Abstract
We have recently described a competitive binding assay for rat insulin-like growth factor-binding protein-3 (IGFBP-3) based on the ability of IGFBP-3 to form a ternary complex with the acid-labile subunit (ALS) in the presence of IGF-I. Using this assay we studied groups of male (n = 6) and female rats (n = 6) at 20, 30, 40, 50, 60, 80, and 130 days of age. Nonfasting serum levels of IGFBP-3 were compared with those of total (extractable) IGF-I (tIGF-I) and ALS as well as IGFBP-3 determined by ligand blotting. Additionally, we studied the relationship between ultrafiltered free IGF-I (fIGF-I) and immunoassayable IGFBP-1.
IGFBP-3 was dependent on age only (P < 0.0001), but tended to be higher in males than in females (P= 0.06); between 20–130 days levels increased from 6.5 ± 1.7 to 73.6 ± 7.2 nmol/liter in males and from 5.4 ± 1.6 to 51.3 ± 8.0 nmol/liter in females. IGFBP-3 correlated positively with tIGF-I (r = 0.90; P < 0.0001), ALS (r = 0.92; P < 0.0001), and IGFBP-3, as determined by ligand blotting (r = 0.88; P < 0.0001). The molar ratio of IGFBP-3 to tIGF-I increased from 0.23± 0.04 to 0.76 ± 0.04 (P < 0.0001) without any sex dependence. An age- and sex-dependent decrease in IGFBP-1 was observed (P < 0.0001), from 10.9 ± 2.5 to 1.2 ± 0.2 nmol/liter in females and from 8.9 ± 0.7 to 0.2 ± 0.04 nmol/liter in males. Free IGF-I (fIGF-I) increased with age (from 0.7 ± 0.2 to 7.1 ± 0.5 nmol/liter; P < 0.0001), and levels were inversely correlated with IGFBP-1 (r = −0.80; P < 0.0001). In young rats, IGFBP-1 circulated in a 10-fold molar excess over the level of fIGF-I, whereas in older rats, fIGF-I exceeded IGFBP-1 by an average of 9-fold in females and by up to almost 60-fold in males.
We conclude that in rats 1) IGFBP-3 and fIGF-I are strongly age dependent; 2) IGFBP-3 correlates positively with ALS and tIGF-I; and 3) fIGF-I and IGFBP-1 are inversely correlated. This is in accordance with clinical findings. However, in humans the adult level of fIGF-I rarely exceeds 0.3 nmol/liter, and IGFBP-1 usually circulates in excess of fIGF-I. Thus, our results also imply species differences in the IGF systems of humans and rats.
THE MAJORITY of circulating insulin-like growth factor I (IGF-I) is carried within a ternary complex together with IGF-binding protein-3 (IGFBP-3) and the non-IGF-binding acid-labile subunit (ALS) (1). Formation of the ternary complex extends the half-life of IGF-I from approximately 10 min in its free form to 18 h, thereby stabilizing and maintaining plasma IGF-I levels (2–4) and preventing the egress of IGF-I from the circulation (5, 6). Thus, ternary complex formation is believed to play a pivotal role in the regulation of IGF-I bioavailability in vivo.
To study the regulation of the ternary complex, specific immunoassays for human (h) IGFBP-3, hALS, and rat (r) ALS were developed several years ago (7–9), whereas no quantitative assay for rIGFBP-3 has been available. Previously, determinations of rIGFBP-3 have been performed using ligand blotting, which is at best semiquantitative (10). We have, however, recently developed a novel competitive binding assay for determination of intact rIGFBP-3 based on the ability of IGFBP-3 to form a ternary complex with ALS in the presence, but not in the absence, of IGF-I (11).
Only a minor fraction of the circulating IGF-I is bound to IGFBP-1. This binding protein has, however, received considerable attention, because it is the only IGFBP with a rapid dynamic regulation (12). IGFBP-1 is primarily regulated by insulin, which reduces the hepatic gene and messenger RNA expression of IGFBP-1 and suppresses the circulating IGFBP-1 levels (12). Conversely, substrate deprivation stimulates IGFBP-1 production by a cAMP-dependent mechanism (13). These observations have led to the hypothesis that IGFBP-1 regulates IGF-I bioactivity in vivo by adjusting the level of free IGF-I (fIGF-I) according to the actual fuel supply. In support of this, we found an inverse correlation between circulating levels of fIGF-I and IGFBP-1 in several clinical cross-sectional and longitudinal studies in fasting and nonfasting situations (14–17). No such data are available in rats.
