ABC294640

Experimental Osteoarthritis in Rats Is Attenuated by ABC294640, a Selective Inhibitor of Sphingosine Kinase-2

Leo R. Fitzpatrick  Cecelia Green  Lynn W. Maines  Charles D. Smith

 

b

a

 
Department of Pharmacology, Penn State College of Medicine, and  Apogee Biotechnology Corporation,

Hershey, Pa., USA

Key Words
Osteoarthritis Sphingosine kinase Apoptosis

⦁ bstract
⦁ ackground/Aims: Osteoarthritis (OA) is a progressive de- generative disease characterized by cartilage degradation and chondrocyte apoptosis, which may involve aberrant sphingolipid metabolism. ABC294640 is a compound that selectively inhibits sphingosine kinase-2, a key enzyme in the sphingolipid pathway. Our goal was to assess the phar- macological effects of ABC294640 in the monosodium iodo- acetate (MIA) model of OA. Methods: MIA (3 mg) was inject- ed into the right knee joint to induce osteoarthritis in rats. Subsequently, the rats were treated with vehicle, ABC294640 or tramadol over a 28-day period. To assess pain, incapaci- tance readings were obtained weekly. MIA-injected knee joints were evaluated for histological damage, cartilage deg- radation and chondrocyte apoptosis (terminal deoxynucle- otidyl transferase-mediated dUTP nick end labeling histo- chemistry). Results: The percent weight bearing in vehicle/ MIA rats significantly (p !0.01) decreased from 48.8 80.8 (day 0) to 41.9 82.9 (day 28). In contrast, these values in ABC294640-treated rats were virtually the same on days 0 and 28. Knee joint histology scores were less severe in

© 2011 S. Karger AG, Basel
0031–7012/11/0874 –0135$38.00/0
Fax +41 61 306 12 34

ABC294640-treated rats. Cartilage proteoglycan staining was more prominent in ABC294640/MIA animals than in ve- hicle/MIA rats. The percentage of apoptotic chondrocytes was decreased from 39.5% (vehicle treatment) to 25.8% (ABC294640 treatment). Conclusion: ABC294640 attenuat- ed the knee joint histological damage and pain associated with MIA-induced OA in rats. Copyright © 2011 S. Karger AG, Basel

Introduction

Osteoarthritis (OA) is a highly prevalent form of ero- sive arthritis that is increasing in frequency [1]. OA is characterized by cartilage degradation, subchondral bone disruption and pain [1, 2]. Moreover, pain that is worsened by weight bearing activity is one of the classic symptoms of OA [3].
A commonly used model of OA involves the injection of monosodium iodoacetate (MIA) into the knee joint of rats [4 –6]. The injection of MIA leads to the inhibition of glycolysis, which results in cell death, cartilage degrada- tion and bone disruption [7, 8]. Recently, both human and MIA-induced OA have been shown to involve increased chondrocyte apoptosis [9, 10].

Leo R. Fitzpatrick, PhD
Department of Pharmacology, Penn State College of Medicine
1214 Research Boulevard

E-Mail [email protected]
www.karger.com
Accessible online at:
www.karger.com/pha
Hummelstown, PA 17036 (USA)
Tel. +1 717 531 0393, E-Mail lfitzpatrick  @ psu.edu Downloaded by:
Columbia University
128.59.222.12 – 3/2/2015 6:41:06 AM

Sphingosine kinase (SK) exists as SK1 and SK2 iso- forms [11, 12]. Both isoforms catalyze the formation of sphingosine-1-phosphate from sphingosine [11]. Notably, only the SK2 isoenzyme induces apoptosis in cells [12, 13]. Synovial fluid sphingosine-1-phosphate levels are report- edly higher in patients with rheumatoid arthritis than OA [14]. In this regard, recent data suggests that the inhibition of SK can effectively attenuate disease in murine models of rheumatoid arthritis [15, 16]. However, the efficacy of SK inhibitors has not yet been tested in an animal model of OA. ABC294640 is a novel chemical compound that selectively inhibits SK2 and also improves experimental rheumatoid arthritis in rodents [17, 18]. Therefore, we hy- pothesized that ABC294640 might also attenuate various features of MIA-induced OA, including apoptosis [10].
With this background information, the aims of our study were to assess the effects of ABC294640 on carti- lage and bone destruction in the knee joints of MIA-treat- ed rats, as well as the resultant pain response in these animals. We also analyzed the efficacy of ABC294640 for attenuating MIA-induced chondrocyte apoptosis in the knee joints of rats.

