Pathway Analysis Hints Towards Beneficial Effects of Long-Term Vibration on Human Chondrocytes Open Access

Chondrocytes or cartilage cells are the only cell type detectable in articular cartilage. In healthy cartilage, they produce and are embedded in a stable but elastic extracellular matrix (ECM) [1]. This ECM composition with collagens and proteoglycans provides resistance against mechanical forces and friction for the joints and determines the overall properties and functionality of the cartilage [2]. Articular cartilage is important for joint mobility and serves as a mechanical, frictionless cushion during movement [3].

Whole-body vibration (WBV) has been used to treat musculoskeletal diseases like osteoarthritis (OA), but the direct effect of vibration on joint cartilage is not clear. A recent study showed that WBV induced cartilage degeneration in mice [4]. Higher frequencies of WBV (30 Hz and 40 Hz) in a rabbit model had a negative influence on cartilage volume and cartilage resorption, whereas lower frequencies (10 Hz and 20 Hz) decreased cartilage resorption, accelerated cartilage formation, and delayed cartilage degradation especially at the 20 Hz regimen [5].

Little is known about the direct impact of vibration on chondrocytes in vitro. It has been demonstrated that mechanical vibrations can elevate the proliferation of chondrocytes in monolayer cultures [6]. These vibration effects seem to be limited to the early stages where ECM (collagen and proteoglycans) accumulation is at a minimum [6].

A prolonged stay in orbit affects the health of humans in space. Bone loss and cartilage breakdown in microgravity occurs after a long-term spaceflight [7]. Due to the poor regenerative capacity of cartilage tissue, this degradation may disturb the flight crews’ mobility and may negatively influence mission activities [8]. Certainly, microgravity is a key factor for these degradation processes, but also other factors such as cosmic radiation and vibration might be important to alter cartilage tissue.

To receive more information about the time point when these changes on the cellular level occur, chondrocytes were investigated during short-term microgravity obtained during a parabolic flight (PF) [9]. As a control, a corresponding short-term, two-hour period of vibration as it occurs during the PF was affecting cartilage cells (upregulated inflammatory IL6 and CXCIL8 mRNAs, downregulated growth factor genes like EGF, VEGF, FGF17) [9]. However, no significant changes in morphology and gene expression levels of cytoskeletal genes, cell adhesion molecules, integrins, as well as of TGFB, CAV1 and SOX9 were observed during vibration [10].

To our knowledge no information exists concerning the impact of long-term vibration on chondrocytes in vitro. Therefore, these experiments are designed to determine the influence of a 24-hour vibration-exposure of human chondrocytes on gene expression and protein content of selected factors important for various biological processes like cytoskeleton, ECM, cell adhesion, apoptosis, growth and others. Vibrations may influence the cells and the human body during the launch of a rocket, during the flight of a plane in parabolic campaigns vibration [11, 12] or during normal passenger flights. Therefore, it is important to increase our knowledge in this field and to investigate whether long-term vibration has a detrimental or even beneficial effect for chondrocyte growth in vitro. Moreover, the results of this study are of high importance for regenerative medicine and orthopedics.

Human chondrocytes derived from hip joint cartilage were purchased from Provitro (Berlin, Germany). They were cultured in Chondrocyte Growth Medium (CGM; Provitro®, Berlin, Germany) supplemented with 10 % fetal calf serum (Provitro®, Berlin, Germany) and antibiotics – 100 IU penicillin/mL and 100 μg streptomycin/mL (Provitro®, Berlin, Germany). The cells from frozen stocks (passage 1) were grown in T175 cell culture flasks (175 cm2; Sarstedt, Nümbrecht, Germany) until sub-confluent layers were obtained. Afterwards, the cells were split (passage 2) in 5 T175 cell culture flasks and after reaching confluence 4 T175 were split (passage 3) in 80 T25 cell culture flasks (25 cm2; Sarstedt, Nümbrecht, Germany) for the vibration experiments (n=40 for vibrated samples and n= 40 as 1g-control samples).

