Human Pathology (2014) 45, 1995–2005
www.elsevier.com/locate/humpath
Original contributions
Identification of putative pathogenic microRNA and its downstream targets in anaplastic lymphoma kinase–negative anaplastic large cell lymphoma☆ Meenakshi Mehrotra PhD, L. Jeffrey Medeiros MD, Rajyalakshmi Luthra PhD, Rachel L. Sargent MD, Hui Yao PhD, Bedia A. Barkoh MS, Rajesh Singh PhD, Keyur P. Patel MD, PhD ⁎ Department of Hematopathology, The University of Texas MD Anderson Cancer Center, Houston, Texas 77030 Received 7 January 2014; revised 10 June 2014; accepted 19 June 2014
Keywords: Anaplastic large cell lymphoma; microRNA; Expression profiling; RT-PCR; In situ hybridization
Summary Anaplastic large cell lymphomas (ALCL) are tumors of T/null-cell lineage characterized by uniform CD30 expression. The 2008 World Health Organization classification subdivided ALCLs into 2 groups: anaplastic lymphoma kinase (ALK)–positive (established entity) and ALK-negative (proposed new entity) ALCL. The genetic basis for the pathogenesis of newly categorized ALK− ALCL is poorly understood. In this study, we used microRNA microarray analysis to identify differentially expressed microRNAs in ALK+ and ALK− ALCL. ALK− ALCL showed significantly higher expression of miR155 (0.888 ± 0.228) compared with ALK+ ALCL (0.0565 ± 0.009) on microarray and by quantitative real-time polymerase chain reaction in ALK− ALCL compared with ALK+ ALCL (P b .05) with a strong correlation between the 2 platforms (R = 0.9, P b .0003). A novel in situ hybridization method allows direct visualization of expression patterns and relative quantitation of miR-155 (mean score, 2.3 versus 1.3; P = .01) for the first time in tissue sections of ALCL. Among computationally predicted targets of miR-155, we identified ZNF652 (r = −0.57, P = .05), BACH1 (r = 0.88, P = .02), RBAK (r = 0.81, P = .05), TRIM32 (r = 0.92, P = .01), E2F2 (r = 0.81, P = .05), and TP53INP1 (r = −0.31, P = .03) as genes whose expression by quantitative real-time polymerase chain reaction correlated significantly with the level of miR155 in ALCL tumor tissue. © 2014 Elsevier Inc. All rights reserved.
1. Introduction Anaplastic large cell lymphoma (ALCL) is a non-Hodgkin lymphoma of T/null-cell lineage that is characterized by uniform ☆
Competing interests: The authors declare no conflicts of interest. ⁎ Corresponding author at: Department of Hematopathology, Unit 149, 1515 Holcombe Blvd, MD Anderson Cancer Center, Houston, Texas 77030. E-mail address: [email protected] (K. P. Patel). http://dx.doi.org/10.1016/j.humpath.2014.06.012 0046-8177/© 2014 Elsevier Inc. All rights reserved.
CD30 expression. In the 2008 World Health Organization (WHO) classification of lymphomas, 2 distinct entities of ALCL were proposed: anaplastic lymphoma kinase (ALK)– positive ALCL and ALK-negative ALCL [1]. ALK+ ALCL is an established entity that carries a t(2;5)(p23;q35) or variant ALK abnormalities resulting in aberrant expression of ALK with ALCL [2-4] and shows favorable prognosis when treated with chemotherapy. In contrast, ALK− ALCL is currently
1996 considered as a provisional category in the WHO classification [5] that closely resembles ALK+ ALCL morphologically and expresses CD30 with similar intensity; however, these tumors lack ALK translocations or overexpress ALK, and affected patients have a poorer prognosis when treated with chemotherapy [2-4]. Significantly, no definitive pathogenic mechanisms have been identified for ALK− ALCL. Genomic and expression profiling studies have shown different genetic aberrations and messenger RNA (mRNA) expression patterns in ALK+ and ALK− ALCL [6-8]. The functional importance of these findings and their roles in pathogenesis are far better understood in ALK+ ALCL than in ALK− ALCL. Improved understanding of ALK− ALCL pathogenesis may lead to identification of potential therapeutic targets that could guide therapy or predict outcome. MicroRNAs (miRNAs) are a class of approximately small (20-22 nucleotides) noncoding RNA molecules that play crucial roles in regulating gene expression, by either inducing mRNA degradation or inhibiting translation. These noncoding RNAs can control a wide range of biologic functions including differentiation, proliferation, and apoptosis. Deregulation of miRNA expression has been shown to play an important role in oncogenesis [9-11]. Deregulated expression of specific miRNAs has been described in solid cancers, leukemias, and lymphomas [12-16], and there are limited studies that have shown different miRNA profiles in ALK+ and ALK− ALCLs [17-20]. These approaches, to date, however, have been applied to the study of whole tumor samples including neoplastic cells and the microenvironment. It would be helpful to assess the tissue and subcellular distribution pattern of individual miRNAs in ALCLs using in situ hybridization (ISH), which could provide important development and homeostasis information as well as a convenient diagnostic tool [15,21,22]. In this study, we used miRNA arrays to identify miRNA that could underlie pathogenesis of ALK− ALCL by screening in ALK+ ALCL and ALK− ALCL tumor samples. miR-155 was found to be expressed at significantly higher level in ALK− ALCL. We demonstrated tissue and subcellular distribution of miR-155 using ISH in ALK+ ALCL and ALK− ALCL tumor samples. We also identified ZNF652, BACH1, RBAK, BACH1, E2F2, and TP53INP1 as potential targets of miR-155, which were significantly correlated with the level of miR-155 in ALCL tumor tissue samples. The ability to visualize miRNA expression patterns and perform relative quantitation in routinely processed, paraffin-embedded tissue sections provides a convenient diagnostic tool to assess differential expression of miRNAs in hematopoietic neoplasms such as ALCL.
2. Materials and methods 2.1. Patient samples This study was performed with approval from the institutional review board. Formalin-fixed, paraffin-embedded
M. Mehrotra et al. (FFPE) tissue blocks of cases of ALK+ ALCL (n = 17), ALK− ALCL (n = 18), and reactive lymph node controls (n = 4) were retrieved from the archives of the Department of Hematopathology, The University of Texas MD Anderson Cancer Center. The diagnosis of ALCL was based on morphological and immunophenotypic criteria as described in the WHO classification [1]. Original data were received and confirmed by 2 different pathologists.
2.2. Cell lines and reagents The Karpas 299, SUDHL-1, and SUPM2 cell lines (all ALK+ ALCL) were a gift from Suzanne Turner (Division of Molecular Histopathology, Department of Pathology, University of Cambridge, Addenbrooke's Hospital, Cambridge, United Kingdom). The FE-PD and Mac2A cell lines (both ALK− ALCL) were gifts from Annarosa del Mistro (Veneto Oncology Institute, Padua, Italy) and Marshall Kadin (Department of Dermatology and Skin Surgery, Roger Williams Medical Center, Boston University School of Medicine, Providence, RI). Cells were cultured in RPMI-1640 medium (catalog no. 30-2001; American Type Culture Collection, Manassas, VA, USA) containing 1% penicillin and streptomycin and 10% fetal bovine serum (FBS) (catalog no. 30-2020; American Type Culture Collection) (for Karpas 299, SUPM2, and FE-PD), 15% FBS (for Mac2A), and 20% FBS (for SUDHL-1).
2.3. RNA isolation Total RNA was isolated from microdissected FFPE tumor samples using Recover All Total Nucleic Acid Isolation Kit for FFPE Tissues (Applied Biosystems, Foster City, CA). Because this kit allows isolation of both miRNA and mRNA, the same sample can be used for miRNA profiling and miRNA and mRNA validation studies. RNA concentrations were determined using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE).
2.4. miRNA microarray experiments Human miRNA Microarray version 3 (Agilent Technologies, Santa Clara, CA) chips containing oligonucleotide probes for 866 human miRNA and 89 human viral miRNAs in a 15 K format, where each target is represented 10 times on a hybridization array, were used to generate miRNA profiles of 11 study cases, specifically 6 ALK+ ALCL and 5 ALK− ALCL according to manufacturer’s protocol (detailed methods for microarray experiments are explained in Supplementary Material). Commercially available universal human miRNA pool (Agilent Technologies) was used as a technical control for hybridization in the miRNA profiling experiments.
miR-155 expression and its targets in ALCL Table 1
1997
Differentially expressed miRNAs in ALK− ALCL and ALK+ ALCL by microarray analysis
miRNAs with higher expression in ALK− ALCL
miRNAs with lower expression in ALK− ALCL
Gene ID
Fold change
P
Gene ID
Fold change
P
hsa-miR-155 hsa-miR-29a hsa-miR-29b-1-5p hsa-miR-720
4.89966 3.02282 2.7880540 1.3649024
.001 .005 .01 .03
hsa-miR-486-5p hsa-miR-500-3p hsa-miR-629
−0.9231489 −0.4229108 −3.2318022
.02 .03 .05
2.5. miRNA quantitative real-time polymerase chain reaction assays Expression levels of selected mature miRNAs were assessed by quantitative real-time polymerase chain reaction (qRT-PCR). Complementary DNA was prepared from 100-ng total RNA using Taqman MicroRNA Reverse Transcription Kit (Applied Biosystems). Predesigned Taqman MicroRNA assays for miR-155 (ID 002623), miR-29a (ID 002447), miR-500-3p (ID 002428), miR486-5p (ID 001278), and miR-629 (ID 002436) were purchased from Applied Biosystems. qRT-PCR was performed using an Applied Biosystems 7900 fast real-time polymerase chain reaction (PCR) system. All reactions were run in triplicate and normalized to the expression of miR-16 (ID 000391). Relative expression levels were determined with the ΔΔCT method and reported as 2−ΔΔCT .
