The healing after rotator cuff surgery is still dissatisfying, and increased muscle fatty infiltration even more impairs the healing success. To achieve sufficient healing after rotator cuff reconstructions, the use of growth factors may be one possibility. The aim of the study was to identify a possible relationship between fatty infiltration of the supraspinatus muscle and cellular biological characteristics and stimulation potential of tenocyte-like cells (TLCs). TLCs of 3 donor groups differing in grade of muscle fatty infiltration were analyzed for their cellular characteristics and were stimulated with BMP-2 or BMP-7 in a 3D scaffold culture. The cell count and potency for self-renewal were significantly decreased in TLCs from donors with high muscle fatty infiltration compared to the lower fatty infiltration groups. Cell count and collagen-I expression as well as protein synthesis were stimulated by growth factors. Interestingly, TLCs of the high fatty infiltration group exhibited a weaker stimulation potential compared to the other groups. TLCs from donors with high muscle fatty infiltration generally revealed inferior characteristics compared to cells of lower fatty infiltration groups, which may be one reason for a weaker healing potential and may represent a possible starting point for the development of future treatment options. © 2013 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 32:129–137, 2014.
The amount of fatty infiltration in rotator cuff muscles is a predictor of poor outcome by means of increased incidence of re-tears or non-healing.[1-4] The quantity of fatty infiltration was graded by Goutallier et al. into five grades with the ratio of muscle to fat (Table 1) according MRI scans. A Goutallier grade >1 was shown to be the cut off between tendon integrity and recurrent defects. Muscle fatty infiltration may occur due to a reduced mechanical stimulation of the rotator cuff muscle as a result of pain and therefore limited active movement. A relationship between cellular biological characteristics of rotator cuff tenocytes and the grade of muscle fatty infiltration has not been proven so far. Previously we showed that the age and sex of patients can influence the properties of tenocytes, while cells of aged as well as female donors had inferior characteristics.[7, 8] This may explain the weaker healing potential in these patient groups as demonstrated by several clinical or radiographic follow-up studies.[3, 9-11] In order to improve the clinical outcome after rotator cuff surgery, biological approaches have been proposed to additionally increase tendon cell activity. Growth factors such as bone morphogenetic protein-2 (BMP-2) or BMP-7, which are already clinically used for bone repair, were shown to positively influence tendon or ligament cell cultures.[7, 8, 13-15] Furthermore, animal studies revealed a positive effect of BMP-2 and BMP-7 on tendon integration, by means of an increased biomechanical strength at the tendon-bone insertion site.[16-19] This study aims to analyze the stimulation potential of rotator cuff tenocyte-like cells (TLCs) of patient groups with varying grades of muscle fatty infiltration as a precondition for the development of a possible future treatment option.
|Grade||Ratio of Muscle to Fat (Based on Computed Tomography Scan)|
|0||Completely normal muscle|
|1||Some fatty streaks in the muscle|
|2||Amount of muscle is greater than fatty infiltration|
|3||Amount of muscle is equal to fatty infiltration|
|4||Amount of fatty infiltration is greater than muscle|
MATERIAL AND METHOD
TLCs were isolated from Supraspinatus (SSP) tendon biopsies from patients undergoing arthroscopic or open shoulder surgery. Biopsies were obtained 3–5 mm from the torn proximal tendon edge. All patients gave their written informed consent and the local ethics committee authorized the anonymous use of tendon samples, which would otherwise be discarded (Ethic number: EA1/060/09).
Biopsies were taken from male donors younger than 65 years of age. Donors were grouped by fatty infiltration of the SSP muscle according to Goutallier et al. as follows: grade 0 or 1: low fatty infiltration (n = 6, mean age 45.3), grade 2: middle fatty infiltration (n = 6, mean age 59.7), grade 3 or 4: high fatty infiltration (n = 6, mean age 55.7). The data of the low-fatty infiltration group were published earlier as young male group for analysis of age-related differences.
