Mineral Distribution Spatially Patterns Bone Marrow Stromal Cell Behavior on Monolithic Bone Scaffolds

Abstract

Interfaces between soft tissue and bone are characterized by transitional gradients in composition and structure that mediate substantial changes in mechanical properties. For interfacial tissue engineering, scaffolds with mineral gradients have shown promise in controlling osteogenic behavior of seeded bone marrow stromal cells (bMSCs). Previously, we have demonstrated a ‘top-down’ method for creating monolithic bone-derived scaffolds with patterned mineral distributions similar to native tissue. In the present work, we evaluated the ability of these scaffolds to pattern osteogenic behavior in bMSCs in basic, osteogenic, and chondrogenic biochemical environments. Immunohistochemical (IHC) and histological stains were used to characterize cellular behavior as a function of local mineral content. Alkaline phosphatase, an early marker of osteogenesis, and osteocalcin, a late marker of osteogenesis, were positively correlated with mineral content in basic, osteogenic, and chondrogenic media. The difference in bMSC behavior between the mineralized and demineralized regions was most pronounced in an basic biochemical environment. In the mineralized regions of the scaffold, osteogenic markers were clearly present as early as 4 days in culture. In osteogenic media, osteogenic behavior was observed across the entire scaffold, whereas in chondrogenic media, there was an overall reduction in osteogenic biomarkers. Overall, these results indicate local mineral content of the scaffold plays a key role in spatially patterning bMSC behavior. Our results can be utilized for the development of interfacial tissue engineered scaffolds and understanding the role of local environment in determining bMSC behavior.

Graphical Abstract

Introduction

Soft tissue-to-bone interfaces are present throughout the mammalian musculoskeletal system. These interfacial tissues play essential roles in transferring load between unmineralized soft tissue and mineralized bone [1]. The transition from soft tissue to bone normally corresponds to gradations in mineral content, biochemical composition, and hierarchical structure in the extracellular matrix (ECM) in addition to multiple cellular populations [1]. Collectively, these gradations mediate transitions between tissues of distinct stiffness and reduce stress concentrations at the interface [2]. Operative interventions are often needed to restore the function of injured interfacial tissue due to its limited regenerative ability. Common surgical practices such as reattaching soft tissue to the bone remain challenging, because the transitional ECM and complex cell phenotypes at tissue interfaces are not preserved, which can limit robustness of the attachments [3–5]. As such, interfacial tissue engineering has emerged as a promising strategy to functionally reconstitute healthy transitional tissue.

Engineering interfacial tissues requires the recapitulation of the graded distribution of ECM components found at the interface. One of the crucial features is the gradation of hydroxyapatite (HAp) mineral content within the ECM across the bone-to-soft tissue interface. In native interfaces, the presence of HAp in the stiff bone matrix can direct osteogenic differentiation of progenitor stem cells in the bone marrow and enhance new bone formation [6–9]. Studies have revealed a positive relationship between mineral content and osteogenic behavior of cells in vitro as well [10–14]. For instance, bone marrow stromal cells (bMSCs) seeded in poly(lactic-co-glycolic acid) (PLGA) microspheres loaded with 20 wt% HAp showed upregulated expression of osteogenic markers compared to bMSCs seeded on less mineralized or unmineralized scaffolds [13]. Other researchers have demonstrated that bMSCs developed a cuboidal osteoblast-like phenotype when cultured on mineralized collagen-based scaffolds but display an elongated phenotype when cultured in unmineralized scaffolds [14]. Taken together, these studies and others show potential for controlling osteogenic behavior of bMSCs in scaffolds by tuning local mineral content [10–17].

While the relationship between mineral content and osteogenic behavior of bMSCs is well-studied [10–17], challenges remain in constructing monolithic scaffolds with graded mineral distributions and structural resemblance to native interfacial tissue. Some researchers have focused on ‘bottom-up’ approaches in which graded mineral distributions are deposited on a single scaffold to reproduce the ECM environment of an interfacial tissue [18–23]. For example, a mineral gradient was patterned on a PLGA nanofiber scaffold by partially submerging the unmineralized scaffold in a solution of calcium and phosphate ions to nucleate apatitic mineral [20]. Although the stem cells seeded on the scaffolds exhibited graded osteogenic expression, the scaffolds lacked the native-like transitional structure from bone to soft tissue. Recently, our group demonstrated a ‘top-down’ method to create mineral gradients in decellularized trabecular bone cores by removing the HAp with ethylenediaminetetraacetic acid (EDTA), a calcium chelator, in a spatially controlled manner [24]. This “top-down” method generated scaffolds with continuous gradients in mineral, resembling those observed in native tissues, thereby providing a potential platform for analysis of the effects of material cues on stem cell behavior in a unique biochemical environment. Additionally, using decellularized trabecular bone as scaffolds has several advantages over synthetic materials. For example, it maintains the native architecture of mineralized collagen, and thusly the mechanical properties of bone [25,26]. When properly decellularized, the ECM-associated growth factors in the bone matrix can be preserved and used to further instruct osteogenic behavior of cells on subsequently seeded scaffolds [27].

