Mesenchymal stem cells overexpressing BMP-9 by CRISPR-Cas9 present high in vitro osteogenic potential and enhance in vivo bone formation

Abstract

Cell therapy is a valuable strategy for the replacement of bone grafts and repair bone defects and mesenchymal stem cells (MSCs) are the most frequently used cells. This study was designed to genetically edit MSCs to overexpress bone morphogenetic protein 9 (BMP-9) using Clustered Regularly Interspaced Short Palindromic Repeats/associated nuclease Cas-9 (CRISPR-Cas9) technique to generate iMSCs-VPR^BMP-9+, followed by in vitro evaluation of osteogenic potential and in vivo enhancement of bone formation in rat calvaria defects. Overexpression of bone morphogenetic proteins (BMP)-9 was confirmed by its gene expression and protein expression, as well as its targets Hey-1, Bmpr1a, and Bmpr1b, Dlx-5, and Runx2. Protein expression of SMAD1/5/8 and pSMAD1/5/8 iMSCs-VPR^BMP-9+ displayed significant changes in the expression of a panel of genes involved in TGF-β/BMP signaling pathway. As expected, overexpression of BMP-9 increased the osteogenic potential of MSCs indicated by increased gene expression of osteoblastic markers Runx2, Sp7, Alp, and Oc, higher ALP activity, and matrix mineralization. Rat calvarial bone defects treated with injection of iMSCs-VPR^BMP-9+ exhibited increased bone formation and bone mineral density when compared with iMSCs-VPR- and phosphate buffered saline (PBS)-injected defects. This is the first study to confirm that CRISPR-edited MSCs overexpressing BMP-9 effectively enhance bone formation, providing novel options for exploring the capability of genetically edited cells to repair bone defects.

1. INTRODUCTION

Life expectancy has been increasing globally and the number of health complications due to aging, such as bone tissue impairment, has also been increasing. It is well-known that bone tissue has the capacity of regeneration depending on the nature of the injury and the extension of bone loss¹. Diseases such as osteoporosis, diabetes, cancer, and extensive trauma can negatively affect the skeletal system, generating bone defects that require additional surgical procedures to promote healing². Current treatments for these complications, including autografts and allografts, present limitations of availability, adverse effects, potential infection, and low efficiency. In the search of strategies for overcoming these drawbacks, cell therapy has emerged as a promising approach.

Our research group has been investigating bone repair using cell therapy supported by injection of either bone marrow or adipose tissue-derived mesenchymal stem cells (MSCs) and osteoblasts^3–6. Despite the promising results of all these cells in increasing bone formation, none support the competence to completely regenerate bone defects. Here we explored the hypothesis that modifying cells by making them competent to overexpress growth factors can enhance their ability to support repair of bone defects.

Several studies have shown that growth factors, including bone morphogenetic proteins (BMPs), vascular endothelial growth factor (VEGF), insulin-like growth factors 1 and 2 (IGF-1 and IGF-2), and platelet-derived growth factor, can facilitate osteogenesis^7–10. BMPs belong to the transforming growth factor-beta (TGF-β) family and some of them are potent osteoinductive agents¹¹. More than twenty BMPs have been identified and grouped into classes based on their sequences and functions, including those widely recognized by their influences on bone tissue¹². Among them, BMP-9, also designated as growth differentiation factor 2, is considered one of the most osteogenic, but the mechanisms and clinical applications remain unclear.

BMP-9 induces osteoblastic differentiation of stem cells by activating the Smad-dependent signaling pathway^{13, 14} and provides higher osteogenic potential when compared with BMP-2, −4, −6, and −7^{15, 16}. BMP-9 is more resistant to the inhibitory action of the extracellular BMP-antagonist noggin, which can contribute to its increased osteoinductive effectiveness¹⁵. Based on these findings, several groups have evaluated the potential of BMP-9 to repair bone defects. BMP-9 was directly incorporated into different biomaterials^17–19 and BMP-9-loaded scaffolds were combined with cells²⁰ or BMP-9-transduced cell lineages were loaded into scaffolds^21–24. Consistent with expectations, all studies reported that BMP-9 increased bone formation; however, its effect could be impaired by the composition of the scaffold¹⁷. Transfected cells modified by Clustered Regularly Interspaced Short Palindromic Repeats/associated nuclease Cas-9 (CRISPR-Cas9) and/or bone marrow MSCs have not been evaluated. CRISPR-Cas9 is an effective gene editing tool that is more efficient and faster than already existing tools, such as zinc-finger nucleases and transcription activator-like effector nucleases^{25, 26}. By either activation (CRISPRa) or inhibition (CRISPRi), CRISPR-Cas9 has been used to modify the genome of organisms ranging from bacteria to vertebrates²⁷.

In the present study, CRISPRa was used to edit MSCs to overexpress BMP-9. Initially, we immortalized mouse bone marrow MSCs (iMSCs) that were further edited to overexpress BMP-9. After confirming that iMSCs overexpress BMP-9 (iMSCs-VPR^BMP-9+) and that this editing increases their osteogenic potential, we evaluated the effect of locally injected iMSCs-VPR^BMP-9+ on bone formation in rat calvarial defects.

