Revealing the involvement of miR-376a, miR-432 and miR-451a in infantile ascending hereditary spastic paralysis by microRNA profiling in iPSCs

Infantile-onset ascending hereditary spastic paralysis (IAHSP) is a rare, early onset, autosomal recessive motor neuron disease characterized by progressive weakness and spasticity. Several mutations in the alsin 2 gene (ALS2) have been described in IAHSP patients; however, a relevant subset of patients is ALS2 mutationnegative, and pathogenic events causing the disease are unknown. The present study aimed at better understanding the molecular mechanisms underlying motor neuron loss in IAHSP patients by identifying microRNAs (miRNAs) potentially implicated in neuronal differentiation. Using the human induced pluripotent stem cell (iPSC) technology, we developed a patient-specific in vitro cellular model and performed miRNome profiling in fibroblasts, iPSCs and iPSCs-derived neurons obtained from an ALS2 mutation-negative IAHSP patient and a healthy control. The selected differentially expressed miRNAs were also analyzed in fibroblasts, iPSCs and iPSCs-derived neurons from two patients affected by other motor neuron diseases, two patients with other neurological disease, and three healthy controls. We found that miR-376a, miR-432 and miR-451a expression was altered in cell cultures obtained from the IAHSP patient compared to the other patients and controls. In addition, the hierarchical clustering analysis revealed that miR-451a was differentially expressed in fibroblasts and iPSCs, whereas miR-376a and miR432 in neuronal cells. These results, together with the miRNA/mRNA target analysis, were indicative of a significant involvement of miR-451a in stem cell biology processes, and of miR-376a and miR-432 in the establishment of the neuronal phenotype. Our overall findings identified miR-376a, miR-432 and miR-451a as molecules involved in neuronal differentiation, and potentially in IAHSP pathogenesis, which could provide cues for future development of patient-specific miRNAbased therapeutic strategies for IAHSP or other motor neuron diseases.


Introduction
Motor neuron diseases (MNDs) are a heterogeneous group of neurological disorders characterized by a progressive and selective loss of motor neurons in the motor cortex, brainstem, and spinal cord [1], variably affected depending on the specific MND [2]. The spectrum of MNDs extends from childhood to adulthood and the variety of clinical manifestations reflects the numerous genetic alterations identified in patients with familial and sporadic MND [3,4].
MNDs with juvenile-onset include the infantile-onset ascending hereditary spastic paralysis (IAHSP), a rare fatal disorder characterized by weakness and spasticity due to a selective severe degeneration of the pyramidal tract, with onset at lower limbs within the first two years of life, and slowly progressive ascending evolution [5]. Familial febrile seizures plus condition -GEFS + , and severe myoclonic epilepsy at infancy -SMEI). Our findings identified a miRNA pattern associated with neuronal differentiation whose expression was altered in IAHSP cell cultures, thus providing new insights on miRNAs involved in IAHSP pathogenesis that could be further explored for future development of patient-specific miRNA-based therapeutic strategies for this disorder, and other MNDs.

Materials and methods
Reagents and source companies are listed in Supplemental Table 1.

Patients and clinical data
The study included: a 16-year-old boy affected by IAHSP (OMIM #607225) clinically defined, negative for ALS2 gene mutations; two patients affected by MNDs, one 35-year-old female patient with ALS (OMIM#105400) and one 3-year-old male patient with SMA (OMIM#253300); one 40-year-old female patient affected by generalized epilepsy with febrile seizures plus condition (GEFS + ) (OMIM#604233), and one 40-year-old male with severe myoclonic epilepsy at infancy (SMEI) (OMIM#607208); three healthy females (36,55 and 60 years old) and one male (50 years old) with no neurological diseases.
Written informed consent was obtained from IAHSP and SMA patients' parents, ALS, GEFS + and SMEI patients, and from healthy controls. The study was approved by the local Ethics Committee and performed according to the amended Declaration of Helsinki.
Fibroblasts, iPSCs and iPSCs-derived neurons were obtained from all patients and healthy controls. Cells derived from the IAHSP patient and one healthy control were included in miRNome profiling and miRNA validation. Cells obtained from the patients affected by ALS, SMA, GEFS + and SMEI, and the other three healthy controls were included in miRNA validation experiments.

