PCMD » Scientific Symposia » June 2004 Symposium Retreat

June 2004 Symposium Retreat

Type III Collagen-Deficient Mice Exhibit Altered Skeletal and Craniofacial Development
S. L. Adams, S. Shah and A. J. Cohen

Type III collagen (Col3) is found mainly in extensible tissues such as blood vessels, skin and lung, in heterotypic fibrils with type I collagen. A very small amount of Col3 is also present in type I collagen-containing fibrils in bone. There have been no studies on the involvement of Col3 in skeletal and craniofacial development. Our objective was to investigate the role of Col3 in skeletal development using mice in which one copy of the Col3 gene was inactivated. No abnormal phenotype has been reported for these heterozygous mice; however, we recently observed that they have a hunchback phenotype, suggesting a potential role for Col3 in skeletal development. Fourteen wild-type and ten Col3 heterozygous mice, ranging in age from eight days to eleven months, were analyzed using X-rays, microCT, and histochemical staining of skeletons with alcian blue and alizarin red, which designate areas of cartilage and bone, respectively. The heterozygous mice displayed several unexpected skeletal abnormalities. In heterozygous mice, the posterior frontal suture remained open until at least 5.5 months, while in wild-type mice, this suture typically closes around day 35. Furthermore, the foramen magnum and the entire spinal column were more ovoid in shape in heterozygotes, both 8-day-old and 8-month-old animals. In addition, the morphology of the spinous processes was altered in heterozygous mice, a change that was visible as early as day 8. In older heterozygotes, the spine also exhibited a pronounced dorsal curvature or hunchback. These skeletal phenotypes, which were consistent among all heterozygous mice examined, indicate the unexpected finding that type III collagen plays a role in skeletal and craniofacial development, even though it is present in very small amounts in skeletal tissues.


Murine Transcript Profiling at the Penn Microarray Facility
Jianhua Wang, Kristine Korzow, Fanyi Zeng, Marie Scearce, Grace Straszewski and Don Baldwin

The Penn Microarray Facility provides instrumentation and expertise for RNA transcript profiling (see www.med.upenn.edu/microarr). DNA microarrays provide a highly parallel means of measuring the abundance of RNA for targeted genes in a biological sample. The Cancer Center is a co-founder of the Facility, which has been operational since September 2001. A wide clientele within and outside the University of Pennsylvania utilizes services that include sample preparation, Affymetrix GeneChip processing, and printing and hybridization of custom microarrays derived from cDNAs, oligonucleotides and BAC clones. Data analysis software associated with the Facility's instrumentation is available for client use. Tools for additional statistical analyses, database management, pattern recognition, and inter-project collaboration are available via the Facility's close ties to the Biomedical Informatics Facility (www.genomics.upenn.edu). This poster provides examples of several microarray assays using mouse systems.


Mechanical Differences between Mouse Lumbar and Tail Motion Segments are Largest in the Transition Zone
Sarver, JJ; Allen, SE; Bowles, RD; Elliott, DM

Rodents are widely used models for studying disc degeneration. We recently conducted a validation study which confirmed that, when normalized for geometry, mouse and human lumbar mechanics were similar [1]. We also found differences in compression stiffness between lumbar and tail segments. Given the wide use of the rodent tail as a model for studying mechanical pathways of degeneration [2-6], expanded analyses of the lumbar and tail mechanical differences are needed. The objective of this study was to examine the mechanical differences between lumbar and tail motion segments of the mouse under axial compression and tension loading. We hypothesized that the transition zone would be larger (i.e. more axial displacement) for tail segments due to the greater physiological motion experienced by the tail as compared to the lumbar spine. We also hypothesized that tail motion segments would be less stiff in compression due to the relatively higher loading of the lumbar spine in compression during ambulation [7].

A total of 8 two-month old C57BL/6 mice were utilized according to an approved IACUC protocol. Lumbar and tail spines were dissected, facet joints removed, and microradiographed to calculate disc height (anterior and posterior) and width (lateral and anterior-posterior), which were used to normalize mechanical data. Motion segments were prepared by making axial cuts through the L1/L2 (lumbar) and C7/C8 (tail) vertebrae.
The vertebrae were gripped by customized micro-vises and placed in a PBS bath in an uniaxial mechanical testing system (Instron 5542). Each motion segment was loaded in axial compression and tension according to the following protocol: the segment was cycled 20 times between 0.5 N of compression and 0.5 N of tension at 0.1 Hz (tension-compression cycle), compressed to 0.5N over 100 seconds (slow ramp), returned to zero load, and finally compressed to 0.25N over 1.0 second, and held for 30 minutes (creep).
The nonlinear mechanical data from the loading portion of the 20th cycle were analyzed using a tri-linear model to calculate the stiffness in three loading zones: compression, transition, and tension, as well as the displacement within the transition zone, which is analogous to the 'neutral zone' (Fig 1). Stiffness (S) was normalized using the equation: S*h/A, where h is disc height and A is area, assumed to be elliptical. Displacement (D) was normalized to the original disc height and reported as stretch = (h + D)/h. The stiffness of the slow ramp test was determined from a linear regression, and the decrease in disc height during the creep test was calculated. The lumbar and tail levels were compared using unpaired t-tests. Significance was set at p<0.05.

The largest differences between lumbar and tail mechanics were in the transition zone, where lumbar segments were 2 times stiffer than the tail, and half the displacement length of the tail (Fig 1, Fig 2). In compression of the cyclic loading, there was a trend for lumbar stiffness to be greater than tail (3.65 0.98 and 2.48 1.09 MPa, respectively, p=0.1) and this difference was significant for the slow-ramp test (1.87 0.5 and 1.21 0.5 MPa). Lastly, stiffness in tension was not significantly different between lumbar and tail segments (2.54 0.44 and 2.16 0.95 MPa). FIG 1: Average mechanical behavior of lumbar and tail segments in the tension-compression cycle. During the creep test, lumbar motion segments compressed significantly less than tail following the initial application of load (77% of original disc height compared to 44% respectively, as seen in Fig 3). At the end of the creep test, the lumbar segments were also compressed much less than the tail (41% compared to 8% of initial disc height). Interestingly, both segments had the same average decrease in disc height (36%) at the same rate over the 30 minute creep test as seen in Fig 3. FIG 2: Transition zone normalized stiffness (Sn, left) and displacement (right) for lumbar and tail segments (mean SD) FIG 3: Average creep behavior for lumbar and tail segments with the initial and final % of disc height noted.

This study demonstrated that lumbar and tail motion segments of the mouse were significantly different in elastic and viscoelastic mechanics. Figure 1 clearly illustrates the large differences between lumbar and tail transition zone behavior, namely that the lumbar transition zone is much smaller and stiffer than the tail. Furthermore, these transition zone differences carry over into the creep results, in that the stretch at the beginning of the creep test for the lumbar segments (77%) was much larger than for the tail segments (44%). The additional 36% decrease in disc height due to creep for both segments, resulted in a final stretch of 41% for the lumbar and only 8% for the tail. Thus, the larger and less stiff transition zone of the tail segments, led to nearly a total loss of disc height after 30 minutes of loading at 1X body weight. These findings are particularly relevant to rodent tail models where chronic loads are applied in vivo to study mechanical pathways of degeneration [2-6]. Specifically, our results indicate that a static load of 1X body weight will produce drastically different strains in the lumbar and tail discs, which should be taken into account when interpreting results from in vivo studies. It is also important to keep in mind that our results are from lumbar segments without transverse processes and therefore differences reported here are likely to be even larger in vivo where intact lumbar processes would reduce the amount of creep under static loading. Therefore, while accessibility will continue to make the rodent tail an attractive model, our results show, based on creep and transition zone mechanical behavior, that the mouse lumbar motion segment more closely resembles the human.

