BMP SIGNALING SUPPORTS PRIMORDIAL GERM CELL

DEVELOPMENT BY REGULATING KIT LIGAND

BY

BRIAN MASON DUDLEY

Submitted in partial fulfillment of the requirements

for the degree of Doctor of Philosophy

Dissertation Adviser: Kathleen Molyneaux, Ph.D.

Department of Genetics

CASE WESTERN RESERVE UNIVERSITY

August 2010 CASE WESTERN RESERVE UNIVERSITY SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of ______Brian Mason Dudley______Candidate for the ___Ph.D.____ degree*.

(Signed) ___Ron_Conlon, Ph.D.______(Chair of Committee)

___Michiko Watanabe, Ph.D.______

___Brian Bai, Ph.D.______

___Kathleen Molyneaux, Ph.D.______(Thesis Advisor)

(Date) ___June 24, 2010______

* We also certify that written approval has been obtained for any proprietary material contained therein.


Dedication

I would like to dedicate this work to the following people: My parents Michael and Paula for all of their support throughout my education, they have always stressed the importance of education and helped motivate me through this process. To my brother

Philip and sister Erin for moral support and for visiting whenever they had the chance.

Also to my in laws, all of my friends, and Jasmine for providing entertainment away from school. Finally, and most of all I would like to dedicate this to my wife Colleen for being so supportive during my graduate work. She has been there to read over drafts, listen to talks (which only occasionally put her to sleep), keep me motivated, and help me relax and have fun. Getting though grad school was really a team effort and I could not have done any of it without the love and support of Colleen. I thank her and all my friends and family for ensuring that I will always have fond memories of my time as a Ph.D. student.


Table of Contents

" ((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((,
%"!(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((2
!" "(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((3
 "
+'

" #"
"
 

$ "((((((((((((((((((((((((((((((((((((((((((( ++
+(+

$ "
)

$ $%(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((+,
0+0+0

 $ +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 01
0+0+1

"$ 

  *$ 

$
$
#+++++++++++++++++++++++++++++++++++++++++++++++++++++ 03
0+0+2
(%
"$$ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 05
+(,

!


 "
"!


$ "((((((((((((((((((((((((((+2
0+1+0
 #

$
,.
% ")++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 07
0+1+1
 #
"
"%
$
$
#$$)

&" %#

&" $#+++++++++++++++++++++++++++++++++ 10
0+1+2
 #

") 
$$"

$
$
"#+ ++++++++++++++++++++++++++++++++++++++++++++++++++++ 13
0+1+3
 #
)
 "
" #

%$ 
#$#

" " %$&
& $+++++++++++++++++ 14
+(-

!




# $$

 "(((((((((((((((((((((((((((((((((((((((((((((((((((((((((,1
0+2+0
$


"

& $++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 16
0+2+1
$- $
") 
( "## ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 16
0+2+2

#$"%$%"#

-  ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 17
0+2+3
- 
#
$')#++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 20
0+2+4
$#

$
%$$ #


& $+ +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 21
0+2+5
 

$


"$ ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 22
0+2+6
, 
# + ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 23
 "
,'


#"!

# !

""&


#"# (((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((( -0
,(+
!" "(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((-1
,(,
" #" ((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((-2
,(-
" !

"! ((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((.*
1+2+0
"
%$%"+ +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3/
1+2+1
% #$+ +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 30
1+2+2
,!%$$$&
,+ +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 30
,(.
!#"!(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((.,
1+3+0
#

# $
#
( "##
 
" $ "#

#
%"
"

"$ + ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 31
1+3+1
"
#

 ,#
$"
'$
$
#  "
#)
$
8+4

0/+4++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 32
1+3+2
"$%"
 
#
$#

%"#

8+4
#
%$%"++++++++++++++++++++++++ 35
1+3+3

$"$$
#% "###

 $$)+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 37
1+3+4
(  %#
 3

 4
&$
( "## 

$

0

"
%$%"+38
1+3+5

$"$$
"%#
0

$
( "## 
'$
$
%" $
"+ + 41
,(/
!#!! ((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((/,
,(0
%"!(((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((/2
,(1
# "
"  ((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((((/3

3
 #!
0*
 

 #! "
 !# 


! !
!


&#
#
!(
#
"++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 3-
0+.
"#!#+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++3.
0+/
#! $#  ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++3/
0+0
#!"

# " ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++31
/)/)-


#'
#"))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 20
/)/).

$#"

$"#"#!())))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 20
/)/)/

' !""
(""))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 23
/)/)0
* ")))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 23
/)/)1
##"#")))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 24
0+1
"$#"+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++36
/)0)-

"
 #"
!
' !""

