(This is a sample cover image for this issue. The actual cover is not yet available at this time.)

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial research

and education use, including for instruction at the author'sinstitution and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier's archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

lable at ScienceDirect

Toxicon 121 (2016) 1e9

Author's Personal Copy

Contents lists avai

Toxicon

journal homepage: www.elsevier .com/locate/ toxicon

RNA sequence analyses of r-Moj-DM treated cells: TXNIP is required toinduce apoptosis of SK-Mel-28

Terri D. McBride, U. Andrew, Nicko Ly, Julio G. Soto*

Biological Sciences Department, San Jos�e State University, One Washington Square, San Jos�e, CA 95192-0100, USA

a r t i c l e i n f o

Article history:Received 4 June 2016Received in revised form25 July 2016Accepted 23 August 2016Available online 24 August 2016

Keywords:RNA sequencingRecombinant RGD disintegrinTXNIP knockdownsCell viabilityTXNIP protein localization

* Corresponding author.E-mail address: Julio.Soto@sjsu.edu (J.G. Soto).

http://dx.doi.org/10.1016/j.toxicon.2016.08.0140041-0101/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

RNA sequencing of untreated and r-Moj-DM treated SK-Mel-28 cells was performed after 6 h, to beginunraveling the apoptotic pathway induced by r-Moj-DM. Bioinformatic analyses of RNA sequencing datayielded 40 genes that were differentially expressed. Nine genes were upregulated and 31 were down-regulated. qRT-PCR was used to validate differential expression of 13 genes with known survival orapoptotic-inducing activities. Expression of BNiP3, IGFBP3, PTPSF, Prune 2, TGF-ß, and TXNIP werecompared from cells treated with r-Moj-DN (a strong apoptotic inducer) or r-Moj-DA (a non-apoptoticinducer) for 1 h, 2 h, 4 h, and 6 h after treatment. Our results demonstrate that significant differencesin expression are only detected after 4 h of treatment. In addition, expression of TXNIP (an apoptoticinducer) remains elevated at 4 h and 6 h only in r-Moj-DN treated cells. Based on the consistency ofelevated TXNIP expression, we further studied TXNIP as a novel target of disintegrin activation. Confocalmicroscopy of anti-TXNIP stained SK-Mel-28 cells suggests nuclear localization of TXNIP after r-Moj-DMtreatment. A stable TXNIP knockdown SK-Mel-28 cell line was produced to test TXNIP' role in theapoptotic induction by r-Moj-DM. High cell viability (74.3% ±9.1) was obtained after r-Moj-DM treatmentof TXNIP knocked down SK-Mel-28 cells, compared to 34% ±0.187 for untransduced cells. These resultssuggest that TXNIP is required early in the apoptotic-inducing pathway resulting from r-Moj-DM bindingto the av integrin subunit.

© 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Apoptosis can result from the detachment of cells from theextracellular matrix or adjacent cells in tissues (Gilmore, 2005). Cellshrinkage and blebbing, chromatin condensation, chromosomalfragmentation, and formation of apoptotic bodies are morpholog-ical characteristics associated with apoptosis (Lockshin and Zakeri,2004).When normal cells detach, an extrinsic pathway of apoptosisis initiated (Gilmore, 2005). However, cancer cells can efficientlyavoid apoptosis by a variety of mechanisms (Hanahhan andWeinberg, 2011). Because of this apoptotic avoidance, apoptoticinduction of cancer cells has been a target for the development ofanti-cancer treatments (Meng et al., 2006).

Disintegrins induce apoptosis of normal and cancer cell lines byantagonizing integrins (Ramos et al., 2016; Huveneers et al., 2007;Barja-Fidalgo et al., 2005). Examples of apoptotic inducing dis-integrins include: r-Moj-DL (Ramos et al., 2016), r-Moj-DN (Ramos

et al., 2016), r-Moj-DM (Ramos et al., 2016), DisBa-01 (Ribeiro et al.,2014), r-Rub (Carey et al., 2012), vicrostatin (Minea et al., 2010),agkistin-s (Ren et al., 2006), echistatin (Alimenti et al., 2004;Brassard et al., 1999), rhodostomin (Wu et al., 2003), salmosin(Hong et al., 2003), contortrostatin (Zhou et al., 1999), and accutin(Yeh et al., 1998).

The complete elucidation of signal transduction pathwaysresulting in apoptotic induction after disintegrin treatment remainsan important, mostly uncharted line of research. Moreover, somecytoplasmic and nuclear targets have been identified. FAK, JNK,ERK, and PI-3K are important targets of integrin-induced apoptosis(Chiarugi and Giannoni, 2008) and are also common targets ofseveral disintegrins (Park et al., 2012; Selistre-de-Araujo et al.,2010). For instance, echistatin treatment results in the activation ofcaspase-3 and the Tyr phosphorylation inhibition of FAK (Alimentiet al., 2004). Rhodostomin inhibits theMAPK pathway by inhibitingthe phosphorylation of p38, JNK, and ERK (Hsu et al. 2010, 2016).Other disintegrins inhibit NF-kB (Jain and Kumar, 2012), over-express TGF-b (Ribeiro et al., 2014) or TNF-a (Calderon et al., 2014),or suppress IL-8 expression (Kim et al., 2007).

T.D. McBride et al. / Toxicon 121 (2016) 1e92

Author's Personal Copy

We have previously demonstrated that r-Moj-DM inducesapoptosis of SK-Mel-28 cells by antagonizing the av integrin sub-unit (Ramos et al., 2016). There are two distinct recombinantmojastin disintegrins, one that has a 10 amino acid truncation atthe amino terminus (r-Moj-WN) and the other one with a full-length sequence (r-mojastin-1). The first recombinant version (r-Moj-WN) was originally derived from a truncated cDNA clone thatwas synthesized from mRNA isolated from a Crotalus scutulatusscutulatus (Mohave) rattlesnake venom gland (Soto et al., 2007).The Mohave venom contains two disintegrin isoforms namedmojastin 1 and 2 (Sanchez et al., 2006). Both mojastins inhibitplatelet aggregation and T24 cell binding to fibronectin. Recombi-nant mojastin 1 disintegrin also inhibits metastasis and tumorcolonization (Lucena et al., 2011). The recombinant r-Moj-DMcontains two mutations in the binding loop (WN to DM) makingthis disintegrin different than r-mojastin-1 (Seoane et al., 2010).