For histological investigations, the chondrocytes were seeded in slide flasks one day before the experiments (n=40 each group; Thermo Fisher Scientific, Waltham, MA, USA).

The Vibraplex vibration platform (frequency range 0.2 Hz-14 kHz) was used to create vibrations like to the ones occurring during parabolic plane flights. The device is driven by an amplified wave signal. The injected wave signal is equal to a 1/f² noise (red noise). The signal was generated with the software WaveLab 4.0 from Steinberg. The device was described in detail in [10].

F-actin was visualized by means of rhodamine-phalloidin staining (Molecular Probes) [10]. For this purpose, seeded cells were fixed for 30 min with 4% PFA (in DPBS), washed twice with DPBS, incubated with 5 mg/mL fluorescent phalloidin-conjugate solution in DPBS with 1% bovine serum albumin for 20 min at room temperature, and then washed 3 times with DPBS to remove unbound phalloidin conjugate. Afterwards, the nuclei were stained with Hoechst 33342 (Molecular Probes) for 5 min and washed twice with DPBS. For evaluation, the samples were mounted with Vectashield (Vector Laboratories) and analyzed microscopically.

Hematoxylin–eosin- and Elastica-van-Gieson-staining procedures were used to evaluate the cell morphology of the chondrocytes. The methods were published earlier [13-15]. All samples were visualized by light microscopy. The samples were investigated by microscopy using a LEICA DM2000 microscope equipped with a Leica DFC310 FX digital CCD color camera.

F-actin stained cells were analyzed with a Leica TCS SP8 confocal laser scanning microscope (Leica Microsystems, Wetzlar, Germany).), equipped with a Plan-Apochromat 363 1.4 objective. Excitation and emission wavelengths were λ_exc = 488 nm and λ_em = 505 nm for FITC. All samples were analyzed with the help of the image analysis program Scion Image (Version 1.63 MacOs; Scion Corporation, Frederick, MD, USA). Phase contrast microscopy was done with a LEICA DM2000 microscope equipped with a Leica DFC310 FX digital CCD color camera.

TdT-mediated dUTP-biotin nick end labeling (TUNEL) staining was done according to the manual provided by the manufacturer (Calbiochem®, FragEL^TM DNA Fragmentation Detection Kit, Fluorescent - TdT Enzyme). Cells for 4’,6-diamidino-2-phenylindole (DAPI) staining were fixed with 3.7% formaldehyde (room temperature, 10 min) and incubated in 1 μg/mL DAPI in PBS for 15 min (Invitrogen/Molecular Probes, Darmstadt, Germany). Stained cell samples (VIB and static CON) were investigated utilizing a Leica DM 2000 microscope connected to an external light source, Leica EL 6000 (Leica Microsystems GmbH, Wetzlar, Germany). Palatine tonsil cross-sections were taken as positive control using an objective with a calibrated magnification of 400x.

Western blotting, immunoblotting, and densitometry were performed employing standard protocols [14, 15]. We used the Biorad ChemiDoc XRT+ device. The number of investigated samples/per group was n=4.

Equal amounts of 20 µL lysate, containing 3 µg/µL protein, were loaded onto precast TGX stain-free gels (Bio-Rad, Munich, Germany). Each western blot was performed 3 times. Anti-beta-tubulin, and anti-vimentin, were used at a dilution of 1: 1000 (Cell Signaling Technology, Inc.); anti-osteopontin antibody was used at a dilution of 1: 1000 (Rockland Immunochemicals Inc., Limerick, PA, USA) as well as the anti-integrin-beta1 antibody (Epitomics, Burlingame, CA, USA). The anti-SOX9 antibody was applied at a dilution of 1: 500 (Life Technologies). In addition, we used an anti-caveolin-1 antibody (1: 1000; Abcam).

The anti-ICAM-1 antibody was purchased from Santa-Cruz Biotechnologies (Heidelberg, Germany) and diluted 1: 500. The secondary horseradish peroxidase-linked antibody was used at a dilution of 1: 4000 (Cell Signaling Technology, Inc.).

