Osteoarthritis and Cartilage (2009) 17, 1485e1493

Crown Copyright ª 2009 Published by Elsevier Ltd on behalf of Osteoarthritis Research Society International. All rights reserved.doi:10.1016/j.joca.2009.05.006

InternationalCartilageRepairSociety

Intra-articular magnesium sulfate (MgSO4) reduces experimentalosteoarthritis and nociception: association with attenuationof N-methyl-D-aspartate (NMDA) receptor subunit 1phosphorylation and apoptosis in rat chondrocytesC. H. Leeya, Z. H. Wenza, Y. C. Changz, S. Y. Huangz, C. C. Tangx, W. F. Chenk, S. P. Hsieh{,C. S. Hsieh# and Y. H. Jeanyy*yDepartment of Orthopedics and Traumatology, Taipei Medical University Hospital, Taipei, TaiwanzDepartment of Marine Biotechnology and Resources, Asia-Pacific Ocean Research Center, National SunYat-Sen University, Kaohsiung, TaiwanxDepartment of Early Childhood Education, National Pingtung University of Education, TaiwankDepartment of Neurosurgery, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Taiwan{Section of Pathology, Pingtung Christian Hospital, Pingtung, Taiwan# Section of Pediatric Surgery, Pingtung Christian Hospital, Pingtung, TaiwanyySection of Orthopedic Surgery, Pingtung Christian Hospital, Pingtung, Taiwan

Summary

Objective: To study the effects of intra-articular injection of magnesium sulfate (MgSO4) on the development of osteoarthritis (OA) and toexamine concomitant changes in the nociceptive behavior of rats.

Methods: OA was induced in Wistar rats with intra-articular injection of collagenase (500 U) in the right knee; the left knee was left untreated.In the OAþMgSO4 group (n¼ 7), the treated knee was injected with 500-mg (0.1-ml) MgSO4 twice a week for 5 consecutive weeks starting at1 week after collagenase injection; in the OA group (n¼ 7), the same knee was injected with the same amount of physiological normal saline.In the MgSO4 group (n¼ 6), na€ıve rats received only MgSO4 injections; in the control group (n¼ 6), na€ıve rats received only physiologicalnormal saline injections. Nociceptive behavior (mechanical allodynia and thermal hyperalgesia) on OA development was measured beforeand at 1, 2, 4, 6, and 8 weeks after collagenase injection, following which the animals were sacrificed. Gross morphology and histopathologywere examined in the femoral condyles, tibial plateau, and synovia. Immunohistochemical analysis was performed to examine the effect ofMgSO4 on N-methyl-D-aspartate (NMDA) receptor subunit 1 phosphorylation (p-NR1) and apoptosis in the articular cartilage chondrocytes.

Results: OA rats receiving intra-articular MgSO4 injections showed a significantly lower degree of cartilage degeneration than the rats receiv-ing saline injections. MgSO4 treatment also suppressed synovitis. Mechanical allodynia and thermal hyperalgesia showed significant improve-ment in the OAþMgSO4 group as compared to the OA group. Moreover, MgSO4 attenuated p-NR1 and chondrocyte apoptosis in OA-affectedcartilage.

Conclusions: Our results indicate that local intra-articular administration of MgSO4 following collagenase injection in an experimental rat OAmodel (1) modulates chondrocyte metabolism through inhibition of cell NMDA receptor phosphorylation and apoptosis, (2) attenuates thedevelopment of OA, and (3) concomitantly reduces nociception.Crown Copyright ª 2009 Published by Elsevier Ltd on behalf of Osteoarthritis Research Society International. All rights reserved.

Key words: Magnesium sulfate, Osteoarthritis, Nociception, NMDA, Chondrocyte apoptosis.

Introduction

Osteoarthritis (OA), the most common cause of pain anddisability in the elderly, is considered to be caused bylong-term mechanical disturbance the aging process1. Re-cent studies have revealed the role of inflammation in thepathogenesis of OA2, and potential mediators have been

aC. H. Lee and Z. H. Wen contributed equally to this paper.*Address correspondence and reprint requests to: Yen-Hsuan

Jean, Section of Orthopedic Surgery, Pingtung Christian Hospital,#60, Da-Lan Road, Pingtung 900, Taiwan. Tel: 886-8-7368686;Fax: 886-8-7338536; E-mail: jean.tang@msa.hinet.net

Received 15 January 2009; revision accepted 4 May 2009.

1485

described3. Currently available pharmacological therapiesfor OA mainly target palliation of pain and include analge-sics, intra-articular therapy, and topical treatment4. Pain in-hibitors targeting both joint pain and secondaryhyperalgesia and allodynia may therefore represent the op-timal approach for reducing OA-induced pain.

Magnesium, the fourth-common cation in the body, hasnumerous physiological activities, including activation ofmany enzymes involved in energy metabolism and proteinsynthesis5. Magnesium sulfate (MgSO4) is used as a phar-macological agent in a variety of clinical situations: tachyar-rhythmia, myocardial and neuronal ischemia, asthma,spasmophilia, preeclampsia, tocolysis, and post-anesthesiashivering6. Magnesium has been shown to exert

1486 C. H. Lee et al.: Intra-articular MgSO4 attenuates OA and nociception in rats

a physiological block of the ion channel on the N-methyl-D-aspartate (NMDA) receptor, preventing extracellular cal-cium ions from entering the cell and contributing to second-ary neuronal changes7. The NMDA receptor subunit 1(NR1) is considered an essential component of all func-tional NMDA receptors8. Increased phosphorylation ofNR1 (p-NR1), occurring via intracellular signaling pathways,has been recognized as a major mechanism contributing tothe regulation of NMDA receptor function9. Bondok and El-Hady10 showed that intra-articular magnesium reduces thepostoperative analgesic requirement of patients after arthro-scopic knee surgery. MgSO4 is an old and well-known ther-apeutic agent used in clinical practice. However, the role ofMgSO4 in the development of OA and OA-induced nocicep-tive behavior has not been well elucidated. In the presentstudy, we investigated the effects of intra-articular MgSO4

injection on cartilage degeneration and nociception ina rat model of collagenase-induced OA. Immunohistochem-ical examination was also performed to determine the effectof MgSO4 on p-NR1 and chondrocyte apoptosis in the artic-ular cartilage.

