Main

Growth is an inherent property of life. Normal somatic growth requires the integrated function of many of the hormonal, metabolic, and other growth factors involved in the hypothalamo-pituitary-growth axis. Short stature associated with GH deficiency has been estimated to occur in about 1/4 000 to 1/10 000 of the population in various studies(1–4). Although most cases are sporadic and are believed to result from environmental cerebral insults or developmental anomalies, 3-30% of cases are reported to have an affected first degree relative, suggesting a genetic etiology(2, 3). Inasmuch as magnetic resonance examinations detect only 12-20% anomalies of either hypothalamus or pituitary gland in patients suffering from IGHD, one might assume that many genetic defects may not be diagnosed, and a significantly higher proportion of sporadic cases may have a genetic cause(5).

The GH gene cluster consists of five very similar genes in the order 5′ [GH-1, CSHP (chorionic somatomammotropin pseudogene), CSH-1(chorionic somatomammotropin gene), GH-2, CSH-2] 3′ encompassing a distance of about 65 000 bp (65 kb) on the long arm of chromosome 17 at bands q22-24(6). The extensive homology (92-98%) between the immediate flanking, intervening, and coding sequences of these five genes suggests that this multigene family arose through a series of duplicational events(7). Mature GH encoded by the GH-1 gene is a single polypeptide chain consisting of 191 amino acids, has a molecular mass of 22 kD, and four cysteines which form two intramolecular disulfide bridges.

Familial IGHD is associated with at least four Mendelian disorders; two forms that have autosomal recessive inheritance (IGHD types IA and IB) as well as autosomal dominant (IGHD type II) and X-linked (IGHD III) forms(8). Complete lack of endogenous GH production can cause the formation of anti-GH-antibodies after GH replacement therapy(9). These cases are mainly caused by GH-1 gene deletions(10–13). However, there are recent reports of GH-1 gene defects other than deletions causing the phenotype of GH deficiency. These defects include point mutations/deletions causing stop codons, frameshifts, and/or splicing errors(14–18). Because the proportion of IGHD cases that are caused by all these GH-1 gene defects is unknown, the aim of this study was to determine the prevalence of GH-1 gene defects among different populations of patients suffering from severe forms of IGHD (height <-4.5 SD score) by sequencing the whole GH-1 gene after PCR amplification.

METHODS

Patients. One hundred fifty-one subjects with severe growth retardation due to IGHD were studied. Diagnostic criteria included a pretreatment height of <-4.5 SD score, [except for one baby aged 3 mo (-3.4 SD score)], decreased height velocity, retarded bone age, normal karyotype, normal prolactin levels and thyroid function, and peak GH levels below 4 ng/mL after a stimulation test (insulin-induced hypoglycemia). Details of the patients and their families are shown in Tables 1 and 2. The individuals were broadly grouped into three populations: Northern Europeans (north of the Alps), Mediterraneans (south of the Alps), and Asians.

Table 1 Patients and families studied
Full size table
Table 2 Clinical details*
Full size table

DNA isolation. Genomic DNA was isolated from peripheral leukocytes of subjects and their relatives as previously reported(19). The concentration of each sample was determined by measuring the OD of the purified DNA at 260 and 280 nm.

Restriction endonuclease digestion, Southern blot analysis, and PCR amplifications (GH-gene deletions). Samples of DNA (5 μg) were digested to completion with the restriction enzymes BamHI and Hin dIII. After electrophoresis, Southern blotting was performed as previously described(19, 20). To screen for gene human GH-1 gene deletions, PCR amplification was done as reported elsewhere(20).

