Doxy

Matrix metalloproteinases (MMPs) may have a role in the process of aseptic loosening. Doxycycline has been shown to inhibit MMPs. Our aim was to investigate the potential pharmacological effect of doxycycline on aseptic loosening. We used radiolabelled mouse calvariae cultured with human interface membrane cells from aseptically loosened hips.

Bone resorption was confirmed in this model. The effect of doxycycline was assessed by culturing dead radiolabelled bone discs with cells from the interface membrane with doxycycline. The control group consisted of the same culture system without doxycycline. Supernatant 45calcium and the total 45calcium remaining in the bone discs at the completion of the culture were used to measure osteolysis.

We found that doxycycline can inhibit osteolysis at the interface membrane of aseptically loosened hips. This may have therapeutic implications for the treatment of patients with aseptic loosening of total joint replacements.

J Bone Joint Sw;q IRr] 2003;85-B:456-6 1. Received /6 Jat-x 2002; Accepted 14 June 2002

Total joint replacement is commonly performed for patients with end-stage arthritis. The most common cause of longterm failure in total joint replacement is aseptic loosening and this usually requires complex revision surgery. The most consistent feature of aseptic loosening is the presence of a soft-tissue membrane at the prosthesis-bone interface which was first described in 1976 by Harris et al" in a series of case reports. This membrane was found to contain macrophages, multinucleated giant cells, fibroblasts and wear particles.2 Goldring et al-t showed that in tissue culture it had the ability to produce prostaglandin E2 and collagenase, which can influence bone resorption. They proposed that the interface membrane played a major role in bone lysis and suspected that debris particles may be involved in its formation.3

Wear particles including metal,4,5 polyethylene,6,7 polymethylmethacrylate 8 and ceramic 9 can activate macrophages. Cytokines are released during the phagocytosis of these wear particles by macrophages.10 They are able to stimulate osteoclastic bone resorption 10, 11 and also to convert macrophages into osteoclast-like bone-resorting cells.12

Osteoclasts resort bone by acidification and dissolution of crystals of hydroxyapatite, followed by proteolysis of the matrix of the bone collagen.13 Matrix metalloproteinases (MMPs) are a family of at least 16 homologous zincdependent enzymes which are able to degrade almost all the components of the extracellular matrix. 14 They are found in osteoclasts and are believed to participate in both physiological and pathological bone resorption.14,15 MMPs are also found in abundant quantity in the interface membrane of aseptically loosened hip replacements and are believed to play a major role in the periprosthetic osteolysis which leads to loosening. 16-18

Previous studies have shown that subantimicrobial doses of tetracyclines can inhibit MMPs 19 and doxycycline was found to be the most potent inhibitor.20 It works by chelating zinc which prevents the activation of MMPs.21 Doxycycline has also been shown to inhibit bone resorption by chick osteoclasts.22 MMPs may represent the final common pathway in bone resorption irrespective of the type of wear particle. Aseptic loosening could potentially be prevented or treated by inhibiting the pathway of MMPs. As there has been no study of the potential of doxycycline as an inhibitor of MMPs in the treatment of aseptic loosening, our aim was to investigate the ability of doxycycline to inhibit bone resorption using an in vitro radiolabelled mouse calvaria model.23.24 The first part of the experiment was designed to validate our model and the second part to test the effect of doxycycline on bone resorption.