Methods

ANIMAL MODEL (COLLAGENASE INJECTION FOR OA

INDUCTION)

The experimental protocol was approved by the Animal Care and UseCommittee of National Sun Yat-Sen University and conformed to the Na-tional Institutes of Health guidelines for the care and use of animals in re-search. Two-month-old male Wistar rats (body weight¼ 275e310 g) wereused. OA was induced by intra-articular injection of collagenase (Clostridiumhistolyticum type II, enzyme activity¼ 333 U/mg; Sigma, St. Louis, MO) intothe right knee joint. The enzyme was dissolved in saline and filtrated witha 0.22-mm membrane, and administered in a volume of 0.1-ml using a 27-gauge, 0.5-inch needle. After shaving and sterilizing, the right knee jointwas injected intra-articularly with 0.1-ml of collagenase (500 U) or saline so-lution. Injection was performed twice on Days 1 and 4 according to themethod of Kikuchi et al.11. The left knee was left untreated.

EXPERIMENTAL DESIGN AND MgSO4 INJECTION

One week after the first collagenase injection, the animals were dividedinto four groups. Rats in the OAþMgSO4 group (n¼ 7) were injected in-tra-articularly with 500-mg (0.1-ml) MgSO4 (Sigma) in the collagenase-treatedknee twice a week for 5 consecutive weeks. Rats in the OA group (n¼ 7)were injected with 0.1-ml sterile physiological normal saline on the sameschedule. Na€ıve rats in the MgSO4 group (n¼ 6) and control group (n¼ 6)received only intra-articular MgSO4 injections and 0.1-ml intra-articular nor-mal saline injections, respectively, in the right knee twice a week for 5 con-secutive weeks. For the intra-articular MgSO4 or saline injections, rats wereanesthetized with 3% isoflurane in an oxygeneair mixture (1:1) at a flow rateof 0.5 L/min, and injection was performed under aseptic conditions by pass-ing a 27-gauge needle attached to a tuberculin syringe through the joint cap-sule lateral to the patellar tendon. Several extensions and flexions wereperformed after injection to ensure equal distribution of the injected materialthroughout the joint cavity.

ASSESSMENT OF NOCICEPTION

Nociception at the knee joint was assessed as the change in mechanicalallodynia12 and thermal hyperalgesia13. The nociceptive behaviors of the an-imals were tested before (baseline) and at 1, 2, 4, 6, and 8 weeks after colla-genase injection with the observer blinded to the type of injection receivedby the animals. For nociceptive testing, the animals were placed in plasticboxes with a transparent Perspex surface and allowed 30 min for habituation.

MECHANICAL ALLODYNIA

The diameters of the filaments corresponded to a logarithmic scale of theforce exerted, and thus the perceived intensity could be measured on a linearand interval scale. The withdrawal threshold was determined by Chaplan’s‘‘upedown’’ method involving the use of alternate large and small fibers todetermine the 50% withdrawal threshold12. Each von Frey hair was applied

to the plantar surface of the paw for 5 s. Briefly, when the rat lifted its paw inresponse to the pressure, the filament size was recorded, and a weaker fil-ament was used subsequently. Conversely, in the absence of a response,a stronger stimulus was used. A sequence of such responses was therebygenerated, and the 50% response threshold was calculated using a responsevariable spreadsheet. The von Frey filament was applied to each paw for fivetrials at approximately 3-min intervals.

THERMAL HYPERALGESIA

Thermal hyperalgesia was assessed by placing the hind paw on a radiantheat source; the paw-withdrawal latency at low-intensity heat was measuredwith an IITC analgesiometer (IITC Inc., Woodland Hills, CA) using a previ-ously described method13. The heat stimulus was applied until the animalwithdrawal, defined as lifting, licking, or flinching of the paw, was observed.The latency in foot withdrawal response to heat was recorded. A 30-s cutoffwas used to prevent soft tissue damage in the absence of a response. Foreach rat, five stimuli were applied with a stimulus interval of 10 min.

INFLAMMATION, GROSS MORPHOLOGY, AND

HISTOPATHOLOGICAL EXAMINATION OF THE KNEE JOINTS

The severity of knee joint inflammation was reflected by an increase in thehind-limb knee joint width14. The width of the bilateral hind-limb knee jointswas measured from the medial to the lateral aspect of the joint line by usingcalipers before (baseline) and 1, 2, 4, 6, and 8 weeks after the collagenaseinjection. The gross morphological changes in the cartilage of the femoralcondylar and tibial plateau were examined according to previously describedmethods15. A macroscopic total joint score was also obtained by adding themean scores of the cartilage lesions from the medial and lateral femoral con-dyles together with those from the medial and lateral tibial plateaus. Thejoints were sectioned 0.5 cm above and below the joint line, fixed in 10%neutral buffered formalin for 3 days, and then decalcified for 2 weeks in buff-ered 12.5% ethylenediaminetetraacetic acid (EDTA) and formalin solution.The joints were then sectioned mid-sagittally, washed under running tap wa-ter, and paraffin-embedded in an automatic processor (Autotechnicon Mono2; Technion Co., Chauncey, NY). The cartilage was stained with hematoxy-lineeosin (H & E) and Safranin-O/fast green stains to assess the generalmorphology and matrix proteoglycans. Microscopic examination of the artic-ular cartilage of the medial and lateral femoral condyles and the tibial plateauwere graded according to Mankin’s grading system16. A representative spec-imen of the synovial membrane from the medial and lateral compartments ofthe knee was dissected from the underlying tissues for histological examina-tion, as previously described17.

