Genetic Deletion of Menin in Mouse Mesenchymal Stem Cells: An Experimental and Computational Analysis

ABSTRACT Loss‐of‐function mutations in the MEN1 tumor‐suppressor gene cause the multiple endocrine neoplasia type 1 syndrome. Menin, the MEN1 gene product, is expressed in many tissues, including bone, where its function remains elusive. We conditionally inactivated menin in mesenchymal stem cells (MSCs) using paired‐related homeobox 1 (Prx1)‐Cre and compared resultant skeletal phenotypes of Prx1‐Cre;Men1 f/f menin‐knockout mice (KO) and wild‐type controls using in vivo and in vitro experimental approaches and mechanics simulation. Dual‐energy X‐ray absorptiometry demonstrated significantly reduced bone mineral density, and 3‐dimensional micro‐CT imaging revealed a decrease in trabecular bone volume, altered trabecular structure, and an increase in trabecular separation in KO mice at 6 and 9 months of age. Numbers of osteoblasts were unaltered, and dynamic histomorphometry demonstrated unaltered bone formation; however, osteoclast number and activity and receptor activator of NF‐κB ligand/osteoprotegerin (RANKL/OPG) mRNA profiles were increased, supporting increased osteoclastogenesis and bone resorption. In vitro, proliferative capabilities of bone marrow stem cells and differentiation of osteoblasts and mineralization were unaltered; however, osteoclast generation was increased. Gross femur geometrical alterations observed included significant reductions in length and in mid‐metaphyseal cross‐sectional area. Atomic force microscopy demonstrated significant decreases in elasticity of both cortical and trabecular bone at the nanoscale, whereas three‐point bending tests demonstrated a 30% reduction in bone stiffness; finite element analysis showed morphological changes of the femur microgeometry and a significantly diminished femur flexural rigidity. The biomechanical results demonstrated the detrimental outcome of the accelerated osteoclastic bone resorption. Our studies have a twofold implication; first, MEN1 deletion from MSCs can negatively regulate bone mass and bone biomechanics, and second, the experimental and computational biomechanical analyses employed in the present study should be applicable for improved phenotypic characterization of murine bone. Furthermore, our findings of critical menin function in bone may underpin the more severe skeletal phenotype found in hyperparathyroidism associated with loss‐of‐function of the MEN1 gene. © 2022 The Authors. JBMR Plus published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research.


Introduction
T he autosomal dominant multiple endocrine neoplasia type 1 (MEN1) syndrome in which tumors arise in select endocrine tissues, including the parathyroid glands, and in nonendocrine tissues is caused by loss-of-function mutations in the tumor suppressor gene, MEN1. (1,2) The MEN1 gene product, menin, is widely expressed from early in fetal development and onward. (3)(4)(5)(6) In various endocrine cell lines, menin has been shown to directly bind to Smad3 and mediate the cell proliferation and differentiation actions of transforming growth factor-β (TGF-β). (7)(8)(9) In bone cell lines, menin interacts physically and functionally with bone morphogenic protein-2 (BMP-2), Smads1/5, and Runx2 to promote mesenchymal stem cell commitment into the osteoblast lineage, and with TGF-β/Smad3 to maintain osteoblast function and differentiation. (10)(11)(12)  The Men1 gene deletion is present in tissues of mesenchymal origin. Genomic DNA from WT and KO mouse tissues was used as a template for multiplex PCR amplification with primers A, C, and D. Product A + D = 638 bp (recombinant Δ3-8 Men1 allele); product A + C = 239 bp (target Men1flox allele). (C) Gene expression levels of menin in WT and KO tissues determined by qPCR and normalized to GAPDH expression levels (n = 7). (D) Representative photograph of WT and KO 9 months of age, closely mimicking the human MEN1 disorder. (13)(14)(15) Homozygous null (Men1 À/À ) mice die in utero at mid-gestation from development failure of multiple organs. The fetuses exhibit craniofacial defects, suggesting that menin may be critical for bone development in vivo. (13,14) Conditional knockout in mice of the MEN1 gene in Pax3-or Wnt1-expressing cells of the neural crest (which are precursors of mesenchymal stem cells) leads to perinatal death with cleft palates and other craniofacial defects. (16) We have previously shown, using an Osteocalcin (OC)-Cre-mediated recombination strategy to conditionally disrupt the MEN1 gene in mature osteoblasts, that menin was required for optimal bone formation and maintenance of bone mass. (17) Deletion of menin in early osteoblasts and osteocytes using the osteoblast transcription factor Runx2 and the DMP1 (dentin matrix acidic phosphoprotein 1) promoters revealed a role for menin in controlling osteoclastogenesis via the osteoblast lineage. (18) In recent years, there has been growing interest in molecular mediators of bone pathways to identify novel therapeutic targets for low bone mass disorders such as osteoporosis. Menin has emerged as one such potential therapeutic target. (17) In mesemchymal stem cells, represented by the bone marrow cell line ST2, in vitro knockdown of menin expression achieved by transfection of menin antisense cDNA resulted in decreased bone morphogenetic protein (BMP)-2-induced alkaline phosphatase activity and osteocalcin and Runx2 mRNA expression. Menin was co-immunoprecipitated with Smad1/5 and inactivation of menin antagonized BMP-2-induced transcriptional activity of Smad1/5 in ST2 cells but not in more mature MC3T3-E1 osteoblastic cells. Menin also co-immunoprecipitated with the key osteoblast regulator, Runx2, and anti-sense menin diminished Runx2 transcriptional activity and the ability of Runx2 to stimulate alkaline phosphatase activity only in ST2 cells. (11) These studies suggested that deletion of menin in mesenchymal cell precursors may produce distinctive actions from deletion in more mature cells of the osteoblast lineage.
In the present study, we therefore conditionally inactivated the MEN1 gene in mouse MSCs in vivo using the Cre-LoxP recombination system, and used high-resolution imaging, and histological and biomechanical techniques to investigate the role of menin in the regulation of bone development in vivo. We demonstrate herein that menin functions critically in mesenchymal stem cells in vivo to regulate bone mass, bone structure, and bone mechanics, including microstructural and macroscopic geometric features as well as overall and local mechanical properties.