Given the complex set of cellular phenotypes found across interfacial tissues, e.g., osteoblasts in bone, hypertrophic chondrocytes in calcified cartilage, and chondrocytes in cartilage, establishing similar cellular populations is key to engineering interfacial tissues [1,28,29]. Since bMSCs are multipotent and can differentiate into both osteogenic and chondrogenic lineages in response to local mineral content, our partially demineralized scaffolds can be used to investigate the response of bMSCs to spatially patterned mineral [30–34]. Additionally, cells in native interfaces are exposed to a variety of soluble factors released from adjacent cell populations [35,36]. Particularly in interfacial tissues, these adjacent populations can exist in distinct material environments but in similar biochemical environments. We recreate this complex scenario by subjecting bMSC-seeded scaffolds to global media cues, e.g., osteogenic and chondrogenic. This setup has the potential to reveal the effects of crosstalk between cellular population exposed to different material environments, as commonly observed in vivo.

In the current study, we investigate the effect of mineral distribution within our scaffolds on the behavior of bMSCs in vitro in basic, osteogenic, and chondrogenic biochemical environments. We hypothesize that the osteogenic behavior will be positively correlated with mineral distribution in the scaffolds. As other studies have shown that the HAp can suppress chondrogenesis in bMSCs [22,33,34], we also hypothesize that chondrogenic behavior will be inversely correlated with the mineral distribution in the scaffolds. The outcomes of this study will advance the development of tissue engineered interfaces through the spatial patterning of stem cell behavior and the analysis of this behavior in a native-like environment.

Materials and Methods

1. Scaffold Fabrication

Cores of trabecular bone were extracted and decellularized as previously described [37]. Briefly, bone cores were explanted from the distal femurs of twelve, 1-to-3-day old neonatal bovids (Gold Medal Packing, Inc., Rome, NY) with a 6 mm diameter coring bit. The articular cartilage of the extracted cores was removed, and the trabecular portion was sectioned into 10 mm long cylindrical cores. The cellular debris and bone marrow of the cores were removed with a high velocity stream of deionized water. The bone cores were soaked sequentially in 0.1 w/v% EDTA (TCI, Tokyo, Japan) in phosphate buffered saline (PBS) (Corning, Manassas, VA) for 1 hour at room temperature, hypotonic buffer (10 mM Trizma base (TCI, Tokyo, Japan), 0.1 w/v% EDTA in PBS) at 4°C for 24 hours, and detergent (10 mM Trizma base, 0.5 w/v% sodium dodecyl sulfate (SDS) (Sigma, St. Louis, MO) in PBS) at 4°C for 24 hours [38]. Following washes, cores were rinsed thoroughly with PBS. The decellularized bone cores were skewered halfway along the long axis of the cylinder on a surgical needle and partially submerged in a 9.5 w/v% EDTA in PBS (pH = 7.4) for 4.5 hours to achieve partial demineralization, according to our previously published procedure for fully demineralizing ~50% of the length of the bone core (Figure 1A) [24].

Figure 1.

A) The mineral gradient is produced by demineralizing part of trabecular bone cores in EDTA solution [24]. Von kossa stained thin section shows the mineralized region of the scaffold in black (top) and the unmineralized region in red (bottom). B) bMSCs are extracted from trabecular bone of neonatal bovids, cultured, and expanded to Passage 3. C) Bone scaffolds were seeded with bMSCs in a spinner flask supplied with basic media, osteogenic media, or chondrogenic media without TGF-β1 for 48 hrs. D) Seeded scaffolds were cultured in well plates supplied with basic media, osteogenic media, or chondrogenic media with TGF-β1 for 4, 7, or 14 days. E) Biomarkers were examined by immunohistochemical staining and histological staining.

2. bMSC Extraction and Expansion

BMSCs were isolated from the trabecular bone marrow in distal femurs of ten, 1-to-3-day old neonatal bovids as previously described [30]. Briefly, the trabecular region of the femur was washed with heparin supplemented media, and the extracted solution was centrifuged at $300\times \overline{\text{g}}$ [39]. Pelleted cells were plated on culture flasks and the non-adherent cell population was washed off after 48 hours. Isolated bMSCs were plated at a density of 2000 cells/cm² and expanded to passage 3 in growth media consisting of Dulbecco’s modified Eagle’s medium without sodium pyruvate (DMEM) (Corning, Manassas, VA), 10 vol% fetal bovine serum (FBS) (Gemini Bio-Products, West Sacramento, CA, 1 ng/mL basic fibroblastic growth factor (bFGF) (BD Biosciences, Franklin Lakes, NJ) and 100 IU/mL penicillin (Figure 1B).

