraffin blocks. B, Each block, consisting of a single sample, was sectioned at five different levels into the FFPE tissue at levels 0, 1, 2, 3, and 4. IHC was performed to detect GFP + ASCs (red staining, AEC). C, The sections were analyzed to quantify the area of positive staining using image processing software. Each section was divided into six zones, where zone 0 = ASCs 1 (ASC injection site 1), zone 1, 2, 02, and 01 = wound divided into four equal regions, and zone 00 = ASCs 2 (ASC injection site 2). The summed area of positive staining in each of the different zones over the five levels and for both tissues was quantified to obtain a value per foot. This value was normalized to the summed area for all the levels, zones, and tissues per foot.

| Locally administered ASCs survived and became distributed within the wound bed
When administered locally, the luminescent signal remained strongly detectable for 7 days at the injection site ( Figure 4C,D) with no signal being detectable in the lungs ( Figure 4C). Interestingly, histological analysis of animals that received ASCs locally revealed that GFP staining was not restricted to the ASC injection sites. When 15% of the wound was analyzed as represented by the five levels at which sections were assessed, GFP positive ASCs became distributed within the wound bed as early as 24 hours after administration ( Figure 6A-E). At the injection site (zone 0, 00), ASCs remained detectable until d7 ( Figure 6B). Within the wound bed (zone 1, 2, 02, and 01), a significant increase in GFP staining was found from 24 to 72 hours (0.039% vs 0.057%, P = .0087) with no significant change on d7 ( Figure 6C). This significant increase in GFP staining from 24 to 72 hours was found in both the outer (zone 1 and 01; 0.031% vs 0.047%, P = .0043) and inner wound sites (zones 2 and 02; 0.010% vs 0.033%, P = .0173). However, even though the increase was significant in the inner wound site between 24 and 72 hours (0.013% vs 0.043%, P = .0173), and 48 and 72 hours (0.014% vs 0.043%, P = .0152), GFP staining remained lower than in the outer wound site. The outer wound site showed no significant changes in GFP staining from 72 hours to d7, whereas in the inner wound site a F I G U R E 4 In vivo tracking of systemically and locally administered Fluc-GFP + ASCs during wound repair under physiological conditions. ASCs transduced to express both Fluc and GFP were injected either systemically (2 × 10 6 ASCs) via the tail vein or locally (2 × 10 5 ASCs) into two sides around each wound in rats 24 hours after bilateral full-thickness wounds had been created on the dorsal aspect of the feet. Prior to imaging, rats received 150 mg/kg body weight of D-luciferin. A single representative image of a, A, systemically and, C, locally treated rat and BLI control rat (received non-transduced cells) imaged at 3, 24, 48, and 72 hours. An enlarged area of the feet of the treated animals is also shown. The luminescence signal in the, B, lungs of systemically treated and the, D, feet of locally treated rats was quantified and recorded as total flux over time (gray bars). Data shown as mean ± SD. Abbreviations: ASCs, adipose-derived mesenchymal stromal cells; BLI, bioluminescence imaging; Fluc, firefly luciferase; GFP, green fluorescent protein significant decrease was found from 72 hours to d7 (0.043% vs 0.010%, P = .0152; Figure 6D,E).