Methods

Chemicals
Sodium iodoacetate and tramadol hydrochloride were ob- tained from Sigma Chemical Company (St. Louis, Mo., USA). ABC294640 that was synthesized under Good Manufacturing Practice conditions was provided by Apogee Biotechnology Cor- poration (Hummelstown, Pa., USA). The ApopTag peroxidase in situ apoptosis detection kit was obtained from Millipore (Bil- lerica, Mass., USA). Diaminobenzidine was obtained from Vector Laboratories (Burlingame, Calif., USA).

Vertebrate Animals
Male Wistar rats were obtained from Charles River Laborato- ries (Wilmington, Mass., USA). Animals were housed under stan- dard laboratory conditions with a 12-hour light-dark cycle. The MIA-induced OA protocol was approved by the International An- imal Care and Use Committee at the Penn State College of Medi- cine (Hershey, Pa., USA).

Induction of MIA-Induced OA
OA was essentially induced by MIA according to the method described by previous investigators [4, 19]. Briefly, adult male Wistar rats were anesthetized with ketamine (75 mg/kg)/xyla- zine (7.5 mg/kg). Rats received a single injection of MIA (3 mg, 50- l volume) with a 27-gauge needle into the right knee joint. A control group of animals received 50 l of saline into the right knee. All rats received a 50- l injection of saline into the left knee. Rats were allowed to fully recover from the arthritis induc- tion procedure and then administered test compounds, as de- scribed above.
MIA-treated rats (n = 22), weighing initially between 220 and 270 g, were randomized into 3 treatment groups. The mean body weight of the ABC294640 treatment group (n = 9) was used as a reference to prepare twice-daily doses (approx. 50 mg/kg per dose) of this SK inhibitor. ABC294640 was dosed by orogastric gavage in size 9 gelatin capsules (Torpac, Maharashatra, India) during the initial 7 study days. Subsequently (study days 8–27), ABC294640 was dissolved in 0.5 ml of aqueous vehicle (46.7% polyethylene glycol 400, 46.7% saline and 6.6% ethanol) and ad- ministered by oral gavage. Another group of MIA animals (n = 4) received tramadol (100 mg/kg) only on study days on which weight bearing measurements were taken (see below). The third group of MIA-treated rats (n = 9) underwent a sham dosing pro- cedure (days 1–8), or were dosed with aqueous vehicle (days 8–27). These animals were designated as the vehicle/MIA treatment group. The animals injected with saline into both knee joints (n = 4) were dosed similarly to the vehicle/MIA group. These rats were considered as a vehicle/saline treatment group.

Measurement of Weight Bearing (MIA-Associated Pain)
The assessment of MIA-associated pain was performed using a hind limb weight bearing apparatus (Incapacitance Analgesia Meter, model No. 57210, Stoelting, Wood Dale, Ill., USA). The rats were acclimated to the testing apparatus before experimental readings were conducted. Dosing with ABC294640 was 2 h before the weight bearing measurements, while tramadol was dosed 1 h before these measurements. These time points were chosen, based on the relevant literature [5, 17]. Once animals had assumed a cor- rect postural stance in the testing apparatus, readings were taken over a 3-second period [8]. The percent of weight borne on the right leg was calculated as: weight on right leg/combined weight on both legs !100 [4]. Typically, 4 –6 readings were taken for each rat, on the applicable study days (days 0, 2, 7, 14, 20 and 28). The averages of these readings were determined and then used to cal- culate the treatment group average on a specific study day.

Determination of Histological Parameters (MIA OA Model) On study day 28, rats were euthanized by carbon dioxide in-
halation followed by adjunct cervical dislocation. The right knee joints were collected and stored in 10% buffered formalin. Subse- quently, sections of the knee joint were collected and stained by either hematoxylin and eosin for routine histological analysis, or by safranin O stain for measuring proteoglycan staining [6, 7].
Knee joint histology scores were determined on coded slides with a 25-mm ocular grid, at a magnification of !100. Three sec- tions of both the femur and tibia were analyzed, for a total of 6 areas per slide. These sites corresponded to the anterior, interme- diate and posterior areas, as defined by other investigators [20]. For each of these knee joint areas, histological damage was deter- mined using scoring systems adapted from previous publications [6, 7]. Cartilage degradation was scored as: 0 = none, 1 = superficial zone, 2 = middle zone, 3 = deep zone, and 4 = tidemark. Therefore, the total cartilage score for each rat was 0 –24. Bone distortion/ disruption was scored as: 0 = !10%, 1 = 10 –30%, 2 = 30 –50%, 3 = 50 –70%, and 4 = 170%. The total bone score for each rat was 0–24. Therefore, the total joint histology score for each animal was with- in the range of 0 –48. Individual rat joint histology scores were then averaged to determine a mean value for each treatment group.
The scoring of joint proteoglycan and safranin O red staining was done from the same 6 areas of the knee joint as described