The bMSCs showed multipotency through a triple lineage test (Fig. S1). The triple lineage test was carried out following a published procedure [40]. Briefly, the bMSCs were seeded in 6-well plates at a density of 50,000 cells per well and cultured in growth media until 100% confluent. The confluent cells were treated with osteogenic media (MEM-α, 10 vol% FBS, 100 IU/mL penicillin, 2 mM L-glutamine (VWR, Brooklyn, NY), 0.1 μM β-glycerolphosphate (MP, Santa Ana, CA), 50 μM ascorbic acid (Sigma, St. Louis, MO), 0.1 μM dexamethasone (Sigma, St. Louis, MO)) or adipogenic media (MEM-α, 10 vol% FBS, 100 IU/mL penicillin, 2.0 mM L-glutamine (VWR, Brooklyn, NY), 1.0 μM dexamethasone, 10% insulin (Sigma Aldrich), 500 μM 3-isobutyl-1-methylxanthine (IBMX) (Sigma Aldrich)). The cells were cultured for 3 weeks and the media was replaced every other day. At the end of culture, cells were washed in PBS 2 times, fixed in buffered formalin for 15 minutes, rinsed in DI water 2 times and stained by Oil-red-O or Alizarin red. For chondrogenic differentiation, the pellet culture method was used [40]. Briefly, the bMSCs were suspended in growth media at 8 million cells per mL. 30 μL of the suspension containing 240,000 cells were dropped onto the well plate. The droplets were incubated at 37°C, 5% CO₂ for 2 hours until cell pellets formed. The cell pellets were then treated with chondrogenic media (DMEM, 10 vol% FBS, 100 IU/mL penicillin, 1mM non-essential amino acids (Invitrogen, Thermo Fisher, Waltham, MA), 0.4 mM L-proline (Sigma, St. Louis, MO), 50 μg/mL ascorbic acid) and cultured for 4 weeks. The resulting micromasses of cells were collected, fixed in formalin, embedded in paraffin, sliced, and stained with Alcian blue and nuclear fast red. bMSCs cultured in basic media (Minimum Essential Medium-α (MEM-α) (ThermoFisher, Waltham, MA), 10 vol% FBS, 100 IU/mL penicillin) served as negative control for chondrogenic differentiation.

3. bMSC Seeding and Culturing on Partially Demineralized Bone Scaffolds

The partially demineralized scaffolds were soaked in 70 vol% ethanol for 30 minutes, followed by a 1.5 hour wash in PBS and a 2 hour wash in either basic media composed of Minimum Essential Medium-α (MEM-α) (ThermoFisher, Waltham, MA), 10 vol% FBS, 100 IU/mL penicillin, osteogenic media composed of MEM-α, 10 vol% FBS, 100 IU/mL penicillin, 2 mM L-glutamine (VWR, Brooklyn, NY), 0.1 μM β-glycerolphosphate (MP, Santa Ana, CA), 50 μM ascorbic acid (Sigma, St. Louis, MO), 0.1 μM dexamethasone (Sigma, St. Louis, MO), or chondrogenic media composed of DMEM, 10 vol% FBS, 100 IU/mL penicillin, 1mM non-essential amino acids (Invitrogen, Thermo Fisher, Waltham, MA), 0.4 mM L-proline (Sigma, St. Louis, MO), 50 μg/mL ascorbic acid. The bone scaffolds were seeded with bMSCs as previously described [24]. Briefly, bone scaffolds were skewered and suspended in a spinner flask. The flask was supplied with 150 mL of basic media, osteogenic media, or chondrogenic media at an initial cellular density of 500,000 cells/scaffold and incubated for 48 hours at 37°C, 5% CO₂. After seeding, the scaffolds were transferred into basic media, osteogenic media, or chondrogenic media. At this time, 10 ng/mL transforming growth factor β1 (TGF-β1) (Thermo Fisher Scientific, Waltham, MA) was added to the chondrogenic media. The scaffolds were then cultured statically for 4, 7 or 14 days (Fig. 1C, D). The cultured scaffolds were fixed in formalin for 24 hours and stored in 70 vol% ethanol. Triplicate samples were run for each media condition and timepoint; in total 27 samples were collected and analyzed.

4. Histology

Seeded bone scaffolds were fixed, decalcified, dehydrated, embedded in paraffin, sectioned, and stained. Hematoxylin and eosin (H&E) and picrosirius red with hematoxylin were used to visualize the ECM and collagen matrix. Since samples went through decalcification prior to being embedded in paraffin, the mineralized part of the sample could not be distinguished from the demineralized part through brightfield microscopic imaging. As previously described, mineralized regions showed greener colors when stained with picrosirius red and observed through crossed polarizers [24]. Therefore, polarized light images of slides stained by picrosirius red were used to distinguish between mineralized and demineralized regions of the scaffolds. Alcian Blue with nuclear fast red was used to visualize glycosaminoglycans (GAGs) deposited by bMSCs in scaffolds cultured with chondrogenic media. Sections were imaged using an Aperio Scanscope slide scanner (Aperio Technologies, Inc., Vista, CA) under brightfield illumination. Picrosirius red stained slides were imaged under cross-polarizers with a Nikon Eclipse TE2000-S microscope (Nikon Instruments, Melville, NY) and a SPOT RT camera (Diagnostic Instruments, Steriling Heights, MI) to view the alignment of collagen fibrils deposited by bMSCs. Images of Alcian blue and nuclear fast red stains were quantified by ImageJ. Briefly, an optimized color deconvolution threshold was created to split the original images into three 8-bit images containing only Alcian blue stain, nuclear fast red stain and a residual image respectively [41,42]. The optimization of the color deconvolution was done by using Alcian blue staining of GAGs on native neonate bovine patella osteochondral tissue as references of positive Alcian blue staining (Fig. S2). The intensity of the Alcian blue stain was evaluated by the IHC profiler plugin for ImageJ [43]. 20 measurements from images of interest (200 px*200 px, 96 dpi) were collected from each of the mineralized or demineralized regions of 3 seeded scaffolds for each media condition and timepoint. For instance, we measured 6, 6, and 7 images from samples 1, 2, and 3, respectively. Resultant scores were generated to describe the amount of GAGs stained by Alcian blue in the region of interest.

