Nutrition and somatomedin XVI: Somatomedins and somatomedin inhibitors in fasted and refed rats

Nutrition

and Somatomedin

XVI: Somatomedins and Somatomedin in Fasted and Refed Rats

L.S. Phillips, S. Goldstein,

Inhibitors

and J.R. Gavin III

Nutritional deprivation is associated with poor growth and decreased levels of net circulating somatomedin activity, as measured by bioassay. Since somatomedin activity reflects the contributions both of somatomedins (which stimulate cartilage) and of somatomedin inhibitors (which antagonize the ability of the somatomedins to stimulate cartilage), we asked if changes in net somatomedin activity could involve progressive underlying alterations in levels of both somatomedins end somatomedin inhibitors. Groups of rats were killed during three days of fasting and 24 hours of refeeding. Fasting was associated with a rise in serum &hydroxybutyrate from 1.6 to 12.6 mmol/L after one day, followed by a decline to 4 mmol/L at three days. Somatomedins (low-MW) were separated from somatomedin inhibitors (high-MW) by gel permeation chromatography at acid pH on Sephadex G-50 and TSK-2000 HPLC. Somatomedins fell 35% after one day of fasting, and decreased to 59% below control levels after three days (P < .05). Somatomedins did not change with six hours of refeeding, but then rose rapidly, reaching control levels after 24 hours. Somatomedins were correlated with Inhibitors rose to 195% above change in weight (r = .41, P c .05), but not with glucose or &hydroxybutyrate. control-levels after one day of fasting, and continued to rise to 375% above control after three days (P -C .Ol ). In contrast to the delayed change in somatomedins with refeeding, there was an abrupt fall in inhibitors (41% below three-day fasted levels after six hours), returning to control levels after 24 hours. Levels of somatomedin inhibitors were correlated significantly with both &hydroxybutyrate and change in weight (both P < .05). Somatomedin responses appeared to be due in part to changes in circulating IGF-I, but inhibitor responses could not be attributed to changes in serum IGF binding activity. We conclude that circulating levels of both somatomedins and somatomedin inhibitors respond progressively to changes in nutritional status: apparent more rapid responsiveness of somatomedin inhibitors to refeeding may indicate noncoordinate regulation of production/clearance of somatomedins v inhibitors. B 1988 by Grune & Stratton, Inc.

W

HEN NUTRITION is inadequate, growth is poor despite growth hormone levels that may be normal’ or elevated,2 suggesting impairment of growth hormone action on skeletal elongation. This idea is supported by the inability of exogenous growth hormone (GH) to stimulate growth in malnourished children’ or to increase epiphyseal width in malnourished rats.4 Since the skeletal growth-promoting effects of GH appear to be mediated by the somatomedins, anabolic factors that stimulate cartilage proliferation directly, it is possible that impairment of GH action could involve defective somatomedin generation and/or action. Studies from our laboratory’ that circulating somatomeand by other workers637Indicate . din activity falls in fasted rats and rises with refeeding. The decrease in somatomedin activity cannot be prevented by GH administration, suggesting regulation by nutritional status. Net circulating somatomedin activity as measured by bioassay reflects the presence of both somatomedins and somatomedin inhibitors, factors that antagonize somatomedin action.839 In the present studies, we examined circulating somatomedins and somatomedin inhibitors during fasting and refeeding in rats in order to define dynamic changes in these factors and their potential contributions to altered somatomedin activity and growth in states of undernutrition. MATERIALS General Experimental

AND METHODS

Design

Rats were fasted for one, two, or three days, or fasted for three days and refed over 24 hours. Serum somatomedins and somatomedin inhibitors were separated on the basis of differences in molecular

Merabolism, Vol

37, No 3 (March), 1988: pp 209-216

size, measured separately by bioassay, and compared to indices of nutritional status.

Animals

and Animal

Serum

Male, Sprague-Dawley CD rats (Charles River Laboratories, Boston) weighing 108 to 144 g were randomized to seven groups of ten to 14 animals each. All had free access to water and were fed Purina Rat Chow ad libitum for three days prior to study. Rats were housed six per plastic cage on bedding that was changed after the beginning of fasting so that feces were removed from the cages. Food was withdrawn at 8AM and rats were fasted for 1, 2, or 3 days, or fasted 3 days and refed for 6, 12, or 24 hours. Control animals were allowed free access to food. At death, blood was obtained by aortic puncture under pentobarbital anesthesia, and serum frozen at - 20°C. Since individual serum volumes were insufficient for analysis, sera from two or more animals within each feeding group were combined to form five to seven pools for study of growth-related factors and metabolic intermediates. @-hydroxybutyrate, somatomedins, and somatomedin inhibitors were measured in the pooled samples, and glucose was determined in individual sera; body weight

From the Division of Endocrinology and Metabolism, Department of Medicine, Emory University School of Medicine. Atlanta, and the Metabolism Division, Department of Medicine, Washington University School of Medicine, St Louis. Supported in part by Research Grant Nos. DK-33475, DK-34785. AM-20579, and AM-05105 from the National Institutes of Health. Address reprint requests to Lawrence S. Phillips, MD, Department of Medicine, Emory University School of Medicine, 69 Butler Street, SE, Atlanta, GA 30303. Q 1988 by Grune & Stratton, Inc. 0026-0495/88/3703-0002$03.00/O

209

210

PHILLIPS, GOLDSTEIN, AND GAVIN

and glucose contributions to each pool were assessed on a weighted basis, according to the volumes of individual sera in each pool.

