|Home | About | Journals | Submit | Contact Us | Français|
Recent studies using twins and inbred strains of mice reveal evidence for genetic mechanisms contributing to variation in circulating levels of IGF-I, IGF-II, and IGF binding protein (IGFBP)-3. To examine the hypothesis that serum IGFBP-5 levels have a strong heritable component, we intercrossed two inbred strains of mice, MRL/MpJ and SJL, which exhibit 79% difference in serum IGFBP-5 levels (554 ± 68 vs. 309 ± 51 ng/ml respectively, P < 0.001). A genome-wide scan was carried out using 137 polymorphic markers in 633 F2 female mice. Serum IGFBP-5 levels in the F2 progeny showed a normal distribution with an estimated heritability of 74%. Whole genome-wide scans for cosegregation of genetic marker data with high or low serum IGFBP-5 levels revealed six different quantitative trait loci (QTL) in chromosomes 1, 9 (two), 10, and 11 (two), which together explained 24% of F2 variance. Chromosome 11 QTL exhibited the highest LOD score (7.5). Based on the past findings that IGFBP-5 is an important bone formation stimulator, we predicted IGFBP-5 to contribute to bone mineral density variation in F2 mice. Accordingly, we found two of the six IGFBP-5 QTLs (Chrs 1 and 11) identified for serum IGFBP-5 phenotype also showed significant association with total body bone mineral density phenotype (measured by dual energy x-ray absorptiometry) in the F2 mice.
IGF-I and -II are growth factors that have both mitogenic and metabolic actions, and participate in the growth, survival, and differentiation of a number of cell types including osteoblasts. The importance of IGFs in regulating bone metabolism is evident from recent studies on skeletal changes using mice with disruption of IGF-I or IGF-II (1–3). The functions of IGFs depend not only on the amount of IGF produced but also on the level of IGF binding proteins (IGFBPs). IGFBPs exert the traditional functions of binding proteins, whereby they modulate the half-life and activity of IGFs. In addition, some of the IGFBPs have been shown to act via mechanisms independent of IGFs (4–6). Of the six high-affinity IGFBPs produced by osteoblasts, IGFBP-5 has several distinct features that suggest it is a key component of the IGF system in bone: 1) IGFBP-5 is the most abundant IGFBP stored in bone, where it is bound to hydroxyapatite and extracellular matrix proteins, binding that provides a mechanism to fix IGFs in bone for subsequent release in a regulatable manner; 2) IGFBP-5 has consistently been shown to stimulate bone formation parameters in vitro and in vivo; 3) IGFBP-5 shows considerable changes in clinical disease states and correlates with changes in bone formation; and 4) IGFBP-5 can also function as a growth factor in addition to its role as a traditional binding protein as evident from recent studies using IGF-I knockout mice (7–16).
In clinical studies, we found evidence that serum level of IGFBP-5 correlated positively with bone mineral density (BMD) (14, 16, 17). Furthermore, it was found that there was considerable interindividual variation in circulating levels of IGFBP-5 in normal human individuals (14, 16, 17). In terms of potential factors that could contribute to variation in circulating levels of IGFBP-5, a number of possibilities exist, including differences in GH secretion, nutritional uptake, proteolysis, and genetic makeup (5, 16, 18). Of these variables, the genetic component is proposed to exert a significant influence on circulating levels of IGFBP-5 based on several findings. First, Kao et al. (19) demonstrated for the first time that the variation in IGF-I levels in healthy twin children is almost completely of genetic origin. Second, Harrela et al. (20) showed that, in adults, there is a substantial genetic contribution responsible for interindividual variation of the circulating levels of IGF-I, IGF-II, and IGFBP-3. Third, Rosen et al. (21) showed that serum IGF-I levels in C57BL/C3HHeJ F2 mice are inherited as a polygenic trait. Based on these data, we proposed that serum IGFBP-5 level, like IGF-I, has a strong heritable component and is controlled by one or more genes. To evaluate this hypothesis, we performed quantitative trait loci (QTL) studies using two inbred strains of mice that exhibit extreme differences in serum levels of IGFBP-5.
