Role of Region-Specific Brain Decellularized Extracellular Matrix on In Vitro Neuronal Maturation

Abstract

Recent advancements in tissue engineering suggest that biomaterials, such as decellularized extracellular matrix (ECM), could serve to potentiate the localization and efficacy of regenerative therapies in the central nervous system. Still, what factors and which mechanisms are required from these ECM-based biomaterials to exert their effect are not entirely understood. In this study, we use the brain as a novel model to test the effects of particular biochemical and structural properties by evaluating, for the first time, three different sections of the brain (i.e., cortex, cerebellum, and remaining areas) side-by-side and their corresponding decellularized counterparts using mechanical (4-day) and chemical (1-day) decellularization protocols. The three different brain subregions had considerably different initial conditions in terms of cell number and growth factor content, and some of these differences were maintained after decellularization. Decellularized ECM from both protocols was used as a substrate or as soluble factor, in both cases showing good cell attachment and growth capabilities. Interestingly, the 1-day protocol was capable of promoting greater differentiation than the 4-day protocol, probably due to its capacity to remove a similar amount of cell nuclei, while better conserving the biochemical and structural components of the cerebral ECM. Still, some limitations of this study include the need to evaluate the response in other biologically relevant cell types, as well as a more detailed characterization of the components in the decellularized ECM of the different brain subregions. In conclusion, our results show differences in neuronal maturation depending on the region of the brain used to produce the scaffolds. Complex organs such as the brain have subregions with very different initial cellular and biochemical conditions that should be considered for decellularization to minimize exposure to immunogenic components, while retaining bioactive factors conducive to regeneration.

Impact statement

The present study offers new knowledge about the production of decellularized extracellular matrix scaffolds from specific regions of the porcine brain, with a direct comparison of their effect on in vitro neuronal maturation. Our results show differences in neuronal maturation depending on the region of the brain used to produce the scaffolds, suggesting that it is necessary to consider the initial cellular content of the source tissue and its bioactive capacity for the production of an effective regenerative therapy for stroke.

Introduction

An increase in the elderly population worldwide and an associated higher incidence of noncommunicable diseases (NCDs), especially in developing countries, is driving demand for therapies that can address degenerative tissue loss.¹ Stroke is the leading neurological disorder in the NCD list and the second cause of death around the world after ischemic heart disease.^2,3 It is also the main neurological cause of adult disability in terms of disability-adjusted life-years, and thus, carries a heavy societal and economic burden.^2,4 Although relatively small in number, the few clinical trials for regenerative therapies in stroke are using stem cells injected intravenously or directly into the brain of the patient, showing good safety and promising but limited efficacy for these therapies.^5–7 Stroke is a convenient neurological disease target for regenerative therapies because of its burden on the population and its localized nature within the brain. Recent advancements in tissue engineering suggest that biomaterials, such as organic scaffolds and hydrogels, could serve to potentiate the localization and efficacy of regenerative therapies in the central nervous system (CNS).^8–11

The use of tissue-derived biomaterials based on extracellular matrix (ECM) is not a new concept, with one of the first reports published in the late 40s,¹² but it still occupies an important place in the regenerative medicine field due to its relative bioactive advantages over synthetic and single-biomolecule materials. Several extensive reviews about methods for decellularization, sources of tissues, and areas of application have been published in the literature.^13–17 Since it can be derived from the tissue of interest, ECM-based biomaterials are capable of providing tissue-specific cues¹⁴ and controlling several key cellular processes such as chemotaxis, mitogenesis, and differentiation.^18,19 In most cases, the harvested raw material needs to be processed or “decellularized,” to remove immunogenic intracellular components in the native tissue, while preserving the structure and desirable biochemical cues of the ECM. A functional evaluation criterion does not exist to define a successful decellularization process, but attempts have been made to establish residual DNA content as the critical, but indirect, immunogenicity threshold for in vitro and in vivo studies.¹⁴

While ECM products derived from amniotic, dermal, and urinary bladder tissue have found several clinical applications,²⁰ ECM derived from the brain is still confined to academic research laboratories. Possible reasons for this paucity include the complexity of treating neurological disorders, the inherent structural and biochemical complexity of the nervous system, and the mechanical properties of cerebral tissue, as the brain is a viscoelastic material and one of the softest tissues in the body.²¹ Cerebral ECM is mainly composed of hyaluronic acid, glycosaminoglycans, and proteoglycans, as opposed to fibrous proteins (e.g., collagen and fibronectin) that characterize connective tissue composition.²² As a result, successful decellularization for cerebral tissue is particularly challenging to achieve compared to other tissues because of its loose mechanical structure leading to more material loss in the process.