2.6. Locked nucleic acid probe and ISH miRCURY LNA miRNA Detection (FFPE) Optimization Kit 5 containing a known cell-specific miRNA miR-126 (hsa-miR-126 GCATTATTACTACGGTACGA, predicted melting temperature Tm 84°C), positive control (U6 small nuclear RNA, hsa/mmu/rno CACGAATTTGCGTGTCATCCTT, predicted Tm 84°C), and negative control (22-mer scrambled probe with a random sequence GTGTAACACGTCTATACGCCCA, predicted Tm 87°C) with no known complementary sequence target among human transcripts were purchased from Exiqon (Vedbaek, Denmark). We used double digoxigenin-labeled miRCURY locked nucleic acid (LNA)–modified probes for miR-155 (hsa-miR-155 ACCCCTATCACGATTAGCATTAA, predicted Tm 83°C). All LNA oligos were digoxigenin labeled at the 5′ and 3′ ends. ISH analysis was performed according to manufacturer’s instructions along with optimization for proteinase K treatment in terms of concentration and exposure time, probe concentration, and hybridization conditions in terms of hybridization temperature (see Supplementary Material and Supplementary Figs. S1 and S2). The results of ISH were interpreted and scored by 2 independent observers using a 0 to 3+ scoring system based on intensity as follows: 0, negative; 1, weak; 2, moderate; and 3, strong and diffuse. Median values and average score for each case were calculated.
2.7. miRNA target gene expression From hundreds of predicted targets of miR-155 by the currently available major prediction programs, including TargetScan (http://www.targetscan.org), DIANA microT v3 (http://diana.cslab.ece.ntua.gr/microT/), Miranda (http:// www.microrna.org), and PicTar (http://pictar.bio.nyu.edu), we selected genes individuated by means of more than 2 database and by applying knowledge from literature (ZNF652, BACH1, E2F2, KRAS, TRIM32, TP53INP1, PIK3CA, IGF2, JARID2, RBAK, FBXO30, and GDF6). We used qRT-PCR to determine RNA levels of genes listed above using customdesigned Taqman array cards (assay on demand; Applied Biosystems). All qRT-PCR reactions were run in triplicate on an Applied Biosystems 7900 fast real-time PCR system. The qRTPCR data were quantified using the SDS 2.3 software (Applied Biosystems) and normalized using the 18S as endogenous control. To calculate the relative changes in gene expression among different sample, we used the 2−ΔΔCT method.
2.8. Statistical analysis The modified t test, implemented in the R package limma (version 2.18.2) Vienna, Austria [23] was used to identify differentially expressed miRNAs between ALK+ ALCL versus ALK− ALCL cases. The Benjamini-Hochberg method was applied to control false discovery rate [24]. Hierarchical clustering analysis with euclidean distance metric and the ward linkage rule was used to discover the sample clusters. The robustness of resulted clusters was examined by a bootstrapping method with 300 iterations. Principle component analysis (PCA) was also performed to examine the structure of multiple miRNA expression levels. All statistical analyses were performed using statistical R software (version 2.9.0) (R Foundation for Statistical Computing, Vienna, Austria) [24].
3. Results 3.1. Patient characteristics The study group included 17 patients with ALK+ ALCL and 18 patients with ALK− ALCL (Supplementary Table 1).
1998
M. Mehrotra et al.
Fig. 1 A, Hierarchical clustering analysis of ALK+ ALCL (A1-A6) and ALK− ALCL (B2-B6) samples based on their miRNA expression patterns. The figure shows hierarchical clustering of 386 miRNAs, processed by feature extraction background correction and quantile normalization. Colors in the heat map represent normalized and log2-transformed miRNA expression levels, with black representing 0, pure red representing +5 and higher expression, and pure green representing −5 and lower expression. Blue: ALK− ALCL; orange: ALK+ ALCL. B, Enlarged view of region showing differentially expressed miR-155. C, PCA showing separation of differentially expressed miRNA patterns in ALK+ and ALK− ALCL samples. Blue: ALK− ALCL; orange: ALK+ ALCL. D, Box plots of significantly differentially expressed miRNAs (miR-155, miR-29a, miR-486-5p, miR-500, and miR-629) in ALK+ ALCL and ALK− ALCL tumor samples validated by qRT-PCR using the miR-16 as normalizer.
The median age of the patients with ALK+ ALCL was 36 years (range, 15-52 years) with an approximately equal male/ female ratio. Most ALK+ ALCL cases carried t(2;5)(p23; q35)/NPM1-ALK translocation (n = 7) and t(2;17)/NPM1-
ALK translocation (n = 2) as shown by immunohistochemistry (nuclear and cytoplasmic pattern of ALK expression for NPM1-ALK), reverse-transcription PCR, or long-range PCR. The median age of the patients with ALK− ALCL was 45
miR-155 expression and its targets in ALCL Table 2 ISH scoring of ALK+ ALCL (A1-A17) and ALK− ALCL (B1-B18) samples by 2 independent observers Sample
ISH scoring observer 1
ISH scoring observer 2
A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 A13 A14 A15 A16 A17 B1 B2 B3 B4 B5 B6 B8 B9 B10 B11 B16 B17 B18
0 2 0 N/A 1 2 2 1 N/A 1 1 1 N/A N/A 1 N/A N/A 1 3 3 0 3 2 3 1 N/A N/A 3 3 2
1 2 1 N/A 1 1 2 2 N/A 2 2 1 N/A N/A 1 N/A N/A 2 3 3 1 3 2 3 2 N/A N/A 3 3 2
NOTE. Staining intensity: 0, negative; 1, weak; 2, moderate; and 3, strong and diffuse. Abbreviation: N/A, not available.
years (range, 33-80 years) with a male/female ratio of 3.5:1. No cases showed evidence of ALK expression or ALK gene rearrangements. ALK+ ALCL and ALK− ALCL samples were categorized in 2 subsets. There was a discovery set of 6 cases of ALK+ ALCL and 5 cases of ALK− ALCL in which microarray profiling was performed to identify differentially expressed miRNAs. A second subset of 11 cases of ALK+ ALCL and 13 cases of ALK− ALCL was used to validate differentially expressed miRNAs identified in the discovery subset and for quantitative PCR (qPCR) and ISH studies.
3.2. miRNA expression patterns in ALCL tumors We scanned global miRNA expression profiles in 6 cases of ALK+ ALCL and 5 cases of ALK− ALCL. Hierarchical clustering analysis for downstream analysis of the preprocessed data revealed that 386 of 866 miRNAs were expressed above background (P b .05, Fisher exact test). We identified differentially expressed miRNAs between
1999 ALK+ ALCL and ALK− ALCL based on the preprocessed data (Fig. 1A and B). PCA on differentially expressed miRNAs explored patterns determined by simultaneous expression of multiple miRNAs, which confirmed the results from clustering analysis that these 2 lymphomas can be distinguished by miRNA expression (Fig. 1C).
3.3. Identification of differentially expressed miRNAs Differentially expressed miRNAs were identified using normalized miRNA expression profiling data (Fig. 1A and B). Unsupervised hierarchical clustering showed no distinct clustering of ALK+ ALCL and ALK− ALCL. Comparing ALK+ ALCL with ALK− ALCL, miR-486-5p, miR-500-3p, and miR-629 were expressed at significantly higher levels in ALK+ ALCL. By contrast, miR-155, miR-29a, miR-29b-15p, and miR-720 were at significantly higher levels in ALK− ALCL (Fisher exact test, P b .05) with a false discovery rate of less than 0.1 (Table 1). Multivariate analysis showed an association between ALK translocations and the miRNA expression patterns in ALCL.
3.4. Confirmation of miRNA array results by qRT-PCR in ALCL tumors and cell lines qRT-PCR analysis was performed for 7 differentially expressed miRNAs using the discovery subset of ALK+ ALCL and ALK− ALCL samples. The miRNAs assessed included hsamiR-155, hsa-miR-29a, hsa-miR-29b-1-5p, hsa-miR-486-5p, hsa-miR-500-3p, hsa-miR-629, and hsa-miR-720. We also assessed ALCL cell lines and normal reactive lymph node (control) specimens. Two miRNAs, miR-155 (R = 0.9, P b .0003), miR486-5p (R = 0.85, P b .00003) showed a strong and significant correlation between the microarray and qRT-PCR– based results as assessed by Pearson correlation. The remaining 3 miRNAs, miR-29a, miR-629, and miR-500-3p, showed only a moderate (R = 0.74, P b .005) although significant correlation, and 2 miRNAs, miR-729 (R = −0.64, P b .017) and miR29b-1 (R = −0.3, P b .113) did not show a significant correlation. Fig. 1D shows box plots for 5 selected miRNAs. Combining the qPCR results from the validation subset to the discovery subset of 17 ALK+ and 13 ALK− ALCL tumor samples showed that, of the 7 miRNAs, only miR-155 showed significantly different expression between ALK− ALCL (0.888 ± 0.228) and ALK+ ALCL (0.0565 ± 0.009), followed by miR-29a. The combination of miR-155 and miR-29a allowed discrimination between ALK+ ALCL and ALK− ALCL tumor samples (Supplementary Table 2). The remaining 5 miRNAs did not show a significant difference in expression by qPCR. In ALCL cell lines, the results were similar to tumor samples. miR-155 expression was 2.508 (±0.27, P b .004) in the Mac2A cell line as compared with ALK+ ALCL cell lines that showed negligible expression of miR-155, except in SUDHL-1 (0.0866 ± 0.03) (Fig. 2A). miR-29a expression was also higher in Mac2A (2.77 ± 0.35, P b .005) and FE-PD
2000
M. Mehrotra et al.
Fig. 2 Validation of miRNA array results by qRT-PCR in ALK+ ALCL (Jurkat, Karpas, SUDHL-1, SUPM2) and ALK− ALCL (Mac2A and FE-PD) cell lines using the miR-16 as normalizer. Values are expressed as relative quantitation (RQ 2−ΔΔCT). A, miR-155 expression higher in Mac2A cell line as compared with ALK+ ALCL cell lines that showed negligible expression of miR-155 at P b .004. B, miR-29a expression higher in Mac2A as compared with ALK+ ALCL cell lines at Pb .005. C and D, miR-629 and miR-500-3p expression higher in ALK+ ALCL cell lines, respectively. E, miR-486-5p negligible expression in Karpas 299, SUDHL-1, and FE-PD cell lines.