Ponceau S red staining was used as an alternative to housekeeping proteins as loading controls. The membranes were analyzed using ImageJ software (U.S. National Institutes of Health, Bethesda, MD, USA; http://rsb.info.nih.gov/ij/), for densitometric quantification of the bands. Ponceau S was evaluated according to standard protocols.

RNA isolation was performed using the RNeasy Mini Kit (Qiagen, Hilden, Germany), with an additional DNase digestion step (Qiagen) in order to eliminate residual DNA contaminations. Subsequently, the amount of RNA was quantified via a Photometer Ultrospec2010 (Amersham Biosciences, Freiburg, Germany). The first strand cDNA synthesis kit (Thermo Fisher Scientific, Waltham, US) was used for reverse transcription. qPCR was performed using the 7500 Fast Real-Time PCR System (Applied Biosystems, Darmstadt, Germany) according to routine protocols [16, 17]. cDNA-selective-primers were synthesized by TIB Molbiol (Berlin, Germany) and are listed in Table 1. The primers were designed using Primer Express (Applied Biosystems, Darmstadt, Germany) to have a T_m of ∼ 60°C and to span exon-exon boundaries. All samples were measured in triplicate. For normalization, 18S rRNA was used as a housekeeping gene. The comparative C_T (ΔΔC_T) method was used for relative quantification of transcription levels and 1g was defined as 100 % for reference.

The Illumina HumanHT-12 v4 Expression BeadChip arrays have been normalized using the GenomeStudio V2011.1. with Gene Expression Module 1.9.0. and quantile normalization without background correction. For each group (control and vibration), six independent microarrays were analyzed. After quantile normalization and exclusion of low or not expressed genes (minimum Illumina detection p-value > 0.05) differences between chondrocyte control, and vibrated chondrocytes was analyzed using Student’s t-test. None of the probes reached a 5% FDR level [18].

Differentiation of the expression profiles was performed using hierarchical clustering over 13 probes with a nominal p-value < 0.001. The cluster analysis was done using Partek Genomic Suite 6.6 applying hierarchical average linkage clustering on standardized log2 signal values.

To investigate the mutual regulation of genes and to visualize localization and interactions between proteins, we entered relevant UniProtKB entry numbers in the Pathway Studio v.11 software (Elsevier Research Solutions, Amsterdam, The Netherlands). Graphs were generated for gene expression and protein regulation and binding. The method was described previously [19, 20].

All statistical analyses were performed using SPSS 16.0 software (SPSS, Inc., Chicago, IL, USA). We used the Mann-Whitney-U-test to assess differences, which were considered significant at the level of p < 0.05. All data are presented as means ± standard deviation.

After 24 h of vibration, no morphological differences between static control cells (Fig. 1A) and vibrated chondrocytes (Fig. 1B) were visible. Both samples showed regular cell morphologies without any detectable aberration.

Both static control (Fig. 1C) as well as vibrated cells (Fig. 1D) displayed a normal distribution and structure of actin fibers with no signs of damage, no disarrangements or deposits of F-actin at the cellular membranes.

To characterize the cell morphology of both control as well as vibrated chondrocytes, we additionally performed hematoxylin and eosin (HE) as well as Elastica van Gieson (EVG) staining. HE staining revealed no signs of necrosis, but normal morphology in both groups. Furthermore, proliferation rates did not differ. Collagen and elastin fibers content as well as distribution were similar in both control and vibrated cells, as demonstrated by EVG staining.

Cells subjected to vibration as well as static control cells were analyzed with DAPI and TUNEL staining to assess the extent of apoptosis. Cross-sections of palatine tonsils were used as positive control. No apoptotic cells could be detected in both vibrated and control samples (Fig. 1E-H).

We selected proteins of interest which are involved in several crucial biological processes such as the organization of the cytoskeleton, cell adhesion or cell-cell-contact in order to investigate the influence of vibration on the stability and integrity of the single cells as well of the whole cell layer.