The unrealized potential of most decellularized ECM approaches may hinge around the lack of in-depth understanding of the basic cellular and molecular mechanisms triggered by the integration of a complex ECM-based biomaterial.²³ While the regenerative capacity of these scaffolds has been correlated with their ability to modulate the host macrophage phenotype from an M1 to an M2 response,²⁴ the growth factors or structural cues in the ECM that play a direct role in the cellular response toward regeneration remain to be elucidated. To complicate the situation, changes in any of the variables of a decellularization protocol can result in final products with vastly different compositions, and most research groups use their own customized protocols.^25–27 Therefore, we do not have a clear picture of what factors are critically required after a decellularization process, what can be lost or removed, and how to assess these elements in a reproducible way. Even fundamental questions about the advantages of using tissue-specific matrices have not been answered conclusively. For example, no studies to our knowledge have evaluated the cellular response to cerebral ECM from the cortex compared to that of the cerebellum or the remaining areas. Such information could help tailor the design of regenerative therapies for insults in these different subregions of the brain. Information about tissue specificity and corresponding bioactivity is valuable both for more classic decellularized matrix scaffold preparation,²⁵ as well as newer applications in tissue engineering such as three-dimensional (3D) printed scaffolds.²⁸ Thus, the field continues its pursuit of the optimal biomaterial for tissue regeneration.^29,30

Although less studied compared to other tissues, cerebral decellularized ECM has been capable of differentiating PC12 cells and different types of stem cells into neuronal cells when exposed to the biomaterial as a substrate^25,26 or as a soluble factor.³¹ The decellularization protocols used in these studies have been distinctly customized, probably leading to different biochemical compositions, but a parameter in common is that they process the entire brain as a homogeneous organ. However, the cellular content and matrix composition of the different subregions of the brain can vary considerably,^32,33 which could be important when trying to develop the most effective regenerative therapies for global and local ischemic stroke. In this study, we compare two brain decellularization methods based on mechanical or chemical forces, and develop a novel in vitro model to test cell response to slight changes in decellularized ECM properties by processing the brain in three separate subregions: cortex, cerebellum, and the remaining areas (limbic/medulla system). Our main goal is to better understand the differences between the ECM from the three subregions that could lead to particular cell responses. We report here the characterization of the resulting decellularized tissues and their capacity to promote neuronal attachment and growth through in vitro studies with PC12 cells.

Materials and Methods

Collection of tissue from animals

All animal tissue handling was performed with approval from the Institutional Animal Care and Use Committee (IACUC) of INDICASAT AIP, which is in compliance with the NIH Guide for Care and Use of Laboratory Animals. Porcine brains were collected fresh at the Macelo S. A. abattoir, in Panama City, Panama. The animals were euthanized through electric shock stunning and cardiac arrest using electrodes set at 0.5 A and 220 V. Brain tissue is a byproduct of the operations of the abattoir, so the material used does not pose a risk to any additional animals. The pigs selected for the study were female Sus scrofa domesticus, around 6 to 8 months old and ∼118 kg average weight. Handling of all animals in the abattoir was supervised from arrival until sacrifice by two official veterinary physicians, one from the Ministry of Health and the other from the Ministry of Agriculture, to ensure proper health record, humane handling, and good manufacturing practices in the abattoir.

Brain halves were removed from the skull using the foramen magnum as the inferior limit of the medulla oblongata, immediately after animals were euthanized and cut in half. The tissue was either kept fresh for isotropic fractionator or flash-frozen in crushed dry ice for decellularization, placed in sealable plastic bags, and transported to the laboratory in a cooler with ice or dry ice, respectively, for further processing.

Preparation of decellularized extracellular matrix biomaterial

The flash-frozen native tissue was stored in a freezer at −80°C for at least 24 h to include at least one freeze-thaw cycle before processing, to aid with the decellularization process. Once thawed, the porcine brains were similarly dissected, with some interfacial regions discarded, into the three regions of interest: cortex, cerebellum, and remaining areas (limbic/medulla system) (Fig. 1A). All dissections were minced into small pieces of less than 1 cm³, transferred to pyrogen-free 50 mL conical tubes (Sarstedt, Nümbrecht, Germany), and processed separately in subsequent steps (Supplementary Fig. S1). Two previously described methods were used to produce decellularized ECM for the material characterization and in vitro studies, one based on mechanical forces and the other on chemical forces^25,26 (see Supplementary Data for details).

FIG. 1.

Overview of the decellularization process and quantification of DNA levels in native and decellularized tissues. (A) Porcine brains were dissected into cortex, cerebellum, and remaining areas. (B) Each section underwent the decellularization process separately, was lyophilized, and, finally, subjected to microbiological and mycological analyses to assess sterility of the biomaterial. Subsequently, DNA levels were measured in (C) cortex, (D) cerebellum, and (E) remaining areas before decellularization and every day of processing in the 4-day protocol, as well as (F) after the 1-day protocol for the same brain sections. *refers to a statistically significant p-value below 0.05 versus native from the same brain subregion; ^#refers to a statistically significant p-value below 0.05 versus native from another subregion. Color images are available online.

Microbiological analysis

See Supplementary Data for details.

DNA quantification assay

DNA extraction from native and decellularized brain samples was done following a previously published protocol, with some minor modifications.²⁶ Initially, 1 mg of lyophilized tissue was resuspended in buffer containing 1% sodium dodecyl sulfate, 100 mM NaCl (Merck), 10 mM Tris (Sigma-Aldrich, St. Louis, MO), 25 mM EDTA (Merck), and 0.1 mg/mL proteinase K (QIAGEN, Duesseldorf, Germany), and samples were incubated at 37°C for 60 h. Then samples were treated with phenol:chloroform:isoamyl alcohol (AppliChem, Maryland Heights, MO) and centrifuged to remove proteins. The aqueous layer was collected and placed in a new tube with ice-cold 3 M sodium acetate (AppliChem) (0.1 volumes) and 100% ethanol (2.5 volumes). Samples were stored at −80°C overnight to facilitate DNA pellet precipitation. The DNA pellet was washed with 70% ethanol, dried, resuspended with nuclease free water (Sigma-Aldrich) and stored at −20°C until analysis. Double-stranded DNA was quantified using the Quant-iT™ PicoGreen^® assay (Invitrogen) as per kit instructions.