(1.632 ± 0.186, P b .004) compared with Karpas 299, SUPM2, and SUDHL-1 (Fig. 2B). In contrast, miR-500-3p and miR-629 showed higher expression levels in ALK+ ALCL cell lines than in ALK− ALCL cell lines (Fig. 2C and D). miR486-5p was not detected in the Karpas 299, SUDHL1, and FE-PD cell lines (Fig. 2E; Supplementary Table 3).
3.5. ISH in control tissue and cell lines Following controls on FFPE tissue of normal lymph node were run with each experiment. A scrambled RNA control, which helps in excluding nonspecific staining (Fig. 3A), was used. For positive controls, 2 small nucleolar RNA probes
miR-155 expression and its targets in ALCL Table 3
Potential targets of miR-155 in ALCL
Gene symbol Gene name ZNF652 FBXO30 JARID2 IGF2 BACH1 KRAS TP53INP1 E2F2 PIK3CA TRIM32 GDF6 RBAK
Zinc finger protein 652 F-box protein 30 Jumonji, AT-rich interactive domain Insulinlike growth factor type 2 Basic region leucine zipper transcriptional regulator Kirsten rat sarcoma viral oncogene homolog Tumor protein p53 inducible nuclear protein 1 E2F transcription factor 2 Phosphatidylinositol-4,5-bisphosphate 3-kinase Tripartite motif containing 32 Growth differentiation factor 6 RB-associated KRAB zinc finger
Note. Genes that are computationally predicted to be potential targets of miR-155 and whose expression is known to be altered with the change in expression of miR-155.
that show uniform expression in nuclei (Fig. 3B) and a miR126 probe that shows uniform cytoplasmic expression (Fig. 3C) were used to ensure tissue integrity as well as to exclude false-positive and-negative results in each experiment. Reactive lymph nodes showed no expression of miR155 (Fig. 3E and F).
2001 We performed ISH analysis using fixed, paraffinembedded sections of cell pellets of ALK+ ALCL (Karpas 299 and SUDHL-1) and ALK− ALCL (Mac2A) cell lines. Strong and uniform cytoplasmic expression of miR-155 was detected in Mac-2a cells (Fig. 4C and D) as compared with negligible expression in Karpas 299 and SUDHL-1 cells (Fig. 4A and B).
3.6. ISH in ALCL tumors ISH analysis for miR-155 was performed on FFPE sections ALK+ and ALK− ALCLs (Fig. 4; Table 2). miR155 expression in ALK+ ALCL was found to be negative or weak and focal (Fig. 4F). In comparison, miR-155 expression in ALK− ALCL tumors was strong and uniform (Fig. 4H). The intensity varied from case to case with intense expression in tumor areas in contrast with normal lymphoid cells. The tissue distribution of miR-155 expression correlated with morphological involvement by lymphoma. Statistical evaluation of miR-155 ISH scoring showed significantly higher expression in ALK− ALCL as compared with ALK+ ALCL (median score, 3 versus 1; mean score, 2.3 versus 1.3; P = .01) (Fig. 4E-H). A low correlation coefficient of +0.41 in ALK+ ALCL and +0.29 for ALK− ALCL tumors was observed between ISH scoring and relative quantitation values established by qPCR for miR-
Fig. 3 ISH for controls, conditions were identical as with miRCURY LNA miR-155 probe. A, ALCL paraffin sections of hybridization with a scrambled probe substituted for miR-155 used as a negative control with no detectable hybridization signal. B, 5-nM small nucleolar control LNA U6 small nuclear RNA probe with uniform expression in nuclei. C, 100-nM miRCURY LNA positive control miR-126 known to express universally in the cytoplasm. D, Hematoxylin and eosin (H&E) staining of reactive lymph node. E and F, No detectable expression of hybridization signal for 150-nM miR-155 probe was detected in the benign lymphocytes from reactive lymph nodes.
2002
M. Mehrotra et al.
miR-155 expression and its targets in ALCL
2003
Fig. 5 A, Predicted base complementarity of miR-155 to TRIM32, JARID2, ZNF652, KRAS, TP53INP1, RBAK, E2F2, FBXO30, IGF2, BACH1, and PIK3CA. The base complementarity of miR-155 to 3′-UTR binding sites of TRIM32, JARID2, ZNF652, KRAS, TP53INP1, RBAK, E2F2, FBXO30, IGF2, and BACH1, as predicted by Sanger miRNA database (http://microrna.sanger.ac.uk/sequences/). B, Relative expression of target genes quantified using 16+ Taqman custom array cards of predesigned available assay in ALK+ ALCL and ALK− ALCL tumor tissue samples.
155 (Fig. 4I and J). The qRT-PCR results reflect miR-155 expression in the entire tissue that includes additional nontumoral cells using a sensitive PCR technique, whereas
the ISH score reflects expression in tumor cells using a manual interpretation of the staining intensity. These differences weaken the strength of correlation coefficient
Fig. 4 ISH using 150-nM double digoxigenin-labeled miRCURY LNA miR-155 probe in FFPE cell pellets of ALK+ ALCL and ALK− ALCL cell lines and tissue samples. A, Karpas299. B, SUDHL-1 showing undetectable expression of miR-155 on hybridization. C and D, ALK− ALCL cell line Mac2A showing positive hybridization signal with high cytoplasmic expression of miR-155. E and F, Paraffin sections of ALK+ ALCL tissue sample. E, H&E staining. F, Weak and focal hybridization signal for miR-155 with absence of background. G and H, Paraffin section of ALK− ALCL tissue sample. G, H&E staining. H, strong and uniform hybridization signal in cytoplasm of tumor cells for miR-155 expression. I and J, Correlation between the mean relative quantitation expression (RQ value) by qPCR and mean ISH scores generated by ISH in ALK+ ALCL (I) and ALK− ALCL (J) tumor samples.
2004 between 2 analyses; however, the overall direction of change, that is, increased or decreased relative expression, was still maintained.
3.7. miRNA target genes expression To validate the findings of miRNA expression associated with genetic categories, we tested the mRNA levels of selected target genes in 3 FFPE human tissue samples of each ALK+ ALCL and ALK− ALCL sample. From hundreds of predicted targets of miR-155 by the currently available major prediction programs, including TargetScan S, Miranda, DIANA microT v 3.0, and PicTar, we selected genes individuated by means of more than 2 databases and with appealing knowledge from literature (Fig. 5A; Table 3). In particular, we choose to evaluate association of these target genes for ALK+ ALCL and ALK− ALCL correlated to miR-155 level by Pearson correlation coefficient. ZNF652 (r = −0.57, P = .05) and TP53INP1 (r = −0.31, P = .03) showed inverse correlation with up-regulation of miR-155 level in ALCL, whereas E2F2 (r = 0.81, P = .05), TRIM32 (r = 0.92, P = .01), BACH1 (r = 0.88, P = .02), and RBAK (r = 0.81, P = .05) showed positive correlation with up-regulation of miR-155 at statistically significant level; whereas FBXO30 (r = 0.56, P = .25), JARID2 (r = 0.65, P = .17), IGF2 (r = 0.75, P = .08), KRAS (r = 0.53, P = .53), and PIK3CA (r = 0.76, P = .08) also showed positive correlation with the up-regulation of miR155 level in ALCL, these are not found to be statistically significant (Fig. 5B; Supplementary Fig. S3). GDF6 did not show any expression. We confirmed the direct or indirect interaction of these potential target genes with miR-155 by Ingenuity pathway analysis software Qiagen, Redwood city, CA, USA with a conclusion that deregulation of miR-155 levels leads to change in expression of studied target genes, which are indirectly or directly involved in regulation of PIK3/AKT pathway, JAK-STAT signaling, and DNA mismatch repair pathway (Supplementary Fig. S4).