136

Pharmacology 2011;87:135 –143

Fitzpatrick /Green /Maines /Smith Downloaded by:
Columbia University
128.59.222.12 – 3/2/2015 6:41:06 AM

a
b

Vehicle/Saline (n = 4)
Vehicle/MIA (n = 9)

ABC/MIA (n = 9)
Tramadol/MIA (n = 4)

50

55
b
ABC

b
b
45

50
b

45

a

a

a

a a
40

Sham

40
a
a
a
a
35

35

30

30

Day 0

Day 2 Day 5 Day 7 Day 14 Day 20 Day 28
b
0
2
4
6
8
12 16
10 14 18
Study day
20
22
24
26
28

a

Fig. 1. a Effects of ABC294640 (ABC; 100 mg/kg/day) or trama- dol (100 mg/kg, single dose) on weight bearing, as an indicator of MIA-induced pain; n = 4 or 9 rats per group as indicated on the graph.  p !0.05 versus day 0 by repeated measures 2-way ANOVA and Bonferroni posttest analyses;  p !0.05 versus vehicle/MIA

above. The scoring system was according to the method of Janusz et al. [7]: 0 = normal pattern of staining, 1 = minimal staining loss, 2 = mild staining loss, 3 = moderate staining loss, and 4 = severe loss of staining. Therefore, the maximum proteoglycan staining score for each animal was 24. Individual scores were then aver- aged to determine a mean value for each treatment group.
Photographs of all the applicable rat knee joints were captured at a magnification of !40, using an Olympus IMT-2 microscope (Olympus Corporation, Lake Success, N.Y., USA) and EPIX- XCAP image capture software program (Buffalo Grove, Ill., USA). After picture quality and size optimization by Microsoft Office picture manager (Microsoft Corporation, Redmond, Wash., USA), the cartilage thickness in centimeters was measured with an electronic screen ruler (Spadix Software, Austin, Tex., USA). These measurements were performed on anterior, interme- diate and posterior areas of both the femur and tibia. Individual cartilage thickness values for each rat were obtained, and then mean values determined for the various treatment groups. The final cartilage thickness values in millimeters were calculated as the measured cartilage thickness value in centimeters !10, di- vided by the photographic magnification (!40).
Assessment of MIA-Induced Chondrocyte Apoptosis
The terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) method was used to determine chondrocyte apoptosis in the knee joint cartilages of MIA- or saline-injected rats. Essentially, we utilized the method de- scribed by Watrin-Pinzano et al. [10] with some modifications. Slides were deparaffinized with xylene. After washes in succes- sively lower concentrations of ethanol, the sections were per- meabilized with 0.3% Triton X-100/PBS, followed by proteinase

by 2-way ANOVA and Bonferroni posttest analyses. b The same data as in a , but analyzed by linear regression; an ANCOVA anal- ysis showed that the slopes of the regression lines for the vehicle/ MIA and ABC294640/MIA treatment groups were significantly (p = 0.0029) different.

K. Subsequently, the sections were quenched in 3% hydrogen peroxide, followed by successive applications of various compo- nents (equilibration buffer, terminal deoxynucleotidyl transfer- ase and antidigoxigenin conjugate) provided with the apoptosis detection kit. After application of diaminobenzidine substrate (10 min), and brief counterstaining with Gill ’s hematoxylin (30 s), aqueous mounting solution was applied to the slides. Chondrocyte apoptosis was determined by counting the num- ber of positively stained (brown) nuclei as a percentage of the total number of cells, at a magnification of !400 [10]. As in pre- vious histological-based evaluations in this study, we evaluated 6 total areas (3 areas from the femur and 3 areas from the tibia). From the individual percent apoptosis values for each rat, the mean percent of TUNEL positively stained cells was then calcu- lated for each treatment group.

Statistical Analyses
All data are presented as the mean 8SEM. Statistical analyses were conducted with a Graph-Pad Prism software program (San Diego, Calif., USA). Potential data outliers were analyzed by Grubb ’s test. For the weight bearing data, we used 2-way ANOVA followed by a Bonferroni posttest analysis to compare different treatment groups at the same times. To compare the same treat- ment group on different treatment days, we used a repeated mea- sures 2-way ANOVA followed by a Bonferroni posttest analysis. For other data, unpaired test analyses were typically used to compare two different treatment groups. Differences between linear regres- sion lines were analyzed by ANCOVA analyses. For these statistical analyses, p !0.05 was considered to be statistically significant.