Cellular Biological Characteristics
The analysis of cellular biological characteristics and the growth factor stimulation was performed in direct comparison to previously published work for age-related difference in TLCs of male and female donors.[7, 8]
Cell Isolation and Analysis of Cell Density
SSP tendon biopsies were weighted under sterile conditions. TLCs were isolated by a 2 h digestion with 0.3% collagenase type CLS II (5 ml collagenase solution per 20 mg tendon). Digested tendon material was plated onto a culture flask with a ratio of tendon material (mg)/growth area (cm2) from 0.2 to 0.3. The cells were cultured in DMEM/Ham's F12 with 10% FCS and 1% penicillin/streptomycin (all Biochrom AG, Berlin, Germany) at 37°C, under standard conditions. After 7 days of culture, metabolic activity of the cells was analyzed by Alamar Blue assay (Biozol, Eching, Germany). Using a standard curve with defined cell numbers the cell count was calculated. The approximate cell density was calculated by normalization of the cell count to the weight of the tendon biopsy.
Analysis of Cell Count Over 14 Days
TLCs at passage 2 were seeded with 2.5 × 103 vital cells per well in a 48-well plate, and in triplicates. At day 1, 4, 7, and 14 after seeding, cell activity was analyzed by Alamar Blue assay and cell count was calculated using a standard curve method. The cell count of day 4, 7, and 14 was referred to cell count of day 1 by subtraction.
Gene Expression Analysis
At passage 2, RNA was isolated from the cells with the NucleoSpin RNA II Kit (Macherey-Nagel, Düren, Germany). cDNA was synthesized from 100 ng of RNA using the qScript cDNA supermix (Quanta BioSciences, Gaithersburg, MD) and an Epgradient Mastercycler (Eppendorf, Hamburg, Germany). Cells were characterized by analyzing gene expression of collagen-I (Col-I), Col-II, Col-III, decorin, osteocalcin, scleraxis, tenomodulin, and mohawk relative to the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (primer sequences see Table S1). Real-Time PCR (qRT-PCR) was performed in triplicates with the following components per well: 1.25 ng cDNA, 12.5 µl Sybr Green Supermix (Quanta BioSciences), 1 µl primermix (10 µM), and 6.5 µl RNase/DNase-free water. The following amplification protocol was used with the Realplex Mastercycler (Eppendorf): initial denaturation at 95°C, amplification was repeated for 40 cycles (95°C for 15 s, annealing temperature for 45 s, and 72°C for 30 s) and finished with a melting curve. The relative gene expression levels were calculated with the method from mean values of triplicates for each gene and donor (ΔCt = mean cycle threshold [Ct] of target gene − mean Ct of housekeeping gene).
Analysis of Col-I Protein Synthesis
The MicroVue C1CP EIA Kit (TecoMedical, Bünde, Germany) was used to determine the Col-I protein synthesis from cell culture supernatant of day 14 of cell count analysis. The level of Col-I protein synthesis was normalized to total protein content, which was analyzed with the Coomassie Plus™ protein assay (Thermo Fisher Scientific, Dreieich, Germany).
Analysis of Stem Cell Phenotype
For analyzing the potential for self-renewal, 1,000 TLCs in passage 2 were cultured for 11 days in a 10 cm petri dish in triplicates with normal growth medium and a medium change 3 times a week. For quantification, colonies were stained with 1% methylene blue in boratbuffer/1% azure in dH2O (1:1, Sigma–Aldrich) for 10 min. Number and average size of colonies (range: 1–10 mm2) were analyzed using an image analyzing system with an adaptive threshold (ImageJ 1.44i, Wayne Rasband, National Institute of Health, Bethesda, MD).
Stem cell phenotype was analyzed by FACS analysis according to an established stem cell panel from the core unit of the Berlin-Brandenburg Center for Regenerative Therapies. A total of 2.5–5 × 105 vital cells in passage 1 were stained with a Live/Dead reagent and antibodies against CD29, 44, 73, 90, 105 and a negative mix consisting of CD11b, 14, 19, 34, and 45 for 25 min at 4°C (details see Table S2). Cells were fixed with 1% PFA and measured with the BD FACS Canto II system (BD Biosciences, Heidelberg, Germany) and FACS Diva software. Data were analyzed using FlowJo 8.8.6 software. Isotype controls and unstained cells served as controls.
