Pediatr Blood Cancer
Circulating Serum miRNAs as Potential Biomarkers for Nephroblastoma Dr. Nicole Ludwig,
1 PhD, *
Nasenien Nourkami-Tutdibi, MD,2 Christina Backes, PhD,3 Hans-Peter Lenhof, Norbert Graf, MD,2 Andreas Keller, PhD,3 and Eckart Meese, PhD1
Background. Nephroblastoma (or Wilms tumor—WT) is the most common childhood kidney cancer. In Europe, nephroblastoma is treated with preoperative chemotherapy without histological confirmation by biopsy. Therefore, minimal-invasive diagnostic markers confirming nephroblastoma diagnosis are highly warranted. Procedure. In our study, we aim to identify circulating miRNAs with diagnostic potential for differentiating nephroblastoma from controls. We determined the level of 19 miRNAs in serum of 32 patients with nephroblastoma and 12 controls with quantitative real-time PCR. Three miRNAs were further tested in an independent validation set including sera of patients with renal tumors other than Wilms. Results. In total, 14 miRNAs showed significantly higher abundance in serum of patients with nephroblastoma than in controls. The miRNAs with highest diagnostic potentials included miRs-130b-3p,
PhD,
4
-100-5p, and -143-3p with an AUC of 0.94, 0.90, and 0.89, respectively. A signature based on these three miRNAs to differentiated patients from controls with an accuracy of 84.58%, a sensitivity of 76.67%, and a specificity of 92.5%. Higher expression of miRs100-5p and -130b-3p was confirmed in an independent validation set. The signature based on miRs-100-5p and -130b-3p differentiated patients with nephroblastoma from healthy controls with an accuracy, sensitivity, and specificity of 79.6%, 69.2%, and 90.0%, respectively. Conclusion. In summary, we provide first evidence that serum miR-100-5p and -130b-3p hold potential as biomarker for WT irrespective of the subtype and that expression level of these miRNA in serum is unaffected by differences in serum collection. Pediatr Blood Cancer # 2015 Wiley Periodicals, Inc.
Key words: circulating miRNA; miR-130b-3p; miR-100-5p; miR-143-3p; serum; Wilms tumor
INTRODUCTION Wilms tumor (WT), or nephroblastoma, is the sixth most common human childhood malignancy and the most common renal tumor in children [1,2]. In up to 20% WTs are diagnosed incidentally during regular check-up visits as they generate no specific symptoms. Patients are treated in prospective randomized multicenter trials mainly conducted by the Children’s Oncology Group (COG) in North America and the Socie´te´ Internationale d’Oncologie Pe´diatrique (SIOP) in Europe. Current treatment according to SIOP includes preoperative chemotherapy for reduction of tumor volume, followed by tumor resection and postoperative adjuvant therapy stratified according to the local stage and the histological subtype [3]. After preoperative chemotherapy WT can be classified into low-, intermediate-, and high-risk tumors according to the revised histological Stockholm classification [4]. Prognosis of nephroblastoma is mainly depending on age, stage, histological subtype, and preoperative tumor volume. According to the SIOP approach treatment is solely based on imaging techniques without biopsies. Therefore blood borne biomarkers to assure WT diagnosis are of utmost importance and are highly warranted. MiRNAs are short (about 22 nucleotides), single stranded RNAs that regulate gene expression at a posttranscriptional level by either inhibiting translation or promoting mRNA degradation [5]. There is increasing evidence of supposedly tumor-derived miRNAs circulating in serum or plasma, and that these circulating miRNAs have diagnostic or prognostic potential [6–12]. In serum, miRNAs are either protein-bound or included in microvesicles or exosomes, which offers a certain protection against degradation by RNases [13]. Therefore they are remarkably stable in serum, which is one prerequisite for any potential biomarker. Here, we analyzed the expression of 19 miRNAs in serum of patients with WT before onset of therapy and non-tumor controls with quantitative real-time PCR (qRT-PCR). We tried to identify miRNAs differentially expressed between patients with WT and controls and between patients with regressive, triphasic and blastemal histology and to validate potential candidates in an independent sample set. Furthermore, we tested the influence of C
2015 Wiley Periodicals, Inc. DOI 10.1002/pbc.25481 Published online in Wiley Online Library (wileyonlinelibrary.com).
storage time and temperature of blood vials on the stability of the potential marker miRNAs in a cohort of controls.