In the present study we investigated the age- and sex-related changes in the circulating levels of fIGF-I, total (extractable) IGF-I (tIGF-I), IGFBP-1, IGFBP-3, and ALS in normal Wistar rats. We provide evidence that the circulating IGF system shows similarities as well as dissimilarities when comparing rats and humans.
Materials and Methods
Animals
Male and female Wistar rats were used. They were housed two or three per cage in a room with a 12-h (0630–1830 h) artificial light cycle, a temperature of 21 ± 2 C, and a humidity of 55 ± 2%, with free access to tap water and standard fodder (Altromin, Lage, Germany).
Protocol
The animals were randomized to seven groups, containing six male and female rats each, and were killed at the ages of 20, 30, 40, 50, 60, 80, and 130 days. All animals were studied in the morning in the nonfasted state. This decision was made because young rats have a higher relative fodder consumption than old rats, and fasting overnight might therefore affect young and old rats differently, causing a less reliable basis for comparison. On the day of the study, animals were weighed, anesthetized (pentobarbital; 50 mg/kg, ip), and bled through the retrobulbar venous plexus using heparinized capillary tubes. All blood samples were collected within 15 min after anesthesia. Serum samples were kept at −20 C for later analysis.
Assays
All samples were analyzed in duplicate within the same assay unless otherwise stated. IGF-I was determined by an in-house IGF-I RIA using a polyclonal rabbit antibody (Nichols Institute Diagnostics, San Capistrano, CA) and recombinant hIGF-I as standard (Amgen Biologicals, Thousand Oaks, CA). Monoiodinated hIGF-I ([125I]Tyr31-hIGF-I) was obtained from Novo Nordisk (Bagsværd, Denmark). Total IGF-I was determined in acid-ethanol serum extracts with within- and between-assay coefficients of variation (CVs) averaging 5% and 10%, respectively (18). Free IGF-I (fIGF-I) was determined as previously described, using centrifugal ultrafiltration under conditions approaching those in vivo (19, 20). In brief, Amicon YMT 30 membranes and MPS-1 supporting devices were used (Amicon Division, Beverly, MA). Before centrifugation, serum samples were adjusted to pH 7.4 by gassing with CO2, after which aliquots of 400 μl were applied to the membranes, incubated (30 min at 37 C), and centrifuged (1500 rpm at 37 C; model Rotixa/RP, Hettich Zentrifugen, Tuttlingen, Germany). Serum fIGF-I was determined directly in the ultrafiltrates with a within-assay CV that, including ultrafiltration and RIA, averaged 26%.
Rat IGFBP-3 was determined by a novel competitive binding assay based on the ability of rIGFBP-3 to form a ternary complex with ALS in the presence of IGF-I (11). A defined amount of hALS was bound to Maxisorb test tubes (Roskilde, Denmark) precoated with hALS antibody. The assay depends on competition between a covalent complex of[ 125I]hIGF-I and hIGFBP-3, added as tracer, and hIGFBP-3 or rIGFBP-3 in standards and test samples for binding to the immobilized hALS. Purified natural hIGFBP-3 served as the standard. hIGFBP-3 and rIGFBP-3 were able to compete for tracer binding in the presence, but not in the absence, of IGF-I, and they diluted in parallel. Before the assay, rat serum samples were acidified to denature endogenous ALS. Samples were analyzed in two assays with within- and between-assay CVs of 13% and 17%, respectively.
rIGFBP-3 was also determined using SDS-PAGE and ligand blotting analysis, which were performed according to the method of Hossenlopp et al. (10). Autoradiographs of ligand blots were scanned using a laser densitometer (model CS 90001PC, Shimadzu Europe, Duisburg, Germany), and the relative densities of the bands were expressed in arbitrary absorbency units per mm2. Serum samples from the various groups were equally distributed on each gel.
rALS and rIGFBP-1 were determined by specific RIA methods as previously described (9, 21).
Insulin was determined by RIA (Novo Nordisk) using iodinated recombinant human insulin ([125I]recombinant human insulin) as tracer, purified rat insulin as standard, and a polyclonal guinea pig antibody. Samples were analyzed in triplicate in two assays.
