ARTHRITIS & RHEUMATISMVol. 50, No. 4, April 2004, pp 1216–1225DOI 10.1002/art.20149© 2004, American College of Rheumatology

Mediation of Spontaneous Knee Osteoarthritis by ProgressiveChondrocyte ATP Depletion in Hartley Guinea Pigs

Kristen Johnson,1 Camilla I. Svensson,1 Deborah Van Etten,1 Soumitra S. Ghosh,2

Anne N. Murphy,2 Henry C. Powell,1 and Robert Terkeltaub1

Objective. Because articular chondrocytes residein a hypoxic milieu, anaerobic glycolysis is central ingenerating ATP to support chondrocyte matrix synthe-sis and viability, with mitochondrial oxidative phos-phorylation possibly providing physiologic reserve ATPgeneration. Nitric oxide (NO) potently suppresses mito-chondrial oxidative phosphorylation. Because enhancedcartilage NO generation occurs in osteoarthritis (OA),we systematically tested for mitochondrial dysfunctionin the pathogenesis of OA.

Methods. We assessed chondrocytes for ATP de-pletion and for in situ changes in mitochondrial ultra-structure prior to and during the evolution of sponta-neous knee OA in male Hartley guinea pigs, a model inwhich chondrocalcinosis also supervenes.

Results. Spontaneous NO release from knee car-tilage samples in organ culture doubled between ages 2months and 8 months as knee OA developed. Concom-itantly, chondrocyte intracellular ATP levels declined by�50%, despite a lack of mitochondrial ultrastructureabnormalities in knee chondrocytes. As ATP depletionprogressed with aging in knee chondrocytes, an in-creased ratio of lactate to pyruvate was observed, con-sistent with an adaptive augmentation of glycolysis tomitochondrial dysfunction. Furthermore, we observedprogressive elevation of chondrocyte ATP-scavenging

nucleotide pyrophosphatase/phosphodiesterase (NPP)activity and extracellular levels of the NPP enzymaticend product inorganic pyrophosphate (PPi), whichstimulate chondrocalcinosis.

Conclusion. Profound chondrocyte ATP depletiondevelops in association with heightened NO generationin guinea pig knee OA. Increased NPP activity andconcordant increases in extracellular PPi, which arestrongly associated with human aging-associated degen-erative arthropathy and directly stimulate chondrocal-cinosis, may be primarily driven by chondrocyte ATPdepletion. Our findings implicate a decreased mitochon-drial bioenergetic reserve as a pathogenic factor in bothdegenerative arthropathy and chondrocalcinosis in ag-ing.

Articular cartilage chondrocytes must survive andmaintain tissue integrity in an avascular, hypoxic milieu,which requires adaptively increased anaerobic glycolysisto support ATP synthesis (1–7). However, oxygen doesdiffuse into articular cartilage, and articular chondro-cytes possess mitochondria and respire in vitro (8).Furthermore, mitochondrial oxidative phosphorylation(OXPHOS) may account for up to 25% of the ATPproduced in articular cartilage (6).

Physiologic generation of ATP by chondrocytessupports chondrocyte viability and collagen synthesis, aswell as the formation of phosphoadenosine sulfate,which is involved as a sulfate donor in the synthesis ofsulfated proteoglycans (PGs) (1,2,9). Hence, it has beenhypothesized that impaired mitochondrial ATP synthe-sis is a factor in the pathogenesis of OA (1). Observa-tions supporting this notion include, first, the capacity ofdirect inhibition of the mitochondrial electron transportchain or of mitochondrial ATP synthase in culturedchondrocytes to depress collagen and PG synthesis (2).Second, nitric oxide (NO), a pathogenic factor in OA,suppresses chondrocyte OXPHOS (2), an effect that is

Dr. Terkeltaub’s work was supported by grants from theDepartment of Veterans Affairs, the University of California Biotech-nology STAR program, and the NIH (P01-AG-07996, AR-47347). Dr.Powell’s work was supported by a grant from the NIH (5-R01-NS-18715).

1Kristen Johnson, BA, CA, Camilla I. Svensson, MSc, Debo-rah Van Etten, BS, Henry C. Powell, MD, Robert Terkeltaub, MD:Veterans Affairs Medical Center, and University of California, SanDiego; 2Soumitra S. Ghosh, PhD, Anne N. Murphy, PhD: Mitokor,San Diego, California.

Address correspondence and reprint requests to Robert Ter-keltaub, MD, VA Medical Center, 3350 La Jolla Village Drive, SanDiego, CA 92161. E-mail: rterkeltaub@ucsd.edu.

Submitted for publication August 15, 2003; accepted inrevised form December 22, 2003.

1216

enhanced under low oxygen tension (3). Third, signifi-cant impairment of the function of both complex IIand complex III of the mitochondrial electrontransport chain was observed in knee chondrocytescultured from human subjects with advanced OA (4).Fourth, chronic inflammation and NO can promotechronic oxidant stress in chondrocytes that appears tohave a significant functional impact in experimental OA(10–13). Last, oxidant stress–mediated damage to chon-drocyte DNA within OA cartilage has been reported(14), and oxidant damage to mitochondrial DNA andproteins appears to mediate a variety of degenerativediseases, consonant with the “mitochondrial hypothesisof aging” (15–17).