METHODS Selection of Candidate miRNAs for WT For serum testing, we selected 19 miRNAs that were overexpressed in WT tissue in general or in specific subtypes as published in literature or in our own data. In detail, miRs-17-5p, -24-3p, -1005p, and -223-3p have been previously described as deregulated in tissue samples of high-risk WTs compared to intermediate risk WTs and miRs-17-5p, -18a-5p, -130b-3p, -181b-5p, -320a, and -335-5p are upregulated in WT tissue compared to normal kidney or other renal tumors [14,15]. Furthermore, we analyzed miRNA microarray data of WT tissue produced in our lab for the study of Wegert et al. [16] that has been recently accepted for publication in “Cancer Cell” (GEO dataset GSE57370). We found overexpression of let-7a5p, let-7f-5p and miRs-17-5p, -18a-5p, -24-3p, -101-3p, -126-3p, Additional Supporting Information may be found in the online version of this article at the publisher’s web-site. 1
Department of Human Genetics, Saarland University, Homburg/Saar, Germany; 2Department of Pediatric Oncology and Hematology, Medical School, Saarland University, Homburg, Germany; 3Chair for Clinical Bioinformatics, Saarland University, Saarbru¨cken, Germany; 4 Center for Bioinformatics, Saarland University, Saarbru¨cken, Germany Grant sponsor: European Union’s Seventh Framework Programme for Research, Technological Development and Demonstration; Grant number: 600841 Conflict of interest: Nothing to declare. Andreas Keller and Eckart Meese contributed equally as senior author.
Correspondence to: Nicole Ludwig, Department of Human Genetics, Medical Faculty, Saarland University, Building 60, Kirrberger Strasse, D-66421 Homburg/Saar, Germany. E-mail: [email protected] Received 9 October 2014; Accepted 28 January 2015
2
Ludwig et al.
-143-3p, -195-5p, -320a, -335-5p, and -3651 and downregulation of miRs-181b-5p, -223-3p, -638, and -1825 in blastemal WT compared to regressive WT (fold change >2; FDR-adjusted unpaired t-test, P < 0.05). MiR-572 was also more than twofold downregulated in blastemal WT, but did not reach statistical significance.
Patient Samples To identify miRNA signatures in serum of patients with WT and to determine whether or not the signatures differ between subtypes, we selected a cohort of 32 patients before preoperative chemotherapy and 12 control subjects without malignant disease (termed initial sample set). All serum samples were taken with patient’s informed consent. Serum of patients was collected as part of the SIOP 2001 trial in Germany, serum of controls was collected at the Department of Pediatric Oncology and Hematology at Saarland University. Serum was stored at 80˚C. The research was approved ¨ rztekammer by the local ethical committee (Ethikkommission der A des Saarlandes, No. 136/01). A second, independent set of 65 serum samples (termed validation set), including sera of 43 patients with WT, 9 patients with renal tumors other than WT and 13 healthy controls, was used to validate differential expression of candidate miRNAs. Summary of clinical data of initial and validation set are included as Supplemental Table SI.
miRNA Isolation Total RNA including miRNAs was isolated using miRNeasy Serum/Plasma kit (Qiagen, Hilden, Germany) according to manufacturer’s instructions. In short, serum was centrifuged 10 min at 12,000 rpm to remove possible cellular contaminations. For each patient, 200 ml serum was mixed with Qiazol lysis reagent and incubated 5 min at RT. After addition of 3.5 ml cel-miR-39 spike-in (1.6 108 copies/ml), 200 ml chloroform was added, the sample was vigorously shaken and incubated at RT for 3 min followed by centrifugation 20 min at 12,000 rpm and 4˚C for phase separation. The aqueous phase (exactly 600 ml) was mixed with 3 ml glycogen (20 mg/ml) and placed in a Qiacube instrument for automated purification using the standard protocol for Serum/ Plasma kit (15 ml elution volume). RNA was stored at 80˚C.