To facilitate a direct comparison between the different peptide concentrations, all levels are given in nanomoles per liter, using the following molecular masses: ALS, 85 kDa; IGFBP-3, 42 kDa; IGFBP-1, 25 kDa; and IGF-I, 7.7 kDa.
Statistics
The effect of sex was investigated using two-way ANOVA, and the effect of age was studied using one-way ANOVA. Linear regression analyses were used to assess the relationship between the measured variables. All data were log transformed before analysis to improve normality and variance homogeneity. P < 0.05 was considered statistically significant. All values are given as the mean ± sem
Results
Changes in body weights were highly dependent on sex and age (P < 0.0001; Fig. 1). In males, body weight continued to increase during the observation period, whereas in females, it reached a plateau after 80 days.
Changes in body weights in male (•) and female rats (○). A significant age dependence was observed for both male and female rats (P < 0.0001). The result of the ANOVA testing for sex dependence is stated in the figure.
The impact of sex was different for ALS, IGFBP-3, and tIGF-I. No significant effect of sex was observed for ALS (Fig. 2A) or IGFBP-3 (Fig. 2B). Levels of IGFBP-3 tended, however, to be higher in males than in females (P = 0.06), and the difference became significant when comparing adult animals 80 days of age or older (P < 0.01). Total IGF-I (tIGF-I) was higher in males than in females (Fig. 2D; P < 0.0001). All three peptides increased significantly with age (P < 0.0001). During the study period, ALS increased from 121 ± 5 to 569 ± 25 nmol/liter (20 vs. 130 days), whereas IGFBP-3 increased from 6.5± 1.7 to 73.6 ± 7.2 nmol/liter in males and from 5.4 ± 1.6 to 51.3 ± 8.0 nmol/liter in females. Levels of tIGF-I increased from 25.2 ± 1.1 to 65.8 ± 2.3 nmol/liter in females and from 28.4 ± 2.2 to 97.1 ± 6.7 nmol/liter in males.
Changes in molar levels of ALS (A), IGFBP-3 determined using the competitive binding assay (B), IGFBP-3 determined using ligand blotting (C; ∼38/42-kDa band), tIGF-I (D), ultrafiltered fIGF-I (E), and IGFBP-1 (F) in male (•) and female rats (○). In all cases a significant age dependence was observed for both male and female rats (P < 0.0001). The results of the ANOVA testing for sex dependence are stated in each figure.
We also used ligand blotting to estimate levels of IGFBP-3, assuming that the 38/42-kDa double band represented the intact form of IGFBP-3. The results were in accordance with those obtained using the competitive binding assay (Fig. 2C), i.e. levels of the 38/42-kDa band were dependent on age (P < 0.0001), and a significant sex dependence was observed only in adult animals 80 days of age or older (P < 0.02). After log transformation of raw data, a significant positive linear correlation between the two assays was observed (r = 0.88; P < 0.0001).
fIGF-I was independent of sex, but was dependent on age (P < 0.0001) (Fig. 2E) and increased from 0.7 ± 0.2 to 7.1 ± 0.5 nmol/liter (20 vs. 130 days; P < 0.0001). Levels of IGFBP-1 decreased significantly with age in both groups (P < 0.0001), in males from 8.9 ± 0.7 to 0.2 ± 0.04 nmol/liter and in females from 10.9 ± 2.5 to 1.2 ± 0.2 nmol/liter (Fig. 2F). In contrast to fIGF-I, IGFBP-1 was dependent on sex, with levels being higher in females than in males (P < 0.0001).
No effect of sex was observed for the molar ratio of IGFBP-3 to tIGF-I (Fig. 3A), in contrast to the sex dependence of the ratios of IGFBP-3 to ALS (P < 0.01; Fig. 3B) and ALS to tIGF-I (P < 0.0001; Fig. 3C). All three ratios increased significantly with age (P < 0.0001).
Changes in molar ratios of IGFBP-3 to tIGF-I (A), IGFBP-3 to ALS (B), and ALS to tIGF-I (C) in male (•) and female rats (○). In all cases, a significant age dependence was observed for both male and female rats (P < 0.0001). The results of the ANOVA testing for sex dependence are stated in each figure.