Our present work aims to unravel the signal transductionpathway that is activated upon r-Moj-DM treatment. Instead ofexamining the expression of known pro-apoptotic or survival fac-tors, we identified the expression of gene targets using RNAsequencing after 6 h of r-Moj-DM treatment of SK-Mel-28 cells.Several insufficiently studied gene targets were identified. In thepresent work, we demonstrate that TXNIP, an apoptotic inducerprotein linked to endoplasmic reticulum (ER) damage (Lerner et al.,2012; Oslowski et al., 2012) and sugar metabolism (Ji et al., 2016;Shen et al., 2015), is required for the induction of apoptosis of SK-Mel-28 after r-Moj-DM treatment.

2. Materials and methods

2.1. Expression and purification of r-Moj-DA, r-Moj-DM, r-Moj-DN,and GST-Moj-DM peptides

Expression and purification of r-Moj-DA, r-Moj-DM, and r-Moj-DN peptides were performed with a method described earlier(Ramos et al., 2016). GST-Moj-DM peptides used in confocal mi-croscopy experiments were expressed and purified with themethod described by Seoane et al. (2010).

2.2. Cell culture conditions

SK-Mel-28 cells were grown in Eagle's minimum essential me-dium (ATCC). Media was supplemented with 10% fetal bovineserum (FBS) and penicillin-100 (IU/mL), streptomycin (0.1 mg/mL),and amphotericin B (0.25 mg/mL). Cells were grown at 37 �C in ahumidified incubator with 5% CO2.

2.3. RNA isolation and RNA sequencing

Onemillion SK-Mel 28 cells were treated with r-Moj-DM for 6 h,and total RNA extracted using the RNeasy Mini Kit (Qiagen). RNAwas quantified using a NanoDrop spectrophotometer. One thou-sand nanograms of total RNAwere isolated from any contaminatinggenomic DNA with the use of gDNA Wipeout Buffer from theQuantiTect Reverse Transcription Kit (Qiagen). The RNA waschecked for quality on the Agilent Bioanalyzer. Libraries were madewith Illumina TruSeq according to the manufacturer's protocol.Two RNA samples per sequenced, one from treated cells and onefrom untreated cells. Sequencing was performed 2 � 100 PE on theHiSeq2000 by Centrillion Technologies.

2.4. Bioinformatics analyses

Sequencing results (reads) were analyzed using the softwaretools tophat (Trapnell et al., 2009), bowtie (Langmead et al., 2009),

and cufflinks (Roberts et al., 2011). All mapped reads were analyzedusing cufflinks to identify transcripts and the FPKM (Fragments PerKilobase of transcript per Million) was then calculated. Differentialexpression levels between treated and control samples wereanalyzed using the software tool cuffdiff (Roberts et al., 2011).

2.5. Validation of RNA sequencing results using quantitative RT-PCR

Three independent cell cultures were treatedwith r-Moj-DM for6 h and the total RNA isolated as described in the methods section2.3. The isolated RNA was reverse transcribed at 42 �C for 15 minusing Quantiscript Reverse Transcriptase (Qiagen). Reactions wereprepared for each sample cDNA with correlating primers. Samplereactions containing 2.5 ml of cDNA and 1.0 mM of forward andreverse primers (Supplemental Table 1) were prepared in triplicatewells with the use of QuantiTect SYBR Green PCR kit (Qiagen). ThePCR conditions were 1 cycle of 95 �C for 5 min, followed by 40cycles of amplification (95 �C for 10 s and 60 �C for 30 s), and a finalextension cycle at 95 �C for 15 s and 60 �C for 1min. Amelting curvewas used to verify the presence of one specific peak for the gene.Average Ct was used to calculate relative fold change. Normaliza-tion of the data was performed by the use an endogenous control,HPRT primers (forward: ATGACCAGTCAACAGGGGAC, and reverse:GGTCCTTTTCACCAGCAAGC), and an untreated control sample.

2.6. Detection of six apoptotic gene expression by qPCR after r- Moj-DA or r-Moj-DN peptide treatment

One million SK-Mel 28 cells were treated with r- Moj-DA or r-Moj-DN peptides for 1 h, 2 h, 4 h, and 6 h. Total RNA was extractedfrom each sample and purified using the method described insection 2.1. The isolated RNA was reverse transcribed at 42 �C for15 min using Quantiscript Reverse Transcriptase (Qiagen). Re-actions were prepared for each sample cDNA with correlatingprimers as described in the methods section 2.5.

2.7. TXNIP cellular localization using confocal microscopy

Two hundred thousand SK-Mel-28 cells were seeded onto tissueculture microscope slides containing 2 mL of media. Cells wereincubated with 3 mM of GST-Moj-DM or with 1X PBS for 24 h. Thencells were washed with warm 1X PBS and fixed using 3% EM gradeformaldehyde for 1 h. This was followed by a 10 min incubationwith 1 mL of 0.2% triton x-100, and then a 20 min incubation with1 mL of goat serum. Three slides were incubated with anti-GST IgGgoat antibody conjugated to DyLight® 650 (ab117497) at a 1:200dilution, and three slides were incubated with anti-VDUP1 (TXNIP)(H-2) mouse IgG (sc-271328), at a 1:1000 dilution, overnight at 4 �Cwith constant shaking. After this incubation, the slides were rinsedwith warm 1X PBS. Rabbit anti-mouse IgG conjugated to FITC(diluted at 1:1000, sc-358946) was added to the slides that werepreviously incubated with the anti-VDUP (TXNIP) antibody. Cellsthen were incubated for 45 min (in the dark) at room temperaturewith constant shaking. The secondary antibody excess wasremoved with three washes of 1X PBS. Cells were mounted andtheir nuclei stained with Vectashield mounting mediumwith DAPI.Slides were viewed at 100X, oil immersion with a confocal ZeissLSM 700 microscope.

2.8. TXNIP gene expression knockdown

Inhibition of TXNIP expression was performed in SK-Mel-28 cells using VDUP (TXNIP) shRNA (h) lentiviral particles (SantaCruz Biotech, sc-44943-V). Control shRNA lentiviral particles (sc-108080) were used as a negative scrambled shRNA sequence

T.D. McBride et al. / Toxicon 121 (2016) 1e9 3

Author's Personal Copy

control. Cells were plated in 12-well plates, 24 h prior to viralinfectionwith 1 mL of complete EMEMmedia. Then, the mediawaschanged with complete media containing 5 mg/mL of Polybrene (sc-134220). TXNIP shRNA (h) lentiviral particles were thawed andadded to the cells, mixed gently, and incubated overnight in a hu-midified 5% CO2 air incubator at 37 �C. After 24 h, the media waschanged with 1 mL of complete media without Polybrene andincubated overnight. To select stable clones, cells were split in theratio of 1:5 and incubated for 24 h in complete media. After 48 h,stable clones expressing TXNIP shRNA were selected using Puro-mycin dihydrochloride (sc-108071) at 5 mg/ml. The cells wereexpanded and grown in puromycin containing complete mediauntil ready for use.