Quantitative Real-time PCR We performed reverse transcription and quantitative real-time PCR using miScript II RT kit (Qiagen) and miScript SYBR1-Green PCR Kit (Qiagen) in a StepOnePlus instrument (Life Technologies, Carlsbad, CA) according to manufacturers instructions with minor variations. In detail, reverse transcription reaction was diluted 1:5 instead of 1:10 in RNase-free water and 1 ml of the dilution was used for quantitative real-time PCR in a 20ml instead of 25 ml PCR volume. PCR reactions were set up in duplicates and mean Ct and standard deviation was computed with StepOnePlus Software. All Mean Ct values with standard deviations >0.5 were omitted from further analysis, all “Ct undetermined” values were set to 40. Raw Ct values and a MIQE compliant description of the experiment are attached as Supplemental Material. Since there is no standardized internal control for normalization of serum miRNAs, we used the spiked-in cel-miR-39 as reference for DCt calculation (DCtmiRNA-X ¼ CtmiRNA-X – Ctcel-miR-39). Fold change between two groups of samples (control, triphasic, regressive, or blastemal) was calculated with FC ¼ 2DDCt. For all sera used in the study, DCtmiR-23a-3p–miR-451a as indicator of risk of hemolysis was
Circulating Serum miRNAs as Potential Biomarkers for Nephroblastoma Dr. Nicole Ludwig,
1 PhD, *
Nasenien Nourkami-Tutdibi, MD,2 Christina Backes, PhD,3 Hans-Peter Lenhof, Norbert Graf, MD,2 Andreas Keller, PhD,3 and Eckart Meese, PhD1
Background. Nephroblastoma (or Wilms tumor—WT) is the most common childhood kidney cancer. In Europe, nephroblastoma is treated with preoperative chemotherapy without histological confirmation by biopsy. Therefore, minimal-invasive diagnostic markers confirming nephroblastoma diagnosis are highly warranted. Procedure. In our study, we aim to identify circulating miRNAs with diagnostic potential for differentiating nephroblastoma from controls. We determined the level of 19 miRNAs in serum of 32 patients with nephroblastoma and 12 controls with quantitative real-time PCR. Three miRNAs were further tested in an independent validation set including sera of patients with renal tumors other than Wilms. Results. In total, 14 miRNAs showed significantly higher abundance in serum of patients with nephroblastoma than in controls. The miRNAs with highest diagnostic potentials included miRs-130b-3p,
PhD,
4
-100-5p, and -143-3p with an AUC of 0.94, 0.90, and 0.89, respectively. A signature based on these three miRNAs to differentiated patients from controls with an accuracy of 84.58%, a sensitivity of 76.67%, and a specificity of 92.5%. Higher expression of miRs100-5p and -130b-3p was confirmed in an independent validation set. The signature based on miRs-100-5p and -130b-3p differentiated patients with nephroblastoma from healthy controls with an accuracy, sensitivity, and specificity of 79.6%, 69.2%, and 90.0%, respectively. Conclusion. In summary, we provide first evidence that serum miR-100-5p and -130b-3p hold potential as biomarker for WT irrespective of the subtype and that expression level of these miRNA in serum is unaffected by differences in serum collection. Pediatr Blood Cancer # 2015 Wiley Periodicals, Inc.
Key words: circulating miRNA; miR-130b-3p; miR-100-5p; miR-143-3p; serum; Wilms tumor
INTRODUCTION Wilms tumor (WT), or nephroblastoma, is the sixth most common human childhood malignancy and the most common renal tumor in children [1,2]. In up to 20% WTs are diagnosed incidentally during regular check-up visits as they generate no specific symptoms. Patients are treated in prospective randomized multicenter trials mainly conducted by the Children’s Oncology Group (COG) in North America and the Socie´te´ Internationale d’Oncologie Pe´diatrique (SIOP) in Europe. Current treatment according to SIOP includes preoperative chemotherapy for reduction of tumor volume, followed by tumor resection and postoperative adjuvant therapy stratified according to the local stage and the histological subtype [3]. After preoperative chemotherapy WT can be classified into low-, intermediate-, and high-risk tumors according to the revised histological Stockholm classification [4]. Prognosis of nephroblastoma is mainly depending on age, stage, histological subtype, and preoperative tumor volume. According to the SIOP approach treatment is solely based on imaging techniques without biopsies. Therefore blood borne biomarkers to assure WT diagnosis are of utmost importance and are highly warranted. MiRNAs are short (about 22 nucleotides), single stranded RNAs that regulate gene expression at a posttranscriptional level by either inhibiting translation or promoting mRNA degradation [5]. There is increasing evidence of supposedly tumor-derived miRNAs circulating in serum or plasma, and that these circulating miRNAs have diagnostic or prognostic potential [6–12]. In serum, miRNAs are either protein-bound or included in microvesicles or exosomes, which offers a certain protection against degradation by RNases [13]. Therefore they are remarkably stable in serum, which is one prerequisite for any potential biomarker. Here, we analyzed the expression of 19 miRNAs in serum of patients with WT before onset of therapy and non-tumor controls with quantitative real-time PCR (qRT-PCR). We tried to identify miRNAs differentially expressed between patients with WT and controls and between patients with regressive, triphasic and blastemal histology and to validate potential candidates in an independent sample set. Furthermore, we tested the influence of C
2015 Wiley Periodicals, Inc. DOI 10.1002/pbc.25481 Published online in Wiley Online Library (wileyonlinelibrary.com).
storage time and temperature of blood vials on the stability of the potential marker miRNAs in a cohort of controls.