No changes in the protein content between control and vibrated cells were detected for integrin β-1 (ITGB1, Fig. 2A), SOX-9 (SOX9, Fig. 2B), caveolin-1 (CAV1, Fig. 2C), intercellular adhesion molecule 1 (ICAM1, Fig. 2F) and intact osteopontin (OPN, Fig. 2G).

The proteins vimentin (VIM, Fig. 2D), beta-1 tubulin (TUBB, Fig. 2E), and the MMP-cleavage derived 32 kDa fragment of osteopontin (OPN, Fig. 2H), on the other hand, were significantly reduced under vibration.

In order to further characterize the influence of vibration on cultured chondrocytes, we examined the gene expression of genes found to be differentially regulated in chondrocytes subjected to simulated or real microgravity. In summary, vibration did not have an impact on the expression of our chosen set of genes comprising ACTA2, ACTB, BMP2, CAV1, ICAM1, INTA10, INTB1, KRT8, SOX5, SOX6, SOX9, SPP1, TGFB1, TUBB, and VIM (Fig. 3 A-O).

From the total of 28855 detected probes, only 13 were found to be differentially expressed (p < 0.001) between control and vibrated samples CNOT7, COMMD2, RPS26L, BAIAP2, HNRPUL1, Hs.445274, and LOC653157 were downregulated in vibrated chondrocytes, while TBX15, LOC641365, CYBASC3, PSMD4, GPR112, and ACO2 were upregulated in vibrated chondrocytes in comparison to control samples (Fig. 4A). A short-annotated overview of the regulated genes is given in Table 2. A further STRING analysis [21] (Version 10.5, using confidence levels of both 0.4 and 0.15, available from https://string-db.org) revealed, that the regulated genes do not interact with one another and do not share a common pathway in which they are involved (Fig. 4B).

However, a part of the genes and their products (CNOT7, PSMD4, TBX15) recognized by microarray technology had an influence on expression and function of the proteins selected for the western blot analysis (Fig. 2) which show strong interaction activity at gene (Fig. 4C) and gene product levels (Fig. 4D). On the gene level the nuclear protein CNOT7 is able to regulate ICAM1 and CXCIL8 expression (Fig. 4C, [22]). It is known that CNOT7 silencing stabilized ICAM1 and CXCIL8 mRNAs and increased ICAM-1 and IL-8 production following TNF-α stimulation [23]. It is of special interest as osteopontin, a factor that up-regulates ICAM-1 (Fig. 4C, [24]) is totally abolished after vibration. Moreover, VIM which is often stabilized by OPN [25] is significantly reduced. This fact could contribute to enhanced stability of ICAM-1 [26].

Chondrocytes are the major cells distributed in the dense ECM of articular cartilage. They exhibit a well-differentiated phenotype with unique physiological functions determining the properties of the load-bearing cartilage tissue. Diseases of the cartilage comprise OA which increased in prevalence across the world due to the aging population and it is one of the most frequent causes of chronic pain. Because there is no causal treatment of OA, the major objective of physicians is to use current available therapies and to slow down the progression of this disease. Therefore, new research focusing on the chondrocytes is necessary. A recent study determined the effects of low-magnitude whole WBV on cartilage degradation, bone/cartilage turnover, and OA joint function in a rabbit knee OA model [27]. The authors reported that lower frequency (20 Hz) WBV can improve bone microstructure, increase bone turnover, delay cartilage degeneration and improve limb function in this rabbit OA model [27]. Little is known about vibration effects on the cellular level.

In this study, we investigated the long-term (24 h) effect of vibrations on human chondrocytes in vitro. It is known that mechanical vibrations increase the proliferation of articular chondrocytes in high-density cultures of isolated bovine articular chondrocytes [6]. In addition, vibration promoted differentiation of rat pheochromocytoma (PC12) cells, which supported the hypothesis that effects of the physical environment can influence cellular differentiation [28].

Chondrocytes embedded in the ECM of the cartilage are known to be responsive to mechanical forces such as pressure, shear stress or vibrations [29-34]. During OA progression, the ECM is actively remodeled by chondrocytes under inflammatory conditions [35]. These alterations affect the biomechanical environment of chondrocytes, which leads to disease progression. ECM remodeling in OA progression also influences the contribution of mesenchymal stem cells in the repair process by inhibiting their chondrogenic differentiation [35].