ABC294640 Attenuates Osteoarthritis in
Rats

Pharmacology 2011;87:135 –143

137Weight bearing (% R/R + L leg)
Weight bearing (% R/L leg)
Downloaded by:
Columbia University
128.59.222.12 – 3/2/2015 6:41:06 AM

a b c

d e f

50
Fig. 2. Representative photographs are shown from sections of rat

*

right knee joints. Hematoxylin and eosin. !100. a Normal femur, vehicle/saline treatment. b Cartilage and bone loss of the femoral condyle were evident (upper boxed area), as well as cartilage deg- radation (lower boxed area), vehicle/MIA treatment. c An area of femoral cartilage degradation (boxed area), ABC294640 (100 mg/ kg/day)/MIA treatment; the arrow indicates an area of bone loss. d Normal tibia, vehicle/saline treatment. e Cartilage degradation (boxed area) and bone disruption (arrows) in the tibial plateau, vehicle/MIA treatment. f A focal area of tibial cartilage degrada-
40

30

20

10

tion (boxed area) and bone disruption (arrow), ABC294640 (100 mg/kg/day)/MIA treatment. g A summary of the knee joint his-
0

Vehicle/saline

Vehicle/MIA

ABC/MIA

tology data; ABC designates ABC294640 treatment. * p ! 0.05
g
(n = 4)
(n = 9)
(n = 9)

versus vehicle/saline by unpaired t test analysis.

Results

Effects of ABC294640 on MIA-Associated Pain Vehicle/saline-treated rats had mean weight bearing
values near or above 50% throughout the study (fig. 1a). Moreover, tramadol/MIA-treated animals showed a sim-

ilar data profile, thereby demonstrating the expected an- algesic effect of this drug [5]. As shown in figure 1a, sig- nificant decreases in weight bearing values on the right (injected) leg occurred during study days 0 –14 in both vehicle/MIA- and ABC294640/MIA-treated rats. Inter- estingly, between days 14 and 28, there was a progressive

138

Pharmacology 2011;87:135 –143

Fitzpatrick /Green /Maines /Smith Total joint histology score (0 –48)
Color version available online
Downloaded by:
Columbia University
128.59.222.12 – 3/2/2015 6:41:06 AM

65
60
55
50

n = 22
r = –0.5941
p = 0.0036

0.5

0.4

NS

45
40
35
0.3

0.2

0.1

*

30
25

0

10 20 30 40 50

60

0

Vehicle/saline
(n = 4)

Vehicle/MIA
(n = 9)

ABC/MIA
(n = 9)

Histology score

Fig. 3. Linear regression analysis (n = 22) shows a significant in- verse association (r = –0.5941, p = 0.0036) between joint histology scores and weight bearing with the rat MIA OA model; black squares = vehicle/saline treatment, red diamonds = vehicle/MIA treatment and blue circles = ABC294640 (100 mg/kg/day)/MIA treatment; quartiles (separated by dotted lines) in the graph illus- trate data clustering for treatment group samples.

increase in the weight bearing response in ABC294640, but not vehicle-treated animals. However, statistically there was no difference between the weight bearing val- ues in the ABC294640/MIA and vehicle/MIA treatment groups on these study days (fig. 1a).
When this weight bearing data was plotted in a dif- ferent format (comparison of slopes; fig. 1b), there was clear evidence that the MIA-associated weight bearing time course response was significantly (p = 0.0029) im- proved in ABC294640-treated rats. Specifically, these data show that the slope of the regression line (percent weight bearing per day) in ABC294640/MIA-treated rats (value of 0.477 80.033) was significantly different than in vehicle/MIA-treated animals (value of 0.158 8 0.068).

Effects of ABC294640 on MIA-Associated Histological Parameters
Representative histological photographs (femur and tibia) are shown in figure 2a–f . Clear evidence of patho- logical changes including cartilage degradation/loss, as well as bone disruption/loss (boxed areas, arrows), were observed in the vehicle/MIA-treated rat (fig. 2b, e), when compared to the vehicle/saline-treated animal (fig. 2a, d). In contrast, although some pathological changes (e.g. cartilage degradation, bone disruption) were apparent in the ABC294640/MIA-treated animal (fig. 2c, f, boxed

Fig. 4. Effect of ABC294640 (ABC; 100 mg/kg/day) on cartilage thickness (rat knee joint); n = 4 or 9 per treatment group. * p ! 0.05 versus vehicle/saline by unpaired t test analysis. NS = Not significant by unpaired t test analysis.

