Origin of tendon stem cells in situ

Abstract

BACKGROUND: Adult stem cells are surveillance repositories capable of supplying a renewable source of progenitors for tissue repair and regeneration to maintain tissue homeostasis throughout life. Many tissue-resident stem cells have been identified in situ, which lays the foundation for studying them in their native microenvironment, i.e. the niche. Within the musculoskeletal system, muscle stem cells have been unequivocally identified in the mouse, which have led to considerable advances in understanding their role in muscle homeostasis and regeneration. On the other hand, for bone and tendon progenitor cells, mesenchymal stem cells have been used as the main in vitro cell model as they can differentiate into osteogenic, chondrogenic and tenogenic fates. Despite considerable efforts and employment of modern tools, the in vivo origins of bone and tendon stem cells remain debated. Tendon regeneration via stem cells is understudied and deserves attention as tendon damage is noted for a bleak, time-consuming recovery and the repaired tendon seldom regains the structural integrity and strength of the native, uninjured state.

OBJECTIVE: Here we review the past efforts and recent studies toward defining adult tendon stem cells and understanding tendon regeneration instead of tendon development. The focus is on adult tendon resident cells in situ and the uncertainty of their roles in regeneration.

METHODS: A systematic literature search using the Pubmed search engine was conducted encompassing the seminal papers in the tendon field.

CONCLUSIONS: Investigation of tendon stem cells in situ is in its infancy mainly due to lack of necessary tools and standardized injury model. We propose a concerted effort toward establishing a comprehensive cell atlas of the tendon, making genetic tools and choosing a reliable injury model for coordinated studies among different laboratories. Increasing our basic understanding should aid future therapeutic innovations to shorten and enhance the tendon repair/regeneration process.

Keywords

Tendon / stem cells / midsubstance / sheath / injury

Introduction

Tendons are specialized, dense connective tissue that connects skeletal muscles and bones. They are mechanical, force-transmitting conduits for muscles to affect the position of their attached bones for posture and locomotion. As such, they are continually under tension and capable of withstanding tremendous forces. For example, the Achilles tendon can bear up to 7 times the weight of the human body (^{Shah et al., 2015}). A related tissue type that connects the ends of bones together is ligament. In the clinical setting, over-stretching and tearing of tendons and ligaments present from recreational overuse, trauma, or age-associated degenerative diseases. Reconstruction surgery of the anterior cruciate ligament is a commonly known procedure to restore knee function (^{Ateschrang et al., 2017}; Petersen et al., 2017; ^{Zampeli et al., 2017}). Continuous improvement of orthopedic surgical procedures for tendons and ligaments seemingly outpaces our basic understanding for the natural process of their repair/regeneration. Given a long and productive history of studying tendon development and contrasting that to the tremendous advancement of the stem cell field at large, we are at the time to implement powerful techniques used in other developmental and stem cell systems to define and investigate tendon stem cells in vivo. With the constant mechanical stresses tendons bear, we imagine that they and their stem cells use unique strategies to maintain a homeostatic state as well as their repair/regeneration mechanisms. Studying them should therefore gain new biological principals and concepts of tissue repairs.

Tendon development

‘Regeneration recapitulates development’ is a generally held concept (^{Hoffman and Cleveland, 1988}; ^{Imokawa and Yoshizato, 1997}). Herein, we recount a few developmental events that may provide a frame of references to the process of adult tendon regeneration. For comprehensive developmental accounts, please see other reviews (Edom-Vovard and Duprez, 2004; ^{Huang et al., 2015}; Subramanian and Schilling, 2015; ^{Gaut and Duprez, 2016}). An important milestone in tendon development is the cloning and characterization of the Scleraxis gene (Scx; ^{Cserjesi et al., 1995}; ^{Schweitzer et al., 2001}; ^{Brent et al., 2003}). Scx encodes a bHLH transcription factor. Its expression defines the syndetome, a unique compartment located in the paraxial mesoderm that gives rise to the entire trunk musculoskeletal system. In the somite, myotome FGF signaling prefigures the Scx expression domain in the sclerotome (^{Brent et al., 2003}), which is regulated by the restricted expression of Pea3 and Erm (^{Brent and Tabin, 2004}). In the limb, Scx is expressed the subectodermal mesenchyme in early limb bud stages in the mouse embryo. Scx expression prefigures future tendon and ligament cells and continues into differentiated cells in these perspective tissues. It has since been used as a definitive marker for tendon progenitors and tenocytes. While not pivotal for induction in vivo, TGFb signaling, particularly through type II TGFb receptor, is essential for maintaining Scxexpression in tendon progenitors (^{Pryce et al., 2009}).

To follow Scx-expressing cells, a transgenic mouse line that harbors the regulatory elements of Scx to drive GFP expression, i.e. the ScxGFP mouse, was made (^{Pryce et al., 2007}). GFP-expression pattern of this transgene generally corresponds to endogenous Scx expression and this transgenic mouse has been widely used in the field. The authors did note that ScxGFP expression appears in a slightly larger domain than that of the endogenous Scx gene in early tenogenic regions. Subsequently, ScxGFP expression has been found in tissues outside of the tendon (^{Levay et al., 2008}; Agarwal et al., 2017). These data reflect either ectopic expression of the transgene or differential expression levels (i.e. detection sensitivity) between endogenous gene versus transgene. As transcriptional regulation is tightly controlled at the endogenous locus and miRNA-mediated post-transcriptional regulation via 3′ untranslated region (3′UTR) is prevalent, a mouse knock-in (KI) allele with direct fusion between Scx and GFP (or with an intervening 2A-peptide) and intact 3′UTR would be a better tool and deserves to be made in the future.

To investigate the role of Scx in tendon development, germ line mutant and conditional knockout of Scx have both been reported. The first report used the Prx1Cre driver, which directs Cre activity in the limb mesenchyme, to inactivate the Scxflox allele (Murchison et al., 2007). This limb-specific conditional mutant reveals the essential role of Scx in the maturation/differentiation of force-transmitting and intramuscular tendons, but not muscle-anchoring tendons. When ScxGFP was used to monitor mutant cells, it appears that they arrive at the normal location but arrested at a progenitor stage.

To selectively manipulate tendon/ligament progenitors by genetic means, two transgenic Scx-Cre drivers (^{Sugimoto et al., 2013}) as well as a KI ScxCre allele have been generated (^{Yoshimoto et al., 2017}). The transgenic Scx-Cre lines were said to be more effective in marking early progenitors than the ScxCre KI allele as assessed by Cre reporter expression. Because the ScxCre allele is also a null allele, ScxCre/Cremutant mice were analyzed for phenotypes outside of the limbs (constrained by Prx1Cre in the first report). Not only were the authors able to confirm previously described defects in the limb, they also found defects in tendons and ligaments elsewhere, including those in the diaphragm and intervertebral disks, supporting a general requirement of Scx in tendon and ligament development. Using the cell marking capacity of this ScxCre null allele, it was again shown that marked mutant cells were present at the tendon location and reduced in cell number but did not express a full complement of tendon differentiation matrix genes. Thus, Scx function is not important for specifying early progenitors, but rather for terminal differentiation of tenocytes. Should this KI ScxCre allele become publicly available, the field at large will be able to investigate gene function in tendons and ligaments systematically.

Subsequently, two other genes, Mohawk (Mkx, encoding a homeobox containing transcription factor; ^{Anderson et al., 2006}; ^{Ito et al., 2010}) and Early growth response 1 (Egr1, encoding a zinc-finger transcription factor; ^{Léjard et al., 2011}; ^{Guerquin et al., 2013}) were identified as markers for differentiating tenocytes. Mkx and Egr1 mutant mice develop tendons with reduced mass and reduced expression of tendon matrix proteins. Intriguingly, Egr1 and Scx expression are regulated by mechano-loading (^{Maeda et al., 2011}; ^{Gaut et al., 2016}). Together with Scx, their temporal expression and genetic function define two main stages of tendon development: Scx for the transition from progenitor to differentiated tenocytes, while Mkx and Egr1 for tenocyte to be fully mature in producing matrix proteins. Whether there are additional tendon developmental stages, especially progenitor and stem cell states, is an open question.

The transcriptional properties and target genes of these three transcription factors have been studied primarily in mesenchymal stem cells and reviewed extensively, and not a focus herein. Suffice it to say that they participate in activating genes that encode tendon extracellular matrix proteins, notably Col1a1 (collagen 1a1; Col1), Col1a2 (collagen 1a2) and Tnmd (tenomodulin) (^{Shukunami et al., 2006}; ^{Léjard et al., 2007}; Bagchi and Czubryt, 2012). For example, Tnmd expression is almost completely absent in ScxCre/Cre mice, and diminished in Mkx and Egr1 mutant mice. Curiously, Col1 is still detectable in ScxCre/Cre mice at the correct location. It is possible each of these transcription factors can partially compensate for each other in activating terminal differentiation matrix genes. In addition, the Smad3 transcription factor acting downstream of TGFb signaling has been shown to be indispensable for tenogenic gene expression during tenocyte differentiation (^{Berthet et al., 2013}). Despite the advances and wealth of information accumulated to date from studying these key transcriptional regulators, the primary molecular driver(s), beyond TGFb signaling and Scx, for tendon progenitor specification and expansion during embryonic development remain elusive.

