EP4 agonist KMN-159 induces osteogenesis in rodent and porcine dental models Authors: Thomas A Owen, Phillip Rzeczycki, Natalia Da Ponte Leguizamon, Qiming Jin, Chandni Patel, Shanqiao Wei, Stephen D Barrett, Andre J van Wijnen, María Inés Morano Published: 2026 ABSTRACT AbstractThe selective EP4 receptor (PTGER4) agonist KMN-159 stimulates osteoblast function and may support applications in dental bone tissue regeneration and tissue engineering. The goal of this study was to further validate stimulation of osteogenic differentiation ex vivo and in vivo by KMN-159. We tested KMN-159 in cultured human bone marrow-derived mesenchymal stem/stromal cells (BM-MSCs) and rat BM-MSCs grown in a three-dimensional deproteinized bone explant culture model. The effects of KMN-159 treatment in vivo were also investigated in two different alveolar bone repair models (i.e., male rat tooth socket model with and without a dental implant), as well as in a pilot study with a dental repair model in which human implants were insert into the mandibles of male minipigs. Our results show that KMN-159 promotes osteogenic differentiation of human BM-MSCs, (e.g., ALPL enzyme and mineralization) and the temporal expression of osteoblast markers (e.g., RUNX2, SP7, BMP2, TNFRSF11B) and extracellular matrix (ECM) proteins (e.g., BGLAP). KMN-159 also stimulates osteogenesis in rat BM-MSCs in bone-derived 3D scaffolds by accelerating proliferation and modulating expression of osteoblast phenotype markers. In the in vivo studies of bone repair KMN-159 increases bone volume in the tooth socket or around the coronal aspect of the implant in the rat maxillary extraction models. Similar results were obtained in the porcine dental implant repair model. We conclude that KMN-159 is a viable pharmacotherapeutic candidate to promote bone mass accrual or to support bone tissue engineering in dental, craniofacial and skeletal applications.Keywords: Tooth regeneration/transplantation, Receptors, Prostaglandin E, Mesenchymal stem cells, Osteoblasts, Osteoclasts, Gene expression Introduction Achieving better osseointegration of dental implants remains a challenge. Although the success rate for implants is high, patients with comorbidities such as uncontrolled diabetes [1] and osteoporosis, as well as older patients [2], smokers [3] or those with other conditions that negatively affect bone homeostasis, are at greater risk for implant failure due to poor bone quality and/or quantity [4,5]. Poor alveolar bone quality in the maxilla or mandible leads to inadequate primary implant stability and ultimately to early implant failure. When implants fail, there is significant hardship for patients, both financially and in loss of quality of life. In addition to aesthetic losses and time spent undergoing corrective procedures, loss of teeth as a result of implant failure leads to further impaired masticatory function and has been shown to be a risk factor for cognitive decline and frailty. One way to circumvent the issue of inadequate bone quantity and quality in patients would be to develop a therapeutic agent administered at the time of implantation. The ideal therapeutic would enhance both the quantity of bone in the local area and improve osseointegration by increasing the contact between the implant and the surrounding bone. While there are therapeutic agents available to aid local bone formation in the jaw, these agents are relatively expensive, protein-based receptor ligands such as recombinant human bone morphogenetic protein 2 (rhBMP2), platelet-derived growth factor (PDGF), and parathyroid hormone (PTH) [6–8]. The most widely used of these agents, rhBMP2, has raised safety concerns including ectopic bone formation, seroma formation, and alteration of cytokine balance in the treatment area [9]. These economic and safety considerations together limit the widespread use of protein-based bone anabolic agents in dental applications. The use of stem cells for bone regeneration in dental medicine is also being explored, but like the protein-based anabolics, their general application for implant osteointegration has multiple limitations [10]. In contrast, small molecule therapeutic agents (molecular weight 〈1000) are in general manufactured more cost-effectively, have superior stability and shelf life, and have greater versatility in methods of delivery. The naturally occurring fatty acid metabolites Prostaglandins E1 and E2 (PGE1 and PGE2) are bioactive compounds that have long been known to play a crucial role in regulating bone metabolism by modulating both bone resorption [11] and bone formation [12]. Subsequent studies demonstrated that both PGE1 and PGE2 can induce bone formation in mammalian alveolar bone repair models (i.e., rodents and dogs) [13,14] as well as in humans [15]. The physiological effects of PGE1 and PGE2 are mediated through four G protein-coupled receptors designated EP1 through EP4 (gene symbols: PTGER1, PTGER2, PTGER3, and PTGER4) [16,17], but it is well documented that their activation of the EP4 (PTGER4) receptor signaling is responsible for the majority of the bone anabolic effects observed [18,19]. Clinical use of PGE1 or PGE2 as therapeutic agents is limited by undesirable side effects including vasodilation and flushing, lethargy, and diarrhea that render them unacceptable as bone anabolic agents [20]. These effects are due to the pleiotropic pan-activation of all four EP receptors. To address this problem, we developed a series of novel small molecule difluorolactam agonists selective for the EP4 (PTGER4) receptor. We selected one of these agonists, designated KMN-159, as our lead compound to take advantage of the robust bone stimulatory properties of activated EP4 receptors (PTGER4) while excluding interactions with the other EP receptors which are responsible for inducing the side effects (i.e., PTGER1, PTGER2, and PTGER3). Beyond the biological activity of KMN-159 as a potent and selective EP4 agonist that stimulates bone formation, this compound acts locally and is then rapidly cleared, thus avoiding the systemic exposure that might cause EP4 receptor-associated toxicity. KMN-159 is known to induce bone formation in several in vivo models including calvarial repair [21], fracture repair [22], and lumbar spinal fusion [23] and offers superior chemical stability and lower cost of goods when compared to protein-based agents [24]. In this study, we examined the stimulatory activity of KMN-159 during osteogenic differentiation ex vivo and in vivo. We tested KMN-159 in cultured human bone marrow-derived mesenchymal stem/stromal cells (BM-MSCs) and rat BM-MSCs grown in a three-dimensional deproteinized bone explant culture model. The effects of KMN-159 treatment in vivo were also investigated in two different alveolar bone repair models (i.e., male rat tooth socket model with and without a dental implant), as well as in a pilot study with a dental repair model in which human implants were insert into the mandibles of male minipigs. Our results demonstrate that this potent small molecule can stimulate osteoblast-related gene expression in both human and rat MSCs in vitro, as well as stimulate bone formation in two in vivo rat models and a porcine dental model. Materials and methods RNA sequencing and bioinformatic analysis RNA-seq was performed as reported using RNA samples derived from mesenchymal cell types that were previously described [25]. Sequencing data for polyadenylated mRNAs were obtained for dental pulp derived stem cells (DPSCs) [26], adipose tissue derived mesenchymal stem cells (AMSCs) [27,28] and bone marrow derived stem cells (BMSCs) [29–32]. RNA was purified with RNA Sample Prep Kit v2 and transcript libraries were generated using TruSeq SBS Kit v3 on the Illumina HiSeq 2000 (Illumina, San Diego, CA). Transcript reads were processed using HCS v2.0.12 software, MAPRSeq (v.1.2.1), TopHat 2.0.6 [33] and HTSeq [34]. All samples were analyzed using the same bioinformatic pipeline. Transcript reads were analyzed as paired-end reads and normalized as fragments per kilobase per million (FPKM). EP4 receptor binding The rat and human EP4 receptors were separately overexpressed in HEK293T cells and cell membranes prepared. Competition binding assays were conducted from 30 nM to 30 pM KMN-159 against [3H]-PGE2 (180 Ci/mmol), which was held constant at 0.5 nM. Assays were performed in 10 mM 2-(N-morpholino) ethanesulphonic acid (MES) buffer containing 1 mM EDTA and 10 mM MnCl2, pH 6.0, during a 120 min incubation at room temperature. The membrane fractions were recovered on glass fiber filters using a cell harvester, and after washing with cold assay buffer solution, radioactivity was measured with a liquid scintillation analyzer. The individual IC50 values were determined using non-linear regression analysis of the binding-concentration curves of at least 13 data points (n = 3 for each point in each of 3 experiments) using GraphPad Prism 10 software (Dotmatics, Boston, MA). The inhibition constants Ki values were calculated using the Cheng Prusoff equation Ki = IC50 / (1 ± L/Kd) where L is the concentration of the ligand in the assay, and Kd is the affinity of the radioligand for the receptor. A Scatchard plot was used to determine Kd. Culture