IMMUNOHISTOCHEMISTRY FOR p-NR1

Cartilage specimens were processed for immunohistochemical analysisas described in previous studies18. Briefly, sections (2 mm) of the paraffin-embedded specimens were placed on slides, deparaffinized with xylene,and dehydrated in a graded series of ethanol, following which the endoge-nous peroxidase activity was quenched by 30-min incubation in 0.3% hydro-gen peroxide. The antigen was retrieved by enzymatic digestion withproteinase K (20 mM; Sigma) in phosphate-buffered saline (PBS) for20 min. The slides were incubated with the primary antibody against eitherin rabbit anti-p-NR1 antibody (1:1000; Upstate, Lake Placid, NY) at 4�C for48 h in a humidified chamber. Thereafter, sections were treated with theavidinebiotin complex (ABC) technique by using an ABC kit (VectastainABC kit; Vector Labs, Burlingame, CA). The images were viewed using a Le-ica DM-2500 microscope (Leica, Heiderberg, Germany) and captured usinga SPOT CCD RT-slider integrating camera (Diagnostic Instruments Inc.,Sterling Heights, MI). For the statistical analysis of the immunohistochemicalfindings, the presence of different antigens in the cartilage was detected bya modification of the method of Pelletier et al.19. Prior to evaluation, it wasensured that each OA specimen had an intact cartilage surface that couldbe detected and used as a marker for the validation of morphometric analy-sis. The data obtained from the medial and lateral femoral condyles and me-dial and lateral tibial plateaus were considered together for the purpose ofstatistical analysis. Each slide was reviewed by two independent readers(Lee CH and Hsieh SP) who were blinded to the treatment groups.

CHONDROCYTE APOPTOSIS IN THE ARTICULAR CARTILAGE

Apoptotic chondrocytes were detected by terminal deoxynucleotidyl trans-ferase-mediated deoxyuridine triphosphate (d-UTP) nick end-labeling (TU-NEL) assays. Reaction, labeling, and detection of all samples wereperformed by using In Situ Cell Death Detection Kits, AP (Roche, Indianap-olis, IN, USA) according to the manufacturer’s protocols. The cells in the ap-optotic stage were quantified using the method proposed by Diaz-Gallegoet al.20. After TUNEL staining, the sections were subjected to the following

Time (weeks)

0 2 4 6 8

Mec

hani

cal t

hres

hold

(g)

2

4

6

8

10

12

14

16

18 control (n = 6)OA (n = 7)OA + MgSO4 (n = 7)MgSO4 (n = 6)

*

**

*

*

* *

**

*

##

Fig. 1. Time course of the antiallodynic effect of MgSO4 in the OAnociception model. In the mechanical allodynia test, the force re-quired to elicit hind-paw withdrawal in the OA þMgSO4 groupwas significantly improved compared with OA rats at 6 and 8weeks after the collagenase injection. Data shown represent themean� S.E.M. of the ipsilateral hind paw of each group. *P< 0.05

vs the control group; #P< 0.05 vs the OA group.

Time (weeks)0 2 4 6 8

Dif

fere

nt o

f w

idth

fro

mno

n-in

jury

kne

e (m

m)

-0.5

0.0

0.5

1.0

1.5

2.0

control (n = 6)OA (n = 7)OA + MgSO4 (n = 7)MgSO4 (n = 6)

*

*

***

*

***#

# #*

Fig. 3. Time course of joint width changes after the collagenase in-jection. The widths of the bilateral hind-limb knee joints were mea-sured in each rat before and at 1, 2, 4, 6, and 8 weeks afterinjection. Data (mean� S.E.M.) are expressed as the difference inknee width between the values at each time point and at timezero (before injection). *P< 0.05 vs the control group; #P< 0.05

vs the OA group.

1487Osteoarthritis and Cartilage Vol. 17, No. 11

procedures. The color was developed by incubation of all samples for 5 minwith a new fuchsin solution. The apoptotic chondrocyte ratio (%) was mea-sured as the ratio between the total number of TUNEL-positive nuclei andthe total number of cells present in six randomly chosen microscope fields.

DATA AND STATISTICAL ANALYSIS

All data are presented as the mean� standard error of the mean (S.E.M.).The data were analyzed by using one-way analysis of variance (ANOVA),followed by StudenteNewmaneKeuls method post hoc test. A P-value ofless than 0.05 was considered significant.

Results

On sacrifice, all collagenase-injected knees showed OA.No signs of drug toxicity such as ataxia, muscle weakness,and lethargy were noted in the rats treated with MgSO4.

Time (weeks)0 2 4 6 8

Paw

wit

hdra

wal

late

ncy

(sec

)

14

16

18

20

22

24

26

28

30

32

control (n = 6)OA (n = 7)OA + MgSO4 (n = 7)MgSO4 (n = 6)

*

#

*

*

**

*

*

**#

#

Fig. 2. In the thermal hyperalgesia test, ipsilateral paw-withdrawallatencies in the OAþMgSO4 group were significantly improvedcompared with the OA group at 4, 6, and 8 weeks after the collage-nase injection. Data shown represent the mean� S.E.M. of the ipsi-lateral hind paw of each group. *P < 0.05 vs the control group;

#P< 0.05 vs the OA group.