METHODS Selection of Candidate miRNAs for WT For serum testing, we selected 19 miRNAs that were overexpressed in WT tissue in general or in specific subtypes as published in literature or in our own data. In detail, miRs-17-5p, -24-3p, -1005p, and -223-3p have been previously described as deregulated in tissue samples of high-risk WTs compared to intermediate risk WTs and miRs-17-5p, -18a-5p, -130b-3p, -181b-5p, -320a, and -335-5p are upregulated in WT tissue compared to normal kidney or other renal tumors [14,15]. Furthermore, we analyzed miRNA microarray data of WT tissue produced in our lab for the study of Wegert et al. [16] that has been recently accepted for publication in “Cancer Cell” (GEO dataset GSE57370). We found overexpression of let-7a5p, let-7f-5p and miRs-17-5p, -18a-5p, -24-3p, -101-3p, -126-3p, Additional Supporting Information may be found in the online version of this article at the publisher’s web-site. 1
Department of Human Genetics, Saarland University, Homburg/Saar, Germany; 2Department of Pediatric Oncology and Hematology, Medical School, Saarland University, Homburg, Germany; 3Chair for Clinical Bioinformatics, Saarland University, Saarbru¨cken, Germany; 4 Center for Bioinformatics, Saarland University, Saarbru¨cken, Germany Grant sponsor: European Union’s Seventh Framework Programme for Research, Technological Development and Demonstration; Grant number: 600841 Conflict of interest: Nothing to declare. Andreas Keller and Eckart Meese contributed equally as senior author.
Correspondence to: Nicole Ludwig, Department of Human Genetics, Medical Faculty, Saarland University, Building 60, Kirrberger Strasse, D-66421 Homburg/Saar, Germany. E-mail: [email protected] Received 9 October 2014; Accepted 28 January 2015
2
Ludwig et al.
-143-3p, -195-5p, -320a, -335-5p, and -3651 and downregulation of miRs-181b-5p, -223-3p, -638, and -1825 in blastemal WT compared to regressive WT (fold change >2; FDR-adjusted unpaired t-test, P < 0.05). MiR-572 was also more than twofold downregulated in blastemal WT, but did not reach statistical significance.
Patient Samples To identify miRNA signatures in serum of patients with WT and to determine whether or not the signatures differ between subtypes, we selected a cohort of 32 patients before preoperative chemotherapy and 12 control subjects without malignant disease (termed initial sample set). All serum samples were taken with patient’s informed consent. Serum of patients was collected as part of the SIOP 2001 trial in Germany, serum of controls was collected at the Department of Pediatric Oncology and Hematology at Saarland University. Serum was stored at 80˚C. The research was approved ¨ rztekammer by the local ethical committee (Ethikkommission der A des Saarlandes, No. 136/01). A second, independent set of 65 serum samples (termed validation set), including sera of 43 patients with WT, 9 patients with renal tumors other than WT and 13 healthy controls, was used to validate differential expression of candidate miRNAs. Summary of clinical data of initial and validation set are included as Supplemental Table SI.
miRNA Isolation Total RNA including miRNAs was isolated using miRNeasy Serum/Plasma kit (Qiagen, Hilden, Germany) according to manufacturer’s instructions. In short, serum was centrifuged 10 min at 12,000 rpm to remove possible cellular contaminations. For each patient, 200 ml serum was mixed with Qiazol lysis reagent and incubated 5 min at RT. After addition of 3.5 ml cel-miR-39 spike-in (1.6 108 copies/ml), 200 ml chloroform was added, the sample was vigorously shaken and incubated at RT for 3 min followed by centrifugation 20 min at 12,000 rpm and 4˚C for phase separation. The aqueous phase (exactly 600 ml) was mixed with 3 ml glycogen (20 mg/ml) and placed in a Qiacube instrument for automated purification using the standard protocol for Serum/ Plasma kit (15 ml elution volume). RNA was stored at 80˚C.
Quantitative Real-time PCR We performed reverse transcription and quantitative real-time PCR using miScript II RT kit (Qiagen) and miScript SYBR1-Green PCR Kit (Qiagen) in a StepOnePlus instrument (Life Technologies, Carlsbad, CA) according to manufacturers instructions with minor variations. In detail, reverse transcription reaction was diluted 1:5 instead of 1:10 in RNase-free water and 1 ml of the dilution was used for quantitative real-time PCR in a 20ml instead of 25 ml PCR volume. PCR reactions were set up in duplicates and mean Ct and standard deviation was computed with StepOnePlus Software. All Mean Ct values with standard deviations >0.5 were omitted from further analysis, all “Ct undetermined” values were set to 40. Raw Ct values and a MIQE compliant description of the experiment are attached as Supplemental Material. Since there is no standardized internal control for normalization of serum miRNAs, we used the spiked-in cel-miR-39 as reference for DCt calculation (DCtmiRNA-X ¼ CtmiRNA-X – Ctcel-miR-39). Fold change between two groups of samples (control, triphasic, regressive, or blastemal) was calculated with FC ¼ 2DDCt. For all sera used in the study, DCtmiR-23a-3p–miR-451a as indicator of risk of hemolysis was
