nat. med. 17, 726–731 (2011) defective wnt … › original › nature-assets › nm › ...nat....

Nat. Med. 17, 726–731 (2011)

Defective Wnt-dependent cerebellar midline fusion in a mouse model of Joubert syndromeMadeline A Lancaster, Dipika J Gopal, Joon Kim, Sahar N Saleem, Jennifer L Silhavy, Carrie M Louie, Bryan E Thacker, Yuko Williams, Maha S Zaki & Joseph G Gleeson

In the version of this supplementary file originally posted online, the Supplementary Discussion was missing. The Supplementary Discussion is now provided as of 8 July 2011.

co r r e c t i o n n ot i c e

Supplementary Information

Defective Wnt-dependent cerebellar midline fusion in a mouse model of Joubert syndrome

Madeline A. Lancaster, Dipika J. Gopal, Joon Kim, Sahar N. Saleem, Jennifer L. Silhavy,

Carrie M. Louie, Bryan E. Thacker, Yuko Williams, Maha S. Zaki, Joseph G. Gleeson

Nature Medicine doi:10.1038/nm.2380

Supplemental Figure 1

a

Ahi1+/+ Ahi1–/–

15 m

onth

s

b c

Ahi1+/+ Ahi1–/– Ahi1–/–

Rostral

Caudal

Control Ahi1–/–

0

0.05

0.1

0.15

0.2

0 10 20 30 40 50 60

Control Ahi1–/– Control Ahi1–/–

Total brain size Relative cerebellum size

Rel

ativ

e ar

ea (a

.u.) ***

d

e

Control Ahi1–/–

Are

a (m

m2 )

Supplemental Figure 1. Later phenotype but intact brainstem structures. a. Average total brain area size measurements for Ahi1 mutants compared with littermate controls (left). Average cerebellum area size relative to total brain area (a.u.=arbitrary units) from Ahi1 mutants compared with littermate controls (right). *P<0.05, **P<0.0005 Student’s t-test, n=3 for each genotype. b. Representative 15 month old littermate sagittal C-V stained sections revealing persistent size and foliation defects (arrows). c. Top, C-V stained coronal sections revealing intact superior cerebellar peduncles (arrow) in Ahi1–/– compared with littermate control. Bottom, C-V stained mediolateral sagittal sections revealing intact brainstem structures in Ahi1 mutants. Arrowheads=deep cerebellar nuclei, arrow=superior cerebellar peduncle, open arrows=descending trigeminal tract, open arrowheads=dorsal column nuclei. d. C-V stained sagittal paramedical sections of cerebella from Ahi1 mutant and control littermate at higher magnification revealing deep cerebellar nuclei (arrowheads). e. Biotinylated dextran amine (BDA) tracing assay in two Ahi1 mutants and littermate control. Coronal sections through brainstem and spinal cord reveal crossing of pyramidal axons from rostral to caudal sections. Boxes outline DAB positive axons and lines reveal crossing of axons.



Ahi1+/– Ahi1–/–3

wee

ks

IGL PC ML IGL PC ML

b

a

c Ahi1+/– Ahi1–/–

Ahi1+/– Ahi1–/–

Cal

b H

oech

stG

AB

A-A

Rα6

Supplemental Figure 2. Normal layering in Ahi1 mutant cerebella. a. High magnification image of layering within the cerebellar vermis of Ahi1 mutant and littermate control as visualized by C-V staining of sagittal sections. Layers are shown: inner granule layer (IGL), Purkinje cell layer (PC), molecular layer (ML). b. Staining for cerebellar layers appears intact in Ahi1 mutants. Calbindin (green) labels Purkinje neurons, GABA-A receptor α6 (GABA-ARα6, red) labels inner granule layer. Hoechst (blue) labels nuclei. c. Golgi-Cox staining revealing dendritic tree architecture of Purkinje neurons in Ahi1–/– and Ahi1+/– cerebella.



a

0

0 .005

0 .01

0 .015

0 .02

0 .025R

elat

ive

PH

3+ cel

ls

Con KO

E16.5

b


Ptc

1P

tc1

Hoe

chst

*

+/– –/–Ahi1:N-Myc

Tubulin

c

Supplemental Figure 3. Early proliferation defects but intact Shh signaling in Ahi1 mutants. a. Average number of phospho-histone H3 (PH3) stained external granule layer neurons (EGL) at E16.5 relative to total cells in EGL. *P<0.05, Student’s t-test, n=3. b. Western blot on whole cerebellum lysates from P5 littermates. N-myc antibody recognizes a primary band at approximately 70kDa. Tubulin is the loading control. c. Ptc1 (green) staining in P5 midline sagittal sections from Ahi1–/– andAhi1+/– littermates. Hoechst (blue) labels nuclei.


