**1. Introduction**

Collagen is the most abundant protein in animals; in particular, it is involved in the formation of a variety of fibrillar networks, providing the structural integrity of the tissues. This molecule, and its derived gelatin, is used as a biomaterial for tissue engineering and regenerative medicine (TERM) purposes, thanks to its low immunogenicity and mechanical properties. Collagen-derived biomaterials can be obtained in two ways: (i) by preserving as much as possible the shape of the original tissue via isolation of a de-cellularized extracellular matrix (ECM); or, (ii) by in vitro polymerization of soluble tropo-collagen, alone or combination with other ECM components, such as glycosaminoglycans (GAGs) [1,2], elastin [3,4] or chitosan [5,6]. These additions are meant to improve the enzymatic resistance and mechanical properties of the derived biomaterial. According to their

application, collagen-based biomaterials can be used in the form of (i) collagen sponges, employed as three-dimensional (3-D) scaffolds in bone and cartilage repair; (ii) injectable hydrogels for drug delivery; and, (iii) two-dimensional (2-D) thin films or membranes suitable in wound dressing, dural closure, reinforcement of a compromised tissue and guided tissue regeneration [7]. The sources of collagen for these applications are mainly waste of bovine and porcine skin and bones. However, the risk of BSE (bovine spongiform encephalopathy), TSE (transmissible spongiform encephalopathy), and religious constraints (e.g., avoidance of porcine derivatives), has recently led to considering different sources of collagen. The most intriguing and promising ones come from the marine environment (for a review see [8]). In order to obtain economically sustainable amounts of these molecules, particular attention has been paid to collagens derived from fish or mollusks waste of food industry [9,10]. Besides, studies have also been done on collagens that were extracted from marine invertebrates, such as jellyfishes, generating real "blooms", in particular, environmental or seasonal conditions [11,12].

Marine sponges (Porifera) are the oldest metazoan group still extant on our planet [13]. Unlike other animal groups, they are characterized by a simple level of organization, they lack real tissues and organs, and are only formed by specialized cells types (e.g., choanocytes and pinacocytes [14]). These cells are embedded in a complex 3-D matrix network that is rich in collagen, which, in the Demospongiae class, is generically referred to as spongin [15]. It is formed by proteoglycans, glycoproteins, conventional striated collagen fibrils, and thin collagen microfilaments forming large structures that may be associated to siliceous spicules, or, in horny sponges, to foreign materials [16]. Although still being poorly characterized at the molecular level [17,18], spongin extracts [19], or sponge intact skeletons [20–22] provide a suitable framework for mammalian cell attachment and for migration and the proliferation of osteoprogenitor cells. These features reveal a promising biomaterial for bone repair. Furthermore, new fields of application of spongin are also arising, such as its use as supports for the immobilization of dyes [23–33]. Interestingly, in the Hexatinellida, a new unexpected structural role of collagen has been recently described. Indeed, this protein was identified within the basal siliceous spicules of the glass sponge *Hyalonema sieboldii* Gray, 1835 [34,35], stimulating further research on sponge collagen for a better comprehension of the skeletal structures of these deep-sea organisms.

Sponges are also organisms with a high biotechnological potential, since they are a relevant marine source of bioactive compounds. These animals, in fact, have developed a large variety of secondary metabolites for their defense from competitors, predators, and pathogens. The types of molecules produced are mainly nucleosides, toxins, and terpenoids with antibacterial, antifungal, or antiviral activity [36]. Notably, of the 18,000 marine natural bioactive products that have described until now in the pharmaceutical field, more than 30% have been isolated from sponges [37]. Unfortunately, wild sponge populations are frequently insufficient or inaccessible to produce sustainable commercial quantities of metabolites of interest. Thus, various mariculture [38,39] and in vitro cell culture [40,41] methods have been developed to fully exploit their interesting pharmaceutical features. In this perspective, the waste of sponge biomasses remaining after bioactive compound extraction could be used as a source of marine collagen.

One of the most studied sponge collagen is derived from the *Chondrosia reniformis* species, which is a member of the Demospongiae class. Its marine habitat includes a broad geographic (and bathymetric) range of marine environments, among which is the Mediterranean Sea. The body of *C. reniformis* is mainly made of tightly packed collagenous fibres that are lacking the typical needle-like silica spicules that strengthen the body of most other species of the same class. Its collagen is characterized by a unique dynamic plasticity due to the ability to reversibly alter its viscoelastic properties in an extraordinary short-time span. These features seem to be under the control of a calcium-dependent extracellular aggregating factor [42,43]. In the past, this collagen was partially biochemically characterized [44,45], and, more recently, some of its gene sequences were identified [46,47], as well as that of a collagen-maturation enzyme [48]. Moreover, an ultra-structural study was performed thanks to its partial solubilization [49]. Its biocompatibility on human skin has already been evaluated [50], and its use in the form of nanoparticles as carriers and coatings for drug preparations has been

also described [51,52]. Notwithstanding, the use of *C. reniformis* collagen for membranous scaffold production in TERM applications has never been reported.

