Search of miRNAs critical for medulloblastoma formation using MiRaGE method

Search of miRNAs critical for medulloblastoma formation using MiRaGE method

○Y-h. Taguchi（Dept. Phys., Chuo Univ.）Jun Yasuda（Tohoku Univ.)

Present address:Cancer Research Inst. Ariake, Tokyo

2

Experiments　microRNA vs tumor

Target genes

Computer oriented prediction (uncertain)

genome

microRNA

microRNA

mRNA

mRNA

3

4

miRNA1 ○ × ○ ○ ○ ○ × ×miRNA2 ○ × ○ ○ × × ○ ○miRNA3 × ○ ○ × ○ ○ × ×miRNA4 ○ ○ ○ × ○ ○ × ×

Gen

e1G

ene2

Gen

e3G

ene4

Gen

e5G

ene6

Gen

e7G

ene8

miRNA target gene list (計算機による予測）simple seed match

VS

Murine Medulloblastoma(MB)Murine Medulloblastoma(MB) (Dr. Tetsuo Noda’ group(Dr. Tetsuo Noda’ group

(The JFCR-Cancer Institute)).(The JFCR-Cancer Institute)).

gene1gene2gene3

予測

miRNA1

5

Materials

P6P6＝6 days after birth, normal but growingP30P30＝30 days after birth, normal and not growingMBMB＝a few month after birth, malignant neoplasm30% of the Ptc1 +/- mice suffers from MB.

6

mRNA/miRNA expression byArray: AgilentatP6, P30 and MB

log(xg[mRNA/miRNA:MB or P6]) vs

log(xg[mRNA/miRNA:P30])

xg: mRNA/miRNA expression

Target gene list: simple seed match

7

log(xg[miRNA:P6/MB]) vs log(miRNA:xg[P30])

of considered miRNA(*)

(*) each miRNA is measured by multiple probe

t te s t fo r m iRNA e x p re s s io n

8

log(xg[mRNA:P6/MB]) – log(xg[mRNA:P30]) in target genes of considered miRNA

t test for miRNA target genes (MiRaGE method)

log(xg[mRNA:P6/MB]) – log(xg[mRNA:P30]) in target genes of

any of other miRNA

VS

9

P30 P6/MB

Gether the information of miRNA targets

Compare the expressions of targets for each miRNAs

Calculate False Discovery Rate

Generate ranking

10

miRNA Targets Down P-value FDR

miR-a 54 3 0.5 0.4

miR-b 120 54 0.0001 0.005

miR-c 36 1 0.5 0.7

... ... ... ... ...

miR-X 60 18 0.001 0.007

Reject miR-a & c because the FDR > 0.05

Filtrate with miRNA expression profiles

Ranking

MiRaGEMiRaGE

1 1

1 mmu-miR-25mmu-miR-25 1 1

2 mmu-miR-466i-5p 1 1

3 mmu-miR-92ammu-miR-92a 0.75 1

4 mmu-miR-19ammu-miR-19a 1 0.69

5 mmu-miR-19bmmu-miR-19b 1 0.69

6 mmu-miR-3082-5p 1 0.56

7 mmu-miR-130a 1 0.5

8 mmu-miR-130b 1 0.5

9 mmu-miR-15b 1 0.5

10 mmu-miR-2861 1 0.5

11 mmu-miR-3096-5p 1 0.5

12 mmu-miR-32 0.5 1

13 mmu-miR-322 1 0.5

14 mmu-miR-721 1 0.5

15 mmu-miR-149* 0.5 0.88

16 mmu-miR-3081* 1 0.38

17 mmu-miR-574-5p 1 0.31

18 mmu-miR-669n 0.5 0.81

19 mmu-miR-1187 1 0.25

20 mmu-miR-182mmu-miR-182 0.5 0.75

miRNA 　　 P30<<MBtarget gene P30>>MB

miR-17~92 cluster family miR-17~92 cluster family membersmembers are ranked in top 5 by combination of MiRaGE methods and miRNA expression profiling.

selected bymiRNA miRNA expression / MiRaGE

suggested contribution suggested contribution to cancer formationto cancer formation

12

mmu-miR-100 1 1

mmu-miR-126-3p 1 1

mmu-miR-29cmmu-miR-29c 1 1

mmu-miR-376ammu-miR-376a 1 1

mmu-miR-451 1 1

mmu-miR-99b 1 1

mmu-miR-136* 1 0.9375

mmu-miR-299* 0.75 1

mmu-miR-26ammu-miR-26a 1 0.5

mmu-miR-26bmmu-miR-26b 1 0.5

mmu-miR-29ammu-miR-29a 0.5 1

mmu-miR-7a-1*mmu-miR-7a-1* 1 0.5

mmu-miR-3107 1 0.4375

mmu-miR-340-5p 1 0.3125

mmu-miR-369-5p 1 0.3125

mmu-let-7ammu-let-7a 1 0.25

mmu-let-7emmu-let-7e 1 0.25

mmu-let-7gmmu-let-7g 1 0.25

mmu-let-7immu-let-7i 1 0.25

mmu-miR-467b 0.25 1

selected bymiRNA miRNA expression / MiRaGE

tumor-suppressive miRNAs tumor-suppressive miRNAs

neuron-specific miRNAsneuron-specific miRNAs

miRNA 　　 P30>>MBtarget gene P30<<MB

Some of the neuron-neuron-specific miRNAspecific miRNAs and tumor-suppressive tumor-suppressive miRNAsmiRNAs seem to contribute to the gene expression profiles of P30.

