Chemically Modified Synthetic microRNA-205 Inhibits the ... - Cell Press

original article

© The American Society of Gene & Cell Therapy

Chemically Modified Synthetic microRNA-205 Inhibits the Growth of Melanoma Cells In Vitro and In Vivo Shunsuke Noguchi1, Junya Iwasaki1,2, Minami Kumazaki1,2, Takashi Mori3,4, Kohji Maruo3,4, Hiroki Sakai 4,5, Nami Yamada6, Kazuyuki Shimada7, Tomoki Naoe7, Yukio Kitade2 and Yukihiro Akao1,4 United Graduate School of Drug Discovery and Medical Information Sciences, Gifu University, Gifu, Japan; 2Graduate School of Engineering, Gifu University, Gifu, Japan; 3Department of Veterinary Clinical Oncology, Faculty of Applied Biological Sciences, Gifu University, Gifu, Japan; 4Comparative Cancer Center, Gifu University, Gifu, Japan; 5Department of Veterinary Pathology, Faculty of Applied Biological Sciences, Gifu University, Gifu, Japan; 6 United Graduate School of Veterinary Medical Sciences, Gifu University, Gifu, Japan; 7Graduate School of Medicine, Nagoya University, Nagoya, Japan 1

microRNA (miR)-205 is downregulated and acts as a tumor suppressor in human melanoma cells. Previously, for clinical application, we added aromatic benzene- pyridine (BP-type) analogs to the 3′-overhang region of the RNA-strand and changed the sequences of the passenger strand in the miR-143 duplex. Here, we demonstrated the antitumor effect in vitro and in vivo of miR-205 that was also chemically modified by BP and had altered passenger sequence. In in vitro experiments, transfection with the synthetic miR-205 (miR-205BP/S3) significantly inhibited the growth of human melanoma cells. Exogenous miR-205BP/S3 suppressed the protein expression levels of E2F1 and VEGF, which are validated targets of miR-205-5p, and BCL2, a transcribed molecule of E2F1, as did Pre-miR-205, used as a miR-205 mimic having the wild-type sequence. On the basis of the results of a luciferase activity assay, miR-205BP/S3 directly targeted E2F1, as did Pre-miR-205. However, miR-205BP/S3 was much more resistant to RNase than Pre-miR-205 in fetal bovine serum and to RNase in mice xenografted with human melanoma tissues. In addition, the intratumoral injection of miR-205BP/S3 exhibited a significant antitumor effect compared with the case of control miRNA or Pre-miR-205 in human melanoma cell-xenografted mice. These findings indicate that miR-205BP/S3 is a possible promising therapeutic modality for melanoma. Received 22 December 2012; accepted 20 March 2013; advance online publication 30 April 2013. doi:10.1038/mt.2013.70

INTRODUCTION

Malignant melanoma is one of the most common skin cancers in humans.1 The incidence of this cancer continues to rise more rapidly than that of all other malignancies, except for lung cancer.2 Malignant melanoma is a spontaneous, highly aggressive neoplasm that can readily metastasize to other organs. In recent studies on it, both MAPK and PI3K/AKT signaling pathways have been a matter of focus, the signaling of which are known to be

constitutively activated through multiple mechanisms.3 BRAF inhibitors, MEK inhibitors, and ipilimumab (an anti-CTLA-4 monoclonal antibody) are promising new therapies for human malignant melanoma.4 However, safer and more effective therapies are to be desired. microRNAs (miRNA; miR) have emerged recently as a large group of short (18–25 nucleotides), noncoding, small RNA molecules that negatively regulate gene expression.5–7 Over 1,300 miRNAs are predicted to exist in humans (miRBase; http://www. mirbase.org/). Nevertheless, the latest study indicates that these previous estimates of conserved miRNAs might be inflated.8 Although the specific biological functions of most miRNAs remain largely unknown, these molecules are believed to constitute a large gene regulatory network that can modulate the expression of up to 30% of total cellular proteins. There is increasing experimental evidence supporting the role of miRNAs in the regulation of a wide range of physiological or pathophysiological responses, including development,9 cellular apoptosis,10 differentiation,11 cell proliferation,12 and cancer.13–18 Several studies by us and other groups have shown that miR205 is downregulated and acts as an antioncomir by targeting E2F1, ZEB2, and ERBB3 in human and canine melanomas.19–22 It has been suggested that the downregulation of miR-205 in human melanoma is associated with the loss of chromosome 1q32, on which miR-205 is located.22 Also, Liu et al. reported that miR-205 regulates the migration of melanoma cells and that the expression level of miR-205 is associated with melanoma progression.21 Thus, miR-205 is one of the most important antioncomirs in the case of canine and human melanomas. Accordingly, we consider the application of miR-205 for melanoma treatment to be reasonable and promising. However, for clinical application, the problem of degradation by RNA nuclease must be solved, and more efficient modification of miRNAs and their drug delivery system must be devised. For replacement therapy using antioncomirs, retroviral expression vectors are mainly evaluated.23 On the other hand, we recently showed that miRNA with chemical modification at its 3′-overhang portion acquires resistance to RNase and that when its passenger sequence is altered the miRNA exhibits a greater