2. MATERIALS AND METHODS

All animal procedures were approved and are in accordance with the animal care guidelines of the Institutional Animal Care and Use Committee of the University of Vermont, College of Medicine and the Committee of Ethics in Animal Research of the School of Dentistry of Ribeirão Preto (Protocol # 2018.1.30.58.8).

2.1. Isolation and culture of MSCs

Four-week-old C57BL/6 male mice were euthanized with overdose of carbon dioxide (CO₂). The distal and proximal ends of the femurs were cut and the bone marrow was extracted by flushing with growth medium (GM) containing α-MEM, 20% fetal bovine serum (Gibco-Life Technologies, Grand Island, NY, USA), 50 μg/mL gentamicin (Gibco-Life Technologies), and 0.3 μg/mL fungizone (Gibco-Life Technologies), as previously described²⁸. The MSCs were pooled and grown in GM in culture flasks (Corning Incorporated, Corning, NY, USA) until sub-confluence to allow selection of MSCs by adhesion to polystyrene. The cells were kept in incubators with a controlled environment (37°C, 5% CO₂). The culture medium was changed three times per week.

2.2. Immortalization of MSCs

To obtain iMSCs, after sub-confluence, MSCs were enzymatically released, seeded in 6-well culture plate (1×10⁵ cells/well), and infected with lentiviral vector pLOX-TERT-ires TK (Addgene plasmid # 12245, Addgene, Cold Spring Harbor, NY, USA; a gift from Dr. Didier Trono), which expresses human telomerase (TERT). Afterward, real-time polymerase chain reaction (RT-PCR) and immunofluorescence assays were used to confirm the expression of TERT in iMSCs. In addition, flow cytometry, cell proliferation, and osteoblastic differentiation of iMSC were assessed to confirm the stemness potential of the MSCs.

2.3. iMSCs characterization

2.3.1. Real-time PCR to evaluate the gene expression of TERT

Both MSCs and iMSCs at their third passage were seeded in 6-well culture plates (1×10⁵ cells/well) and cultured in GM. On day 3, Trizol reagent (Invitrogen, Carlsbad, CA, USA) was used to extract their total RNA and the concentration and purity of the extracted RNA was analyzed. Afterward, a reverse transcription reaction (Kit High Capacity, Invitrogen, USA) was performed to synthetize complementary DNA (cDNA) using 1 μg of total RNA of each sample. The RT-PCR was done in triplicates (n = 3) using 5 μL of SYBR® Green PCR Master Mix and 4.5 μL of cDNA (11.25 ng) in a Step One Plus Real-Time PCR (Gibco-Life Technologies). Beta-actin (β-actin) and glyceraldehyde-3-phosphate dehydrogenase (Gapdh) were used as housekeeping genes. Relative gene expression of TERT (Table 1) was calculated by the 2^−ΔΔCT method and β-actin was selected for normalization based on its cycle threshold value.

Table 1.

Primers for real-time PCR

Gene	Forward	Reverse
Human TERT	GGCTTCAAGGCTGGGAGGAAC	AGCACACATGCGTGAAACCTG
Mouse Bmp-9	CAGAACTGGGAACAAGCATCC	GCCGCTGAGGTTTAGGCTG
Mouse Hey-1	GGCCTGCTTGGCTTTTCT	CCAAGTGCAGGCAAGGTC
Mouse Bmpr1a	TTTCCAGCCCTACATCATGGC	GCTCCAACTTACTTCATCGCT
Mouse Bmpr1b	CCCTCGGCCCAAGATCCTA	CAACAGGCATTCCAGAGTCATC
Mouse Dlx-5	GCCCCTACCACCAGTACG	TCACCATCCTCACCTCTGG
Mouse Runx2	CGACAGTCCCAACTTCCTGT	CGGTAACCACAGTCCCATCT
Mouse Sp7	ATGGCGTCCTCTCTGCTTG	TGAAAGGTCAGCGTATGGCTT
Mouse Alp	CCAACTCTTTTGTGCCAGAGA	GGCTACATTGGTGTTGAGCTTTT
Mouse Oc	CTGACAAAGCCTTCATGTCCAA	GCGCCGGAGTCTGTTCACTA
Mouse β-actin	GGCTGTATTCCCCTCCATCG	CCAGTTGGTAACAATGCCATGT
Mouse Gapdh	AGGTCGGTGTGAACGGATTTG	TGTAGACCATGTAGTTGAGGTCA

2.3.2. Immunofluorescence assay to evaluate the protein expression of TERT

Both MSCs and iMSCs at their third passage were seeded in 24-well culture plates (1×10⁴ cells/well) on coverslips and cultured in GM. After 24 h, both cells were fixed for 20 min, permeabilized for 5 min, and blocked for 1 h at room temperature using 10% buffered neutral formalin, 0.5% Triton X-100/PBS, and 1% bovine serum albumin. Both cells were incubated with primary monoclonal antibody IgM anti-TERT (1:200; Thermo Fisher Scientific, MA5-16034, Waltham, MA, USA) and Alexa-Fluor conjugated secondary antibody (1:2000; A-11094, Invitrogen) for 1 h and 30 min. Nuclei were stained with 4′, 6-diamidino-2-phenylindole dihydrochloride (DAPI, Thermo Fisher Scientific). The coverslips were assembled on glass slides and images of the cells were acquired with a Zeiss Apotome 2 fluorescence microscope (Carl Zeiss Inc., Oberkochen, Germany).