Mitochondrial disorders are clinically hetero-geneous and can involve such organ systems as thecentral and peripheral nervous systems, eye, skeletalmuscle, heart, liver, and kidney (17). Distinctive ultra-structural changes represent a key morphologic featurein certain metabolic and degenerative disorders, such asmitochondrial myopathy and inclusion body myositis(18,19). In these conditions, paracrystalline intramito-chondrial aggregates are observed under electron mi-croscopy, chiefly in the subsarcolemmal region (18,19).In contrast, biochemical investigations have identifieddisorders of mitochondrial metabolism without any evi-dence of morphologic alterations (20,21). Scattered ob-servations of OA cartilage have suggested that mito-chondrial swelling and loss of mitochondria may bestructural features of chondrocytes in situ late in the OAprocess (22–24). But, a systematic controlled study ofchondrocyte ATP levels or of mitochondrial ultrastruc-ture and numbers during OA evolution has not beenperformed.

Our primary objective in this study was to sys-tematically test for ATP depletion and mitochondrialalterations in knee cartilage chondrocytes of Hartleyguinea pigs, which spontaneously develop OA, with theearliest lesions beginning by 3–4 months of age (25–29).Significantly, Hartley guinea pigs also spontaneouslydevelop pathologic meniscal cartilage calcification by 6months of age (30), akin to the development of chon-drocalcinosis in human aging and OA cartilage (31,32).Thus, our secondary objectives included testing forpotential adaptive changes to chondrocyte ATP deple-tion that promote chondrocalcinosis, including an in-crease in ATP-scavenging nucleotide pyrophosphatase/phosphodiesterase (NPP) activity that generatesinorganic pyrophosphate (PPi) and promotes calcifica-tion in aging and OA cartilage (31,32). Our findingssuggest that substantial decrements in mitochondrial

ATP generation contribute to the pathogenesis of OAand to specific metabolic alterations in OA and agingchondrocytes that promote chondrocalcinosis.

MATERIALS AND METHODS

Reagents. Recombinant human interleukin-1� was ob-tained from R&D Systems (Minneapolis, MN). All otherreagents were obtained from Sigma (St. Louis, MO), unlessindicated otherwise.

Cartilage tissue collection, chondrocyte isolation, andculture. Hartley guinea pigs ages 2, 4, 6, and 8 months werepurchased from Charles River (Wilmington, MA). The ani-mals were euthanized, and after dissection of each hind knee,uncalcified cartilage was removed from the tibial plateau, andsamples were taken from 6 individual zones, using a samplingstrategy based on a previously described method (26). Sliceswere cut from the medial and lateral central hyaline cartilagezones not covered by menisci (width � length 2 � 3 mm), themedial and lateral peripheral hyaline cartilage zones coveredby menisci (1 � 2.5–3 mm), the anterior and posterior peri-pheral hyaline cartilage zones covered by menisci (0.5–1.5 �0.5–1.5 mm), and the medial and lateral menisci (3 � 3 mm).The thickness of all slices was �1 mm. Each sample ofcartilage obtained from these zones, including the menisci, waswashed twice with Dulbecco’s modified Eagle’s medium(DMEM) (high glucose), supplemented with 1% fetal calfserum (FCS), 1% L-glutamine, 100 units/ml of penicillin, and50 �g/ml of streptomycin (Omega Scientific, Tarzana, CA).

Further processing of cartilage samples for experi-ments on isolated cells was initiated by digestion for 4 hours inDMEM (high glucose) containing 1% FCS, 1% L-glutamine,100 units/ml of penicillin, 50 �g/ml streptomycin, and 2 mg/mlof collagenase (Worthington, Lakewood, NJ) at 37°C. Thechondrocytes were centrifuged at 800g for 5 minutes, washedtwice, and plated in growth medium consisting of DMEM(high glucose) containing 10% FCS, 1% L-glutamine, 100units/ml of penicillin, and 50 �g/ml of streptomycin. Approxi-mately 1 � 106 cells per guinea pig hind knee were obtainedinitially. These cells were grown for 8–10 days (�3-fold–greater number) and used in the first passage for the biochem-ical studies described below. These studies were done atconfluence in medium composed of DMEM (high glucose)containing 1% FCS, 1% L-glutamine, 100 units/ml of penicillin,and 50 �g/ml of streptomycin.

Electron and light microscopy and evaluation of mito-chondrial ultrastructure. The cartilage was removed from thetibial plateau of the left hind knee of 2 male Hartley guineapigs and divided into medial and lateral sections. Forelectron microscopy, only cartilage from the medial tibialplateau (zone 1 in data presented below) was studied, and thesamples were placed in modified Karnovsky’s fixative (2%paraformaldehyde, 2.5% glutaraldehyde in 0.1M sodium caco-dylate buffer, pH 7.4) for 18 hours at 4°C followed bytreatment in 1% OsO4 in 0.1M sodium cacodylate buffer, pH7.4. Samples were subsequently dehydrated using a gradedseries of ethanol solutions followed by propylene oxide andinfiltration with epoxy resin (Scipoxy 812; Energy Beam Sci-

ATP DEPLETION IN OSTEOARTHRITIS 1217

ences, Agawam, MA). After polymerization overnight at 65°C,thin sections of the superficial to middle zones of the cartilagewere cut and then stained with uranyl acetate (4% uranylacetate in 50% ethanol) followed by bismuth subnitrate. Gridsthat contained both superficial and middle zones of the medialcartilage were examined at an accelerating voltage of 60 kVusing a Zeiss EM10B electron microscope (Zeiss, Wetzlar,Germany).

For light microcopy, the right hind knee of the sameanimals used for the electron microscopy studies was fixed for7 days in Streck tissue fixative (Streck Laboratories, Omaha,NE), a proprietary, non–formalin-based fixative that includesdiazolidinyl urea and 2-bromo-2-nitropropane-1,3-diol, decal-cified overnight in 20% formic acid, and processed using astandard paraffin technique. The 6-�m sections were stainedwith hematoxylin and eosin.