Correspondence: Shunsuke Noguchi, 1-1 Yanagido, Gifu 501–1193, Japan. E-mail: [email protected]

1204

www.moleculartherapy.org vol. 21 no. 6, 1204–1211 june 2013


Synthetic microRNA-205 Suppresses Melanoma Growth

tumor-suppressive effect than unaltered wild-type miRNA because of increased affinity for Argonaute 2 (Ago2) protein.24–26 In this study, we showed that the addition of an aromatic benzene-pyridine (BP) analog to the 3′-overhang region of the double-stranded miR-205 and alteration of the passenger sequence (miR-205BPs) resulted in resistance to RNase in vitro and long-term stability in vivo compared with those properties of Pre-miR-205 purchased from Applied Biosystems (Foster City, CA). Furthermore, one of the synthetic miR-205BPs showed a significant tumor-suppressive effect on mice xenografted with human melanoma cells in vivo. Here, we discussed the tumor-suppressive mechanism of the synthetic miR-205BP and differences between

Pre-miR-205 and the synthetic miR-205BP. These findings suggest that the synthetic miR-205BP may have a potential therapeutic effectiveness against melanoma.

RESULTS Synthetic miR-205BP/S3 showed the greatest growth-suppressive effect on melanoma cells among eight different synthetic miR-205BPs

We synthesized eight different types of miR-205BPs having different structures of the double strand for the experiment on cell viability (Figure 1). The structure of BP is shown in Supplementary Figure S1.

Control miRNA 5′-

-BP-3′

3′-

-5′

Mismatched portion miR-205/BP-mm1,2,3,4,5,6 (miR-205BP/S1)

Wild-type miR-205 (miR-205/S1) u

uc

c

g

u

uc

g

c

g

g

5′-

cuuca

uccac

g

a

ucug

5′-

cuuca

uccac

g

a

3′-

gaagu

aggug

c

u

agac

3′-BP

gaagu

aggug

c

u

g

uu

a

u

miR-205/BP-mm1,2,3,5 (miR-205BP/S2) u g uc 5′agucug -BP uccaccg cuuca agguggc

gaagu

3′-BP

g

uu

agguggccucagac

uc

3′-BP

aagu

agguggccucagac

cuucauuccaccggagucug- BP

5′-

gaaguaagguggccucagac

3′-BP

g

uu miR-205/BP-mm3,4,5,6 (miR-205BP/S7) c g g u 5′- uccuuca uccac g a ucug -BP

miR-205/BP-mm0 (miR-205BP/S6) uccuucauuccaccggagucug -BP

3′-BP aggaagu

aggaaguaagguggccucagac

miR-205/BP-mm3,4,5,6 (miR-205BP/S8) c g 5′-

agac u

miR-205/BP-mm1,2 (miR-205BP/S5)

u uccaccggagucug -BP

5′-

u

g

uu

uuca

3′-BP

gaagu

3′-BP

ucagac u

5′uuc

a

miR-205/BP-mm1,2,3 (miR-205BP/S3) u uc 5′cuuca uccaccggagucug -BP

miR-205/BP-mm1,2,3 + 1 (miR-205BP/S4) ucc

g

uu

u

ucug -BP

uccuucauuccac

g

a

c

u

a

u

c a

u u

agac u

miR-205/-mm1,2,3 (miR-205/S3) u uc 5′cuuca uccaccggagucug

g

3′-BP aggaaguaaggug

aggug g

ucug -BP

gaagu

3′-

agac

uu

u

agguggccucagac g

Figure 1 Sequences of the variants of miR-205 with or without BP used in this study. The sequence of Pre-miR-205 is same as that of the wildtype miR-205 (miR-205/S1). 120

A2058

120 100

**

80 60

**

*

*

**

* **

40 20 0

Viable cell number (% of cont)

Viable cell number (% of cont)

140

Mewo

100

**

80

**

60 40 20

S8

S7

S6

S5

S4

S3

S2

S1

5 -2 0

Pre-miR-205

miR-205BP/S3

iR

Cont

Pr e-

m

C

on

t

0

Figure 2 Number of viable A2058 and Mewo cells transfected with control miRNA, Pre-miR-205, or each synthetic miR-205BP (10 nmol/l). Cell counts were performed at 96 (A2058) or 120 (Mewo) hours after the transfection. *P < 0.05, **P < 0.01. P values were determined for differences between the cells transfected with control miRNA and those transfected with Pre-miR-205 or each miR-205BP tested. Means + SD indicated by error bars are shown.

