Aging Affects the Efficacy of Platelet-Rich Plasma Treatment for Osteoarthritis

Objective

Despite the increased use of platelet-rich plasma in the treatment of osteoarthritis, whether and how age of the platelet-rich plasma donor affects therapeutic efficacy is unclear.

Design

In vitro, male osteoarthritic human chondrocytes were treated with platelet-rich plasma from young (18–35 yrs) or old (≥65 yrs) donors, and the chondrogenic profile was evaluated using immunofluorescent staining for two markers of chondrogenicity, type II collagen and SOX-9. In vivo, we used a within-subjects design to compare Osteoarthritis Research Society International scores in aged mouse knee joints injected with platelet-rich plasma from young or old individuals.

Results

In vitro experiments revealed that platelet-rich plasma from young donors induced a more youthful chondrocyte phenotype, as evidenced by increased type II collagen (P = 0.033) and SOX-9 expression (P = 0.022). This benefit, however, was significantly blunted when cells were cultured with platelet-rich plasma from aged donors. Accordingly, in vivo studies revealed that animals treated with platelet-rich plasma from young donors displayed a significantly improved cartilage integrity when compared with knees injected with platelet-rich plasma from aged donors (P = 0.019).

Conclusions

Injection of platelet-rich plasma from a young individual induced a regenerative effect in aged cells and mice, whereas platelet-rich plasma from aged individuals showed no improvement in chondrocyte health or cartilage integrity.

What Is Known

• Platelet-rich plasma (PRP) is a common—albeit polarizing—treatment of many musculoskeletal conditions, including osteoarthritis. Despite equivocal evidence in the literature, PRP elicits benefit in a subset of patients. The controversy is, however, confounded, as patient characteristics that may contribute to the efficacy of PRP have not been studied.

What Is New

• Through in vitro and in vivo studies, we demonstrated that PRP derived from young individuals demonstrates a rejuvenative effect, as evidenced by increased chondrogenicity and improved cartilage integrity. However, our studies revealed that age attenuated the beneficial effect of PRP on chondrocyte health.

Platelet-rich plasma (PRP) is a highly sought treatment in multiple fields, from cardiac surgery to dermatology, for its vast clinical applications in stimulating tissue healing in multiple anatomical areas.¹ However, it remains controversial in its acceptance by the musculoskeletal field. Platelet-rich plasma is defined as an autologous mixture of highly concentrated platelets derived from whole blood.² It has been proposed that PRP provides a regenerative stimulus, as it acts as a reservoir of concentrated circulating factors, such as growth factors and cytokines, that target the healing cascade to reduce inflammation, promote proliferation, and support tissue remodeling.

Given its use among professional athletes, PRP has especially garnered a strong interest in the field of musculoskeletal medicine through widespread media attention.³ However, its applications have not been limited to the young and elite sportsperson, as PRP has been used widely for the treatment of degenerative cellular and tissue processes, particularly osteoarthritis (OA).⁴ Osteoarthritis is one of the most common causes of disability worldwide and is frequently implicated in the loss of physical mobility in aging individuals.⁵ Age-related musculoskeletal decline significantly hinders the ability of individuals to live independently and maintain an acceptable quality of life.⁶ The degenerative process of OA leads to pathological changes in several musculoskeletal tissues including cartilage, meniscus, ligament, and synovium.⁷ In the treatment of OA, PRP is typically injected into inflamed tissue with the goal of improving cartilage structure and function.^4,8

Recently, two double-randomized clinical trials investigated the efficacy of PRP for the treatment of OA in the knee (average age 62 yrs) and ankle (average age 56 yrs). The trials recruited individuals with pain secondary to the diagnosis of OA and evaluated pain on an analog scale at 12-mo and 26-wk time points, respectively. The trials found that intra-articular PRP injections did not induce a significant difference in symptoms, structure, or function in the ankle or knee when compared with placebo injections.^9,10 The results from these studies, however, were met with concerns, most of which focused on the fact that the PRP preparation protocol was not one that is standardly used in clinic, thereby raising concerns about insufficient platelet numbers.^11,12 Despite the many concerns raised, one question remained both unasked and unanswered: Do the patient characteristics play a role in the efficacy of PRP?

