Collagen Network Formation in In Vitro Models of Musculocontractural Ehlers–Danlos Syndrome

Loss-of-function mutations in carbohydrate sulfotransferase 14 (CHST14) cause musculocontractural Ehlers–Danlos syndrome-CHST14 (mcEDS-CHST14), characterized by multiple congenital malformations and progressive connective tissue fragility-related manifestations in the cutaneous, skeletal, cardiovascular, visceral and ocular system. The replacement of dermatan sulfate chains on decorin proteoglycan with chondroitin sulfate chains is proposed to lead to the disorganization of collagen networks in the skin. However, the pathogenic mechanisms of mcEDS-CHST14 are not fully understood, partly due to the lack of in vitro models of this disease. In the present study, we established in vitro models of fibroblast-mediated collagen network formation that recapacitate mcEDS-CHST14 pathology. Electron microscopy analysis of mcEDS-CHST14-mimicking collagen gels revealed an impaired fibrillar organization that resulted in weaker mechanical strength of the gels. The addition of decorin isolated from patients with mcEDS-CHST14 and Chst14−/− mice disturbed the assembly of collagen fibrils in vitro compared to control decorin. Our study may provide useful in vitro models of mcEDS-CHST14 to elucidate the pathomechanism of this disease.


Introduction
Ehlers-Danlos syndrome (EDS) comprises a clinically and genetically heterogeneous group of heritable connective tissue disorders characterized by joint hypermobility, skin hyperextensibility, and tissue fragility [1,2]. Currently, EDS is classified into 14 subtypes based on clinical, molecular and biochemical features according to the 2017 International Classification and more recent updates [3,4]. Dominant negative effects or haploinsufficiency of mutant procollagen α-chain genes or deficiency of collagen-processing enzymes have been identified as the basis for various types of EDS [4,5].

Preparation of Fibroblast-Embedded Collagen Gel
The collagen gels were prepared as described previously with minor modifications [22]. Briefly, 1 × 10 6 dermal fibroblasts were suspended in 4 mL of DMEM containing FBS, sodium bicarbonate, penicillin, and streptomycin. Then, 4 mL of the cell suspension was mixed with 2 mL of bovine type I collagen acidic solution (3 mg/mL, pH 3.0, Nitta Gelatin Co., Osaka, Japan, #IAC-30). Then, 5 mL of the mixture was placed in a 6-well plastic dish and incubated, at 37 • C, for allowing gelation of collagen. The resultant collagen gels containing fibroblasts were detached from the surrounding brim of plastic dishes and maintained, at 37 • C, in air supplemented with 5% CO 2 . The gel contraction was evaluated by measuring the diameter of the gels at the indicated time. We used the 0.1% collagen gels for gel contraction assay and transmission electron microscopy analysis and 0.22% collagen gels for measuring the breaking strength and scanning electron microscopy analysis.

Transmission Electron Microscopy
Transmission electron microscopy was performed as described previously [16,17]. Briefly, the collagen gels were prefixed in 3% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) for 2 h, at room temperature, and postfixed in 1% osmium tetroxide in phosphate buffer for 1 h, at room temperature. After washing with water, gels were dehydrated in ethanol and transferred to QY-1 (Nisshin EM, Tokyo, Japan). Samples were embedded in the epoxy-resin mixture (Nisshin EM), cut into ultrathin sections, and observed by transmission electron microscopy (JEM-1220; JEOL, Tokyo, Japan).

Scanning Electron Microscopy
Scanning electron microscopy was performed as described previously [17]. Briefly, the collagen gels were prefixed in 3% glutaraldehyde in phosphate buffer for 2 h, at room temperature, and postfixed in 1% osmium tetroxide in phosphate buffer for 1 h, at room temperature. After washing with phosphate buffer, the gels were incubated with 1.0% tannic acid in water for 30 min. The gels were again fixed in 1% osmium tetroxide for 1 h, at room temperature, and washed with phosphate buffer. The samples were then dried by the t-butyl alcohol freeze-drying method, mounted on metal stubs, coated with platina using an ion sputter (JUC-5000; JEOL), and observed under a scanning electron microscope (JSM-5200; JEOL).

Mechanical Strength Measurement of Collagen Gel
Mechanical strength of the collagen gel was analyzed using a creep meter (RE-33005, Yamaden) as described previously [25]. The breaking strength was measured in 60% humidity, at 25 • C, using a cylindrical probe (diameter 5 mm) moving into the gel at a speed of 5 mm/min. The penetration of the probe was stopped halfway in the whole thickness of the gel specimen. The peak top of the stress-strain curve was defined as the breaking point of a gel. The measurement was repeated 8 times for each gel.