2.9. Detecting a reduction of TXNIP transcript and protein onTXNIP-knocked downed SK-Mel-28 cells

The success of the stable TXNIP stable shRNA knockdown wasdetermined by QPCR and densitometry analysis of western blots.Three types of cells were used in these experiments: stable TXNIPknocked down, scrambled shRNA control, and untransduced cells.Both types of experiments were performed in triplicates. First,QPCR was used to detect TXNIP transcript levels after 24 h of r-Moj-DM treatment of 1 � 106 cells. Total RNA of treated cells was per-formed as described in section 2.3. QPCR was performed asdescribed in section 2.5. Second, whole cell lysates were obtainedfrom cells after 24 h of r-Moj-DM treatment and separated in a4e12% NuPAGE BiseTris gel under reducing conditions. Proteinswere then transferred to a nitrocellulose membrane. Membraneswere blocked in 5% skim milk containing 5% BSA (bovine serumalbumin) overnight at 4 �C with constant shaking. Then themembrane was incubated with anti-VDUP1 (TXNIP) (H-2) mouseIgG (sc-271328), at a 1:1000 dilution, overnight at 4 �C with con-stant shaking. Membranes were rinsed with 1X TBS for 30 min,with wash changes every 5 min. Goat anti-mouse IgG conjugatedwith HRP (sc-2005) was added at a concentration of 1:1500.Membranes were incubated at room temperature for 1 h withconstant shaking. Membranes were rinsed as described above andprocessed colometrically. Densitometry analyses of protein bandsobtained from the western blots were performed using Image-Jversion 1.51a.

2.10. Cell viability assay

The effects of r-Moj-DM peptides on cell viability were deter-mined using the WST-1 assay (GenDEPOT). Fifty thousand of TXNIPknocked down, control shRNA scrambled, and untransduced SK-Mel-28 cells/well were added to a 96-well plate grown for 24 h at37 �C in a 5% CO2 incubator. After 24 h, growthmediawas aspirated.Then, 100 mL of culture media containing 3.5 mM of r-Moj-D_peptide were added to their respective wells. Negative controlsconsisted of cells treated with equal volume of PBS. After 24 h allwells were then treated with 10 mL of WST-1 cell proliferationassay reagent for 1 h. The same volume of culture medium andWST-1 reagent used to treat the cells was used as the control blankfor the microplate reader. The formation of the formazan productwas measured at 450 nm. The reference wavelength was 630 nm. Adual wavelength reading was used to obtain the results, where thereferencewavelength was subtracted from the reading wavelength,450e630 nm. Percentage of cell viability was calculated using thefollowing formula: [(Abs450 nme600 nm of treated O Abs450nme600 nm of untreated) x 100]-100. Experiments were per-formed in triplicate.

2.11. Chromatin condensation

TXNIP knocked down and control scrambled shRNA SK-Mel28 cells were grown on chambered slides. One and a half millioncells were grown in 2 mL of media for 24 h. After initial incubation,cells were treated with 5 mM r-Moj-DM peptides for 24 h. Themedia was aspirated and the cells washed twice with 1X PBS. Cellswere fixed on the slide with 3% formaldehyde for 1 h, followed bytwo washes with 1X PBS. Microtubules were stained with anti-tubulin IgG antibody (sc-23948) overnight at 4 �C with constantshaking. After this incubation, the slides were rinsed with warm 1XPBS, followed by rabbit anti-mouse IgG conjugated to FITC (sc-358946). Cells were mounted and their nuclei stained with Vecta-shield mounting medium with DAPI. Slides were viewed at 100�,oil immersion with a confocal Zeiss LSM 700 microscope.

3. Results

3.1. RNA sequencing and bioinformatics analyses of r-Moj-DMtreated SK-Mel-28 cells

FASTQ analysis was done to examine the quality and filtering ofthe raw sequencing reads. RNA sequencing quality control datashow that sequencing quality was good with all mean Phred scoresat base positions 1e100 > 30. Both samples mean per sequencequality scores were 38. Sequence comparisons between untreatedand treated cells were done in order to identify genes that weredifferentially expressed in r-Moj-DM treated cells. Density andvolcano plots were obtained to identify those sequences. Thesequencing coverage was calculated to be 8.33. RNA sequencingwas done at single reads of 100 bp with approximately 250 millionread pairs per sample.

3.2. Differential gene expression after r-Moj-DM treatment

Forty genes were differentially expressed after 6 h of r-Moj-DMtreatment (Table 1). Nine genes were upregulated and 31 weredownregulated. The expression of 13 of those genes was validatedby qRT-PCR (Fig. 1). CXCL1, IL-8, TXNIP, BirC3, and Hist1-H2BJ wereupregulated in both the RNA sequencing data and qRT-PCR. IGFBP3,Prune2, BNiP3, HLA-B, RGS1, PTPRF, TGFß-1, and PTPRD were down-regulated in both methods.

3.3. Apoptotic genes are upregulated as a result of r-Moj-DN butnot r-Moj-DA treatment

R-Moj-DN is a strong apoptotic inducer of SK-Mel-28 cells, whiler-Moj-DA is not (Ramos et al., 2016). Differential expression wasexamined by qRT-PCR of the genes BNiP3, IGFBP3, PTPRF, Prune2,TGFß-1, and TXNIP (Fig. 2). There were no differences in expressionin between cells treated with r-Moj-DN or r-Moj-DA early aftertreatment (1 he2 h). However, higher expression was detected forall genes examined in r-Moj-DN treated cells from 4 h to 6 h aftertreatment.

3.4. TXNIP protein localization dynamics after GST-Moj-DMtreatment

Confocal microscopy results from immunostaining of anti-GSTantibody staining of GST-Moj-DM treated cells suggest that GST-Moj-DM results in integrin clustering. In addition, anti-TXNIP an-tibodies suggest that some TXNIP protein is localized to the nucleiof treated cells (Fig. 3).

Table 1Differentially expressed genes between r-Moj- DM treated and control SK-Mel-28 cells at 6 h and identified with RNA-SEQ (n ¼ 40).