5. Immunohistochemical (IHC) Staining

IHC staining was performed on fixed samples embedded in paraffin to study the cellular behavior and the matrix deposited by the bMSCs. Following deparaffinization, antigen retrieval was performed by treating the sections with aqueous citric acid buffer at 60°C overnight, followed by 3 vol% H₂O₂ at room temperature for 10 minutes. The slides were incubated in blocking serum with 1 wt% of BSA at room temperature for 1 hour to inhibit unspecific binding of antibodies. The slides were rinsed 5 times in Tris buffer and incubated overnight with rabbit anti-COL1 (1:500) (Abcam, ab34710), rabbit anti-COL2 (1:500) (Abcam, ab34712), rabbit anti-COLX (1:1000) (Abcam, ab58632), mouse anti-alkaline phosphatase (1:250) (Abcam, ab116592) and mouse anti-osteocalcin (1:250) (ab13420) primary antibodies at 4°C. Negative controls were incubated in Tris buffer without primary antibodies (Fig. S3). After rinsing off the primary antibody or negative control solution, the slides were incubated in secondary antibody (either rabbit or mouse IgG) solution for 1 hour and avidin-biotinylated horseradish peroxidase for 30 minutes (Vector Laboratories, PK-4001, PK-4002). Following the incubation, the slides were treated with 3,3’-diaminobenzidine (DAB) peroxidase substrate (Vector Laboratories, Burlingame, CA) until brown precipitate was observed in the droplet. The reaction was ceased by rinsing specimens with deionized water. Hematoxylin was used as counterstain to visualize the original bone matrix and cell nuclei. Images of slides stained by DAB were quantified in ImageJ using plugins and macros. Briefly, an optimized color deconvolution threshold was created to split the original images into 3 8-bit images containing only DAB brown stain, hematoxylin stain and a residual image respectively [41,42]. The optimization of the color deconvolution was done by using DAB staining of ALP and OCN on native neonate bovine patella osteochondral tissue as references of positive ALP and OCN staining (Fig. S4). The intensity of the brown DAB staining of the samples was evaluated by the IHC profiler plugin in ImageJ [43]. 20 measurements from images of interest (200 px*200 px, 96 dpi) were collected from each of the mineralized or demineralized regions of 3 seeded scaffolds for each media condition and timepoint. For instance, we measured 6, 6, and 7 images from samples 1, 2, and 3, respectively. Resultant scores were generated from the measurements and averaged to describe the amount of specific antigen presenting in the images.

6. Measurement of Scaffold Porosity

The porous area of the mineralized and demineralized regions in the scaffolds was measured to determine the effect of EDTA treatment on the porosity of the scaffolds. Images of slides stained by H&E were processed by the color deconvolution plugin in ImageJ to separate the regions stained by eosin [41]. The eosin images were converted to 8-bit images, thresholded to remove cellular contents and binarized to images only colored in black and white. Since the porous area is the space not occupied by trabecular bone, the trabecular area was measured by BoneJ [44] and subtracted from the total area ascribed by the scaffold to obtain the porous area.

7. Measurement of Local Seeding Density

Cell seeding density on the scaffolds was quantified using the samples obtained on Day 4 of culture. We compared the seeding density on the mineralized and demineralized regions of the scaffolds to determine the effect of demineralization on seeding density. Images of slides stained by H&E were processed by the color deconvolution plugin in ImageJ to separate the regions stained by hematoxylin and eosin. The hematoxylin images were converted to 8-bit images, masked to remove the scaffold contents and binarized to images only colored in black and white. The number of cells was counted using the particle analyze function with a threshold of 10 – 60 μm² in ImageJ and confirmed manually. The number of cells was normalized by the trabecular area calculated by the BoneJ using the method described in section “6. Measurement of Scaffold Porosity”.

8. Statistical Analysis

All quantitative data were tested for normality and expressed as mean and standard deviation within groups. Pairwise comparisons between treatment groups were conducted using the ANOVA test and Tukey’s post-hoc test.

Results

The objective of this work was to control the osteogenic and chondrogenic behavior of bMSCs in bone scaffolds with spatially patterned distribution of mineral. To obtain spatially patterned bone scaffolds, we submerged half of the length of decellularized trabecular bone cores into an EDTA solution (9.5 w/v%) for 4.5 hours. As we have published previously, this technique for partial demineralization reliably produces consistent mineral gradients within the bone cores [24]. The resulting scaffolds have a fully demineralized region (mean size of ~400 μm), a middle region of partially demineralized trabeculae (mean size of ~300 μm), and a fully mineralized region (mean size of ~400 μm) [24]. After seeding with bMSCs, these scaffolds were cultured in basic media, osteogenic media, or chondrogenic media. We evaluated seeding density, osteogenic, hypertrophic, and chondrogenic biomarkers for bMSCs, as a function of location within the scaffolds, by histological and IHC staining, coupled with statistical analysis of staining profiles.

Initially, we examined cell seeding density as a function of location across the scaffolds to confirm that we could compare the resulting behavior of cells for the different environments within a scaffold. To examine seeding density, we firstly compared local porosity in the scaffolds to confirm that the seedable area was the same for all regions within a given scaffold. The porosities in the mineralized and demineralized regions of the scaffolds were not significantly different indicating that the demineralization treatment did not affect scaffold porosity (Fig. S5). The seeding density of the bMSCs was then quantified and normalized to trabecular area of the scaffolds at Day 4 of culture for both mineralized and demineralized regions in all groups (Fig. S6). Overall, we did not detect any significant differences in seeding density within a scaffold for any group (Fig. S6A–C). When the scaffolds were cultured in osteogenic media, however, there was a lower cell density in the demineralized regions as compared to the mineralized region, although the difference was not statistically significant (Fig. S6C). Overall, bMSCs cultured in osteogenic media showed higher cell density compared to those cultured in basic and chondrogenic media (Fig. S6D).

