Tissue was received from three individual donors. These varied in terms of absolute size, yet appreciable vessels were identified in all of them with similar cross-sectional diameters of ~ 1 mm. We found the easiest way to identify the perforator (Perf) blood vessels was to look around the cut border of the tissue within the subcutaneous fat to find an appreciable lumen. If no lumen was visible, sequential full thickness cuts were made in the tissue, working in from the borders a few millimetres at a time, until a lumen could be seen. Once the lumen was identified, it was secured using fine forceps and the tissue surrounding the vessel gradually dissected away using a scalpel blade to reveal the length of the vessel.

The length of vessel that was isolated from each tissue sample varied between 1 and 3 cm. Once removed from the tissue bed, the adventitia was removed using fine forceps and a scalpel. Despite the small lumen size, vessels were able to be opened longitudinally and gently scraped to remove the endothelium – care was taken here not to apply too much pressure that the vessel was cut. Irrespective of the starting length, the opened vessels were cut into smaller ~ 1 cm lengths. Each of these was cut into 1–2 mm3 cubes and put into separate flasks. There was no difference in the speed of explant between the donated tissue samples, with SMC becoming visible within 1 week of explant and confluency being reached within 4–6 weeks (Fig. 2a).

Following explant, the identity of the cells was verified using co-staining for SM-MHC (a SMC marker protein) and α-SMA; a protein enriched in SMC but that can also be expressed in other cell types including myofibroblasts (Owens et al. 2004). All cells stained positively, confirming a pure population of SMC that was not contaminated by other vascular cell types such as endothelial or adventitial fibroblast cells (Fig. 2b).

Isolation and validation of perforator smooth muscle cell identity. (a) Human skin tissue was collected from the lower limb of patients undergoing amputation. The skin was dissected, dermis side up, and the perforator vessels identified and removed. These were covered in growth media and the adventitia was dissected off. The vessels were opened longitudinally, the endothelial layer was scraped away, and the resulting medial tissue was dissected into ~ 2 mm3 pieces. These were transferred to a 25 cm2 tissue culture flask and maintained in a tissue culture incubator. After approximately 1 week, cells could be seen migrating from the edges of the tissue and the flask was confluent within 1 month. (b) The identity of the cells as smooth muscle cells was confirmed using co-staining for smooth muscle myosin heavy chain (SM-MHC), alpha smooth muscle actin (α-SMA) and DAPI nuclear stain at passage 1. Scale bar = 200 μm; images are representative of n = 3 independent isolations

After confirming that the explanted cells were SMC, we wanted to compare their appearance and function with SMC from alternative sources that have already been characterised. For this, we chose the umbilical artery (UA) as an alternative arterial source, and saphenous vein (SV) as an alternative venous source. We have extensive experience in using and characterising these cells (Hemmings et al. 2021; Riches et al. 2014a, b).

The spread cell area of Perf-SMC was 10,295 ± 764 µm2. Although larger than UA-SMC (6,092 ± 930 µm2) and SV-SMC (6,994 ± 407 µm2), the difference was not significant. The spindle morphology of the cells was also consistent across SMC from each tissue type with no difference in average circularity (Fig. 3a). SMC function was assessed using 7-day proliferation curves. Over the course of 7 days, Perf-SMC increased their cell number by 3.3-fold. This was similar to UA-SMC, but less than SV-SMC proliferation rates where cell number increased by 5.6-fold. This was significantly higher than both Perf- and UA-SMC (Fig. 3b).

We also evaluated the stress of the SMC by looking at the proportion of multinucleated cells as a marker of potential DNA damage. Only 3.33 ± 0.33% of Perf-SMC exhibited more than one nucleus per cell. UA-SMC were slightly higher at 5.04 ± 0.51%, and SV-SMC higher still at 5.98 ± 1.22%. However, there was no significant difference between the source tissue types (Fig. 3c). From this data, we have shown that Perf-SMC have a comparable in vitro cell morphology, proliferative capacity and multinucleation to UA-SMC and, with the exception of proliferative rate, SV-SMC.