1. Elliott DM, et al, Spine 2003(submitted); 2. Ching CT, et al, Clin Biomech 2003(18):p182; 3. Iatridis JC, et al, Spine 1999(24):p996; 4. Lotz J, et al, Spine 1998(23):p2493; 5. Stokes IA, et al, J Spinal Dis 1998(11):p261; 6. Kim CH, et al, TORS, 2003; 7. Smit TH, Eur Spine J 2002(11):p137.

ACKNOWLEGEMENTS: The Whitaker Foundation


The Murine Amelogenin/ARHGAP6 Nested Gene Deletion
Gibson CW, Wright JT, Boyd C, Prakash SK, Li Y, Yuan ZA, Abrams WR, Kuehl MA, Van den Veyver IB

A Anested@ gene is located within the boundaries of a larger gene, often within an intron and in the opposite orientation. These gene structures are common in bacteria and viruses, but have also been described in higher species such as Drosophila, chick, rat and human. The human amelogenin gene, which encodes the most abundant enamel protein, is located within an intron of the ARHGAP6 gene. ARHGAP6 encodes a RhoGAP protein, which regulates activity of RhoA, a small G protein involved in intracellular signal transduction.

OBJECTIVES: This study was done to describe the structure of the murine amelogenin/ARHGAP6 gene, and the nature of the enamel defect when the complex gene structure is deleted.

METHODS: A mouse was generated in which the entire ARHGAP6 gene was deleted, which also removed the nested amelogenin gene. Teeth from these mice were analyzed using light and scanning electron microscopy.

RESULT: Enamel from these mice was chalky white in color, and teeth showed excessive wear. The enamel layer was hypoplastic and nonprismatic. This phenotype is similar to that reported for amelogenin null mice, and resembles amelogenesis imperfecta in humans.

CONCLUSION: Analysis of the enamel from the amelogenin/ARHGAP6 null mice verifies that the amelogenin gene is nested within the ARHGAP6 gene in mice. The previously reported amelogenin null mice have a short deletion that removed the region surrounding the translation initiation site. This study shows that removal of the entire amelogenin gene leads to an identical phenotype, that corroborates the earlier amelogenin null mouse model results.

Supported by NIDCR grant DE11089.


Effect of Refuse on Mechanical Performance of 3 External Fixator Clamps
Snyder, DM, Gilley RS; Smith GK; Boston RC; Radin A; Kapatkin AS

Introduction: External skeletal fixation is a popular method of fracture repair in veterinary orthopedics. Although it is accepted that new fixation pins and connecting bars should be used for each patient and fracture, the reuse of external fixator clamps is common in clinical practice. The effect of clamp reuse on external fixator clamp performance is presently unknown. Thus, the purpose of this study was to determine the effects of clamp reuse on the mechanical performance of 3 types of external fixator clamps. These are as follows: the Kirschner-Ehmer (K-E) clamp; Securos clamp; and the IMEX SK system clamp.

Materials and Methods: Specially designed fixtures were used to test fixator pin and connecting bar axial slippage at 7.68 N-m of clamp bolt-tightening torque. The fixtures were mounted on an Instron Electrical Mechanical Testing Machine 4206. Force was applied at a constant rate of 0.25 mm/sec. The input loading element was the fixator pin or connecting bar, and the output element was the clamp body. Each movement parameter was tested six times using the same clamp with new connecting bars or fixator pins. Load displacement curves were analyzed using a variant of generalized least squares incorporating an iterated Gauss Newton optimization scheme, estimates of the load response (linearly estimated) and slippage point (nonlinearly estimated) were found. Statistical analysis was performed using Stata 8.1.

Results: The axial force required to cause slippage of the fixator pin in the Securos and IMEX SK clamp was comparable for each test. The axial force required to cause slippage of the fixator pin in the K-E clamp increased during testing. However, the axial force required to cause slippage of the connecting bar in the Securos, K-E, and IMEX SK clamp decreased with repeated testing.

Discussion/Conclusion: This study suggests that repeated use of external fixator clamps results in degradation of clamp performance in connecting bar axial loading. Therefore, for optimal clamp performance the reuse of the external fixator clamps evaluated in this study cannot be recommended. Clinicians should reuse external fixator clamps only with the understanding that reuse may decrease clamp performance.

Acknowledgement: This study was funded by Departmental Research Grant provided by the University of Pennsylvania School of Veterinary Medicine and donation of materials by the IMEX and Securos companies.


Progenitor Cells Responsible for Post-Natal Tendon Repair and Ectopic Bone Formation
Glaser, DL; Gupta, R; Ramachandran, R; Shore, EM; Goldhamer, DJ; Soslowsky, LJ; Kaplan, FS

Introduction: A fundamental question in the cellular pathogenesis of postnatal tendon repair and ectopic bone formation is the identity of progenitor cells responsible for new tissue. In Fibrodysplasia Ossificans Progressiva (FOP), heterotopic ossification affects both tendon and muscle. While several stem cell populations could contribute cells to this process, the heritages of these cells remain undefined. The purpose of this study is to identify the progenitor cells responsible for post-natal tendon repair and ectopic bone formation.

Methods: To generate animal models with tagged cells of specific lineages, transgenic mice that express Cre recombinase under control of a tissue-specific promoter/enhancer were then mated to Rosa26-LacZ reporter mice, resulting in cell-specific and permanent LacZ expression. Once activated, LacZ expression remains constitutive for the life of the cell and its descendants, regardless of phenotypic changes. Transgenic mouse lines with the following promoters (and cell specific expression) were generated: MyoD (satellite cells); Smooth Muscle Myosin Heavy Chain (pericytes); and Tie2 (endothelial cells, hematopoietic stem cells, side population cells). To determine the role of bone marrow-derived cells, chimeric mice were produced by transplanting bone marrow (BMT) from constitutive Rosa26-LacZ mice into irradiated C57BL/6 hosts. The fates of the lacZ tagged cells were followed during BMP induced ectopic ossification and in a patellar tendon injury model. Animals were examined histologically at various intervals from 4 days to 12 weeks. Specimens were stained for b-galactosidase (b-gal) activity using X-gal as the substrate to identify participation of specific cell lineages.

Results: In the MyoD-cre/R26R mice, labeled cells were consistently found in cartilage and bone but not in tendon. In the SMMHC-cre/R26R mice, labeled cells were found in healing tendon but not in the heterotopic ossification model. In the Tie2-cre/R26RA, labeled cells were abundant in the fibroproliferative tissue in both bone and tendon formation. In the BMT animals, labeled inflammatory cells were observed early in both models, however, fibroproliferative, cartilage, bone and tendon cells were not labeled. The bone marrow elements that repopulated the bone ossicle were labeled.

Discussion: Taken together, these lineage tracing studies provide the first definitive evidence of the heritage of cells in postnatal tendon repair and ectopic ossification and provide direct cellular targets for therapeutic approaches to regulating the regenerative response in these tissues.