The ultrastructural features of mitochondria and theoverall appearance of chondrocytes in the medial tibial plateauwere examined in a minimum of 20 cells/section for each of theanimals (3 sections per animal), with analyses at ages 2, 4, 6,and 8 months. Cells considered suitable for evaluation werethose in which the section passed through the nucleus and inwhich no part of the cell was covered by the grid bars. Anelectron micrograph was taken of each cell. The mitochon-drion was graded as ultrastructurally abnormal when inclusionbodies or crystalline profiles were present within the organelle.The numbers of mitochondria per chondrocyte were alsocounted on electron micrographs of cells that met the criteriafor evaluation.

Measurement of glycosaminoglycan (GAG) and NOrelease from cartilage slices. Spontaneous release of GAGsand NO was determined from each of the cartilage sectionsafter a 24-hour incubation at 37°C in complete growth me-dium. For the GAG release assay, the conditioned media werecollected, mixed with 46 �M dimethylmethylene blue, 40 mMglycine, and 40 mM NaCl (pH 3.0), immediately analyzedspectrophotometrically at 525 nm, and compared against astandard curve of 1–50 �g/ml of chondroitin sulfate. Tomeasure NO release, the Greiss reaction was used, as previ-ously described (31). Results were reported as micrograms ofGAG released or, for NO, as micromolar nitrate/nitrite re-leased per mg of cartilage weight.

Measurement of chondrocyte ATP levels and lactate-to-pyruvate ratios. For determinations of cellular ATP levelsand lactate-to-pyruvate ratios, aliquots of 1 � 105 primarychondrocytes were plated in a 15-mm dish with 1.0 ml of theDMEM (high glucose) containing 1% FCS described above.After 24 hours, the conditioned media were collected foranalysis of lactate and pyruvate levels using Sigma diagnostickits according to the manufacturer’s instructions. For cellularATP levels, the cells were snap frozen, extracted in trichloro-acetic acid, neutralized in KH2CO3, and the supernatant wasadded to the Sigma luciferase ATP assay kit. Emitted light wasmeasured by luminometry (2).

Measurement of NPP, transglutaminase (TG), andPPi levels. For determination of NPP, TG, and PPi levels,aliquots of 3 � 105 primary chondrocytes were plated in a35-mm dish with 1.5 ml of experimental media. After 24 hours,the conditioned media were collected, and after treatment for10 minutes at 65°C, PPi was determined by differential adsorp-tion on activated charcoal of UDP-D-[6-3H]-glucose (Amer-

sham, Chicago, IL) from the reaction product 6-phospho-[6-3H]-gluconate (32). PPi was normalized to the DNA content ofeach well, which was determined chromogenically followingprecipitation in perchlorate. To determine the specific activityof NPP, the cells were harvested and the lysates were assayedfor hydrolysis of p-nitrophenyl thymidine monophosphate atalkaline pH, as previously described (32). Units of NPP weredesignated as micromoles of substrate hydrolyzed per hour(per �g of protein in each sample). Cell lysate TG-specificactivity was determined via incorporation of a biotinylated

Figure 1. Histochemical changes in articular cartilage during thespontaneous development of osteoarthritis (OA) in male Hartleyguinea pig knees. A, At 2 months of age, the articular cartilage had agrossly normal appearance. B, At 4 months of age, early OA changeswere associated with a decrease in cartilage cellularity, particularly inthe superficial zone (asterisks). C, By 6 months of age, surface lesions(arrow) in the medial tibial plateau, and foci of chondrocyte cloning(asterisk) and hypertrophy (arrowhead) were evident. D, At 8 monthsof age, further loss of cellularity (asterisks) in the superficial zoneoccurred in association with surface defects (arrow) in both the medialand lateral tibial plateaus. (Hematoxylin and eosin–stained 6-�msections; original magnification � 63.)

1218 JOHNSON ET AL

pentylamine into N,N-dimethylcasein and detection by bindingof streptavidin–alkaline phosphatase conjugate. Activity wasquantified kinetically at 410 nm over 15 minutes, as describedelsewhere (31).

Statistical analysis. Statistical analyses were per-formed using analysis of variance (ANOVA; multifactor withreplication).

RESULTS

Pathologic changes in cartilage in spontaneousguinea pig knee OA. We first evaluated the developmentand progression of knee OA in Hartley guinea pigs. Indoing so, our evaluation included light microscopy ofarticular cartilage sections from animals of progressivelyolder ages. At 2 months of age, the articular cartilage inthe knees demonstrated a normal histologic appearance(Figure 1A). By 4 months of age, the superficial zone ofthe articular cartilage of the medial and lateral tibialplateaus appeared less cellular (Figure 1B), and minorcartilage surface irregularity was detectable grossly (re-sults not shown). By 6 months of age, chondrocyte“cloning” and the development of chondrocyte hyper-trophy were detectable near cartilage surface defects inthe medial tibial plateau (Figure 1C). At this age,

surface lesions were grossly detectable in medial tibialplateau cartilage, and osteophytes had begun to develop(results not shown). By 8 months, a particularly markeddecrease in cellularity was seen in the superficial zonesnear areas of erosion (Figure 1D), and the formation ofvertical columns of chondrocytes (palisades) was detect-able in the middle and deep zones (results not shown).On gross examination, surface erosions were readilydetected in both the medial and lateral tibial plateaus(results not shown).

