RESEARCH ARTICLE Identification of serum microRNAs as potential biomarkers in Pompe disease Ana Carrasco-Rozas1 , Esther Fernandez-Simo  n1, Maria Cinta Lleixa 1, Izaskun Belmonte2, Irene 2 2 Pedrosa-Hernandez , Elena Montiel-Morillo , Claudia Nun ~ ez-Peralta3, Jaume Llauger Rossello3, Sonia Segovia , Noemı De Luna , Xavier Suarez-Calvet , Isabel Illa1,4, Pompe Spanish Study group, Jordi 1,4 1,4 1,4 Dıaz-Manera1,4* & Eduard Gallardo1,4* 1 noma de Barcelona, Barcelona, Spain Neuromuscular Disorders Unit, Neurology Department, Hospital de la Santa Creu i Sant Pau, Universitat Auto 2 Rehabilitation and Physiotherapy Department, Hospital de la Santa Creu i Sant Pau, Universitat Autonoma de Barcelona, Barcelona, Spain 3 noma de Barcelona, Barcelona, Spain Radiology Department, Hospital de la Santa Creu I Sant Pau, Universitat Auto 4 n Biomedica en Red en Enfermedades Raras, Valencia, Spain Centro de Investigacio Correspondence Abstract Jordi Dıaz-Manera and Eduard Gallardo, Neuromuscular disorders Unit, Hospital de la Objective: To analyze the microRNA profile in serum of patients with Adult Santa Creu i Sant Pau, C/ Sant Antoni Mª Onset Pompe disease (AOPD). Methods: We analyzed the expression of 185 Claret 167, Barcelona 08025, Spain. microRNAs in serum of 15 AOPD patients and five controls using microRNA Tel: +34-935565986; Fax: +34-935565602; PCR Panels. The expression levels of microRNAs that were deregulated were E-mails: jdiazm@santpau.cat; further studied in 35 AOPD patients and 10 controls using Real-Time PCR. egallardo@santpau.cat Additionally, the skeletal muscle expression of microRNAs which showed signif- icant increase levels in serum samples was also studied. Correlations between Funding information microRNA serum levels and muscle function test, spirometry, and quantitative This study was supported by a grant from muscle MRI were performed (these data correspond to the study NCT01914536 Sanofi-Genzyme (GZ-2015-11342) to Dr. at ClinicalTrials.gov). Results: We identified 14 microRNAs that showed differ- Gallardo and has been registered in ent expression levels in serum samples of AOPD patients compared to controls. Clinicaltrials.gov (identifier NCT03045042). We validated these results in a larger cohort of patients and we found increased levels of three microRNAs, the so called dystromirs: miR-1-3p, miR-133a-3p, Coinvestigators – The Spanish Pompe Study Group: Miguel Angel Barba-Romero (Hospital and miR-206. These microRNAs are involved in muscle regeneration and the General de Albacete, Albacete, Spain), Joseba expression of these was increased in patients’ muscle biopsies. Significant corre- Barcena (Hospital Universitario Cruces, lations between microRNA levels and muscle function test were found. Inter- Baracaldo, Spain), Marıa Rosario Carzorla pretation: Serum expression levels of dystromirs may represent additional (Hospital Puerta de Hierro, Majadahonda, biomarkers for the follow-up of AOPD patients. Spain), Carlota Creus (Hospital Virgen de las Nieves, Granada, Spain), Jaume Coll-Cantı (Hospital Germans Tries i Pujol, Badalona, Spain), Noemı de Luna (Neuromuscular Disorders Unit, Neurology Department, Hospital de la Santa Creu i Sant Pau, Universitat Autonoma de Barcelona, Barcelona, Spain, Centro de Investigacio n Biomedica en Red en Enfermedades Raras (CIBERER), Spain), Manuel Dıaz (Hospital de Cabuen ~es, Gijo n, Spain), Cristina Domınguez (Hospital 12 de Octubre, Madrid & Insituto de n i+12, Madrid, Spain), Roberto Investigacio Fern andez Torro n (Hospital Universitario Donostia, Spain), Marıa Jose Garcıa Antelo (Hospital Universitario A Corun ~a, A Corun~a, Spain), Josep Marıa Grau (Hospital Clınic, Barcelona, Spain), Marıa Teresa Go mez Caravaca (Hospital de Co rdoba, Spain), Juan 1214 ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. A. Carrasco-Rozas et al. microRNAs in Pompe Disease Carlos Leon Hern andez (Hospital Universitario Nuestra Sen ~ora de la Candelaria, Tenerife, Spain), Adolfo Lo pez de Munain (Hospital Universitario Donostia, Spain), Francisco Antonio Martınez-Garcıa (Hospital Clınico Universitario Virgen de la Arrixaca, Murcia, Spain), Yolanda Morgado (Hospital Universitario Virgen de Valme, Sevilla, Spain), Antonio Moreno (Hospital Universitario Morales Meseguer, Murcia, Spain), German Morıs (Hospital Universitario de Asturias, Oviedo, Spain), Miguel Angel Mun ~oz-Blanco (Hospital Gregorio Maran ~o n, Madrid, Spain), Andres Nascimento (Hospital Virgen del Rocıo, Sevilla, Spain), Carmen Paradas (Hospital Virgen del Rocıo, Sevilla, Spain), Jose