ATP and Phosphate Dependence of Single Rabbit Skeletal Actomyosin Interactions Under Differing Loads
Yasuharu Takagi, Earl E. Homsher, Yale E. Goldman, Henry Shuman

The isometric force clamp (Takagi et al., Biophys. J. 78:235A, 2000) was used to maintain the position of an actin filament constant, within a finite time, in response to a myosin interaction by feedback positioning of a "transducer bead" in a dual beam optical trap. By changing the feedback gain, the response time, ?r , of the force clamp, and therefore the dynamic load on actomyosin, could be adjusted. The durations of actomyosin interactions and their dependence on ATP and Pi concentrations were measured at two values of ?r : 1 and 10 ms. With high dynamic load (?r = 1 ms), the average duration of isometric actomyosin events of rabbit skeletal myosin II was approximately 10 ms and was virtually independent of [ATP] and [Pi]. Moreover, these events were much shorter than durations measured without feedback (Takagi et al., Biophys. J. 82:373A, 2002). At moderate dynamic load (?r = 10 ms), the event durations were intermediate between the high dynamic load and no feedback. Unlike those at high dynamic load, they exhibited a dependence on both [ATP] (24.9 and 47.8 ms for 10 and 1 ?M, respectively) and [Pi] (25.2 and 47.8 ms for 10 mM and <2 ?M, respectively; [ATP] = 1 ?M). These results suggest that mechanical work by actomyosin can occur before Pi release, that the work can be reversed by an applied load, and that Pi release is load dependent. Supported by NIH grants HL15835 and AR45990.


Defective BMP Receptor Internalization in Fibrodysplasia Ossificans Progressiva
P.C. Billings, L. Serrano de la Pea, J.L. Fiori, R. Caron, E.M. Shore and F.S. Kaplan

Fibrodysplasia ossificans progressiva (FOP) is a catastrophic condition characterized by congenital malformations of the great toes and progressive heterotopic ossification of connective tissues. While the precise genetic alteration that induces bone formation in FOP is unknown, BMP4 is upregulated in FOP cells and lesional tissue suggesting that the molecular defect, giving rise to FOP, resides in the BMP signal transduction cascade. We have demonstrated previously that although steady state BMP receptor IA (BMPRIA) mRNA levels are similar in control and FOP cells, BMPRIA protein is elevated ~6 fold in FOP cells compared with cells from unaffected individuals. Using [35S]-Met metabolic labeling and immunoprecipitation, we have now determined that the rate of synthesis of BMPRIA protein is the same in normal and FOP cells. However, in the presence of BMP4 ligand, BMPRIA is internalized and rapidly degraded in control cells, but little or no receptor internalization is observed following exposure of FOP cells to BMP4. These results suggest that BMP-receptor trafficking is defective in FOP and supports the hypothesis that promiscuous BMP signaling in FOP cells results from increased BMPRIA density on the cell surface, thus giving rise to bone formation. Analysis of the molecular pathology of the human BMP4 pathway in FOP will elucidate the mechanism of abnormal ossification in this condition.


Myostatin Propeptide Mediated Amelioration of Dystrophic Pathophysiology
Sasha Bogdanovich, Kelly J. Perkins, Thomas O.B. Krag, Lisa-Anne Whittemore and Tejvir S. Khurana

Mutations in myostatin (GDF8) cause marked increases in muscle mass, suggesting that this novel transforming growth factor- (TGF-) superfamily member negatively regulates muscle growth. Myostatin blockade therefore has the potential of reversing muscle wasting in Duchenne's muscular dystrophy (DMD) without resorting to genetic manipulation. Here, we demonstrate that pharmacological blockade, using a myostatin propeptide fused to an IgG-Fc, ameliorates the dystrophic pathophysiology in the mdx mouse model of DMD. Functional benefits evidenced by specific force improvement exceeded those reported previously using myostatin antibody-mediated blockade. More importantly, use of a propeptide blockade strategy obviates possibilities of anti-idiotypic responses that could potentially limit the effectiveness of antibody-mediated myostatin blockade strategies over time. This study provides a novel, pharmacological approach using an endogenous inhibitor of myostatin for treatment of diseases associated with muscle wasting such as DMD, and should help circumvent technical hurdles and toxicity associated with conventional gene or cell based therapies.


Mutations in NIPBL, the Human Homologue of the Drosophila Nipped-B Gene, Cause Cornelia de Lange Syndrome.
Ian D Krantz, Jennifer McCallum, Cheryl DeScipio, Maninder Kaur, Lynette A Gillis, Dinah Yaeger, Lori Jukofsky, Nora Wasserman, Armand Bottani, Colleen A Morris, Malgorzata JM Nowaczyk, Helga Toriello, Michael J Bamshad, John C Carey, Eric Rappaport, Shimako Kawauchi, Arthur D Lander, Anne L Calof, Hui-hua Li, Marcella Devoto, Laird G Jackson

The Cornelia de Lange syndrome (CdLS), also termed the Brachmann-de Lange syndrome (BDLS) (OMIM# 122470) is a dominantly inherited multisystem developmental disorder. The phenotype consists of growth retardation, neurodevelopmental delay, characteristic facial features, hirsutism, striking abnormalities of the upper extremities ranging from subtle changes in the phalanges and metacarpal bones to oligodactyly and phocomelia with only minimal involvement of the lower extremities manifesting as small feet with or without syndactyly of the second and third toes. Additional findings include gastroesophageal dysfunction. Prevalence is estimated to be as high as 1 in 10, 000.

We performed genome-wide linkage exclusion analysis in 12 CdLS families and identified 4 candidate regions, with chromosome 5p13.1 giving the highest multipoint LOD score of 2.7. This information, together with the previous identification of a CdLS child with a de novo t(5;13)(p13.1;q12.1) allowed delineation of a 1.1 Mb critical region for the CdLS gene on chromosome 5. Within this region, mutations in one gene, which we have named NIPBL, were identified in sporadic and familial CdLS cases. To date mutations in this gene have been identified in over 40% of individuals with CdLS. We have characterized the genomic structure of human NIPBL and find that it is widely expressed in fetal and adult tissues. NIPBL is the human homolog of the Drosophila Nipped-B gene. The Drosophila Nipped-B gene is expressed at all stages of development in the fly, with the highest levels being seen in newly laid embryo. Although its function in mammalian systems has not been elucidated, nipped-B has been shown to be an essential regulator of cut, Ultrabithorax, and Notch receptor signaling. Nipped-B's homology to a family of chromosomal adherins and its identification through a screen for mutations that reduce activation of a distal wing margin enhancer suggest an architectural role for this protein between enhancers and promoters to facilitate their interactions.


Dll3-Notch Signaling Pathway Regulation of Somitogenesis
Kenro Kusumi, Stacey A. Stevens, Dorian M. Hall, Mizuho S. Mimoto, Dilusha A. William, Megan L. O'Brien, Alyssa A. Schaffer, Kathleen M. Loomes, and Sally L. Dunwoodie

Genes in the notch signaling pathway regulate somitogenesis. Mutations in the notch ligand Dll3 disrupt somite patterning, leading to congenital vertebral disorders in mouse and humans. We have tested several notch pathway double mutant combinations to identify the key genetic interactions regulating this process. We have found that Dll3neo/+; Notch1tm1Con/+ double mutants display vertebral, rib and craniofacial defects at full penetrance, while Notch1 or Dll3 single heterozygous animals do not display these phenotypes. Histological and in situ hybridization analysis of neonates and somite-stage embryos are currently being conducted.