Fibroblast cultures and iPSC generation
Fibroblasts were isolated from skin biopsies as previously described [37]. Cells were maintained and expanded until 3-4 passages in Dulbecco's modified Eagle medium (DMEM) supplemented with 20% fetal bovine serum (FBS), 1% L-Glutamine, 1% non-essential amino acids, 100 mM sodium pyruvate and 1% penicillin/streptomycin (P/S). Fibroblasts (2x10 4 cells) were infected by using the STEMCCA Cre-Excisable Constitutive Polycistronic Lentivirus reprogramming juvenile amyotrophic lateral sclerosis (ALS), in which both upper and lower motor neurons are involved [6]. However, in a notable number of IAHSP patients no mutation was found in the coding sequence of ALS2 [7], indicating that IAHSP is a genetically heterogeneous syndrome. Pathophysiological mechanisms leading to progressive loss of motor neurons in IAHSP patients have not been clarified yet. Further genetic and molecular studies are needed to understand IAHSP pathogenesis, especially in those patients in whom a disease-causing mutation was not found.
An approach to disclose molecular mechanisms underlying MNDs is to establish in vitro models starting from human samples. The recently developed techniques allowing neuronal differentiation from induced pluripotent stem cells (iPSCs) obtained from patients' skin fibroblasts represent a valuable system for in vitro modeling of neurological diseases and for the discovery of gene regulatory networks [8][9][10]. Several studies have shown that iPSCs from patients affected by neurological diseases, including ALS and spinal muscular atrophy (SMA), represent an ideal new patient-specific in vitro model to study and reveal molecular and cellular disease mechanisms, or to develop novel drugs [11][12][13][14][15][16][17][18][19][20][21]. Such an iPSC modeling for IAHSP has never been developed so far, thus its development promises to be relevant for understanding IAHSP pathogenesis.
In the context of stem cell biology, microRNAs (miRNAs) are of particular interest since they are involved in modulating stem cell self-renewal, pluripotency and neuronal cell differentiation [22][23][24][25][26][27]. MiRNAs have been identified as markers of cell identity and their expression profiles clearly distinguish different cell types [28][29][30]. In addition, the implication of miRNAs in neurodegenerative diseases has become evident in the last few years [31][32][33][34][35], particularly in the field of MNDs [31,34,35]. Growing evidence suggests that miRNAs are ideal candidates as biomarkers for neurodegenerative diseases, both for disease diagnosis and prognosis, and as potential therapeutic targets [36].
In this study we performed miRNome profiling in fibroblasts, iPSCs and iPSCs-derived neurons obtained from an ALS2 mutationnegative IAHSP patient and a healthy control, in order to disclose those miRNAs associated with neuronal differentiation stages, relevant for stem cell pluripotency and neuronal phenotype. The selected miRNAs, differentially expressed in IAHSP cells compared to control cells, were further investigated in fibroblast, iPSC and neuronal cell cultures from additional healthy controls, patients affected by ALS, SMA, and by neurological disorders other than MNDs (generalized epilepsy with  and Nucleofector™ Kits, as previously described [38]. At day 25 post-infection, at least 10 iPSC clones were manually picked and cultured on mouse embryonic fibroblast feeder layer and maintained in human embryonic stem cell (ESC) medium containing 20 ng/mL basic fibroblast growth factor (bFGF), as described elsewhere [39], and expanded up to 3 passages for pluripotent status evaluation and analysis of morphology, and up to 20 passages for neuronal cell differentiation induction. All the experiments were performed on three iPSC lines obtained from three iPSC clones derived from each patient and healthy control.

RNA isolation from iPSCs and cDNA synthesis
Total RNA was extracted using TRIzol reagent from 1 to 2 × 10 6 iPSCs, and its quality was checked using a 2100Nano Bioanalyzer (Agilent Technologies, Waldbron, Germany). Total RNA (100 ng) was retro-transcribed using SuperScript VILO cDNA Synthesis kit.

PCR analysis for pluripotency markers
To demonstrate the pluripotency of iPSCs, cDNA obtained from iPSC lines was amplified by PCR using primers specific for Oct4, Sox2, Nanog and Lin-28 genes, all markers of pluripotency, as previously described [40]. GAPDH was examined as housekeeping gene. Sequences of primers are reported in Supplemental Table 1. PCR products were run on DNA stain clear G-stained 1% agarose gel by electrophoresis.

Real-time PCR for AKT1 and Bcl2 in iPSCs
To analyze AKT1 and Bcl2 mRNA expression in iPSCs, cDNA was amplified by Taqman gene expression assays specific for AKT1 and Bcl2 gene on the ViiA7 Real time PCR system (Thermo Fisher Scientific, Waltham, MA, USA). 18s ribosomal RNA was stably expressed in both control and IAHSP cells (standard deviation of Ct values: < 0.5) and was used as endogenous control. Transcriptional levels of the target genes were expressed as relative values (2 -ΔCt ) normalized towards 18s levels.

Flow cytometry of apoptotic cell death in iPSCs
Flow cytometry analysis of apoptotic events was performed in two iPSC lines from the IAHSP patient and a healthy control using the Attune NxT Acoustic Focusing Cytometer (Thermo Fisher Scientific). iPSCs were stained using annexin V-FITC apoptosis detection kit following manufacturer's instructions. A minimum of 40,000 events was acquired for each sample. Annexin V-positive events were considered apoptotic cells. Analysis was performed using FlowJo vX.0.7 data analysis software (Tree Star, Ashland, Oregon, USA).

Generation of neuronal precursors
iPSCs were resuspended in human ESC medium without bFGF and cultured as spheres in low binding non-treated dishes to form embryoid bodies (EBs). After 5-6 days, EBs were collected and plated in Matrigel-coated dishes and maintained for 4-5 days in neuronal induction medium (DMEM/F12) plus 1% N-2 supplement and 20 ng/ mL bFGF, to obtain neural rosettes. The medium was then replaced with the Neurobasal medium plus 1% N-2 supplement and after 10 days neuronal precursors and neurons were obtained. To obtain motor neurons, the rosettes, maintained in Neurobasal medium plus 1% N-2 supplement, were manually picked, plated in Matrigel-coated dishes and cultured with 1 μM cyclic adenosine 3′,5′-monophosphate (cAMP), 1% P/S, 0.1 μM retinoic acid (RA) and 50 ng/mL Sonic Hedgehog (SHH) [39]. Adherent neural progenitors were dissociated with 1X accutase and plated on Matrigel-treated dishes for further analysis.