4. Discussion We performed comprehensive screening for deregulated miRNA expression in ALK+ ALCL and ALK− ALCL tumor samples and ALCL cell lines using FFPE specimens. We observed significant differences in the miRNA expression profiles between ALK+ ALCL and ALK− ALCL. In this study, we assessed the distribution of miR-155 in tissues involved by ALCL by ISH method, a marker validated as being overexpressed in ALK− ALCL. We also identified ZNF652, BACH1, RBAK, TRIM32, E2F2, and TP53INP1 as potential targets of miR-155, significantly correlated with miR-155 levels in ALCL tissue samples. Hierarchical clustering analysis has shown that ALK+ ALCL has a miRNA expression profile that is distinct from ALK− ALCL. In the initial set of cases, we identified a set of 7 miRNAs that are expressed differently between ALK+
M. Mehrotra et al. ALCL and ALK− ALCL. Four of these miRNAs, miR-155, miR-29a, miR-29b-1-5p, and miR-720, are expressed at higher levels in ALK− ALCL. In contrast, 3 miRNAs including miR-29b-1, miR-500-3p, miR-629, and miR-4865p were expressed at higher levels in ALK+ ALCL. PCA confirmed the results of cluster analysis. We selected miR-155 for further downstream target analysis due to its role in a variety of hematopoietic malignancies and solid tumors [15]. miR-155 is also involved in hematopoiesis, inflammation, and immunity [25]. Previous miRNA studies attempting to distinguish ALK+ and ALK− ALCL have shown miR-155 as a differentially expressed miRNA [18-20,26]. Upregulation of miR-155 in malignant cells of various types indicates that miR-155 likely has a major role as an oncogene; however, a possible tumor suppression function also has been suggested [27]. In the current study, BACH1, ZNF652, E2F2, TRIM 32, RBAK, and TP53INP1 are identified as potential miR-155 targets. These findings are consistent with documentation of BACH1 and ZNF652 as miR-155 targets in earlier studies; whereas functional studies involving each of the targets are beyond the scope of the article, preliminary literature evidence suggests roles of these targets in lymphomagenesis. BACH1, ZNF652, and E2F2 are known transcriptional regulators. ZNF652 is shown to repress transcription of CBFA2T3, which, in turn, has been shown to be lost in human lymphoma [28-30]. RBAK is shown to be expressed in CD34+ hematopoietic stem cells [31]. The expression of TP53INP1, a known mediator of P53 antioxidant function and a possible mediator of oxidative stress–related lymphoma, was significantly inversely correlated to high expression of miR-155 in ALK− ALCL [32,33]. A role for TRIM32 in oncogenesis is proposed through involvement in a regulatory feedback loop with P73 [34]. On a larger scale, involvement of these genes in regulation of PIK3/AKT pathway, JAK-STAT signaling, and DNA mismatch repair pathway (Supplementary Fig. S4) needs further evaluation. We report direct visualization of miRNA expression in FFPE sections of ALCL tumors by ISH method for the first time. In earlier studies, a limitation of identifying and validating potential biomarkers has been difficult in large part because of protein and RNA degradation in archival tissue [9,35,36]. The small size of miRNAs offers a unique advantage because miRNA molecules are oncogenic biomarkers that are less susceptible to enzymatic and mechanical degradation. [9-12,15,16,36-38]. The ISH results confirmed our miRNA microarray results and, more importantly, allow for the assessment of spatial miRNA expression patterns. The miRNA signature of “contaminating” nontumor cells can affect the analysis of miRNA profiling data derived from using whole tissue extracts. Bioinformatics approaches of removing these data raise questions about the efficacy of this approach for discovery of a potential biomarker. Our study overcomes that limitation by providing a useful tool for the validation of biomarkers [38,39]. We showed by ISH analysis higher miR-155 expression levels in ALK− ALCL compared with ALK+ ALCL.
miR-155 expression and its targets in ALCL Furthermore, the tissue distribution of miR-155 signal correlated well with morphological involvement by lymphoma. No specific ISH signal was observed in benign lymphocytes of reactive lymph nodes, supporting the specificity of this method. In addition, we need to further validate the potential targets in large study cohort of ALCL tumor tissue samples and perform functional studies to determine their role in lymphomagenesis.
Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.humpath.2014.06.012.
References [1] Delsol G, Falini B, Muller-Hermelink HK, et al. Anaplastic large cell lymphoma (ALCL), ALK-positive. In: Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA, Stein H, Thiele J, Vardiman JW, editors. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. IARC Lyon; 2008. p. 312-6. [2] Lai R, Ingham RJ. The pathobiology of the oncogenic tyrosine kinase NPM-ALK: a brief update. Ther Adv in Hematol 2013;4:119-31. [3] Kinney MC, Higgins RA, Medina EA. Anaplastic large cell lymphoma: twenty-five years of discovery. Arch Pathol Lab Med 2011;135:19-43. [4] Medeiros LJ, Elenitoba-Johnson KS. Anaplastic large cell lymphoma. Am J Clin Pathol 2007;127:707-22. [5] Mason DY, Harris NL, Delsol G, et al. Anaplastic large cell lymphoma, ALK-negative. In: Swerdlow SH, Campo E, Harris NL, et al, editors. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. IARC Lyon; 2008. p. 317-9. [6] Kutok JL, Aster JC. Molecular biology of anaplastic lymphoma kinasepositive anaplastic large-cell lymphoma. J Clin Oncol 2002;20:3691-702. [7] Wellmann A, Thieblemont C, Pittaluga S, et al. Detection of differentially expressed genes in lymphomas using cDNA arrays: identification of clusterin as a new diagnostic marker for anaplastic large-cell lymphomas. Blood 2000;96:398-404. [8] Lamant L, de Reynies A, Duplantier MM, et al. Gene-expression profiling of systemic anaplastic large-cell lymphoma reveals differences based on ALK status and two distinct morphologic ALK+ subtypes. Blood 2007;109:2156-64. [9] Calin GA, Croce CM. MicroRNA-cancer connection: the beginning of a new tale. Cancer Res 2006;66:7390-4. [10] Mirnezami AH, Pickard K, Zhang L, et al. MicroRNAs: key players in carcinogenesis and novel therapeutic targets. Eur J Surg Oncol 2009;35:339-47. [11] Melo SA, Esteller M. Dysregulation of microRNAs in cancer: playing with fire. FEBS Lett 2011;585:2087-99. [12] Zhao H, Wang D, Du W, et al. MicroRNA and leukemia: tiny molecule, great function. Crit Rev Oncol Hematol 2010;74:149-55. [13] Zhao JJ, Lin J, Lwin T, et al. MicroRNA expression profile and identification of miR-29 as a prognostic marker and pathogenetic factor by targeting CDK6 in mantle cell lymphoma. Blood 2010;115:2630-9. [14] Calin GA, Ferracin M, Cimmino A, et al. A MicroRNA signature associated with prognosis and progression in chronic lymphocytic leukemia. N Engl J Med 2005;353:1793-801. [15] Donnem T, Eklo K, Berg T, et al. Prognostic impact of MiR-155 in non-small cell lung cancer evaluated by in situ hybridization. J Transl Med 2011;9:6. [16] Mueller DW, Rehli M, Bosserhoff AK. miRNA expression profiling in melanocytes and melanoma cell lines reveals miRNAs associated with formation and progression of malignant melanoma. J Invest Dermatol 2009;129:1740-51.
2005 [17] Drexler HG, MacLeod RA. Malignant hematopoietic cell lines: in vitro models for the study of anaplastic large-cell lymphoma. Leukemia 2004;18:1569-71. [18] Matsuyama H, Suzuki HI, Nishimori H, et al. miR-135b mediates NPM-ALK-driven oncogenicity and renders IL-17-producing immunophenotype to anaplastic large cell lymphoma. Blood 2011;118:6881-92. [19] Merkel O, Hamacher F, Laimer D, et al. Identification of differential and functionally active miRNAs in both anaplastic lymphoma kinase (ALK)+ and ALK- anaplastic large-cell lymphoma. PNAS 2010;107:16228-33. [20] Liu C, Iqbal J, Teruya-Feldstein J, et al. MicroRNA expression profiling identifies molecular signatures associated with anaplastic large cell lymphoma. Blood 2013;122:2083-92. [21] Nelson PT, Wilfred BR. In situ hybridization is a necessary experimental complement to microRNA (miRNA) expression profiling in the human brain. Neurosci Lett 2009;466:69-72. [22] Song R, Ro S, Yan W. In situ hybridization detection of microRNAs. Methods Mol Biol 2010;629:287-94. [23] Smyth GK. Linear models and empirical Bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol 2004;3:1544-6115. [24] Benjamini Y, Hochberg Y. Controlling the false discovery rate - a practical and powerful approach to multiple testing. J R Stat Soc Ser B Methodol 1995;57:289-300. [25] Faraoni I, Antonetti FR, Cardone J, et al. miR-155 gene: a typical multifunctional microRNA. Biochim Biophys Acta 2009;1792: 497-505. [26] Sandhu SK, Croce CM, Garzon R. Micro-RNA expression and function in lymphomas. Adv Hematol 2011;2011:347137. [27] Dorsett Y, McBride KM, Jankovic M, et al. MicroRNA-155 suppresses activation-induced cytidine deaminase-mediated Myc-Igh translocation. Immunity 2008;28:630-8. [28] Lamprecht B, Walter K, Kreher S, et al. Derepression of an endogenous long terminal repeat activates the CSF1R proto-oncogene in human lymphoma. Nat Med 2010;16:571-9. [29] Kumar R, Manning J, Spendlove HE, et al. ZNF652, a novel zinc finger protein, interacts with the putative breast tumor suppressor CBFA2T3 to repress transcription. Mol Cancer Res 2006;4:655-65. [30] Holm R, Knopp S, Kumar R, et al. Expression of ZNF652, a novel zinc finger protein, in vulvar carcinomas and its relation to prognosis. J Clin Pathol 2008;61:59-63. [31] Gomes I, Sharma TT, Edassery S, et al. Novel transcription factors in human CD34 antigen-positive hematopoietic cells. Blood 2002;100:107-19. [32] Cano CE, Gommeaux J, Pietri S, et al. Tumor protein 53-induced nuclear protein 1 is a major mediator of p53 antioxidant function. Cancer Res 2009;69:219-26. [33] Gironella M, Seux M, Xie MJ, et al. Tumor protein 53-induced nuclear protein 1 expression is repressed by miR-155, and its restoration inhibits pancreatic tumor development. PNAS 2007;104:16170-5. [34] Gonzalez-Cano L, Hillje AL, Fuertes-Alvarez S, et al. Regulatory feedback loop between TP73 and TRIM32. Cell Death Dis 2013;4:e704. [35] Ludwig JA, Weinstein JN. Biomarkers in cancer staging, prognosis and treatment selection. Nat Rev Cancer 2005;5:845-56. [36] Nielsen BS, Jorgensen S, Fog JU, et al. High levels of microRNA-21 in the stroma of colorectal cancers predict short disease-free survival in stage II colon cancer patients. Clin Exp Metastasis 2011;28:27-38. [37] Philippidou D, Schmitt M, Moser D, et al. Signatures of microRNAs and selected microRNA target genes in human melanoma. Cancer Res 2010;70:4163-73. [38] Kopp KL, Ralfkiaer U, Nielsen BS, et al. Expression of miR-155 and miR126 in situ in cutaneous T-cell lymphoma. APMIS 2013;121:1020-4. [39] Rossing M, Borup R, Henao R, et al. Down-regulation of microRNAs controlling tumourigenic factors in follicular thyroid carcinoma. J Mol Endocrinol 2012;48:11-23.
