Osteoarthritis (OA) is the most prevalent degenerative joint disorder and a main reason for chronic pain and disability in the elderly population . While multiple factors, such as genetic disposition, mechanical stress (caused by overload, obesity, or joint malalignment), previous injuries, or age are known to increase the risk to develop OA , the precise etiology of this multifactorial disorder remains elusive. Characteristically, OA is defined by progressive cartilage degeneration , alterations of the cartilage–bone interface , inflammation of the synovial tissue , formation of osteophytes, and bone sclerosis .
Recently, the cartilage underlying subchondral bone has been ascribed a crucial role in the progression of OA [7,8]. While at earlier phases of OA subchondral bone volume is decreased , its substantial thickening, the so-called sclerosis, characterizes later stages of OA . The temporal sequence of bone and cartilage changes might vary according to the disease phenotype or the animal model, however it is commonly agreed that the interplay between these tissues is crucially involved in the development of OA [10,11]. In this regard, in vitro studies have demonstrated that osteoblasts (OB) from patients with OA significantly influenced the metabolic activity of human chondrocytes [12,13], e.g., by upregulating the expression of extracellular matrix degrading enzymes, such as matrix metalloproteinase-3 (MMP-3) and matrix metalloproteinase-13 (MMP-13) [14,15]. This can be most possibly due to the altered characteristics of OB from osteoarthritic tissues, e.g., in terms of their abnormal production of extracellular matrix components  or of other factors, such as cartilage degrading mediators and growth factors, as compared to OB from healthy subchondral bone [17,18,19].
Cellular aging is defined as the progressive decline in the resistance of cells to stress, characterised by distinct hallmarks, such as the aggregation of reactive oxygen species (ROS), the loss of protein homeostasis, mitochondrial dysfunction, DNA damage, and epigenetic alterations . The accumulation of these events will ultimately result in senescent cells, which cease proliferation and start expressing pro-inflammatory and extracellular matrix-degrading factors . Senescent cells, typically occurring during natural aging, have been found to highly accumulate in musculoskeletal diseases, including OA and osteoporosis [22,23]. While replicative senescence has been described in osteoblasts of OA patients in vitro , a biological role of senescence on bone formation was addressed only very recently, demonstrating that the elimination of senescent cells in mice prevented not only age-related disorders in general , but also bone loss . However, the contribution of specific senescent cells, such as OB, to this phenomenon remains unclear because of their systemic clearance. For chondrocytes, which are another cell type present in the synovial joint, the role of senescence and its possible function on age-associated modifications of cartilage, and thus on the development of osteoarthritis, has been addressed in some more studies [26,27,28]. In particular, such premature chondrosenescence is proposed to contribute to a pro-inflammatory and catabolic environment , altering intercellular communication and compromising the surrounding extracellular matrix thus promoting the onset of OA [30,31]. In line with these findings, a recent study impressively demonstrated that the local clearance of senescent cells in vivo attenuated the development of post-traumatic osteoarthritis in mice .
Free radicals and the associated oxidative stress are known as major inducers of cellular senescence  and also seem to play a major role in the process of impaired functionality and decreased number of osteoblasts during aging, as shown in vitro and in vivo in numerous studies [34,35,36]. Moreover, it has been highlighted that a decrease in new bone formation and osteoblastic activity is caused by increased levels of free radicals and is associated in vitro with increased ROS levels leading to the activation of p53 and p66shc, key components influencing apoptosis pathways and lifespan .
Since the production of free radicals such as ROS has been proposed to cause senescence, counteracting oxidative stress could represent a possible mean to increase the life span of OB. Accordingly, it has been demonstrated that the supplementation of antioxidants such as ascorbic acid could have a beneficial effect on the reproduction and differentiation of osteoblastic cell lines, enhancing mesenchymal stem cell proliferation and maintaining their characteristic phenotype in vitro [38,39]. However, it remains unaddressed whether the supplementation of antioxidants can reduce or delay the number of senescent cells in primary OB derived from OA patients, while maintaining their phenotypical characteristics.