of human BM-MSCs Human bone marrow-derived mesenchymal stem cells (BM-MSCs; American Type Culture Collection (ATCC), Manassas, VA) were plated at 2500 cells/cm2 in mesenchymal stem cell basal media with growth supplements (ATCC) in either 24-well plates (for measurement of alkaline phosphatase activity) or 6-well plates (for analysis of gene expression) and cultured overnight in a humidified 37 °C/5% CO2 incubator. The media was then removed and replaced with plating media containing either 100 nM KMN-159 or 0.1% DMSO vehicle. Media was subsequently replaced every 3 days with fresh osteogenic medium [basal growth medium plus 10 mM β-glycerophosphate (βGP), 10 nM dexamethasone (dex), and 100 μg/mL L-ascorbic acid; ATCC, Manassas, VA] containing 100 nM KMN-159 or 0.1% DMSO vehicle. For measurement of the dose response of alkaline phosphatase enzyme activity, cells were harvested on Day 14 (Fig. 1) and on Day 15 in the experiments where cells were harvested for gene expression analysis (Fig. 2). Alkaline phosphatase enzyme activity was assessed as described below. For analysis of gene expression, cells were harvested on Days 5, 7, 11, 15, and 21 after plating and analyzed by RT-qPCR as described below. The EP4 receptor (PTGER4) is robustly expressed in human mesenchymal stem cell types relevant to dental tissue regeneration and binds to and activates both human and rat EP4. Panel A: Diagram showing selective activation of EP4 receptor by KMN-159 (right) relative to the endogenous agonist PGE2 which regulates all four receptors and activates multiple second messengers [18,21,24]. KMN-159 is known to activate an osteogenic pathway that induces the cAMP-PKA-CREB signaling cascade [21,24]. Panels B & C: Either human PTGER4 (B) or rat Ptger4 (C) was overexpressed in human embryonic kidney cells (HEK293) to maximize the levels of the EP4 receptor. Competition binding assays with a radioligand (3H-PGE2) show that both the human (B) and rat (C) EP4 receptors similarly bind KMN-159 with Ki values of 211 ± 42 pM (n = 3) and 181 ± 74 pM (n = 3), respectively. All results are expressed as mean ± SD of at least 3 replicates/concentration in each of 3 experiments. Curve fitting of data from competition binding assays using 3H-PGE2 and various doses of KMN-159 was used to determine the KMN-159 concentration that causes half-maximal inhibition (inhibitor constant, Ki) for each receptor, with curves representative of the mean shown. Panels D & E: The effects of KMN-159 on either human donor bone marrow-derived MSCs (D) or rat BM-MSCs (E) were compared. Primary human BM-MSCs were plated and then fed with fresh media containing various doses of KMN-159 the next day and every third day until they were harvested on Day 14. Total primary rat bone marrow cells were treated at plating with various doses of KMN-159 and harvested on Day 8. Alkaline phosphatase activity was quantitated as described in the Appendix. In each case, KMN-159 has a dose-dependent effect on ALPL activity with EC50 values for the human and rat cells, respectively, of 1.14 ± 0.75 nM (n = 3–7 wells for each point in each of 3 experiments) and 1.24 ± 0.72 nM (n = 8 wells for each point in each of 8 experiments). EC50 curves representative of the mean values are shown and demonstrate that KMN-159 is a potent activator of the early stages of osteogenic differentiation in both human and rat BM-MSCs upon high affinity binding to the EP4 receptor. KMN-159 induces the expression of bone phenotype marker genes in human BM-MSCs in 2D culture. We tested the effects of KMN-159 (green closed triangles) compared to vehicle (black closed circles) on the differentiation of human osteogenic cells using expression of genes involved in osteogenesis as a readout. Primary human BM-MSCs were plated in growth media and then fed with fresh media containing 100 nM KMN-159 or vehicle (0.1% DMSO) the next day and every third day. Beginning on Day 5, osteogenic media (growth medium plus 10 mM β-glycerophosphate (βGP), 10 nM dexamethasone (dex), and 100 μg/mL L-ascorbic acid) was used to feed the cells. RNA was isolated from cells on Days 5, 7, 11, 15, and 21 and expression of bone phenotype marker mRNAs was analyzed by RT-qPCR. Gene expression values were calculated using the ΔCq method and data were normalized to the expression of the housekeeping gene GAPDH. Alkaline phosphatase activity was also determined on Day 15. Panels A & B: measurements of ALPL enzymatic activity (A) and alkaline phosphatase (ALPL) mRNA (B) as early markers of osteogenesis show the expected transient upregulation during differentiation on Day 15 that is significantly enhanced although modestly by KMN-159 (<50%; left panel). Panels C-J: These panels show the effects of KMN-159 on mRNA levels of (C,D) bone related extracellular matrix proteins (e.g., COL1A1, BGLAP), (E,F) osteogenic transcription factors (RUNX2, SP7), (G,H) secreted differentiation factors (BMP2, VEGFA), and (I,J) secreted cell/cell communication factors (TNFSF11 (RANKL) and TNFRSF11B (OPG)). Results are expressed as mean ± SD of n = 3–4 replicates group. Statistical analysis in panels B-J was performed using 2-way ANOVA followed by multiple comparisons Sidak's test. A significant time effect (P < 0.001) was identified in all gene expression panels. Effect of treatment with respect to the corresponding vehicle control on the same day is indicated as * P < 0.01, ** P < 0.001, *** P < 0.0001. Panel A was evaluated using unpaired t-test. Measurement of alkaline phosphatase activity in human BM-MSCs Alkaline phosphatase activity in human BM-MSCs was measured as follows. Alkaline phosphatase lysis buffer (150 mM Tris pH 8.4, 0.1 mM ZnCl2, 0.1 mM MgCl2, 1% Triton X-100) was added to the cell layers on ice following one wash with phosphate buffered saline (PBS). Lysates were frozen on dry ice, quickly thawed in a 37 °C water bath, and centrifuged at 13,000 ×g for 10 min at 4 °C. 10 μL of each supernatant was moved in triplicate to a 96-well assay plate, 50 μL of Atto Glow AP Substrate (Promega, Madison, WI) was added to each well, and chemiluminescence measured using a Flex Station 3 microplate reader (Molecular Devices, San Jose, CA). The individual EC50 values in the dose-response experiments were determined using non-linear regression analysis of the signal-concentration curves of n = 3 wells each for at least 14 data points in 3 experiments (GraphPad Prism 10 software, Dotmatics, Boston, MA). Analysis of gene expression in human BM-MSCs On Days 5, 7, 11, 15, and 21 after plating, media was aspirated from cultured human BM-MSCs and the cell layers were rinsed 1× with PBS. RNA was then isolated using the Zymo Quick-RNA kit as described by the manufacturer (Zymo Research, Irvine, CA) and 300 ng was reverse transcribed using iScript Reverse Transcription Supermix for RT-qPCR as described by the manufacturer (Bio-Rad, Hercules, CA). cDNA (9 ng) and gene specific primer pairs (IDT, Coralville, IA) were then used to perform qPCR using SsoAdvanced Universal SYBR Green Supermix and the CFX Opus 96 Real-Time PCR System (Bio-Rad, Hercules, CA). Data were normalized to the housekeeping gene GAPDH and gene expression values were calculated using the ΔCq method. Results are expressed as mean ± SD of n = 3–4 replicates / group. The primer pairs used are specified in (see Supplementary Table 1). Isolation and culture of rat BM-MSCs Total rat bone marrow cells (rat BM-MSCs) were isolated from the femurs and tibiae of 8-week old male Sprague-Dawley rats as described [21]. Cells were plated in T75 cell culture flasks at 6.4 × 105 cells/cm2 in complete osteogenic media (MEMα without nucleotides (Gibco/ThermoFisher, Waltham, MA) supplemented with 10% fetal calf serum (FCS), 10 nM dex, and 50 μM L-ascorbic acid) and either 1 μM KMN-159, 100 ng/mL recombinant human bone morphogenetic protein-2 (rhBMP2; GenScript, Piscataway NJ), or 0.1% DMSO vehicle. Cultures were maintained for 7 days in a humidified 37 °C/5% CO2 incubator prior to experimental use. Measurement of alkaline phosphatase activity in rat BM-MSCs in 2D culture For measurement of the dose response of alkaline phosphatase enzyme activity to KMN-159 in 2D cell cultures of rat BM-MSCs, cells were harvested on Day 8 after plating (Fig. 1). Alkaline phosphatase lysis buffer (150 mM Tris pH 8.4, 0.1 mM ZnCl2, 0.1 mM MgCl2, 1% Triton X-100) was added to the cell layers on ice following one wash with phosphate buffered saline (PBS). Lysates were frozen overnight at −80 °C, quickly thawed in a 37 °C water bath, and centrifuged at 13,000 ×g for 10 min at 4 °C. 20 μL of each supernatant was moved in triplicate to a 96-well assay plate, mixed with 50 μL of 1 mg/mL 4-nitrophenyl phosphate solution (pNPP; Cayman Chemical, Ann Arbor, MI), and incubated for 60 min at 37 °C. Alkaline phosphatase activity was quantitated by the formation of the p-nitrophenol product as determined by absorbance at 405 nm using a Flex Station 3 microplate reader (Molecular Devices, San Jose, CA). Data were normalized to total protein in the lysate determined using the 660 nm Protein Assay Kit (Pierce/ThermoFisher, Waltham, MA) according to the manufacturer's instructions. The individual EC50 values in the dose-response experiments were determined using non-linear regression analysis of the signal-concentration curves of n = 3 wells each for at least 11 data points in 8 experiments (GraphPad Prism 10 software, Dotmatics, Boston, MA). Production of collagen plug drug delivery device 24 h prior to seeding of rat BM-MSCs onto deproteinized bovine bone graft blocks as described below, (5×5×5 mm; DSI Dental Solutions, Ashdod, Israel), rat type I collagen plugs loaded with either 30 μg KMN-159, 5 μg rhBMP2, or 7.5 mM Tris (pH 9.0) vehicle were aseptically prepared. To accomplish this, Rat Tail Collagen for 3D Hydrogels and its neutralization buffer ((RatCol); Advanced Biomatrix, San Diego, CA) were mixed with either KMN-159, rhBMP2, or Tris (tris(hydroxymethyl) aminomethane) buffer vehicle in a total volume of 1 mL each, with the drug or vehicle being 15% of the total volume. 