Biology of platelet-rich plasma and its clinical application in cartilage repair

In almost all published studies, PRP has been shown to have a strong positive effect on chondrocyte proliferation in vitro (Table 1) [29‐34]. When adult porcine chondrocytes were cultured in alginate beads in the presence of 10% PRP releasate (PRPr), 10% platelet-poor plasma (PPP) releasate or 10% fetal bovine serum (FBS) for 72 hours, the increase in DNA content in the PRP group was significantly higher than that in the PPP group and that in the FBS group [29]. The stimulatory effect of PRP on chondrocyte proliferation was also observed after 4, 8 and 12 days of in vitro culture, in either a monolayer or three-dimensional environment [30]. Additional studies revealed a dose- and time-dependent enhancement of chondrocyte proliferation by PRPr, even during the 20 day monolayer culture [31, 32].

There is, however, less concordance in the effect of PRP on chondrocyte differentiation. Akeda and colleagues [29] reported that 10% PRP treatment significantly increased proteoglycan and Col II synthesis compared to treatments with 10% PPP or 10% FBS, with the major profiles of proteoglycan and collagens being similar to those seen in cells cultured with FBS, indicating maintenance of a stable chondrocyte phenotype with PRP exposure. Similar effects were also noted in human osteoarthritic chondrocytes [30]. The supernatant from platelet-rich fibrin (PRF) up-regulated the mRNA expression of Col II and aggrecan and increased synthesis of glycosaminoglycan and proteoglycan by chondrocytes cultured either on the two-dimensional surface of fibrin scaffolds or in three-dimensional scaffolds compared to controls without addition of exudate. Furthermore, a proteomic study revealed that PRPr supplementation could also induce the expression of proteins associated with chondrocyte differentiation [31]. In particular, PRPr increased the expression of aggrecan and Sox9, without increasing the expression of Col X and alkaline phosphatase. Increased Sox9 expression has been shown to be associated with the chondrocytic re-differentiation process [35]. These results show that PRP had differential effects on chondrocytes; that is, promoting the synthesis of hyaline cartilage matrix while cellular progression to terminal hypertrophy is not facilitated or at least delayed. Another study further demonstrated that the chondrogenic differentiation and maturation induced by PRP treatment was related to the up-regulated expression of cannabinoid receptor 1 and 2 [36, 37]. However, a few authors have argued that PRP treatment was unable to induce the deposition of typical cartilage matrix components [32, 34], that there was no difference in the enhancement of extracellular matrix (ECM) production between the groups with PRP, PPP, whole blood or fibrin glue added into gelatin-based microcarriers [38], and that PRP treatment could in fact induce a dedifferentiation of chondrocytes towards a fibroblast-like phenotype [33]. This lack of consistency among the published reports may be attributable to the heterogeneity of study designs, variations in PRP preparations, and differences in PRP delivery. For instance, some studies used platelet lysate through repeated freeze-thawing after centrifugation, some employed the exudate after clot formation without addition of external activators, while others adopted PRPr collected after thrombin activation [29‐31].

PRP has also been demonstrated recently to have anti-inflammation potential in an osteoarthritic milieu. Human osteoarthritic chondrocytes were cultured with 10 ng/ml IL-1β to mimic an osteoarthritic environment [39] in medium with or without 10% PRPr. After 48 hours, IL-1β inhibited Col II and aggrecan gene expression and concomitantly increased expression of a disintegrin and metalloproteinase with thrombospondin motifs-4 and prostaglandin-endoperoxide synthase-2, whereas PRPr supplementation reduced these IL-1β-mediated effects. In addition, the IL-1β-induced activation of nuclear factor kappa B (NF-κB), a major pathway involved in the pathogenesis of OA, could be completely inhibited by PRP (P < 0.001). Further study revealed that PRPr inhibited NF-κB activation through increasing gene expression of hepatocyte growth factor (HGF) and TNF-α [40]. HGF has been shown to increase NF-κB inhibitor-α expression, thus impairing p65 translocation to the nucleus, which is necessary for NF-κB activation [41, 42]. TNF-α enhances p50 homodimer formation and its binding to DNA to inhibit NF-κB pathway activation [43]. It is noteworthy that the NF-κB pathway is not the only one involved in the PRP anti-inflammatory activities. IGF-1 and PDGF-bb present in PRP could also suppress the activation of the Src/PI3K/AKT pathway, thus inhibiting chondrocyte apoptosis and inflammation induced by IL-1β [44].