The inbred strains, MRL/MpJ (MRL) and SJL/J (SJL) were purchased from The Jackson Laboratory (Bar Harbor, ME) and maintained under 14-h light, 10-h dark cycles. MRL females were mated to SJL males to produce the F1 progeny, which were then intercrossed to produce F2 progeny. These inbred strains of mice were selected from our study of twenty different strains that were used for ear regeneration studies (22, 23). Parental strains, F1, and F2 mice were killed at 7 wk of age and used for phenotypic and genotypic measurements. GH-deficient lit/lit and control lit/+ mice were generated using breeding pairs of lit/+ mice kindly provided by Dr. L. R. Donahue (The Jackson Laboratory). Due to a spontaneous mutation in GHRH receptor molecule, the lit/lit mice (C57BL/6J) are deficient in GH and are 50% smaller than that of corresponding control lit/+ mice (24). Serum samples were collected from 8-wk-old lit/lit and lit/+ mice and frozen at −70 C for IGFBP-5 measurements. The experimental protocols were in compliance with animal welfare regulations and approved by the Jerry L. Pettis VA Medical Center.
Serum IGFBP-5 was measured by a RIA that has been validated in our laboratory for measurement of serum IGFBP-5 levels in mice. Recombinant human IGFBP-5 was used as a standard and tracer. Antibodies against recombinant human IGFBP-5 were raised in guinea pigs as described previously (25). IGFBP-5 antiserum that showed cross-reactivity with mouse IGFBP-5 was selected for RIA. Mouse serum samples were diluted 1:10 before assay. The inter- and intraassay coefficient of variation for this assay is less than 10%. The sensitivity of the assay is 10 ng/ml. None of the other IGFBPs showed significant cross-reactivity in this assay.
Total body BMD was measured using PIXIMUS densitometer (Lunar Corp., Madison, WI) as described previously (27). The precision of PIXIMUS for measurement of total body BMD is less than 2%.
Genomic DNA was extracted from liver using a commercial DNA extraction kit (Promega Corp., Madison, WI). Genotyping of individual mouse DNAs was accomplished by PCR with oligonucleotide primer pairs for microsatellite markers that were purchased from Research Genetics, Inc. (Huntsville, AL). A total of 600 primer pairs were tested for polymorphism between the two inbred strains to identify 137 markers with distinguishable polymorphisms that were used for genotyping. The conditions for PCR cycling were set as described previously (23, 27). PCR products were separated on 6% polyacrylamide gel, stained with ethidium bromide and visualized by the ChemiImager 4400 Low Light Imaging System (α Innotech, San Leandro, CA). Alleles derived from the MRL/MpJ parent were designated “A”, SJL/J-derived alleles designated as “B,” and MRL/SJL heterozygotes as “H” in data analyses (23, 27).
Genotype data were analyzed using a MAPQTL (4.0) program (DLO Center for Plant Breeding and Reproduction Research, Wageningen, The Netherlands). MAPQTL interval mapping was used for QTL mapping and the LOD score significance thresholds were calculated using MAPQTL permutation test (23, 27). The LOD score is a statistical test for measuring the probability that there is a linkage of loci with a given phenotype. QTL with an LOD score of more than 3.5 is considered to be significant while those QTL with a LOD score of more than 2.7 are considered to be suggestive (28). The broad sense of heritability was estimated by using variances obtained from parental strains, F1 and F2 mice as previously described (23, 27).
To determine whether human IGFBP-5 RIA could be applied for measurement of immunoreactive IGFBP-5 in mouse serum, we tested various IGFBP-5 antisera raised in guinea pigs (25) for competition experiments. We found that one out of the five guinea pig IGFBP-5 antisera tested showed significant cross-reactivity with IGFBP-5 present in mouse serum (data not shown). Figure 1 shows that mouse serum inhibited the binding of [125I]IGFBP-5 to IGFBP-5 antibody raised against human IGFBP-5 in guinea pig in a parallel manner. This antiserum detected IGFBP-5 in human serum as expected. To further validate the IG-FBP-5 RIA for measurement of IGFBP-5 in mouse serum, we compared IGFBP-5 levels in GH-deficient lit/lit and corresponding control lit/+ mice. Based on previous findings that GH is a major regulator of serum IGFBP-5 levels, we predicted serum levels of IGFBP-5 to be significantly low in lit/lit mice. Figure 2 shows that serum IGFBP-5 level was reduced by 70% in lit/lit mice compared with corresponding age-matched control mice.