Luis Parajua Pozo (Hospital de Can Mises, Ibiza, Spain), Luis Querol (Neuromuscular Disorders Unit, Neurology Department, Hospital de la Santa Creu i Sant Pau, Universitat Auto noma de Barcelona, Barcelona, Spain, Centro de n Biom Investigacio edica en Red en Enfermedades Raras (CIBERER), Spain), Arturo Robledo-Strauss (Hospital Juan Ramo n Jim enez, Huelva, Spain), Ricard Rojas Garcıa (Neuromuscular Disorders Unit, Neurology Department, Hospital de la Santa Creu i Sant Pau, Universitat Autonoma de Barcelona, Barcelona, Spain, Centro de Investigacio n Biom edica en Red en Enfermedades Raras (CIBERER), In ~igo Rojas-Marcos (Hospital Virgen de Macarena, Sevilla, Spain), Jose Antonio Salazar (Hospital Regional Universitario de Malaga, Spain), Mercedes Uso n (Hospital de Son Llatzer, Palma de Mallorca, Spain). Received: 14 March 2019; Revised: 3 May 2019; Accepted: 4 May 2019 Annals of Clinical and Translational Neurology 2019; 6(7): 1214–1224 doi: 10.1002/acn3.50800 *These authors equally contributed to this work. Patients are classified as infantile (IOPD) or adult onset Introduction Pompe disease (AOPD) depending on the age at onset of Pompe disease is an autosomal recessive disorder pro- symptoms.3 Classic IOPD patients have a rapidly progres- duced by mutations in the GAA gene that encodes for the sive disorder characterized by hypotonia, muscle weak- lysosomal enzyme acid alpha-glucosidase.1 This enzyme is ness, respiratory insufficiency, and hypertrophic essential for the degradation of glycogen to glucose in cardiomyopathy that lead patients to death before the first lysosomes.2 Mutations in the GAA gene cause absence or year of life if untreated.4 AOPD patients have a more deficiency of the enzyme leading to an accumulation of heterogeneous clinical picture ranging from asymptomatic glycogen in the lysosomes of several tissues particularly HyperCKemia to weakness involving limb, axial, and res- cardiac, skeletal, and smooth muscle.3 piratory muscles.5 Natural history studies have shown that ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 1215 microRNAs in Pompe Disease A. Carrasco-Rozas et al. AOPD is a slowly progressive disorder in which most of Material and Methods the patients will present moderate to severe motor dis- ability developing respiratory insufficiency and needing Patient and study design noninvasive ventilation.6 Enzyme replacement therapy (ERT) with recombinant This study is part of an ongoing prospective open-label alpha-glucosidase has changed natural history of the dis- study in which we are following up a group of symp- ease. In infantile patients, the treatment reduces the risk of tomatic and nonsymptomatic AOPD patients using mus- death improving cardiomyopathy and respiratory func- cle function tests, muscle MRI, and blood analysis. This tion.7,8 In AOPD patients, ERT improves and stabilizes study has been registered in Clinicaltrials.gov (identifier muscle function for a period of 2–3 years and afterwards NCT01914536). The study was approved by The Ethics motor function slowly deteriorates9. There is an active dis- Committee of Hospital de la Santa Creu i Sant Pau cussion in the scientific community to know which tests (HSCSP) in Barcelona. All participants signed an appro- are most sensitive to identify changes in muscle perfor- priate informed consent form. All patients included ful- mance in AOPD patients after shorter periods of time10,11. filled the diagnostic criteria of Pompe disease established ERT is not approved for asymptomatic AOPD patients, by the European Pompe Consortium.17 therefore these patients are being followed up using motor We defined a patient as symptomatic when we identi- function tests to identify changes in motor performance fied muscle weakness in clinical examination using the that can support starting ERT. However it is possible that Medical Research Council score (MRC) or when Forced irreversible muscle degeneration has already started with- Vital Capacity (FVC) seated was lower than 85%. We out influencing motor function. The identification of new included in this study a total of 35 patients: 21 symp- tools to follow-up progression of patients is one of the tomatic patients treated with ERT, nine symptomatic hottest topics in the neuromuscular disorders field. In this patients nontreated with ERT and five asymptomatic sense, quantitative muscle MRI (qMRI) has revealed to be patients. Clinical and genetic data of this Pompe cohort useful in several muscle disorders including Pompe dis- have been previously