Immunofluorescence
Immunostaining analysis was performed on iPSCs, EBs, neural rosettes and neuronal cells plated onto Matrigel-treated glass coverslips. The cells were fixed in 4% paraformaldehyde at room temperature for 20 min, permeabilized with 0.1% Triton X-100 and treated with 10% normal goat serum in PBS to block non-specific binding. iPSCs and EBs were incubated with a combination of the following primary antibodies: rabbit anti-Nanog, mouse anti-Sox2 and rabbit anti-Oct-4 antibodies, mouse anti-TRA-1-60, all markers of pluripotency [40]. To assess the proliferative activity, iPSCs were stained with the rabbit anti-Ki67 antibody. Neural rosettes were immunostained with mouse anti-nestin antibody. Neuronal cells were stained with a combination of the following primary antibodies: mouse anti-β-tubulin III (β-TubIII), mouse anti-microtubule associated protein 2 (MAP2) and rabbit anti-homeobox gene Hb9 (Hb9), mouse anti-neuronal nuclei (NeuN), markers of immature motor neurons. Immunopositivity was revealed with Alexa Fluor 488-conjugated goat anti-mouse IgG and Cy3-conjugated goat anti-rabbit IgG secondary antibodies. Cells were counterstained with 4',6-diamidino-2-phenylindole (DAPI) and the coverslips were mounted with FluorSave. Confocal fluorescence images were obtained with a laser-scanning microscope (Eclipse TE 2000-E, Nikon, Tokyo) and analyzed using EZ-C1 3.70 imaging software (Nikon). Antibody details and working dilutions are summarized in Supplemental Table 1. For alkaline phosphatase colorimetric detection, iPSC cultures were stained with 5-Bromo-4-chloro-3-indolyl phosphate and Nitro Blue Tetrazolium.

Quantification of iPSCs, EBs and neuronal cells
For iPSC diameter measurement, one hundred DAPI-positive iPSC clones were acquired for each of the three IAHSP and three control iPSC lines using EZ-C1 3.70 imaging software (Nikon). The diameter of each iPSC clone, defined as the widest transversal distance, was measured and expressed as mean ± standard deviation (SD). The number of DAPI-positive cells forming each iPSC clone was counted using Image Pro-Plus (Media Cybernetics, Silver Spring, MD, USA) and expressed as mean ± SD. To assess the proliferative activity, fifty iPSC clones were examined for each of the three IAHSP and three control iPSC lines after immunofluorescence staining specific for the Ki67 antigen followed by cell counting using Image Pro-Plus (Media Cybernetics). The percentage of Ki67-positive cells in each iPSC clone was calculated in relation to the total number of DAPI-positive cells/ field and expressed as mean ± SD for each examined group.
For EB diameter measurement, ten DAPI-positive EBs were acquired for each of the three EB lines derived from the three IAHSP and three control iPSC lines, using EZ-C1 3.70 imaging software (Nikon), and the size was reported as mean ± SD. β-TubIII-and MAP2-positive neurons were counted on four randomly selected fields per coverslip. Two coverslips were analyzed for each of the three IAHSP and three control iPSC-derived neuron cell lines. The total number of β-TubIII-and MAP2-positive cells/field was expressed as mean ± SD for each examined group. For neurite length measurement, β-TubIII and MAP2-positive neurons were manually selected on four randomly selected fields per coverslip, and the longest neurite of each selected cell was measured using EZ-C1 3.70 imaging software (Nikon) and expressed as mean ± SD. Two coverslips were analyzed for each of the three IAHSP and control iPSC-derived neuron cell lines. For Hb9-positive motor neuron quantification a mean of one hundred cells/field were counted on four randomly selected fields per coverslip.
Two coverslips were analyzed for each of the three IAHSP and three control iPSC-derived motor neuron cell lines for a total number of thirty neural rosettes. The percentages of Hb9-positive motor neurons were calculated in relation to the total number of DAPI-and β-TubIIIpositive cells/field and expressed as mean ± SD for each examined group.

miRNome profiling and data analysis
Total RNA was extracted from 1 to 2 × 10 6 fibroblasts, iPSCs and neuronal cells obtained from the IAHSP patient and one healthy control and reverse transcribed using Megaplex RT primers Human Pool A and B and MultiScribe Reverse Transcriptase Kit. cDNA, corresponding to 500 ng total RNA, was combined with TaqMan Universal PCR Master Mix and dispensed into each port of the TaqMan Human MicroRNA array card A and B v2.0, following the manufacturer's instructions. The arrays were run on the Viia 7 Fast Real-Time PCR System (Thermo Fisher Scientific) in triplicate. Human array A and B cards contained primers for 381 miRNAs, including 3 positive control miRNAs and 1 negative control. All the results were normalized against U6 endogenous control [41,42] and relative miRNA expression levels were calculated using the ΔCt method.