Isolation of Decorin
Human decorin was isolated from the cultured medium of normal and the patient's fibroblasts. Serum-free cultured medium (700-900 mL) was collected, freeze-dried, and dissolved in the extraction buffer (7 M urea, 50 mM Tris-HCl (pH 7.4), 0.1 M NaCl, and protease inhibitors). The extract was applied to a DEAE-Toyopearl 650 M column (Tosoh Corp., Tokyo, Japan) equilibrated with the extraction buffer. Decorin was eluted with stepwise increases of 0.3, 0.5, and 1.0 M NaCl. Fractionated samples were precipitated with ethanol and dissolved in 0.4 mL of water. For mouse decorin, skin powder was delipidated by mixing with ethanol and extracted with 4 M guanidine hydrochloride solution containing 50 mM Tris-HCl (pH 7.4), 0.1 M NaCl, 5 mM benzamidine hydrochloride, and 10 mM EDTA, at 4 • C, for 72 h with rotating. After centrifugation, the supernatant was dialyzed against the extraction buffer and applied to a DEAE column as described.
To detect decorin, we separated samples by acrylamide gel electrophoresis, transferred them onto a polyvinylidene difluoride membrane (Millipore, Burlington, MA, USA), and incubated them with the following primary antibodies: anti-human decorin (mouse IgG1, 1:1000, R&D #115402), and anti-mouse decorin (goat IgG, 1:1000, R&D #AF1060). After washing, the membranes were incubated with the appropriate peroxidase-labeled secondary antibodies and developed using a chemiluminescent peroxidase substrate (Millipore). For human decorin, the 0.5 M NaCl fraction was collected, and for mouse decorin, the 0.3 M and 0.5 M NaCl fractions were pooled and used for subsequent experiments.

Decorin-Mediated Fibrillar Organization of Type I Collagen Gels
Isolated decorin was quantified by a direct enzyme-linked immunosorbent assay using the anti-decorin antibodies. Decorin was incubated with bovine type I collagen acidic solution (collagen:decorin = 500:1 (w/w)) for 2 h in a neutral solution. The resultant collagen gels were analyzed by scanning electron microscopy as described.

Statistical Analysis
All data are presented as the mean ± SE. An unpaired, two-tailed Student's t-test was used to compare results from two groups. One-way ANOVA with Tukey-Kramer's test was used for multiple-comparison test. A p-value < 0.05 was considered to be significant. Details of statistical analyses, including the statistical tests used and p values, may be found in the relevant figures and figure legends.