published.10 Nontreated symp- ease.10,12 However, qMRI is not available in all centers tomatic patients were visited before starting ERT. We also and, experience is needed to analyze it properly. Therefore, studied 10 age and sex-matched healthy controls. other potential biomarkers are needed and serum mole- cules such as microRNAs (miRNAs) could be useful. miRNA isolation in serum sample A good biomarker should be detected in a sample that can be obtained by minimally invasive procedures for the Serum miRNAs were isolated from serum with DANA- patient. In this sense, serum or plasma is the best candi- GENE microRNA and Cell free RNA Kit (DANAGEN, dates.13 miRNAs are small noncoding RNAs that exert their Badalona, Spain) using MS2 RNA (Roche, Basel, Switzer- effect on gene expression at the posttranscriptional level. land) as a carrier to improve the RNA isolation and They have a length of approximately 22 nucleotides and, in exogenous controls (Unisp2,4 and 5). miRNAs were general, they inhibit protein synthesis or favor mRNA retrotranscribed into a cDNA with the miRCURY LNATM degradation by binding to the UTR 3’ regions of those Universal RT microRNA PCR (EXIQON). mRNAs.14 Each miRNA regulates several target genes and a given gene can be regulated by multiple miRNAs.15 miR- Analysis of miRNA expression in serum NAs regulate several processes in skeletal muscle such as muscle growth, atrophy, or regeneration. A specific group miRNA expression studies consisted two different phases: of miRNAs, known as dystromirs are distinctly expressed in discovery and validation. For the discovery phase, we serum of patients with muscle dystrophies. They have been selected five serum samples from four different groups: studied in Duchenne muscle dystrophy (DMD), Becker controls, asymptomatic AOPD, symptomatic nontreated muscle dystrophy of facio-scapulo-humeral muscle dystro- AOPD and symptomatic AOPD-treated patients. For the phy (FSHD), but they have not been analyzed in serum validation phase, we used samples from 10 healthy con- samples of patients with AOPD.16 trol and 35 Pompe patients. In this project, we studied the serum expression of 190 miRNAs, including the so called dystromirs, in a group of 35 Discovery phase AOPD patients and 10 controls with the following aims: (1) identify miRNAs that were up- or downregulated in AOPD miRNAs were measured by qPCR using Serum/Plasma patients; (2) investigate if they could be used as biomarkers Focus microRNA PCR 384 wells Panels, (V4.M) (EXI- of progression of the disease and; (3) understand the phys- QON). Results were analyzed with software available iopathology of muscle degeneration in Pompe disease. online by Applied Biosystems (SDS 2.4). Mean and SD 1216 ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. A. Carrasco-Rozas et al. microRNAs in Pompe Disease were calculated and a statistical analysis of the raw data Muscle imaging was performed by ANAXOMICS, a consulting company expert in the analysis of miRNA data. The relative levels All patients were examined in a Philips Achieva XR 1.5 of selected serum miRNAs were normalized by using cel- Teslas located at HSCSP. Three-dimensional 3-point miR-39-3p and Unisp6 as spike-in18 and the miRNAs Dixon images were acquired and ROIs were manually has-miR-126-3p, has-miR-151a-3p, and has-miR-21-5p drawn in the areas of interest as previously described10. were identified by applying the geNorm algorithm19, a Mean thighs fat fraction was calculated from the data pair-wise stability-based method previously described.20 acquired and used for the statistical analysis. Validation phase Data analysis The miRNAs identified in the discovery phase were subse- Discovery phase quently studied in the validation phase. Although miR- ANOVA tests were performed to identify differences in 206 was not included in the panels used for the discovery miRNA expression levels between cohorts. P-values phase but was also included in this phase because it is were adjusted for multiple testing using Tukey-Kramer one of the already described dystromirs, together with correction. Dataset with all the variables from the miR-1-3p and miR-133a-3p.21 miR-126-3p was used as individual patients was analyzed using a data science an endogenous normalizer. ExiLENT SYBR Green master strategy based on data mining approach, in which, the mix (QIAGEN, Venlo, Netherlands) and miRCURY LNA following methods were used: as feature selection and miRNA PCR Assays (QIAGEN) to study each miRNAs. extraction algorithms CHOW-LIU, MRMR, RELIEFF, We