miRNA validation by real-time PCR
miRNAs differentially expressed between IAHSP and control fibroblasts, iPSCs and neuronal cells were selected for validation by real-time PCR based on fold change ≥ 3 criterion and literature data [43˗51]. Total RNA was extracted from 1 to 2 × 10 6 fibroblasts, iPSCs and neuronal cells from the IAHSP patient, the four healthy controls and the ALS, SMA, GEFS + and SMEI patients using TRIzol reagent. RNA quality was checked by 2100Nano Bioanalyzer (Agilent Technologies). RNA was reverse transcribed using TaqMan MicroRNA Reverse Transcription Kit with primers specific for the selected miRNAs: miR-155, miR-376a, miR-432, miR-451a, miR-490, miR-520, miR-629, miR-657, and miR-1289. U6 was used as endogenous control [41,42]. cDNA (corresponding to 15 ng of total RNA) was amplified in triplicates by real-time PCR, using Universal PCR master mix and pre-designed TaqMan MicroRNA assays on the Viia 7 Fast Real-Time PCR System (Thermo Fisher Scientific). All the results were normalized against U6 and the relative miRNA expression levels were calculated using the ΔCt method.

Predicted miRNA-mRNA network construction
In silico prediction of miRNA targets was performed using the DIANA-microT-CDS algorithm from the DIANA miRNA database using the default score cut-off [52]. Predicted annotation of gene targets of validated miRNAs was made using Cytoscape, version 2.8.2 (http://www.cytoscape.org/), with the ClueGO plugin [53] and using the metabolic pathways (KEGG, Biocarta) and gene ontology biological process terms. ClueGO displays the functional terms as nodes, and the relationships between the terms, based on the similarity of their associated genes, are shown as edges. The degree of connectivity between terms was calculated by Cohen's kappa statistics using a threshold cutoff n = 0.4.

Real-time PCR for mRNA targets
mRNA targets were selected for validation by real-time PCR considering three genes for each miRNA with a default score cut-off equal to 0.9-0.8. RNA, previous extracted from 1 to 2 × 10 6 fibroblasts, iPSCs and neuronal cells from the IAHSP patient, the ALS, SMA, GEFS + and SMEI patients and the four healthy controls, was retro-transcribed using SuperScript VILO cDNA Synthesis kit. cDNA was amplified by Taqman gene expression assays specific for CAB39, DNAJA1, EIF4B,  KMTC2, NDUFAB1, PSMB8, TSC1, TSG101, SMURF1. 18s gene was used as endogenous control. cDNA (corresponding to 10 ng of total RNA) was amplified in duplicates by real-time PCR, using TaqMan Fast Master Mix and TaqMan gene assays on the Viia 7 Fast Real-Time PCR System (Thermo Fisher Scientific). All the results were normalized against 18s and the relative gene expression levels were calculated using the 2 -ΔCt method.

Statistical analysis
Data analysis was performed with the R Statistical environment (version 3.0.2.) (www.r-statistics.org). Mann-Whitney test was applied to compare experimental data derived from IAHSP and control cells. Bonferroni correction was applied to control for false discovery rate. P value < 0.05 was considered significant. MiRNA data analysis was performed by hierarchical clustering of miRNAs and samples, using the mean values of relative expression converted to a log2 scale. An heat map diagram was obtained to show clustering results. Relationship between the expression of each miRNA and its mRNA targets was evaluated by obtaining "before-after graphs" showing the miRNA/ mRNA correlation trend in the IAHSP patient cells, and cells from all the other patients and healthy controls.

Generation of IAHSP iPSC lines
We generated iPSC lines by reprogramming fibroblasts from an IAHSP patient and a healthy control. Both IAHSP and control iPSCs were positive for alkaline phosphatase ( Figure 1A), normally detected in undifferentiated pluripotent stem cells [38], and expressed the pluripotency markers Nanog, Lin-28, Sox2 and OCT4 [40] (Figure 1B/C), thus indicating that they were pluripotent. The IAHSP iPSCs showed a smaller size, as estimated by diameter measurement, and a lower number of DAPI-positive cells than control iPSCs (p < 0.001) ( Figure 1D). To assess possible apoptotic events, we evaluated the transcriptional levels of the apoptosis regulator Bcl2 by real-time PCR (Supplemental Figure 1) and estimated the percentage of annexin V-positive apoptotic cells by flow cytometry (Supplemental Figures 2): Bcl2 was over-expressed in IAHSP iPSCs compared to control cells and the percentage of annexin V-positive cells was higher in IAHSP than in control iPSCs, but both values did not reach significance. By immunofluorescence, we found a decreased proportion of IAHSP iPSCs expressing the proliferation-associated Ki67 antigen compared to control cells (p < 0.001) ( Figure 1E). Moreover, the expression levels of AKT1, a serine-threonine protein kinase functioning as a critical regulator of cell proliferation and survival [54], was also decreased in IAHSP iPSCs compared to control cells (p < 0.05) ( Figure 1F), thus suggesting a reduced proliferation ability and survival of iPSCs derived from the IAHSP patient.