Materials and Methods
Generation of conditional knock-out mice A conditional knockout model in which the Men1 gene is deleted in mesenchymal stem cells (MSCs) was generated. Mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA) and housed in a Canadian Council on Animal CARE (CCAC)accredited pathogen-free standard animal facility at the Research Institute of McGill University. Mice were fed normal 18% protein chow diet (Envigo [TEKLAD global], Madison, WI, USA; product #T.2918.15) containing 1.0% calcium, 0.7% phosphorus, and 1.5 IU vitamin D 3 /g. All experimental procedures were performed following an animal use protocol approved by the Facility Animal Care Committee (FACC) of McGill University and met the ARRIVE guidelines. All personnel working with the mice were qualified and met all training requirements set by the FACC.

Gene expression analysis
For gene analysis, quantitative reverse transcrition PCR (qRT-PCR) was performed on RNA from long bones, calvaria, cartilage, white adipose tissue (WAT), muscle, and kidneys. Total RNA was male mice and dissected femurs at 9 months of age. (E) Representative femur X-ray radiographs, and femur lengths of 6-month-old male and female WT and KO mice (n = 7). (F) Body weights of male (n = 8) and female (n = 7) WT and KO mice from 3 weeks to 21 weeks of age. (G) Echo-MRI analysis of 12-month-old WT and KO male mice showing total body weight, fat mass, and lean mass (n = 8). (H) Grip strength of mouse forelimbs was measured, using a grip strength meter, in 6-to 7-month-old male (n = 6) and female (n = 7) mice, by allowing individual mice to grip a metal grid with only their forelimbs. Grip strength values (G-Force) were normalized to body weight (g). Data are presented as mean AE SD; statistical significance indicated on plots: **p < 0.01, ***p <0.001. extracted by the standard TRIzol method (Thermo Fisher Scientific Life Sciences, Waltham, MA, USA).
Briefly, tissues were collected from 6-to 7-month-old male and female mice and placed in RNAlater solution (Thermo Fisher-Ambion, Carlsbad, CA, USA) for subsequent RNA isolation. Epiphyses and soft tissues were removed from bones and the marrow was flushed. Bones were crushed in liquid nitrogen and reconstituted inTRIzol reagent. Other tissues were homogenized in TRIzol reagent using the Brinkmann POLYTRON PT 300 Blade-Type Homogenizer (Kinematica, Luzern, Switzerland), and subsequent steps were performed as per the manufacturer's instructions to isolate total RNA. Two micrograms of RNA was used for cDNA synthesis using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA). Complementary DNA (cDNA) was diluted by a 1:1 ratio, and quantitative real-time PCR (qPCR) performed thereafter on 3 μL of cDNA for each 20 μL reaction. Power SYBR Green PCR Master Mix and the ViiA 7 thermal cycler system machine were used for this step (Applied Biosystems); each sample was run in triplicate. Conditions for the reactions were as follows: 40 cycles: 95 C for 15 seconds, 56 to 58 C for 45 seconds, and 72 C for 60 seconds. GAPDH was used as the housekeeping gene reference, and expression levels were calculated using the ΔΔCt method. To avoid amplification of genomic DNA, the primers chosen to amplify menin RNA crossed 2 introns: exon 2 forward primer 5 0 -TTCCTGGCTGTCAACCGTG-3 0 and exon 4 reverse primer 5 0 -TTCCTGGCTGTCAACCGTG-3 0 . The housekeeping GAPDH primers used for normalization were: forward 5 0 -CACCATCTTCCAGGAGC-GAG-3 0 GAPDH reverse 5 0 -CCTTCTCCATGGTGGTGAAG-3 0 .

Forelimb grip strength test
A forelimb grip strength test was used to assess changes in skeletal muscle strength in 6-to 7-month-old male and female wildtype and knockout mice as previously described. (20) Each mouse was lifted by its tail over the top of a grid so that it could grasp the grid platform with its front paws only. With its torso oriented in a horizontal position, the mouse was then gently pulled backward until it released the grid. The maximal force digitally applied by the mouse before releasing its grasp of the grid was displayed and recorded in grams on a grip strength meter (Bioseb, Chaville, France). The observer was blind to genotype and the mean of 10 consecutive tests per mouse was taken as an index of forelimb grip strength. The body weight of each individual mouse was determined at the end of the experiment to normalize the mean grip strength force to body weight.

