Mol Cell Biochem DOI 10.1007/s11010-016-2904-x

hnRNPA1 autoregulates its own mRNA expression to remain non-cytotoxic Hiroaki Suzuki1 • Masaaki Matsuoka1,2

Received: 17 August 2016 / Accepted: 3 December 2016 Ó Springer Science+Business Media New York 2016

Abstract Heterogeneous nuclear ribonucleoprotein (hnRNP)A1, a member of the hnRNP family, is involved in a variety of RNA metabolisms. The hnRNPA1 expression is altered in some human diseases and mutations of the hnRNPA1 gene cause amyotrophic lateral sclerosis and multisystem proteinopathy. It has been therefore assumed that the dysregulation of hnRNPA1 is linked to the pathogenesis of the diseases. However, the mechanism underlying the regulation of the hnRNPA1 expression remains unknown. In this study, using cell-based models, we have found that hnRNPA1 negatively regulates its own mRNA expression by inhibiting the intron10 splicing of hnRNPA1 pre-mRNA. This mechanism likely serves as an autoregulation of the hnRNPA1 expression. We have also found that a low-grade excess of hnRNPA1 expression causes cytotoxicity by activating the mitochondrial apoptosis pathway. Collectively, these data suggest that the level of hnRNPA1 is strictly controlled to be within a certain range by the mRNA autoregulation in the physiological condition so that the cytotoxicity-causative alteration of hnRNPA1 expression does not take place. Keywords hnRNPA1  Autoregulation  Splicing  Cell death

& Masaaki Matsuoka [email protected] 1

Department of Pharmacology, Tokyo Medical University, 6-1-1 Shinjuku, Shinjuku-Ku, Tokyo 160-8402, Japan

2

Department of Dermatological Neuroscience, Tokyo Medical University, 6-1-1 Shinjuku, Shinjuku-Ku, Tokyo 160-8402, Japan

Introduction Heterogeneous nuclear ribonucleoprotein (hnRNP)A1 is an abundant ubiquitous nuclear protein belonging to the hnRNP family [1]. The HNRNPA1 gene encodes two splicing transcripts, a shorter isoform hnRNPA1a (320 a.a.) and a longer isoform hnRNPA1b (372 a.a.). The level of hnRNPA1a expression is 20-times more abundant than that of hnRNPA1b in most tissues [1]. hnRNPA1 has numerous functions related to the RNA metabolisms, including the regulation of transcription, splicing, RNA transport, RNA stability, translation, and miRNA biogenesis [1]. It has been shown that the altered expression of hnRNPA1 and the mutation of the hnRNPA1 gene are associated with some human diseases. The expression of hnRNPA1 is upregulated in a variety of cancers, such as gliomas, lung cancers, colorectal cancers, and liver cancers [2–6]. Missense mutations in the HNRNPA1 gene cause amyotrophic lateral sclerosis (ALS) and multisystem proteinopathy in an autosomal dominant fashion [7]. It was also shown that the expression of hnRNPA1 may be downregulated in patients with Alzheimer’s disease and ALS [8, 9]. These observations suggest that the altered regulation of hnRNPA1 may be linked to the pathogenesis of the human diseases. However, little is known about the mechanism underlying the regulation of the hnRNPA1 expression and how the abnormal regulation of the hnRNPA1 expression is linked to the pathogenesis of diseases. In this study, we show that the overexpression of hnRNPA1 negatively regulates its own mRNA expression by inhibiting the intron10 splicing of hnRNPA1 premRNA. We also show that enforced low-grade overexpression of hnRNPA1 by less than two times over the endogenous level causes cytotoxicity in cancer and

123

Mol Cell Biochem

neuronal cells. These data together suggest that the expression level of hnRNPA1 is tightly autoregulated at the mRNA level, and the autoregulation of hnRNPA1 mRNA likely serves as a self-defense system against the cytotoxicity, caused by abnormal expression of hnRNPA1.

Cell death assay and cell viability assay Cell death assay and cell viability assay were as described previously [11]. Western blot analysis