Gene Function Foldchange

Gene Function Foldchange

IL8 Proinflammatory heparin-binding protein, negativeregulation cell proliferation, locomotion

5.52 STC1 Glycoprotein, angiogenesis, bone and muscle development, cellularmetabolism

�3.70

BIRC3 Negative regulation of apoptosis 5.44 RGS1 Attenuates g-protein activity �3.72TXNIP Positive regulation of apoptosis, negative regulation cell

proliferation, redox regulation4.98 BNIP3 Positive regulation of apoptosis, cell projection �3.72

CXCL1 Melanoma Growth factor, negative regulation cellproliferation

4.93 C2orf72 Uncharacterized protein �3.83

HIST1H2AG Chromatin dynamics 3.37 FLNB Vascular injury repair �3.98HIST1H2BB Chromatin dynamics 3.34 FGFBP2 Plasma protein, cytotoxic lymphocyte- mediated immunity �4.15HIST1H2BJ Chromatin dynamics 3.08 DDX41 DEAD box protein family member, function not determined �4.41HIST1H2AH Chromatin dynamics 3.00 NRN1 Neuritogenesis, glutamate target gene �4.51HIST1H1B Chromatin dynamics 2.83 TGFB1 Positive regulation of apoptosis, growth factor, cell projection,

negative regulation cell proliferation, locomotion�5.01

A2M Extracellular glycoprotein, plasma protein, proteaseinhibitor

�2.78 PPFIA4 Tyrosine phosphatase, hypoxia-induced gene, cell-cell adhesion �5.25

ANKZF1 Ankyrin repeat and zinc finger domain- containingprotein 1, protein-protein interactions

�2.81 LOXL2 Activation of TGFB1 signaling, disease- associated fibroblasts, andgrowth factors

�5.61

TNS1 Focal adhesions, crosslinks actin filaments, Src homology2 (SH2) domain

�2.84 DOK3 Adaptor protein, TLR signaling, inhibits Ras-Erk pathway �5.66

IGFBP3 Positive regulation of apoptosis, negative regulation cellproliferation, locomotion

�2.94 PTPRD Tyrosine phosphatase �5.71

VGF Growth factor, cell projection �3.02 CACNA1H Voltage-gated calcium channel, cell projection �5.92PRUNE2 Positive regulation of apoptosis �3.07 C20orf46 Transmembrane protein �6.27NDRG1 Growth arrest, cell differentiation �3.25 ISM2 ECM protein �6.29HLA-B Immune system, membrane macroglobulin, antigen

presenting�3.43 SEC14L5 Secretory protein �6.33

RORA Nuclear hormone receptors �3.54 SLC8A2 Sodium-calcium exchanger �6.70PTPRF Positive regulation of apoptosis, negative regulation cell

proliferation�3.58 BEND5 Uncharacterized protein, BEN domain �7.08

EN02 Glycolysis, enolase �3.63 CA9 Carbonic anhydrase, cell projection �7.40

Fig. 1. Thirteen genes that were significantly differentially expressed in RNA sequencing experiment were validated using qRT-PCR. Cells for the qRT-PCR analysis were treated for6 h.

T.D. McBride et al. / Toxicon 121 (2016) 1e94

Author's Personal Copy

Fig. 2. (A) Gene expression of pro-apoptotic genes after r-Moj-DN treatment of SK-Mel-28 cells. (B) Gene expression of pro-apoptotic genes after r-Moj-DA treatment of SK-Mel-28 cells.

Fig. 3. Confocal Microscopy of GST-Moj-DM treated (top left and right panels) and untreated (bottom left and right panels) SK-Mel-28 cells. Nuclei were stained with DAPI. Ar-rowheads in the top anti-GST panel indicate integrin crosslinking. Arrowheads on the anti-TXNIP panel indicate chromatin condensation. The white arrow in the anti-TXNIP panelsuggest increased nuclear and cytoplasmic localization. Similar results were obtained when cells were treated with r-Moj-DM and stained with anti-TXNIP antibodies.

T.D. McBride et al. / Toxicon 121 (2016) 1e9 5

Author's Personal Copy

3.5. TXNIP knockdown

A stable knockdown of TXNP resulted in a significant (p < 0.001)reduction of TXNIP transcript (Fig. 4A) and protein (Fig. 4B) aftertreatment with r-Moj-DM for 24 h.

3.6. TXNIP is required for the apoptotic induction initiated by r-Moj-DM/integrin signaling

Cell viability experiments were performed to determine the role

of TXNIP in the apoptotic inducing activity originated from the r-Moj-DM/integrin pathway. High cell viability (74.3% ±9.1) was ob-tained after r-Moj-DM treatment of TXNIP knocked down SK-Mel-28 cells, compared to 49.8% ±0.58 for control shRNA cells and 34%±0.187 for untransduced cells (Fig. 5A).

Chromatin condensation, an apoptotic event, was observed us-ing confocal microscopy (Fig. 5B). After r-Moj-DM peptide treat-ment, chromatin condensation was observed in the scrambledshRNA control cells, bur not in the TXNIP stabled knocked down SK-Mel-28 cells.

Fig. 4. (A) TXNIP RNA levels are significantly reduced (p < 0.001) in TXNIP knocked down SK-Mel-28 cells. RNA levels were measured by QPCR of cells treated after 24 h with r-Moj-DM. (B) TXNIP protein levels are significantly reduced (p < 0.001) in TXNIP knocked downed SK-Mel-28 cells. Densitometry analyses were performed from western blots of cellstreated after 24 h with r-Moj-DM.

T.D. McBride et al. / Toxicon 121 (2016) 1e96

Author's Personal Copy

4. Discussion

4.1. Gene expression changes as a result of r-Moj-DM treatment ofSK-Mel-28 cells

Microarrays were used to examine gene expression changes inHUVECs after Crotalus atrox or Bothrops jararaca venom treatment(Gallagher et al., 2003). However, RNA sequencing demonstratingchanges in gene expression after cancer cell line treatment withsnake venom or purified toxins have not been reported. RNAsequencing has been used to determine snake venom tran-scriptome profiles of several species (Rokyta et al., 2013; Rokytaet al., 2012; Durban et al., 2011) and to examine molecular evolu-tion (Reeks et al., 2016; Aird et al., 2015; Junqueira-de-Azevedoet al., 2015; McGivern et al., 2014; Hargreaves et al., 2014; Tanakaet al., 2013; Moura-da-Silva et al., 2011).