Molecular Therapy vol. 21 no. 6 june 2013

1205



A2058

nt

E2F1

Co

P

1.15

5

20

iR-

m re-

b

Mewo

3

P/S

5B

20 iR-

nt

5

20

iR-

m re-

m

Co

P

m

0.25

0.43

1.45

0.27

0.94

1.74

1.07

1.61

2.01

1.45

1.51

0.74

0.14

0.29

0.35

0.17

0.21

3

P/S

5B

20 iR-

Apoptotic cell number (%)

a

VEGF BCL2 PARP-1

-Proform 4.40 0.029

1.71 0.030

2.27 0.025

4.83

3.02

3.63

20 18 16 14 12 10 8 6 4 2 0

*

nt

Co

-Cleaved form

5

20

R-

mi

Pre

3

P/S

5B

20 iR-

m Treatment (concentration; 10 mmol/l)

β-actin

Figure 3 The effects of transfections with Pre-miR-205 or miR-205BP/S3 on the protein expression levels of target genes of miR-205 and the related genes. (a) Protein expression levels of validated miR-205 target genes, E2F1 and VEGF, downstream of E2F1, BCL2, and PARP-1 in A2058 and Mewo cells. The numerical values under the panels indicate the densitometric value (/β-actin). Extraction of protein was performed at 96 (A2058) or 120 (Mewo) hours after the transfection. (b) Number of apoptotic cells, which were identified by fragmentation or aggregation of nuclei under Hoechst33342 staining, in A2058 cells transfected with each miRNA. *P < 0.05. P values were determined for differences between the cells transfected with each miRNA. Hoechst33342 staining was performed at 96 hours after the transfection. Means + SD indicated by error bars are shown.

miR-205BP/S3 suppressed the expression levels of E2F1, VEGF, and BCL2 in melanoma cells and induced apoptosis in A2058 cells as did Pre-miR-205 As E2F1 and VEGF were earlier verified to be targets of miR-205-5p in melanoma and breast cancer cells,22,27 we validated that miR205BP/S3 suppressed the expression levels of these genes in melanoma. Expectedly, transfection with miR-205BP/S3 suppressed the protein expression levels of E2F1 and VEGF; although the effect of miR-205BP/S3 was less potent than that of Pre-miR-205 (Figure 3a). In addition, the expression level of antiapoptotic BCL2, which is a molecule transcribed downstream of E2F1, was also decreased by the treatment with Pre-miR-205 or miR-205BP/S3. It was reported previously that miR-205 induces apoptosis.22 Our data also revealed that the transfection with miR-205BP/S3 or Pre-miR-205 led to PARP-1 cleavage in A2058 cells, not but in Mewo cells. The number of apoptotic cells judged by Hoechst33342 staining was significantly increased by the treatment with Pre-miR-205 and tended to increase in the case of miR-205BP/S3 (Figure 3b). miR-205BP/S3 directly targeted E2F1 as did Pre-miR-205 To validate whether miR-205BP/S3 targeted the same gene as PremiR-205 did, we performed a luciferase activity assay for E2F1. Expectedly, compared with that of the control, the luciferase activity of the wild-type pMIR-E2F1 was significantly inhibited after the introduction of Pre-miR-205 or miR-205BP/S3 into A2058 1206

Position 285-291 of E2F1 3′ UTR 5′ hsa-miR-205

3′

Mut-E2F1 3′UTR -1 (Mut-1) Mut-E2F1 3′UTR -2 (Mut-2)

*

*

*

120 Luciferase activity (% of cont)

First, we examined the growth-suppressive effect of these synthetic miR-205BPs on melanoma cells to determine which type was the most effective. Among them, miR-205BP/S2, miR205BP/S3, miR-205BP/S5, miR-205BP/S6, miR-205BP/S7, and miR-205BP/S8 significantly suppressed the growth of A2058 cells; although Pre-miR-205 had the greatest effect of all miR-205s tested (Figure 2). miR-205BP/S3 showed the greatest growth suppression among the synthetic miR-205BPs. The growth-suppressive activity of miR-205BP/S3 was almost the same in Mewo cells as in A2058 cells (Figure 2).

