SHOX Variations in Idiopathic Short Stature in North India and a Review of Cases from Asian Countries

Abstract

Objective:

Short stature homeobox (SHOX) haploinsufficiency underlies idiopathic short stature (ISS) and Leri-Weill dyschondrosteosis. The worldwide prevalence of SHOX variations in ISS varies from 2.5% to 15.0%. The aim of this study was to assess the implication of SHOX variation in ISS in North Indians and compare this with other cases of SHOX variations from Asian population.

Methods:

SHOX gene analysis was carried out by multiplex ligation-dependent probe amplification followed by Sanger sequencing in 54 patients with variable phenotypes. Comparison with other reports in a meta-analysis comprising the current study and 11 previous studies (n=979) was performed.

Results:

SHOX analysis resulted in 12.9% positivity (7.4% deletions and 5.5% duplications). SHOX association was seen significantly related to gender, with predominance in females (p=0.047). Short arms and forearms were the only significantly associated trait seen in 51.9% of children. The overall prevalence of SHOX variation was 15.2% in Asians with ISS. No significant difference was found in geographical region-specific analysis.

Conclusion:

This study summarises findings from the last decade and provides an updated picture of the prevalence of SHOX variations in Asians, emphasizing their potential as therapeutic targets in ISS patients. Further high quality, large investigations including functional validation is warranted to validate this association.

What is already known on this topic?

The worldwide prevalence of SHOX variations in idiopathic short stature (ISS) varies from 2.5% to 15%. SHOX variants are one of the most common causes of ISS. SHOX deletions are the most common type of variation encountered worldwide.

What this study adds?

This study highlights the role and prevalence of SHOX variations in ISS in a North Indian population. In addition, a meta-analysis is presented which compiled findings from the last decade and provides an updated insight into the prevalence of SHOX variations in the whole Asian population, underscoring their potential as therapeutic targets in ISS patients.

Introduction

Idiopathic short stature (ISS) is defined as height which is more than two standard deviations (>2 SD) below the corresponding average height in a given age, sex, and population without features suggestive of specific causative disorders (1). ISS affects around 2.0% of children and is a genetically heterogeneous condition. With the advance in techniques, a battery of genes is known to be involved in longitudinal growth in human beings, but despite this, genetic causes constitute only a small proportion of cases of short stature (2). One of the genes closely associated with human short stature is the short-stature homeobox-containing gene (SHOX), alternatively known as pseudoautosomal homeobox-containing osteogenic gene (3). The importance of this gene is highlighted it is one of the most common genetic causes leading to short stature in children, with either isolated or familial short stature.

The location of SHOX is characteristic and is responsible for the variable phenotype. Being located over the tip of the short arms of both sex chromosomes X and Y inside the telomeric region of the pseudoautosomal region 1 (PAR1), named due to its presence over both these chromosomes (4). The location is also pivotal as genes in this area escape the process of lyonization. SHOX has been shown to function in a dose-dependent manner, that is loss-of-function variation which involves only a part of the gene, thus involving one of the SHOX alleles, technically known as haploinsufficiency, can lead to a wide variety of phenotypes and resulting in short stature of varying degrees (5). Furthermore, SHOX deletions are the most common type of variations encountered worldwide, known to account for 80.0% of all variations (6). These deletions can be of varying sizes and can involve either all or part of the gene or regions above and below it, which contain a regulatory enhancer region (7). Other types of variations reported include nonsense and missense sequence alterations. Most common variations are seen in exons 3 and 4, which are known to encode the functional part of this pivotal homeodomain (7).

In different studies from around the world, estimates of the presence of SHOX gene variation leading to ISS range from 2.5% to 15.0% (8,9,10,11). However, the prevalence of ISS in India due to SHOX variations is not known. In the last few decades, many studies have been done on the association between different gene variations and ISS. The prevalence of SHOX variation depends on ethnicity and selection of patients. Some studies from Asian populations have reported important frequencies in comparison to the worldwide populations. In the present study, individuals with ISS, with height <2 SD of population reference range height, with no identified cause were included. These selected individuals had detailed anthropometric evaluation, dysmorphism evaluation and skeletal survey by radiographs. Multiplex ligation-dependent probe amplification (MLPA) and Sanger sequencing of SHOX exons 2-6 were done to identify deletions and sequence variations, respectively. In addition, a meta-analysis of other cases from Asia with SHOX variations was performed to estimate and evaluate the relationship between SHOX variation and ISS in these populations.