The aim of our study was to investigate the effect of the antioxidant Ascorbic acid-2-phosphate (AA) supplementation on the isolation efficiency, expansion, and differentiation potential of human OB from osteoarthritic subchondral bone. We furthermore examined the onset of senescence in the presence and absence of AA and studied the entailed modulation of the transcriptomic profiles of primary osteoarthritic OB.
2.1. Effect of Ascorbic Acid on the Outgrowth and Proliferation Rate of Human Osteoarthritic OB
First, we assessed the influence of AA on the outgrowth efficiency of human primary OB harvested from sclerotic (Sc_OB) and non-sclerotic (N_OB) osteoarthritic subchondral bone. A significantly increased number of OB (per gram of tissue) was observed to grow out from bone chips in the presence of AA (2.5 ± 1.1-fold increase, p = 0.0072 for Sc_OB, and 2.1 ± 0.6-fold increase; p = 0.0135 for N_OB) as compared to the number of OB growing out from the corresponding bone chips cultured in standard culture medium (CM) (Figure 1A). Conversely, statistically significant differences were not detected between the two OB types regarding the mean number of outgrowing cells either at control conditions (CM) or in the presence of 0.1 mM AA (Figure 1A).
The expression of the osteogenic cell markers alkaline phosphatase (ALP) and osteocalcin (OC) was confirmed in standard culture conditions by fluorescence-activated cell sorting (FACS) analysis for the outgrowth of Sc_OB (7.5% of the CD45 negative population) and N_OB (27.7% of the CD45 negative population). In the presence of AA (+AA), slightly larger percentages of CD45−/ALP+/OC+ cells could be identified for Sc_OB (20.0%) and N_OB (38.4%) (Figure 1B).
At the gene expression level, significant differences were not observed between Sc_OB and N_OB in the presence or absence of AA, for OC, Col1A1 and 1A2, or TGFβ1 by quantitative RT-PCR. Irrespective of their sclerotic or non-sclerotic origin, OB demonstrated a significantly increased capacity to proliferate in the presence of AA (1.6 ± 0.6 and 1.9 ± 0.9-fold increase for Sc_OB at P1 (p = 0.0313) and P2 (p = 0.0469), respectively; 1.7 ± 0.5-fold increase for N_OB at P2, p = 0.0156) (Figure 1C). This trend of enhanced proliferation rate was maintained in the following passages when AA was supplemented to the culture medium, finally resulting in a significantly increased total number of accumulated population doublings by P4 (10.03 ± 3.2 PD for Sc_OB vs 12.38 ± 3.8 for Sc_OB+AA, p = 0.0156; 11.8 ± 3.6 PD for N_OB vs 14.3 ± 3.2 PD for N_OB+AA, p = 0.0313) (Figure 1D). This effect positively correlated with the increased expression of a master regulator of proliferation, Ki67, both at the gene expression level (up to 60-fold) (Figure 1E) and at the protein expression level (Figure 1F).
2.2. Attenuation of Human Osteoarthritic OB Senescence in the Presence of Ascorbic Acid
To evaluate a possible correlation between the demonstrated enhanced proliferation capacity of osteoarthritic OB and the amount of senescent cells occurring in the presence of AA, OB were stained at each passage in the course of expansion for senescence-specific β-galactosidase (shown representatively for P1 and P4 in Figure 2A). The number of positively stained cells was quantified for each condition and time point, and was presented individually for each donor as the percentage of senescent cells in the total population (Figure 2B). The application of AA significantly decreased the number of senescent cells. While at single-donor level some significant differences between Sc_OB and N_OB were determined when the OB were cultured in CM (for donor 1 at P2 and for donor 3 at P1 and P3), no such differences were found at any culture condition once the data from these three donors were analysed together. In fact, the number of senescent cells at each passage was always at least 2.4-fold higher for Sc_OB as well as for N_OB, when no AA was supplemented, while the strongest effect of AA was observed at P2 (17.04 ± 20.1 for Sc_OB, p = 0.0424) (Figure 2C).
