Load-deformation characteristics of acellular human scalp: assessing tissue grafts from a material testing perspective

Acellular matrices seem promising scaffold materials for soft tissue regeneration. Biomechanical properties of such scaffolds were shown to be closely linked to tissue regeneration and cellular ingrowth. This given study investigated uniaxial load-deformation properties of 34 human acellular scalp samples and compared these to age-matched native tissues as well as acellular dura mater and acellular temporal muscle fascia. As previously observed for human acellular dura mater and temporal muscle fascia, elastic modulus (p = 0.13) and ultimate tensile strength (p = 0.80) of human scalp samples were unaffected by the cell removal. Acellular scalp samples showed a higher strain at maximum force compared to native counterparts (p = 0.02). The direct comparison of acellular scalp to acellular dura mater and temporal muscle fascia revealed a higher elasticity (p < 0.01) and strain at maximum force (p = 0.02), but similar ultimate tensile strength (p = 0.47). Elastic modulus and ultimate tensile strength of acellular scalp decreased with increasing post-mortem interval. The elongation behavior formed the main biomechanical difference between native and acellular human scalp samples with elastic modulus and ultimate tensile strength being similar when comparing the two.


Scientific Reports
| (2020) 10:19243 | https://doi.org/10.1038/s41598-020-75875-z www.nature.com/scientificreports/ as two other representative scaffolds recently investigated by our group. A biomechanical comparison of these scaffolds will facilitate a thorough assessment of potential graft materials for potential new future applications. Based on the previous observations on ADM and ATMF, where the collagenous backbone that is determining the load-deformation behavior of soft tissues was largely unaffected by the acellularization procedure with sodium dodecyl sulphate (SDS) we stated the following hypothesis: the biomechanical parameters of acellular scalp are non-different from native counterparts.

