- Email: [email protected]
Stimulation of sulfate and thymidine incorporation into hypophysectomized rat cartilage by growth hormone and insulin-like growth factor-I in vitro: The somatomedin hypothesis revisited
Stimulation of sulfate and thymidine incorporation into hypophysectomized rat cartilage by growth hormone and insulin-like growth factor-I in vitro: The somatomedin hypothesis revisited WILLIAM D. SALMON, JR., and V. JEAN BURKHALTER NASHVILLE,TENNESSEE
These studies were undertaken in an attempt to demonstrate sensitivity of cartilage from hypophysectomized rats to growth hormone (GH) in physiologic concentrations in vitro and to compare the actions of the hormone and extrinsic insulin-like growth factor-I (IGF-I). A HEPES-buffered, amino acid-glucose nutrient medium with a low concentration of bovine serum albumin, costal cartilage explants from such animals was responsive to purified or recombinant bovine GH (bGH) and recombinant human GH (hGH) with a sensitivity of less than 100 ng/ml. Sulfate incorporation into proteoglycans and thymidine incorporation into DNA were increased. The effect on sulfate incorporation was near maximal by 24 hours, but the peak effect on thymidine incorporation was delayed. Comparison of the dose-response relationships of bGH and IGF-I demonstrated a greater effect of IGF-I with increasing concentrations of each. The action of bGH was inhibited by an immunoglobulin G fraction of an IGF-I antiserum. These results are consistent with previously available evidence that IGF-I has an autocrine/paracrine action on cartilage, which is regulated by GH. However, the limited cartilage response to the direct action of GH supports the hypothesis of Daughaday that IGF-I also has an endocrine role in the stimulation of skeletal growth by the pituitary hormone. (J Lab Clin Med 1997;129: 430-8)
Abbreviations: b G H = bovine growth hormone; BSA = bovine serum albumin; D N A = deoxyribonucleic acid; G H = growth hormone; HEPES = N-2-hydroxyethypiperazine-N'-2-ethanesulfonic acid; h G H = h u m a n growth hormone; IGF = insulin-likegrowth factor; IgG = i m m u n o globulin G; thr-59 = threonine-59
he somatomedin hypothesis was based on observations in William Daughaday's laboratory nearly 40 years ago. 1 Hypophysectomy in rats had been shown to cause a reduction of sulfate incorporation into cartilage proteoglycans, and G H
T
From Medical and Research Services, Veterans Administration Medical Center, and the Department of Medicine, Vanderbilt University School of Medicine. Supported by the Department of Veterans Affairs. Submitted for publication Sept. 5, 1996; accepted Oct. 28, 1996. Reprint requests: William D. Salmon, Jr., MD, Medical Service, VA Medical Center, Nashville, TN 37212-2637. 0022-2143/97 $5.00 + 0 5/1/79944
430
treatment corrected the abnormality. Then it was found that G H in physiologic concentrations in vitro had no significant effect on sulfate incorporation into cartilage from hypophysectomized rats, but a cartilage-stimulating factor was detected in serum after injection of the h o r m o n e into such animals. Insulin had an action similar to the serum factor, but high concentrations were required. It was postulated that the action of G H on cartilage was mediated by the serum sulfation factor, redesignated somatomeAn editorial relevant to this article appears on p. 398 of this issue of the Journal.
J Lab Clin Med Volume 129, Number 4
din and now IGF-I 2 after the recognition of other effects including stimulation of thymidine incorporation into D N A of cartilage from hypophysectomized rats. 3 Challenges to the hypothesis resulted initially from in vivo experiments that demonstrated rat cartilage is more sensitive to G H than revealed by the original in vitro studies. 4-6 Additional evidence has indicated that the effects of G H are mediated partly by local production of IGF-I, 7 whereas the serum level of IGF-I is maintained primarily by hepatic production under G H regulation, s A modification of the somatomedin hypothesis has been necessary because of this new information, and doubt remains about the role of the circulating growth factor or the endocrine versus the autocrine/ paracrine function of IGF-I. Despite the tremendous expansion of knowledge in this area, the insensitivity of hypophysectomized rat cartilage to G H in vitro in previous studies has remained a curious anomaly. It has been suggested that unfavorable tissue incubation conditions may be the explanation. 9 Also, the possibility that inhibitors 1° may have been present in certain G H preparations or serum supplements is now appreciated. Therefore the direct action of G H has been reexamined with purified or recombinant preparations of the hormone in a modified incubation medium. In this system G H in physiologic concentrations increased sulfate and thymidine incorporation into costal cartilage from hypophysectomized rats, and those effects were inhibited by an IgG fraction of an IGF-I antiserum. Quantitative considerations of the direct effects of G H and extrinsic IGF-I support an endocrine and an autocrine/paracrine role for the growth factor as postulated originally by Daughaday. 5 METHODS Hormones, growth factors, chemicals. Purified bGH (from NIH-GH-B17) was provided by Drs. Leslie Holladay and David Puett (Department of Biochemistry, Vanderbilt University School of Medicine), recombinant bGH by Dr. Thomas Holzman (The Upjohn Co., Kalamazoo, Mich.), and recombinant hGH by Dr. Linda Fryklund (KabiVitrum Peptide Hormones AB, Stockholm, Sweden). Recombinant IGF-I (or thr-59 analog) was obtained from AmGen (Thousand Oaks, Calif.). Amino acids and HEPES buffer components were obtained from CalBiochem Corp. (San Diego, Calif.), and [35S]sulfuric acid and [methyl-3H]thymidine (2.0 Ci/mmol) were obtained from New England Nuclear (Boston, Mass.). A rat serum fraction with somatomedin (IGF) activity was prepared from serum of normal animals. 11 Bioactivity determined by the hypophysectomized rat cartilage assay for sulfation-promoting activity was 54 U/mg of protein (1 U = activity equivalent to 1 ml normal rat serum).
Salmon and Burkhalter
431
IGF-I immunoreactivity was 155 ng/U; IGF-II was undetectable. The IGF assays were provided by Drs. Judson Van Wyk and Marjorie Svoboda (Department of Pediatrics, University of North Carolina School of Medicine). Antibody preparation. A rabbit antiserum (UBK 487) directed against somatomedin-C (IGF-I) was supplied by Dr. Louis Underwood (Department of Pediatrics, University of North Carolina School of Medicine). An IgG fraction of this antiserum (0.25 ml) was prepared by (1) nonspecific adsorption on a column (0.7 x 2.5 cm) of Sepharose 4 Fast Flow (Pharmacia Biotech, Piscataway, N.J.) equilibrated with phosphate-buffered saline solution (0.13 mol/L NaC1-0.02 mol/L NazHPO4, pH 7.0), (2) specific adsorption of the effluent and washings (11 ml phosphate-buffered saline solution) on a column (0.7 × 2.5 cm) of Protein-G Sepharose 4 Fast Flow (Pharmacia Biotech), which was equilibrated also with phosphate-buffered saline solution, (3) washing with 10 ml phosphatebuffered saline solution, (4) elution with 10 ml of 0.1 mol/L glycine-HC1 buffer, pH 2.6, (5) neutralization and dilution of the eluate with HEPES buffers, and (6) dialysis in Spectrapor 2 tubing against a HEPES-buffered saline solution (see following text). The final concentration of this fraction in assay media was equivalent to a 1:200 dilution of the original antiserum. Incubation medium. The defined basal medium for cartilage incubation was constituted with a HEPES-buffered saline solution: NaC1 124, KC1 4.6, NaeHPO 4 1.1, CAC12-2 H20 1.25, MgSO 4 1.0, HEPES 11, and HEPES.Na 14 mmol/L. The other components were amino acids in the formulation of Eagle's basal medium for HeLa cells~eexcept for the addition of serine13: arginine 0.1, cystine 0.05, glutamine 2.0, histidine 0.05, isoleucine 0.2, leucine 0.2, lysine.HC1 0.2, methionine 0.05, phenylalanine 0.1, serine 0.1, threonine 0.2, tryptophane 0.02, tyrosine 0.1, and valine 0.2 mmol/L; glucose 2 mg/ml; penicillin 100 U/ml; gentamicin 50 ~g/ml; and crystalline BSA 0.2 mg/ml (ICN Pharmaceuticals, Inc., Cleveland Ohio, or CalBiochem Corp., the latter preparation designated nuclease- and protease-free). Although crystalline BSA from other commercial sources was tested and found suitable, one preparation was inhibitory, emphasizing the importance of a preliminary test of individual lots. The final pH was adjusted to 7.45 at 37° C with 1 N NaOH or HC1. Cartilage incubation and analysis. This project was approved by the Subcommittee on Animal Studies of the Research and Development Committee, Nashville VA Medical Center. Sprague-Dawley male rats (Hormone Assay Laboratories, Inc., Chicago, II1.,or Zivic-Miller Laboratories, Inc., Zelienople, Pa.) were hypophysectomized at 3 weeks of age and used for experiments after a postoperative interval of at least 9 days. Individual segments of costal cartilage from nine rats were assigned to treatment groups of a given experiment in a randomized block design. Unless specifically stated to the contrary, each segment was incubated in 1.0 ml basal medium without or with additions at 37° C for 24 hours. Sulfate or thymidine incorporation was determined by transfer of each tissue sample from the test media to 1.0 ml basal
J Lab Clin Med April 1997
Salmon and Burkhalter
432
®
b
Table I. Effects of purified bGH, r e c o m b i n a n t bGH,
IGF-I. *~
and r e c o m b i n a n t hGH on sulfate incorporation into cartilage from hypophysectomized rats
X
1.5
IGF-I
;~
Exp.
