Clinical Characteristics of Pathogenic ACAN Variants and 3-Year Response to Growth Hormone Treatment: Real-World Data

Hormone Research in Paediatrics Research Article Horm Res Paediatr DOI: 10.1159/000535651 Received: September 5, 2023 Accepted: November 29, 2023 Published online: January 17, 2024 Clinical Characteristics of Pathogenic ACAN Variants and 3-Year Response to Growth Hormone Treatment: Real-World Data Judith S. Renes a, b Ardine M.J. Reedijk a Monique Losekoot c Sarina G. Kant d Manouk Van der Steen e Danielle C.M. Van der Kaay f Anita C.S. Hokken-Koelega a, f Hermine A. Van Duyvenvoorde c Christiaan de Bruin g a Dutch Growth Research Foundation, Rotterdam, The Netherlands; b Department of Pediatrics, Albert Schweitzer Hospital, Dordrecht, The Netherlands; c Department of Clinical Genetics, Leiden University Medical Center, Leiden, The Netherlands; d Department of Clinical Genetics, Erasmus University Medical Center, Rotterdam, The Netherlands; e Department of Pediatric Endocrinology, Radboud University Medical Center, Nijmegen, The Netherlands; f Department of Pediatrics, Subdivision of Endocrinology, Erasmus University Medical Center, Sophia Children’s Hospital, Rotterdam, The Netherlands; g Division of Endocrinology, Department of Pediatrics, Willem-Alexander Children’s Hospital, Leiden University Medical Center, Leiden, The Netherlands Keywords ACAN ·Aggrecan ·Growth hormone ·Treatment response Abstract Introduction: Heterozygous variants in the ACAN gene may underlie disproportionate short stature with characteristically accelerated bone age (BA) maturation and/or early-onset osteoarthritis (OA). Methods: The objective of this study was to describe phenotype, analyze genotype-phenotype correlations, and assess the response of growth hormone (GH) treatment in children with a heterozygous ACAN variant. Thirtysix subjects (23 boys, 13 girls) with ACAN deficiency and treated for ≥1yearwithGHwereidentified in the Dutch National Registry of GH treatment in children. Results: We identified 25 different heterozygous ACAN variants in 36 subjects. Median (interquartile range) height SDS at start of GH was −2.6 SDS (−3.2 to −2.2). Characteristic features such as disproportion, advanced BA, early-onset OA, and dysmorphic features like midface hypoplasia and brachydactyly were present in the majority of children, but in ~20%, no specific features were reported. Subjects with a truncating ACAN variant had a shorter height SDS compared to subjects with a non-truncating variant (−2.8 SDS and −2.1 SDS, respectively, p= 0.002). After 3 years of GH, height gain SDS in prepubertal children was 1.0 SDS (0.9–1.4). In pubertal children, height SDS remained relatively stable. Conclusion: The phenotype of subjects with pathogenic heterozygous ACAN variants is highly variable, and genetic testing for ACAN deficiency should be considered in any child with significant short stature, even in the absence of disproportion, specific dysmorphic features, or BA advancement. Furthermore, children with ACAN deficiency may benefitfrom GH with a modest but significant response, which is sustained during 3 years of treatment. © 2024 S. Karger AG, Basel Hermine A. van Duyvenvoorde and Christiaan de Bruin contributed equally to this work. [email protected] www.karger.com/hrp © 2024 S. Karger AG, Basel Correspondence to: Judith S. Renes, j.renes @ kindengroei-lrg.nl Downloaded from http://karger.com/hrp/article-pdf/doi/10.1159/000535651/4159203/000535651.pdf by ESPE, European Society for Paediatric Endocrinology user on 29 January 2024 Introduction Normal growth depends not only on the growth hormone (GH)-insulin-like growth factor (IGF) I axis but also on a normal functioning growth plate [1, 2]. Long bones grow by a process of