Epimorphic regeneration approach to tissue replacement in adult mammals

Abstract

Urodeles and fetal mammals are capable of impressive epimorphic regeneration in a variety of tissues, whereas the typical default response to injury in adult mammals consists of inflammation and scar tissue formation. One component of epimorphic regeneration is the recruitment of resident progenitor and stem cells to a site of injury. Bioactive molecules resulting from degradation of extracellular matrix (ECM) have been shown to recruit a variety of progenitor and stem cells in vitro in adult mammals. The ability to recruit multipotential cells to the site of injury by in vivo administration of chemotactic ECM degradation products in a mammalian model of digit amputation was investigated in the present study. Adult, 6- to 8-week-old C57/BL6 mice were subjected to midsecond phalanx amputation of the third digit of the right hind foot and either treated with chemotactic ECM degradation products or left untreated. At 14 days after amputation, mice treated with ECM degradation products showed an accumulation of heterogeneous cells that expressed markers of multipotency, including Sox2, Sca1, and Rex1 (Zfp42). Cells isolated from the site of amputation were capable of differentiation along neuroectodermal and mesodermal lineages, whereas cells isolated from control mice were capable of differentiation along only mesodermal lineages. The present findings demonstrate the recruitment of endogenous stem cells to a site of injury, and/or their generation/proliferation therein, in response to ECM degradation products.

Regeneration in adult mammals is generally limited to select tissues including the bone marrow (1), intestinal mucosa (2), superficial layers of the skin (3), nailbeds (4), and the liver (5). The default response to injury in most other tissues involves the processes of inflammation and scar tissue formation (i.e., repair). Epimorphic regeneration occurs after tissue injury in early human fetal development (6) and in urodeles (e.g., the newt and salamander) of all ages and has been divided into two relatively broad categories. The first category of epimorphic regeneration involves the formation of a blastema. A blastema is a preprogrammed accumulation of multipotential cells that spontaneously proliferate, migrate, differentiate, and spatially organize in three dimensions to form a perfect phenocopy of the missing or injured body part (7). The second category of epimorphic regeneration does not involve a blastema and instead utilizes mechanisms that include: (i) transdifferentiation of cells to replace the missing tissue; (ii) limited dedifferentiation and proliferation of cells; and/or (iii) proliferation and differentiation of stem and progenitor cells in the injured tissue (8). The specific origins of the cells, the genetic profile, and the upstream molecular signals that participate in the two categories of epimorphic regeneration are only partially known, and microenvironmental cues and cell populations required to initiate and sustain the process are gradually and systemically being identified (9, 10). The equivalent of a blastema does not occur in adult mammals, but nonblastemal regeneration does occur in certain organs of adult mammals such as the liver (11), bone marrow (1), and deer antler (12). In the context of regenerative medicine strategies for adult mammals, including humans, nonblastemal epimorphic approaches may be more plausible for the restoration of functional tissues and organs.

Bioactive peptide molecules derived from mammalian extracellular matrix (ECM) have been shown to have potent chemoattractive and mitogenic effects upon endogenous progenitor and stem cells (13 –16). Although full characterization of these bioactive peptides has not been completed to date, ex vivo enzymatic, chemical, and physical methods (17 –19), and in vivo physiologic methods (20), have generated a heterogeneous population of ECM peptides (13, 21) with both chemotactic and mitogenic properties for a variety of stem and progenitor cells.

The objective of the present study was to determine the ability to recruit multipotential cells to the site of injury, and/or endogenously generate such cells, by in vivo administration of chemotactic ECM degradation products in an established mammalian model of digit amputation (22). Conceptually, such a process would represent an essential component of a nonblastemal-based approach to epimorphic regeneration of tissues and organs.

Results

ECM Degradation Products Are Chemotactic for Human Perivascular Cells.

Chemotactic activity of ECM degradation products was confirmed in vitro by using human multipotent perivascular cells (15). Chemokinetic and haptotactic effects were ruled out by control wells that contained varying concentrations of the same ECM degradation products in both the top and bottom chambers. Results showed that cells selectively migrated only when a concentration gradient was present (Fig. 1A), confirming the chemotactic activity of the ECM degradation products.

Fig. 1.

The in vitro migration of perivascular stem cells from the upper chamber of a Boyden assay to the lower chamber only in the presence of an ECM gradient confirms chemotactic and not chemokinetic activity of the degradation products. Error bars are SD among three wells, with a similar trend seen on three separate occasions (A). Gross appearance of digits in mice on day 14 after amputation and injection of ECM degradation products (B) or no injection (C). Arrows denote the amputated digit. Trichrome staining of digits in mice on day 14 postamputation and injection of ECM degradation products (D) or no injection (E). Histologic Trichrome-stained sections are at a magnification of ×40 and ×200.