Materials and methods Antibodies and a compound The following antibodies were purchased from suppliers: hnRNPA1, cleaved-caspase-3, cleaved-PARP, and GAPDH, Cell Signaling TECHNOLOGY (Beverly, MA, USA); Bcl-xS/L, Santa Cruz (Santa Cruz, CA, USA); actin, Sigma (St Louis, MO, USA); Xpress, Invitrogen (Carlsbad, CA, USA); and horseradish peroxidase (HRP)-conjugated goat anti-rabbit secondary antibody and HRP-conjugated goat anti-mouse secondary antibody, Bio-Rad (Hercules, CA, USA). Cycloheximide was purchased from Calbiochem (Darmstadt, Germany). Plasmid constructs Human hnRNPA1a cDNA was provided by Dr. Naoyuki Kataoka (Kyoto University, Graduate School of Medicine). Human hnRNPA1b cDNA was provided by Dr. Adrian R. Krainer (Cold Spring Harbor). The hnRNPA1a cDNA was subcloned into the pEF1/Myc-His vector (Invitrogen) with a native stop codon to construct non-tagged hnRNPA1a. Adenoviral vector-mediated expression The systems of adenovirus expression vectors were purchased from TaKaRa (Shiga, Japan). LacZ, Cre, and CreBcl-xL adenovirus vectors were as described previously [10]. cDNAs encoding non-tagged hnRNPA1a-wt and nontagged hnRNPA1b-wt were inserted into the SwaI site of a cosmid adenoviral vector, pAxCALNLw. In this vector, a stuffer DNA fragment, sandwiched by two loxP sequences, is located just upstream of cDNA and interferes with gene expression. If an adenovirus vector expressing Cre-recombinase is co-introduced into the cells, the stuffer is removed and gene begins to be expressed. To keep the total amounts of viruses constant, appropriate amounts of LacZ adenovirus vectors were added for each infection. All samples were co-infected with Cre-recombinase adenovirus vector at a multiplicity of infection (moi) of 40. Cell culture and transfection Cell culture and transfection were as described previously [11].

123

Western blot analysis was as described previously [11]. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or actin was visualized as an internal control. Intensities of immunodetected signals were densitometrically measured with an ImageJ software. The intensity of a band was normalized with that of actin or GAPDH. hnRNPA1 intron10 splicing assay Total RNA was extracted from NSC34 cells that had been introduced with hnRNPA1a-encoding adenovirus vector together with a reporter vector using the RNeasy Plus mini kit (Qiagen, Valencia, CA, USA) with DNase treatment (Qiagen). First-strand cDNAs were synthesized from total RNA using Sensescript reverse transcriptase (Qiagen). PCR amplification with KOD-plus-ver. 2 (Toyobo, Osaka, Japan) was performed under denaturation at 98 °C for 10 s, annealing at 59 °C for 30 s, and elongation at 68 °C for 45 s, repeated 30 cycles for hnRNPA1 fragment and 19 cycles for GAPDH. The sequences of forward and reverse primers are as follows: hnRNPA1 fragment, sense: 50 -CTAACACCGAGTTCGTGAAG-30 , antisense: 50 -CACCACTGTGCTTGGCTG-30 ; GAPDH, sense: 50 ACCACAGTCCATGCCATCAC-30 , antisense: 50 -TCCA CCACCCTGTTGCTGTA-30 . The hnRNPA1 forward and reverse primers are located within the Renilla luciferase gene and the hnRNPA1 exon11 region, respectively. The amplified DNA fragments were verified by sequence analysis. Quantitative real-time PCR analysis Total RNA was extracted from HeLa cells or NSC34 cells infected with indicated adenovirus vectors using RNeasy Plus mini kit (Qiagen) with DNase treatment (Qiagen) or ISOGEN (Wako, Osaka, Japan). Real-time PCR was as described previously [11]. The pairs of primers and the Taqman probes for target mRNAs were designed based on human or mouse mRNA sequences using TaqMan Gene Expression Assays (Applied Biosystems, Carlsbad, CA, USA, Assay ID: human hnRNPA1, Hs01656228_s1; human GAPDH, Hs02758991_g1; mouse hnRNP A1, Mm02528230_g1; mouse FUS, Mm01271304_m1; and mouse GAPDH, Mm99999915_g1). The target sequence of human hnRNPA1 and mouse hnRNPA1 probes is located within 50 untranslated region (UTR) of

Mol Cell Biochem

human hnRNPA1 and 30 UTR region of mouse hnRNPA1, respectively, and these probes can detect only endogenous hnRNPA1a and hnRNPA1b mRNA expression.

Results Overexpression of hnRNPA1 downregulates its own mRNA expression

Luciferase assay The luciferase reporter plasmid vector (psiCHECK2) (provided by Dr. Sean P. Ryder, University of Massachusetts Medical School) is a reporter plasmid containing the target sequence cloned into the multiple cloning site (MCS) located just downstream of the Renilla luciferase translational stop codon [12]. The cloned target sequence is transcribed as a part of Renilla luciferase mRNA. The firefly luciferase gene, which expresses firefly luciferase under the control of distinct promoter from the Renilla luciferase gene, is in tandem inserted downstream of the target sequence to monitor transfection efficiency. 50 UTR, 30 UTR, and intron regions of the human HNRNPA1 gene were cloned from total RNA or genomic DNA of HeLa and HL60 cells (Takara). The mouse Hnrnpa1 gene containing from exon6 to intron10 region was derived from the pmA1 vector (provided by Dr. Benoit Chabot, Universite´ de Sherbrooke) [13]. NSC34 cells, seeded onto 24-well plates at 4 9 104 cells/ well, were transfected with the reporter plasmid together with an hnRNPA1a-encoding vector. At 48 h after transfection, luciferase assays were performed with Dual-Luciferase Reporter Assay (Promega, Madison, WI, USA). Calculated luciferase activities were normalized by transfection efficiency. siRNA-mediated knock-down siRNAs against hnRNPA1 and non-targeting control siRNA (siControl) were purchased from RNAi Co., Ltd. (Tokyo, Japan). The siRNA sequences for hnRNPA1-#1 and -#2 are 50 -CUUUGGGUUUGUCACAUAUGC-30 and 50 -GGCUAU AAUGGAUUUGGCAAU-30 , respectively. NSC34 cells, seeded on 24-well plates at 4 9 104 cells/well, were transfected with 5 nM siRNA in association with a reporter plasmid and the Lipofectamine2000 reagent (Invitrogen) according to the manufacturer’s reverse transfection protocol. Statistical analysis All values in figures are shown as mean ± SD. All experiments that were statistically analyzed were performed with N = 3. Statistical analysis was performed with Student’s t test. *p \ 0.05 n.s.: not significant.