endochondral ossification. This complex process is regulated by various endocrine and paracrine signals as well as interactions between intracellular proteins and extracellular matrix [3]. Pathogenic variants in genes involved in any of these pathways can result in a diverse group of skeletal dysplasias. One group of skeletal dysplasias results from variants in the gene coding for aggrecan (ACAN) [4]. Aggrecan is the primary proteoglycan of the cartilage growth plate [3]. Homozygous pathogenic ACAN variants lead to a severe form of skeletal dysplasia (spondyloepimetaphyseal dysplasia, aggrecan type; MIM 612813), whereas heterozygous ACAN variants can lead to much milder phenotypes, such as spondyloepiphyseal dysplasia, Kimberley type (MIM 608361), familial osteochondritis dissecans (OD) (MIM 165800), or idiopathic short stature [5–18]. Typically, the short stature is disproportionate and associated with accelerated bone age (BA) maturation, resulting in early cessation of growth and an unexpectedly short adult stature [7, 12, 15]. Other clinical characteristics that have been described are midface hypoplasia, a flat nasal bridge, brachydactyly, broad great toes, wide feet, and an increased arm span to height ratio [7, 14, 15]. However, as more individuals are being diagnosed with ACAN deficiency, the clinical phenotype has expanded in recent years. Until now, only limited data on the effects of GH treatment or other growth-promoting interventions in children and adolescents with ACAN deficiency have been reported [14, 16, 19–22]. Muthuvel et al. [19] recently described improved linear growth in 10 patients with ACAN deficiency during 1 year of GH treatment. Prior to that, van der Steen et al. [14] had shown that GH treatment increased height SDS in combination with 2 years of gonadotrophin-releasing hormone analog (GnRHa) treatment in a number of ACAN-deficient patients. Based on these early data, a standardized, national GH treatment guideline for ACAN deficiency was introduced in The Netherlands with collection of long-term follow-up data. For this study, we selected patients with pathogenic ACAN variants from the Dutch National Registry of GH Treatment in Children, described the clinical characteristics, and evaluated possible genotype-phenotype correlations. In addition, the longer-term response to GH treatment with or without GnRHa or aromatase inhibitor (AI) was evaluated. Patients and Methods Subjects The Dutch National Population-Based Registry of GH Treatment in Children was searched for children with a heterozygous ACAN gene variant. Children were included if: (1) the ACAN gene variant was classified as likely pathogenic (class IV) or pathogenic (class V) according to the ACMG guidelines [23] (throughout the manuscript, we will use the term pathogenic variant for class IV and V); and (2) at least 1 year of GH treatment was given. Children were excluded if the ACAN gene variant was classified as a variant of unknown significance (class III), additional chromosomal defects or syndromes were present, and growth failure could have been influenced by other conditions (e.g., abnormal endocrine evaluation or severe chronic illness). The ACAN variants were identified by either next-generation sequencing or Sanger sequencing (NM_001369268.1) (see online suppl. Table 1 for comparison with the previous reference sequence NM_013227.3; for all online suppl. material, see https:// doi.org/10.1159/000535651). The variants found were defined as a heterozygous deletion of the entire gene or ≥1 complete exon, a truncating variant (frameshift or nonsense) or splice-site variant located less than 2 base pairs from the exon/intron boundary, a missense variant (not found in public databases, predicted to be damaging