www.elsevier.com/locate/humpath
Original contributions
Identification of putative pathogenic microRNA and its downstream targets in anaplastic lymphoma kinase–negative anaplastic large cell lymphoma☆ Meenakshi Mehrotra PhD, L. Jeffrey Medeiros MD, Rajyalakshmi Luthra PhD, Rachel L. Sargent MD, Hui Yao PhD, Bedia A. Barkoh MS, Rajesh Singh PhD, Keyur P. Patel MD, PhD ⁎ Department of Hematopathology, The University of Texas MD Anderson Cancer Center, Houston, Texas 77030 Received 7 January 2014; revised 10 June 2014; accepted 19 June 2014
Keywords: Anaplastic large cell lymphoma; microRNA; Expression profiling; RT-PCR; In situ hybridization
Summary Anaplastic large cell lymphomas (ALCL) are tumors of T/null-cell lineage characterized by uniform CD30 expression. The 2008 World Health Organization classification subdivided ALCLs into 2 groups: anaplastic lymphoma kinase (ALK)–positive (established entity) and ALK-negative (proposed new entity) ALCL. The genetic basis for the pathogenesis of newly categorized ALK− ALCL is poorly understood. In this study, we used microRNA microarray analysis to identify differentially expressed microRNAs in ALK+ and ALK− ALCL. ALK− ALCL showed significantly higher expression of miR155 (0.888 ± 0.228) compared with ALK+ ALCL (0.0565 ± 0.009) on microarray and by quantitative real-time polymerase chain reaction in ALK− ALCL compared with ALK+ ALCL (P b .05) with a strong correlation between the 2 platforms (R = 0.9, P b .0003). A novel in situ hybridization method allows direct visualization of expression patterns and relative quantitation of miR-155 (mean score, 2.3 versus 1.3; P = .01) for the first time in tissue sections of ALCL. Among computationally predicted targets of miR-155, we identified ZNF652 (r = −0.57, P = .05), BACH1 (r = 0.88, P = .02), RBAK (r = 0.81, P = .05), TRIM32 (r = 0.92, P = .01), E2F2 (r = 0.81, P = .05), and TP53INP1 (r = −0.31, P = .03) as genes whose expression by quantitative real-time polymerase chain reaction correlated significantly with the level of miR155 in ALCL tumor tissue. © 2014 Elsevier Inc. All rights reserved.
1. Introduction Anaplastic large cell lymphoma (ALCL) is a non-Hodgkin lymphoma of T/null-cell lineage that is characterized by uniform ☆
Competing interests: The authors declare no conflicts of interest. ⁎ Corresponding author at: Department of Hematopathology, Unit 149, 1515 Holcombe Blvd, MD Anderson Cancer Center, Houston, Texas 77030. E-mail address: [email protected] (K. P. Patel). http://dx.doi.org/10.1016/j.humpath.2014.06.012 0046-8177/© 2014 Elsevier Inc. All rights reserved.
CD30 expression. In the 2008 World Health Organization (WHO) classification of lymphomas, 2 distinct entities of ALCL were proposed: anaplastic lymphoma kinase (ALK)– positive ALCL and ALK-negative ALCL [1]. ALK+ ALCL is an established entity that carries a t(2;5)(p23;q35) or variant ALK abnormalities resulting in aberrant expression of ALK with ALCL [2-4] and shows favorable prognosis when treated with chemotherapy. In contrast, ALK− ALCL is currently
1996 considered as a provisional category in the WHO classification [5] that closely resembles ALK+ ALCL morphologically and expresses CD30 with similar intensity; however, these tumors lack ALK translocations or overexpress ALK, and affected patients have a poorer prognosis when treated with chemotherapy [2-4]. Significantly, no definitive pathogenic mechanisms have been identified for ALK− ALCL. Genomic and expression profiling studies have shown different genetic aberrations and messenger RNA (mRNA) expression patterns in ALK+ and ALK− ALCL [6-8]. The functional importance of these findings and their roles in pathogenesis are far better understood in ALK+ ALCL than in ALK− ALCL. Improved understanding of ALK− ALCL pathogenesis may lead to identification of potential therapeutic targets that could guide therapy or predict outcome. MicroRNAs (miRNAs) are a class of approximately small (20-22 nucleotides) noncoding RNA molecules that play crucial roles in regulating gene expression, by either inducing mRNA degradation or inhibiting translation. These noncoding RNAs can control a wide range of biologic functions including differentiation, proliferation, and apoptosis. Deregulation of miRNA expression has been shown to play an important role in oncogenesis [9-11]. Deregulated expression of specific miRNAs has been described in solid cancers, leukemias, and lymphomas [12-16], and there are limited studies that have shown different miRNA profiles in ALK+ and ALK− ALCLs [17-20]. These approaches, to date, however, have been applied to the study of whole tumor samples including neoplastic cells and the microenvironment. It would be helpful to assess the tissue and subcellular distribution pattern of individual miRNAs in ALCLs using in situ hybridization (ISH), which could provide important development and homeostasis information as well as a convenient diagnostic tool [15,21,22]. In this study, we used miRNA arrays to identify miRNA that could underlie pathogenesis of ALK− ALCL by screening in ALK+ ALCL and ALK− ALCL tumor samples. miR-155 was found to be expressed at significantly higher level in ALK− ALCL. We demonstrated tissue and subcellular distribution of miR-155 using ISH in ALK+ ALCL and ALK− ALCL tumor samples. We also identified ZNF652, BACH1, RBAK, BACH1, E2F2, and TP53INP1 as potential targets of miR-155, which were significantly correlated with the level of miR-155 in ALCL tumor tissue samples. The ability to visualize miRNA expression patterns and perform relative quantitation in routinely processed, paraffin-embedded tissue sections provides a convenient diagnostic tool to assess differential expression of miRNAs in hematopoietic neoplasms such as ALCL.
2. Materials and methods 2.1. Patient samples This study was performed with approval from the institutional review board. Formalin-fixed, paraffin-embedded
M. Mehrotra et al. (FFPE) tissue blocks of cases of ALK+ ALCL (n = 17), ALK− ALCL (n = 18), and reactive lymph node controls (n = 4) were retrieved from the archives of the Department of Hematopathology, The University of Texas MD Anderson Cancer Center. The diagnosis of ALCL was based on morphological and immunophenotypic criteria as described in the WHO classification [1]. Original data were received and confirmed by 2 different pathologists.
2.2. Cell lines and reagents The Karpas 299, SUDHL-1, and SUPM2 cell lines (all ALK+ ALCL) were a gift from Suzanne Turner (Division of Molecular Histopathology, Department of Pathology, University of Cambridge, Addenbrooke's Hospital, Cambridge, United Kingdom). The FE-PD and Mac2A cell lines (both ALK− ALCL) were gifts from Annarosa del Mistro (Veneto Oncology Institute, Padua, Italy) and Marshall Kadin (Department of Dermatology and Skin Surgery, Roger Williams Medical Center, Boston University School of Medicine, Providence, RI). Cells were cultured in RPMI-1640 medium (catalog no. 30-2001; American Type Culture Collection, Manassas, VA, USA) containing 1% penicillin and streptomycin and 10% fetal bovine serum (FBS) (catalog no. 30-2020; American Type Culture Collection) (for Karpas 299, SUPM2, and FE-PD), 15% FBS (for Mac2A), and 20% FBS (for SUDHL-1).
2.3. RNA isolation Total RNA was isolated from microdissected FFPE tumor samples using Recover All Total Nucleic Acid Isolation Kit for FFPE Tissues (Applied Biosystems, Foster City, CA). Because this kit allows isolation of both miRNA and mRNA, the same sample can be used for miRNA profiling and miRNA and mRNA validation studies. RNA concentrations were determined using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE).
2.4. miRNA microarray experiments Human miRNA Microarray version 3 (Agilent Technologies, Santa Clara, CA) chips containing oligonucleotide probes for 866 human miRNA and 89 human viral miRNAs in a 15 K format, where each target is represented 10 times on a hybridization array, were used to generate miRNA profiles of 11 study cases, specifically 6 ALK+ ALCL and 5 ALK− ALCL according to manufacturer’s protocol (detailed methods for microarray experiments are explained in Supplementary Material). Commercially available universal human miRNA pool (Agilent Technologies) was used as a technical control for hybridization in the miRNA profiling experiments.
miR-155 expression and its targets in ALCL Table 1
1997
Differentially expressed miRNAs in ALK− ALCL and ALK+ ALCL by microarray analysis
miRNAs with higher expression in ALK− ALCL
miRNAs with lower expression in ALK− ALCL
Gene ID
Fold change
P
Gene ID
Fold change
P
hsa-miR-155 hsa-miR-29a hsa-miR-29b-1-5p hsa-miR-720
4.89966 3.02282 2.7880540 1.3649024
.001 .005 .01 .03
hsa-miR-486-5p hsa-miR-500-3p hsa-miR-629
−0.9231489 −0.4229108 −3.2318022
.02 .03 .05
2.5. miRNA quantitative real-time polymerase chain reaction assays Expression levels of selected mature miRNAs were assessed by quantitative real-time polymerase chain reaction (qRT-PCR). Complementary DNA was prepared from 100-ng total RNA using Taqman MicroRNA Reverse Transcription Kit (Applied Biosystems). Predesigned Taqman MicroRNA assays for miR-155 (ID 002623), miR-29a (ID 002447), miR-500-3p (ID 002428), miR486-5p (ID 001278), and miR-629 (ID 002436) were purchased from Applied Biosystems. qRT-PCR was performed using an Applied Biosystems 7900 fast real-time polymerase chain reaction (PCR) system. All reactions were run in triplicate and normalized to the expression of miR-16 (ID 000391). Relative expression levels were determined with the ΔΔCT method and reported as 2−ΔΔCT .