Two biomarkers of osteogenic differentiation expressed by the bMSCs, alkaline phosphatase (ALP) and osteocalcin (OCN) were used to assess bMSC behavior within the scaffolds. ALP signifies maturing osteoprogenitors and indicates an immature osteogenic stage for the cells. OCN marks the beginning of mineralization of the ECM and indicates bMSC differentiation towards an osteoblastic phenotype [8]. Other osteogenic biomarkers were not used since they either lack clear indication of the specific stage in osteogenic differentiation (e.g., Runx2) or defined detectable ranges (e.g., osteopontin) [8].

Using IHC staining for ALP and OCN, the osteogenic behavior of bMSCs cultured in basic media was characterized to understand the effect of mineral distribution on the differentiation of bMSCs at 4, 7, and 14 days after seeding (Fig. 2, for additional replicates, please see Fig. S7, Column A and Fig. S8, Column A). The IHC staining intensity was quantified using ImageJ and compared as a function of location (mineralized vs. demineralized) and culture time (Fig. 3). In all the samples, the demineralized regions of the scaffolds showed a more intensive staining profile of ALP but less OCN. Mineralized regions of the scaffolds exhibited inverse trends of ALP and OCN staining profiles. By 4 days of static culture in basic media, more intense ALP stain was observed within the matrix deposited by bMSCs in the mineralized region of the scaffold than in the demineralized region of the scaffold. ALP production by bMSCs in both sides of the scaffold was maintained until day 14 when a decrease of ALP staining was observed on both sides (Fig. 2, Column B). In parallel with the ALP staining, stronger OCN staining was observed within bMSCs residing on the mineralized side of the scaffold, as compared to the demineralized side, as early as day 4 of culture, and the staining intensified at day 14. In contrast, very low staining of OCN was observed on the demineralized side of the scaffold until day 14 when a small amount of OCN staining was observed within the cellular matrix deposited by bMSCs (Fig. 2, Column C). The visual observations matched with the ALP and OCN scores analyzed by ImageJ (Fig. 3A, B). ALP scores decreased over time in mineralized regions of scaffolds, while maintained in demineralized region. OCN scores increased over time in bother mineralized and demineralized regions. In addition, no significant differences were observed in terms of Alcian blue, Col II, and Col X stainings on bMSCs residing within the mineralized and demineralized regions of the scaffolds cultured in basic media (Fig S9, Column B, Fig S10, Column A, Fig S11, Column A, Fig S12, Column A, Fig. S13B, D, F).

Figure 2.

Light microscopy images of polarized picrosirius red (for brightfield images, see Fig. S14, Column A) and DAB-stained, seeded scaffolds, cultured in basic media. Each pair of mineralized (M) and demineralized (DM) images are from a single scaffold, different regions. Column A: Seeded scaffolds stained by picrosirius red, observed through crossed polarizers. Mineralized regions showed greener color. Demineralized regions showed redder color [24]. Column B: Seeded scaffolds stained for ALP. Column C: Seeded scaffolds stained for OCN. Images in the same row from Columns (B) and (C) were taken from serial sections of the same scaffolds from the regions highlighted in Column (A). “S” denotes scaffolds, “Mx” denotes matrices deposited by the bMSCs. Demineralized regions of the scaffolds show a more intensive staining profile of ALP while less intensive staining profile of OCN. Mineralized regions of the scaffolds show inverse trends of staining profiles.

Figure 3.

ALP and OCN scores of seeded scaffolds cultured in basic, osteogenic and chondrogenic media. A) and B) Plots of the ALP and OCN IHC scores as a function of culture time in basic media. Significant differences in OCN scores on mineralized and demineralized regions show on Day 4, 7, and 14, where mineralized regions scored higher. C) and D) Plots of the ALP and OCN IHC scores as a function of culture time in osteogenic media. Significant differences of OCN scores show on Day 4 of culture, where mineralized regions scored higher. E) and F) Plots of the ALP and OCN scores as a function of culture time in chondrogenic media. Numerical data are presented in Table S1. Significance of scores between different static culture time points is denoted by ‘#’. Significance of scores between mineralized and demineralized regions is denoted by ‘*’. P values less than 0.05 are indicated by single symbol and P values less than 0.01 are indicated by double symbols. (n = 20 for each of the mineralized or demineralized regions)

To analyze the effects of osteogenic growth factors on bMSC behavior as a function of mineral content, bMSCs were cultured in osteogenic media. Both ALP and OCN IHC stains were observed at high intensity in mineralized and demineralized regions of the scaffolds at days 4 and 7 (Fig. 4, for additional replicates, please see Fig. S7, Column B and Fig. S8, Column B). Notably, both ALP and OCN showed strong staining profiles in the mineralized region of the scaffold (Fig. 4, Column C, highlighted in red circle). The bMSCs in this highlighted area condensed into a cellular mass with intense staining of ALP and OCN, which may indicate formation of new trabecular bone like that observed in native bone tissue [45] (Fig. S4). The intensity of both stains reduced by day 14 (Fig. 4, Columns B, C). The IHC scores suggest that the intensity of ALP and OCN stains in the mineralized regions continuously decreased over time while the OCN stains in the demineralized regions peaked at day 7 (Fig. 3C, D). This result indicates delayed osteogenic behavior of bMSCs in the demineralized regions.