using prediction software, de novo or segregates with short stature in the family), or an in-frame insertion, duplication, or deletion of ≥1 amino acid meeting the same criteria as missense variants [23]. Of 52 eligible children, 36 were included in the analysis. Sixteen children were not eligible and therefore not included for the following reasons: n=12<1 year of GH treatment, and n= 4 had additional chromosomal defects or syndromes (n= 1 Turner syndrome, n= 1 Wiedemann-Steiner syndrome, n= 1 SHOX haploinsufficiency, and n= 1 22q11 duplication). Four children included in this study were previously reported by van der Steen et al. [14]. The Dutch National Registry of GH Treatment was initiated in 1997; the nationwide coverage of the registry is 100%. Data in this registry are provided by members of the Dutch Society for Pediatric Endocrinology. In the registry, all data are pseudonymized to comply with rigorous privacy guidelines. For data collection in this registry, informed consent and ethical approval were not required according to Dutch law [24–26]. Measurements Height, weight, and Tanner stage were extracted from the Registry at start and subsequently yearly. Body mass index was calculated as weight divided by height squared (kg/m 2 ). Target height (TH) was calculated as TH = 44.5 + 0.376 × paternal height (cm) + 0.411 × maternal height (cm) for boys and TH = 47.1 + 0.334 × paternal height (cm) + 0.364 × maternal height (cm) for girls [27]. Height, weight, body mass index, and TH were expressed in SDS, adjusting for sex and age according to Dutch reference data [28]. Adult height SDS was calculated based on Dutch reference data at age 21 years [28]. The onset of puberty was defined as breast development stage 2 for girls and a testicular volume ≥4 mL for boys according to Tanner [29]. BA was determined either manually according to 2Horm Res Paediatr DOI: 10.1159/000535651 Renes et al. Downloaded from http://karger.com/hrp/article-pdf/doi/10.1159/000535651/4159203/000535651.pdf by ESPE, European Society for Paediatric Endocrinology user on 29 January 2024 Greulich and Pyle by the treating physician or automated using BoneXpert ® [30, 31]. An advanced BA was defined as a BA minus calendar age (CA) of at least 0.5 years [14]. Treatment As described in the Dutch national ACAN guideline, children with a class IV or V ACAN variant were treated with biosynthetic GH with a dose of 1.4 mg/m 2 /day (~0.046 mg/kg/day) subcutaneously [19, 32]. GH treatment was started if height SDS was <−2.5 SDS and/or predicted adult height was expected to be <−2.5 SDS and/or an affected parent had a height <−2.5 SDS. Patients visited the outpatient clinic every 4 months. After each visit, the dose was titrated according to the calculated body surface area, growth velocity, and/or serum IGF-I levels. IGF-I levels were determined at least yearly. Eight children were treated with a GH dose of 1 mg/m 2 /day (~0.033 mg/kg/day) because of short stature after small for gestational age birth, while at that moment, the ACAN deficiency had not yet been diagnosed. Two children were treated with a GH dose of 2 mg/m 2 /day (~0.067 mg/kg/day) according to the GH study design in which they were included [33]. Some patients came to medical attention when they were already in early puberty. In boys and girls, with a relatively short height at onset of puberty defined as a height <140 cm, postponement of puberty for 2 years using a GnRH analog was added to GH treatment (n= 9). Additionally, in 3 boys with an advanced BA and therefore an expected adult height <−2.5 SDS, adjunctive treatment with an AI was started to delay further advancement of BA. GnRHa and/or AI treatment was discussed and started at the discretion of the treating physician. Statistical