It should also be noted that some PRP formulations could be pro-inflammatory [45]. The presence of concentrated leukocytes increased the levels of catabolic and pro-inflammatory signaling molecules, including MMPs and IL-1β [46]. In addition, activated platelets could produce IL-1β to mediate pro-inflammatory signaling [47]. However, the most represented pro-inflammatory cytokine, IL-1β, only showed a slight increase after platelet activation, whereas the anti-inflammatory molecules, such as IL-4 and IL-10, increased more than five times [40]. A recent study confirmed the dual effect of platelet lysate on human chondrocytes - a transient pro-inflammatory activity followed by an inflammation resolution [48]. Although the net results of PRP are variable owing to compositional heterogeneity, the anti-inflammatory effect is likely to predominate in PRP formulations in which the presence of leukocytes is substantially reduced.

Among their various tissue sources, MSCs derived from bone marrow (BMSCs) are employed most extensively in cartilage engineering. Many researchers have found that PRP exhibits a mitogenic effect on MSCs (Table 2) [34, 52‐55]. When human BMSCs in monolayer were cultured with 10% inactivated autologous PRP, a five-fold increase in cellular proliferation was seen at day 7 relative to the control without PRP supplementation [56]. While both chondrogenic and osteogenic gene markers were up-regulated in the presence of PRP, the chondrogenic markers, including Sox9 and aggrecan, increased much more (over 10-fold increase) than RUNX2 (less than 2-fold increase) [56], a marker of early osteogenic differentiation [57]. The authors concluded that PRP could enhance the proliferation and chondrogenic differentiation of BMSCs. In this study, however, BMSCs were cultured for only 7 days, and the long-term effects of PRP on MSCs thus remained unclear. In another study with 21 days of BMSC culture, expression of Sox9, aggrecan and Col II increased significantly at both the mRNA and protein levels in the presence of inactivated PRP, compared to the FBS control [21]. However, Col I gene and protein expression was also increased concomitantly by PRP. This non-selective up-regulation of both chondrogenic and osteogenic genes by PRP in multipotent BMSCs in vitro must be considered with caution in cartilage tissue engineering, but may be beneficial and applicable for repairing osteochondral defects.

Besides BMSCs, the effect of PRP on MSCs derived from fat, muscle and subchondral bone has also been preliminarily studied. Rabbit adipose-derived MSCs responded to PRP stimulation in a manner similar to BMSCs, in that cell proliferation, gene and protein expression of Sox9, aggrecan, Col I and Col II were enhanced significantly compared to the FBS controls [21]. When muscle-derived MSCs were cultured in the presence of PRP, their ability to proliferate, adhere and migrate was significantly promoted [58]. Although chondrogenic gene expression was not up-regulated, the number of cells producing Col II was increased markedly. Meanwhile, cellular apoptosis was also increased in vitro, but in vivo study yielded contrary results showing that apoptosis was suppressed in the presence of PRP [58]. MSCs from the subchondral bone are considered the main cell sources responsible for the repair of cartilage defects in the clinical procedure of microfracture [59]. A study reported that P-PRP treatment could stimulate the vertical migration of subchondral progenitors and cartilaginous matrix accumulation, including proteoglycan and Col II [60]. More importantly, while chondrogenic differentiation of the progenitor cells was induced significantly by PRP treatment, osteogenic and adipogenic differentiation were not affected. These findings suggest that PRP might accelerate the migration of the subchondral progenitors to repair cartilage defects with the formation of hyaline cartilage.

In addition to the positive effects on MSC proliferation, differentiation and migration, PRP may also provide a three-dimensional substrate for cell seeding by virtue of the presence of fibrinogen, which gives rise readily to fibrin gel upon thrombin or calcium activation. In a recent study, about 1 × 10⁵ rabbit BMSCs were mixed with 60 μl ultra-filtered platelet lysate and the composite was then activated by thrombin and CaCl₂ to form a three-dimensional cell-laden scaffold, followed by in vitro culturing in chondrogenic induction medium for 21 days [61]. At 1 week, round-shaped chondrocyte-like cells were found homogeneously distributed inside lacunae and some cells clustered together within the scaffold. Histological analysis at 3 weeks further confirmed the presence of these chondrocyte-like cells and the accumulation of cartilaginous ECM deposition. However, the details of the structure of this scaffold were not investigated.