v z
_o 1.0
1
n,0
o t~ z
icJ
0.5
CONTROL
I 8
I 16
I 24
I 24
I 36
I 48
INCUBATION TIME (h)
Fig. 1. Time courses of actions of recombinant bGH (100 ng/ml) and IGF-I (10 ng/ml; thr-59 analog in Exp. A) on sulfate incorporation into cartilage from hypophysectomized rats. Cartilage samples were incubated in basal medium with additions as indicated for times shown (Exp. A: 8 to 24 hours; Exp. B: 24 to 48 hours), then transferred to basal medium containing 38804 and incubated for 1 or 2 hours (Exp. A and B, respectively). Significance of difference from corresponding control (by analysis of variance and F test): *p < 0. 02; **p < 0. 001.
medium containing 35804 (1 ~Ci/ml) or tritiated thymidine (4 to 5 ixCi/ml); incubation was continued for 2 hours as a general rule. Methods for cartilage analysis of sulfate and thymidine incorporation have been described. 14'15 Analysis of conditioned media for IGF-I. Cartilage samples (10 from each of 12 rats) were preincubated individually in 1 ml basal medium for 2 hours and then transferred to 1 ml of fresh basal medium without or with bGH (100 ng/ml) for an additional 36 hours of incubation. The combined cartilage dry weights for the control and bGH treatments were 94.26 and 94.58 mg, respectively. Thymidine incorporation was measured in a separate pair of cartilage samples from each rat. Pooled media from each treatment individually were analyzed for IGF-I by Drs. Judson Van Wyk and Marjorie Svoboda. Samples (32 to 35 ml) were concentrated by filtration on an Amicon YM5 membrane, and the retentate and washings were evaporated to dryness. The residue was dissolved in trimethylamine-acetic acid buffer, pH 2.8; IGF-I was separated on a Waters Protein-Pak 125 column by elution with the same buffer and measured by radioimmunoassay. The estimate of recovery of IGF-I added to unconditioned medium (2.4 ng/ml) was approximately 75%. Statistical methods. Results are shown as the mean + SEM. The data were submitted to a two-way analysis of variance. If significant nonadditivity of treatment effects was shown by Tukey's test (invariably the case for measurements of thymidine incorporation), a logarithmic transformation of the raw data was applied before the final analysis was performed. 16 The significance of differences between treatments was determined by an F test. A p value of 0.05 or less was considered significant.
2
Addition to incubation medium (ng/ml) Control Purified bGH
10 100 1000 Recombinant bGH 10 100 1000 Control Recombinant bGH 10 100 Recombinant hGH 10 100
38804 incorporation (cpm/mg -+ S.E.) 923 1306 1435 1697 1312 1517 1594 1463 2091 2505 2045 2381
+ 38 -+ 108" -+ 1011_+ 1021_+ 94* _+ 1017 _+ 1 2 8 t + 112 + 156~: _+ 2041+ 197§ _+ 2541-
Significance of difference from corresponding control (by analysis of variance and F test). §p < 0.02. :~p < 0.01. *p < 0.005. tP < 0.001.
RESULTS Stimulation of c a r t i l a g e sulfate incorporation by G H a n d
IGF-I in vitro. With the defined basal medium for incubation of costal cartilage from hypophysectomized rats, the effects of purified b G H and recombinant b G H (Table I, Exp. 1) and then recombinant b G H and recombinant h G H (Table I, Exp. 2) were compared. There was significant stimulation of cartilage sulfation by each, and there was no significant difference between the two b G H preparations (10 to 1000 ng/ml) or between the recombinant bovine and h u m a n preparations (10 to 100 ng/ml). Subsequent experiments were performed with recombinant b G H unless specifically stated. The stimulation of sulfate incorporation at increasing times of cartilage incubation with b G H (100 ng/ml) or I G F - I (10 ng/ml) was determined. Cartilage was incubated without or with the horm o n e or growth factor for 8 to 24 hours and was then pulsed with 35804 in basal medium (Fig. 1, Exp. A). Sulfate incorporation increased on prolonging the time of incubation with either addition to the medium, and the rate of increase accelerated during the 16-to 24-hour period in the presence of b G H . W h e n the incubation times were extended from 24 to 48 hours (Fig. 1, Exp. B), sulfate incorporation was near maximal for either treatment by 24 hours. The 24-hour time of cartilage incubation was selected for dose-response comparisons. The results of three independent experiments illustrate a range of tissue responses to b G H and I G F - I (Fig. 2). I G F - I in a concentration of 1 to 2 ng/ml or higher had a significant stimulating effect. The results with b G H were m o r e variable. A significant effect was sometimes seen at a concentration less than 10 ng/ml
J Lab Clin Med Volume 129, Number 4
Salmon and Burkhalter
200
I
I
i
433
I
o
Oo
.J
O
IGF-I ~ / " ~
4
o
I-Z O
150
GF-I
UJ
o
:
W
2
W U 7
I
0 X
®
o
o i
Z
Q
°
E ¢,, o
o o
E
v
IO0
IGF-I
4
50
3
8 Z
7,7,7, ~
Z
0
bGH'~
:
o / o i
oo/¥
~/~
.
i
~
•
0 Q.
g.
2
•
.
: "
I
I
I
I
I
IO
I00
I000
IGF-I OR bGH [ng/ml)
0 Q) Z
Fig. 3. Pooled results of 42 independent experiments in which one
0
0 U)
-)('1~
4
I I I I0 IGF-I OR bGH
I I00 {no/ml)
Fig. 2. Dose responses of actions of recombinant b G H and IGF-I
(thr-59 analog in Exp. A) on sulfate incorporation into cartilage from hypophysectomized rats in three independent experiments. Broken line in each panel indicates control sulfate incorporation. Significance of difference from control (by analysis of variance and F test): *p < 0. 05; **p < 0. 02; ***p < 0. 001.
(Exp. A and B), but stimulation was reproducible at 50 ng/ml or greater. In each case the dose response was lower with bGH. Because of variations among experiments, the results of all experiments in which one or more concentrations of purified or recombinant b G H and IGF-I were tested at the same time were pooled (Fig. 3). Data were available from 42 separate experiments with a total of 71 and 76 points for b G H
or more concentrations of purified or recombinant b G H and IGF-I (or thr-59 analog) were tested for effects on sulfate incorporation into cartilage from hypophysectomized rats. Each point is mean percent increase of sulfate incorporation over control at given concentration of hormone or growth factor. Open circles = IGF-I; closed circles = bGH. Regression lines were determined by method of least squares (n = 71 and 76 for b G H and IGF-I, respectively; r 2 = 0. 46 and 0. 69 for b G H and IGF-I, respectively).
and IGF-I, respectively (each point representing the mean response at one concentration from an individual experiment). Regression lines (p < 0.001 for both b G H and IGF-I) showed the same disparity of slope (ratio IGF-I/bGH = 69.5/20.4) observed in individual experiments. The difference between slopes was highly significant (p < 0.001). Stimulation of cartilage thymidine incorporation by GH and IGF-I in vitro. G H stimulated cartilage thymidine
incorporation under similar in vitro conditions (Table II). There was no significant difference between purified and recombinant b G H (10 to 1000 ng/ml). The time courses of the actions of recombinant b G H and IGF-I were compared (Fig. 4). First, cartilage was incubated for 24 to 48 hours in basal medium with b G H (100 ng/ml) or IGF-I (10 ng/ml) and was then pulsed with tritiated thymidine in basal medium (Exp. A). Stimulation of thymidine incorporation by IGF-I was evident by 24 hours, but in an exception to the general rule the effect of b G H was not significant at the time. Stimulation by both b G H and IGF-I was significant at 36 hours. At 48 hours
434
Table II. Effects of purified bGH and recombinant bGH on thymidine incorporation into cartilage from hypophysectomized rats
®
Z 0 I-
U
J Lab Clin M e d April 1997
Salmon a n d Burkhalter
Addition to incubation medium (ng/ml)
X
Control Purified bGH -r
24
48 24 INCUBATION TIME (h)
48
Fig. 4. Time courses of actions of recombinant bGH (100 ng/ml) and IGF-I (10 ng/ml) on thymidine incorporation into cartilage from hypophysectomized rats. Cartilage samples were incubated in basal medium with additions as indicated for times shown (Exp. A: 24 to 48 hours; Exp. B: 12 to 60 hours), then transferred to basal medium containing tritiated thymidine. Significance of difference from corresponding control (by analysis of variance and F test): *p < 0. 002; **p < 0. 001.