areas, arrow), by comparison, the overall severity was less pronounced. The mean histology scores (horizontal lines in fig. 2g) were significantly (p !0.05) increased in vehicle/MIA-treated rats (30.9 83.3), as compared to non-MIA-treated animals (4.8 81.4). Treatment with ABC294640 over a 1-month period reduced the mean histology score to 24.1 84.0.
Interestingly, while analyzing the individual data val- ues from this study (fig. 2g), it was evident that there was one aberrantly high histology score value (circled trian- gle) in the ABC294640 treatment group. A strict statisti- cal analysis for data outlier values (Grubb ’s test) did not support exclusion of this data point. However, if this val- ue was excluded from the statistical analyses, there was a statistically significant (p !0.05) reduction in the mean joint histology score of ABC294640/MIA-treated rats compared to the vehicle/MIA treatment group. Figure 3 shows that a significant inverse association (r = –0.5941, p = 0.0036) was detectable between the right knee joint histology scores and weight bearing data.
The representative histology pictures (fig. 2) suggested that the knee joint cartilage thickness was reduced in the vehicle/MIA-treated rats, while some degree of normal- ization seemed apparent in ABC294640/MIA-treated an- imals. So, in conjunction with this study, we also quanti- fied the mean cartilage thickness (fig. 4). As shown in this figure, cartilage thickness measurements (in millime-

ABC294640 Attenuates Osteoarthritis in
Rats

Pharmacology 2011;87:135 –143

139Weight bearing (%)
Color version available online
Joint thickness (mm)
Downloaded by:
Columbia University
128.59.222.12 – 3/2/2015 6:41:06 AM

a b c

Fig. 5. Representative safranin-O-stained samples from rat tibias are shown at a magnification of !100. a Nor- mal staining pattern (vehicle/saline treatment). b There is a loss of staining throughout most of the tibial plateau; vehicle/MIA treatment. c The loss of staining was only present in a portion of the tibia; ABC294640 (100 mg/ kg/day)/MIA treatment.

a b c

55

45

35

25

Fig. 6. Representative TUNEL histochemistry samples from the rat tibia. Stained nu-

clei (arrows) represent examples of TUNEL-positive cells. !300. a Vehicle/saline treatment. b Vehicle/MIA treatment. c ABC294640 (100 mg/kg/day)/MIA treatment.
15

Vehicle/saline

Vehicle/MIA

ABC/MIA

d The data summary is shown for the mean percentages of TUNEL-positive cells in
d
(n = 4)
(n = 9)
(n = 9)

the rat knee joints. ABC = ABC294640 treatment; n = 4 or 9 per treatment group.

ters) were 0.21 80.03 (vehicle/saline), 0.36 80.003 (ve- hicle/MIA) and 0.30 80.05 (ABC294640/MIA). There- fore, the mean cartilage thickness in the knee joints of ABC294640/MIA-treated rats was not statistically differ- ent from nondisease (vehicle/saline) control animals, whereas that of vehicle/MIA rats was significantly de- creased (p !0.05) in comparison. However, it should be noted that there was no statistical difference between the cartilage thickness values in the ABC294640/MIA and vehicle/MIA treatment groups (fig. 4).
Safranin O staining is often used as a histological method for assessing cartilage proteoglycan in conjunc-

tion with the MIA OA model [6, 7]. As shown in figure 5, less safranin O staining (more proteoglycan loss) was evident in the knee joint of the vehicle/MIA-treated rat, as opposed to the ABC294640-treated animal. To be specific, loss of staining was evident throughout most of the tibial plateau with vehicle treatment, but was only present in a portion of the tibia from the ABC294640- treated rat.
Quantification of this data resulted in the follow- ing knee joint proteoglycan scores: 3.0 81.8 (vehi- cle/saline), 16.0 81.3 (vehicle/MIA) and 12.7 81.6 (ABC294640/MIA). Overall, there was increased pro-

140

Pharmacology 2011;87:135 –143

Fitzpatrick /Green /Maines /Smith TUNEL-positive cells (%)
Downloaded by:
Columbia University
128.59.222.12 – 3/2/2015 6:41:06 AM
Color version available online
Color version available online

teoglycan staining in the knee joints of ABC294640- treated rats. However, statistical significance compared to vehicle/MIA treatment was not evident in these drug- treated animals.

Effects of ABC294640 on MIA-Induced Chondrocyte Apoptosis
TUNEL histochemistry was used to evaluate chondro- cyte apoptosis in the right knee joints of animals, 28 days after the injection of saline or MIA. Representative pho- tographs (tibia) are shown in figure 6a–c.
Quantification of these results (fig. 6d) resulted in the following values for the mean percents of positive TUNEL cells: 20.7 85.6 (vehicle/saline), 39.5 87.4 (vehicle/MIA), and 25.8 84.7 (ABC294640/MIA). Therefore, chondro- cyte apoptosis was reduced in ABC294640-treated rats. Due to experimental variability, statistical significance compared to vehicle treatment was not attained.