In another study, Kang and colleagues [22] examined the components and microstructure of L-PRF, which formed naturally during the single step centrifugation of whole blood without anti-coagulants. They found that there were two distinct zones in the PRF scaffold, the platelet zone and the fibrin zone. The marked advantage of L-PRF is the ease of the procedure and the absence of additional chemicals. Nonetheless, the PRF preparation may not allow cells to seed evenly inside, and its two-zone microstructure implies possible large variations in the release of growth factors.

In our recent study, we adopted the traditional double spinning method to prepare liquid PRP first. Then MSCs were distributed in the PRP before it was activated by CaCl₂ to form a three-dimensional scaffold. Histological and scanning electron microscopy evaluations revealed that the PRP matrix had a honeycomb microstructure with platelets aggregated along the fibrin skeleton to which MSCs adhered (Figure 2) [21]. Another pilot study demonstrated that equine BMSCs cultured in the three-dimensional PRP gel for 3 weeks resulted in enhanced cell proliferation and proteoglycan synthesis compared to those in the fibrin gel alone [3]. Unlike L-PRF, however, both the platelet lysate and the double-spun PRP activated by CaCl₂ usually have a low density of fibrin and weak polymerization, and thus dissolve quickly, similar to fibrin glue.

It was reported that commercially available fibrin gel would disintegrate within 7 days, but such a rapid degradation could be delayed by seeding cells and adding fibrinolytic inhibitors, such as tranexamic acid, aprotonin, and galardin [62‐64]. PRP-tranexamic acid gel seeded with chondrocytes could maintain stability in vitro for 4 weeks without any shrinkage, while cells remained viable and were able to migrate [64]. The seeded chondrocytes likely produced abundant ECM, in a manner commensurate with scaffold degradation [62]. Inhibition of fibrin degradation would also mean slower and more extended release of growth factors to produce better reparative results [65]. The low mechanical property of the PRP fibrin scaffold may be improved through genipin cross-linking [66], ruthenium-catalyzed photo cross-linking [67], or adjusting the content of fibrinogen [68]. An optimized fibrin gel could resist dynamic compression and shear at the tissue site, while the embedded chondrocytes continued to produce a coherent cartilaginous ECM, containing proteoglycan and Col II [66, 68]. In addition, PRP gel may also be employed with other biomaterials to enhance its mechanical properties.

Among the case reports illustrated in Table 6[92‐99], Wang-Saegusa and colleagues [99] reported the largest case series of knee OA treated with P-PRP intra-articular injections. A total of 808 patients were treated and 261 were evaluated after strictly applying the inclusion and exclusion criteria. At the end of 6 months, pain was significantly reduced while knee function and quality of life were improved, without occurrence of adverse events. The authors concluded that P-PRP injections had local, effective and temporal effects reducing pain and restoring function. However, the lack of a placebo group or conservatively managed control undermined the conclusions of these studies.

In a retrospective cohort of 30 patients sustaining chronic knee pain, the efficacy of PRP injections was compared with the more common, well-recognized HA treatment [100]. Patients received three intra-articular injections of inactivated double-spun PRP or HA every 3 weeks and were followed for 6 months after the final injections. Both groups showed significant improvement in IKDC, Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and Lequesne Index, but the PRP-treated patients exhibited better results at 6 months than the HA-treated patients. Sánchez and colleagues [101] reported a similar cohort study to compare the therapeutic effect of P-PRP and HA for treatment of knee OA. Each group included 30 patients matched according to their age, sex, body mass index and OA severity on radiography. Results were only considered successful when a reduction of at least 40% from baseline in WOMAC pain scores occurred. In the PRP group, the success rate reached 33.4% at 5 weeks compared to only 10% in the HA group (P = 0.004). In addition, with respect to HA treatment, P-PRP injections significantly reduced the physical function subscale and the overall WOMAC (P = 0.043, P = 0.010, respectively) in favor of PRP treatment.