Our gene expression analysis yielded several unexpected re-sults. First, genes associated to the TNFR (tumor necrosis factorreceptor) involved in the activation of extrinsic apoptosis orintegrin-induced apoptosis were not differentially expressed after

r-Moj-DM treatment. These targets include FAK, ERK, NF-kB, andJNK (Sharma et al., 2014; Ashkenazi and Dixit, 1998). Second, mostof the identified genes have not been studied extensively. Differ-entially expressed genes after r-Moj-DM treatment are groupedinto seven categories: pro-apoptotic (TXNIP, CXCL1, IGFBP3, Prune2,PTPRF, TGF-ß, BNiP3, and DOK3), survival (BIRC3), inflammation (IL-8), chromatin dynamics (HIST1H2AG, HIST1H2BB, HIST1H2BJ, HIS-T1H2AH, andHIST1H1B), tumormetastasis inhibition (NDRG1), focaladhesion (TNS1), and hypoxia-induced (PPIA4). There were 22genes with no known connection to apoptosis or survival.

Gene expression changes of pro-apoptotic and anti-apoptoticgenes after treatment can be confusing and sometimes contradic-tory. However, it has been proposed that the expression ratio be-tween pro-apoptotic and anti-apoptotic proteins may be moreimportant than the overexpression of particular pro-apoptoticproteins in the cellular decision to undergo apoptosis(Zhivotovsky and Orrenius, 2006). Two of the pro-apoptotic genesidentified in the RNA sequencing of r-Moj-DM treated SK-Mel-28 cells were upregulated (TXNIP, and CXCL1), while the rest in thiscategory were downregulated. In contrast, qRT-PCR experiments

Fig. 5. (A) Cell viability increased in r-Moj-DM treated TXNIP knocked down cells. The increase was significant (p < 0.001) compared to the cell viability obtained in r-Moj-D treatedscrambled shRNA control and treated untransduced SK-Mel-28 cells. (B) SK-Mel-28 cells that contained a stable TXNIP knockdown failed to show chromatin condensation after r-Moj-DM treatment. Cells with a stable scrambled shRNA control showed chromatin condensation after r-Moj-DM treatment. Chromatin was stained with DAPI and cells werestained with anti-tubulin antibodies. Representative cells are shown.

T.D. McBride et al. / Toxicon 121 (2016) 1e9 7

Author's Personal Copy

indicated that six pro-apoptotic genes (BNiP3, IGFBP3, PTPRF,Prune2, TGFß-1, and TXNIP) were upregulated after 4 h of r-Moj-DNtreatment. Similar results were not obtained with r-Moj-DA, asupregulation of those genes was not significantly high after 4 h.Ramos et al. (2016) demonstrated that in SK-Mel-28 cells, r-Moj-DNis a stronger apoptotic inducer than r-Moj-DM. Recombinant r-Moj-DA does not induce apoptosis in SK-Mel-28 cells (Ramos et al.,2016).

4.2. TXNIP localizes to the nucleus after r-Moj-DM treatment

For this study we focused on TXNIP, since it is a protein withdifferent cellular roles that intersect different aspects of cell deathinitiation (Zhou and Chng, 2013; Goldberg et al., 2003). TXNIP (alsoknown as VDUP and thioredoxin binding protein 2, or TBP-2) wasfirst identified as a vitamin D binding protein and can act as atranscription factor (Masutani et al., 2012; Han et al., 2003).Therefore, it is not surprising that TXNIP translocates to the nucleusafter pathway activation (Fig. 3). Our confocal experiments withGST-Moj-DM peptides suggest that Moj-DM binding to the integrinresults in integrin clustering, increase in TXNIP mRNA expression,and nuclear translocation of TXNIP (Fig. 3).

It is surprising that TXNIP expression was upregulated in r-Moj-DM treated cells as a result of an integrin-induced pathway. TXNIPupregulation has been detected after ER damage (Lerner et al.,2012; Oslowski et al., 2012), hypoxic conditions (Li et al., 2015),glucocorticoid incubation (Wang et al., 2006), ROS production (Gao

et al., 2015; Zhou and Chng, 2013), or inhibition of glucose uptakeand aerobic respiration (Shen et al., 2015).

4.3. Knockdown of TXNIP revealed its significance in the apoptoticinduction by r-Moj-DM mutant peptides

Stable knockdowns of TXNIP and scrambled shRNA controlswere used to examine the role of TXNIP in the apoptosis of SK-Mel-28 cells. In TXNIP knocked down SK-Mel-28 cells, cell survivalincreased significantly and chromatin condensation (a morpho-logical characteristic of apoptosis) was inhibited, suggesting thatTXNIP is involved in the apoptotic pathway triggered by r-Moj-DMupon antagonizing the av integrin subunit.

Increased survival of cancer cells has been associated withreduced levels of TXNIP (Zhou et al., 2011). In triple-negative breastcancers, c-myc inhibits TXNIP expression resulting in the increase ofcell survival and poor tumor prognosis (Shen et al., 2015).Conversely, high levels of TXNIP correlate to increased apoptosis.Apoptosis of neuroblastomas (Su et al., 2015) and the lung cancercell line A549 (Liu et al., 2015) was activated by ROS increase andupregulation of TNIXP expression.

4.4. Potential relationship between cell proliferation inhibition andTXNIP

TXNIP gene upregulated expression was also detected in r-Moj-DA treated cells. Recombinant r-Moj-DA is not an apoptotic inducer,

T.D. McBride et al. / Toxicon 121 (2016) 1e98

Author's Personal Copy

but a strong cell proliferation inhibitor of SK-Mel-28 cells (Ramoset al., 2016). This finding is not surprising since TXNIP has beenshown to be upregulated after signals that inhibit cell proliferation(Han et al., 2003). Furthermore, it has been proposed that TXNIPinhibits cell proliferation by inhibiting metabolic pathways that arenecessary for cell division (Elgort et al., 2010).

4.5. Conclusions

Our study is the first demonstration of TXNIP gene upregulationinduced by a disintegrin/integrin pathway. The qRT-PCR resultsfrom gene expression profile of r-Moj-DN-treated cells provide atestable hypothesis that may explain this disintegrin's apoptoticpotency. It would be of interest to identify the proteins and signaltransduction pathway(s) that transmit the signal from the integrinreceptor and results in the upregulation of TXNIP, and ultimatelyapoptosis. Our study was limited to gene expression changes at ashort time range (1e6 h). It would be of interest to identify genetargets at later times (8e24 h). Finally, other candidate genesidentified from our study are promising. These include CXCL1,BIRC3, and IL-8. Functional analysis of those gene targets and pro-tein expression levels are possibilities for future studies. Finally, ourstudy focused on gene expression changes and perhaps phos-phorylation/inhibition of phosphorylation events of known pro-apoptotic or survival signals are as important as TXNIP ininducing SK-Mel-28 apoptosis induction by r-Moj-DM. These pos-sibilities remained to be studied.