Generation of IAHSP neurons and motor neuron precursors
At first, we induced the formation of EBs from IAHSP and control iPSCs (Figure 2A). By immunofluorescence, EBs were positive for the pluripotency markers Sox2, Nanog and Oct4 [40] (Figure 2B/C). They were maintained in appropriate conditions for five days and subsequently differentiated in neural rosettes (Figure 2A). Neural rosettes showed immunoreactivity to nestin, a specific marker of neural stem cells [39], in both groups examined ( Figure 2D). IAHSP and control EBs and neural rosettes had similar morphology and did not show significant differences in size ( Figure 2E). Representative images of control (upper panels) and IAHSP (lower panels) iPSC colonies obtained by phase contrast microscopy (left panels) and bright field microscopy after staining with phosphatase alkaline (right panels) (A). Expression analysis by PCR of the pluripotency markers Nanog, Lin-28, Sox2, and Oct4, and of the housekeeping gene GAPDH, in control and IAHSP iPSC clones (B). Confocal microscopy images of control and IAHSP iPSC clones positive for Nanog (red) and Sox2 (green) (upper panels), and for Oct4 (red) and Sox2 (green) (lower panels), with cell nuclei stained with DAPI (blue) (C). Confocal microscopy images of control and IAHSP iPSC clones showing cell nuclei stained with DAPI (blue) (D, panels in left side). Quantification of diameter, defined as the widest transversal distance and espressed as mean ± standard deviation (SD), of one hundred DAPI-positive iPSC clones obtained from control (n=3) and IAHSP (n=3) iPSC lines (D, left graph in right side); number of DAPI-positive cells forming the iPSC clones expressed as mean ± SD (D, right graph in right side). Confocal microscopy images of control and IAHSP iPSC clones after immunostaining specific for the proliferation marker Ki67 (red) and counterstaining with DAPI (blue) (E, panels in left side). Mean percentage of Ki67-positive cells ± SD calculated in relation to the total number of DAPI-positive cells/field for each clone in fifty iPSC clones from control (n=3) and IAHSP (n=3) iPSC lines (E, graph in right side). Real-time PCR analysis of AKT1 gene expression levels in control (white bar), and IAHSP (black bar) iPSCs (n = 3 iPSC lines per group), *p < 0.05; ***p < 0.001: Mann-Whitney test (F). Magnification bars: 80 μm in A, C and D; 30 μm in E.
EB-derived neural rosettes were then maintained for 10 days in N2 supplement, to obtain neuronal cells (Figure 2A). Interestingly, IAHSP neural rosettes generated a greater proportion of MAP2-and β-TubIIIpositive neurons than those from control cultures (p < 0.05) ( Figure  3A/B). However, neurons derived from IAHSP iPSCs displayed shorter neurites than control neurons, as shown by measurement of the neurite length (p < 0.001) ( Figure 3C).
To obtain spinal motor neurons, IAHSP and control neural rosettes were manually picked and cultured with RA to establish a caudal position, and subsequently with the addition of SHH to promote ventral position according to the protocol previously described by Erceg and colleagues (2010) (Figure 2A). By day 50 of the differentiation protocol, IAHSP and control cultures were characterized by a substantial number of cells displaying a motor-specific expression pattern, i.e. positivity for Hb9, a transcription factor induced during motor neuron differentiation, and β-TubIII ( Figure 3D). No significant differences in the percentage of Hb9-positive motor neurons were observed between IAHSP and control cultures ( Figure 3E).

miRNome profiling in IAHSP fibroblasts, iPSCs and neuronal cells
To identify miRNAs relevant for stem cell biology and neuronal differentiation, potentially implicated in IAHSP, we performed a miRNome expression profiling, including 381 human miRNAs, in fibroblasts, iPSCs and neuronal cells derived from the IAHSP patient and a healthy control. In the discovery phase, we revealed that 70 miRNAs were differently expressed in IAHSP compared to control fibroblasts, with 24 miRNAs being up-regulated and 46 down-regulated ( Figure 4A). In IAHSP iPSCs the expression of 179 miRNAs was dysregulated compared to control iPSCs, including 77 miRNAs up-regulated and 102 down-regulated ( Figure 4B). Moreover, 262 miRNAs were differentially expressed in IAHSP neuronal cells compared to control cells, including 32 miRNAs up-regulated and 230 down-regulated ( Figure 4C).
Differences in miRNA expression between IAHSP and control cells were not related to differences in the efficiency of fibroblast isolation and reprogramming, since fibroblast morphology and the number of iPSC colonies obtained from 2x10 4 reprogramming fibroblasts (8 to 10 colonies) were comparable among the examined sample groups (~ 0.2% efficiency).