The molar ratio of fIGF-I to IGFBP-1 was dependent on sex and age (P < 0.0001; Fig. 4A). In young rats, IGFBP-1 circulated in a molar excess of fIGF-I, whereas in rats 40 days of age or older, the concentration of fIGF-I exceeded that of IGFBP-1 by an average of 9-fold in females and by up to almost 60-fold in males. The molar ratio of fIGF-I to tIGF-I (the percentage of fIGF-I) was dependent on age only (P < 0.0001) and increased from approximately 2% to a relatively stable level of 9% (Fig. 4B). The molar ratio of fIGF-I to IGFBP-3 was independent of sex and age, and averaged 0.13 ± 0.01.
Changes in molar ratios of fIGF-I to IGFBP-1 (A) and fIGF-I to tIGF-I (B; i.e. the percentage of fIGF-I) in male (•) and female rats (○). In all cases a significant age dependence was observed for both male and female rats (P < 0.0001). The results of the ANOVA testing for sex dependence are stated in each figure.
Linear regression analysis was used to determine any association among the included variables and was performed on males and females separately. However, as sex-dependent differences were not observed, we show results for pooled data only (Table 1). In addition, the linear regressions of IGFBP-3 vs. ALS, IGFBP-3 vs. tIGF-I, and fIGF-I vs. IGFBP-1 are shown in Fig. 5, A–C. As shown in Table 1, fIGF-I was positively correlated with tIGF-I, IGFBP-3, and ALS and was inversely correlated with IGFBP-1. However, due to multicollinarity among the independent variables (tIGF-I, IGFBP-3, ALS, and IGFBP-1), the multiple linear regression analysis was not able to identify the most important predictors for fIGF-I.
Linear regressions of IGFBP-3 vs. ALS (A), IGFBP-3 vs. tIGF-I (B), and IGFBP-1 vs. fIGF-I (C). The x- and y-axes are logarithmic. The r and P values are stated in each figure. Refer to Table 1 for more information.
Linear regression analysis was performed after log transformation of data
| Variables | n | r value | P value |
|---|---|---|---|
| IGFBP-3 vs. ALS | 76 | 0.92 | <0.0001 |
| Total IGF-I vs. IGFBP-3 | 76 | 0.90 | <0.0001 |
| Total IGF-I vs. ALS | 76 | 0.92 | <0.0001 |
| Free IGF-I vs. IGFBP-1 | 75 | −0.80 | <0.0001 |
| Free IGF-I vs. IGFBP-3 | 75 | 0.65 | <0.0001 |
| Free IGF-I vs. total IGF-I | 83 | 0.87 | <0.0001 |
| Free IGF-I vs. ALS | 75 | 0.90 | <0.0001 |
| Variables | n | r value | P value |
|---|---|---|---|
| IGFBP-3 vs. ALS | 76 | 0.92 | <0.0001 |
| Total IGF-I vs. IGFBP-3 | 76 | 0.90 | <0.0001 |
| Total IGF-I vs. ALS | 76 | 0.92 | <0.0001 |
| Free IGF-I vs. IGFBP-1 | 75 | −0.80 | <0.0001 |
| Free IGF-I vs. IGFBP-3 | 75 | 0.65 | <0.0001 |
| Free IGF-I vs. total IGF-I | 83 | 0.87 | <0.0001 |
| Free IGF-I vs. ALS | 75 | 0.90 | <0.0001 |
Due to the shortage of serum, it was not possible to perform all analyses in every animal. The total number of animals was 84 (42 males and 42 females).
Linear regression analysis was performed after log transformation of data
| Variables | n | r value | P value |
|---|---|---|---|
| IGFBP-3 vs. ALS | 76 | 0.92 | <0.0001 |
| Total IGF-I vs. IGFBP-3 | 76 | 0.90 | <0.0001 |
| Total IGF-I vs. ALS | 76 | 0.92 | <0.0001 |
| Free IGF-I vs. IGFBP-1 | 75 | −0.80 | <0.0001 |
| Free IGF-I vs. IGFBP-3 | 75 | 0.65 | <0.0001 |
| Free IGF-I vs. total IGF-I | 83 | 0.87 | <0.0001 |
| Free IGF-I vs. ALS | 75 | 0.90 | <0.0001 |
| Variables | n | r value | P value |
|---|---|---|---|
| IGFBP-3 vs. ALS | 76 | 0.92 | <0.0001 |
| Total IGF-I vs. IGFBP-3 | 76 | 0.90 | <0.0001 |
| Total IGF-I vs. ALS | 76 | 0.92 | <0.0001 |
| Free IGF-I vs. IGFBP-1 | 75 | −0.80 | <0.0001 |
| Free IGF-I vs. IGFBP-3 | 75 | 0.65 | <0.0001 |
| Free IGF-I vs. total IGF-I | 83 | 0.87 | <0.0001 |
| Free IGF-I vs. ALS | 75 | 0.90 | <0.0001 |
Due to the shortage of serum, it was not possible to perform all analyses in every animal. The total number of animals was 84 (42 males and 42 females).