Echo-MRI
Body composition of adult male mice was assessed at 12 months of age on live animals using the Echo-MRI system (model ET-040, version 11,06,22) at the McGill Mouse Metabolic Platform. Total body fat percentage was calculated as: [fat mass (g) / total body weight (g)] * 100% and total lean mass percentage as: [lean mass (g) / total bodyweight (g)] * 100%.
Bone mineral density (BMD) analysis by dual-energy X-ray absorptiometry (DXA) DXA analysis of 7-to 8-month-old male and female mouse femurs was performed at the Centre for Bone and Periodontal Research Core Facility at McGill University as previously described. (17) Samples were placed prone on a specimen tray and the Kubtec Xpert 80 radiography system (KUB Technologies Inc., Milford, CT, USA) was used. Energy used was 21 kV/500 μA. Width/center used was 23862/4903 and magnification was 3Â. BMD was calculated with the manufacturer's software.

Micro-computed tomography (micro-CT)
Femurs from 7-to 9-month-old male and female mice were dissected free of soft tissue and frozen in PBS until analysis. Highresolution images of the distal femur were acquired with a desktop microtomographic imaging system (SkyScan 1272, Antwerp, Belgium) in accordance with the recommendations of the American Society for Bone and Mineral Research (ASBMR). (21) Images were captured at 45 kV and 222A, every 0.9 through a 180 rotation, using a detection pixel size of 5.625 μm, or every 0.9 through a 360 rotation, using a detection pixel size of 11.25 μm. Reconstruction of raw images was performed using NRecon software (SkyScan), while trabecular volumes of interest (VOI) were defined, modeled, and analyzed using the CTAn software (SkyScan). Trabecular regions were taken to include all cancellous bone in the metaphysis, defined within a 2 mm region that starts 1 mm below the growth plate. Three-dimensional renderings of the distal 3.5 mm of each femur were created using the CTVox software (SkyScan).

Histomorphometry
Static and dynamic histomorphometry were performed on undecalcified and decalcified femurs from 6-to 7-month-old male and female mice as previously described. (17) Calcein (Sigma-Aldrich, St. Louis, MO, USA; C-0875) was injected intraperitoneally at a dose of 25 μg/g body weight on days 10 and 3 before euthanization, defined as day 0, when mice reached exactly 3 months of age. Mineral apposition rate (MAR) was defined as the average distance between two calcein fronts on a double-label surface, divided by the time span between the two calcein doses (7 days). Mineralizing surface/bone surface (MS/BS) was assessed using the formula (dL + ½(sL) / BS, where dL and sL are the total double-and single-labeled bone surfaces, respectively, and BS is the total bone surface in the analyzed field. Bone formation rate/bone surface (BFR/BS) was calculated as the product of the MAR and MS/BS measurements.

Femur histology
Femurs from 3-and 7-month-old male and female mice were isolated and fixed in 4% paraformaldehyde solution and decalcified using 10% EDTA for 2 weeks. Samples were then dehydrated using a graded ethanol series and were paraffin-embedded. Coronal sections (5 μm) were obtained using the Leica RM 2255 automated microtome (Leica Biosystems, Wetzlar, Germany), and stained for alkaline phosphatase (ALP) or tartrate-resistant acid phosphatase (TRAP), for osteoblast and osteoclast visualization, respectively. High-resolution slide scans were acquired using the Aperio At Turbo Digital Pathology Scanner (Leica Biosystems). For each randomly numbered sample, 3 to 5 sections were analyzed in full for bone cell numbers, at 4Â magnification, using the ImageJ software (NIH). Undecalcified bones were embedded in methyl methacrylate (MMA), and 1-μm sections were cut on an ultramicrotome. These sections were stained with the Goldner trichrome method as previously described. (22) All histological analyses were conducted as per the recommendations of the American Society for Bone and Mineral Research Histomorphometry Nomenclature Committee. (23) Serum markers of bone metabolism Male and female mice (5 to 7 months old) were fasted for 12 hours and serum samples were obtained from blood collected by cardiac puncture immediately subsequent to euthanasia. Serum measurements of osteocalcin, procollagen 1 intact Nterminal propeptide (P1NP), C-telopeptide degradation products from type 1 collagen (CTx), and TRAP were performed using ELISA kits (Biomedical Technologies Inc., Stoughton, MA, USA, and Immunodiagnostic Systems (IDS), Bolton, UK, respectively), according to the manufacturer's intructions.

Three-point bending tests
Biomechanical testing was performed on femurs isolated from 6and 9-month-old male mice using the Instron 5943 machine (Instron, High Wycombe, UK). A downward bending load was applied at a rate of 0.05 mm/s to the posterior midshaft with a supporting distance of 7 mm from each side. A loaddisplacement curve was generated and used to measure biomechanical parameters such as stiffness (N/mm), ultimate force (N), ultimate displacement (mm), and work to ultimate fracture (N. mm). Young's modulus was defined as the mean value of the maximum slope of the elastic portion of the load-displacement curve. The initial slope was disregarded because it is not representative of the femur response; rather it can be attributed to the testing apparatus and the unengaged specimen.

