Dexamethasone acts locally bone growth in rabbits JEFFREY BARON, JOHN D. BACHER,
to inhibit
longitudinal
ZE HUANG, KAREN E. OERTER, AND GORDON B. CUTLER, JR.
Developmental Endocrinology Branch, National Institute of Child Health and Human Development, and Surgery, Radiology, and Pharmacy Section, Veterinary Resources Program, National Center for Research Resources, National Institutes of Health, Bethesda, Maryland 20892 Baron, Jeffrey, Ze Huang, Karen E. Oerter, John D. Bather, and Gordon B. Cutler, Jr. Dexamethasone acts locally to inhibit longitudinal bone growth in rabbits. Am. J. Physiol. 263 (Endocrinol. Metab. 26): E489-E492, 1992.Excess glucocorticoid is a potent inhibitor of epiphysial growth. Several mechanisms have been suggested to explain this growth inhibition, including both direct local effects of glucocorticoid on the epiphysial growth plate and indirect systemic effects. Previous studies do not distinguish which of these proposed mechanisms is actually responsible for the growth suppression in vivo. To resolve this controversy, we developed a method for delivering glucocorticoid directly into the rabbit epiphysial growth plate and for accurately measuring the resulting epiphysial growth rate. Five-week-old male rabbits received a local infusion of dexamethasone phosphate (80 ng/pl, 1 pi/h) into one proximal tibia1 growth plate and an infusion of vehicle into the contralateral growth plate. Growth rate was determined by inserting metal pins into the bone immediately adjacent to the growth plate and measuring the change in distance between pins on serial radiographs. This method permitted growth rates to be measured over intervals as short as 3 days, with an error of m 5%. Local dexamethasone administration decreased proximal tibia1 growth rate by 77% compared with the contralateral vehicle-treated tibia (P < 0.0001). We conclude that excess glucocorticoid causes a rapid potent inhibition of growth by a direct local action on the growth plate. glucocorticoids;
growth
plate; epiphysis
GLUCOCORTICOID IS A POTENT inhibitor ofepiphysial growth in both animals and humans (15, 19). Several mechanisms have been suggested to explain this growth inhibition, including both direct local effects of glucocorticoid on the epiphysial growth plate as well as indirect systemic effects. The proposed indirect systemic effects include decreased growth hormone secretion (2)) decreased circulating insulin-like growth factor I (IGF-I) activity (6), and increased circulating IGF inhibitors (27). On the other hand, a local direct effect on epiphysial growth is suggested by in vitro studies in which supraphysiological concentrations of glucocorticoid decreased DNA and proteoglycan synthesis by growth plate chondrocytes (9, 12, 14, 26). Previous studies do not distinguish which of these proposed mechanisms are actually responsible for the growth suppression in vivo. To resolve this controversy, we developed a method to deliver glucocorticoid directly into the rabbit epiphysial growth plate and to measure accurately the resulting epiphysial growth rate. EXCESS
METHODS Growth plate infusion. Five 4-wk-old male New Zealand White rabbits were purchased from Hazleton Research Products (Denver, PA). The animals received National Institutes of Health Open Formula Rabbit Ration (NIH32) and water ad libitum. At 5 wk of age, they were given a single prophylactic
intramuscular dose of penicillin G (44,000 IU/kg) and premedicated with intramuscular glycopyrrolate (0.02 mg), ketamine hydrochloride (33 mg/kg), and xylazine (6.6 mg/kg). They were then intubated and placed under general anesthesia with isoflurane. An incision was made to expose the proximal anteromedial tibia. Sterile infusion sets, which consisted of a 27-gauge stainless steel needle attached to a flexible plastic catheter (Baxter Healthcare, Hookset, NH), had been prepared beforehand for implantation. The needle had been bent to place a right angle 2.5 mm from the sharp tip. This tip was then inserted into the proximal tibia1 growth plate close to the junction with the bony epiphysis (Fig. 1). The shaft of the needle was bent into a “U” shape, the distal portion of which was sutured to the periosteum of the tibia. The plastic catheter was then tunneled subcutaneously to a second incision over the lower abdomen on the ipsilateral side. The free end of the catheter was attached to an osmotic pump (model 2001; Alza, Palo Alto, CA) that was placed in the subcutaneous space. The pump was designed to administer 1 pi/h for 7 days. To determine whether the osmotic pump was capable of delivering a solution into the growth plate, we had previously infused methylene blue dye using this system and found that the dye stained the entire growth plate. In other experiments, we had measured the residual fluid remaining in the pump at the end of 6 days and demonstrated that the pump rate in vivo did match the manufacturer’s specification (data not shown). An osmotic pump and infusion set were placed on each of the two sides of the animal. One osmotic pump contained a solution of dexamethasone phosphate. This solution had been prepared by diluting dexamethasone sodium phosphate (equivalent to 4 mg/ml dexamethasone phosphate in water containing 1 mg/ml anhydrous sodium sulfite, 19.4 mg/ml anhydrous sodium citrate, and 0.01 ml/ml benzyl alcohol, pH 7.0-8.5; Elkins-Sinn, Cherry Hill, NJ) in Hanks’ balanced salt solution (HBSS) (3) without phenol red to yield a final concentration of 80 ng/pl dexamethasone phosphate. The contralateral pump contained HBSS, to which had been added sodium sulfite, sodium citrate, and benzyl alcohol, such that the concentrations of these substances were equal in the two pumps. Growth rate measurement. Three metal pins were also placed in each proximal tibia. The pins were prepared beforehand by breaking 30-gauge hypodermic needles (Becton Dickinson, Rutherford, NJ) 5 mm from the tip. The 5-mm fragments were then subjected to steam sterilization. All three pins were inserted parallel to the growth plate, directed posteriorly (Fig. 1). One pin was placed in the bony metaphysis -2-3 mm distal to the growth plate and just medial to the anterior tibia1 tuberosity. A second pin was placed -5 mm distal to the first pin and also directed posteriorly. The final pin was placed -5 mm proximal to the first pin. This last pin was thus 2-3 mm proximal to the growth plate and within the bony epiphysis. Serial radiographs of the tibiae were obtained on postoperative days 0, 5, 8, 12, 15, 19, and 22. For the radiograph, the animal was sedated with injection of ketamine hydrochloride (33 mg/kg im) plus xylazine (6.6 mg/kg). The rabbit was then placed in the lateral position on an X-ray cassette. The hindleg of interest was placed closer to the cassette and anterior to the E489
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.092.098.118) on August 14, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
E490
GLUCOCORTICOIDS Proximal
AND
BONE
GROWTH
growth rate on the dexamethasone-treated side to the growth rate on the vehicle-treated side differed significantly from one. RESULTS
Dlstal Fig. 1. Diagram of rabbit proximal tibia after surgical implantation of infusion set and measurement pins (represented by 3 horizontal lines opposite arrows). Rate of change in distance A represented epiphysial growth rate. Rate of change in distance B was termed apparent metaphysial growth rate.