Discussion

Pain in OA is purportedly related to degeneration of articular cartilage and subchondral bone [4, 19]. During the past decade, animal models of OA have been used to study the pain associated with the arthritic process in ro- dents. Various investigators have utilized the MIA model to study pain responses by different methods [5, 6, 19]. One commonly used technique involves the use of an in- capacitance meter to measure pain, by means of differ- ences in hind limb weight bearing [4, 6, 8]. Pomonis et al. [4] showed that MIA caused a concentration-dependent decrease in weight bearing, with an optimal response oc- curring with 3 mg of MIA. Based on these results, we chose this concentration of MIA for use in our study. As anticipated from the relevant literature, MIA caused a clear reduction in weight bearing that was most evident on study day 2, but persisted throughout the study [4, 8]. A similar weight bearing pattern was shown previously by other investigators [4].
Tramadol is an analgesic drug, which has been shown to effectively attenuate the altered weight bearing associ- ated with the injection of MIA into the knee joints of rats [5]. In a similar fashion, we showed that the acute oral administration of tramadol (100 mg/kg) tended to nor- malize weight bearing responses in rats (fig. 1). These re- sults support the validity of our model for measuring OA- associated pain in rats.
Interestingly, the effects of ABC294640 on weight bearing were more pronounced at later time points in the
study (days 20 and 28). Notably, by day 28, there was no difference in mean weight bearing values between vehi- cle/saline- and ABC294640/MIA-treated rats. A further analysis of the time course data (fig. 1b) showed that the slopes of the regression lines were significantly different in sham and ABC294640-treated rats. These results sug- gest an enhanced threshold to weight bearing related pain in ABC294640-treated animals that developed over the course of the study.
In a similar fashion, there was evidence of less carti- lage degradation and bone disruption in ABC294640- treated rats on day 28 (fig. 2). Quantification of these re- sults showed a reduction in the mean knee joint histology score of ABC294640-treated rats as compared to vehicle/ MIA-treated animals. Of note, this reduction in the joint histology score reached statistical significance, if 1 ABC294640-treated animal with an aberrantly high his- tology score was removed from the analysis (fig. 2g). The reason(s) for this single aberrant response remain unde- termined, but could relate to either a true pharmacologi- cal nonresponder or a rat with nonoptimal distribution of the drug to the knee joint. As with the joint histology score data, other measured histological parameters in this study also attained statistical significance, if data from this specific animal were excluded from the statisti- cal analyses (data not shown).
Our data (fig. 3) show a significant inverse association between the joint weight bearing and histology data in this study. Interestingly, 5 of the ABC294640-treated rats had values that clustered in the left upper quadrant of the graph, along with the non-MIA-treated animals. Cluster- ing in this quadrant suggests normal patterns of biologi- cal responses. In contrast, 5 of the vehicle/MIA-treated rats clustered towards the bottom right of the graph. These animals had evidence of enhanced pain responses, associated with significant underlying joint pathology. It is likely that the pain-relieving effect of ABC294640 in the MIA model was secondary to a reduction in underly- ing cartilage and/or bone destruction. This statement is based on the observation that SK inhibition reportedly enhances pain responses in rats [21]. Therefore, it is un- likely that ABC294640 had any inherent analgesic effects on weight bearing in rats. However, we did not directly test that possibility in this particular study.
Due to the cartilage degradation associated with MIA administration to rats, it is well known that the articular cartilage thickness decreases during the development of OA [4, 7]. Results from our study (fig. 4) show that treat- ment of MIA-injected rats with ABC294640 over a 1-month period partially reversed the significant de-

ABC294640 Attenuates Osteoarthritis in
Rats

Pharmacology 2011;87:135 –143

141Downloaded by:
Columbia University
128.59.222.12 – 3/2/2015 6:41:06 AM

crease in cartilage thickness that was apparent in vehicle/ MIA-treated animals. Previous investigators have also used safranin O staining in conjunction with the MIA OA model [22, 23]. Our results show a reduction in saf- ranin O staining (fig. 5) and the calculated cartilage pro- teoglycan score of vehicle/MIA-treated rats. The reduc- tion in the cartilage proteoglycan score was partially at- tenuated in the ABC294640 treatment group. In this regard, there was a highly significant association (r = 0.9502, p ! 0.001) between the cartilage histology and proteoglycan scores in our study. Taken as a whole, these data suggest that ABC294640 attenuates the cartilage and proteoglycan loss associated with MIA-induced OA.
ABC294640 is a selective SK2 inhibitor [17]. Overex- pression of the SK isoforms (SK1 and SK2) has opposing biological actions with regard to cellular proliferation and apoptosis [12, 13]. Specifically, SK2 overexpression is reported to inhibit cell growth and enhance apoptosis [12, 13]. Mechanistically, Lee et al. [24] have shown that MIA-induced chondrocyte apoptosis is mediated by the enhanced expression of TNF-related apoptosis-inducing ligand (TRAIL) and its associated death receptor DR4. Interestingly, TRAIL-induced apoptosis in rhabdomyo- sarcoma cells was shown to be dependent on components of the sphingolipid-signaling pathway [25]. Therefore, a