100 80 60 40 20 0

nt

Co

5

20

iR-

-m Pre

R

mi

Wild

3

P/S

5B

-20

nt

Co

5

20

iR-

-m Pre

R

mi

Mut-1

3

P/S

5B

-20

nt

Co

5

20

iR-

-m Pre

R

mi

3

P/S

5B

-20

Mut-2

Figure 4 Luciferase activities after cotransfection with control miRNA, Pre-miR-205, or miR-205BP/S3 and each of the sensor vectors with the wild or mutated 3′-UTR of E2F1. The upper panel shows the regions of the 3′-UTR of human E2F1 mRNA complementary to mature miR-205. The red box indicates the predicted binding sites for miR-205. *P < 0.01. P values were determined for differences between the bracketed samples. Data are expressed as the means + SD indicated by the error bars.

cells (Figure 4). Mutations of the E2F1 3′-UTR-binding site (Mut-1 and Mut-2) markedly abolished the ability of miR-205BP/ S3 to regulate luciferase expression and tended to lessen that of Pre-miR-205 to do so. These results demonstrate that both PremiR-205 and miR-205BP/S3 directly targeted E2F1.

Silencing E2F1 inhibited the cell growth and induced apoptosis in A2058 cells E2F1 is one of the key molecules involved in cell proliferation and plays a central role in melanoma progression.28,29 Therefore, www.moleculartherapy.org vol. 21 no. 6 june 2013


b

120

c

100

Cont

80

E2F1

60

BCL2

*

40

siR-E2F1

PARP-1

-Proform

20 0 Cont siR-E2F1 Treatment (concentration; 10 nmol/l, 72 hours)

-Cleaved form β-actin

Apoptotic cell number (%)

viable cell number (% of cont)

a


18 16 14 12 10 8 6 4 2 0

*

Cont siR-E2F1 Treatment (concentration; 10 nmol/l)

Figure 5 The effect of E2F1 knockdown in A2058 cells. (a) Comparison of the number of viable cells between A2058 cells transfected with control miRNA and those treated with siR-E2F1. (b) Protein expression levels of E2F1 and the predicted signaling molecule downstream of E2F1, BCL2, and PARP-1 in A2058 cells transfected with control miRNA or siR-E2F1. (c) Number of apoptotic cells, which were identified by fragmentation or aggregation of nuclei by Hoechst33342 staining, among A2058 cells transfected with control miRNA or siR-E2F1. The assay was performed at 96 hours after the transfection. *P < 0.01. Means + SD indicated by error bars are given.

Tumor-suppressive effect of chemically modified synthetic miR-205BP (miR-205BP/S3) in vivo To examine the antitumor effect of miR-205 in vivo, we administered Pre-miR-205, miR-205BP/S3 or control miRNA by intratumoral injection into nude mice bearing A2058 tumors. As a result, at the second administration, a significant growth-suppressive effect was observed in the group injected with miR-205BP/S3 compared Molecular Therapy vol. 21 no. 6 june 2013

Remaining rate of miR-205s (0 minutes value as 100%)

miR-205BP/S3 was much more stable than PremiR-205 in fetal bovine serum and in vivo We showed earlier that the addition of an aromatic compound to the 3′-overhang region of the RNA-strand enhances its resistance against nuclease attack.24,26 Therefore, we validated that miR-205BP/S3 was also more resistant to RNase compared with Pre-miR-205. Consistent with our previous studies, the addition of BP to the 3′-overhang region of miR-205 (miR-205BP/S1 and miR-205BP/S3) significantly improved the resistance against RNA nucleases after 10 minutes compared with the resistance shown by miR-205/S1, miR-205/S3, and Pre-miR-205, which were unmodified by BP (Figure 6a). Importantly, the percentage of miR-205BP/ S3 remaining was markedly higher than that of miR-205BP/S1. In addition, we evaluated the stability of miR-205BP/S3 and Pre-miR-205 in vivo. As shown in Figure 6b, the absolute expression level of miR-205 at 24, 48, and 96 hours after the tumor had been injected with miR-205BP/S3 was significantly higher compared with that for the tumor injected with Pre-miR-205; although the expression level of miR-205 at 96 hours after the tumor injected with Pre-miR-205 was markedly higher (>10,000-fold) than that for the tumor injected with control miRNA (the absolute expression level of miR-205; 5.07 ± 1.02 attomol versus 0.18 ± 0.055 zeptomol). Based on these findings, we determined that the frequency of injection should be two times per week and therefore used this schedule in the following antitumor experiment using a mouse model.

a

140

Pre-miR-205

120

miR-205BP/S3 miR-205/S1

100

miR-205BP/S1 miR-205/S3

80 60

§*

40

*

20 0 0

b Absolute miR-205 expression level (pmol)

we confirmed that E2F1 contributes to the growth of A2058 cells by using siRNA. As a result, silencing E2F1 significantly inhibited the growth, consistent with suppression of the E2F1 expression level (Figure 5a,b). In addition, the expression level of BCL2 was decreased; and apoptosis, which was indicated by PARP-1 cleavage and Hoechst33342 staining, was also induced by silencing E2F1 (Figure 5b,c).