contralateral limb. The extremities were fixed to the cassette with tape during the X-ray. The X-ray source was carefully placed just above the tibia of interest at a distance of 91 cm. Each hindleg was X-rayed twice at each time point. The animal was then turned onto its other side to X-ray the contralateral leg. The radiographs were examined under a dissecting microscope, and the distances between the pins were measured using a micrometer. In some cases, contact prints of the radiographs were made to improve contrast. The most reliable measurements were obtained by measuring distances parallel to the long axis of the tibia between the blunt ends of the pins. Measurements were made by a single observer who was blinded as to which side received dexamethasone and which received vehicle. Statistical analysis. The measurements from the two replicate radiographs at each time point were averaged. The increase in distance between the proximal two pins represented the cumulative epiphysial growth. The increase in distance between two radiographic measurements, divided by the intervening time interval, represented the epiphysial growth rate (Fig. 1). The distance between the two metaphysial pins was expected to remain constant with time, because the pins did not span a growth plate. The change in this distance divided by time, which we termed the apparent metaphysial growth rate, was therefore expected to approximate zero. Any deviation of this rate from zero was interpreted as measurement error. To evaluate the magnitude of the measurement error, we calculated the SD of the apparent metaphysial growth rates, measured in all five rabbits, in both tibiae, throughout the 3-wk period. To evaluate the effect of dexamethasone, the cumulative epiphysial growth was calculated for each tibia and each time point. To test statistical significance, we performed a multivariate analysis of variance (ANOVA) (22), in which the side of the animal (and hence treatment) was considered a within-subject variable. We then performed post hoc two-tailed paired t tests using the Bonferroni adjustment to determine at which time points the difference between the cumulative growth on the vehicle-treated side and the cumulative growth on the dexamethasone-treated side differed significantly from zero. Similarly, the epiphysial growth rate was calculated for each tibia and each time interval, and we performed a multivariate ANOVA (22) in which the side of the animal (and hence treatment) was considered a within-subject variable. We then performed post hoc two-tailed paired t tests using the Bonferroni adiustment to determine at which time intervals the ratio of the
Growth rate measurement error. The mean apparent metaphysial growth rate, which was expected to approximate zero, was 0.003 mm/day for all 10 tibiae measured serially over the 3-wk period (Fig. 2). The SD of this rate, which represented the measurement error, was 0.02 mm/ day. The mean epiphysial growth rate of the vehicletreated growth plates was 0.40 mm/day (Fig. 2). Thus the measurement error expressed as a fraction of the epiphysial growth rate in the baseline state was on the order of 0.02/0.40 or 5%. Effect of dexamethasone. The local effect of dexamethasone on growth was examined bY comparing the cumulative eP iphysial growth rate of the dexamethasonetreated growth plates to the contralateral vehicle-treated growth plates (Fig. 2). We observed a significant effect of treatment (P c 0.0001) as well as a significant timetreatment interaction (P < 0.0001). The suppression of cumulative growth on the dexamethasone-treated side achieved statistical significance at 8 days (P ~0.02) and remained statistically significant at all later time points. The corresponding epiphysial growth rates also demonstrated the growth-suppressing effect of dexamethasone (Fig. 3). We observed a significant effect of treatment (P < 0.0001) as well as a significant time-treatment interaction (P < 0.05). During the first 5 days of treatment, the growth rate of the dexamethasone-treated growth plate averaged 51% of the growth rate of the contralateral vehicle-treated growth plate (P ~0.03). From days 5 to 8, the growth rate of the dexamethasone-treated side reached its nadir of 23% (P < 0.03). By day 19, the
.
A
0
5
10
15
20
25
Day Fig. 2. Cumulative proximal tibia1 growth during and after local dexamethasone infusion. Dashed line, mean t SE epiphysial growth of dexamethasone-treated tibiae of 5 rabbits. Solid line, mean epiphysial growth of contralateral vehicle-treated tibiae. Dotted line, mean change in distance between lower 2 measurement pins (apparent metaphysial growth), an index of measurement error, in both tibiae of 5 rabbits. Error bars on dotted line were omitted because they were smaller than symbol. Bar, period of dexamethasone infusion. * P < 0.02, ** P < 0.005, between dexamethasone-treated and contralateral vehicle-treated
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.092.098.118) on August 14, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
GLUCOCORTICOIDS
AND
LQ,
;E is G ;z
40-
1
20-
0(3
b
Dexamethasone
0 0
5
I
1
I
10
15
20
Time
(days)
1
25
Fig. 3. Mean t SE proximal tibia1 growth rate during and after local dexamethasone infusion. Epiphysial growth rate of dexamethasonetreated proximal tibiae was expressed as percent of growth rate of contralateral vehicle-treated tibiae. Growth rate before surgical procedure (time 0) was not measured but was assumed to be equal in 2 limbs. * P < 0.03 between dexamethasone-treated and contralateral vehicletreated epiphysial growth rate.