References
selective SK2 inhibitor (e.g. ABC294640) was hypothe- sized to attenuate the enhanced chondrocyte apoptosis associated with MIA administration to rats [10, 24].
The chronic oral administration of ABC294640 to MIA-injected rats resulted in a reduction in the mean percentage of TUNEL-stained chondrocytes (fig. 6 a). Due to the inherent variability in measuring chondrocyte apoptosis with the TUNEL method [10], statistical sig- nificance compared to sham/MIA treatment was not at- tained (fig. 6b). In another study, treatment with a non- selective SK inhibitor blocked apoptosis of cells. It is fea- sible that a similar mechanism may have been operative in this study, which could explain the mitigating effect of ABC294640 on chondrocyte apoptosis [24, 25]. More- over, it is possible that the reduction of apoptosis by ABC294640 contributed to the reduced joint histological pathology with ABC294640 treatment (fig. 3, 5 and 6).
In summary, treatment of rats with a selective SK2 inhibitor (ABC294640) for a 4-week period attenuated various parameters of knee joint histological damage. ABC294640 treatment also improved the weight bearing pain response associated with the MIA-induced OA model in rats. Therefore, based on this preclinical data, future clinical studies of this agent may be warranted in patients with OA.

⦁ Altmann RD: Early management of osteo- arthritis. Am J Manag Care 2010; 16(suppl): S41–S47.
⦁ Martel-Pelletier J, Boileu C, Pelletier JP, Roughley PJ: Cartilage in normal and osteo- arthritic conditions. Best Pract Res Clin Rheumatol 2008; 22: 351–384.
⦁ Chandran P, Pai M, Blomme EA, Hsieh GC, Decker MW, Honore P: Pharmacological modulation of movement-evoked pain in a rat model of osteoarthritis. Eur J Pharmacol 2009; 613: 39 –45.
⦁ Pomonis JD, Boulet JM, Gottshall SL, Phil- lips S, Sellers R, Bunton T, Walker K: Devel- opment and pharmacological characteriza- tion of a rat model of osteoarthritis pain. Pain 2005; 114: 339 –346.
⦁ Combe R, Bramwell S, Field MJ: The mono- sodium iodoacetate model of osteoarthritis: a model of chronic nociceptive pain in rats? Neurosci Lett 2004; 370: 236 –240.
⦁ Fernihough J, Gentry C, Malcangio M, Fox A, Rediske J, Pellas T, Kidd B, Bevan S, Win- ter J: Pain related behavior in two models of osteoarthritis in the rat knee. Pain 2004; 112: 83 –93.
⦁ Janusz MJ, Hoofkin EB, Heitmeyer SA, Woessnert JF, Freemont AJ, Hoyland JA, Brown KK, Hsieh LC, Almstead NG, De B, Natchus MG, Pikul S, Taiwo YO: Moderation of iodoacetate-induced experimental osteo- arthritis in rats by matrix metalloproteinase inhibitors. Osteoarthritis Cartilage 2001; 9: 751–760.
⦁ Bove SE, Calcatera SL, Brooker RM, Huber CM, Guzman RE, Juneau PL, Schrier DJ, Kilgore KS: Weight bearing as a measure of disease progression and efficacy of anti-in- flammatory compounds in a model of mono- sodium iodoacetate-induced osteoarthritis. Osteoarthritis and Cartilage 2003; 11: 821– 830.
⦁ Del Carlo M, Loeser RF: Cell death in osteo- arthritis. Curr Rheumatol Rep 2008; 10: 37– 42.
⦁ Watrin-Pinzano A, Etienne S, Grossin L, Gaborit N, Cournil-Henrionnet C, Mainard D, Netter P, Gillet P, Galois L: Increased apoptosis in rat osteoarthritic cartilage cor- responds to degenerative chondral lesions and concomitant expression of caspase-3. Biorheology 2006; 43: 403 –412.
⦁ Olivera A, Spiegel S: Sphingosine kinase: a mediator of vital cellular functions. Prosta- glandins 2001; 64: 123 –134.
⦁ Maceyka M, Sankkala H, Hait NC, Le Stunff H, Liu H, Toman R, Collier C, Zhang M, Sa- tin LS, Merrill AH, Milstein S, Spiegel S: Sphk1 and SphK2, sphingosine kinase isoen- zymes with opposing functions in sphingo- lipid metabolism. J Biol Chem 2005; 280: 37118–37129.
⦁ Karliner JS: Sphingosine kinase regulation and cardioprotection. Cardiovasc Res 2009; 82: 184 –192.
⦁ Limaye V, Xia P, Hahn C, Smith M, Vadas MA, Pitson SM, Gamble JR: Chronic in- creases in sphingosine kinase-1 activity in- duce a pro-inflammatory, pro-angiogenic phenotype in endothelial cells. Cell Mol Biol Lett 2009; 14: 424 –441.
⦁ Lai WQ, Inwan AW, Goh HH, Howe HS, Yu DT, Valle-Onate R, McInnes IB, Melendez AJ, Leung BP: Anti-inflammatory effects of sphingosine kinase modulation in inflam- matory arthritis. J Immunol 2008; 181: 8010 – 8017.