Assessment of Somatomedins, and Somatomedin Binding

Somatomedin

Inhibitors,

Circulating somatomedins and somatomedin inhibitors were determined operationally according to the presence of bioassayable somatomedin activity and somatomedin inhibitory activity in active fractions of sera separated by gel permeation chromatography; this approach assured biologic relevance. Levels of somatomedin-C/ IGF-1 and IGF binding protein activity in unfractionated sera were also evaluated via specific radioligand assays to determine potential contributions to alterations in bioassayable somatomedins and somatomedin inhibitors, respectively.

Fractionation of Pooled Serum Samples Total somatomedins in serum pools were separated by gel filtration chromatography on Sephadex G-50 (20 to 80 ~1,Pharmacia, Piscataway, NJ), as described previously.“’ Sera (1 mL) were applied to 1.2 x 100 cm columns and eluted with 1% formic acid at 4°C at a rate of 2.4 mL/h. Void volume (V,) was monitored by the appearance of protein (absorbance at 280 nm) and total volume (V,) was determined by the appearance of added Na’“’ in the eluate. The column effluent between the leading edges of the V, and V, was divided into eight equal fractions; somatomedins were detected between K,, 0.38 and 0.63. Somatomedin inhibitor(s) were separated by size exclusion HPLC” as described previously. Pooled sera (480 pL) were acidified to pH 3.0 with 20 ILLformic acid for 60 minutes at 2YC, and 200 PL applied to a Toyo Soda TSK 2000 SW size-exclusion separation column (7.5 x 600 mm) and eluted with 0.5 mol/L ammonium formate, pH 3.0, at 0.71 mL/min. Twelve equal fractions were collected between V, and the salt peak. When compared to normal rat serum, fasted rat serum exhibited increased inhibitory activity only in fractions of K,, 0.42 to 0.67 (fractions 6 to 8), and only this fraction was analyzed for inhibitor activity. All fractions were lyophilized and relyophilized either once (somatomedins) or twice (inhibitors) after addition of water. The residues were resuspended in bioassay buffer (2 mL for somatomedins, 400 rL for inhibitors) and stored at - 20°C until assay.

Assay of Somatomedins

and Inhibitors

Costa1 cartilage from hypophysectomized rats is very sensitive to specific stimulation by somatomedins, and in these studies was used to estimate the activity of circulating somatomedins and somatomedin inhibitors in experimental animals. Although parallel-line analysis is difficult because of the limited number of pieces of rat cartilage that can be included in a single assay, potency is well reflected by cartilage responses to single concentrations lying within the linear portion of the dose-response curve. s,‘*Thus, single concentrations of serum fractions were tested for somatomedin activity and somatomedin inhibitory activity in the rat cartilage bioassay system.13 Assay conditions for somatomedin and inhibitor activities were selected to maximize sensitivity” and minimize potential contributions from the presence of contamination.““’ The activity of somatomedins in serum fractions was evaluated at 1% (v/v) final concentration, and expressed relative to sulfate uptake by cartilage incubated in buffer alone (% of buffer). The activity of somatomedin inhibitors in serum fractions was determined at 4% (v/v) final concentration by the ability to blunt cartilage stimulation in the presence of 1% pooled normal rat serum (NRS), added to provide uniform stimulation by somatomedins. Inhibitory activity was expressed as the percentage of cartilage stimulation by NRS that

was blocked by the addition of test fractions: % NRS stimulation inhibited = loo x (NRS SO, uptake) - ([NRS + sample] SO4 uptake) (NRS SO4 uptake) - (buffer SO., uptake)

Somatomedin-C

RIA

One-microgram quantities of purified biosynthetic Sm-C/IGF-1 (Amgen Biological Co., Thousand Oaks, CA) were iodinated with lactoperoxidase’4 to specific activities of 100 to 280 &i/pg. Iodinated IGF-1 was filtered through Sephadex G-50 (20 to 80 p) 1.5 x 50 cm in 0.05 mol/L Tris-HCl, pH 7.6, and 0.25% bovine serum albumin (BSA; Sigma Chemical Co, St Louis). The fractions eluting as the monomeric form (K, 0.35 to 0.45) were used for assay. Aggregates were removed by repeat gel filtration after 2 to 3 weeks of standing at 4“C. Somatomedins in sera were extracted with acid-ethanol as described previously.” To 0.2 mL aliquots of serum were added 0.8 mL quantities of 87.5% ethanol/l2.5% 2N HCI. After standing for 30 minutes at room temperature, tubes were centrifuged at 1,860 x g for 30 minutes and 0.5 mL aliquots of supernatant transferred to 12 x 75 mm polypropylene tubes, and neutralized with 0.2 mL of 0.855 mol/L Tris base. An additional 1:lO dilution in assay buffer was made before RIA, which included blanks containing acid/ethanol/Tris in the same amounts as present in the serum extracts. The incubation mixture for the Sm-C/IGF-1 RIA contained 50 PL of neutralized acid-ethanol solution, 250 rL buffer (0.05 mol/L Tris-HCI, pH 7.6, 0.25% BSA), and 100 pL antihuman somatomedin-C antiserum (kindly provided by the National Hormone and Pituitary Program) diluted 1:4000 in assay buffer, with 100 pL ‘2SI-IGF-l in assay buffer (10 to 20,000 cpm). After overnight incubation at 4°C 1-mg rabbit IgG (Miles Laboratory, Elkhart, IN) in 0.1 mL assay buffer was added to each tube, followed by 0.6 mL 25% ice-cold polyethylene glycol 8,000 (Sigma). After standing 30 minutes at 4OC, tubes were centrifuged at 1,860 x g for 30 minutes 4OC, the pellet was washed with 1 mL of 12.5% polyethylene glycol and recentrifuged, and radioactivity in the pellets was measured in a Packard Autogamma counter. Purified biosynthetic IGF-I was used as the standard to define the dose-response displacement curve.

