COX-2抑制剂 在人骨折愈合中的作用

BrielReport

Cyclooxygenase-2 Inhibitors in HUD1an

Skeletal Fracture Healing

AARON DALUISKI MD; KERI E. RAMSEY BSC; YUEXIAN SHI PHD; MATHIAS P. BOSTROM   MD;

BRYAN J. NESTOR MD; GEORGE MARTIN   PHD; ROBERT HOTCHKISS   MD; DIETRICHA. STEPHAN   PHD

This article identifiesthe underlying molecular events responsible for fracture nonunions ina subset of fracture patients. Expression profiling of fracture callus tissue from both uneventful fracture repair and nonunion outcomes showed a decrease of COX-2 expression and an inability to mount an im mune response in nonunion fractures. Validation in vitro with Saos-2 osteo­ progenitor cell lines showed a decrease in osteogenesis potential after the cells were treated with celecoxiba COX-2 specific inhibitor and anti-inflam

matory agent. This articl e recapitulates that an initial immune response is

crucial to fracture healing and suggests limitedusage of COX-2 inhibitors in patients with healing fractures.

ib. Th   ce11s w r th   Iys d by fr   :ze/ thaw twic     andtotal alkalìn   phospha­tase actlvJty was m asured by Sigma Di­ agnostics Alkalìn   Phosphatas   kit #245

(Sigma   DiagnosticsSt LouisMo)   at

405 nm on an ELISA plat read r. Alka­ lìn phosphatas is a s nsitivemarker for osteogen sis. Each experimental condi­tlOn was r plicated six times and  th   n­

t       xperim nt was rep atedtwice with

gus (NSA ID)s aωsw e11   aωs   l                                                                            mal f:racture healing and bon   formation  Softwar    San Diego   Calif) to a signifi­

newer cyclooxygenase-2 (COX-     in rod nts.S.6 Reduction in activity   of          cance level of P<.O l.

2) specific inhibitors (rofecoxib   valde-     th COX-2   nzyme   ith r through di-

coxib   and cel coxib) ar   used for th     r ct genetic deletion or   nzymatic iü-         Expression Profiling

treatmentof paìn and ìnflammation ìn pa-   bition reduc s osteogenesis in multipl           N onunion and fractur   ca110us tissu tients with a multitud   of conditions. OVI :r     assays.S-7   Furtherτnor     bon   morpho-   sampl s w :re removed   ìntraop ratively

50   mi11ion pr scriptions   for Celebrex     genetic   prot in-ìnduc d   ostωg nesis     and immediately flash frozen in liquid

(Pfizer N w York NY) andVioxx (M rck     in rod nt models works in part through     nitrog n. Sp cim ns wer   tak n at the

& Co Inc Whitehouse Station   N J) wer       dir ct ìnduction of COX-2ρ                       dir ct site of nonunion or at th   Slt of fi11ed ìn 2001   gen ratìng over5 bi11ion                                                                               th fracture ca11us and not th   surround- do11ars in revenue in that y ar alone. A   MATERIALS AND METHODS

rhBMP2

Cel   l<ib 100mM                                                                           +

Figure: The effect of COX-2 specific inhibition on human osteogenic cellline Saos-2. Baseline ALP levels shown in first column indicate the osteogenic   natu   of the cells. A significant induction of ALP occurs in   sponse to rhBMP-2 (200 mcg/mL) (P<.001;  one way ANOVA). A dose-dependent reduction in ALP occurs in the presence   of the COX-2 inhibitorcelecoxib at 50 and 100 mM (P<.001). A modest but statistically significant difference exists between cells alone (column1) and cells in the presence   of celecoxib at 50 mM (P<.01) This diffe nceIS mo   pronounced with increased celecoxibdosing (100 mM) (see column 1 versus column

6; P<.001). Abbreviations: ALP=alkaline phosphatase mM=millimolar  and rhBMP2=recombinant human bone morphogenic protein-2

ing muscle tissu avoiding all non-callus tissu.The timing of samp1ecoll ctionfor th   nonunion fractur   s was at tím of surgerytypically 13-33 months. The acut   fractur sw r coll ct d within 3 w   ks of injury. Norma1 tissu   fromfive individua1s (mix d sX     s)atanav rag ag   of 63 Y   ars (rang   :35-81 Y   ars) and fiv   nonunion   tissu s (mix d sX     s) at an av rag ag of 55 y ars(rang   :21-80 y ars) wer   coll ct d. Ribonucleic acid(RNA) was iso1atd from   ach samp1e using TRIzo1 R ag nt (Invitrog n Corp Car1sbadCalif) using th         protoco1 sug­ g st d by th   manufactur r.Two micro­ grams of tota1 RNA wer poo1d from th       nonunion samp1 s and a s parat poo1was construct d from th   norma1tíssu   for a tota1 of 10 micrograms in

ach pool. This   pooling strat gy was

perform d again to produc   two replicat s for   achclinica1 stat.Tota1 RNA was th n 1abe1 d using   stablish   d Af

th r is a r markab1 incr as in sp ci­

Abbreviations: A=absentALP=alkaline phosphataseM =marginaland P=present.

*Comparing the differencein gene expression levels betweenhealing fractures and fracture nonunions. Tissue taken from   e fractu   site expressed bone-specc markers such as ALP with onminimaldíe                                         nceinRNA1 el. BMP2two of       e BMP type 1   ceptorsand

sev     1 of the necessaη downstream targets are expressed       th only minimal dijferences

in levels between the two patientgroups. The COX-2 transcript is expressed 13 times lower

linctu                           nonunionpatients as co'ared to patients with unctu healing

RESULTS

Th   failur   of a fractur   to unít   and hea1 is a significant caus   of morbidity. To t st whether or not th r   ar 1ll VIVO differ nc s in COX-2 1V     ls in humans W     coll ct d fra cture sit   tissu  from

COX-2 I NHI BITORS I  N FRACT U RE HEALlN G I     DALUISKI ET AL

COX-2 is an inflammatory mediator involvedin fracture healingin rodentsand the target of widely presribed anti-inflammatoryagents.