Protein content assay

All tissue samples were pretreated as described previously to facilitate the measurement of proteins.²⁶ In brief, 40 mg of lyophilized native and decellularized ECM samples were suspended in 600 μL of urea-heparin extraction buffer consisting of 2 M urea (Sigma-Aldrich) and 5 mg/mL heparin (Sigma-Aldrich) in 50 mM Tris with protease inhibitors (Roche Applied Science, Penzberg, Germany) containing 1 mM Phenylmethylsulphonyl Fluoride (PMSF), 5 mM Benzamidine, and 10 mM N-Ethylmaleimide at pH 7.4. The extraction mixture was rocked at 4°C for 24 h and then centrifuged at 12,000 g for 30 min at 4°C. Supernatants were collected and 6 mL of freshly prepared urea-heparin extraction buffer was added to each pellet. Pellets with extraction buffer were again rocked at 4°C for 24 h, centrifuged at 12,000 g for 30 min at 4°C, and supernatants were collected and mixed, respectively, with those previously collected. Protein content was determined with the bicinchoninic acid (BCA) protein assay (Pierce, Rockford, IL) following the manufacturer's protocol.

Brain-derived neurotrophic factor (BDNF) and nerve growth factor (NGF) proteins were quantified with enzyme-linked immunosorbent assays (ELISA). ELISA for BDNF and NGF (Duoset ELISA; R&D Systems, Minneapolis, MN) were performed following manufacture instructions. For BDNF and NGF, 15 mg of lyophilized native and decellularized tissue was added per mL of RIPA lysis buffer (Santa Cruz Biotechnology, Dallas, TX) plus protease inhibitors. Samples were homogenized on ice using an Omni Tissue Homogenizer (Omni International, Kennesaw, GA), centrifuged at 12,000 g at 4°C for 15 min. Supernatants were individually collected, placed in new tubes, and stored at −80°C until use.

Histology

See Supplementary Data for details.

Scanning electron microscopy

See Supplementary Data for details.

Surface coating experiments

For in vitro experiments, decellularized tissue was pretreated and handled as described previously.^26,34 In brief, 1 mg lyophilized ECM was comminuted (maximum dimension <1.0 mm), solubilized by digesting it with 1 mg/mL pepsin (Sigma-Aldrich) in 0.01 N HCl (Thermo Fisher Scientific), neutralized to pH 7.4 with 0.1 N NaOH (Merck) isotonically balanced with 10 × phosphate-buffered saline (PBS), diluted to the desired stock concentrations with 1 × PBS, and stored at −20°C until use. In vitro experiments were performed using 12 mm sterile coverslips (Neuvitro) in 24-well plates. Coverslips were rinsed twice with injectable water, covered with 250 μL poly-l-lysine (Trevigen, Gaithersburg, MD), as control substrate, or digested ECM samples diluted to the final experimental concentrations, and left at 4°C overnight. The following day, the excess coating solution was discarded, coverslips were rinsed twice with PBS, and left drying at 37°C for 2 h before starting experiments. To characterize the decellularized ECM in vitro model (i.e., surface adsorption process), two experimental strategies were used:

• Fluorescence Gradient: Lyophilized decellularized ECM was prepared at different concentrations: C1: 1 mg/mL, C2: 250 μg/mL and C3: 100 μg/mL. Aliquots of 250 μL of the different experimental groups were combined with 10 μL FM1-43 (Thermo Fisher Scientific), a green fluorescent dye, and served in each respective well for 5 min at 37°C. Then, the samples were left overnight in 24-well plate at 4°C. The following day, the excess solution was discarded, and the coverslips were rinsed with PBS twice, then dried at room temperature for 2 h and imaged with an Axio Scope fluorescence microscope (Carl Zeiss).

• Protein Surface Adsorption: the amount of protein adsorbed on the surface of the coverslips was determined by calculating the difference in protein content in the coating solutions before (pre-) and after (post-) coating.³⁴ Protein content was measured using the BCA protein content assay. Samples were prepared at a concentration of 100 μg/mL, analyzed for initial protein content (precoating), and served onto coverslips in a 24-well plate for 2 to 4 h at 37°C. After incubation, the supernatants were collected and evaluated for final protein content (postcoating) as well.

In vitro model: PC12 cell: biomaterial interaction and differentiation analysis

PC12 cells are a pheochromocytoma cell line cloned from the rat adrenal medulla that has become one of the most used and cited culture systems for neurodegeneration studies as a unidirectional in vitro model of neuronal proliferation, differentiation, and function.^35–37 An important feature of PC12 cells is that, under nonstimulating conditions, they grow in colonies of spherical shape that remain in suspension without attaching to the surface. However, when exposed to stimulating agents, specifically NGF, they respond with a dramatic change in phenotype and acquire a number of properties characteristic of sympathetic neurons.³⁵ Following similar protocols as previous studies that have used the PC12 cell line to evaluate the effectiveness of ECM adhesion proteins and neurotrophic factors,³⁸ as well as more complex decellularized ECM from different sources, including the brain,²⁶ we used these cells as an in vitro model to assess the cell–biomaterial interface and the differences between three different brain subregions. PC12 cells (cat# CRL-1721; ATCC, Manassas, VA) were expanded in 25 mm² Falcon^® culture flasks (Corning, Corning, NY) with growth medium made of complete RPMI-1640 medium (cat# 30-2001; ATCC) containing 5% fetal bovine serum (cat# 30-2020; ATCC), 10% heat-inactivated horse serum (cat# 30-2040; ATCC), and 1% penicillin/streptomycin. Cultures were maintained according to standard protocols at 37°C in a 95% humidified incubator with 5% CO₂. Only passages above fifth generation were used for experiment purposes.