Somatomedin

Binding Protein Assay

Binding protein activity was determined as described previouslyn with minor modifications. Aliquots of whole serum (10 rL) from experimental animals were incubated overnight at 4“C with 10 to 15,000 cpm of ‘251-IGF-II (more sensitive to binding protein than IGF-1) in a total of 300 rL 0.025 mol/L Tris buffer, pH 7.6. Tubes were then supplemented with 600 pL of 0.5% dextran-coated activated charcoal (Sigma), vortexed, and centrifuged at 1,860 x g for 40 minutes at 4OC. Supernatants were aspirated and pellet radioactivity determined. Normal human serum (10 pL) was used as an assay control (-25% binding), whereas blanks consisted of labeled peptide in the presence of buffer alone. Results are expressed in terms of per cent tracer bound per 10 FL serum or equivalent.

Metabolic Measurements Glucose in individual sera was determined with a Beckman Model 2 glucose analyzer. @-Hydroxybutyrate in pooled serum samples was determined by the method of Olsen.”

Statistics Significance of differences between experimental groups was evaluated by two-tailed, unpaired t-tests and by ANOVA. Relationships between somatomedins, somatomedin inhibitors, and nutri-

SOMATOMEDINS/INHIBITORS

IN FASTED/REFED

211

RATS

Table 1. indices of Nutritional

Status in Fasted and Refed Rats

Days Fasted Control

Index

-11+2

Weight (g) Glucose (mg/dL) B~HB (mmol/L)

142 k 3 1.59 i- 0.03

Hours Refed

2

1 -20

6

3 + 1

54 + 2

81 k9

12.6 k 0.7

11.3 * 2.1

-3Ok

1

103 + 7 3.7 + 0.7

-15

24

12 t 2

-9

179 + 4

196 t

1.77 + 0.07

-8

+ 2

k 2

133 + 3

16

1.60 t 0.10

1.71 * 0.03

Weight and serum levels of metabolic fuels were determined in fasted and refed rats at death. For all animals, initial weight was 125 c 3 g. Changes

weight were

in

expressed relative to weight in the same animals at the beginning of the experiment; thus, the three-day fasted weight change reflects

contributions from fasted as well as refed animals (four groups), while the 24-hour refed weight change reflects a single group. Values are given as mean f SEM for five or six serum pools in each group.

tional indices were examined by linear regression and correlation analysis. RESULTS

Nutritional Indices in Fasted and Refed Rats Metabolic

indices during fasting and refeeding

are shown

in Table 1. Animals weighed 125 g at the beginning of the experiment, and controls gained 7 f 1 g/d (mean * SEM) over the next four days when fed ad libitum. At death, control animals had average serum glucose 142 mg/dL and P-hydroxybutyrate 1.59 mmol/L. When food was withheld, the rats lost between 8 and 11 g/d and averaged 24 5 3% below initial weight after three days of fasting. Serum glucose fell 62 T 1% below control levels after one day of fasting (P c .OOl), then rose gradually in the two- and three-day fasted animals. Serum P-hydroxybutyrate rose eightfold above control levels after one day of fasting (P < .OOl), changed little after a second day of fasting, then declined to 3.7 mmol/L after three days of fasting, approximately twice control levels. With refeeding, animals regained weight rapidly, 14 * 1 g above fasted levels after six hours, increasing gradually to 20 + 1 g above fasted levels after 24 hours. Presumably, such a substantial rise reflects intestinal contents and interstitial fluid retention analogous to refeeding edema,“.19 as well as tissue accretion. Serum glucose rose significantly above both control and three-day fasted levels with six hours of refeeding (P < .OOl), then declined after 24 hours to control levels. In contrast, serum /3-hydroxybutyrate fell to control levels with six hours of refeeding, and did not change thereafter. The relationship between nutrition, weight gain, and growth regulatory factors was studied in this model of fasting and refeeding. After separation from pooled sera by sizeexclusion chromatography, somatomedins and somatomedin inhibitors were examined by bioassay during progressive stages of food deprivation and subsequent refeeding.

tinued fasting, somatomedins fell more slowly, and were 44% and 59% below control at days 2 and 3, respectively. After three days of fasting, serum somatomedins no longer provided cartilage stimulation above buffer levels (P > .05). The fall in somatomedins with fasting was significant when examined by ANOVA (P < .05). After six hours of refeeding, somatomedins remained comparable to three-day fasted levels, and cartilage stimulation was not significantly above buffer (P > .l). Somatomedins rose with continued refeeding 6 1% and 18 1% above fasted levels after 12 and 24 hours, respectively; after 24 hours somatomedins in refed animals were 14% greater than levels in control animals (NS), and assay cartilage stimulation was significantly greater than buffer (P -C ,001). Correlations of Somatomedins