Atomic force microscopy (AFM)
Femurs from 9-month-old male mice were dissected and embedded in methyl methacrylate. Longitudinal and transverse sections of 50 to 100 μm thickness were then cut using an ultramicrotome (Leica RM 2165) using a diamond blade, with minimum surface roughness of <5 nm for AFM imaging and indentation. (24,25) A JPK Atomic Force Microscope (NanoWizard3 BioScience, JPK, Berlin, Germany) was used for imaging and force spectroscopy. Indentation force measurements were made at representative points on the longitudinal and transverse sections. Super-sharp standard force modulation mode reflex coating (FMR) cantilevers (Nanotools USA LLC, Henderson, NV, USA) with diamond-like carbon nano-tip (radius 2 to 3 nm, nominal resonance frequency 75 kHz, and length 225 μm) were used for contact mode imaging. For indentation measurements, non-contact high resonance (NCHR) cantilevers (Nanotools USA LLC) (integrated spherical tips of radius 300 nm [AE10%], nominal resonance frequency 330 kHz, and length 125 μm) were used.
Before indentation tests, the deflection sensitivity of the AFM cantilever was calibrated by engaging the probe on a clean glass slide. To calibrate the spring constant of the cantilever, the first free resonance peak of the cantilever was fitted to the equation of the simple harmonic oscillator so as to measure the power spectral density of the thermal noise fluctuation in air. The longitudinal and transverse elastic moduli (E l and E t ) of the samples were obtained from the force-indentation depth curves through the Hertzian contact mechanics model, where is the relation between the elastic modulus E and the applied indenting load F with ν being the Poisson's ratio of the sample, R the radius of the spherical probe, and δ the indentation depth. Any strain below the elastic limit was assumed infinitesimal, a condition satisfied with the use of an indentation depth below 50 nm that rules out the effect of the substrate as well as any inelastic behavior. (24,26) AFM data analysis was performed with the native JPK data processing software.

Finite element analysis model creation
Micro-CT scan data was segmented and a 3D rendered surfaces of the femurs were created using the ITK-SNAP software (open source software, www.itk-snap.org) (27) ; three-dimensional solid geometries of femurs of the WT and conditional menin KO mice were then reconstructed by using SolidWorks software package (Dassault Systems SolidWorks Corporation, Waltham, MA, USA), and then discretized into volume mesh elements. Elastic moduli obtained from AFM indentation of KO and WT femur sections were assigned to cortical and trabecular bone models. Both KO and WT samples were assumed isotropic. This simplification did not yield noticeable diversions from the results of the orthotropic constitutive model. (28,29) The Hounsfield unit values (HU) were measured from the micro-CT images and were used to assign the elastic properties to cortical and trabecular bones, as well as bone marrow. The mean values of longitudinal elastic moduli, El, of cortical and trabecular bones for KO and WT femurs were used, with a value for Poisson's ratio of 0.3. (29,30) Bone marrow was assumed as an incompressible isotropic material with elastic modulus E = 20 MPa with a Poisson's ratio of ν = 0.499. (29) Primary MSC isolation and culture Six-month-old male and female KO and WT mice were euthanized, and their femurs and tibias were carefully dissected under sterile conditions. Epiphyses of each bone were cut and the marrow cells were harvested as previously described. (17) The cells were cultured in a 5% CO 2 incubator at 37 C in α-minimal essential medium (αMEM, Invitrogen, Carlsbad, CA, USA; 12000-022) supplemented with 10% FBS (Wisent, Saint-Jean-Baptiste, Canada; 080152), with 1% sodium pyruvate (Wisent; 600-110-EL), and 1% penicillinstreptomycin (Wisent; 450-201-EL) and 0.02 m sodium bicarbonate (Sigma; S5761-500G). Medium was replaced every 2 to 3 days.