Figure 4.

Light microscopy images of polarized picrosirius red (for brightfield images, see Fig. S14, Column B) and DAB-stained, seeded scaffolds, cultured in osteogenic media. Each pair of mineralized (M) and demineralized (DM) images are from a single scaffold, different regions. Column A: Seeded scaffolds stained by picrosirius red, observed through crossed polarizers. Mineralized regions showed greener color. Demineralized regions showed redder color [24]. Column B: Seeded scaffolds stained for ALP. Column C: Seeded scaffolds stained for OCN. Images in the same row from Columns (B) and (C) were taken from serial sections of the same scaffolds from the regions highlighted in Column (A). “S” denotes scaffolds, “Mx” denotes matrices deposited by the bMSCs. Demineralized regions of the scaffolds show more intensive staining profile of ALP while less intensive staining profile of OCN. Mineralized regions of the scaffolds show inverse trend of staining profiles.

Chondrogenic behavior of bMSCs cultured in media containing TGF-β1 was assessed at days 4, 7, and 14 of culture. Alcian blue was used to evaluate the production of GAGs by bMSCs. IHC stains of collagens type II and X were used to examine for cartilage-like matrix deposition and hypertrophic differentiation of bMSCs, respectively. Throughout the 14 days of culture, we did not observe significant Alcian blue staining within the scaffold (Fig. S9, S12A, B). As positive controls, we observed strong Alcian blue staining in both native cartilage tissue (Fig. S2) and in bMSCs cultured in well plates with the same chondrogenic media (Fig. S1C). The Alcian blue scores in the ECM deposited by the bMSCs in the scaffolds were even lower than those in the ECM of native trabecular bone (Fig. S2D). Increased staining for both collagens type II and X was detected within the cellular matrix deposited within both regions of the scaffolds over time (Fig. 5, Columns B, C, for additional replicates, please see Fig. S10, Column B and Fig. S11, Column B). IHC stains of OCN and ALP were also performed on seeded scaffolds cultured with chondrogenic media. Surprisingly, we observed more intense OCN staining but less intense ALP staining in the mineralized regions of the scaffold by day 4 of culture (Fig. 5, Columns D, E, for additional replicates, please see Fig. S7, Column C and Fig. S8, Column C). ALP and OCN stains in both regions of the scaffolds decreased over time. bMSCs within the demineralized region of the scaffold eventually showed muted OCN staining but slightly positive ALP staining by day 14. In mineralized regions, bMSCs maintained positive OCN staining. These observations matched with the scores analyzed by ImageJ (Fig. 3E, F). Both ALP and OCN scores show a generally decreasing trend overtime. Prominent differences of OCN scores show on day 4 of culture, where mineralized regions scored higher. Collectively, we observed hypertrophic differentiation of bMSCs throughout the scaffolds, as well as osteogenic differentiation of bMSCs in the mineralized regions of the scaffolds even when they were subjected to a chondrogenic biochemical environment.

Figure 5.

Light microscope images of polarized picrosirius red (for brightfield images, see Fig. S14, Column C) and DAB-stained, seeded scaffolds, cultured in chondrogenic media. Each pair of mineralized (M) and demineralized (DM) images are from a single scaffold, different regions. Column A: Seeded scaffolds stained by picrosirius red, observed through crossed polarizer. Mineralized regions showed greener color. Demineralized regions showed redder color [24]. Column B: Seeded scaffolds stained for Col II. Column C: Seeded scaffolds stained for Col X. Column D: Seeded scaffolds stained for ALP. Column E: Seeded scaffolds stained for OCN. Images in the same row from Column (B), (C), (D) and (E) were taken from serial sections of the same scaffold for the regions highlighted in Column (A). “S” denotes scaffolds, “Mx” denotes matrices deposited by the bMSCs. Demineralized regions of the scaffolds show more intensive staining profile of ALP while less intensive staining profile of OCN. Mineralized regions of the scaffolds show inverse trend of staining profiles.

In addition to characterizing the differentiation behavior of bMSCs, the deposition of aligned collagen fibers was observed on samples stained with picrosirius red, imaged through crossed polarizers (Fig. 6) and collagen type I IHC staining (Fig. S15). Type I collagen was observed in the matrix deposited by bMSCs cultured in osteogenic media (Fig. S15). On day 4, no oriented collagen was observed in either side of the scaffold as indicated by samples stained with picrosirius red observed through crossed polarizers (Fig. 6, Column A). By days 7 and 14, the demineralized side showed birefringent collagen fibers, which were absent from the mineralized side (Fig. 6, Columns B, C). The birefringence observed from these collagen fibers indicate the formation of aligned fibrous tissue.

Figure 6.

Picrosirius red-stained, seeded scaffolds, cultured in osteogenic media stained by picrosirius red observed through Column A: bright field, Column B: crossed polarizers. Mineralized regions show greener color. Demineralized regions show redder color [24]. “F” denotes aligned collagen deposited by bMSCs. “S” denotes scaffolds. Each pair of mineralized (M) and demineralized (DM) images are different regions of the same scaffold.