Multipotent differentiation was analyzed in TLCs at passage 2. For adipogenic and osteogenic differentiation, cells were cultured until confluence in a 24-well plate. For chondrogenic differentiation, 2.5 × 105 pelleted cells were differentiated in a 15 ml Falcon tube. Cells were incubated in differentiation medium or normal growth medium (control) for 3 weeks with a medium change every 3–4 days. Multipotent differentiation and stainings were performed in a modified version as previously described (Table S3).[20, 21] To validate the results for multipotent differentiation, lineage specific markers were analyzed by qRT-PCR for n = 2 cultures per group. For osteogenic and adipogenic differentiation, RNA was directly isolated from the 24-well plates with the NucleoSpin RNA II Kit after 3 weeks. Chondrogenic differentiated cell pellets (2 per donor) were pooled after 3 weeks of culture and homogenized using peqGOLD TriFast (Peqlab, Erlangen, Germany) with the Precellys system (251.4 mm and 32.8 mm ceramic pellets, Peqlab) at 5,000 rpm for 3 times 30 s. RNA was extracted to the aqueous phase using chloroform and diluted 1:1 with 75% ethanol. RNA was then purified using NucleoSpin RNA II kit according to the manufacturer's manual. cDNA synthesis and qRT-PCR was performed as described above. Details for the primers are listed in Table S1. Relative expression levels were normalized to ribosomal protein L13 (RPL13) ( method).
Application of Growth Factors
A total of 2 × 104 vital cells were seeded with a drop-on method into a macro-porous collagen scaffold, consisting of highly oriented porcine Col-I (6 mm diameter × 3 mm height, Matricel, Herzogenrath, Germany). Twenty-four hours before growth factor stimulation, cells were starved with medium without FCS. TLCs were then stimulated with 0.2 µg/ml or 1 µg/ml rhBMP-2 (Wyeth, New York), or rhBMP-7 (R&D Systems, Wiesbaden, Germany) in DMEM/HAM's F12 (1:1) supplemented with 5% FCS and 1% penicillin/streptomycin. The cell count was analyzed by Alamar Blue assay, using a standard curve with defined cell number, before stimulation (day 0), and at days 3, 5, and 7. Gene expression and Col-I protein synthesis was analyzed only at the endpoint of stimulation (day 7) as described above. Relative gene expression levels of Col-I, -II, -III and osteocalcin were normalized to GAPDH and to the untreated control using the method (ΔΔCt = ΔCt of stimulation − ΔCt unstimulated control).
Statistics were performed for n = 18 values (N = 6 donors per group in triplicates) for each donor group. For qRT-PCR analysis, RNA of triplicates was pooled (n = 6). All box plots and values in the text are given as median with 25 and 75 percentiles, except for Figure 1, which represent means ± SD. Statistical analysis was done using SPSS 20 (IBM, Armonk, NY). Significant differences were analyzed with the Mann–Whitney U-test to compare the donor groups with each other or the stimulation groups with the control. The level of significance was set at p < 0.05 and adjusted with the Bonferroni–Holm correction. For the analysis of stimulation potential an additional level of significance was investigated to indicate high significant values (p ≤ 0.001).
The tendon retraction according to Patte showed a linear correlation to the grade of muscle fatty infiltration. The tear size according to Bayne and Bateman was equal between low and middle graded fatty infiltration, but increased with grade 3–4 muscle fatty infiltration (Fig. 1).