Differentiation assays of MSC
Bone marrow cells were washed three times with sterile PBS to remove nonadherent cells. The adherent cell fraction was allowed to expand until 80% confluence was reached and then plated in 6-well plates to determine alkaline phosphatase activity and calcium apposition. Cells were cultured in osteogenic medium containing 50 μg/ml L-ascorbic acid (Sigma-Aldrich) and 10 mM β-glycerophosphate (Sigma-Aldrich) for 14 days and then fixed with 100% ethanol.
To assess calcium deposition, cells were stained with 40 mM Alizarin Red-S (Sigma-Aldrich) at pH 4.2 and incubated for 30 minutes at room temperature with gentle shaking. Mineral accumulation was quantified as followed: Alizarin Red-S was extracted by destaining with 10 mM HCl in 70% ethanol and the intensities quantified at 520 nm in 96-well plates using a plate reader (Synergy H4 Hybrid Multi-Mode Microplate Reader, BioTek, Winooski, VT, USA). ALP activity was determined by histochemical staining using NBT/BCIP tablets (Roche Diagnostics, Indianapolis, IN, USA). One ALP tablet was dissolved in 10 mL of distilled water and then added onto fixed cells followed by direct incubation at 37 C for 30 minutes in the dark. Stained wells were washed and air-dried. To quantify ALP staining, 10Â microscopic images were taken using the Evos XL Core light microscope (Life Technologies, Carlsbad, CA, USA) and relative staining was determined by calculating the mean gray area using the ImageJ software (NIH).
Methyl thiazolyl tetrazolium (MTT) dye viability assay Primary MSCs were seeded at a density of 2-3 Â 10 4 cells per well in 24-well tissue cultures plate and cultured for 48 hours. MTT reagent (Roche Diagnostics) was then added, and incubation was continued for 4 hours. Formed crystals were eluted in 500 μL of DMSO and absorbance was read at 570 nm using a microplate reader (Synergy H4 Hybrid Multi-Mode Microplate Reader, BioTek).
Osteoblast/osteocyte-osteoclast co-cultures Osteoblasts/osteocytes from long bones of 6-to 7-month-old KO and WT mice were isolated. Epiphyses were cut, marrow was flushed, and bones were transferred onto a Petri dish containing sterile PBS. Soft tissues were removed with a scalpel and the bones were chopped into small fragments. The bone fragments were incubated for 30 minutes at 37 C with shaking in 2 mg of collagenase II/ml αMEM (Bioshop, Burlington, Canada). The supernatant was discarded, and the procedure was repeated. Bones were then incubated for 30 minutes at 37 C with shaking in a 0.25% trypsin, 0.1% EDTA solution. The supernatant was discarded and replaced by fresh collagenase II solution for the fourth and final incubation step of 30 minutes at 37 C. Three subsequent washes with αMEM containing 10% FBS were done to remove all traces of Trypsin. The bone fragments were then transferred to T75 flasks and incubated in αMEM containing 10% FBS and 50 μg/mL l-ascorbic acid without moving for 3 to 4 days. Adult mouse bone cells migrated from the bone chips after 48 hours, and after 5 to 6 days, the cells were counted and seeded (1 Â 10 3 ) in transwell insert devices for 24-well plates (Thermo Fisher Scientific) with αMEM containing 10% FBS and ascorbic acid. Osteoblasts/osteocytes were allowed to reach 80% confluency (~5 days).
WT and KO osteoclast precursor cells were isolated from bone marrow according to Boraschi-Diaz and colleagues. (31) Briefly, long bones were cut in half, inserted into cut 1 mL pipette tips placed inside Eppendorf tubes, and centrifuged at 12,000g, three times for 30 seconds. The marrow was resuspended into αMEM by repeated pipetting up and down, transferred to a conical tube containing 10 mL of culture medium, and centrifuged at 1000g for 5 minutes. Red blood cells from the pellet were lysed by resuspending in red blood cell (RBC) lysis buffer (Sigma-Aldrich). After incubation and washing, the final bone marrow cells pellet was resuspended in culture medium, plated in a 75 cm 2 flask, and incubated overnight. The next day, nonadherent cells were pelleted by centrifuging the collected cell medium at 1000g for 5 minutes. The pellet was resuspended in 10 mL of culture medium and viable precursors were counted using Trypan blue (Sigma-Aldrich). Precursors were then seeded (6 Â 10 6 per well) in 24-well plates with αMEM (10% FBS) and allowed to settle for 12 to 16 hours. The transwell devices with grown osteoblasts/ osteocytes were then inserted in each well containing osteoclast precursors and cultured in αMEM containing different concentrations (0, 5, 10, or 25 ng/mL) of RANKL neutralizing antibody (PA5-47641, Thermo Fisher Scientific). Cells were further cultured for another 7 days, at which point, osteoclasts were observed and fixed with 10% paraformaldehyde in PBS, pH 7.4, for 10 minutes at room temperature and stained for tartrate-resistant acid phosphatase (TRAP, Sigma; 387A).
Osteoclast numbers were quantified by counting TRAP-positive cells with three or more nuclei and intact membranes. (32) Statistical analysis Results are expressed as mean AE standard deviation (SD). Statistical comparisons were made using the PRISM analysis software version 9.0 (GraphPad Software Inc., San Diego, CA, USA). One-way analysis of variance was used to discriminate any differences among means of groups of data. Unpaired two-tailed Student's t tests were used to assess significance; a p value <0.05 was taken to indicate a statistically significant difference.

Results
Disruption of the Men1 gene with PRX1-Cre is specific to tissues of mesenchymal origin PCR amplification of tissue genomic DNA confirmed that Prx1mediated Cre recombination occurred in tissues of mesenchymal origin such as bone (femur, tibia, calvaria), white adipose tissue (WAT), cartilage, and muscle but not in kidney of the KO Prx1-Cre; Men1 f/f , mice (Fig. 1B). No recombination was observed in Men1 f/f (WT) mice.

Deletion of menin in mesenchymal stem cells results in altered bone morphology and reduced muscle strength
Gene expression levels of menin were measured by qRT-PCR and were found to be reduced by about 60% in femurs and tibias of KO mice; in contrast, MEN1 expression was not altered in kidney (Fig. 1C). Menin mRNA levels were also reduced by~65% in cartilage, muscle, and white adipose tissue (Fig. 1C). Conditional inactivation of MEN1 in mouse MSCs led to a decrease in overall size of KO mice relative to control WT mice (Fig. 1D). By radiography, femur epicondyles were shown to be morphologically altered, particularly at the level of the greater and lesser trochanters (Fig. 1E), and quantitative assessment of femur length at 6 months, taken as the diagonal distance between the femoral head and the lateral condyle, demonstrated a statistically significant reduction in femur length in the KO mice relative to WT (Fig. 1E). The weight of KO mice was significantly reduced postnatally through 21 weeks of age in both males and females (Fig. 1F), and Echo-MRI analysis of KO males at 12 months of age demonstrated a significant 11% reduction in lean mass weight relative to WT (Fig. 1G). To determine whether changes observed in body metabolism and in lean mass content could affect skeletal muscle strength, a forelimb grip strength test was performed (Fig 1H). Individual values obtained from the grip strength test were normalized to total body weight of each mouse. Six-month-old KO male and female mice demonstrated a significant decrease of~30% in forelimb grip strength in comparison to control WT animals. The lean mass and grip strength phenotypes observed were not significantly different between sexes.
Deletion of menin in mesenchymal stem cells results in reduced bone mineral density but intact cartilaginous growth plates BMD, determined by DXA, was shown to be significantly decreased in KO mice by approximately 11% in males and approximately 13% in females ( Fig. 2A). Qualitatively, reconstructed distal femur models generated from micro-CT scans showed reductions in trabecular bone in femurs of KO mice (Fig. 2B). Nevertheless, cartilaginous growth plates of WT and KO mice were not different and both showed similar intact growth plates (Fig. 2C), indicating that chondrocyte alteration did not contribute substantially to observed changes in postnatal bone mass acquisition. The phenotypes observed were similar in both sexes.