Methods

Subjects

This was a prospective study conducted in a tertiary hospital from July 2020 to March 2022, with approval from the Ethics Committee of the Postgraduate Institute of Medical Education & Research (PGIMER) (approval no: NK/6656/MS/315, date: 01.12.2020). A detailed physical examination of children with ISS was performed to look for dysmorphic features, such as micrognathia, high-arched palate, short forearm, cubitus valgus, short hands, characteristic Madelung deformity, dislocation of ulna at the elbow, short lower leg, bowing of tibia, genu valgus, short feet, scoliosis, and/or muscular hypertrophy. Rappold scoring was done for better delineation of the clinical expression in our cohort (12). The children included in the study were measured for the following anthropometric parameters using standardized techniques and instruments, in the growth laboratory/clinic: height (cm); weight (kg); sitting height (cm); and arm span (cm). Body mass index (BMI; kg/m²) and arm span to height ratio, as well as upper segment to lower segment ratio, was also calculated. Written informed consent was obtained from each patient or their parents or legal guardians, in the case of minors after explaining the purpose and type of all procedures used.

Molecular Analysis

Genomic DNA was extracted from peripheral blood by QIAamp DNA Blood Midi Kit (Qiagen, Hilden, Germany) for subsequent molecular analysis. SHOX analysis was performed using MLPA technique using a SALSA P018-G2 kit (MRC-Holland, Netherlands) as per the manufacturer’s instructions. The entire coding region of the SHOX gene (exons 2-6) and splice junctions were sequenced by Sanger sequencing using an automatic sequencer, the ABI PRISM 3100 Genetic Analyzer (Applied Biosystems, California). Primer sequences are available upon request.

Meta-analysis

Meta-analysis of prevalence was done using the MedCalc software, version 20.111 (https://www.medcalc.org/). This study followed the PRISMA guidelines (13). To identify the relevant publications concerning SHOX gene variations in ISS in Asia, the databases PubMed, Scopus and Web of Science were investigated. The search string was a combination of the following keywords: “Idiopathic Short Stature”, “SHOX”, “deletions/duplications/variations”, “copy number variations”. All the studies published in last 10 years (2012-2022) were selected for this study. The included studies met the following eligibility criteria: region - Asia; period - last 10 years; language - English; and peer-reviewed publication. Included studies were original research articles, related to SHOX prevalence, variations and/or incidence, along with the number of individuals included. The authors independently assessed the risk of bias for the included studies, using the Newcastle-Ottawa Scale (NOS) tool for quality assessment (14).

Statistical Analysis

The strength of the association between SHOX variation and ISS in the Asian population was evaluated by calculating the odds ratio using the Mantel-Haenszel statistics method. A fixed/random-effect model was applied, along with a corresponding 95% confidence interval (CI). Fixed-effects models were used to assess the pooled prevalence of genes for results with low heterogeneity (I² ≤50%). Otherwise, random-effects models were applied for the analyses. MedCalc software was used for all data synthesis and statistical analyses. Chi² and I² statistics were used to calculate heterogeneity across individual studies and subgroups. Population bias was assessed by Funnel-plot analysis. In addition, subgroup analysis was performed based on ethnicity or regional variation between subgroups. All statistical analyses were performed by using IBM Statistical Package for the Social Sciences (SPSS) Statistics for Windows, version 22.0 Armonk, NY: IBM Corp.

Results

A total of 54 patients were enrolled. Of these, 5 (9.3%) were in the age group 1-5 years, 20 (37.0%) were in the age group 5-10 years, 26 patients (48.0%) were 10-15 years of age and 3 patients (5.5%) were above 15 years of age. The median (interquartile range) age of children in this study was 11years (8 years, 13 years). The study included 26 females (48.1%) and 28 males (51.9%). Phenotypic characteristics of enrolled patients are given in Table 1. Short upper limbs (arms and forearms) was the most consistent feature (28/54, 51.85%) and was more common than short lower limbs (24/54, 44.4%). Short upper limbs was a significant finding in the variation positive subjects (Table 1). Mean, SD and range for the measured anthropometric parameters among the study cohort is shown in Supplementary Table 1. Most of the anthropometric parameters analyzed using SPSS software represented a moderately skewed distribution (almost towards normal distribution for practical purposes) with a skewness index ranging between +1 to -1. Detailed physical growth and pubertal characteristics of study subjects who were positive for SHOX variants (n=7, 13%) are depicted in Table 1. Interestingly, four out of these seven cases had initially presented to the endocrinology OPD of Department of Pediatrics and were advised growth hormone therapy. However, due to financial constraints or being lost to follow-up, growth hormone was not initiated in these children, except for one case (Patient 3) in whom a marked increase in height gain was evident [from -6.4 SD score (SDS) at 8 years to -3.3 SDS at 13 years].

Table 1. Association of phenotypic features with SHOX mutation.