The relationship between fractionated serum somatomedins and nutrition-related indices is shown in Fig 2. No correlation was apparent between somatomedins and serum glucose or &hydroxybutyrate over the entire period of study. Since the potential relationship between somatomedins and P-hydroxybutyrate could have been confounded by the decline in P-hydroxybutyrate after three days of fasting, correlation analysis was repeated after the three-day values were omitted and found still to be insignificant (r = .21, P < .5). In contrast, levels of somatomedins correlated sig-

I

Somatomedins

After dissociation from carrier proteins by gel filtration on Sephadex G-50 at acid pH, significant stimulation by somatomedins in rat sera was found only at K,, 0.38 to 0.63. Somatomedins in sera from control animals (Fig 1) provided cartilage stimulation 165 + 47% above buffer (P-C .OOl), comparable to that found in our previous studies with this system.” After one day of deprivation, somatomedins fell 35% below control levels, and assay cartilage stimulation was still significantly above buffer levels (P < .OOl). With con-

With Nutritional Indices

Inr1 1

CONTROL

2

DAYS

i

3

FASTED

6

HOURS

12 REFED

Fig 1. Levels of somatomadins in fasted and refed rats. Sara from animals at death were fractionated on Saphadex G-SO at pH 2.4, and somatomadin activity in fractions at K, 0.38 to 0.63 was determined by rat costel cartilage bioassay according to the ability to promote cartilage ?30, uptake. Sulfate uptake is expressed as % above buffer levels. Mean + SEM for five or six serum pools at each point.

212

PHILLIPS, GOLDSTEIN, AND GAVIN

&

.

. r-o.025

%

q

Y 5

pd4s

300. a

200.

.

l

r--.192

p-w

300-

300

200.. : 200

l

R ,OO8 3

0.

F0.41 P
l

L

l **Zt

’

.v

100.;

:,. ‘*@a

=

,:

O_;mm

l

it! -LOOI 0

100

8. .

0

-LOOk__

200

0

5

10

-100

15

rnM

v/d1

‘T

.s

;

”

-40

Correlations Between Somatomedin Inhibitors and Metabolic Indices

Relationships between levels of somatomedin inhibitors and nutritional indices are shown in Fig 4. There was no apparent relation to serum glucose in animals at death. Although both serum somatomedin inhibitor levels and

h 3

0~~s

FASTED

6

-20

0

9

Inhibitors

CONTROL

.

.

Figure 3 shows fluctuations in circulating somatomedin inhibitors during fasting and refeeding. Experimental sera were fractionated at pH 3.0 by size-exclusion HPLC, with elution of somatomedin inhibitors at K, 0.42 to 0.67. Somatomedin inhibitors in control sera blunted cartilage stimulation by somatomedins in normal rat serum by 17% (NS). After one day of fasting, somatomedin inhibitors rose, blunting cartilage stimulation by 50 + 18% (P < .Ol v buffer). With continued deprivation of food, there was little further increase in somatomedin inhibitors (55 f 12% and 63 k 11% after two and three days, respectively, both P < .Ol v levels in control animals). With refeeding, there was a prompt fall in somatomedin inhibitors, blunting cartilage stimulation by 37 k 16% after six hours, no longer above buffer levels, with a continued fall to control levels after 24 hours of refeeding.

r

v’

I/

nificantly with change in weight (r = .41, P < .05). Lean body mass was not measured. Somatomedin

v : vv

.

.;a

100

I

12

HOURS

’ v

Fig 2. Correlation between levels of som8tomedins (expressed as in Fig 1). and metabolic and nutritional indices 8t the time of death in fasted and refed rsts. Glucose end &hydroxybutyrate were determined. 8nd weight at death ~8s expressed relative to initial weight.

serum P-hydroxybutyrate tended to rise with fasting and fall with refeeding, correlations between these two variables were positive but not significant when all experimental groups were included. Since substrate depletion may have contributed to a decrease in P-hydroxybutyrate after three days of fasting, the analysis was repeated after exclusion of this group, revealing a significant correlation, consistent with observations in diabetes” (r = .37, P < .05). Levels of somalomedin inhibitors were also correlated inversely with change in weight in our experimental animals (P < .05). Interrelationships of Somatomedins and Somatomedin Inhibitors

Alterations in circulating somatomedins and somatomedin inhibitors during fasting and refeeding are shown combined in Fig 5. For purposes of comparison, values during fasting and refeeding were expressed as percent of mean values in control animals. Marked changes in both somatomedins and somatomedin inhibitors occurred with one day of fasting. Somatomedins fell by 35% while somatomedin inhibitors rose by almost 200%. This pattern persisted as long as fasting and weight loss continued; after three days of fasting, somatomedins were 41% of control levels, and inhibitors were 375% of control levels. In our model, six hours of refeeding produced rapid increases in body weight and serum glucose, with a return of @-hydroxybutyrate to control levels. Despite a weight gain of 14 g, serum somatomedins did not change over this period. In contrast, somatomedin inhibitors fell by 41% compared to values in animals fasted for three days. With continued refeeding, serum somatomedins rose and EOHB

50

24

a

REFED

Levels of somatomedin inhibitors in fasted 8nd refed Fig 3. rats. At death, ser8 were fractionated by size exclusion HPLC on TSK-2000 SW et pH 3.0. and SOmatOmedininhibitor activity 8t K, 0.42 to 0.67 w8s determined as the sbility of samples to decrease cartilage stimulation provided by normal rat serum. Inhibition was expressed 8s percent decrease in cartilage SO, uptake, mean r SEM for five or six serum pools in each group.