ECM Degradation Products Promote the Accumulation of Mononuclear Cells in Vivo.

Twenty-four adult, 6- to 8-week-old female C57/BL6 mice were randomly divided into two equal groups (Table 1). After midsecond phalanx amputation of the third digit of the right hind foot, animals in group 1 were injected with ECM degradation products while animals in group 2 were left untreated. Preliminary studies in mice showed that injection of India ink into the footpad resulted in the presence of ink at the site of amputation, demonstrating that channels of flow exist between the site of injection and the site of amputation (Fig. S1). Before 14 days after amputation, no macroscopic differences were noted between the two groups. At 14 days, the amputation site in mice from both groups showed a closed wound with a bulbous soft tissue accumulation at the site of amputation (Fig. 1 B and C). Masson’s trichrome stained sections of digits from group 1 mice showed an accumulation of mononuclear cells distal to the site of amputation and a thin, keratinized epithelium (Fig. 1D). The number of accumulated cells varied between animals in the ECM-treated group, but the accumulation of cells was consistently limited to the soft tissue distal to the cut end of the bone and extended to the digit tip. Mice from the control group showed only scar tissue and an overlying thick keratinized epithelium at the digit amputation site (Fig. 1E).

Table 1.

Surgical groups for the present study

Group	Surgical procedure	Animals, n
1	Amputation and injection of ECM degradation products	12
2	Amputation and no treatment	12

Accumulated Mononuclear Cells Are Multipotent.

To identify the phenotype of the accumulated cells, tissue sections were stained for markers of stem and progenitor cells, including Sox2 (23), Rex1 (24), and Sca1 (25). Mice treated with ECM degradation products accumulated a greater number of cells positive for stem cell markers Sox2 and Rex1 distal to the site of digit amputation compared to untreated mice and compared to normal, uninjured mice (Fig. 2). To confirm the multipotency of accumulated cells, isolated cells from six mice in the treated and untreated injured groups were pooled, expanded to 70% confluence, and evaluated for their ability to differentiate along ectodermal and mesodermal lineages in vitro. Following culture in conditions of neuroectodermal differentiation, cells from ECM treated mice showed heterogeneous morphologies including spindle-like cells with long processes that expressed β3-tubulin and NeuN, markers of differentiating neurons (26), and stellate shaped cells that expressed glial fibrillary acidic protein (GFAP), a marker of glial cells that is also expressed by some neural stem cells (27, 28) (Fig. 3). Cells from untreated mice cultured in conditions of neuroectodermal differentiation showed mainly spindle-shaped cells consistent with a fibroblast or mesenchymal phenotype, and these cells did not give rise to cells expressing β3-tubulin, NeuN, or GFAP (Fig. 3). However, when exposed to an adipogenic differentiation environment, cells from mice in both groups showed vacuoles that stained positive for Oil Red O, a specific stain for the presence of lipid accumulations (Fig. 3). When subjected to conditions of osteogenic differentiation, cells from mice in both groups acquired a round morphology and stained positive for Alizarin Red, a stain that confirms the deposition of calcium by cells (Fig. 3).

Fig. 2.

Histologic sections of cell accumulations distal to the site of amputation in mice 14 days after amputation and injection of ECM degradation products or no treatment after staining for markers of multipotency. All images were taken at a magnification of ×400. Cell counts are displayed in units of number of cells. *P < 0.05; **P < 0.005 (between treatment and no treatment, or treatment and uninjured controls). Error bars are SEM.

Fig. 3.

In vitro lineage differentiation potential of cell isolated distal to the site of amputation in digits of mice treated with ECM degradation products and untreated. Neuroectodermal differentiation was confirmed via expression for neuroectodermal markers (β-tubulin-III, NeuN, and GFAP). Adipogenic differentiation was confirmed via Oil Red O staining for the presence of lipid vacuoles. Osteogenic differentiation was confirmed via Alizarin Red staining for calcium deposition.

Discussion

The present study showed that degradation products of ECM promote the accumulation of a population of cells to a site of injury that have the capacity for differentiation into ectodermal and mesodermal phenotypes. The accumulated cells express Sca1, Sox2, and Rex1, all markers that have been associated with multipotential differentiation (23 –25, 29, 30). Whereas cells isolated from both treated and control animals showed the capacity for differentiation along the mesodermal lineage, only cells that accumulated at the site after injection of ECM degradation products were capable of differentiation toward a neuroectodermal lineage.