NOCICEPTIVE BEHAVIORS IN OA (MECHANICAL ALLODYNIA

AND THERMAL HYPERALGESIA)

The force required for hind-paw withdrawal in theOAþMgSO4 group was significantly increased comparedwith the OA group (P< 0.05) at 6 and 8 weeks (7.5� 0.6vs 4.4� 0.5 g and 6.2� 0.3 vs 4.1� 0.4 g, respectively) af-ter collagenase-induced OA (Fig. 1). The OAþMgSO4

group showed increased von Frey thresholds at 6 and 8weeks after collagenase injection compared with the OAgroup (Fig. 1). Eight weeks after the induction of OA, themechanical threshold in the hind paw of the side contralat-eral to the side of intra-articular injection was 12.1� 1.7,12.8� 0.7, 11.8� 0.7, and l3.2� 1.2 g in the OA,OAþMgSO4, MgSO4, and control groups, respectively. Be-fore intra-articular collagenase injection (baseline values),the mechanical threshold in the hind paw on the side con-tralateral to the side of intra-articular injection was12.3� 0.9, 12. 4� 0.5, 12.6� 0.4, and l2.4� 0.8 g for theOA, OAþMgSO4, MgSO4, and control groups, respec-tively. No significant difference was detected in the mechan-ical allodynia in the contralateral limbs among the fourexperimental groups. In the thermal hyperalgesia test, thepaw-withdrawal latency in the right hind paw was reducedat 1, 2, 4, 6, and 8 weeks after collagenase injection ascompared to the control groups (Fig. 2). The OAþMgSO4

Table IMacroscopic evaluation of the articular cartilage on the femoral

condyles and tibial plateau

Group Macroscopic score

OA (n¼ 7) 2.68� 0.08*OAþMgSO4 (n¼ 7) 1.45� 0.23*,yControl (n¼ 6) 0.18� 0.03

Data are expressed as the mean� S.E.M. For the macroscopic

score, refer to Methods. OA: collagenase-induced OA knee treated

with normal saline; OAþMgSO4: collagenase-induced OA knee

treated with MgSO4; control: saline injection in na€ıve rat knee.

*P< 0.05 vs the control group. yP< 0.05 vs the OA group.

Fig. 4. Histopathological evaluation of the articular cartilage of the femoral condyles and tibial plateau. (A) In the control group, the surface ofthe superficial cartilaginous layer is smooth, and the cartilage matrix is consistently well stained with Safranin-O/fast green. (B) The specimenfrom the OA group shows a decrease in the cartilage thickness, disappearance of the surface layer cells (arrow), a fissure extending into thetransitional and radial zones, and chondrocyte hypocellularity in the transitional and radial zones. (C) The specimen from the OAþMgSO4

group shows mild irregularity of the surface layer, fibrillation of and fissures in the superficial cartilaginous layer (arrow), and slight diffuse hy-percellularity in the transitional and radial zones. (D) In the MgSO4 group, similar changes are observed as in the control group. (E) In the OAgroup, osteophyte formation is evident at the medial cartilage margins of the tibial plateau (arrow). (F) The synovium of the OA group shows

subsynovial lining cell fibrosis and moderate mononuclear inflammatory cellular infiltration, and neovascularization. Scale bar¼ 500 mm.

1488 C. H. Lee et al.: Intra-articular MgSO4 attenuates OA and nociception in rats

group showed increased paw-withdrawal latency as com-pared to the OA group at 4, 6, and 8 weeks after collage-nase injection (22.1� 0.3 vs 17.2� 0.5 s, 23.5� 0.5 vs18.3� 0.4 s, and 24.7� 0.1 vs 20.5� 0.6 s, respectively;Fig. 2). Eight weeks after the induction of OA, the hind-paw-withdrawal latency on the side contralateral to the

side of intra-articular injection was 25.8� 0.9, 26. 7� 1.8,27.5� 1.6, and 27.1� 0.8 s in the OA, OAþMgSO4,MgSO4, and control groups, respectively. Before intra-artic-ular collagenase injection (baseline values), the hind-paw-withdrawal latency on the side contralateral to the side of in-tra-articular injection was 26.1� 1.2, 26. 4� 0.8,

Table IIHistological evaluation scores of the articular cartilage and synovial

tissue by light microscopy

Location Group

OA(n¼ 7)

OAþMgSO4

(n¼ 7)Control(n¼ 6)

Femoral condyleMedial side 8.8� 0.8* 3.5� 0.4*,y 0.5� 0.2Lateral side 8.4� 0.3* 3.2� 0.3*,y 0.8� 0.1

Tibial plateauMedial side 8.2� 0.5* 2.9� 0.3*,y 0.5� 0.3Lateral side 7.5� 0.6* 2.4� 0.2*,y 0.6� 0.1

Synovial tissue 6.6� 0.5* 3.9� 0.4*,y 1.6� 0.1

Data are expressed as the mean� S.E.M. For the Osteoarthritic

score (Mankin) and synovitis score, refer to Methods. OA: collage-

nase-induced OA knee treated with normal saline; OAþMgSO4:

collagenase-induced OA knee treated with MgSO4; control: saline

injection in na€ıve rat knee. *P< 0.05 vs the control group.

yP< 0.05 vs the OA group.

1489Osteoarthritis and Cartilage Vol. 17, No. 11

27.1� 1.2, and 26.5� 1.2 s in the OA, OAþMgSO4,MgSO4, and control groups, respectively. There was no sig-nificant difference in thermal hyperalgesia in the contralat-eral limbs among the four experimental groups. Therewas no significant difference in mechanical allodynia andthermal hyperalgesia between the MgSO4 and controlgroups at any time point during the study (Figs. 1 and 2).