the effect of IGF-I had declined, but that of bGH had risen further. Next, to determine whether the effect of b G H continued to increase, the incubation period was extended from 12 to 60 hours (Exp. B). In this study thymidine incorporation was significantly increased by bGH at 24 hours. Peak stimulation was at 36 to 48 hours, and there was a decrease from that level by 60 hours. Therefore a rise followed by a decline is characteristic of the action of both b G H and IGF-I in vitro, a pattern that Daughaday and Reeder 3 described from in vivo experiments with GH as "synchronous activation of DNA synthesis." Because the time to achieve maximal stimulation of thymidine incorporation may differ, a dose-response comparison of GH and IGF-I with an arbitrary incubation period is questionable. However, in a determination at 36 hours in three independent experiments, the relatively low dose-response slope produced by b G H was similar to the pattern observed with sulfate incorporation as the index of tissue response (Fig. 5). The effective dose range for stimulation of sulfate and thymidine incorporation by b G H was comparable. Inhibition of GH effects on cartilage by antibody to
IGF-I. The possibility of an increased accumulation of
IGF-I in the medium during incubation of cartilage with bGH (100 ng/ml) was assessed in an experiment of 36 hours' duration. After corrections for estimated losses were made, the concentrations of IGF-I in the control and b G H media were 24 and 36 pg/ml, respectively. The difference in IGF-I levels did not provide convincing evidence in explanation
10 100 1000 Recombinant bGH 10 100 1000
3H-thymidineincorporation (cpm/mg -+ S.E.) 218 934 1121 1308 921 1450 1499
-+ 70.9 _+ 220* -+ 273* + 339* + 210" + 294* -+ 500*
Significance of difference from control (by analysis of variance and F test). *p < 0.001
of the difference in thymidine incorporation, which was 5.3 times the control in the b G H treatment group. Therefore experiments were performed with an IgG fraction of a rabbit antiserum to IGF-I based on the reported inhibition of the local action of GH on hypophysectomized rat cartilage in vivo by an antiserum directed against IGF-I. 7 Hormones and growth factors in concentrations that produced highly significant (p < 0.001) stimulation of sulfate and thymidine incorporation into hypophysectomized rat cartilage were tested at the same time in combination with the IgG fraction (Fig. 6). The IgG fraction significantly inhibited the effects of IGF-I, a rat serum fraction containing IGF-I, and bGH, but not insulin. The pattern of inhibition was similar when the effects on sulfate (Exp. A) and thymidine (Exp. B) incorporation were compared. As an additional control, an IgG fraction prepared from normal rabbit serum was tested in a protocol comparable to Exp. A. Again, the effects of hormones and growth factors were highly significant (p < 0.001). The increase of cartilage sulfate incorporation above the level achieved in basal medium was as follows: control 0%/-10%, IGF-I 58%/54%, rat serum IGF-I fraction 43%/49%, b G H 52%/49%, and insulin 30%/39% (without/with IgG). The effect of the IgG addition in each case was not significant. DISCUSSION
The original hypophysectomized rat cartilage assay system used a phosphate-buffered incubation medium with glucose as the only nutrientJ A slight stimulation of cartilage sulfate uptake by b G H 50 txg/ml was observed, but concentrations of the hormone up to 5 izg/ml were ineffective. The selection of the experimental conditions was fortuitous, because the relative insensitivity of cartilage to GH under those circumstances provided the opportunity
J Lab Clin M e d V o l u m e 129, N u m b e r 4
Salmon a n d Burkhalter
435
ANTI-IGF-I
®
r-I-
cpmlmg X I0 "3
4
Z
3
[]+
u
o
oz n-
2
O Z
tO I
O
io
E E
¢-i
(/)
~rq---
X
6
_®
z o
.J
o
0
~2 0 U
Z
_o
I-nO 11. nO (D
IGF'~
4
-~
_=
~
I,,d
zw 0
I
=E
~/
..r
~ ' T
2
-r-
CONTROL IGF-I
Z
£3
6
_©
k-
4-
tO
4
IGF-I
bGH .~ I
2
SM
bGH
INSULIN
Fig. 6. Effects of IgG fraction of rabbit antiserum to IGF-I on actions of IGF-I (2. 5 ng/ml), rat serum fraction containing IGF-I (SM: 0.01 U/ml), recombinant bGH (100 ng/ml), and insulin (50 ng/ml). Cartilage samples were preincubated in 1 ml of basal medium for 2 hours, then transferred to 0. 5 ml of basal medium with additions as indicated _+ IgG fraction. Incubation was continued for 24 or 36 hours, at which time cartilage samples were transferred to basal medium containing 35SO4 or tritiated thymidine, respectively. Significance of difference from corresponding treatment without IgG (by analysis of variance and F test): *p < 0. 01; **p < 0. 001.
bJ Z
>I
®
_
I I I I I0 I00 IGF-I OR bGH (ng/ml) Fig. 5. Dose responses of actions of recombinant bGH and IGF-I on thymidine incorporation into cartilage from hypophysectomized rats in three independent experiments. Cartilage samples were incubated in basal medium with additions as indicated for 36 hours, then transferred to basal medium containing tritiated thymidine. Broken line in each panel indicates control thymidine incorporation. Significance of difference from control (by analysis of variance and F test): *p < 0. 05; **p < 0. 002; ***p < 0. 001,
to recognize a GH-dependent serum factor, now designated IGF-I. Almqvist a7 used a similar medium and found no effect of hGH in concentrations up to 32 ixg/ml. However, he noted a slight stimulation of
cartilage sulfation with hGH (2 or 8 ixg/ml) in the presence of hypophysectomized human serum or hGH (0.25 or 1 txg/ml) with normal human serum. A previous experiment 1 showed no effect of b G H (140 ng/ml) in a medium containing a high concentration of plasma from hypophysectomized rats, but we have found (unpublished experiments) that a nondialyzable component of hypophysectomized rat serum inhibits the stimulating effects of both b G H and IGF-I on sulfate incorporation by hypophysectomized rat cartilage. In retrospect, the variable effects of GH in the early experiments may have been clues to the possibility of both inhibitory and enhancing factors in serum or as impurities in certain natural GH preparations. Cartilage assay media subsequently have included amino acids, because it was recognized that demonstration of optimal effects of the serum sulfation factor required their presence. 12'~3 However, previous attempts to reproduce the stimulating effects of
436
Salmon and Burkhalter
GH in vivo or normal rat serum in vitro on sulfate incorporation into proteoglycans and thymidine incorporation into DNA of hypophysectomized rat cartilage explants by adding GH in concentrations of at least 1 ~g/ml to an amino acid-glucose nutrient medium were unsuccessful.3'1s A simple modification of such a medium, which consisted of substitution of HEPES for a phosphate buffer and addition of a low concentration of BSA, has allowed the demonstration of effects of purified or recombinant bGH and recombinant hGH with a sensitivity of less than 100 ng/ml. Renewed interest in the local action of GH on cartilage followed soon after the demonstration that multiple tissues of the fetal mouse produced IGF-119 and that GH increased the production of IGF-I in cultured human fibroblasts.2°'2~ An innovative approach was followed by lsaksson et al.22 in Sweden; they demonstrated that direct injections of hGH into the proximal tibial epiphyseal cartilage of hypophysectomized rats stimulated local bone growth. In follow-up studies the lowest effective dose of hGH was 0.05 Cg daily for 5 days. 23 Because the injection volume was 20 txl per dose, the concentration of GH at this dose was 2500 ng/ml, which is much greater than the sensitivity threshold of costal cartilage in vitro. Subsequent experiments by these and other investigators showed that unilateral arterial infusion of rat GH 24 or recombinant hGH 2s into a hindlimb of hypophysectomized rats in doses insufficient to produce a systemic effect stimulated tibial epiphyseal growth in that extremity. Schlechter et al.7 made the key observation that the local action of rat GH could be blocked by simultaneous infusion of a rabbit antiserum to IGF-I, whereas infusion of normal rabbit serum was ineffective. From this result it was concluded that GH stimulated long bone growth by inducing local production of IGF-I, but a role for the circulating growth factor was not excluded. Further studies in the rat showed that cultured tibial epiphyseal chondrocytes contained GH receptors26; GH increased mRNA for IGF-I in rib growth plate, 27 GH increased IGF-I-containing cells in tibial epiphyseal cartilage,2s and GH increased sulfate and thymidine incorporation into cultured tibial epiphyseal chondrocytes.29 Also, fetal rat tibial diaphyseal bone increased the release of IGF-I when cultured in a medium with GH, 3° and GH-stimulated growth of cultured newborn rat calvarial osteoblasts was inhibited by IgG from an antiserum to IGF-I. 3~ Inhibition of the in vitro effects of GH on sulfate and thymidine incorporation into hypophysectomized rat cartilage by IgG from an IGF-I an-
J Lab Clin Med April 1997
tiserum further supports the conclusion that increased local production of IGF-I is involved in the direct action of GH on skeletal tissues. Despite this conclusion a case for an endocrine and an autocrine/paracrine role of IGF-I has been presented. 5'6'32 Daughaday5 has emphasized that the tibial epiphyseal growth stimulation from direct local injections of GH is only a fraction of the maximal response to systemic GH treatment, which is accompanied by increased circulating IGF-I. The relatively low dose-response slope of bGH compared with IGF-I in experiments with costal cartilage explants may be analogous, and this parallel from in vivo and in vitro studies suggests that a suboptimal nutrient medium was not responsible for the limited tissue response to GH in vitro. Potency of IGF-I as an endocrine growth factor was shown by the tibial epiphyseal growth in hypophysectomized rats during systemic therapy with a recombinant preparation. 33 Although the systemic effect of recombinant hGH on the tibial epiphysis was disproportionate to the increase of serum IGF-I,33 the responses of cultured normal rat chondrocytesz9 and hypophysectomized rat cartilage explants to GH and IGF-I were greater with IGF-I. This result suggests that the action of GH to increase skeletal growth in hypophysectomized rats may involve systemic effects in addition to the stimulation of IGF-I production such as increased insulin levels, 34 modulation of inhibitory IGF-binding protein(s), 35 and possibly the expression of a potentiator 15 in the face of reduced inhibitor activity. A quantitative reassessment of the direct action of GH on cartilage from hypophysectomized rats is now proposed. Costal cartilage sulfate incorporation in Sprague-Dawley male rats is reduced by hypophysectomy to less than half of normal control levels, a and restoration follows acute GH treatment of the animals. GH levels in the serum of unstressed normal rats are pulsatile, but the overall mean and peak levels for a 6-hour period in Sprague-Dawley male rats 1 month of age or older have been reported to average 60 to 86 and 217 to 434 ng/ml, respectively.36's7 Normal rat serum IGF-I levels are more than 400 ng/ml.6 The dose-response regression from our pooled assay results shows that bGH concentrations of 100 to 1000 ng/ml increased sulfate incorporation by 50% to 70%, which is roughly equivalent to one third to one half the restoration of a normal state. These effects of bGH were comparable to those of 2.5 to 5 ng/ml IGF-I, and 50 ng/ml IGF-I was twice as effective as 1000 ng/ml bGH. Therefore the direct action of GH in high concentrations in vitro failed to increase cartilage sulfate
J Lab Clin Med Volume 129, Number 4
incorporation to a normal level, but tissue exposure to extrinsic IGF-I was successful. This evidence is consistent with the hypothesis of Daughaday 5'6 that the effects of G H on skeletal growth represent a combination of direct and indirect actions, with an intermediary role of IGF-I in either case. The authors are grateful to Drs. Leslie Holladay, David Puett, Thomas I-Iolzman, and Linda Fryklund for gifts of various GH preparations. We are indebted to Drs. Judson Van Wyk, Marjorie Svoboda, and Louis Underwood for the determinations of IGF-I and -II in the serum fraction and IGF-I in conditioned media, a generous supply of IGF-I antiserum, and their interest and advice.