Quantitative micro-CT demonstrates altered trabecular bone in knockout mice
Marked reductions were observed In both male and female KO mice, relative to WT mice, in trabecular bone volume (BV/TV) (75% in both males and females, Fig 3A), trabecular bone surface (BS/TV) (75% in males and 52% in females, Fig 3B), and trabecular number (Tb.N) (77% in males and 72% in females, Fig. 3C); in contrast, increases were observed in both male and female KO mice in trabecular thickness (Tb.Th) (10% in both males and females, Fig. 3D) and trabecular separation (Tb.Sp) (65% increase in males and 72% increase in females, Fig. 3E) relative to WT mice. The increase in structure model index (SMI; 20% increase in both males and females, Fig. 3F) suggested abnormal trabecular structure with potentially compromised mechanical integrity in KO mice. Cortical thickness (Ct.Th) (Fig. 3G) was not altered. These findings suggest that menin functions critically at early stages of the osteoblast lineage to regulate trabecular bone development and maintenance in vivo.

Menin inactivation in mesenchymal stem cells results in unaltered osteoblast numbers and bone formation but increased osteoclastogenesis in vivo and in vitro
Serum bone marker determinations showed normal aminoterminal propeptide of type 1 procollagen (P1NP) and osteocalcin levels (Fig. 5A, B) but a~50% increase in both tartrate-resistant acid phosphatase (TRAP) and C-telopeptide of type I collagen (CTX-1) levels (Fig. 5C, D) in KO mice, consistent with increased osteoclastic activity. Decalcified and paraffin-embedded femur sections from 6-to 7-month-old WT and KO mice were stained with ALP (Fig. 5E) and TRAP (Fig. 5G) for the assessment of osteoblast numbers and osteoclast numbers, respectively. Osteoblast numbers were unchanged in KO mice relative to WT littermates ( Fig. 5F); however, significant increases in the number of osteoclasts per bone perimeter were observed (100% increase, Fig. 5H).
Dynamic histomorphometric parameters were assessed by double calcein labeling in 6-to 7-month-old female mice (Fig. 6A). Trabecular (Fig. 6B) and cortical (Fig. 6C) mineral apposition rate (MAR) and trabecular bone formation rate/bone surface (BFR/BS) (Fig. 6D) were unaltered in KO mice relative to WT. Results in male mice were similar (data not shown). These findings indicate that the significant alterations in bone microscopic and macroscopic parameters observed in KO mice appear to result from an increase in bone resorption, rather than a decrease in formation.
To further assess potential differences in osteoblast function responsible for the osteopenic phenotype observed in KO mice, MSCs were isolated and studied in vitro. MTT assay revealed no significant differences in proliferative capabilities of MSC cultures from KO mice relative to WT at 48 hours after first passage (Fig. 7A). When MSC cultures were treated for 14 days with osteoblast differentiation media, no difference was observed in differentiation and mineralization capabilities as illustrated by alkaline phosphatase (ALP) (Fig. 7B) and Alizarin Red (AR) staining (Fig. 7C), respectively. Taken together, these findings demonstrate that differentiation and mineralization of mature osteoblasts are unaltered in vitro after the deletion of menin in the mesenchymal stem cell. In contrast, the capacity to generate osteoclasts in bone marrow cultures in vitro was significantly increased in KO mice and a blocking antibody to receptor activator of NF-κB ligand (RANKL) reduced the increased osteoclast numbers observed (Fig. 7D).
To further evaluate the molecular basis of the increase in osteoclast numbers, RANKL and osteoprotegerin (OPG) mRNA levels were examined in long bones of WT and KO mice (Fig. 7E, F). Quantitative real-time PCR conducted on total mRNA isolated from the femurs of 6-to 7-month-old animals demonstrated significant increases of RANKL mRNA levels in KO relative to controls (approximately 100%, Fig. 7E). OPG mRNA levels were not significantly elevated (Fig. 7E) and RANKL/OPG levels were increased (approximately 100% increase, Fig. 7F). Deletion of menin in mesenchymal stem cells impairs femur trabecular and cortical microstructure, mechanics, and nano-dynamic elastic properties Atomic force microscope (AFM) indentation was conducted on cortical and trabecular bone samples using both transverse and longitudinal sections, as previously described. (24) Qualitative microscopic geometric differences in cortical bone structure were observed in femur sections obtained from KO mice, most notably, a more porous architecture that is less monolithic than that observed in WT controls (Fig. 8A). The anisotropy ratios of femurs (ie, the ratios of the average values of the longitudinal -month-old WT and KO male mice stained with alkaline phosphatase for assessment of osteoblasts numbers. Scale bar = 300 μm quantitation of osteoblasts (F), including osteoblast surface relative to bone surface and osteoblast numbers normalized to bone perimeter (N.OB/b.pm). (G) Representative paraffin sections of femurs from 6-month-old WT and KO male mice stained with TRAP for assessment of osteoclast numbers. Scale bar = 300 μm. Each value is the mean AE SD of 10 and 11 determinations in WT and KO mice, respectively. Quantitation of osteoclasts (H), including osteoclast surface