-50hEInM

-50

c

-40

Fig 4. Correlations between levels of sorn8tomedin inhibitors 8nd metabolic and nutritional indices at death in fssted and refed rats, expressed 8s in Fig 2.

SOMATOMEDINS/INHIBITORS

IN FASTED/REFED

213

RATS

BIOLOGICALACTIVITy,% OF CONTROL 200

Table 2.

500

Levels of Somatomedin-C/IGF-I in Experimental

and IGF Binding Protein

Animals

Sm-C/lGF-I Animal Groups

CO

04 0

1

4 3 REFED-)

2

FAST-__--_-_--_-)

and IGF

Specific radioligand assay approaches permitted assessment of total somatomedin-C/IGF-I determined by RIA after acid-ethanol extraction of whole serum, and binding protein activity according to the ability of serum samples to bind ‘2sI-IGF-II. As shown in Table 2, normal rats had somatomedin-C/IGF-I levels of 366 ng/ml, which decreased 64% after one day and 82% after three days of fasting (both P < .02); with refeeding, levels rose significantly above but were still 60% below

I

0

loot

0

Ll

60 40 80 t 20 7 0

04 X

-20

I

Ll 0

‘I

-40

.~--100

t---r--t--_-r 0

; 100

200

s

;

300

:

400

SOMATOMEDINS SO4 UPTAKE, % ABOVE BUFFER Fig 6. inhibitors,

Correlations between including all groups.

45.1

somatomedins

2 1.3

Fasting 1 d

131 t- 20*t

45.8

+ 0.6b

Fasting 2 d

66 k 23”

39.2

+ 1.5’

Fasting 3 d

40 + 4’

39.6

+ 1.0’

Refeeding 6 h

72 k 1.t

34.2

+ 0.3”.b

Refeeding 12 h

124 + 36*

34.9

+ 2.6”

Refeeding 24 h

148 f

43.1

f 2.0

14.t

randomly selected serum pools per group. Binding activity is

expressed in terms of per cent added ‘%GF-II

serum somatomedin inhibitors continued to fall, both returning to control levels after 24 hours. Due in part to discrepant temporal fluctuations, somatomedins and somatomedin inhibitors were not correlated during the entire period of study (Fig 6, r = .25, P =_ .l). However, correlations also were not significant during either fasting or refeeding alone (r = -.25 and r = -.30, respectively, both P > .l).

fasted animals

366 k 62

Values are given as mean + SEM. Determinations were performed in

Fig 5. Levels of somatomedins and inhibitors in fasted and refed rats. In order to show relative changes in factors that promote and antagonize promotion of cartilage stimulation. respectively. somatomedins and inhibitors in sera from animals with altered nutrition (from Figs 1 and 3) were expressed as percent of values in fed control animals. Shown are mean values only for five or six serum pools in each group; SEM corresponds to values in Figs 1 and 3.

values in three-day

(%)

Normal rats

three

TIME (DAYS)

Circulating Somatomedin-C/IGF-I Binding Protein Activity

IGF Binding

fng/mL)

and somatomedin

bound per 10 pL serum

added. *Significantly different from normal (control) levels. tsignificantly different from three-day fasted levels by ANOVA, P .c .05.

normal levels after 24 hours (P < .05). Serum IGF binding activity decreased from 45% in normal rats to 40% in three-day fasted animals (P < .OS), continued to fall despite six hours of refeeding, and was partially restored with 24 hours of refeeding. DISCUSSION

Current understanding of the relationship between somatomedins and nutrition began with observations by Daughaday and Kipnis*’ and Salmon** that fasting in rats produces a decrease in the ability of their serum to stimulate sulfate uptake by cartilage in vitro, confirmed in more detailed studies with this model.’ Similar decreases in bioassayable somatomedin activity were found in malnourished children,23-2s as well as in patients with anorexia nervosa26 and coeliac disease.‘? Attempts to model human malnutrition by altering diet composition in rats indicated that protein is particularly important for maintaining the level of somatomedin activity.4*6.7,28 We’ and others4 have observed that the malnutrition-induced decrease in somatomedin activity is refractory to GH administration, consistent with high levels of GH in malnourished children with low serum somatomedin activity’**; this indicates that the decline in somatomedin activity cannot be attributed to inadequate GH secretion, and must be due to nutrition per se. Studies of serum somatomedin activity indicate that stimulation of cartilage in vitro reflects the presence both of somatomedins (bound to carrier proteins) and somatomedin inhibitors, which antagonize somatomedin action.8.v The predominant somatomedin inhibitors in NRS have apparent MW 20 to 40,000 (K,, 0.42 to 0.67 corresponds to MW 12 to 42,000, logarithmic mean around 22,400), consistent with apparent MW -30,000 of an inhibitor isolated from human plasma fractions*’ and intermediate between that of carrierbound somatomedins (- 150,000), and free hormones (-7,500). Whole serum and serum fractions containing a relative excess of inhibitor activity can antagonize both somatomedin and insulin action on cartilage, fat, and muscle.” The opposing properties of somatomedins and somatomedin inhibitors suggest that altered somatomedin/