0

10 0

24

20 48

30 72

96

(Minutes)

(Hours)

0.1 0.01 Pre-miR-205 0.001 0.0001

miR-205BP/S3

*

*

0.00001 0.000001

*

Figure 6 Decay of the various variants of miR-205 in (a) fetal bovine serum or (b) in mice xenografted tumor tissues, evaluated by performing real-time reverse transcriptase PCR using the TaqMan miRNA assay. (a) The % of miR-205 remaining is expressed for each sample indicated in the figure. (b) The absolute expression level of miR-205 based on the standard curve method is expressed. *Significant difference between the cells transfected with Pre-miR-205-5p and *those with miR-205BP/S3, § those with miR-205/S3 and those with miR-205BP/S3, and ¶those with miR-205BP/S1 and those with miR-205/S1. *,§,¶P < 0.01. Means + SD indicated by error bars are given.

with the growth in the control miRNA-injected group (Figure 7a). Furthermore, at the third administration, miR-205BP/S3 showed a significant growth-suppressive effect compared with Pre-miR-205. In addition, the tumor weight at the end of the experiment was also significantly lower in the group injected with miR-205BP/S3 than in that treated with the control miRNA (Figure 7b). Importantly, PremiR-205 did not exhibit growth suppression of either tumor volume 1207



b

a

0

7

14

21 days S

0.6 0.4 0.2 0

¶

*

¶

0

7

10

14

17

c

0.6

2

VEGF BCL2 c-MYC Cyclin E/A etc.

0.4 0.3 0.2

*

0.1

Cell survival Cell proliferation

0

21 days Cont

1

E2F1

0.5

¶

*

§

*

0.7

Tumor weight (g)

Cont Pre-miR-205 miR-205BP/S3

0.8

miR-205

0.8

Pre-miR-205 miR-205BP/S3

Cont

Pre-miR-205

3

4

Relative expression level of VEGF

Tumor volume (cm3)

T

P = 0.1856

0.9

1.2 1

d

* 1

0.55

5

6

7

8

9

miR-205BP/S3 10

11

12

13

14

15

E2F1 VEGF BCL2

0.8 0.6 0.4 0.2 0.0

t

n Co

5 -20

3 P/S

Pre

R mi

B 05

m

2 iR-

Relative expression level of BCL2

Relative expression level of E2F1

β-actin

0.50 0.45 0.40 0.35 0.30

t

n Co

5 -20

3 P/S

R mi

Pre

B 05

2 iR-

m

0.5 0.4 0.3 0.2 0.1 0.0

nt

Co

5

20

iR-

-m Pre

3

P/S

5B

20 iR-

m

Figure 7 Tumor-suppressive effects of miR-205/S3 in mice bearing xenografted A2058 cells. (a) Time course of tumor size in mice injected with control miRNA, Pre-miR-205 or miR-205BP/S3. P values were determined for differences *between the tumors injected with control miRNA and those injected with miR-205/S3, ¶between the tumors injected with Pre-miR-205 and those injected with miR-205BP/S3 or §between the tumors injected control miRNA and those injected with Pre-miR-205. *,¶,§P < 0.05. (b) Comparison of weight of tumors removed from mice at the day of sacrifice, with miRNAs injected into the tumors indicated on the abscissa. *Significant difference between bracketed samples (*P < 0.05). Means + SD indicated by error bars are given. (c) Expression levels of E2F1, VEGF, and BCL2 in the tumors injected with each miRNA are shown in the upper panel. The graphs in the lower panel show the relative expression level of E2F1, VEGF, or BCL2 normalized to β-actin. Means ± SD indicated by error bars are given. (d) A flow diagram on the miR-205 on its target genes and the related pathways. S, sacrifice; T, transplantation; triangles, intratumoral injection.

or weight, although the tumor volume of the group injected with Pre-miR-205 was significantly smaller than that of control miRNA group at the second administration. In addition, the expression levels of E2F1 and BCL2 showed a tendency to decrease in the Pre-miR-205 and miR-205BP/S3 groups compared with their levels in the control group; although the effect was not significant (Figure 7c). However, if one of the samples (No. 1) was excluded, the expression level of E2F1 in the groups of Pre-miR-205 and miR-205BP/S3 was significantly lower than that in the control (control versus Pre-miR-205, P = 0.048; control versus miR-205BP/S3, P = 0.027). We also pathologically examined the tumor-suppressive effects on the tumor tissues. However, unfortunately, no significant findings such as apoptosis, necrosis or suppression of invasion by miR-205BP/S3 administration were observed compared with these findings for the other groups (Supplementary Figure S2). No adverse effects on the local skin or stroma around the tumor tissues were detected. 1208