Taken together, the findings were comparable toprevious descriptions of the model (25–29). We there-fore went on to characterize NO production and pro-gressive biochemical changes in this system.

Concurrent progression of NO and GAG releasein cartilage explants. To define and compare biochem-ical changes in distinct cartilage zones during the devel-opment of OA, we modified a previously describedanalytic strategy (26), dividing the hyaline articularcartilage in the tibial plateau into 6 zones for sampling(Figure 2). The meniscal fibrocartilage was designatedthe seventh zone in this system.

First, spontaneous NO and GAG release weremeasured in cartilage samples from each zone (Figure3). By 8 months of age, NO release from hyalinecartilage sampled in each of zones 1–6 had steadilyincreased to �30-fold the amount released by samplesobtained at 2 months of age (Figure 3A), with GAGrelease increasing �8-fold (Figure 3B) in zones 1–6 overthe same time period. As analyzed by ANOVA, thechanges in NO and GAG release were not significantlydifferent in hyaline cartilage zones 1–6 over ages 2–8months, except that GAG release was lower at age 6months in the anterior peripheral zone (zone 5) (Figure3B).

Marked increases in the spontaneous release ofNO and GAG from meniscal fibrocartilage also oc-curred between 2 and 8 months of age, although themeniscal fibrocartilage samples studied released signifi-cantly less NO and GAG per tissue weight than did the6 hyaline cartilage zones sampled at times when OA waswell-established (Figures 3A and B).

Chondrocyte mitochondrial ultrastructure dur-ing OA evolution. Having determined the time course ofboth the OA and the marked increases in NO generationin Hartley guinea pig knees, we next examined byelectron microscopy changes in chondrocyte mitochon-drial ultrastructural features in association with aging inthe articular cartilage of the medial tibial plateau. Someintramitochondrial inclusions were detectable in chon-drocytes from cartilage of the medial tibial plateau at

Figure 2. Zones of the tibial plateau of male Hartley guinea pig kneesfrom which hyaline articular cartilage explants were derived. Sampleswere taken from 6 zones (boxed areas) of the tibial plateau: the medialcentral (zone 1), lateral central (zone 2), medial peripheral (zone 3),lateral peripheral (zone 4), anterior peripheral (zone 5), and posteriorperipheral (zone 6) zones. Samples were also taken from the menisci(shaded areas). The dimensions (width � length) of each area sampledwere as follows: slices from the medial and lateral central zones were2 � 3 mm, those from the medial and lateral peripheral zones were 1 �2.5–3 mm, those from the anterior and posterior peripheral zones were0.5–1.5 � 0.5–1.5 mm, and those from the menisci were 3 � 3 mm. Allslices were �1 mm in thickness.

ATP DEPLETION IN OSTEOARTHRITIS 1219

ages 4 and 6 months (Figure 4). But, �5% of thechondrocyte mitochondria we studied displayed ultra-structure abnormalities, such as inclusion bodies orcrystalline deposits. Overall, the fine structural morphol-ogy of �95% of the mitochondria in chondrocytes fromthe medial tibial plateau was within normal limits

throughout the development of OA between 2 and 8months of age (Figure 4). Nevertheless, by the age of 8months and beyond, the electron microscopic appear-ance of chondrocytes of the medial tibial plateau becamevariably altered, with fewer intracellular organelles seenin some chondrocytes but a more normal appearance inothers (Figure 4D).

Electron microscopic assessment of the numbersof mitochondria per cell in chondrocytes in situ from themedial tibial plateau indicated an absence of substantialincreases in mitochondrial numbers per chondrocyte asthe OA developed, including time points at which dis-ease was established (Table 1).

Changes in chondrocyte ATP levels and TG andNPP activities. Next, we evaluated contemporaneouschanges in NO generation and intracellular ATP levelsin primary knee articular chondrocytes during the evo-lution of OA. To do so, we extracted and culturedchondrocytes from cartilage samples of the entire tibialplateau at 2–8 months of age. The isolated chondrocytesdemonstrated a significant increase in NO generation atages 4, 6, and 8 months, relative to cells isolated from2-month-old animals (Figure 5A). Concurrently, by 4–8months of age, there was a steadily progressive decreasein the level of intracellular ATP (Figure 5B). Underthese conditions, the chondrocytes demonstrated a pro-gressively increasing lactate-to-pyruvate ratio (Figure5C), representing a compensatory change in glycolysisthat is characteristically seen in the presence of de-pressed mitochondrial oxidative phosphorylation (3).

In view of the progressive depletion of intracell-ular ATP and the progressively increasing lactate-to-pyruvate ratio linked to OA in the guinea pig kneechondrocytes, we concluded the study by testing forother potential chondrocyte adaptive changes related tothe ATP depletion. In this context, impairment of neu-ronal cell OXPHOS causing ATP depletion inducesactivity of TG (33), a direct promoter of chondrocytematrix calcification (31). We observed that by 6 monthsof age, there was a significant rise in TG-specific activityin guinea pig knee chondrocytes (Figure 6A). In addi-tion, Hartley guinea pig knee chondrocytes demon-strated more than doubling of NPP-specific activity (34)by 8 months of age (Figure 6B), and chondrocyteextracellular PPi levels were nearly doubled by 8 monthsof age (Figure 6C).