The source of cells that compose this multipotential cell cluster was not investigated in this study. However, local tissue progenitor cells, multitpotent perivascular cells, neural crest derived cells (31), and circulating mulitpotential cells should all be considered (15, 32 –34). Previous studies have shown that marrow-derived cells participate in the constructive remodeling of tendon tissue (14). However, the number of marrow-derived cells in the reconstructed tendon was relatively low compared to the number of cells observed at the amputation site in the present study. Transdifferentation or dedifferentiation are also plausible explanations for the presence of the multipotent cells, but these phenomena have not been shown to occur in large numbers in vivo in mammalian tissue (35 –38). The possibility that the observed multipotent cells were generated in situ after lineage reprogramming cannot be ruled out. However, few studies have demonstrated such a phenomenon in vivo and the mechanisms underlying such processes are not well understood (39).

The generation of matricryptic peptides with biologic activity from ECM is not novel. Antimicrobial activity has been attributed to such peptides in the form of defensins (40), cecropins (41, 42), and magainins (43). Angiogenic and antiangiogenic activity has been shown to be caused by maticryptic peptides derived from a variety of collagen molecules (18, 44). Peptides derived from extracellular matrix molecules have also been shown to modulate inflammation (45). It would not be surprising, therefore, perhaps even logical, that ECM degradation would result in the recruitment of cells capable of tissue repair and reconstruction. ECM degradation would expectedly occur at times of injury and, thus, the bioactivity intrinsic to resulting matricryptic peptides would represent a competitive survival and evolutionary advantage.

The mechanisms by which matricryptic peptides recruit stem cells in vivo are as of yet unknown. Additionally, the extent to which matricryptic peptides remain active in vivo is not known. Because ECM scaffolds consist of various molecules such as collagen and fibronectin, proteoglycans, glycoproteins, growth factors, and cytokines (46), degradation of these ECM scaffolds releases a heterogeneous set of molecules (13, 21), each with varying biologic properties in vivo. Subsets of these peptides have been found to have different bioactive properties in vitro (13, 17, 19, 44, 47, 48). Although a subset of generated peptides is clearly chemotactic for stem cells, the overall contribution of these peptides to stem cell recruitment in vivo is as of yet unknown. The contribution of otherwise inert proteins in this mixture to the chemotactic process is not known with certainty, but further degradation of these inert molecules by local proteases at the site of injury may produce new, active matrcryptic residues (44, 49).

It is possible, and in fact likely, that stem cell recruitment in vivo requires a heterogeneous set of bioactive peptides. A heterogeneous set of matricryptic peptides may, through varying effects upon different cell populations locally (Fig. S1) and systemically (50), promote a local microenvironmental niche that facilitates stem cell recruitment and/or maintenance (51 –53). An independent contributing factor to the host response is the effect of ECM scaffolds on tissue macrophages (33, 54 –56) and T cells (57). Macrophages are critical determinants of the overall regenerative response in vivo (56). Inhibition of ECM scaffold degradation in vivo alters macrophage polarization toward a proinflammatory phenotype (54). The present study demonstrated that digit amputation itself increases stem cell recruitment, suggesting that some aspect of the host response to local tissue injury plays a facilitating role in mediating stem cell recruitment.

Because it is obvious that the cells recruited to the site of injury after treatment with ECM degradation products do not spontaneously regenerate the missing body part, these cells do not fit the classic definition of a blastema. The response observed in the present study would be most consistent with the initial phase of a nonblastemal, epimorphic regenerative response (7, 8). This type of regeneration is characterized by cell proliferation and subsequent regeneration of site appropriate tissue, the classic example of which in mammals is liver regeneration after acute liver injury (5, 11). Considered in this light, it is plausible that the recruitment of endogenous stem and progenitor cells to a site of injury that would not normally be populated by such cells, at least in such great numbers, could promote a regenerative response. The present study showed that the recruited cell population has multipotential differentiation capability. The logical next step is to identify strategies that can promote site appropriate differentiation, spatial organization, and patterning of these cells.

Establishment of an optimal microenvironmental niche is critical for appropriate differentiation of multipotential cells. If the microenvironment could be controlled sufficiently by factors such as pH, state of hydration, electric field potential, and nutrient and growth factor availability among others, it may be possible to promote the development of site appropriate tissues in adult mammals.

Materials and Methods

Preparation of ECM Degradation Products.