DISCUSSION

We focused on the antitumor effects of miR-205 in melanoma, because the previous studies by us and others revealed that miR205 is downregulated in these tumors and functions as an antioncomir.20,22,30 In this study, we successfully showed that a synthetic miR-205, miR-205BP/S3, which was chemically modified by adding BP to the 3′ portion of the double strands and altering the passenger sequence, exhibited significant tumor-suppressive effects on human melanoma cells both in vitro and in vivo. In the in vitro experiments, however, the antitumor effect of miR-205BP/S3 was less potent than that of Pre-miR-205, which is a commercially available miR-205 mimic from Applied Biosystems. The addition of BP to wild-type miR-205 unfortunately resulted in loss of function of miR-205; because miR-205BP/S1, which was chemically modified by BP but not altered in its passenger sequence, could not exert the growth-suppressive effect in the melanoma www.moleculartherapy.org vol. 21 no. 6 june 2013


cells tested. Also, the suppressive effect of miR-205/S3, which was unmodified by BP, on the E2F1 expression level was greater than that of miR-205BP/S3, although the growth-suppressive effect of miR-205/S3 was almost equivalent to that of miR-205BP/S3 (Supplementary Figure S3). These results suggest that BP modification reduced the function of miR-205 and that the passenger sequence must be altered to rescue the function of the synthetic miR-205 chemically modified by BP. However, we could not find any relationship between the activity and structures of the double strands by changing the passenger sequence. Furthermore, the melting temperature of each synthetic miR-205BP also did not show any relationship with the degree of the growth-suppressive effect (data not shown). The antitumor mechanisms of miR-205BP/S3 were similar to those of Pre-miR-205, because both miRNAs targeted E2F1. These findings indicate that the guide strand of miR-205BP/S3 functioned like that of Pre-miR-205. It is reported that miR-205 overexpression in melanoma cells induces cellular senescence and apoptosis with caspase-3 activation and cytochrome c release.22 Consistent with that previous report, we also observed apoptosis in A2058 cells, but not in Mewo cells, by treatment with Pre-miR-205 or miR-205BP/S3. The apoptosis induction was associated with E2F1 downregulation as the siRNA experiment indicated. However, in this study, obvious cellular senescence, which was examined by senescence associated-β-galactosidase staining, and caspase-9 activation were not observed in A2058 and Mewo cells (data not shown). Accordingly, the mechanisms of tumor-suppressive effect including apoptosis and/or senescence induction depend on the type of melanoma cell lines examined. On the other hand, miR-205BP/S3 exhibited a greater tumorsuppressive effect than Pre-miR-205 in vivo, in spite of the smaller antitumor effects of miR-205BP/S3 in vitro. This difference could be related to resistance to RNase. miR-205BP/S3 was obviously more stable than Pre-miR-205 both in vitro and in vivo. The results shown in Figure 6a indicate that BP modification attenuated the degradation of RNA by exonuclease and that the change in double-stranded structure by alteration of the passenger sequence prevented the degradation by endonuclease. In the in vivo experiment, at the time of the second administration, both Pre-miR-205 and miR-205BP/S3 showed a significant tumor-suppressive effect compared with control miRNA. However, after the second administration, Pre-miR-205 did not exhibit a tumor-suppressive effect any more. RNase is abundant in a tumor tissue including tumor cells and stroma, and the amount of RNase may increase as a tumor tissue grows. We consider that this increasing amount of RNase would have afforded easy degradation of the Pre-miR-205. No differences in the degree of antitumor effects such as apoptosis, necrosis or the suppression of vascularization and invasion following each miRNA administration were pathologically observed in the treated tumor tissues (Supplementary Figure S2a). In addition, innate immune responses such as invasion by natural killer cells or lymphocytes by the treatment with exogenous miRNA were not observed. Also, the induction of interferons and interferon-stimulated genes such as OAS1 was not observed in the cells transfected with Pre-miR-205 or miR-205BP/S3 (Supplementary Figure S2b). The expression level of E2F1, a miR-205 target gene, was significantly decreased by the administration of Pre-miR-205 Molecular Therapy vol. 21 no. 6 june 2013


or miR-205BP/S3 if the No.1 sample in Figure 7c, which was injected with control miRNA, was excluded from the experiment. These findings indicate that miR-205BP/S3 also functioned like miR-205 in vivo. The E2F1 expression level of the sample from No. 1 was exceptionally low. The region used for protein extraction in the No.1 tumor tissue might have included much stroma or necrotic region compared with that used for other samples in the control group. In conclusion, miR-205 functioned as an antioncogenic miRNA by targeting E2F1 in human melanoma cells (Figure 7d), consistent with an earlier study.22 Furthermore, the synthetic miR205BP/S3 that was generated by chemical modification by adding BP to the 3′ region of the double strands and alteration of the passenger sequence acted like wild-type miR-205 and showed a tumor-suppressive effect on mice xenografted with human melanoma cells, which effect probably reflected resistance to RNase. This study suggests that miR-205BP/S3 can be a new therapeutic modality for the treatment of melanoma by local injection. In a further study based on the present results, we will consider the clinical application of miR-205BP/S3 for canine oral melanoma, in which miR-205 is downregulated and miR-205 acts as an antioncomir, as it does in human melanoma.20