2.3.3. Evaluation of surface markers of MScs by Flow cytometry

MSCs at their second passage and iMSCs at their tenth passage were grown in GM in culture flasks (Corning Incorporated) until sub-confluence. Afterward, the cells were detached from the polystyrene flasks using 0.25% trypsin (Gibco-Invitrogen), washed with PBS, Gibco-Invitrogen), and incubated with the following anti-mouse antibodies: anti-SCA1, CD29, CD31, CD44, CD45, CD105, CD117 and CD150 (BD Biosciences, San Jose, CA, USA). Flow cytometry was performed in FACSCanto^™ system (BD Biosciences).

2.3.4. Cell proliferation assay

The iMSCs at passages 10, 20, and 30 were seeded in 6-well culture plates (1×10⁴ cells/well) and cultured in GM. On day 3, 7, and 10, the cells were detached from the polystyrene plates using 0.25% trypsin (Gibco-Invitrogen) and counted using an automatic cell counter (Countess Cell Counter, Invitrogen).

2.3.5. Osteoblastic differentiation of iMSCs

2.3.5.1. RT-PCR to evaluate the gene expression of bone markers

The iMSCs at passage 30 were seeded in 6-well culture plates (1×10⁵ cells/well) and cultured in GM or osteogenic medium (OM), which is GM containing 50 μg/mL ascorbic acid (Gibco-Life-Technologies), 7 mM β-glycerophosphate (Sigma-Aldrich, St. Louis, MO, USA), and 10⁻⁷ M dexamethasone (Sigma-Aldrich). On day 7, RNA extraction and RT-PCR were performed as described above (item 2.3.1) to evaluate the relative gene expression of runt-related transcription factor 2 (Runx2), osterix (Sp7), alkaline phosphatase (Alp), and osteocalcin (Oc) (Table 1).

2.3.5.2. In situ ALP activity

The iMSCs at passages 10, 20, and 30 were seeded in 6-well culture plates (1×10⁵ cells/well) and cultured in GM or OM. On day 14, in situ ALP activity was detected in cells stained with a solution containing 1.8 mM Fast Red-TR 1,5-naphthalenedisulfonate salt (Sigma-Aldrich) and 0.9 mM Naphtol AS-MX phosphate (Sigma-Aldrich). The culture plates were kept in incubators for 30 min, solution was removed, and samples were examined in a Leica MZ6 modular stereomicroscope (Leica Microsystems, Bensheim, Germany) attached to a digital camera (DFC310 FX camera, Leica Microsystems).

As described above, we immortalized primary MSCs and confirmed the gene and protein expression of TERT at passage 3 after cell transduction, which was considered “passage zero.” The following experiments were performed after this passage.

2.4. CRISPR-Cas9 construction

2.4.1. iMSCs expressing a constitutive dCas9-VPR for CRISPRa

The iMSCs at passage 4 were seeded in 6-well culture plates (1×10⁵ cells/well) and infected with lentiviral vector Lenti-EF1a-dCas9-VPR-Puro (Addgene plasmid # 99373; a gift from Dr. Kristen Brennand) as previously described²⁹. After transduction, the iMSCs expressing a constitutive dCas9-VPR (iMSCs-VPR) were selected through puromycin resistance test for 6 days. Afterward, the protein expression of dCas9-VPR was evaluated by Western blot, as described below. The non-transduced iMSCs were used as control.

2.4.2. Western blot assay to evaluate the protein expression of dCas9-VPR

The iMSCs and iMSCs-VPR at passage 6 were seeded in 6-well culture plates (1×10⁵ cells/well) and cultured in GM. On day 3, the proteins of each group were transferred to PVDF membrane, which was incubated with primary monoclonal antibody IgG1 anti-CRISPR (Cas9) (1:500; Biolegend, 844301, San Diego, CA, USA) and secondary antibody goat anti-mouse IgG1-HPR (1:3000; Santa Cruz Biotechnology, sc-2060, Dallas, TX, USA) overnight at 4°C and 1 h at room temperature, respectively. The control was beta-actin protein detected with primary monoclonal antibody IgG1 anti-beta-actin (1:1000; Cell Signaling Technology, 8H10D10, Danvers, MA, USA) and secondary antibody goat anti-mouse IgG1-HPR (1:3000; Santa Cruz Biotechnology, sc-2060). The protein bands were detected using a Western Lightning Plus Kit (PerkinElmer–Life Sciences, Waltham, MA, USA).

2.4.3. Single guide RNA design

Single guide RNAs (sgRNAs) were chosen to target 1000 bp upstream to transcriptional start site (Figure S1) using Benchling platform (http://blenchling.com). Four guides for BMP-9 overexpression were designed and sgRNAs with twenty base pairs were selected considering the best score for “on target” and “off target.” The oligonucleotide sequences were purchased from Invitrogen (Gibco-Life-Technologies); all guides were cloned into Lenti_sgRNA_EFS_GFP vector (Addgene plasmid # 65656; a gift from Christopher Vakoc) and lentiviral particles were produced.

2.4.4. Transduction of iMSCs-VPR with lentiviral particles containing sgRNA to overexpress BMP-9

The iMSCs-VPR at passage 6 were infected with lentiviral particles containing sgRNAs to overexpress BMP-9 in order to generate iMSCs-VPR^BMP-9+. As the lentiviral particles also contained a sequence of green fluorescent protein (GFP), cell sorting was performed to select the iMSCs-VPR^BMP-9+, which are also GFP positive cells.