Given the autologous nature of the PRP, it is unclear whether and how individual patient characteristics, and most notably age, affect the efficacy of the treatment. This is surprising in the light of pivotal studies over the last two decades that have rigorously and repeatedly demonstrated that circulating factors from young animals play a supportive role in tissue regenerative cascades, but that the benefit is lost with aging.¹³ The rejuvenating effects of “young” circulatory factors are demonstrated through the heterochronic parabiosis model, where two animals are surgically sutured to develop a single shared circulatory system. In this model, the systemic effects of a young circulation can be evaluated on an old animal and vice versa. In one study that implemented this model, investigators found that old chondrocytes displayed a rejuvenated regenerative capacity when exposed to a young circulation.¹⁴ Beyond heterochronic parabiosis, even murine studies using heterochronic blood exchange, where an old animal is transfused with young blood and vice versa, have demonstrated that circulating factors play a crucial role in the regulation of tissue homeostasis and function.^13,15 These studies indicate that the systemic environment has a pervasive effect on organ and tissue function.

Although PRP has been studied extensively, there is a significant gap in the literature on the study of patient characteristics, such as age, which we posit contributes to the conflicting reports. To test our hypothesis, we used a murine model of age-associated OA to evaluate whether the beneficial impact of PRP on cartilage integrity is dependent on age of the PRP donor.

METHODS

Chondrocyte Isolation and Culture

Isolation

Frozen, uncultured (P0, passage 0) cadaveric human chondrocytes were purchased from an external vendor (StemBioSys, San Antonio, TX). Briefly, chondrocytes were isolated from the femurs of young and healthy (Y/H), old and healthy (O/H), and old and osteoarthritic (O/OA) males after obtaining consent from a relative of the deceased individual. To minimize inclusion of immune cells in the chondrocyte samples, O/OA chondrocytes were extracted from the tissue adjacent to, but not directly over, macroscopically visible lesion. Osteoarthritis status was confirmed by microscopic inspection based on a scale modified from Outerbridge (1961).¹⁶ Nonosteoarthritic articular chondrocytes were isolated from donors with no suspected OA and minimal fibrosis in the knee joint (≤ grade 2 OA). Osteoarthritic articular chondrocytes were isolated from donors with overt fibrosis and/or erosion/major lesions in the knee joints (≥ grade 3 OA). After isolation, cells were cryopreserved at ≥1 × 10⁶ cells/vial. A certificate of analysis included donor serology, microplasma, and endotoxin results. We received the samples blinded from the vendor, and as such, institutional board review was not required and consent was waived.

Cell Culture

Frozen chondrocytes were thawed at 37°C for 1 min. One milliliter (ml) of chondrocyte culture medium (Dulbecco’s Modified Eagle Medium low glucose medium [Millipore Sigma-D 6046] with 15% fetal bovine serum [Gibco, 10468-26], 1% penicillin-streptomycin [Gibco 15140-122], and 1% Glutamax [Gibco-3505006]) was added to thawed chondrocytes, and the cells were transferred to a 15-ml tube with 10 ml of medium. The tube was then centrifuged for 5 mins at 1300g. Once the pellet was visualized, the medium was aspirated, and the chondrocytes were resuspended in fresh chondrocyte medium. Chondrocytes were then seeded in a T150 flask with an additional 30 ml of culture medium and incubated the cells at 37°C, 5% CO₂. The chondrocyte medium was changed every 2–3 days until the chondrocytes reached approximately 70% confluence (approximately 7–14 days). Brightfield images of the cells were then obtained to confirm visualization of typical chondrocyte stellate morphology, and purity of the samples was confirmed by an investigator upon receipt of the samples. Cells were dissociated with 0.25% trypsin for 2 mins, seeded at 5000 cells/well onto 12-well polycarbonate transwells (Thermo Fisher Scientific, Inc, Waltham, MA) and grown for 48 hrs before treatment.

Platelet-Rich Plasma Procurement

Handling of PRP

Platelet-rich plasma from young (18–35 yrs) and old (older than 65 yrs) males (n = 6 per group) was obtained from an external vendor (Innovative Research, Novi, MI). Samples were shipped frozen on dry ice and subsequently transferred to a −80°C freezer to avoid thawing and activation of the platelets in the PRP until further use.