Effect of Fibroblasts from mcEDS-CHST14 Patients on Contraction, Fibrillar Organization, and Mechanics of Type I Collagen Gels
We used dermal fibroblast derived from two patients with mcEDS-CHST14 whose pathology was reported in a previous study [7]. Similarly to patient 3, whose GAGs were previously analyzed [7], the cultured medium of fibroblast from patient 6 contained CS but not DS (Supplemental Figure S1, Supplemental Table S1). We then examined the effects of fibroblasts from patients with mcEDS-CHST14 on the collagen gel contraction in vitro. Culturing control fibroblasts in a free-floating collagen gel resulted in a contraction of the gel in a time-dependent manner (Figure 1a,b). By measuring the diameter of the gel at different time points, we found that contraction of the collagen gel mainly occurred within the first 6 h under our experimental conditions. No gel shrinkage was observed in the absence of fibroblasts. The collagen gels cultured with two patient-derived fibroblasts showed significantly larger gel diameters and delayed contraction than that with the healthy control at 1 and 3 h (Figure 1a,b). After 24 h, the diameters of the gels were comparable regardless of the origin of the fibroblasts, suggesting that patient-derived fibroblasts delayed the progression of the initial phase of collagen gel contraction. of the gel in a time-dependent manner (Figure 1a,b). By measuring the diameter of the gel at different time points, we found that contraction of the collagen gel mainly occurred within the first 6 h under our experimental conditions. No gel shrinkage was observed in the absence of fibroblasts. The collagen gels cultured with two patient-derived fibroblasts showed significantly larger gel diameters and delayed contraction than that with the healthy control at 1 and 3 h (Figure 1a,b). After 24 h, the diameters of the gels were comparable regardless of the origin of the fibroblasts, suggesting that patient-derived fibroblasts delayed the progression of the initial phase of collagen gel contraction. We observed fibrillar organization in the collagen gel by transmission electron microscopy analysis (Figure 2a). The assembly of collagen fibrils was evaluated by counting fibrils attached to other fibrils in the transverse sections. Three days after culturing with control fibroblasts, multiple fibrils were assembled to form collagen fibers (Figure 2a). In contrast, the assembly of fibrils was less frequent when cultured with patients-derived fibroblasts. The percentage of assembled fibrils in the total fibrils was 68.2% (n = 36 fibrils) in the control gel, and it was 38.9% and 50.0% in the two patient-mimicking gels, respectively (n = 36 and 16 fibrils for patient 3 and 6, respectively). The fibrillar organization of the gels was further examined using scanning electron microscopy ( Figure 2b). The diameter size of fibrils was significantly increased by co-culturing with fibroblasts ( Figure 2c). Although there is no apparent difference in the mean diameter of fibrils between control We observed fibrillar organization in the collagen gel by transmission electron microscopy analysis (Figure 2a). The assembly of collagen fibrils was evaluated by counting fibrils attached to other fibrils in the transverse sections. Three days after culturing with control fibroblasts, multiple fibrils were assembled to form collagen fibers ( Figure 2a). In contrast, the assembly of fibrils was less frequent when cultured with patients-derived fibroblasts. The percentage of assembled fibrils in the total fibrils was 68.2% (n = 36 fibrils) in the control gel, and it was 38.9% and 50.0% in the two patient-mimicking gels, respectively (n = 36 and 16 fibrils for patient 3 and 6, respectively). The fibrillar organization of the gels was further examined using scanning electron microscopy ( Figure 2b). The diameter size of fibrils was significantly increased by co-culturing with fibroblasts ( Figure 2c). Although there is no apparent difference in the mean diameter of fibrils between control and patientmimicking, the diameter size distribution of fibril was different between two groups. The diameter of fibrils in the control gels appeared more homogeneous than those in patientmimicking gels (Figure 2d). These results suggest that patient-derived fibroblasts disturbed fibrillar organization of the collagen gels in vitro. and patient-mimicking, the diameter size distribution of fibril was different between two groups. The diameter of fibrils in the control gels appeared more homogeneous than those in patient-mimicking gels (Figure 2d). These results suggest that patient-derived fibroblasts disturbed fibrillar organization of the collagen gels in vitro.  We investigated mechanical properties of the collagen gels by measuring the breaking strength using a creep meter. The stress-strain curves of the collagen gels indicated that the breaking strength of the collagen gel was markedly increased in the presence of control and patient-derived fibroblasts compared with fibroblast-free gels (Figure 3a). We found that the patient-mimicking gels reached the failure points at a lower stress than the control gels. Consistently, the patient-mimicking gels showed significantly lower maximum stress than the control gels ( Figure 3b). These data indicated that the impaired fibrillar organization in patient-mimicking gels resulted in weaker mechanical strength of the gels than the control gels.
Genes 2023, 14, 308 7 of 15 We investigated mechanical properties of the collagen gels by measuring the breaking strength using a creep meter. The stress-strain curves of the collagen gels indicated that the breaking strength of the collagen gel was markedly increased in the presence of control and patient-derived fibroblasts compared with fibroblast-free gels (Figure 3a). We found that the patient-mimicking gels reached the failure points at a lower stress than the control gels. Consistently, the patient-mimicking gels showed significantly lower maximum stress than the control gels ( Figure 3b). These data indicated that the impaired fibrillar organization in patient-mimicking gels resulted in weaker mechanical strength of the gels than the control gels.