Salmon and Burkhalter
15.
16.
17.
18. REFERENCES
1. Salmon WD Jr, Danghaday WH. A hormonally controlled serum factor which stimulates sulfate incorporation by cartilage in vitro. J Lab Clin Med 1957;49:825-36. 2. Daughaday WH, Hall K, Salmon WD Jr, Van den Brande JL, Van Wyk JJ. On the nomenclature of the somatomedins and insulin-like growth factors. J Clin Endocrinol Metab 1987;65: 1075-6. 3. Danghaday WH, Reeder C. Synchronous activation of DNA synthesis in hypophysectomized rat cartilage by growth hormone. J Lab Clin Med 1966;68:357-68. 4. Isaksson OGP, Lindahl A, Nilsson A, Isgaard J. Mechanism of the stimulatory effect of growth hormone on longitudinal bone growth. Endocr Rev 1987;8:426-38. 5. Daughaday WH. A personal history of the origin of the somatomedin hypothesis and recent challenges to its validity. Perspect Biol Med 1989;32:194-211. 6. Daughaday WH, Rotwein P. Insulin-like growth factors I and II. Peptide, messenger ribonucleic acid and gene structures, serum, and tissue concentrations. Endocr Rev ~989; 10:68-91. 7. Schlechter NL, Russell SM, Spencer EM, Nicoll CS. Evidence suggesting that the direct growth-promoting effect of growth hormone on cartilage in vivo is mediated by local production of somatomedin. Proc Natl Acad Sci U S A 1986;83:7932-4. 8. Werner H, Adamo M, Roberts CT Jr, LeRoith D. Molecular and cellular aspects of insulin-like growth factor action. Vitam Horm 1994;48:1-58. 9. Isaksson OGP, Ed6n S, Jansson J-O. Mode of action of pituitary growth hormone on target cells. Annu Rev Physiol 1985;47:483-99. 10. Salmon WD Jr. Interaction of somatomedin and a peptide inhibitor in serum of hypophysectomized and starved, pituitary-intact rats. Adv Metab Disord 1975;8:183-99. 11. Salmon WD Jr, Holladay LA, Burkhalter VJ. Partial characterization of somatomedin inhibitor in starved rat serum. Endocrinology 1983;112:360-70. 12. Salmon WD Jr, Daughaday WH. The importance of amino acids as dialyzable components of rat serum which promote sulfate uptake by cartilage from hypophysectomized rats in vitro. J Lab Clin Med 1958;51:167-73. 13. Koumans J, Daughaday WH. Amino acid requirements for activity of partially purified sulfation factor. Trans Assoc Am Physicians 1963;76:152-62. 14. Sullivan JN, Bomboy JD Jr, Salmon WD Jr. Stimulation of uridiue, leucine, and sulfate incorporation into rat cartilage
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
437
macromolecules by adenosine 3',5'-monophosphate. Endocrinology 1984;115:793-800. Bomboy JD Jr, Salmon WD Jr. Effects of cyclic nucleotides on deoxyribonucleic acid synthesis in hypophysectomized rat cartilage: stimulation of thymidine incorporation and potentiation of the action of somatomedin by analogs of adenosine 3',5'-monophosphate or a cyclic nucleotide phosphodiesterase inhibitor. Endocrinology 1980;107:626-32. Kirk RE. Fundamental assumptions in analysis of variance. In: Experimental design: procedures for the behavioral sciences. Belmont, CA: Brooks-Cole, 1968:35-67. Almqvist S. Studies on sulfation factor (SF) activity of human serum. Evaluation of an improved method for the in vitro bioassay of SF and the effect of glutamine and human growth hormone in the system. Acta Endocrinol (Copenh) 1961;36: 31-50. Phillips LS, Weiss LJ. Circulating growth factor studies in growth plate versus resting cartilage in vitro: tissue responsiveness. Endocrinology 1982;111:1255-62. D'Ercole AJ, Applewhite GT, Underwood LE. Evidence that somatomedin is synthesized by multiple tissues in the fetus. Dev Biol 1980;75:315-28. Atkison PR, Weidman ER, Bhaumick B, Bala RM. Release of somatomedin-like activity by cultured WI-38 human fibroblasts. Endocrinology 1980;106:2006-12. Clemmons DR, Underwood LE, Van Wyk JJ. Hormonal control of immunoreactive somatomedin production by cultured human fibroblasts. J Clin Invest 1981;67:10-9. Isaksson OGP, Jansson J-O, Gause/AM. Growth hormone stimulates longitudinal bone growth directly. Science 1982; 216:1237-9. Isgaard J, Nilsson A, Lindahl A, Jansson J-O, Isaksson OGP. Effects of local administration of GH and IGF-I on longitudinal bone growth in rats. Am J Physiol 1986;250(Endocrinol Metab 13):E367-72. Schlechter NL, Russell SM, Greenberg S, Spencer EM, Nicoll CS. A direct growth effect of growth hormone in rat hindlimb shown by arterial infusion. Am J Physiol 1986; 250(Endocrinol Metab 13):E231-5. Nilsson A, Isgaard J, Lindahl A, Peterson L, Isaksson O. Effects of unilateral arterial infusion of GH and IGF-I on tibial longitudinal bone growth in hypophysectomized rats. Calcif Tissue Int 1987;40:91-6 Nilsson A, Lindahl A, Ed6n S, Isaksson OGP. Demonstration of growth hormone receptors in cultured rat epiphyseal chondrocytes by specific binding of growth hormone and immunohistochemistry. J Endocrinol 1989;122:69-77. Isgaard J, M611er C, Isaksson OGP, Nilsson A, Mathews LS, Norstedt G. Regulation of insulin-like growth factor messenger ribonucleic acid in rat growth plate by growth hormone. Endocrinology 1988;122:1515-20. Nilsson A, Isgaard J, Lindahl A, Dahlstr6m A, Skottner A, Isaksson OGP. Regulation by growth hormone of number of chondrocytes containing IGF-I in rat growth plate. Science 1986;233:571-4. Ohlsson C, Nilsson A, Isaksson OGP, Lindahl A. Effect of growth hormone and insulin-like growth factor-I on DNA synthesis and matrix production in rat epiphyseal chondrocytes in monolayer culture. J Endocrinol 1992;133:291-300. Stracke H, Schulz A, Moeller D, Rossol S, Schatz H. Effect of growth hormone on osteoblasts and demonstration of somatomedin-C/IGF I in bone organ culture. Acta Endocrinol (Copenh) 1984;107:16-24. Ernst M, Froesch ER. Growth hormone dependent stimula-
438
Salmon and Burkhalter
tion of osteoblast-like cells in serum-free cultures via local synthesis of insulin-like growth factor I. Biochem Biophys Res Commun 1988;151:142-7. 32. Gluckman PD, Douglas RG, Ambler GR, Breier BH, Hodgkinson SC, Koea JB, et al. The endocrine role of insulin-like growth factor I. Acta Paediatr Scand (Suppl) 1991; 372:97-105. 33. Guler H-P, Zapf J, Scheiwiller E, Froesch ER. Recombinant human insulin-like growth factor I stimulates growth and has distinct effects on organ size in hypophysectomized rats. Proc Natl Acad Sci U S A 1988;85:4889-93. 34. Gosteli-Peter MA, Winterhalter KH, Schmid C, Froesch ER, Zapf J. Expression and regulation of insulin-like growth factor-I
J Lab Clin Med April 1997
(IGF-I) and IGF-binding protein messenger ribonucleic acid levels in tissues of hypophysectomized rats infused with IGF-I and growth hormone. Endocrinology 1994;135:2558-67. 35, Rechler MM. Insulin-like growth factor binding proteins. Vitam Horm 1993;47:1-114. 36. Ed6n S. Age- and sex-related differences in episodic growth hormone secretion in the rat. Endocrinology 1979;105:55560. 37. Terry LC, Martin JB. The effects of lateral hypothalamicmedial forebrain stimulation and somatostatin antiserum on pulsatile growth hormone secretion in freely behaving rats: evidence for a dual regulatory mechanism. Endocrinology 1981;109:622-7.