214

inhibitor balance could contribute to the depressed circulating somatomedin activity and altered tissue anabolic/ catabolic balance associated with malnutrition, and preliminary methodologic studies ‘O*”have indicated that starvation produces a fall in somatomedins with a rise in inhibitors. The present studies show that circulating somatomedin inhibitors and somatomedins both respond progressively to alterations in nutritional status. Fasted animals exhibited a prompt increase in somatomedin inhibitors, together with a more gradual fall in somatomedins. These changes were reversed when the animals were refed. With six hours of refeeding, there was a striking fall in somatomedin inhibitors but little change in somatomedins, despite a gain in weight and normalization of serum ,&hydroxybutyrate. With continued refeeding, both somatomedins and somatomedin inhibitors returned to control levels. With both fasting and refeeding, the somatomedin inhibitors appeared to respond more rapidly than the somatomedins. The decline in bioassayable somatomedins associated with food withdrawal in the present study is consistent with reported changes in somatomedin A3’,‘* and somatomedin C.33 (Since somatomedin A appears to be identical to IGF1,94 the somatomedin A radioreceptor assay may reflect changes in somatomedin C/IGF-1). Both somatomedin C and somatomedin A fell 50% to 75% in rats fasted for one day, with a modest further decline after two additional days of fasting. With refeeding, there was little increase in somatomedin A after six hours and only partial restoration after 24 hours, consistent with studies of somatomedin-C. RIA measurement of somatomedin-C/IGF-1 in our experimental sera confirmed the decrease in somatomedin-C/ IGF-1 with fasting, with only partial restoration upon refeeding. Thus, normalization of bioassayable somatomedins with 24 hours of refeeding in the present studies must be attributed to bioassay sensitivity to both IGF-1 and IGF-2, as well as to other biologically important factors which have not yet Although Merimee et al” noted been fully characterized. 3s*36 that IGF-2 fell less than IGF-1 with short-term fasting in normal subjects, there have been no comparisons with refeeding. It should be noted that refeeding-associated normalization of bioassayable somatomedins is accompanied by a marked increase in cartilage growth activity, presumably a reflection of somatomedin action in vivo.5 The lack of rise in fractionated somatomedins after six hours of refeeding contrasts with our findings of a significant elevation in net bioassayable serum somatomedin activity over this interval, and suggests that such an increase in net somatomedin activity might reflect a fall in inhibitors rather than a rise in somatomedins. To date, we are aware of no other studies of circulating somatomedin inhibitors during homeostatic perturbation imposed by fasting and reversed with refeeding. While net inhibitory activity in whole serum has been noted in models of hypopituitarism3* and uncontrolled diabetes3g as well as starvation,40 there has been little examination of circulating inhibitors after fractionation to separate these factors from somatomedins, aside from our delineation of inhibitors as elevated in conditions of uremia,4’ glucocorticoid excess,42

PHILLIPS, GOLDSTEIN, AND GAVIN

and ketoacidosis.” In the present studies, the prompt inhibitor rise with fasting and rapid fall with refeeding indicate that inhibitors very likely contribute to inverse changes in net circulating somatomedin activity seen with fasting and refeeding5; underlying mechanisms (synthesis/degradation, etc) are unknown. It should be noted that the observed fluctuations in somatomedin inhibitors are opposite from changes in somatomedin binding proteins, which appear to be depressed in fasted animals.43 In our study, somatomedin inhibitors were quantified on the basis on antagonism of cartilage stimulation by whole serum, rather than isolated somatomedins; since whole serum contains binding proteins, the assay should be relatively insensitive to binding proteins in serum inhibitor fractions. We found a decrease in serum IGF-II binding activity with fasting, partially restored (after a lag) with refeeding. These observations of IGF binding compare well with determinations of mixed IGF-I/IGF-II binding in fasted animals as studied by Miller et al43 and indicate that changes in bioassayable somatomedin inhibitors are unlikely to be due to underlying changes in binding proteins. Nutrition-associated changes in circulating somatomedins and somatomedin inhibitors may reflect altered release of these factors by the liver. This organ both contains and releases factors that have somatomedin activity and somatomedin inhibitory activity,“-48 and Vassilopoulou-Sellin et al49 have shown that net somatomedin activity (reflecting both somatomedins and somatomedin inhibitors) in liver extracts and liver perfusates declines more rapidly than circulating somatomedin activity during fasting or streptozotocin-induced diabetes in rats. In combination, such findings buttress the notion of hepatic regulation. The asynchrony between somatomedins and somatomedin inhibitors observed with fasting and refeeding in the present study is consistent with findings in models of diabetes” or glucocorticoid excess:* suggesting that lack of correlation cannot be attributed simply to bioassay variability and more likely reflects differences in underlying control processes. Given the anabolic actions of somatomedins and the antianabolic actions of the somatomedin inhibitors, the net impact of the fluctuations observed in the present studies might be to promote conservation of metabolic fuels during fasting, with restoration of anabolic fuel utilization during refeeding.j’ In this way, regulated alterations in both factors could contribute to homeostasis during periods of limited availability of nutrients. The potential role of somatomedins in such homeostatic responses, and the apparent sensitivity of somatomedins to changes in nutritional status,5’-53has led to interest in the utility of somatomedin measurements as markers of nutritional status; the present studies indicate that determinations of levels of circulating somatomedin inhibitors may also be useful in nutritional assessment. ACKNOWLEDGMENT

We gratefully acknowledge the technical assistance of J. Lamb, L. Stivaletta, C.K. Matheson, K.M. Keery, and M. Kapadia. We thank Aaliyah manuscript.

