Silva Lab

Matthew Silva, Ph.D

Julia and Walter R. Peterson Orthopaedic Research Professor

Department of Orthopaedics Washington University
BJC - Institute of Health
11th floor - RM 11619
Phone: (314) 747-3772

Click here to visit the Musculoskeletal Structure and Strength website.


Response of the Osteoporotic Skeleton to In Vivo Loading: We are interested in determining if the skeleton has diminished responsiveness to mechanical loading with age. We are subjecting young-adult and aged mice to two types of loading: 1) high-amplitude, low-frequency tibial compression, and 2) low-amplitude, high-frequency whole-body vibration (WBV). These studies will clarify whether or not there is reduced mechanoresponsiveness in the aged/osteoporotic skeleton.

Whole-body vibration setup

Figure. Whole-body vibration setup. Mice are placed in the plastic housing and the plate is accelerated in the vertical direction at low-amplitude (0.1 – 1.0 g), high frequency (30-90 Hz) for 15 min/day.

MicroCT images of the tibia

Figure. MicroCT images of the tibia from 7 and 22 month old BALB/C mice illustrating differences in bone structure. At 22-months of age (~75 years in human) there is less trabecular bone in the proximal tibia compared to 7-months (~25 years in human). In addition, the medullary cavity is expanded and the cortex is thinner (arrows) in the diaphysis of the 22-mo mice compared to 7-mo. These age-related changes mimic what is seen in humans.

Sections of the tibial diaphysis from control

Figure. Sections of the tibial diaphysis from control (non-loaded) and loaded legs of 7- and 22-mo old mice after 5 days of tibial compression loading. Sites of bone formation are labeled with calcein green and alizarin red. There is little bone formation in control tibias. In contrast, the loaded tibias of both 7- and 22-mo mice have abundant new bone formation both on the endocortex (Ec) and the periosteum (Ps). We found no deficit in the ability of aged mice to respond to high-amplitude tibial compression. (B-bone; Ma-marrow)

Osteogenic and Angiogenic Responses to Loading: We are characterizing the skeletal response to damaging loading, with a focus on woven bone formation and associated changes in vascularity and gene expression. We refined the rat forelimb compression model to produce several levels of sub-fracture bone damage using either fatigue or creep displacement. We discovered that the osteogenic response to fatigue loading is proportional to the level of damage, with increasing amounts of woven bone formed with increasing damage (dose-response). In the first week after loading, woven bone area increases, leading to a partial recovery of bone strength; in the second week, the woven bone area is unchanged but it becomes denser and more mineralized, leading to full recovery of strength. We then determined that static creep loading that generates damage also induces a woven bone dose-response, indicating that damage is the predominant stimulus for the woven bone response seen in skeletal fatigue. In terms of vascular changes, both vessel area and number increase prior to increases in bone area, and their spatial distribution matched the subsequent pattern of bone formation. Angiogenic genes (e.g., VEGF) are upregulated 1 hr after fatigue loading, and BMP-2 is localized to vascular cells at this early time. Other known osteogenic genes (e.g., BSP, Osx) are then upregulated starting on day 1. Thus, vascular responses occurred in the immediate stages after fatigue loading, followed by osteogenesis.

MicroCT scans of ulnae loaded cyclically

Figure: MicroCT scans of ulnae loaded cyclically to different levels of sub-fracture displacement (expressed as % of fracture displacement) illustrate progressive increases in crack severity at the midshaft that corresponded to progressive decreases in ulnar strength.

Combined PET/CT image from cover of Bone

Figure: Combined PET/CT image from cover of Bone (August 2006) illustrating increased uptake of fluoride after creation of a stress fracture.

Ulna 7 days after fatigue loading

Figure. Ulna 7 days after fatigue loading, illustrating stress fracture (arrowheads) which forms where the strain magnitude is greatest, on the medial aspect of the ulna near its midpoint. The pattern of woven bone corresponds to the location of the stress fracture, with the greatest amount of new bone near the midpoint on the medial side.