studied 21 miRNA: 14 problem miRNAs, one nor- RFE-SVM, SFFS, and Wilcoxon with correlation and malizer, two hemolysis controls and four exogenous con- as the base classifier GLM_Binomal and Na€ıve trols. Relative miRNA expression was analyzed using Bayes.24–29 In order to prioritize the generalization comparative Ct method with the SDS 2.4 software.22 capability of the conclusion, a K-fold validation analy- sis was performed, yielding cross-validated quality Analysis of miRNA expression in muscle measures (such as accuracy) for each biomarker. The study was carried out with and without imputed miss- We analyzed the expression levels of miRNAs in muscles ing values. The imputation method applied used the obtained from AOPD patients (n = 4) and healthy con- k-nearest neighbor averaging, using Euclidean metric trols (n = 5). Muscle miRNAs were isolated using in the space of each miRNA by cohort to impute the DANAGENE microRNA Kit (DANAGEN). For this tech- missing values elements.18 nique, we also used MS2 (MS2 RNA, Roche) and exoge- nous controls for the RNA isolation. For miRNA expression analysis, ExiLENT SYBR Green master mix Validation phase (QIAGEN) and different miRCURY LNA miRNA PCR We used nonparametric tests for the statistical analysis of Assays (QIAGEN) were used for each of the miRNAs. the variables. We used the Mann–Whitney U test to We used U6 snRNA (RNU6-2) as an endogenous con- investigate whether there were significant differences in trol to normalize the data.23 the variables between groups (asymptomatic vs symp- tomatic/controls vs Pompe). Muscle function tests The following muscle function tests were performed: Clinical correlation 6 min walking test (6MWT), time to walk 10 meters, We used Spearman’s rank correlation (coefficient timed up-and-go test, time to climb up and down four reported as) to investigate whether there was a correlation steps, and motor function measure-20 items scale (MFM- between serum miRNA expression and the clinical results. 20). Muscle strength was studied using both MRC and As we run multiple correlations, a Bonferroni test was hand-held myometry. ACTIVLIM, INQoL, and SF-36 performed to avoid type 1 error. The results of all statisti- were used as patient reported outcome measures. Forced cal studies were considered significant if P was lower than vital capacity seated and in lying position was obtained 0.05. Statistical studies were performed using IBM SPSS with a spirometer Carefusion Microlab ML 3500 MK8 Statistics software version 21. (Carefusion, Yorba Linda, CA, USA). ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 1217 microRNAs in Pompe Disease A. Carrasco-Rozas et al. significant differences between asymptomatic and symp- Results tomatic-treated AOPD patients (Mann–Whitney U test, P < 0.05) (Fig. 1C). Although we identified other poten- Description of the cohort tial miRNAs in the discovery phase none of them showed We included in the study 35 AOPD patients. Thirty significant differences in the validation phase (Fig. 2).Nor- patients were symptomatic and five were asymptomatic. malized relative expression levels of serum miRNAs in Twenty-one out of the 30 symptomatic patients were Pompe patients. Data are presented as mean  SD. already receiving ERT when the first sample was obtained. Mann–Whitney U test. In the remaining nine symptomatic patients, blood sam- ples were obtained before ERT was started. Demographic Expression of dystromirs in the progression and clinical data of the patients are described in Table 1. of the disease We also included samples from 10 age and sex-matched controls. We decided to study if starting ERT treatment modified serum levels of dystromirs (miR-1-3p, miR-133a-3p and miR-206). We studied samples for three patients that miRNA expression in the discovery phase were obtained before and 6 months to 1 year after start- We observed different expression levels between the ing ERT treatment. Dystromirs serum levels increased in groups of study in nine of the 185 miRNAs included in two patients while they decreased in one of the patients the discovery assay (ANOVA, P < 0.05). Expression levels (Fig. 3). However, we did not apply any statistical study of these nine miRNAs were further analyzed in the valida- due to the small size of our cohort. tion phase. These miRNAs included: miR-106b-5p, miR- 199a-5p, miR-652-3p, miR-423-5p, miR-1-3p, miR-133a- miRNA expression in muscle 3p, miR-186-5p, miR-338-3p, and miR-23a-3p. We also included in the validation phase a group of miRNAs that We observed a statistically