growth rate of the dexamethasone-treated epiphysis had recovered completely to 101% of that of the control side. DISCUSSION
We have developed both a new method to infuse an agent locally into the epiphysial growth plate of the rabbit and an improved method to measure epiphysial growth rate accurately over short periods of time. With these methods we tested the hypothesis that dexamethasone can inhibit epiphysial growth via local direct actions, without the obligatory involvement of indirect systemic effects. Consistent with this hypothesis, we observed a marked (77%) suppression of the epiphysial growth rate of the dexamethasone-treated epiphysis compared with the contralateral vehicle-treated epiphysis. An indirect effect of glucocorticoid, mediated by other systemic signals such as growth hormone, circulating IGF-I, or IGF inhibitors could not account for this effect, since the contralateral growth plate would also have been affected. Several mechanisms have previously been suggested to explain the growth inhibitory effect of glucocorticoid. In humans, high doses of glucocorticoid decrease growth hormone secretion (2). However, this effect is more readily demonstrated in adults than in children (2 1). Strickland et al. (25) studied four patients with Cushing’s disease, using insulin and arginine provocative tests, and found that attenuation of growth hormone secretion was a late and less frequent occurrence when compared with growth retardation. However, with frequent sampling, diminished spontaneous growth hormone secretion has been demonstrated in children with Cushing’s disease (5), again suggesting the possibility that decreased growth hormone could be responsible for the growth inhibition. Alterations in circulating IGF-I activity have also been suggested to play a role. In rats, glucocorticoid decreases IGF-I mRNA in the liver (17). Circulating IGF-I levels are maintained in patients with Cushing’s syndrome, (6)
BONE
GROWTH
E491
but somatomedin activity, as measured by embryonic chick cartilage bioassay, falls (6). This effect may be explained by an increase in circulating IGF inhibitors (27). Our data do not exclude the possibility that alterations in circulating growth hormone, IGF-I , or IGF inhibitors or other indirect effects play a role in the growth inhibition caused by glucocorticoid. However, our data do suggest that the growth effects of excess glucocorticoid can be explained without invoking these indirect mechanisms. The local effects of glucocorticoid demonstrated in these experiments may have been due to a direct effect of the dexamethasone on the growth plate chondrocytes. High-affinity binding sites for glucocorticoid have been demonstrated in epiphysial cartilage (13)) and glucocorticoids have also been shown to act on growth plate chondrocytes in vitro. Glucocorticoids, in near-physiological concentrations, help maintain the chondrocytic phenotype and actually stimulate proteoglycan synthesis (9, 12, 14, 26). However, higher doses of glucocorticoid cause decreased synthesis of proteoglycans relative to the optimal concentration (9, 14, 26). Physiological concentrations of glucocorticoid have been reported both to increase DNA synthesis (26) and to decrease DNA synthesis (9, 12). Supraphysiological concentrations inhibit DNA synthesis in growth plate chondrocytes relative to the optimal concentration (9, 12, 26). We do not know the actual concentration of dexamethasone achieved within the growth plate during our infusion. We wished to achieve a clearly supraphysiological glucocorticoid concentration to cause growth suppression. We chose a perfusate dexamethasone concentration of 80 pg/ml, which is - 2O,OOO-fold above physiological (18), because previous studies in soft tissues (16, 23) suggest that the concentration of an agent falls off rapidly with distance from the needle tip. Based on these previous studies, we estimated that the concentration of dexamethasone within most of the growth plate would be loto loo-fold less than that of the perfusate, or -2OO- to 2,OOO-fold above physiological levels. There are several mechanisms by which glucocorticoid might act on the chondrocyte to decrease growth. Altered transcription of growth-regulating intracellular factors is possible. Alternatively, glucocorticoid might alter expression of autocrine or paracrine growth factors or their receptors. In particular, dexamethasone might act by altering local expression of IGF-I in the growth plate. IGF-I is produced in proliferative chondrocytes in the growth plate (20). This expression is positively regulated by growth hormone (11, 20). Luo and Murphy (17) have shown that systemic administration of dexamethasone decreases the growth hormone-responsive IGF-I mRNA in the proximal tibiae of hypophysectomized rats. Because IGF-I increases epiphysial growth (lo), the growthsuppressing effect of glucocorticoid might be mediated by decreased IGF-I expression within the growth plate. Although to our knowledge the infusion method that we have developed has not been employed previously, a related approach was used by Isgaard et al. (10) who gave daily local injections of growth hormone or IGF-I to hypophysectomized rats. Injection of growth hormone into the proximal tibia1 growth plate, into the articular space
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.092.098.118) on August 14, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
E492
GLUCOCORTICOIDS
of the knee joint, or through a cannula implanted in the bony epiphysis all increased growth rates. Our infusion method avoided the trauma to the growth plate of repeated injections. Also, the method replaced daily bolus injections with a continuous infusion. We also used the rabbit rather than the rat in our studies. We selected the rabbit because it appears to be a closer model for human growth than the rat. Unlike the rat, the rabbit growth plate fuses at the time of puberty, probably in response to sex steroids (4). The method for growth rate measurement developed for these studies appears to have advantages over previously described methods (1, 24, 28). By inserting fine reference points (3O-gauge needles) in the bone immediately adjacent to the growth plate, we were able to make precise measurements with a micrometer. As a result, growth rates over intervals as short as 3 days were measured with an error of 45%. An alternative method for assessing short-term longitudinal bone growth in small laboratory animals involves administration of a tetracycline to label newly formed bone (7, 8, 10). Compared with the tetracycline-labeling method, our method has the advantages of taking less time and of allowing serial measurements in the same animal. In summary, we have shown that dexamethasone, a glucocorticoid, acts locally on the epiphysial growth plate to suppress growth. Additionally, the methods developed in this study provide an approach that could be used to explore the local actions of other growth-altering hormones and paracrine growth factors in epiphysial growth. J. Baron, K. E. Oerter, J. D. Bather, and G. B. Cutler, Jr., are Commissioned Officers in the United States Public Health Service. Present address of Z. Huang: Dept. of Physiology, Univ. of Maryland School of Medicine, Baltimore, MD 2 1201. Address for reprint requests: J. Baron, Bldg. 10, Rm. lON262, National Institutes of Health, Bethesda, MD 20892. Received
30 January
1992; accepted
in final
form
10 April
1992.
