Cell-Derived Extracellular Matrix for Tissue Engineering and Regenerative Medicine

Cell-derived extracellular matrices (CD-ECMs) captured increasing attention since the first studies in the 1980s. The biological resemblance of CD-ECMs to their in vivo counterparts and natural complexity provide them with a prevailing bioactivity. CD-ECMs offer the opportunity to produce microenvironments with costumizable biological and biophysical properties in a controlled setting. As a result, CD-ECMs can improve cellular functions such as stemness or be employed as a platform to study cellular niches in health and disease. Either on their own or integrated with other materials, CD-ECMs can also be utilized as biomaterials to engineer tissues de novo or facilitate endogenous healing and regeneration. This review provides a brief overview over the methodologies used to facilitate CD-ECM deposition and manufacturing. It explores the versatile uses of CD-ECM in fundamental research and therapeutic approaches, while highlighting innovative strategies. Furthermore, current challenges are identified and it is accentuated that advancements in methodologies, as well as innovative interdisciplinary approaches are needed to take CD-ECM-based research to the next level.


INTRODUCTION
The extracellular matrix (ECM) is the non-cellular component present in all connective tissues and has a composition specific for each tissue. It is comprised of a complex and highly organized three-dimensional macromolecular network of biomolecules. These include fibrous proteins (such as collagens) and glycosaminoglycan (GAG)-based components. Fibrous ECM components form the backbone of the polymer network, thereby providing shape/stability and tensile strength to tissues. They also regulate cell adhesion and support cell migration. GAG-based components fill the interstitial space, ensuring hydration and lubrication of tissues, and acting as a reservoir and modulator of cytokine signaling (Theocharis et al., 2016;Yong et al., 2020).
ECM-driven communication arises from a complex combination of biochemical, topological and biomechanical cues, facilitating a reciprocal dialogue with cells, which can respond via remodeling of the ECM. This multi-dimensional signaling enables the ECM to guide intricate cellular and tissue processes such as homeostasis, healing and regeneration (Kaukonen et al., 2017).

ECM AS A BIOMATERIAL
The ECM is a biomaterial designed by nature that underwent over 600 million years of material optimization (Ozbek et al., 2010). It serves as a blueprint for many man-made biomimetic biomaterials. Nonetheless, these materials represent oversimplified versions of the ECM that are not able to replicate its complex bioactivity (Kaukonen et al., 2017). As a result, ECM derived from decellularized tissues, remains one of the most successful biomaterials in clinics (Hussey et al., 2018).
Unfortunately, tissue-derived ECM faces various challenges to its clinical application. The limited availability of human cadaveric tissue leads to the use of animal tissue-derived ECM as an alternative source. Especially the incomplete decellularization of tissue carries the risk of disease transmission and immunological rejection. Some ECMs are plainly not available, since some specific tissues are hard to isolate (e.g., stem cell niches). Further, tissue-derived ECM is set in its composition, therefore cannot be customized in its bioactivity toward a specific application (Aamodt and Grainger, 2016).
As cell-derived ECM (CD-ECM) partially recapitulates the complex biological machinery of native tissue (Ahlfors and Billiar, 2007), it can address many of the tissue-derived ECM's limitations. It can be derived from human cell cultures by gentle decellularization to remove immunogenic components, while preserving its bioactivity. ECM-synthesizing cells can be standardized and pre-screened (Sharma et al., 2020), minimizing the risk of disease transmission. Deriving ECM in vitro provides the opportunity to select appropriate ECM-producing cell types, further modify them (e.g., genetically) and expose them to specific stimuli, thus enabling the creation of ECM with desired properties (Maia et al., 2020). CD-ECM is therefore an incredibly versatile material to be used in physiological studies and therapeutic approaches.

Facilitating ECM Deposition in vitro
Slow ECM assembling kinetics in vitro necessitate long cell culture periods up to several weeks to harvest sufficient CD-ECM amounts for the desired application (Bourget et al., 2012). This can be improved by adjusting culture conditions (Hoshiba, 2017).
The most essential supplement for robust ECM deposition is ascorbate, a cofactor of lysyl hydroxylase and prolyl hydroxylase, essential enzymes in collagen fibrillogenesis (Pinnell, 1985). Collagen type I is the most prominent ECM component and its deposition increases the overall yield of CD-ECM and improves its mechanical properties. Nonetheless, rapid degradation of ascorbate (Grinnell et al., 1989) calls for frequent media changes, thereby discarding the not-yet deposited ECM components. A stable form of ascorbate (2-phospho-L-ascorbate) can reduce the frequency of medium replacements .
The yield of deposited ECM can be amplified by introducing macromolecules, which emulate the crowded conditions present in vivo. The biophysical principle of macromolecular crowding (MMC) relies on macromolecules occupying space, thereby increasing the effective concentration of other molecules and the thermodynamic activity of the system. This has profound effects on protein folding, molecular interactions and enzyme kinetics . In particular, under MMC more ECM can be deposited within 1 week than after several weeks under non-crowded conditions. Most commonly used "crowders" are Ficoll, carrageenan, polyvinylpyrrolidone and dextran sulfate (Lareu et al., 2007;Lu et al., 2011;Blocki et al., 2015;Gaspar et al., 2019), albeit dextran sulfate was recently found to act as a precipitating agent, independent of MMC (Assunção et al., 2020).

