Abstract

Context

   There is growing evidence of the role of epigenetic regulation of
   growth, and miRNAs potentially play a role.

Objective

   The aim of this study is to identify changes in circulating miRNAs
   following GH treatment in subjects with isolated idiopathic GH
   deficiency (IIGHD) after the first 3 months of treatment, and verify
   whether these early changes can predict growth response.

Design and Methods

   The expression profiles of 384 miRNAs were analyzed in serum in 10
   prepubertal patients with IIGHD (5 M, 5 F) at two time points before
   starting GH treatment (t−3, t0), and at 3 months on treatment (t+3).
   MiRNAs with a fold change (FC) >+1.5 or <-1.5 at t+3 were considered as
   differentially expressed. In silico analysis of target genes and
   pathways led to a validation step on 8 miRNAs in 25 patients. Clinical
   and biochemical parameters were collected at baseline, and at 6 and 12
   months. Simple linear regression analysis and multiple stepwise linear
   regression models were used to explain the growth response.

Results

   Sixteen miRNAs were upregulated and 2 were downregulated at t+3 months.
   MiR-199a-5p (p = 0.020), miR-335-5p (p = 0.001), and miR-494-3p (p =
   0.026) were confirmed to be upregulated at t+3. Changes were
   independent of GH peak values at testing, and levels stabilized after
   12 months. The predicted growth response at 12 months was considerably
   improved compared with models using the common clinical and biochemical
   parameters.

Conclusions

   MiR-199a-5p, miR-335-5p, and miR-494-3p changed after 3 months of GH
   treatment and likely reflected both the degree of GH deficiency and the
   sensitivity to treatment. Furthermore, they were of considerable
   importance to predict growth response.

   Keywords: miR-199a-5p, miR-335-5p, miR-494-3p, growth, GH deficiency,
   GH treatment

1 Introduction

   During the last decade, knowledge on epigenetics has increased, and in
   this context, microRNAs (miRNAs) have attracted the interest of
   researchers given their role as key regulators of multiple biological
   processes. MiRNAs are endogenous small non-coding RNAs that act as
   transcriptional ([47]1, [48]2) and post-transcriptional regulators
   ([49]3). Multiple changes in miRNA abundance can occur, where
   simultaneously up- and downregulated miRNAs can target the same gene
   with a range of predicted effects and, vice versa, a single miRNA can
   regulate several target genes ([50]4). To date, the miRNA network is
   considered of fundamental importance for the regulation of gene
   expression ([51]5). MiRNAs are key regulators of metabolic pathways
   ([52]6–[53]9), and are currently studied as biomarkers of disease and
   response to drug administration ([54]10, [55]11). Evidence on
   longitudinal growth regulation by miRNAs has been reported in different
   in vitro and animal models ([56]12, [57]13) and miRNAs have been
   described to contribute to the regulation of the
   hypothalamic–pituitary–IGF axis and to growth plate function ([58]12).
   Currently, only one study has shown that miRNAs change under conditions
   of dysregulated growth hormone (GH) levels in humans ([59]14). One in
   vitro study highlighted that the GH receptor can be regulated by
   specific miRNAs, suggesting that this regulatory system could be of
   importance for the GH axis ([60]15). Finally, one recent study
   described that circulating miRNAs in adult patients and mice with
   congenital GH deficiency were regulated in relationship with aging
   ([61]16). However, so far, no studies have investigated the changes in
   miRNA circulating levels in response to GH treatment in childhood.

   The growth response in patients on GH treatment is variable depending
   both on the patient’s basal conditions and on personal innate
   sensitivity to therapy ([62]17). Often, the measured growth rate does
   not coincide with the expected one and the degree of correlation
   between clinical–auxological parameters and dose and GH peak vary
   enormously, both inter- and intra-individually during treatment. In
   this context, some patients run the risk of receiving an excessively
   low or high GH dose ([63]18–[64]20). Currently, in the attempt to
   improve the growth response, some medical devices on web platforms have
   become available in clinical practice ([65]21). However, these
   interactive tools use universal algorithms based on growth prediction
   models built by collecting clinical data stored in international
   databases, and they can be used only after the first year of treatment.