REFERENCES 1. Bowen, C. V. A., B. M. O’Brien, and G. J. Gumley. Experimental microvascular growth plate transfers. J. Bone Jt. Surg. Br. Vol. 70B: 311-314, 1988. 2. Frantz, A. G., and M. T. Rabkin. Human growth hormone: clinical measurement, response to hypoglycemia and suppression by corticosteroids. N. Engl. J. Med. 271: 1375-1381, 1964. 3. Freshney, R. I. Culture of Animal Cells: A Manual of Basic Technique. New York: Liss, 1983, p. 70. 4. Gilsanz, V., T. F. Roe, D. T. Gibbens, E. E. Schulz, M. E. Carlson, 0. Gonzalez, and M. I. Boechat. Effect of sex steroids on peak bone density of growing rabbits. Am. J. Physiol. 255 (Endocrinol. Metab. 18): E416-E421, 1988. 5. Gomez, M. T., and G. P. Chrousos. GH secretion in endogenous Cushing’s syndrome before and following correction of hypercortisolism (Abstract). Pediatr. Res. 27: 76A, 1990. 6. Gourmelen, M., F. Girard, and M. Binoux. Serum somatomedin/insulin-like growth factor (IGF) and IGF carrier levels in patients with Cushing’s syndrome or receiving glucocorticoid therapy. J. Clin. Endocrinol. Metab. 54: 885-892, 1982. 7. Hansson, L. I. Daily growth in length of diaphysis measured by oxytetracycline in rabbit normally and after medullary plugging.
AND
BONE
GROWTH
Acta Orthop. Stand. Suppl. 101: 34-84, 1967. 8. Hansson, L. I., K. Menander-Sellman, A. Stenstrom, and K.-G. Thorngren. Rate of normal longitudinal bone growth in the rat. Calcif. Tissue Res. 10: 238-251, 1972. 9. Hill, D. J. Effects of cortisol on cell proliferation and proteoglycan synthesis and degradation in cartilage zones of the calf costochondral growth plate in vitro with and without rat plasma somatomedin activity. J. Endocrinol. 88: 425-435, 1981. 10. Isgaard, J., A. Nilsson, A. Lindahl, J. 0. Jansson, and 0. G. P. Isaksson. Effects of local administration of GH and IGF-I on longitudinal bone growth in rats. Am. J. Physiol. 250 (Endocrinol. Metab. 13): E367-E372, 1986. 11. Isgaard, J., C. Moller, 0. G. P. Isaksson, A. Nilsson, L. S. Mathews, and G. Norstedt. Regulation of insulin-like growth factor messenger ribonucleic acid in rat growth plate by growth hormone. Endocrinology 122: 1515-1520, 1988. 12. Itagane, Y., H. Inada, K. Fujita, and G. Isshiki. Interactions between steroid hormones and insulin-like growth factor-I in rabbit chondrocytes. Endocrinology 128: 1419-1424, 1991. 13. Kan, K. W., R. L. Cruess, B. I. Posner, H. J. Guyda, and S. Soloman. Hormone receptors in the epiphysial cartilage. J. Endocrinol. 103: 125-131, 1984. 14. Kato, Y., and D. Gospodarowicz. Stimulation by glucocorticoid of the synthesis of cartilage-matrix proteoglycans produced by rabbit costal chondrocytes in vitro. J. Biol. Chem. 260: 23642373, 1985. 15. Loeb, J. N. Corticosteroids and growth. N. Engl. J. Med. 295: 547-552, 1976. 16. Lum, J. T., T. Nguyen, and L. P. Felpel. Drug distribution in solid tissue of the brain following chronic local perfusion utilizing implanted osmotic minipumps. J. Pharmacol. Methods 12: 141147, 1984. 17. Luo, J., and L. J. Murphy. Dexamethasone inhibits growth hormone induction of insulin-like growth factor I (IGF-I) messenger ribonucleic acid (mRNA) in hypophysectomized rats and reduces IGF-I mRNA abundance in the intact rat. Endocrinology 125: 165-171, 1989. 18. Meikle, A. W., and F. H. Tyler. Potency and duration of action of glucocorticoids. Am. J. Med. 63: 200-207, 1977. 19. Mosier, H. D., and R. A. Jansons. Rats stunted by high-dose glucocorticoid treatment are capable of undergoing catch-up growth after fasting. Pediatr. Res. 25: 373-376, 1989. 20. Nilsson, A., J. Isgaard, A. Lindahl, A. Dahlstrom, A. Skottner, and 0. G. P. Isaksson. Regulation by growth hormone of number of chondrocytes containing IGF-I in rat growth plate. Science Wash. DC 233: 571-574, 1986. 21. Phillips, L. S., and T. G. Unterman. Somatomedin activity in disorders of nutrition and metabolism. Clin. Endocrinol. Metab. 13: 145-189, 1984. 22. SAS Institute Inc. SAS User’s Guide: Statistics (5th ed.). Cary, NC: SAS Institute, 1985. 23. Sendelbeck, S. L., and J. Urquhart. Spatial distribution of dopamine, methotrexate and antipyrine during continuous intracerebral microperfusion. Brain Res. 328: 251-258, 1985. 24. Sissons, H. A. Experimental determination of rate of longitudinal bone growth. J. Anat. 87: 228-236, 1953. 25. Strickland, A. L., L. E. Underwood, S. J. Voina, F. S. French, and J. J. Van Wyk. Growth retardation in Cushing’s syndrome. Am. J. Dis. Child. 123: 207-213, 1972. 26. Takano, T., M. Takigawa, and F. Suzuki. Stimulation by glucocorticoids of the differentiated phenotype of chondrocytes and the proliferation of rabbit costal chondrocytes in culture. J. Biochem. 97: 1093-1100, 1985. 27. Unterman, T. G., and L. S. Phillips. Glucocorticoid effects on somatomedins and somatomedin inhibitors. J. Clin. Endocrinol. Metab. 61: 618-626, 1985. 28. Wolohan, M. J., and D. J. Zaleske. Hemiepiphyseal reconstruction using tissue donated from fetal limbs in a murine model. J. Orthop. Res. 9: 180-185, 1991.