MRL mice exhibit 79% higher serum levels of IGFBP-5 compared with SJL mice (Fig. 3). Interestingly, serum IGF-I levels were also significantly higher in the MRL mice compared with SJL mice (579 ± 89 vs. 425 ± 61 ng/ml, n = 20 per group; P < 0.001). F1 mice have a mean serum IGFBP-5 level closer to the MRL parent than to the intermediate level between MRL and SJL parents (Fig. 3). These data suggest that the MRL parent may contain a dominant gene that contributes to high serum IGFBP-5 levels.
Frequency distribution of serum IGFBP-5 in F2 mice is not significantly different from the theoretical normal distribution (Fig. 4). Whereas a majority of the F2 mice exhibit serum IGFBP-5 levels between the two parental strains, few F2 mice exhibit higher or lower serum IGFBP-5 levels than MRL and SJL, respectively. These data suggest that serum IGFBP-5 is a quantitative trait and that both MRL and SJL strains contain genetic loci that contribute to variation in serum IGFBP-5 in F2 mice. The estimated broad sense of heritability for serum IGFBP-5 was 74%.
The results of genome-wide scans are presented in Fig. 5. Whole-genome scans with marker regression revealed highly significant peaks on Chrs 1, 9, 10, and 11, of which Chrs 9 and 11 contain two QTL, whereas Chrs 1 and 10 contain one QTL. Table 1 provides the list of markers for various QTL that show significant linkage with serum IGFBP-5 levels and the percent of F2 variance explained by individual QTL. Of the six QTL, chromosome 11 QTL (D11Mit36) exhibited the highest LOD score and contributed to 6.7% of the variation in serum IGFBP-5 levels seen in the MRL/SJL F2 mice. The six identified QTL explained 24% of variation in serum IGFBP-5 levels seen in the MRL/SJL F2 mice.
Because IGFBP-5 has been shown to stimulate bone formation in mice in part by an IGF-independent mechanism (13), and because serum IGFBP-5 levels show significant positive correlation with BMD in clinical studies (14, 16, 17), we predicted common genetic mechanisms regulating serum IGFBP-5 and BMD phenotypes. Accordingly, we found that two of the identified QTLs (Chrs 1 and 11) for serum IGFBP-5 also showed significant linkage with total body BMD in the MRL/SJL F2 mice (Table 2).
In this study, we found that heritability for serum IGFBP-5 levels was high (74%) and that at least 6 QTLs for serum IGFBP-5 level had LOD scores greater than 3.5. These loci are distributed in four different chromosomes and accounted for 24% of the variance in serum IGFBP-5 levels among the 633 MRJ/SJL F2 mice. In addition, we have shown that two of the QTLs identified for serum IGFBP-5 also exhibit significant association with total body BMD phenotype.
In previous studies using MRL and SJL inbred strains of mice, we demonstrated significant QTLs for body length, lean body mass, muscle size, femur length, bone density, and wound regeneration (23, 27, 29–31). Because we and others have shown that IGFBP-5 is a stimulator of both osteoblast and myoblast differentiation (10, 32), we predicted IGFBP-5 levels to be different between the two inbred strains of mice. Accordingly, we found serum IGFBP-5 level to be 79% higher in MRL (good healer strain) compared with SJL (poor healer strain). In this study, we have shown that the serum IGFBP-5 variation in MRL/SJL F2 mice has a strong heritable component and that multiple genetic loci contribute to variation in serum IGFBP-5 levels. Although 74% of variation in serum IGFBP-5 levels can be explained on the basis of heredity, the six identified QTL explained only 24% of variation in IGFBP-5 levels seen in F2 mice. There are a number of potential explanations for this low contribution by the QTLs identified in this study: 1) serum IGFBP-5 levels could be under the influence of many genetic loci and we were able to detect only the major QTLs in our study; and 2) there could be significant epistatic interaction between genetic loci that could play an important role in the regulation of serum IGFBP-5 levels. This interaction could occur between the genetic loci that regulate serum IGFBP-5 levels or between IGFBP-5 QTLs and other QTLs (e.g. QTLs for other IGF system components). Further studies are needed to determine the other variables that contribute to the remaining two thirds of the observed variation in serum IGFBP-5 levels in the MRL/SJL F2 mice.