Two recent prospective cohort studies further confirmed the superiority of PRP over HA injections for treatment of knee OA [102, 103]. Spaková and colleagues [102] reported 120 knee OA patients who were randomly treated with three injections of L-PRP or HA, one per week. At 3 and 6 months follow-up, better results in WOMAC and Numeric Rating Scale were obtained in the L-PRP group (P < 0.01). The other study included 150 patients suffering knee cartilage degenerative lesions, who were divided evenly into three groups [103]. These three homogeneous groups received three intra-articular injections of double-spun PRP, low-molecular-weight HA or high-molecular-weight HA. At 6 months follow-up, the best results in terms of IKDC, visual analogue scale and patient satisfaction were achieved in the PRP group (P < 0.005), in particular for the younger patients affected by cartilage lesions or early OA.

In an RCT comparing P-PRP and HA treatment in a total of 32 patients suffering talar osteochondral lesions [104], the mean AOFAS score was significantly improved from 68 and 66 before injection to 92 and 78 after 28 weeks in P-PRP and HA groups (P < 0.0001), respectively, favoring P-PRP treatment (P < 0.05). Better results for the visual analogue scale and other subjective function scores were also noted in the P-PRP group (P < 0.01).

A recent RCT compared PRP with placebo for the treatment of knee OA [105]. Seventy-eight patients (156 knees) were randomly divided into three groups. Group A received a single injection of P-PRP, group B received two injections of P-PRP 3 weeks apart, and group C received a single saline infiltration. After 6 months, all the subscores in WOMAC improved in groups A and B, but worsened in group C. These results support the short-term effectiveness of P-PRP injections over placebo for relieving pain and improving knee function.

The efficacy of PRP in the treatment of OA was also compared with HA administration. Cerza and colleagues [106] reported on 120 gonarthrosis patients undergoing 4 randomized intra-articular injections of P-PRP or HA. Patients in two groups were matched in terms of age, gender, severity of knee arthrosis and pre-treatment WOMAC scores. All patients were WOMAC evaluated before the infiltration and at 4, 12 and 24 weeks after the first injection. While post-treatment WOMAC scores in both groups significantly improved compared to before the infiltration, the improvement was more significant in the P-PRP-treated group than the HA group at each time point. In addition, the trend continued during the 24-week follow-up in the P-PRP group, but began to attenuate at 4 weeks in the HA group. These results indicated that P-PRP had a stronger and longer effect on the attenuation of OA with respect to HA treatment. In a multicenter, double-blind RCT, a total of 176 patients with symptomatic knee OA were randomly assigned to receive P-PRP or HA infiltrations [107]. The groups were well balanced for age, gender, body mass index, percentage of patients with primary arthritis, daily consumption of analgesics, radiographic grade, and WOMAC and Lequesne scores. The primary outcome measure was a 50% decrease in knee pain from baseline to week 24; according to this, the rate of response was 14.1% higher in P-PRP-treated patients compared to the HA-treated group. Regarding the secondary outcome measures assessing pain, stiffness and physical function, PRP also yielded better results than HA, albeit not reaching significance. Another recent RCT confirmed the superiority of P-PRP over HA in the alleviation of knee pain and stiffness and the improvement of physical function at both 24 and 48 weeks [108].

On the other hand, a single-center, double blind RCT including 109 matched patients demonstrated that PRP treatment did not lead to statistically significant differences in all scores evaluated with respect to HA injections at 12-month follow-up [109]. Further analysis showed a tendency favoring PRP in patients with less degenerated joints at 6 months and 12 months, although no significant difference was reached (P = 0.08 and P = 0.07, respectively). However, unlike the aforementioned RCTs, which used fresh P-PRP, this trial prepared PRP manually by double-spinning followed by freezing and thawing. Although the accurate concentration of leukocytes was unreported, it was estimated that this preparation concentrated leukocytes together with platelets and the final product was likely to contain much higher levels of pro-inflammatory signaling cytokines than P-PRP [46, 110].

The unfavorable effect of concentrated leukocytes in PRP is also reflected in post-injection reactions. After intra-articular P-PRP injections, undetectable or only mild, self-resolved adverse events were observed, comparable to that observed with HA administration [106‐108], but the double-spun PRP induced a significantly higher rate of pain reaction than the HA treatment (P = 0.039) [109]. A RCT comparing single- and double-spun PRP confirmed that the latter produced more pain and swelling reaction than the former, in which leukocytes were less concentrated [111].