Cellular Biological Characteristics
The approximate cell density was 5,902 (3,232–7,905) cells per mg tendon biopsy in the low, 7,895 (2,822–17,194) cells per mg biopsy in the middle, and 2,803 (1,894–8,434) cells per mg in the high fatty infiltration group without significant differences between all groups. When culturing the cells over 14 days, a significantly decreased cell count was found in TLCs of the high fatty infiltration group compared to the low- and middle fatty infiltration group (Fig. 2A). TLCs expressed high amounts of Col-I, Col-III and decorin, no Col-II and low amounts of osteocalcin. Cells of the low fatty infiltration group showed highest Col-I, Col-III, and osteocalcin expression, which was significantly different to the other groups after Mann–Whitney U-test, but not Bonferroni–Holm correction. They expressed scleraxis and mohawk, but almost no tenomodulin. Mohawk expression was significantly decreased in the high fatty infiltration group compared to both other groups, but not after Bonferroni–Holm correction (Table 2). The Col-I protein synthesis was not affected by the grade of muscle fatty infiltration (data not shown). The TLCs showed stem cell like properties. They were able to form colonies, while the amount of colonies was significantly reduced in the high fatty infiltration group compared to both other groups. In addition, the colony area was significantly decreased in the high muscle fatty infiltration group compared to the middle fatty infiltration group (Fig. 2B). The stem cell markers CD29, 44, 73, 90, and 105 were found on >95% of the viable cells in all groups. Less than 2% of the viable TLCs expressed the negative markers CD11b, 14, 19, 34, and 45 (Table S2, Fig. 3). A portion of cells were able to differentiate into multiple directions (Fig. 4). The quantitative analysis of the Oil Red O staining revealed that adipogenic differentiation was significantly increased in the middle fatty infiltration group compared to both other groups. Similarly, the osteogenic differentiation was strongest in the middle fatty infiltration group with significant differences compared to the low fatty infiltration group, analyzed by solubilized Alizarin Red S staining (Fig. 2C). The TLCs of all donor groups showed a chondrogenic differentiation with a positive Alcian Blue staining of the cell pellets compared to the untreated control pellets (Fig. 4). qRT-PCR analysis revealed that lineage specific markers were highly upregulated in the differentiated cells compared to the controls. For adipogenic differentiation PPARγ (peroxisome proliferator-activated receptor γ), LPL (lipoprotein lipase), and FABP (fatty acid binding protein) were strongly increased. Increased Runx2 (runt-related transcription factor 2), and ALPL (alkaline phosphatase tissue-nonspecific isozyme) expression underlined osteogenic differentiation and chondrogenic differentiation was proved by upregulated Col-II, COMP (cartilage oligomeric matrix protein) and aggrecan expression (Fig. 4).
|Relative Gene Expression ||Muscle Fatty Infiltration|
|Col-I||2.7 (2.4–3.2)a||1.6 (1.3–2.1)||2.1 (1.7–2.7)|
|Col-III||0.5 (0.4–0.6)b||0.2 (0.2–0.4)||0.3 (0.3–0.4)|
|Osteocalcin [10−4]||9.0 (5.3–12.5)b||4.3 (3.9–6.3)||4.7 (4.2–5.0)|
|Decorin||0.1 (0.1–0.3)||0.3 (0.1–0.3)||0.2 (0.1–0.2)|
|Scleraxis [10−3]||1.6 (1.0–3.4)||1.4 (0.7–2.0)||0.9 (0.3–1.9)|
|Mohawk [10−2]||1.6 (1.2–2.0)||1.3 (0.9–2.1)||0.5 (0.2–1.1)b|
The cell count was significantly increased by application of BMP-7 (1 µg/ml) at day 3, 5, and 7 in TLCs of the low and middle fatty infiltration group (Fig. 5A and C), but only at day 7 in the high muscle fatty infiltration group (Fig. 5E). At day 7, cell count was additionally increased with 0.2 µg/ml BMP-7 in the middle fatty infiltration group (Fig. 5C) and the 0.2 µg/ml BMP-2 concentration in the low fatty infiltration group (Fig. 5A). Col-I expression and protein synthesis were significantly increased by both factors and concentrations in TLCs of the low- and middle fatty infiltration groups (Fig. 5B and D). In the high muscle fatty infiltration group, Col-I protein synthesis was only increased by BMP-7 application (Fig. 5F). The expression of Col-III was significantly enhanced after application of 1 µg/ml BMP-7 in TLCs of the low- and middle fatty infiltration group by 2.2 (1.6–2.7) and 2.5 (1.9–2.6) fold, while it was increased with both factors and concentrations in cells of donors with high fatty infiltration (1.4–2.1-fold). The expression of osteocalcin and Col-II were not affected by growth factor stimulation (data not shown).