As outlined above many studies have been centered around the Scx expressing (Scx) cells, but there are other cells in and around the tendon. Tubulin Polymerization-Promoting Protein Family 3 (Tppp3)has been observed by in situ hybridization to be expressed in a thin layer surrounding individual tendons (^{Staverosky et al., 2009}) and its expression persists in adulthood (^{Wang et al., 2017}). Additionally, RNA-seq on FACS-isolated ScxGFP⁺ cells from various staged developing mouse limbs uncovered over 1500 genes with expression changes at distinct time windows (^{Liu et al., 2015}). In the future, it will be interesting to see how they coordinate with Scx⁺ cells during development to generate the overall tendon structure as well as any role they may have during regeneration.

Mature tendon is compartmentalized

To date, most of our understanding of tendon organization is derived from electron microscopy data (^{Franchi et al., 2007}; ^{Richardson et al., 2007}; ^{Starborg et al., 2013}; ^{Buschmann and Bürgisser, 2017}) (Fig. 1). Despite the generally agreed upon structural organization and anatomical positions of resident cells, quite a deal of uncertainty remains over the function of these cell types as well as whether more cell types exist. During development, tendon and ligament progenitors migrate to their respective positions and condense into discrete elements/nodules during mid-gestation stages in the mouse (^{Perez et al., 2003}; ^{Watson et al., 2009}). At late embryogenesis, each of these elements take shape. They continue to grow in size after birth until adulthood, indicative of active progenitors at perinatal stages (Liu et al., 2012; ^{Dyment et al., 2014}; ^{Dyment et al., 2015}). A fully formed tendon element contains two main sub-compartments: the midsubstance in which the matrix producing tenocytes reside, and the tendon sheath which encases the midsubstance. The two ends of a tendon are further specialized: the end that connects to the bone is the enthesis, while the end that connects to the skeletal muscle, the myotendinous junction. Although midsubstance is the main compartment of a given tendon, injuries can occur to any of these compartments, and each of them may have different strategies for regeneration/repair. Below is a brief overview of these compartments.

The midsubstance

Within the midsubstance lies 90-95% of the tendon cellular components—tenoblasts and tenocytes (Kirkendall and Garrett, 1997), responsible for the synthesis and deposition of matrix precursors—collagen, elastin and proteoglycans. Tenoblasts are immature tendon cells. As they mature, tenoblasts elongate and transform into tenocytes. Tenocytes have a lower metabolic demand and are noted for their decreased nuclear to cytoplasmic ratio. The collagens produced by tenocytes form dense fibers that align approximately parallel to the tendon’s long axis, which is the direction of mechano-transduction and tenocyte alignment (Kirkendall and Garrett, 1997). The structural unit of a tendon is a collagen fibril, which is bundled hierarchically into a fascicle. Groups of fascicles compose a tendon (^{Elliott, 1965}).

The peritenon/sheath, surrounding loose connective tissue layers

The tendon midsubstance is surrounded by a loose sheath of cells called the peritenon, or simply the sheath. This peritenon sheath is thought to allow tendons to glide against surrounding tissues (^{Elliott, 1965}). Peritenon can be subdivided into an outer paratenon and an inner epitenon. While the paratenon cells are loosely organized, the epitenon is a thin layer of cells directly covering the midsubstance and continuing into its interior as the endotenon. The endotenon envelops individual fascicles, carrying small blood vessels, lymphatics and nerves (passing through the outer paratenon) (Kirkendall and Garrett, 1997). Intriguingly, perivascular cells associated with the blood vessels were recently shown to be a source of expandable progenitors via CTGF delivery during repair (^{Lee et al., 2015}). Whether these cells naturally participate in the healing process requires lineage-based studies (see below). Depending on the tendon, the peritenon varies in proportion and organization of para- versus endo-tenons. These variations likely reflect the mechanical loads of a given tendon to support the movement of the particular bone and muscle they insert into (Kirkendall and Garrett, 1997). Peritenon/epitenon has been documented to alter Col1 synthesis in response to mechanical loading (^{Mendias et al., 2012}), perhaps due to its close association with the force bearing midsubstance. In general, peritenon is thought to insulate and cushion the tendon from friction and acts as an interface between tendon proper and surrounding tissues as well as systemic factors.

The enthesis, where tendon meets the bone

Entheses are the interface between tendons (or ligaments) and bones. These interconnecting joints are essential for movement. The bone has a mechano-modulus on the order of 200 GPa whereas the tendon (or ligament) has a modulus on the order of 200 MPa (^{Lu and Thomopoulos, 2013}). To alleviate/absorb such a large modulus mismatch, entheses have specialized transitional zones with a gradual change in structure, composition and mechanical behavior (Thomopoulos et al., 2003). They are classified in two categories based on how their collagen fibers attach to the bone (^{Benjamin and Ralphs, 1998}). Direct insertions, also known as fibrocartilaginous entheses, are composed of four zones from tendon, uncalcified fibrocartilage, calcified fibrocartilage, and directly to the apophysis of the bone. Examples of these include patellar and Achilles tendons. On the other hand, indirect insertions, also known as fibrous entheses, have little structural/compositional transitions and attach to periosteal components on the shaft (i.e. metaphysis or diaphysis) of long bones, e.g. the deltoid tendon (^{Lui et al., 2010}).

Myotendinous junctions, where tendon meets the muscle

Myotendinous junctions (MTJs) are the dedicated force transmission interface where skeletal muscle engages tendon. At the MTJ, tendon collagen fascicles match to muscle fiber fascicles for structural engagement. From the muscle side, finger-like projections of sarcolemma composed of bundled actin-rich microfilaments, emanate from the last Z-line toward the tendon. This is thought to increase the contacting surface area of the MTJs for effective force transmission. Two major muscle transmembrane systems have been described for engagement with the protein matrix between myofibers and tendons, the dystrophin-associated protein complex (^{Ibraghimov-Beskrovnaya et al., 1992}) and a7b1 integrin complex (^{Bao et al., 1993}; ^{Miosge et al., 1999}). Both constitute a structural link between cytoplasmic actin and tendinous ECM proteins via Laminin 211. These ECM proteins tether the basement membrane to the collagen fibril matrix (^{Hall et al., 2007}; Charvet, Ruggiero, Le Guellec, 2012). Secreted by muscle cells, Thrombospondin 4 (Thbs4; ^{Subramanian et al., 2007; Subramanian and Schilling, 2014}; ^{Frolova et al., 2014}) is a specialized ECM protein for the MTJ (^{Tidball and Lin, 1989}). Whether there exists unique ECM protein(s), other than the generic Col1 and Tenascin-C, made by tenocytes is unknown.

Considering four domains of a given tendon, the questions naturally arisen are 1) do they have the same developmental cell origin? If so, what drives them to diversify and adopt specialized roles? 2) If they have different cell origins, where are they and how to find them? For the purpose of this review, we pose related questions: 1) How do tendon stem cells become a reserved population of cells during embryonic development to adult transition? 2) For tendon regeneration after injury, are all four domains capable of regeneration? If so, do they use the same source of stem cells? 3) If each domain has a distinct stem cell population, how do these stem cells differ in nature? 4) Perhaps the most fundamental questions are, where are they located in situ, how do they behave during regeneration, and what molecular underpinnings control their regenerative potential? Efforts toward answering these questions have been probed and discussed below.

Tendon stem/progenitor cells defined by clonogenic assays in vitro

Clonal expansion followed by lineage-specific differentiation of cells isolated from a specific tissue is considered the basic level of success in uncovering putative stem/progenitor cells (^{Liu and Martin, 2003}; ^{Bajpai et al., 2012}). Prospective cell isolation based on surface markers by FACS, followed by transplantation into the same or a different host has long been used to define a stem cell population, prior to using engineered genetic tools. Culturing cells from dissociated tendons has a long history. In these cultures, fibroblastic cells eventually become a dominant population. In specialized culture medium, mouse and human tendon cells dissociated from the midsubstance were able to form colonies in the midst of the fibroblastic cells (^{Bi et al., 2007}). These colonies contained multi-potential stem/progenitor cells that can be induced to differentiate into osteogenic, adipogenic, and tenogenic cells in vitro. They were coined as tendon progenitor/stem cells, TPSCs. Profiling of these cells using a large panel of surface markers by FACS revealed slightly different marker atlas between mouse and human TPSCs. In particular, ~ 96% of mouse TPSCs were positive for Sca1 (Sca1⁺), and ~ 75% of them, CD44⁺ . Importantly, these cells can form ectopic tendon-like structures after transplantation together with bone matrix into the mouse. They express Biglycan and Fibromodulin, two proteoglycans of the midsubstance. Mice mutant for these genes have smaller tendons (^{Bi et al., 2007}). Curiously, these mutants develop ossification near the tendon, presumably due to increased BMP2 signaling that induces them toward osteogenic fate. How they respond to tendon injury was not reported. ^{Zhang and Wang (2010)} used rabbit Achilles and patellar tendon cells to extend such in vitro system using additional markers. They found that the cultured TPSCs were reactive to Nucleostemin, Oct-4, and SSEA-4 through continuous passage, while tenocytes were unreactive to all three markers.