Several RNA-binding proteins, belonging to the hnRNP family, are known to regulate their own mRNA expression [14–19]. To examine whether hnRNPA1 also autoregulates its own mRNA expression, human hnRNPA1a and hnRNPA1b were overexpressed in human adenocarcinoma HeLa cells and mouse motor neuronal NSC34 cells, a hybrid cell line established from a mouse neuroblastoma cell line and mouse embryo spinal cord cells, using adenoviral vectors. In HeLa cells, overexpression of hnRNPA1a decreased the expression of endogenous hnRNPA1b (long exposure of Fig. 1a, lanes 1, 2) and overexpression of hnRNPA1b decreased the expression of endogenous hnRNPA1a (Fig. 1a, lanes 1, 3). We are technically incapable of knowing whether the overexpression of exogenous hnRNPA1a (or hnRNPA1b) decreased the expression of endogenous hnRNPA1a (or hnRNPA1b) because endogenous hnRNPA1a (or hnRNPAb) is indistinguishable from exogenous hnRNPA1a (or hnRNPA1b). Similarly, in NSC34 cells, overexpression of hnRNPA1a or hnRNPA1b reduced the expression of endogenous hnRNPA1b or hnRNPA1a, respectively (Fig. 1b). Furthermore, quantitative real-time PCR analysis showed that the overexpression of hnRNPA1a or hnRNPA1b downregulated endogenous levels of mRNA of hnRNPA1 consisting of both hnRNPA1a and hnRNPA1b in both HeLa cells (Fig. 1c) and NSC34 cells (Fig. 1d). A similar result was repeatedly obtained using primary cultured neurons (PCNs) (Fig. 1e). These data led to the conclusion that the overexpression of hnRNPA1 downregulates its own mRNA expression. Downregulation of hnRNPA1 mRNA expression is not mediated by the destabilization of mature hnRNPA1 mRNA nor the nonsense-mediated mRNA decay pathway We next examined the molecular mechanism underlying the overexpressed hnRNPA1-mediated downregulation of the hnRNPA1 mRNA. Previous studies have shown that hnRNPA1 are involved in the regulation of stability and splicing of RNA [1, 20]. We therefore hypothesized that the overexpression of hnRNPA1 autoregulates the level of hnRNPA1 mRNA by affecting these processes. We first examined the involvement of 50 UTR, the coding region, and 30 UTR of hnRNPA1 mRNA in the

123

Mol Cell Biochem

Fig. 1 hnRNPA1 negatively regulates its own mRNA expression. a, c HeLa cells, seeded on 6-well plates at 5 9 104 cells/well, were infected with the indicated adenovirus vectors at a moi of 100. At 48 h after the infection, the cell lysates were subject to immunoblot analysis using the indicated antibodies (a) and quantitative real-time PCR analysis of hnRNPA1 mRNA was performed (c). *p \ 0.05. b, d NSC34 cells, seeded on 6-well plates at 1 9 105 cells/well, were infected with the indicated adenovirus vectors at a moi of 400. At

48 h after the infection, the cell lysates were subject to immunoblot analysis using the indicated antibodies (b) and quantitative real-time PCR analysis of hnRNPA1 mRNA was performed (d). *p \ 0.05. e PCNs, seeded on 6-well plates at 1 9 106 cells/well, were infected with the indicated adenovirus vectors at a moi of 200. At 72 h after the infection, the cell lysates were subject to immunoblot analysis using the indicated antibodies. Intensities of immunodetected signals were densitometrically estimated with an ImageJ software