Ethical statement

Our manuscript has not been published and is not underconsideration for publication in other journals. All co-authors in themanuscript carefully read the manuscript, carried-out the experi-ments, and agreed on the content presented in the manuscript.

Conflict of interest statement

The authors declare that there is no conflict of interest.

Acknowledgments

We thank Stephanie Mandal for the preparation and labeling ofcomposite figures. Funding for this project was provided by NSF-REU Grant #DBI1004350, NIH Grant #R25GM071381, NIH Grant#5R25GM71381, and NIH Grant #5T34GM008253-27. The writingof this manuscript was supported while serving at the NationalScience Foundation. Any opinion, findings, and conclusions orrecommendations expressed in this material are those of the au-thor(s) and do not necessarily reflect the views of the NationalScience Foundation.

Transparency document

Transparency document related to this article can be foundonline at http://dx.doi.org/10.1016/j.toxicon.2016.08.014.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.toxicon.2016.08.014.

References

Aird, S.D., Aggarwal, S., Villar-Briones, A., Tin, M.M., Terada, K., Mikheyev, A.S., 2015.Snake venoms are integrated systems, but abundant venom proteins evolve

more rapidly. BMC Genomics. http://dx.doi.org/10.1186/s12864-015-1832-6.Alimenti, E., Tafuri, S., Scibelli, A., D'Angelo, D., Manna, L., Pavone, L.M.,

Belisario, M.A., Staiano, N.N., 2004. Pro- apoptotic signaling pathway activatedby echistatin in GD25 cells. Biochim. Biophys. Acta 1693, 73e80.

Ashkenazi, Dixit, V.M., 1998. Death receptors: signaling and modulation. Science281, 1305e1308.

Barja-Fidalgo, C., Coelho, A.L., Saldanha-Gama, R., Helal-Neto, E., Oliveira, A.M.,Freitas, M.S., 2005. Disintegrins: integrin selective ligands which activateintegrin-coupled signaling and modulate leukocyte functions. Braz. J. Med. Biol.Res. 38, 1513e1520.

Bhupender, S., Sukhdev, S., Kanwar, S.S., 2014. Methionase: a therapeutic enzyme totreat malignancies. BioMed Res. Int. http://dx.doi.org/10.1155/2014/506287.

Brassard, D.L., Maxwell, E., Malkowski, M., Nagabhushan, T.L., Kumar, C.C.,Armstrong, L., 1999. Integrin alpha(v)beta(3)-mediated activation of apoptosis.Exp. Cell. Res. 251, 33e45.

Calderon, L.A., Sobrinho, J.C., Zaqueo, K.D., de Moura, A.A., Grabner, A.N.,Mazzi, M.V., Marcussi, S., Nomizo, A., Fernandes, C.F.C., Zuliani, J.P.,Carvalho, B.M.A., da Silva, S.L., St�abeli, R.G., Soares, A.M., 2014. Antitumoralactivity of snake venom proteins: new trends in cancer therapy. BioMed Res.Int. http://dx.doi.org/10.1155/2014/203639.

Carey, C., Bueno, R., Gutierrez, D.A., Petro, C., Lucena, S.A., Sanchez, E.E., Soto, J.G.,2012. Recombinant rubistatin (r-Rub), an MVD disintegrin, inhibits cell migra-tion, proliferation, and it is a strong apoptotic inducer of SK-Mel-28. Toxicon 59,241e248.

Chiarugi, P., Giannoni, E., 2008. Anoikis: a necessary death program for anchorage-dependent cells. Biochem. Pharmacol. 76, 1352e1364.

Durban, J., Ju�arez, P., Angulo, Y., Lomonte, B., Flores-Diaz, M., Alape-Gir�on, A.,Sasa, M., Sanz, L., Guti�errez, J.M., Dopazo, J., Conesa, A., Calvete, J.J., 2011.Profiling the venom gland transcriptomes of Costa Rican snakes by 454 pyro-sequencing. BMC Genomics. http://dx.doi.org/10.1186/1471-2164-12-259.

Elgort, M.G., O'Shea, J.M., Jiang, Y., Ayer, D.E., 2010. Transcriptional and translationaldownregulation of thioredoxin interacting protein is required for metabolicreprogramming during G1. Genes & Cancer 1, 893e907.

Gallagher, P.G., Bao, Y., Serrano, S.M., Kamiguti, A.S., Theakston, R.D., Fox, J.W., 2003.Use of microarrays for investigating the subtoxic effects of snake venoms: in-sights into venom-induced apoptosis in human umbilical vein endothelial cells.Toxicon 41, 429e440.

Gao, K., Chi, Y., Zhang, X., Zhang, H., Li, G., Sun, W., Takeda, M., Yao, J., 2015. A novelTXNIP-based mechanism for Cx43-mediated regulation of oxidative drug injury.J. Cell Mol. Med. 19, 2469e2480.

Gilmore, A.P., 2005. Anoikis. Cell Death Differ. 12, 1473e1477.Goldberg, S.F., Miele, M.E., Hatta, N., Takata, M., Paquette-Straub, C., Freedman, L.P.,

Welch, D.R., 2003. Melanoma metastasis suppression by chromosome 6: evi-dence for a pathway regulated by CRSP3 and TXNIP. Cancer Res. 63, 432e440.

Han, S.H., Jeon, J.H., Ju, H.R., Jung, U., Kim, K.Y., Yoo, H.S., Lee, Y.H., Song, K.S.,Hwang, H.M., Na, Y.S., Yang, Y., Lee, K.N., Choi, I., 2003. VDUP1 upregulated byTGF-b1 and 1,25-dihydoryvitamin D3 inhibits tumor cell growth by blockingcell cycle progression. Oncogene 22, 4035e4046.

Hanahhan, D., Weinberg, R.A., 2011. Hallmarks of cancer: the next generation. Cell144, 646e674.

Hargreaves, A.D., Swain, M.T., Logan, D.W., Mulley, J.F., 2014. Testing the Toxicofera:comparative transcriptomics casts doubt on the single, early evolution of thereptile venom systems. Toxicon. http://dx.doi.org/10.1016/j.toxicon.2014.10.004.

Hong, S., Lee, H., You, W., Chung, K., Kim, D., Song, K., 2003. The snake venomdisintegrin salmosin induces apoptosis by disassembly of focal adhesions inbovine capillary endothelial cells. Biochem. Biophys. Res. Commun. 302,502e508.