Porcine urinary bladders were harvested from euthanized market weight (240–260 lb) pigs. The basement membrane and underlying lamina propria were isolated and harvested as described (58). After peracetic acid, ethanol, deionized H₂O, and PBS treatment (19), lyophilized sheets were comminuted and digested in pepsin and 0.01 M HCl for 48 h before neutralization and dilution in PBS to yield a 5 mg/mL solution. The soluble protein concentration of ex vivo-generated ECM degradation products was found to be 1.68 ± 0.17 μg/mL, and SDS/PAGE of these products in previous studies demonstrated a heterogeneous population of peptides (13, 21).

Confirmation of Chemoactivity of ECM Degradation Products.

To confirm the chemotactic activity associated with ECM degradation products (19), human perivascular stem cells isolated from human skeletal muscle (15) were assayed for a chemotactic response in a Boyden chamber. Perivascular stem cells were prepared for chemotactic assay as described (15). After starvation, cells were resuspended in DMEM at a concentration of 6 × 10⁵ cells/mL for 1 h. Polycarbonate PFB filters (Neuro Probe, Gaithersburg, MD) with 8-μm pores were coated with 0.05 mg/mL Collagen Type I (BD Biosciences, San Jose, CA). ECM degradation products were placed in the top and bottom chambers of Neuro Probe 48-well chemotaxis chambers (Neuro Probe, Gaithersburg, MD) at varying concentrations with 3 × 10⁴ cells in the top chamber, and incubated at 37 °C in 95% O₂/5% CO₂ for 3 h. Migrated cells were fixed in methanol, stained with 500 nM DAPI (Sigma; D9564), imaged on a Nikon E600 microscope, and counted by using ImageJ (NIH, Bethesda, MD).

Animal Model of Digit Amputation.

All methods were approved by the Institutional Animal Care and Use Committee at the University of Pittsburgh and performed in compliance with NIH Guidelines for the Care and Use of Laboratory Animals. Adult female 6- to 8-week-old C57/BL6 mice were obtained from Jackson Laboratories (Bar Harbor, ME). After induction of surgical plane anesthesia with isofluorane (1–2%), each mouse was subjected to aseptic midsecond phalanx amputation of the third digit of the right hind foot. At 0, 24, and 96 h after surgery 20 μg of chemotactic ECM degradation products were injected at the base of the amputated digit in group 1 animals via a 30-gauge needle (Fig. S1). Animals in group 2 were left untreated. Animals were killed via cervical dislocation under deep isoflurane anesthesia (5–6%) at 14 days after surgery. Six mice in each group at day 14 were used for immunolabeling studies, and the remaining six were used for cell isolation.

Tissue Immunolabeling.

Harvested mouse digits were fixed in 10% neutral buffered formalin and decalcified for 2 weeks in 5% formic acid before being paraffin embedded and sectioned and stained for Sox2, Rex1, or Sca1.

Sox2 staining used a primary antibody from Abcam ab15830 (Abcam, Cambridge, MA). Antigen retrieval was in 10 mM Citrate Buffer (Citrate: C1285; Spectrum, New Brunswick, NJ) for 20 min at 96 °C. After antigen retrieval, slides were placed in TBS + 0.05% Tween-20 for 5 min, rinsed in PBS for 5 min twice, blocked for 1 h in 1.5% BSA/PBS, and incubated overnight with primary antibody diluted in 1.5% BSA/PBS (1:100). Slides were then rinsed in PBS, treated with 3% hydrogen peroxide solution in methanol for 30 min, rerinsed in PBS and incubated with HRP conjugated secondary antibody for 1 h (Rabbit anti-rat IgG-HRP; Dako, Carpinteria, CA; P0450), rinsed again in PBS, and developed with 3,3′-diaminobenzidine (DAB) (Vector Labs, Burlingame, CA).

Sca-1 staining used the primary antibody ab25196 (Abcam, Cambridge, MA). Antigen retrieval was for 10 min at 93 °C in R&D Systems Antigen Retrieval Reagent Universal (R&D Systems CTS015, Minneapolis, MN) by following manufacturer's protocol. R&D Systems HRP-DAB Cell and Tissue Staining Kit for goat primary IgG antibodies (R&D Systems CTS008, Minneapolis, MN) was used for subsequent staining by following manufacturer's instructions with the following three changes: (i) slides were blocked in 1.5% BSA/PBS (ii), primary antibody was diluted 1:1000 in 1.5% BSA/PBS and (iii) incubations with biotinylated secondary antibody and subsequent washes were deleted because ab25196 is biotin-conjugated.