These studies were undertaken in an attempt to demonstrate sensitivity of cartilage from hypophysectomized rats to growth hormone (GH) in physiologic concentrations in vitro and to compare the actions of the hormone and extrinsic insulin-like growth factor-I (IGF-I). A HEPES-buffered, amino acid-glucose nutrient medium with a low concentration of bovine serum albumin, costal cartilage explants from such animals was responsive to purified or recombinant bovine GH (bGH) and recombinant human GH (hGH) with a sensitivity of less than 100 ng/ml. Sulfate incorporation into proteoglycans and thymidine incorporation into DNA were increased. The effect on sulfate incorporation was near maximal by 24 hours, but the peak effect on thymidine incorporation was delayed. Comparison of the dose-response relationships of bGH and IGF-I demonstrated a greater effect of IGF-I with increasing concentrations of each. The action of bGH was inhibited by an immunoglobulin G fraction of an IGF-I antiserum. These results are consistent with previously available evidence that IGF-I has an autocrine/paracrine action on cartilage, which is regulated by GH. However, the limited cartilage response to the direct action of GH supports the hypothesis of Daughaday that IGF-I also has an endocrine role in the stimulation of skeletal growth by the pituitary hormone. (J Lab Clin Med 1997;129: 430-8)
Abbreviations: b G H = bovine growth hormone; BSA = bovine serum albumin; D N A = deoxyribonucleic acid; G H = growth hormone; HEPES = N-2-hydroxyethypiperazine-N'-2-ethanesulfonic acid; h G H = h u m a n growth hormone; IGF = insulin-likegrowth factor; IgG = i m m u n o globulin G; thr-59 = threonine-59
he somatomedin hypothesis was based on observations in William Daughaday's laboratory nearly 40 years ago. 1 Hypophysectomy in rats had been shown to cause a reduction of sulfate incorporation into cartilage proteoglycans, and G H
T
From Medical and Research Services, Veterans Administration Medical Center, and the Department of Medicine, Vanderbilt University School of Medicine. Supported by the Department of Veterans Affairs. Submitted for publication Sept. 5, 1996; accepted Oct. 28, 1996. Reprint requests: William D. Salmon, Jr., MD, Medical Service, VA Medical Center, Nashville, TN 37212-2637. 0022-2143/97 $5.00 + 0 5/1/79944
430
treatment corrected the abnormality. Then it was found that G H in physiologic concentrations in vitro had no significant effect on sulfate incorporation into cartilage from hypophysectomized rats, but a cartilage-stimulating factor was detected in serum after injection of the h o r m o n e into such animals. Insulin had an action similar to the serum factor, but high concentrations were required. It was postulated that the action of G H on cartilage was mediated by the serum sulfation factor, redesignated somatomeAn editorial relevant to this article appears on p. 398 of this issue of the Journal.
J Lab Clin Med Volume 129, Number 4
din and now IGF-I 2 after the recognition of other effects including stimulation of thymidine incorporation into D N A of cartilage from hypophysectomized rats. 3 Challenges to the hypothesis resulted initially from in vivo experiments that demonstrated rat cartilage is more sensitive to G H than revealed by the original in vitro studies. 4-6 Additional evidence has indicated that the effects of G H are mediated partly by local production of IGF-I, 7 whereas the serum level of IGF-I is maintained primarily by hepatic production under G H regulation, s A modification of the somatomedin hypothesis has been necessary because of this new information, and doubt remains about the role of the circulating growth factor or the endocrine versus the autocrine/ paracrine function of IGF-I. Despite the tremendous expansion of knowledge in this area, the insensitivity of hypophysectomized rat cartilage to G H in vitro in previous studies has remained a curious anomaly. It has been suggested that unfavorable tissue incubation conditions may be the explanation. 9 Also, the possibility that inhibitors 1° may have been present in certain G H preparations or serum supplements is now appreciated. Therefore the direct action of G H has been reexamined with purified or recombinant preparations of the hormone in a modified incubation medium. In this system G H in physiologic concentrations increased sulfate and thymidine incorporation into costal cartilage from hypophysectomized rats, and those effects were inhibited by an IgG fraction of an IGF-I antiserum. Quantitative considerations of the direct effects of G H and extrinsic IGF-I support an endocrine and an autocrine/paracrine role for the growth factor as postulated originally by Daughaday. 5 METHODS Hormones, growth factors, chemicals. Purified bGH (from NIH-GH-B17) was provided by Drs. Leslie Holladay and David Puett (Department of Biochemistry, Vanderbilt University School of Medicine), recombinant bGH by Dr. Thomas Holzman (The Upjohn Co., Kalamazoo, Mich.), and recombinant hGH by Dr. Linda Fryklund (KabiVitrum Peptide Hormones AB, Stockholm, Sweden). Recombinant IGF-I (or thr-59 analog) was obtained from AmGen (Thousand Oaks, Calif.). Amino acids and HEPES buffer components were obtained from CalBiochem Corp. (San Diego, Calif.), and [35S]sulfuric acid and [methyl-3H]thymidine (2.0 Ci/mmol) were obtained from New England Nuclear (Boston, Mass.). A rat serum fraction with somatomedin (IGF) activity was prepared from serum of normal animals. 11 Bioactivity determined by the hypophysectomized rat cartilage assay for sulfation-promoting activity was 54 U/mg of protein (1 U = activity equivalent to 1 ml normal rat serum).
Salmon and Burkhalter
431
IGF-I immunoreactivity was 155 ng/U; IGF-II was undetectable. The IGF assays were provided by Drs. Judson Van Wyk and Marjorie Svoboda (Department of Pediatrics, University of North Carolina School of Medicine). Antibody preparation. A rabbit antiserum (UBK 487) directed against somatomedin-C (IGF-I) was supplied by Dr. Louis Underwood (Department of Pediatrics, University of North Carolina School of Medicine). An IgG fraction of this antiserum (0.25 ml) was prepared by (1) nonspecific adsorption on a column (0.7 x 2.5 cm) of Sepharose 4 Fast Flow (Pharmacia Biotech, Piscataway, N.J.) equilibrated with phosphate-buffered saline solution (0.13 mol/L NaC1-0.02 mol/L NazHPO4, pH 7.0), (2) specific adsorption of the effluent and washings (11 ml phosphate-buffered saline solution) on a column (0.7 × 2.5 cm) of Protein-G Sepharose 4 Fast Flow (Pharmacia Biotech), which was equilibrated also with phosphate-buffered saline solution, (3) washing with 10 ml phosphatebuffered saline solution, (4) elution with 10 ml of 0.1 mol/L glycine-HC1 buffer, pH 2.6, (5) neutralization and dilution of the eluate with HEPES buffers, and (6) dialysis in Spectrapor 2 tubing against a HEPES-buffered saline solution (see following text). The final concentration of this fraction in assay media was equivalent to a 1:200 dilution of the original antiserum. Incubation medium. The defined basal medium for cartilage incubation was constituted with a HEPES-buffered saline solution: NaC1 124, KC1 4.6, NaeHPO 4 1.1, CAC12-2 H20 1.25, MgSO 4 1.0, HEPES 11, and HEPES.Na 14 mmol/L. The other components were amino acids in the formulation of Eagle's basal medium for HeLa cells~eexcept for the addition of serine13: arginine 0.1, cystine 0.05, glutamine 2.0, histidine 0.05, isoleucine 0.2, leucine 0.2, lysine.HC1 0.2, methionine 0.05, phenylalanine 0.1, serine 0.1, threonine 0.2, tryptophane 0.02, tyrosine 0.1, and valine 0.2 mmol/L; glucose 2 mg/ml; penicillin 100 U/ml; gentamicin 50 ~g/ml; and crystalline BSA 0.2 mg/ml (ICN Pharmaceuticals, Inc., Cleveland Ohio, or CalBiochem Corp., the latter preparation designated nuclease- and protease-free). Although crystalline BSA from other commercial sources was tested and found suitable, one preparation was inhibitory, emphasizing the importance of a preliminary test of individual lots. The final pH was adjusted to 7.45 at 37° C with 1 N NaOH or HC1. Cartilage incubation and analysis. This project was approved by the Subcommittee on Animal Studies of the Research and Development Committee, Nashville VA Medical Center. Sprague-Dawley male rats (Hormone Assay Laboratories, Inc., Chicago, II1.,or Zivic-Miller Laboratories, Inc., Zelienople, Pa.) were hypophysectomized at 3 weeks of age and used for experiments after a postoperative interval of at least 9 days. Individual segments of costal cartilage from nine rats were assigned to treatment groups of a given experiment in a randomized block design. Unless specifically stated to the contrary, each segment was incubated in 1.0 ml basal medium without or with additions at 37° C for 24 hours. Sulfate or thymidine incorporation was determined by transfer of each tissue sample from the test media to 1.0 ml basal
J Lab Clin Med April 1997
Salmon and Burkhalter
432
®
b
Table I. Effects of purified bGH, r e c o m b i n a n t bGH,
IGF-I. *~
and r e c o m b i n a n t hGH on sulfate incorporation into cartilage from hypophysectomized rats
X
1.5
IGF-I
;~
Exp.