PCR amplification of genomic DNA and sequencing. The GH-1 genes of affected and unaffected families members were PCR-amplified as previously reported by Cogan et al.(17). Briefly, 1μg of genomic DNA was added to a 100-μl reaction mixture of 10 mmol/L Tris-HCl (pH 8.4), 50 mmol/L KCl, 1.5 mmol/L MgCl2, 0.1% Triton X-100, 200 mmol/L of each dNTP, 0.1 μmol/L of each primer, and 4 U Taq DNA polymerase (Boehringer Mannheim; Rotkreuz, Switzerland). The sense and antisense primers used corresponded to nucleotides 5101-5136(5′-GTGGGGGCAACAGTGGGAGAGAAGGGGCCAGGTATA-3′) and the complement of 7255-7226 (5′-TGTCCCACCGGTTGGGCATGGCAGGTAGCC-3′) of the GH gene cluster(21). In Figure 1 the exon-intron structure of the genomic DNA, the cDNA and the human GH peptide is presented. The PCR reaction mix was denatured for 5 min at 94°C and cycled for 30 times (94°C, 1 min; 61°C, 1 min; and 72°C, 1 min) followed by a 10-min extension at 72°C. The resulting GH-1 PCR products (2155 bp) were cleaned by filtration with a Qiagen PCR purification kit (Qiagen, Basel, Switzerland) and used as templates for direct sequencing. Sequencing from both ends was performed by the dideoxy method(22) using dsDNA cycle sequencing system (Boehringer Mannheim, Rotkreuz, Switzerland; GATC GmbH, Konstanz, Germany). Furthermore, the 2155-bp GH-1 PCR amplification products were used as templates for a second PCR (direct cycle sequencing), which had a final 100-μL reaction mixture of 50 mmol/L KCl, 10 mmol/L Tris-HCl (pH 8.4), 1.5 mmol/L MgCl2, 0.1% Triton-X-100, 200 mmol/L of each dNTP, 0.3 μmol/L of each primer, and 4 U of Taq DNA polymerase (Boehringer Mannheim). Reaction mixtures were denatured for 6 min at 94°C, cycled 32 times (94°C, 1 min; 63°C, 45 s; and 72°C, 1 min), followed by a 10-min extension at 72°C. For exon 1 and 2 sense and antisene primers were used as follows: sense: (5101-5136) 5′- GTGGGGGCAACAGTGGGAGAGAAGGGGCCAGGTATA-3′; antisense: (5738-5715) 5′-AGTCTCCTCCTCTTATTTCCCAG-3′(21). Exons 3 and 4 primers were partly overlapping: sense: (5715-5738) 5′- CTGGGAAATAAGAGGAGGAGACT-3′; antisense: (6302-6279) 5′-CTAACACAGCTCTCAAAGTCAGTG-3′(21). The resulting PCR products were treated and sequenced as stated above. The different mutations found were confirmed by restriction digestion analysis as previously described(15–18). Briefly, filtration followed by sequencing using the dideoxy method and the double-stranded DNA cycle sequencing system(22).

Figure 1
figure 1

Schematic representation of the human GH gene/genomic DNA, cDNA, and protein structure. Genomic DNA: exons(I-V), introns (IVS), and nontranslated (NT) sequences are depicted by solid, open, and shaded rectangles. Nucleotides are indicated according to the sequence published by Chen et al.(21). Exon I contains 60 bp of 5′-nontranslated sequences (5′NT) and 10 bp of coding sequence. The sizes of the coding sequences (bp) of the exons are shown. Exon V codes for amino acids 127-191, and 112 bp of 3′-nontranslated (3′NT) region. cDNA and GH protein structure: Mature protein, signal peptide, and NT regions are indicated in solid and shaded rectangles. For the protein structure amino acids (aa) and the length encoded by different exons are shown.

Full size image

RESULTS

The overall prevalence of GH-1 gene alterations and the different locations of the various mutations found are shown in Table 3 and Figure 2. Mutations associated with stop codons, frameshifts, and splicing defects are indicated. In Table 4, the prevalences of GH-1 gene alterations are presented in relation to the GH peak levels after GH stimulation tests.

Table 3 Overall prevalence of GH-1 gene alterations in IGHD
Full size table
Figure 2
figure 2

GH-1 gene mutations. Schematic representation of the GH-1 gene showing the locations of various point mutations. Mutations associated with frameshifts, stop codon, splicing defects and nonsense mutations are indicated by closed arrows, open arrows, closed circles, and open circles, respectively. (a) All the mutations known so far are shown(14–18). (b) The mutations found in our study on 151 subjects with severe restricted growth due to IGHD.

Full size image
Table 4 Prevalence of GH-1 gene alterations in IGHD based on GH peak levels following stimulation-test
Full size table

Total frequency of GH-1 gene alterations and family studies. The 151 affected individual studied belonged to a total of 83 families. In addition, 155 unaffected family members were also studied. GH-1 gene deletions were present in 12.5% (19/151) of familial IGHD, in addition 6% (9/151) were caused by GH-1 gene alterations, such as point mutations and splicing errors. Furthermore, total GH-1 gene alterations varied among different populations from 11.6% in Northern Europe, 14.7% in Mediterranean countries, and 31.2% in Asia. Importantly, all families studied were unrelated, and each had one or more individuals with IGHD (Table 1). In each parent of normal stature the corresponding heterozygous pattern was found. Interestingly, there was a history of consanguinity in all of the families studied suffering from IGHD type IA (Table 4). This was not the case in type IB- and II-affected patients (Table 4). Beside a selection bias of the subjects referred for gene analysis the different degree of consanguinity in the three different population may be the major cause of the higher prevalence of patients in the Asia than in the Mediterranean and North European subpopulation with severe IGHD.