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.092.098.118) on August 14, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
to inhibit
longitudinal
ZE HUANG, KAREN E. OERTER, AND GORDON B. CUTLER, JR.
Developmental Endocrinology Branch, National Institute of Child Health and Human Development, and Surgery, Radiology, and Pharmacy Section, Veterinary Resources Program, National Center for Research Resources, National Institutes of Health, Bethesda, Maryland 20892 Baron, Jeffrey, Ze Huang, Karen E. Oerter, John D. Bather, and Gordon B. Cutler, Jr. Dexamethasone acts locally to inhibit longitudinal bone growth in rabbits. Am. J. Physiol. 263 (Endocrinol. Metab. 26): E489-E492, 1992.Excess glucocorticoid is a potent inhibitor of epiphysial growth. Several mechanisms have been suggested to explain this growth inhibition, including both direct local effects of glucocorticoid on the epiphysial growth plate and indirect systemic effects. Previous studies do not distinguish which of these proposed mechanisms is actually responsible for the growth suppression in vivo. To resolve this controversy, we developed a method for delivering glucocorticoid directly into the rabbit epiphysial growth plate and for accurately measuring the resulting epiphysial growth rate. Five-week-old male rabbits received a local infusion of dexamethasone phosphate (80 ng/pl, 1 pi/h) into one proximal tibia1 growth plate and an infusion of vehicle into the contralateral growth plate. Growth rate was determined by inserting metal pins into the bone immediately adjacent to the growth plate and measuring the change in distance between pins on serial radiographs. This method permitted growth rates to be measured over intervals as short as 3 days, with an error of m 5%. Local dexamethasone administration decreased proximal tibia1 growth rate by 77% compared with the contralateral vehicle-treated tibia (P < 0.0001). We conclude that excess glucocorticoid causes a rapid potent inhibition of growth by a direct local action on the growth plate. glucocorticoids;
growth
plate; epiphysis
GLUCOCORTICOID IS A POTENT inhibitor ofepiphysial growth in both animals and humans (15, 19). Several mechanisms have been suggested to explain this growth inhibition, including both direct local effects of glucocorticoid on the epiphysial growth plate as well as indirect systemic effects. The proposed indirect systemic effects include decreased growth hormone secretion (2)) decreased circulating insulin-like growth factor I (IGF-I) activity (6), and increased circulating IGF inhibitors (27). On the other hand, a local direct effect on epiphysial growth is suggested by in vitro studies in which supraphysiological concentrations of glucocorticoid decreased DNA and proteoglycan synthesis by growth plate chondrocytes (9, 12, 14, 26). Previous studies do not distinguish which of these proposed mechanisms are actually responsible for the growth suppression in vivo. To resolve this controversy, we developed a method to deliver glucocorticoid directly into the rabbit epiphysial growth plate and to measure accurately the resulting epiphysial growth rate. EXCESS
METHODS Growth plate infusion. Five 4-wk-old male New Zealand White rabbits were purchased from Hazleton Research Products (Denver, PA). The animals received National Institutes of Health Open Formula Rabbit Ration (NIH32) and water ad libitum. At 5 wk of age, they were given a single prophylactic
intramuscular dose of penicillin G (44,000 IU/kg) and premedicated with intramuscular glycopyrrolate (0.02 mg), ketamine hydrochloride (33 mg/kg), and xylazine (6.6 mg/kg). They were then intubated and placed under general anesthesia with isoflurane. An incision was made to expose the proximal anteromedial tibia. Sterile infusion sets, which consisted of a 27-gauge stainless steel needle attached to a flexible plastic catheter (Baxter Healthcare, Hookset, NH), had been prepared beforehand for implantation. The needle had been bent to place a right angle 2.5 mm from the sharp tip. This tip was then inserted into the proximal tibia1 growth plate close to the junction with the bony epiphysis (Fig. 1). The shaft of the needle was bent into a “U” shape, the distal portion of which was sutured to the periosteum of the tibia. The plastic catheter was then tunneled subcutaneously to a second incision over the lower abdomen on the ipsilateral side. The free end of the catheter was attached to an osmotic pump (model 2001; Alza, Palo Alto, CA) that was placed in the subcutaneous space. The pump was designed to administer 1 pi/h for 7 days. To determine whether the osmotic pump was capable of delivering a solution into the growth plate, we had previously infused methylene blue dye using this system and found that the dye stained the entire growth plate. In other experiments, we had measured the residual fluid remaining in the pump at the end of 6 days and demonstrated that the pump rate in vivo did match the manufacturer’s specification (data not shown). An osmotic pump and infusion set were placed on each of the two sides of the animal. One osmotic pump contained a solution of dexamethasone phosphate. This solution had been prepared by diluting dexamethasone sodium phosphate (equivalent to 4 mg/ml dexamethasone phosphate in water containing 1 mg/ml anhydrous sodium sulfite, 19.4 mg/ml anhydrous sodium citrate, and 0.01 ml/ml benzyl alcohol, pH 7.0-8.5; Elkins-Sinn, Cherry Hill, NJ) in Hanks’ balanced salt solution (HBSS) (3) without phenol red to yield a final concentration of 80 ng/pl dexamethasone phosphate. The contralateral pump contained HBSS, to which had been added sodium sulfite, sodium citrate, and benzyl alcohol, such that the concentrations of these substances were equal in the two pumps. Growth rate measurement. Three metal pins were also placed in each proximal tibia. The pins were prepared beforehand by breaking 30-gauge hypodermic needles (Becton Dickinson, Rutherford, NJ) 5 mm from the tip. The 5-mm fragments were then subjected to steam sterilization. All three pins were inserted parallel to the growth plate, directed posteriorly (Fig. 1). One pin was placed in the bony metaphysis -2-3 mm distal to the growth plate and just medial to the anterior tibia1 tuberosity. A second pin was placed -5 mm distal to the first pin and also directed posteriorly. The final pin was placed -5 mm proximal to the first pin. This last pin was thus 2-3 mm proximal to the growth plate and within the bony epiphysis. Serial radiographs of the tibiae were obtained on postoperative days 0, 5, 8, 12, 15, 19, and 22. For the radiograph, the animal was sedated with injection of ketamine hydrochloride (33 mg/kg im) plus xylazine (6.6 mg/kg). The rabbit was then placed in the lateral position on an X-ray cassette. The hindleg of interest was placed closer to the cassette and anterior to the E489
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.092.098.118) on August 14, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
E490
GLUCOCORTICOIDS Proximal
AND
BONE
GROWTH
growth rate on the dexamethasone-treated side to the growth rate on the vehicle-treated side differed significantly from one. RESULTS
Dlstal Fig. 1. Diagram of rabbit proximal tibia after surgical implantation of infusion set and measurement pins (represented by 3 horizontal lines opposite arrows). Rate of change in distance A represented epiphysial growth rate. Rate of change in distance B was termed apparent metaphysial growth rate.