DISCUSSION

In the Hartley guinea pig, OA spontaneouslyappears first at 3–4 months of age in the medial tibial

Figure 3. Comparison of the progression of increased nitric oxide(NO) and glycosaminoglycan (GAG) release in each of the 6 hyalinecartilage zones and in the fibrocartilage menisci sampled from thehyaline articular cartilage of male Hartley guinea pig knees at ages 2,4, 6, and 8 months. Hyaline articular cartilage explants from zones 1–6(see Figure 2) and 3 � 3–mm meniscal explants (combined from themedial and lateral menisci) were removed from the tibial plateau andplaced into culture media for a 24-hour recovery period. Freshmedium was then added to each sample, and after 24 hours, theconditioned media were collected and analyzed for A, NO release andB, GAG release. Values are the mean and SD of triplicate cultures foreach explant type (n � 12 knees). � � P � 0.05 for differences in levelsof NO and GAG release between distinct zones of hyaline cartilageand meniscus samples at the various ages, by analysis of variance.

1220 JOHNSON ET AL

plateau cartilage zone that is not covered by meniscalcartilage; this is mediated partly by obesity and by varuspositioning of the hind limbs (25–29). However, this

study revealed that in early, as well as more advanced,OA, the increases in spontaneous release of both NOand GAG were comparable in all the tibial plateauhyaline cartilage zones studied. In addition, NO andGAG release increased markedly in meniscal fibrocarti-lage even early in the disease. These findings suggestwidespread perturbation of chondrocyte function drivenby inflammatory cytokines and oxidant stress withinspontaneous guinea pig knee OA cartilage, where thereis a predisposition to initial OA lesions where guinea pigknee chondrocytes are most challenged biomechanically(25). In this setting, later development of OA in thelateral tibial plateau and other areas likely occurs on abackground of established, widespread chondrocyte dys-function in the knee.

In this study, we demonstrated that marked andprogressive ATP depletion, to �50% of the chondrocyte

Table 1. Numbers of mitochondria per cell in male Hartley guineapig knee chondrocytes of different ages, as examined in situ by electronmicroscopy*

2 months 4 months 6 months 8 months

No. of mitochondria/no.of cells

315/72 249/62 348/75 383/135

Average 4.4 4.0 4.6 2.8

* Mitochondria in chondrocytes in medial tibial plateau zone 1 (seeFigure 2) were examined (minimum of 60 cells/section for each animal; 3sections per animal) at ages 2, 4, 6, and 8 months. An electron micrographwas taken of each cell, and the numbers of mitochondria per chondrocytethat met the criteria for evaluation were counted. Cells consideredsuitable for evaluation were those in which the section passed through thenucleus and in which no part of the cell was covered by the grid bars.

Figure 4. Mitochondrial ultrastructure in chondrocytes from the medial central zone of articular cartilage obtained from the knees of male Hartleyguinea pigs. Articular cartilage from male guinea pigs (n � 2 animals per age group) was collected and prepared for electron microscopy as describedin Materials and Methods. A–C, Samples from animals ages 2, 4, and 6 months, respectively. D, Samples from animals ages 8 months (left) and 12months (right). (Original magnification � 5,000.) In higher-magnification views of the boxed areas, arrows indicate mitochondria and arrowheadsindicate intramitochondrial inclusions. The overall findings for the mitochondrial ultrastructure are discussed in the text. (Original magnification �16,300.)

ATP DEPLETION IN OSTEOARTHRITIS 1221

ATP levels seen at 2 months of age, occurs in articularchondrocytes as OA spontaneously develops and evolveswith aging in the knee joints of Hartley guinea pigs.Significantly, intracellular ATP levels in knee chondro-cytes of aging guinea pigs dropped steadily in spite ofincreased glycolysis, as evidenced by a rising lactate-to-pyruvate ratio with aging.

The capacity of chondrocytes to repair theirmatrix in response to stress appears to require adequategeneration of ATP, and articular chondrocytes maintaina high rate of anaerobic glycolysis to support ATPgeneration within their relatively hypoxic environment(2–7). The generation of ATP through OXPHOS is18-fold more efficient than the generation of ATPthrough glycolysis, but it has not been entirely clear towhat extent OXPHOS supports chondrocyte ATP gen-eration in normal and OA cartilage (1). Here, weobserved that guinea pig knee chondrocytes generated

increasing amounts of NO as both the structural OAprocess and chondrocyte GAG release and ATP deple-tion progressed. NO directly interferes with the activitiesof complexes within the mitochondrial electron trans-port chain and promotes mitochondrial depolarization,thereby suppressing OXPHOS in chondrocytes (1–4).NO also stimulates increased glycolysis as an incompletecompensatory response to suppressed OXPHOS inchondrocytes (3). Taken together, our results suggestedthat chondrocyte ATP depletion was produced via im-paired OXPHOS, which was partly attributable to in-creased NO production.

Chondrocyte mitochondrial swelling and a pau-city of mitochondria have been reported in small obser-vational studies of human OA chondrocytes sampled atthe time of well-established disease (22–24). However, instudies of cells analyzed in primary culture at confluenceseveral weeks after isolation from human knees under-

Figure 5. Changes in nitric oxide (NO) generation, glycolysis, andintracellular ATP with progression of osteoarthritis and with aging inprimary articular chondrocytes from male Hartley guinea pigs. Pri-mary articular chondrocytes from guinea pigs ages 2, 4, 6, and 8months were collected and placed in monolayer culture in 15-mmdishes, as described in Materials and Methods. At confluence, themedium was changed, and chondrocytes were incubated for 48 hours,at which time intracellular ATP levels in the chondrocytes weredetermined (B), and aliquots of conditioned media were collected toanalyze NO release (A) and glycolysis (C) (as the ratio of lactate topyruvate levels). Values are the mean and SD of triplicate cultures(n � 10 animals per age group; 20 knees). � � P � 0.05 versus2-month-old animals, by analysis of variance.