#
#
!"
$!

!#) )))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 25
/)0).
#
""

 !-
"
#
!"

") ))))))))))))))))))))))))))))))))))) 3-
/)0)/
!"
 !-
' !""
!!#"
&#
"


#!#

' !"") )))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 31
/)0)0
$

"
"
#
!"
#!#

"
#
#
#
!") )))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 32
/)0)1
#
 !-

$#
!("
%
&!
!#!(
")))))))))))))))))))))) 34
/)0)2

$#
!("
%
!"
#
!
"#

#) ))))))))))))))))))))))) 4-
0+2
"$""  ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++50
0+3
 &#"+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++54
0+4
$ #
#! ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++55
 #!
1*

  )# 

#
#

% %"
!
 $

#
, 
"  ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ 60
1+.
"#!#+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++61
1+/
#! $#  ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++61
1+0
#!"

# " ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++64
0)/)-
!
$#$!) ))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 53
0)/).

 ") ))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 53
0)/)/

$#$!

"
""(")))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 54
0)/)0
##"#")))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))) 55
1+1
"$#"++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .--
0)0)-
"
! $!
+
$!
!#
#
#
#
!") ))))))))))))))))))))))))))))))))-,,
0)0).
'$"
"$

!$"

$!"

#""$
$#$!))))))))))))))))))))))))))))))-,,
0)0)/

!#
"
"!$ #
&
#!##
&#
"$
 ) )))))))))))))))))))))))-,.
0)0)0
$

#!##
!""

!#
!
-,)1
#
!")))))-,0
0)0)1
!
$

!"
$#$!
"
!
##(
#
"$
 ) ))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))))-,2
1+2
"$""  +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ .-6
1+3
 &#"++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ../
1+4
$ #
#! +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ..0
 #!
2*
"$"" 

$#$!
!# "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++..2
2+.
 
"
$!

  )# 

#
#
!" +++++++++++++++++++++++++++++++ ..3
2+/
 
!$# 

#

#
#
!" ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ./-
2+0

! 

 
!%


' !"" 


"$!%%

!#  ++++++++++++ ./0
2+1
$#$!
& !++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ ./5
!"+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++.0.

1
List of Tables

Table 2-1: PCR Primers used for Real Time RT-PCR ………………………………... 59 Table 3-1: RT-PCR primer sequences ………………………………………………... 92


List of Figures

Figure 1-1: Bone morphogenetic protein (BMP) signaling pathway …………………. 20 Figure 1-2: The structures of KIT and KITL ………………………………………….. 30 Figure 2-1: BMP signaling components are enriched in the soma at E10.5 …………... 44 Figure 2-2: Cells of the mesonephric mesenchyme respond to endogenous BMPs …... 45 Figure 2-3: BMP signaling regulates PGC numbers ………………………………….. 47 Figure 2-4: Noggin treatment slows and randomizes PGC movements ………………. 50 Figure 2-5: BMP treatment induces expression of Kitl and Id1 ……………………… 53 Figure 2-6: Noggin treatment represses Kitl and Id1 expression in the genital ridges .. 54 Figure 3-1: Tools for targeting BMP-signaling during PGC migration ………………. 70 Figure 3-2: Tamoxifen induced Cre recombination of Bmpr1a-fx leads to decreased expression of Bmpr1a and decreased BMP signaling ………………………………….. 73 Figure 3-3: Expression of BMP target genes are reduced in the genital ridges of Bmpr1a conditional knock out embryos ………………………………………………………… 77 Figure 3-4: Conditional loss of Bmpr1a leads to reduced PGC targeting of genital ridges ………………………………………………………………………………………….. 79 Figure 3-5: Reduced BMP signaling leads to decreased numbers of PGCs in vivo ..… 80 Figure 3-6: Conditional loss of Bmpr1a leads to increased somatic cell apoptosis in the mesonephric mesenchyme ……………………………………………………………... 82 Figure 3-S1: A time course of Bmpr1a-fx recombination …………………………….. 88 Figure 3-S2: BMP signaling was not reduced in the E10.5 genital ridge of Pax2-Cre/+ Bmpr1a-fx/Bmpr-s embryos …………………………………………………………… 89 Figure 3-S3: Examples of PGCs exhibiting targeting and non-targeting migration .…. 90 Figure 3-S4: PGC trace times are similar in movies generated from wild type, heterozygous, and conditional KO embryos …………………………………………… 91 Figure 3-S5: Injection with TM at E8.5 does not affect PGC numbers at E9.5 ………. 91 Figure 4-1: Inhibition of KIT by ACK2 reduces PGC numbers in cultured E9.5 slices …………………………………………………………………………………………. 101 Figure 4-2: Exogenous soluble KITL reduces the number of PGCs in organ culture .. 103 Figure 4-3: Manipulation of KIT/KITL signaling with KITL(s) or ACK2 disrupts PGC migration ……………………………………………………………………………… 105 Figure 4-4: PGC survival curves in control and ACK2 treated slices ……………….. 106 Figure 4-5: Exogenous KITL(s) leads to increased PGC emigration from the E10.5 genital ridge …………………………………………………………………………... 107 Figure 4-6: Cultured PGCs adhere more tightly to cells expressing KITL(mb) than cells expressing KITL(s) or no KITL ……………………………………………………… 108 Figure 5-1: Bone morphogenetic protein (BMP) signaling in the E9.5 genital ridge establishes a permissive environment for arriving primordial germ cells ……………. 121 Figure 5-2: Membrane bound (mb) KITL tethers primordial germ cells (PGC) to their somatic neighbors in the genital ridges ……………………………………………….. 127