contralateral limb. The extremities were fixed to the cassette with tape during the X-ray. The X-ray source was carefully placed just above the tibia of interest at a distance of 91 cm. Each hindleg was X-rayed twice at each time point. The animal was then turned onto its other side to X-ray the contralateral leg. The radiographs were examined under a dissecting microscope, and the distances between the pins were measured using a micrometer. In some cases, contact prints of the radiographs were made to improve contrast. The most reliable measurements were obtained by measuring distances parallel to the long axis of the tibia between the blunt ends of the pins. Measurements were made by a single observer who was blinded as to which side received dexamethasone and which received vehicle. Statistical analysis. The measurements from the two replicate radiographs at each time point were averaged. The increase in distance between the proximal two pins represented the cumulative epiphysial growth. The increase in distance between two radiographic measurements, divided by the intervening time interval, represented the epiphysial growth rate (Fig. 1). The distance between the two metaphysial pins was expected to remain constant with time, because the pins did not span a growth plate. The change in this distance divided by time, which we termed the apparent metaphysial growth rate, was therefore expected to approximate zero. Any deviation of this rate from zero was interpreted as measurement error. To evaluate the magnitude of the measurement error, we calculated the SD of the apparent metaphysial growth rates, measured in all five rabbits, in both tibiae, throughout the 3-wk period. To evaluate the effect of dexamethasone, the cumulative epiphysial growth was calculated for each tibia and each time point. To test statistical significance, we performed a multivariate analysis of variance (ANOVA) (22), in which the side of the animal (and hence treatment) was considered a within-subject variable. We then performed post hoc two-tailed paired t tests using the Bonferroni adjustment to determine at which time points the difference between the cumulative growth on the vehicle-treated side and the cumulative growth on the dexamethasone-treated side differed significantly from zero. Similarly, the epiphysial growth rate was calculated for each tibia and each time interval, and we performed a multivariate ANOVA (22) in which the side of the animal (and hence treatment) was considered a within-subject variable. We then performed post hoc two-tailed paired t tests using the Bonferroni adiustment to determine at which time intervals the ratio of the
Growth rate measurement error. The mean apparent metaphysial growth rate, which was expected to approximate zero, was 0.003 mm/day for all 10 tibiae measured serially over the 3-wk period (Fig. 2). The SD of this rate, which represented the measurement error, was 0.02 mm/ day. The mean epiphysial growth rate of the vehicletreated growth plates was 0.40 mm/day (Fig. 2). Thus the measurement error expressed as a fraction of the epiphysial growth rate in the baseline state was on the order of 0.02/0.40 or 5%. Effect of dexamethasone. The local effect of dexamethasone on growth was examined bY comparing the cumulative eP iphysial growth rate of the dexamethasonetreated growth plates to the contralateral vehicle-treated growth plates (Fig. 2). We observed a significant effect of treatment (P c 0.0001) as well as a significant timetreatment interaction (P < 0.0001). The suppression of cumulative growth on the dexamethasone-treated side achieved statistical significance at 8 days (P ~0.02) and remained statistically significant at all later time points. The corresponding epiphysial growth rates also demonstrated the growth-suppressing effect of dexamethasone (Fig. 3). We observed a significant effect of treatment (P < 0.0001) as well as a significant time-treatment interaction (P < 0.05). During the first 5 days of treatment, the growth rate of the dexamethasone-treated growth plate averaged 51% of the growth rate of the contralateral vehicle-treated growth plate (P ~0.03). From days 5 to 8, the growth rate of the dexamethasone-treated side reached its nadir of 23% (P < 0.03). By day 19, the
.
A
0
5
10
15
20
25
Day Fig. 2. Cumulative proximal tibia1 growth during and after local dexamethasone infusion. Dashed line, mean t SE epiphysial growth of dexamethasone-treated tibiae of 5 rabbits. Solid line, mean epiphysial growth of contralateral vehicle-treated tibiae. Dotted line, mean change in distance between lower 2 measurement pins (apparent metaphysial growth), an index of measurement error, in both tibiae of 5 rabbits. Error bars on dotted line were omitted because they were smaller than symbol. Bar, period of dexamethasone infusion. * P < 0.02, ** P < 0.005, between dexamethasone-treated and contralateral vehicle-treated
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.092.098.118) on August 14, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
GLUCOCORTICOIDS
AND
LQ,
;E is G ;z
40-
1
20-
0(3
b
Dexamethasone
0 0
5
I
1
I
10
15
20
Time
(days)
1
25
Fig. 3. Mean t SE proximal tibia1 growth rate during and after local dexamethasone infusion. Epiphysial growth rate of dexamethasonetreated proximal tibiae was expressed as percent of growth rate of contralateral vehicle-treated tibiae. Growth rate before surgical procedure (time 0) was not measured but was assumed to be equal in 2 limbs. * P < 0.03 between dexamethasone-treated and contralateral vehicletreated epiphysial growth rate.