miRNA validation: identification of miR-376a, miR-432 and miR-451a as modulators of neuronal differentiation in IAHSP
To validate data from miRNoma profiling, the 9 selected miRNAs were analysed by real-time PCR in fibroblasts, iPSCs, and neuronal cells derived from the IAHSP patient and the healthy control included in the discovery phase, and in fibroblasts, iPSCs and neuronal cells derived from additional three healthy controls, two patients affected by MNDs -one affected by ALS and one by SMA -, and two patients affected by neurological diseases different from MNDs, GEFS + and SMEI (Supplemental Figure 3/4). The results of the validation phase showed that miR-451a expression was decreased in IAHSP fibroblasts compared to fibroblasts of healthy controls, MND, GEFS + and SMEI patients ( Figure 5), in line with the previous miRNome profiling data, whereas miR-376a and miR-629 levels were similar in all the fibroblast cultures analyzed (Supplemental Figure 5 and Table 1). In iPSCs, miR-451a was up-regulated in IAHSP compared to healthy control, GEFS + and SMEI cells, confirming miRNome profiling results, but it was down-regulated if compared with MND iPSCs ( Figure 5 and Table  1). All the other miRNAs, miR-155, miR-376a, miR-432, miR-520, and miR-490, whose expression was altered in the discovery phase, were not detected or expressed at similar levels in the analyzed iPSC samples (Supplemental Figure 5 and Table 1). Validation of miRNAs in neuronal cells showed a significant reduction of miR-376a and miR-432 in IAHSP compared to all the other samples ( Figure 5 and Table  1), in agreement with the discovery phase data. Contrariwise, miR-155 down-regulation in IAHSP neuronal cells was not confirmed, as well as up-regulation of miR-520, miR-629 and miR-1289. Indeed, miR-155 and miR-520 were expressed at similar levels among IAHSP, SMA and healthy control cells, and were not detected in ALS, GEFS + and SMEI cells; miR-629 expression was similar in IAHSP compared to all the analyzed neuronal samples; miR-1289 was not detected in all cultures (Supplemental Figure 5).
Taken together, these data identified an altered expression profile of miR-376a, miR-432 and miR-451a in IAHSP fibroblasts, iPSCs and neuronal cells compared to the corresponding cell cultures from healthy controls and patients affected by ALS, SMA, SMEI and GEFS + ( Figure 5 and Table 1). Interestingly, the hierarchical clustering analysis showed a distinct miRNA cluster characterizing IAHSP samples compared to healthy control, ALS, SMA, SMEI and GEFS + samples, suggesting that altered expression of the identified miRNAs might contribute to IAHSP pathogenesis. A sub-cluster distinctive of ALS and SMA compared to all the other samples was also disclosed, leading to the hypothesis that a fine balance of miR-376a, miR-432 and miR-451a expression levels may be critical in MNDs. Of note, the hierarchical clustering analysis also revealed two different clusters of miRNAs, one characterizing fibroblasts and iPSCs, which included miR-451a, and the other one being associated with the neuronal cell phenotype, which included miR-376a and miR-432. These results suggested that miR-451a is specifically involved in stem cell biology processes, whereas miR-376a and miR-432 participate in neuronal differentiation processes.

In-silico miRNA target prediction and functional network construction
A network was constructed to visualize the functional connections of miR-376a, miR-432, and miR-451a, and their predicted targets, and the annotations involved in motor neuron disease-related processes ( Figure 6). miR-376 resulted a regulator of: i) RNA transport by modulating EIF4A1 and 2, EIF4B, NUP58 and 160, TGS1 genes; ii) protein processing in endoplasmic reticulum, closely linked with N-Glycan biosynthesis, by modulating MAN1C1, DNAJA1, and SEC24D; and iii) oxidative phosphorylation, connected with phagosome and synaptic vesicle cycle, by modulating ATP6V1G1, NDUFAB1, and SDHD. miR-432 was implicated in: i) GABAergic synapse pathway, including cAMP signaling, via targeting GABRB1 and GABBR1 genes; ii) lysine degradation, via targeting WHSC1L1 and KMT2C genes; and iii) endocytosis, via targeting EPN3, PIP5K1B, TSG101, SMURF. miR-451a, and their putative target genes CAB39 and tuberous sclerosis 1 (TSC1), were implicated in mTOR signaling pathways, closely linked with AMPK and P13K-Akt signaling pathways; the same miRNA, along with miR-376a, was also associated with proteasome via targeting PSMA1 and 6, and PSMB8 genes.