Taken together, Sc_OB and N_OB responded in a similar way to the application of AA by diminishing the number of senescent cells.
We then investigated whether the attenuation of OB senescence in the presence of AA was mediated by the antioxidative function of AA, by quantification of intracellular reactive oxygen species (ROS) production. ROS expression was significantly decreased in the presence of AA by 13-fold ± 7.6 at P1, as compared to CM-expanded OB, and the extent of this reduction successively decreased with each passage from 8.9 ± 3.0 at P2, to 1.9 ± 0.4 at P3, to finally 2.7 ± 1.5 at P4 (Figure 3). Conversely, no significant differences in the mean fluorescence intensity, which was considered to directly correlate with the amount of ROS expression, were observed between Sc_OB and N_OB in the same culture conditions (either with or without AA).
Here, we demonstrated that the enhanced proliferative capacity of osteoarthritic OB in the presence of AA was accompanied by a reduced amount of senescent cells and a decreased expression of ROS.
2.3. Osteogenic Differentiation of Human Osteoarthritic OB in the Presence of Ascorbic Acid
Osteoarthritic OB expanded for two passages in either the absence or presence of AA were cultured in standard 2D osteogenic medium for 21 days (n = 5 donors). Under control conditions, lacking the osteogenic factors dexamethasone and β-glycerophosphate, no mineralisation was detectable by alizarin red staining. Following osteogenic induction, variable degrees of mineralisation were observed for both Sc_OB and N_OB, ranging between no and strong mineralisation. The observed intradonor specific osteogenic differentiation ability of OB remained unchanged in the presence of AA.
To investigate the bone formation potential of osteoarthritic OB in vivo, cells seeded on a ceramic scaffold (Engipore) were implanted in subcutaneous pouches of nude mice. Masson trichrome staining of explanted construct 4 weeks postimplantation did not reveal mature bone formation, however areas of pre-bone were identified (Figure 4A) and quantified by bone histomorphometry (Figure 4B). In general, a similar bone formation capacity was identified for all OB types expanded under different conditions, although significantly more bone-like tissue was found in scaffolds seeded with N_OB+AA as compared to N_OB.
These results indicate that supplementation of OB with AA during outgrowth and expansion did not result in reproducible changes of their osteogenic differentiation capacity.
2.4. Transcriptomic Analysis of Human Osteoarthritic OB in the Absence and Presence of Ascorbic Acid
To assess the effect of AA supplementation on osteoarthritic OB at molecular levels, RNAseq-based transcriptomic profiling was performed following outgrowth (P0) or expansion (at P2 and P4). For this analysis, Sc_OB and N_OB were not discriminated since no major differences were observed between these two osteoarthritic OB populations in the experiments described before.
For three of the four donors (donors 4, 5, and 6), outgrowing OB (P0) were found to group according to their overall expression profile, whereas no such correlation was identified for P2- and P4-expanded OB (Figure 5A). Therefore, for further downstream analysis, P2- and P4-expanded OB were not discriminated, but analysed as one group representing the phenotype of expanded cells in the absence or presence (+AA) of AA. For outgrowing OB cultured in the presence of AA, 24 genes (14 genes downregulated and 10 genes upregulated; Supplemental Table S1) were identified to be differentially expressed as compared to CM-cultured cells (Figure 5B upper plot). In control conditions, among the higher expressed transcripts were PTX3 (pentraxin 3), IGFBP1 (insulin-like growth factor-binding protein 1), and SESN2 (sestrin 2), whereas transcript levels of HAS1 (hyaluronan synthetase 1) and SEMA7A (semaphorin 7A) were decreased. Following expansion, significant differences were found for 44 genes (33 genes downregulated, 11 genes upregulated; Supplemental Table S2) between CM- and AA-expanded OB. Higher levels of transcripts were detected at control conditions, e.g., for LGR5 (Leucine-rich repeat-containing G-protein coupled receptor 5), MEIS3 (Meis homeobox 3), and SMOC2 (SPARC related modular calcium binding 2). Conversely, mRNA levels of LPX (leupaxin) and IGF2BP3 (insulin-like growth factor 2-binding protein 3) were increased by AA (Figure 5B lower plot).