Given that Biglycan and Fibromodulin are globally distributed in the tendon midsubstance (^{Bi et al., 2007}), it is unclear whether all midsubstance cells have regenerative potential and the culturing condition selects a few to expand. Alternatively, only a sub-population has the intrinsic regenerative potential. Immunostaining was performed for Sca1, CD44, Nucleostemin, Oct4, and other markers and found in midsubstance and cells surrounding the midsubstance, i.e. the sheath. Surprisingly, a separate study using sheath-derived cells (^{Wang et al., 2017}), were also found to display similar multi-differentiation potential in vitro, suggesting both midsubstance and sheath are of heterogeneous cell composition and contain stem/progenitors (see below). In some tissues (e.g. liver and intestine), excellent congruence of ‘stem/progenitor cell-identification’ between clonogenic assays and lineage tracing has been documented (^{Dorrell et al., 2011}; ^{Sato et al., 2009}), though in the case of prostate stem cells discordance between these assays has been observed (^{Liu et al., 2010}). Establishing an in vitro culture condition to coax and keep stem cells competent for desired tissue regeneration upon transplantation is a technical achievement toward the goal of cell-based therapy. Whether it teaches the normal physiologic process awaits proof from a more definitive methodology such as in vivo lineage tracing.

Three-dimensional (3D) studies for tendon in vitro

In addition to culturing tendon progenitors on 2D plastic dishes, efforts have been made to culture tendon/ligament struts tethered by silk sutures or hydroxyapatite (^{Calve et al., 2004}; ^{Paxton et al., 2009}; ^{Paxton et al., 2010}) in 3D configuration. This system is useful to model tendon organization and mechano-load. More recently, ^{Chien et al. (2017)} demonstrated the feasibility of coaxing isolated, embryonic fibroblast-derived ScxGFP⁺ cells to form 3D constructs in vitro upon treatment with TGFb2. The authors also provided proof-of-principal studies for gene manipulations using viral vehicles in this setting. Because these are ScxGFP⁺ cells of embryonic origin, the relevance to adult tendon stem cells is unclear. We acknowledge several other in vitro systems, beyond the scope of this review, aimed at therapeutics have been reported. For more details, please see the references provided: engineering biomaterial scaffolds (Wu et al., 2017a; ^{Wu et al., 2017b}) and, 3D hydrogels (^{Rubio-Azpeitia et al., 2015}; ^{Yin et al., 2018}), and directing MSCs toward tendon stem/progenitors (^{Leong and Sun, 2016}; ^{Veronesi et al., 2017}). As a method to identify and isolate definitive adult tendon stem cells becomes available, we envision ‘tendon organoid’ will be a useful in vitro system. The ability to organoids from stem cells in vitro has advanced considerably. For example, ^{Gjorevski et al. (2016)} was able to decipher critical regulators of gut stem cell lineage progression/restriction using polyethylene glycol gels as a supportive material. In the future, similar strategy can be applied to tendon stem cells to screen for regulators of their activities, including proliferation, cell death, self-renewal, and differentiation.

Tendon contains label retaining cells expressing generic stem cell markers in vivo

Prior to engineered tools to define adult stem cells, retention of pulse-labeled DNA during replication in a cell after a long chase period was considered a generic way to identify stem cells in situ (^{Potten and Hendry, 1975}). The assumption for label retaining cells (LRCs) being stem cells is that stem cells divide rarely and therefore retain incorporated DNA labels for a long period, or that stem cells asymmetrically segregate and retain old DNA strands permanently, i.e. the immortal strand hypothesis (^{Cairns, 1975}). Direct testing of these hypothetical proposals by comparing LRCs with definitive stem cell markers in various tissues have, however, discredited LRCs as an inherent property of stem cells (^{Barker et al., 2007}; ^{Snippert et al., 2010}; Li and Clevers, 2010). Nevertheless, several publications have described the presence of LRCs in the tendon (^{Bi et al., 2007}; Kurth et al., 2011; ^{Runesson et al., 2013}). Due to a lack of credibility for LRCs being definitive stem cells, we describe these studies below to highlight the proliferation dynamics during tendon growth and regeneration.

During post-natal growth period, three consecutive pulses of BrdU (a nucleotide analog to label DNA) administration to 3-day old mouse pups labeled ~40% of patellar tendon cells, supporting rapid proliferation during post-natal tendon growth. After chased to 14 weeks (adult), ~6% of cells retained the BrdU-label. Regardless whether these 6% of LRCs may or may not be tendon stem cells, the reported profile of BrdU labeling and retention indicate a robust proliferation of tendon cells after the pulse-labeling period that dilutes the LRCs at 14 weeks. The patterns of LRCs at both time points were not noted to have a particular spatial signature. A subsequent study used the rat patellar tendon for IdU (another nucleotide analog) labeling at 1-day of birth to determine LRCs (^{Tan et al., 2013}). It was found that LRCs become a stable population between 6 and 8 weeks. Here, more LRCs were noted in peritenon and enthesis than in the midsubstance (ranging from ~10-4%), and some associated with the blood vessel (CD146⁺). In the midsubstance, some but not all LRCs are CD44⁺ or Sca1+, but it was unclear on the extent of their overlap. A large population of non-LRCs is also CD44⁺ or Sca1⁺. Whether Sca1⁺CD44⁺ LRCs correspond to the in vitro cultured TPSCs cannot be ascertained.

After injury to the rat patellar tendon at 6 weeks, pre-labeled IdU⁺ LRCs were found to increase by ~5 fold during the first 7 day regeneration (^{Tan et al., 2013}). Sca1⁺ and CD44⁺ populations were not quantified and did not appear concordant by in situ pattern. At 14 days post-injury, LRCs disappeared, likely reflecting dilution of pre-labeled IdU during transient expansion for repair/regeneration. As the IdU detection limit is unknown, the replicative cycles of these LRCs were not known. Many LRCs in uninjured control samples were said to express Scx or Tnmd. Given that Tnmd is a target gene of Scx, these LRCs were presumably both Scx⁺ and Tnmd⁺. Expanding LRCs at 7 days post injury were also reported to be Scx+ Tnmd+. It was unclear whether the Scx⁺Tnmd+ LRCs found at 6 weeks (prior to injury) were differentiated tenocytes incorporating IdU at their last cell cycles, and whether the expanding Scx⁺Tnmd⁺ LRCs found at 7 days post-injury were derived from Scx-Tnmd- LRCs, or whether some Scx⁺Tnmd⁺ tenocytes were proliferative. Without a second replication label and lineage relationship established between these cells, these results should be weighed carefully. Four indicators for generic stem cell markers, Oct4, Nanog, Sox2, and Nucleostemin, was further employed to correlate the stem-ness of LRCs in uninjured and injured-expanding LRCs. CD146, a marker for pericyte, was also examined. None of them were expressed by LRCs in uninjured controls, but all were in expanding LRCs.

When tendon midsubstance was dissociated for culture, expanding cells diluted out IdU, lost CD146, but maintained the other four markers – Sca1 and CD44 were not assayed. It therefore appears that these four generic stem cell markers stained transit amplifying cells instead of quiescent stem cells – that is, they indicate proliferation competence rather than stem-ness per se. Whether a given transit amplifying cell expresses only 1 or all 4 of these markers is unknown. Critically, the ‘expanding’ LRCs at day 7 after injury might not be of midsubstance origin, as there were other pre-labeled LRCs outside the midsubstance that could potentially migrate into the injured region.

In another study to survey the injury/regeneration process in the Achilles tendon, ^{Runesson et al. (2015)} employed immunostaining for CD45 (for leukocytes), Dynamin 2 (a proxy for migrating cells), as well as for Nucleostemin and Oct4. Infiltration of CD45⁺ cells 7 days after injury was found and expected. In contrast to ^{Tan et al. (2013)}, they noted Nucleostemin⁺ and Oct4⁺ cells in uninjured tendon midsubstance. Nucleostemin⁺ cells increased from ~10% in uninjured tendon to ~60% in injured area in the first 2 weeks post injury, and returned to ~10% by 8 weeks, depicting a dynamic pattern of this population of cells. When regional differences of Nucleostemin⁺ cell fractions were quantified, the surrounding connective tissue (CT) region had the smallest, while the ‘alcian blue rich region’ had the largest percentages. Few Oct-4⁺ and Dynamin2⁺ (presumed to label migratory) cells were found in the peritenon and MTJ but not in the midsubstance in uninjured controls, but they appeared in injured/regenerated midsubstance. It was inferred that the Oct4⁺ and/or Dynamin2⁺ cells in the regenerating midsubstance migrated from peritenon or MTJ. Without support from lineage tracing data, it is hard to exclude the possibility that the regenerative cells within the midsubstance turn on these markers as would be suggested by data from ^{Tan et al. (2013)}. Although the above in vivo studies reviewed here have limitations to pinpoint the origin(s) of tendon stem cells in situ, they do provide invaluable spatial and temporal information of markers expression in uninjured and injured tendon compartments.