For qualitative morphological evaluation of cell–biomaterial interactions, PC12 cells were plated on coverslips coated with poly-l-lysine as control or decellularized ECM (100 μg/mL) from the different brain sections, at a density of 30,000 cells/cm². Cells were cultured for 24 h in growth medium and immediately harvested for morphological evaluation with scanning electron microscopy (SEM), as described previously.³⁹

For cell viability and differentiation studies, PC12 cells were plated at a density of 35,000 cells/cm² on poly-l-lysine. PC12 cells were treated with a negative control (i.e., PBS), positive controls (e.g., NGF), or the different biomaterials as soluble factors added to the growth medium (i.e., differentiation media). Viability was assessed 24 h after exposure to the differentiation media with the Calcein AM Cell Viability Assay (R&D Systems) and propidium iodide (R&D Systems) for live-dead evaluation. For differentiation analysis, all groups, including the negative control, positive controls (NGF and BDNF at a concentration of 50 ng/mL each), and experimental groups, stimulated with the biomaterial, brain decellularized ECM as soluble factor were cultured for 5 days and fed every 48 h. At the fifth day, cells were fixed in 4% paraformaldehyde and stained with Phalloidin for morphological visualization under the Axio Scope fluorescence microscope and evaluation using the ImageJ software (NIH, Bethesda, MD).

Statistical analyses

Data from all experiments are presented as the mean ± one standard deviation of all the measurements performed on different samples of the same specimen type. Data from experiments examining cell response are presented as the mean ± one standard deviation, with all experiments having three independent cultures per variable and repeated at least twice to ensure reproducibility. Significant differences were determined between all groups shown by one-way analysis of variance with Tukey-Kramer post hoc analysis. A p-value below 0.05 was considered to indicate a statistically significant difference.

Results

Macroscopic analysis of decellularization process

To evaluate tissue-specific effects of decellularized ECM on cell response in a novel in vitro model, each subregion of the brain (i.e., cortex, cerebellum, and remaining areas) was sectioned and processed separately to perform either the 4- or 1-day decellularization protocol (Fig. 1A). Visually, the native and decellularized tissues were evidently different (Fig. 1B). The native tissue exhibited a typical pink hue that was less pronounced in the white matter of the cortex, the branching pattern of the cerebellum, and most of the remaining areas. However, during either one of the decellularization protocols, the tissues became whiter and lost their distinct morphological features, suggesting removal of molecular and cellular components.

No evident differences were found between the decellularized ECM samples processed for 1 or 4 days, except for a reduction in volume (i.e., more tissue lost) in the samples treated for 4 days, particularly in the cerebellum section (Supplementary Fig. S2). After the decellularization process was completed, a portion of the samples was used for histological analysis, while the remaining material was lyophilized to facilitate handling for the rest of the molecular assays. Microbiological and mycological analyses showed that most samples did not present any evident colony growth (i.e., sterility >90% of samples) for any of the native or decellularized specimens, which was a requisite for the samples to be deemed suitable for further evaluation (Supplementary Fig. S3). Still, the 4-day protocol registered slightly more incidents of microbiological contamination (i.e., lower sterility percent) when compared to the native tissue and the 1-day protocol.

Evaluating the efficiency of decellularization

The DNA content in native brain tissue varied between the different subregions and was reduced considerably after decellularizing with either the 4- or 1-day protocols (Fig. 1C–F). The cerebellum, which was the brain subregion with the highest cellular content, exhibited the greatest decrease in DNA concentration, regardless of the decellularization method. Using the 4-day method, we found that the DNA concentration in the native cortex decreased from an initial level of 1029 ng/mg dry weight to 118, 100, 82, and 54 ng/mg after 1, 2, 3, and 4 days of treatment, respectively (Fig. 1C). Similarly, the cerebellum exhibited a decrease from 1912 ng/mg down to 64 ng/mg after the fourth day, while the remaining areas went from 802 ng/mg to a final concentration of 58 ng/mg (Fig. 1D, E). With the 1-day method, the DNA content in the cortex, cerebellum, and remaining areas decreased to a final concentration of 97, 76, and 91 ng/mg, respectively (Fig. 1F). Our results showed that the 4-day method eliminated most of the DNA content after the first day of treatment (i.e., 76% in the cortex, 71% in the cerebellum, and 68% in the remaining areas), achieving a final reduction of 95% in the cortex, 97% in the cerebellum, and 94% in the remaining areas after the fourth day. In contrast, the 1-day method was more efficient at reaching a lower DNA concentration during the initial 24 h, resulting in a DNA reduction of 89% in the cortex, 90% in the cerebellum, and 90% in the remaining areas, but the final DNA content was higher than for the 4-day protocol. Although the initial conditions of the three brain subregions were variable (i.e., cerebellum had much higher initial DNA concentration), their behavior under either of the two decellularization protocols was similar and both methods yielded considerably low DNA levels of less than 100 ng DNA/mg of dry weight.