Materials and Methods

Radiolabelled calvaria preparation. Newborn CDI mice were injected with 0.005 mCi of^sup 45^ calcium chloride (Amersham Life Science Ltd, Amersham, UK) into the peritoneal cavity on days 1 and 3. The radiolabelled calvariae were harvested on day 5 and the soft tissue carefully removed by dissection. They were halved along the median suture and the parietal bone was isolated from each hemicalvaria. A punch biopsy knife (Stiefel, Bucks, UK) was used to remove two 4 mm bone discs from each parietal bone. These were subjected to three cycles of freezing and thawing to devitalise all the cells in the bone before use. Those from the same parietal bone were used as the experimental group and control group for comparison 24 because of the variability in the uptake of radioactive calcium between different mice. When the same calvaria is used variability in the uptake of 45 calcium is minimal.24

Culture medium. Alpha minimal essential medium (Alpha MEM; Gibco, Paisley, UK) was supplemented with 100 U/ ml of penicillin, 10 mcg/ml of streptomycin, 10 mmol Lglutamine (Sigma, Dorset, UK). 10%lc heat-inactivated fetal calf serum (Gibco), 25 ng/ml of recombinant macrophage colony-stimulating factor (M-CSF; Peprotech, London, UK), 100 ng/ml of RANK ligand (RANKL) (Peprotech), 1,25-dihydroxyvitamin D3 (10^sup -7^ M) (Sigma) and hydrocortisone (10^sup -6^ M) (Sigma).25 In the doxycycline experiment 15 mcg/ml of doxycycline were added to the culture medium. The 1,25-dihydroxyvitamin D^sub 3^ was dissolved in absolute alcohol and stored at -20 deg C.

Cell isolation from the interface membrane of aseptically loose hips. Interface membranes from six aseptically loose total hip replacements were obtained during revision surgery. Table I gives the clinical details of the patients. The tissue specimens were processed within one hour after explantation. Further specimens were sent to the microbiology laboratory for culture and the pathology department for histological examination to exclude infection.

The cell-isolation technique was previously described by Sabokbar et al.25 Tissue specimens were washed thoroughly with phosphate-buffered saline (PBS) to remove all the blood clots. They were then cut into small fragments and digested in alpha MEM containing 1 mg/ml of collagenase type 1 (Sigma) for 30 minutes at 37 deg C. This was followed by a further incubation for one hour in 0.25% trypsin (Sigma). The digested tissue was filtered with a 70 micrometer cell strainer (Falcon, Oxford, UK) and the fluid contaming the cells was centrifuged at 800 g for ten minutes. The cell pellet was resuspended and any red blood cells were lysed by adding 10 ml of ice-cold sterile water for 30 seconds followed by a further 15 ml of ice-cold PBS. The cell mixture was centrifuged again at 800 g for ten minutes and followed by resuspension of the cell pellet. The cells were counted using a haemacytometer.

Cell culture. Culture dishes with 24 wells (Nunclon; Nalge Nunc International, Hereford, UK) were used with incubation conditions set at 37 deg C in air with 5% CO^sub 2^. The experiment was divided into two parts. The first was to show that bone resorption was present in the model and the second to investigate the inhibitory effect of doxycycline. In the first experiment, the test culture consisted of 1 x 10^sup 5^ cells with a radiolabelled bone disc and 2 ml of supplemented culture medium, and the control culture consisted of a bone disc with supplemented culture medium only. In the second experiment, the test culture consisted of 1 x 10^sup 5^ cells with a bone disc and supplemented culture medium with 15 mcg/ ml of doxycycline. The control culture consisted of 1 x 10^sup 5^ cells with a bone disc and supplemented culture medium but no doxycycline.

The test and control cultures were incubated for 14 days. The supernatant was sampled on days 3, 7 and 10 by removing 1 ml of medium for scintillation counting then replenishing the well with an equal amount of fresh medium. At the end of culture on day 14, a final supernatant sample was taken. The bone disc was then removed, washed and dissolved in 1M HCl for 72 hours. Scintillation counting was done on the acid sample to determine the total radioactive calcium remaining in the bone disc. In the first experiment on day 14 after removal of the last supernatant sample and the bone disc, 0.5 ml of 0.25% trypsin (Sigma) was added to the test and control well. This was to release the cells in the test culture well and in the control well to ensure that there was the same amount of dilution factor as the test well. These samples were removed for scintillation counting.