2.5. Gene and protein expression of BMP-9 and its targets

The iMSCs-VPR and iMSCs-VPR^BMP-9+ at passages 8–14 were cultured in GM up to 14 days. The RNA extraction and RT-PCR reactions were performed as described above (item 2.3.1) to evaluate the relative gene expression of Bmp-9 and its target genes Hes related family BHLH transcription factor with YRPW motif 1 (Hey-1), bone morphogenetic protein receptors type 1A and 1B (Bmpr1a and Bmpr1b), distal-less homeobox 5 (Dlx-5), and Runx2 (Table 1). The BMP-9 protein expression was evaluated by enzyme-linked immunosorbent assay (ELISA). In brief, iMSCs-VPR and iMSCs-VPR^BMP-9+ were grown in GM in culture flasks (Corning Incorporated). On day 6, the GM was changed by medium containing only α-MEM (Gibco-Life Technologies). After 24 h, supernatant of iMSCs-VPR and iMSCs-VPR^BMP-9+ cultures were collected and concentrated using Vivaspin^® 20 (5 kDa, Merk). ELISA assay was performed using Mouse BMP-9 ELISA Kit (Abcam, ab267576, Cambridge, UK). The expressions of SMAD1/5/8 and phospho-SMAD1/5/8 (pSMAD1/5/8) were evaluated by Western blot as described above (item 2.4.2) using primary polyclonal antibody IgG anti-SMAD1/5/8 (1:600; Santa Cruz Biotechnology, sc-6031) and primary polyclonal antibody anti-pSMAD1/5/8 (1:500; EDM Millipore, AB3848, Burlington, MA, USA). The control was beta-actin protein detected with primary monoclonal antibody IgG1 anti-beta-actin (1:1000; Cell Signaling Technology, 8H10D10) and secondary antibody goat anti-mouse IgG1-HPR (1:3000; Santa Cruz Biotechnology, sc-2060). The protein bands were detected with a Western Lightning Plus Kit (PerkinElmer–Life Sciences).

2.6. Evaluation of the effect of overexpression of BMP-9 on expression of genes related to TGF-β/BMP signaling pathway

The iMSCs-VPR and iMSCs-VPR^BMP-9+ at passage 10 were seeded in 6-well culture plates (1×10⁵ cells/well), cultured in GM and, on day 3, the total RNA of each culture was extracted as described above (item 2.3.1). A reverse transcription reaction (RT² First Strand Kit (Qiagen, Hilden, Germany) was performed to synthetize cDNA using 1.5 μg of total RNA of each sample. The expression of 84 genes from Custom RT² Profiler PCR array plates (RT² Profiler TM PCR Array Mouse TGF-β/BMP Signaling Pathway, PAMM-035Z, Qiagen) was detected using RT² SYBR Green ROX qPCR Mastermix (Qiagen) and performed in a CFX96 RT-PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA). Five housekeeping genes were used for normalization: β-actin, beta-2 microglobulin (B2m), Gapdh, glucuronidase-beta (Gusb), and heat shock protein 90 alpha (Hsp90ab1). The results were calibrated by iMCS-VPR and analyzed by Qiagen web portal at GeneGlobe Data Analysis Center (https://geneglobe.qiagen.com/us/analyze/).

2.7. Evaluation of the effect of overexpression of BMP-9 on osteoblastic differentiation

The iMSCs-VPR and iMSCs-VPR^BMP-9+ at passage 10 were seeded in 6-well culture plates (1×10⁵ cells/well) and cultured in GM. RNA extraction was performed on days 3, 7, and 14 and RT-PCR was performed as described above (item 2.3.1) to evaluate the relative gene expression of Runx2, Sp7, Alp, and Oc (Table 1). In situ ALP activity was evaluated on days 3, 7, 10, and 14 and extracellular matrix mineralization was evaluated on day 21. In brief, both iMSCs-VPR and iMSCs-VPR^BMP-9+ at passage 10 were seeded in 24-well culture plates (1×10⁵ cells/well), cultured in GM for ALP activity assay, as described above (item 2.3.5) or cultured in OM for extracellular matrix mineralization assay. For this, samples (n = 4) were fixed with 10% buffered neutral formalin for 24 h at 4 °C and stained with Alizarin red staining (Sigma-Aldrich). Afterward, the wells were photographed using a Zeiss Apotome 2 fluorescence microscope (Carl Zeiss Inc.). ImageJ software version 1.5i (National Institutes of Health, Bethesda, MD, USA) was used to quantify the ALP activity and extracellular matrix mineralization by counting the pixels of each group.

2.8. Evaluation of the effect of iMSCs overexpressing BMP-9 on bone formation

2.8.1. Creation of calvarial bone defects

Calvarial bone defects were created in 36 male Wistar rats weighing 150–200 g (n = 12, for each group). This sample size was based on previous studies using cell therapy to regenerate calvarial bone defects^{4, 5}. In brief, rats were anesthetized and all preoperative cares were taken, such as trichotomy and antisepsis of the surgical area. Afterward, an incision was made toward the sagittal suture to expose the parietal bones and create a unilateral 5-mm diameter bone defect. After suturing the surgical wound, the region of bone defect was delimited using permanent markers on the skin to ascertain the exact location for future cell injection. A combination of antibiotics, anti-inflammatories, and anesthetics were administered as postoperative care to avoid animal discomfort and prevent infection.