Morphological and biochemical assessment of the decellularized ECM

Electron micrographs acquired by SEM offered a closer look from the surface topography of native brain ECM and after applying the 1- or 4-day protocols (Supplementary Fig. S4). The native tissue exhibited a complex surface topography for all three sections (cortex, cerebellum, and remaining areas), with a combination of fibers and other protuberances of micro- and submicro-scale size (Supplementary Fig. S4A–C). After both decellularization protocols, the surface topography was mostly conserved without any evident differences between the different brain subregions. However, the 4-day protocol did show smoother, less defined surfaces (Supplementary Fig. S4D–F) than the 1-day protocol (Supplementary Fig. S4G–I), especially when comparing cerebellum and the remaining areas.

The integrity of the brain ECM protein ultrastructure was evidently affected by the decellularization protocols, and the impact seemed to be directly correlated with the number of processing days (Supplementary Fig. S5). Starting from the sample preparation steps, it was clear that the decellularized tissue, either from the 4- or 1-day protocol, was more fragile and prone to rupturing while sectioning and staining. Histological evaluations with Coomassie Blue staining and Hematoxylin counter-staining validated the efficacy of decellularization by highlighting the presence of nuclei in the native tissues (Supplementary Fig. S5A–C), which were almost entirely absent in the decellularized samples of both protocols (Supplementary Fig. S5D–I). These results were corroborated with diamidino-2-phenylindole (DAPI) staining and cell nuclei quantification, resulting in a similar drop in the number of nuclei after decellularization (Supplementary Fig. S6). The Coomassie Blue optical micrographs of the 4-day protocol show that the ECM still contains proteins, although with less ultrastructural organization and with several voids spaces in between the tissue that reflect the expected loss of material during processing (Supplementary Fig. S5D–F). Similarly, the 1-day protocol also exhibits a loss of ultrastructural protein organization, but not as pronounced as the 4-day protocol (Supplementary Fig. S5G–I). We quantified the void spaces in the brain ECM of the different brain sections using image analysis software and found around 10% higher levels of void space (i.e., less ECM ultrastructure preservation) in the specimens proceeding from the 4-day protocol compared to the 1-day method (Supplementary Fig. S5J). No evident differences were found when comparing the three brain subregions within each of the decellularization protocol.

Both decellularization protocols reduce the original protein content considerably, in direct proportion to the days of processing, but their effect on certain growth factors seems to be specific. Using the BCA biochemical assay, we quantified the loss of protein during decellularization. The 4-day protocol reduced the protein levels in the cortex, cerebellum, and remaining areas from 1184, 1802, and 1413 μg/mL to 811, 631, and 573 μg/mL; while the protein content for the 1-day protocol in the cortex, cerebellum, and remaining areas decreased to 993, 720, and 896 μg/mL, respectively. The results of the 1-day decellularization protocol showed higher protein content compared to the 4-day protocol, especially in the cortex and the remaining areas, which maintained 15.4% and 23.1% more protein, respectively (Fig. 2A).

FIG. 2.

Protein content analysis of native and dECM specimens from porcine brains. (A) Total protein content quantified with the bicinchoninic acid assay before (native) and after decellularization with the 4- and 1-day protocols. (B) NGF and (C) BDNF levels were quantified by ELISA on native and decellularized specimens. *refers to a statistically significant p-value below 0.05 versus native from the same brain subregion; ^#refers to a statistically significant p-value below 0.05 versus native from another subregion; ^%refers to a statistically significant p-value below 0.05 versus 4-day protocol from another subregion; ^@refers to a statistically significant p-value below 0.05 versus the same protocol from another subregion. BDNF, brain-derived neurotrophic factor; dECM, decellularized extracellular matrix; ELISA, enzyme-linked immunosorbent assays; NGF, nerve growth factor.

We also quantified the presence of two important neurotrophic factors involved in the neuronal differentiation of the PC12 cell line model we chose to test the decellularized ECM: NGF and BDNF (Fig. 2B, C). The quantification of NGF in the different sections of the brain showed that the cerebellum had almost twice the amount of NGF than the other sections analyzed, with levels of 8 ng/mg dry weight, while the cortex had 4 ng/mg dry weight, and the remaining areas had 5 ng/mg dry weight. However, NGF seemed to be lost during decellularization, as it was not detectable after either one of the protocols (Fig. 2B). BDNF absolute levels were higher than NGF in all brain sections, with native tissue levels of 28, 80, and 156 ng/mg dry weight for cortex, cerebellum, and remaining areas, respectively (Fig. 2C). Interestingly, we observed an increase in BDNF content after decellularization, particularly in the remaining areas, with the 4-day protocol exhibiting levels of 32, 104, and 235 ng/mg dry weight and the 1-day protocol showing levels of 90, 92, and 666 ng/mg dry weight for cortex, cerebellum, and remaining areas, respectively.

