Postnatal mechanical loading drives adaptation of tissues primarily through modulation of the non-collagenous matrix

Reviewer #1:

[…] Overall, the manuscript is well written, and the results are logically and clearly presented. While the authors highlight the importance of TGF-β as a regulator of ECM organization and functional adaptation based on its upregulation in the energy storing tendon following mechanical loading, data establishing the cause and effect relationship between changes in TGF-β, matrix composition and functional adaptation seem lacking and can be addressed in a number of ways including:

1) Performing correlation analyses between changes in TGF-β1 levels, matrix composition and functional changes in tendons during aging, and

2) Evaluating consequence of knockdown of expression or function of TGF-β1 in equine tenocytes on expression levels of key ECM proteins implicated in functional adaptation.

We are thanking the reviewer for the valuable comments and have carried out correlation analyses between TGFB1 mRNA expression, matrix composition and changes in the mechanical properties during development. The correlation analyses revealed some interesting correlations between the matrisome proteins and viscoelastic and yield point properties of the matrix phase (IFM). This work has been added in the Results section and the Supplementary files 5 and 6. We have also conducted TGFB1 knockdown and stimulation experiments to support the pathway analysis carried out with the Ingenuity Pathway analysis software. Knockdown of TGFB1 and stimulation with recombinant TGF-β1 showed downregulation and upregulation, respectively, of key ECM components, BGN, COMP, COL1A2, and COL3A1, supporting a regulatory role for TGF-β1 in tenocytes and strengthening the pathway analysis. This work has been added in the Results and Discussion section and in the Figure 6C-D.

Reviewer #2:

[…] This study is of importance to the tendon community as it clearly shows the heterogeneity in fibrillar (collagenous) and matricellular (non-collagenous) phases of the tendon in composition, cellularity, and mechanical properties. The authors could improve the manuscript further by addressing the following concerns.

1) Previously published work in mice and zebrafish have shown that the matricellular protein- Thbs4 is essential for proper organization and bundling of collagen fibrils and levels of proteoglycans in the tendon (Frolova et al., 2014, Matrix Biology and Subramanian and Schilling, 2014) In addition to Maeda et al., 2011, a recently published work in zebrafish has shown that mechanical force controls cellular morphology and matrix organization in zebrafish tenons (Subramanian et al., 2018). The findings from these papers are relevant to the results shown by the authors and it would benefit the readers if the authors could discuss their results in the context of these published papers.

Thank you for the addition of these excellent papers. In this study, we have been using the term “matrix phase” to refer to the interfascicular matrix interspersed around the fascicles of the tendon and the term “fibre phase” to refer to the tendon fascicles. We realise this could have created some confusion with “matrix phase” being used for all non-collagenous components of the tendon including the intrafascicular and interfibrillar matrix. To clarify this, we have replaced the term “matrix phase” with interfascicular matrix and the term “fibre phase” with fascicles throughout the manuscript.

Having been conducted in the zebrafish and the mouse the above papers are not directly related to the investigation of the interfascicular matrix as the interfascicular matrix is found in the tendons of humans and other larger mammals. However, although not directly investigating the interfascicular matrix phase to which we are referring in this study, these papers offer key additional information and support about the role of TGFB in the regulation of tendon organisation and adaptation as a response to mechanical force as well as the role of non-collagenous ECM components for the organisation of collagen in the fascicles and as such have been added to the Discussion.

2) The authors state that based on their analyses, the changes in composition, functional adaptation, and biochemistry occur following mechanical loading in the matrix phase. The entire study relies on the analysis of tendons from developmental time points in horses. How are the authors ruling out other factors related to the normal development of a horse from birth? To be sure that mechanical force is involved, the muscles associated with the tendons should be paralyzed, which I believe is difficult to achieve in utero. Hence, the authors should modify their discussion accordingly.

Muscle paralysis interventions can be used to demonstrate a causal effect with mechanical force; however, such experiments cannot be conducted in horses and other large mammals. Therefore, here, to help distinguish between the effect of normal development and changes related to loading during development we compared a highly loaded tendon, the superficial digital flexor tendon, and its positional counterpart, the common digital extensor tendon, which is not experiencing the same high loads. This way, divergence in adaptation and mechanical properties between the two tendons was attributed to events related to loading. The limitation of muscle paralysis experiments and how we overcame it was added to the Discussion. In addition, details about the difference in function and mechanical properties of the two tendons were added in the Introduction and Discussion to define the tendons better and make this clearer.

