Over the past 8 years, our group has been continuously improving tendon repair using a functional tissue engineering (FTE) paradigm. This paradigm was motivated by inconsistent clinical results after tendon repair and reconstruction, and the modest biomechanical improvements we observed after repair of rabbit central patellar tendon defects using mesenchymal stem cell-gelsuture constructs. Although possessing a significantly higher stiffness and failure force than for natural healing, these first generation constructs were quite weak compared to normal tendon. Fundamental to the new FTE paradigm was the need to determine in vivo forces to which the repair tissue might be exposed. We first recorded these force patterns in two normal tendon models and then compared these peak forces to those for repairs of central defects in the rabbit patellar tendon model (PT). Replacing the suture with end-posts in culture and lowering the mesenchymal stem cell (MSC) concentration of these constructs resulted in failure forces greater than peak in vivo forces that were measured for all the studied activities. Augmenting the gel with a type I collagen sponge further increased repair stiffness and maximum force, and resulted in the repair tangent stiffness matching normal stiffness up to peak in vivo forces. Mechanically stimulating these constructs in bioreactors further enhanced repair biomechanics compared to normal. We are now optimizing components of the mechanical signal that is delivered in culture to further improve construct and repair outcome. Our contributions in the area of tendon functional tissue engineering have the potential to create functional load-bearing repairs that will revolutionize surgical reconstruction after tendon and ligament injury. ß

The objective of this study was to determine how mechanical stimulation affects the biomechanics and histology of stem cell-collagen sponge constructs used to repair central rabbit patellar tendon defects. Autogenous tissue-engineered constructs were created for both in vitro and in vivo analyses by seeding mesenchymal stem cells from 10 adult rabbits at 0.14x10(6) cells/construct in type I collagen sponges. Half of these constructs were mechanically stimulated once every 5 min for 8 h/day to a peak strain of 4% for 2 weeks. The other half remained in an incubator without mechanical stimulation for 2 weeks. Samples allocated for in vitro testing revealed that mechanically stimulated constructs had 2.5 times the linear stiffness of nonstimulated constructs. The remaining paired constructs for in vivo studies were implanted in bilateral full-thickness, full-length defects in the central third of rabbit patellar tendons. Twelve weeks after surgery, repair tissues were assigned for biomechanical (7 pairs) and histologic (3 pairs) analyses. Maximum force, linear stiffness, maximum stress, and linear modulus for the stimulated (vs. nonstimulated) repairs averaged 70% (vs. 55%), 85% (vs. 55%), 70% (vs. 50%), and 50% (vs. 40%) of corresponding values for the normal central third of the patellar tendons. The average force-elongation curve for the mechanically stimulated repairs also matched the corresponding curve for the normal patellar tendons, up to 150% of the peak in vivo force values recorded in a previous study. Construct and repair linear stiffness and linear modulus were also positively correlated (r = 0.6 and 0.7, respectively). Histologically both repairs showed excellent cellular alignment and mild staining for decorin and collagen type V, and moderate staining for fibronectin and collagen type III. This study shows that mechanical stimulation of stem cell-collagen sponge constructs can significantly improve tendon repair biomechanics up to and well beyond the functional limits of in vivo loading.

The objective of this study was to introduce mesenchymal stem cells (MSCs) into a gel-sponge composite and examine the effect the cells have on repair biomechanics and histology 12 weeks postsurgery. We tested two related hypotheses-adding MSCs would significantly improve repair biomechanics and cellular organization, and would result in higher failure forces than peak in vivo patellar tendon (PT) forces recorded for an inclined hopping activity. Autogenous tissue-engineered constructs were created by seeding MSCs from 15 adult rabbits at 0.1 x 10(6) cells/mL in 2.6 mg/mL of collagen gel in collagen sponges. Acellular constructs were created using the same concentration of collagen gel in matching collagen sponges. These cellular and acellular constructs were implanted in bilateral full-thickness, full-length defects in the central third of patellar tendons. At 12 weeks after surgery, repair tissues were assigned for biomechanical (n = 12 pairs) and histological (n = 3 pairs) analyses. Maximum force and maximum stress for the cellular repairs were about 60 and 50% of corresponding values for the normal central third of the PT, respectively. Likewise, linear stiffness and linear modulus for these cellular repairs averaged 75 and 30% of normal PT values, respectively. By contrast, the acellular repairs exhibited lower percentages of normal PT values for maximum force (40%), maximum stress (25%), linear stiffness (30%), and linear modulus (20%). Histologically, both repairs showed strong staining for collagen types III and V, fibronectin, and decorin. The cellular repairs also showed cellular alignment comparable to that of normal tendon. This study shows that introducing autogenous mesenchymal stem cells into a gel-collagen sponge composite significantly improves tendon repair compared to the use of a gel-sponge composite alone in the range of in vivo loading.