Molecular Analysis Results

SHOX del/dup was found in 12.9% patients (7/54) using MLPA, including heterozygous deletion of exons in 4 (7.4%), and duplications in 3 (5.5%). MLPA results of positive patients are given in Table 2 and Supplementary Figure 1. In the family of Patient 1, her mother and 17-year-old elder sister were also short. Mother was 139 cm (-3.73 SD) and elder sister was 140.5 cm (-3.46 SD) in height; they also had the same heterozygous SHOX exon 4 deletion. Father (171 cm) was also tested but did not have the deletion. In the family of Patient 4, in whom a heterozygous SHOX exon 4 deletion was detected, the mother was 128.5 cm (-5.31 SD). The family however did not consent for further testing in the mother. There was an apparently healthy younger brother with no short stature. In the other five families, two families had no other offspring and the other three families had one more offspring each with no history of short stature. From the above, it is apparent that 1 out of 7 (14%) copy number variations was familial. Sanger sequencing revealed no sequence variations in any of the MLPA negative patients. In two female patients with heterozygous exon 4 deletion on MLPA, Sanger sequencing revealed no sequence variations, thus ruling out point mutation at the ligation site as the cause.

Table 2. Clinical and molecular features of the children with SHOX deficiency.

Clinical Characteristics

Clinical parameters of the seven patients with SHOX variants, their Rappold scores (Supplementary Table 2) and phenotypic features are given in Table 1.

Among the 54 cases, SHOX gene variation was present in 3.5% of boys and 23.1% of girls enrolled. This was a significant relationship between SHOX gene variations and female sex (p=0.047) (Table 1).

From among all the traits which were noted among the enrolled patients, short arms and forearms were significant and related to those with ISS due to SHOX variations, whereas the other phenotypic characteristics were not found to be linked with variation. The phenotypic characters and their respective association are summarized in Table 1. No significant association was found for height, mid-parental height, weight, BMI, upper segment to lower segment ratio, or arm span to linear height ratio (data not shown). Radiological survey was done using X-ray of wrist, arms and forearm, legs, thoracolumbar spine and chest X-ray. However, no significant association was established between abnormal radiological findings and SHOX gene variation.

Meta-analysis

A total of 852 studies were retrieved. After the removal of duplicates (692 studies), 160 studies including the present study were considered potentially eligible for evaluation, but 149 did not meet the inclusion criteria, leaving 11 studies for analysis (Figure 1). This meta-analysis comprised 10 previous studies and the present study with a total of 979 participants. Detailed characteristics of the studies are provided in Table 3 with NOS scoring. Studies were conducted in six different Asian countries, which were further subcategorised into South, West and Eastern Asia. Most of the studies used MLPA as a major method, followed by Sanger sequencing, chromosomal microarray (CMA) and fluorescent in situ hybridisation analysis.

Figure 1.

Flow chart for selection of studies (PRISMA 2020)

Table 3. Characteristics of studies included for analysis.

Variations in SHOX were identified in 83 of 979 patients. Using random effect model, the mean prevalence of SHOX variations was 14.4% (95% CI, 1.0-3.0, p<0.001, I²=95.1%) (Figure 2A, Table 4). The Funnel-plot was asymmetric, suggesting the possibility of publication bias Figure 2B. Further prevalence and association were calculated on the basis of the different regions of Asia, divided into South Asia (patients n=134), West Asia (n=226) and East Asia (n=619). No significant differences were observed in these rates in the different regions of the Asia. The mean frequency of SHOX variations in South Asia was 10.4%, with this rate being higher than the rates in other regions; East Asia (8.4%) and West Asia (7.5%) (Figure 3, Table 5).

Figure 2.

(A) The prevalence of SHOX variations is 14% (95% CI: 5.03-27.0, p<0.01, I²=95%). (B) Funnel-plot for SHOX variation studies to analyse risk bias

CI: confidence interval

Table 4. Prevalence of SHOX mutation in Asian population along with test for heterogeneity and publication bias.

Figure 3.

Prevalence of SHOX variations in ISS. (A) All studies were grouped according to the geographical origins of the patients. On the map of the world are shown the frequencies (%) of SHOX variations in South Asia, East Asia, and West Asia. (B) Dotted lines correspond to the overall prevalence and the 95% CI

ISS: idiopathic short stature, CI: confidence interval, I²: inconsistency index

Table 5. Prevalence of SHOX mutation in South, West and East Asian population.

Discussion

SHOX gene deficiency is one of the single gene disorders of bone metabolism resulting in highly variable osteodysplasia and so affecting the overall height of affected children and adults. However, as understanding of the gene and allelic alterations in the gene have evolved over time, it is now known that isolated deficiency of either the SHOX gene or its modifiers up and down the PAR region are responsible for a spectrum of disorders, ranging from simple ISS to more severe disorders, like Leri-Weill dyschondrosteosis (LWD, MIM 127300) and Langer Mesomelic dysplasia (LMD, MIM 249700)”. In the present study, SHOX variants were found in 12.9% of patients using MLPA and Sanger sequencing, further sub-divided into heterozygous deletion of exons (n=4), and duplications (n=3).