Figure. Photomicrographs illustrating dramatic increase in periosteal vasculature (perfused vessels are black) 3 days after fatigue loading. (B-bone; P-periosteum; M-muscle)

Photomicrographs upregulation of BMP-2, Osx, PCNA and BSP expression

Figure. Photomicrographs upregulation of BMP-2, Osx, PCNA and BSP expression in the periosteum of loaded ulnae after fatigue loading. There is early activation of surface cells in the first day, and by day 3 the nascent woven bone is forming. (B-original cortical bone; P-periosteum; WB-nascent woven bone)

Simulating bone formation using constrained tibial vibration: In order to advance the understanding of how vibrational loading stimulates bone formation and to complement translational studies using whole-body vibration, We have developed a new method of applying vibrational loading to the mouse skeleton – constrained tibial vibration (CTV). Our overall goal is to characterize the osteogenic response of the murine tibia to a range of vibrational conditions. For example, we are determining the bone formation response of the tibia to vibrational loading under a range of conditions designed to produce different levels of tibial strain. We are examining whether or not the loading response is due to vibration per se or due to bone strain induced by vibrational loading. In addition, we will assess molecular responses by examining solute transport and gene expression induced by constrained tibial vibration.

Illustration of CTV loading set-up

Figure- Illustration of CTV loading set-up.

Bone biomechanics, obesity and leptin in mice: In collaboration with Professor Jim Cheverud of the Dept. of Anatomy and Neurobiology ( we are examining dietary and genetic factors affecting bone mass, morphology, and biomechanical properties and the relationships of these features to obesity and leptin in mice. We are utilizing an established mouse model for obesity, diabetes, and dietary response, the cross of LG/J and SM/J mouse strains. We are measuring the level of heritability for bone traits (e.g., cortical area, trabecular bone volume, femoral strength) and their genetic correlations with obesity and leptin levels in the LGXSM Recombinant Inbred (RI) strains (N=512) and Advanced Intercross (AI) Line (N =1000). Animals have been fed either a high or low fat diet allowing us to examine the effects of both genes and environment (dietary fat) on bone characteristics and bone-obesity relationships. We will identify genomic regions (Quantitative Trait Loci, QTLs) affecting bone and its relationship to obesity. Our goal is to identify novel genes and physiological pathways affecting osteoporosis and examine how genetic and environmentally-based obesity affects bone mass, morphology and biomechanical function.

In vivo cartilage changes after mechanical injury: In collaboration with Drs. Joseph Borrelli and Linda Sandell, we have developed a model of cartilage impact to study the effects of cartilage injury on the development of post-traumatic arthritis. A single impact is delivered to the medial femoral condyle of the rabbit knee. Cartilages changes are examined histologically and with creep-indentation testing, while bone changes are examined histologically and by microCT.

cart damage

Figure- Six months after impact there is a dramatic loss of proteoglycan content (Safranin-O, top) and BMP-2 synthesis (bottom) in articular cartilage.

Creep curves from sham and impacted cartilage samples

Figure- Creep curves from sham and impacted cartilage samples 1 month after impact. Impact caused structural damage as illustrated by greatly increased creep deformation.

Connexins in Bone Formation: In collaboration with Drs. Roberto Civitelli ( and Sue Grimston we are examining the role of connexins 43 and 45 in skeletal development and responses to mechanical loading. Connexin proteins form gap junctions which facilitate cell-cell communication. We are utilizing murine models of conditional connexin deletion in osteoblast lineage cells to examine how disruption in connexins influences skeletal mechanoresponsiveness.

Attenuated mechanical stimulation of new bone formation in mice

Figure- Attenuated mechanical stimulation of new bone formation in mice with conditional ablation of Gja1 (Cx43). Double calcein labels are present in the endocortical surface of loaded bones from wild type (WT) mice at the apex and on the opposite endocortical surface, corresponding to the areas of maximum endocortical compression and tension, respectively. Double-labeled surfaces are also seen in heterozygous (HET) tibiae, whereas predominantly single labels are present in knockout (KO) tibial sections. (Image courtesy of Sue Grimston.)

FGFs in skeletal development, vasculogenesis and repair: In collaboration with Professor David Ornitz ( we are examining the role of FGFs in post-natal bone formation induced by mechanical loading. The importance of FGF signaling in skeletal biology is illustrated by the large number of missense mutations in the genes encoding FGF receptors (FGFRs) 1, 2 and 3 that are the etiology of many human craniosynostosis and chondrodysplasia syndromes. Furthermore, loss of function and skeletal-specific conditional loss of function mutations in mouse FGFRs 1,2 and 3 also show specific defects in skeletal development and in the structure and integrity of adult bone. In contrast to our increasing understanding of the function of FGFRs in skelatogenesis, there is little information on the FGF ligands that regulate skeletal development, growth, remodeling, vasculogenesis and repair. We will test the hypothesis that in response to mechanical load, FGF signaling is required for cortical bone formation and associated increased periosteal vascularization.