Characterisation of Perf-SMC phenotype compared to other vascular SMC sources. (a) The area and circularity were measured from fifty proliferating cells per patient donor of SMC isolated from the skin perforator (Perf), umbilical artery (UA) or saphenous vein (SV). Scale bar = 50 μm. (b) The proliferation of SMC from all sources were measured in full growth media over 7 days. (c) The proportion of multinucleated cells in five fields of view per patient donor were quantified from all SMC sources. Scale bar = 50 μm. Insets show an example of mono- or bi-nucleated cells. ***P < 0.001, **P < 0.01, ns = not significant. n = 3 patient donors for each cell source. Small circles represent individual cells, and larger triangles or diamonds represent the summary data. Each donor is presented as a separate colour

SMC are not terminally differentiated, and other studies have suggested that primary SMC can lose their phenotypic characteristics as early as passage 7 (Chang et al. 2014). Thus, once we had established that Perf-SMC were phenotypically comparable to more commonly used SMC, we were interested to see how long they maintained their phenotype in culture. To do this, we serially-passaged the Perf-SMC and compared the morphological, proliferative, and stress-related characteristics at low passage (p2-3) and at high passage (p8-9).

The appearance of Perf-SMC was not different between low and high passage, with comparable cell areas and circularity in both conditions (Fig. 4a). Moreover, the proliferative capacity of low and high passage cells was comparable at early timepoints but was significantly higher for high passage cells at day 7 (Fig. 4b). There was almost double the number of multinucleated cells in high passage SMC compared to low passage SMC (Fig. 4c). This did not reach statistical significance, however we decided to further interrogate markers of cell stress and DNA damage, to confirm or negate the suggestion of increased stress.

The nuclei in cells which have undergone DNA damage or premature in vitro ageing can become deformed (Hänzelmann et al. 2015) and so we quantified the proportion of cells that had an irregularly shaped nucleus. For this, we considered ovoid or round nuclei to be ‘normal’ and any nuclei that had clefts, invaginations, non-ovoid shapes or evidence of micronuclei adjacent to the main nucleus to be ‘aberrant’. At low passage, 8.67 ± 0.33% of cells had aberrant nuclei, and this was similar at high passage which had a frequency of 6.33 ± 0.88% SMC (Fig. 4d). We also looked more deeply at a sub-organelle level and performed immunocytochemistry for γH2AX, which is a marker of double-stranded DNA breaks (Valdiglesias et al. 2013). At low passage, 18.00 ± 2.50% of SMC exhibited at least one γH2AX-positive loci in their nucleus. At high passage, this value was not significantly different at 15.33 ± 1.76% of SMC (Fig. 4e).

Stability of Perf-SMC phenotype across passages. Matched Perf-SMC were cultured and their parameters measured at both low (p2-3) and high (p8-9) passages. a Area and circularity from fifty proliferating cells per patient donor. b Proliferation in full growth media over 7 days. c The proportion of multinucleated cells (scale bar = 20 μm), d aberrantly-shaped nuclei (scale bar = 50 μm; insets show examples of aberrant ‘A’ and normal ‘N’ nuclear morphology using DAPI staining) and e γH2AX-positive nuclei (scale bar = 20 μm; insets show γH2AX as pink foci on DAPI-stained nuclei) was quantified in five fields of view per patient donor for matched low and high passage Perf-SMC. **P < 0.01, ns = not significant. n = 3 patient donors for each cell source. Small circles represent individual cells, and larger triangles represent the summary data. Each Perf-SMC donor is presented as a separate colour shade

Thus, although there was a suggestion of increased DNA damage according to the proportion of multinucleated cells, this was not borne out by further interrogation using different assays. From this, we conclude that the morphology of Perf-SMC is stable across passage up to at least passage 9, and that they are not undergoing replicative senescence or an increased rate of DNA damage and stress at these later passages.