In the second experiment a similar amount of trypsin was added to both wells for approximately 30 seconds and the cells were sampled and stained with Trypan Blue. The number of viable and non-viable cells was counted by using the haemacytometer.26 Cell counting was repeated three times for each well and the average was taken as the final count.

Statistical analysis. All data were analysed statistically using Minitab 10.5 software (Minitab Inc, Pennsylvania). ANOVA (General Linear Model) and the paired t-test were used for analysis of data. Differences in the mean values at a level of p

Results

The results were expressed as scintillation count ratios of the bone disc exposed to cells (BC) over the bone disc only (B) in the first experiment to confirm bone resorption. In the second experiment they were expressed as scintillation count ratios of the bone disc exposed to cells and doxycycline (Doxy) over the bone disc exposed to cells only (BC).

In the preliminary experiments to confirm bone resorption, eight paired bone discs were used.

Supernatant 45calcium. The mean ratios of scintillation counts (BC/B) at days 3, 7, 10 and 14 were 0.83 (0.71 to 0.95), 0.88 (0.70 to 1.04), 0.97 (0.77 to 1.11) and 1.08 (0.86 to 1.46), respectively (Table II). The scintillation count ratio showed an increase in release of supernatant 45calcium with time (ANOVA, p = 0.0001; Fig. 1).

Total 45calcium remaining in bone discs. The mean ratio of scintillation counts of the total 45calcium remaining in the bone discs at the completion of culture (BC/B) was 0.81 (0.59 to 0.94) (Table III). The results show that more bone (19%) remained in the absence of cells compared with the presence of cells (paired t-test, p = 0.02), indicating bone resorption by cells.

45calcium at the bottom of culture wells. The mean ratio of scintillation counts of 45calcium remaining at the bottom of the well at the completion of the culture (BCB) was 1.80 (1.03 to 3.85) (paired t-test, p = 0.02) (Table IV) indicating that more 45calcium was present at the bottom of the culture wells containing interface membrane cells. This could be due to either an increase in calcium uptake by the cells or to the formation of an organic calcium precipitate.

In the second experiment to investigate the effect of doxycycline on bone resorption, again eight paired bone discs were used.

Supernatant 45calcium. The mean ratios of supernatant scintillation counts derived from bone exposed to doxycycline (doxy) over bone disc not exposed to doxycycline (cont) on days 3, 7, 10 and 14 were 0.94 (0.81 to 1.06), 0.88 (0.76 to 1.07), 0.87 (0.65 to 1.01) and 0.81 (0.59 to 0.97), respectively (ANOVA, p = 0.003) (Fig. 2; Table V). There was more 45calcium released into the supernatant of the culture well with no doxycycline suggesting that bone resorption was more active in the absence of doxycycline. Total 45calcium remaining in bone disc. The mean ratio of scintillation count of the total 45calcium remaining in the bone disc at the completion of culture (doxy/cont) was 1.21 (1.01 to 1.49) (Table VI). Thus there was more bone (21%) left in the doxycycline exposed bone discs compared with bone discs not exposed to doxycycline (paired t-test, p = 0.02). This implies that there was less bone resorption in the presence of doxycycline.

Cell viability. Cell viability in both the doxycycline and non-doxycycline cultures was almost the same, 86% (82 to 90) in the doxycycline culture and 85% (78 to 89) in the non-doxycycline culture (ANOVA, p = 0.63). Hence doxycycline was not cytotoxic.