2.6. Locked nucleic acid probe and ISH miRCURY LNA miRNA Detection (FFPE) Optimization Kit 5 containing a known cell-specific miRNA miR-126 (hsa-miR-126 GCATTATTACTACGGTACGA, predicted melting temperature Tm 84°C), positive control (U6 small nuclear RNA, hsa/mmu/rno CACGAATTTGCGTGTCATCCTT, predicted Tm 84°C), and negative control (22-mer scrambled probe with a random sequence GTGTAACACGTCTATACGCCCA, predicted Tm 87°C) with no known complementary sequence target among human transcripts were purchased from Exiqon (Vedbaek, Denmark). We used double digoxigenin-labeled miRCURY locked nucleic acid (LNA)–modified probes for miR-155 (hsa-miR-155 ACCCCTATCACGATTAGCATTAA, predicted Tm 83°C). All LNA oligos were digoxigenin labeled at the 5′ and 3′ ends. ISH analysis was performed according to manufacturer’s instructions along with optimization for proteinase K treatment in terms of concentration and exposure time, probe concentration, and hybridization conditions in terms of hybridization temperature (see Supplementary Material and Supplementary Figs. S1 and S2). The results of ISH were interpreted and scored by 2 independent observers using a 0 to 3+ scoring system based on intensity as follows: 0, negative; 1, weak; 2, moderate; and 3, strong and diffuse. Median values and average score for each case were calculated.
2.7. miRNA target gene expression From hundreds of predicted targets of miR-155 by the currently available major prediction programs, including TargetScan (http://www.targetscan.org), DIANA microT v3 (http://diana.cslab.ece.ntua.gr/microT/), Miranda (http:// www.microrna.org), and PicTar (http://pictar.bio.nyu.edu), we selected genes individuated by means of more than 2 database and by applying knowledge from literature (ZNF652, BACH1, E2F2, KRAS, TRIM32, TP53INP1, PIK3CA, IGF2, JARID2, RBAK, FBXO30, and GDF6). We used qRT-PCR to determine RNA levels of genes listed above using customdesigned Taqman array cards (assay on demand; Applied Biosystems). All qRT-PCR reactions were run in triplicate on an Applied Biosystems 7900 fast real-time PCR system. The qRTPCR data were quantified using the SDS 2.3 software (Applied Biosystems) and normalized using the 18S as endogenous control. To calculate the relative changes in gene expression among different sample, we used the 2−ΔΔCT method.
2.8. Statistical analysis The modified t test, implemented in the R package limma (version 2.18.2) Vienna, Austria [23] was used to identify differentially expressed miRNAs between ALK+ ALCL versus ALK− ALCL cases. The Benjamini-Hochberg method was applied to control false discovery rate [24]. Hierarchical clustering analysis with euclidean distance metric and the ward linkage rule was used to discover the sample clusters. The robustness of resulted clusters was examined by a bootstrapping method with 300 iterations. Principle component analysis (PCA) was also performed to examine the structure of multiple miRNA expression levels. All statistical analyses were performed using statistical R software (version 2.9.0) (R Foundation for Statistical Computing, Vienna, Austria) [24].
3. Results 3.1. Patient characteristics The study group included 17 patients with ALK+ ALCL and 18 patients with ALK− ALCL (Supplementary Table 1).
1998
M. Mehrotra et al.
Fig. 1 A, Hierarchical clustering analysis of ALK+ ALCL (A1-A6) and ALK− ALCL (B2-B6) samples based on their miRNA expression patterns. The figure shows hierarchical clustering of 386 miRNAs, processed by feature extraction background correction and quantile normalization. Colors in the heat map represent normalized and log2-transformed miRNA expression levels, with black representing 0, pure red representing +5 and higher expression, and pure green representing −5 and lower expression. Blue: ALK− ALCL; orange: ALK+ ALCL. B, Enlarged view of region showing differentially expressed miR-155. C, PCA showing separation of differentially expressed miRNA patterns in ALK+ and ALK− ALCL samples. Blue: ALK− ALCL; orange: ALK+ ALCL. D, Box plots of significantly differentially expressed miRNAs (miR-155, miR-29a, miR-486-5p, miR-500, and miR-629) in ALK+ ALCL and ALK− ALCL tumor samples validated by qRT-PCR using the miR-16 as normalizer.
The median age of the patients with ALK+ ALCL was 36 years (range, 15-52 years) with an approximately equal male/ female ratio. Most ALK+ ALCL cases carried t(2;5)(p23; q35)/NPM1-ALK translocation (n = 7) and t(2;17)/NPM1-
ALK translocation (n = 2) as shown by immunohistochemistry (nuclear and cytoplasmic pattern of ALK expression for NPM1-ALK), reverse-transcription PCR, or long-range PCR. The median age of the patients with ALK− ALCL was 45
miR-155 expression and its targets in ALCL Table 2 ISH scoring of ALK+ ALCL (A1-A17) and ALK− ALCL (B1-B18) samples by 2 independent observers Sample
ISH scoring observer 1
ISH scoring observer 2
A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 A13 A14 A15 A16 A17 B1 B2 B3 B4 B5 B6 B8 B9 B10 B11 B16 B17 B18
0 2 0 N/A 1 2 2 1 N/A 1 1 1 N/A N/A 1 N/A N/A 1 3 3 0 3 2 3 1 N/A N/A 3 3 2
1 2 1 N/A 1 1 2 2 N/A 2 2 1 N/A N/A 1 N/A N/A 2 3 3 1 3 2 3 2 N/A N/A 3 3 2
NOTE. Staining intensity: 0, negative; 1, weak; 2, moderate; and 3, strong and diffuse. Abbreviation: N/A, not available.
years (range, 33-80 years) with a male/female ratio of 3.5:1. No cases showed evidence of ALK expression or ALK gene rearrangements. ALK+ ALCL and ALK− ALCL samples were categorized in 2 subsets. There was a discovery set of 6 cases of ALK+ ALCL and 5 cases of ALK− ALCL in which microarray profiling was performed to identify differentially expressed miRNAs. A second subset of 11 cases of ALK+ ALCL and 13 cases of ALK− ALCL was used to validate differentially expressed miRNAs identified in the discovery subset and for quantitative PCR (qPCR) and ISH studies.
3.2. miRNA expression patterns in ALCL tumors We scanned global miRNA expression profiles in 6 cases of ALK+ ALCL and 5 cases of ALK− ALCL. Hierarchical clustering analysis for downstream analysis of the preprocessed data revealed that 386 of 866 miRNAs were expressed above background (P b .05, Fisher exact test). We identified differentially expressed miRNAs between
1999 ALK+ ALCL and ALK− ALCL based on the preprocessed data (Fig. 1A and B). PCA on differentially expressed miRNAs explored patterns determined by simultaneous expression of multiple miRNAs, which confirmed the results from clustering analysis that these 2 lymphomas can be distinguished by miRNA expression (Fig. 1C).
3.3. Identification of differentially expressed miRNAs Differentially expressed miRNAs were identified using normalized miRNA expression profiling data (Fig. 1A and B). Unsupervised hierarchical clustering showed no distinct clustering of ALK+ ALCL and ALK− ALCL. Comparing ALK+ ALCL with ALK− ALCL, miR-486-5p, miR-500-3p, and miR-629 were expressed at significantly higher levels in ALK+ ALCL. By contrast, miR-155, miR-29a, miR-29b-15p, and miR-720 were at significantly higher levels in ALK− ALCL (Fisher exact test, P b .05) with a false discovery rate of less than 0.1 (Table 1). Multivariate analysis showed an association between ALK translocations and the miRNA expression patterns in ALCL.
3.4. Confirmation of miRNA array results by qRT-PCR in ALCL tumors and cell lines qRT-PCR analysis was performed for 7 differentially expressed miRNAs using the discovery subset of ALK+ ALCL and ALK− ALCL samples. The miRNAs assessed included hsamiR-155, hsa-miR-29a, hsa-miR-29b-1-5p, hsa-miR-486-5p, hsa-miR-500-3p, hsa-miR-629, and hsa-miR-720. We also assessed ALCL cell lines and normal reactive lymph node (control) specimens. Two miRNAs, miR-155 (R = 0.9, P b .0003), miR486-5p (R = 0.85, P b .00003) showed a strong and significant correlation between the microarray and qRT-PCR– based results as assessed by Pearson correlation. The remaining 3 miRNAs, miR-29a, miR-629, and miR-500-3p, showed only a moderate (R = 0.74, P b .005) although significant correlation, and 2 miRNAs, miR-729 (R = −0.64, P b .017) and miR29b-1 (R = −0.3, P b .113) did not show a significant correlation. Fig. 1D shows box plots for 5 selected miRNAs. Combining the qPCR results from the validation subset to the discovery subset of 17 ALK+ and 13 ALK− ALCL tumor samples showed that, of the 7 miRNAs, only miR-155 showed significantly different expression between ALK− ALCL (0.888 ± 0.228) and ALK+ ALCL (0.0565 ± 0.009), followed by miR-29a. The combination of miR-155 and miR-29a allowed discrimination between ALK+ ALCL and ALK− ALCL tumor samples (Supplementary Table 2). The remaining 5 miRNAs did not show a significant difference in expression by qPCR. In ALCL cell lines, the results were similar to tumor samples. miR-155 expression was 2.508 (±0.27, P b .004) in the Mac2A cell line as compared with ALK+ ALCL cell lines that showed negligible expression of miR-155, except in SUDHL-1 (0.0866 ± 0.03) (Fig. 2A). miR-29a expression was also higher in Mac2A (2.77 ± 0.35, P b .005) and FE-PD
2000
M. Mehrotra et al.
Fig. 2 Validation of miRNA array results by qRT-PCR in ALK+ ALCL (Jurkat, Karpas, SUDHL-1, SUPM2) and ALK− ALCL (Mac2A and FE-PD) cell lines using the miR-16 as normalizer. Values are expressed as relative quantitation (RQ 2−ΔΔCT). A, miR-155 expression higher in Mac2A cell line as compared with ALK+ ALCL cell lines that showed negligible expression of miR-155 at P b .004. B, miR-29a expression higher in Mac2A as compared with ALK+ ALCL cell lines at Pb .005. C and D, miR-629 and miR-500-3p expression higher in ALK+ ALCL cell lines, respectively. E, miR-486-5p negligible expression in Karpas 299, SUDHL-1, and FE-PD cell lines.