Establishing an in vitro model for decellularized ECM coatings

To evaluate the biological effects of the decellularized ECM on cell culture both as a soluble factor and a cell culture substrate, we had to characterize the in vitro model and the process of biomaterial adsorption to coverslips for cell attachment. The capacity of the decellularized ECM to adsorb to the surface of coverslips was assessed with the fluorescent marker FM1-43 (Fig. 3A). At low magnification in the fluorescence microscope, the presence of adsorbed cortex from the 1-day protocol on the surface of coverslips can be evidently visualized when a solution of 100 μg/mL is combined with FM1-43 (Fig. 3B) as opposed to when the same coating concentration is not fluorescently tagged (Fig. 3C). High-magnification micrographs of coatings from a concentration gradient (i.e., 100 to a 1000 μg/mL) reveal a direct correlation between the concentration of the coating solution and the fluorescence level detected on the adsorbed coating (Fig. 3D). The high-concentration coating (i.e., 1000 μg/mL) seemed to form heterogeneous clumps on the surface, while the low-concentration coating (i.e., 100 μg/mL) created a relatively homogeneous layer on the surface of the coverslip. Similar results were found with the other two brain subregions and the 4-day protocol (data not shown).

FIG. 3.

Surface adsorption assessment of dECM onto glass coverslips using a fluorescent marker. (A) We stained soluble biomaterial with the nonspecific fluorescent marker FM1-43 to confirm adsorption of the biomaterial to the glass surface using fluorescent microscopy. (B, C) The FM1-43 marker allowed us to detect the fluorescent biomaterial adhered to the coverslip compared to control coverslips with the biomaterial but without FM1-43. (D) We performed a concentration gradient of the dECM + FM1-43 mixture (C1: 1 mg/mL; C2: 500 μg/mL and C3: 100 μg/mL) to determine the concentration that offered the most homogeneous coating. (E) The PLL coatings had a very homogeneous coverage, (F) The dECM coatings showed the formation of small mounds (white arrows) throughout the surface of the coverslip. (G, H) A scratch-test on the adsorbed dECM was used to confirm the presence of the coating in the coverslip and to determine the thickness of the coatings. (I, J) The amount of adsorbed proteins from the 100 μg/mL solution was determined by calculating the difference in the protein levels from the coating solutions pre- and postcoating. No significant differences were found between groups. Color images are available online.

High-magnification micrographs showed that glass coverslips coated with poly-l-lysine had a homogeneously smooth surface with barely any other distinctive characteristics (Fig. 3E). Coating with decellularized ECM from the cortex also resulted in a smooth layer with a few sporadic features probably representing small accumulations of biomaterial (Fig. 3F). A simple scratch-test exposed the biomaterial coating on the coverslip and its relative thinness (Fig. 3G, H). Similar results were obtained with decellularized ECM from all three subregions of the brain.

Finally, the amount of protein that adsorbed to the coverslips from the decellularized ECM biomaterials was also quantified by evaluating the coating solution before and after adsorption (Fig. 3I, J). The results showed that all brain sections exhibited similar protein adsorption levels on the coverslips.

In vitro experiments on decellularized ECM coatings

Qualitative morphological evaluation of cell–biomaterial interactions using SEM showed that the decellularized ECM from the three different brain regions is capable of promoting PC12 cell attachment after only 24 h. Control poly-l-lysine coatings provided a good substrate for PC12 growth and expansion as a monolayer, with minimal effects on the immature, rounded morphology of the cells (Fig. 4A). In the case of the decellularized ECM from the 4- and 1-day protocols, cell attachment seemed to be similarly promoted, while no cellular morphological differences were detectable at this early stage of attachment when the biomaterial was used as a substrate, regardless of the protocol or brain section used (Fig. 4B–G).

FIG. 4.

Adhesion of PC12 cells over dECM coating biomaterial by using scanning electron microscopy. PC12 cells attached to coverslips with (A) poly-l-lysine (B–G) and dECM after 24 h of culture. The scale bar for all images is equal to 10 μm.

In vitro experiments on decellularized ECM as soluble factor

Cell viability monitoring is critical to determine the cytotoxicity effect of a biomaterial. PC12 cell viability was assessed through calcein/propidium iodide staining and image data analysis (Supplementary Fig. S7). Considering that the 4- and 1-day protocols use different solvents and enzymes that could be toxic for the cells, we wanted to know if there were any remnant negative effects on the viability of PC12 cells and any inherent differences comparing the two protocols. Brain decellularized ECM (i.e., cortex, cerebellum, and remaining areas) obtained from the 4-day protocol (Supplementary Fig. S7A–C) and 1-day protocol (Supplementary Fig. S7D–F) was used as a soluble factor to study its cytocompatibility in vitro with PC12 cells. Quantification of the different fields from the fluorescence microscopy images confirmed that PC12 cell viability decreased compared to controls (98% viability) when exposed to any of the brain sections as a soluble factor (Supplementary Fig. S7G). Still, cell viability with all biomaterial groups remained considerably high, above 95% in all cases.

PC12 cells under nonstimulating control conditions grew in colonies of spherical cells with almost no morphological features (Fig. 5A). However, when these cells were in the presence of the positive controls (i.e., purified growth factors NGF and BDNF at a 50 ng/mL each), they underwent morphological changes associated to maturation that led to the generation of neurite-like structures in about 55.65% and 44.75% of the cells, respectively. (Fig. 5B, C). The decellularized ECM biomaterial (100 μg/mL) was also capable of promoting such stimulation of PC12 cells, although to a lesser extent (Fig. 5D–J). Decellularized ECM from the 4-day protocol promoted maturation in 16.3%, 13.8%, and 16.0% of the cells when treated with cortex, cerebellum, and remaining areas, respectively. Meanwhile, the 1-day protocol, facilitated maturation in 24.3%, 20.4%, and 27.4% of cells treated with cortex, cerebellum, and remaining areas, respectively.