20 μL droplets were dispensed by pipet onto the surface of a sterile 100 mm cell culture dish and gelation allowed to proceed for at 37 °C for 1 h in a humidified incubator, after which they were allowed to cure overnight at 4 °C in a sealed humidified container. 3D culture of rat BM-MSCs in deproteinized bovine bone graft blocks A deproteinized bovine bone graft block was placed in each well of a 24-well cell culture dish and wet with 300 μL of complete osteogenic media to promote cell attachment and infiltration. Rat BM-MSCs, cultured as described above, were detached from their flasks by trypsinization and counted. A small defect was made in each bone block using a sterile needle and 2 × 105 cells in osteogenic media were seeded into the defect. A collagen plug with the desired drug or vehicle (see above) was then placed into the defect and a 1 × 2 mm 3D printed titanium implant pressed into place (lab of Dr. Haseung Chung; Michigan State University). An additional 500 μL of complete osteogenic media containing 3 × 105 cells was then dispersed onto the outer surfaces of the bone block to seed additional cells. This volume of media submerged the bone block just over its top surface. Bone blocks were cultured in 5% CO2 in a humidified 37 °C incubator for either 14 or 28 days and fed every 3 days by media replacement with fresh complete osteogenic media but with no additional compound dosing. Measurement of alkaline phosphatase enzyme activity in rBM-MSC 3D culture Alkaline phosphatase activity of the rat BM-MSCs growing in bovine bone blocks (6–9 blocks per treatment) was measured on Day 14 (Fig. 4). Media was aspirated from each block which was then rinsed with PBS. Blocks were snap-frozen in liquid nitrogen and then pulverized in a manual stainless steel tissue homogenizer which had been precooled on dry ice. Each pulverized sample was scraped into a tube containing 0.5 mL alkaline phosphatase lysis buffer (see section on 2D culture above), vigorously vortexed for 1 min, and then incubated on ice for 15 min. Remaining insoluble debris was pelleted by centrifugation at 13,000 ×g for 10 min at 4 °C. 20 μL of each supernatant was moved in triplicate to a 96-well assay plate, mixed with 50 μL of 1 mg/mL 4-nitrophenyl phosphate solution (pNPP; Cayman Chemical, Ann Arbor, MI), and incubated for 60 min at 37 °C. Alkaline phosphatase activity was quantitated by the formation of the p-nitrophenol product as determined by absorbance at 405 nm using a Flex Station 3 microplate reader (Molecular Devices, San Jose, CA). Data were normalized to total protein in the lysate determined using the 660 nm Protein Assay Kit (Pierce/ThermoFisher, Waltham, MA) according to the manufacturer's instructions. KMN-159 induces expression of bone phenotype marker genes in rat BM-MSCs in 3D bone block cultures. The biological effects of KMN-159 on osteogenic differentiation markers were examined in rat BM-MSCs that were differentiated in osteogenic differentiation media over a 4-week culture period. Gene expression values were calculated using the ΔCq method and data were normalized to the expression of the housekeeping gene GAPDH. RT-qPCR analyses show levels of the proliferative marker Mki67 (A) and Alpl mRNA (B), in parallel to ALPL enzyme activity on Day 14 (M). RT-qPCR (C) and ELISA (N) show measurements for the late differentiation marker Bglap (OC) mRNA and protein, respectively. RT-qPCR (D) and ELISA (O) show measurements for the ECM marker Col1a1 mRNA and protein, respectively, and (E) shows RT-qPCR data for MGP (encoding an ECM protein). The remaining panels show the effects of KMN-159 and BMP2 on mRNA levels of secreted differentiation-related factors Bmp7 (F), Sost (G) and Vegfa (H); fibrogenic factors Tgfb1 (I) and Ccn2 (CTGF) (J); and the secreted osteoblast/osteoclast communication factors Tnfsf11 (RANKL) (K) and Tnfrsf11B (OPG) (L). The ratio of RANKL to OPG in differentiated rat (Day 14) and human BM-MSCs (Day 15) was plotted to show how KMN-159 increases the RANKL/OPG ratio to support remodeling-dependent bone formation (P). Statistical analysis of mRNA expression in panels A-L (n = 5–12 replicates/group) and protein expression in panels N and O (n = 5–7 replicates/group) was performed using 2-way ANOVA followed by multiple comparisons Tukey's test. A significant time effect (P < 0.001) was identified in all panels except for panel O. Effect of treatment with respect to the corresponding vehicle control on the same day is indicated as * P < 0.01, ** P < 0.001, *** P < 0.0001. One-way ANOVA followed by multiple comparisons Tukey's test was used for panel M statistical evaluation (n = 6–9 replicates/group). Panel P was evaluated using unpaired t-test. Histological analysis of rat BM-MSCs in 3D culture On Day 28, 3 bone blocks from the vehicle treatment group were rinsed three times in PBS and then fixed in 4% paraformaldehyde in PBS for 10 min at room temperature. Following fixation, one block was rinsed three times in PBS and nuclei were stained with DAPI hydrochloride (1 μg/mL, Cayman Chemical, Ann Arbor, MI) for 10 min, followed by three rinses in PBS. After staining and rinsing, thin sections of the bone block (~0.5 mm thick) were cut using a new scalpel blade and nuclei imaged using a Cytation 5 Cell Imaging Multimode Reader (BioTek Instruments/Agilent, Santa Clara, CA). The other two fixed bone blocks were embedded in paraffin and sectioned using standard methods. Sections from the middle of each block were stained with hematoxylin/eosin (H&E) for cellularity or stained for mineral deposition by the method of von Kossa using 5% silver nitrate. Additional sections were analyzed for the localization of alkaline phosphatase activity by incubation with naphthol AS-MX phosphate and fast blue RR salt (Sigma-Aldrich, St. Louis, MO). Analysis of gene expression in rBM-MSCs in 3D culture After either 14 or 28 days in culture, bone blocks were rinsed with PBS, snap-frozen intact in liquid nitrogen, and stored at −80 °C until analysis. The titanium implant was carefully removed from the bone block and samples were mechanically pulverized using a manual stainless steel tissue homogenizer which had been precooled on dry ice to prevent RNA degradation. RNA was then isolated from each sample using a Monarch Total RNA Miniprep Kit (New England Biolabs; Ipswich, MA) as described by the manufacturer. RNA concentration and purity was determined after isolation using a Nanodrop One Microvolume UV–Vis Spectrophotometer. RNA (300 ng) was reverse transcribed using iScript Reverse Transcription Supermix for RT-qPCR as described by the manufacturer (Bio-Rad, Hercules, CA). cDNA (9 ng) and gene specific primer pairs (IDT, Coralville, IA) were then used to perform qPCR using SsoAdvanced Universal SYBR Green Supermix and the CFX Opus 96 Real-Time PCR System (Bio-Rad, Hercules, CA). Gene expression values were calculated using the ΔCq method and data were normalized to the housekeeping gene Gapdh. Results are expressed as mean ± SD of n = 5–12 replicates/group. The primer pairs used are specified in Supplementary Table 2. Osteocalcin and collagen 1α1 ELISA assays For quantification of BGLAP (osteocalcin) and COL1A1 (collagen 1α1) protein on Days 14 and 28, three bone blocks for each treatment at each timepoint were pooled for each biological replicate. These blocks were rinsed with PBS, snap-frozen intact in liquid nitrogen, and stored at −80 °C until analysis. For analysis, the titanium implants were carefully removed from the bone blocks and samples were mechanically pulverized using a manual stainless steel tissue homogenizer which had been precooled on dry ice. Lysates were generated by scraping each pulverized pellet into a tube containing 300 μL of M-PER Mammalian Protein Extraction Reagent (Pierce/ThermoFisher, Waltham, MA), vigorously pipetting, and then centrifuging at 13,000 × g for 10 min. The supernatants were then aliquoted and stored at −80 °C until analyzed. Protein concentrations were determined using the 660 nm Protein. Rat study approval, blinding and analysis The experimental protocols for the rat tooth socket and dental implant models were both approved by the Institutional Animal Care and Use at the University of Michigan. The Unit for Laboratory Animal Medicine facilitated all animal procurement and care at the University of Michigan. The studies conformed to ARRIVE guidelines 2.0 for preclinical studies. These pilot experiments