Effect of Decorin Isolated from Normal or Patient's Fibroblasts on Type I Collagen Gel Properties
Various factors are involved in the fibroblast-mediated collagen gel contraction, making it challenging to identify the molecules that regulate the collagen fibrillar organization. We sought to examine whether the disturbed collagen network in the gels co-cultured with patient-derived fibroblasts was due to abnormalities in decorin. To this end, we isolated decorin from the cultured medium of control and the patient's fibroblasts and investigated its effects on collagen fibrillogenesis in vitro. Anion exchange chromatography followed by immunoblot analysis indicated that decorin was highly enriched in the 0.5 M NaCl elution fraction (Figure 4a). The apparent molecular weight of decorin was undistinguishable between normal and patient fibroblasts.
Type I collagen was incubated with or without decorin (collagen:decorin = 500:1 (w/w)) for 2 h in a neutral solution, and the resultant collagen gels were analyzed by scanning electron microscopy (Figure 4b and Supplemental Figure S2). In the absence of decorin, the majority of collagen fiber was composed of two or three fibrils. Thicker collagen fibers were formed in the presence of decorin derived from normal fibroblasts, confirming the instructive roles of decorin in collagen fibrillar organization. Observation at higher magnification visualized GAG chains of decorin attached to collagen fibrils ( Figure  4c and Supplemental Figure S2). GAG chains of control decorin were evenly spaced and vertically oriented on the collagen fibers. In contrast, GAG chains of patient-derived decorin were misoriented and did not tightly attach to collagen fibrils. The diameter of collagen fibers was significantly smaller in the collagen gels incubated with patient-derived decorin than normal decorin (Figure 4d). We also found that the diameter of each fibril was smaller when incubated with patient-derived decorin compared to the control (Figure 4e). The diameter of fibrils was more variable in the collagen gel incubated with

Effect of Decorin Isolated from Normal or Patient's Fibroblasts on Type I Collagen Gel Properties
Various factors are involved in the fibroblast-mediated collagen gel contraction, making it challenging to identify the molecules that regulate the collagen fibrillar organization. We sought to examine whether the disturbed collagen network in the gels co-cultured with patient-derived fibroblasts was due to abnormalities in decorin. To this end, we isolated decorin from the cultured medium of control and the patient's fibroblasts and investigated its effects on collagen fibrillogenesis in vitro. Anion exchange chromatography followed by immunoblot analysis indicated that decorin was highly enriched in the 0.5 M NaCl elution fraction (Figure 4a). The apparent molecular weight of decorin was undistinguishable between normal and patient fibroblasts.
Type I collagen was incubated with or without decorin (collagen:decorin = 500:1 (w/w)) for 2 h in a neutral solution, and the resultant collagen gels were analyzed by scanning electron microscopy (Figure 4b and Supplemental Figure S2). In the absence of decorin, the majority of collagen fiber was composed of two or three fibrils. Thicker collagen fibers were formed in the presence of decorin derived from normal fibroblasts, confirming the instructive roles of decorin in collagen fibrillar organization. Observation at higher magnification visualized GAG chains of decorin attached to collagen fibrils (Figure 4c and Supplemental Figure S2). GAG chains of control decorin were evenly spaced and vertically oriented on the collagen fibers. In contrast, GAG chains of patient-derived decorin were misoriented and did not tightly attach to collagen fibrils. The diameter of collagen fibers was significantly smaller in the collagen gels incubated with patient-derived decorin than normal decorin (Figure 4d). We also found that the diameter of each fibril was smaller when incubated with patient-derived decorin compared to the control (Figure 4e). The diameter of fibrils was more variable in the collagen gel incubated with patient-derived decorin than with controls ( Figure 4f). These results directly demonstrated an impaired function of patient-derived decorin in fibrillar organization of type I collagen.
patient-derived decorin than with controls ( Figure 4f). These results directly demonstrated an impaired function of patient-derived decorin in fibrillar organization of type I collagen.

Effect of Fibroblasts and Decorin Derived from Chst14-Deficient Mice on the Fibrillar Organization of Type I Collagen Gels
A mcEDS-CHST14 model mouse lacking functional CHST14 protein has been established recently. We examined the effect of fibroblasts from Chst14-deficient mice on the fibrillar organization of type I collagen gels in vitro. Scanning electron microscopy analysis showed the assembly of collagen fibers composed of multiple fibrils (Figure 5a,b and Supplemental Figure S3). The diameter of fibrils was markedly increased by co-culturing with fibroblasts compared with the cell-free condition (Figure 5a,b). We found that the diameter of fibrils was smaller in the collagen gels cultured with fibroblasts derived from Chst14 −/− mice compared to those from Chst14 +/+ and Chst14 +/− mice (Figure 5c). Higher magnification views indicated that GAG chains were not tightly aligned on fibrils in Chst14-deficient condition (Figure 5b).