v z
_o 1.0
1
n,0
o t~ z
icJ
0.5
CONTROL
I 8
I 16
I 24
I 24
I 36
I 48
INCUBATION TIME (h)
Fig. 1. Time courses of actions of recombinant bGH (100 ng/ml) and IGF-I (10 ng/ml; thr-59 analog in Exp. A) on sulfate incorporation into cartilage from hypophysectomized rats. Cartilage samples were incubated in basal medium with additions as indicated for times shown (Exp. A: 8 to 24 hours; Exp. B: 24 to 48 hours), then transferred to basal medium containing 38804 and incubated for 1 or 2 hours (Exp. A and B, respectively). Significance of difference from corresponding control (by analysis of variance and F test): *p < 0. 02; **p < 0. 001.
medium containing 35804 (1 ~Ci/ml) or tritiated thymidine (4 to 5 ixCi/ml); incubation was continued for 2 hours as a general rule. Methods for cartilage analysis of sulfate and thymidine incorporation have been described. 14'15 Analysis of conditioned media for IGF-I. Cartilage samples (10 from each of 12 rats) were preincubated individually in 1 ml basal medium for 2 hours and then transferred to 1 ml of fresh basal medium without or with bGH (100 ng/ml) for an additional 36 hours of incubation. The combined cartilage dry weights for the control and bGH treatments were 94.26 and 94.58 mg, respectively. Thymidine incorporation was measured in a separate pair of cartilage samples from each rat. Pooled media from each treatment individually were analyzed for IGF-I by Drs. Judson Van Wyk and Marjorie Svoboda. Samples (32 to 35 ml) were concentrated by filtration on an Amicon YM5 membrane, and the retentate and washings were evaporated to dryness. The residue was dissolved in trimethylamine-acetic acid buffer, pH 2.8; IGF-I was separated on a Waters Protein-Pak 125 column by elution with the same buffer and measured by radioimmunoassay. The estimate of recovery of IGF-I added to unconditioned medium (2.4 ng/ml) was approximately 75%. Statistical methods. Results are shown as the mean + SEM. The data were submitted to a two-way analysis of variance. If significant nonadditivity of treatment effects was shown by Tukey's test (invariably the case for measurements of thymidine incorporation), a logarithmic transformation of the raw data was applied before the final analysis was performed. 16 The significance of differences between treatments was determined by an F test. A p value of 0.05 or less was considered significant.
2
Addition to incubation medium (ng/ml) Control Purified bGH
10 100 1000 Recombinant bGH 10 100 1000 Control Recombinant bGH 10 100 Recombinant hGH 10 100
38804 incorporation (cpm/mg -+ S.E.) 923 1306 1435 1697 1312 1517 1594 1463 2091 2505 2045 2381
+ 38 -+ 108" -+ 1011_+ 1021_+ 94* _+ 1017 _+ 1 2 8 t + 112 + 156~: _+ 2041+ 197§ _+ 2541-
Significance of difference from corresponding control (by analysis of variance and F test). §p < 0.02. :~p < 0.01. *p < 0.005. tP < 0.001.
RESULTS Stimulation of c a r t i l a g e sulfate incorporation by G H a n d
IGF-I in vitro. With the defined basal medium for incubation of costal cartilage from hypophysectomized rats, the effects of purified b G H and recombinant b G H (Table I, Exp. 1) and then recombinant b G H and recombinant h G H (Table I, Exp. 2) were compared. There was significant stimulation of cartilage sulfation by each, and there was no significant difference between the two b G H preparations (10 to 1000 ng/ml) or between the recombinant bovine and h u m a n preparations (10 to 100 ng/ml). Subsequent experiments were performed with recombinant b G H unless specifically stated. The stimulation of sulfate incorporation at increasing times of cartilage incubation with b G H (100 ng/ml) or I G F - I (10 ng/ml) was determined. Cartilage was incubated without or with the horm o n e or growth factor for 8 to 24 hours and was then pulsed with 35804 in basal medium (Fig. 1, Exp. A). Sulfate incorporation increased on prolonging the time of incubation with either addition to the medium, and the rate of increase accelerated during the 16-to 24-hour period in the presence of b G H . W h e n the incubation times were extended from 24 to 48 hours (Fig. 1, Exp. B), sulfate incorporation was near maximal for either treatment by 24 hours. The 24-hour time of cartilage incubation was selected for dose-response comparisons. The results of three independent experiments illustrate a range of tissue responses to b G H and I G F - I (Fig. 2). I G F - I in a concentration of 1 to 2 ng/ml or higher had a significant stimulating effect. The results with b G H were m o r e variable. A significant effect was sometimes seen at a concentration less than 10 ng/ml
J Lab Clin Med Volume 129, Number 4
Salmon and Burkhalter
200
I
I
i
433
I
o
Oo
.J
O
IGF-I ~ / " ~
4
o
I-Z O
150
GF-I
UJ
o
:
W
2
W U 7
I
0 X
®
o
o i
Z
Q
°
E ¢,, o
o o
E
v
IO0
IGF-I
4
50
3
8 Z
7,7,7, ~
Z
0
bGH'~
:
o / o i
oo/¥
~/~
.
i
~
•
0 Q.
g.
2
•
.
: "
I
I
I
I
I
IO
I00
I000
IGF-I OR bGH [ng/ml)
0 Q) Z
Fig. 3. Pooled results of 42 independent experiments in which one
0
0 U)
-)('1~
4
I I I I0 IGF-I OR bGH
I I00 {no/ml)
Fig. 2. Dose responses of actions of recombinant b G H and IGF-I
(thr-59 analog in Exp. A) on sulfate incorporation into cartilage from hypophysectomized rats in three independent experiments. Broken line in each panel indicates control sulfate incorporation. Significance of difference from control (by analysis of variance and F test): *p < 0. 05; **p < 0. 02; ***p < 0. 001.
(Exp. A and B), but stimulation was reproducible at 50 ng/ml or greater. In each case the dose response was lower with bGH. Because of variations among experiments, the results of all experiments in which one or more concentrations of purified or recombinant b G H and IGF-I were tested at the same time were pooled (Fig. 3). Data were available from 42 separate experiments with a total of 71 and 76 points for b G H
or more concentrations of purified or recombinant b G H and IGF-I (or thr-59 analog) were tested for effects on sulfate incorporation into cartilage from hypophysectomized rats. Each point is mean percent increase of sulfate incorporation over control at given concentration of hormone or growth factor. Open circles = IGF-I; closed circles = bGH. Regression lines were determined by method of least squares (n = 71 and 76 for b G H and IGF-I, respectively; r 2 = 0. 46 and 0. 69 for b G H and IGF-I, respectively).
and IGF-I, respectively (each point representing the mean response at one concentration from an individual experiment). Regression lines (p < 0.001 for both b G H and IGF-I) showed the same disparity of slope (ratio IGF-I/bGH = 69.5/20.4) observed in individual experiments. The difference between slopes was highly significant (p < 0.001). Stimulation of cartilage thymidine incorporation by GH and IGF-I in vitro. G H stimulated cartilage thymidine
incorporation under similar in vitro conditions (Table II). There was no significant difference between purified and recombinant b G H (10 to 1000 ng/ml). The time courses of the actions of recombinant b G H and IGF-I were compared (Fig. 4). First, cartilage was incubated for 24 to 48 hours in basal medium with b G H (100 ng/ml) or IGF-I (10 ng/ml) and was then pulsed with tritiated thymidine in basal medium (Exp. A). Stimulation of thymidine incorporation by IGF-I was evident by 24 hours, but in an exception to the general rule the effect of b G H was not significant at the time. Stimulation by both b G H and IGF-I was significant at 36 hours. At 48 hours
434
Table II. Effects of purified bGH and recombinant bGH on thymidine incorporation into cartilage from hypophysectomized rats
®
Z 0 I-
U
J Lab Clin M e d April 1997
Salmon a n d Burkhalter
Addition to incubation medium (ng/ml)
X
Control Purified bGH -r
24
48 24 INCUBATION TIME (h)
48
Fig. 4. Time courses of actions of recombinant bGH (100 ng/ml) and IGF-I (10 ng/ml) on thymidine incorporation into cartilage from hypophysectomized rats. Cartilage samples were incubated in basal medium with additions as indicated for times shown (Exp. A: 24 to 48 hours; Exp. B: 12 to 60 hours), then transferred to basal medium containing tritiated thymidine. Significance of difference from corresponding control (by analysis of variance and F test): *p < 0. 002; **p < 0. 001.