Acknowledgements

I would like to acknowledge many people who have helped me throughout my graduate student career. First I would like to acknowledge the current and past members of the Molyneaux lab who provided critical help and training, especially Congli Cai,

Ph.D. I would also like to thank the members of my committee for their guidance throughout this process. I never had a single difficulty planning a meeting thanks to their great accessibility. I would like also like to acknowledge the members of the developmental biology group organized by Brian Bai, Ph.D. Through our weekly meeting I gained excellent experience in presenting my data and critically following presentations by others. Additionally, this group provided excellent suggestions and necessary critiques. I would also like to thank Alex Siebold, Ph.D., for acting like a student mentor while we were both at CWRU, Johnnie Chau for being my unofficial journal club partner, and Mike Schnetz, Ph.D. and Gabe Zentner for all their help with the unreported ChIP work. I would also like to express my gratitude for all the help Patty

Conrad provided through imaging training, retraining, and troubleshooting. Nearly every image in this thesis was collected using a microscope supported by Patty. Lastly, I would like to thank my advisor Kathy Molyneaux, Ph.D., for her support, training, advice, and everything she has done to get me to this point, including hiring me as a research assistant six years ago. I absolutely would not be where I am now without her guidance.


BMP Signaling Supports Primordial Germ Cell Development

by Regulating Kit Ligand

Abstract

By

BRIAN MASON DUDLEY

Many species have evolved to propagate through sexual reproduction. This process provides reliable transmission of genetic material to offspring while allowing for the accumulation of mutations that contribute to diversity. Sexual reproduction requires the development of specialized cell types. In mammals, the germ cells are specified early in embryogenesis and undergo an elaborate process of migration before colonizing the nascent gonads. There, they undergo gradual differentiation, culminating in the production of gametes in the adult. Germ cell development relies on communication between germ cells and surrounding somatic cells. These signals control germ cell survival, proliferation, migration, and sexual differentiation, without which, life as we know it would cease.

For my thesis, I examined the signals that control primordial germ cell (PGC) migration from the hindgut to the genital ridges. Using tissue culture and conditional gene targeting in mice, I demonstrated that bone morphogenetic proteins (BMPs) play a pivotal role in PGC migration. I discovered that BMPs expressed in the pronephric


compartment of the E9.5 genital ridge establish and maintain a PGC nice within the

nascent gonads. BMP signaling supports somatic survival in the mesonephric

mesenchyme and represses expression of Scarb1, an epithelial marker. In the coelomic epithelium, BMP signaling promotes the expression of the PGC survival factor Kitl and

putative chemo-attractant Sdf1a, creating a niche environment for arriving PGCs.

Integral to the effectiveness of the genital ridge niche is high expression of Kitl.

During PGC migration KITL is present as a gradient with the highest levels in the genital

ridges. Disruption of this through KIT receptor inhibition, decreased Kitl expression, or

exogenous soluble KITL led to decreased PGC survival and defects in PGC targeting

resulting in increased ectopic accumulation of PGCs. I propose a model where PGCs

interact with their somatic neighbors through an adhesive interaction between somatic

expressed membrane bound (mb) KITL and PGC expressed KIT receptor. PGCs crawl

along a track of KITL(mb) to high genital ridge KITL(mb) levels that tether PGCs to the

target tissue. Through this mechanism BMPs promote a permissive environment within

the genital ridge for arriving PGCs.