KNEE JOINT WIDTH AND GROSS MORPHOLOGIC CHANGES

An increase in the width of the hind-limb knee joint, was sig-nificant at 1, 2, 4, 6, and 8 weeks after the collagenase injec-tion in the OAþMgSO4 and OA groups. As shown in Fig. 3,with a significant difference (P< 0.05) between the controlgroup and both OAþMgSO4 and OA groups as well as be-tween the OAþMgSO4 group and the OA group. TheOAþMgSO4 group showed a greater decrease in joint in-flammation than the OA group (Fig. 3). In the OA group, grosscharacteristics of cartilage degeneration, such as fibrillation,erosion and ulcer formation, and osteophyte formation, wereseen in the femoral condyle and tibial plateau. Markedly lessseverity of cartilage damage was seen in the OAþMgSO4

group. In the control and MgSO4 groups, the cartilage of thefemoral condyle and tibial plateau was macroscopically nor-mal, with a glistening, smooth surface, and no cartilage de-fects or osteophytes were observed. Table I lists the grossevaluation scores for each group. A significant difference(P< 0.05) in gross morphologic score was found betweenthe OA group and both OAþMgSO4 and control groups (Ta-ble I), but not between the control group and the MgSO4 group(P¼ 0.49). The grade of cartilage damage in theOAþMgSO4 group was significantly lower than that in theOA group. Synovia from the OA group were hypertrophicand showed a reddish-yellow discoloration, whereas in theOAþMgSO4 group, they were thinner and the discolorationwas less intense. Synovia from the control and MgSO4

groups had a white luster and transparent appearance, withno hyperemia or evidence of synovitis.

MICROSCOPIC FINDINGS

The cartilage of the control and MgSO4 groups had a nor-mal histological appearance. A thin, glistening, smooth lam-ina filled with flattened chondrocytes was observed, and no

loss of proteoglycan was seen in the matrix on Safranin-Ostaining [Fig. 4(A and D)]. Specimens from the OA groupshowed obvious histological changes, including completedisorganization, moderate-to-severe hypocellularity, proteo-glycan reduction on Safranin-O/fast green staining, and de-nudation of articular surface and fissures extending into thedeep zones [Fig. 4(B)]. Osteophytes were present at themedial margins of the femoral condyle and tibial plateau[Fig. 4(E)]. In the OAþMgSO4 group, there was marked re-duction in the severity of the femoral condyle and tibial pla-teau lesions: only fibrillation and fissures extending into thesuperficial layer of cartilage were observed [Fig. 4(C)]. Theaverage scores obtained for the above findings based onthe evaluation criteria are shown in Table II. Significant dif-ferences (P< 0.05) were found between the control andboth OA and OAþMgSO4 groups (Table II). Mankin’sscore for the OAþMgSO4 group was significantly lowerthan that for the OA group. Cartilage degeneration wasmore severe at the medial sides than at the lateral sidesof the femoral condyle and tibial plateau in the OA group(Table II). In the present study, the gross and histopatholog-ical appearance of both the hip and ankle joints (n¼ 4) didnot show any significant changes among the experimentalgroups at 8 weeks after collagenase injection (data notshown). Synovia from the OA group were thick, had focalvilli, and showed hyperplasia of the lining cells and moder-ate infiltration of mononuclear inflammatory cells [Fig. 4(F)].The histology of the synovia from the control and MgSO4

groups was within normal limits. The synovitis scores on mi-croscopic evaluation are shown in Table II; significant differ-ences (P< 0.05) were found between the control group andboth OA and OAþMgSO4 groups and between theOAþMgSO4 group and the OA group (Table II). The syno-vitis score was lower for the OAþMgSO4 group than theOA group, suggesting that synovial inflammation was lesssevere in the OAþMgSO4 group.

IMMUNOHISTOCHEMISTRY OF p-NR1 IN ARTICULAR CARTILAGE

The immunolocalization of p-NR1 protein expression incartilage specimens from the control, OA, OAþMgSO4,and MgSO4 groups, and the negative control were exam-ined [Fig. 5(AeE)]. Little or no p-NR1 protein was observedin chondrocytes of specimens from the control and MgSO4

groups [Fig. 5(A and D)]. The p-NR1 protein was seen inchondrocytes of the superficial and transitional cartilaginouszones of the OA group [Fig. 5(B) inset, arrow]. Comparedwith the OA group, OAþMgSO4 specimens showed a no-ticeable decrease in the number of p-NR1-positive chondro-cytes [Fig. 5(C)]. From the four individual observations, nostaining was observed in the negative control [Fig. 5(E) in-set]. Quantitative analysis showed that MgSO4 significantlyreduced the collagenase-induced increase in the number ofp-NR1-positive chondrocytes in knee cartilage [Fig. 5(F)].

INHIBITION OF CHONDROCYTE APOPTOSIS BY MgSO4

In general, in the cartilage of the control group, onlya very small number of cells stained positive on TUNELstaining [Fig. 6(A inset and E)]. Compared with the controlgroup, the OA group showed a noticeable and statisticallysignificant increase in the number of TUNEL-positivechondrocytes [Fig. 6(B inset and E)]. Notably, the collage-nase-induced increase in the number of TUNEL-positivechondrocytes in articular cartilage was statistically reduced

Fig. 5. Distribution of p-NR1 protein immunoreactivity in the cartilage of the control, OA, OAþMgSO4, and MgSO4 groups. Positive immuno-reactivity of the p-NR1 protein is indicated by the red-brown color. Distribution of anti-p-NR1 immunoreactivity in the cartilage of the (A) control,(B) OA, (C) OAþMgSO4, and (D) MgSO4 groups. All were stained with antibodies against the p-NR1 protein. (E) Sample from the negativecontrol group incubated without primary antibody for p-NR1 showing no specific staining. Scale bar¼ 100 mm. (F) Quantitative analysisshowed that MgSO4 significantly reduced the collagenase-induced increase in the number of p-NR1-positive chondrocytes in cartilage of

the OA knee. Scale bar¼ 100 mm. *P< 0.05.

1490 C. H. Lee et al.: Intra-articular MgSO4 attenuates OA and nociception in rats

by the MgSO4 injection [seen in the OAþMgSO4 group;Fig. 6(C inset and E)]. The MgSO4 group did not have a sig-nificantly changed number of TUNEL-positive cells, com-pared with the control group [Fig. 6(D and E)].

