Reticulocalbin 3 is involved in postnatal tendon development by regulating collagen fibrillogenesis and cellular maturation

Tendon plays a critical role in the joint movement by transmitting force from muscle to bone. This transmission of force is facilitated by its specialized structure, which consists of highly aligned extracellular matrix consisting predominantly of type I collagen. Tenocytes, fibroblast-like tendon cells residing between the parallel collagen fibers, regulate this specialized tendon matrix. Despite the importance of collagen structure and tenocyte function, the biological mechanisms regulating fibrillogenesis and tenocyte maturation are not well understood. Here we examine the function of Reticulocalbin 3 (Rcn3) in collagen fibrillogenesis and tenocyte maturation during postnatal tendon development using a genetic mouse model. Loss of Rcn3 in tendon caused decreased tendon thickness, abnormal tendon cell maturation, and decreased mechanical properties. Interestingly, Rcn3 deficient mice exhibited a smaller collagen fibril distribution and over-hydroxylation in C-telopeptide cross-linking lysine from α1(1) chain. Additionally, the proline 3-hydroxylation sites in type I collagen were also over-hydroxylated in Rcn3 deficient mice. Our data collectively suggest that Rcn3 is a pivotal regulator of collagen fibrillogenesis and tenocyte maturation during postnatal tendon development.

www.nature.com/scientificreports/ stability to the collagen triple helix through hydrogen bonds and water bridges 20,21 . Several tissue-specific prolyl 3-hydroxylation modification sites have been identified in tendon type I collagen, including α1(I) Pro986, α1(I) Pro707, α2(I) Pro707, and a C-terminal (GPP)n motif that are mainly modified by P3H2 and α1(I) Pro986 that is modified by P3H1 22 . The total prolyl 3-hydroxylation content in the adult Achilles tendon collagen was higher than that in fetal Achilles tendon collagen 23 . Several studies suggested that prolyl 3-hydroxylation is involved in collagen fibril assembly 22,24,25 . Reticulocalbin 3 (Rcn3), a 45 kDa ER lumen protein, is a member of the CREC (Cab45/reticulocalbin/ERC45/ calumenin) family of multiple EF-hand Ca 2+ -binding proteins, including Rcn1, ERC-55 (Rcn2), Rcn3, Cab45, and calumenin 26 . Proteomic studies have suggested that Rcn3 functions as a chaperone protein in the secretory pathway and has emerged as a new potential regulator of collagen production 27,28 . Recent mouse genetic studies have demonstrated that Rcn3 plays a critical role in perinatal lung maturation by regulating synthesis and secretion of surfactant and is involved in pulmonary injury remodeling via regulation of alveolar epithelial cell (AEC) apoptosis and ER stress 29,30 . Despite this emerging role of Rcn3 in the secretory pathway, the physiological function of Rcn3 in various tissues producing a large amount of extracellular matrix is not yet fully established.
In this study, we investigated the physiological function of Rcn3 in tendon by conditionally deleting Rcn3 in mouse tendon cells. Loss of Rcn3 caused impaired postnatal tendon maturation with markedly decreased tendon thickness. Consistently, structural properties were reduced in Rcn3 deficient mice, but material properties were mostly unchanged. We further found that Rcn3 deficiency resulted in impaired collagen fibrillogenesis with over-modification of type 1 collagen and reduced fibril diameter distribution. Collectively, these findings, for the first time, establish the biological function of Rcn3 in postnatal tendon development and provide an exciting mouse genetic model for studying collagen fibrillogenesis and modification in tendon.