To identify genes downstream of notch signaling in somitogenesis, we have used microarray expression screens to compare Dll3 homozygous mutant and wild-type microdissected somite tissue and whole embryos. Mdfi, an early myogenic factor, is decreased 13-fold in Dll3 mutant somites. In situ hybridization analysis of 10.5 dpc embryos shows that Mdfi, Myod1, Myf5, and myogenin are disrupted in expression in Dll3 mutant embryos, suggesting Dll3-notch signaling may play a role in expression of early myogenic factors.

This research was supported by the Burroughs Wellcome Fund Hitchings-Elion Fellowship, F.R.C. Murray Fellowship, and NIH-AR-050687 (KK), and the Pharmacia Foundation of Australia Fellowship and NHMRC Project Grant (SLD).


Genetic Analysis of Congenital Scoliosis
MK Maisenbacher, J-S Han, SA Stevens, MS Mimoto, ML O'Brien, D Hall, B Erol, MR Tracy, KM Loomes, EH Zackai, JP Dormans, K Kusumi

Congenital vertebral malformations occur in the context of genetic syndromes such as VATER association and Goldenhar syndrome, as well as non-syndromic congenital scoliosis and kyphosis. The genetic etiology of congenital scoliosis is largely unknown. Recently, mutations in the notch ligand DLL3 and modulator MESP2 have been identified in the vertebral disorder, spondylocostal dysostosis, and studies in the mouse have identified that notch pathway disruptions lead to vertebral defects.

We are working to identify the genetic etiology of congenital scoliosis. First, we are defining diagnostic subgroups, using radiological, clinical genetic, and molecular data from 74 infants and children with congenital vertebral malformations. Patients were divided into 32 syndromic (including VATER and Goldenhar) and 42 nonsyndromic (congenital scoliosis and kyphosis) cases. Cases were also classified by the extent of congenital malformation, into vertebral dysostosis type 1 (global), type 2 (multiple vertebrae affected in 1 region), and type 3 (single). Clinical genetic analysis of the 42 nonsyndromic congenital scoliosis patients identified a high rate of nonskeletal associations, including genitourinary (14%), cardiovascular (10%), and neurological (17%) systems.

Molecular sequence analysis of notch pathway genes was carried out for 18 syndromic patients, and sequence polymorphisms were identified: DLL3 (8 cases with synonymous and 4 with amino acid changes), and HES7 (8 cases with syn. changes). Molecular analysis of 27 nonsyndromic patients identified polymorphisms: DLL3 (13 cases with syn. and 4 with amino acid changes), and HES7 (5 cases with syn. and 1 with amino acid changes). To elucidate the significance of these findings, comparison with the general population and identification of multiplex families is underway.


Effect of Severe Chronic Obstructive Pulmonary Disease (COPD) on Sarcoendoplasmic Recticulum Calcium ATPase (SERCA) Expression in Human Diaphragm
Taitan Nguyen, Neal A Rubinstein, Larry Kaiser, Joseph Shrager, Sanford Levine

The human diaphragm expresses both a fast (SERCA1) and a slow (SERCA2) SERCA isoform. To assess the effect of COPD on SERCA expression in the human diaphragm, we obtained intra-operative biopsies of the costal diaphragm in 6 patients with severe COPD and 6 patients with essentially normal pulmonary function tests (i.e., controls). We then used isoform specific monoclonal antibodies (i.e., clone VE12G9 for SERCA1 and clone IID8 for SERCA2) to characterize diaphragm fibers with respect to the expression of SERCA isoforms. COPD diaphragms contained a lower proportion of fibers expressing only SERCA1 (175 vs 462%, p<0.001) and greater proportions of fibers expressing only SERCA2 (514 vs 432%, p<0.05) and hybrid fibers (i.e., those expressing both SERCA isoforms) (326 vs 112%, p<0.01). Subsequently, SDS-PAGE and immunoblot experiments-carried out on crude sarcoplasmic reticulum (SR) extracts-indicated that (a) both SERCA1 and SERCA2 exhibited an identical molecular mass of 116 kDa; and (b) COPD diaphragms exhibited a 30% reduction from control diaphragms in the 116 kDa protein content. These results indicate that severe COPD is associated with diaphragm remodeling characterized by (a) a fast-to-slow SERCA fiber-type transformation; and (b) a probable decrease in SERCA content.

Supported by Department of Veterans Affairs Merit Review funds.


Induction of Osteogenic Differentiation in Human Dermal Fibroblasts in the Presence of 1 ,25-Dihydroxyvitamin
D3 Hee, C K; Jonikas, M A; Nicoll, S B

Stem cells derived from multiple tissues are known to possess the ability to differentiate into specialized cell types, including osteoblasts, chondrocytes, and adipocytes [1,2]. Recently, murine cells committed to the fibroblast lineage were shown to express phenotypic markers consistent with cells of an osteoblastic phenotype [3]. However, the osteogenic differentiation potential of human fibroblasts requires further investigation. Therefore, the objective of this study was to investigate the ability of human dermal fibroblasts to adopt an osteogenic phenotype under chemically defined culture conditions.

Cell Culture: Human neonatal foreskin fibroblasts (Cascade Biologics, Portland, OR) were plated at a density of 2.5 x 104 cells/ml in either serum containing medium consisting of minimum essential medium (MEM) with 10% fetal bovine serum (FBS) and antibiotics (S) or serum containing medium with ?-glycerophosphate (?GP) (10mM), ascorbic acid (50 ?g/ml), and 1?,25 dihydroxyvitamin D3 (100 nM) (S+vitD). After 24 hours, the scaffolds were rinsed three times with Dulbecco's phosphate buffered saline (DPBS) and the medium was replaced with chemically defined medium in which 1% Insulin-Transferrin-Selenium (ITS) was substituted for serum (I or I+vitD). Control cultures were maintained in serum containing medium (S) for the duration of the study. Cultures were fed every other day and were analyzed at 7, 14, and 21 days for gene expression, alkaline phosphatase activity, DNA content, and subjected to histological and immunohistochemical characterization. RT-PCR: At each time point, total cellular RNA was extracted using the TRIZOL isolation system (Invitrogen, Carlsbad, CA). RT-PCR was performed on total RNA extracts using the Superscript First-Strand Synthesis System for RT-PCR (Invitrogen). Oligonucleotide primers specific for human osteocalcin (OC) and osteopontin (OP) were as published [4] while those for alkaline phosphatase (AP) and ?-Actin (BA) were designed using computer-aided software based on sequences deposited in GenBank. Biochemistry: Protein and DNA were extracted in 1% Triton X-100 in DPBS. Alkaline phosphatase activity was determined using a p-nitrophenol phosphate colorimetric assay (Sigma, St. Louis, MO). DNA content was quantified by the Hoechst 33258 fluorescent assay [5] adapted for a 96-well microplate. Histology: Alkaline phosphatase localization was visualized using an alkaline dye solution of Fast Blue RR salt and Naphthol AS-MX phosphate (Sigma). Anti-fibroblast monoclonal antibodies (TE-7, Chemicon, Temecula, CA) were used with Alexa Fluor 568-conjugated secondary antibodies (Molecular Probes, Eugene, OR). Alexa Fluor 488-conjugated phalloidin was used to stain F-actin. Propidium iodide or DAPI were used for nuclear counterstaining. Non-immune controls were performed without primary antibody. Statistical Analysis: A two-way ANOVA with a Bonferroni post-hoc test was performed to determine the effect of culture time and culture medium. Data represent mean standard deviation (n=3).