In previous studies conducted earlier the prevalence of SHOX variations in children with ISS ranges from 2.0-15.0% (8,9,10,11). In a study by Hirschfeldova et al. (10), MLPA analysis detected SHOX gene anomalies in 13.7% (n=7), including 4 deletions, 1 duplication with an “ambiguous effect” and 2 SHOX gene point variations. In the same study, single isolated enhancer duplication was also observed in the LWD group. In the present study, three duplications were detected, with no apparent enhancer effect. In the present study, SHOX gene association was significantly (p=0.047) related to gender, with predominance in female patients but there was no significant age correlation. Short arms and forearms were the only trait seen in 51.9% of the children and this was significantly (p=0.024) related to SHOX gene variation, with the remaining phenotypic traits were not being significantly associated. The seven children with SHOX variants were investigated for phenotypic-genotypic correlation. Copy number changes involving the conserved non-coding DNA element (CNE) both upstream and downstream have been described in ISS (15). Duplications are very rare. In the present study, there were three heterozygous duplications, one involving the whole gene, one the first three exons and upstream and the third, downstream. All duplications gave rise to ISS. The transcriptional regulation of SHOX is highly complex. Genomic studies have identified multiple CNEs downstream of SHOX with there being a few upstream CNEs. CNEs have been reported to not always be highly conserved (5).

For better delineation of the clinical expression in our cohort, Rappold scoring was performed. Interestingly, the scores were very variable with a median (range) of 3 (0-10). Short forearm was the most commonly observed feature (62.5%), while cubitus valgus was observed in two patients. All reported BMIs were less than 50^th percentile. Unexpectedly, abnormal arm span/height ratio was uncommon (28.5%), whereas bowing of forearm and tibia was present in only one patient. Sitting height/height ratio (>55.5%), dislocation of ulna at the elbow and muscular hypertrophy were absent in our cohort.

In different studies conducted in different populations, the prevalence of SHOX variations in ISS children again varies widely. Stuppia et al. (16) reported around 7.0% of patients with ISS having SHOX gene deletions, while Musebeck et al. (17) investigated 35 patients with ISS, none of whom had the SHOX deletions.

This meta-analysis aimed to investigate the association between the SHOX variation and ISS in Asia. Data on 979 ISS cases was analysed. The overall prevalence of SHOX variation in Asia was 14.3%. Subgroup analysis showed the presence of SHOX variations in 10.4% of the patients with ISS in South Asia. Among South Asians, Kumar et al. (18) reported pathogenic heterozygous variants in 4 children (6.5%) out of 61, exon 5 duplication, splice site variation c.278-1G>C, one partial deletion and complete deletion of SHOX. Singh et al. (19) showed SHOX haploinsufficiency in two patients (10.5%), while one patient (5.2%) had mosaic gain in SHOX out of 19 patients.

In a West Asian population, Delil et al. (20) identified one patient (2.6%) with SHOX variation. In contrast, Bakır et al. (21) found no variation in the SHOX gene in 37 patients. Alharthi et al. (22) found only one variation in exon 4 of SHOX while rest of patients had polymorphisms in exons 1, 2, 4, and 6. Gürsoy et al. (23) found three point variations and one whole gene deletion in 15 patients from four different families. The overall prevalence of SHOX variants in ISS in West Asia was around 7.5%.

In Eastern Asia, Fukami et al. (5), reported six rare CNVs in PAR1 in 245 patients. Lee et al. (24), confirmed SHOX deficiency in 23 patients from 15 unrelated families. In a study by Tung et al. (25), SHOX intragenic deletions were found in five patients, one deletion in the regulatory region, and a missense variation at exon 5. Prevalence of variation in SHOX gene varied between different geographical regions of Asia, being highest in South Asia. Shima et al. (26) reported SHOX abnormalities in 3.8% of ISS and 50.0% of LWD cases. These results indicate the difference in the prevalence of the SHOX variations based on selection criteria, methodology, different sample size, and ethnicity.

Study Limitations

This was a hospital-based study and results should be interpreted in that context. Our study had a relatively small sample size, as it was a single centre study, and as it was a dissertation study, had to be concluded over a limited time period.

Conclusion

In the cohort of North Indian children with ISS, the prevalence of SHOX variants was 12.9%. This was consistent with the subgroup analysis of studies from this region. The meta-analysis, a compilation of findings from the last decade across western, southern and eastern Asia, presented an updated picture of overall prevalence of SHOX variations in Asians, underscoring its potential as a main target in ISS patients. Further investigations of higher quality, large cohort size with functional validation are warranted to validate this association.

Click the link to access Supplementary Figure 1, Supplementary Tables 1, 2.

Access link