Results
Loss of Rcn3 caused decreased tendon thickness. The endogenous Scleraxis (Scx) begins to be expressed in the lateral plate mesoderm and the body wall from E9.5, and its expression is restricted to the developing tendons at E14.5 31 . Besides, Scx is expressed in tenocytes and directly regulates the expression of type I collagen (Col1a1) that is highly expressed in differentiated tenocytes [32][33][34] . Therefore, to understand the function of Rcn3 in tendons, we generated Scx-Cre; Rcn3 fl/fl mice in which Rcn3 was specifically deleted in Scx-expressing cells that are abundant in tendons 35,36 . The growth of Scx-Cre; Rcn3 fl/fl mice was normal compared with wildtype mice, which was evident with normal body weight (Fig. 1A). To access the knockout efficiency of Rcn3 in tendons, we performed immunohistochemistry analysis against Rcn3 (Fig. 1B). Rcn3 was highly expressed in patellar and Achilles tendons in wild-type mice at the early postnatal stage, including P5, P21, and P30, but its expression was decreased at a later stage, P60 (Fig. 1B). Immunohistochemistry analysis demonstrated that the deletion of Rcn3 in Scx-Cre; Rcn3 fl/fl tendons were efficient in both patellar and Achilles tendons throughout the postnatal maturation (Fig. 1B). In the patellar tendon, we observed residual Rcn3 expression at P5 and P21, but the expression was not detected at P30 and P60. In the Achilles tendon, we found residual expression till P30,  www.nature.com/scientificreports/ but the expression was not detected at P60. These data suggest that the deletion of Rcn3 is efficient in the patellar and Achilles tendons. We performed histological analyses on patellar and Achilles tendons to assess the tendon phenotypes in Scx-Cre; Rcn3 fl/fl mice. Although Scx-Cre; Rcn3 fl/fl mice had no difference in body weight compared to wild-type mice (Fig. 1A), Scx-Cre; Rcn3 fl/f mice clearly displayed decreased tendon thickness at P30 compared to wild-type littermate in both patellar ( Fig. 2A) and Achilles tendon (Fig. 2C). Our quantification results confirmed the 40% reduced thickness in the patellar tendon and 20% reduced thickness in the Achilles tendon (Fig. 2B,D). We also found that the cellularity of tendons in Scx-Cre; Rcn3 fl/fl mice was higher than wild-type mice, with threefold and 1.5-fold increases in the patellar and Achilles tendons, respectively (Fig. 2B,D). The increased cell density without the change in total cell number (Fig. 2B,D) prompted us to investigate cellular proliferation in Scx-Cre; Rcn3 fl/fl mice. We found no differences in cell proliferation of Scx-Cre; Rcn3 fl/fl mice (Fig. 2E,F), suggesting that reduced matrix formation likely caused increased cell density. These results collectively suggest that Rcn3 is critical for postnatal tendon development for both patellar and Achilles tendons.
Loss of Rcn3 reduced collagen fibril diameter. The decreased tendon thickness prompted us to examine collagen fibril ultrastructure because collagen fibrillogenesis is critical for tendon growth. We performed transmission electron microscopy (TEM) using patellar and Achilles tendons at 2 months of age to measure collagen fibril diameter. Scx-Cre; Rcn3 fl/fl mice exhibited decreased fibril diameter in both patellar (Fig. 3A) and Achilles tendons (Fig. 3C) compared with wild-type mice. Specifically, quantification results showed that the distribution of collagen fibril diameter in patellar tendon ranged from 20 to 210 nm in wild-type mice, whereas Scx-Cre; Rcn3 fl/fl mice showed a narrower collagen fibril distribution (10-170 nm) in the patellar tendon lacking the larger fibrils found in WT mice (Fig. 3B). In the patella tendon, control mice had the peak number of collagen fibril diameter at 170 nm, and Scx-Cre; Rcn3 fl/fl mice had the peak number at 90 nm. Consistent with the patellar tendon, the Achilles tendon also showed the same trend with a shifted distribution in the mutants lacking the larger fibrils (i.e., 190-260 nm) found in WT mice (Fig. 3D). In the Achilles tendon, control mice had the These results suggest that Rcn3 is required for proper collagen fibrillogenesis, and decreased tendon thickness could be due to the smaller diameter of collagen fibril in Scx-Cre; Rcn3 fl/fl mice.

Loss of Rcn3 increased collagen modification. Collagen post-translational modifications and cross-
linking analysis were investigated in Scx-Cre; Rcn3 fl/fl mice at 1 month of age. The modification state of several key residues was examined using peptide mass spectrometry. Notable findings include the C-telopeptide crosslinking lysine from α1(I) chain, which was slightly over-hydroxylated in Scx-Cre; Rcn3 fl/fl patellar (20% more hydroxylated) and Achilles (9% more hydroxylated) tendons compared to wild-type mice (Table 1A). These data support that partial telopeptide lysine hydroxylation in tendon is being shifted to fuller hydroxylation in Scx-Cre; Rcn3 fl/fl mice. This is a useful mark of altered collagen cross-linking and is consistent with an increase in pyridinoline cross-links. Indeed, the cross-linking analysis revealed an increase in hydroxylysylpyridinoline in both tendons of the Scx-Cre; Rcn3 fl/fl mice. It should be noted that while hydroxylysylpyridinoline provides a useful measure of stable (irreversible) cross-links, it is not a measure of the total cross-links (Table 1B). The pattern of collagen over-modification was also observed across the sites of proline 3-hydroxylation in type I collagen in Scx-Cre; Rcn3 fl/fl Achilles and patellar tendons (Table 1). A function has yet to be discovered for this rare modification, but it has been suggested that 3-hydrxyproline can regulate the postnatal lateral fusion of smaller fibrils in developing tendons 23 .