Human fracture healing operates througha bottleneck in COX-2 that when inac­

tiveresultsin chronicnonunion fractures.

Patients should not be prescribed COX-2 inhibitors in the first weeks after fracture.

SUPPLEMENTARY INFORMATION

• The Neurogenomics Dívísíon ofTGen advocates timely díssemínation ofdata from genomíc screens. Thís web síte provídes bo   supplementary and raw data files for publícations from the Dívísíon. The web address where the data for   ís study ís hosted ís: http://ww\ tgen.orglneurogenomicsldata. Another resourcethat ís of use íse NIH Neuroscíence.

•Mícroarray Consortium ( arrayconsortium.tgen.org). The consortium ís m de up of fours síteslocated at T GenYaleDuke and UCLA.These four genome ínstitutes

F ovíde expression profiling and single-nucleotide polymorphism scanning exper­

tise and servicesto a user base of approxim tely 10    o scíentists aαoss e coun­

I 1. All data is eely díssemínated to further scientific inquiry at six months after

project completion

•The Gene Expressíon Omníbus (GEO) is anothercentral dataclearinouse for ex­presslOn proling experiments for le research communíty. This resource can be found athttp://www.ncbi.nlm.nih.govlgeol.

patients with un V ntful fracture h aling    in th   human ost oprogenitor c 11 1in and   patients with   fracture non lons       Saos-2. Th s   C l1s  show   an induction xpression profi1ed th   two classesand        of a1ka1in phosphatase in r sponse to then   p rforrn d a stringent supervised    tr atment with rhBMP2 (Figur). In th

ana1ysis.τn       nonunion patients had           presence of ce1 coxiba significant (d

a 13-fo1d reduction in 10ca1 1eve1s of                                                            dose-dep nd nt)   reduction in a1ka1inCOX-2   xpression (Supp1 m ntary In-                     phosphatas   occurs. This inhibition a1so formation). This nascent1ack of COX-2    occurs toth   a1kaline phosphatase base-presumab1y cu1minat   s in an inability to                        lin   l V 1 in th S   c l1sindicating that hea1 throughan  inabi1ity to initiat   a 10-                             COX-2 inhibition reduc s th   ost   og   - ca1 immune r spons This findingwas        ic potentia1 of human osteoprog nitor fo11owed by in vitro validation.                         ce11s. ThusBMP-induced ostωg   eSlS

r quires prostag1andin synth SlS   Vla

DISCUSSION                                                      COX-2 in humans similarto that found

To t st wh th :r COX-2 function is      in mic   and rats.5

n C   ssary   for bone mohogenic pro-

t in   (BMP)-induc d osteog n sis   w         CONCLUSION

m asur   d a1kalin   phosphatas function          Inhibition of COX-2 caus   s a r duc-

tion in ost   ogenesis inhumanfractures. Bas   d on our datawe sugg      st limitingth   us   of traditiona1 NSAIDsas w 11 as th   more se1ectiveCOX-2 inhibitors in patients with healingfractur   ?sp cia11y among e1d r1y pati nts with co­ morbid conditions. Prosp ctive clinica1tria1s are n C ssary to validate the effect of thes   m dications on fractur   h a1-ing rat        to det rrnin   th tru impact on this 1arg   S gm nt of  the popu1ation. :!;

REFERENCES

1. Hochberg MCMelin JMReicinA.Cox-2 inhibitors and fracture healing: an   gu­ ment agalllst su h an effect.J Bone Miner Res. 2003; 18:583

2. Einhorn TA. Doinhibitors of cyclooxygen ase-2 impair bone healing?J Bone Miner Res. 2002; 17:977 978

3. Einhorn TA.Cox-2 inhibitorsand fracture healing: argumentsfor such an e ect and a suggestion about cautio in use. J Bone Miner Res.2003; 18:584

4. O'Connor Jp.Cox-2 inhibitors and fracturehealing guments for such an effect and a suggestion about caution in use. J Bone MinerRes. 2003; 18:585.587

5. Simon AM   1anigrasso MB O 'Connor JP Cyclo-oxygenase 2 function is essentialf01 bone fracture healing.J BoneMiner Res

20ω; 17:963-976

6. Zhang X SchwarzEMYoungDA PuzasJE Rosier RN O'Keefe RJ. Cyclooxygenase-2 regula s mesenchymal cell differentiation into the osteoblast lineage and is critically involved in bone rep r. J Clin Invest. 2002;

109:1405-1415

7. Goodman S Ma TTrindade M et al. COX-2selective NSAID decreasesbone ingrowth in vivo. J Orthop Res. 2002; 20:1164-1169

8. Chikazu D Li X Kawaguchi H et al. Bone morphogenetic protein2 induces cyclo-oxy gena   2 in osteoblasts viaa Cbfa1 binding site: role in effects of bone morphogenetic prot n 2 in vitro and in vivo. J Bone Miner Res. 2002; 17:1430-1440

9. Kloen P Doty SB Gordon E RubelIFGou­ mans MJHe!fet DL.Expression and ac tivation of the B 1P   gnaling components in human fracture nonunions. J Bone Joint SurgAm. 2002; 84:1909-1918

10.  Chen Y V Zhao P Borup R Hoffman EP Expression pro ling in the muscular dys trophies: identi cation of novel aspectsof molecular pathophysiology. J Cell Biol

2000; 151:1321-1336