Discussion

Scintillation count ratios in the supernatant would theoretically equal 1 if there was no difference in the 45calcium release between the test and control cultures. Any rise in the ratio of test over control would suggest bone resorption. Assessment at zero time would theoretically produce 0/0 which would be difficult to interpret and thus scintillation counts at zero time were not taken. The overall trend of increased supernatant 45calcium in the test cultures compared with the control cultures with time is shown in Figure 1, and is in keeping with bone resorption. The initial supernatant scintillation count ratio was less than one and initially seemed opposite to that expected. Could this represent bone formation? However, sampling of the bottom of the wells on day 14 showed scintillation counts which were 80% higher in the wells with cells compared with those without cells. The initial decrease in ratio can be explained by either the intake of 45calcium by the cells or organic calcium precipitates. At the later part of the culture the amount of 45calcium released by bone resorption overtook the total amount of absorbed or precipitated 45calcium which resulted in a net increase in the supernatant 45calcium ratio above 1 on day 14.

In our study the paired bone discs from each test-control comparison were from the same parietal bone. The total amount of 45calcium in each bone disc should be very nearly the same at the start of the culture.24 The total 45calcium remaining in the bone discs which were exposed to cells was 19% less than in the control discs which were incubated alone. We believe that these results confirm that bone resorption was present in our model.

In the second part of the study, the scintillation counts for supernatant 45calcium in the test well (doxycycline, cells and bone disc) were consistently higher compared with the controls (cells and bone disc) throughout the entire experiment, This was reflected in the ratio of supernatant 45calcium remaining below I as shown in Figure 2. Any calcium uptake or organic precipitate should be the same in test and control cultures and the confusing results seen on the bone resorption part of the study would not be expected in this latter part of the study examining the effect of doxycycline. The total 45calcium remaining in the bone discs on day 14 was higher in the doxycycline-containing culture wells compared with the wells without doxycycline (Table VI). Thus the presence of doxycycline inhibited bone resorption by interface membrane cells. This inhibitory effect was not due to cell toxicity since tests for cell viability were similar in both cultures. It has been shown in previous studies that doxycycline can inhibit osteoclasts both in vitro27 and in vivo.28 This is the first time that it has been shown to inhibit bone resorption by interface membrane cells. We believe that inhibition of MMP by doxycycline is a probable explanation.

Tetracycline is taken up and incorporated into hydroxyapatite crystals together with calcium in the inorganic bone matrix. On degradation of the bone, tetracycline is released. yielding high concentrations of localised tetracycline.29 The concentration of doxycycline in periodontal tissue was about ten times greater when compared with the serum level.30 The mean serum concentration of doxycycline when taken at a dose of 200 mg a day levels out at 1.5 mcg/ml. This is the recommended clinical dose for its antibiotic effect.31 In our study the concentration of doxycycline was set at 15 mcg/ml based on the ability of bone to concentrate doxycycline. Furthermore, a concentration of as low as 5 mcg/ml has been shown to inhibit bone resorption by chick osteoclasts.22 Previous studies have shown that pharmacological inhibition of aseptic loosening may be a realistic prospect. Bisphosphonates have been shown to inhibit bone resorption by activated macrophages 32 and osteoclasts when co-cultured with activated macrophages.24 They also reduce bone lysis in a canine model of aseptic loosening.33 Pentoxyfilline (Tentral), an inhibitor of a tumour necrosis factor alpha, has been shown to inhibit the activation of human monocytes by particulate debris 34 and of debrismediated bone lysis in a mouse model. Doxycycline is another possible agent which could influence aseptic loosening either alone or in combination with other agents. It is relatively inexpensive and is used extensively as an antibiotic with minimal reported adverse reactions. Further in vitro and in vivo studies are required to confirm the effectiveness of doxycycline in the treatment of aseptic loosening.

We wish to thank the Wishbone Trust for funding this study and Pfizer for donating the purified doxycycline.

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

References

1. Malchau H, Herberts P. SWerman P. Oden A. Prognosis of total hip replacement: update and validation of results from the Swedish National Hip Arthroplasty Register 1979-1998. 67th Annual Meeting American Academy of Orthopaedic Surgeons. 2000.

2. Harris WH, Schiller AL, Scholler JM, Freiberg RA, Scott R. Extensive localized bone resorption in the femur following total joint replacement. J Bone Joint Surg,/Am/1976:58-A:612-8.