Loss of Rcn3 caused abnormal tendon cell morphology.
To investigate the cellular phenotype caused by the deletion of Rcn3, we analyzed tendon cell morphology using confocal microscopy after Zo1 and phalloidin double staining (Fig. 4A). Tendon cell areas became largest at P30 and dramatically decreased at P60 in wild-type mice (Fig. 4C, white bar). Interestingly, tendon cells in Scx-Cre; Rcn3 fl/fl mice exhibited increased cell area at P10 and P21 but significantly decreased cell areas at P30 compared to wild-type mice (Fig. 4C). We also found changes in protrusion number in tendon cells of Scx-Cre; Rcn3 fl/fl mice. These protrusions from the cell body are critical for tendon cell interaction with adjacent cells and the ECM. The protrusion number, in these transverse sections, decreased during postnatal tendon maturation in wild-type mice (Fig. 4D, white bar). The Scx-Cre; Rcn3 fl/fl mice showed a significantly increased protrusion number compared with wild-type mice at all stages of postnatal tendon development (Fig. 4D). Collectively, these results suggest that the function of Rcn3 is indispensable for the morphological maturation of tendon cells.
Loss of Rcn3 increased the gene expression related to tenocyte differentiation. To further understand the cellular phenotype at the molecular level, we performed qRT-PCR using Achilles tendon from wild-type mice and Scx-Cre; Rcn3 fl/fl mice. First, we confirmed the efficient deletion of Rcn3 in the Achilles tendon of Scx-Cre; Rcn3 fl/fl mice (Fig. 5). The expression of tenogenic markers, such as Scx, Mkx, Col1a1, and Tnmd, were significantly increased in Scx-Cre; Rcn3 fl/fl mice compared with wild-type mice (Fig. 5). These results suggest that the loss of Rcn3 enhanced tendon cell differentiation or delayed full tenocyte maturation. tendons, we performed uniaxial biomechanical testing of the Achilles tendon at 2 months of age. Consistent with our histological analysis, cross-sectional area (CSA) was significantly reduced in Scx-Cre; Rcn3 fl/fl tendons (Fig. 6A). Concomitantly, stiffness and failure load were also significantly lower, suggesting a loss of structural capability in the Scx-Cre; Rcn3 fl/fl mice (Fig. 6B,C). No difference was observed in percent stress relaxation between the two groups, suggesting that Rcn3 does not perturb the viscoelastic response of the Achilles tendon (Fig. 6D). In addition, no differences were observed in failure stress or tissue elastic modulus, which indicates that loss of Rcn3 did not cause an inherent material change in the tendon tissue (Fig. 6E,F). These data demonstrate that reduced mechanical properties caused by loss of Rcn3 are due to decreased structural properties but not material properties.

Discussion
We investigated the function of Rcn3 in postnatal tendon development by characterizing a tendon-specific Rcn3 knockout mouse (Scx-Cre; Rcn3 fl/fl ) model. The Scx-Cre; Rcn3 fl/fl mice exhibited impaired postnatal tendon development, including decreased tendon thickness, increased cellularity, impaired tenocyte maturation, and reduced collagen fibril diameter with altered collagen modification. Our study identified the novel function of Rcn3 in tendon and provides genetic approaches that can be used to investigate the underlying mechanisms regulating cellular and matrix maturation in tendon. Rcn3 is known as an ER protein regulating the secretion of proteins, but its critical molecular function is not clear. Our study showed that Rcn3 deficiency caused impaired collagen fibrillogenesis with over-modification of collagen. These results suggest that Rcn3 is a critical regulator of collagen fibrillogenesis via regulation of collagen modification. However, the precise molecular mechanism by which Rcn3 regulates collagen modification is still not clear. One possibility is that Rcn3, as a chaperon protein, stabilizes collagen for proper modification. Previous studies also suggested that Rcn3 functions as a chaperon protein 27,28 . The other possibility is that Rcn3 mediates the binding of collagen modification enzymes to collagen molecules. Further molecular analyses will be required to elucidate the precise function of Rcn3 in collagen modification in tendon.