relative to bone surface and osteoclast numbers normalized to bone perimeter (N.OC/b.pm). Each value is the mean AE SD of 10 and 11 determinations in WT and KO mice, respectively. Statistical significance indicated on plots: *p < 0.05, ***p < 0.001. elastic modulus El, over the transverse elastic modulus Et) for KO samples ranged from 1.3 to 1.5, in contrast to values of 1.2 to 2 for WT mice. The mean values of cortical bone elastic moduli in the transverse (Et) (Fig. 8B) and longitudinal (El) (Fig. 8C) directions were found to be 51% and 43% larger in WT femur samples, respectively, relative to their corresponding KO littermates. Similar significant differences were observed for trabecular bone; mean Et and El values in WT trabecular samples were found to be 25% and 33% greater than what was observed in KO, respectively (Fig. 8D, E). Statistically significant differences were found between the mean values of Et and El in WT cortical bone.
To compare the dynamic elastic behavior of WT and KO samples, additional indentation tests were performed at given frequencies from 1 to 100 Hz. (24,26) For each sample examined, within the cortical (Fig. 8F) and trabecular (Fig. 8G) bone compartments indented, a logarithmic fit of the correlation between the longitudinal elastic modulus El and the indentation frequency was established, as presented in Fig. 6F, G. Depending on the frequency, the mean value of El in WT cortical bone was found to be 15% to 38% greater than in KO counterparts (Fig. 8F); the mean El value in WT trabecular samples was 23% to 35% greater in WT than in KO mice (Fig. 8G).
Taken together, these findings indicate that deletion of menin early in the osteoblast lineage impairs femur trabecular and cortical microarchitecture and compromises its mechanical and dynamic elastic properties at the nanoscale.
Computational and experimental analyses demonstrate significantly reduced mechanical alterations in menindeficient femurs Fig. 9A shows the geometric portion of flexural rigidities (ie, the capacity of the beams to support lateral forces) for femurs of WT and KO mice. Bending coefficients Imax/C1 and Imin/C2 (with Imax and Imin being the principal second moments of inertia of the femur lateral cross section, and C1 and C2 the farthest distances of material elements in each cross section from the two principal bending axes of that cross section) were obtained using the micro-CT data of three pairs of KO and WT mice at 9 months of age. As demonstrated in Fig. 9A, geometric bending stiffness was significantly reduced along the femoral shaft by up to 50% at the level of the diaphysis in femurs of KO mice. This geometric feature, together with the longitudinal and transverse elastic moduli, E l and E t (obtained through AFM indentation testing)    describes the overall bending stiffness of the femoral bone as a comparative tool to highlight the differences in bending capacities of WT and KO samples.
Three-point bending tests were then conducted to assess differences in macroscopic elastic properties at the level of the middiaphysis. The work of fracture (Fig. 9B), load at the yield point (Fig. 9C), and load at ultimate failure (Fig. 9D) were all significantly reduced in KO femurs relative to WT. Thus, the work of fracture in WT samples was 40% larger than the corresponding value in KO samples, the load at yield in WT samples was 60% larger than that in KO ones, and the ultimate load in WT samples was 60% larger than that in KO ones. The average value of the maximum load supported by WT samples was 50% greater than that measured in KO mice, at 27 N and 18 N, respectively (Fig. 9E). Young's modulus, taken as the mean value of the maximum slope of the force-displacement curve, was measured as 180 N/m in WT femurs, a value 30% greater than that in KO mice (140 N/m) (Fig. 9F). The average value of the maximum strain up to failure supported by WT samples was 0.57, which was found to be 29.5% greater than that measured in KO mice with maximum strain of 0.44 (Fig. 9G, H).
Elastic 3-point bending tests were then replicated using a computational approach as per the American Society for Testing and Materials (ASTM) standards for mechanical testing; femurs were modeled as simply supported beams with a 5 N concentrated load applied at the midspan. Femurs were then arranged such that condyles were positioned inferiorly to replicate the 3PBT experimental design. Simulation outcomes were cast into performance metrics to capture changes in the mechanical behavior in response to the applied loads for given boundary conditions. In particular, distributions of von Mises stress and von Mises strains were obtained in three pairs of WT and KO samples, with a representative pair presented in Fig. 10, for the anterior, posterior, transverse, and sagittal views. Quantitatively, KO femurs showed a 44% higher von Mises stresses than control samples, consistent with lower flexural stiffness in response to bending. In addition, KO femurs exhibited 75% larger von Mises strains relative to controls.
Taken together, these findings demonstrate that for the prescribed boundary conditions simulated and experimentally evaluated, deletion of menin in mesenchymal stem cells significantly reduced global bending stiffness and compromised macroscopic femur biomechanical strength.