3. Goldring SR, Schiller AL, Roelke M, et al. The synovial-like membrane at the bone cement interface in loose total hip replacement and its proposed role in bone lysis. J Bone Joint Surg /Am/1983;65A:575-84.

4. Haynes DR, Rogers SD, Hay S, Pearch MJ, Howie DW. The differences in toxicity and release of bone-resorbing mediators induced by titanium and cobalt-chrome alloy particles. J Bone Joint Surg [Ami] 1993:75-A:799-801.

5. Pazzaglia UE, Cecilliani I,, Wilkinson MJ, Dell'Orbo C. Involvement of metal particles in loosening of metal-plastic total hip prostheses. Arch Orthop Trauma Sung 1985:104:164 74.

6. Boynton E, Waddell JP, Morton J, Gardiner GW. Aseptic loosening in total hip implants: the role of polyethylene wear debris. Can J Surg 1991:34:599-605.

7. Willert HG, Bertram G, Buchhorn GH. Osteolysis in alloarthroplasty of the hip: the role of ultra high molecular weight polyethylene wear particles. Clin Orthop 1990:258:95-107.

8. Horrowitz SM, Doty SB, Lane JM, Burstein AH. Studies of the mechanism by which the mechanical failure of polymethylmethacrylate leads to bone resorption. J Bone Joint Surg JAin] 1993:75-A:802-13.

9. Yoon TR, Rowe SM, Jung ST, Seon KJ, Maloney WJ. Osteolysis in association with total hip arthroplasty with ceramic bearing surfaces. J Bone Joint Surg [Ain / 1998:80-A: 1459-68.

10. Horrowitz SM, Purdon MA. Mechanisms of cellular recruitment in aseptic loosening of prosthetic joint implants. Calcif Tissue ]nt 1995;57:301-5.

11. Pollice PF, Silverton S, Horrowitz SM. Polymethylmethacrylate stimulated macrophages increase osteoclast precursor recruitment through their effect on osteoblasts in vitro. J Orthop Res 1995;13:325-34.

12. Athanasou NA, Quinn J, Bulstrode CJK. Resorption of bone by inflammatory cells derived from the joint capsules of hip arthroplasties. J Bone Joint Surg [Br] 1992:74-B:57-62.

13. Sims N, Baron R. Bone cells and their function. In: Ernesto Canalis, ed. Skeletal growth factors. Lippincott Williams & Wilkins, 2000:1 16.

14. Nagase H, Woessner JF, Jr. Matrix metalloproteinases. J Biol Chem 1999;274:21491-4.

15. Okada Y, Naka K, Kawamura K, et al. Localization of matrix metalloproteinase 9 (92-kilodalton gelatinase/type IV collagenase = gelatinase B) in osteoclasts: implications for bone resorption. Lab Investig 1995:72:311-22.

16. Takagi M, Konttinen Y, Santavirta S, et al. Extracellular matrix metalloproteinases around loose total hip prostheses. Acta Orthop Scand 1994;65:281-6.

17. Yokohama Y, Matsumoto T, Hirakawa M, et al. Production of matrix metalloproteinases at the bone-implant interface in loose total hip replacements. Lab Invest 1995;72:899-91 1.

18. Yu Y, Walsh WR, Sonnabend DH, et al. Cytokines and matrix metalloproteinases mRNA expression in archival human tissues from failed hip arthroplasty using in situ hybridization and color video image analysis. Bull Hosp J Dis 1998:57:23-9.

19. Greenwald RA, Golub LM, Lavietes B, et al. Tetracyclines inhibit human synovial collagenase in vivo and in vitro. J Rheumatol 1987;14:28-32.

20. Golub LM, Ramamurthy NS, McNamara TF, Greenwald RA, Rifkin BR. Tetracyclines inhibit connective tissue breakdown: new therapeutic implications for an old family of drugs. Crit Rev Oral Biol Med 1991;2:297-322.