The aim of the present work was to evaluate the performance of various types of *C. reniformis* collagen extracts for the production of membranous scaffolds for TERM strategies. For this purpose, intact collagen fibrillar extracts were obtained, using four different methods that were previously described in literature [46,53–55]. Following their physico-chemical and biological characterization, the four extracts were used for the generation of collagenous membranes. Finally, the obtained membranes were characterized, and their biocompatibility evaluated, in order to establish the best extraction procedure to generate suitable tools for TERM applications.

#### **2. Results**

#### *2.1. Fibrillar Collagen Suspensions Extraction and Characterization*

Fibrillar collagen suspensions (FS) were extracted from *C. reniformis* tissue using four different methods, as indicated in Scheme 1, in order to evaluate the more suitable for collagenous membrane production. The total soluble proteins, collagen and GAGs content were then quantified in each FS. Table 1 reports the total yields obtained in each extract (from F1 to F4) in terms of FS, collagen, and GAGs, expressed as total mg/g of dry sponge tissue. Conversely, Table 2 shows a comparison of the composition by the quantification of total soluble proteins, collagen, and GAGs, which were calculated as weight percentage with respect to the total FS content of each extract. Moreover, a collagen/GAGs ratio (RC/GAG) was calculated in order to establish the optimal collagen/sugar composition for the further membrane production.

**Table 1.** Yield value for each of the four extraction procedures (F1–F4) of the total FS (column 3), collagen (column 4) and glycosaminoglycan (GAGs) (column 5), expressed as mg/g dry sponge tissue. <sup>1</sup> From [53]; <sup>2</sup> from [54]; <sup>3</sup> from [46]; <sup>4</sup> from [55].


**Table 2.** Percentage composition of total soluble proteins (column 2), GAGs (column 4) and collagen (column 5) in the four FSs expressed as mg/100 mg of total FS. Percentage composition of collagen (column 3) in the four FSs expressed as mg/100 mg of total soluble proteins. RC/GAG values (column 6) obtained from the ratio between the collagen content and GAG content in the four extracts.


**Scheme 1.** Extraction methods used. Schematic representation of the four extraction procedures used to obtain four different sponge fibrillar suspensions (F1–F4) as detailed in Section 4.1. Method 1 from [53]; method 2 from [54]; method 3 from [46]; and, method 4 from [55].

F1 extract was obtained, as described by Gross et al., 1956 [53], by an overnight trypsin digestion at 37 ◦C in order to remove non collagenous proteins as much as possible. The digested sponge tissue was then subjected to two rounds of water extraction at 5 ◦C. After the second step of water extraction, the resulted sponge tissue completely dissolved. The collagen extract resulted viscous and light colored (Figure 1A). The yield of F1 suspension had the lowest values with respect to the yields of the other FS extracts (Table 1, column 3). The yield in collagen had intermediate values, while the GAGs had the lowest values of extraction with respect to the values that were obtained in the other extracts (Table 1, columns 4 and 5, respectively). Conversely, in terms of percentages, F1 suspension resulted composed almost all of the soluble proteins, half of which were collagen (Table 2, columns 2 and 3, respectively). Finally, the RC/GAG was of 14.95, the highest value observed with respect to the other FS extracts (Table 2).

**Figure 1.** Fibrillar suspensions (FS) appearance and viscosity test. (**A**) Appearance of F1-F4 sponge FSs obtained through the four different extraction methods summarized in Scheme 1; (**B**) Flow sweep curves obtained in the viscosity tests of the four FSs by rheological measurements (for methods see Section 4.3.6). Curves were fitted by the Carreau-Gahleitner model (Mezger, 2006) as shown in Equation: <sup>η</sup>−η<sup>∞</sup> <sup>η</sup>0−η<sup>∞</sup> <sup>=</sup> <sup>1</sup> (1+(a· . γ) b ) <sup>P</sup> where η is the shear viscosity, η<sup>∞</sup> is the infinity-shear viscosity, η<sup>0</sup> is the zero-shear viscosity, a is the Carreau constant, b is the Gahleitner exponent, and P is the Carreau exponent. The experimental values of η<sup>0</sup> and η<sup>∞</sup> are shown in Table 3.