Material/methods
Sample retrieval and processing. A total of 68 human scalp samples were harvested at the Institute of Legal Medicine, University of Leipzig, Germany during forensic autopsies. The tissues were evenly distributed into two groups for further processing, an acellular scalp (AS) and a native scalp group (NS). The AS group (13♀, 21♂) had a mean age of 56 ± 22 years (range 17 to 87 years) and a post-mortem interval (PMI) of 77 ± 31 h (range 11 to 120 h). The NS group (13♀, 21♂) had a comparable mean age (range 18 to 87 years) and a PMI of 65 ± 24 h (range 32 to 111 h). The samples were taken from the temporal, fronto-parietal and occipital region. Only macroscopically intact scalp samples without scars and with no known history of systemic dermatologic disorders were selected for this study. The samples were predominantly taken from cadavers that died from sudden events such as an acute myocardial infarction, traumatic brain injury, suicide by hanging, vertebral artery dissection or polytrauma (excluding damage of the used scalp region). The ADM 12 and ATMF 13 sample properties for statistical comparison were taken from recent own publications and these data were obtained with similar methods. All methods were carried out in accordance with relevant guidelines and regulations. The protocol has been approved by the Ethics Committee of the University of Leipzig, Germany (protocol number 486/ 16-ek) and in line with the Saxonian Death and Funeral Act of 1994 (third section, paragraph 18 item 8). As per German law, the state attorney as the legally authorized representative approved the here used tissue samples for research purposes in each individual case. Following the harvesting, the samples were precooled and then transferred to − 80 °C for storage. When defrosted for further processing, the hair was shaved off all samples to assure an even surface for in-plane strain measurements by digital image correlation. Subsequently, the samples were cropped into a dog bone shape by means of a template, adapted from the ISO 527-2 standard (DIC; Fig. 1) 14 .
All samples were orientated in an anterior-posterior direction as performed previously 15 assuring a high level of comparability between studies. Cell removal for the 34 scalp samples of the AS group was performed as described previously 2 . The samples were submerged in an SDS solution of 1 wt.% (Roth, Karlsruhe, Germany) When loaded, the upper clamp moved away from the lower clamp (II), causing a deformation of the area of parallel measurement length. In the last phase of the experiment, the sample fails in the area of parallel measurement length (red dotted circle) after reaching the maximum force (III). Mechanical testing. For the determination of the cross-sectional area, which was required to calculate the ultimate tensile strength (UTS) and the elastic modulus (E mod ), casts of the samples were produced. Therefore, following the tapering, polysiloxane material (medium-bodied, Exahiflex; GC Corporation, Tokyo, Japan) was used to cast a negative impression of the scalp samples in the parallel measurement area. The resulting casts were scanned at a resolution of 1200-dpi (Perfection 7V750Pro Scanner; Seiko Epson Corporation, Suwa, Japan) with a reference scale before their cross-sectional areas were determined using Measure 2.1d software (DatInf, Tübingen, Germany). To allow for optical strain measurements, a randomly-distributed speckle pattern was created on the samples' surface using graphite powder. A universal testing machine (AllroundLine Table- Histology, immunolabeling and scanning electron microscopy. Six representative scalp samples (three NS and three AS) were fixed in 10% neutral-buffered paraformaldehyde solution (Sigma-Aldrich, Munich, Germany), dehydrated and subsequently embedded in paraffin. A H&E stain was performed for an overview stain to observe general microscopic changes between native and acellular scalp samples. An Alcian Blue stain was performed to assess the sulphated glycosaminoglycans and a Weigert's Resorcin-fuchsin stain to visualize elastic fibers in native and acellular samples. Moreover, specific immunostains were performed to assess the following components between native and acellular scalp samples: collagen type I and III, fibronectin and decorin. The staining protocol for the aforementioned stains is outlined below. Sections of 7 μm were deparaffinized with xylene and rehydrated in a descending ethanol series. For hematoxylin and eosin (H&E) staining deparaffinized sections were incubated for six minutes in Harry's hematoxylin (Sigma-Aldrich), rinsed in tap water and counterstained for four minutes in eosin (Carl Roth, Karlsruhe, Germany). For the performance of Alcian blue (AB) staining, rehydrated sections were incubated for three minutes in one percent acetic acid and then immersed for 30 min in 1% AB staining solution (Carl Roth). Following the rinsing in 3% acetic acid and a 2 min washing step in distilled water, cell nuclei were counterstained for 5 min using nuclear fast red aluminium sulphate solution (Carl Roth). Rehydrated sections were stained using Weigert's Resorcin-fuchsin (Waldeck GmbH, Münster, Germany) for 5 min to detect elastic fibers. Thereafter, the sections were differentiated in hydrochloric acid/ethanol for 1 min. After washing with tap water for 10 min, an additional washing step in distilled water was conducted. Nuclear fast red aluminum sulphate solution was applied to counterstain the cell nuclei in the Resorcin-fuchsin stains. All stained sections were entellan covered (Merck-Millipore, Darmstadt, Germany). Images were taken using a DM1000 LED light microscope (Leica, Wetzlar, Germany). For the immunostainings, the above-mentioned samples were washed three times with Tris buffered saline (TBS: 0.05 M Tris, 0.015 M NaCl, pH 7.6), before being incubated with a 0.1% Pronase ready to use solution (BioLogo, Kronshagen, Germany) to demask epitopes for 15 min at 37° C. Subsequently, the sections were re-rinsed with TBS before being incubated with protease-free donkey serum (5% diluted in TBS with 0.1% Triton X 100 for cell permeabilization) for 20 min at room temperature. From here immunostainings for type I/III collagen, decorin and fibronectin were performed as described previously, including counterstaining with 4′,6-diamidino-2-phenylindole (DAPI) 12 . Details of the antibodies used are shown below, (Table 1) Labeled sections were then washed three times for 5 min with TBS before being mounted with Fluoromount mounting medium (Southern Biotech, Eching, Germany) and digitalized using a SPEII confocal laser scanning microscope (Leica). Data assessment and statistical analysis. The mechanical properties of the scalp samples were evaluated from the DIC data and synchronized force readings using MATLAB R2017b software (Mathworks, Natick, MA, USA). Engineering stress-strain curves were calculated under inclusion of the cross-sectional areas. E mod , UTS and strain at maximum force (SF max ) were assessed. The E mod was determined in the linear region at the beginning of each stress-strain curve and UTS as the maximum stress value of each graph. SF max was set as the corresponding engineering strain when reaching the maximum force or UTS, respectively. Excel Version 16.15 (Microsoft Corporation, Redmond, USA) and GraphPad Prism software version 8 (GraphPad Software, La Jolla, CA, USA) were used for the statistical evaluation. The D' Agostino & Pearson normality test was used to test the Gaussian distribution of the samples. Parametric data of samples were then tested using an ordinary one-way ANOVA. A Kruskal-Wallis test was used for nonparametric data. A Fisher's LSD (least significant difference) test for ANOVA and uncorrected Dunn's test for Kruskal-Wallis has been applied. Pearson and Spearman correlation coefficients were reported for normally and non-normally distributed values, respectively. P values equal to or smaller than 0.05 were considered statistically significant. Mean values ± standard deviations are reported in the text.  Fig. 2). The SF max of AS (24.7 ± 3.6%, median = 24%) was significantly higher compared to NS (18.3 ± 3.4%, median = 18.5%, p = 0.021), ATMF (11.1 ± 3.1%, median = 10.6%, p < 0.001) and ADM (13.2 ± 1.8%, median = 13.4%, p < 0.001; Fig. 2).