   This study aimed to identify changes in circulating miRNAs following GH
   treatment in children with isolated idiopathic GH deficiency (IIGHD)
   after the first 3 months of treatment, to explore their associations
   with clinical and biochemical parameters during the first 12 months of
   therapy, and to test the ability of early changes in these selected
   miRNAs to predict the clinical outcome in terms of growth on GH
   treatment.

2 Patients and Methods

2.1 Patients

   Ten prepubertal children at diagnosis of idiopathic isolated GH
   deficiency were enrolled for a preliminary profiling step
   [chronological age (CA): 8.80 ± 2.60 years; 5 male patients (M) and 5
   female patients (F)]. Twenty-five prepubertal patients were included in
   the following validation step (CA: 9.08 ± 3.05 years); the main
   characteristics of the study cohort at diagnosis are reported in
   [66]Table 1. All subjects were diagnosed with isolated idiopathic
   growth hormone deficiency (IIGHD) according to the official indications
   ([67]22) and remained prepubertal throughout this 12-month study. At
   diagnosis, 24 subjects underwent an arginine stimulation test, and 1
   underwent a clonidine stimulation test as the first test. Nineteen
   underwent a glucagon stimulation test and six underwent a clonidine
   stimulation test as the second test. All the subjects underwent a
   magnetic resonance imaging (MRI) scan of the hypothalamus and pituitary
   gland. Patients were enrolled at the pediatric endocrine centers in
   Reggio Emilia and Modena. Patients with ascertained or probable genetic
   syndromes (e.g., skeletal dysplasia, Silver-Russell syndrome) and/or
   obesity were excluded to further reduce the chances of confounding
   factors. The patients were treated with GH, according to the
   indications of the Italian Regulatory Agency (AIFA Note 39) and
   underwent routine practice for treatment. Both biosimilar and
   recombinant human GH were used.

Table 1.

   Auxological and biochemical features of patients at baseline, and at 6
   and 12 months of GH treatment.
   Baseline 6 months of treatment 12 months of treatment
   Sex, M/F 17/8
   CA, years 9.08 ± 3.05
   Target height, cm 166.1 ± 8.12
   Target height SDS −0.96 ± 0.81
   Highest GH peak N<5 ng/ml/N>5 ng/ml 7/18
   GH peak at first test, ng/ml 4.23 ± 2.06
   GH peak at second test, ng/ml 5.14 ± 2.01
   Bone age, years 7.35 ± 2.77 8.41 ± 2.88
   Height, cm 119.22 ± 16.79 124.22 ± 16.88 127.96 ± 16.78
   Height SDS −1.92 ± 0.37 −1.61 ± 0.39^* −1.48 ± 0.39^#
   Weight, kg 23.17 ± 7.39 25.38 ± 8.51 27.39 ± 9.10
   Weight SDS −1.79 ± 0.74 −1.62 ± 0.79^* −1.45 ± 0.79^#
   BMI, kg/m^2 15.85 ± 1.48 15.92 ± 1.78 16.21 ± 1.93
   BMI SDS −0.49 ± 0.79 −0.61 ± 0.88 −0.55 ± 0.91
   Growth velocity, cm/year ^§ 4.49 ± 1.59 8.03 ± 1.69 7.19 ± 1.51
   Growth velocity SDS ^§ −1.60 ± 1.04 2.78 ± 1.99^* 1.82 ± 2.40^#
   IGF-I, ng/ml 150.96 ± 62.86 281.22 ± 127.69 284.33 ± 113.92
   IGF-I SDS −0.03 ± 0.59 0.88 ± 0.76^* 0.79 ± 0.63^#
   Fasting blood glucose, mg/dl 81.44 ± 4.75 85.44 ± 8.61 86.07 ± 7.14
   Insulin, µU/ml 7.73 ± 4.41 8.83 ± 5.30
   HbA1c, mmol/mol 33.18 ± 3.00 32.83 ± 2.96
   Alkaline phosphatase, U/L 267.43 ± 136.97 603.59 ± 216.75 448.33 ±
   227.44
   Drug, N recombinant/N biosimilar 11/14
   GH dose, mg/kg/day 0.028 ± 0.004 0.023 ± 0.004 0.023 ± 0.004
   Hypothalamus–pituitary MRI N = 1 Rathke’s cleft cyst
   N = 2 small pituitary gland
   N = 22 normal
   [68]Open in a new tab