Multiple experimental paradigms for tendon injuries

To study tendon regeneration in vivo, many experimental paradigms have been devised to injure tendons (outlined in Table 1). In a systematic evaluation across many tendons, the patellar tendon has shown to be the least variable by numerous metrics (^{Beason et al., 2012}). From a basic research standpoint, a reproducible experimental paradigm is of utmost importance. With the variety of injury paradigms and lack of agreement on model tendons to study, collective progress is difficult as reliable comparisons across laboratories cannot be made. We propose the field adopt the biopsy punch injury for the Patellar tendon as the initial standardized paradigm for it faithfully yields consistent, reproducible repair (^{Lin et al., 2006}). More importantly, this paradigm provides a focal and precise context to study the repair/regeneration process in an otherwise intact tendon. By contrast, some injury models elicit fibrotic scar formation instead of tendon regeneration. For example, longitudinal mid-1/3 patellar tendon excision appeared to generate a ‘collagenous bridge’ covering the surface over the wounded region without generating tendon fibrils (^{Dyment et al., 2014}; and see below). Complete horizontal transection of the Achilles tendon caused fibrotic tissue formation between the cut stumps without regeneration (Howell et al., 2017). Fibrotic tissue accumulation and scar formation after tendon injury are not unexpected. In fact, fibrosis likely presents a challenge toward uncovering stem cells in situ. Using punch injury, where fibrosis is limited, may pose an opportunity to identify them. This paradigm should serve as a platform to rigorously tease out the dynamics of repair/regeneration and tendon stem cell biology. Once determined, this knowledge should be applied to the other ‘failure-to-regenerate’ injury paradigms and may help explain why stem cells are unable to regenerate the tendon back to its original state.

Both robust and failed tendon regeneration models should greatly aid the findings for stimulatory and prohibitory factors, respectively, and help translate the findings toward alleviating unfavorable conditions for tendon regeneration clinically. In addition, exploiting the genetic basis for scar-free tendon regeneration in MRL/MpJ (^{Lalley et al., 2015}) and LG/J (^{Arble et al., 2016}) mice may help uncovering molecular players that circumvent regenerative hurdles encountered by tendons.

Lineage tracing studies for tendon regeneration

As mentioned above, lineage tracing is a way to track stem/progenitor cell behavior. It has been a gold standard in embryological studies to map the origins of different cell types (reviewed in Kretzchmar and Watt, 2012). Multiple lineage tracing methods have been invented and the best to date is the Cre-loxP system for indelible reporter expression (^{Soriano, 1999}). For tendon development, ScxCre driver has been used to label the tendon and ligament lineage described above. In addition to tendon and ligament, a few cells in cartilage elements are also found. Using Gdf5Cre together with the Confetti reporters to mark descendant cells with multiple fluorescent proteins, the developmental pattern of fibrocartilage cells in the enthesis has also been documented (^{Dyment et al., 2015}). In the adult, Gdf5-lineage occupies cartilage, fibrocartilage at the tendon enthesis, and many cells in the tendon midsubstance, and all cells in the ligament (^{Dyment et al., 2014}). Because constitutive Cre-drivers such as Gdf5Cre records all cells at any given time expressing Gdf5 prior to the assayed time, the adult lineage data can either be that there is a common Gdf5⁺ cell population for all described cell types, or that cells of different origins express Gdf5 at different times. By extension, even though ScxCre and Gdf5Cre mark tendon, ligament, and cartilage, one can interpret that cells in the three tissues express these genes at one point of their life, instead of their common cell origin. Intriguingly, the Bgn/Fmod mutant mice develop ectopic bones near the tendon, and ScxCre labeled cells in mice harboring a specific form of ACVR1 directly contribute to ectopic ossification, indicating that tendon cells are capable of transfating to cartilage/bone cells under specialized conditions in vivo (^{Agarwal et al., 2017}). Conditions that allow transfate may be captured in vitro where TPSCs can be differentiated to tendon, fat, cartilage, and bone cells (^{Bi et al., 2007}).

An advanced Cre system, the CreERT or the second-generation CreERT2 system utilizes a tamoxifen inducible Cre for temporal control, enabling cell marking at a specified time (^{Feil et al., 1996}; Feil et al., 1997). The advantage of using the inducible CreER over the constitutive Cre in lineage tracing has been demonstrated many times in the literature. In the developing joint, the original study using Gdf5Cre(^{Rountree et al., 2004}) assigned all joint cells of a single Gdf5-expressing progenitor source, while the later study using Gdf5CreER (^{Schwartz et al., 2016}) revealed that different joint cell lineages are derived from newly recruited Gdf5-expressing cells at different times. For the enthesis, a Gli1CreERT2 driver was activated at a selected time point to mark and trace Hedgehog-responding fibrochondrocytes that mature from an unmineralized to mineralized state during tendon growth (^{Dyment et al., 2015}), confirming prior work that Hedgehog signaling (Gli1 is a direct target of Hedgehog signaling) is critical for enthesis development. When Gli1-expressing cells in the immature enthesis were marked at post-natal day 5 (P5), and the enthesis was injured at P7, proliferative pre-marked cells were found at P14 and presumed to regenerate enthesis (^{Schwartz et al., 2017}). Cells pre-marked at P5 did not proliferate when injury was performed at P42, indicating that they progressed to a differentiated state and lost their regenerative potential by P42. However, when tamoxifen was used to mark Gli1-expressing cells after P42, small clusters of marked cells were found near the injury site, suggesting that they proliferated little (not directly assessed). These studies support that Hedgehog signals to early enthesis progenitors for their proliferation and differentiation, and that this signaling is diminished in adult enthesis, which may underlie their poor regenerative property.

This inducible lineage tracing strategy was also applied to document the tendon regeneration process using a transgenic aSMACreERT2 driver. This driver was originally developed to mark mesenchymal progenitors of adult bone (^{Grcevic et al., 2012}), but was found to also mark a subset of cells in paratenon, midsubstance, and myotendinous junction (^{Dyment et al., 2014}). Following 1/3 longitudinal excision of patellar tendon (described above), aSMA-lineage marked cells first accumulate over the injury site considerably, followed by expression of the ScxGFP transgene. Marked cells residing in the surface made collagen fibrils not characteristic of tendon cells, whereas those entering the injury site did not show longitudinally aligned collagen second harmonic generation (SHG) signal typical of tendon midsubstance. The structure formed by these aSMACreERT2 marked cells were referred to as a ‘collagenous bridge’. The presence of aSMA-descending ScxGFP⁺ cells suggest that mesenchymal cells can be coaxed to express ScxGFP in this injury paradigm, but the lack of detectable SHG signal from these cells would suggest that they are not tendon cells, at least not fully differentiated. Furthermore, as this CreERT2 driver marks multiple cell types in and around the tendons, the origin of the ‘collagenous bridge’ cells is difficult to ascertain.

More recently, two other lineage tracing strategies were documented (^{Zhang et al., 2017}; ^{Wang et al., 2017}). One used a constitutive Cre, Osteocalcin (Bglap)-Cre, which labeled sheath cells in the adult tendon (^{Wang et al., 2017}). While they confirmed extrinsic cells in the tendon can proliferate in response to injury, suggested by data from treadmill-induced mechanical loading of ScxGFP tendons (^{Mendias et al., 2012}), it is important to note that lineage marking by constitutive Cre can occur at any time up until assay. They demonstrated Bglap is upregulated in response to injury, thus, the stem cell, progenitor, differentiated cell, or a mixture of these could be labeled. While intriguing, future exploration of this gene using an inducible lineage tracing will offer greater precision and reach a definitive answer. Keratocan (Kera) has been reported to also be expressed in tenocytes (^{Rees et al., 2009}; ^{Zhang et al., 2017}). Using the tetracycline (tet)-inducible system, ^{Zhang et al. (2017)} demonstrated a Keratocan-IRES-rtTA (KerartTA) in combination TH2B-EGFP reporter line enables spatiotemporal monitoring of tenocytes during development and adulthood. The role these cells play in regeneration, and whether they overlap with Scx⁺ cells remain to be explored. Caution must be taken when using H2B-GFP as a lineage tracer given the serial dilution of GFP following each cell division.

A ScxCreERT2 driver has also been generated (Agarwal et al., 2017). Exactly how this allele was made is unclear in the literature. Nevertheless, it was used to monitor Scx-expressing cell lineage after complete horizontal transection to the Achilles tendon (Howell et al., 2017). As mentioned above, adult Achilles tendon failed to regenerate in this injury paradigm. A few unexpected and illuminating findings were reported in this study. First, neonates that underwent the same complete transection procedure did regenerate. In the regenerated region, proliferated and Scx-lineage marked cells were found, likely reflecting the proliferative potential of peri-natal progenitors needed for tendon growth. There appeared many ScxGFP⁺ cells that are not lineage marked by ScxCreERT2 (by a red color lineage tracer). Whether they are derived from non-Scx-lineage or result from inefficient lineage marking is unclear. Second, increased ScxGFP signal was detected at the severed stump in adult Achilles. Intuitively, these areas should experience the least mechanical force after severing, suggesting that either Scx expression can be induced by low mechano-force, or by a biochemical factor(s) at the injured site. Third, in the fibrotic area, aSMA⁺ cells were widespread and aSMA+ blood vessels contain ScxGFP⁺ cells.

Comparing this result to the 1/3 patella excision paradigm by ^{Dyment et al. (2014)} in which the aSMA-CreERT2 marked lineage turns on ScxGFP after covering the surface of the injury site, it seems that a aSMA lineage is capable of activating ScxGFP expression in the ‘failed regeneration’ condition, and that ScxGFP expression in itself does not signify tendon regeneration. Therefore, aSMA population(s) more likely represent cell source(s) for fibrotic scar formation. Lastly, in the tendon area next to the stump, ectopic cartilages formed with ScxGFP-expressing cells as well as ScxCreERT2 lineage marked cells. Thus, in this failed regeneration condition, Scx-lineage can transdifferentiate into chondrocytes and maintain ScxGFP expression. Given that there are no regenerated tendons in this scenario, whether Scx-lineage can regenerate tendon cells in a permissive condition is not known. If Scx-lineage can regenerate tendon cells, whether all or selected few Scx⁺ cells can do so, will be difficult to determine using this CreERT2 driver. As to why adult tendon in this paradigm does not regenerate, we suspect that low mechanical load and small size of the neonatal Achilles tendon allows the actively proliferating progenitor cells to migrate and join together, whereas without suturing the transected adult Achilles, the tendon stumps were pulled apart too far and without any linkage support, failed to regenerate.