Decellularization and Processing of CD-ECM
CD-ECMs are usually generated in a small format, permitting gentle decellularization methods with focus on maintaining architecture and bioactivity. Most methods use detergents, enzymes, chelating agents, mechanical approaches and combinations thereof ( Figure 1A; Woods and Gratzer, 2005;Faulk et al., 2014;Levorson et al., 2014;Gilpin and Yang, 2017). Complete decellularization is further achieved by removing genetic material with nucleases to prevent host immune reaction, as can be observed in tissue-derived ECMs (Crapo et al., 2011).

APPLICATIONS OF CD-ECM
Numerous applications have been explored for CD-ECMs including the improvement of cellular functions, seen in tailored cellular niches, the study of ECM in a physiological FIGURE 1 | Methodologies to generate CD-ECM in different formats. (A) CD-ECMs are synthesized by different cell types (i.e., MSCs and fibroblasts). Culture conditions are adjusted to facilitate ECM deposition by e.g., introducing MMC or hypoxia into cell culture. The assembled ECM is then gently decellularized, while preserving the ECM's integrity as much as possible. The resulting CD-ECM can be used in its original format or further processed. (B) The arrangement in which the ECM producing cells are cultured determines the format of the CD-ECM material. Which type of presentation is most advantageous depends on the desired application. Easiest decellularization can be achieved in 2D cultures. The resulting CD-ECM is suitable for coating of culture dishes and biomaterial surfaces or can be further processed and integrated with other biomaterials, such as hydrogels. Integration of CD-ECM with other materials provides the opportunity to combine the bioactivity of the ECM with desired geometries and mechanical properties. CD-ECM assembled in 3D recapitulates native cellular niches more closely. It can thus be utilized to engineer improved tissue models and ECM 3D scaffolds with desired geometries. Various techniques exist that enable the construction of 3D scaffolds based on CD-ECM. and pathophysiological context, and the application in tissue engineering and regenerative medicine (TERM) (Figure 2).

Stem Cell Niches
The emulation of the native cellular microenvironment in culture is a prerequisite to maintain the cells' phenotype and function. This is especially true for sensitive cell types, such as stem cells, which are known to undergo senescence and lose their stemness ex vivo (Hoshiba et al., 2016).
Various studies demonstrated that MSC-derived ECM can recapitulate the stem cell niche sufficiently to protect reseeded MSCs from oxidative stress, promote their proliferation, and conserve their stemness (Chen et al., 2007;Lai et al., 2010;Liu et al., 2016;Xing et al., 2020). CD-ECMs were also shown to maintain the native phenotype of neural progenitor cells Hoshiba et al., 2018), embryonic stem cells (ESCs) (Klimanskaya et al., 2005), periodontal ligament stem cells (Xiong et al., 2019) and hematopoietic stem cells (Prewitz et al., 2013). Furthermore, ECMs derived from younger MSCs were shown FIGURE 2 | CD-ECM applications in fundamental research, pathophysiological studies and renegerative medicine. The ease with which CD-ECM can be modified, makes it the ideal platform to study detailed ECM mechanisms or the role of cellular niches under physiological and pathophysiological conditions. Specialized engineered cellular niches can be further utilized to improve cellular functions in vitro, such as stemness. In TERM, CD-ECMs can be created with specific mechanical and biological properties to be used on their own (i.e., as vascular grafts) or to enhance the performance of (semi-) synthetic biomaterials.
to rejuvenate in vitro-aged and chronologically-aged MSCs (Pei et al., 2011;Sun et al., 2011;Lin et al., 2012). These effects were tightly linked to the biological profile of the ECM (reviewed in Sart et al., 2020).
Similarly to adult stem cells, CD-ECMs synthesized by differentiating ESCs were able to promote early differentiation of ESCs, even without external factors (Goh et al., 2013). ECM produced by an endoderm-inducing cell line and ECM from liver progenitor cells promoted differentiation of pluripotent cells into insulin-expressing pancreatic β-cells (Higuchi et al., 2010) and hepatic cells (Kanninen et al., 2016), respectively.
Hence, CD-ECMs can be utilized to tailor cell and tissuespecific niches to promote cellular functions and study cell-niche interactions in detail.