the effect of IGF-I had declined, but that of bGH had risen further. Next, to determine whether the effect of b G H continued to increase, the incubation period was extended from 12 to 60 hours (Exp. B). In this study thymidine incorporation was significantly increased by bGH at 24 hours. Peak stimulation was at 36 to 48 hours, and there was a decrease from that level by 60 hours. Therefore a rise followed by a decline is characteristic of the action of both b G H and IGF-I in vitro, a pattern that Daughaday and Reeder 3 described from in vivo experiments with GH as "synchronous activation of DNA synthesis." Because the time to achieve maximal stimulation of thymidine incorporation may differ, a dose-response comparison of GH and IGF-I with an arbitrary incubation period is questionable. However, in a determination at 36 hours in three independent experiments, the relatively low dose-response slope produced by b G H was similar to the pattern observed with sulfate incorporation as the index of tissue response (Fig. 5). The effective dose range for stimulation of sulfate and thymidine incorporation by b G H was comparable. Inhibition of GH effects on cartilage by antibody to
IGF-I. The possibility of an increased accumulation of
IGF-I in the medium during incubation of cartilage with bGH (100 ng/ml) was assessed in an experiment of 36 hours' duration. After corrections for estimated losses were made, the concentrations of IGF-I in the control and b G H media were 24 and 36 pg/ml, respectively. The difference in IGF-I levels did not provide convincing evidence in explanation
10 100 1000 Recombinant bGH 10 100 1000
3H-thymidineincorporation (cpm/mg -+ S.E.) 218 934 1121 1308 921 1450 1499
-+ 70.9 _+ 220* -+ 273* + 339* + 210" + 294* -+ 500*
Significance of difference from control (by analysis of variance and F test). *p < 0.001
of the difference in thymidine incorporation, which was 5.3 times the control in the b G H treatment group. Therefore experiments were performed with an IgG fraction of a rabbit antiserum to IGF-I based on the reported inhibition of the local action of GH on hypophysectomized rat cartilage in vivo by an antiserum directed against IGF-I. 7 Hormones and growth factors in concentrations that produced highly significant (p < 0.001) stimulation of sulfate and thymidine incorporation into hypophysectomized rat cartilage were tested at the same time in combination with the IgG fraction (Fig. 6). The IgG fraction significantly inhibited the effects of IGF-I, a rat serum fraction containing IGF-I, and bGH, but not insulin. The pattern of inhibition was similar when the effects on sulfate (Exp. A) and thymidine (Exp. B) incorporation were compared. As an additional control, an IgG fraction prepared from normal rabbit serum was tested in a protocol comparable to Exp. A. Again, the effects of hormones and growth factors were highly significant (p < 0.001). The increase of cartilage sulfate incorporation above the level achieved in basal medium was as follows: control 0%/-10%, IGF-I 58%/54%, rat serum IGF-I fraction 43%/49%, b G H 52%/49%, and insulin 30%/39% (without/with IgG). The effect of the IgG addition in each case was not significant. DISCUSSION
The original hypophysectomized rat cartilage assay system used a phosphate-buffered incubation medium with glucose as the only nutrientJ A slight stimulation of cartilage sulfate uptake by b G H 50 txg/ml was observed, but concentrations of the hormone up to 5 izg/ml were ineffective. The selection of the experimental conditions was fortuitous, because the relative insensitivity of cartilage to GH under those circumstances provided the opportunity
J Lab Clin M e d V o l u m e 129, N u m b e r 4
Salmon a n d Burkhalter
435
ANTI-IGF-I
®
r-I-
cpmlmg X I0 "3
4
Z
3
[]+
u
o
oz n-
2
O Z
tO I
O
io
E E
¢-i
(/)
~rq---
X
6
_®
z o
.J
o
0
~2 0 U
Z
_o
I-nO 11. nO (D
IGF'~
4
-~
_=
~
I,,d
zw 0
I
=E
~/
..r
~ ' T
2
-r-
CONTROL IGF-I
Z
£3
6
_©
k-
4-
tO
4
IGF-I
bGH .~ I
2
SM
bGH
INSULIN
Fig. 6. Effects of IgG fraction of rabbit antiserum to IGF-I on actions of IGF-I (2. 5 ng/ml), rat serum fraction containing IGF-I (SM: 0.01 U/ml), recombinant bGH (100 ng/ml), and insulin (50 ng/ml). Cartilage samples were preincubated in 1 ml of basal medium for 2 hours, then transferred to 0. 5 ml of basal medium with additions as indicated _+ IgG fraction. Incubation was continued for 24 or 36 hours, at which time cartilage samples were transferred to basal medium containing 35SO4 or tritiated thymidine, respectively. Significance of difference from corresponding treatment without IgG (by analysis of variance and F test): *p < 0. 01; **p < 0. 001.
bJ Z
>I
®
_
I I I I I0 I00 IGF-I OR bGH (ng/ml) Fig. 5. Dose responses of actions of recombinant bGH and IGF-I on thymidine incorporation into cartilage from hypophysectomized rats in three independent experiments. Cartilage samples were incubated in basal medium with additions as indicated for 36 hours, then transferred to basal medium containing tritiated thymidine. Broken line in each panel indicates control thymidine incorporation. Significance of difference from control (by analysis of variance and F test): *p < 0. 05; **p < 0. 002; ***p < 0. 001,
to recognize a GH-dependent serum factor, now designated IGF-I. Almqvist a7 used a similar medium and found no effect of hGH in concentrations up to 32 ixg/ml. However, he noted a slight stimulation of
cartilage sulfation with hGH (2 or 8 ixg/ml) in the presence of hypophysectomized human serum or hGH (0.25 or 1 txg/ml) with normal human serum. A previous experiment 1 showed no effect of b G H (140 ng/ml) in a medium containing a high concentration of plasma from hypophysectomized rats, but we have found (unpublished experiments) that a nondialyzable component of hypophysectomized rat serum inhibits the stimulating effects of both b G H and IGF-I on sulfate incorporation by hypophysectomized rat cartilage. In retrospect, the variable effects of GH in the early experiments may have been clues to the possibility of both inhibitory and enhancing factors in serum or as impurities in certain natural GH preparations. Cartilage assay media subsequently have included amino acids, because it was recognized that demonstration of optimal effects of the serum sulfation factor required their presence. 12'~3 However, previous attempts to reproduce the stimulating effects of
436
Salmon and Burkhalter
GH in vivo or normal rat serum in vitro on sulfate incorporation into proteoglycans and thymidine incorporation into DNA of hypophysectomized rat cartilage explants by adding GH in concentrations of at least 1 ~g/ml to an amino acid-glucose nutrient medium were unsuccessful.3'1s A simple modification of such a medium, which consisted of substitution of HEPES for a phosphate buffer and addition of a low concentration of BSA, has allowed the demonstration of effects of purified or recombinant bGH and recombinant hGH with a sensitivity of less than 100 ng/ml. Renewed interest in the local action of GH on cartilage followed soon after the demonstration that multiple tissues of the fetal mouse produced IGF-119 and that GH increased the production of IGF-I in cultured human fibroblasts.2°'2~ An innovative approach was followed by lsaksson et al.22 in Sweden; they demonstrated that direct injections of hGH into the proximal tibial epiphyseal cartilage of hypophysectomized rats stimulated local bone growth. In follow-up studies the lowest effective dose of hGH was 0.05 Cg daily for 5 days. 23 Because the injection volume was 20 txl per dose, the concentration of GH at this dose was 2500 ng/ml, which is much greater than the sensitivity threshold of costal cartilage in vitro. Subsequent experiments by these and other investigators showed that unilateral arterial infusion of rat GH 24 or recombinant hGH 2s into a hindlimb of hypophysectomized rats in doses insufficient to produce a systemic effect stimulated tibial epiphyseal growth in that extremity. Schlechter et al.7 made the key observation that the local action of rat GH could be blocked by simultaneous infusion of a rabbit antiserum to IGF-I, whereas infusion of normal rabbit serum was ineffective. From this result it was concluded that GH stimulated long bone growth by inducing local production of IGF-I, but a role for the circulating growth factor was not excluded. Further studies in the rat showed that cultured tibial epiphyseal chondrocytes contained GH receptors26; GH increased mRNA for IGF-I in rib growth plate, 27 GH increased IGF-I-containing cells in tibial epiphyseal cartilage,2s and GH increased sulfate and thymidine incorporation into cultured tibial epiphyseal chondrocytes.29 Also, fetal rat tibial diaphyseal bone increased the release of IGF-I when cultured in a medium with GH, 3° and GH-stimulated growth of cultured newborn rat calvarial osteoblasts was inhibited by IgG from an antiserum to IGF-I. 3~ Inhibition of the in vitro effects of GH on sulfate and thymidine incorporation into hypophysectomized rat cartilage by IgG from an IGF-I an-
J Lab Clin Med April 1997
tiserum further supports the conclusion that increased local production of IGF-I is involved in the direct action of GH on skeletal tissues. Despite this conclusion a case for an endocrine and an autocrine/paracrine role of IGF-I has been presented. 5'6'32 Daughaday5 has emphasized that the tibial epiphyseal growth stimulation from direct local injections of GH is only a fraction of the maximal response to systemic GH treatment, which is accompanied by increased circulating IGF-I. The relatively low dose-response slope of bGH compared with IGF-I in experiments with costal cartilage explants may be analogous, and this parallel from in vivo and in vitro studies suggests that a suboptimal nutrient medium was not responsible for the limited tissue response to GH in vitro. Potency of IGF-I as an endocrine growth factor was shown by the tibial epiphyseal growth in hypophysectomized rats during systemic therapy with a recombinant preparation. 33 Although the systemic effect of recombinant hGH on the tibial epiphysis was disproportionate to the increase of serum IGF-I,33 the responses of cultured normal rat chondrocytesz9 and hypophysectomized rat cartilage explants to GH and IGF-I were greater with IGF-I. This result suggests that the action of GH to increase skeletal growth in hypophysectomized rats may involve systemic effects in addition to the stimulation of IGF-I production such as increased insulin levels, 34 modulation of inhibitory IGF-binding protein(s), 35 and possibly the expression of a potentiator 15 in the face of reduced inhibitor activity. A quantitative reassessment of the direct action of GH on cartilage from hypophysectomized rats is now proposed. Costal cartilage sulfate incorporation in Sprague-Dawley male rats is reduced by hypophysectomy to less than half of normal control levels, a and restoration follows acute GH treatment of the animals. GH levels in the serum of unstressed normal rats are pulsatile, but the overall mean and peak levels for a 6-hour period in Sprague-Dawley male rats 1 month of age or older have been reported to average 60 to 86 and 217 to 434 ng/ml, respectively.36's7 Normal rat serum IGF-I levels are more than 400 ng/ml.6 The dose-response regression from our pooled assay results shows that bGH concentrations of 100 to 1000 ng/ml increased sulfate incorporation by 50% to 70%, which is roughly equivalent to one third to one half the restoration of a normal state. These effects of bGH were comparable to those of 2.5 to 5 ng/ml IGF-I, and 50 ng/ml IGF-I was twice as effective as 1000 ng/ml bGH. Therefore the direct action of GH in high concentrations in vitro failed to increase cartilage sulfate
J Lab Clin Med Volume 129, Number 4
incorporation to a normal level, but tissue exposure to extrinsic IGF-I was successful. This evidence is consistent with the hypothesis of Daughaday 5'6 that the effects of G H on skeletal growth represent a combination of direct and indirect actions, with an intermediary role of IGF-I in either case. The authors are grateful to Drs. Leslie Holladay, David Puett, Thomas I-Iolzman, and Linda Fryklund for gifts of various GH preparations. We are indebted to Drs. Judson Van Wyk, Marjorie Svoboda, and Louis Underwood for the determinations of IGF-I and -II in the serum fraction and IGF-I in conditioned media, a generous supply of IGF-I antiserum, and their interest and advice.