Platelet-Rich Plasma Preparation Protocol

The PRP samples were collected using a previously published centrifugation protocol, yielding a sample that is similar to what is obtained using the commercially available PRP kits—GPSIII, SmartPrep, and Magellan.^17,18 Whole blood was collected in a citrated anticoagulant and immediately processed for PRP without chilling the blood at any time using previously published method.¹⁸ The blood was first centrifuged using a “soft” spin (400g for 10 mins, 4°C), yielding the following three layers: RBCs, buffy coat, and platelet fraction (PRP 1). The supernatant plasma containing platelets (PRP 1) was then transferred into another sterile tube (without anticoagulant). The PRP 1 fraction was then centrifuged using a “hard” spin (2000g–3000g for 20 mins), yielding two layers: the lower one third is PRP and upper two third is platelet-poor plasma. The PRP layer was then extracted and frozen for transport. Platelet counts were not obtained by the external vendor; however, the PRP preparation protocol used in these experiments has been frequently analyzed and has been shown to yield approximately 7 billion platelets, which exceeds United States Food and Drug Administration’s minimum platelet concentration for PRP and meets Marx’s definition (i.e., 2–3 times higher platelet concentrations compared with whole blood).^17–19

Given our PRP samples arrived frozen, we completed preliminary studies comparing fresh and frozen PRP (Supplemental File "Fresh vs Frozen PRP", Supplemental Digital Content 3, http://links.lww.com/PHM/B937), as the PRP used in the clinical setting is fresh and typically injected within minutes of extraction. Methodological details are described in the Supplemental Figure 1 (Supplemental Digital Content 1, http://links.lww.com/PHM/B901). We found no significant differences in the collagen II expression of O/OA chondrocytes treated with fresh or frozen PRP derived from young healthy individuals (P = 0.15).

In Vitro Co-culture Experiments

The O/OA human chondrocytes were cultured in the presence of young or old PRP. We administered 10% PRP (diluted in cell culture media) to the bottom of the transwells plated with chondrocytes for period of 4 days in vitro.

Evaluation of Chondrogenic Profile-Immunofluorescence Staining

Transwells seeded with cells were fixed with warm 2% paraformaldehyde for 15 mins and then washed with phosphate-buffered saline (PBS) 3 times. The samples were then permeabilized with 0.1% Triton-X for 15 mins after which the samples were blocked with 3% bovine serum albumin, and 0.1% Triton-X for 45 mins. Membranes were removed from the transwells using a scalpel, and the membranes were then sectioned into four pieces. The samples were incubated overnight at 4°C with primary antibodies against markers of chondrogenicity in an antibody solution made out of 3% +5% goat serum +0.1% Triton-X at the dilutions noted in Table 1. The cells were then assayed for chondrogenic profile within 3 days of fixation.

TABLE 1.

Primary antibodies used in immunofluorescence staining

Primary Antibody	Host Species	Product Number	Dilution
Type II collagen	Rabbit	Abcam ab-34712	1:500
SOX-9	Rabbit	Cell Signaling 82630	1:300

After incubation overnight, membranes were washed with PBS 5 times for 2 mins while maintaining the membranes on a rocker in between washes. Membranes were then stained with host- and isotype-specific secondary antibodies for 1 hr at the dilutions noted in Table 2.

TABLE 2.

Secondary antibodies used in immunofluorescence staining

Antibody	Host Species	Product Number	Dilution
AlexaFluor 594	Rabbit	Life Technologies, A11012	1:400
AlexaFluor 488	Rabbit	Life Technologies, A11034	1:500

The membranes were then washed with PBS 5 times for 2 mins each. Lastly, membranes were stained with DAPI (Invitrogen D1306) at 1:500 dilution in PBS for 2 mins, followed by three PBS washes for 2 mins each. Membranes were then mounted on the glass slides with Gelvatol (Source: Center for Biologic Imaging, University of Pittsburgh), and images were taken on Zeiss Axiovision microscope. A blinded investigator quantified the degree of immunohistochemical labelling for each of the dependent variables by quantifying protein intensity per cell across experimental groups.