2.8.2. Cell injection

To mimic a pre-existing bone defect, local injection was performed 2 weeks post-defect creation. For this, rats (n = 12 for each treatment) were anesthetized and randomly allocated to receive local injection of 50 μL of PBS (Gibco-Life-Technologies) containing 5×10⁶ iMSCs-VPR or iMSCs-VPR^BMP-9+ or without cells using a micropipette coupled to a 21-gauge needle. The needle insertion site was approximately 1 cm away from the edge of the bone defect. Afterward, it was moved in the anteroposterior direction of the animal, tangent to the skullcap, until its tip reached the center of the bone defect, with the bevel facing ventrally. Four weeks after the injection, the rats were euthanized, calvarias with the defects were harvested, fixed in 4% paraformaldehyde, and processed to evaluate the bone formation by micro-computed tomography (μCT) and histological analyses. To allow comparison of the bone formed by iMSCs-VPR^BMP-9+ with the native bone, 5-mm diameter fragments of contralateral calvarias were harvested and processed for μCT analysis.

2.8.3. μCT analysis

The morphometric analysis was performed using the SkyScan 1172 system (Bruker-Skyscan, Kontich, Belgium) by a single blinded investigator. The parameters of the μCT scan acquisition were voxel size (9.92 μm³), source voltage (59 kVp), source current (165 mA), filter (Al 0.5 mm), exposure (610 ms), rotation step (0.3 deg), frame averaging (4n), and random movement (10). The images were reconstructed using NRecon software (Bruker-Skyscan), with smoothing set at 2, ring artifact correction set at 5, and beam hardening correction set at 20%. The volume of interest (VOI) selected to determine the borders and limits of the defects was 5-mm diameter and 0.5-mm thick. After the VOI delimitation, bone segmentation within the defect was defined in the gray histogram (0–255) between 70 and 255 and the morphometric data bone volume (BV), bone volume/total volume (BV/TV), bone surface (BS), trabecular thickness (Tb.Th), trabecular number (Tb.N), and bone mineral density (BMD) were automatically generated using 3D Ctan software (Bruker-Skyscan), as previously described³⁰.

2.8.4. Histological analysis

After the μCT, the calvarial samples were processed for histological analysis. In brief, they were decalcified, dehydrated, and diaphanized using 4% ethylenediaminetetraacetic acid (EDTA, Merk, Darmstadt, Germany), a sequence of alcohols, and xylol, respectively. Afterward, the samples were embedded in paraffin, cut to 5 μm thickness in a microtome, mounted on glass slides, and stained with hematoxylin and eosin and Masson’s Trichrome. The images were acquired with a light microscope attached to a digital camera DFC310 FX (Leica Microsystems) and a blinded reader performed the histological description.

2.9. Statistical analysis

In vitro experiments were performed in triplicates and repeated three times. The presented data are from one representative experiment. Student’s t test was used to analyze the RT-PCR data (n = 3 for each gene), ELISA assay (n = 3), RT-PCR array (n = 3), and extracellular matrix mineralization (n = 4). Two-way Analysis of Variance followed by Tukey’s post hoc test was used to analyze the data obtained from cell proliferation assay (n = 5), RT-PCR for evaluating the effect of overexpression of BMP-9 on osteoblastic differentiation (n = 3 for each gene), and ALP activity (n = 4). Kruskal-Wallis, followed by Dunn’s test and correlation analysis using Spearman’s correlation coefficient, were performed to evaluate the morphometric parameters generated by μCT (n = 12). Statistical significance was set at 5% (p ≤ 0.05).

3. RESULTS

3.1. Immortalization of MSCs

After transduction with a lentivirus vector expressing TERT, primary MSCs were efficiently immortalized to generate a high percentage of cells that kept similar expression of classical MSCs surface markers (Figure S2) and fibroblast-like morphology. Proliferative and osteoblastic differentiation capacities were sustained up to passage 30 (Figure S3).

3.2. Efficiency of BMP-9 overexpression

The Western blot analysis confirmed the presence of dCas9-VPR in iMSCs (iMSCs-VPR, Figure 1A). The iMSCs-VPR were infected with a “pool” of sgRNA to overexpress BMP-9 (iMCS-VPR^BMP-9+) and, after 24 h, the GFP signal of sgRNA was detected in iMCS-VPR^BMP-9+ (Figure 1B–C). The gene expression of Bmp-9 was higher on days 3, 7, and 10 (p = 0.001, for all) in iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR, without significant difference on day 14 (p = 0.197) (Figure 1D). iMCS-VPR^BMP-9+ secreted 83.52 pg/mL of BMP-9 protein in 24 h, while iMSCs-VPR did not secrete detectable amounts of it (Figure 1E). On day 3, the gene expression of BMP-9 targets, Hey-1 (p = 0.001), Bmpr1a (p = 0.001), Bmpr1b (p = 0.012), Dlx-5 (p = 0.001), and Runx2 (p = 0.001) was higher in iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR (Figure 1F). Confirming these results, the protein expression of SMAD1/5/8 and pSMAD1/5/8 was higher in iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR (Figure 1G).