Serum levels of insulin were independent of sex, but were changed significantly by age (P < 0.0001; data not shown). A significant inverse correlation between insulin and IGFBP-1 (r = −0.65; P < 0.0001) was observed, whereas insulin was not significantly correlated with fIGF-I.
Discussion
The aim of the present study was to describe the age- and sex-related changes in the circulating IGF system in normal male and female rats, with special reference to IGFBP-3 and the relationship between fIGF-I and IGFBP-1.
rIGFBP-3 was determined by use of a competitive binding assay based on the ability of IGFBP-3 to complex with ALS in the presence of IGF-I (11). Using this assay, levels of IGFBP-3 reached a plateau of approximately 51 nmol/liter in females and 74 nmol/liter in males. As IGFBP-3 stabilizes IGF-I in a complex with ALS, the significant sex difference in IGFBP-3 in adult animals may at least partly explain the higher levels (∼30 nmol/liter) of IGF-I in males. ALS changed in accordance with the original findings in rats (9), and as previously observed in humans, ALS always remained in molar excess of IGFBP-3 (8).
The strong age dependence of IGFBP-3, tIGF-I, and ALS, and the tight positive correlations among the three GH-dependent peptides were expected, and these findings are in accordance with observations in humans (4, 8, 22–24). Thus, the present results may be regarded as a further validation of the new rIGFBP-3 assay (11). Levels of rIGFBP-3 were positively correlated with those obtained by ligand blotting. A difference between the two methods was observed, however. As seen in Fig. 2, B and C, ligand blotting yielded a relative increase from 20–130 days of age that was much larger than that observed using the competitive binding assay, whereas the two methods showed comparable relative changes from 40–130 days of age. Therefore, it seems as if ligand blotting underestimates levels in young rats, and this may be explained by the semiquantitative nature of ligand blotting.
During human puberty, tIGF-I and IGFBP-3 reach peak levels about 1 yr earlier in girls than in boys, but in adult life, the sex difference is absent (23, 25–27). In rats, it appears to be different; during puberty, levels of IGFBP-3 and tIGF-I (28, 29) increase concomitantly in males and females, whereas both peptides circulate in higher concentrations in mature males than in mature females. This may be related to the well described sexual dimorphism in GH secretion, which becomes evident at the entry of puberty; in male rats, the pituitary content, rate of synthesis, and pulse amplitude of GH are higher than those in females (29, 30). However, studies in ovariectomized rats, in which estradiol suppresses and testosterone increases the circulating levels of tIGF-I, imply that sex steroids may also be of importance (31). In this context it is interesting that ALS showed no sex difference; if ALS, IGFBP-3, and IGF-I responded similarly to changes in GH secretion, the molar ratios would be expected to remain more or less unaltered during development. This was not the case, however, and further studies are needed to investigate the precise mechanisms that differentiate the regulations of ALS, IGFBP-3, and IGF-I.
Our study implies that IGFBP-3 is the main carrier of circulating IGF-I in mature rats, and this is consistent with findings in human serum (7). Interestingly, during the postweanling period, rats have a much lower IGFBP-3/tIGF-I ratio, suggesting that other IGFBPs (probably IGFBP-1 and IGFBP-2) (32) have a more important role in IGF transport before puberty. There has been some discrepancy concerning the true level of tIGF-I in rats (33, 34). However, Lee et al. (33) using a homologous assay (based on rat calibrators and antibodies to rat IGF-I) found a serum tIGF-I level of approximately 120 nmol/liter in adult male rats, corresponding well with our concentration of 100 nmol/liter; therefore, we believe that the calculated ratios between IGFBP-3 and tIGF-I are valid.
The present study also aimed at examining the concomitant changes in IGFBP-1 and fIGF-I in rats using specific and validated methods (19, 21). The age-related changes in IGFBP-1 were in accordance with previous findings in rats (21) as well as in humans (12). The sex difference is in accordance with clinical studies (35, 36), whereas in rats, levels of IGFBP-1 were reported to be higher in fasted, but not in nonfasted, female animals, compared with those in male littermates (21). The latter study included, however, only eight rats, which may explain the discrepancy with the present study.