Salmon and Burkhalter
15.
16.
17.
18. REFERENCES
1. Salmon WD Jr, Danghaday WH. A hormonally controlled serum factor which stimulates sulfate incorporation by cartilage in vitro. J Lab Clin Med 1957;49:825-36. 2. Daughaday WH, Hall K, Salmon WD Jr, Van den Brande JL, Van Wyk JJ. On the nomenclature of the somatomedins and insulin-like growth factors. J Clin Endocrinol Metab 1987;65: 1075-6. 3. Danghaday WH, Reeder C. Synchronous activation of DNA synthesis in hypophysectomized rat cartilage by growth hormone. J Lab Clin Med 1966;68:357-68. 4. Isaksson OGP, Lindahl A, Nilsson A, Isgaard J. Mechanism of the stimulatory effect of growth hormone on longitudinal bone growth. Endocr Rev 1987;8:426-38. 5. Daughaday WH. A personal history of the origin of the somatomedin hypothesis and recent challenges to its validity. Perspect Biol Med 1989;32:194-211. 6. Daughaday WH, Rotwein P. Insulin-like growth factors I and II. Peptide, messenger ribonucleic acid and gene structures, serum, and tissue concentrations. Endocr Rev ~989; 10:68-91. 7. Schlechter NL, Russell SM, Spencer EM, Nicoll CS. Evidence suggesting that the direct growth-promoting effect of growth hormone on cartilage in vivo is mediated by local production of somatomedin. Proc Natl Acad Sci U S A 1986;83:7932-4. 8. Werner H, Adamo M, Roberts CT Jr, LeRoith D. Molecular and cellular aspects of insulin-like growth factor action. Vitam Horm 1994;48:1-58. 9. Isaksson OGP, Ed6n S, Jansson J-O. Mode of action of pituitary growth hormone on target cells. Annu Rev Physiol 1985;47:483-99. 10. Salmon WD Jr. Interaction of somatomedin and a peptide inhibitor in serum of hypophysectomized and starved, pituitary-intact rats. Adv Metab Disord 1975;8:183-99. 11. Salmon WD Jr, Holladay LA, Burkhalter VJ. Partial characterization of somatomedin inhibitor in starved rat serum. Endocrinology 1983;112:360-70. 12. Salmon WD Jr, Daughaday WH. The importance of amino acids as dialyzable components of rat serum which promote sulfate uptake by cartilage from hypophysectomized rats in vitro. J Lab Clin Med 1958;51:167-73. 13. Koumans J, Daughaday WH. Amino acid requirements for activity of partially purified sulfation factor. Trans Assoc Am Physicians 1963;76:152-62. 14. Sullivan JN, Bomboy JD Jr, Salmon WD Jr. Stimulation of uridiue, leucine, and sulfate incorporation into rat cartilage
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
437
macromolecules by adenosine 3',5'-monophosphate. Endocrinology 1984;115:793-800. Bomboy JD Jr, Salmon WD Jr. Effects of cyclic nucleotides on deoxyribonucleic acid synthesis in hypophysectomized rat cartilage: stimulation of thymidine incorporation and potentiation of the action of somatomedin by analogs of adenosine 3',5'-monophosphate or a cyclic nucleotide phosphodiesterase inhibitor. Endocrinology 1980;107:626-32. Kirk RE. Fundamental assumptions in analysis of variance. In: Experimental design: procedures for the behavioral sciences. Belmont, CA: Brooks-Cole, 1968:35-67. Almqvist S. Studies on sulfation factor (SF) activity of human serum. Evaluation of an improved method for the in vitro bioassay of SF and the effect of glutamine and human growth hormone in the system. Acta Endocrinol (Copenh) 1961;36: 31-50. Phillips LS, Weiss LJ. Circulating growth factor studies in growth plate versus resting cartilage in vitro: tissue responsiveness. Endocrinology 1982;111:1255-62. D'Ercole AJ, Applewhite GT, Underwood LE. Evidence that somatomedin is synthesized by multiple tissues in the fetus. Dev Biol 1980;75:315-28. Atkison PR, Weidman ER, Bhaumick B, Bala RM. Release of somatomedin-like activity by cultured WI-38 human fibroblasts. Endocrinology 1980;106:2006-12. Clemmons DR, Underwood LE, Van Wyk JJ. Hormonal control of immunoreactive somatomedin production by cultured human fibroblasts. J Clin Invest 1981;67:10-9. Isaksson OGP, Jansson J-O, Gause/AM. Growth hormone stimulates longitudinal bone growth directly. Science 1982; 216:1237-9. Isgaard J, Nilsson A, Lindahl A, Jansson J-O, Isaksson OGP. Effects of local administration of GH and IGF-I on longitudinal bone growth in rats. Am J Physiol 1986;250(Endocrinol Metab 13):E367-72. Schlechter NL, Russell SM, Greenberg S, Spencer EM, Nicoll CS. A direct growth effect of growth hormone in rat hindlimb shown by arterial infusion. Am J Physiol 1986; 250(Endocrinol Metab 13):E231-5. Nilsson A, Isgaard J, Lindahl A, Peterson L, Isaksson O. Effects of unilateral arterial infusion of GH and IGF-I on tibial longitudinal bone growth in hypophysectomized rats. Calcif Tissue Int 1987;40:91-6 Nilsson A, Lindahl A, Ed6n S, Isaksson OGP. Demonstration of growth hormone receptors in cultured rat epiphyseal chondrocytes by specific binding of growth hormone and immunohistochemistry. J Endocrinol 1989;122:69-77. Isgaard J, M611er C, Isaksson OGP, Nilsson A, Mathews LS, Norstedt G. Regulation of insulin-like growth factor messenger ribonucleic acid in rat growth plate by growth hormone. Endocrinology 1988;122:1515-20. Nilsson A, Isgaard J, Lindahl A, Dahlstr6m A, Skottner A, Isaksson OGP. Regulation by growth hormone of number of chondrocytes containing IGF-I in rat growth plate. Science 1986;233:571-4. Ohlsson C, Nilsson A, Isaksson OGP, Lindahl A. Effect of growth hormone and insulin-like growth factor-I on DNA synthesis and matrix production in rat epiphyseal chondrocytes in monolayer culture. J Endocrinol 1992;133:291-300. Stracke H, Schulz A, Moeller D, Rossol S, Schatz H. Effect of growth hormone on osteoblasts and demonstration of somatomedin-C/IGF I in bone organ culture. Acta Endocrinol (Copenh) 1984;107:16-24. Ernst M, Froesch ER. Growth hormone dependent stimula-
438
Salmon and Burkhalter
tion of osteoblast-like cells in serum-free cultures via local synthesis of insulin-like growth factor I. Biochem Biophys Res Commun 1988;151:142-7. 32. Gluckman PD, Douglas RG, Ambler GR, Breier BH, Hodgkinson SC, Koea JB, et al. The endocrine role of insulin-like growth factor I. Acta Paediatr Scand (Suppl) 1991; 372:97-105. 33. Guler H-P, Zapf J, Scheiwiller E, Froesch ER. Recombinant human insulin-like growth factor I stimulates growth and has distinct effects on organ size in hypophysectomized rats. Proc Natl Acad Sci U S A 1988;85:4889-93. 34. Gosteli-Peter MA, Winterhalter KH, Schmid C, Froesch ER, Zapf J. Expression and regulation of insulin-like growth factor-I
J Lab Clin Med April 1997
(IGF-I) and IGF-binding protein messenger ribonucleic acid levels in tissues of hypophysectomized rats infused with IGF-I and growth hormone. Endocrinology 1994;135:2558-67. 35, Rechler MM. Insulin-like growth factor binding proteins. Vitam Horm 1993;47:1-114. 36. Ed6n S. Age- and sex-related differences in episodic growth hormone secretion in the rat. Endocrinology 1979;105:55560. 37. Terry LC, Martin JB. The effects of lateral hypothalamicmedial forebrain stimulation and somatostatin antiserum on pulsatile growth hormone secretion in freely behaving rats: evidence for a dual regulatory mechanism. Endocrinology 1981;109:622-7.