By 7 days, staining of the actin cytoskeleton showed a change from a fibroblast-like, spindle-shaped morphology to a polygonal shape, typical of osteoblasts (Fig. 1). Similarly, after 7 days in culture, anti-fibroblast immunohistochemistry revealed intense staining in monolayers cultured in S and I medium, while there was only slight staining in the monolayers cultured in I+vitD medium. Alkaline phosphatase RNA (normalized to ?-Actin) was expressed in all cultures with no statistical difference (Fig. 2A), however, alkaline phosphatase activity (nmoles pNitrophenol released/min/?g DNA) was significantly higher (p < 0.02) in cells cultured in I+vitD compared to all other cultures at 21 days (Fig 2D). Both osteocalcin and osteopontin gene expression (normalized to ?-Actin) were significantly increased (p < 0.02) in I+vitD cultures over S and I cultures at both 14 and 21 days (Fig 2B,C). Histologic staining for alkaline phosphatase was seen in all cultures, with the most intense staining in the I+vitD cultures (Fig. 3, arrows).

By seven days, monolayer fibroblast cultures in chemically defined medium with vitamin D (I+vitD) began to adopt a polygonal morphology and displayed marginal fibroblast staining, suggesting a change to an osteoblastic phenotype. This was followed at 14 days by a significant increase in gene expression for the osteoblast markers osteocalcin and osteopontin, and alkaline phosphatase staining. The temporal expression profile observed here closely parallels previous reports of osteoblast differentiation in response to treatment with vitamin D [6]. Moreover, these findings are consistent with earlier studies describing the importance of ITS in maintaining the differentiated phenotype of osteoblasts [7] and chondrocytes [8] from mesenchymal stem cells and dermal fibroblasts, respectively. Overall, the results of this study suggest that human dermal fibroblasts possess osteogenic differentiation potential and may be of use as an alternate source of cells for bone repair.

Figure 1: Actin cytoskeletal staining of (A) S medium, (B) I medium, and (C) I+vitD cultures at 7 days.
Figure 2: (A) alkaline phosphatase, (B) osteocalcin, and (C) osteopontin gene expression, and (D) alkaline phosphatase activity for fibroblast monolayers cultured in S, I, and I+vitD medium. *: sign. diff. compared to S cultures; #: sign. diff. compared to I cultures
Figure 3: Histological staining of alkaline phosphatase in (A) S medium, (B) I medium, and (C) I+vitD at 14 days.

[1] Pittenger et al. (1999) Science 284:143-147. [2] Young et al. (2001) Anat. Rec. 264:51-62. [3] Shui and Scutt (2002) J Cell. Phys. 193:164-172. [4] Halvorsen et al. (2001) Tiss. Eng. 7:729-741. [5] Kim et al. (1988) Anal. Biochem. 174:168-176. [6] Owen et al. (1991) Endocrinology 128 :1496-1504 [7] Osyczka et al. (2003) Trans. ORS 28:932. [8] Nicoll et al. (2002) Biochem. Biophys. Res. Comm. 292:819-825.

Differential Osteogenic Potential in Bone Marrow and Soft Tissue Stromal Cells Mediated by the Gnas/GNAS1 Gene
R.J. Pignolo, S. Blythe, J. Gilmore, F.S. Kaplan, and E.M. Shore

The Gnas gene, encoding the murine alpha-subunit of the stimulatory G-protein of adenylyl cyclase (Gs?), may function as a regulator of osteoblast differentiation in target cells that are, as yet, unidentified. Heterozygous inactivating mutations of the human GNAS1 gene have been identified in patients with progressive osseous heteroplasia (POH), a disorder in which ectopic osteoblast differentiation and bone formation occurs within soft connective tissues. We have examined osteoblast differentiation in bone marrow and soft tissue stromal cells derived from wild-type mice and mice that are heterozygous for a Gnas null allele. Bone marrow derived mesenchymal stem cells (MSCs) were isolated by aspiration of femoral cavities after removal of the metaphyses. Soft tissue stem cells were partially purified from subcutaneous fat tissue. In both cases, cells were placed into osteogenic media and assayed for their ability to form bony nodules. Bone marrow stromal cells from heterozygous mice were found to have a reduced ability to differentiate into osteoblasts. However, soft tissue stromal cells from heterozygous mice appeared to have greater osteogenic potential compared to wild-type cells. These findings suggest that Gnas differentially influences cellular differentiation and bone formation based on the type and location of target cells. Thus, inactivating mutations in human GNAS1 that cause developmental disorders of heterotopic bone formation may depend on the differential response of specific precursor cells to G-protein mediated stimulation as well as their location in soft tissue. The cellular mechanism(s) by which soft tissue stem cells give rise to ectopic bone formation remain to be elucidated.


MRI Diagnostic Criteria for Interosseous Membrane Injury Following Forearm Trauma
Neil Roach MD, Karl Limmer, Brendon Hopgood MD, Simon Chao, Scott Kozin MD

Interosseous membrane (IOM) disruption typically occurs following traumatic forearm injury. Unrecognized IOM injury ultimately results in longitudinal forearm instability with chronic pain and reduced range of motion. There currently is no definitive criteria for diagnosing IOM lesions. The purpose of this study was to define criteria for IOM injury diagnosis using MRI and examine their effectiveness in identifying IOM disruption in a double-blinded in-vitro study of forearm trauma.

Methods and Materials:
Sixteen cadaver arms were subjected to longitudinal trauma in an in-vitro forearm injury model. Each specimen was tested in elbow extension and three rotational positions (supination n = 6, neutral n = 6, pronation n = 4). A 27 kg axial load was applied from a height of 90 cm onto the distal radius. Prior to and following the creation of each lesion, longitudinal and axial MRI images were obtained. Subsequently, the specimens were dissected by a blinded investigator, describing the extent, location, and type of IOM injury. The MRI images were examined by a blinded radiologist. The procedure and criteria used for examination of IOM competence include the following: 1) identify the extent of the oblique cord from the ulna directed proximally to the radius, 2) examine the forearm beginning just distal to the proximal radio-ulnar joint identifying the IOM origin along the radius, 3) identify the IOM mid-substance traversing the interosseous space in the central third of the forearm, 4) examine the distal third of the forearm, identifying the IOM insertion along the distal ulna, 5) measure the interosseous space between the radius and ulna at a point proximal and distal in the forearm and at its largest central dimension.

Dissection revealed IOM injury in eight specimens. Seven demonstrated complete IOM disruption from its ulnar insertion and one revealed a mid-substance tear. Four specimens also had a concomitant tear of the oblique cord of the IOM. Using the above criteria, MRI analysis identified the IOM injury in seven forearms. The correct injury location was identified in six. When injured, the oblique cord of the IOM was not identifiable using MRI.

Overall, MRI demonstrated a positive predictive value of 100%, and a negative predictive value of 89% with a 87.5% sensitivity and 100% specificity. This study was the first to outline criteria for examining IOM competence using MRI. Clinically, MRI of the forearm should be obtained in any instance of axial trauma to determine the IOM status prior to definitive treatment.


Gene Expression Profiles of Extraocular Muscle Layers by Using Laser Capture Microscopy
Murat Budak, Olga Lozynska, Tejvir Khurana, and Neal Rubinstein

Extraocular muscle (EOM) is fundamentally distinct from other skeletal muscles. Moreover, one of the most striking anatomical features of all six EOMs is a distinct laminar organization, consisting of a thin outer orbital layer (OL) and a thicker inner global layer (GL). The consistency across species of this compartmentalization of the EOMs into layers suggests unique functions for each layer; unfortunately, these unique functions have not yet been identified. We have used laser capture microscopy (LCM) to isolate the OL and GL layers of rat EOMs and to analyze the differences in their gene-expression profiles.