Analyses Clinical characteristics and growth results are presented as median (interquartile range [IQR]), unless stated otherwise. Normal distribution of variables was tested using KolmogorovSmirnov test. Differences in characteristics between subgroups were evaluated using independent sample ttest for normal distributed variables and Mann-Whitney test otherwise. Changes in height SDS and BA over time were analyzed with repeated measure regression analysis to correct for multiple testing and missing data. For this analysis, the group was divided into children who were prepubertal at start of GH treatment and adolescents who were pubertal at start of GH. Two children in the prepubertal group started puberty during the first years of GH treatment; therefore, only their prepubertal measurements were included. The analysis for gain in height SDS was adjusted for GH dose and in the pubertal group also for GnRHa treatment. Analyses were performed using the statistical package SPSS for Windows (version 26; SPSS Inc., Chicago, IL, USA) and R, R Core Team (2022; Posit Team; RStudio: Integrated Development Environment for R. Posit Software, Boston, MA, USA). A pvalue <0.05 was considered significant. Results In total, 36 individuals (23 boys and 13 girls) from 25 different pedigrees were identified and included in the present study (Table 1). ACAN Gene Variants and Phenotypic Characteristics A total of 25 different ACAN variants were identified. There was 1 subject with a heterozygous deletion of >1 complete exon, 8 subjects with a frameshift variant, 17 subjects with a nonsense variant, 2 subjects with a splice-site variant, 7 subjects with a missense variant, and 1 subject with an in-frame insertion (online suppl. Table 1). Most children were born full term, and 6 children were born preterm (<37 weeks gestational age). Eleven were born small for gestational age, while the other ones were born with normal birth weight and/or birth length. Median (IQR) birth length tended to be in the lower part of the normal range −1.7 SDS (−2.7 to −0.4) (Table 2). Median height SDS at start of GH treatment was −2.6 SDS (−3.2 to −2.2) for the total group. Height SDS was not correlated with age (r= 0.27, p= 0.11). Sitting height-to-height ratio was ≥2.0 SDS in only 10 subjects (28%) and ≥1.0 SDS in 20 subjects (56%). In 4/ 13 (30%) children with a known arm span, the arm span (cm) was at least 3 cm larger compared to the height (cm). Dysmorphic features are described in Table 1. Midface hypoplasia was reported in 9 children. Brachydactyly was reported in 6 children, and wide feet in 5 children. Radiologically confirmed OD was present in 2 children, one had a variant in the glycosaminoglycan attachment region (subject 11a) and one in the G3 domain (subject 24). In subject 22, with a variant in the G3 domain, OD was diagnosed after start of GH treatment. OD was also described in parents with a pathogenic variant in other domains of the ACAN gene (Table 1). In ~20% (6/36) of the children, no specific clinical (advanced BA and/or increased sitting height and/or increased arm span to height ratio) or dysmorphic features were reported. In 18 children (50%), before start of GH treatment or any other growth-promoting medication, BA was advanced (≥0.5 years), of whom only 6 had a markedly advanced BA (2 years or more). BA advancement was inversely correlated with age (r=−0.39, p= 0.02), also after correcting for height SDS and year of starting GH treatment. Genotype-Phenotype Associations The genetic variants were widely distributed over the ACAN gene (Fig. 1). The missense variants were located in the G1 and G3 domains. There were 26 subjects with a truncating variant, 8 subjects with a non-truncating variant, and 2 subjects with a