hnRNPA1-mediated downregulation of hnRNPA1 mRNA using a dual-luciferase reporter assay system. In this reporter assay system, a target sequence is cloned into the multiple cloning site (MCS), located just downstream of a Renilla luciferase translational stop codon (Fig. 2a) so that the cloned target sequence is transcribed as a part of 30 UTR of Renilla luciferase mRNA. The target sequence was the mature whole mRNA of hnRNPA1a consisting of 50 UTR, the coding region, and 30 UTR of hnRNPA1a (Fig. 2a, Mature mRNA of hnRNPA1a). After we co-transfected a reporter vector that expresses Renilla luciferase mRNA fused to the target sequence, together with the hnRNPA1aencoding vector, into NSC34 cells, we found that the overexpression of hnRNPA1a did not reduce luciferase activity (Fig. 2b). This result indicates that the hnRNPA1mediated downregulation of hnRNPA1 mRNA is not mediated by the exon-derived regions including 50 UTR, the protein-coding, and 30 UTR. Given that most hnRNP family proteins including hnRNPA1 are involved in the regulation of splicing [1, 20], it is also possible that the overexpression of hnRNPA1 causes skipping of an exon such as exon4 or exon10 of HNRNPA1 gene that produces a premature termination codon. This possibility prompted us to examine whether the hnRNPA1-induced downregulation of hnRNPA1 mRNA is mediated by the nonsense-mediated mRNA decay (NMD). However, as shown in Fig. 2c (left panel) and d, treatment of cells with cycloheximide, an NMD inhibitor, did not inhibit the hnRNPA1-mediated downregulation of hnRNPA1 although it decreased the basal levels of hnRNPA1 mRNA by an undefined mechanism. This result excludes the possibility that the downregulation of hnRNPA1 mRNA is mediated by NMD. On the other hand, treatment of cells with cycloheximide increased the

mRNA expression of fused in sarcoma (FUS) whose transcript variants are known to be a substrate for NMD (Fig. 2c, right panel) [19].

123

Downregulation of hnRNPA1 mRNA is mediated by inhibition of intron10 splicing of hnRNPA1 premRNA Based on the finding that the hnRNPA1-mediated downregulation is observed in both human and mouse cells (Fig. 1), we next examined whether the hnRNPA1 mRNA autoregulation is mediated by some sequences of hnRNPA1 introns that are conserved between the human HNRNPA1 and the mouse Hnrnpa1 genes. A BLASTmediated homology search indicated that the nucleotide sequence of the intron10 is conserved between human and mouse hnRNPA1 at the highest homology level (Fig. 3a), prompting us to examine whether the hnRNPA1-induced downregulation of hnRNPA1 mRNA is mediated by the intron10 sequence. To examine this possibility, we constructed a reporter vector that contains the 30 portion of the exon10, the intron10, and the exon11 of hnRNPA1, inserted downstream of the Renilla luciferase region (Fig. 2a, E10/I10/E11). As a putative negative control, we also constructed a reporter vector that contains the exon10 and the exon11 downstream of the Renilla luciferase region (Fig. 2a, E10/E11). Notably, the overexpression of hnRNPA1a significantly decreased luciferase activity in the cells in which the E10/I10/E11 reporter vector was introduced, but not in those in which the E10/E11 reporter vector was introduced (Fig. 3b). These results indicate that the hnRNPA1-mediated downregulation of hnRNPA1 is dependent on the intron10 and suggest that hnRNPA1 regulates the splicing of the intron10 of hnRNPA1 pre-

Mol Cell Biochem

Fig. 2 Downregulation of hnRNPA1 mRNA expression is not mediated by the destabilization of mature hnRNPA1 mRNA nor the nonsense-mediated mRNA decay pathway. a Schematic illustration of the human HNRNPA1 gene and the core of the luciferase reporter vector. The human HNRNPA1 gene is composed of exons 1-11 and introns 1-10. Exon8 is spliced out in hnRNPA1a. In the luciferase reporter vector, target sequences can be cloned into the multiple cloning site (MCS) located just downstream of the Renilla luciferase translational stop codon. The firefly luciferase gene, which expresses firefly luciferase under the control of distinct promoter from the Renilla luciferase gene, is in tandem inserted downstream of the target sequence to monitor transfection efficiency. b NSC34 cells,

seeded on 24-well plates at 4 9 104 cells/well, were transfected with 0.0625 lg/well of the indicated reporter vector in association with 0.025 lg/well of pEF1-Myc/His-vec or pEF1-hnRNPA1a. At 48 h after transfection, luciferase activity was measured using dualluciferase assays. c, d NSC34 cells, seeded on 6-well plates at 1 9 105 cells/well, were infected with the indicated adenovirus vectors at a moi of 800. At 24 h after infection, media were replaced with DMEM/N2 supplement. At 8 h from the replacement of media, cells were treated with 100 lg/mL cycloheximide (CHX) for 16 h. Then quantitative real-time PCR analysis of hnRNPA1 and FUS mRNA was performed (c), and the cell lysates were subject to immunoblot analysis using the indicated antibodies (d) *p \ 0.05

mRNA. In agreement, after we performed semi-quantitative PCR analysis using the E10/I10/E11 reporter vector as an artificial hnRNPA1 minigene, we found that the overexpression of hnRNPA1a inhibited the intron10 splicing (Fig. 3c, d). Both unspliced and spliced DNA sequences were confirmed by sequence analysis (Fig. 3e). These results show that hnRNPA1 negatively regulates its own mRNA expression by inhibiting the intron10 splicing, finally followed by a reduction in mature mRNA. We also examined whether other regions are involved in the hnRNPA1 autoregulation using an E6-I10 reporter vector that contains mouse Hnrnpa1 intron7, exon8, and intron8, the second most conserved regions between the human HNRNPA1 and the mouse Hnrnpa1 genes. As shown in