were designed to test KMN-159 efficacy using doses previously established as efficacious in vivo in other bone regeneration models [21], fracture repair [22], and lumbar spinal fusion [23]. The rats were sacrificed using CO2 inhalation followed by pneumothorax (AVMA Guidelines for the Euthanasia of Animals: 2020 Edition) and maxillae were harvested 14 or 28 days following implant placement and fixed in 10% neutral buffered formalin. μCT images were captured using a Scanco μCT 100 micro-computed tomography system (SCANCO Medical AG, Brüttisellen, Switzerland) by the MicroCT Core of the University of Michigan School of Dentistry and analyzed using Dragonfly software. All rats used in our studies were included in statistical analyses. Rat tooth socket model Male Sprague Dawley rats (24 rats total, 200–220 g; Charles River Laboratories, Wilmington, MA) were acclimated for at least 7 days after arrival, pair housed, and given free access to standard rodent chow and water. Each rat was considered an experimental unit. The assignment of rats to the experimental groups was done on the first day of surgery and based on body weights so that the mean body weight of each experimental group was within 10% of the overall mean body weight of all animals. For extraction of the 1st molar, carprofen (5 mg/kg; s.c.) was administered 20 min before surgery. Under a general anesthesia induced by i.p. administration of ketamine (70 mg/kg) / xylazine (7 mg/kg), the rat was laid on the back on a surgery table and a custom mouth opener was inserted. A spoon excavator (#18) was used to separate gingival tissue from the maxillary 1st molar and then inserted into the interproximal area between the maxillary 1st and 2nd molars to luxate the 1st molar. The 1st molar was then extracted. A few animals with residual roots were excluded from the study and analysis. After allowing the extraction socket to heal for one month, and again under ketamine/xylazine anesthesia as done above, an osteotomy with a dimension of 1 mm in diameter and 2 mm in length was created surgically at the extraction site. Care was taken to limit perforation of the maxillary sinus floor as the proximity of the limited bone height is such that perforation occasionally occurs during osteotomy preparation. Helistat™ collagen sponges (Integra Life Sciences, Princeton, NJ) were cut into cubes of 2 mm × 2 mm × 3 mm in a sterile field using a scalpel. Pieces of sponge were then loaded with either 30 or 60 μg of KMN-159 by absorption from a concentrated stock solution in 100% ethanol. A set of sponges were also loaded with an equivalent volume of 100% ethanol as a vehicle control. All sponges were air dried following loading. One loaded sponge was inserted into each test osteotomy and sutured in place. The assignment of groups of rats to specific treatments was performed using a random number generator (Microsoft Excel) by a person who was not involved in any other aspect of the study and this blind was not broken until after the analyses were completed The total number of rats included was 21: 7 for vehicle control (4 for Day 14 and 3 for Day 28), 6 for 30 μg KMN-159 (3 for Day 14 and 3 for Day 28), 8 for 60 μg KMN-159 (4 for Day 14 and 4 for Day 28). Rat dental implant model Male Sprague Dawley rats (50 rats total, 200–220 g; Charles River Laboratories, Wilmington, MA) were acclimated for at least 7 days after arrival, pair housed, and given free access to standard rodent chow and water. Each rat was considered an experimental unit. Group size was based on published work using this model and the expected magnitude of effect. The assignment of rats to the experimental groups was done on the first day of surgery and based on body weights so that the mean body weight of each experimental group was within 10% of the overall mean body weight of all animals. For extraction of the 1st molar, carprofen (5 mg/kg; s.c.) was administered 20 min before surgery. Under a general anesthesia induced by i.p. administration of ketamine (70 mg/kg) / xylazine (7 mg/kg), the rat was laid on the back on a surgery table and a custom mouth opener was inserted. A spoon excavator (#18) was used to separate gingival tissue from the maxillary 1st molar and then inserted into the interproximal area between the maxillary 1st and 2nd molars to luxate the 1st molar. The 1st molar was then extracted. A few animals with residual roots were excluded from the study and analysis. After allowing the extraction socket to heal for one month, and again under ketamine/xylazine anesthesia as done above, an osteotomy was created with a custom drill-bit (apical portion: 0.95 mm diameter, 1 mm long; coronal aspect: 2.2 mm diameter, 1 mm long). The apical part of the osteotomy was for initial fixation while the coronal part of the drill bit created a circumferential osseous defect (1 mm in diameter and 2 mm in length). A rat collagen plug with either 1.2 μg, 6 μg, or 30 μg KMN-159; 5 μg rhBMP2; or 0.1% ethanol vehicle was placed in the osteotomy and a 1 × 2 mm 3D-printed titanium dental implant press-fit through the collagen plug and into the osteotomy. The assignment of groups of rats to specific treatments was performed using the random number generator in Microsoft Excel by an employee of Cayman Chemical who was not involved in any other aspect of the study and this blind was not broken until after the analyses were completed. The total number of rats included was 45: 11 for vehicle control, 8 for 1.2 μg KMN-159, 10 for 6 μg KMN-159, 8 for 30 μg KMN-159, and 8 for 5 μg rhBMP-2. Porcine dental implant model We applied an established porcine dental repair model to assess if KMN-159 can be scaled to larger animals [35,36]. Six male Yucatan minipigs ranging in age from 14 to 26 months were used for this study and were acclimated to the study facility for at least a week before study initiation. This study was performed by Veranex Preclinical Services (Atlanta, GA) under the supervision of the Pathology Principal Investigator with surgery performed by an experienced human implant surgeon. All aspects of the study were approved by Veranex's Institutional Animal Care and Use Committee (IACUC) which granted approval prior to initiation of work as Veranex PCS-SOP-0139. The facility is accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care, International (AAALAC) and registered with the United States Department of Agriculture to conduct research in laboratory animals, ensuring the quality and integrity of the study and its histopathological analyses. Digital images presented in this study are illustrative representations of the histological preparations and analysis results. The treatment groups encompass animals that were treated with KMN-159 at either of two doses (i.e., 200 μg or 2000 μg) or vehicle control (500 mM sodium bicarbonate with 5% ethanol). On Day 0, animals had their 3rd and 4th premolars and 1st molar extracted on each side of the mandible and immediately received three dental implants on each side together with the low dose (200 μg; n = 2 pigs, 12 implants) or high dose (2000 μg; n = 2 pigs, 12 implants) of KMN-159 or with DMSO Vehicle Control (n = 2 pigs, 12 implants) delivered via MatrixOss bone graft substitute (Regenity Biosciences, Paramus, NJ). On the day of extractions and dosing, Naxcel (ceftiofur; 3 mg/kg; IM) was administered as a prophylactic antibiotic and the animals were sedated with Telazol (5 mg/kg; IM), Xylazine (1 mg/kg; IM), and Propofol (4 mg/kg; IV) and intubated. After sedation, an indwelling jugular catheter was placed for introduction of fluids (lactated Ringer's solution; 5 mL/kg/h) and IV medications. The animals were maintained on gas anesthesia (Isoflurane; 5% in O2) and positive ventilation (use of ventilator). Once appropriate anesthetic depth was achieved, each animal was positioned in lateral recumbency. The mouth was braced open to provide adequate surgical access to the mandible. Any calculus and debris from the tooth surfaces was removed and the surgical area rinsed with 3% hydrogen peroxide followed by disinfection with povidone-iodine. Local nerve block was performed by a single injection of bupivacaine (1 mg/kg) at the mandibular foramen. The left mandibular 3rd and 4th premolars and 1st molar were removed using elevation and traction. Extraction sockets were then widened using a 5 mm drill bit and a 3 mm deep osteotomy was then created in the underlying bone using a 2.5 mm drill bit for primary implant stability. A 3 mm wide × 8–10 mm long (tooth dependent) titanium dental implant (Biohorizon) was then placed in the socket and anchored into the mandibular bone osteotomy using an implant motor. For each implant, one vial of the appropriate dose of KMN-159 or vehicle was opened and the dosing solution drawn into a new MatrixOss syringe to the level of the granules visible in the syringe (0.25 mL). The liquid was pressed out of the syringe back into the vial and this draw/ejection repeated 2 additional times. Following this, 0.25 mL of the dosing solution was drawn into the syringe and the now combined dosing solution and MatrixOss granules placed around the implant by the surgeon. Following dosing, the sites were covered by a mucoperiosteal flap which was sutured in place. The