3) Since eLife is a journal with wider readership, it would benefit the readers if there is a diagrammatic representation showing the respective tendons in the forelimb of the horse. The authors also have not defined the abbreviation CDET (subsection “Mechanical adaptation is localised to the matrix phase”, fourth paragraph).

A definition for the CDET was added and details about the difference in function and mechanical properties of the two tendons were added in the Introduction and Discussion. In addition, we added a schematic of the equine forelimb with the SDFT and CDET highlighted in Figure 1—figure supplement 1A to aid understanding.

4) The authors mention that based on pathway analysis an ECM-Integrin-Cytoskeleton to nucleus signaling paradigm was shown to be in action and they chose to focus on TGF-β signaling. The authors need to elaborate on how they arrived at TGF-β as the likely candidate considering that many ECM-Integrin interactions are possible in the tendon including several possible mechanotransduction pathways. Without a functional assay to support the hypothesis, the authors should explain if other components and pathways were considered and how they centered on TGF-β.

The ECM-integrin-cytoskeleton to nucleus signalling is known to be involved in mediating cell responses to mechanical stimuli and integrin and actin cytoskeleton signalling were canonical pathways predicted to be activated in our dataset by pathway analysis (IPA). TGFB1 was not singled out from the ECM-integrin-cytoskeleton signalling but separately predicted to be an upstream regulator activated in our dataset by pathway analysis (IPA). The predicted canonical pathways and predicted upstream regulators were therefore two separate elements of the analysis rather than us choosing TGFB1 from the ECM-integrin-cytoskeleton signalling. Other components and pathways might be involved in tendon functional adaptation, however, based on our matrix phase proteomic dataset these pathways and regulators were highlighted in the analysis. This clarification was also added in the Results section.

Reviewer #3:

[…] The authors should consider the following concerns:

1) The functional adaptation and tissue composition in superficial digital flexor tendon (SDFT) and common digital extensor tendon (CDET) were compared throughout the study. Figure 2 and Figure 3 show they have different adaptions after birth. More description/discussion should be included in the Introduction and Discussion sections on the main differences between these two types of tendon and why they show different adaptions.

More details defining the function and the mechanical properties of the SDFT and CDET were added in the Introduction and Discussion and a schematic of the equine forelimb with the SDFT and CDET highlighted was added in Figure 1—figure supplement 1A to aid understanding.

2) The matrix phase of SDFT shows more significant changes in stress relaxation, failure force, failure extension than those of CDFT among all groups after birth. However, only PRG4 and ELN show remarkable differences in 1-2 years in SDFT and CDET, respectively. Explanation should be provided on the discrepancy.

Selected proteins were chosen for immunohistochemistry based on differences identified in mass spectrometry and on biologic importance. However, more proteins identified in our mass spectrometry dataset show differences between age groups that can account for changes in mechanical properties. In addition, more subtle differences may not be identified using immunohistochemistry where quantitation was based on scoring, with the exception of elastin which was quantified with a colourimetric assay and PRG4 which was quantified with the HistoQuest Analysis Software."

3) The authors claimed that TNC is predominantly in the matrix phase of tendons. However, the data in Figure 3 does not support this conclusion. The peak score of TNC is around 2 in SDFT and CDET, which is comparable to scores of DCN and FMOD.

Yes, TNC is predominantly found in the matrix phase/IFM and not found in the fibre phase/fascicles, rather than TNC is the predominant protein in the matrix phase/IFM. We modified this sentence in manuscript to make it clearer (subsection “Structural adaptation is localised to the IFM”).

4) The histology data in Figure 3A and C are in poor quality. Higher resolution or higher power images are needed. In addition, the quantitation of the histology data are lacking in Figure 3A. It is not clear how the "more elongated nuclei" and "more linear organisation" were obtained. It would be helpful if the authors provide valid criteria for the measurements.

The poor quality is due to the generation of the PDF and the actual figures have a higher resolution. Still, we split Figure 3 into two figures to allow for larger and clearer images for the previously labelled panels 3A and C. The full quantitation data are in Figure 3—figure supplement 1 along with the scoring criteria which are also found in Supplementary file 1. The method for obtaining the quantitation data is described in the Materials and methods (subsection “Histology scoring”).

5) Detailed description /explanations are needed for results from Figure 4C. It would be helpful if more information are included in the figure legends or in the text.

More information about the pathway analysis was added both in the legend of Figure 4C (now Figure 6A) and in the Results section.

6) Figure 3B and Figure 4C were not correctly cited in the text.

Amended manuscript citations for Figures 3B and 4C (now Figure 3B and Figure 6A).