Fig. 3f, the overexpression of hnRNPA1a did not decrease luciferase activity in the cells in which the E6-I10 reporter vector was introduced (Fig. 3f). Finally, we asked whether the hnRNPA1 autoregulation occurs physiologically. For this purpose, we knocked down the endogenous hnRNPA1 using two types of specific siRNA against mouse hnRNPA1 in NSC34 cells and examined their effects on luciferase activity. Reduced expression of endogenous hnRNPA1 decreased luciferase activity in the cells in which the backbone reporter vector was introduced by an undefined mechanism (Fig. 3g, lanes 1, 2, 3). On the other hand, reduced expression of endogenous hnRNPA1 decreased luciferase activity to a lesser extent or did not reduce it in the cells in which the

123

Mol Cell Biochem

E10/I10/E11 reporter vector was introduced (Fig. 3g, lanes 4, 5, 6). If the value of luciferase activity in the cells in which the E10/I10/E11 reporter vector was introduced was adjusted by the average value of luciferase activity in the cells in which the backbone reporter vector was introduced, it is apparent that the luciferase activity, derived from the E10/I10/E11 reporter vector, was upregulated in the cells in which hnRNPA1 siRNAs were introduced (Fig. 3h). This result indicates that knock-down of endogenous hnRNPA1

123

expression increased the luciferase activity (Fig. 3h, i) and that hnRNPA1 autoregulation physiologically occurs. Overexpression of hnRNPA1 induces cell death We then asked whether dysregulation of hnRNPA1 autoregulation affects cell viability. To test this, we examined the effect of overexpression of hnRNPA1 on cell viability. Using lactate dehydrogenase (LDH) release cell

Mol Cell Biochem b Fig. 3 Downregulation of hnRNPA1 mRNA is mediated by inhibi-

tion of intron10 splicing of hnRNPA1 pre-mRNA. a Sequence alignment between the human HNRNPA1 (Upper) and the mouse Hnrnpa1 (Lower) intron10. Asterisk indicates the position that has conserved base between human and mouse. b NSC34 cells, seeded on 24-well plates at 4 9 104 cells/well, were transfected with 0.0625 lg/ well of the indicated reporter vector in association with 0.025 lg/well of pEF1-Myc/His-vec or pEF1-hnRNPA1a. At 48 h after transfection, luciferase activity was measured using dual-luciferase assays. *p \ 0.05. n.s. not significant. c, d NSC34 cells, transiently overexpressed hnRNPA1a by adenoviral infection at a moi of 800 together with psiCHECK2-hnRNPA1-E10/I10/E11 reporter plasmid by transfection, were harvested at 48 h after transfection and the prepared total RNA were used for splicing assays (c). Reverse transcription (RT) (-) was used as negative control for monitoring of PCR amplification from plasmid DNA. intron10 (?) or (-) indicated the amplicon that includes or does not include intron10 region, respectively. The cell lysates were subject to immunoblot analysis using the indicated antibodies (d). e Sequence analysis of splicing assay (c) was performed. Top panel: the sequence of intron10 (?) fragment containing exon10, intron10, and exon11; Bottom panel: the sequence of intron10 (-) fragment containing exon10 and exon11. f NSC34 cells, seeded on 24-well plates at 4 9 104 cells/well, were transfected with 0.0625 lg/well of the indicated reporter vectors in association with 0.025 lg/well of pEF1-Myc/His-vec or pEF1-hnRNPA1a. At 48 h after transfection, luciferase activity was measured using dualluciferase assays. n.s.: not significant. g, h NSC34 cells, seeded on 24-well plates at 4 9 104 cells/well, were transfected with 5 nM control siRNA (siControl), hnRNPA1-#1, or -#2 siRNA together with the indicated reporter plasmids using Lipofectamine 2000 reagent. At 72 h after transfection, luciferase activity was measured using dualluciferase assays (g). For normalization, the value of luciferase activity obtained from E10/I10/E11 reporter vector was divided by the average of luciferase activity obtained from backbone reporter vector (h). *p \ 0.05. i NSC34 cells, seeded on 12-well plates at 8 9 104 cells/well, were transfected with 5 nM control siRNA (siControl), hnRNPA1-#1, or -#2 siRNA together with the indicated reporter plasmids using Lipofectamine 2000 reagent. At 72 h after transfection, immunoblot analysis was performed using the indicated antibodies

death assay, we found that increase in expression of hnRNPA1a by less than two times over the endogenous level caused cell death in HeLa cells (Fig. 4a, b) and in NSC34 cells (Fig. 4c, d) in an expression level-dependent manner. Performing WST-8 cell viability assay, we also found that overexpression by less than 1.5 times over the endogenous level decreased cell viability of PCNs (Fig. 4e, f). The overexpression of hnRNPA1b also caused cell death (Fig. 4g, h). The hnRNPA1a-induced cell death was associated with increased cleavage of caspase-3 and cleavage of poly (ADP-ribose) polymerase (PARP), one of the caspase substrates (Fig. 4i). Co-expression of Bcl-xL completely inhibited hnRNPA1a-induced cell death and reversed the hnRNPA1a-mediated increase in the cleavage of caspase-3 and PARP (Fig. 4i, j). These results together indicate that overexpression of hnRNPA1 causes cytotoxicity through the mitochondrial apoptotic pathway.