| Both systemic and local administration lead to enhanced WC time
As a secondary aim, we determined whether ASC administration might lead to a change in WC time. In comparison to a total WC time of 26 days in the NaCl treated group, WC was significantly faster when ASCs were administered locally (19 days, P = .0108) and systemically (21 days, P < .0024; Figure 7A). Both ASC treated groups showed an increase in wound size after 3 days before reducing in size. At 7 days, wound size was significantly smaller in the control group compared with the ASC systemically treated group (P = .03013); however, this was not maintained. Comparison of ASC treated groups showed significantly smaller wounds with local treatment at 10 days (P = .0102; Figure 7B).  and led to improvement in growth rates of the children as well as their ability to synthesize intact bone. [53][54][55] A study in rats showed that BM-MSCs administered either systemically or locally were able to promote fracture healing. 56 Although the proportion of BM-MSCs that migrated into the injury sites following systemic injection was significantly less compared with locally injected BM-MSCs, both groups promoted fracture healing equally well, indicating that systemically injected BM-MSCs may have contributed in an indirect manner. 56 We set out to compare the fate and distribution of systemically and locally administered ASCs in an in vivo model of physiological wound repair. The first concern with using transduced ASCs was that this could change their characteristics. Two parameters currently used to identify MSCs were assessed, namely their immunophenotype and differentiation capacity. 30,57 Transduction of ASCs did not change their immunophenotype or differentiation capacity compared with non-transduced ASCs and transduced ASCs maintained GFP expression in vitro over several passages. This confirmed previous studies showing that MSCs transduced to express GFP 58,59 or Fluc 33 maintain F I G U R E 5 Detection of systemically administered ASCs during physiological wound repair by IHC. ASCs transduced to express both Fluc and GFP were injected systemically (2 × 10 6 ASCs) via the tail vein in rats 24 hours after bilateral fullthickness wounds had been created on the dorsal part of the feet. Lungs and feet were processed for IHC to detect GFP positive ASCs (red staining, AEC). A, Images of the lung at 3 hours, 24 hours, 48 hours, 72 hours, and at d7 are shown. The right panel is a magnified view of the black rectangle showing positively stained ASCs. B, A representative image of the feet is shown where panel (i) is a magnified view of the epithelium lip and (ii) of the wound bed. Tissue sections were counterstained with hematoxylin (blue staining). Abbreviations: AEC, aminoethyl carbazole; ASCs, adiposederived mesenchymal stromal cells; Fluc, firefly luciferase; GFP, green fluorescent protein; IHC, immunohistochemistry their characteristics in vitro. Here, we report that by using a single lentivector, both GFP and Fluc can be used in combination without changing the characteristics of ASCs.
Systemically administered ASCs showed no detectable homing to the site of injury. We suggest two reasons for this: first, ASCs could not bypass entrapment in the lung and re-enter the circulation; and second, the inflammatory response from the site of injury was not strong enough to activate ASC homing during physiological wound repair. We postulate that pulmonary passage was the major reason for the lack of detectable ASC homing. Entrapment in the lungs has been a recurring problem with systemic administration of MSCs. 60,61 Of note, lung entrapment in our rat model was not associated with any visible signs of distress or deterioration of health.
To overcome this pulmonary first-pass effect and to ensure that administered cells reach the wound bed at the injured site in a timely manner and in sufficient numbers, alternative routes of administration have been suggested such as regional or local injection. 27,29,55 We investigated the fate of locally administered ASCs and found that F I G U R E 6 Locally administered ASCs were detected within the wound over time. A total of 2 × 10 5 ASCs transduced to express both Fluc and GFP were injected locally into two sides around each wound 24 hours after bilateral full-thickness wounds had been created on the dorsal part of the feet. A, a single tissue section from the wound in the feet 24 hours post injection stained for GFP (red staining, AEC). Magnified images of the (i) injection site, (ii) outer wound site, and (iii) inner wound site. The distribution of ASCs at the different sites of the wound was quantified over time showing, B, The injection site, outer, and inner wound site in a single graph, C, the wound site (zone 1, 2, 02, 01), D, the outer wound site (zone 1, 01), and, E, the inner wound site (zone 2, 02). Data shown as mean ± SD. Significance is shown where *P < .05; **P < .01. Abbreviations: AEC, aminoethyl carbazole; ASCs, adipose-derived mesenchymal stromal cells; E, epithelium; F, hair follicles; Fluc, firefly luciferase; G, granulation tissue; GFP, green fluorescent protein; T, tendons they remained viable at the injection sites for 7 days. This is in line with another study reporting that rat BM-MSCs transplanted into traumatized skeletal muscle could be tracked in vivo until day 7. 33 At the histological level, we found a significant increase in ASCs local- Wounds can close by two mechanisms, contraction, or re-epithelialization. Contraction occurs predominately as a rapid repair mechanism, whereas re-epithelialization takes longer and is closer to regeneration. Our animal model allowed for measurements of wound size, time to complete healing, and ratios of WC by contraction and epithelialization. Interestingly, an enhancement in WC time was found with ASC treatment independently of the route of administration.
Both administration routes resulted in more rapid WC by reepithelialization rather than contraction as compared with the control group. The enhancement of WC by re-epithelialization rather than contraction is of clinical relevance, as it can have a significant impact on the quality of the scar by decreasing scar retraction.