Expression analysis of mRNA targets of miR-376a, miR-432 and miR-451a
Transcriptional levels of selected putative target genes of miR-376a, miR-432 and miR-451a were analyzed in fibroblasts, iPSCs and neuronal cells from the IAHSP patient, the ALS, SMA, SMEI, and GEFS + patients, and the healthy controls. Specifically, in iPSCs and fibroblasts we analyzed CAB39, TSC1 and PSMB8, as targets of miR-451a ( Figure 7A and B); in neuronal cells, DNAJA1, EIF4B and NDUFAB1 were tested as targets of miR-376a, and SMURF, TSG101 and KMT2C as targets of miR-432 ( Figure 7C). miR-451a expression showed a negative relationship with the expression of its targets CAB39 and TSC1 in fibroblasts and iPSCs (only TSC1) of ALS, SMA, SMEI, and GEFS + patients and controls, but not in the corresponding cells of the IAHSP patient, where CAB39 and TSC1 levels were higher than those of the miRNA, indicative of a positive correlation trend. In all fibroblast samples, PSMB8 mRNA levels were lower than those of miR-451a, with IAHSP fibroblasts having the lower values of miR-451a expression levels ( Figure 7A and B). miR-376a expression displayed a positive relationship with the expression of its targets DNAJA1 and NUDFAB1 in neuronal cells of the IAHSP patient, but not in the corresponding cells of ALS, SMA, SMEI, and GEFS + patients and controls ( Figure 7C). In the same cells from all patients and controls, a negative correlation trend was observed between EIF4B mRNA levels and those of miR-376a. However, neurons of the IAHSP patient had the lower expression levels of both miR-376a and EIF4B. A positive correlation trend between miR-432 and TSG101expression levels was showed in neuronal cells of the IAHSP patient, but not in cells from the other patients and controls. A trend indicative of negative correlation was found between miR-432 levels and those of SMURF and KMT2C in all samples, with neurons of the IAHSP patient showing the lower expression levels of the miRNA and its targets ( Figure 7C).

Discussion
MNDs, including IAHSP, are complex neurodegenerative disorders in terms of genetic background and clinical presentations. Several mutations in ALS2 locus have been described in IAHSP patients, but no genetic alterations were found in a relevant subset of patients in whom mechanisms underlying motor neuron deficiency remain to be clarified. Notably, even if no mutations in ALS2 gene were tracked in the IAHSP patient under study in this manuscript, the gene transcript expression levels in IAHSP-derived neuronal cells was down regulated compared to all the other disease and healthy controls (Supplemental Figure 6). This here suggests that an altered transcriptional regulation might occur, underlying the complexity of combinatorial gene control mechanisms in IAHSP, and other ALS2 gene-related diseases.
Reprogramming fibroblasts from MND patients into iPSCs, and differentiating them into neurons, could provide a patient-specific cellular model for studying diseases' molecular mechanisms [18]. As far as we know iPSCs from IAHSP patients have never been generated and studied to investigate pathogenic mechanisms or possible therapeutic interventions, therefore it is still unknown whether there are key molecules associated with IAHSP iPSC identity or involved in neuronal differentiation and pathogenic processes in the disease. Due to their well-known involvement in stem cell function and neurodegeneration [22,55,56], miRNAs represent one of the most plausible molecular candidates suspected to play a role in the pathogenic events underlying MNDs. However, their role in IAHSP remains unexplored. Here, we performed a miRNome profiling, including 381 human miRNAs, in fibroblasts, iPSCs and iPSC-derived neuronal cells from an ALS2 mutation-negative IAHSP patient and a healthy control, to disclose miRNAs differentially expressed in IAHSP cells and potentially implicated in the disease. Validation of selected miRNAs in cell cultures from additional healthy controls, patients affected by different MNDs (ALS and SMA), and neurological disorders other than MNDs (SMEI and GEFS + ), revealed a miRNA pattern specifically associated with neuronal differentiation and potentially implicated in IAHSP.
Our molecular analysis of the developmental axis "fibroblast-iPSCneuronal cell" strongly demonstrated the relevance of characterizing patient-derived cell types to identify the molecular underpinning of IAHSP and other MNDs.
iPSCs generated from IAHSP patient showed reduced size, accompanied by decreased ability to proliferate compared to control iPSCs, suggesting disease-related differences in intrinsic survival properties of patient-derived cells. However, such differences were not reflected in a different efficiency in reprogramming of fibroblasts; indeed, fibroblast morphology, and iPSC-derived colonies' number were similar in the two examined groups.
The analysis of miRNome in IAHSP and control fibroblasts, and the further validation of selected miRNAs in fibroblast cultures of additional disease-controls -ALS, SMA, SMEI and GEFS + patientsled to identify a reduced expression of miR-451a in IAHSP fibroblasts compared to all the other samples. Contrariwise, this miRNA was incremented in the IAHSP fibroblast-derived iPSCs compared to control cells and cells from SMEI and GEFS + patients, but not in comparison with cells from ALS and SMA patients. By exploring in-silico miR-451 function through miRNA-mRNA target network construction, we found that this miRNA putatively regulates mTOR, AMPK and PI3K-Akt signaling pathways, which are relevant for cell cycle regulation [27], via targeting TSC1 gene. Accordingly, our mRNA expression analyses showed a positive trend correlation between miR-451a and TSC1 gene level in iPSCs. The regulation of this pathway by miR-451a is corroborated by studies in the field of cancer which demonstrated that, by activating mTOR/AMPK/PI3K-Akt signaling, via direct targeting of TSC1, miR-451a enhances stemness features [57]. Based on these literature data, and the observed mRNA target trend, miR-451a up-regulation in IAHPS iPSCs, compared to the fibroblast stage, could be related to its ability to promote self-renewal capacity and expression of pluripotent cell markers, actually proper of iPSC stage of differentiation. Our gene expression analysis also demonstrated an inverse relationship between miR-451a and AKT1 mRNA levels, suggesting that altered cell size, and decreased proliferative activity, we observed in IAHSP iPSCs compared to controls, might be linked to a miR-451a-dependent decreased expression of AKT1, which is known to promote proliferation, and increase cell survival [54]. Moreover, our miRNA-mRNA study for miR-451 identified proteasome subunit beta 8 (PSMB8) gene as another potential target. PSMB8 has a major role in the removal of misfolded and damaged proteins, controlling cell integrity and cell viability [58]. Interestingly, in IAHSP, as well as in ALS and SMA cells, miR-451a expression levels were higher compared to control samples and other neurological disease samples, suggesting an altered regulation in protein quality control, which is actually known to play a crucial role in MNDs [59]. miRNA analysis in iPSC-derived neurons showed that miR-376a and miR-432 were reduced in IAHSP neuronal cells compared to all the other cells. Considerable data suggest that RNA processing and intracellular transport are altered in MNDs associated with mutations in TDP-43, a protein required for miRNA biogenesis [60,61]. Interestingly, the putative relationship between miR-376a and genes implicated in RNA transport pathways (i.e. EIF4A1 and 2, EIF4B, NUP58 and 160, TGS1), suggests a contribution of this miRNA in altering RNA metabolism, not only in TDP-43-associated MNDs, but also in other MNDs as IAHSP. Furthermore, by predicted functional network, we found that oxidative phosphorylation pathways are additional potential targets of miR-376a. In accordance, NDUFAB1 gene, which codes for a subunit of Complex I of the mitochondrial respiratory chain, showed a positive trend correlation with miR-376a. This observation is of particular interest, due to the well-known role of oxidative stress, and mitochondrial function, in MNDs pathogenesis, including ALS, particularly considering that anti-oxidants are potential therapeutic agents for the disease [62,63].
Substantial studies suggested that a better knowledge of protein processing in endoplasmic reticulum pathways should be valuable in understanding the pathogenesis and eventually in designing therapy for neurodegenerative diseases [64,65]. Our prediction analysis showed a putative relationship between miR-376a and genes implicated in endoplasmic reticulum-associated pathways (i.e. MAN1C1, DNAJA1, and SEC24D). This was supported by a miRNA/mRNA positive relationship with DNAJA1 gene in IAHSP neuronal cells, suggestive of a possible implication also for miR-376a in misfolded and damaged protein repair mechanisms in MNDs. Of note, Jovicic, et al. demonstrated that miR-376a was enriched in functional neurons and induced stem cell differentiation towards neuronal phenotype in rat primary cultures [66]. Relying on these data, and on the iPSC-derived neuronal cell characterization, we hypothesized that the reduction of miR-376a expression levels may be responsible for the altered neuron phenotype characterized by impaired neurite outgrowth. Regarding this aspect, further investigations are needed to understand whether this defect is related to RNA metabolism alterations or to an altered axonal transport not mediated by alsin.
Regarding miR-432, our network analysis showed that GABAergic synapse, lysine degradation, and endocytosis pathways, are putative miR-432-regulated mechanisms. Interestingly, a contribution of miRNAs in modulating the interaction between endosomal toxicity and synaptic vesicle recycling pathways has been described [67,68]. Here, we suggested that miR-432 might be critically involved in these pathways, which are widely implicated in MND, as ALS, onset and progression [68,69]. In addition, it has been shown that miR-432 induces neurite outgrowth, inhibition of proliferation, and cell cycle arrest in human neuroblastoma cells by targeting nestin, corepressor 1 and methyl cpg binding protein 2 [49]. These results suggest an involvement of miR-432 in neuritogenesis processes, and indicate that defective expression of this miRNA, perhaps together with miR-376, might negatively influence neuronal cell differentiation by impairing neurite growth. Finally, the reduced expression of miR-432 we found in IAHSP neuronal cells, in accordance with its positive relationship with TSG101 gene, which acts as a negative growth regulator, might be responsible for the observed different number of neurons in the IAHSP cultures compared to controls.