Adult Wistar rats were sacrificed by CO2 asphyxiation. Fresh 10microM thick cryo-sections were prepared and stored by RNase free techniques .OL and GL tissues were captured from several sections by PixCell II LCM with recommended methods of Arcturus manual and Histogene LCM frozen section staining kit . After RNA isolation with Pico Pure RNA isolation kit; isolated RNA was used for T7-based RNA amplification by using the RiboAmp OA RNA Amplification kit. Second round amplification of cRNA and biotin labeling of the cRNA were performed with RiboAmp OA RNA Amplification kit and ENZO BioArrayTM HighYieldTM RNA Transcript (biotin) Labeling Kit as described in Affymetrix protocols. Biotin labeled cRNAs were fragmented to about 200-bp size. Then, 15 g of biotin-labeled cRNA samples were used for hybridization to RAE230A GeneChips for 16 h on an Affymetrix fluidics station 400. Affymetrix rat RAE230A series oligonucleotide-based arrays (GeneChips) were screened using guidelines provided by the manufacturer. Data was analyzed using GeneSpring and S-plus (statistic) softwares. Some of the structural significant genes were validated by Q-PCR and immunohistochemistry at the tissue level. Also some glucose metabolism related gene differences were validated by PAS staining of fixed tissue.

Of 15,866 micro array probe sets, 102 transcripts (including 57 known genes) were enriched in OL (e.g., Embryonic myosin-Myh3), while 111 transcripts (including 61 known genes) were GL-enriched (e.g., Actinin alpha 3-Actn3). Genes related to the metabolic pathways comprised 27.11 % ( 32 of 118) and structural genes in muscle and nerve 25.41 %( 30 of 118) are the largest functional groups of all differentially expressed genes. Other groups include genes encoding intracellular signaling 14.41%(17/118), chemo protection and Ca++ homeostasis 12.71% (15/118), growth and regeneration 10.17% (12/118), channels and membrane 7.63% (9/118), and immune response 2.54% (3/118). There were no vascular gene differences in OL and GL. We compared these results with a previous comparison of EOM vs. leg muscle which was created in our lab using the rat gene chip (RG_U34A) using GeneSpring software. Intersection gene list of RG_U34A EOM-up regulated and OL-up regulated genes had ~89% (8 out of 9) same correlation. Also, intersection gene list of RG_U34A leg-up regulated genes) and GL- up regulated genes (common 17 genes) seemed to have ~88% (15 out of 17) same correlation. .

1) The differences between two layers are predominantly in structural and metabolic genes, and as expected OL has more embryonic gene expression than GL. 2) The other interesting observation is the high correlation of heart related genes in EOM. 3) OL has more glycogen storage capacity than GL. 4) The genes that give the EOMs unique feature when they were compared with leg muscle in Fischer at all `s work seems mostly to come from OL layer.


Expression of mRNA and Protein for the Stimulatory G Protein Alpha Subunit (Gs ) of Adenylyl Cyclase in Progressive Osseous Heteroplasia (POH) Patients with Inactivating GNAS1 Mutations.
Meiqi Xu, Suzanne Jan de Beur, Chuanzhao Li, Nader Hebela, Steve Fitzgerald, Michael A. Levine, Frederick S. Kaplan, and Eileen M. Shore

Progressive osseous heteroplasia (POH) is an autosomal dominant disorder of extensive dermal ossification during childhood followed by disabling ossification of skeletal muscle and deep connective tissue. POH is caused by heterozygous inactivating GNAS1 mutations. Although most identified cases of POH appear to be due to spontaneous mutations, all known examples of genetic transmission occur by a paternal inheritance pattern. To date, we have identified GNAS1 mutations in 18 of 27 POH families with one or more affected members. These mutations include small (1-4 nucleotide) insertions and deletions, tandem duplications, and a single nucleotide substitution that disrupts a splice site junction. Each of these mutations predicts a protein reading frame shift. The GNAS1 gene consists of 12 common exons plus 4 alternative first exons and generates transcripts for XL s, NESP55, 1A, and Gs . Although the GNAS1 gene is reciprocally imprinted, the most well-characterized product, the Gs protein, is biallelically expressed in most cell types. Examination of Gs mRNA expression by RT-PCR in cultured lymphoblasts from POH patients with GNAS1 mutations revealed expression of only normal transcripts if the mutation occurred prior to exon 13, but both normal and mutant transcripts were detected if the mutation affected the last exon. This result was confirmed by RNA blot analyses that detected reduced or normal amounts of hybridizing Gs mRNA for patients with pre or post exon 13 mutations, respectively. No Gs transcripts of altered sizes were identified. Immunoblot analysis showed reduced Gs protein in cultured lymphoblast membranes from POH patients as compared to controls. These data suggest that haploinsufficiency of Gs protein may be important in directing the clinical phenotype of POH. However, paternal inheritance of GNAS1 mutations in POH is similar to that observed in pseudopseudohypoparathyroidism, a form of Albright hereditary osteodystrophy (AHO) that is clinically distinct from POH but which is also caused by paternally inherited mutations of the GNAS1 gene. We conclude that the explanation for differences between these two disorders may not be solely dependent on the level of Gs expression but may be influenced by the tissue-specific expression of the additional GNAS1 transcript forms (Nesp55, XLas, 1A) and/or by involvement of other genes.


Effects of a Restricted Feeding on Radiographic and Histopathologic Hip OA: A Lifelong Study in Laborador Retrievers
Smith,GK; Powers, MY; Biery DN; Shoffer, FS; Gregor, TP; Ballam, JM; Mantz SL; Evans, RH; Lawler, DF; Kealy, RD.

The diagnosis of canine hip dysplasia (CHD) by convention is based on subjective radiographic findings of subluxation of the coxofemoral joint, or secondary osteoarthritis (OA) on the hip-extended, ventrodorsal radiographic projection of the pelvis. It has been generally accepted that this hip phenotype at 1 or 2 years of age accurately reflects the true phenotype and genotype of the dog. The purpose of this investigation was to test the influence of food restriction on ultimate hip phenotype, and to compare 2 years of age and end-of-life hip phenotypes of subjective hip scores (OFA), OA scores and PennHIP scores with histopathology. Methods: 48, 8-week old Labrador retriever puppies from 7 litters were allotted by pairing to 2 groups of 24 dogs each. The control-fed (CF) group was fed for 15 minutes, and each member of the other group (restricted-fed, RF) was always offered 25% less of the same food given to the control-fed pair mate. Hip radiographs were made when the dogs were 30, 42, and 54 weeks of age, then yearly until end of life. The hips were evaluated for CHD and OA using criteria of the OFA scoring system. At 2 years of age, PennHIP distraction indexes were done. Histopathology of the hips was done on 45/48 dogs when they succumbed to geriatric conditions.