Engineering ECM in Disease
The ECM has a long-implicated role in disease development and progression, although the exact mechanisms often remain elusive. While the CD-ECM platform provides the opportunity to manipulate ECM and study it in detail, few studies utilized CD-ECM to study ECM mechanisms in disease (Raghunathan et al., 2018), most of them related to cancer.
It is currently well accepted that the tumor microenvironment plays a pivotal role in cancer cell behavior, including proliferation, invasiveness, metastasis and drug resistance (Serebriiskii et al., 2008). CD-ECMs provide the prospect to improve cancer models by recreating the cancer microenvironment using standard 2D, 3D cultures or more complex, organ-on-a-chip strategies (Gioiella et al., 2016;Kaukonen et al., 2017;Hoshiba, 2018). Indeed, culture of cancer cells on tumor CD-ECMs led to more physiologically relevant cancer cell phenotypes, as observed in various carcinoma (Serebriiskii et al., 2008;Eberle et al., 2011;Kaukonen et al., 2017), breast (Castelló-Cros et al., 2009;Hoshiba and Tanaka, 2013), and colon (Hoshiba, 2018) cancer models. Increased malignancy and drug resistance of cells was observed on invasive cancer CD-ECMs, in comparison to non-invasive cancer CD-ECMs (Hoshiba and Tanaka, 2013;Hoshiba, 2018). In contrast, upon culture on MSC-derived ECM, cancer cells proliferated less (Marinkovic et al., 2016) and showed reduced tumorgenecity upon implantation . Differences in cancer cell behavior were attributed not only to the biochemical composition of the tumorassociated ECM, but also to changes in stiffness (Kaukonen et al., 2016;Hoshiba, 2018) and a decreased cell adhesion (Hoshiba and Tanaka, 2013).

Engineering and Characterization of CD-ECM to Study ECM Physiology
The ease of manipulating CD-ECM in vitro provides the opportunity to examine the reciprocal relationship between cells and their ECM.
Biochemical ECM re-engineering could be achieved through direct addition of functional groups (Xing et al., 2015) or exogenous factors (Sart et al., 2020), genetic modification (Higuchi et al., 2010) or growth factor stimulation of ECM-synthesizing cells (Wolchok and Tresco, 2012). Other changes in culture conditions, such as hypoxic cultures, were also shown to affect ECM properties and bioactivity (Hielscher, 2013).
Mechano-physical re-engineering could be achieved by culturing ECM-secreting cells in 3D sacrificial hydrogels (Yuan et al., 2018), on micro-molds (Schell et al., 2016), and micro-and nano-grooves (Ozguldez et al., 2018;Almici et al., 2020;Yang et al., 2020), forcing cell reorganization and leading to ECM assemblies with unique architectures (i.e., parallel fiber alignment). ECM postprocessing, such as cross-linking, could further alter ECM stiffness (Subbiah et al., 2016) or the overall presentation of CD-ECM.In particular, cross-linking of pepsin-solubilized CD-ECM with genipin resulted in the formation of hydrogels (Nyambat et al., 2020), Changes in biochemical and mechano-physical properties of the ECM let to changes in gene expression and behavior of reseeded cells (Kim et al., 2015;Ozguldez et al., 2018;Sart et al., 2020).
CD-ECM characterization and correlation with specific bioactivities can contribute to the mechanistic understanding of the ECM. ECM ultrastructure can be generally studied by scanning electron microscopy or atomic force microscopy (Kaukonen et al., 2016;Raghunathan et al., 2018). The latter method can also be used for biomechanical characterization (Prewitz et al., 2013;Assunção et al., 2020). Identification of proteins of interest is best performed by antibody-based assays such as immunocytochemistry or western blotting (Sart et al., 2020). Proteomic analysis based on mass spectroscopy enables the simultaneous identification of many components, however also faces challenges based on the insolubility and high complexity of the ECM (Ragelle et al., 2017;Senthebane et al., 2018;Silva et al., 2019). Furthermore, additional methods, such as Raman microscopy, can be used for biochemical characterization (Brauchle and Schenke-Layland, 2013).