Surface plasmon resonance (SPR) is a collective charge density oscillation that occurs at a metal-dielectric interface when light passes through a substrate and is reflected by the metal-dielectric interface [1,2]. If the wave-vector component of the incident light that is oriented parallel to the interface matches the propagation constant of the surface plasmon wave, SPR then occurs. In this case, most of the incident energy is coupled into the surface plasmon mode field, which results in shifts in the resonance angle and wavelength, along with changes in the intensity and phase of the reflected light.
Since SPR was first used for biosensing purposes by Liedberg in 1983 , research interest in SPR biosensors has increased rapidly [1,4,5] and these sensors have found wide-ranging applications in fields including drug discovery , nucleic acid detection , food safety , and environmental monitoring . Several methods have been used to date to monitor the excitation of SPR, including angle [10,11], wavelength [12,13], intensity [14,15] and phase [16,17] interrogation techniques. As many researchers have demonstrated [18,19,20,21], the phase of the SPR reflected light changes much more abruptly than its intensity. Therefore, phase-interrogated SPR biosensors are the most sensitive excitation monitoring method.
Because the optical configurations of phase-interrogated SPR sensors are more complex than those of the other sensor types, we believe that it is necessary to review SPR phase detection techniques and to compare the performances of these techniques with the other SPR monitoring methods. This paper focuses on phase-interrogated SPR biosensing technology, reviews the fundamentals of SPR sensing and the associated methods of phase interrogation, and discusses developmental advances and emerging trends in the field.
Prior to the main discussion, it is important to clarify the terminology defining sensitivity characteristics of SPR sensors. The sensitivity of SPR biosensors is composed of chemical and physical sensitivity [10,17]. The chemical sensitivity depends on surface chemistry and assay format (“direct”, “sandwich”, “competitive”, “inhibition” etc.), and the physical sensitivity depends on plasmonic transduction modality, optical configuration, and the level of instrumental and environmental noises. In this review, we will be focused on the methods of SPR response interrogation and thus just consider the physical sensitivity. In addition, the term “sensitivity” usually implies the shift of SPR response (angular or spectral position of the SPR dip, intensity, phase) over the variation of the refractive index (RI). But it is difficult to compare the sensitivity for different interrogation schemes. So here we use the term “sensitivity” to characterize the minimal measurable variation of RI. More exactly, it is should be termed as “limit of detection” (LOD). But since the term “sensitivity” is more prevalent for SPR sensors, we term it as “sensitivity” to compare SPR sensors of different interrogation schemes.
2. Principles of SPR Sensing
The intensity and the phase of the reflected light both change when there is a change in the refractive index of the sample solution in the vicinity of the gold film. Intensity monitoring offers the advantage of simple optical configuration requirements, and has been utilized in many commercial devices [25,26]. With a more complex optical configuration, the phase interrogated SPR sensors can obtain a higher sensitivity, since the phase of the reflected light undergoes a more abrupt change than the intensity [18,19,20,27,28,29].
3. Optical Configurations for SPR Phase Interrogation
While light intensity measurement is a straightforward process, the high-frequency oscillations (of the order of 1014 Hz) of light cannot be observed directly. Complex optical configurations are thus required to retrieve SPR-induced phase changes, with methods including heterodyne detection, ellipsometry and various interferometry techniques [17,30].
3.1. Heterodyne Detection
The heterodyne method is commonly used in phase detection [31,32,33]. The fundamental aspect of this method is the generation of two identical laser beams that include two orthogonally polarized components at slightly different frequencies. Combination of these orthogonally polarized components produces an interference signal with a “beat” frequency that is lower than the detector’s response frequency. Extraction of the phase from low-frequency signals then becomes much easier.
The typical optical configuration that is used for SPR sensing is shown in Figure 2