significant increase in the con- have been involved in the muscle degeneration process: centration of miR-206 in muscle biopsies from AOPD miR-206, miR-145-5p, miR-181-5p, miR-486 and miR- patients taken before the treatment was started (n = 4) 29a-3p.21 compared to controls (n = 5) (Fig. 4C). Although the dif- ferences observed in miR-1-3p and miR-133a-3p muscle expression were not statistically significant, there was a miRNA expression in the validation phase trend to be increased in patients compared to controls We observed a statistically significant increase in the con- (Fig. 4A and B). centration of miR-1-3p and miR-206 between AOPD patients and controls (Mann–Whitney U test, P < 0.05) Correlations between clinical features and (Fig. 1A and C). We also analyzed if there were differ- miRNAs serum levels ences between the different subgroups of study: asymp- tomatic AOPD, symptomatic nontreated AOPD, We identified a statistically significant correlation between symptomatic-treated AOPD and controls. We observed serum levels of miR-1-3p, miR-133a-3p, and miR-206a in significant differences between symptomatic-treated AOPD patients (Spearman test, P < 0.05 (Table 2). We AOPD patients and controls in the expression of miR-1- also investigated if there were correlations between serum 3p, miR-133a-3p and miR-206 (Mann–Whitney U test, levels dystromirs and clinical features of the patients. We P < 0.05), (Fig. 1). In miR-206, we also observed did not identify any correlation between age at study, age Table 1. Clinical characteristics of the control group and the Pompe patients included in the study. Mean and standard deviation are provided for the quantitative variables. Clinical characteristics Control n = 10 Asymptomatic n = 5 Nontreated n = 9 Treated n = 21 Age 41.2  16.4 24  8.2 50.6  11.4 50.81  8.6 Sex (male/female) 6/4 2/3 6/3 7/15 Time from onset of symptoms – – 9.4  6.6 15  9.1 Time on ERT – – – 5.1  2.5 Aids for walking – – 0 10 Ventilation – – 2 9 1218 ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. A. Carrasco-Rozas et al. microRNAs in Pompe Disease Figure 1. Normalized relative expression levels of validated serum miRNAs in Pompe patients. Data are presented as mean  SD. Mann–Whitney U test *P < 0.05, **P < 0.01 and ***P < 0.001. at start of ERT, time on ERT or delay in the start of known as dystromirs. Dystromirs are only expressed by ERT, and serum levels of miRNAs. In contrast, we identi- heart and skeletal muscle and, therefore, they can poten- fied significant correlations (Spearman test, P < 0.05) tially be a valuable tool to follow muscle pathology in between dystromirs’ serum levels and some of the muscle patients with muscle disorders. The function of dystro- function tests (Table 2). However, the correlation coeffi- mirs in the process of muscle degeneration and regenera- cients were lower than 0.6. There was a correlation tion has been investigated. While miR-133 enhances between mean thighs fat fraction analyzed using quantita- proliferation of myoblasts, miR-1 and miR-206 promote tive muscle MRI and miR-206 serum levels, although as it differentiation from myoblasts to myotubes. Dystromirs happened with muscle function tests, the correlation coef- expression is regulated by MyoD, Myf5, MEF2, SRF, and ficient was lower than 0. other transcription factors involved in muscle develop- ment.37,38 The process of muscle degeneration in Pompe disease is not completely understood. Reduced expression Discussion of alpha-glucosidase leads to a progressive accumulation The aim of this work was to study miRNA profiles in of glycogen inside the lysosomes of muscle fibers disrupt- Pompe disease patients and evaluate their role as potential ing the physiologic process of autophagy. As a result, biomarkers. We studied the expression of 190 miRNAs in glycogen charged lysosomes and autophagic vesicles are AOPD patients and controls and we found increased found in the sarcoplasm, disrupting the contractile prop- expression levels in AOPD patients in three miRNAs: erties of skeletal muscle fibers. Eventually, the muscle miR-1-3p, miR-133a-3p, and miR-206. These differences fiber dies and is replaced by fat and fibrous tissue resem- were more robust when we compared serum levels of bling what happens in muscle dystrophies. In view of this symptomatic patients with controls. Moreover, serum fact, it is reasonable to find increased serum levels of dys- levels of miR-206 were higher in symptomatic than tromirs in AOPD patients, as happens in other muscle presymtomatic