Restricted-feeding had a profound positive effect on the hip phenotype of Labrador retrievers. RF dogs had significantly lower incidence and severity of CHD and OA compared to CF pairmates. This benefit continued for the life of the dogs. In the pooled sample of 48 dogs, the prevalence of hip OA increased linearly throughout the study, from 15% at 2 years of age to 67% at end-of-life. For the CF dogs, end-of-life OA prevalence was 83% and for the RF dogs, 50%. At two years of age, OFA-type scoring judged 19 of the 48 dogs to be 'dysplastic' while 29 dogs were scored as 'normal'. The 19 dysplastic dogs remained dysplastic for life, with OA increasing in severity for many of the dogs. However, of the 29 dogs scored 'normal', 16 (55%) were scored radiographically dysplastic by end-of-life, equating to a 46% false-negative rate of diagnosis at 2 years of age. Twenty-four of 26 dogs scored normal at 2 years of age went on to develop histopath OA, equating to 8% negative predictive value using the OFA type scoring method. In contrast, PennHIP results showed that all the dogs in this study were susceptible to OA (DI's >0.36, range 0.36 - 0.92). Kaplan-Meier curves of disease-free interval showed that dogs with DI 0.4 had a median disease free interval of 12 years of age, compared to dogs with DI >0.6 whose median disease free interval was only 3 years of age.

This lifelong study showed that by keeping dogs lean the onset of OA was delayed and its severity and prevalence was reduced significantly. Fifty percent of RF dogs (mean body condition score of 4.6) had OA compared to 83% of the CF dogs ( 6.7 mean body score). The linear increase in OA incidence over the life of these Labrador retrievers refutes the accepted beliefs that the radiographic onset of hip dysplasia is negligible after 3 years of age and that idiopathic hip OA is a separate disease entity. This life-long study provides conclusive evidence, at least for Labs, that the designation of "normal" hips at 2 years of age was incorrect more than it was correct. The negative predictive value of the OFA type score at 2 years of age compared to histopathology was 8%, meaning that of the dogs scored 'normal', 92% went on to develop OA. The PennHIP DI indicated that all dogs in this study were susceptible to OA and therefore genotypically abnormal. Diet did not influence the distraction index.

ACKNOWLEDGMENT: - Nestle Purina PetCare Research, St Louis, MO

Changes if Cell Morphology and Gene Expression by Annulus Fibrous Cells Cultured in 2-D and 3-D Environments
Chou, A I; Reza, A T; Miller, G R; Nicoll, S B

Despite the prevalence and high costs associated with degenerative disc disease, relatively little is known about the behavior of the cells that comprise the intervertebral disc (IVD). The phenotype of IVD cells in vitro has been shown to be affected by environmental factors, such as cell-cell interactions and topographical features of the material substrate on which the cells are seeded (i.e., 2-D vs. 3-D) [1]. However, the precise role that these external cues play in governing the phenotype of IVD cells is unclear. Therefore, the objective of this study was to characterize the cellular morphology and gene expression of outer annulus fibrosus (AF) cells grown in three distinct environments: monolayer, micromass and 3-D polymer scaffold cultures. We hypothesized that outer AF cells would display morphological changes that would translate into differential collagen expression between the v

Primary Cell Isolation: Lumbar intervertebral discs were harvested from the spines of adult sheep. The outer AF was separated, diced and placed onto tissue culture-treated petri dishes and supplemented with Dulbecco's minimum essential medium (DMEM) w/ 20% fetal bovine serum (FBS), 2.5 g/mL Fungizone reagent and antibiotics. Cells were allowed to grow out from the tissue explants [2] and designated passage 0. 3-D Polymer Scaffold: Scaffolds consisted of a 1.1 mm thick poly(glycolic acid) (PGA) fiber mesh (Biomedical Structures, Slatersville, RI) that was reinforced with a 3% solution of 50 kDa poly(L-lactic acid) (PLLA). Polymers were pretreated with 1N NaOH, fashioned into 5 mm squares, and sterilized with ethanol prior to seeding. Cell Culture: All outer AF cells used in this study were from passages 1-3. Cultures were expanded in DMEM w/ 10% FBS and antibiotics. Monolayer cultures were seeded at a density of 2.5x105 cells/mL in 10 mL of medium while micromass cultures were plated at a density of 2x104cells/ L in 10 L as previously described [3]. A 40 L volume of a 5x104 cells/ L cell suspension was seeded per polymer scaffold. Cells were allowed to adhere for 1 hr and then flooded with media. Cultures were fed every 2 days and analyzed at days 3 and 7 for cell morphology and gene expression. Cell morphology: The morphology of seeded cells was assessed using a Zeiss Axiovert 200 inverted phase contrast microscope equipped with Hoffman modulation contrast optics and AxioVision image capturing software. RT-PCR: Total RNA was extracted at days 3 and 7 using the TRIZOL isolation system (Invitrogen, Carlsbad, CA). Reverse transcription (RT) was performed using the Superscript First-Strand Synthesis System for RT-PCR (Invitrogen). Oligonucleotide primers for type II collagen (COL II) were as published [4] while primers for type I collagen (COL I) and -actin (BA) were designed using computer-aided software based on sequences deposited in GenBank. Statistical Analysis: Two-way ANOVA with a Fisher's LSD post-hoc test was performed to determine effect of culture time and condition. Data represent mean standard deviation (n=3).

Outer AF cells exhibited differences in cell morphology depending on culture condition (Figure 1). Outer AF cells cultured in monolayer displayed an elongated, fibroblast-like morphology (Figure 1A) while those in micromass culture were more dense and possessed an irregular, polygonal shape (Figure 1B). Cells seeded on 3-D polymer scaffolds readily adhered to the fiber mesh and exhibited a spindle-shaped morphology typical of fibroblasts (Figure 1C, arrows). With respect to gene expression, cells on 3-D polymer scaffolds showed a significant increase in type I collagen expression (normalized to -actin) at 3 and 7 days in comparison to monolayer and micromass cultures (Figure 2A). There was no significant difference in type I collagen expression between monolayer and micromass cultures at either time point. However, a significant decrease in type II collagen expression was observed in all cultures over time (Figure 2B). In addition, the ratio of type II to type I collagen decreased between 3 and 7 days in all cultures. By 7 days, only outer AF cells seeded on 3-D polymer scaffolds showed a difference in relative collagen expression, with type I increased over type II.

Figure 1: Cell morphology of outer AF cells cultured in (A) monolayer (phase), (B) micromass (phase), and (C) 3-D polymer scaffolds (Hoffman) at 7 days in culture.
Figure 2: Gene expression for (A) Type I Collagen and (B) Type II Collagen for outer AF cells cultured in monolayer, micromass and 3-D polymer scaffolds *: sign. diff. compared to monolayer and micromass (p <0.02) #: sign. diff. compared to corresponding culture at day 7 (p <0.02)

A distinct morphological difference between outer AF cells grown in monolayer and micromass did not translate into differences in gene expression. This was surprising since this culture technique is routinely used to enhance cell-cell interactions, which promote the chondrogenic phenotype [3]. However, outer AF cells cultured in micromass did not maintain or increase type II collagen expression over time. A possible explanation for this finding is the use of outer AF cells, which are known to synthesize more type I collagen than type II [5]. In future studies, inner AF cells which inherently express more type II collagen will be cultured in micromass to determine if this cell culture method is appropriate for preservation of type II collagen expression in IVD cells in vitro. The increase in type I collagen expression by outer AF cells seeded on 3-D polymer scaffolds supports previous studies using PGA/PLLA scaffolds to engineer IVD tissue [6]. The elevated collagen expression may be a response to the scaffold chemistry (i.e., PGA/PLLA) or to the 3-D cellular arrangement that the scaffold provides. Future studies will examine outer AF cells seeded on 2-D films of PGA/PLLA to determine the contribution of scaffold chemistry and 3-D format to the behavior of AF cells in vitro. Overall, this study supports previous work that demonstrates the effect of culture environment on regulating the native IVD cellular phenotype [1,7]. REFERENCES: [1] Maldonado et al. (1992) J. Orthop. Res. 10:677-690. [2] Gruber et al. (1997) Exp. Cell. Res. 235:13-21. [3] Ahrens et al. (1977) Dev. Biol. 60:69-82. [4] Waggett et al. (1998) Matrix Biol. 16:457-470. [5] Hayes et al. (2001) Matrix Biol. 20:107-121. [6] Mizuno et al. (2001) Trans. Orthop. Res. Soc. 26:78. [7] Lee et al. (2001) Spine 26:2316-2322. ACKNOWLEDGMENTS: Louis J. Soslowsky for providing animal tissue.