21. Ryan ME, Ramamurthy NS, Golub LM. Matrix metalloproteinases and their inhibition in periodontal treatment. Cur Opin Periodontol 1996;3:85-96.

22. Chowdhury MH, Moak SA, Rifkin BR, Greenwald RA. Effect of tetracyclines which have metalloproteinase inhibitory capacity on basal and heparin stimulated bone resorption by chick osteoclasts. Agents Actions 1993;40:124-8.

23. Reynolds JJ, Dingle JT. A sensitive in vitro method for studying the induction and inhibition of bone resorption. Calcif Tissue Res 1970;4:339-49.

24. Horowitz SM, Algan SA, Purdon MA. Pharmacologic inhibition of particulate-induced bone resorption. .1 Biomed Mat Res 1996:31:91-6.

25. Sabokhar A, Fujikawa Y, Neale S, Murray DW, Athanasou NA. Human arthroplasty derived macrophages differentiate into osteoclastic bone resorbing cells. Ann Rheum Dis 1997;56:414-20.

26. Doyle A, Griffiths JB. Haemocytometer cell counts and viability studies. In: Cell and tissue culture for medical research. John Wiley & Sons Ltd. 2000:12-6.

27. Greenwald RA, Golub LM, Ramamurthy S, et al. In vitro sensitivity of the three mammalian collagenases to tetracycline inhibition: relationship to bone and cartilage degradation. Bone 1998;22:33-8.

28. Ramamurthy N, Bain S, Liang CT, et al. A combination of subtherapeutic doses of chemically modified doxycyline (CMT-8) and a bisphosphonate (clodronate) inhibits bone loss in the ovariectomized rat: a dynamic histomorphosphonate and gene expression study. Curr Med Chem 2001:8:295-303.

29. Green J. The physiochemical structure of bone: cellular and non cellular elements. Miner Electrolyte Metab 1994:20:7-15.

30. Vernillo AT, Ramamurthy NS, Golub LM, Rifkin BR. The nonantimicrobial properties of tetracycline for the treatment of periodontal disease. Curr Opin Periodontol 1994:111-8.

31. Doxycycline. In: Dollery C, ed. Therapeutic drugs. 2nd edition. Churchill Livingstone 1999.

32. Pandey R, Quinn JM, Sabokbar A, Athanasou NA. Bisphosphonate inhibition of bone resorption induced by particulate biomaterial-associated macrophages. Acta Orthop Scand 1996:67:221-8.

33. Shanbhag AS, Hasselman CT, Rubash HE. Inhibition of wear debris mediated osteolysis in a canine total hip arthroplasty model. Shanbhag AS, Hasselman CT, Rubash HE. Inhibition of wear debris Orthop 1997;344:33-43.

34. Pollice PF, Rosier RN, Looney RJ, et al. Oral pentoxyfylline inhibits release of turnout necrosis factor-alpha from human peripheral blood monocytes: a potential treatment for aseptic loosening of total joint components. J Bone Joint Surg [an] 2001:83-A:1057-61.

35. Schwarz EM, Benz EB, LIu AP, et al. Quantitative small animal surrogate to evaluate drug efficacy in preventing wear debris induced osteolysis. J Orthop Res 2000; 18:849-55.

S. M. Ong, G. J. S. Taylor

From Glenfield Hospital, Leicester, England

S. M. One, FRCS G, Orthopaedic Research Fellow

G. J. S. Taylor, FRCS Orth Ed, Consultant Orthopaedic Surgeon Department of Orthopaedics, Glenfield Hospital, Groby Road, Leicester LE3 9 QP, UK.

Correspondence should be sent to Mr S. M. Ong.

Copyright British Editorial Society of Bone & Joint Surgery Apr 2003
Provided by ProQuest Information and Learning Company. All rights Reserved