splice-site variant (online suppl. Table 1). ACAN Variants and 3-Year Response to Growth Hormone Treatment Horm Res Paediatr DOI: 10.1159/000535651 3 Downloaded from http://karger.com/hrp/article-pdf/doi/10.1159/000535651/4159203/000535651.pdf by ESPE, European Society for Paediatric Endocrinology user on 29 January 2024 Table 1. Subjects with a heterozygous pathogenic ACAN variant: clinical and dysmorphic features Subject Sex ACAN variant Height SDS father Height SDS mother Age at start GH Height SDS at start GH BACA at start GH SH/H ratio SDS Dysmorphic and clinical features at start of GH treatment Pubertal at start GH GnRHa/AI treatment 1 M c.1_9del p.Met1? −0.2 § −0.6 § 14.8 −3.3 −0.8 0.9 No Yes AI 2a M c.130G>A p.(Gly44Arg) −1.7 −3.8 6.1 −2.5 0.8 1.3 No No No 2b M c.130G>A p.(Gly44Arg) −1.7 −3.8 10.0 −2.3 0.6 2.8 No No No 3a F c.454+1G>A p.? −0.5 −2.7 4.2 −3.0 1.9 2.5 No No No 3b M c.454+1G>A p.? −1.7 −1.7 4.3 −3.8 0.0 1.1 Frontal bossing No No 4 M c.547G>A p.(Ala183Thr) 0.7 § 1.8 § 7.1 −1.6 1.7 −0.6 Arm span>height No No 5a F c.706C>T p.(Arg236*) 1.0 −4.2 7.8 −2.2 1.6 0.9 Mild midface hypoplasia, wide feet No No 5b M c.706C>T p.(Arg236*)NA −4.1 8.7 −2.3 0.0 1.7 Stocky build, mild midface hypoplasia, arm span>height, wide feet, brachydactyly No No 5c M c.706C>T p.(Arg236*) 1.0 −4.2 9.2 −3.1 0.0 1.8 Mild midface hypoplasia, wide feet No No 6 M c.775_784delins17 p.(Thr259 fs) −3.9 # −2.5 14.9 −5.0 −2.8 1.8 No Yes GnRHa 7 F c.1526C>A p.(Ser509*)−0.5 § −1.1 § 6.4 −3.3 2.0 0.8 No No No 8a M c.1608C>A p.(Tyr536*)−1.0 −3.4 5.0 −3.2 0.9 2.7 No No No 8b F c.1608C>A p.(Tyr536*)−1.4 −4.9 # 5.0 −3.7 0.0 2.2 Midface hypoplasia, broad great toes, and short thumbs No No 8c M c.1608C>A p.(Tyr536*)−1.4 −4.9 # 11.9 −2.4 0.6 1.4 Midface hypoplasia, mild posteriorly rotated ears, broad great toes, absent left kidney Yes GnRHa 9 F c.1631dup p.(Cys545Metfs*4) 0.9 0.4 11.2 −3.1 −1.2 0.5 Mild midface hypoplasia, Osgood Schlatter, pes plano valgus, patella dislocation, mild scoliosis Yes GnRHa 4Horm Res Paediatr DOI: 10.1159/000535651 Renes et al. Downloaded from http://karger.com/hrp/article-pdf/doi/10.1159/000535651/4159203/000535651.pdf by ESPE, European Society for Paediatric Endocrinology user on 29 January 2024 Table 1 (continued) Subject Sex ACAN variant Height SDS father Height SDS mother Age at start GH Height SDS at start GH BACA at start GH SH/H ratio SDS Dysmorphic and clinical features at start of GH treatment Pubertal at start GH GnRHa/AI treatment 10 M c.2423C>G p.(Ser808*)−1.5 § 1.3 § 7.7 −2.7 −0.3 0.7 No No No 11a F c.2735_2736delins31 p.(Leu912fs) −4.1 # −0.5 4.2 −4.8 −2.1 2.7 OD knees No No 11b F c.2735_2736delins31 p.(Leu912fs) 1.3 −2.9 # 11.9 −2.2 1.4 0.3 Brachydactyly dig I hand, sandal gap Yes GnRHa 12 M c.2770G>T p.(Glu924*) 1.2 § −0.7 § 3.3 −2.6 0.8 2.0 Frontal bossing, brachydactyly dig 5 hand, Baker’s cyst both sides No No 13 F c.4573del p.(Leu1525Serfs*11) 0.5 § −1.1 § 11.2 −2.9 1.0 0.0 Arm span>height Yes GnRHa 14 F c.4700del p.(Ser1567Metfs*10) −2.3 −1.4 6.0 −3.7 0.0 2.0 Brachydactyly dig 5 hand No No 15 M c.4762_4765del p.(Gly1588Cysfs*26) −3.7 # 0.6 12.5 −2.8 1.0 1.6 Midface hypoplasia, broad great toes, posteriorly rotated ears Yes GnRHa 16 M c.5219C>A p.(Ser1740*)−1.2 −4.9 13.4 −1.9 1.0 2.1 Lumbar lordosis, genu vara, arm span>height Yes No 17 M c.5344dup p.(Ile1782Asnfs*3) 0.1 0.0 9.3 −2.1 0.0 0.9 Almond shaped eyes, downslant, offset ears, cafe-au-lait spots No No 18a M c.6673C>T p.(Gln2225*)−2.7 −1.6 10.9 −2.2 0.3 2.4 Stocky build No No 18b M c.6673C>T p.(Gln2225*)−2.7 −1.6 12.6 −2.0 0.2 0.9 No Yes No 19 M c.7072T>C p.(Cys2358Arg) 0.9 § 0.9 § 10.1 −1.4 2.9 1.4 Wide feet, coarse bone structure hands and feet, stocky build No No 20 M c.7096C>T p.