Discussion
Expression of the homeobox gene Prx1 is restricted to the mesoderm during embryonic development (33) and can regulate differentiation of mesenchymal precursors. Using the Prx1 cre driver, conditional deletion of Men1 in mesenchymal stem cells demonstrated significant reductions in overall weight throughout life, as well as a reduction in bone mineral density and femur length compared with WT littermates. Furthermore, lean weight and grip strength were reduced. This may reflect Prx1-Cre expression in muscle as well as bone in keeping with the mesenchymal origin of both tissues (Fig. 1C) and the consequent deletion of menin in both tissues. Nevertheless, mice with conditional Men1 knockout in mesenchymal stem cells showed intact epiphyseal growth plates (Fig. 2C). Micro-CT imaging showed a decrease in trabecular bone volume, altered trabecular structure, and increase in trabecular separation but unaltered cortical bone thickness. Similar, although less severe, changes were observed in heterozygous mice with menin deletion. Dynamic histomorphometry demonstrated unaltered in vivo bone formation metrics and histologically, the number of osteoblasts was unchanged. In vitro proliferation of bone marrow stem cells isolated from KO mice, and differentiation and mineralizing capabilities of osteoblasts were unaltered compared with WT. Nevertheless, there was a marked increase in osteoclast number, consistent with our in vivo findings of increased circulating TRAP and CTX-1 levels and increased RANKL/OPG mRNA profiles in bone. These findings supported increased osteoclastogenesis and bone resorption signaled by osteoblasts in reducing bone mass in menin-deficient mice.
Increasingly, it has been recognized that measurement of BMD alone may not accurately demonstrate bone fragility and predisposition to fracture, and additional approaches to complement BMD measurements in osteoporosis have been reported. (34) We, therefore, provided a detailed analysis of the biomechanical properties of bone caused by deletion of menin early in the osteoblast lineage. Thus, the mechanical behavior of bone is strongly dependent on its complex organization and hierarchical structure. (35,36) At the micro scale, the local response of bone to mechanical deformation is mainly a function of its material properties, namely, the spatial organization of hydroxyapatite crystals and collagen. (37) At the macroscopic scale, its microstructural material properties and macro-geometry regulate its mechanical behavior. (38) Given the significant skeletal phenotype observed, we examined femur biomechanics and found that elastic properties were decreased at the femoral mid-diaphysis with significant reductions in stiffness, maximum load, the work of fracture, the ultimate load, and the load at yield point.
Macroscopic testing such as three-point bending tests cannot elucidate the microscopic differences in bone resulting from a gene knockout, (25,27,35) nor does it provide information on the internal distribution of stresses and strains within the bone tissue in response to given loading conditions. (27,35) Differences in elastic properties along both the longitudinal and transverse axes were measured in both KO and WT femur samples. This confirmed a high level of local anisotropy in both trabecular and cortical bone tissues. (24) Compared with the elastic moduli of cortical and trabecular bones in the transverse plane, a 20% to 50% increase was observed in the longitudinal direction in both KO and WT strains. The transverse section of femoral bone showed higher longitudinal elastic moduli compared with the transverse elastic moduli in the longitudinal section in both WT and KO samples. Furthermore, elastic moduli were up to 70% greater for cortical bone than for trabecular bone, with the lowest elastic modulus being that of trabecular bone in the transverse direction in both KO and WT bones. Also, in both KO and WT samples, El was found to be greater than Et at the same anatomical location within a given femur. Elastic moduli were greater at higher indentation frequencies in both KO and WT mice and in both cortical and trabecular bone compartments. For the same indentation depth, δ (ie, smaller than 50 nm), the indenting force, F, increased with the loading rate, thus demonstrating the dependence of bone stiffness on loading rate. This variation can be attributed to the proteinaceous nature of collagen fibrils at the molecular scale, (26) and the direct correlation between elastic modulus and frequency is a result of the viscoelastic nature of bone. (39) However, Et and El values of cortical bone were found to be significantly smaller in KO femur samples compared with their corresponding values in WT samples. Concomitantly, Et and Et values in KO trabeculae were observed to be significantly smaller than their counterparts in WT trabeculae. The sensitivity of this technique is thus paramount for an adequate assessment of the biomechanical phenotype of bone murine models such as the current one.
AFM experiments confirmed significant reductions in trabecular and cortical bone stiffness along both longitudinal and transverse planes in femurs of KO mice relative to controls. Computational analysis and simulation of the three-point bending tests, using elastic properties obtained via AFM indentation, corroborated the low flexural stiffness observed experimentally. AFM proved valuable not only in demonstrating differences in nanoscale mechanical properties of KO femurs but also in documenting and establishing differences in bone mechanics within the same femur sample at different locations and orientations. (24) The experimental, computational, and simulation biomechanical analyses adopted in this study thus unveil properties and performance metrics that would otherwise be dismissed with classical testing at the component level. This multiscale approach can therefore be used and applied in future studies as part of the phenotypic characterization of other murine models of bone development and bone physiology, providing useful insight pertaining to the integrity of bone and its behavior and extending the clinical utility of what BMD, X-ray, and micro-CT scan data can presently offer.
The tumor suppressor menin can impact several downstream signaling sytems, eg, menin can directly interact with Smad3 protein, and inactivation of menin blocks TGF-β and activin signaling in anterior pituitary cells, which can antagonize their growth-inhibiting properties. (7,8) Mediated by the Smads and the transcription factor, Pit-1, menin can also cause activininduced inhibition of prolactin expression. (8) In parathyroid cells, menin reduction can lead to loss of TGF-β inhibition of parathyroid cell proliferation and parathyroid hormone (PTH) secretion. (12) Menin was shown to regulate canonical Wnt/β-catenin, and overexpression of menin reduced nuclear accumulation of β-catenin and suppressed its transcriptional activity. Menin may also control endocrine cell proliferation, apoptosis, and DNA repair, at least in part by regulating histone modifications and chromatin structure, leading to altered gene transcription. (40) Menin also epigenetically represses Hedgehog signaling, a proproliferative and oncogenic pathway. (41) Although menin can suppress tumorigenesis in the endocrine lineage, it can also promote oncogenic function in other tissues, such as breast. (42) In the hematopoietic lineage, menin serves as a molecular adaptor that physically links the mixed-lineage leukemia (MLL) histone methyltransferase with lens epitheliumderived growth factor (LEDGF), a chromatin-associated protein implicated in leukemia. (43) Binding of menin to MLL1 leads to upregulation of Hox gene transcription and leukemia in MLLrearranged acute myeloid leukemia (AML) and acute lymphocytic leukemia (ALL). (44)