Inducible lineage tracing coupled with molecular markers offers unparalleled power to define spatiotemporal relationships between stem cells and their descending cell types. In the case of using Gli1CreERT2, the constraint lies in Hedgehog-responsive cells. Non-Hedgehog responsive tendon enthesis stem cells, if exist, could not be examined. The ScxCreERT2 driver is invaluable, but it has only been used in limited ways. The combination of ScxCreERT2 and Scxflox alleles should allow future investigation of Scx function in maintaining tendon midsubstance and regenerating tendons after injury in the adult. However, one or two CreERT2 alleles are unlikely sufficient to understand tendon regeneration comprehensively, especially considering that the tendon has multiple compartments, not to mention numerous cells types already revealed by staining patterns of various antibodies.

Therapeutic interventions to improve tendon healing

Due to clinical relevance, many methods have been sought to improve tendon healing by supplying natural or synthetic biomaterials to serve as a structural scaffold (^{Sundar et al., 2009}; ^{Hexter et al., 2017}). In addition, delivery of growth factors, such as IGF-1 (^{Kurtz et al., 1999}; ^{Dahlgren et al., 2002}), VEGF (^{Kaux et al., 2014}), bFGF (^{Fukui et al., 1998}; ^{Chan et al., 2000}), TGF-b (^{Chang et al., 2000}), PDGF (^{Letson and Dahners, 1994}; ^{Hildebrand et al., 1998}), GDF5/6/7 (^{Wolfman et al., 1997}; Rickert et al., 2001), and platelet-rich plasma (PRP; ^{Kajikawa et al., 2008}; ^{Lyras et al., 2009}; ^{Baksh et al., 2013}), have been applied with variable degrees of success (^{Molloy et al., 2003}). Whether and how the endogenous tendon stem cells respond to the above factors are not known due to the elusive identity of these cells. It is possible that these factors act in other supportive cells that are beneficial to tendon healing. Even so, they are still clinically relevant and inform us on new dimensions to investigate their roles in supporting tendon stem cell activity. How they may influence the proliferation and differentiation properties of the in vitro defined transplantation competent TSPCs (^{Bi et al., 2007}) should be examined for a better understanding. Along this line, a different population of TSPCs identified by another surface marker CD146 has been found to be responsive to connective tissue growth factor (CTGF) toward tenogenic differentiation in vitro (Lee et al., 2015). A direct comparison between these new CD146⁺ TSPCs (Lee et al., 2015) and the Sca1⁺/CD44⁺ mouse TPSCs (^{Bi et al., 2007}) is also lacking. Interestingly, ~ 91% of the in vitro defined human TSPCs are CD164⁺ (^{Bi et al., 2007}). Nevertheless, CTGF delivery in vivo enhanced healing, suggestive of a stem-like role of CD146⁺ cells. CD146+ cells appeared rare in the tendon tissue in situ and were only found around CD31+ endothelial cells, which is consistent with CD146 being an ascribed pericyte marker (^{Shih IM, 1999}; ^{Covas et al., 2008}). Could pericytes or a sub-population of them be the elusive tendon stem cells in situ, or they are coaxed to transfate to tendon progenitors by CTGF treatment? Unfortunately, CTGF knockout mice develop chondrodysplasia and are lethal, so the endogenous role of CTGF in adult tendon repair/regeneration is yet to be determined in vivo (^{Ivkovic et al., 2003}). For therapeutic designs that target tendon stem cells to enhance their regenerative capacity, definitive identification of these cells, followed by molecular characterization, is very much needed.

Concluding remarks

Because molecular pursuit of tendon stem cells is at the beginning stage, the field should take this as an opportunity to plan future research strategically. As a starting point, one must tackle the current unmet needs in tendon stem cell biology: 1) attain a global assessment of gene expression patterns in adult tendon compartments and assign spatially distinct cell types; 2) generate tendon compartment-specific CreERT2 drivers to investigate the role of each compartment in tendon growth, injury and disease; 3) selective gene inactivation studies to determine crucial players in various aspects of tendon biology. To understand the biological complexity of cell types residing in the adult tendon, a gene expression atlas is of crucial importance. Gene expression of ScxGFP⁺ cells has been profiled at different stages in the embryonic limb (^{Liu et al., 2015}). We propose that a comprehensive profiling of all tendon cell populations, rather than one cell type, is needed. In addition to bulk RNA-seq, single cell RNA-seq has been shown to be powerful for uncovering novel markers and assembling single cell lineage trees (^{Zheng et al., 2017}). Applying this technology to tendon and combined with marker analysis, a topographic expression map of the tendon can be constructed. Once a molecular atlas of tendon is drawn, CreERT2 alleles driven by genes in different compartments or specialized sub-compartment should then be made as essential genetic tools. For the making of new CreERT2 alleles, a knock-in design that preserves the 5′ and 3′ UTR of the gene to recapitulate post-transcriptional regulation is desirable. These alleles then can be used to determine cell location (marking) and the fate(s) of their descendant cells (tracing), i.e. whether any of them mark tendon stem cells for maintaining homeostasis throughout life and/or for regeneration after injury. Alternatively, screening pre-existing CreERT2 alleles deposited to JAX may be an immediate solution to obtain lineage marking tools for the tendon. Finally, the function of a given gene of interest should be tested by conditional inactivation in conjunction with lineage marking by choosing from the bank of tendon-specific CreERT2 alleles. Combination of cell marking and gene inactivation should allow for determining the critical action of the mutated gene.

We have proposed above to standardize injury paradigms, which should incentivize the re-examination of tendon injury-regeneration process with extensive time points by RNA-seq. A repository of these data sets will provide the molecular foundation to stage regenerative progression, instead of only assessing a few selected tendon-specific genes by qRT-PCR. Isolating selective lineage-marked cells after injury/regeneration for comprehensive expression profiling and comparing their profile to the tendon expression atlas in the future should substantially increase the accuracy to determine regenerated cell type(s). By developing new tools and applying modern genomic technologies, tremendous strides will be made to tackle the fascinating biology of the tendon. Of utmost importance, definitive proof of stem cell existence in situ by way of lineage tracing and transciptome analysis will revolutionize the tendon field. Should the elusive tendon stem cell(s) be found, stem cell-based therapies may change the current clinical approaches for treating developmental, functional and disease-related issues.

References

[1] Agarwal S, Loder S J, Cholok D, Peterson J, Li J, Breuler C, Cameron Brownley R, Hsin Sung H, Chung M T, Kamiya N, Li S, Zhao B, Kaartinen V, Davis T A, Qureshi A T, Schipani E, Mishina Y, Levi B (2017). Scleraxis-lineage cells contribute to ectopic bone formation in muscle and tendon. Stem Cells, 35(3): 705–710

[2] Anderson D M, Arredondo J, Hahn K, Valente G, Martin J F, Wilson-Rawls J, Rawls A (2006). Mohawk is a novel homeobox gene expressed in the developing mouse embryo. Dev Dyn, 235(3): 792–801

[3] Arble J R, Lalley A L, Dyment N A, Joshi P, Shin D G, Gooch C, Grawe B, Rowe D, Shearn J T (2016). The LG/J murine strain exhibits near-normal tendon biomechanical properties following a full-length central patellar tendon defect. Connect Tissue Res, 57(6): 496–506

[4] Ateschrang A, Ahmad S S, Stöckle U, Schroeter S, Schenk S, Ahrend M D (2017). Recovery of ACL function after dynamic intraligamentary stabilization is resultant to restoration of ACL integrity and scar tissue formation. Knee Surg Sports Tramatol Arthrosc

[5] Bagchi R A and Czubryt M P (2012). Synergistic roles of scleraxis and Smads in the regulation of collagen 1a2 gene expression. Biochim Biophys Acta, 1823(10): 1936–1944

[6] Bajpai V K, Mistriotis P, Andreadis S T (2012). Clonal multipotency and effect of long-term in vitro expansion on differentiation potential of human hair follicle derived mesenchymal stem cells. Stem Cell Res (Amst), 8(1): 74–84

[7] Baksh N, Hannon C P, Murawski C D, Smyth N A, Kennedy J G (2013). Platelet-rich plasma in tendon models: a systematic review of basic science literature. Arthroscopy, 29(3): 596–607

[8] Bao Z Z, Lakonishok M, Kaufman S, Horwitz A F (1993). Alpha 7 beta 1 integrin is a component of the myotendinous junction on skeletal muscle. J Cell Sci, 106(Pt 2): 579–589

[9] Barker N, van Es J H, Kuipers J, Kujala P, van den Born M, Cozijnsen M, Haegebarth A, Korving J, Begthel H, Peters P J, Clevers H (2007). Identification of stem cells in small intestine and colon by marker gene Lgr5. Nature, 449(7165): 1003–1007