(1.632 ± 0.186, P b .004) compared with Karpas 299, SUPM2, and SUDHL-1 (Fig. 2B). In contrast, miR-500-3p and miR-629 showed higher expression levels in ALK+ ALCL cell lines than in ALK− ALCL cell lines (Fig. 2C and D). miR486-5p was not detected in the Karpas 299, SUDHL1, and FE-PD cell lines (Fig. 2E; Supplementary Table 3).
3.5. ISH in control tissue and cell lines Following controls on FFPE tissue of normal lymph node were run with each experiment. A scrambled RNA control, which helps in excluding nonspecific staining (Fig. 3A), was used. For positive controls, 2 small nucleolar RNA probes
miR-155 expression and its targets in ALCL Table 3
Potential targets of miR-155 in ALCL
Gene symbol Gene name ZNF652 FBXO30 JARID2 IGF2 BACH1 KRAS TP53INP1 E2F2 PIK3CA TRIM32 GDF6 RBAK
Zinc finger protein 652 F-box protein 30 Jumonji, AT-rich interactive domain Insulinlike growth factor type 2 Basic region leucine zipper transcriptional regulator Kirsten rat sarcoma viral oncogene homolog Tumor protein p53 inducible nuclear protein 1 E2F transcription factor 2 Phosphatidylinositol-4,5-bisphosphate 3-kinase Tripartite motif containing 32 Growth differentiation factor 6 RB-associated KRAB zinc finger
Note. Genes that are computationally predicted to be potential targets of miR-155 and whose expression is known to be altered with the change in expression of miR-155.
that show uniform expression in nuclei (Fig. 3B) and a miR126 probe that shows uniform cytoplasmic expression (Fig. 3C) were used to ensure tissue integrity as well as to exclude false-positive and-negative results in each experiment. Reactive lymph nodes showed no expression of miR155 (Fig. 3E and F).
2001 We performed ISH analysis using fixed, paraffinembedded sections of cell pellets of ALK+ ALCL (Karpas 299 and SUDHL-1) and ALK− ALCL (Mac2A) cell lines. Strong and uniform cytoplasmic expression of miR-155 was detected in Mac-2a cells (Fig. 4C and D) as compared with negligible expression in Karpas 299 and SUDHL-1 cells (Fig. 4A and B).
3.6. ISH in ALCL tumors ISH analysis for miR-155 was performed on FFPE sections ALK+ and ALK− ALCLs (Fig. 4; Table 2). miR155 expression in ALK+ ALCL was found to be negative or weak and focal (Fig. 4F). In comparison, miR-155 expression in ALK− ALCL tumors was strong and uniform (Fig. 4H). The intensity varied from case to case with intense expression in tumor areas in contrast with normal lymphoid cells. The tissue distribution of miR-155 expression correlated with morphological involvement by lymphoma. Statistical evaluation of miR-155 ISH scoring showed significantly higher expression in ALK− ALCL as compared with ALK+ ALCL (median score, 3 versus 1; mean score, 2.3 versus 1.3; P = .01) (Fig. 4E-H). A low correlation coefficient of +0.41 in ALK+ ALCL and +0.29 for ALK− ALCL tumors was observed between ISH scoring and relative quantitation values established by qPCR for miR-
Fig. 3 ISH for controls, conditions were identical as with miRCURY LNA miR-155 probe. A, ALCL paraffin sections of hybridization with a scrambled probe substituted for miR-155 used as a negative control with no detectable hybridization signal. B, 5-nM small nucleolar control LNA U6 small nuclear RNA probe with uniform expression in nuclei. C, 100-nM miRCURY LNA positive control miR-126 known to express universally in the cytoplasm. D, Hematoxylin and eosin (H&E) staining of reactive lymph node. E and F, No detectable expression of hybridization signal for 150-nM miR-155 probe was detected in the benign lymphocytes from reactive lymph nodes.
2002
M. Mehrotra et al.
miR-155 expression and its targets in ALCL
2003
Fig. 5 A, Predicted base complementarity of miR-155 to TRIM32, JARID2, ZNF652, KRAS, TP53INP1, RBAK, E2F2, FBXO30, IGF2, BACH1, and PIK3CA. The base complementarity of miR-155 to 3′-UTR binding sites of TRIM32, JARID2, ZNF652, KRAS, TP53INP1, RBAK, E2F2, FBXO30, IGF2, and BACH1, as predicted by Sanger miRNA database (http://microrna.sanger.ac.uk/sequences/). B, Relative expression of target genes quantified using 16+ Taqman custom array cards of predesigned available assay in ALK+ ALCL and ALK− ALCL tumor tissue samples.
155 (Fig. 4I and J). The qRT-PCR results reflect miR-155 expression in the entire tissue that includes additional nontumoral cells using a sensitive PCR technique, whereas
the ISH score reflects expression in tumor cells using a manual interpretation of the staining intensity. These differences weaken the strength of correlation coefficient
Fig. 4 ISH using 150-nM double digoxigenin-labeled miRCURY LNA miR-155 probe in FFPE cell pellets of ALK+ ALCL and ALK− ALCL cell lines and tissue samples. A, Karpas299. B, SUDHL-1 showing undetectable expression of miR-155 on hybridization. C and D, ALK− ALCL cell line Mac2A showing positive hybridization signal with high cytoplasmic expression of miR-155. E and F, Paraffin sections of ALK+ ALCL tissue sample. E, H&E staining. F, Weak and focal hybridization signal for miR-155 with absence of background. G and H, Paraffin section of ALK− ALCL tissue sample. G, H&E staining. H, strong and uniform hybridization signal in cytoplasm of tumor cells for miR-155 expression. I and J, Correlation between the mean relative quantitation expression (RQ value) by qPCR and mean ISH scores generated by ISH in ALK+ ALCL (I) and ALK− ALCL (J) tumor samples.
2004 between 2 analyses; however, the overall direction of change, that is, increased or decreased relative expression, was still maintained.
3.7. miRNA target genes expression To validate the findings of miRNA expression associated with genetic categories, we tested the mRNA levels of selected target genes in 3 FFPE human tissue samples of each ALK+ ALCL and ALK− ALCL sample. From hundreds of predicted targets of miR-155 by the currently available major prediction programs, including TargetScan S, Miranda, DIANA microT v 3.0, and PicTar, we selected genes individuated by means of more than 2 databases and with appealing knowledge from literature (Fig. 5A; Table 3). In particular, we choose to evaluate association of these target genes for ALK+ ALCL and ALK− ALCL correlated to miR-155 level by Pearson correlation coefficient. ZNF652 (r = −0.57, P = .05) and TP53INP1 (r = −0.31, P = .03) showed inverse correlation with up-regulation of miR-155 level in ALCL, whereas E2F2 (r = 0.81, P = .05), TRIM32 (r = 0.92, P = .01), BACH1 (r = 0.88, P = .02), and RBAK (r = 0.81, P = .05) showed positive correlation with up-regulation of miR-155 at statistically significant level; whereas FBXO30 (r = 0.56, P = .25), JARID2 (r = 0.65, P = .17), IGF2 (r = 0.75, P = .08), KRAS (r = 0.53, P = .53), and PIK3CA (r = 0.76, P = .08) also showed positive correlation with the up-regulation of miR155 level in ALCL, these are not found to be statistically significant (Fig. 5B; Supplementary Fig. S3). GDF6 did not show any expression. We confirmed the direct or indirect interaction of these potential target genes with miR-155 by Ingenuity pathway analysis software Qiagen, Redwood city, CA, USA with a conclusion that deregulation of miR-155 levels leads to change in expression of studied target genes, which are indirectly or directly involved in regulation of PIK3/AKT pathway, JAK-STAT signaling, and DNA mismatch repair pathway (Supplementary Fig. S4).