In Vivo Experiments: Osteoarthritis Research Society International Scoring and Synovial Membrane Analysis

All animal experiments were performed with previous approval from the Institutional Animal Care and Use Committee of the University of Pittsburgh. Aged mice (C57/BL6; 21–24 months old; male; NIA colony) received PRP injections using 6 of the 12 PRP samples used in the in vitro studies above (three young and three aged samples). To mitigate the effects of between-animal variability, we used a within-subjects design (n = 20; 12 experimental, 8 control). In the experimental group, each mouse received injections of one of the three young PRP samples into the right knee and an injection of one of the three aged PRP samples into the left knee. Four biological replicates were analyzed for each sample. In the control group, each mouse received a saline injection into the right knee, while the left knee remained untreated.

For PRP administration in mice, we developed an intra-articular knee injection protocol using a lateral injection approach to mimic that standard of care in humans. Each mouse knee was held in a slightly flexed position and a 31-gauge needle was then introduced anterolaterally beneath the patellar ligament and into the intra-articular space. Ten microliter of either saline or PRP was injected into the knee joint. The knee was then cycled through flexion and extension 3 times, and the mouse was returned to its cage (Supplemental Fig. 2, Supplemental Digital Content 2, http://links.lww.com/PHM/B902). Before the start of experimental studies, two independent investigators confirmed the intra-articular localization of the injection using Trypan Blue dye injections into 12 mice.

Fourteen days after the injections, mice were euthanized. Knee joints were harvested, fixed, and decalcified before paraffin embedding. Paraffin-fixed knee joints were subsequently sectioned and stained with Safranin O/fast green/hematoxylin to evaluate cartilage integrity using the Osteoarthritis Research Society International (OARSI) score. Statistical outliers in the OARSI scores as determined by the JMP statistical software were excluded from this study. Synovial thickness of each knee joint was measured to assess for synovial inflammation using ImageJ software (National Institutes of Health). All analyses were performed by an investigator blinded to the treatment group.

Statistical Analysis

For the in vitro experiments, a nonparametric Kruskal-Wallis analysis of variance with Dunn's multiple comparison test was used to compare means between three or more groups. When comparing two groups, a nonparametric Mann Whitney U test was performed. For the in vivo experiments, a two-sided paired t test (young PRP vs. old PRP) or two-sided unpaired t test (young PRP vs. saline, old PRP vs. saline) was used to compare means between groups. A linear mixed effect model was used to assess the relationship between OARSI score and synovial thickness. P < 0.05 was considered to be significant.

RESULTS

In Vitro Studies

Baseline Demographics/Characteristics of Human Chondrocytes and PRP

Human cadaveric male chondrocytes were obtained from donors aged 19 yrs (Y/H), 68 yrs (O/H), and 73 yrs (O/OA; Table 3).

TABLE 3.

Baseline characteristics of articular chondrocytes

Description	Passage	Cells/Vial	Donor Age
CELLvo human chondrocytes-articular (HC-a), P0; <40 age; healthy	P0	1,000,000	19
CELLvo human chondrocytes-articular (HC-a), P0; 50–70 age; healthy	P0	1,000,000	68
CELLvo human chondrocytes-articular with OA (HC-a), P0; 50–70 age	P0	1,000,000	73

Platelet-Rich Plasma

Platelet-rich plasma samples derived from young individuals were obtained from donors aged 18 to 35 yrs, while PRP samples derived from older individuals were obtained from male donors 65 yrs or older. Race, age, and sex of the donors are listed in Table 4.

TABLE 4.

Baseline properties of PRP isolated from male humans

PRP Descriptor	Race	Donor Age
Young_1 (Y1)	Hispanic	30
Y2	Hispanic	26
Y3	Black	32
Old_1 (O1)	Caucasian	71
O2	Hispanic	65
O3	Hispanic	66

In Vitro Co-culture Cell Models

We quantified type II collagen and SOX-9, two established markers of chondrogenicity.^20,21 Consistent with previous reports,^22,23 we found O/H chondrocytes displayed a modest decrease in type II collagen and SOX-9 expression when compared with young counterparts (Figs. 1A–C). The O/OA chondrocytes displayed a significantly reduced type II collagen and SOX-9 expression when compared with Y/H counterparts (Figs. 1A–C).

FIGURE 1.

Osteoarthritic cells displayed reduced expression of chondrogenic markers. A, Immunofluorescent imaging of collagen II (green), SOX-9 (red), and nuclei (DAPI, blue) in Y/H, O/H, and O/OA chondrocytes. Scale: 50 μm. Quantification of (B) collagen II protein and (C) SOX-9+ cells (%) across the groups (*P < 0.05, nonparametric Kruskal-Wallis analysis of variance with Dunn's multiple comparison test, n = 3 wells/group).