To measure morphological changes of tendon cells at a single-cell level, we developed a novel confocal imaging method (Fig. 4). We found that the protrusion number was decreased throughout the postnatal tendon development in wild-type mice. Interestingly, this result is different from the current finding using serial block face-scanning electron microscopy (SBF-SEM) 12 . Kalson et al. showed that a similar number of protrusions contacting adjacent cells between the tendons of newborn and 6-week-old mice. There are several variations between our and Kalson's study. First, we have analyzed the patellar tendon, but they analyzed the tail tendon. Table 1. Collagen cross-link and mass spectral analysis to investigate the hydroxylation levels (A) of lysine and proline in type 1 collagen (n = 1 for patellar and n = 2 for Achilles tendon) and the collagen cross-linking levels (B) (* indicates P < 0.05 between genotypes, n = 1 for patellar and n = 3 for Achilles tendon) from wildtype mice and Scx-Cre; Rcn3 fl/fl (tendon-specific Rcn3 loss-of-function model) littermates at P30.  www.nature.com/scientificreports/ Second, the age is different. We have analyzed age from P10 to P60. They analyzed newborn and 6-week-old mice. These age and tendon types could cause different results of tendon phenotype in each study. Further studies analyzing various tendon types and ages using our confocal imaging technique and SBF-SEM will be necessary to fully understand the morphological maturation of tenocytes. During normal tendon development, the average cross-sectional cell area became largest at P30 in wildtype (Fig. 4). However, Scx-Cre; Rcn3 fl/fl mice showed a prematurely increased cell area with increased cell area at P10 and P21, while significantly decreased cell area at P30 compared to wildtype. Interestingly, the protrusion numbers in mutant mice are significantly increased at all stages, suggesting the delay of tenocyte maturation. These cellular changes correspond with the molecular analysis (Fig. 5) that shows increased expression of early tenogenic markers in Rcn3 deficient mice. It is also consistent with the immature growth of collagen fibril. Based on these data, we carefully suggest that Rcn3 positively regulates the maturation of differentiated tenocytes.  www.nature.com/scientificreports/ Therefore, deletion of Rcn3 caused the delay of full maturation of tenocytes which cause increased protrusion number, increase expression of differentiated tenocyte genes, and immature growth of collagen fibril diameter. Investigating the mechanism by which Rcn3 regulates tenocyte maturation will be an interesting future direction of this study. Scx-Cre; Rcn3 fl/fl mice displayed an increase in post-translational modifications and cross-linking in type I collagen. Specifically, Scx-Cre; Rcn3 fl/fl mice showed an increase in telopeptide lysine hydroxylation as well as hydroxylysylpyridinoline in both patellar and Achilles tendons. Hydroxylysylpyridinoline is a stable collagen cross-link, typically found in adult bone and tendon, that is known to increase in concentration as these tissues mature. Besides, Scx-Cre; Rcn3 fl/fl mice exhibited over-hydroxylated proline 3-hydroxylation in type I collagen. The function of this rare modification is not clear, but it has been suggested that 3-hydrxyproline can regulate the postnatal lateral fusion of smaller fibrils in developing tendons 23 . Based on these collagen modification data, we carefully suggest that the reduced collagen fibril diameter in Scx-Cre; Rcn3 fl/fl mice is due to the inhibition of lateral fusion caused by over-modification on type I collagen.
In conclusion, our results uncover the novel function of Rcn3 in postnatal tendon development. This study will be the basis of future mechanistic studies for tenocyte maturation, collagen modification, and fibrillogenesis by providing useful mouse genetic models.

Methods
Animals. All studies were approved by the Institutional Animal Care and Use Committee (IACUC) and University Laboratory Animal Resource (ULAR) at the University of Pennsylvania (Philadelphia, Pennsylvania, USA) and we confirm that all the methods in current studies were performed in accordance with the relevant guidelines and regulations following the IACUC approved protocol. The study was carried out in compliance with the ARRIVE guidelines 37 . The Scx-Cre mouse line was previously described 38 . Mouse sperms carrying Knockout-first Rcn3 allele (Rcn3 tm1a(EUCOMM)Hmgu ) were obtained from the Knockout Mouse Programme (KOMP), and mouse carrying Rcn3 tm1a(EUCOMM)Hmgu was generated by in vitro fertilization at BaylorCollege of Medicine. Rcn3 tm1a(EUCOMM)Hmgu mice were crossed with Rosa26-Flippase (Flp) mice to delete β-gal and the neo cassette in order to generate the Rcn3 conditional knockout allele (Rcn3 fl/fl ).