[10] Beason D P, Kuntz A F, Hsu J E, Miller K S, Soslowsky L J (2012). Development and evaluation of multiple tendon injury models in the mouse. J Biomech, 45(8): 1550–1553

[11] Benjamin M and Ralphs J R (1998). Fibrocartilage in tendons and ligaments--an adaptation to compressive load. J Anat, 193(4): 481–494

[12] Berthet E, Chen C, Butcher K, Schneider R A, Alliston T, Amirtharajah M (2013). Smad3 binds Scleraxis and Mohawk and regulates tendon matrix organization. J Orthop Res, 31(9): 1475–1483

[13] Bi Y, Ehirchiou D, Kilts T M, Inkson C A, Embree M C, Sonoyama W, Li L, Leet A I, Seo B M, Zhang L, Shi S, Young M F (2007). Identification of tendon stem/progenitor cells and the role of the extracellular matrix in their niche. Nat Med, 13(10): 1219–1227

[14] Brent A E, Schweitzer R, Tabin C J (2003). A somitic compartment of tendon progenitors. Cell, 113(2): 235–248

[15] Brent A E, Tabin C J (2004). FGF acts directly on the somitic tendon progenitors through the Ets transcription factors Pea3 and Erm to regulate scleraxis expression. Development, 131(16): 3885–3896

[16] Buschmann J, Bürgisser G M ( 2017). Biomechanics on tendons and ligaments. Zurich: Elsevier, Print

[17] Cairns J (1975). Mutation selection and the natural history of cancer. Nature, 255(5505): 197–200

[18] Calve S, Dennis R G, Kosnik P E 2nd, Baar K, Grosh K, Arruda E M (2004). Engineering of functional tendon. Tissue Eng, 10(5-6): 755–761

[19] Chan B P, Fu S, Qin L, Lee K, Rolf C G, Chan K (2000). Effects of basic fibroblast growth factor (bFGF) on early stages of tendon healing: a rat patellar tendon model. Acta Orthop Scand, 71(5): 513–518

[20] Chang J, Thunder R, Most D, Longaker M T, Lineaweaver W C (2000). Studies in flexor tendon wound healing: neutralizing antibody to TGF-beta1 increases postoperative range of motion. Plast Reconstr Surg, 105(1): 148–155

[21] Charvet B, Ruggiero F, Le Guellec D (2012). The development of the myotendinous junction. A review. Muscles Ligaments Tendons J, 2(2): 53–63

[22] Chien C, Pryce B, Tufa S F, Keene D R, Huang A H (2017). Optimizing a 3D model system for molecular manipulation of tenogenesis. Connect Tissue Res, 22: 1–14

[23] Covas D T, Panepucci R A, Fontes A M, Silva W A Jr, Orellana M D, Freitas M C, Neder L, Santos A R, Peres L C, Jamur M C, Zago M A (2008). Multipotent mesenchymal stromal cells obtained from diverse human tissues share functional properties and gene-expression profile with CD146+ perivascular cells and fibroblasts. Exp Hematol, 36(5): 642–654

[24] Cserjesi P, Brown D, Ligon K L, Lyons G E, Copeland N G, Gilbert D J, Jenkins N A, Olson E N (1995). Scleraxis: a basic helix-loop-helix protein that prefigures skeletal formation during mouse embryogenesis. Development, 121(4): 1099–1110

[25] Dahlgren L A, van der Meulen M C, Bertram J E, Starrak G S, Nixon A J (2002). Insulin-like growth factor-I improves cellular and molecular aspects of healing in a collagenase-induced model of flexor tendinitis. J Orthop Res, 20(5): 910–919

[26] Dorrell C, Erker L, Schug J, Kopp J L, Canaday P S, Fox A J, Smirnova O, Duncan A W, Finegold M J, Sander M, Kaestner K H, Grompe M (2011). Prospective isolation of a bipotential clonogenic liver progenitor cell in adult mice. Genes Dev, 25(11): 1193–1203

[27] Dyment N A, Breidenbach A P, Schwartz A G, Russell R P, Aschbacher-Smith L, Liu H, Hagiwara Y, Jiang R, Thomopoulos S, Butler D L, Rowe D W (2015). Gdf5 progenitors give rise to fibrocartilage cells that mineralize via hedgehog signaling to form the zonal enthesis. Dev Biol, 405(1): 96–107

[28] Dyment N A, Hagiwara Y, Matthews B G, Li Y, Kalajzic I, Rowe D W (2014). Lineage tracing of resident tendon progenitor cells during growth and natural healing. PLoS One, 9(4): e96113

[29] Edom-Vovard F, Duprez D (2004). Signals regulating tendon formation during chick embryonic development. Dev Dyn, 229(3): 449–457

[30] Elliott D H (1965). Structure and Function of Mammalian Tendon. Biol Rev Camb Philos Soc, 40(3): 392–421

[31] Feil R, Wagner J, Metzger D, Chambon P (1997). Regulation of Cre recombinase activity by mutated estrogen receptor ligand-binding domains. Biochem Biophys Res Commun, 237(3): 752–757

[32] Feil R, Brocard J, Mascrez B, LeMeur M, Metzger D, Chambon P (1996). Ligand-activated site-specific recombination in mice. PNAS 93: 10887–10890

[33] Franchi M, Trirè A, Quaranta M, Orsini E, Ottani V (2007). Collagen structure of tendon relates to function. Sci World J, 7: 404–420

[34] Frolova E G, Drazba J, Krukovets I, Kostenko V, Blech L, Harry C, Vasanji A, Drumm C, Sul P, Jenniskens G J, Plow E F, Stenina-Adognravi O (2014). Control of organization and function of muscle and tendon by thrombospondin-4. Matrix Biol, 37: 35–48

[35] Fukui N, Katsuragawa Y, Sakai H, Oda H, Nakamura K (1998). Effect of local application of basic fibroblast growth factor on ligament healing in rabbits. Rev Rhum Engl Ed, 65(6): 406–414

[36] Gaut L, Duprez D (2016). Tendon development and diseases. Dev Biol, 5(1): 5–23

[37] Gaut L, Robert N, Delalande A, Bonnin M A, Pichon C, Duprez D (2016). EGR1 regulates transcription downstream of mechanical signals during tendon formation and healing. PLoS One, 11(11): e0166237

[38] Gjorevski N, Sachs N, Manfrin A, Giger S, Bragina M E, Ordóñez-Morán P, Clevers H, Lutolf M P (2016). Designer matrices for intestinal stem cell and organoid culture. Nature, 539(7630): 560–564

[39] Grcevic D, Pejda S, Matthews B G, Repic D, Wang L, Li H, Kronenberg M S, Jiang X, Maye P, Adams D J, Rowe D W, Aguila H L, Kalajzic I (2012). In vivo fate mapping identifies mesenchymal progenitor cells. Stem Cells, 30(2): 187–196

[40] Guerquin M J, Charvet B, Nourissat G, Havis E, Ronsin O, Bonnin M A, Ruggiu M, Olivera-Martinez I, Robert N, Lu Y, Kadler K E, Baumberger T, Doursounian L, Berenbaum F, Duprez D (2013). Transcription factor EGR1 directs tendon differentiation and promotes tendon repair. J Clin Invest, 123(8): 3564–3576

[41] Gumucio J P, Phan A C, Ruehlmann D G, Noah A C, Mendias C L (2014). Synergist ablation induces rapid tendon growth through the synthesis of a neotendon matrix. J Appl Physiol (1985), 117(11): 1287–1291

[42] Hall T E, Bryson-Richardson R J, Berger S, Jacoby A S, Cole N J, Hollway G E, Berger J, Currie P D (2007). The zebrafish candyfloss mutant implicates extracellular matrix adhesion failure in laminin 2-deficient congenital muscular dystrophy. Proc Natl Acad Sci USA, 104(17): 7092–7

[43] Hexter A T, Pendegrass C, Haddad F, Blunn G (2017). Demineralized Bone Matrix to Augment Tendon-Bone Healing: A Systematic Review. Orthop J Sports Med, 5(10): 2325967117734517

[44] Hildebrand K A, Woo S L, Smith D W, Allen C R, Deie M, Taylor B J, Schmidt C C (1998). The effects of platelet-derived growth factor-BB on healing of the rabbit medial collateral ligament. An in vivo study. Am J Sports Med, 26(4): 549–554

[45] Hoffman P N, Cleveland D W (1988). Neurofilament and tubulin expression recapitulates the developmental program during axonal regeneration: induction of a specific b-tubulin isotype. Proc Natl Acad Sci USA, 85(12): 4530–4533

[46] Howell K, Chien C, Bell R, Laudier D, Tufa S F, Keene D R, Andarawis-Puri N, Huang A H (2017) Novel model of tendon regeneration reveals distinct cell mechanisms underlying regenerative and fibrotic tendon healing. Sci Rep, 7: 45238

[47] Huang A H, Lu H H, Schweitzer R (2015). Molecular regulation of tendon cell fate during development. J Orthop Res, 33(6): 800–812

[48] Ibraghimov-Beskrovnaya O, Ervasti J M, Leveille C J, Slaughter C A, Sernett S W, Campbell K P (1992). Primary structure of dystrophin-associated glycoproteins linking dystrophin to the extracellular matrix. Nature, 355(6362): 696–702

[49] Imokawa Y, Yoshizato K (1997). Expression of Sonic hedgehog gene in regenerating newt limb blastemas recapitulates that in developing limb buds. Proc Natl Acad Sci USA, 94(17): 9159–9164