FIG. 5.

Cell maturation promoted by dECM biomaterials. PC12 neuronal maturation induced by brain ECM was correlated to neurite extension. Immunofluorescence of the maturation of PC12 cells was compared using the following soluble factors: (A) PBS as a negative control, (B) NGF and (C) BDNF as positive control, (D–F) cortex, cerebellum, and remaining areas from brain dECM (100 μg/mL) obtained from the 4-day protocol and (G–I) cortex, cerebellum, and remaining areas from brain dECM (100 μg/mL) obtained from the 1-day protocol. (J) Histogram of the maturation levels of PC12 cells in the presence of soluble factors, with n > 250 cells for each condition. The scale bar for all images is equal to 25 μm. *refers to a statistically significant p-value below 0.05 versus negative control (i.e., PBS); ^#refers to a statistically significant p-value below 0.05 versus NGF; ^$refers to a statistically significant p-value below 0.05 versus BDNF; ^%refers to a statistically significant p-value below 0.05 versus 4-day cortex; ^@refers to a statistically significant p-value below 0.05 versus 4-day cerebellum; ^&refers to a statistically significant p-value below 0.05 versus 4-day remaining areas; ^refers to a statistically significant p-value below 0.05 versus 1-day cerebellum. ECM, extracellular matrix; PBS, phosphate-buffered saline. Color images are available online.

Discussion

In order for decellularized ECM from the brain to become an option in the treatment of cerebral ischemia and other neurological diseases, basic questions regarding the structural and biochemical properties of the different subregions and the importance of using tissue-specific biomaterials need to be answered. In this study, we used the complexity of the brain as a novel model to explore structural and biochemical differences between three brain sections (i.e., cortex, cerebellum, and remaining areas) and compared the process of two different decellularization protocols on these separate brain sections. Our results highlight differences among the brain sections that should be considered during decellularization and that eventually have an effect on promotion of cellular maturation. Still, all brain sections regardless of the decellularization protocol were able to establish a homogeneous coating on coverslips, adequate cell adhesion, and high cell viability using a unidirectional cell differentiation model based on the PC12 cell line.

Quantification of the genetic material left on the decellularized ECM biomaterial after processing (i.e., DNA content) is an indirect but key parameter to evaluate the safety and efficacy of a decellularization protocol.¹⁴ Although most commercially available decellularized ECM products contain some amount of remnant DNA content, the immune response elicited tends to be anti-inflammatory leading to an environment conducive to regeneration.^40–42 Still, ineffective removal of cellular material from the ECM has been linked to adverse immunogenic reactions upon in vivo implantation ^43,44 Interestingly, one study confirmed a serum antibody response after decellularized ECM implantation in a nonhuman primate model, but without any adverse effect on tissue remodeling.⁴⁰ Complete removal of these immunogenic factors is extremely difficult and, thus, a more realistic expectation has been to keep undesirable cellular remnants below a threshold level.⁴⁵ A quantitative threshold of residual dsDNA of 50 ng per mg of lyophilized ECM has been established based on studies that achieved in vivo tissue remodeling without noticeable adverse host reactions.^14,46 However, functional assays based on the activation of immune cells and tissue remodeling for different types of tissues are required to test this threshold and allow adjustments depending on their immunogenic activity.²⁹

Efficient decellularization is a trade-off between the duration/strength of the decellularization protocol and the integrity/maintenance of the ECM proteins. We evaluated a 4-day, simple decellularization protocol that used mild concentrations of detergents and depended more on physical forces (i.e., shaking),²⁵ compared to a shorter, 1-day protocol that used a stronger combination of reagents and enzymes²⁶ for the removal of cells. Both protocols were effective at removing close to 90% or more of the DNA present but, as expected, the longer processing time resulted in a decellularized ECM with lower remnant DNA content.

Morphological evaluations of the decellularized tissue using electron microscopy and histology confirmed that both decellularization protocols were able to maintain similar topographical properties and structural integrity above 80%. The 4-day protocol did cause a significant decrease in tissue ultrastructural integrity probably due to the longer exposure to chemical and mechanical forces. This is in agreement with reports in the literature that show that mechanical forces may have a more detrimental effect on tissue integrity than chemical forces.¹⁴ The loss of structural integrity promoted by 4-day protocol was also coupled to a greater volumetric/weight loss and a slightly higher propensity to contamination throughout the process. Sterility is an important consideration for the usability of decellularized ECM scaffolds that is still not commonly included in the characterization of most studies in the literature.^25–27 Aseptic technique during processing, as used in our study, is probably the most common method used to ensure sterility of the biomaterial, taking advantage of the harsh reagents (e.g., detergents, acids) and conditions (e.g., low pH during digestion), which are not conducive to microbiological growth.⁴⁷ Still, contamination risks seem to increase proportionally to the duration of the processing protocol, as our results showed.