patients, suggesting that it could be a valid dystrophies, reinforcing the idea that Pompe disease and biomarker of disease progression. Despite this finding, we muscle dystrophies may share a pathologic process of identified weak correlations between muscle function tests muscle degeneration. and serum levels of miRNAs. As dystromirs are related with the muscle regeneration Several reports have demonstrated that miRNAs are process, it is not surprising that they are increased in potential biomarkers of different diseases, such as cancer, muscle disorders, especially in those patients in whom the liver injury, or heart failure.30–32 The expression profile of process of muscle degeneration is active. We observed miRNAs has also been studied in different muscle diseases that dystromirs were more expressed in symptomatic than such as DMD, myotonic dystrophy, and FSHD.16,33–36 In asymptomatic AOPD patients. Specifically, the expression these studies, several miRNAs have a different pattern of of miR-206 was significantly higher in serum samples of expression when compared to controls. Three of them, symptomatic compared to asymptomatic AOPD patients. the miR-1-3p, miR-133a-3p and miR-206, are persistently Therefore, miR-206 serum levels may have a role as a increased in muscular dystrophies and for this reason are potential biomarker to follow-up presymptomatic patients ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 1219 microRNAs in Pompe Disease A. Carrasco-Rozas et al. Figure 2. Normalized relative expression levels of serum miRNAs in Pompe patients. Data are presented as mean  SD. Mann–Whitney U test. 1220 ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. A. Carrasco-Rozas et al. microRNAs in Pompe Disease Figure 3. Dystromir expression changes in Pompe patients in response to the treatment. Data are presented as fold change values in post-ERT patients regarding their pre-ERT samples. Figure 4. Normalized relative expression levels of muscle miRNAs in Pompe patients. Data are presented as mean  SD. Mann–Whitney U test *P < 0.05. Table 2. Correlation between microRNAs serum levels and clinical features. miR-1-3p miR-133a-3p miR-206 Sig CC Sig CC Sig CC miR-1-3p – – 0.007 0.457** 0 0.761** miR-133a-3p 0.007 0.457** – – 0.003 0.493** miR-206 0.000 0.761** 0.003 0.493** – – MRC score 0.089 – 0.094 – 0.035 0.385* Myometry score 0.899 – 0.803 – 0.337 – 6MWT 0.803 – 0.073 – 0.584 – Time to walk 10 m 0.514 – 0.043 0.378* 0.151 – Timed up-and-go test 0.446 – 0.061 – 0.589 – Time to climb up 4 steps 0.386 – 0.053 – 0.119 – Time to climb down 4 steps 0.855 – 0.12 – 0.211 – MFM-20 0.388 – 0.039 0.373* 0.064 – CVF seated 0.465 – 0,025 0.416* 0.243 – CVF lying 0.286 – 0.057 – 0.535 – Activlim 0.628 – 0.986 – 0.462 – qMRI fat fraction 0.107 – 0.205 – 0.013 0.420* Sig, bilateral significance; CC, correlation coefficient. *Spearman Rho test P < 0.05. **Spearman Rho test P < 0.01. ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 1221 microRNAs in Pompe Disease A. Carrasco-Rozas et al. over the years in order to identify patients in whom the Acknowledgments process of muscle degeneration has started and are at risk of developing symptoms of muscle weakness. We did not We thank the people of the Neuromuscular Disorders observe differences in the serum levels of dystromirs Unit for their comments and continuous support. This between treated and nontreated symptomatic patients. study was supported by a grant from Sanofi-Genzyme However, a higher number of patients should be analyzed (GZ-2015-11342) to Gallardo and has been registered in to draw conclusions on the effect of ERT on the muscle Clinicaltrials.gov (identifier NCT03045042). regeneration process in symptomatic Pompe disease. ERT has been shown to decrease glycogen accumulation in muscle fibers of preclinical animal models of the disease Author Contributions and also in patients. However, recent studies demonstrate ACR performed the experimental part, analyzed the that muscle weakness of treated AOPD patient progresses results, performed the statistical analysis and prepared the despite the treatment and that muscle fatty replacement, figures, and wrote the paper. EFS and CL designed the which is related to muscle degeneration process, increases protocol, analyzed the results, and reviewed the paper. IB, over the years.9,10 Our finding supports that the process IP. and EM designed the protocol, visited the patients, of muscle degeneration remains active despite the starting and acquired motor function data. CNP and JL visited of the