Discussion

To our knowledge, this is the first study to demonstratethat intra-articular MgSO4 injection can attenuate the devel-opment of OA and associated nociceptive behavior (me-chanical allodynia and thermal hyperalgesia) in anexperimental rat OA model. More interestingly, MgSO4 in-hibited p-NR1 and chondrocyte apoptosis in the articularcartilage in the experimental OA model.

A collagenase injection into the knee joint is thought tonot only directly destroy the cartilage but also cause an in-flammatory reaction in synovial tissues and accelerates car-tilage degeneration21. According to the study of van derKraan et al.22 and Kikuchi et al.11, intra-articular administra-tion of collagenase-induced OA of the knee in mice andrabbit models, respectively. In the present study, we dem-onstrated that intra-articular collagenase administration in-duced characteristic OA changes, including cartilagedegradation, synovial inflammation, and osteophyte forma-tion, in the knee joints of rats (Fig. 4). In the present study,both the macroscopic grading score and the Mankin scorewere significantly lower in the OAþMgSO4 group than inthe OA group (Fig. 4, Tables I and II). MgSO4 injection sig-nificantly reduced the severity of cartilage degradation in

Fig. 6. Inhibition of chondrocyte apoptosis by MgSO4 in the articular cartilage of rat knee. (A) In general, in the cartilage of the control andMgSO4 groups, only a very small number of cells stained positive on TUNEL staining (inset and E). (B) In the OA group, a statistically sig-nificant increase was observed in the number of TUNEL-positive chondrocytes (inset and E). (C). In the OAþMgSO4 group, the numberof TUNEL-positive chondrocytes was significantly reduced (inset and E). (D). In the MgSO4 group, no significant change in the number of

TUNEL-positive cells was observed when compared with the control group (inset and E). Scale bar¼ 100 mm. *P< 0.05.

1491Osteoarthritis and Cartilage Vol. 17, No. 11

the OA knee. In OA, synovial inflammation also plays an im-portant role in the disease process2. In the present study,moderate synovitis was noted in the collagenase-inducedOA knee. The synovitis score was lower in theOAþMgSO4 group than in the OA group (Table II). Someauthors had found that joint width could be measured to de-termine the extent of tissue swelling as an index of inflam-mation14,23. In the early stage, the increase in the kneejoint width is accompanied with joint damage. In the presentstudy, the OAþMgSO4 group showed a greater decreasein the knee joint width as compared to the OA group (Fig. 3).

Animal models of OA have been widely used to study thepathophysiology and progression of joint damage, but withlittle characterization of the associated pain. Mechanical

allodynia and hyperalgesia, i.e., increased touch-evokedsensitivity to harmless and noxious stimuli, respectively,are common symptoms of inflammation24. While the pain as-sociated with OA is primarily localized to the joint, it is becom-ing increasingly apparent that a number of patients exhibitincreased nociception in adjacent or even remote bodyareas25,26. Bradley et al.26 showed that patients with kneeOA demonstrated increased sensitivity to pressure pain inthe hip as compared to healthy age-matched controls. Ferni-hough et al.14 reported that both iodoacetate injection andpartial meniscectomy in the knee joint of the rats induced his-tological change, and mechanical allodynia and thermal hy-peralgesia pain-related behaviors characteristic of humanOA. In the present study, we measured mechanical allodynia

1492 C. H. Lee et al.: Intra-articular MgSO4 attenuates OA and nociception in rats

and thermal hyperalgesia as analgesic effects on intra-artic-ular injection of MgSO4 in the OA knee. Mechanical allodyniaand thermal hyperalgesia were significantly improved in theOAþMgSO4 group as compared to the OA group.

In the inflamed state of arthritic knees, an increase in glu-tamate concentration is observed not only in axons in theinflamed region27 but also in the synovial fluid28,29. Intra-ar-ticular injection of glutamate into the knee joint results inthermal hyperalgesia and mechanical allodynia, which areattenuated by local injection of NMDA or non-NMDA recep-tor antagonists30. Dietary restriction of magnesium intakelowers the mechanical nociceptive thresholds in rats, whichis reversed by the NMDA receptor antagonist, MK-80131. Adecrease in magnesium levels can enhance nociceptivesigns by increasing the binding between glutamate andNMDA receptors32. The Mgþ2 had preventive effects onthe development of inflammation in adjuvant-induced rheu-matoid arthritis (RA) rats33. It has recently been found thatthe NMDA receptors on human synoviocytes may contrib-ute to joint inflammation and destruction in RA34. A previousstudy has reported the expression of NR1 and NR2 in nor-mal human articular chondrocytes and has suggested thatNMDA receptors may have a role in chondrocyte mechano-transduction35. The same team further reported the expres-sion of the NMDA receptor, NR2B, in OA, but not in normalchondrocytes36. The extent of Mgþ2 blockade depends onthe structural arrangement of the NMDA receptor-channelcomplex, and channels containing NR2D are less sensitiveto Mgþ2 than those containing NR2B37. Magnesium defi-ciency upregulates interleukin (IL)-1a and IL-6, pleiotropiccytokines implicated in acute phase responses and inflam-mation38. Zhang et al.39 reported that the IL-1 receptorantagonist attenuates bone cancer pain, and can inhibitp-NR1, suggesting that spinal IL-1b enhances p-NR1 to fa-cilitate bone cancer pain in rats. In our present study, p-NR1occurred in the articular cartilage of the OA andOAþMgSO4 groups (Fig. 5), and MgSO4 could reducep-NR1 in OA rat knees. This is the first presentation thata functional NR1 in chondrocytes can be attenuated byMgSO4 treatment in an OA rat model. We proposed thatinhibition of the NMDA receptor by Mg2þ subsequentlydecreased the entry of extracellular calcium into the cells,resulting in the prevention of chondrocyte damage in OA.However, the exact mechanism for this feature is yetunclear and requires further investigation.