Salim

for assistance

in the

preparation

of the

SOMATOMEDINS/INHIBITORS

IN FASTED/REFED

215

RATS

REFERENCES

I Raghuramalu N, Jaya Rao KS: Growth hormone secretion in protein-calorie malnutrition. J Clin Endocrinol Metab 38:176-180, 1974 2. Beas F, Muzzo S: Growth hormone and malnutrition: The Chilean experience, in Gardner LI, Amacher P (eds): Endocrine Aspects of Malnutrition. Santa Ynez, CA, Kroc Foundation, 1973, pp I-18 3. Hadden DR, Rustishauser IHE: hormone in kwashiorkor and marasumus. 1967

Effect of human growth Arch Dis Child 42:29-33,

4. Shapiro B, Pimstone B: The histology and in vivo sulphate uptake of epiphyseal cartilage in protein-depleted rats. Br J Nutr 38:513-518, 1977 5. Phillips LS, Young HS: Nutrition and Somatomedin. I. Effect of fasting and refeeding on serum somatomedin activity and cartilage growth activity in rats. Endocrinology 99:304-3 14, 1976 6. Price DA, Wit JM, van Buul-Offers S, et al: Serum somatomedin activity and cartilage metabolism in acutely fasted, chronically malnourished, and refed rats. Endocrinology 105:851-861, 1979 7. Reeves RD, Dickinson composition on somatomedin 109:613-620. 1979

L. Lee J, et al: Effects of dietary activity in growing rats. J Nutr

8. Phillips LS, Belosky DC, Young HS, et al: Nutrition and somatomedin VI. Somatomedin activity and somatomedin inhibitory activity in serum from normal and diabetic rats. Endocrinology 104:1519-1524. 1979 9. Salmon WD Jr, Holliday LA, Burkhalter VJ: Partial characterization of somatomedin inhibitor in starved rat serum. Endocrinology II 2:360-370, 1983 IO. Phillips LS, Bajaj VR, Fusco AC, et al: Nutrition and Somatomedin. XII. Fractionation of somatomedins and somatomedin inhibitors in normal and diabetic rats. Int J Biochem 17:597-603. 1985 Il. Goldstein S. Stivaletta LA, Phillips LS: Separation of somatomedins and somatomedin inhibitors by size exclusion highperformance liquid chromatography. J Chromatogr 339:388-393, 1985 12. Phillips LS, Bajaj VR, Fusco AC, et al: Nutrition and somatomedin. Xl. Studies of circulating somatomedin inhibitors in rats with streptozotocin-induced diabetes. Diabetes 32:117-l 125, 1983 13. Phillips LS, Belosky DC, Reichard LA: Nutrition and somatomedin. V. Action and measurement of somatomedin inhibitor(s) in serum from diabetic rats. Endocrinology 104:1513-1518, 1979 14. Tait JF. Weinman SA, Bradshaw RA: Dissociation kinetics of ‘*‘I-nerve growth factor from cell surface receptors. Acceleration by unlabeled ligand and its relationship to negative cooperativity. J Biol Chem 256:I 1086-l 1092, 1981 15. Daughaday WH, Mariz IK. Blethen SL: Inhibition of access of bound somatomedin to membrane receptor and immunobinding sites: A comparison of radioreceptor and radioimmunoassay of somatomedin in native and acid-ethanol-extracted serum. J Clin Endocrinol Metab 5 I :78 1-788, 1980 16. Cohen DA, Blethen SL: Somatomedin binding proteins in GH-deficient children with normal plasma somatomedin levels. J Clin Endocrinol Metab 56:461-469, 1983 17. Olsen C: An enzymatic fluorimetric micromethod for the determination of acetoacetate, B-hydroxybutyrate, pyruvate and lactate. Clin Chim Acta 33:293-300, 1971 18. Yang MU, Van ltallie TB: Composition of weight lost during short-term weight reduction: Metabolic responses of obese subjects

to starvation and low-calorie Clin Invest 58:722-730, 1976

hetogenic

and nonketogenic

diets. J

19. Gamble JL, Ross GS, Tisdall FE: The metabolism base during fasting. J Biol Chem 57:633-695, 1923

of fixed

20. Phillips LS, Fusco AC, Unterman TG: Nutrition and somatomedin. XIV. Altered levels of somatomedins and somatomedin inhibitors in rats with streptozotocin-induced diabetes. Metabolism 34:765-770, 1985 21. Daughaday WH, Kipnis DM: The growth promoting anti-insulin actions of growth hormone. Recent Prog Horm 22:49-99, I966