Supplemental Figure 4Ahi1+/– Ahi1–/–

Ant

erio

rP

oste

rior

AP

E13

.5

V

D

Supplemental Figure 4. Roof plate widening in Ahi1 mutants. Matched transverse C-V stained sections from littermates arranged in a series from anterior to posterior revealing the progressive posterior widening of the roof plate in Ahi1 mutant (bars). Diagram inset depicts approximate locations of anterior and posterior sections.



a

PGK-neo

knockout allele

Cre recombination

wildtype allele

PGK-neotargeting vector HSV-TK

homologous recombination in ES cells

Chr. 10129/SvJ

targeted allele

36 37 3839 4035343331 32

1 kb

383936 3735343331 32

35343331 32 36 37 3839

383935343331 32

b

0

2

4

6

8

10

Control Cep290–/–

Are

a (m

m2 )

02468

1012

Aver

age

lobu

le n

umbe

r



c

Ahi1+

/–

Ahi1–

/–

Sup

erio

rIn

ferio

r

Sup

erio

rIn

ferio

rMouse E16.5

NS

*

Supplemental Figure 5. Cep290 targeting and adult phenotype. a. Schematic of targeting approach used to generate Cep290 null mice described in methods. Homologous recombination is represented by dashed crossing lines, and resulting alleles are shown. Positions for genotyping primers are shown: Closed arrows are wild-type primers, open arrows are knock-out primers. b. Left, representative midline sagittal C-V stained sections from adult Cep290–/– and littermate control. Arrow points to mild foliation defect in Cep290 mutants. Right, quantification of average vermis area in Cep290–/– and littermate controls. NS=not significant, Student’s t-test, n=3 for each genotype. Bottom, quantification of average number of lobules in Cep290–/– littermate controls. *P<0.05, Student’s t-test, n=3 for each genotype. Error bars are S.E.M. c. Transverse C-V stained sections from Ahi1 mutant and littermate control shown in series from superior to inferior. Arrow points to fusion defect.



Vecto

rW

TR7

23Q

H896

RV4

43D

GFP

α-tubulin

25 kDa

50

150

a

b

Jbnα-tubulin

WT

R723

QH8

96R

V443

D

150

50

R723Q H896R V443D

Per

cent

age

of C

ells

(%)

Ciliary localization Nonciliary (diffuse) localization

Vector WT V443DR723Q H896R

WT

0102030405060708090

100

c

Supplemental Figure 6. Expression analysis and localization of disease mutations. a. Western blot from 293T cells transfected with mutant constructs and assayed for GFP levels. Tubulin represents the loading control. b. Western blot from IMCD cells transfected with mutant constructs and assayed for Jbn levels. Tubulin represents the loading control. c. Localization of Jbn disease mutation constructs (green) in IMCD cells stained for cilia (acetylated tubulin, red). Hoechst labels nuclei. Below, quantification of cilia localization in 100 cells for each construct.



0

0 .5

1

1 .5

2

a

bVector WT R723Q H896R V443D

+ + + + +W3A:

c

GFP EV

GFP-JbnW

T

GFP-R723Q

GFP-H896R

GFP-V443D

GFP EV

GFP-JbnW

T

GFP-R723Q

GFP-H896R

GFP-V443D

IP: anti-GFP Lysates

β-Catenin

GAPDH

100 kDa

37

GFP

WB:

25

25

100 kDa

37

25

25

150150

5μm

Ac-Tubulin Jbn-GFP Tubulin Jbn Hoechst

Ac-Tubulin α-Jbn Tubulin Jbn Hoechst

5μm

α-Jbn β-catenin-GFP Jbn β-cat Tubulin

5μm

Ac-Tubulin α-β-catenin Tub β-Cat Hoechst

5μm

Rel

ativ

e lu

cife

rase

*

NSNS

NS

Supplemental Figure 7. Disease mutations fail to potentiate Wnt signaling. a. Luciferase activity in fibroblasts transfected with wild-type Jbn or mutant constructs and treated with Wnt3A (W3A) conditioned media. NS=not significant, *P<0.05, Student’s t-test, n=3. Error bars are S.E.M. b. Co-immunoprecipitation in 293T cells of Jbn GFP constructs pulled down with GFP antibody and western blot for endogenous β-catenin which reveals interaction between β-catenin and wild-type Jbn as well R723Q and H896R but not with V443D. GFP western blot reveals efficiency of immunoprecipitation, while GAPDH is a negative control. c. Localization of Jbn-GFP construct (green, first row) or endogenous Jbn (green, second row) to cilia (arrows) stained for acetylated tubulin (red) in CGNs isolated from wild-type P5 cerebella. Ciliary localization of endogenous β-catenin (green, third row) and colocalization of β-catenin-GFP construct (green, fourth row) with endogenous Jbn (blue) at the cilium (acetylated tubulin, red). Hoechst (blue) labels nuclei in the first three rows.


Supplemental Figure 8a

Ac-

Tubu

lin H

oech

st

Ahi1+/+ Ahi1–/–


Arl1

3b H

oech

st

20μm 20μm

b

c

Ahi1+

/–Ahi1–

/–

Gli1 Ptc1 Merged

100μm

100μm

Supplemental Figure 8. Cilia staining in Ahi1 mutant cerebellar neurons. a. Cilia staining (arrows, acetylated tubulin, red) in isolated CGNs from P5 littermates. Hoechst labels nuclei (blue). b. Shh target gene staining in E14.5 transverse sections from Ahi1–/– and Ahi1+/– littermates. Gli1 (red) is not expressed at this stage. Ptc1 (green) is also weakly expressed primarily in the future EGL as well as at the site of fusion. This expression appears equally low in the Ahi1 mutant. Hoechst (blue) labels nuclei. c. Arl13b staining (red) to visualize cilia (arrows) on transverse sections of Ahi1–/– and littermate control cerebella at E13.5. Hoechst (blue) labels nuclei.


Supplemental Figure 9β-

Gal

acto

sida

se H

oech

st

Control Ahi1–/–

NaCl

LiCl

a

b

Con

trol

Ahi1-/-

Untreated LiClc

Control Ahi1–/– Control Ahi1–/–Untreated Lithium

Ki6

7 H

oech

st

Supplemental Figure 9. Wnt activity staining in LiCl treated embryos. a. β-galactosidase (green) staining of transverse sections from E14.5 BATgal+ embryos treated with NaCl or LiCl reveals increased Wnt activity (arrows) with LiCl treatment in both control and Ahi1–/– embryos compared with NaCl treated embryos. Hoechst (blue) labels nuclei. b. Ki67 staining (red, arrows) in Ahi1–/– and Ahi1+/– littermate controls treated with LiCl compared with untreated controls which was quantified in Figure 4f. Hoechst labels nuclei. c. Cresyl-violet sagittal sections from P11 littermates treated with LiCl at E12.5 and E13.5 or untreated.



1.E + 03

1.E + 04

1.E + 05

1.E + 06

1.E + 07

1.E + 08

1.E + 09

1.E + 10

1.E + 11

E11.

5E1

2.5

E13.

5E1

4.5

E15.

5

E16.

5E1

7.5

E18.

5E1

9.5

E20.