treatment in Pompe patients. In accordance, we the patients and reviewed the paper. SS coordinated the did not observe a significant reduction in miRNA expres- study, visited the patients and reviewed the paper. NDL sion levels in serum samples obtained before and after the and XSC designed the protocol and reviewed the paper. treatment. Recently, Tarallo et al studied the profile of II designed and discussed the protocol. JDM designed the miRNAs in Pompe disease and observed high expression protocol, visited the patients, analyzed the images, per- levels of miR-133a-3p both in IOPD and AOPD formed the statistical analysis, and wrote the paper. EG patients.39 These results are in accordance with our inves- designed the protocol, analyzed the results, wrote the tigations, and suggest that miRNAs could be potential paper, and obtained funding for the study. biomarkers of Pompe disease. Other miRNAs have a role in skeletal muscle home- ostasis both in health and in disease-regulating myogene- Conflict of Interest sis, muscle atrophy or structural and metabolic changes This study (GZ-2015-11342) has been performed with related to aging or exercise21. These miRNAs include funding from SANOFI/Genzyme. miR-23a, miR-29, miR-146a, miR-486, and miR-181 among others. We have studied serum expression levels References of these miRNAs, but we did not find significant differ- ences between AOPD patients and controls. However, 1. van der Ploeg AT, Reuser AJ. Pompe’s disease. Lancet although our cohort included 35 patients, it is possible 2008;372:1342–1353. that higher number of samples are needed to identify sub- 2. Nascimbeni AC, Fanin M, Masiero E, et al. The role of tle changes in miRNA expression in serum. In this sense, autophagy in the pathogenesis of glycogen storage disease analysis of miRNA expression using skeletal muscle biop- type II (GSDII). Cell Death Differ 2012;19:1698–1708. sies should show higher differences. Nevertheless, muscle 3. Dasouki M, Jawdat O, Almadhoun O, et al. Pompe biopsies of patients are not always available. We studied disease: literature review and case series. Neurol Clin miRNA expression of muscle biopsies and we confirmed 2014;32:751–776. higher expression levels of miR-206 in AOPD patients 4. Manganelli F, Ruggiero L. Clinical features of Pompe than in controls. Although miR-1-3p and miR-133a-3p disease. Acta Myol 2013;32:82–84. levels were increased in AOPD muscle biopsies, we did 5. Desnuelle C, Salviati L. Challenges in diagnosis and not find significant differences with controls, mainly due treatment of late-onset Pompe disease. Curr Opin Neurol to a high variability of miRNA expression. 2011;24:443–448. In conclusion, we have identified a group of miR- 6. van der Beek NA, de Vries JM, Hagemans ML, et al. Clinical features and predictors for disease natural NAs with increased levels of expression in blood sam- progression in adults with Pompe disease: a nationwide ples of patients with Pompe disease. The role of these prospective observational study. Orphanet J Rare Dis miRNAs in muscle regeneration has been previously 2012;7:88. reported and suggests that there is a common patho- 7. Nicolino M, Byrne B, Wraith JE, et al. Clinical outcomes logical pathway between Pompe disease and muscle after long-term treatment with alglucosidase alfa in infants dystrophies. 1222 ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. A. Carrasco-Rozas et al. microRNAs in Pompe Disease and children with advanced Pompe disease. Genet Med the 2(-Delta Delta C(T)) Method. Methods 2001;25:402– 2009;11:210–219. 408. 8. Katzin LW, Amato AA. Pompe disease: a review of the 23. Sauer E, Babion I, Madea B, Courts C. An evidence current diagnosis and treatment recommendations in the based strategy for normalization of quantitative PCR era of enzyme replacement therapy. J Clin Neuromuscul data from miRNA expression analysis in forensic organ Dis 2008;9:421–431. tissue identification. Forensic Sci Int Genet 2014;13: 9. Kuperus E, Kruijshaar ME, Wens SCA, et al. Long-term 217–223. benefit of enzyme replacement therapy in Pompe disease: a 24. Peng H, Long F, Ding C. Feature selection based on 5-year prospective study. Neurology 2017;89:2365–2373. mutual information: criteria of max-dependency, max- 10. Figueroa-Bonaparte S, Segovia S, Llauger J, et al. Muscle relevance, and min-redundancy. IEEE Trans Pattern Anal MRI findings in childhood/adult onset Pompe disease Mach Intell 2005;27:1226–1238. correlate with muscle function. PLoS ONE 2016;11: 25. Kira K, Rendell LA. The feature selection