When analyzing the stimulation potential with regard to the muscle fatty infiltration, a significantly reduced stimulation of cell count (days 5 and 7) and Col-I protein synthesis was found in the high fatty infiltration group compared to the low (p = 0.002–0.001) and middle (cell count: p < 0.001, Col-I synthesis: p = 0.025) fatty infiltration groups after application of 1 µg/ml BMP-7. In the 0.2 µg/ml BMP-2 stimulation group, a reduced cell count was found at day 7 for the high fatty infiltration group compared to the low fatty infiltration group (p = 0.002). For the stimulation with 0.2 µg/ml BMP-7, the strongest increase in cell count at day 7 was seen in cells from donors with middle fatty infiltration. This was significantly increased compared to the low (p = 0.005) and the high (p = 0.002) muscle fatty infiltration group.
TLCs of all donor groups had a tenocyte specific gene expression profile, while expressing high amounts of Col-I, lower amounts of Col-III and decorin. Furthermore, cells expressed tendon-related genes like scleraxis and mohawk homeobox. Cells of donors with high muscle fatty infiltration seem to lose their tenogenic potential, with mohawk expression being decreased in these cells. Nearly no tenomodulin expression was measured in TLCs, which might be caused by higher passage and the 2D culture.[25, 26] TLCs expressed no Col-II and only low amounts of osteocalcin, which served as controls for an osteogenic or chondrogenic lineage. A portion of the TLCs may be tendon stem cells, because of their colony forming ability, which was present in 13%, 11.8%, and 6.7% of cells from donors with low, middle, and high fatty infiltration, respectively. Cells showed a clear stem cell phenotype by expressing stem cell markers, but no negative markers. A portion of TLCs of all groups showed multipotent differentiation. The differentiation toward an osteogenic phenotype was relatively weak, when considering the Alizarin Red S staining. We assume to have a mixture of tenocytes and tendon stem cell in culture, which have a weaker differentiation potential than reported for pure stem cells. However, the multipotent differentiation potential of the TLCs was proved by analysis of lineage specific markers, which revealed a high upregulation of the osteogenic markers Runx2 and ALPL, the adipogenic markers PPARγ, LPL, and FABP4 and the chondrogenic markers Col-II, COMP and aggrecan.
The adipogenic and osteogenic differentiation of the TLCs was strongest in the middle fatty infiltration group. It was hypothesized that stem cell populations in skeletal muscle are responsible for the increased adipogenic differentiation and subsequently the development of fat within the muscle. It is possible that a population of tenocytes or tendon stem cells as well have this increased differentiation ability. The decreased differentiation capacity in cells from the high fatty infiltration group compared to the middle fatty infiltration group may be due to the fact that these cells are already less potent to show a stronger differentiation.
Many biomechanical studies focused on testing different suture techniques for mechanically optimized rotator cuff repair.[28-30] However, from the clinical perspective, the surgical method, such as single versus double row repair, seems to have no important impact on the mid-term outcome.[31, 32] Therefore, other approaches for improved rotator cuff repair should be taken into consideration. The present study along with previous studies showed that important cellular characteristics can be augmented with BMP-2 and BMP-7 in all analyzed donor groups.[7, 8, 13] Analyzed parameters like cell proliferation and Col-I protein synthesis may be key factors for the healing of rotator cuff tears in vivo. The BMP-7 treatment revealed the strongest results and may therefore represent a possible treatment option for future rotator cuff repair. The stimulation of TLCs with BMP-2 had a weaker effect compared to BMP-7, as shown previously.[7, 8, 13] Rui et al. showed that BMP-2 stimulation leads to adipogenic, osteogenic, and chondrogenic differentiation of tendon-derived stem cells, but repressed the differentiation into the tenogenic direction. In the present study, no analysis of multipotent differentiation after BMP stimulation was performed, but would represent an interesting approach for future investigations. However, as investigated by Rui et al. for BMP-2 stimulation, we showed previously that decorin expression was significantly decreased after BMP-7 stimulation in TLCs of young and aged male donors, and the young male group representing the low fatty infiltration group in the present study. It was furthermore claimed that the application of osteoinductive growth factors to tendons could lead to calcification, due to an increased mineralization potential of the cells.[14, 35] This may negatively influence the biomechanical properties of the tendon bone insertion site. On the other hand, the additional stimulation of osteoblasts at the interface may improve the integration. In general, most animal studies, applying BMP-2 or BMP-7 to the tendon bone insertion site showed better biomechanical outcome.[16-19] We showed that BMP-2 or BMP-7 did not increase the expression of osteocalcin as an osteogenic marker. However, the effect of the BMPs in tendon bone repair in vivo could not be anticipated from the present study.