GH-1 gene alteration and clinical features according to the type of IGHD . IGHD type IA. In response to exogenous GH, the children had a strong initial anabolic response that was frequently followed by the development of anti-GH antibodies in a sufficient titer to cause an arrest of response to GH.

The basic observation in these patient was the deficiency of GH due to a deletion of the GH-1 gene, and none of the remaining genes of the gene cluster could make up for the deficit. However, there were variabilities in long-term responses to GH treatment and in the development of antibodies to the exogenous GH. Two of our patients (7.6-kb gene deletion) developed anti-GH antibodies to administered GH; unremarkable compared with other patients without GH-1 gene deletion(23). The frequency of GH-1 gene deletions as a cause for GH deficiency varied among different populations and the criteria and definition of short stature chosen(Table 2). Analyzing patients with severe IGHD (<-4 to-4.5 SD score) the frequency found was 8.7% (Northern Europeans), 11.8%(Mediterraneans), and 18.7% (Asians) (Table 3). The sizes of the deletions were heterogeneous with the most frequent (78%) being 6.7 kb. The remaining deletions described include 7.6 and 45 kb. At a molecular level, these deletions involved unequal recombination and crossing over within the GH gene cluster at meiosis(13, 24). In addition, in one patient with the IGHD type IA phenotype, PCR amplification products detected a new point mutation in codon 23 of the GH signal peptide converting GAG to a TAG (stop codon; E23X) resulting in an early termination of translation.

IGHD type IB. This phenotype was quite variable. Some affected children had very severe growth restriction resembling IGHD type IA with no detectable GH after stimulation test, and others grew relatively normal up to childhood. In contrast to type IA subjects those with type IB had neither a detectable GH-1 gene deletion nor a functionally absent peptide(e.g. mutation/deletion in signal peptide) and responded well to exogenous GH replacement. Using DNA sequence analysis of PCR amplification products only in two patients, a point mutation causing the same G → C transversion (GT → CT) was detected. This mutations altered the first base of the donor splice site of intron IV(21).

IGHD type II. Type II is autosomal dominantly inherited. Sequence analysis revealed that the mutations causing intron III donor splice site alterations are common among the population studied (6/6, 100%; in more detail: Tables 3 and 4, and Fig. 2).

IGHD type III. In one patient the X-linked form of IGHD combined with hypogammaglobulinemia was found. Although there is an X-linked recessive mode of inheritance in that disorder, we studied the GH-1 gene in this family to make certain that there was no additional GH-1 gene defect present. As expected, no GH-1 gene alteration was found.

DISCUSSION

The GH gene cluster consists of five very similar genes in the order 5′ (GH-1, CSHP, CSH-1, GH-2, CSH-2) 3′ encompassing a distance of about 65 kb on the long arm of chromosome 17 at bands q22-24(6, 7). The GH-1 gene encodes GH, a single polypeptide of 191 amino acids, that is synthesized by the somatotropic cells of the anterior pituitary gland. Previous population studies suggested that GH-1 gene deletions represent about 13% of familial IGHD cases, whereas the proportion due to other GH-1 gene alterations has not been studied yet(11, 12). The aim of our study was to evaluate in a large cohort the prevalence of GH-1 gene alterations among different populations with familial IGHD. Familial IGHD can be inherited as an autosomal recessive (type I; subdivided into type IA and IB), autosomal dominant (type II) or X-linked (type III) trait(8). In addition, these disorders differ not only in their modes of inheritance but also in their degrees of GH deficiency and responses to exogenous GH therapy(8).

