Mechanical properties of randomly amplitude-modulated tendons
The primary goal of this study was to test the importance of amplitude modulation for the mechanical stimulation of linearly-oriented tissues. We found significant differences in tensile stiffness between the stochastically loaded and the cyclic, sinusoidal loaded tendons (with equivalent RMS amplitude) in the first two days of loading (Figure 4). Generally, the stiffness of tendons of the stochastically stretched group was reduced, compared to the cyclically loaded tendons (Figure 4). We cannot explain this difference in stiffness strictly by biological changes, such as cell viability or activity of tenocytes (Figures 5 and 6), since we did not see any significant changes in cell viability, activity or matrix production. Furthermore, it is unlikely that metabolic changes would immediately result in observable matrix degradation. Thus, the differences are probably purely mechanical, by microfracture of collagen fibers. This should be investigated in future experiments using histological analysis at the μm scale or by SEM.
Biological response of tenocytes
Relative to the day 0 control, all three groups of tenocytes responded with a minor down-regulation of ACAN and collagen type 1 (Figure 6). However, tenocytes of the stochastic loading regime tended to down-regulate ACAN, collagen type I, ADAMTS4 and MMP13 relative to the cyclic_RMS and cyclic_high group. An increase in collagen I with cyclic loading was also found by Wang et al. and Parkinson 2010 et al.  observed that there is a net proteoglycan content increase in injured tendons, due to an altered metabolism rather than due to changes in gene expression levels. However, there was no difference between the loading groups and the unloaded control in the present experiments, which is also true for the up-regulation of other genes. It should be mentioned that the measured gene expression is possibly a mixture of tenocytes and progenitor cells due to the relatively young age of the rabbits.
Culture time was certainly a limit of the study; it is possible that changes to the extracellular matrix (ECM) cannot be seen with a culture period of only twelve days. On the one hand, any changes in gene expression should be still detectable, since RNA changes can be found within hours upon mechanical loading . The timing of the culture start (here allowing an equilibration period of 3 days) will most likely have a detrimental influence on the mRNA transcript level, not so for col 1, but definitely for MMP3 and MMP13; these transcript levels have been shown to increase over time in an explant model of rat tail tendon fascicles . With respect to tissue homeostasis, we did not find any significant differences among the three loading regimes. On the other hand, it may also be that the sampling window for gene expression was delayed and thus, no changes in RNA could be detected after the stimuli. However, it has been reported that changes in mRNA persist after 24h incubation time [23, 25]. The time point of harvest after the loading regime still seems to be critical, there have been significant changes found if tissue is analyzed after 1h or longer time periods
An up-regulation of the pro-inflammatory genes that could lead to apoptosis was not evident at the RNA level in our study. Further investigation by histology or scanning electron microscopy would allow inferring definitive conclusions on the microstructure of the tendons. During the first 1–2 days of mechanical loading, 8 out of 40 tendons experienced partial ruptures and had to be re-clamped (3 in the stochastic; 2 in the cyclic_RMS and 3 in the cyclic_high group). Thus, the re-clamping of the tendons might have had an influence on the stiffness results, but may also indicate the accumulation of micro-damage. Improved clamping techniques may allow a more unbiased comparison of the clamped versus unclamped regions .
The cell density of the tendon is relatively low compared to other musculoskeletal tissues. This also includes the vascular cells and synovial cells of the tendon sheath that encloses each tendon . The tissue is sparsely vascularized and the main constituent is collagen type 1  Collagen is the main component of most organic matrices like bones, ligaments, tendons and the intervertebral disc . A remarkable 60-85% of a tendons dry weight is assigned to type I collagen. A small, mechanically important portion (2%) is elastin and 4-5% are different proteins. The extracellular substance is dominated by proteoglycans (PG) and, in combination with water, they are thought to have a spacing and lubricating role for tendon [27–29]. The mechano-biological response might be masked by the generally very rich culture media, which has an abundance of growth factors, high glucose content and vitamins. Results from the matrix production at the protein level should also be reflected by the gene expression data. For all 11 genes studied, there were no statistical and biologically significant changes amongst the loading groups. These results are consistent with studies in human achilles tendon, where no changes in the expression for genes of the major collagens and proteoglycans could be found . The same study also did not see any change for ADAMTS-4, MMP3, MMP13 and TIMP3 with the exception of the up-regulation of TIMP1. The authors hypothesize that the matrix turnover is favored for degeneration rather than matrix generation. However, another limitation of this study is that we did not look at tenocyte specific transcription factors such as scleraxis, which have been shown to respond to mechanical stimulation, especially with increased cyclic compression [30–32] nor did we look at tenomodulin and tenascin-C , two marker genes, which are important for maintaining tenocyte phenotype [34, 35]. For MMP1 and MMP3, it was found that cyclic mechanical loading inhibits their expression [6, 36]. It is generally accepted that training promotes both synthesis and degeneration and the process is highly dynamic . It is important to state that by analysis of only RNA expression-levels, conclusions on protein expression are limited. Translation efficiency, post-translational modification and -activation, protein turnover rates or inhibitory proteins that may have a large influence on how much protein is actually synthesized. MMP could be present in the tissue as pro-MMP, and thus in an inactive form, or they might be bound to TIMPs. An up-regulation of a MMP does therefore not necessarily mean matrix degeneration . Due to these potential effects it would be crucial to also include quantification on the protein level to support real-time PCR data if longer loading / culture times will be chosen in further experiments.