Figure 6. Increases in articular chondrocyte transglutaminase (TG)–specific and nucleotide pyrophosphatase/phosphodiesterase (NPP)–specific activities as well as extracellular inorganic pyrophosphate(PPi) levels in association with the development of osteoarthritis inmale Hartley guinea pig knees. Primary articular chondrocytes col-lected from the tibial plateau of guinea pigs ages 2, 4, 6, and 8 monthswere plated in 35-mm dishes (3 � 105 cells/dish) and incubated for 24hours after reaching confluency. The cell lysates were collected todetermine TG-specific (A) and NPP-specific (B) activities, and theconditioned media were collected to determine extracellular PPi levels(C), as described in Materials and Methods. Values are the mean andSD of triplicate cultures (n � 10 animals per age group; 20 knees). �

� P � 0.05 versus 2-month-old animals, by analysis of variance.

1222 JOHNSON ET AL

going total joint replacement, Maneiro et al (4) observedthat chondrocyte mitochondrial mass actually appearedto be modestly increased in human OA chondrocytes. Inthe present study, in situ assessment by electron micros-copy of knee chondrocyte mitochondrial ultrastructurebetween 2 months and 8 months of age revealed onlyrare abnormalities. Changes in the numbers of mito-chondria per chondrocyte in the medial tibial plateauonly appeared by 8 months of age, when the numbers ofmitochondria per cell decreased by �30% relative tobaseline levels. As such, chondrocyte ATP depletion,which developed prior to the appearance of the mostadvanced cartilage structural lesions, was not attribut-able to either overt mitochondrial architectural anoma-lies or decreased mitochondrial mass.

It remains to be determined whether the capacityof NO to induce mitochondrial biogenesis also promotesthe preservation of mitochondrial numbers in chondro-cytes during the early evolution of OA (35). Mitochon-drial dysfunction without any evidence of alteration ofcell morphology is not without precedent, as illustratedby studies of muscle pathology in pyruvate dehydroge-nase deficiency (20,21). A remaining question is whetheraltered mitochondrial ATP generation without overtultrastructural changes in chondrocyte mitochondriaduring the evolution of spontaneous guinea pig knee OAreflects at least partial reversibility of mitochondrialdysfunction in the disease.

Compromised mitochondrial function can lead toapoptosis, and NO donors stimulate apoptosis in cul-tured articular chondrocytes (1,36). Furthermore, apo-ptotic chondrocytes are found in human OA cartilage(36–39). However, the extent of chondrocyte apoptosisin OA cartilage and the contribution of apoptosis to thepathogenesis of OA remain subjects of controversy(40,41). It should be noted that certain steps in theapoptotic process, including activation of specificcaspases, are ATP-dependent (42–45). Hence, it alsowill be of interest to determine whether the chondrocyteATP depletion manifested in spontaneous knee OA inthe guinea pig influences the comparative rates of loss ofcell viability by apoptosis or necrosis.

Pathologic knee meniscal calcification typicallydevelops in spontaneous OA in the Hartley guinea pigby 6 months of age (30). Here, we observed thatprogressive chondrocyte ATP depletion was paralleledby apparent adaptive changes to the ATP depletion,including elevations in chondrocyte NPP activity, PPilevels, and TG activity. Increased levels of each of thesemediators can directly promote calcification by chondro-

cytes in vitro (31,32,46). Human aging, OA, and chon-drocalcinotic knees have increased NPP activity (32,46),the physiologic function of which is partly to retrievepurine nucleoside bases for ATP scavenging via nucleo-side triphosphate pyrophosphohydrolase (NTPPPH) ac-tivity (32,34,46). PPi is generated as a byproduct ofNTPPPH activity, and extracellular PPi increases signif-icantly in aging and chondrocalcinotic joints, favoringcartilage deposition of calcium pyrophosphate dihydratecrystals (32,46). Cartilage aging and OA also are asso-ciated with progressive increases in cartilage TG activity(31,47), and TG activity can be induced in neuronal cellsby primary disruption of OXPHOS (33).

Taken together, the observed elevations in NPPactivity, extracellular PPi levels, and TG activity inparallel with ATP depletion in the guinea pig spontane-ous knee OA model suggest that pathways that potentlypromote chondrocalcinosis are activated in aging andOA cartilage, in part as a dysfunctional adaptive re-sponse to chondrocyte ATP depletion. The rise in TGactivity in guinea pig OA chondrocytes holds furthersignificance because of the major role of TG activity inpromoting both cytokine-induced matrix calcificationand chondrocyte hypertrophic differentiation, which re-sults in a dysregulated matrix-reparative response in OA(48,49).

In conclusion, substantial decrements in mito-chondrial ATP generation appear to contribute to thepathogenesis of OA and to several metabolic alterationsin aging chondrocytes that promote chondrocalcinosis.The association of elevated NO production and chon-drocyte ATP depletion in chondrocytes observed in thisstudy provides new evidence for the role of sustained,low-grade inflammatory stress mediated by NO in theimpairment of ATP-regulated chondrocyte remodelingand repair functions in OA. These effects may contrib-ute to an insufficiency of matrix synthesis relative tomatrix degradation that mediates OA progression(26,27). Last, our results suggest that adaptive increasesin TG activity and PPi generation secondary to chon-drocyte intracellular ATP depletion promote pathologiccartilage calcification in aging and OA.