In fact, it has been proposed that mechanical stress may be a useful tool for tissue engineering of articular cartilage [36]. Here, we focused on vibrations, as we have already conducted short-term studies on chondrocytes earlier [9, 10], which hinted towards a negative influence of vibration on cultured chondrocytes after 2 h.

We have prolonged the vibration time up to 24 h and in contrast to our earlier findings, we did not see a negative effect on chondrocytes in vitro. It is interesting to note, that none of the pro-inflammatory genes (such as IL6 and CXCIL8) which were differentially regulated after 2 h of vibration were detected as differentially expressed after 24 h employing a whole genome microarray analysis. Moreover, none of the previously identified microgravity-sensitive genes [10] were found to be regulated in a separate qPCR analysis. An overview of the candidate proteins analyzed by Western blot, qPCR, and whole genome microarray is given in Table 3. These results indicate that the chondrocytes adapt to vibration with time, which is an interesting finding, especially with regard to long-distance flights.

Only 13 genes have been identified to be differentially expressed after 24 h. The most interesting one was BAIAP2 (BAI1 associated protein 2, also known as IRSp53). The corresponding protein constitutes an insulin receptor substrate [37], and it has been shown that it is involved in the regulation of lamellipodia formation via activation of the WAVE2 protein [38]. Of note, it has recently been observed, that the formation of filopodia and lamellipodia in thyroid cancer cells were affected by the conditions of a TEXUS sounding rocket flight, which also includes vibration during the take-off phase [39]. Some of the identified genes listed in Table 2, if already characterized, were involved in unrelated general metabolic processes in the cell such as regulation of transcription, oxidation and reduction processes, generation of precursor metabolites and energy, mRNA splicing or DNA-templated transcription. This hints towards an unspecific activation of the vibrated cells as a general answer to the altered culture conditions, but offers no further insight into possible specific reactions of chondrocytes to long-term vibrations. Overall, these results seem to indicate that chondrocytes are able to adapt to vibrational stress over time. While short-term vibrations clearly exhibited a detrimental effect on cultured chondrocytes, our current study shows that the cells are obviously able to counteract these initial perturbations and revert to a normal state with only minor and more generalized reactions on the molecular level.

We found a downregulated T box transcription factor TBX15 in vibrated chondrocytes. High amounts of this transcription factor were found in human aortic valve interstitial cells (hAVICs) compared to human mitral valve interstitial cells [40]. Sun et al. identified TBX15 as a potential calcification-preventing factor in hAVICs [40].

PSMD4, or 26S proteasome non-ATPase regulatory subunit 4, is differentially regulated and is interacting with ICAM-1 (Fig. 4 ). This protein is of interest because the proteasome and its subunits are of clinical significance. A dysfunction can be associated with the underlying pathophysiology of specific diseases. The ubiquitin proteasome complex is also involved in the regulation of inflammatory responses. This activity is usually attributed to the role of proteasomes in the activation of NF-κB which further regulates the expression of pro-inflammatory cytokines and cell adhesion molecules like ICAM-1 and others [41].

Another interaction was found for CNOT7 (CCR4-NOT transcription complex subunit 7) and ICAM1. The carbon catabolite repressor protein 4 (CCR4)-negative on TATA (NOT) complex includes multiple subunits and is detecTable in human cells [22]. CCR4-associated factor 1 (CAF1) is involved in the regulation of ICAM1 and CXCIL8 expression. This interesting interaction reveals that mechanical vibration might influence cell adhesion and survival of the chondrocytes. In human pulmonary microvascular endothelial cells CNOT7/hCAF1 are involved in ICAM-1 and IL-8 regulation by tristetraprolin [23].

An interesting finding was the reduced amount of osteopontin in the human chondrocytes after 24 hours, which indicates a cell-protective influence of long-term vibration on the human chondrocytes. Osteopontin is associated with the severity and progression of OA [42].