Hallack40 showed that intra-peritoneal injection of MgSO4

increased the Mg2þ level in both cerebrospinal fluid and dif-ferent brain regions in rats. Significant changes in the Mg2þ

concentration were observed in the spinal cord after a sys-temic injection and could lead to a potential central effectsuch as ataxia, muscle weakness, and lethargy41.

Chondrocyte apoptosis plays an important role in the pro-gression of OA42. Yeh et al.43 reported that intra-articular in-jection of collagenase could induce chondrocyte apoptosisin the lumbar facet joint of rats. In the present study, theOAþMgSO4 group showed a significantly reduced per-centage of TUNEL-positive staining of chondrocytes com-pared with the OA group (Fig. 6). This is also a novelfinding that MgSO4 could reduce chondrocyte apoptosis inan experimental OA rat model.

In conclusion, our results indicate that local intra-articularadministration of MgSO4 following collagenase injection inan experimental rat OA model (1) modulates chondrocyte me-tabolism through inhibition of cell NMDA receptor phosphory-lation and apoptosis, (2) attenuates the development of OA,and (3) concomitantly reduces nociception. These findingsmay pave the way for further investigations on MgSO4 as

a potentially therapeutic agent in the treatment of the inflam-matory component of OA. Further research is needed notonly to better define the effect of MgSO4 on OA but also to clar-ify the roles of the NMDA antagonist in OA treatment.

Conflict of interest

The authors acknowledge that there are no conflicts ofinterest pertaining to this manuscript.

Acknowledgments

This study was supported by the National Science Councilof Taiwan (NSC-97-2314-B-475-001-MY2) and the Ping-tung Christian Hospital, Taiwan (PS-96014).

References

1. Hardingham T, Bayliss M. Proteoglycans of articular cartilage:changes in aging and in joint disease. Semin Arthritis Rheum1990;20:12e33.

2. Pelletier JP, Martel-Pelletier J, Abramson SB. Osteoarthritis, an inflam-matory disease: potential implication for the selection of new thera-peutic targets. Arthritis Rheum 2001;44:1237e47.

3. Abramson SB, Attur M, Amin AR, Clancy R. Nitric oxide and inflamma-tory mediators in the perpetuation of osteoarthritis. Curr RheumatolRep 2001;3:535e41.

4. American College of Rheumatology Subcommittee on OsteoarthritisGuidelines. Recommendations for the medical management of osteo-arthritis of the hip and knee. Arthritis Rheum 2000;43:1905e15.

5. James MFM. Clinical use of magnesium infusions in anesthesia. AnesthAnalg 1992;74:129e36.

6. Delhumeau A, Granry JC, Monrigal JP, Costerousse F. Indications forthe use of magnesium in anesthesia and intensive care. Ann FrAnesth Reanim 1995;14:406e16.

7. Fawcet WJ, Haxby EJ, Male DA. Magnesium: physiology and pharma-cology. Br J Anaesth 1999;83:302e20.

8. Zou X, Lin Q, Willis WD. Role of protein kinase A in phosphorylation ofNMDA receptor 1 subunits in dorsal horn and spinothalamic tract neu-rons after intradermal injection of capsaicin in rats. Neuroscience2002;115:775e86.

9. Raymond LA, Tingley WG, Blackstone CD, Roche KW, Huganir RL.Glutamate receptor modulation by protein phosphorylation. J PhysiolParis 1994;88:181e92.

10. Bondok RS, El-Hady AMA. Intra-articular magnesium is effective forpostoperative analgesia in arthroscopic knee surgery. Br J Anaesth2006;97(3):389e92.

11. Kikuchi T, Sakuta T, Yamaguchi T. Intra-articular injection of collage-nase induces experimental osteoarthritis in mature rabbits. Osteoar-thritis Cartilage 1998;6:177e86.

12. Chaplan SR, Bach FW, Pogrel JW, Chung JM, Yaksh TL. Quantitativeassessment of tactile allodynia in the rat paw. J Neurosci Methods1994;53:55e63.

13. Hargreaves K, Dubner R, Brown F, Flores C, Joris J. A new and sensi-tive method for measuring thermal nociception in cutaneous hyperal-gesia. Pain 1998;32:77e88.

14. Fernihough J, Gentry C, Malcangio M, Fox A, Rediske J, Pellas T, et al.Pain related behaviour in two models of osteoarthritic in the rat knee.Pain 2004;112:83e93.

15. Pelletier JP, Mineau F, Raynauld JP, Woessner Jr JF, Gunja-Smith Z,Martel-Pelletier J. Intraarticular injections with methyl-prednisoloneacetate reduce osteoarthritic lesions in parallel with chondrocyte stro-melysin synthesis in experimental osteoarthritis. Arthritis Rheum1994;37:414e23.

16. Mankin HJ, Dorfman H, Lippiello L, Zarins A. Biochemical and metabolicabnormalities in articular cartilage from osteoarthritic human hips. IICorrelation of morphology with biochemical and metabolic data.J Bone Joint Surg 1971;53A:523e37.

17. Yoshimi T, Kikuchi T, Obara T, Yamaguchi T, Salcalcibara Y, Itoh H,et al. Effects of high-molecular-weight sodium hyaluronate on experi-mental osteoarthritis induced by the resection of rabbit cruciateligament. Clin Orthop Relat Res 1994;298:296e304.

18. Jean YH, Wen ZH, Chang YC, Hsieh SP, Tang CC, Lin JD, et al.Increase in excitatory amino acid concentration and transporters ex-pression in osteoarthritic knees of anterior cruciate ligament trans-ected rabbits. Osteoarthritis Cartilage 2008;16:1442e9.

19. Pelletier JP, Lascau-Coman V, Jovanovic D, Fernandes JC, Manning P,Currie MG, et al. Selective inhibition of inducible nitric oxide synthase

1493Osteoarthritis and Cartilage Vol. 17, No. 11

in experimental osteoarthritis is associated with reduction in tissuelevels of catabolic factors. J Rheumatol 1999;26:2002e14.