REFERENCES

1. Terkeltaub R, Johnson K, Murphy A, Ghosh S. The mitochon-drion in osteoarthritis. Mitochondrion 2002;1:301–19.

2. Johnson K, Jung A, Murphy A, Andreyev A, Dykens J, TerkeltaubR. Mitochondrial oxidative phosphorylation is a downstreamregulator of nitric oxide effects on chondrocyte matrix synthesisand mineralization. Arthritis Rheum 2000;43:1560–70.

3. Tomita M, Santo EF, Nishikawa M, Yamano Y, Inoue M. Nitric

ATP DEPLETION IN OSTEOARTHRITIS 1223

oxide regulates mitochondrial respiration and functions of articu-lar chondrocytes. Arthritis Rheum 2001;44:96–104.

4. Maneiro E, Martin MA, de Andres MC, Lopez-Armada MJ,Fernandez-Sueiro JL, del Hoyo P, et al. Mitochondrial respiratoryactivity is altered in osteoarthritic human articular chondrocytes.Arthritis Rheum 2003;48:700–8.

5. Stefanovic-Racic M, Stadler J, Georgescu HI, Evans CH. Nitricoxide and energy production in articular chondrocytes. J CellPhysiol 1994;159:274–80.

6. Lee RB, Urban JP. Evidence for a negative Pasteur effect inarticular cartilage. Biochem J 1997;321:95–102.

7. Tushan FS, Rodnan GP, Altman M, Robin ED. Anaerobicglycolysis and lactate dehydrogenase (LDH) isozymes in articularcartilage. J Lab Clin Med 1969;73:649–50.

8. Gonsalves M, Barker AL, Macpherson JV, Unwin PR, O’Hare D,Winlove CP. Scanning electrochemical microscopy as a local probeof oxygen permeability in cartilage. Biophys J 2000;78:1578–88.

9. Kanwar Y, Yoshinaga Y, Liu Z, Wallner E, Carone F. Biosyntheticregulation of proteoglycans by aldohexoses and ATP. Proc NatlAcad Sci U S A 1992;89:8621–5.

10. Clancy RM, Abramson SB, Kohne C, Rediske J. Nitric oxideattenuates cellular hexose monophosphate shunt response tooxidants in articular chondrocytes and acts to promote oxidantinjury. J Cell Physiol 1997;172:183–91.

11. Pelletier JP, Martel-Pelletier J, Abramson SB. Osteoarthritis, aninflammatory disease: potential implication for the selection ofnew therapeutic targets. Arthritis Rheum 2001;44:1237–47.

12. Clancy R, Rediske J, Koehne C, Stoyanovsky D, Amin A, Attur M,et al. Activation of stress-activated protein kinase in osteoarthriticcartilage: evidence for nitric oxide dependence. OsteoarthritisCartilage 2001;9:294–9.

13. Pelletier JP, Lascau-Coman V, Jovanovic D, Fernandes JC, Man-ning P, Connor JR, et al. Selective inhibition of inducible nitricoxide synthase in experimental osteoarthritis is associated withreduction in tissue levels of catabolic factors. J Rheumatol 1999;26:2002–14.

14. Zhu J, Yue Z, Wang J. Oxidative DNA damage of cartilage inosteoarthritis. Hunan I Ko Ta Hsueh Pao Bull Hunan MedUnivers 1998;23:438–40.

15. Takayuki O. Genetic and functional changes in mitochondriaassociated with aging. Physiol Rev 1997;77:442–64.

16. Ozawa T. Mitochondrial genome mutation in cell death and aging.J Bioenerg Biomembr 1999;31:377–90.

17. Wallace DC. Mitochondrial diseases in man and mouse. Science1999;283:1482–8.

18. Stadhouders AM, Jap PH, Winkler HP, Eppenberger HM, Wal-limann T. Mitochondrial creatine kinase: a major constituent ofpathological inclusions seen in mitochondrial myopathies. ProcNatl Acad Sci U S A 1994;91:5089–93.

19. Askanas V, Engel WK. Unfolding story of inclusion-body myositisand myopathies: role of misfolded proteins, amyloid-�, choles-terol, and aging. J Child Neurol 2003;18:185–90.

20. Sengers RC, Trijbels JM, Bakkeren JA, Ruitenbeek W, JanssenAJ, Stadhouders AM, et al. Demyelination and disturbed metab-olism of pyruvate: a case report. Eur J Pediatr 1983;140:127–30.

21. Chung SJ, Asoh S, Yamanaka T, Okamura-Oho Y, Toshima K,Woo M, et al. Muscle involvement in pyruvate dehydrogenasecomplex (PDHC) deficiency. Brain Dev 1987;9:9–15.

22. Weiss C. Ultrastructural characteristics of osteoarthritis. Fed Proc1973;32:1459–66.

23. Weiss C, Mirow S. An ultrastructural study of osteoarthritischanges in articular cartilage of human knees. J Bone Joint SurgAm 1972;54:954–72.

24. Chai BF. Relation of ultrastructural changes of articular cartilageand the arthroscopic classification in osteoarthritic knee. Zhong-hua Wai Ke Za Zhi 1992;30:18–20.

25. Bendele A. Animal models of osteoarthritis. J Musculoskel Neu-ron Interact 2001;1:363–76.

26. Wei L, Svensson O, Hjerpe A. Correlation of morphologic andbiochemical changes in the natural history of spontaneous osteo-arthrosis in guinea pigs. Arthritis Rheum 1997;40:2075–83.

27. Wei L, Svensson O, Hjerpe A. Proteoglycan turnover duringdevelopment of spontaneous osteoarthrosis in guinea pigs. Osteo-arthritis Cartilage 1998;6:410–6.