Several studies have analyzed the prevalence of GH-1 gene deletions, and therefore IGHD type IA, in patients with severe IGHD (<-4 to -4.5 SD score) and reported summarized frequencies of 9.4% (Northern Europe; n = 32/340); 13.6% (Mediterranean; n = 22/162); 16.6% (Turkey;n = 24/144); 38% (Oriental Jews; n = 13/34), 12% (Chinese;n = 26/217); 0% (Japanese; n = 10), respectively(11, 12, 25, 26). The recent frequencies we found are similar and, although the sizes of the deletions were heterogeneous with the most frequent (78%) being 6.7 kb, none of our patient studied presented with a 7.0-kb human GH-1 gene deletion. Additionally, in one subject a new mutation in the signal peptide coding region was found causing IGHD type IA phenotype.

Patients with IGHD type IB are characterized by either an apparent lack of endogenous GH as determined by RIA or by low but detectable levels of GH. Furthermore, there is a failure to produce anti-GH antibodies in response to GH treatment in these patients, implying that some GH, although defective, is being or has been made. Clinically, these patients are common and present a more variable phenotype. In some families, the disorder resembles IGHD type IA with a very severe growth restriction, whereas in others, growth may be relatively normal and not noted until mid-childhood. This heterogeneity of clinical and laboratory findings support the idea that IGHD type IB might not be caused by GH-1 gene alterations only. This hypothesis is supported by our overall data, presenting only two subjects (1.3%; 2/151) with a GH-1 gene mutation responsible for the phenotype classified as IGHD type IB, whereas 12.6% of the patients studied (19/151) presented a GH-1 gene deletion causing IGHD type IA. Therefore, the whole hypothalamo-pituitary-growth axis has to be studied to define other genetic defects causing this disorder. Candidate genes include GHRH, GHRH-receptor, and perhaps many other genes interfering directly and specifically with that axis.

In contrast, IGHD type II is autosomal dominantly inherited, which may make a genetic analysis easier and more successful. All patients with a dominantly inherited GH deficiency carried at the molecular level a splicing mutation(6/6, 100%). However, there are reports of sporadic cases of severe idiopathic IGHD due to heterozygous de novo splice site mutations in the GH-1 gene causing IGHD type II(18). Therefore, to prevent any biased assignment of patients to a specific group based on the mode of inheritance the affected subject were further divided into two different groups: 1) absent GH secretion on a stimulation test (GH <0.3 ng/mL) and 2) low peak GH secretion after stimulation (>0.3 ng/mL to <4 ng/mL) (Table 4). Importantly, using this selection criteria, 4.8% of subjects studied (6/126) with deficient but detectable amounts of GH after provocative stimuli presented a GH-1 gene mutation responsible for IGHD type II. All the different forms of mutations responsible for IGHD type II have been previously reported to affect the donor splice site of intron III, resulting in an alternative use of the donor splice site of intron II in conjunction with the acceptor site of intron III(16, 17). This alternative splicing pattern deletes exon III resulting in the loss of amino acids 32 to 71, including cysteine at position 53, from the corresponding mature GH-1 product. All the affected subjects were heterozygous for the intron III base change. The mechanism by which the mutant GH allele inactivated the normal GH allele seems to involve formation of GH heterodimers through cysteine bond formation and, therefore, disruption of regular protein secretion by trapping GH in the secretory granules(17).

IGHD type III has an X-linked mode of inheritance, and the patients mainly have agammaglobulinemia associated with their disorder(8). Only one of our patients presented this pattern of GH deficiency. Although expected to be normal, the GH-1 gene was studied in this subject in order that no genetic defect within the GH-1 gene was missed.

In conclusion, we studied a large cohort of patients and their families presenting with IGHD and analyzed the frequency of GH gene deletions and mutations associated with frameshifts, splicing defects, or stop codons by sequencing the whole GH-1 gene. GH-1 gene alterations in the most severe growth-restricted patients due to IGHD varied among different populations from 11.6% in Northern Europe, 14.7% in Mediterranean countries, and 31.2% in Asia. Most striking is the low frequency rate of GH-1 gene mutations in the most common form of IGHD, namely type IB. Only in two patients (1.7%, 2/119), presenting deficient by detectable GH secretion after stimulation test (<4 ng/mL), was a mutation found. This finding highlights the need to focus on the promoter region of the GH-1 gene and, therefore on cis-acting elements and trans-acting factors. In addition, other candidate genes specific for the GH axis have to be studied to define genetically the IGHD type IB phenotype in more detail. Some of the components of the GH pathway are unique to GH, whereas others are shared by many others. In patients with IGHD, mutational changes in genes specific to GH are important,e.g. GHRH, GHRH-receptor, and last but not least the locus control region (LCR) of the GH-1 gene itself(27).