problem: e0163493. traditional methods and a new algorithm. Proc of AAAI- 11. Schoser B, Laforet P, Kruijshaar ME, et al. 208th ENMC 92 1992:122–126. International Workshop: formation of a European 26. Guyon I, Weston J, Bamhill S, Vapnik V. Gene selection Network to develop a European data sharing model and for cancer classification using support vector machines. treatment guidelines for Pompe disease Naarden, The Machine Learn 2002;46:389–422. Netherlands, 26–28 September 2014. Neuromuscul Disord 27. Ververidis D, Kotropoulos C. Fast and accurate sequential 2015;25:674–678. floating forward feature selection with the Bayes classifier 12. Figueroa-Bonaparte S, Llauger J, Segovia S, et al. applied to speech emotion recognition. Signal Process Quantitative muscle MRI to follow up late onset Pompe 2008;88:2956–2970. patients: a prospective study. Sci Rep 2018;8:10898. 28. Christin C, Hoefsloot HC, Smilde AK, et al. A critical 13. Cox TM. Biomarkers in lysosomal storage diseases. 2006. assessment of feature selection methods for biomarker 14. Chen JF, Callis TE, Wang DZ. microRNAs and muscle discovery in clinical proteomics. Mol Cell Proteomics disorders. J Cell Sci 2009;122(Pt 1):13–20. 2013;12:263–276. 15. Guller I, Russell AP. MicroRNAs in skeletal muscle: their 29. Russell SJ, Norvig P. Artificial intelligence: a modern role and regulation in development, disease and function. approach. Pearson. Education 2003. J Physiol 2010;588(Pt 21):4075–87. 30. Romero-Cordoba SL, Salido-Guadarrama I, Rodriguez- 16. Matsuzaka Y, Kishi S, Aoki Y, et al. Three novel serum Dorantes M, Hidalgo-Miranda A. miRNA biogenesis: biomarkers, miR-1, miR-133a, and miR-206 for Limb- biological impact in the development of cancer. Cancer girdle muscular dystrophy, Facioscapulohumeral muscular Biol Ther 2014;15:1444–1455. dystrophy, and Becker muscular dystrophy. Environ 31. McDaniel K, Herrera L, Zhou T, et al. The functional role Health Prev Med 2014;19:452–458. of microRNAs in alcoholic liver injury. J Cell Mol Med 17. van der Ploeg AT, Kruijshaar ME, Toscano A, et al. 2014;18:197–207. European consensus for starting and stopping enzyme 32. Katz MG, Fargnoli AS, Williams RD, et al. MiRNAs as replacement therapy in adult patients with Pompe disease: potential molecular targets in heart failure. Future Cardiol a 10-year experience. Eur J Neurol 2017;24:768-e31. 2014;10:789–800. 18. Marabita F, de Candia P, Torri A, et al. Normalization of 33. Zaharieva IT, Calissano M, Scoto M, et al. Dystromirs as circulating microRNA expression data obtained by serum biomarkers for monitoring the disease severity in quantitative real-time RT-PCR. Brief Bioinform Duchenne muscular Dystrophy. PLoS ONE 2013;8: 2016;17:204–212. e80263. 19. Perkins JR, Dawes JM, McMahon SB, et al. ReadqPCR 34. Perfetti A, Greco S, Cardani R, et al. Validation of plasma and NormqPCR: R packages for the reading, quality microRNAs as biomarkers for myotonic dystrophy type 1. checking and normalisation of RT-qPCR quantification Sci Rep 2016;6:38174. cycle (Cq) data. BMC Genom 2012;13:296. 35. Koutsoulidou A, Kyriakides TC, Papadimas GK, et al. 20. Vandesompele J, De Preter K, Pattyn F, et al. Accurate Elevated muscle-specific miRNAs in serum of myotonic normalization of real-time quantitative RT-PCR data by dystrophy patients relate to muscle disease progress. PLoS geometric averaging of multiple internal control genes. ONE 2015;10:e0125341. Genome Biol 2002;3:RESEARCH0034. 36. Perfetti A, Greco S, Bugiardini E, et al. Plasma microRNAs 21. Sharma M, Juvvuna PK, Kukreti H, McFarlane C. Mega as biomarkers for myotonic dystrophy type 1. roles of microRNAs in regulation of skeletal muscle health Neuromuscul Disord 2014;24:509–515. and disease. Front Physiol 2014;5:239. 37. Kim HK, Lee YS, Sivaprasad U, et al. Muscle-specific 22. Livak KJ, Schmittgen TD. Analysis of relative gene microRNA miR-206 promotes muscle differentiation. J expression data using real-time quantitative PCR and Cell Biol 2006;174:677–687. ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association. 1223 microRNAs in Pompe Disease A. Carrasco-Rozas et al. 38. Rosenberg MI, Georges SA, Asawachaicharn A, et al. 39. Tarallo A, Carissimo A, Gatto F, et al. microRNAs as MyoD inhibits Fstl1 and Utrn expression by inducing biomarkers in Pompe disease. Genet Med 2019;21:591– transcription of miR-206. J Cell Biol 2006;175:77–85. 600. 1224 ª 2019 The Authors. Annals of Clinical and Translational Neurology published by Wiley Periodicals, Inc on behalf of American Neurological Association.