and Res

22. Salmon WD Jr: Investigation with a partially purified preparation of serum sulfation factor: Lack of specificity for cartilage sulfation, in Pecile A, Muller EE (eds): Proceedings 2nd Int Symp on Growth Hormone. Amsterdam, Excerpta Medica, Int Congr Ser 244:180-191, 1972 23. Grant DB, Hambley J, Becker D, et al: Reduced sulphation factor in undernourished children. Arch Dis Child 48:596-600, 1973 24. Van den Brande JL, Du Caju MVL: Plasma somatomedin activity in children with growth disturbances, in Raiti S (ed): Advances in Human Growth Hormone Research. Washington DC, DHEW Publication No. (NIH) 74-612, 1974, pp 98-126 25. Hintz RL, Suskind R, Amatayakul K, et al: Plasma somatomedin and growth hormone values in children with proteincalorie malnutrition. J Pediatr 92: 153-l 56, 1978 26. Rappaport R, Prevot C, Czernichow P: Somatomedin activity and growth hormone secretion I. Changes related to body weight in anorexia nervosa. Acta Paediatr Stand 69:37-41, 1980 27. Lecornu M, David L, Francois R: Low serum somatomedin activity in celiac disease. Helv Paediatr Acta 33:509-516, 1978 28. Phillips LS, Orawski AT, Belosky DC: Somatomedin and nutrition. IV. Regulation of somatomedin activity and growth cartilage activity by quantity and composition of diet in rats. Endocrinology 103:121-127, 1978 29. Ahmad M, Goldstein S, Shirley M, et al: Purification of somatomedin inhibitor from human plasma. Proceedings of the 69th Annual Meeting of the Endocrine Society, Indianapolis, 101 (abstr) 30. Phillips LS, Scholz TD: Nutrition and somatomedin. IX. Blunting of insulin-like activity by inhibitor in diabetic rat serum. Diabetes 31:97-106, 1982 31. Takano K, Hizuka N, Kawai K, et al: Effect of growth hormone and nutrition on the level of somatomedin A in the rat. Acta Endocrinol87:485-494, 1978 32. Takano K, Hizuka N, Shizume K, et al: ElTect of nutrition on growth and somatomedin A levels in the rat. Acta Endocrinol 94:321-326, 1980 33. Maes M, Underwood LE, Ketelslegers JM: Plasma somatomedin-C in fasted and refed rats: close relationship with changes in liver somatogenic but not lactogenic binding sites. J Endocrinol 97~243-252, 1983 34. Enberg G, Carlquist M, Jornvall H, et al: The characterization of somatomedin A, isolated by microcomputer-controlled chromatography, reveals an apparent identity to insulin-growth factor I. Eur J Biochem 143:117-124, 1984 35. Herington AC, Kuffer AD: Insulin-like growth factor characteristics of an acidic non-suppressible insulin-like activity. Biochem J 223:89-96, 1984 36. Kuffer AD, Herington AC: Proteolytic conversion of insulinlike growth factors to an acidic form(s). Biochem J 223:97-103, 1984

216

37. Merimee T, Zapf J, Froesch ER: Insulin-like growth factors in the fed and fasted states. J Clin Endocrinol Metab 55999-1002, 1982 38. Salmon WD Jr: Effects of somatomedin on cartilage metabolism: Further observations on an inhibitory serum factor, in Raiti S (ed): Advances in Human Growth Hormone Research. Washington, DC, DHEW Pub. No. (NIH) 74-612,1974, pp 76-97 39. Phillips LS, Young HS: 1976 Nutrition and somatomedin. II. Serum somatomedin activity and cartilage growth activity in streptozotocin-diabetic rats. Diabetes 25516-527, 1976 40. Salmon WD Jr: Interaction of somatomedin and a peptide inhibitor in serum of hypophysectomized and starved, pituitaryintact rats. Adv Metab Dis 8:183-199, 1975 41. Phillips LS, Fusco AC, Unterman TG, et al: Somatomedin inhibitor in uremia. J Clin Endocrinol Metab 59:764-772, 1984 42. Unterman TG, Phillips LS: Glucocorticoid effects on somatomedins and somatomedin inhibitors. J Clin Endocrinol Metab 61:618-626, 1985 43. Miller LL, Schalch DS, Draznin B: Role of the liver in regulating somatomedin activity: Effects of streptozotocin diabetes and starvation on the synthesis and release of insulin-like growth factor and its carrier protein by the isolated perfused rat liver. Endocrinology 108:1265-1271, 1981 44. Vassilopoulou-Sellin R, Phillips LS, Reichard LA: Nutrition and somatomedin. VII. Regulation of somatomedin activity by the perfused rat liver. Endocrinology 106:260-267, 1979 45. Vassilopoulou-Sellin R, Oyedeji CO, Samaan NA: Extrac-

PHILLIPS, GOLDSTEIN, AND GAVIN

tion of cartilage sulfation inhibitors and somatomedins from rat liver. Endocrinology 114:567-581, 1984 46. Schalch DS, Heinrich UE, Draznin B, et al: Role of the liver in regulating somatomedin activity: Hormonal effects on the synthesis and release of insulin-like growth factor and its carrier protein by the isolated perfused rat liver. Endocrinology 104:1143-1151, 1979 47. Rechler MM, Eisens HJ, Higa OZ, et al: Characterization of a somatomedin (insulin-like growth factor) synthesized by fetal rat liver organ cultures. J Biol Chem 254:7942-7950, 1979 48. Binoux M, Hossenlopp P, Lassarre C, et al: Somatomedin production by rat liver in organ culture. I. Validity of the technique. Influence of the released material on cartilage sulphation. Effects of growth hormone and insulin. Acta Endocrinol93:73-82, 1980 49. Vassilopoulou-Sellin R, Phillips LS, Oyedeji CO, et al: Metabolic regulation of somatomedin activity in rat liver. J Endocrinol 101:257-261, 1984 50. Phillips LS: Nutrition, somatomedins, and the brain. Metabolism 35:78-87, 1986 51. Clemmons DR, Klibanski, Underwood LE, et al: Reduction of plasma immunoreactive somatomedin C during fasting in humans. J Clin Endocrinol Metab 53:1247-1250,198l 52. Unterman TG, Vazquez RM, Salas AJ, et al: Nutrition and somatomedin. XIII. Usefulness of somatomedin-C in nutritional assessment. Am J Med 78:228-234, 1985 53. Clemmons DR, Underwood LE, Brown RO, et al: Use of plasma somatomedin-C/insulin-like growth factor I measurements to monitor the response to nutritional repletion in malnourished patients. Am J Clin Nutr 41:191-198, 1985