The ability of decellularized ECM to elicit a positive regenerative response is, in part, due to a synergistic combination of components, including growth factors and other important biochemical cues, such as bioactive structural proteins and cryptic peptides.^48,49 The 4- and 1-day decellularization protocols had very similar behavior regarding their capacity to maintain levels of desirable growth factors (i.e., neurotrophic factors). In this study, we focused on the presence of NGF and BDNF because of their direct involvement in the neuronal differentiation of the PC12 cell line,^50–52 which were the cells chosen to test the decellularized ECM. Freely soluble factors such as NGF were easily removed from the ECM by either protocol, while vesicle- and synaptosome-bound factors such as BDNF^53,54 were maintained, and unexpectedly enriched, by the decellularization process, particularly in the 1-day protocol.

Neurotrophic factors promote the development and regeneration of neurons.⁵⁵ The neurotrophins family includes NGF, BDNF, NT-3, and NT-4, all of which present considerable structural homology and bind to similar receptors of the tyrosine receptor kinase (trk) type with high affinity.⁵⁶ In conjunction with the trk receptors, neurotrophins also bind to the low affinity coreceptor p75.⁵⁷ We continue to explore the presence of other neurotrophins, such as NT-3, NT-4, CNTF, and GDNF,^57,58 as well as other growth factors shown to be retained in cerebral decellularized ECM, such as VEGF and bFGF.²⁶

Still, a more detailed characterization of the components in the decellularized ECM of the different brain subregions, such as proteoglycans and other small proteins that have critical functions in the normal functioning of the brain, is required to better understand the mechanisms of the cellular responses. For example, chondroitin sulfate proteoglycans have been shown to act as regulators of growth of neuronal processes and neural patterning,^59–61 and affect migration events in CNS tissue.^62–64 Similarly, Tenascins (e.g., TNC and TNR) have been shown to exhibit adhesive properties on neurons and other CNS cell types.^65–67

Finally, we characterized the use of subregion-specific cerebral decellularized ECM as an in vitro substrate and evaluated their effect at the cellular level using the PC12 cell line. Cerebral decellularized ECM has been used as a soluble factor,²⁶ as a two-dimensional coating²⁵ and as a 3D hydrogel⁶⁸ to study its role in proliferation, migration, and differentiation of neural cells and stem cells. However, to our knowledge, no studies have looked at subregion-specific effects of brain decellularized ECM, considering the large variability in cellular and protein content in each subregion of the brain. Evaluation of these different types of presentations is important, as decellularized ECM scaffold delivery in vivo probably elicits differential effects depending on the initial conditions of the biomaterial and the presentation format. In agreement with most studies in the literature,^25,26,68,69 our results showed that the decellularized ECM did not affect cell viability, regardless of the protocol used. However, we did observe higher levels of PC12 differentiation with the biomaterial from the 1-day protocol compared with the 4-day protocol. No differences were evident when comparing biomaterial from the different brain subregions for each protocol, even though BDNF levels in some of the regions were different.

PC12 cells react to neurotrophins, such as NGF and BDNF, resulting in neuronal differentiation due to the expression of the trk and p75 receptors.^50,70 The morphological changes seen in the cells after exposure to the decellularized biomaterials was comparable to the changes caused by BDNF alone. These results are consistent with the finding that the biomaterials retained BDNF but not NGF after decellularization, with the 1-day protocol doing so at higher concentrations than the 4-day protocol. Further studies are required, however, to reach this conclusion, since PC12 cells are an immortalized cell line, with their inherent biological limitations, and require that these results be confirmed with primary cells, in in vivo models or in the human disease. In addition, other important neurotrophic factors (e.g., NT-3, NT-4, CTNF) and others receptors (i.e., GFRα-1, GFRα-2) not evaluated in this study could be involved.⁵⁷ The distribution of neurotrophins and their receptors is wide-spread but nonhomogeneous throughout the nervous system, with presence in the hippocampus, cortex, and cerebellum of the CNS^56,58 as our results also showed. In addition, other structural and bioactive components (e.g., laminin, proteoglycans) complicate the puzzle,^71–73 as they can have distinct biochemical and biomechanical roles, with studies showing that properties such as stiffness in different brain areas can be involved in the processes of neuronal migration, glial dynamics, and axonal regeneration.^21,74 This highlights the importance that decellularized ECM biomaterials retain these growth factors to achieve the best biological performance.

Conclusions

Neurological disorders are considered as a major future pandemic, in part, due to higher incidence in the aging worldwide population and lack of effective treatments in most cases. The limited number of studies focused on brain decellularized ECM and the absence of guidance on best practices and standard methods requires an evaluation of current decellularization protocols to find the optimal and most efficient methodologies. Our results show that different decellularization protocols can reach adequate decellularization levels according to current standards, yet, their effects in vitro can be quite different depending on their capacity to retain the tissue ultrastructure and growth factor content. The protocol based on chemical decellularization favored higher levels of neurotrophic factors that promoted higher cell differentiation and was dependent on the subregion of the brain from which the biomaterial was sourced. Further studies are required to optimize the decellularization protocol to retain and characterize the most bioactive molecules and evaluate their mechanisms of action.

Disclosure Statement

No competing financial interests exist.

Funding Information

This work was supported by UK Foreign Commonwealth Office (award number PAMAS14PNP000012, for RAG, DR); NIH/NINDS Fogarty International Center (grant number 1R21NS098896-02, for MAPP, RAG); SNI-SENACYT (contracts 91–2015 and 146–2017, for RAG) and SENACYT (grants number ITE15-016 and FID14-066).

Supplementary Material

Supplementary Data

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5

Supplementary Figure S6

Supplementary Figure S7