MATERIALS AND METHODS

Cell culture and cell viability. Human malignant melanoma cell lines A2058

and Mewo were purchased from Health Science Research Resources Bank (Osaka, Japan), and the cells were maintained according to the manufacturer’s protocol. The number of viable cells was determined by performing the trypan blue dye exclusion test.

Cell transfection with miRNA or siRNA. A2058 or Mewo cells were seeded into six-well plates at the concentration of 0.5 × 105 cells per well the day before transfection. We used Pre-miR-205 (Applied Biosystems), which has the same sequences as wild-type miR-205 (miR-205/S1) and is generally available as a representative miR-205 mimic. Pre-miR-205 or synthetic miR-205s (Hokkaido System Sciences, Sapporo, Japan) with an aromatic BP analog added to their 3′-overhang region (miR-205BP; Supplementary Figure S1) and with the passenger sequence of the mismatched portion between passenger and guide strands changed to matched ones (miR-205BP/S) was used for the transfection of the cells. Transfection was achieved by using cationic liposomes, Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA) at a concentration of 10 nmol/l, according to the manufacturer’s Lipofection protocol. The sequences of the synthetic miR-205BPs used in this study were shown in Figure 1. All combinations to match each of the six mismatched portions are considered to be 720 combinations. To generate typical synthetic miR-205BPs, we chose eight different types; because there were too many for us to generate all types. In detail, we generated miR-205BP with non-altered sequences, one with all mismatched portions matched, and ones matched mainly mismatched portions at the 3′ or 5′ region of the guide strand. Short interfering RNA (siRNA) for E2F1 (10 nmol/l) was also used for transfection of A2058 cells. Its sequence was 5′-UCGGCACCUGAGAAGCCUCUUGAAA-3′ (siR-E2F1; Invitrogen). To determine a precise BP modified control miRNA, we examined the effects of two types of control miRNA on the cell growth and E2F1 expression. As shown in Supplementary Figure S3, the effects of the control miRNA designated as cont/19 mer, which was used in our previous studies,25,26 and another control miRNA termed cont/22 mer showed no obvious difference on the growth of A2058 cells. However, the transfection with cont/22 mer tended to decrease E2F1 expression level compared with cont/19 mer. Therefore, we decided to use cont/19 mer.

1209



Both of the control miRNAs were obtained from Hokkaido System Sciences. The sequences of the non-specific double-stranded control miRNAs used in this study are given in Figure 1. Western blotting. Total protein was extracted from whole cells by the

procedure described previously.25 Protein contents were measured with a DC Protein Assay Kit (Bio-Rad, Hercules, CA). Ten micrograms of lysate protein for Western blotting was separated by SDS-PAGE using polyacrylamide gels and electroblotted onto a PVDF membrane (PerkinElmer Life Sciences, Boston, MA). Details of the method used after blotting were described earlier.25 The antibodies used in this study were antihuman E2F1 mouse monoclonal antibody, antihuman VEGF mouse monoclonal antibody, antihuman BCL2 mouse monoclonal antibody, and antihuman PARP-1 mouse monoclonal antibody (Santa Cruz, Santa Cruz, CA), all of which were properly diluted with TBS-T containing 2% bovine serum albumin and 0.01% sodium azide. The loading control was prepared by reincubating the same membrane with antihuman β-actin antibody (Sigma, St Louis, MO).

Quantitative reverse transcriptase PCR using real-time PCR. Total RNA was isolated from cells by the phenol/guanidium thiocyanate method with DNase I treatment. To determine the expression of miRNAs, we used TaqMan MicroRNA Assay (hsa-miR-205; Applied Biosystems) to reverse transcribe the mature miRNA sequences to its cDNA. The PCR procedure was performed by real-time PCR. Briefly, after reverse transcription of 25 ng of total RNA, cDNA was generated. The PCR reaction consisted of 40 cycles (95 °C for 5 seconds, 60 °C for 30 seconds) after an initial denaturation step (95 °C for 10 seconds). The threshold cycle (Ct) was defined as the fractional cycle number at which the fluorescence passes a fixed threshold. The expression level of miR-205 in each sample was measured in terms of Ct value. The relative expression level of miR-205 was calculated by the ΔΔCt method, and the absolute expression level of miR-205 was calculated based on the standard curve of Pre-miR-205 or miR-205BP/S3. Hoechst33342 staining. For assessment of the morphological characteristics of apoptosis, A2058 cells were collected at 96 hours after the transfection. The cells were stained with Hoechst33342 (5 μg/ml) at 37 °C for 1 hour, washed once with phosphate-buffered saline, resuspended, pipetted dropwise onto a glass slide, and examined by fluorescence microscopy using an Olympus microscope (Tokyo, Japan) equipped with an epiilluminator and appropriate filters. The cells with condensed and/or fragmented nuclei stained with Hoechst33342 were assessed to be apoptotic, and the number of apoptotic cells among 1,000 cells was counted. Assay for luciferase activity. We constructed the sensor vector by joining