Conclusions
Our overall data identified miR-451a, miR-376a, and miR-432 as closely implicated in the differentiation program from fibroblasts to iPSC-derived neurons, and as molecules whose altered expression might significantly affect signaling pathways relevant in MND-related processes. Of interest, hierarchical clustering analysis of miRNA expression levels in the axis "fibroblast-iPSC-neuron" revealed two clusters of miRNAs: the first, including miR-451a, was distinctive of fibroblasts and iPSCS, indicating a contribution of miR-451 to stem cell biology processes; the second, which included miR-376a and miR-432, was associated with the neuronal cell phenotype, highlighting a role for these miRNAs in modulating neuronal differentiation programs. Our findings on dysregulation of miR-451a, miR-376a, and miR-431 expression pattern in the iPSC-based in vitro model of IAHSP provide new insights into the IAHSP pathogenesis, pointing out the possibility that altered neuronal differentiation processes, depending on these miRNAs, could eventually lead to an altered availability of motor neurons in the IAHSP disease. Further studies in genetically defined, or not defined, IAHSP patients, and in other MNDs, could help to deeply understand the synergistic effect of miR-376a, miR-432 and miR-451a in disease onset and/or progression, towards future development of patient-specific miRNA-based therapeutic strategies for IAHSP and other MNDs.