Autogenous tissue-engineered constructs were fabricated at four cell-to-collagen ratios (0.08, 0.04, 0.8, and 0.4 M/mg) by seeding mesenchymal stem cells (MSCs) from 16 adult rabbits at one of two seeding densities (0.1 x 10(6) and 1 x 10(6) cells/mL) in one of two collagen concentrations (1.3 and 2.6 mg/mL). The highest two ratios (0.4 and 0.8 M/mg) were damaged by excessive cell contraction and could not be used in subsequent in vivo studies. The remaining two sets of constructs were implanted into bilateral full-thickness, full-length defects created in the central third of the patellar tendon (PT). At 12 weeks after surgery, repair tissues were assigned for biomechanical (n = 13) and histological (n = 3) analyses. A second group of rabbits (n = 6) received bilateral acellular implants with the same two collagen concentrations. At 12 weeks, repair tissues were also assigned for biomechanical (n = 4) and histological (n = 2) analyses. No significant differences were observed in any structural or material properties or in histological appearance among the two cell-seeded and two acellular repair groups. Average maximum force and maximum stress of the repairs were approximately 30% of corresponding values for the central one-third of normal PT and higher than peak in vivo forces measured in rabbit PT from one of our previous publications. However, average repair stiffness and modulus were only 30 and 20% of normal PT values, respectively. Current repairs achieved higher maximum forces than in previous studies and without ectopic bone, but will need to achieve sufficient stiffness as well to be effective in the in vivo range of loading.

Rats were trained on a mixed fixed-interval schedule in which stimulus A (tone or light) indicated food availability after 10 s and stimulus B (the other stimulus) indicated food availability after 20 s. Testing consisted of non-reinforced probe trials in which the stimulus was A, B, or the compound AB. On single-stimulus trials, rats responded with a peak of activity around the programmed reinforced time. On compound-stimulus trials, rats showed a single scalar peak of responding at a time midway between those for stimulus A and B. These results suggest that when provided with discrepant information regarding the temporal predictability of reinforcement, rats compute an average of the scheduled reinforcement times for the A and B stimuli and use this average to generate an expectation of reward for the compound stimuli.Keywords peak-interval procedure; interval timing; compound stimulus; vector averaging; time perceptionThe ability of organisms to behave in accordance with the temporal relationships between events in the world is important for survival (Bateson, 2003;Brunner, Kacelnik, & Gibbon, 1992), and has recently been proposed to play a central role in associative learning (Gallistel & Gibbon, 2000;Kirkpatrick & Church, 2000;Miller & Barnet, 1993). Timing in the seconds to minutes range, interval timing, has been demonstrated in a wide range of species, ranging from honeybees to humans (Boisvert & Sherry, 2006;Henderson, Hurly, Bateson, & Healy, 2006;Rakitin et al., 1998), and shows characteristic features such as flexibility (i.e., animals are able to time stimuli that have randomly occurring onsets) and relativity (i.e., the magnitude of errors is directly proportional to the interval being timed).In order to account for this perceptual and behavioral capacity, models of interval timing generally consist of three components; a clock, memory and decision process (Church, 1997). Briefly, the clock process provides some representation of the amount of time that has elapsed since an event (e.g., a linear accumulator), the memory process stores the current clock representation whenever a biologically relevant event takes place (e.g., upon reinforcement), and the decision process determines whether the current representation of elapsed time is sufficiently close to a memory of past event times to warrant responding in a "time's up" manner. As such, interval timing models correctly predict that timing-related behavior will, on average, reflect the temporal aspects of past experience. Further, even in situations in which the programmed duration varies in a systematic manner across trials (e.g., the time of reinforcement is progressively later on each trial), the addition of a weighting parameter that