growth rate of the dexamethasone-treated epiphysis had recovered completely to 101% of that of the control side. DISCUSSION
We have developed both a new method to infuse an agent locally into the epiphysial growth plate of the rabbit and an improved method to measure epiphysial growth rate accurately over short periods of time. With these methods we tested the hypothesis that dexamethasone can inhibit epiphysial growth via local direct actions, without the obligatory involvement of indirect systemic effects. Consistent with this hypothesis, we observed a marked (77%) suppression of the epiphysial growth rate of the dexamethasone-treated epiphysis compared with the contralateral vehicle-treated epiphysis. An indirect effect of glucocorticoid, mediated by other systemic signals such as growth hormone, circulating IGF-I, or IGF inhibitors could not account for this effect, since the contralateral growth plate would also have been affected. Several mechanisms have previously been suggested to explain the growth inhibitory effect of glucocorticoid. In humans, high doses of glucocorticoid decrease growth hormone secretion (2). However, this effect is more readily demonstrated in adults than in children (2 1). Strickland et al. (25) studied four patients with Cushing’s disease, using insulin and arginine provocative tests, and found that attenuation of growth hormone secretion was a late and less frequent occurrence when compared with growth retardation. However, with frequent sampling, diminished spontaneous growth hormone secretion has been demonstrated in children with Cushing’s disease (5), again suggesting the possibility that decreased growth hormone could be responsible for the growth inhibition. Alterations in circulating IGF-I activity have also been suggested to play a role. In rats, glucocorticoid decreases IGF-I mRNA in the liver (17). Circulating IGF-I levels are maintained in patients with Cushing’s syndrome, (6)
BONE
GROWTH
E491
but somatomedin activity, as measured by embryonic chick cartilage bioassay, falls (6). This effect may be explained by an increase in circulating IGF inhibitors (27). Our data do not exclude the possibility that alterations in circulating growth hormone, IGF-I , or IGF inhibitors or other indirect effects play a role in the growth inhibition caused by glucocorticoid. However, our data do suggest that the growth effects of excess glucocorticoid can be explained without invoking these indirect mechanisms. The local effects of glucocorticoid demonstrated in these experiments may have been due to a direct effect of the dexamethasone on the growth plate chondrocytes. High-affinity binding sites for glucocorticoid have been demonstrated in epiphysial cartilage (13)) and glucocorticoids have also been shown to act on growth plate chondrocytes in vitro. Glucocorticoids, in near-physiological concentrations, help maintain the chondrocytic phenotype and actually stimulate proteoglycan synthesis (9, 12, 14, 26). However, higher doses of glucocorticoid cause decreased synthesis of proteoglycans relative to the optimal concentration (9, 14, 26). Physiological concentrations of glucocorticoid have been reported both to increase DNA synthesis (26) and to decrease DNA synthesis (9, 12). Supraphysiological concentrations inhibit DNA synthesis in growth plate chondrocytes relative to the optimal concentration (9, 12, 26). We do not know the actual concentration of dexamethasone achieved within the growth plate during our infusion. We wished to achieve a clearly supraphysiological glucocorticoid concentration to cause growth suppression. We chose a perfusate dexamethasone concentration of 80 pg/ml, which is - 2O,OOO-fold above physiological (18), because previous studies in soft tissues (16, 23) suggest that the concentration of an agent falls off rapidly with distance from the needle tip. Based on these previous studies, we estimated that the concentration of dexamethasone within most of the growth plate would be loto loo-fold less than that of the perfusate, or -2OO- to 2,OOO-fold above physiological levels. There are several mechanisms by which glucocorticoid might act on the chondrocyte to decrease growth. Altered transcription of growth-regulating intracellular factors is possible. Alternatively, glucocorticoid might alter expression of autocrine or paracrine growth factors or their receptors. In particular, dexamethasone might act by altering local expression of IGF-I in the growth plate. IGF-I is produced in proliferative chondrocytes in the growth plate (20). This expression is positively regulated by growth hormone (11, 20). Luo and Murphy (17) have shown that systemic administration of dexamethasone decreases the growth hormone-responsive IGF-I mRNA in the proximal tibiae of hypophysectomized rats. Because IGF-I increases epiphysial growth (lo), the growthsuppressing effect of glucocorticoid might be mediated by decreased IGF-I expression within the growth plate. Although to our knowledge the infusion method that we have developed has not been employed previously, a related approach was used by Isgaard et al. (10) who gave daily local injections of growth hormone or IGF-I to hypophysectomized rats. Injection of growth hormone into the proximal tibia1 growth plate, into the articular space
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.092.098.118) on August 14, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
E492
GLUCOCORTICOIDS
of the knee joint, or through a cannula implanted in the bony epiphysis all increased growth rates. Our infusion method avoided the trauma to the growth plate of repeated injections. Also, the method replaced daily bolus injections with a continuous infusion. We also used the rabbit rather than the rat in our studies. We selected the rabbit because it appears to be a closer model for human growth than the rat. Unlike the rat, the rabbit growth plate fuses at the time of puberty, probably in response to sex steroids (4). The method for growth rate measurement developed for these studies appears to have advantages over previously described methods (1, 24, 28). By inserting fine reference points (3O-gauge needles) in the bone immediately adjacent to the growth plate, we were able to make precise measurements with a micrometer. As a result, growth rates over intervals as short as 3 days were measured with an error of 45%. An alternative method for assessing short-term longitudinal bone growth in small laboratory animals involves administration of a tetracycline to label newly formed bone (7, 8, 10). Compared with the tetracycline-labeling method, our method has the advantages of taking less time and of allowing serial measurements in the same animal. In summary, we have shown that dexamethasone, a glucocorticoid, acts locally on the epiphysial growth plate to suppress growth. Additionally, the methods developed in this study provide an approach that could be used to explore the local actions of other growth-altering hormones and paracrine growth factors in epiphysial growth. J. Baron, K. E. Oerter, J. D. Bather, and G. B. Cutler, Jr., are Commissioned Officers in the United States Public Health Service. Present address of Z. Huang: Dept. of Physiology, Univ. of Maryland School of Medicine, Baltimore, MD 2 1201. Address for reprint requests: J. Baron, Bldg. 10, Rm. lON262, National Institutes of Health, Bethesda, MD 20892. Received
30 January
1992; accepted
in final
form
10 April
1992.