Discussion

Our data demonstrate spatially patterned osteogenic behavior in bMSCs cultured on engineered bone scaffolds, as a function of mineral distribution. We utilized different culture media, i.e., biochemical cues from solution, to drive this behavior. Specifically, the influence of the mineral content in the scaffold is most pronounced when bMSC-seeded scaffolds were cultured in a basic biochemical environment (i.e., basic media). In basic media, increased OCN and decreased ALP were observed in the regions with high mineral content over time. In contrast, bMSCs in the demineralized regions did not express osteocalcin until day 14 of the culture while showing consistently high production of ALP. In osteogenic media (e.g., with dexamethasone, β-glycerophosphate, and ascorbic acid), biochemical cues, coupled with the bone-derived scaffold, enhanced osteogenic behavior of bMSCs as well as increased the seeding density across the entire scaffolds. The larger cell populations residing in the scaffolds could lead to increased production of osteogenic biomarkers and/or an accelerated differentiation process. As demonstrated by ALP and OCN IHC staining scores, we still noticed that osteogenic behavior of bMSCs was slightly delayed in the demineralized regions of the scaffold at day 4. The difference in terms of osteogenic behavior between bMSCs in the mineralized and demineralized regions of the scaffold, however, diminished by day 7. As such, we observed spatial control over bMSC osteogenic behavior as patterned by the mineral distribution in our bone scaffolds.

Osteogenic differentiation of stromal cells promoted by the presence of mineral has been consistently reported by others. Researchers have shown that osteogenesis by bMSCs in terms of calcium deposition and ALP activities was enhanced by the presence of mineral in poly (L-lactic acid)(PLLA)/alginate scaffolds [44]. Others have demonstrated the enhanced OCN expression and cuboidal phenotype of adipose-derived stem cells (ACSs) cultured on a mineralized PLLA/Poly-benzyl-L-glutamate/Collagen scaffold [14]. Our results, for both basic and osteogenic media, align with these findings by other researchers. For example, increasing osteogenesis by ASCs was found along a mineral gradient deposited in a PLGA scaffold [20]. Interestingly, compared to stem cells cultured within these synthetic scaffolds, bMSCs cultured on our bone scaffolds exhibited more rapid osteogenesis in basic media without any osteogenic supplements. Other studies have reported bMSCs and ASCs in the mineralized regions of their scaffolds showing osteogenic markers including OCN and ALPs by day 7 of the cultures in osteogenic media [19,20,22]. We detected OCN expression within cells and cellular matrix by day 4 of the culture in the mineralized regions of the scaffolds. In addition, the bMSCs in the demineralized regions of the scaffold showed OCN at day 14 of static culture in contrast to the lack of OCN expression by ASCs in the unmineralized regions of the PLGA scaffold as previously reported [20]. The earlier onset of osteogenic behavior in our current work may be related to the presence of ECM-associated growth factors in the decellularized bone matrix, which can potentially promote induction of osteogenic behavior. The preservation of these growth factors, however, depends on the processing of the bone cores [27]. We have previously shown our decellularization and demineralization process maintains the underlying protein matrix of the trabecular bone [24]. Therefore, the ECM-associated growth factors were likely preserved in our scaffolds as well.

We also examined how the patterned mineral distribution in our scaffolds affected the bMSC behavior in a chondrogenic environment. We firstly examined cellular phenotype and two chondrogenic biomarkers, collagen type II and GAGs. Although bMSCs exhibited increasing collagen type II production over time, they deposited low levels of GAGs in both the mineralized and demineralized regions of the scaffold. Meanwhile, collagen type X was observed within the matrix deposited by the bMSCs, which indicates hypertrophic differentiation. In addition, the presence of ALP and OCN expression in the bMSC-deposited matrix suggests osteogenic behavior of bMSCs in a chondrogenic biochemical environment. Both of the osteogenic biomarkers decreased over time in mineralized and demineralized regions of the scaffolds. These results could be due to the addition of TGF-β1, which is known to suppress osteogenic behavior in bMSCs [47]. Despite showing an overall reduction in osteogenic biomarkers, bMSCs in the mineralized regions of the scaffolds still showed more intense OCN staining than those in the demineralized regions by the end of the culture period. This observation implies that the osteoinductive properties of mineralized regions in our scaffolds were partially maintained even in a chondrogenic biochemical environment.

Chondrogenesis by bMSCs could be influenced by the presence of mineral along with other physical or biochemical cues. Researchers have shown that the chondrogenesis by bMSCs was enhanced when cultured in a hyaluronic acid (HA)/gelatin hydrogel containing 10wt% HAp [33]. Lower concentrations of HAp in the hydrogel resulted in a downregulation of collagen type II and GAG production while higher concentration of HAp led to hypertrophic or osteogenic differentiation [33]. Our observation of cellular behavior in mineralized regions of the scaffolds agrees with their findings as both hypertrophic (e.g., collagen II and X) and osteogenic biomarkers (e.g., ALP and OCN) were detected. We did not observe evidence, however, of robust chondrogenic behavior (e.g., GAGs) even in the demineralized regions of the scaffold. Notably, other research has shown chondrogenic differentiation of bMSCs depends on the chemistry and structure of the scaffold. Hydrogel scaffolds have been shown to possess a superior ability for inducing chondrogenesis compared to porous scaffolds such as demineralized bone matrix [48]. The injection of a hydrogel material, such as collagen, to the pores of our bone-derived scaffold may enhance chondrogenic behavior of bMSCs given that hydrostatic pressure can reportedly improve the chondrogenic characteristics of demineralized bone matrix [49].