Hsu, C.C., Chuang, W.J., Chung, C.H., Chang, C.H., Peng, H.C., Huang, T.F., 2016. Snakevenom disintegrin inhibits the activation of Toll-like receptors and alleviatessepsis through integrin alphavbeta3 blockade. Sci. Rep. http://dx.doi.org/10.1038/srep23387.

Hsu, C.C., Chuang, W.J., Chang, C.H., Tseng, Y.L., Peng, H.C., Huang, T.F., 2010. Im-provements in endotoxemic syndromes using a disintegrin, rhodostomin,through integrin avb3dependent pathway. J. Thromb. Haemost. 9, 593e602.

Huveneers, S., Truong, H., Danen, H.J., 2007. Integrins: signaling, disease, andtherapy. Int. J. Radiat. Biol. 83, 743e751.

Jain, D., Kumar, S., 2012. Snake venom: a potent anticancer agent. Asian Pac. J.Cancer Prev. 13, 4855e4860.

Junqueira-de-Azevedo, I.L.M., Mancini, C., Bastos, V., Ho, P.L., Luna, M.S.,Yamanouye, N., Casewell, N.R., 2015. Venom-related transcripts from Bothropsjararaca tissues provide novel molecular insights into the production andevolution of snake venom. Mol. Biol. Evol. 32, 754e766.

Kim, D.S., Jang, Y.J., Jeon, O.H., Kim, D.S., 2007. Saxatilin inhibits TNF-alpha-inducedproliferation by suppressing AP-1-dependent IL-8 expression in the ovariancancer cell line MDAH 2774. Mol. Immuno 44, 1409e1416.

Langmead, B., Trapnell, C., Pop, M., Salzberg, S.L., 2009. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. GenomeBiol. 10, R25.

Lerner, A.G., Upton, J.P., Praveen, P.V.K., Gosh, R., Nakagawa, Y., Igbaria, A., Shen, S.,Nguyen, V., Backes, B.J., Helman, M., Heintz, N., Greengard, P., Hui, S., Tang, Q.,Trusina, A., Oakes, S.A., Papa, F.R., 2012. IRE1a induces thioredoxin-interactingprotein to activate the NLRP3 inflammasome and promote cell death underirremediable ER stress. Cell Metab. 16, 250e264.

Ji, S., Qin, Y., Liang, C., Huang, R., Shi, S., Liu, J., Jin, K., Liang, D., Xu, W., Zhang, B.,Liu, L., Liu, C., Xu, J., Chiao, P.J., Li, M., Yu, X., 2016. FBW7 (F-box and WD repeat

T.D. McBride et al. / Toxicon 121 (2016) 1e9 9

Author's Personal Copy

domain-containing 7) negatively regulates glucose metabolism by targeting thec-myc/TXNIP (thioredoxin binding protein) axis in pancreatic cancer. Clin.Cancer Res. http://dx.doi.org/10.1158/1078-0432.CCR-15-2380.

Li, Y., Miao, L.Y., Huang, M., Yu, M., Meng, K., Cal, H.R., 2015. Hypoxia induced highexpression of thioredoxin interacting protein (TXNIP) in non-small cell lungcancer and its prognostic effect. Asian Pac J. Cancer Prev. 16, 2953e2958.

Liu, W., Gu, J., Qi, J., Zeng, X.N., Ji, J., Chen, Z.Z., Sun, X.L., 2015. Lentinan exertssynergistic apoptotic effects with pacilitaxel in A549 cells via activating ROS-TXNIP-NLRP3 inflammasome. J. Cell Mol. Med. 19, 1949e1955.

Lockshin, R.A., Zakeri, Z., 2004. Apoptosis, autophagy, and more. Int. J. Biochem. CellBiol. 36, 2405e2419.

Lucena, S., Sanchez, E.E., Perez, J.C., 2011. Anti-metastatic activity of the recombi-nant disintegrin, r-mojastin 1, from the Mohave rattlesnake. Toxicon 57,794e802.

Masutani, H., Yoshihara, E., Masaki, S., Chen, Z., Yodoi, J., 2012. Thioredoxin bindingprotein (TBP)-2/TXNIP and a-arrestin proteins in cancer and diabetes mellitus.J. Clin. Biochem. Nutr. 50, 23e34.

McGivern, J.J., Wray, K.P., Margres, M.J., Couch, M.E., Mackessy, S.P., Rokyta, D.R.,2014. RNA-seq and high-definition mass spectrometry reveal the complex anddivergent venoms of two rear-fanged colubrid snakes. BMC Genomics. http://dx.doi.org/10.1186/1471-2164-15-1061.

Meng, X.W., Lee, S.H., Kaufmann, S.H., 2006. Apoptosis in the treatment of cancer: apromise kept? Curr. Opin. Cell Biol. 18, 668e676.

Minea, R.O., Helchowski, C.M., Zidovetzki, S.J., Costa, F.K., Swenson, S.D.,Markland, F.S., 2010. Vicrostatin e an anti-invasive multi-integrin targetingchimeric disintegrin with tumor anti-angiogenic and pro-apoptotic activities.PLoS ONE 5 (6), e10929. http://dx.doi.org/10.1371/journal.pone.0010929.

Moura-da-Silva, A.M., Furlan, M.S., Caporrino, M.C., Grego, K.F., Portes-Junior, J.A.,Clissa, P.B., Valente, R.H., Magalh~aes, G.S., 2011. Diversity of metalloproteinasesin Bothrops neuwiedi snake venom transcripts: evidence for recombinationbetween different classes of SVMPs. BMC Genet. http://dx.doi.org/10.1186/1471-2156-12-94.

Oslowski, C.M., Hara, T., O'Sullivan-Murphy, B., Kanekura, K., Lu, S., Hara, M.,Ishigaki, S., Zhu, L.J., Hayashi, E., Hui, S.T., Greiner, D., Kaufman, R.J., Bortell, R.,Urano, F., 2012. Thioredoxin-interacting protein mediates ER stress-induced ßcell death through initiation of the inflammasome. Cell Metab. 16, 266e273.

Park, M.H., Jo, M.R., Won, D., Sueb, S., Han, S.B., Song, M.J., Tae, J., 2012. Snake venomtoxin from vipera lebetina turanicainduces apoptosis of colon cancer cells viaupregulation of ROS- and JNK-mediated death receptor expression. BMC Cancer.http://dx.doi.org/10.1186/1471-2407-12-228.