Insulin is the primary regulator of IGFBP-1 (12), and insulin levels were significantly inversely correlated with those of IGFBP-1. However, no sex difference in levels of insulin was observed, pointing to additional regulatory mechanisms of IGFBP-1. Previous studies have shown that estrogen increases and testosterone decreases levels of IGFBP-1 in humans (37–39), and exposure to sex steroids may contribute to the observed difference, which, accordingly, seemed to be most pronounced after commencement of the pubertal growth surge.
Serum fIGF-I was determined using ultrafiltration, as previously described (19, 20). In the present study, the imprecision of fIGF-I averaged 26%, which is higher than usual. Most likely, this does not affect changes between groups, but it may be of importance when performing linear regression analyses, which are based on comparison of individual values. Thus, it may explain why the multiple regression analysis could not isolate the most important predictors for fIGF-I among tIGF-I, ALS, IGFBP-3, and IGFBP-1.
In humans, levels of ultrafiltered fIGF-I were found to be highly age dependent, with concentrations increasing to 1 nmol/liter during puberty (17), after which a steady age-dependent decline was noticed, reaching levels as low as 0.05 nmol/liter in old subjects (19). Thus, in both humans and rats, fIGF-I is strongly age dependent. Similarly, levels of fIGF-I were significantly inversely correlated with IGFBP-1, again consistent with our previous observations in humans (14–17).
However, our results of fIGF-I and IGFBP-1 also point to interesting species differences. In adult humans, the concentration of fIGF-I rarely exceeds 0.3 nmol/liter (14–16, 19), and in overnight fasting samples, IGFBP-1 nearly always circulates in a molar excess over the level of fIGF-I (14, 15). Even after a 4-h hyperinsulinemic clamp with very low levels of IGFBP-1, the ratio remains in favor of IGFBP-1 (16). In rats, the nonfasted levels of fIGF-I are severalfold higher than those of IGFBP-1, and we have data (Frystyk, J., P. J. D. Delhanty, C. Skjærbæk, and R. C. Baxter, unpublished) showing that this is also true after 24 h of fasting (ratio of fIGF-I/IGFBP-1 ∼3:1). It has previously been argued that the rat circulation must contain a high concentration of bioavailable IGF-I, because a high proportion of injected hIGFBP-3 was found in ternary complex with ALS within 2 min after injection (40). Our observation of a high level of fIGF-I in adult rat serum provides a simple explanation for this observation, compatible with the view that IGFBP-3 binding to ALS is very weak in the absence of IGFs. In contrast to the high concentration of readily available IGF-I in rat serum, the very low levels of fIGF-I in the adult human circulation suggest a much tighter control of IGF-I bioavailability.
The mechanisms underlying this species difference are not clear, but as the half-life of fIGF-I is very short compared with that of bound IGF-I (2), the high levels of fIGF-I in rats imply a much higher production rate. Making the assumptions that fIGF-I in plasma is the primary determining fraction for the metabolism of circulating IGF-I, and that the half-life and volume of distribution [set at 10 min (2) and 230 ml/kg (41), respectively] are independent of age, the estimated production rate of IGF-I increases during the study period from approximately 10 to 200 nmol/kg·day. These values may be compared with the reported adult human IGF-I production rate of approximately 20 nmol/kg·day (2, 42).
In conclusion, the present study has revealed similarities as well as dissimilarities in the circulating IGF system when comparing rats and humans. In accordance with clinical findings, rat IGFBP-3 was strongly age dependent, was positively correlated with ALS and tIGF-I, and circulated in a molar concentration less than that of ALS. The age dependence of ultrafiltered fIGF-I and its inverse correlation with IGFBP-1 are also in accordance with findings in humans. However, the impact of sex on levels of tIGF-I and IGFBP-3 appears to differ between humans and rats. Finally, levels of fIGF-I are severalfold higher in rats than in humans.
We are indebted to Mrs. K. Nyborg Rasmussen, Mrs. S. Sørensen, Mrs. J. Hansen, Mrs. I. Bisgaard, Mrs. N. Rosenqvist, and Mrs. K. Mathiassen for skilled technical assistance.