[50] Ito Y, Toriuchi N, Yoshitaka T, Ueno-Kudoh H, Sato T, Yokoyama S, Nishida K, Akimoto T, Takahashi M, Miyaki S, Asahara H (2010). The Mohawk homeobox gene is a critical regulator of tendon differentiation. Proc Natl Acad Sci USA, 107(23): 10538–10542

[51] Ivkovic S, Yoon B S, Popoff S N, Safadi F F, Libuda D E, Stephenson R C, Daluiski A, Lyons K M (2003). Connective tissue growth factor coordinates chondrogenesis and angiogenesis during skeletal development. Development, 130(12): 2779–2791

[52] Kajikawa Y, Morihara T, Sakamoto H, Matsuda K, Oshima Y, Yoshida A, Nagae M, Arai Y, Kawata M, Kubo T (2008). Platelet-rich plasma enhances the initial mobilization of circulation-derived cells for tendon healing. J Cell Physiol, 215(3): 837–845

[53] Kaux J F, Janssen L, Drion P, Nusgens B, Libertiaux V, Pascon F, Heyeres A, Hoffmann A, Lambert C, Le Goff C, Denoël V, Defraigne J O, Rickert M, Crielaard J M, Colige A (2014). Vascular Endothelial Growth Factor-111 (VEGF-111) and tendon healing: preliminary results in a rat model of tendon injury. Muscles Ligaments Tendons J, 4(1): 24–28

[54] Kirkendall D T and Garrett W E (1997). Function and biomechanics of tendons. Scand J Med Sci Sports, 7(2): 62–66

[55] Kretzschmar K and Watt F M (2012). Lineage tracing. Cell, 148(1-2): 33–45

[56] Kurth T B, Dell’Accio F, Crouch V, Augello A, Sharpe P T, De Bari C (2011). Functional mesenchymal stem cell niches in adult mouse knee joint synovium in vivo. Arthritis Rheum, 63(5): 1289–1300

[57] Kurtz C A, Loebig T G, Anderson D D, DeMeo P J, Campbell P G (1999). Insulin-like growth factor I accelerates functional recovery from Achilles tendon injury in a rat model. Am J Sports Med, 27(3): 363–369

[58] Lalley A L, Dyment N A, Kazemi N, Kenter K, Gooch C, Rowe D W, Butler D L, Shearn J T (2015). Improved biomechanical and biological outcomes in the MRL/MpJ murine strain following a full-length patellar tendon injury. J Orthop Res, 33(11): 1693–1703

[59] Lee C H, Lee F Y, Tarafder S, Kao K, Jun Y, Yang G, Mao J J (2015). Harnessing endogenous stem/progenitor cells for tendon regeneration. J Clin Invest, 125(7): 2690–2701

[60] Léjard V, Blais F, Guerquin M J, Bonnet A, Bonnin M A, Havis E, Malbouyres M, Bidaud C B, Maro G, Gilardi-Hebenstreit P, Rossert J, Ruggiero F, Duprez D (2011). EGR1 and EGR2 involvement in vertebrate tendon differentiation. J Biol Chem, 286(7): 5855–5867

[61] Léjard V, Brideau G, Blais F, Salingcarnboriboon R, Wagner G, Roehrl M H, Noda M, Duprez D, Houillier P, Rossert J (2007). Scleraxis and NFATc regulate the expression of the pro-a1(I) collagen gene in tendon fibroblasts. J Biol Chem, 282(24): 17665–17675

[62] Leong D J, Sun H B (2016). Mesenchymal stem cells in tendon repair and regeneration: basic understanding and translational challenges. Ann N Y Acad Sci, 1383(1): 88–96

[63] Letson A K, Dahners L E (1994). The effect of combinations of growth factors on ligament healing. Clin Orthop Relat Res, (308): 207–212

[64] Levay A K, Peacock J D, Lu Y, Koch M, Hinton R B Jr, Kadler K E, Lincoln J (2008). Scleraxis is required for cell lineage differentiation and extracellular matrix remodeling during murine heart valve formation in vivo. Circ Res, 103(9): 948–956

[65] Li L and Clevers H (2010). Coexistence of quiescent and active adult stem cells in mammals. Science, 327(5965): 542–545

[66] Lin T W, Cardenas L, Glaser D L, Soslowsky L J (2006). Tendon healing in interleukin-4 and interleukin-6 knockout mice. J Biomech, 39(1): 61–69

[67] Liu C F, Aschbacher-Smith L, Barthelery N J, Dyment N, Butler D, and Wylie C (2012). Spatial and temporal expression of molecular markers and cell signals during normal development of the mouse patellar tendon. Tissue Eng Part A, 18(5-6): 598–608

[68] Liu H, Xu J, Liu C F, Lan Y, Wylie C, Jiang R (2015). Whole transcriptome expression profiling of mouse limb tendon development by using RNA-seq. J Orthop Res, 33(6): 840–848

[69] Liu R, Zhang Z, Xu Y (2010). Downregulation of nucleostemin causes G1 cell cycle arrest via a p53-independent pathway in prostate cancer PC-3 cells. Urol Int, 85(2): 221–227

[70] Liu Z, Martin L J (2003). Olfactory bulb core is a rich source of neural progenitor and stem cells in adult rodent and human. J Comp Neurol, 459(4): 368–391

[71] Lu H H, Thomopoulos S (2013). Functional attachment of soft tissues to bone: development, healing, and tissue engineering. Annu Rev Biomed Eng, 15(1): 201–226

[72] Lui P, Zhang P, Chan K, Qin L (2010). Biology and augmentation of tendon-bone insertion repair. J Orthop Surg, 5(1): 59

[73] Lyras D N, Kazakos K, Verettas D, Botaitis S, Agrogiannis G, Kokka A, Pitiakoudis M, Kotzakaris A (2009). The effect of platelet-rich plasma gel in the early phase of patellar tendon healing. Arch Orthop Trauma Surg, 129(11): 1577–1582

[74] Maeda T, Sakabe T, Sunaga A, Sakai K, Rivera A L, Keene D R, Sasaki T, Stavnezer E, Iannotti J, Schweitzer R, Ilic D, Baskaran H, Sakai T (2011). Conversion of mechanical force into TGF-b-mediated biochemical signals. Curr Biol, 21(11): 933–941

[75] Mendias C L, Gumucio J P, Bakhurin K I, Lynch E B, Brooks S V (2012). Physiological loading of tendons induces scleraxis expression in epitenon fibroblasts. J Orthop Res, 30(4): 606–612

[76] Miosge N, Klenczar C, Herken R, Willem M, Mayer U (1999). Organization of the myotendinous junction is dependent on the presence of alpha7beta1 integrin. Lab Invest, 79(12): 1591–1599

[77] Molloy T, Wang Y, Murrell G (2003). The roles of growth factors in tendon and ligament healing. Sports Med, 33(5): 381–394

[78] Murchison N D, Price B A, Conner D A, Keene D R, Olson E N, Tabin C J, Schweitzer R (2007). Regulation of tendon differentiation by scleraxis distinguishes force-transmitting tendons from muscle-anchoring tendons. Development, 134: 2697–2708

[79] Paxton J Z, Donnelly K, Keatch R P, Baar K (2009). Engineering the bone-ligament interface using polyethylene glycol diacrylate incorporated with hydroxyapatite. Tissue Eng Part A, 15(6): 1201–1209

[80] Paxton J Z, Grover L M, Baar K (2010). Engineering an in vitro model of a functional ligament from bone to bone. Tissue Eng Part A, 16(11): 3515–3525

[81] Perez A V, Perrine M, Brainard N, Vogel K G (2003). Scleraxis (Scx) directs lacZ expression in tendon of transgenic mice. Mech Dev, 120(10): 1153–1163

[82] Petersen J R, Agarwal S, Brownley R C, Loder S J, Ranganathan K, Cederna P S, Mishina Y, Wang S C, Levi B (2015). Direct mouse trauma/burn model for heterotopic ossification. J Vis Exp (102): 52880

[83] Petersen W, Fink C, Kopf S (2017). Return to sports after ACL reconstruction: a paradigm shift from time to function. Knee Surg Sports Traumatol Arthrosc, 25(5): 1353–1355

[84] Potten C S, Hendry J H (1975). Differential regeneration of intestinal proliferative cells and cryptogenic cells after irradiation. Int J Radiat Biol Relat Stud Phys Chem Med, 27(5): 413–424

[85] Pryce B A, Brent A E, Murchison N D, Tabin C J, Schweitzer R (2007). Generation of transgenic tendon reporters, ScxGFP and ScxAP, using regulatory elements of the scleraxis gene. Dev Dyn, 236(6): 1677–1682

[86] Pryce B A, Watson S S, Murchison N D, Staverosky J A, Dünker N, Schweitzer R (2009). Recruitment and maintenance of tendon progenitors by TGFbeta signaling are essential for tendon formation. Development, 136(8): 1351–1361

[87] Rees S G, Waggett A D, Kerr B C, Probert J, Gealy E C, Dent C M, Caterson B, Hughes C E (2009). Immunolocalisation and expression of keratocan in tendon. Osteoarthritis Cartilage, 17(2): 276–279

[88] Richardson S H, Starborg T, Lu Y, Humphries S M, Meadows R S, Kadler K E (2007). Tendon development requires regulation of cell condensation and cell shape via cadherin-11-mediated cell-cell junctions. Mol Cell Biol, 27(17): 6218–6228