   BMI, body mass index; CA, chronological age; F, females; FBG, fasting
   blood glucose; GH, growth hormone; HbA1c, glycated hemoglobin; IGF-I,
   insulin-like growth factor 1; M, males; N, number; SD, standard
   deviation; SDS, standard deviation score; U, units. ^§calculated on the
   previous 6 months. *p < 0.0001 baseline vs. 6 months on treatment. ^#p
   < 0.0001, baseline vs. 12 months on treatment. Data are reported as
   mean ± SD.

   For the purpose of analyses, patients were also subdivided according to
   GH peak concentrations (highest peak > or < 5 ng/ml) at testing, and
   based on response to GH treatment [change in height (Ht) SDS after 12
   months on treatment > or < +0.3 SDS, defined as responders and
   non-responders, respectively) ([69]23).

   The study was approved by the local Ethical Committee (Study title:
   “Role of miRNAs as predictors of response to growth hormone (GH) in
   patients with GH deficiency” Prot no. 2016/0002409) at the
   Institutions. Written informed consent was obtained from all
   participants and their parents as appropriate.

   The general workflow of the study is described in [70]Figure 1.

Figure 1.

   Figure 1
   [71]Open in a new tab

   Study workflow. In the discovery step, a profiling approach was used,
   and all those miRNAs showing spontaneous variations independent of
   treatment (−3 and 0 months) were excluded from further analyses.
   Eighteen miRNAs were found to show changes at +3 months. Based on their
   function, 8 miRNAs were selected out of this initial pool of 18. The
   validation phase used TaqMan Real-Time qRT-PCR approach and studied the
   response of these 8 miRNAs at 3 months on GH treatment in 25 patients.
   Three specific miRNAs were found to change significantly on treatment,
   and were included, together with other clinical and biochemical
   variables, to contribute to explain growth after 1 year of treatment.
   To further understand miRNA changes on treatment, these were further
   measured in serum at 12 months. Green stars in the figure represent
   time points in which clinical and biochemical data have been recorded
   (t0, t+6, t+12). BA, bone age; CA, chronological age; F, females; GH,
   growth hormone; M, males; n, number; pts, patients; yr, years. Created
   with [72]BioRender.com.

2.2 Sample Processing and Total RNA Isolation

   Whole blood was drawn in BD Vacutainer Serum Separator Tubes, and it
   was processed within 2 h from collection and after overnight fasting,
   between 7:30 and 8:30 a.m. Whole blood was then centrifuged at 2,000 g
   for 10 min at 4°C. Serum was aliquoted in 1.5-ml sterile RNase-free
   tubes and further centrifuged at 2,500 g for 10 min at 4°C to remove
   any contaminant cells and debris. Serum was then collected in sterile
   RNase-free tubes and stored at −80°C until use. Blood samples were
   collected at two time points before the beginning of treatment (3
   months before, t−3, and just before the treatment, t0), and at 3 and 12
   months after the beginning of the treatment (t+3 and t+12) in the
   context of routine controls. Total RNA was isolated from 400 μl of
   serum using the miRVana PARIS kit (Invitrogen Cat No. AM1556) according
   to the manufacturer’s protocol and the eluate was stored at −80°C. RNA
   was reverse-transcribed using the TaqMan™ Advanced miRNA cDNA Synthesis
   Kit (Applied Biosystems Cat No. A28007) following the manufacturer’s
   instructions.