13

miRNA 　　 P30<<P6target gene P30>>P6

miR-17~92, miR-17~92, mir-106b-25mir-106b-25 ,,mir-106a-363mir-106a-363cluster family members are ranked in top 5 by combination of MiRaGE methods and miRNA expression profiling.

selected bymiRNA miRNA expression / MiRaGE

1 mmu-miR-106bmmu-miR-106b 1.00 1.00

2 mmu-miR-130a 1.00 1.00

3 mmu-miR-130b 1.00 1.00

4 mmu-miR-15b 1.00 1.00

5 mmu-miR-17mmu-miR-17 1.00 1.00

6 mmu-miR-20ammu-miR-20a 1.00 1.00

7 mmu-miR-20bmmu-miR-20b 1.00 1.00

8 mmu-miR-301b 1.00 1.00

9 mmu-miR-322 1.00 1.00

10 mmu-miR-721 1.00 1.00

11 mmu-miR-93mmu-miR-93 1.00 1.00

12 mmu-miR-542-3p 1.00 0.94

13 mmu-miR-3081* 1.00 0.88

14 mmu-miR-335-3p 1.00 0.88

15 mmu-miR-199a-5p 1.00 0.81

16 mmu-miR-199b* 1.00 0.81

17 mmu-miR-19ammu-miR-19a 1.00 0.81

18 mmmmu-u-mmiiRR--1199bb 1.00 0.81

19 mmu-miR-148a 0.75 0.94

20 mmu-miR-214 1.00 0.63

14

selected bymiRNA miRNA expression / MiRaGE

tumor-suppressive miRNAs tumor-suppressive miRNAs

neuron-specific miRNAsneuron-specific miRNAs

miRNA 　　 P30>>P6target gene P30<<P6

Some of the neuron-neuron-specific miRNAspecific miRNAs and tumor-suppressive tumor-suppressive miRNAsmiRNAs seem to contribute to the gene expression profiles of P30.

mmu-miR-29c 1.00 1.00

mmu-miR-376ammu-miR-376a 1.00 1.00

mmu-miR-451 1.00 1.00

mmu-let-7bmmu-let-7b 1.00 0.94

mmu-let-7emmu-let-7e 1.00 0.94

mmu-let-7gmmu-let-7g 1.00 0.94

mmu-let-7immu-let-7i 1.00 0.94

mmu-miR-98 1.00 0.94

mmu-miR-126-3p 0.75 1.00

mmu-miR-299* 0.75 1.00

mmu-miR-29a 0.75 1.00

mmu-let-7ammu-let-7a 0.75 0.94

mmu-miR-3070b-3p 1.00 0.69

mmu-miR-138 1.00 0.63

mmu-miR-3107 1.00 0.56

mmu-miR-181a-1* 0.50 1.00

mmu-let-7dmmu-let-7d 0.50 0.94

mmu-miR-1937b 0.25 1.00

mmu-miR-1937c 0.25 1.00

mmu-miR-337-5p 1.00 0.25

15

MiRaGE method + miRNA expression successfully pick up biologically important miRNAs. Further (wet) experiments which supress miRNA expression with tiny LNA is now planed.

If it is successful, our method can find miRNAs which control tumor formation.

16

Significance of reciprocal relationship between miRNA and its target genes.

t.test of P-values between top n miRNAs and others:

P30 →　MBP(mRNA:down|miRNA:up)P(mRNA:up|miRNA:down) P(miRNA:down|mRNA:up) P(miRNA:up|mRNA:down)

P=0.05

⟨ log10P⟩

n

17

Significance of reciprocal relationship between miRNA and its target genes.

t.test of P-values between top n miRNAs and others:

P30 →　P6P(mRNA:down|miRNA:up)P(mRNA:up|miRNA:down) P(miRNA:down|mRNA:up) P(miRNA:up|mRNA:down)

P=0.05

⟨ log10P⟩

n

18

MiRaGE method + miRNA expression satisfy reciprocal relationship very well. In our knowledge, this is for the first time to do this for such a large number of miRNAs

19

Discussion: What causes successful achievement?

Point 1:Usage of “good” maicroarryAffymetric ☓Agilent ○

Point 2:Negative set = genes not targeted by considered miRNA but done by other miRNAs

20

As for Points:Although we do not know the reason, off-target genes targeted by other miRNAs are more expressive.

let-7a transfection (Taguchi & Yasuda, 2010)

off target target

21

Conclusion:

MiRaGE (MiRNA Ranking by Gene Expression) method is very simple, but

1) can successfully pickup biologically important genes

and

2) can detect reciprocal relationship between miRNAs and their target genes (mRNA).

22

Acknowledgements:Acknowledgements: We thank Drs. Tetsuo Noda We thank Drs. Tetsuo Noda and Katsuyuki Yaginuma for and Katsuyuki Yaginuma for

providing reagents.providing reagents.These works were supported These works were supported by KAKENHI (23300357) .by KAKENHI (23300357) .

23

Related Talk:TomorrowTomorrow

(38)/SIG-BIO 15:30 - 15:45Gene expression regulation during differentiation from murine ES cells due to microRNA○Masato Yoshizawa,Y-h. Taguchi(Chuo Univ.)