right mandibular 3rd and 4th premolars and 1st molar were then removed, implanted, and dosed as described for the left side teeth using the same dose as for the left side. Upon closure of all incisions (implant sites and central venous line), animals were recovered under veterinary staff care and returned to routine care. Following extraction and implantation surgery, animals had free access to water and food (swine chow softened with water) for the duration of the study. Two weeks prior to euthanasia, one animal from each of the three dose groups was sedated and infused with calcein green. Specifically, animals were sedated with Telazol, Xylazine, and Propofol (4 mg/kg; IV) as described for surgery (above); intubated; and maintained on inhaled isoflurane (0–5%) to maintain appropriate anesthetic depth. Calcein green (Cayman Chemical) was prepared in 2% sodium bicarbonate (pH 7.4), sterilized by 0.2 uM filtration in a cell culture hood, dispensed into sterile 50 mL tubes and stored at 4 °C prior to use. The absence of endotoxin in the dosing solution was confirmed by Cayman's QC group using a Nexgen-pts analyzer (Charles River). Calcein green was administered i.v. at a dose of 15 mg/kg over the course of a one-hour infusion. For animals being terminated at 29 days, infusion occurred on Day 13 post-implantation and for animals being terminated at 40 days, infusion occurred on Day 23 post-implantation. Three days prior to euthanasia, one animal from each of the three dose groups was sedated and infused with Alizarin complexone. Specifically, animals were sedated with Telazol, Xylazine, and Propofol (4 mg/kg; IV) as described for surgery (above); intubated; and maintained on inhaled isoflurane (0–5%) to maintain appropriate anesthetic depth. Alizarin complexone (Cayman Chemical) was prepared in 2% sodium bicarbonate (pH 7.4) at 15 mg/mL, sterilized by 0.2 uM filtration in a cell culture hood, dispensed into sterile 50 mL tubes and stored at 4 °C prior to use. The absence of endotoxin in the dosing solution was confirmed by Cayman's QC group using a Nexgen-pts analyzer (Charles River). Alizarin complexone was administered i.v. at a dose of 30 mg/kg over the course of a one-hour infusion. For animals being terminated at 29 days, infusion occurred on Day 26 post-implantation and for animals being terminated at 40 days, infusion occurred on Day 37 post-implantation. On the designated days post-implantation, animals were euthanized by administration of Euthasol (100 mg/kg; IV) according to the guidelines of the American Veterinary Medical Association. The mandible was dissected out from the overlying tissues in all implanted animals and the mandibular foramen and mental foramen were identified. Using an appropriately sized needle, 25–50 mL of 10% neutral buffered formalin (NBF) was infused through vessels of the left and right mandibular or mental foramen using gentle pressure. The left and right mandibles were then separated, but in all samples, the implants were left in place to minimize damage to the surrounding tissues. MicroCT and morphometry analysis of porcine specimens MicroCT analysis (Bruker SkyScan 1173) of mandibular implant sites was performed to generate computed axial cross section slice images using NRecon software. These axial slice images were visualized in 2-D and 3-D with Bruker's DataViewer and CTAn software and analyzed for bone volume and bone mineral density (BMD) within each defect site. Prior to analysis, using DataViewer, each dataset for analysis was aligned orthogonally with the defect axis parallel to the dataset Z (depth) axis. Datasets from each sample were saved individually for each implant site, numbered as 1, 2, and 3 for the 3rd and 4th premolars and 1st molar, respectively. Using CTAn, a volume of interest (VOI) with a 3.5 mm diameter at the bone surface was saved from each site with the VOI centered on the titanium dental implant and manually contoured to a depth and width for each implant site following the site morphology and visible defect boundaries. Using CTAn, the VOIs had a median filter applied and the images were segmented to exclude the titanium dental implant from the analysis and to separate the volume of bone (BV) from the background surrounding each dental implant. Data for each analyzed image set was collected using CTAn's batch manager tool (“BatMan”). Bone content within the defect VOI was performed using global thresholding values, selected by visual matching with the greyscale images. Threshold regions within the ROIs were output as image sets being saved for visual reference of analysis results. Histological analysis Following MicroCT analysis, samples were radiographed and subjected to histological processing. Samples were trimmed further as applicable, processed and embedded in Methyl Methacrylate (MMA) plastic resin. Resulting plastic blocks were sectioned twice through the implant and left unstained for dynamic morphometry analysis. Following dynamic morphometry analysis, one slide from each block was stained with Hematoxylin and Eosin (H&E) and the second slide was stained with Stevenel's Blue (SB). All resulting slides were digitally scanned. Semi-quantitative histopathology evaluation of histology slides was performed by the Study Pathologist, Serge Rousselle, DVM, DACVP in a manner blinded to treatment. After completion of the slide evaluation, the blinded data were documented and saved. The data set was unblinded, sorted and summarized to complete the histopathological interpretation. Slides were screened for all changes, and evaluated for a range of parameters, including relative size of the involved area, implant material intervening presence, resorption/remodeling, bone regrowth/formation (osteoblastic activity), suture site/granuloma, fibrosis ingrowth within defect site, hemosiderin, overall inflammation, inflammatory cell types, polymorphonuclear cells, lymphocytes, plasma cells, macrophages, giant cells, necrosis, fibrosis, fatty infiltrate, and vascularization. Statistical analysis All experiments were performed at least in triplicates as indicated in the corresponding figure legends, and all numerical data were presented as mean ± standard deviation (SD) or mean ± standard error of the mean (SEM) as noted in the respective figures. Temporal gene and protein expression analysis was performed by two-way analysis of variance (ANOVA) followed by multiple comparisons Sidak's or Tukey's tests (Figs. 2 & 4, respectively). Alkaline Phosphatase activity was analyzed using an unpaired t-test (Fig. 2A) or a one-way ANOVA followed by multiple comparisons Tukey's test (Fig. 4M). Analysis of results for the in vivo studies was performed by two-way or one-way ANOVA (Fig. 5B, C & 6D, respectively) with Tukey's post hoc test utilized in multiple comparisons. All the statistical analysis was conducted with GraphPad Prism 10 software (Dotmatics, Boston, MA). P < 0.05 was considered statistically significant. KMN-159 increases bone formation in rat maxillary molar extraction models. Panel A: We tested KMN-159 for bone formation potential in two rat dental repair strategies (Models #1 & #2). In both models, the first maxillary molar was unilaterally extracted and the socket was allowed to heal for 28 days without treatment. Osteotomies were performed by generating lesions that were 1 mm deep × 1 mm wide in the apical portion and 1 mm deep × 2 mm wide in the coronal portion. Treatment proceeded in Model #1 (without implants) or with a Ti-based dental implant (1 × 2 mm) in Model #2). In Model #1, the osteotomy was filled with an absorbable collagen matrix (Helistat®) that was loaded with either KMN-159 or ethanol vehicle. In model #2, the osteotomy received the 3D-printed titanium implant that was inserted through a gelatinous collagen plug. Panel B: KMN-159 increases bone volume in a dose-dependent manner in a tooth socket repair model. MicroCT analysis of the alveolar bone revealed a significant increase on bone volume (BV) relative to total volume (TV) at Day 28 after administration. Results are expressed as mean ± SEM of n = 3–4 animals/group. Statistical analysis of the differences was performed using 2-way ANOVA followed by multiple comparisons Tukey's test. P < 0.01 effect of time and treatment. * P < 0.05 and ** P < 0.01 as compared to any treatment on Day 14. No significant differences were observed during treatment on Day 14. Panel C: KMN-159 increases bone volume in a dose-dependent manner in the alveolar bone in a dental implant model. MicroCT analysis was performed on alveolar bone that was treated with a range KMN-159 concentrations compared to recombinant human BMP2 (rhBMP2) as a benchmark. The data show a robust enhancement in BV/TV at Day 21 after administration with the 30 μg dose of KMN-159. This level of stimulation is comparable to that observed for 5 μg of rhBMP2. Results are expressed as mean ± SEM of n = 8–10 animals/group. Statistical analysis of the differences was performed using 1-way ANOVA followed by multiple comparisons Tukey's test. * P < 0.05 as compared to vehicle control and 1.2 μg KMN-159 treatment. KMN-159 stimulates alveolar bone formation in a porcine dental implant model. Panel A: The experimental design of our studies includes implantation (at Day 0) of dental screws with two doses of KMN-159 (200 μg or 2000 μg per implant) or vehicle (control; 0 μg), with fluorescent labeling of bone surfaces with calcein (green arrows; Day 13 or Day 23) or Alizarin Red (red arrows; Day 26 or D37) and sacrifice (gray arrows; Day 29 or Day 40) for fluorescence microscopy. Panels B &C: Representative MicroCT (B) and histological images (C) of mandibles with dental screws (3 mm; n = 6 with three implants on either side of the mandible). Panels D, E & F: Dynamic histomorphometry results (D) and representative fluorescent images of single tooth sockets treated with vehicle control (E) or a high dose (2000 μg) of KMN-159 for 6 weeks (F) after labeling with calcein (green channel) or Alizarin Red (red channel). The graph (D) shows mean measured distances between adjacent green and red signals (n = 3–5 determinations near the dental screws) at 4 or 6 weeks after implantation. Two-Way ANOVA statistical analysis indicates a significant treatment effect (P < 0.005). The treatment with high dose KMN-159 has significant bone growth effects with respect to control (****P < 0.001) and low dose KMN-159 treatment (***P < 0.003) at 6 weeks of treatment. Results KMN-159 binds and activates the EP4 receptor that is robustly expressed in osteogenic stem cells KMN159 is a selective agonist of the EP4 receptor (PTGER4) that activates the cAMP/PKA pathway, unlike PGE2 which activates all four E prostaglandin receptor (Fig. 1A). We first addressed whether EP4 is expressed in cell types relevant to dental bone tissue regeneration and tissue engineering. PTGER4 (EP4) mRNA is expressed at higher levels in human dental pulp stem cells (DPSCs) and bone marrow stromal cells (BMSCs) compared to adipose tissue derived cells (AMSCs) (Fig. S1). The expression of EP4 in DPSCs, BMSCs and AMSCs indicates that these cell types may be responsive to EP4 agonists such as KMN-159. To determine the affinity of KMN-159 for EP4, we overexpressed both human and rat PTGER4 in HEK293 cells. The data from a radioligand (3H-PGE2) competition binding assay (Fig. 1B & C) show that both the human and the rat EP4 receptors bind KMN-159 with high affinity (Ki = 211 pM and 181 pM, respectively), consistent with previous studies that additionally demonstrated KMN-159-induced activation of cyclic AMP and cAMP-response element binding protein (CREB) in cell-based assays [21,24]. It is known that activation of cAMP by EP4 agonists such as PGE2 mediates osteogenic activation of downstream events, including enhancement of alkaline phosphatase (ALPL) gene expression and enzyme activity (ALP) [19,21]. Indeed, activation of endogenous EP4 by KMN-159 increases secreted ALP in human and rat BM-MSCs (Fig. 1D & E, respectively). These data show that both exogenously and endogenously expressed EP4 levels in both human and rat are activated by KMN-159 at nanomolar or lower concentrations, corroborating previous findings for the rat EP4 receptor [21]. Because ALP is a marker for osteoblast differentiation, these data indicate that KMN-159 is a viable candidate for a pharmacotherapeutic to promote bone mass accrual or to support bone tissue engineering. KMN-159 temporally stimulates osteogenic differentiation, transcription factors and paracrine ligands in human BM-MSCs KMN-159 promotes osteoblastogenesis by enhancing expression of osteoblast biomarkers in classic 2-dimensional human BM-MSCs cultures in osteogenic media. KMN-159 increases the levels of secreted ALPL activity on Day 15 (Fig. 2A), while the expression curves for ALPL mRNA between Days 7 and 15 are shifted towards earlier time points (Fig. 2B). Shifts in expression towards earlier timepoints are also observed for bone-related extracellular matrix (ECM) proteins, transcription factors and secreted protein ligands (Fig. 2C-J), including the bone master regulators RUNX2 and SP7 (osterix) (Fig. 2E&F) that control BGLAP (osteocalcin; OC) gene transcription [37]. Thus, KMN-159 may enhance osteogenesis via increased expression of key transcription factors required for the expression of ECM proteins critical for biomineralization. KMN-159 stimulates expression of secreted paracrine factors that control osteogenesis or osteoclast differentiation by supporting cell/cell communications between differentiating human BM-MSCs, between bone-forming pre-osteoblasts and bone-resorbing osteoclasts, or between differentiating MSCs and vascular endothelial cells. KMN-159 increases the temporal expression of BMP2 mRNA (2- to 4-fold increase after Day 7) (Fig. 2G) but temporal effects on the angiogenic protein VEGFA are less pronounced (Fig. 2H). The mRNA for the osteoclast differentiation factor TNFSF11 (RANK ligand; RANKL) is increased by KMN-159 (Fig. 2I) while the mRNA for the osteoclast-inhibitory protein TNFRSF11B (osteoprotegerin; OPG) is downregulated by KMN-159 (Fig. 2J). Thus, KMN-159 may promote expression of osteogenic factors to generate positive paracrine feedback loops that sustain signals beyond the initial KMN-159 stimulus to control osteogenesis and/or bone remodeling. KMN-159 stimulates osteogenic differentiation of rat BM-MSCs on a 3D bone scaffold ex vivo To extend our findings into an in vitro model more comparable to the in vivo system, we tested KMN-159 in an ex vivo 3D cell culture model for osseointegration using rat BM-MSCs. The culture system is composed of deproteinized and demineralized 3D bone-derived scaffolds (‘bone blocks’) containing a needle-induced osteotomy and 3D-printed titanium implant which are submerged in cell culture medium. Rat BM-MSCs were first pretreated for a week with either BMP2 (100 ng/mL), KMN-159 (1 μM) or vehicle in 2D cultures in order to enrich for responsive stem cells. Following seeding of these rat BM-MSCs onto bone blocks, they were then treated with a single dose of the same agents used for pretreatment and were analyzed on Days 14 or 28 (Fig. 3A). DAPI and H&E staining show that the bone-derived 3D scaffolds permit cell attachment and support expansion of MSCs. Positive staining for both ALP enzyme activity and mineral deposition (von Kossa staining) clearly demonstrates that this bone block culture system supports the osteoblastic differentiation of the rat BM-MSCs (Fig. 3B). Development of a novel 3D bone block system for studying implant osseointegration. Panel A: Schematic diagram of cell, implant, and treatment components in the bone block 3D cell culture system. Rat BM-MSCs were harvested from rat tibiae and femurs and plated in cell culture flasks with either KMN-159 (1 μM), rhBMP2 (100 ng/mL), or vehicle (0.1% ethanol) for 7 days. Cells were then trypsinized and seeded onto 5×5×5 mm pre-wet 3D scaffolds (deproteinized, demineralized bovine bone; 20% cortical bone/80% cancellous bone; ‘bone blocks’). A needle was used to penetrate the bone block to generate an osteomic defect. A 20 μL collagen plug containing either KMN-159 (30 μg), rhBMP2 (5 μg), or vehicle (5 μL ethanol) was placed on top of the defect and a 1 × 2 mm titanium implant was press-fit through the collagen plug into the defect. Blocks were then cultured to endpoints (14 or 28 days). Panel B: Images showing the osteoblastic differentiation of rat bone marrow cells in representative sections from the 3D bone block model system over a 28-day culture period. Representative blocks were fixed on Day 28 and sectioned fresh for DAPI staining of nuclei or paraffin embedded and sectioned for histological analysis (H&E; alkaline phosphatase; von Kossa). DAPI staining of nuclei and H&E staining for overall cellularity after 28 days in culture on bone blocks demonstrate significant cell penetration throughout the block. Staining for ALP enzyme activity and for mineralization (von Kossa) demonstrate the ability of this 3D cell culture system to support the osteoblastic differentiation of the rat BM-MSCs. Temporal stimulation of osteogenic differentiation by KMN-159 in rat BM-MSCs in 3D culture RT-qPCR analysis of rat BM-MSCs in the 3D bone block culture system shows that KMN-159 increases and decreases mRNA levels of the proliferation marker Mki67 (Ki67), on Day 14 and Day 28, respectively (Fig. 4A). Hence, KMN-159 stimulates MSC proliferation early during the time course resulting in decreased proliferation later as cells reach confluence sooner. KMN-159 increases expression of Alpl mRNA on Day 14 and this positive effect subsides in mature cultures at Day 28 (Fig. 4B). The stimulatory effects of KMN-159 on Alpl mRNA is reflected by increased ALPL enzyme activity on Day 14 (Fig. 4M). These data with rat BM-MSCs on 3D bone blocks corroborate findings obtained with human BM-MSCs in 2D culture (see Figs. 2A&B). KMN-159 has stimulatory, albeit not significant, effects on Bglap mRNA (Fig. 4C) but demonstrates significant stimulation of its secreted protein levels on both Days 14 and 28 in rat BM-MSCs grown on 3D scaffolds (Fig. 4N). Significant increases for Col1a1 mRNA (Fig. 4D) and protein (Fig. 4O) were observed only after KMN-159 treatment on Day 28. In contrast, on Day 28 Mgp (matrix gla protein) mRNA was decreased by KMN-159 or BMP-2 treatments (Fig. 4E). The KMN-159 stimulation of Bglap and Col1a1 mRNA and protein levels indicates that KMN-159 is a very potent osteogenic