IGFBP levels in serum and other biological fluids depend on the rate of synthesis and degradation rate (18). In terms of regulatory molecules involved in regulating IGFBP-5 levels, several lines of evidence demonstrate that GH is a major regulator of serum IGFBP-5 levels. First, serum IGFBP-5 levels were significantly reduced in GH-deficient children as well as in adults (15, 16). Second, GH treatment of GH-deficient children and adults increased serum IGFBP-5 levels to levels similar to those seen in corresponding age-matched normal subjects (15, 16). Third, children with GH receptor insensitivity syndrome exhibited serum IGFBP-5 levels that are 80% less compared with normal children (33). Fourth, serum levels of IGFBP-5 exhibit significant positive correlation with other GH responsive IGF system components (e.g. IGF-I, IGFBP-3, and ALS) but not others (e.g. IGFBP-4) (34). Finally, GH increased expression of IGFBP-5 in bone (35). To determine if differences in the production and/or actions of GH could contribute, in part, to a variation in serum IGFBP-5 levels, we evaluated if any of the genetic loci identified for the regulation of serum IGFBP-5 levels contain candidate genes known to be involved in the production and actions of GH. In this regard, it is worth noting that GH gene is located in the chromosome 11 QTL (IGFBP5-6) that is linked to serum IGFBP-5 level (Fig. 4D). Because the identified QTL region in chromosome 11 is rather large, further studies using additional microsatellite markers in this region are needed to further define the location of the GH gene in relationship to the chromosomal region that contributes to variation in serum IGFBP-5 levels.
Our findings that serum IGF-I levels are also elevated in MRL mice compared with SJL mice seem to suggest that variations in the production and/or actions of GH could in part contribute to the observed differences in serum IGFBP-5 levels between MRL and SJL mice via increasing the local and endocrine actions of IGF-I (18). We did not measure serum GH levels in this study because GH is secreted in a pulsatile manner and that a single time measurement of GH has not been shown to be a valid measurement of overall GH secretion. In any case, if the prediction involving GH turns out to be true, we would expect the chromosome 11 IGFBP-5 QTL to be a shared QTL with GH, IGF-I, IGFBP-3, and ALS.
It is known that degradation of IGFBP-5 by IGFBP proteases is an important mechanism by which IGFBP-5 levels in serum and other biological fluids are controlled. In this regard, we and others have shown that a number of proteins including ADAM-9, ADAM-12, pregnancy-associated plasma protein-A2, and complement C1s degrade IGFBP-5 specifically (36–39). Of these various IGFBP-5 proteases, the pregnancy-associated plasma protein-A2 gene is found to be located within the chromosome 1 IGFBP-5 QTL region. Based on these and previous findings, it is likely that alterations in both synthesis and degradation of IGFBP-5 could contribute to a variation in serum levels of IGFBP-5 between MRL and SJL mice.
IGFBP-5 is unique in that only IGFBP-5 has been consistently shown to potentiate IGF actions in a variety of cell types including osteoblasts, chondroblasts, and smooth muscle cells. Studies on the mechanisms by which IGFBP-5 stimulates osteoblasts have shown that IGFBP-5 itself is a growth factor capable of stimulating bone formation in the absence of its ligand (13). Accordingly, serum levels of IGFBP-5 have been shown to correlate with BMD in several clinical studies (14, 16, 17). If IGFBP-5 is an important variable regulating BMD, we predicted common genetic determinants for BMD and serum IGFBP-5 phenotypes. In this regard, we found that two of the six IGFBP-5 QTLs are shared QTLs with BMD and serum IGFBP-5 phenotypes. It remains to be established whether the shared QTLs for the BMD and serum IGFBP-5 harbor a single gene that regulates IGFBP-5 levels and, consequently, BMD or if they contain two genes, one that regulates IGFBP-5 levels and consequently BMD and the second that regulates BMD phenotype independently of IGFBP-5.
In conclusion, our study provides the first identification of QTLs regulating serum IGFBP-5 levels and genetic evidence for a binding protein regulating BMD. Future identification of one or more genes regulating extreme differences in serum IGFBP-5 levels between MRL and SJL inbred strains of mice could lead to not only identification of molecular pathways of IGFBP-5 regulation but also of genes regulating BMD differences.
All work was performed in the facilities provided by the Department of Veterans Affairs. We would like to thank Alice Kramer for technical assistance and Sean Belcher for secretarial assistance.
This work was supported by the National Institutes of Health (AR-31062) and by the Assistance Award No. DAMD17-99-1-9571. The U.S. Army Medical Research Acquisition Activity (Fort Detrick, MD) 21702-5014 is the awarding and administering acquisition office for the DAMD award. The information contained in this publication does not necessarily reflect the position or the policy of the Government, and no official endorsement should be inferred.