Effect of Fibroblasts and Decorin Derived from Chst14-Deficient Mice on the Fibrillar Organization of Type I Collagen Gels
A mcEDS-CHST14 model mouse lacking functional CHST14 protein has been established recently. We examined the effect of fibroblasts from Chst14-deficient mice on the fibrillar organization of type I collagen gels in vitro. Scanning electron microscopy analysis showed the assembly of collagen fibers composed of multiple fibrils (Figure 5a,b and Supplemental Figure S3). The diameter of fibrils was markedly increased by co-culturing with fibroblasts compared with the cell-free condition (Figure 5a,b). We found that the diameter of fibrils was smaller in the collagen gels cultured with fibroblasts derived from Chst14 −/− mice compared to those from Chst14 +/+ and Chst14 +/− mice (Figure 5c). Higher magnification views indicated that GAG chains were not tightly aligned on fibrils in Chst14deficient condition (Figure 5b). We digested the skin lysate with two chondroitinases to investigate GAG modification on decorin. Chondroitinase ABC degrades both DS and CS chains, whereas chondroitinase B acts only on DS. Decorin from Chst14 +/+ mice was shifted to lower molecular weight after digestion with chondroitinase ABC and chondroitinase B (Figure 6a). In contrast, decorin from Chst14 −/− mice was partially resistant to chondroitinase B but was digested by chondroitinase ABC. This result indicated that DS chains on decorin were markedly reduced and replaced by CS chains in Chst14 −/− mice, as previously reported in patients with mcEDS-CHST14 [7]. We isolated decorin from cultured media of mouse fibroblast using anion exchange chromatography (Figure 6b) and examined its effect on the fibrillar organization of type I collagen gels. Again, we found that the collagen gels We digested the skin lysate with two chondroitinases to investigate GAG modification on decorin. Chondroitinase ABC degrades both DS and CS chains, whereas chondroitinase B acts only on DS. Decorin from Chst14 +/+ mice was shifted to lower molecular weight after digestion with chondroitinase ABC and chondroitinase B (Figure 6a). In contrast, decorin from Chst14 −/− mice was partially resistant to chondroitinase B but was digested by chondroitinase ABC. This result indicated that DS chains on decorin were markedly reduced and replaced by CS chains in Chst14 −/− mice, as previously reported in patients with mcEDS-CHST14 [7]. We isolated decorin from cultured media of mouse fibroblast using anion exchange chromatography (Figure 6b) and examined its effect on the fibrillar organization of type I collagen gels. Again, we found that the collagen gels incubated with decorin isolated from two Chst14 −/− mice had significantly smaller fibril diameters compared to decorin from Chst14 +/+ and Chst14 +/− mice (Figure 6c,d). These results indicate that fibroblast-and decorin-mediated fibrillar organization of collagen are impaired by functional loss of CHST14 in both humans and mice.
Genes 2023, 14, 308 10 of 15 incubated with decorin isolated from two Chst14 −/− mice had significantly smaller fibril diameters compared to decorin from Chst14 +/+ and Chst14 +/− mice (Figure 6c,d). These results indicate that fibroblast-and decorin-mediated fibrillar organization of collagen are impaired by functional loss of CHST14 in both humans and mice.