Discussion In this study, we have shown that hnRNPA1 negatively autoregulates its own mRNA expression and inhibits the intron10 splicing of hnRNPA1 pre-mRNA (Figs. 1, 2, 3). Given that unspliced pre-mRNA is rapidly degraded by nucleases such as Xrn2 exonuclease [21], these results suggest that hnRNPA1 negatively regulates its own mRNA expression by inhibiting the intron10 splicing, potentiating the degradation of premature mRNA, and/or stalling of the splicing, finally followed by a reduction in mature mRNA. Regarding the mechanism underlying the hnRNPA1 autoregulation of its own mRNA, we have also addressed other possibilities. Our experiments have indicated that the splicing of introns 7 and 8 of hnRNPA1 that are the second most conserved regions between the human HNRNPA1 and the mouse Hnrnpa1 genes was not downregulated by hnRNPA1a overexpression (Fig. 3f) and that the autoregulation of the hnRNPA1 mRNA by hnRNPA1 does not occur through 50 UTR and 30 UTR (Fig. 2). The latter finding excludes the possibility that a previously identified RNAbinding protein named Quaking, which binds to 30 UTR of hnRNPA1 mRNA and affects the stabilization of hnRNPA1 mRNA [12], is involved in the hnRNPA1 autoregulation. It remains undetermined how hnRNPA1 inhibits the intron10 splicing of hnRNPA1 pre-mRNA. It is possible that hnRNPA1 binds to hnRNPA1 mRNA and directly inhibits the intron10 splicing. Alternatively, hnRNPA1 may indirectly inhibit the splicing of the intron10 through an undetermined mRNA-binding protein including some spliceosomal proteins. The detailed mechanism underlying this regulation remains to be investigated. Interestingly, the overexpression of hnRNPA1 by less than two times over the endogenous level causes cytotoxicity (Fig. 4). As a consequence, we could assume that the level of hnRNPA1 is tightly regulated by the mRNA autoregulation to be less than a certain level so that a cytotoxicity-causative excess of protein expression does not occur. Currently, the mechanism underlying the hnRNPA1-induced cell death remains undefined. hnRNPA1 has numerous functions on the RNA metabolisms, including the regulation of transcription, splicing, RNA transport, RNA stability, translation, and miRNA biogenesis [1]. The overexpression of hnRNPA1 may induce cytotoxicity through multiple processes of the RNA metabolism. Previously, we found that the low-grade overexpression of two other members of the hnRNP family, transactive response DNA-binding protein-43 (TDP43) and FUS, induces cell death [10, 22]. Previous studies by other researchers have shown that, similar to hnRNPA1, TDP-43 and FUS possess their own mRNA autoregulation systems [17–19] and suggested that the impairment of

123

Mol Cell Biochem

Fig. 4 Overexpression of hnRNPA1 induces cell death. a, b HeLa cells, seeded on 6-well plates at 5 9 104 cells/well, were infected with hnRNPA1a adenovirus vector at mois of 0–200. At 18 h after infection, media were replaced with DMEM/N2 supplement. At 24 h from the replacement of media, LDH release was measured (a) and immunoblot analysis was performed with the indicated antibodies (b). *p \ 0.05. c, d NSC34 cells, seeded on 6-well plates at 1 9 105 cells/ well, were infected with hnRNPA1a adenovirus vector at mois of 0–800. At 24 h after infection, media were replaced with DMEM/N2 supplement. At 24 h from the replacement of media, LDH release was measured (c) and immunoblot analysis was performed with the indicated antibodies (d). *p \ 0.05. e PCNs, seeded on 96-well plates at 5 9 104 cells/well, were infected with LacZ or hnRNPA1a adenovirus vector at a moi of 200. At 72 h after the infection, WST-8 cell viability assay was performed. *p \ 0.05. f PCNs, seeded on 6-well plates at 1 9 106 cells/well, were infected with LacZ or hnRNPA1a adenovirus vector at a moi of 200. At 72 h after the