Results
Elastic modulus and ultimate tensile strength of human acellular scalp decrease with increasing post-mortem interval. Apart from the significant correlations between PMI and both the E mod as well as UTS of AS samples, no other significant correlations between biomechanical properties of AS samples and potential confounders such as age, sex and PMI were found (Fig. 3).
Sodium dodecyl sulphate removed cellular components from human scalp samples, whereas antibody staining intensities partly varied between acellular and native scalp samples. The comparison of SDS-treated NS and AS samples revealed the absence of cell nuclei and stained chromatin in the latter in the performed H&E, AB (both Fig. 4), Resorcin-fuchsin (Fig. 5), anti-collagen type I and III (Fig. 6), anti-decorin and anti-fibronectin (both Fig. 7) staining. This indicates an overall successful acellularization of the samples. Also, the SEM indicated cell removal after scalp sample treatment with SDS (Fig. 8). Acellularization appeared to weaken the AB stain (Fig. 4), but lead to more intense immunolabeling for collagens type I and III (Fig. 6) and fibronectin (Fig. 7) but not to a clear effect on decorin. In acellular samples, both the anti- www.nature.com/scientificreports/ collagen type I and III stainings seemed to be more prominent in the stratum papillare compared to the stratum reticulare, whereas in NS the intensity of anti-collagen staining appeared to be uniformly distributed over the different layers (Fig. 6). In the anti-collagen-I and III immunostaining as well as the anti-decorin and -fibronectin stains the native skin shows an abundance of cells, indicated by blue DAPI-stained chromatin (Figs. 6 and 7). In contrast, the absence of stained chromatin in the acellular samples depicts an encompassing acellularization with an accompanied dissolving of the epidermis (Fig. 6 and 7). Whereas collagen types I and III appear to be uniformly distributed throughout the dermal layer in the NS, they seem to be more intense in the superficial stratum papillare of the dermis (white dotted squares labeled with number 1 in Fig. 6) compared to the deeper stratum reticulare (white dotted squares labeled with number 2 in Fig. 6) in AS. After cell removal, elastic fibers seem to be visible more clearly (Fig. 5). In SEM images the superficial structure of the skin became discernable after cell removal (Fig. 8).