Ramos, C.J., Gutierrez, D.A., Aranda, A.S., Koshlaychuk, M.A., Carrillo, D.A.,Medrano, R., McBride, T.D., U A, Medina, S.M., Lombardo, M.C., Lucena, S.E.,Sanchez, E.E., Soto, J.G., 2016. Functional characterization of six aspartate (D)recombinant mojastin mutants (r-Moj): a second aspartate amino acid carboxylto the RGD in r-Moj-D_ peptide is not sufficient to induce apoptosis in SK-Mel-28 cells. Toxicon 118, 36e42.

Reeks, T., Lavergne, V., Sunagar, K., Jones, A., Undheim, E., Dunstan, N., Fry, B.,Alewood, P.F., 2016. Deep venomics of the Pseudonaja genus reveals inter- andintra-specific variation. J. Proteomics. http://dx.doi.org/10.1016/j.jprot.2015.11.019.

Ren, A., Wang, S., Cai, W., Yang, G., Zhu, Y., Wu, X., Zhang, Y., 2006. Agkistin-s, adisintegrin domain, inhibits angiogenesis and induces BAECs apoptosis. J. CellBiochem. 99, 1517e1523.

Ribeiro, L.C.A., Massimino, L.C., Durante, A.C., Tansini, A., Urbaczek, A., Selistre-deAraujo, H.S., Carlos, I.Z., 2014. Recombinant disintegrin targets a(v) b(3) integrinand leads to mediator production. Cell Adhesion Migr. 8, 60e65.

Roberts, A., Pimentel, H., Trapnell, C., Pachter, L., 2011. Identification of noveltranscripts in annotated genomes using RNA-Seq. Bioinformatics. http://

dx.doi.org/10.1093/bioinformatics/btr355.Rokyta, D.R., Lemmon, A.R., Margres, M.J., Aronow, K., 2012. The venom gland

transcriptome of the eastern diamondback rattlesnake (Crotalus adamanteus).BMC Genomics. http://dx.doi.org/10.1186/1471-2164-13-312.

Rokyta, D.R., Wray, K.P., Margress, M.J., 2013. The genesis of an exceptionally lethalvenom in the timber rattlesnake (Crotalus horridus) revealed throughcomparative venom-gland transcriptomics. BMC Genomics 14. http://dx.doi.org/10.1186/1471-2164-14-394.

Sanchez, E.E., Galan, J.A., Russell, W.K., Soto, J.G., Russell, D.H., Perez, J.C., 2006.Isolation and characterization of two disintegrins inhibiting ADP-induced hu-man platelet aggregation from the venom of Crotalus scutulatus scutulatus(Mohave Rattlesnake). Toxicol. Appl. Pharmacol. 212, 59e68.

Selistre-de-Araujo, H.S., Pontes, C.L.S., Montenegro, C.F., Martin, A.C.B.M., 2010.Snake venom disintegrins and cell migration. Toxins 2, 2606e2621.

Seoane, A.I., Tran, V.L., Sanchez, E.E., White, S.A., Choi, J.L., Gayt�an, B., Chavez, N.,Reyes, S.R., Ramos, C.J., Tran, L.H., Lucena, S.E., Sugarrek, M., Perez, J.C.,Mandal, S.A., Ghorab, S., Rodriguez-Acosta, A., Fung, B.K., Soto, J.G., 2010. Themojastin mutant Moj-DM induces apoptosis of the human melanoma Sk-Mel-28, but not the mutant Moj-NN nor the non-mutated recombinant Moj-WN.Toxicon 56, 391e401.

Shen, L., O'Shea, J.M., Kaadige, M.R., Cunha, S., Wilde, B.R., Cohen, A.L., Welm, A.L.,Ayer, D.E., 2015. Metabolic reprogramming in triple-negative breast cancerthrough Myc suppression of TXNIP. PNAS 112, 5425e5430.

Soto, J.G., White, S.A., Reyes, S.R., Regalado, R., Sanchez, E.E., Perez, J.C., 2007. Mo-lecular evolution of PIII-SVMP and RGD disintegrin genes from the genus Cro-talus. Gene 389, 66e72.

Su, C., Shi, A., Cao, G., Tao, T., Chen, R., Hu, Z., Shen, Z., Tao, H., Hu, D., Bao, J., 2015.Fenofibrate suppressed proliferation and migration of human neuroblastomacells via oxidative stress dependent of TXNIP. Biochem. Biophys. Res. Commun.460, 983e988.

Tanaka, Y., Oyama, S., Hori, S., Ushio, K., Shioj, N., Tarada, S., Deshimaru, M., 2013.Accelerated evolution of fetuin family proteins in Protobothrops flavoviridis(habu snake) serum and the discovery of an L1-like genomic element in theintronic sequence of a fetuin-encoding gene. Biosci. Biotechnol. Biochem. 77,582e590.

Trapnell, C., Pachter, L., Salzberg, S.L., 2009. TopHat: discovering splice junctionswith RNA-Seq. Bioinformatics 25, 1105e1111.

Wang, Z., Rong, Y.P., Malone, M.H., Davis, M.C., Zhong, F., Distelhorst, 2006. Thio-redoxin-interacting protein (TXNIP) is a glucocorticoid-regulated primaryresponse gene involved in mediating glucocorticoid-induced apoptosis. Onco-gene 25, 1903e1913.

Wu, W.B., Peng, H.C., Huang, T.F., 2003. Disintegrin causes proteolysis of beta-catenin and apoptosis of endothelial cells. Involvement of cellecell andcelleECM interactions in regulating cell viability. Exp. Cell Res. 286, 115e127.

Yeh, C.H., Peng, H.C., Huang, T.F., 1998. Accutin, a new disintegrin, inhibits angio-genesis in vitro and in vivo by acting as integrin alphavbeta3 antagonist andinducing apoptosis. Blood 92, 3268e3276.

Zhivotovsky, B., Orrenius, S., 2006. Carcinogenesis and apoptosis: paradigms andparadoxes. Carcinogenesis 27, 1939e1945.

Zhou, J., Chng, W.J., 2013. Roles of thioredoxin binding proteins (TXNIP) in oxidativestress, apoptosis and cancer. Mitochondrion 13, 163e169.

Zhou, Q., Nakada, M.T., Arnold, C., Shieh, K.Y., Markland, F.S., 1999. Contortrostatin, adimeric disintegrin from Agkistrodon contortrix contortrix, inhibits angiogen-esis. Angiogenesis 3, 259e269.

Zhou, J., Yu, Q., Chng, W.J., 2011. TXNIP (VDUP-1, TBP-2): a major redox regulatorcommonly suppressed in cancer epigenetic mechanisms. Int. J. Biochem. CellBiol. 43, 1668e1673.