142

Pharmacology 2011;87:135 –143

Fitzpatrick /Green /Maines /Smith Downloaded by:
Columbia University
128.59.222.12 – 3/2/2015 6:41:06 AM

⦁ Baker DA, Barth J, Chang R, Obeid LM, Gilekson GS: Genetic sphingosine 1 defi- ciency significantly decreases synovial in- flammation and joint erosions in murine TNF- -induced arthritis. J Immunol 2010; 185: 2570 –2579.
⦁ French KJ, Zhuang Y, Maines LW, Gao P, Wang W, Beljanski V, Upson JJ, Green CL, Keller SN, Smith CD: Pharmacology and an- titumor activity of ABC294640, a selective inhibitor of sphingosine kinase-2. J Pharma- col Exp Ther 2010; 333: 129 –139.
⦁ Fitzpatrick LR, Green C, Frauenhoffer EE, French KJ, Zhuand Y, Maines LW, Upson JJ, Paul E, Donahue H, Mosher TJ, Smith CD: Attenuation of arthritis in rodents by a novel orally-available inhibitor of sphingosine ki- nase. Inflammopharmacology 2010, E-pub
ahead of print.
⦁ Beyreuther B, Callizot N, Stohr T: Antinoci- ceptive efficacy of lacosamide in the mono- sodium iodoacetate rat model of osteoarthri- tis pain. Arthritis Res Ther 2007; 9:R14.
⦁ Cifuentes CJ, Rocha LG, Silva LA, Brito AC, Rueff-Barosso CR, Porto LC, Pinho RA: De- crease in oxidative stress and histological changes induced by physical exercise cali- brated in rats with osteoarthritis induced by monosodium iodoacetate. Osteoarthritis Cartilage 2010; 18: 1088 –1095.
⦁ Coste O, Brennels C, Linke B, Pierre S, Mau- erer C, Becker W, Schmidt H, Gaio W, Geisslinger G, Scholich K: Sphingosine-1 phosphate modulates spinal nociceptive pro- cessing. J Biol Chem 2008; 283: 32442 –32451.
⦁ Al-Saffar FJ, Ganabadi S, Yaakub H, Fakura- zi S: Collagenase and sodium-iodoacetate- induced experimental osteoarthritis model in Sprague Dawley rats. Asian J Sci Res 2009; 2: 167–179.
⦁ Del Valle L, Ochalon A, Cohen S, Patoka R, Zozulya G, Barer F, Atar E, Pina-Oveido S, Perez-Liz G, Castel D, Fishman P: Induction of anti-inflammatory effect and prevention of cartilage damage in rat knee osteoarthritis by CF101 treatment. Arthritis Rheum 2009; 60: 3061–3071.
⦁ Lee SW, Lee HJ, Chung WT, Choi SM, Rhyu SH, Kim DK, Kim KT, Kim JY, Kim JM, Yoo YH: TRAIL induces apoptosis of chondro- cytes and influences the pathogenesis of ex- perimentally induced rat osteoarthritis. Ar- thritis Rheum 2004; 50: 434 –442.
⦁ Petak I, Vernes R, Szucs KS, Anozie M, Izer- adjene K, Douglas L, Tillman DM, Phillips DC, Houghton JA: A caspase-8-independent component in TRAIL/Apo-2-l-induced cell death in human rhabdomyosarcoma cells. Cell Death Differ 2003; 10: 729 –739.

ABC294640 Attenuates Osteoarthritis in
Rats

Pharmacology 2011;87:135 –143

143Downloaded by:
Columbia University
128.59.222.12 – 3/2/2015 6:41:06 AM