[89] Rickert M, Jung M, Adiyaman M, Richter W, and Simank H G (2001). A growth and differentiation factor-5 (GDF-5)-coated suture stimulates tendon healing in an Achilles tendon model in rats. Growth Factors, 19(2): 115–126

[90] Rountree R B, Schoor M, Chen H, Marks M E, Harley V, Mishina Y, Kingsley D M (2004). BMP receptor signaling is required for postnatal maintenance of articular cartilage. PLoS Biol, 2(11): e355

[91] Rubio-Azpeitia E, Sánchez P, Delgado D, Andia I (2015). Three-dimensional platelet rich plasma hydrogel model to study early tendon healing. Cells Tissues Organs, 200(6): 394–404

[92] Runesson E, Ackermann P, Brisby H, Karlsson J, Eriksson B I (2013). Detection of slow-cycling and stem/progenitor cells in different regions of rat Achilles tendon: response to treadmill exercise. Knee Surg Sports Traumatol Arthrosc, 21(7): 1694–1703

[93] Runesson E, Ackermann P, Karlsson J, Eriksson B I (2015). Nucleostemin- and Oct 3/4-positive stem/progenitor cells exhibit disparate anatomical and temporal expression during rat Achilles tendon healing. BMC Musculoskelet Disord, 16(212): 1

[94] Sato T, Vries R G, Snippert H J, van de Wetering M, Barker N, Stange D E, van Es J H, Abo A, Kujala P, Peters P J, Clevers H (2009). Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature, 459(7244): 262–265

[95] Schwartz A G, Galatz L M, Thomopoulos S (2017). Enthesis regeneration: a role for Gli1+ progenitor cells. Development, 144(7): 1159–1164

[96] Schwartz Y, Viukov S, Krief S, Zelzer E (2016). Joint development involves a continuous influx of Gdf5-positive cells. Cell Reports, 15(12): 2577–2587

[97] Schweitzer R, Chyung J H, Murtaugh L C, Brent A E, Rosen V, Olson E N, Lassar A, Tabin C J (2001). Analysis of the tendon cell fate using Scleraxis, a specific marker for tendons and ligaments. Development, 128(19): 3855–3866

[98] Shah R R, Nerurkar N L, Wang C C, Galloway J L (2015). Tensile properties of craniofacial tendons in the mature and aged zebrafish. J Orthop Res, 33(6): 867–873

[99] Shih I M (1999). The role of CD146 (Mel-CAM) in biology and pathology. J Pathol, 189(1): 4–11

[100] Shukunami C, Takimoto A, Oro M, Hiraki Y (2006). Scleraxis positively regulates the expression of tenomodulin, a differentiation marker of tenocytes. Dev Biol, 298(1): 234–247

[101] Snippert H J, van der Flier L G, Sato T, van Es J H, van den Born M, Kroon-Veenboer C, Barker N, Klein A M, van Rheenen J, Simons B D, Clevers H (2010). Intestinal crypt homeostasis results from neutral competition between symmetrically dividing Lgr5 stem cells. Cell, 143(1): 134–144

[102] Soriano P (1999). Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat Genet, 21(1): 70–71

[103] Starborg T, Kalson N S, Lu Y, Mironov A, Cootes T F, Holmes D F, Kadler K E (2013). Using transmission electron microscopy and 3View to determine collagen fibril size and three-dimensional organization. Nat Protoc, 8(7): 1433–1448

[104] Staverosky J A, Pryce B A, Watson S S, Schweitzer R (2009). Tubulin polymerization-promoting protein family member 3, Tppp3, is a specific marker of the differentiating tendon sheath and synovial joints. Dev Dyn, 238(3): 685–692

[105] Subramanian A and Schilling T F (2014). Thrombospondin-4 controls matrix assembly during development and repair of myotendinous junctions. eLife, 3: e02372

[106] Subramanian A and Schilling T F (2015). Tendon development and musculoskeletal assembly: emerging roles for the extracellular matrix. Development, 142(24): 4191–4204

[107] Subramanian A, Wayburn B, Bunch T, Volk T (2007). Thrombospondin-mediated adhesion is essential for the formation of the myotendinous junction in Drosophila. Development, 134(7): 1269–1278

[108] Sugimoto Y, Takimoto A, Hiraki Y, Shukunami C (2013). Generation and characterization of ScxCre transgenic mice. Genesis, 51(4): 275–283

[109] Sundar S, Pendegrass C J, Blunn G W (2009). Tendon bone healing can be enhanced by demineralized bone matrix: a functional and histological study. J Biomed Mater Res B Appl Biomater, 88B(1): 115–122

[110] Tan Q, Lui P P Y, Lee Y W (2013). In vivo identity of tendon stem cells and the roles of stem cells in tendon healing. Stem Cells Dev, 22(23): 3128–3140

[111] Thomopoulos S, Williams G R, Gimbel J A, Favata M, Soslowsky L J (2003). Variation of biomechanical, structural, and compositional properties along the tendon to bone insertion site. J Orthop Res, 21(3): 413–419

[112] Tidball J G, Lin C (1989). Structural changes at the myogenic cell surface during the formation of myotendinous junctions. Cell Tissue Res, 257(1): 77–84

[113] Urdzikova L M, Sedlacek R, Suchy T, Amemori T, Ruzicka J, Lesny P, Havlas V, Sykova E, Jendelova P (2014). Human multipotent mesenchymal stem cells improve healing after collagenase tendon injury in the rat. Biomed Eng Online, 13(42): 1–15

[114] Veronesi F, Salamanna F, Tschon M, Maglio M, Nicoli Aldini N, Fini M (2017). Mesenchymal stem cells for tendon healing: what is on the horizon? J Tissue Eng Regen Med, 11(11): 3202–3219

[115] Wang Y, Zhang X, Huang H, Xia Y, Yao Y, Mak A F, Yung P S, Chan K M, Wang L, Zhang C, Huang Y, Mak K K (2017). Osteocalcin expressing cells from tendon sheaths in mice contribute to tendon repair by activating Hedgehog signaling. eLife, 6: e30474

[116] Watson S S, Riordan T J, Pryce B A, Schweitzer R (2009). Tendons and muscles of the mouse forelimb during embryonic development. Dev Dyn, 238(3): 693–700

[117] Wolfman N M, Hattersley G, Cox K, Celeste A J, Nelson R, Yamaji N, Dube J L, DiBlasio-Smith E, Nove J, Song J J, Wozney J M, Rosen V (1997). Ectopic induction of tendon and ligament in rats by growth and differentiation factors 5, 6, and 7, members of the TGF-beta gene family. J Clin Invest, 100(2): 321–330

[118] Wu Y, Wang Z, Ying Hsi Fuh J, San Wong Y, Wang W, San Thian E (2017). Direct E-jet printing of three-dimensional fibrous scaffold for tendon tissue engineering. J Biomed Mater Res B Appl Biomater, 105(3): 616–627

[119] Wu Y, Wong Y S, Fuh J Y H (2017). Degradation behaviors of geometric cues and mechanical properties in a 3D scaffold for tendon repair. J Biomed Mater Res A, 105(4): 1138–1149

[120] Yin H, Yan Z, Bauer R J, Peng J, Schieker M, Nerlich M, Docheva D (2018). Functionalized thermosensitive hydrogel combined with tendon stem/progenitor cells as injectable cell delivery carrier for tendon tissue engineering. Biomed Mater, 13(3): 034107

[121] Yoshimoto Y, Takimoto A, Watanabe H, Hiraki Y, Kondoh G, Shukunami C (2017). Scleraxis is required for maturation of tissue domains for proper integration of the musculoskeletal system. Sci Rep, 7: 1-16

[122] Zampeli F, Terzidis I, Espregueiera-Mendes J, Georgoulis J D, Bernard M, Pappas E, Georgoulis A D (2017). Restoring tibiofemoral alignment during ACL reconstruction results in better knee biomechanics. Knee Surg Sports Traumatol Arthrosc, 25(6): 1367-1374

[123] Zhang J and Wang J H C (2010). Characterization of differential properties of rabbit tendon stem cells and tenocytes. BMC Musculoskelet Disord, 11(10): 1

[124] Zhang Y, Kao W W Y, Hayashi Y, Zhang L, Call M, Dong F, Yuan Y, Zhang J, Wang Y C, Yuka O, Shiraishi A, Liu C Y (2017). Generation and characterization of a novel mouse line, Keratocan-rtTA (KeraRT), for corneal stroma and tendon research. Invest Ophthalmol Vis Sci, 58(11): 4800–4808

[125] Zheng G X Y, Terry J M, Belgrader P, Ryvkin P, Bent Z W, Wilson R, Ziraldo S B, Wheeler T D, McDermott G P, Zhu J, Gregory M T, Shuga J, Montesclaros L, Underwood J G, Masquelier D A, Nishimura S Y, Schnall-Levin M, Wyatt P W, Hindson C M, Bharadwaj R, Wong A, Ness K D, Beppu L W, Deeg H J, McFarland C, Loeb K R, Valente W J, Ericson N G, Stevens E A, Radich J P, Mikkelsen T S, Hindson B J, Bielas J H (2017). Massively parallel digital transcriptional profiling of single cells. Nat Commun, 8: 1-12

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag GmbH, part of Springer Nature