Discussion
Previous electron microscopic analyses have shown that collagen fibrils in patients with mcEDS-CHST14 are dispersed in the reticular dermis, in contrast to the regularly and tightly assembled fibrils observed in healthy individuals [7,16]. In addition, patients show small-sized collagen bundles comprising collagen fibrils with variable diameters separated by irregular interfibrillar spaces [8]. In the present study, we established in vitro models of fibroblast-mediated collagen network formation that recapacitate mcEDS-CHST14 pathology. Electron microscopic analysis revealed impaired fibrillar organization in mcEDS-CHST14-mimicking gels, similar to the abnormal collagen network observed in patients with mcEDS-CHST14.
There are two forms of mcEDS resulting from loss-of-function mutations in CHST14 or DSE [2][3][4]. Both forms of mcEDS shares the primary diagnostic criteria, including multiple congenital contractures, characteristic craniofacial features, and characteristic cutaneous features. However, core skin features and joint manifestations are significantly less common in patients with mcEDS-DSE than in patients with mcEDS-CHST14 [10]. In mice, Dse-deficient fibrils have a larger diameter than wild-type fibrils, while Chst14deficient fibrils do not show such an increase [26]. These studies suggest that CHST14 and DSE deficiency have different impacts on the pathogenesis of mcEDS. Our in vitro collagen network formation may be useful for analyzing not only mcEDS-DSE but also other EDS pathologies.
Using decorin isolated from cultured skin fibroblasts of patients and Chst14 −/− mice, we indicated that decorin is responsible for pathological collagen fibrinogenesis. The decorin preparation used in this study may contain proteoglycans other than decorin. However, since decorin is the major carrier protein of DS chains in the dermis, the impaired collagen fibrillar organization in the mcEDS-CHST14-mimicking gels may be due to the abnormal glycosylation of decorin. Further studies using recombinantly expressed decorin may clarify the exact contribution of decorin in mcEDS.
Decorin, a member of small leucine-rich proteoglycans, localizes to the surfaces of collagen fibrils and plays pivotal functions in fibrillogenesis [19,27]. A central domain of decorin is composed of ten leucine-rich repeats flanked by two cysteine-rich regions at the amino-and carboxyl-terminus. The arch-shaped structure of decorin allows interaction with the triple helix of type I collagen molecule [27]. Amino acid sequences located between the fourth and sixth leucine-rich repeats mediate the interaction with collagen [28,29]. Decorin-deficient mice display aberrant organization of collagen fibrils and fragile skin with decreased strength and stiffness [30,31]. The fibril diameter in decorin-deficient mice is more heterogeneous than that in controls, similarly to what was observed in our mcEDS-CHST14 models. Furthermore, an in vitro study also shows that the inclusion of decorin during collagen fibrillogenesis increases the elasticity and tensile strength of resulting collagen gels [32]. These reports are consistent with our conclusion that the abnormal collagen network in the mcEDS-CHST14 models is due to a functional defect of decorin.
The amino-terminal domain of decorin contains a single DS chain. Since the protein core of decorin mainly mediates the interaction with collagen, it is not fully understood how the loss of DS chain affects collagen fibrinogenesis. However, emerging evidence has indicated a link between DS chains on decorin and the collagen network. In the shape module model, the protein core of decorin is associated with collagen fibrils, while DS chains interact with each other and form antiparallel duplexes that bridge adjacent collagen fibrils [33][34][35]. On the other hand, several lines of study have indicated that DS chains associate with collagen fibrils by electrostatic interaction but not from the antiparallel duplexes [36,37]. These two models are not mutually exclusive, and a recently proposed ring mesh model may explain the apparent discrepancy. In this model, multiple DS chains of decorin interact with each other around collagen fibrils and form a ring mesh-like structure [16,17,38]. Each ring surrounds a collagen fibril at its D-band and fuses with adjacent rings to create a planar network. In any case, DS chains on decorin affect the interfibrillar distance between the collagen fibrils and the overall structure of the collagen network.
The DS chains on decorin are replaced with a CS chain in the mcEDS-CHST14 patient [7,16]. We also confirmed the replacement of GAG chains on decorin in Chst14deficient mice. DS chains show unique conformational flexibility because IdoA residues are in an equilibrium of 1C4, 2S0, and 4C1 conformations. In contrast, the structure of CS chains is rigid due to the fixed 4C1 conformation of GlcA in CS chains [39]. Therefore, the replacement from DS to CS chains may reduce the flexibility of the GAG chains on decorin and destabilize the GAG-antiparallel duplex. Recent studies have revealed that DS chains on decorin display a curved shape and tightly adhere to the outer surface of collagen fibrils [16,17]. In CHST14-deficient humans and mice, CS chains on decorin linearly extend from collagen fibrils resulting in a disrupted ring mesh-like surrounding collagen fibrils and spatial disorganization of collagen networks. The present study also demonstrates that the alternation of GAG chains on decorin from DS to CS chains causes impaired collagen fibrinogenesis using in vitro reconstitution models.
It has been reported that the initial interaction between fibroblasts and collagen leads to rearrangement and increased density of the collagen fibrils [20,40]. Collagen gel contraction provides an in vitro model for wound healing, fibrosis, scar contraction, and connective tissue morphogenesis. We found a delayed gel contraction in mcEDS-CHST14 models suggesting that DS supports the initial phase of collagen gel contraction. DS impacts cell proliferation, migration, and tissue morphogenesis through interactions with secreted molecules such as fibroblast growth factors and hepatocyte growth factors [13,[41][42][43][44]. A disturbed pleiotropic function of DS in mcEDS-CHST14 models may result in delayed gel contraction during the early phase of collagen network remodeling.
In conclusion, this study established in vitro collagen network formation models of mcEDS-CHST14 and revealed that the alternation of GAG chains on decorin from DS to CS chains leads to abnormal collagen assembly. This study may provide useful in vitro models of mcEDS-CHST14 to elucidate the pathomechanism of this disease.

Supplementary Materials:
The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/genes14020308/s1, Figure S1: HPLC profiles of the digests of the conditioned medium from skin fibroblasts cultures with chondroitinase AC or B. Figure S2

Conflicts of Interest:
The authors declare no conflict of interest.