Figure 1. Use of CRISPRa to overexpress BMP-9 in immortalized bone marrow-derived MSCs (iMSCs).

Protein expression of Cas9 in iMSCs and iMSCs expressing a constitutive dCas9-VPR (iMSCs-VPR) on day 3 (A). GFP expression of the single guide RNA in iMSCs-VPR overexpressing BMP-9 (iMSCs-VPR^BMP-9+) on day 3 (B, C). Gene expression of Bmp-9 on days 3, 7, 10, and 14 in iMSCs-VPR and iMSCs-VPR^BMP-9+ (D). Protein expression of BMP-9 in iMSCs-VPR and iMSCs-VPR^BMP-9+ on day 7 (E). Gene expression of BMP-9 target genes, Hey-1, Bmpr1a, Bmpr1b, Dlx-5, and Runx2, in iMSCs-VPR and iMSCs-VPR^BMP-9+ on day 3 (F). Protein expression of SMAD1/5/8 and pSMAD1/5/8 in iMSCs-VPR and iMSCs-VPR^BMP-9+ on day 3 (G). Data are presented as mean ± standard deviation. Asterisks (*) indicate statistically significant differences (p ≤ 0.05) between iMSCs-VPR and iMSCs-VPR^BMP-9+. Scale bar: B-C = 200 μm.

3.3. Effect of BMP-9 overexpression on expression of genes related to TGF-β/BMP signaling pathway

Overexpression of BMP-9 resulted in significant changes in expression of a panel of genes involved in TGF-β/BMP signaling pathway. iMSCs-VPR^BMP-9+, in comparison with iMSCs-VPR, exhibited greater than 1.2-fold upregulation of 51 genes and downregulation of 10 genes (p ≤ 0.05, Table S1).

3.4. Effect of BMP-9 overexpression on osteoblastic differentiation

Gene expression of Runx2 was higher on days 3 and 7 (p = 0.001, for both) and lower on day 14 (p = 0.001) in iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR (Figure 2A). Gene expression of Sp7 was higher on days 3 and 7 (p = 0.001, for both) and lower on day 14 (p = 0.008) in iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR (Figure 2B). Gene expression of Alp was higher on days 3, 7, and 14 (p = 0.001, for all) in iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR (Figure 2C). Gene expression of Oc was lower on day 3 (p = 0.001) and higher on days 7 and 14 (p = 0.001, for both) in iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR (Figure 2D). ALP activity was higher on days 7 and 10 (p = 0.047 and p = 0.001, respectively) in iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR and it was not significantly different on days 3 and 14 (p = 0.872 and p = 0.116, respectively) (Figure 2E). On day 21, extracellular matrix mineralization was higher (p = 0.003) in iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR (Figure 2F).

Figure 2. Effect of BMP-9 overexpression on osteoblastic differentiation of immortalized bone marrow-derived MSCs (iMSCs).

Gene expression of the bone markers, Runx2 (A), Sp7 (B), Alp (C), and Oc (D) in iMSCs expressing a constitutive dCas9-VPR (iMSCs-VPR) and iMSCs-VPR overexpressing BMP-9 (iMSCs-VPR^BMP-9+). In situ ALP activity of iMSCs-VPR and iMSCs-VPR^BMP-9+ (E). Extracellular matrix mineralization of iMSCs-VPR and iMSCs-VPR^BMP-9+ on day 21 (F). Data are presented as mean ± standard deviation. Asterisks (*) indicate statistically significant difference (p ≤ 0.05) between iMSCs-VPR and iMSCs-VPR^BMP-9+.

3.5. Effect of iMSCs-VPR ^BMP-9+ on bone formation

The 3D reconstructed μCT images demonstrated that defects injected with iMSCs-VPR^BMP-9+ exhibited higher bone formation compared with iMSCs-VPR and PBS (Figure 3A–C). The morphometric parameters generated by μCT revealed that BV (Figure 3D, p = 0.001), BV/TV (Figure 3E, p = 0.001), BS (Figure 3F, p = 0.002), and BMD (Figure 3I, p = 0.001) were all higher in the calvarial defects treated with iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR or PBS. The Tb.Th (Figure 3G, p = 0.001) was higher in the calvarial defects treated with iMSCs-VPR than PBS. The Tb.N (Figure 3H, p = 0.003) was higher in the calvarial defects treated with iMSCs-VPR^BMP-9+ than iMSCs-VPR. There were positive correlations among the defects injected with PBS, iMSCs-VPR or iMSCs-VPR^BMP-9+ for the following morphometric parameters: BV (Figure 3D, r_s = 0.835, p = 0.001), BV/TV (Figure 3E, r_s = 0.865, p = 0.001), BS (Figure 3F, r_s = 0.413, p = 0.012), Tb.N (Figure 3H, r_s = 0.357, p = 0.033), and BMD (Figure 3I, r_s = 0.662 p = 0.001). In agreement with the μCT data, the histological analyses demonstrated that the defects injected with iMSCs-VPR^BMP-9+ exhibited higher bone formation when compared with iMSCs-VPR and PBS (Figure 4A–C). In general, defects injected with PBS were filled with connective tissue (Figure 4D and G). Irrespective of treatment with either iMSCs-VPR or iMSCs-VPR^BMP-9+, the new bone exhibited similar histological features with the presence of woven bone, osteocytes, osteoblasts (Figure 4E and H), lamellar bone, cement line, and bone lining cells (Figure 4F and I).