compound. KMN-159 enhances expression of multiple paracrine factors including Bmp7 (Fig. 4F), Sost (sclerostin) (Fig. 4G), and Vegfa (Fig. 4H) in 3D rat BM-MSC cultures on Day 14, although the differences observed for Vegfa did not reach statistical significance. Thus, rat BM-MSCs actively produce Bmp7 and Sost mRNAs in this 3D culture system, reflecting feedback control during osteogenic differentiation. The latter finding is consistent with known crosstalk between BMP and WNT signaling [38]. The effects of KMN-159 are less apparent on Day 28. Although on Day 14, KMN-159 had a statistically significant effect on both Bmp7 (fold change>2, P < 0.0001; Fig. 4F) and Sost (fold change>4, P < 0.01; Fig. 4G) mRNA expression, these differences were not observed on Day 28. Basal expression of Sost and Vegfa mRNA are higher on Day 28, which could be one possible reason for the temporal differences in KMN-159 effects. Expression of the pro-fibrotic growth factors Tgfb1 (Fig. 4I) and Ccn2 (connective tissue growth factor; CTGF) (Fig. 4J) is reduced by KMN-159 on Day 28 (Tgfb1: fold change>2, P < 0.01; Ccn2: fold change>2). Both fibrogenic factors are already expressed at low levels on Day 14 and therefore less sensitive to the suppressive effects of KMN-159. On Day 14, basal expression of the osteoclast differentiation factor Tnfsf11 (RANKL) is high and that of the osteoclast-attenuator Tnfrsf11b (OPG) is lower, while these basal expression levels are reversed on Day 28 (Tnfsf11, Fig. 4K; Tnfrsf11b, Fig. 4L). KMN-159 enhances the expression of both factors on both Days 14 and 28, albeit that enhancement of Tnfrsf11 expression on Day 14 did not reach statistical significance. Although expression of both Tnfsf11 (RANKL) and Tnfrsf11b (OPG) is increased by KMN-159 on Day 14, the mRNA ratio of Tnfsf11 to Tnfrsf11b (RANKL to OPG) dramatically increases upon KMN-159 treatment, consistent with similar changes in the ratio in 2D human BM-MSC cultures (Fig. 4P). The collective modulations and ratio changes for the mRNAs of pro- and anti-resorptive factors and bone anabolic ligands suggest that KMN-159 modifies the competency for remodeling dependent bone formation of rat BM-MSCs. KMN-159 increases bone formation in rat maxillary extraction models The effectiveness of KMN-159 in vivo was examined in two rat dental repair models after molar extraction (Fig. 5A). We first chose to do a small pilot experiment based on the results from previous studies in rat spinal fusion models [23] and our in vitro results (see above) to demonstrate the efficacy of KMN-159, selecting doses expected to be efficacious in forming bone. The model in which this was tested (Fig. 5A, Model #1) involves the unilateral removal of the first maxillary molar followed by tooth socket healing and creation of osteotomy that was filled with an absorbable collagen matrix (Helistat®) loaded with either 30 or 60 μg KMN-159 or ethanol vehicle [39]. We observed that KMN-159 did not significantly affect bone volume in the osteotomies on Day 14, but increased bone volume on Day 28 in a statistically significant, dose-dependent manner (Fig. 5B). We then chose to retest the 30 μg dose of KMN-159 as well as lower doses in the second in vivo model (Fig. 5A, Model #2). This study followed the same extraction and healing strategy as model 1 but was extended with the insertion of a 3D-printed titanium implant (same implants as in the 3D bone block culture system). A gelatinous collagen plug was then used to deliver either KMN-159, ethanol vehicle, or rhBMP-2 (positive control for bone healing) to the osteotomy. Lower doses of KMN-159 were used in Model #2 to permit assessment of the lowest effective concentration of KMN-159 and this experiment was terminated on Day 21 post-implantation to ensure that the analysis examined the dynamic phase of bone repair. Results in this model demonstrate that KMN-159 increased bone volume in a dose-dependent manner around the dental implants of the tooth socket and that a single 30 μg dose of KMN-159 yielded an increase in BV/TV equivalent to that observed for rhBMP2 (Fig. 5C). The in vivo data from both models indicate that KMN-159 supports alveolar bone regeneration for dental repair. KMN-159 stimulates alveolar bone growth in a porcine dental repair model To support future pre-clinical studies in which human dental implants are placed in the Yucatan minipig mandible, we set up a minimal study with this model to gain pilot data for the potential safety and efficacy of KMN-159. A secondary objective of this pilot experiment was to assess the intragroup variation among treatment groups and the potential magnitude of treatment effects to permit power calculations and economize the design for a more comprehensive future study. Our experimental design included two doses of KMN-159 (200 μg and 2000 μg per implant) that were delivered with a human-approved porcine bone graft substitute using a bilateral implantation strategy (n = 36 implants) (Fig. 6A). Alveolar bone healing was assessed by MicroCT (Figs. 6B&C) and dynamic histomorphometry (Figs. 6D-F). Due to the small sample size of the experiment (and/or homeostatic compensation of bone formation by resorption), the MicroCT data were inconclusive and did not exhibit statistically significant differences in bone parameters. In contrast, dynamic histomorphometry revealed that the high dose of KMN-159 resulted in a statistically significant increase in bone formation (Fig. 6D) as determined by quantifying the relative spacing between two differently colored bone-bound fluorophores at 6 weeks (Figs. 6E&F). Thus, KMN-159 may directly stimulate bone formation during alveolar bone healing to support implant integration in this porcine repair model. Discussion These studies experimentally demonstrate that the potent small molecule PTGER4 agonist KMN-159 stimulates osteoblastogenesis as reflected by increased markers of biomineralization and osteoblast-related gene expression. At the molecular level, our study shows that PTGER4 is abundantly expressed in stem cells isolated from sources that could be involved in dental tissue regeneration and that KMN-159 binds with high affinity to both human and rat PTGER4. This study shows that PTGER4 signaling results in a dose-dependent increase in the activity of the osteogenic marker enzyme alkaline phosphatase. We also establish by molecular analysis that KMN-159 modulates the expression of osteogenic factors (e.g., BMP2 and BMP7) and the ratio of pro- and anti-resorptive factors (e.g. Tnfsf11 (RANKL) to Tnfrsf11b (OPG)) that may generate positive paracrine feedback loops that sustain signals beyond the initial KMN-159 stimulus to control osteogenesis and bone remodeling. We previously addressed mechanistic aspects of EP4 receptor activation which has been shown to work through a Gs-alpha dependent mechanism that stimulates cAMP and protein kinase A (PKA) (see Fig. 1A) [21,24]. We have shown that KMN-159-induced activation of both human and rat PTGER4 stimulates cAMP signaling and cAMP-responsive reporter gene expression [21,24]. We previously also demonstrated the selectivity of KMN-159 for the EP4 receptor by measuring binding affinity to the human EP receptor subtypes [24]. KMN-159 is >10,000 times more selective for the human EP4 receptor (Ki = 0.38 nM) than for human EP1, EP2, and other prostanoid receptors. The compound is also >1000 times more selective for the human EP4 than the human EP3 receptor upon transient expression in human cells. Binding of EP4 agonists activates adenylate cyclase resulting in rapid increased formation of intracellular cAMP and activation of the transcription factor cAMP-responsive element-binding protein (CREB) downstream of the signal transduction cascade [24]. We showed that KMN-159 activates cAMP and CREB in human cells transiently expressing the human EP4 receptor (EC50 = 2.6 and 0.09 nM, respectively). More than 10,000 higher concentrations of KMN-159 are required to activate similar biological activity in overexpressed human EP2 receptor. Similarly, results were observed upon transient expression of rat EP4 and rat EP2 receptors in a follow up study [21]. The biological activity of KMN-159 is 10,000 more selective for rat EP4 than rat EP2 with remarkable sub-nanomolar potency in both cAMP accumulation and cAMP responsive element (CRE) reporter activation. To further demonstrate the selectivity of the EP4 receptor, we investigated the activation of alkaline phosphatase (ALP) in rat bone marrow cells with EP receptor-selective agonists [21]. Only KMN-159 was able to elicit a dose-dependent ALP response in contrast to the EP2 and EP3 specific agonists. Moreover, pre-treatment of the BMCs with selective EP1, EP2, and EP4 antagonists 30 min before KMN-159 treatment indicated that only the EP4 antagonist significantly shifted the dose response [21]. Hence, the mechanism of action of KMN-159 is well-established and involves the activation of the cAMP-PKA-CREB pathway to drive osteogenesis, similar to the bone anabolic actions of PTH and related agents. The regenerative effects of KMN-159 can be generated by either osteoinductive or osteoconductive mechanisms [40–42]. Osteoinduction as defined by the recruitment and differentiation of progenitor