When O/OA cells were cultured in the presence of young versus aged PRP, we found that administration of young PRP on O/OA cells rejuvenated the chondrogenic profile of the treated cells, as evidenced by increased type II collagen and SOX-9 expression (Figs. 2A, B). However, the benefit was attenuated in the presence of PRP derived from aged donors (Figs. 2A, B).

FIGURE 2.

Administration of PRP had an age-dependent effect on chondrogenicity of O/OA cells. A, Immunofluorescence imaging and quantification of collagen II expression in old and young PRP-treated cells compared with expression profile of O/OA cells. B, Immunofluorescence imaging and quantification of SOX-9 expression in old and young PRP-treated cells compared with expression profile of O/OA cells. Scale: 50 um.

Platelet-Rich Plasma Injections In Vivo Using a Mouse Model

The data thus far suggest that PRP derived from young, but not old, male donors enhances the chondrogenic profile of O/OA cells. To better understand the physiological relevance of these findings, we administered PRP to the knees of aged male mice. Consistent with in vitro data, in vivo findings revealed that knees treated with young male PRP displayed significantly lower OARSI scores and decreased cartilage surface roughness when compared with knees that received old male PRP (Fig. 3). We additionally found that PRP derived from aged male donors demonstrated increased cartilage surface disruption and numerous chondrocyte lacunes (Fig. 3). Young, but not aged, PRP also displayed a significantly decreased OARSI score compared with saline controls (P = 0.042 for young vs. saline; P = 0.922 for aged vs. saline; unpaired two-tailed Student t test). Of note, there were two outliers that had OARSI scores greater than two SD above the mean, indicating high-grade/high-stage OA. These animals were not included in the analyses, because it is well established that PRP is not used in the clinical setting for individuals with severe OA.

FIGURE 3.

Treatment with young but not old PRP enhanced cartilage integrity in osteoarthritic knees. A, Schematic of experimental design. B, Histological imaging of Safranin O–stained knee sections from young and old PRP-injected knees. Quantification of (C) degree of OA progression (OARSI scores) and (D) cartilage surface roughness. Scale: 50 μm, inset scale: 20 μm.

Whereas treatment with young PRP did not affect synovial thickness as compared with saline counterparts, the synovium of knees treated with aged PRP displayed increased synovial membrane thickness, suggesting increased inflammation (Fig. 4). We fit the synovium thickness data and OARSI scores into a linear mixed effect model and found that the synovial thickness and OARSI scores were statistically correlated with young PRP-treated knees displaying lower OARSI scores and less synovium thickness (P < 0.001; Fig. 4C). These results suggest that not only does treatment with old PRP fail to elicit a chondro-protective, but it may also provoke an inflammatory response.

FIGURE 4.

A single injection of old PRP increased synovium inflammation in OA knees. A, Histological imaging of fast green and hematoxylin-stained knee sections from young and old male PRP-injected knees. Quantification of inflammation of the synovium (synovium thickness). Scale: 20 μm. C, Linear mixed effect model correlating synovium thickness and OARSI scores.

DISCUSSION

Platelet-rich plasma remains controversial in many circles. Given the overwhelming rise in clinical use of PRP and anecdotal data indicating improvement in function and pain, PRP has tremendous clinical potential if used in the appropriate patient. In current literature, one of the biggest criticisms for PRP as a treatment is the lack of standardization of the PRP preparation protocol, which can be evidenced by the multiple PRP classification systems that exist, including the Dohan Ehrenfest, Mishra, and PAW classifications, to name a few.^24,25 Much less considered is the impact of patient characteristics, such as sex, activity level, and, of course, age. This led us to the question: Is all PRP the same? and, more importantly: Does age attenuate the effects of PRP? Through a series of in vitro and in vivo studies designed to be a first step to answering these questions, we found that young, but not old, PRP induced restorative properties in osteoarthritic chondrocytes by generating a chondrogenic profile that closely resembled young and healthy chondrocytes. In addition, in vivo studies revealed that injection of young PRP induced a chondro-protective effect in aged mice. Although PRP injection is a popular OA treatment used in the clinic, we found that the efficacy of PRP is attenuated with age and, paradoxically, may even provoke an inflammatory response.