Figure 3. Effect of immortalized bone marrow-derived MSCs (iMSCs) overexpressing BMP-9 on bone formation in rat calvarial defects: μCT analysis.

Three-dimensional reconstructed μCT images and morphometric parameters of new bone tissue formed into rat calvarial defects treated with vehicle phosphate buffered saline (PBS, Control, A), iMSCs expressing a constitutive dCas9-VPR (iMSCs-VPR, B) or iMSCs-VPR overexpressing BMP-9 (iMSCs-VPR^BMP-9+, C) at 4 weeks post-injection. Bone volume (BV, D) bone volume/total volume (BV/TV, E), bone surface (BS, F), trabecular thickness (Tb.Th, G), trabecular number (Tb.N, H) and bone mineral density (BMD, I). Data are presented as mean ± standard deviation. Asterisks (*) indicate statistically significant differences (p ≤ 0.05, n = 12) among PBS, iMSCs-VPR and iMSCs-VPR^BMP-9+. Scale bar: A-C = 500 μm.

Figure 4. Effect of immortalized bone marrow-derived MSCs (iMSCs) overexpressing BMP-9 on bone formation in rat calvarial defects: histological analysis.

Light microscopy of new tissues formed into rat calvarial defects treated with vehicle phosphate buffered saline (PBS, Control, A, D, G), iMSCs expressing a constitutive dCas9-VPR (iMSCs-VPR, B, E, H) or iMSCs-VPR overexpressing BMP-9 (iMSCs-VPR^BMP-9+, C, F, I) at 4 weeks post-injection. New bone tissue was observed in the defects treated with iMSCs-VPR (B, E, H) and iMSCs-VPR^BMP-9+ (C, F, I), while defects treated with PBS were filled with connective tissue (A, D, G). Hematoxylin and eosin staining (A-F) and Masson’s Trichrome staining (H-I). bv: blood vessel; ct: connective tissue; ot: osteocyte; ob: osteoblast; wb: woven bone; lb: lamellar bone; cl: cement line; blc: bone lining cells. Scale bar: A-C = 1.25 mm; D-F = 50 μm; G-I = 100 μm.

4. DISCUSSION

In this study, we generated an immortalized MSCs lineage that was edited to overexpress BMP-9 by CRISPR-Cas9 technology and applied it in cell therapy to repair bone defects. iMSCs were obtained by transfection of bone marrow MSCs with TERT, followed by gene editing to overexpress BMP-9. Overexpression of BMP-9 was validated by its gene and protein expression as well as its targets. This overexpression resulted in modulated expression of a panel of genes involved in TGF-β/BMP signaling pathway and increased potential of osteoblastic differentiation. In vivo results support our hypothesis that locally injected iMSCs-VPR^BMP-9+ can improve bone repair in preexistent bone defects as demonstrated by μCT generated morphometric parameters and histological analysis.

First, our results have shown that MSCs can be efficiently immortalized by TERT, as shown by its gene and protein expression, maintaining the fibroblast-like morphology, displaying classical MSCs surface markers, and sustaining proliferative activity^31–32. In addition, iMSCs were competent to differentiate into osteoblasts, even at high passages, as reflected by gene expression of bone markers and ALP activity. These findings agree with studies that evaluated other cell lines immortalized by overexpression of TERT and demonstrated genetic and phenotypic stability^33,34.

Several studies have shown that overexpression of BMP-9 can be achieved by non-viral gene delivery or adenovirus techniques using ex vivo or in vivo experimental designs^24,35–39. However, the direct gene delivery in vivo is limited due to low transfection efficiency, transient expression, and low selective cell targeting⁴⁰. In this study, using CRISPR-Cas9 in an ex vivo model, it was possible to overexpress BMP-9 in MSCs for up to 10 days and induce bone formation in vivo. The advantage of this strategy was that MSCs produced an osteoinductive factor for a longer half-life and, therefore, may increase the recruitment and differentiation of osteoprogenitor cells, leading to an intensified formation of new bone⁴¹. To document successful gene editing, iMSCs-VPR^BMP-9 were characterized by gene and protein expression of BMP-9, some of the targets of BMP-9, BMP signaling pathway, and osteoblastic differentiation. As anticipated, iMSCs-VPR^BMP-9+ exhibited BMP-9 gene and protein expressions and some of its key targets were upregulated when compared with iMSCs-VPR. Gene expression of Hey-1 (the most significantly upregulated target in response to BMP-9⁴²), Bmpr1a, and Bmpr1b (that both play an essential role in BMP-9-induced osteoblastic differentiation of MSCs⁴³), Dlx-5 (a transcription factor upregulated by BMP-9 through SMAD signaling pathway¹⁵), and Runx2 (a principal transcription factor involved in osteoblastic differentiation^44,45) was upregulated in iMSCs-VPR^BMP-9+. Expression of proteins involved in the canonical BMP pathway SMAD1/5/8 and pSMAD1/5/8⁴⁶ were increased in iMSCs-VPR^BMP-9+.