(Gln2366*) 0.1 −3.6 4.3 −3.5 1.0 2.4 Hypertelorism, cafe-au-lait spots No No 21 F c.7156_7161dup p.(Cys2386_Arg2387dup) −0.7 # −0.4 6.1 −2.4 2.5 1.7 No No No ACAN Variants and 3-Year Response to Growth Hormone Treatment Horm Res Paediatr DOI: 10.1159/000535651 5 Downloaded from http://karger.com/hrp/article-pdf/doi/10.1159/000535651/4159203/000535651.pdf by ESPE, European Society for Paediatric Endocrinology user on 29 January 2024 Height SDS was not correlated with the location of the variant on the gene (r= 0.23, p= 0.19). However, median height was −2.8 SDS (−3.4 to −2.3) in subjects with a truncating variant, compared to a median height of −2.1 SDS (−2.4 to −1.6) in subjects with a non-truncating variant (p= 0.002). BA advancement (BA-CA) was not correlated with the location of the variant on the gene (r=−0.12, p= 0.50). Median BA advancement was 0.1 years (−0.4 to 1.0) in the group with a truncating variant and 0.7 years (−0.5 to 2.3) in the non-truncating group (p= 0.21). Response to GH Treatment At start of GH treatment, median age was 9.7 years (6.0–12.4) (Table 2). Twenty-two children were prepubertal, and 14 were pubertal at start of GH treatment. Prepubertal Children Children who were prepubertal at start of GH treatment showed a significant median (IQR) increase in height SDS (ΔHSDS) of 0.6 SDS (0.5–0.8, p<0.01) after 1 year and 1.0 SDS (0.9–1.4, p<0.01) after 3 years of GH (Fig. 2). Table 1 (continued) Subject Sex ACAN variant Height SDS father Height SDS mother Age at start GH Height SDS at start GH BACA at start GH SH/H ratio SDS Dysmorphic and clinical features at start of GH treatment Pubertal at start GH GnRHa/AI treatment 22 M c.7204C>T p.(Gln2402*)−2.0 −5.6 # 12.0 −2.8 0.5 1.4 Midface hypoplasia, mild prognathism, broad great toes, exaggerated lumbar lordosis Yes GnRHa 23a F c.7255G>C p.(Asp2419His) −3.0 # −1.6 11.6 −1.6 0.5 −0.6 Cafe-au-lait spots, arm span>height Yes GnRHa 23b M c.7255G>C p.(Asp2419His) −3.0 # −1.6 14.3 −2.1 −1.0 −0.6 Hypertension, arm span>height Yes No 24 F c.7408T>G p.(Cys2470Gly) 0.2 −1.1 12.4 −2.0 −0.8 0.2 Mild midface hypoplasia, OD knees Yes No 25a F c.7602G>A p.(Trp2534*)−1.2 −3.8 # 5.9 −2.5 0.0 0.5 Frontal bossing, hypertelorism, wide feet, brachydactyly No No 25b M c.7602G>A p.(Trp2534*) 1.9 −2.2 # 14.3 −2.4 −0.5 −0.6 Offset ears No No 25c M c.7602G>A p.(Trp2534*)−0.3 −2.2 # 15.7 −3.1 −2.5 −0.7 Brachydactyly, arm span>height Yes No If one of the parents has the same ACAN gene variant as the patient, height SDS of the affected parent is in bold. If the mutation is de novo, this is indicated by a §. Persistent joint problems of the knees and hips and/or knee or hip replacement at a young age due to osteoarthritis/osteochondritis dissecans in parents are indicated by a #. If BA is advanced (≥0.5 years), this is underlined. If the SH/H ratio is increased (≥1.0), this is underlined. GnRHa/AI treatment during the first 3 years of GH treatment. AI, aromatase inhibitor; BA, bone age; CA, calender age; F, female; GH, growth hormone; GnRHa, gonadotrophin-releasing hormone analog; H, height; M, male; NA, not available; SDS, standard deviation score; SH, sitting height. 6Horm Res Paediatr DOI: 10.1159/000535651 Renes et al. Downloaded from http://karger.com/hrp/article-pdf/doi/10.1159/000535651/4159203/000535651.pdf by ESPE, European Society for Paediatric Endocrinology user on 29 January 2024 During these 3 years, the GH dose was increased in 1 subject because of a poor response to GH treatment and reduced in 5 subjects because of serum IGF-I levels >2 SDS. Adjusting ΔHSDS for GH dose did not significantly change the results (β: 0.23, p= 0.34). One subject (No. 11a) showed a suboptimal response to GH with a ΔHSDS of 0.4 SDS after 1 year and 0.3 SDS after 3 years, due to noncompliance. Fig. 1. Distribution of variants