Upon a closer look at the demographics of the two recent PRP clinical trials on knee and ankle OA, consideration of patient characteristics in the efficacy of the treatment is notably lacking. The mean age of patients in the knee OA trial was 61.9 yrs and the mean age of patients in the ankle OA trial was 56 yrs.^9,10 To the best of our knowledge, there are no equivalent comparators in the literature evaluating the efficacy of PRP in a younger population. A literature search revealed only one additional study evaluating the effects of age on PRP.²⁶ Specifically, investigators evaluated the composition of inflammatory mediators and growth factors within leukocyte-poor PRP in young and healthy males as compared with older males with severe knee OA. Gene expression profiling of cells revealed that treatment with PRP derived from older individuals not only suppressed chondrocyte matrix synthesis but also promoted macrophage inflammation in vitro as evidenced by upregulation of tumor necrosis factor α and matrix metallopeptidase 9.²⁶ This is consistent with our in vivo data suggesting that PRP provokes an inflammatory response, as indicated by thickening of the synovial membrane and invasion by immune cells.

In addition to age, which is implicated in the pathogenesis of OA, other patient characteristics, such as adiposity and sex, should also be considered in patient characteristics that require evaluation before PRP injection. Increased adiposity is correlated with an increased incidence of OA. Indeed, murine preclinical studies have previously demonstrated that adiposity has detrimental effects on cartilage health and integrity independent of mechanical impacts.²⁷ In a German working group position statement, experts in the field found that PRP treatments of younger patients with a lesser severity of OA and lower body mass index tended to have better patient-reported outcomes.²⁸ Previous studies have also demonstrated significant differences in the composition of PRP between men and women with evaluation of cytokine and growth factor levels.^29,30 Male patients had higher cytokine and growth factor levels in their PRP compared with female counterparts, including both inflammatory and anti-inflammatory cytokines. Given the complex interplay between inflammatory (interleukin 1β, tumor necrosis factor α) and anti-inflammatory cytokines (interleukin-1 receptor antagonist protein) in the healing cascade, sex may play an important role in the efficacy of PRP. Accordingly, older women tend to be at a higher risk for OA, which may be more difficult to treat with autologous PRP.

There are limitations to the current studies that should be noted. The sample size of PRP tested was small. To account for this, we used the same samples in in vitro and in vivo experiments, thereby enhancing the rigor of the work. In addition, we used multiple biological replicates of the PRP samples in our in vivo studies. Given that our samples arrived from our external vendor frozen on dry ice, we were unable to obtain platelet concentrations in the PRP. However, as noted previously, the PRP preparation protocol used in these experiments has been frequently analyzed and has been shown to yield approximately 2–3 times higher platelet concentrations compared with whole blood. Finally, this study used a freeze-thaw method, as we purchased commercially available PRP for these experiments. While our studies comparing fresh and frozen PRP revealed no significant differences between groups when comparing the effects on chondrocyte health, future studies would ideally use freshly isolated PRP to more closely mimic clinical protocols.

CONCLUSIONS AND FUTURE DIRECTIONS

Our findings suggest that age may be an important patient characteristic determining the efficacy of PRP in the treatment of OA. Therefore, evaluation of patient characteristics, such as age, deserves further investigation as predictors for the efficacy of PRP treatments in OA. Taken together, our findings suggests that young, but not old, human PRP rejuvenates O/OA human chondrocytes in vitro and O/OA murine chondrocytes in vivo, ultimately enhancing cartilage integrity. Furthermore, we found that PRP derived from older individuals may paradoxically cause harm in the knee joint.

More studies are needed to understand factors in the age-dependent effects of PRP in mitigating OA. Future work will assess the potential for long-term regenerative effects of PRP in osteoarthritic joints. In addition, by establishing a mechanism of action for PRP in the future, we hope to develop strategies to augment the efficacy of autologous PRP. Finally, we anticipate that the findings from these studies may lay the groundwork for the development of future testing platforms in which clinicians can evaluate the efficacy of PRP before injection. Such a platform would allow clinicians to assess the quality of PRP on a case-by-case basis to determine whether PRP is the most appropriate treatment. Accordingly, clinicians may be able to target the appropriate patient with the appropriate PRP formulation for OA.