infection, the cell lysates were subject to immunoblot analysis using the indicated antibodies. Intensities of immunodetected signals were densitometrically estimated with an ImageJ software. g, h HeLa cells, seeded on 6-well plates at 5 9 104 cells/well, were infected with hnRNPA1a or hnRNPA1b adenovirus vector at moi of 100. At 18 h after infection, media were replaced with DMEM/N2 supplement. At 24 h from the replacement of media, LDH release was measured (g) and immunoblot analysis was performed with the indicated antibodies (h). *p \ 0.05. i, j NSC34 cells, seeded on 6-well plates at 1 9 105 cells/well, were co-infected with LacZ or hnRNPA1a adenovirus vector at a moi of 600 in association with LacZ or BclxL adenovirus vector at a moi of 200. At 24 h after infection, media were replaced with DMEM/N2 supplement. At 24 h from the replacement of media, LDH release was measured (j) and immunoblot analysis was performed with the indicated antibodies (i). *p \ 0.05

mRNA autoregulation in TDP-43 and FUS may be linked to the pathogenesis of some neurodegenerative diseases [17–19]. Analogously, the impairment of mRNA autoregulation of hnRNPA1 may contribute to the onset and the progression of some hnRNPA1-linked diseases. hnRNPA1 expression is upregulated in some cancer cells [2–6]. These cancer cells may require higher levels of hnRNPA1 than normal counterpart cells because of their accelerated RNA metabolism, associated with high cell proliferation. If this is true, it is highly likely that the slight increase in the hnRNP level similarly causes cytotoxicity even in such cancer cells. Alternatively, these cancer cells may be fundamentally resistant to the cytotoxicity, caused by the overexpression of hnRNPA1. Furthermore, it could

be also hypothesized that the upregulation of hnRNPA1 causes tumorigenesis through the RNA metabolism. For example, it has been reported that hnRNPA1 facilitates the production of pre-miR-18a that potentially acts as an oncogene [23]. Missense mutations in the HNRNPA1 gene cause autosomal dominant ALS and multisystem proteinopathy [7]. The disease-linked mutations enhance hnRNPA1 fibrillization and recruitment of hnRNPA1 to RNA granules [7]. In multisystem proteinopathy patients with an hnRNPA1 mutation, hnRNPA1 mislocalizes in the cytoplasm and the nuclear level of hnRNPA1 is decreased in some fibers in muscle. These pathological features are also observed in both familial and sporadic inclusion body myositis cases without

123

Mol Cell Biochem

the hnRNPA1 mutations [7, 24]. Currently, it is unknown how the mutations of hnRNPA1 are linked to cytotoxicity and whether these pathological features are causes of, independent co-incidences to, or results of hnRNPA1-linked cytotoxicity. Another previous clinical studies also demonstrated that the levels of hnRNPA1 are downregulated in Alzheimer’s disease patients and sporadic ALS patients with TDP-43 aggregates [8, 9]. These observations suggest that the downregulation of hnRNPA1 may be linked to the pathogenesis of Alzheimer’s disease and some ALS cases. In support for this hypothesis, it has been shown that knockdown of hnRNPA1 causes neurological dysfunction [8]. Another study has also shown that knock-down of both hnRNPA1 and hnRNPA2 causes cytotoxicity in a variety of cells [25]. In this study, however, we were unable to show that siRNA-mediated knock-down of hnRNPA1 results in cytotoxicity (unpublished observation). In summary, our results suggest that hnRNPA1 autoregulates its own mRNA expression and the hnRNPA1-induced autoregulation of hnRNPA1 mRNA likely serves as a self-defense system against the cytotoxicity, caused by abnormal expression of hnRNPA1. Acknowledgements We are especially grateful to Takako Hiraki and Tomoko Yamada for essential assistance throughout the study. We thank Dr. Naoyuki Kataoka, Dr. Adrian R. Krainer, Dr. Sean P. Ryder, and Dr. Benoit Chabot for providing plasmids and Dr. Neil Cashman for providing NSC34 cells. This work was supported in part by the Grant-in-Aid for Scientific Research (B) [grant number 15H04689] to [M.M.] and for Scientific Research (C) [grant number 25460342] to [H.S.] from Japan Society for the Promotion of Science, by Japan ALS association (JALSA) to [H.S.], by Naito Foundation Natural Science Scholarship to [M.M.], and by Akaeda Medical Research Foundation to [M.M.]. Compliance with ethical standards Conflict of interest The authors declare that they have no conflict of interest.

References 1. Jean-Philippe J, Paz S, Caputi M (2013) hnRNP A1: the Swiss army knife of gene expression. Int J Mol Sci 14:18999–19024 2. Ushigome M, Ubagai T, Fukuda H, Tsuchiya N, Sugimura T, Takatsuka J et al (2005) Up-regulation of hnRNP A1 gene in sporadic human colorectal cancers. Int J Oncol 26:635–640 3. David CJ, Chen M, Assanah M, Canoll P, Manley JL (2010) HnRNP proteins controlled by c-Myc deregulate pyruvate kinase mRNA splicing in cancer. Nature 463:364–368 4. Boukakis G, Patrinou-Georgoula M, Lekarakou M, Valavanis C, Guialis A (2010) Deregulated expression of hnRNP A/B proteins in human non-small cell lung cancer: parallel assessment of protein and mRNA levels in paired tumour/non-tumour tissues. BMC Cancer 10:434 5. Zhou ZJ, Dai Z, Zhou SL, Fu XT, Zhao YM, Shi YH et al (2013) Overexpression of HnRNP A1 promotes tumor invasion through regulating CD44v6 and indicates poor prognosis for hepatocellular carcinoma. Int J Cancer 132:1080–1089