in exonic location of ACAN (NM_001369268.1) and domain structure of the aggrecan proteoglycan. The structure of the ACAN gene is shown in the upper row with exon numbers. The structure of the aggrecan protein is shown in the lower row with crucial domains drawn approximately to scale. Two structures (G1 and G3) are linked by dashed line to indicate the exonic locations of respective domains. Variants above are reported in this study. Figure adapted from Hu et al. [18]. G, globular domain; IGD, interglobular domain; KS, keratan sulfate; CS, chondroitin sulfate; CLD, C-type lectin domain; CRP, complement regulatory-like domain; EGF, epidermal growth factor-like domain. Table 2. Baseline characteristics Total group (n= 36) Prepubertal (n= 22) Pubertal (n= 14) pvalue Boys/girls 23/13 14/8 9/5 0.98 Gestational age, weeks 40.0 (38.7–41.2) 39.7 (38.2–41.3) 40.1 (38.5–41.2) 0.86 Birth weight SDS −0.3 (−1.2 to 0.1) −0.4 (−1.4 to 0.0) 0.1 (−1.1 to 0.2) 0.35 Birth length SDS −1.7 (−2.7 to −0.4) −2.0 (−4.1 to −0.2) −1.3 (−2.4 to −0.5) 0.26 TH SDS −1.2 (−1.7 to −0.3) −1.2 (−1.7 to −0.1) −1.4 (−2.3 to −0.3) 0.35 At start of GH treatment Age, years 9.7 (6.0–12.4) 6.2 (4.8–9.3) 12.5 (11.8–14.4) <0.01 Height SDS −2.6 (−3.2 to −2.2) −2.6 (−3.5 to −2.3) −2.6 (−3.1 to −2.0) 0.52 TH –height SDS 2.0 (0.6–2.5) 2.1 (1.2–2.4) 1.6 (0.1 to 2.6) 0.20 BMI SDS 0.4 (−0.2 to 1.4) 0.8 (−0.1 to 1.5) 0.1 (−0.7 to 1.0) 0.25 Sitting height/height ratio SDS 1.4 (0.5–2.0) 1.7 (0.9–2.4) 0.7 (−0.2 to 1.5) 0.01 BA –CA, years 0.4 (−0.2 to 1.0) 0.5 (0.0–1.6) 0.4 (−1.1 to 1.0) 0.06 Values expressed as median (IQR). BMI, body mass index; GH, growth hormone; SDS, standard deviation score; TH, target height. ACAN Variants and 3-Year Response to Growth Hormone Treatment Horm Res Paediatr DOI: 10.1159/000535651 7 Downloaded from http://karger.com/hrp/article-pdf/doi/10.1159/000535651/4159203/000535651.pdf by ESPE, European Society for Paediatric Endocrinology user on 29 January 2024 During GH treatment, BA did not accelerate. At start of GH treatment, BA/CA was 1.02 (1.0–1.2), after 1 year of GH treatment, 1.05 (1.0–1.2, p= 0.88), and after 3 years, 1.06 (1.0–1.1, p= 0.83). Pubertal Children Pubertal children showed an increase in HSDS (ΔHSDS) compared to baseline of 0.3 SDS (0.2–0.3, p< 0.01) after 1 year of GH treatment and 0.5 SDS (0.2–0.7, p= 0.03) after 3 years (Fig. 3, 4). In this group, 8 patients were also treated with GnRHa. In 3 patients, the GH dose during these 3 years could not be increased or was even reduced because of serum IGF-I levels >2SDS.AdjustingtheΔHSDS for GnRHa use and GH dose did not significantly change the results (β:0.07,p=0.83). One subject (No. 6) showed a suboptimal response to GH treatment with a decrease in height SDS. He was 14.9 years old at start of GH and was treated with GH 1.4 mg/m 2 /day and GnRHa simultaneously. Adult Height Ten patients reached adult height, namely, subjects 8b, 8c, 9, 10, 15, 18b, 22, 23a, 23b, and 25c (Fig. 5). Four subjects reached an adult height ≥−2SDS,andin5 subjects, adult height SDS was greater than the affected parent. In subject 10, the variant was de novo, and his final height was −2.4 SDS. Subjects 10, 18b, 23b, and 25c were treated with GH only, and they reached a height of −2.4 SDS, −1.6 SDS, −2.3 SDS, and −1.9 SDS, respectively. Subjects 8b, 9, 15, and 23a were treated with GH and GnRHa, and they reached an adult height of −4.0 SDS, −2.0 SDS, −2.6 SDS, and −2.4 SDS, respectively. Subjects 8c and 22 were both treated with GH 2 mg/m 2 /day, GnRHa, and AI, and they reached an adult height of −1.3 SDS and −3.1 SDS, respectively. Safety GH treatment was overall well tolerated. Two pubertal subjects developed a thoracic lumbar scoliosis during treatment: in one subject, this was mild and no intervention was needed, and in the other subject, a brace was prescribed. Serum IGF-I values varied during GH treatment. Values >2 SDS were observed in 9 individuals (6 prepubertal and 3 pubertal). The IGF-I levels normalized without intervention or after reduction of the GH dose. Fig. 2. Gain in height SDS during 3 years of GH treatment in prepubertal children (≥1 year of GH in 22 subjects, ≥2 years n= 17, and ≥3 years n= 10). GH, growth hormone; SDS, standard deviation score. 