We interrogated the consequences of BMP-9 overexpression on TGF-β/BMP signaling pathway. It is well-known that BMP-9 exerts several biological effects through at least ten different pathways, but the TGF-β/BMP canonical pathway is a significant driver in osteoblastic differentiation⁴⁷. Noteworthy, there was activation of several components of this pathway, since the expression of Tgf-β-1, -2, and -3, Bmp-1, -2, -5, -6, -7, and -9, their receptors, and intracellular signaling molecules (Smad-2, -3, and -4) were all increased. Transcription factors prominently involved in osteoblastic differentiation, such as include Runx2 and Sox4, were upregulated. Interestingly, the expression of Igf-1, which is related to the enhancement of the osteogenic potential of BMP-9⁴⁸ and Igfbp3 (insulin-like growth factor-binding protein 3) that modulates the anabolic activities of BMPs and Igf-1^49,50, were upregulated. Expression of Bglap2 (osteocalcin) was downregulated potentially due to the increased of Tgf-β-1 expression⁵⁰.

The effects of BMP-9 overexpression on osteoblastic differentiation were confirmed by expression of bone markers including Runx2, Sp7, Alp and Oc^44,51,52, ALP activity, and extracellular matrix mineralization. The iMSCs-VPR^BMP-9+ showed higher gene expression of Alp from 3 to 14 days, while gene expression of Runx2 and Sp7 was upregulated only in the early periods of the osteoblastic differentiation process and Oc was upregulated in the late periods. Overall, the profiles of expression of these markers of bone formation are consistent with their known properties during osteogenesis and osteoblast differentiation. Corroborating the enhanced osteoblastic differentiation potential of BMP-9 that has been reported^15,16, both ALP activity and extracellular matrix mineralization were higher in iMSCs-VPR^BMP-9+ when compared with iMSCs-VPR. These findings are consistent with reported increases in phenotypic bone markers in mouse adipose-derived MSCs infected with Ad-BMP-9⁵³.

To address the clinical application, iMSCs-VPR^BMP-9+ were used in cell therapy to repair bone defects. For this protocol, we created a defect and, after 2 weeks, the bone defects were treated by injection of 5×10⁶ cells through a 21-G needle that does not affect the cell viability⁴. Because we directly injected cells overexpressing BMP-9 into the defects without any biomaterial 2 weeks after the defect creation, the newly formed connective tissue could act as a natural scaffold to retain the cells in the bone defects. Comparison of the effects of iMSCs-VPR^BMP-9+ and iMSCs-VPR on bone formation demonstrated that overexpression of BMP-9 enhances bone formation. Five out of six parameters were increased in defects treated with iMSCs-VPR^BMP-9+ and this finding is corroborated by the positive correlations among PBS, iMSCs-VPR, and iMSCs-VPR^BMP-9+. Also, by normalizing BV/TV of iMSCs-VPR^BMP-9+ to its respective PBS, the rate (2.97) is almost 2-fold higher than the rate of iMSCs-VPR (1.48) or even of the rates of our previous study using MSCs derived from either bone marrow (1.79) or adipose tissue (1.42)⁴, indicating that gene-editing iMSCs to overexpress BMP-9 is highly effective to increase bone formation. Regarding the extent of in vivo bone formation induced by BMP-9, these results are similar with some data in literature despite using different approaches, since MSCs treated with chemically modified RNA encoding BMP-9 associated with a collagen scaffold were used elsewhere⁵⁴. In addition, we compared the bone tissue formed by iMSCs-VPR^BMP-9+ with native bone of the calvaria. The native bone presented around 64% of BV/TV, while our results showed that the bone formed by iMSCs-VPR^BMP-9+ presented around 26% (Figure S4). It is worthy to note that the Tb.N and BMD of bone formed by cells overexpressing BMP-9 were very similar to the native bone, which is biologically and clinically relevant, because BMD is a reference for evaluating bone quality^55–56. Our results indicate that in addition to increasing the amount of bone formation, iMSCs-VPR^BMP-9+ induce a bone tissue similar in quality to the preexistent bone. These μCT findings were corroborated by histological analysis, which confirmed that more bone tissues are present in the iMSCs-VPR^{BMP-9 +}-injected defects when compared with iMSCs-VPR and PBS without inflammatory indicators in any of the histological sections. These results support previous reports on the positive effect of BMP-9 on bone formation either directly incorporated into scaffolds or in combination with scaffolds loaded with BMP-9-transduced cells^20,21,23.

The many possible mechanisms by which the overexpression of BMP-9 leads to enhanced bone formation is yet to be addressed. However, considering that the healing effects of MSCs rely on their secreted products (cocktails of growth factors, cytokines, and hormones), release of extracellular vesicles, and cell–cell interactions⁵⁷, it is reasonable to anticipate that these mechanisms may mediate the ability of iMSCs-VPR^BMP-9+ to increase and/or accelerate bone formation. The presence of iMSCs-VPR^BMP-9+ in bone defects may have resulted in a limited inflammatory response, which facilitates osteogenesis and angiogenesis, thus contributing physiologically to bone repair⁵⁸.

In conclusion, to the best of our knowledge, this is the first demonstration that MSCs can be genetically modified to overexpress BMP-9, thus generating a cell lineage with increased potential of osteoblastic differentiation, at least in part, due to activation of TGF-β/BMP signaling pathway, enabling the significant induction of enhanced bone formation. Our findings support the development of cell therapy with gene edited cells for regeneration of bone tissue in challenging sites.