cells into the osteoblast lineage through activation of the EP4 prostaglandin receptor is most consistent with our data. It is less likely that KMN-159 acts by osteoconduction, because there is no evidence that this soluble ligand becomes incorporated into a passive scaffold that supports bone growth. Rather, our data show that EP4 signaling operates in cell culture to promote recruitment of marrow stromal progenitor cells, osteogenic differentiation, survival of osteoblast-lineage cells, angiogenic coupling, and anabolic bone formation. Mechanistically, KMN-159 activation stimulates the cAMP-PKA-CREB signaling pathway [21,24] that converges on induction of osteogenic transcriptional programs, including Runx2/Cbfa1 expression. Previous studies demonstrated that EP4-selective agonists like KMN-159 induce de novo bone formation, increase osteoblast number and osteoid formation, stimulate mineralized nodule formation, and enhance osteoblast recruitment from bone marrow stromal cells, all hallmark features of osteoinduction rather than osteoconduction [21,24,40–42]. Thus, our findings support the concept that KMN-159 functions as a soluble extracellular osteoinductive ligand that promotes bone regeneration through EP4-dependent anabolic signaling. While we do not exclude the possibility that EP4 activity may also be influenced by interactions with extracellular matrix components, such mechanisms remain largely unexplored and warrant future investigation. One question that arises is whether KMN-159 is capable of driving osteblastogenesis in culture without other osteogenic agents in the medium. Previous studies on the osteoblast differentiation sequence have clearly demonstrated that maturation of cultured osteoblasts only proceeds in the presence of Vitamin C and an exogenous phosphate source such as beta-glycerol phosphate [43–45]. The requirement for Vitamin C is linked to necessary epigenetic events that rely on Vitamin C dependent hydroxylation of DNA, demethylation of repressive histone modifications and cross-linking of collagen [43,44]. Beta glycerol phosphate is necessary as a source of organic phosphate to permit formation of hydroxyapatite crystals [43,45]. This osteogenic process is co-stimulated by agents (e.g., PTH, Abaloparatide, PGE2) capable of stimulating cAMP signaling. Because KMN-159 induces cAMP signaling, it is not predicted to promote osteogenesis in the absence of other osteogenic cues. We employed two rat maxillary molar extraction models to demonstrate alveolar bone regeneration during a 14-to-28-day time-course without or with the engraftment of a titanium implant. Strikingly, a single in vivo treatment of KMN-159 is similarly effective to a single dose of rhBMP2 in the Ti implant engraftment model, reflecting the considerable osteogenic potential of KMN-159 in bone anabolic applications. Systemic activation of the EP4 receptor is known to result in bone formation throughout the skeleton [46] and the local use of KMN-159 has been shown to stimulate bone formation in rat models for a calvarial defect [21], a femoral defect [22], and posterolateral spinal fusion [23]. In all these reports of the anabolic activity of KMN-159 in the skeleton, no side effects were noted at efficacious doses and only minimal and transient side effects were noted at higher doses. These compelling in vivo data indicate that KMN-159 is a promising therapeutic that supports alveolar bone regeneration, and this biological activity may have practical implications for clinical scenarios requiring effective osseointegration of dental implants. The apparent efficacy of KMN-159 in our rat model provided the experimental foundation for pilot studies towards using the Yucatan minipig model in which matrix-based formulations of KMN-159 (e.g., engineered to support burst-release or extended drug release) were delivered to improve bone-dental implant contact and osseointegration. Dynamic histomorphometry data from these studies are encouraging and also suggest that KMN-159 enhances bone formation. Parathyroid hormone (PTH) and its derivatives (e.g., Teriparatide and Abaloparatide), as well as anti-sclerostin antibodies are FDA-approved bone anabolic agents that have been widely used for clinical treatment of osteoporosis, despite some reported side effects [47,48]. These bone stimulatory agents and anti-resorptives such as the bisphosphonates and anti-RANKL antibodies, have not been fully explored in dental regenerative procedures, and an increased incidence of osteonecrosis of the jaw has been observed following use of bisphosphonates [47,48]. These agents are also dosed systemically, exposing patients to potentially undesired effects across the entire body when the desired action is only for increased osteogenesis in the jaw. In contrast, KMN-159 is dosed in a delivery matrix from which it activates PTGER4 and is then rapidly metabolized (T½ <15 min) and eliminated from the body [21,24]. Local release of KMN-159 at the desired site of action in the tooth socket should substantially minimize the risk of systemic issues. Our studies are not without limitations. The formation and remodeling of bone is a dynamic, multi-cellular process that requires multiple cell types, most notably, monocytes and other immune cells derived from hematopoietic stem cells within the context of a complex oral microbiome. The experiments described here use cells of mesenchymal origin, and the inflammatory processes regulated by monocytes and their progeny are crucial in the formation of healthy bone. Agents that stimulate osteogenesis in osteoblasts or MSCs will indirectly affect osteoblast/osteoclast communication. Our data show that KMN-159 upregulates the RANKL/OPG ratio in MSCs. This modulation should alter the balance between osteoblast and osteoclast activities. This finding is consistent with our previous results showing that KMN-159 in the presence of RANKL and M-CSF increases maturation of osteoclast precursors from bone marrow in culture [21]. Early-stage inflammation during bone formation or healing is a necessary step to bring in chemokines, cytokines, and growth factors which promote processes such as angiogenesis. Since monocytes are the precursors of osteoclasts and MSCs secrete paracrine factors that control monocyte differentiation, it remains to be established in vivo how KMN-159 modulates the crosstalk between monocyte/macrophage lineage cells, osteoclasts, osteoblasts and osteocytes, because these cell types together regulate, respectively, bone resorption and bone deposition [49]. In conclusion, our results demonstrate that the potent small molecule PTGER4 agonist KMN-159 can stimulate osteoblast-related gene expression in both human and rat BM-MSCs in 2D and 3D culture models in vitro and can effectively stimulate in vivo bone formation in two rat models of alveolar bone regeneration. Furthermore, a pilot study in minipigs suggests that KMN-159 stimulates alveolar bone formation around dental implants. These data indicate that KMN-159 holds promise as a local therapeutic agent capable of supporting alveolar bone regeneration with minimal off-target or systemic effects, resulting in the more effective osseointegration of dental implants and the potential for use in many other dental applications. Supplementary Material Acknowledgments Research reported in this publication was supported by the National Institute of Dental & Craniofacial Research of the National Institutes of Health under Award Number R44-DE030370 and by Cayman Chemical Company, Inc. We also thank Dr. Karl Winter for his surgical expertise and carrying out porcine implant surgeries. We thank Dr. Matthew Galliher for assistance with the formulation of KMN-159 used in this study. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.bone.2026.117983. Footnotes Declaration of competing interest Authors Owen, Rzeczycki, Patel, Wei, Barrett, and Morano either are or were employees of Cayman Chemical Company, Inc. All other authors declared that they have no conflicts of interest with respect to the research, authorship, and/or publication of this article. CRediT authorship contribution statement Thomas A. Owen: Writing – review & editing, Writing – original draft, Validation, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization. Phillip Rzeczycki: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Conceptualization. Natalia Da Ponte Leguizamon: Writing – review & editing, Validation, Methodology, Investigation. Qiming Jin: Writing – review & editing, Validation, Supervision, Methodology, Investigation. Chandni Patel: Writing – review & editing, Validation, Methodology, Investigation, Conceptualization. Shanqiao Wei: Writing – review & editing, Validation, Methodology, Investigation, Conceptualization. Stephen D. Barrett: Writing – review & editing, Conceptualization. Andre J. van Wijnen: Writing – review & editing, Writing – original draft, Validation, Methodology. María Inés Morano: Writing – review & editing, Writing – original draft, Validation, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Data availability Data will be made available on request. References Associated Data Supplementary Materials Data Availability Statement Data will be made available on request.