20. Diaz-Gallego L, Prieto JG, Coronel P, Gamazo LE, Gimeno M,Alvarez AI. Apoptosis and nitric oxide in an experimental model ofosteoarthritis in rabbit after hyaluronic acid treatment. J Orthop Res2005;23:1370e6.

21. van Osch GJ, van der Kraan PM, Blankevoort L, Huiskes R, van denBerg WB. Relation of ligament damage with site specific cartilageloss and osteophyte formation in collagenase induced osteoarthritisin mice. J Rheum 1996;23(7):1227e32.

22. van der Kraan PM, Vitters EL, van Beuningen HM, van de Putte LB, vanden Berg WB. Degenerative knee joint lesions in mice after a singleintra-articular collagenase injection. A new model of osteoarthritis.J Exp Pathol 1990;71:19e31.

23. Jou IM, Shiau AL, Chen SY, Wang CR, Shieh DB, Tsai CS, et al. Throm-bospondin 1 as an effective gene therapeutic strategy in collagen-in-duced arthritis. Arthritis Rheum 2005;52(1):339e44.

24. Wall PD. Neurophysiological mechanisms of referred pain and hyperalge-sia. In: Vecchiet L, Albe-Fessard D, Lindblom U, Giamberardino MA,Eds. New Trends in Referred Pain and Hyperalgesia. Amsterdam:Elsevier; 1993:3e12.

25. Bajaj P, Baja P, Graven-Nielsen T, Arendt-Nielsen L. Osteoarthritis andits association with muscle hyperalgesia: an experimental controlledstudy. Pain 2001;93:107e14.

26. Bradley LA, Kersh BC, DeBerry JJ, Deutsch G, Alarcon GA, McLain DA.Lessons from fibromyalgia: abnormal pain sensitivity in knee osteoar-thritis. Novartis Found Symp 2004;260:258e70 (discussion 270e9).

27. Carlton SM, Coggeshall RE. Inflammation-induced changes in periph-eral glutamate receptor populations. Brain Res 1999;820:63e70.

28. Jean YH, Wen ZH, Chang YC, Lee HS, Hsieh SP, Wu CT, et al. Hyalur-onic acid attenuates osteoarthritis development in the anterior cruciateligament-transected knee: associated with excitatory amino acidrelease in the joint dialysate. J Orthop Res 2006;24:1052e61.

29. Jean YH, Wen ZH, Chang YC, Hsieh SP, Tang CC, Wang YH, et al.Intra-articular injection of cyclooxygenase-2 inhibitor parecoxib atten-uates osteoarthritis progression in anterior cruciate ligament-transec-tion knee in rats: role of excitatory amino acids. OsteoarthritisCartilage 2007;15:638e45.

30. Lawand NB, Willis WD, Westlund KN. Excitatory amino acid receptorsinvolvement in peripheral nociceptive transmission in rats. Eur J Phar-macol 1997;324:169e77.

31. Dubray C, Alloui A, Bardin L, Ratssiguier Y, Eschallier A. Magnesiumdeficiency induces an hyperalgesia which is reversed by a NMDA an-tagonist, dizocilpine (MK-801) (Abstract). In: VII World Congress onPain, 1996, pp. 41e2.

32. Begon S, Pickering G, Eschalier A, Dubray C. Magnesium and MK-801have a similar effect in two experimental models of neuropathic pain.Brain Res 2000;887:436e9.

33. Nagai N, Fukuhata T, Ito Y, Tai H, Hataguchi Y, Nakagawa K. Preven-tive effect of water containing magnesium ion on paw edema in adju-vant-induced arthritis rat. Biol Pharm Bull 2007;30(10):1934e7.

34. Flood S, Parri R, Williams A, Duance V, Mason D. Modulation of inter-leukin-6 and matrix metalloproteinase 2 expression in human fibro-blast-like synoviocytes by functional ionotropic glutamate receptors.Arthritis Rheum 2007;56:2523e34.

35. Salter DM, Wright MO, Millward-Sadler SJ. NMDA receptor expressionand roles in human articular chondrocyte mechanotransduction. Bio-rheology 2004;41:273e81.

36. Ramage L, Martel MA, Hardingham GE, Salter DM. NMDA receptor ex-pression and activity in osteoarthritic human articular chondrocytes.Osteoarthritis Cartilage 2008;16(12):1576e84.

37. Monyer H, Burnashev N, Laurie DJ, Sakmann B, Seeburg PH. Develop-mental and regional expression in the rat brain and functional proper-ties of four NMDA receptors. Neuron 1994;12:529e40.

38. Bernardini D, Nasulewicz A, Mazur A, Maier JA. Magnesium and micro-vascular endothelial cells: a role in inflammation and angiogenesis.Front Biosci 2005;10:1177e82.

39. Zhang RX, Liu B, Wang L, Ren K, Qiao JT, Berman BM, et al. Interleukin1b facilitates bone cancer pain in rats by enhancing NMDA receptorNR-1 subunit phosphorylation. Neuroscience 2008;154:1533e8.

40. Hallak M. Effect of parenteral magnesium sulfate administration on excit-atory amino acid receptors in the rat brain. Magnes Res 1998;11(2):117e31.

41. Feria M, Abad F, Sanchez A, Abreu P. Magnesium sulfate injected sub-cutaneously suppresses autotomy in peripherally deafferented rats.Pain 1993;53:287e93.

42. D’Lima D, Hermida J, Hashimoto S, Colwell C, Lotz M. Caspase inhib-itors reduce severity of cartilage lesions in experimental osteoarthritis.Arthritis Rheum 2006;54:1814e21.

43. Yeh TT, Wen ZH, Lee HS, Lee CH, Yan Z, Jean YH, et al. Intra-articularinjection of collagenase induced experimental osteoarthritis of thelumbar facet joint in rats. Eur Spine J 2008;17(5):734e42.