2.3 Discovery Step: miRNA Expression Profiling

   The expression profiles of 384 miRNAs were analyzed in 10 patients, of
   which 5 were male patients and 5 were female patients (10 samples
   collected at t−3, 10 samples at t0, and 10 samples at t+3) to avoid
   gender-specific miRNAs, using TaqMan Advanced miRNA Human A Cards
   (Applied Biosystems Cat No. A34714) that contain 384 miRNA assays.
   Briefly, 2 μl of RNA eluate was reverse-transcribed to cDNA, using the
   TaqMan Advanced miRNA cDNA synthesis kit (Applied Biosystems Cat No.
   A28007) following the manufacturer’s instructions in the Thermal Cycler
   T100 (Bio-Rad). The cDNA was loaded into the TaqMan Advanced miRNA
   array Card A and run in a 7900HT Fast PCR system (Applied Biosystems).
   Array data were normalized using the standard internal reference
   hsa-miR-16-5p (Assay ID: 477860_mir) as endogenous control (24); to be
   sure about its validity, we selected hsa-miR-16-5p after assessment of
   its stability in our study cohort and further review of the literature
   (24). The data were analyzed using the 2^−ΔΔCt method and miRNAs with
   Ct > 35 were considered as not expressed and excluded from further
   analysis. Moreover, those miRNAs changing significantly (p-value at
   paired Student’s t-test ≤0.05) in the two time points before starting
   treatment (t−3 and t0) were excluded in order to not consider those
   miRNAs changing for other causes independent of treatment. Considering
   the exploratory purpose of the study and based on the number of
   subjects analyzed, we did not perform FDR correction. The miRNAs having
   a fold change (FC) (log[2]2^−ΔΔCt) >+1.5 or FC (log[2]2^−ΔΔCt) < −1.5
   between baseline and 3 months on treatment were considered as
   differentially expressed.

   An in silico analysis was performed to identify the validated target
   genes and pathways for each differentially expressed miRNA in order to
   select those expected to be involved with growth. In particular, the
   network that represents miRNA target interactions and highlights the
   most impacted pathways was obtained from the miRNet v 2.0 online tool
   ([73]25). This tool collects data from three well-annotated databases,
   miRTarBase v8.0, TarBase v8.0, and miRecords. The significance was set
   at a p-value of 0.05.

2.4 Validation of the Profiling Results by Real-Time qRT-PCR

   MiRNAs for the validation step were chosen based on FC (>+1.5 or <−1.5)
   between baseline and 3 months on treatment, and based on miRNA target
   gene analysis. Eight miRNAs were selected and evaluated by TaqMan
   Advanced miRNA assays (Applied Biosystems): hsa-miR-22-3p (Assay ID:
   477985_mir), hsa-miR-30c-5p (Assay ID: 478008_mir), hsa-miR-106a-5p
   (Assay ID: 478225_mir), hsa-miR-140-5p (Assay ID: 477909_mir),
   hsa-miR-199a-5p (Assay ID: 478231_mir), hsa-miR-335-5p (Assay ID:
   478324_mir), hsa-miR-340-5p (Assay ID: 478042_mir), and hsa-miR-494-3p
   (Assay ID: 478135_mir). cDNA was prepared as described above (Section
   2.3), and miRNA expression was evaluated according to the
   manufacturer’s protocol and run in triplicate in a 7900HT Fast PCR
   system (Applied Biosystems). Data were normalized using hsa-miR-16-5p
   (Assay ID: 477860_mir) as endogenous control ([74]24). The validation
   analysis was performed at two time points t0 and t+3. In addition,
   miRNA levels were analyzed at 12 months on treatment (t+12).

2.5 Auxological Parameters and Bone Age

   A full auxological assessment was done at baseline, 6 months, and 12
   months on treatment and medical history was taken. Body mass index
   (BMI), calculated as weight/height^2 (kg/m^2), and weight were
   converted to standard deviation scores (SDS) using the references of