In vivo analysis of growth hormone receptor signaling domains and their associated transcripts

Abstract

The growth hormone receptor (GHR) is a critical regulator of postnatal growth and metabolism. However, the GHR signaling domains and pathways that regulate these processes in vivo are not defined. We report the first knock-in mouse models with deletions of specific domains of the receptor that are required for its in vivo actions. Mice expressing truncations at residue m569 (plus Y539/545-F) and at residue m391 displayed a progressive impairment of postnatal growth with receptor truncation. Moreover, after 4 months of age, marked male obesity was observed in both mutant 569 and mutant 391 and was associated with hyperglycemia. Both mutants activated hepatic JAK2 and ERK2, whereas STAT5 phosphorylation was substantially decreased for mutant 569 and absent from mutant 391, correlating with loss of IGF-1 expression and reduction in growth. Microarray analysis of these and GHR(-/-) mice demonstrated that particular signaling domains are responsible for the regulation of different target genes and revealed novel actions of growth hormone. These mice represent the first step in delineating the domains of the GHR regulating body growth and composition and the transcripts associated with these domains.

FIG. 1.

Creation of GHR targeted knockin lines. (A) Targeting strategy for both mutant types with altered exon 10 using neomycin resistance for selection in ES cells. A Cre/lox system was adapted to remove the selection cassette. (B) Southern confirmation of heterozygous ES cell clones for mutant 569 (1), mutant 391 (2), and mutant 391 transiently transfected with Cre recombinase (3). (C) Northern blot analysis confirmed that Ghr and Ghrbp (GH-binding protein) mRNA was expressed in hepatic tissue in both mutants 569 and 391 in both heterozygous (H) and homozygous forms. (D) Western blotting of hepatic microsomal membranes with cytoplasmic domain specific antibody detected truncated mutant 569 GHR protein; however, the epitope was absent from mutant 391.

FIG. 2.

Signaling in mutant mice in response to GH injections. (A) Predicted binding of signaling and adaptor molecules to the cytoplasmic domain of the GHR of the WT, mutant 569, and mutant 391. (B to D) Livers from the bGH- and saline-injected 19-day-old mice were used in immunoprecipitation analyses 15 min after the injections. Loading for each of the proteins was confirmed by using antibodies specific for an appropriate protein. (B) Antiphosphotyrosine (PY) immunoblot of JAK2 immunoprecipitated from liver homogenate. (C) Immunoblot for active ERK1/2 (Phospho-p44/42 mitogen-activated protein kinase) from liver homogenate. (D) Antiphosphotyrosine (PY) immunoblot of STAT5 immunoprecipitated from liver homogenate. (E to G) Densitometric quantification of the blots from panels B to D. Signals for the activated JAK2 (E), ERK1/2 (F), and STAT5a/b (G) were normalized for loading. The data for all graphs are presented as means ± the SEM.

FIG. 3.

Postnatal growth reduction of the GHR mutant mice. (A) Photograph of WT and mutant homozygous 60-day-old male mice. (B to D) Color coded as follows: WT (black), mutant 569 (blue), mutant 391 (red), and GHR^−/− (green). (B and C) Postnatal growth curves (B, females; C, males). All mice were weighed over a period of 60 days. The results are presented as mean body weight ± the SEM (n = 8). Growth was significantly reduced in all mutants compared to WT (P < 0.001).(D) Organ weights were recorded at 60 days postnatally. The results are shown as relative to body weight and are expressed as means ± the SEM (n = 6 to 8). The sizes of some organs were significantly reduced (all changes are marked with a single asterisk due to space constraints; P < 0.01 for all changes, except P < 0.05 for spleen in mutant 391 and liver in GHR^−/− mice). (E) Subcutaneous fat pad sizes at 4 (▪) and 10 (□) months in male mice. The results are expressed relative to body weight and are means ± the SEM (n = 3 to 4). ✽✽, P < 0.01; ✽, P < 0.05. (F) Glucose levels in 10-month-old males after a 16 h fasting (n = 3) and are shown as means ± the SEM (✽, P < 0.01).

FIG. 4.

IGF-1 axis in GHR mutant mice. (A) Densitometric quantification of a Northern blot analysis of the IGF-1 levels in GHR mutant mice and WT littermates. Transcripts of 0.9 to 1.2 kb were used for this analysis. The data are shown as percentages of WT and are displayed as means ± the SEM (n = 10). (B) Levels of IGF-1 serum as measured by radioimmunoassay (see Materials and Methods) were significantly decreased in mutant 569, mutant 391, and GHR^−/− mice in comparison to WT (P < 0.001). The results are shown are means ± the SEM (n = 7 to 11). (C) Western ligand blot profile of IGFBP levels in serum in mice at 42 days shows a significant decrease in the levels of IGFBP3 (P < 0.001), with no change detected in the remaining IGFBPs.

FIG. 5.

Regulation of hepatic nuclear factors in GHR mutant mice. (A and B) Western blot analysis of the expression of HNF3β (A) and HNF1α (B) in the livers of 10-month-old mice fasted for 16 h. Both HNFs were immunoprecipitated from the liver homogenates, and immunoglobulin G light chain was used as a loading control. (C and D) Densitometric quantification of the results from panels A and B are shown as means ± the SEM (n = 3). ***, P < 0.001.

FIG. 6.

Global analysis of the GHR mutant mice by using microarrays. (A) Venn diagram showing the number of genes differentially expressed in the three knockin lines compared to WT mice and the number of genes up- and downregulated in each group. Genes expressed in common between the mutants are indicated by overlaps of the circles. (B and C) Hierarchical clustering was performed with DecisionSite 7.2 software (functional genomics module). (B) Mutant 569 is most similar in its pattern of gene expression to the WT, and mutant 391 is similar to these two groups and to GHR^−/− mice cluster separately. This clustering was generated by using the average signal (n = 3) for all probe sets from the arrays, with exclusion of genes absent in all groups. (C) Metabolically, mutant 569 is most similar to WT mice in its gene expression pattern, whereas mutant 391 is most similar to GHR^−/− mice. This clustering was performed by using average signal (n = 3) for probe sets representing genes encoding proteins involved in metabolism. (D) Northern blot confirmation of microarray results. Total RNA from the 42-day-old mice was run on a 1% denaturing gel, blotted onto a nylon membrane, and probed with probes targeting the same region of the gene as the Affymetrix probes. The blot was then reprobed with 18S oligomer probe to confirm loading. (E) Changes in signaling and GH-dependent growth in GHR^−/− mice. A summary of the results shows that in WT mice 100% signaling through all major pathways contributes to 100% of GH-dependent growth, whereas in mutant 569 the loss of 70% of STAT5 signaling results in the loss of 66% of GH-dependent growth, and the absence of active STAT5 pathway in mutant 391 is not adequate to reduce GH-dependent growth to zero; these mice retain 11% of this growth. Metabolic pathways identified by GO analysis and examples of genes belonging to each category that were differentially expressed are shown (http://research.imb.uq.edu.au/∼mwaters/ghr/).