An in vivo Investigation into the Effect of Environment on Wound Healing of Adult and Fetal Tendons Favata M, Beredjiklian PK, Cartmell JS, Mehta S, Blodgett RT, Crombleholme TM, Soslowsky LJ

Post-surgical scarring is a common problem in tendon repair. Previous studies established that fetal skin and cartilage heal regeneratively, with no formation of scar tissue. Recent results in our lab have shown that fetal tendon is also capable of healing scarlessly. Whether this capability is intrinsic to fetal tendon or the result of its environment is not known. Thus, we sought to examine the effect of an adult environment on adult and fetal tendon healing.

Materials and Methods:
Lateral extensor tendons were harvested from 5 ewes and their fetuses. Tendon sections were transplanted subcutaneously into adult SCID mice. After 1 week, all transplants were exposed and a 50% tenotomy was created in the injury groups. Mice with adult transplants were sacrificed 1 week post-injury; mice with fetal transplants 1 and 3 weeks post-injury. Collagen organization was quantified via polarized light microscopy using an automated digital analysis system. For biomechanics, tendons were immersed in a 37C PBS bath and preconditioned. A tensile test to failure at 0.1%/sec was performed. Paired t-tests were used to determine differences within each group.

Histologically (n=3-4), inflammation was observed in all adult, but not in any fetal specimens (Fig 1). Quantitative polarized light analysis (n=2-3) revealed high similarity between injured and control fetal tendons. Biomechanically (n=4-5), differences were found in the stiffness and modulus of the wounded vs. unwounded adult specimens 1 week post-injury (Fig 2). However, no differences were found in the stiffness of the wounded and unwounded tendons in either of the fetal groups, nor in the modulus of the 3-week fetal tendons. The modulus of the wounded 1-week fetal tendons was decreased relative to controls, but to a lesser degree than the wounded adult specimens.

photo photo

The results of our investigation indicate that the adult environment into which the fetal tendons were transplanted had no effect on the ability of this tissue to heal in a regenerative fashion and therefore that this capability is intrinsic to the fetal tendon itself. Future studies will evaluate healing at longer time points, will further examine the role of inflammation in the process, and will begin to evaluate the mechanisms responsible for this regenerative response to tendon injury.

Post-Operative Immobilization Improves Tendon to Bone Healing of the Rotator Cuff Gimbel JA, Van Kleunen JP, Thomopoulos S, Williams GR, Soslowsky LJ

Surgical repair of rotator cuff tears is frequently performed to reduce pain and improve function, but failure of the repair is a common clinical problem. Potential causative factors include chronicity, tendon quality, surgical technique, patient age, tear size, and post-operative rehabilitation. Little information exists regarding the relationship between post-operative activity level and tendon to bone healing [1,2]. Therefore, the objective of this study was to investigate the effect of activity level and time post-repair on insertion site healing in a rat model.

The supraspinatus tendon was surgically detached at its insertion site on the humerus and repaired in 57 Sprague-Dawley rats [3]. The post-operative activity level was controlled in three groups: cast immobilization (IM), cage activity (CA), and moderate treadmill exercise (EX). Animals were sacrificed at four and sixteen weeks. The biomechanical properties of the insertion site were determined using methods described previously [2,4]. Briefly, the cross-sectional area was measured and a stress relaxation (~5% strain, 600sec hold) followed by a ramp to failure experiment (~0.2%/sec) was performed in a heated PBS bath. Tissue strain at the insertion site was measured optically. Differences between activity levels at each time point were evaluated using an ANOVA followed by a Fisher's test.

The stiffness and modulus were not statistically different between activity levels at four weeks, but were increased for the IM group relative to the CA and EX groups at sixteen weeks. The peak load and load ratio were not different between activity levels at either time point, but the equilibrium load was increased for the CA group at sixteen weeks. The area was not different between activity levels at four weeks, but was increased for the EX group relative to the IM and CA groups at sixteen weeks.

Table: Insertion site properties over time with various post-operative activity levels (*p <0.05 vs activity level)
Figure: Change in stiffness with activity level at four weeks and sixteen weeks post-injury and repair

This study investigated the effect of post-operative activity level on the healing tendon to bone insertion site. Immobilization improved healing at sixteen weeks compared to exercise and cage activity, but not at four weeks. Immobilization may protect the remodeling of the tendon to bone interface from excessive loading. Therefore, the addition of a period of immobilization following rotator cuff repair may reduce the re-tear rate. Future studies will investigate the effect of combined activity levels on tendon to bone healing and joint stiffness.

Acknowledgments: This study was supported by the NIH/NIAMS and the OREF.

[1] Gelberman et al: J Hand Surg, 7:170, 1983. [2] Thomopoulos et al: JBME 125(1):106, 2003. [3] Thomopoulos et al: JOR, 20(3):454, 2002. [4] Soslowsky et al: CORR, 304:10, 1994.

Reported Physical Activity and Other Factors Associated with Cortical Bone Dimensions and Strength in School Age Children
B.S. Zemel, A Buison, R.F. Ittenbach, A Vresilovic, J Tetlak, V.A. Stallings, M.B. Leonard

In order to develop strategies to prevent osteoporosis later in life, it is important to identify the modifiable factors in childhood that are associated with fracture risk. Studies in animals and adult humans have identified cortical bone dimensions and the strain-strength index (SSI) as strong indicators of fracture risk. Little is known about the correlates of these measures in children. Peripheral quantitative computed tomography (Stratec, XCT2000) was used to measure cortical area (crt_a) and content (crt_c), and SSI in139 healthy children ages 7 to 10 years of age. Height and weight were measured and BMI z-score (BMIZ) was calculated based on the CDC growth charts. Pubertal development, ethnicity (African-American vs others) calcium intake, physical activity, and fracture history were assessed by questionnaire. Calcium intake was expressed as percent of the recommended adequate intake. Cumulative strain score was calculated based on reported physical activity and classified into low, medium and high strain activity (cum_str). Children were also classified into no high strain activities vs. any high strain activities (any_str). Continuous variables were tested for skewness and log-normalized when appropriate. Multiple regression models were developed to identify the appropriate non-modifiable predictors, such as growth status and ethnicity, so that the effect of physical activity and calcium intake on pQCT measures of bone health could be assessed. Statistical significance was determined at the p <=0.05 level. Tibia length and BMIZ were strong predictors of all measures. After inclusion of these measures, ethnicity was not statistically significant. Gender was an important additional predictor for crt_a and crt_c, but not SSI. Calcium intake and history of fracture were not associated with any of the measures. Any_str was significantly associated with all bone health measures after adjusting for tibia length, BMI and other factors (p <=0.02). Overal R2 values were 0.73, 0.72 and 0.75 for crt_a, crt_c and SSI, respectively. These findings demonstrate (a) the importance of adjusting for growth and BMI effects on bone dimensions and strength, and (2) even small amounts of routine high impact physical activity in children can significantly improve cortical bone strength.