Histological analyses. Mouse hindlimbs were collected from specific stages and fixed in 10% neutral buffered formalin overnight at 4 °C. Samples were paraffin-embedded and sectioned at 6 μm, following decalcification in 10% (w/v) EDTA (pH 7.4) for 2 weeks with daily solution changes. The paraffin sections of the patellar and Achilles tendons were used for Hematoxylin and Eosin (H&E). Histological evaluation was blindly performed by 2 independent individuals.
BrdU assay. For BrdU labeling in tendon, a single stock solution of 10 mg/mL BrdU (Sigma, B5002) and 1.2 mg/mL FdU (Sigma, F0503) was prepared in 1X PBS. BrdU/FdU stock (0.01 ml/g of body weight) was injected into the intraperitoneal space of mice at P8 and P9. And then, mouse hindlimbs were collected at P10 and fixed in 10% neutral buffered formalin overnight at 4 °C. Samples were paraffin-embedded and sectioned at 6 μm, following decalcification in 10% (w/v) EDTA (pH 7.4) for 1 week with daily solution changes. BrdU Immunohistochemistry Kit (Abcam, ab125306) was used to detect BrdU incorporation into the cells according to the manufacturer's instructions.  Collagen cross-linking analysis. Collagen pyridinoline cross-link content was determined by fluorescence monitoring with reverse-phase HPLC. Pyridinoline cross-links were analyzed in mouse tendons by HPLC after acid hydrolysis in 6 M HCl for 24 h at 108 °C. Dried samples were dissolved in 1% (v/v) n-heptafluorobutyric acid for quantitation of hydroxylysyl pyridinoline (HP) by reverse-phase HPLC and fluorescence monitoring as previously described 39 .
Mass spectrometry. Collagen α-chains were cut from SDS-PAGE gels and subjected to in-gel trypsin digestion as previously described 24 . Electrospray mass spectrometry was carried out on the trypsin-digested peptides using an LTQ XL linear quadrupole ion-trap mass spectrometer equipped with in-line Accela 1250 liquid chromatography and automated sample injection (ThermoFisher Scientific). Proteome Discoverer software (ThermoFisher Scientific) was used for peptide identification. Tryptic peptides were also identified manually by calculating the possible MS/MS ions and matching these to the actual MS/MS spectrum using Thermo Xcalibur software. Differences in post-translational modifications were determined manually by averaging the full scan MS over several LCMS minutes to include all the post-translational variations of a given peptide. Protein sequences used for MS analysis were obtained from the Ensembl genome database.
Immunofluorescent staining. For immunofluorescent staining, mouse hindlimbs were collected during postnatal tendon maturation from P10 to P60 and fixed in 4% PFA overnight at 4 °C. Samples were OCT compound embedded and sectioned at 10 μm, following decalcification in 10% (w/v) EDTA (pH 7.4) for 10 days with daily solution changes. The frozen sections of the patellar tendon blocked with blocking solution (Thermo, 8120) for 1 h RT, followed by incubation at 4 °C overnight with antibodies against ZO1 (Fisher, 61-7300) and then incubated for 1 h at RT with Phalloidin Alexa Fluor 488 conjugate. Finally, sections were mounted in antifade solution with DAPI (Invitrogen, P36935).
qRT-PCR. RNA was extracted with Trizol and Direct-zol kit (Zymo Research, R2060) from primary tendon cells. And cDNA was synthesized from 1 μg RNA using iScript Reverse Transcription (Bio-rad, 1708841). qRT-PCR was performed by QuantStudio 6 Flex real-time PCR system (ThermoFisher Scientific) using Fast SYBR Green Master Mix (Pec, 4385612). The sequences of primers are listed in Table 2.
Uniaxial biomechanical test. All mice assigned for mechanical testing were frozen at − 20 °C untill the day of testing. Mice were thawed at room temperature, and calcaneal bone-Achilles tendon-muscle complexes were grossly dissected. Under microscope guidance, all extraneous soft tissues and muscles were finely  R: ACC TGC ACG TAG CCA TCT TT   mScx  F: AAG ACG GCG ATT CGA AGT TAG AAG   R: TCT CTC TGT TCA TAG GCC CTG CTC ATA G   mMkx  F: GAT GGC GAC TCC TGC TCT GA   R: CGG TCT GCC GCC AGC TTT TA   mCol1a1  F: TTG  www.nature.com/scientificreports/