The disappearance of transplanted cells and low engraftment at the injured site while still enhancing WC time suggests that ASCs play a role as initiators of wound repair rather than effectors. 62 This F I G U R E 7 Wound closure assessment in animals treated locally and systemically. Animals were treated with Fluc-GFP + ASCs, injected locally into two sides around each wound (2 × 10 5 ASCs) or systemically (2 × 10 6 ASCs) into the tail vein. Control animals were treated systemically with NaCl. A, Time taken for complete wound closure after treatment. B, Wound area over time after treatment normalized to the original wound area on day 0. C, Calculation of the contraction/epithelialization ratio: the area of the wound is measured directly after surgery. At the day of complete wound closure, the area of hairless skin is measured and considered to correspond to the area of the wound closed by epithelialization. Subtracting the area of epithelialization from the initial wound area provides the area closed by contraction. D, Wound area closed by contraction and by epithelialization. Data shown as mean ± SD. Significance is shown where *P < .05; **P < .01; ***P < .001. Abbreviations: ASCs, adipose-derived mesenchymal stromal cells; Fluc, firefly luciferase; GFP, green fluorescent protein confirms findings from a study using a mouse model of acute acidburn injury. 63 The authors also found that despite short-lived survival of locally transplanted ASCs, WC was enhanced. Another study found that intravenous human MSCs improved myocardial infarction in mice without significant engraftment through the secretion of antiinflammatory protein TSG-6. 64 The mechanism of wound repair by MSCs has been suggested to stem from their paracrine signaling, 65 which mobilizes the host cells to promote wound repair. This leads us to ask; how do MSCs filtered out in the lung influence wound repair mechanisms that are located distally? It has been shown that MSCs secrete various molecules that are modulators of cellular growth, replication, differentiation, and adherence. 66 It has been shown that systemically administered MSC exosomes attenuated lesion size and improved functional recovery post spinal cord injury. 72 A study that examined the role of MSC exosomes in wound healing showed that these structures were able to induce proliferation and migration of fibroblasts as well as enhance angiogenesis in vitro. 73 This study confirms that systemically injected ASCs behave in a manner similar to BM-MSCs, being able to enhance WC despite showing intrapulmonary trapping and limited homing to the wound area. We provide evidence of migration of locally injected ASCs into the wound bed. We suggest that the therapeutic effect of ASCs is due to their paracrine signaling. Clinical studies using ASCs for cutaneous non-healing/chronic wounds are still limited. Instead, animal models are used. The local administration of ASCs for wound repair is favored and is being evaluated for various injuries with/without scaffolds in animal studies of tendon repair, 74-76 cutaneous non-healing wounds, 77-83 burn injuries, 84 spinal cord injuries, 85,86 and for bone defects, 87 among other indications. In the field of plastic and reconstructive surgery, since the discovery of ASCs within adipose tissue, 4,88 the addition of both the SVF and ASCs have been investigated in order to overcome the re-absorption of fat grafts. [89][90][91] A limitation of our work could be the underestimation of ASC homing and engraftment because of our detection techniques and experimental setup. Although BLI allows for in vivo tracking, single cells and luminescence in deep organs cannot be detected. To overcome this, we included GFP staining to detect ASCs at a single cell level and in non-target organs. The animal model may have influenced the experimental outcome since Wistar rats are outbred and thus ASCs were isogenic rather than autologous. The "foreign" or isogenic ASCs might have primed the immune system and enabled a faster response and healing effect. In future studies, the use of alterative systemic routes such as intra-arterial to overcome the pulmonary first-pass effect, or treatment to prevent such filtering for example, by vasodilation, should be considered. 55 Another avenue to explore is the repeated administration of systemic ASCs. Fischer and colleagues hypothesized that the increase in pulmonary passage seen with their study was due to saturation of receptors during the first treatment, allowing more cells to pass the vasculature with the second treatment. 61 For local administration, the use of scaffolds or dressings could enable longer viability and more directed administration of ASCs into the wound bed. The lentiviral transduction of ASCs could alter properties that we were unable to detect. This could be a limitation for acceptance by regulatory authorities as proof of principle for advancing to phase I clinical trials. More in depth characterization of ASCs is needed to confirm that transduction does not affect their characteristics in vivo.
In this study, we provide data showing an accelerating effect of ASCs on cutaneous wound repair under physiological conditions. These data support the use of ASCs in wound management since physiological wound repair is generally considered a naturally effective process with little margin for improvement. In order to determine the clinical usefulness of ASC injection in the management of chronic wounds, we plan to perform further investigations studying their effect in a model of wound repair under pathological conditions.

| CONCLUSION
This study demonstrates that although the survival and homing of systemically administered ASCs was limited, they nonetheless led to enhanced wound repair. Locally administered ASCs survived for 7 days at the transplantation site. They showed the ability to migrate within the wound bed, effectively enhancing wound repair by significantly reducing WC time. Taken together, ASCs have the potential to enhance wound repair without the need for homing to the injured site, possibly by paracrine signaling through the release of exosomes.

ACKNOWLEDGMENTS
We would like to thank the following funders: The Fonds National

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.