A positive correlation between muscle fatty infiltration and the tendon retraction as well as tear size was observed. The differences seen in the cellular biological characteristic and the stimulation potential of TLCs may also be related to changes in tendon retraction and tear size, possibly based upon longer-standing pathologies. Cells with higher grades of muscle fatty infiltration showed lower cell count and potential for self-renewal. Longo et al. reported on histopathological changes in tendons with small compared to larger rotator cuff tears. They found decreased active fibroblast populations in larger tear groups. Accordingly, Matthews et al. reported on reduced cell proliferation in large- and no proliferation in massive rotator cuff tears in tissue sections. These findings are comparable to the present results. Furthermore, the decreased stimulation potential of cell count in the high muscle fatty infiltration group underlined the inferior characteristics of these cells.
From the clinical perspective, a Goutallier grade higher than 1 was found to be the cut off between tendon healing and recurrent defects.[1, 4] On the cellular level, the main significant differences in biological characteristics of TLCs were found between a 2 and 3–4 graded muscle fatty infiltration. As a reason for the generally weaker cellular characteristics, it can be presumed that the reduced mechanical simulation of the tendon and muscle, as a result of discontinuity and pain and with limited active movement,[2, 5] may influence the cells in the tendon and cause the present findings. The mechanoresponsiveness of tenocytes was shown earlier. It was reported that tenocytes isolated from achilles or patellar tendons under dynamic tensile stimulation react with an increased cell proliferation and Col-I synthesis.[38, 39] An increased collagen synthesis was also found by Maeda et al. who applied tensile strain to tendon fascicles. It can be supposed that the muscle fatty infiltration may not have a direct effect on the TLCs, but the missing or constrained movement, which is causing the increased fatty infiltration, may also affect the cells in the tendon.
The age difference in the 3 donor groups varied from a mean of 45.3 years in the low fatty infiltration group to 59.7 and 55.7 years in the middle and high fatty infiltration groups. Previously, we reported that cellular characteristics change with age (mean: 45.3 vs. 71.3 years). Therefore, it could be speculated, that the present findings are also influenced by age differences. However, no significantly different characteristics or stimulation potential was observed between the low and middle fatty infiltration groups, were the age difference was most evident. The loss of cellular capacities occurred only with high muscle fatty infiltration. In the present study, no TLCs of donors from an aged cohort older than 65 years were analyzed. It can be hypothesized that under the age of 65, which is the clinical cut-off for a poorer healing outcome among some clinical studies, the age difference does not cause dramatic changes at the cellular level.
In conclusion, the present study showed an overall weakness of important characteristics of TLCs isolated from SSP tendons from patients with high muscle fatty infiltration compared to lower fatty infiltration groups. This may be a reason for the poorer healing potential at the tendon bone insertion site in this donor group. Due to the stimulation of important parameters like cell count and Col-I protein synthesis with BMP-2 and BMP-7, this may represent an approach for future treatment options in order to improve rotator cuff repair after surgery. Additionally, the recovery of the muscle structure should be addressed for a better healing potential. One possibility could be a continuous elongation of the muscle as investigated by Gerber et al., who showed that this leads to the restoration of normal muscle architecture in sheep. Addressing the BMP-application to stimulate the healing at the tendon bone insertion site together with the recovery of the muscle structure in a rotator cuff repair model would be an interesting approach for future investigations.
We would like to thank Dr. Martina Seifert (BCRT) for kindly providing us with the stem cell panel for FACS analysis. We thank B.Sc. Janosch Schoon for his great support with the multipotent differentiation analysis and Mario Thiele for the image analyzing system for quantifying CFUs. We furthermore would like to acknowledge the excellent assistance of the BCRT flow cytometry lab and the BCRT cell harvesting core unit.