6. Liu X, Zhou Y, Lou Y, Zhong H (2016) Knockdown of HNRNPA1 inhibits lung adenocarcinoma cell proliferation through cell cycle arrest at G0/G1 phase. Gene 576:791–797 7. Kim HJ, Kim NC, Wang YD, Scarborough EA, Moore J, Diaz Z et al (2013) Mutations in prion-like domains in hnRNPA2B1 and hnRNPA1 cause multisystem proteinopathy and ALS. Nature 495:467–473 8. Berson A, Barbash S, Shaltiel G, Goll Y, Hanin G, Greenberg DS et al (2012) Cholinergic-associated loss of hnRNP-A/B in Alzheimer’s disease impairs cortical splicing and cognitivefunction in mice. EMBO Mol Med 4:730–742 9. Honda H, Hamasaki H, Wakamiya T, Koyama S, Suzuki SO, Fujii N et al (2015) Loss of hnRNPA1 in ALS spinal cord motor neurons with TDP-43-positive inclusions. Neuropathology 35:37–43 10. Suzuki H, Lee K, Matsuoka M (2011) TDP-43-induced death is associated with altered regulation of BIM and Bcl-xL and attenuated by caspase-mediated TDP-43 cleavage. J Biol Chem 286:13171–13183 11. Suzuki H, Shibagaki Y, Hattori S, Matsuoka M (2015) Nuclear TDP-43 causes neuronal toxicity by escaping from the inhibitory regulation by hnRNPs. Hum Mol Genet 24:1513–1527 12. Zearfoss NR, Clingman CC, Farley BM, McCoig LM, Ryder SP (2011) Quaking regulates Hnrnpa1 expression through its 30 UTR in oligodendrocyte precursor cells. PLoS Genet 7:e1001269 13. Blanchette M, Chabot B (1997) A highly stable duplex structure sequesters the 50 splice site region of hnRNP A1 alternative exon 7B. RNA 3:405–419 14. Wollerton MC, Gooding C, Wagner EJ, Garcia-Blanco MA, Smith CW (2004) Autoregulation of polypyrimidine tract binding protein by alternative splicing leading to nonsense-mediated decay. Mol Cell 13:91–100 15. Dredge BK, Stefani G, Engelhard CC, Darnell RB (2005) Nova autoregulation reveals dual functions in neuronal splicing. EMBO J 24:1608–1620 16. Rossbach O, Hung LH, Schreiner S, Grishina I, Heiner M, Hui J et al (2009) Auto- and Cross-Regulation of the hnRNP L Proteins by Alternative Splicing. Mol Cell Biol 29:1442–1451 17. Ayala YM, De Conti L, Avendan˜o-Va´zquez SE, Dhir A, Romano M, D’Ambrogio A et al (2011) TDP-43 regulates its mRNA levels through a negative feedback loop. EMBO J 30:277–288 18. Avendan˜o-Va´zquez SE, Dhir A, Bembich S, Buratti E, Proudfoot N, Baralle FE (2012) Autoregulation of TDP-43 mRNA levels involves interplay between transcription, splicing, and alternative polyA site selection. Genes Dev 26:1679–1684 19. Zhou Y, Liu S, Liu G, Oztu¨rk A, Hicks GG (2013) ALS-associated FUS mutations result in compromised FUS alternative splicing and autoregulation. PLoS Genet 9:e1003895 20. Han SP, Tang YH, Smith R (2010) Functional diversity of the hnRNPs: past, present and perspectives. Biochem J 430:379–392 21. Davidson L, Kerr A, West S (2012) Co-transcriptional degradation of aberrant pre-mRNA by Xrn2. EMBO J 31:2566–2578 22. Suzuki H, Matsuoka M (2015) Overexpression of nuclear FUS induces neuronal cell death. Neuroscience 287:113–124 23. Guil S, Ca´ceres JF (2007) The multifunctional RNA-binding protein hnRNP A1 is required for processing of miR-18a. Nat Struct Mol Biol 14:591–596 24. Pinkus JL, Amato AA, Taylor JP, Greenberg SA (2014) Abnormal distribution of heterogeneous nuclear ribonucleoproteins in sporadic inclusion body myositis. Neuromuscul Disord 24:611–616 25. Patry C, Bouchard L, Labrecque P, Gendron D, Lemieux B, Toutant J et al (2003) Small interfering RNA-mediated reduction in heterogeneous nuclear ribonucleoparticule A1/A2 proteinsinduces apoptosis in human cancer cells but not in normal mortal cell lines. Cancer Res 63:7679–7688

123