the regions with a possible binding site from the 3′-UTR of human E2F1 to a luciferase reporter pMIR-control vector (Applied Biosystems) to examine the target sequence recognized by miR-205-5p. For amplification of E2F1 mRNA, total RNA was reverse transcribed with a PrimeScript RT Reagent Kit (TaKaRa, Otsu, Japan). The sequences of the primers used in this study were as follow: E2F1-sense- 1996, TGTGCAATCAGGTGTCTCTC; E2F1antisense- 2350, TTCATCCAGAAGGCTGTGGA. Also, to generate the sensor vectors with 2 or 3 mutations in the binding site for miR-205-5p, we mutated seed regions from AUGAAGG to AUGCCGG (mt1) or AUUCCGG (mt2; PrimeSTAR Mutagenesis Basal Kit; TaKaRa). The sensor vectors with mutations were submitted to Life Science Research Center, Gifu University for DNA sequencing. The cells were seeded in 12-well plates at a concentration of 0.5 × 105/well the day before the transfection. The sensor vector (concentration; 0.5 μg/well) and 20 nmol/l Pre-miR-205, miR-205BP/S3 or non-specific control miRNA (Dharmacon, Tokyo, Japan) was used for the cotransfection of the cells by using cationic liposomes Lipofectamine RNAiMAX. Forty-eight hours after the cotransfection, luciferase activities were measured by using a Dual-Glo Luciferase Assay System (Promega, Madison, WI) according to the manufacturer’s protocol. Firefly luciferase activity was normalized to Renilla luciferase activity.

1210

Assay for stability of miRNA in vitro. To evaluate the stability of miRNA

in vitro, we additionally obtained miR-205/S1 and miR-205/S3 (Figure 1), which were unmodified by BP, from Hokkaido System Sciences. We incubated 20 pmol of Pre-miR-205, miR-205BP/S1, miR-205/S1, miR-205BP/ S3, or miR-205/S3 without Lipofectamine RNAiMAX in 100 μl of fetal bovine serum (Hyclone Laboratories, Logan, UT) at 37 °C for 0, 2, 10, 20, or 30 minutes. Then, total RNA was isolated; and real-time reverse transcriptase PCR using TaqMan microRNA assay was thereafter performed to quantify the expression level of miR-205, which was calculated by the ΔΔCt method.

In vivo tumor model and administration of the miR-205/liposome complexes. BALB/cSlc-nu/nu (nude) mice were obtained from Japan SLC

(Hamamatsu, Japan). A2058 cells were concentrated to 1 × 106 per 100 μl and injected s.c. into the back of each mouse. The tumor volume was calculated by the formula: 0.5236 L1 (L2)2, where L1 is the long axis and L2 is the short axis of the tumor. This formula was described in a previous report.31 First, to decide the frequency of administration of miRNA, at 7 days after inoculation, Pre-miR-205 or miR-205BP/S3 (0.1 nmol) in 50 μl of Opti-MEM (Invitrogen) was mixed with 2 μl of cationic liposomes (Lipofectamine RNAiMAX); and the mixture was injected into the tumors at one time. At 0, 24, 48, 72, or 96 hours after the injection, the mice were killed; and the whole tissues of transplanted tumors were harvested for total RNA extraction and subsequent evaluation of their miR-205 expression level. For evaluation of the antitumor effect of miR-205s in vivo, miR-205BP/ S3, Pre-miR-205, or control miRNA (0.1 nmol per 1 administration) in 50 μl of Opti-MEM was mixed with 2 μl of Lipofectamine RNAiMAX; and the mixture was injected into the tumors twice per week. Each group contained five mice. The evaluation of the tumors was performed after killing the mice at 21 days after transplantation of the tumor cells. The tissue sections from the tumors were used for pathological evaluation, total protein extraction, and total RNA extraction. Animal experimental protocols were approved by the Committee for Animal Research and Welfare of Gifu University. Statistics. Each examination was performed in triplicate. All calculated

data were compared by using Student’s t-test except in the case of the volume and weight of the tumors xenografted into mice, which data were compared by using Mann–Whitney U-test. A P value