5 P0 P1 P2 P3 P4 P5

W TAhi1 KOShh m utant

Stage 1:Wnt

Stage 2:Shh

Num

ber o

f cel

ls (l

og s

cale

)

x(t) = a • 2t/1

x(t) = a • 2t/0.5

x(t) = a • 2t/2

Dorsal view:

Ventral view:

Dorsal view of MHB:Lateral view of embryo:

Control

Ahi1-/-

V

D

V

D

a

L R

P

AV

D

R L

P

AD

V

E13.5

E14.5

mb

cbcb

lrllrlIV

b

Supplemental Figure 10. Fusion model and mathematical modeling of a two stage proliferation model. a. Model of midline cerebellar fusion and its disruption in Ahi1 mutant mice. The embryonic head is shown from the side and a close up view of the mid-hindbrain junction (MHB) is shown from the dorsal surface revealing the midbrain (mb), cerebellar hemispheres (cb), lower rhombic lip (lrl) and fourth ventricle (IV). A dorsal view of the isolated cerebellum is shown with a depiction of the plane of sectioning for transverse sections (shown in a schematic at the right) and axes at the left: anterior (A), posterior (P), left (L), right (R), ventral (V), dorsal (D). The ventral view of the fusing cerebellum reveals the fusion events that occur and are visible on transverse sections. A progression from E13.5 to E14.5 is depicted to provide a sense of the closure that occurs. b. Mathematical modeling of the proliferation defect in Ahi1 knock-out (KO) mice compared with wild-type and a hypothetical Shh mutant. Equations are shown for each segment. See supplemental text.


Supplementary Discussion

Mathematical modeling of the proliferation defect

Our findings suggest that JS gene mutation leads to early proliferation defects in a Wnt-

dependent manner while postnatal proliferation and Shh signaling is relatively preserved.

Using the data we have obtained from BrdU labeling at various time points, we attempted to

mathematically model the proliferation which occurs early and late in cerebellar development.

The rate of proliferation at the midline at E13.5 is quite similar to that we observed at E16.5 in

control mice, and this rate increases to approximately 2-fold higher by P5. We therefore

propose a mathematical model of this proliferation, which occurs in two stages, an early Wnt-

dependent stage and a later Shh-dependent stage.

For this model, the basic equation for exponential growth can be written

x(t) = a · bt/τ

where the quantity x depends upon time t, the constant a is the starting quantity and the

constant b is the growth factor. τ is the time required for x to increase by a factor of b. In our

model of cerebellar growth, x refers to the number of cerebellar cells at a given time t in days

from E11.5 to P5. The constant b is 2 since cerebellar proliferation is a process of doubling,

and we can set a with an arbitrary number of 1000 at E11.5, since we are interested in

relative growth rather than absolute number of cells. In the model for 2-stage growth, the

second stage exhibits a proliferation rate that is 2-fold higher than the initial stage based upon

our BrdU measurements. Therefore, we assign τ the value of 1 for the initial stage from E12.5

to E16.5 whereas this value changes to 0.5 during the second stage from E17.5 to P5. When

plotted on a logarithmic scale, cell growth during the initial stage exhibits a smaller slope

compared with the second stage (Supplemental Fig. 10b).

Next we can model the effect of reducing the proliferation rate of only one stage as in the

Ahi1 mutant mice. Since we have determined that Ahi1 mutants exhibit a reduction in

proliferation of approximately one half during early development, we assign τ the value of 2

during the first stage while τ remains 0.5 during the second stage. This results in a decreased


slope during stage 1 compared with wild-type whereas during stage 2 the slopes are equal.

When compared to a mutant which would affect stage 2 proliferation, such as a Shh mutant, if

we assign the same 2-fold reduction in proliferation rate, a reduction which has been

previously reported with Shh inhibition1, but instead for stage 2, then τ is 1 during the first

stage and remains 1 for the second stage as well. This results in a reduction in slope only

during the second stage whereas the slope is equal to wild-type during stage 1.

The final outcome of this model reveals a relatively mild effect of disruption of stage 1

compared with stage 2. This supports our findings which reveal that the phenotypic result of

loss of Ahi1 is relatively mild compared with mutants which affect Shh-dependent later

proliferation, such as Smo or cilia mutants2. This mathematical model captures the effect

seen in Ahi1 mutants in which vermis growth appears delayed but never fully recovers.

References

1. Wallace, V.A. Purkinje-cell-derived Sonic hedgehog regulates granule neuron precursor cell proliferation in the developing mouse cerebellum. Curr Biol 9, 445-448 (1999).

2. Spassky, N., et al. Primary cilia are required for cerebellar development and Shh-dependent expansion of progenitor pool. Dev Biol 317, 246-259 (2008).


nat. med. 17, 726–731 (2011) defective wnt … › original › nature-assets › nm › ...nat....

Documents