F2 extract obtained, as described previously by Diehl-Seifert et al. (1985) [54], was characterized by sponge tissue extraction under alkaline and reducing conditions in 8 M urea. The concentrated urea was used to improve fibre dissolving, whereas 2-mercapto-ethanol was used to ensure the intra and interchain disulfide collagen bridge break. At the end of the 24 h incubation at room temperature (RT), the sponge tissue was not completely dissolved and the insoluble fraction was removed by centrifugation. The collagen fibres were then precipitated by lowering the pH with the addition of acetic acid. The repeated centrifugation steps during the washings were necessary to neutralize the pH, but made the suspension less homogeneous than the others (Figure 1A), with several aggregates that necessitated a further homogenization step. It is to note that, in these extraction conditions the collagen retained in the F2 suspension had the lowest yield respect to the other extracts (Table 1). Furthermore, less than half of the F2 suspension was composed of soluble proteins, of which only the 16% was collagen (Table 2). Conversely, the percentage of GAGs extracted was the highest with respect to the other FSs, and consequently, the RC/GAG was the lowest (1.18), amounting to less than 1/10 of the F1 RC/GAG (Table 2, columns 4 and 6, respectively).

F3 and F4 extracts were obtained with a similar procedure [46,55], in both cases in fact, sponge fragments were incubated in an orbital shaker disk for four days at 17 ◦C in the presence of 2-mercapto-ethanol and EDTA. The only differences were that in F3 the pH was 7.4, whereas in F4, it was adjusted to 8.0, and that for F3 extraction, 1 M NaCl was used, while for F4, the NaCl concentration was only of 0.5 M. With the F3 extraction method, complete sponge tissue dissolution was observed at the end of the 4 day incubation. Conversely, using the F4 extraction method some insoluble tissue fragments were still present. F3 final extract resulted viscous and lightly colored, quite similar to F1, while F4 resulted clearer than the other FSs (Figure 1A). With these two extraction methods, the total yields of the two suspensions were the highest with respect to the other two methods (Table 1, column 3). Moreover, the F3 method showed the highest collagen and GAGs yields with respect to all of the other extraction procedures (Table 1, columns 4 and 5, respectively). For what concerns protein percentages, with the F3 method, the suspension obtained was composed of more than 60% of soluble proteins, of which nearly the 50% was collagen, while in the F4 method, less than 30% of the suspension was composed of soluble proteins, of which the collagen amounted to 60% (Table 2, column 2 and 3, respectively). Finally, the RC/GAG was higher in the F3 suspension, and was similar to the F1 value, and lower in the F4, pointedly nearly a half with respect to F1 and F3 (Table 2, column 6).

#### *2.2. Viscosity Evaluation*

Figure 1B reports the flow sweep curves fitted by the Carreau-Gahleitner model [56], while the values that were obtained by the experimental rheological measurements of η<sup>0</sup> and η<sup>∞</sup> are shown in Table 3.


**Table 3.** Rheological measurements values of η<sup>0</sup> and η<sup>∞</sup> of the four FSs from which were derived the flow sweep curves shown in Figure 1B.

All of the tested solutions showed a low viscosity and typical gel behaviour, with a yield point in the low-shear range; moreover, an evident shear thinning behaviour was observable. Indeed, the samples were characterized by a shear-dependent viscosity (i.e., η decreases quickly with the increasing of the shear rate). F1, F2, and F3 samples showed similar values of η∞, but significantly different values of zero-shear rate (η0); in particular, in steady state, F3 and F4 appeared to be, respectively, the most and the lowest viscous samples, while F1 and F2 showed intermediate values of η.

#### *2.3. Transmission Electron Microscopy Analysis of the FSs*

To evaluate the state of integrity of the collagen fibrils that was extracted with the four different methods, TEM analyses were performed on the four different FSs.

At the ultrastructural level, all of the negatively-stained FS samples contained long unbranched single fibrils that was characterized by a periodicity pattern (Figure 2). F1, F2, and F4 samples (panel A, B and D, respectively) showed small clots (white arrows) that were closely apposed to the fibrils and thin curled filaments of *bona fide* mucopolysaccharides (MPS), which were completely absent in the F3 preparation (panel C). Of note, the F2 sample showed many curled fibrils, indicating a possible damage of the structural integrity of the fibrils (not shown). No curled collagen fibrils were observed in the other preparations.

**Figure 2.** TEM analysis. Representative negatively stained transmission electron microscopy images of the four FSs (F1–F4) taken at 92,000× magnification (scale bar: 200 nm) with a CM10 Philips transmission electron microscope equipped with Megaview 3 camera and Olympus SIS iTEM software for digital image acquisition. (**A**) F1; (**B**) F2; (**C**) F3; and, (**D**) F4. All of the FSs show long unbranched banded fibrils of uniform size and periodicity band pattern. F1, F2, and F4 show small clots and filaments (white arrows) of putative mucopolysaccharides.