It is secreted by various cell types, including macrophages, lymphocytes, epithelial cells, vascular smooth muscle cells, and even chondrocytes as well as synoviocytes [43-45]. Interestingly, mRNA expression and protein abundance of OPN are associated with the pathogenesis of OA [42]. Pullig et al. had demonstrated that the SPP1 mRNA isolated from human OA cartilage is elevated compared to normal cartilage samples [46]. Elevated OPN plasma levels in OA patients were found [47], indicated that SPP1 mRNA expression is associated with progression of OA. Therefore, we may conclude that the reduced osteopontin content in our experimental setting may indicate a beneficial effect of long-term vibration on chondrocytes.

Recent studies showed that phosphorylated osteopontin increases apoptosis and pro-inflammatory cytokine expression in human knee OA chondrocytes [48]. The authors concluded that inhibition of p-osteopontin may provide a novel effective strategy to slow or halt OA progression [48].

The chondrocyte cytoskeleton consists of actin microfilaments, microtubules and vimentin [49]. On the protein level, vimentin was significantly downregulated under vibration. Vimentin belongs to the intermediate filaments [50] and as such is involved in the maintenance of cell shape, cytoplasm integrity, and stabilization of the cytoskeleton [51]. Furthermore, vimentin proved to be sensitive to exposure to culture conditions on a Random Positioning Machine (RPM), where after 4 h disruptions in the vimentin network occurred [52]. Similar effects have also been observed in samples subjected to 31 parabolas (about 2h, fixation after the 31st parabola) during a parabolic flight [10]. However, after 16 h on the RPM, these disruptions were already partially repaired by the cell, which was accompanied by an increase of vimentin protein production [52]. This indicates, that obviously within a day, chondrocytes are able to adapt to different culture conditions with altered mechanical environments.

Vibration reduced beta-tubulin protein significantly supporting the thesis that the microtubule cytoskeleton is not adapting after 24 h and is still affected. Another study showed comparable data of follicular thyroid cancer cells (ML-1 cell line) exposed to a two-hour period of vibration. The protein content of alpha-tubulin was significantly reduced in ML-1 cells after two hours [53].

Therefore, we speculate, that in this study we can see very late stages of the adaption process of chondrocytes to cultivation under vibration. If any changes occurred in the actin cytoskeleton, they have probably been repaired at the time of observation. The detected downregulation of vimentin could be the consequence of an earlier overproduction of the protein to reconstruct the possibly disrupted vimentin network.

In this study, we have shown that chondrocytes do not undergo any noteworthy morphological changes when cultured under conditions of vibrations on a Vibraplex device for 24 h, simulating the vibration profile occurring during parabolic flight. Therefore, we conclude that vibrations in the frequency range of 0.2 Hz to 14 kHz did not have a negative impact on chondrocytes in vitro. In contrast, the reduction of osteopontin protein, the down-regulation of PSMD4 and TBX15 reveal that long-term vibration even positively influences chondrocytes in vitro.

The authors would like to thank the European Space Agency (ESA; (DG, RH) ESA-CORA-GBF-Project 2014-03 with acronym: Vibrated Chondrocytes), the German Space Agency (DLR; (DG) BMWi project 50WB1524), and Aarhus University, Denmark for funding. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. We also like to thank Roland Hartig and the section “Multidimensional Microscopy and Cellular Diagnostics” for assistance during microscopy.

D.G., R.L., R.H. and M.W. designed the experiments. C.B., S.R., S.K., K.So. and J.S. executed the experiments and collected the material. J.S. and S.K. performed western blot analyses. K.So. and C.B. performed qPCR analyses. K.Sa., N.H. and H.S. performed the gene array and analysis. M.K. and S.R. performed the staining. J.B. performed the pathway analyses. D.G., and M.W. wrote the manuscript. M.I. and R.H. contributed reagents, materials and analysis tools. All authors reviewed the manuscript.

The authors declare no competing financial interests.

D. Grimm and M. Wehland contributed equally to this work.