In previous studies, injectable material such as cellular or acellular collagen gels were integrated with fully mineralized bone scaffolds to fabricate an entire engineered tissue interface [30,37]. Due to insufficient formation of aligned collagen fibers at the bone-gel interface, the mechanical properties of the engineered interfacial tissue remained less desirable compared to those observed in native interfacial tissues [30,37]. In this study, collagen deposited by the bMSCs at both sides of the scaffold exhibited different colors when stained by picrosirius red and observed through crossed polarizers. The red colors of the collagen fibers at the demineralized side indicate collagen bundles of increasing diameter or alignment [50,51]. These results showed that bMSCs seeded in the demineralized side of the scaffold deposit thicker and more aligned collagen compared to those in the mineralized side. Therefore, the demineralized part of these mineral gradient scaffolds might integrate better with injectable gels through improved collagen deposition by bMSCs.

Demineralized bone matrix derived from animal models has been used to study orthopedic applications of tissue engineering in vitro and in vivo [52–54]. In this work, we selected a bovine model of high relevance to research in soft-hard tissue interfaces [55,56] Given that neonatal bovine bone is of comparable bone density to healthy human bone [57], cellular response to local mineral content is likely to be similar to human trabecular bone. In addition, bMSCs derived from bovine bone marrow were used to study cellular interaction with bone-derived scaffolds in this study. Although significant differences in cell surface markers have been observed between bovine bMSCs and human bMSCs [58], similar cellular behaviors in terms of osteogenic and chondrogenic differentiation have been reported when they are seeded on bone matrix derived from bovine models [22,59–62]. Future experiments involving the application of human cells may reveal the feasibility of using these scaffolds to repair interfacial tissues in humans.

One limitation of this study is that the presence of HAp in the scaffold may not be the sole factor that dictated the observed patterning of osteogenic behavior. Studies have revealed that the mineral content and the stiffness of the scaffolds are inherently linked [18,20]. Both variables are known to play roles in inducing the osteogenesis of the bMSCs [63–65]. Using our current fabrication technique, we are not able to isolate the distinct effects of mineral content and scaffold stiffness on bMSC behavior. Furthermore, other characteristics of the scaffold materials including the surface chemistry, topography, and porosity can direct the fate of the bMSCs. For example, studies have shown hydroxyapatite scaffolds with pore sizes between ~100 and ~200 μm upregulated ALP and OCN expressions of bMSCs [66]. Our methods of processing the trabecular bone matrix, however, did not significantly modify the trabecular area and the pore size of the scaffolds (Fig. S5), indicating that pore size did not dictate the spatial differences in cellular behavior we observed. Surfaces with higher roughness, phosphorylated ligands, and bone morphogenic protein coatings can also promote osteogenesis of the adhering cells [67–76]. To reveal any possible connections between these surface features and the bMSCs in our scaffolds, other characterization methods such as scanning electron microscopy (SEM) and profilometry will be used in the future. Finally, to better quantify cellular responses to the mineral graded scaffolds quantitatively, a more powerful tool, for example, biomarker assays will be necessary [19,20].

Conclusion

This work investigated the effect of local mineral content on osteogenic behavior of bMSCs cultured on bone-derived scaffolds with gradients of HAp. We found that the mineral content in these scaffolds can controlled osteogenic behavior of bMSC spatially without the addition of osteogenic biochemical conditions, e.g., in basic media. The bMSCs and associated ECM showed OCN staining as early as 4 days in areas of high mineral content, whereas in regions of low mineral content, weaker OCN staining occurred at 14 days. Interestingly, bMSCs in the demineralized side of the scaffold produced thicker and more aligned collagen fibers, potentially indicating use of this phenomenon for integrating these scaffolds with soft tissues. We also examined the capability of the scaffold to induce chondrogenic behavior in the bMSCs. We found, however, that the addition of chondrogenic biochemical cues were insufficient to promote chondrogenesis in these scaffolds. This work demonstrated that our bone-derived scaffolds with mineral gradients can be used to pattern cellular complexity through localized tuning of osteogenic behavior of bMSCs. These results show promise for diversifying the cellular populations in a monolithic scaffold while maintaining the biologically relevant ECM structure necessary for creating engineered tissue interfaces.

Statement of Significance.

Soft tissue-to-bone interfaces, such as tendon-bone, ligament-bone, and cartilage-bone, are ubiquitous in mammalian musculoskeletal systems. These interfacial tissues have distinct, hierarchically-structured gradients of cellular, biochemical, and materials components. Given the complexity of the biological structures, interfacial tissues present unique challenges for tissue engineering. Here, we demonstrate that material-derived cues can spatially pattern osteogenic behavior in bone marrow stromal cells (bMSCs). Specifically, we observed that when the bMSCs are cultured on bone-derived scaffolds with mineral gradients, cells in contact with higher mineral content display osteogenic behavior at earlier times than those on the unmineralized substrate. The ability to pattern the cellular complexity found in native interfaces while maintaining biologicallyrelevant structures is a key step towards creating engineered tissue interfaces.