8Horm Res Paediatr DOI: 10.1159/000535651 Renes et al. Downloaded from http://karger.com/hrp/article-pdf/doi/10.1159/000535651/4159203/000535651.pdf by ESPE, European Society for Paediatric Endocrinology user on 29 January 2024 Discussion In this study, we present detailed clinical characteristics of a large group of patients with pathogenic heterozygous ACAN variants who were followed longitudinally, thereby expanding our knowledge of this clinical phenotype. Twenty-two of the 25 variants were not previously reported [14]. Although the majority of children displayed mild dysmorphic features, 20% of children did not have any clinical signs suggestive of a skeletal dysplasia and presented with isolated short stature. To our knowledge, this is the first cohort of ACAN-deficient children who were treated according to a standardized, national GH treatment protocol with 3-year follow-up. We found that GH treatment had a modest but significant growthpromoting effect in the majority of children, which was maintained during 3 years of GH treatment. The ACAN gene is located on the long arm of chromosome 15 and consists of 19 exons (NM_001369268.1), with the first exon being noncoding [34, 35]. Only patients with de novo or segregating (likely) pathogenic ACAN variants were included in our study. The 25 variants found were spread throughout the gene and included mostly truncating or splice-site variants (n= 19) and less often a missense or in-frame variant (n=6) (Fig. 1). We found that subjects with a truncating ACAN variant had a significantly shorter stature compared to those without a truncating variant. These findings are in line with the results of Wu et al. [16]. In contrast to Wu et al. [16], we did not observe that truncating variants led to a more advanced BA. Interpretation of variants in the ACAN gene, especially non-truncating variants, can be difficult. Missense variants are often classified as a variant of unknown significance. In line with previous reports, the missense variants found in our cohort were located in the folded, globular domains of the aggrecan core protein (G1 and G3) [36]. This may reflect that a missense variant in any of the glycosaminoglycan attachment sites of the nonfolded parts of aggrecan only has a marginal effect on aggrecan function and would not result in a clear phenotype [36]. On the other hand, the globular domains mediate important interactions for cartilage extracellular matrix assembly and organization [36, 37]. Indeed, Stattin et al. [36] found a link between missense variants affecting the G3 domain of ACAN and OD. We also found that subjects and parents with a pathogenic ACAN variant in the G3 domain presented with OD/early-onset Fig. 3. Gain in height SDS during 3 years of GH treatment and additional GnRHa/AI treatment in pubertal children (≥1yearofGHin14 subjects, ≥2yearsn=11,and≥3yearsn=7). AI, aromatase inhibitor; GH, growth hormone; GnRHa, gonadotrophin-releasing hormone analog; SDS, standard deviation score. ACAN Variants and 3-Year Response to Growth Hormone Treatment Horm Res Paediatr DOI: 10.1159/000535651 9 Downloaded from http://karger.com/hrp/article-pdf/doi/10.1159/000535651/4159203/000535651.pdf by ESPE, European Society for Paediatric Endocrinology user on 29 January 2024