CD-ECM Applications in TERM
CD-ECM uses for TERM have been increasingly explored, either with CD-ECM alone or integrated in biomaterials. 3D scaffolds purely composed of CD-ECM were produced by decellularizing stacked cell sheets (McAllister et al., 2009) and pellets (Zwolinski et al., 2011), or depositing ECM in sacrificial materials, such as hollow tubes (ECM fibers) (Roberts et al., 2017) and foams (ECM porous scaffolds) (Wolchok and Tresco, 2010; Figure 1B).
For applications that require specific mechanical properties of the biomaterials, CD-ECM was integrated with synthetic materials, forming hybrid scaffolds (Schenke-Layland et al., 2009;Carvalho et al., 2019b;Sart et al., 2020). Hybrid materials met mechanical requirements, while providing adequate biochemical stimuli, thus facilitating implant integration and functionality (Silva et al., 2020). Commonly, CD-ECM was utilized as a coating by simply decellularizing cells on the biomaterial surface Junka et al., 2020), although solubilized CD-ECM was also used as a coating (Decaris et al., 2012). A more sophisticated approach introduced azide-modified monosaccharides into culture media, which subsequently were incorporated into the ECM. The CD-ECM could then be covalently "clicked" to material surfaces (Ruff et al., 2017). Alternative approaches directly incorporated CD-ECM into the biomaterial during synthesis (e.g., electro-spinning) (Schenke-Layland et al., 2009;Carvalho et al., 2019b).
CD-ECMs based biomaterials were mainly investigated for skeletal and cardiovascular repair, although other applications such as in skin (Suhaeri et al., 2018) and peripheral nerve repair (Gu et al., 2017) were also explored.
Therapeutic approaches targeting cartilage repair mainly utilized 3D scaffolds purely composed of CD-ECM (Jin et al., 2007;Tang et al., 2013Tang et al., , 2014 or CD-ECM-loaded hydrogels (Yuan et al., 2013). Indeed, 3D scaffolds of chondrocyte-and MSC-derived ECM reseeded with chondrocytes induced ectopic hyaline-like cartilage formation in vivo (Jin et al., 2007;Tang et al., 2013). When applied to an osteochondral defect together with bone marrow stimulation, autologous MSC-derived ECM could enhance cartilage repair (Tang et al., 2014). In another study, a protective effect on the degenerating cartilage could be demonstrated, when collagen I microspheres containing nucleus pulposus CD-ECM and MSCs were injected into a rabbit degenerative disc model (Yuan et al., 2013).
Cardiac patches were composed of fibroblast ECM alone (Schmuck et al., 2014) or combined with a polyvinyl alcohol sheet, resulting in a stretchable scaffold for cell delivery. Application of the latter in a rat myocardial infarct model resulted in improved cardiac remodeling (Kim et al., 2019).
A cardiac valve prototype containing vein-derived fibroblast ECM was implanted in a non-human primate. Albeit valve functionality was reduced, there was a significant improvement in repopulation by host cells, when compared to decellularized human heart valves (Weber et al., 2013).
McAllister et al. (2009) utilized partially devitalized autologous fibroblast/endothelial CD-ECM sheets to form vascular access grafts for dialysis patients. Complete remodeling and repopulation of CD-ECM occurred, although diffuse dilation of the graft was observed (McAllister et al., 2009).
In order to improve this low graft resistance, Syedain et al. (2014) stimulated tubular fibroblast cultures in a pulsed-flowstretch bioreactor. Upon implantation of the decellularized graft into the femoral artery of sheep, no dilation was observed. Once completely recellularized, the grafts resembled native vessels in terms of cellular composition, ECM architecture and mechanical properties (Syedain et al., 2014).

CONCLUSION AND OUTLOOK
Although CD-ECM was continuously explored for over three decades and many safety concerns associated with tissue-derived products can be addressed, relatively slow advancements were made over the years. This can be partially attributed to the low amounts of CD-ECM that can be harvested in vitro, indicating that strategies for upscaling processes as well as manufacturing of larger 3D constructs need to be developed.
In addition, most TERM approaches used unmodified ECM from MSCs or tissue-specific cell types to induce cellular responses in vitro and in vivo. And although various approaches on how to re-engineer the CD-ECM are proposed, relatively few are applied to address scientific questions or to manufacture biomaterials with enhanced desired bioactivities. The reason for the limited progress can be partially attributed to our restricted fundamental understanding of the ECM. Hence, functional studies in combination with CD-ECM characterization will have to be adopted. Another reason is that re-engineering approaches are mainly focused on biological manipulation. Research at the interface to other disciplines such as materials science is indeed required to enable further evolvement of the CD-ECM research field. Future applications could focus on bio-inks with tailormade bioactivities for 3D bioprinting or improved biomimetic cell niches in organ-on-a-chip approaches.
In conclusion, CD-ECM based research is far from its full potential. Advancements in methodologies as well as innovative interdisciplinary approaches are needed to pave the way for an exciting next generation of CD-ECMs for basic research and therapeutic approaches.

AUTHOR CONTRIBUTIONS
All authors contributed to the elaboration of this review.

FUNDING
This study was financially supported by a laboratory start-up grant (8508266) from the Chinese University of Hong Kong (CUHK) and by the Innovation and Technology Commission-Hong Kong (Innovation and Technology Commission of Hong Kong Special Administrative Government) (ITS/116/19) and the Postdoctoral Hub for ITF projects (PiH/007/20).

ACKNOWLEDGMENTS
MA would like to acknowledge School of Biomedical Sciences (The Chinese University of Hong Kong) for her postgraduate scholarship.