REFERENCES 1. Bowen, C. V. A., B. M. O’Brien, and G. J. Gumley. Experimental microvascular growth plate transfers. J. Bone Jt. Surg. Br. Vol. 70B: 311-314, 1988. 2. Frantz, A. G., and M. T. Rabkin. Human growth hormone: clinical measurement, response to hypoglycemia and suppression by corticosteroids. N. Engl. J. Med. 271: 1375-1381, 1964. 3. Freshney, R. I. Culture of Animal Cells: A Manual of Basic Technique. New York: Liss, 1983, p. 70. 4. Gilsanz, V., T. F. Roe, D. T. Gibbens, E. E. Schulz, M. E. Carlson, 0. Gonzalez, and M. I. Boechat. Effect of sex steroids on peak bone density of growing rabbits. Am. J. Physiol. 255 (Endocrinol. Metab. 18): E416-E421, 1988. 5. Gomez, M. T., and G. P. Chrousos. GH secretion in endogenous Cushing’s syndrome before and following correction of hypercortisolism (Abstract). Pediatr. Res. 27: 76A, 1990. 6. Gourmelen, M., F. Girard, and M. Binoux. Serum somatomedin/insulin-like growth factor (IGF) and IGF carrier levels in patients with Cushing’s syndrome or receiving glucocorticoid therapy. J. Clin. Endocrinol. Metab. 54: 885-892, 1982. 7. Hansson, L. I. Daily growth in length of diaphysis measured by oxytetracycline in rabbit normally and after medullary plugging.
AND
BONE
GROWTH
Acta Orthop. Stand. Suppl. 101: 34-84, 1967. 8. Hansson, L. I., K. Menander-Sellman, A. Stenstrom, and K.-G. Thorngren. Rate of normal longitudinal bone growth in the rat. Calcif. Tissue Res. 10: 238-251, 1972. 9. Hill, D. J. Effects of cortisol on cell proliferation and proteoglycan synthesis and degradation in cartilage zones of the calf costochondral growth plate in vitro with and without rat plasma somatomedin activity. J. Endocrinol. 88: 425-435, 1981. 10. Isgaard, J., A. Nilsson, A. Lindahl, J. 0. Jansson, and 0. G. P. Isaksson. Effects of local administration of GH and IGF-I on longitudinal bone growth in rats. Am. J. Physiol. 250 (Endocrinol. Metab. 13): E367-E372, 1986. 11. Isgaard, J., C. Moller, 0. G. P. Isaksson, A. Nilsson, L. S. Mathews, and G. Norstedt. Regulation of insulin-like growth factor messenger ribonucleic acid in rat growth plate by growth hormone. Endocrinology 122: 1515-1520, 1988. 12. Itagane, Y., H. Inada, K. Fujita, and G. Isshiki. Interactions between steroid hormones and insulin-like growth factor-I in rabbit chondrocytes. Endocrinology 128: 1419-1424, 1991. 13. Kan, K. W., R. L. Cruess, B. I. Posner, H. J. Guyda, and S. Soloman. Hormone receptors in the epiphysial cartilage. J. Endocrinol. 103: 125-131, 1984. 14. Kato, Y., and D. Gospodarowicz. Stimulation by glucocorticoid of the synthesis of cartilage-matrix proteoglycans produced by rabbit costal chondrocytes in vitro. J. Biol. Chem. 260: 23642373, 1985. 15. Loeb, J. N. Corticosteroids and growth. N. Engl. J. Med. 295: 547-552, 1976. 16. Lum, J. T., T. Nguyen, and L. P. Felpel. Drug distribution in solid tissue of the brain following chronic local perfusion utilizing implanted osmotic minipumps. J. Pharmacol. Methods 12: 141147, 1984. 17. Luo, J., and L. J. Murphy. Dexamethasone inhibits growth hormone induction of insulin-like growth factor I (IGF-I) messenger ribonucleic acid (mRNA) in hypophysectomized rats and reduces IGF-I mRNA abundance in the intact rat. Endocrinology 125: 165-171, 1989. 18. Meikle, A. W., and F. H. Tyler. Potency and duration of action of glucocorticoids. Am. J. Med. 63: 200-207, 1977. 19. Mosier, H. D., and R. A. Jansons. Rats stunted by high-dose glucocorticoid treatment are capable of undergoing catch-up growth after fasting. Pediatr. Res. 25: 373-376, 1989. 20. Nilsson, A., J. Isgaard, A. Lindahl, A. Dahlstrom, A. Skottner, and 0. G. P. Isaksson. Regulation by growth hormone of number of chondrocytes containing IGF-I in rat growth plate. Science Wash. DC 233: 571-574, 1986. 21. Phillips, L. S., and T. G. Unterman. Somatomedin activity in disorders of nutrition and metabolism. Clin. Endocrinol. Metab. 13: 145-189, 1984. 22. SAS Institute Inc. SAS User’s Guide: Statistics (5th ed.). Cary, NC: SAS Institute, 1985. 23. Sendelbeck, S. L., and J. Urquhart. Spatial distribution of dopamine, methotrexate and antipyrine during continuous intracerebral microperfusion. Brain Res. 328: 251-258, 1985. 24. Sissons, H. A. Experimental determination of rate of longitudinal bone growth. J. Anat. 87: 228-236, 1953. 25. Strickland, A. L., L. E. Underwood, S. J. Voina, F. S. French, and J. J. Van Wyk. Growth retardation in Cushing’s syndrome. Am. J. Dis. Child. 123: 207-213, 1972. 26. Takano, T., M. Takigawa, and F. Suzuki. Stimulation by glucocorticoids of the differentiated phenotype of chondrocytes and the proliferation of rabbit costal chondrocytes in culture. J. Biochem. 97: 1093-1100, 1985. 27. Unterman, T. G., and L. S. Phillips. Glucocorticoid effects on somatomedins and somatomedin inhibitors. J. Clin. Endocrinol. Metab. 61: 618-626, 1985. 28. Wolohan, M. J., and D. J. Zaleske. Hemiepiphyseal reconstruction using tissue donated from fetal limbs in a murine model. J. Orthop. Res. 9: 180-185, 1991.
Downloaded from www.physiology.org/journal/ajpendo by ${individualUser.givenNames} ${individualUser.surname} (137.092.098.118) on August 14, 2018. Copyright © 1992 American Physiological Society. All rights reserved.