Rex1 staining used antigen retrieval with Antigen Unmasking Solution (Vector Laboratories, Burlingame, CA) by boiling under pressure for 2 min. Slides were then incubated in 3% hydrogen peroxide in methanol for 15 min to quench endogenous peroxidase activity. This step was followed by incubation with the anti-mouse Rex1 primary antibody (affinity-purified, polyclonal rabbit antibody, custom-generated and supplied by Alpha Diagnostic, San Antonio, TX; project name: ZFP42-17, peptide 13209) diluted 1:20 in 1.5% goat serum for 1 h at room temperature. HRP-conjugated goat anti-rabbit secondary antibody (Catalog No. 87-9263, SuperPicture; Zymed, San Francisco, CA) was added to each tissue section and incubated at room temperature for 30 min. Staining of the antigen (brown stain) was accomplished by incubation of the tissue sections with DAB chromogen substrate (SuperPicture; Zymed, San Francisco, CA) for 8–10 min at 22 °C. The negative control sections were treated identically to the adjacent sections except that 1.5% goat serum was used in place of primary Rex1 antibody.

After staining with DAB, all slides were counterstained with Harris’ hematoxylin, dehydrated, coverslipped with nonaqueous mounting medium, and imaged on a Nikon E600 microscope, Olympus Provis Microscope, or Olympus Fluoview 1000 Confocal Microscope. Images were taken at ×40 and ×400 magnification. For quantification of the number of cells positive for markers, four images were taken in each sample: distal to the amputated edge of the second phalanx bone, proximal to the tip of the digit, and lateral to the cut edge of the second phalanx bone on either side. The number of positive cells in each image was counted by three independent investigators who were blinded to the treatment group. The mean number of positive cells was compared among various groups by a two-sided, unpaired Student's t test with unequal variance. Significance was determined at P = 0.05 level (α = 0.05, β = 0.2).

Cell Isolation and Differentiation Assays.

The amputated digit was harvested and placed into cold culture medium consisting of DMEM, 10% mesenchymal stem cell grade FBS (Invitrogen, Carlsbad, CA), 100 U/mL penicillin, 100 μg/mL streptomycin, and 0.1 mg/mL ciprofloxacin (USP, Rockville MD; 1134313). Using a microdissection microscope and aseptic technique, the epidermis and dermis were removed and the tissue distal to the amputated second phalanx bone was harvested into serum-free DMEM containing 0.2% Collagenase Type II (Gibco Invitrogen; 17101-015) for 30 min at 37 °C, reconstituted in culture medium, and plated onto four or eight chamber slides.

Mesodermal differentiation.

For adipogenic differentiation, cells were cultured in adipogenic differentiation medium (HyClone, Logan, UT; SH30886.02) for 14 days. Cells were then fixed in 10% neutral buffered formalin (NBF) and incubated in 0.5% Oil Red O (Alfa Aeasar, Ward Hill, MA) for 5 min, extensively washed, and imaged for the presence of lipid vacuoles. For osteogenic differentiation, cells were cultured in osteogenic differentiation medium (HyClone; SH30881.02) for 21 days. After fixation in NBF, cells were incubated in 2% Alizarin Red solution to investigate for the presence of calcium.

Neuroectodermal differentiation.

Cells were cultured in DMEM supplemented with 10% FBS and 20 ng/mL bFGF (Sigma, St. Louis, MO; F0291) for 24 h. Culture medium was then replaced with Neurobasal-A medium supplemented with 20 ng/mL bFGF, 20 ng/mL EGF (Sigma, E9644), and 2 μM all-trans retinoic acid (Sigma, R2625) for 7 days. After 7 days of culture, cells were fixed in 4% paraformaldehyde and prepared for staining. Cells were assessed for the presence of neuron and glial cell specific proteins β-tubulin III, NeuN, and glial fibrillary acidic protein (GFAP). The primary antibodies used were β-tubulin III ab7751 (Abcam), glial fibrillary acidic protein Z0334 (DakoUSA), and NeuN MAB377 (Millipore, Billerica, MA). After fixation, slides were then washed, rehydrated in Tris Buffered Saline with 0.05% Tween 20 (TBST), and incubated for 15 min in permeablization buffer (0.1% Triton X in TBST) at room temperature. Slides were then incubated in blocking solution [2% wt/vol BSA (Sigma; A2153) in TBST] for 1 h at room temperature, followed by incubation in primary antibody for 1 h. All antibodies were used at a concentration of 1:100. After extensive washing, slides were then incubated in fluorophore conjugated secondary antibodies for 1 h at room temperature. Secondary antibodies, including Alexa Fluor 488 conjugated anti-rabbit IgG (Invitrogen; A-11008) and Alexa Flour 546 conjugated anti-mouse IgG (Invitrogen; A-11001), were used at a concentration of 1:250. Slides were counterstained with DAPI before mounting in fluorescent mounting medium (DakoUSA).