Discussion
Following surgical implantation, acellular scaffolds provide cell attachment sites, thereby guiding tissue growth and remodeling 17 . The biomechanical properties of acellular scaffolds seem to be closely interconnected with these functions. This given study showed for the first time that quasi-static biomechanical properties of AS do not differ significantly from the native state when deformed in a uniaxial tensile testing environment. Likewise, the two materials do not differ regarding their capacity to withstand loads when elongated, indicating that the load-bearing collagen backbone of human skin stays intact during acellularization. A similar E mod and tensile strength, when compared to the native tissue, is in line with previous investigations of our group on human iliotibial tract 2 , dura mater 12 and temporal muscle fascia 13 . On the contrary, acellularization of porcine samples revealed decreasing values for the E mod and UTS following an identical sample treatment as for the human tissues. These interspecies differences are unexpected as the porcine skin is morphologically similar to human skin and therefore frequently used as a model for the same 18,19 . The significantly higher elongation of AS compared to NS is likely a consequence of occurring 'free spaces' in the tissue following the cell removal and subsequent loss of many hair shafts in this study, which was proven by using histology, immunohistochemistry and SEM agreeing with previous observations 2,20 . Regarding this, the SF max also increased in acellular iliotibial tract samples 2 , but stayed unaltered in human dura mater 12 , temporal muscle fascia 13 and porcine skin 3 . The increased tissue elongation following acellularization might be a result of both the amount of occurring free spaces formerly occupied by cells and the axis of load application during the uniaxial tensile test. In human dura mater, temporal muscle fascia and iliotibial tract comparatively few cells are scattered within the predominant collagen matrix 2,12,13 . When the highly anisotropic iliotibial tract is tensile loaded in the direction of the dominant collagen orientation following cell removal, the elongation increases as collagens can fill these 'empty spaces' . However, the increase in elongation might not be observed, when the load is applied perpendicular to the axis of the preferred collagen course. In tissues, with a more complex anisotropic organization compared to the iliotibial tract such as dura mater or temporal muscle fascia the increasing effect of acellularization on the SF max might diminish due to varying collagen bundle arrangements in the loading axis of the tested sample, which cannot be detected by the naked eye, but was visualized by SEM recently 21 . Complementing this, for a tissue with a comparatively high number of cellular components in relation to the present collagens, such as the here presented human scalp, the axis of load application is less important as the occurring free spaces following acellularization are too large and consequently inevitably lead to an increased elongation of the acellular sample compared to the native one as shown in this study. All things considered, due to the different strain values of acellular skin compared to www.nature.com/scientificreports/ the native one, the hypothesis has to be rejected and it has to be stated that cell removal does partly influence the biomechanical parameters of human scalp. The here mentioned 'free spaces' caused by the cell removal seemed to expose the epitopes recognized by anti-collagen type I and III as well as fibronectin antibodies resulting in intensified staining in the superficial skin layers, showing the decreased intensity of coloration from superficial to deep when assessed qualitatively. An increased occurrence of e.g. dermal fibronectin in proximity to the cell-rich epidermal layers seems plausible, taking into consideration that fibronectin is of importance for the cell adhesion, migration, differentiation and ingrowth of cells 22 . Removal of the unformed ECM, namely proteoglycans and glycosaminoglycans may be a further explanation for the enhanced visibility of the formed ECM, which is supported by the weakened AB staining in acellular samples. Additionally, in line with previous investigations on human dura mater 12 , acellularization does expose elastic fibers seen in standard histology and of the dermal surface in SEM. The latter could for example be used to analyze the three-dimensional architecture of dermal papillae in different body regions to study their structural adaptation to mechanical loading.
Besides their orthotopic use, acellular skin grafts are frequently used as heterotopic grafts in various body sites 5,[23][24][25] . In this study, we have shown that human AS is significantly more elastic compared to human ATMF and ADM. When an acellular skin graft is used to cover dura mater defects, this increased elasticity of the graft compared to the original tissue should be considered, given that energy storage and dissipation might be altered compared to the original tissue 26 . This particularly applies when the graft will be used for a bridge-like duraplasty as done in surgical cranectomies 27 without the skull limiting the expansion of the scaffold during cerebrospinal fluid pulsations. The given study showed that both AS neither differed from NS nor from compared acellular matrices (ATMF and ADM) regarding its capability to withstand loads before failing. Hypothetically, in cases of duraplasty in which the soft tissue overlying the brain is not covered by the skull 27 , a more compliant graft by being equally load-resistant such as an acellular skin scaffold could allow for greater expansions during brain edema, potentially causing less increase of the intracranial pressure 28 and benefits the patient's outcome after intensive care periods. ATMF and ADM did not statistically differ regarding their tensile strength from the AS scaffolds in this study but revealed differences in strain values and elastic behavior under load application. A thorough knowledge of these biomechanical cross-comparisons will lead to a large biomechanical database, which allows for an individual selection of the most suitable graft material during surgery and in addition aids as a reference for the design of new artificial 3D-printed graft materials.  www.nature.com/scientificreports/ To the best of our knowledge, the influence of both the donor's age as well as the time between graft retrieval and transplanting into the recipient on the biomechanical properties of the graft were not investigated before. Our results reveal that the E mod and the UTS of the human acellular skin scaffold decrease with increasing time after death. Therefore, post-mortem tissue donation should be initialized as soon as possible after the individual's death. The decreasing E mod could be caused by progressing collagen degradation or autolysis after the tissue loses vitality. This might be caused by the decreased pH value occurring in tissues, when oxygen circulation stops and therefore energy metabolism, respectively 29 .

Limitations
The here presented study is limited by the sample size, which had been restricted by the available number of tissues for the given project. Samples in this study were harvested from the scalp region. Biomechanical acellular dermal properties of other body regions likely vary. For the performed tensile tests a quasi-static uniaxial setup was selected, which might differ from the dynamic and multiaxial properties of scalp tissues under native ('traumatic') conditions. Furthermore, an impact of potential skin-affecting diseases on the biomechanical parameters obtained in the given study cannot be excluded although tissues were only harvested from cadavers with no known history of systemic dermatologic disorders and tissues that were macroscopically sound in this regard.

Conclusions
The biomechanical properties of skin samples provided in this study complement previous findings that acellularization of human scalp with SDS does not impact the scaffold's elastic modulus and ultimate tensile strength, when compared to the native state irrespective of the origin of the tissue. The elongation, most likely dependent on the tissue's original cell-collagen-ratio, was found to be the main biomechanical difference between native and acellular human scalp tissues. Skin grafts should be harvested as soon as possible after death to prevent degradation influencing the biomechanical properties (Supplementary information).