28. De Bri E, Reinholt FP, Svensson O. Primary osteoarthrosis inguinea pigs: a stereological study. J Orthop Res 1995;13:769–76.

29. Jimenez PA, Glasson SS, Trubetskoy OV, Haimes HB. Spontane-ous osteoarthritis in Dunkin Hartley guinea pigs: histologic, radio-logic, and biochemical changes. Lab Anim Sci 1997;47:598–601.

30. Kapadia RD, Badger AM, Levin JM, Swift B, Bhattacharyya A,Dodds RA, et al. Meniscal ossification in spontaneous osteoarthri-tis in the guinea-pig. Osteoarthritis Cartilage 2000;8:374–7.

31. Johnson K, Hashimoto S, Lotz M, Pritzker K, Terkeltaub R.Interleukin-1 induces pro-mineralizing activity of cartilage tissuetransglutaminase and factor XIIIa. Am J Pathol 2001;159:149–63.

32. Johnson K, Hashimoto S, Lotz M, Pritzker K, Goding J, Ter-keltaub R. Up-regulated expression of the phosphodiesterasenucleotide pyrophosphatase family member PC-1 is a marker andpathogenic factor for knee meniscal cartilage matrix calcification.Arthritis Rheum 2001;44:1071–81.

33. Lesort M, Tucholski J, Zhang J, Johnson GV. Impaired mitochon-drial function results in increased tissue transglutaminase activityin situ. J Neurochem 2000;75:1951–61.

34. Goding JW, Grobben B, Slegers H. Physiological and pathophys-iological functions of the ecto-nucleotide pyrophosphatase/phos-phodiesterase family. Biochim Biophys Acta 2003;1638:1–19.

35. Nisoli E, Clementi E, Paolucci C, Cozzi V, Tonello C, Sciorati C,et al. Mitochondrial biogenesis in mammals: the role of endoge-nous nitric oxide. Science 2003;7:896–9.

36. Hashimoto S, Takahashi K, Ochs RL, Coutts RD, Amiel D, LotzM. Nitric oxide production and apoptosis in cells of the meniscusduring experimental osteoarthritis. Arthritis Rheum 1999;42:2123–31.

37. Blanco FJ, Guitian R, Vazquez-Martul E, de Toro FJ, Galdo F.Osteoarthritis chondrocytes die by apoptosis: a possible pathwayfor osteoarthritis pathology. Arthritis Rheum 1998;41:284–9.

38. Pelletier JP, Mineau F, Boileau C, Martel-Pelletier J. Diacereinreduces the level of cartilage chondrocyte DNA fragmentation anddeath in experimental dog osteoarthritic cartilage at the same timethat it inhibits caspase-3 and inducible nitric oxide synthase. ClinExp Rheumatol 2003;21:171–7.

39. Kim DY, Taylor HW, Moore RM, Paulsen DB, Cho DY. Articularchondrocyte apoptosis in equine osteoarthritis. Vet J 2003;166:52–7.

40. Aigner T, Kim HA. Apoptosis and cellular vitality: issues inosteoarthritic cartilage degeneration. Arthritis Rheum 2002;46:1986–96.

41. Goggs R, Carter SD, Schulze-Tanzil G, Shakibaei M, MobasheriA. Apoptosis and the loss of chondrocyte survival signals contrib-ute to articular cartilage degradation in osteoarthritis. Vet J2003;166:140–58.

42. Tatsumi T, Shiraishi J, Keira N, Akashi K, Mano A, Yamanaka S,et al. Intracellular ATP is required for mitochondrial apoptoticpathways in isolated hypoxic rat cardiac myocytes. Cardiovasc Res2003;1:428–40.

43. Troyano A, Sancho P, Fernandez C, de Blas E, Bernardi P, AllerP. The selection between apoptosis and necrosis is differentiallyregulated in hydrogen peroxide-treated and glutathione-depletedhuman promonocytic cells. Cell Death Differ 2003;10:889–98.

44. Han BS, Hong HS, Choi WS, Markelonis GJ, Oh TH, Oh YJ.Caspase-dependent and -independent cell death pathways in pri-mary cultures of mesencephalic dopaminergic neurons after neu-rotoxin treatment. J Neurosci 2003;23:5069–78.

1224 JOHNSON ET AL

45. Steinbach JP, Wolburg H, Klumpp A, Probst H, Weller M.Hypoxia-induced cell death in human malignant glioma cells:energy deprivation promotes decoupling of mitochondrial cyto-chrome c release from caspase processing and necrotic cell death.Cell Death Differ 2003;10:823–32.

46. Pay S, Terkeltaub R. Calcium pyrophosphate dihydrate and hy-droxyapatite crystal deposition in the joint: new developmentsrelevant to the clinician. Curr Rheumatol Rep 2003;5:235–43.

47. Rosenthal AK, Derfus BA, Henry LA. Transglutaminase activity

in aging articular chondrocytes and articular cartilage vesicles.Arthritis Rheum 1997;40:966–70.

48. Johnson KA, Van Etten D, Nanda N, Graham RM, TerkeltaubRA. Distinct transglutaminase 2-independent and transglutami-nase 2-dependent pathways mediate articular chondrocyte hyper-trophy. J Biol Chem 2003;278:18824–32.

49. Merz D, Liu R, Johnson K, Terkeltaub R. IL-8/CXCL8 andGRO�/CXCL1 induce chondrocyte hypertrophic differentiation.J Immunol 2003;171:4406–15.

ATP DEPLETION IN OSTEOARTHRITIS 1225