4. Discussion We performed comprehensive screening for deregulated miRNA expression in ALK+ ALCL and ALK− ALCL tumor samples and ALCL cell lines using FFPE specimens. We observed significant differences in the miRNA expression profiles between ALK+ ALCL and ALK− ALCL. In this study, we assessed the distribution of miR-155 in tissues involved by ALCL by ISH method, a marker validated as being overexpressed in ALK− ALCL. We also identified ZNF652, BACH1, RBAK, TRIM32, E2F2, and TP53INP1 as potential targets of miR-155, significantly correlated with miR-155 levels in ALCL tissue samples. Hierarchical clustering analysis has shown that ALK+ ALCL has a miRNA expression profile that is distinct from ALK− ALCL. In the initial set of cases, we identified a set of 7 miRNAs that are expressed differently between ALK+
M. Mehrotra et al. ALCL and ALK− ALCL. Four of these miRNAs, miR-155, miR-29a, miR-29b-1-5p, and miR-720, are expressed at higher levels in ALK− ALCL. In contrast, 3 miRNAs including miR-29b-1, miR-500-3p, miR-629, and miR-4865p were expressed at higher levels in ALK+ ALCL. PCA confirmed the results of cluster analysis. We selected miR-155 for further downstream target analysis due to its role in a variety of hematopoietic malignancies and solid tumors [15]. miR-155 is also involved in hematopoiesis, inflammation, and immunity [25]. Previous miRNA studies attempting to distinguish ALK+ and ALK− ALCL have shown miR-155 as a differentially expressed miRNA [18-20,26]. Upregulation of miR-155 in malignant cells of various types indicates that miR-155 likely has a major role as an oncogene; however, a possible tumor suppression function also has been suggested [27]. In the current study, BACH1, ZNF652, E2F2, TRIM 32, RBAK, and TP53INP1 are identified as potential miR-155 targets. These findings are consistent with documentation of BACH1 and ZNF652 as miR-155 targets in earlier studies; whereas functional studies involving each of the targets are beyond the scope of the article, preliminary literature evidence suggests roles of these targets in lymphomagenesis. BACH1, ZNF652, and E2F2 are known transcriptional regulators. ZNF652 is shown to repress transcription of CBFA2T3, which, in turn, has been shown to be lost in human lymphoma [28-30]. RBAK is shown to be expressed in CD34+ hematopoietic stem cells [31]. The expression of TP53INP1, a known mediator of P53 antioxidant function and a possible mediator of oxidative stress–related lymphoma, was significantly inversely correlated to high expression of miR-155 in ALK− ALCL [32,33]. A role for TRIM32 in oncogenesis is proposed through involvement in a regulatory feedback loop with P73 [34]. On a larger scale, involvement of these genes in regulation of PIK3/AKT pathway, JAK-STAT signaling, and DNA mismatch repair pathway (Supplementary Fig. S4) needs further evaluation. We report direct visualization of miRNA expression in FFPE sections of ALCL tumors by ISH method for the first time. In earlier studies, a limitation of identifying and validating potential biomarkers has been difficult in large part because of protein and RNA degradation in archival tissue [9,35,36]. The small size of miRNAs offers a unique advantage because miRNA molecules are oncogenic biomarkers that are less susceptible to enzymatic and mechanical degradation. [9-12,15,16,36-38]. The ISH results confirmed our miRNA microarray results and, more importantly, allow for the assessment of spatial miRNA expression patterns. The miRNA signature of “contaminating” nontumor cells can affect the analysis of miRNA profiling data derived from using whole tissue extracts. Bioinformatics approaches of removing these data raise questions about the efficacy of this approach for discovery of a potential biomarker. Our study overcomes that limitation by providing a useful tool for the validation of biomarkers [38,39]. We showed by ISH analysis higher miR-155 expression levels in ALK− ALCL compared with ALK+ ALCL.
miR-155 expression and its targets in ALCL Furthermore, the tissue distribution of miR-155 signal correlated well with morphological involvement by lymphoma. No specific ISH signal was observed in benign lymphocytes of reactive lymph nodes, supporting the specificity of this method. In addition, we need to further validate the potential targets in large study cohort of ALCL tumor tissue samples and perform functional studies to determine their role in lymphomagenesis.
Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.humpath.2014.06.012.
References [1] Delsol G, Falini B, Muller-Hermelink HK, et al. Anaplastic large cell lymphoma (ALCL), ALK-positive. In: Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA, Stein H, Thiele J, Vardiman JW, editors. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. IARC Lyon; 2008. p. 312-6. [2] Lai R, Ingham RJ. The pathobiology of the oncogenic tyrosine kinase NPM-ALK: a brief update. Ther Adv in Hematol 2013;4:119-31. [3] Kinney MC, Higgins RA, Medina EA. Anaplastic large cell lymphoma: twenty-five years of discovery. Arch Pathol Lab Med 2011;135:19-43. [4] Medeiros LJ, Elenitoba-Johnson KS. Anaplastic large cell lymphoma. Am J Clin Pathol 2007;127:707-22. [5] Mason DY, Harris NL, Delsol G, et al. Anaplastic large cell lymphoma, ALK-negative. In: Swerdlow SH, Campo E, Harris NL, et al, editors. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. IARC Lyon; 2008. p. 317-9. [6] Kutok JL, Aster JC. Molecular biology of anaplastic lymphoma kinasepositive anaplastic large-cell lymphoma. J Clin Oncol 2002;20:3691-702. [7] Wellmann A, Thieblemont C, Pittaluga S, et al. Detection of differentially expressed genes in lymphomas using cDNA arrays: identification of clusterin as a new diagnostic marker for anaplastic large-cell lymphomas. Blood 2000;96:398-404. [8] Lamant L, de Reynies A, Duplantier MM, et al. Gene-expression profiling of systemic anaplastic large-cell lymphoma reveals differences based on ALK status and two distinct morphologic ALK+ subtypes. Blood 2007;109:2156-64. [9] Calin GA, Croce CM. MicroRNA-cancer connection: the beginning of a new tale. Cancer Res 2006;66:7390-4. [10] Mirnezami AH, Pickard K, Zhang L, et al. MicroRNAs: key players in carcinogenesis and novel therapeutic targets. Eur J Surg Oncol 2009;35:339-47. [11] Melo SA, Esteller M. Dysregulation of microRNAs in cancer: playing with fire. FEBS Lett 2011;585:2087-99. [12] Zhao H, Wang D, Du W, et al. MicroRNA and leukemia: tiny molecule, great function. Crit Rev Oncol Hematol 2010;74:149-55. [13] Zhao JJ, Lin J, Lwin T, et al. MicroRNA expression profile and identification of miR-29 as a prognostic marker and pathogenetic factor by targeting CDK6 in mantle cell lymphoma. Blood 2010;115:2630-9. [14] Calin GA, Ferracin M, Cimmino A, et al. A MicroRNA signature associated with prognosis and progression in chronic lymphocytic leukemia. N Engl J Med 2005;353:1793-801. [15] Donnem T, Eklo K, Berg T, et al. Prognostic impact of MiR-155 in non-small cell lung cancer evaluated by in situ hybridization. J Transl Med 2011;9:6. [16] Mueller DW, Rehli M, Bosserhoff AK. miRNA expression profiling in melanocytes and melanoma cell lines reveals miRNAs associated with formation and progression of malignant melanoma. J Invest Dermatol 2009;129:1740-51.
2005 [17] Drexler HG, MacLeod RA. Malignant hematopoietic cell lines: in vitro models for the study of anaplastic large-cell lymphoma. Leukemia 2004;18:1569-71. [18] Matsuyama H, Suzuki HI, Nishimori H, et al. miR-135b mediates NPM-ALK-driven oncogenicity and renders IL-17-producing immunophenotype to anaplastic large cell lymphoma. Blood 2011;118:6881-92. [19] Merkel O, Hamacher F, Laimer D, et al. Identification of differential and functionally active miRNAs in both anaplastic lymphoma kinase (ALK)+ and ALK- anaplastic large-cell lymphoma. PNAS 2010;107:16228-33. [20] Liu C, Iqbal J, Teruya-Feldstein J, et al. MicroRNA expression profiling identifies molecular signatures associated with anaplastic large cell lymphoma. Blood 2013;122:2083-92. [21] Nelson PT, Wilfred BR. In situ hybridization is a necessary experimental complement to microRNA (miRNA) expression profiling in the human brain. Neurosci Lett 2009;466:69-72. [22] Song R, Ro S, Yan W. In situ hybridization detection of microRNAs. Methods Mol Biol 2010;629:287-94. [23] Smyth GK. Linear models and empirical Bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol 2004;3:1544-6115. [24] Benjamini Y, Hochberg Y. Controlling the false discovery rate - a practical and powerful approach to multiple testing. J R Stat Soc Ser B Methodol 1995;57:289-300. [25] Faraoni I, Antonetti FR, Cardone J, et al. miR-155 gene: a typical multifunctional microRNA. Biochim Biophys Acta 2009;1792: 497-505. [26] Sandhu SK, Croce CM, Garzon R. Micro-RNA expression and function in lymphomas. Adv Hematol 2011;2011:347137. [27] Dorsett Y, McBride KM, Jankovic M, et al. MicroRNA-155 suppresses activation-induced cytidine deaminase-mediated Myc-Igh translocation. Immunity 2008;28:630-8. [28] Lamprecht B, Walter K, Kreher S, et al. Derepression of an endogenous long terminal repeat activates the CSF1R proto-oncogene in human lymphoma. Nat Med 2010;16:571-9. [29] Kumar R, Manning J, Spendlove HE, et al. ZNF652, a novel zinc finger protein, interacts with the putative breast tumor suppressor CBFA2T3 to repress transcription. Mol Cancer Res 2006;4:655-65. [30] Holm R, Knopp S, Kumar R, et al. Expression of ZNF652, a novel zinc finger protein, in vulvar carcinomas and its relation to prognosis. J Clin Pathol 2008;61:59-63. [31] Gomes I, Sharma TT, Edassery S, et al. Novel transcription factors in human CD34 antigen-positive hematopoietic cells. Blood 2002;100:107-19. [32] Cano CE, Gommeaux J, Pietri S, et al. Tumor protein 53-induced nuclear protein 1 is a major mediator of p53 antioxidant function. Cancer Res 2009;69:219-26. [33] Gironella M, Seux M, Xie MJ, et al. Tumor protein 53-induced nuclear protein 1 expression is repressed by miR-155, and its restoration inhibits pancreatic tumor development. PNAS 2007;104:16170-5. [34] Gonzalez-Cano L, Hillje AL, Fuertes-Alvarez S, et al. Regulatory feedback loop between TP73 and TRIM32. Cell Death Dis 2013;4:e704. [35] Ludwig JA, Weinstein JN. Biomarkers in cancer staging, prognosis and treatment selection. Nat Rev Cancer 2005;5:845-56. [36] Nielsen BS, Jorgensen S, Fog JU, et al. High levels of microRNA-21 in the stroma of colorectal cancers predict short disease-free survival in stage II colon cancer patients. Clin Exp Metastasis 2011;28:27-38. [37] Philippidou D, Schmitt M, Moser D, et al. Signatures of microRNAs and selected microRNA target genes in human melanoma. Cancer Res 2010;70:4163-73. [38] Kopp KL, Ralfkiaer U, Nielsen BS, et al. Expression of miR-155 and miR126 in situ in cutaneous T-cell lymphoma. APMIS 2013;121:1020-4. [39] Rossing M, Borup R, Henao R, et al. Down-regulation of microRNAs controlling tumourigenic factors in follicular thyroid carcinoma. J Mol Endocrinol 2012;48:11-23.