Chapter 1: An Introduction to Germ Cell Development

By

Brian Mason Dudley


1.1 PGC Development – An Overview

The germ cell lineage is arguably one of the most important cell lineages for sexually reproducing organisms. These cells, comprised of the sperm and egg, are responsible for forming a new generation and thus continuing the species. Germ cells carry the genetic information from one generation to the next giving them a vital role in evolution. The factors that control development of germ cells ensure that the organism can reproduce successfully.

Early mammalian germ cell development is characterized by a conserved sequence of events: specification, migration, and colonization of the genital ridges. First, multipotent cells receive signals that cause them to differentiate into primordial germ cells (PGCs). These PGCs then undergo a series of active and passive migrations that ultimately bring them to their final destination at the developing gonads. There, PGCs associate with somatic tissue and form either the testis or ovaries. At this point germ cells begin to differentiate and continue on in a sex specific manner, undergoing meiosis in the females and mitotic arrest in males. Errors in any one of these steps can lead to infertility or the formation of germ cell tumors.

1.1.1 PGC Specification.

In the mouse, PGCs can first be identified at embryonic day 7.5 (E7.5) as a cluster of Alkaline Phosphatase positive cells in the posterior primitive streak at the base of the allantois (1). PGC specification occurs when signals from the extraembryonic ectoderm, including members of the bone morphogenetic protein (BMP) family, initiate


differentiation of a subset of proximal epiblast stem cells expressing high levels of

Fragilis (also known as Ifitm3) at approximately E6.25 (2-5). Upon specification, PGCs express B-lymphocyte maturation protein 1 (Blimp1) (also known as Prdm1) which in turn represses mesoderm lineage markers Hoxa1 and Hoxb1 (2, 5-7). Activation of

Blimp1 is followed by expression of the germ cell specific marker Stella (4, 7). Stella positive cells co-label with Alkaline Phosphatase indicating that these cells are in fact

PGCs.

BMP signaling plays a crucial role in early PGC development. Extraembryonic ectoderm derived BMP4 is absolutely necessary for PGC specification with BMP4 null embryos failing to specify PGCs (5). Bmp4 null embryos fail to initiate expression of

Fragilis and Blimp1 indicating an early requirement of BMP4 (4, 7). Loss of Bmp8b leads to a strong reduction in the number of nascent PGCs. BMP8b may act synergistically with BMP4 through formation of heterodimers (2, 8). Additionally, endoderm derived BMP2 is not required for PGC formation but may influence PGC numbers (3).

The transition from multipotent proximal epiblast stem cells to PGCs involves large-scale changes including repression of the somatic program, activation of PGC specific genes, reactivation of pluripotency genes, and reprogramming of the epigenetic state. Upon specification, PGCs already express many genes characteristic of their somatic neighbors that must be turned off in transition to the PGC fate (9). BLIMP1 has been shown to be the key factor responsible for suppression of the somatic program in newly specified PGCs (10). A wide range of somatic markers are gradually repressed including mesodermal markers Evx1, FGF8, and Wnt5b, as well as genes involved in cell


cycle regulation and all genes necessary for DNA methylation (9-11). Blimp1 expression, likely combined with other BLIMP1 independent regulators, leads to the expression of various genes necessary for PGC development such as Dnd1, Stella, Kit, and Nanos3, as well as pluripotency genes including Sox2 and Nanog (10). Evidence suggests that a second PR-domain containing protein, PRDM14, is necessary for epigenetic reprogramming within PGCs, independent of BLIMP1 (11). Newly specified

PGCs share a similar epigenetic state with their somatic neighbors until E8.0 when DNA methylation and dimethylated H3K4 modifications are quickly removed in PGCs (12).

At E8.5, H3K27 trimethylation increases in PGCs while remaining low in the epiblast

(12).

1.1.2 PGC Migration and Colonization of the Genital Ridges.

Newly formed PGCs migrate from the posterior primitive streak into the definitive endoderm at E7.5 where they soon become associated with the invaginating hindgut (13). Not all specified PGCs effectively migrate into the endoderm, instead forming non-motile clusters in the proximal allantois through E8.5 (13). PGCs within the endoderm and hindgut actively migrate throughout the hindgut as the embryo extends

(14). At E9.5, PGCs exit the hindgut, migrate dorsally, enter the body wall, and move away from the gut. Between E9.5 and E10.5, PGCs migrate into the midline and then laterally towards the genital ridges, the future sites of the gonads, in a wave from anterior to posterior. Migration into the body wall occurs before extension of the mesentery and those PGCs occupying the far mesentery or the gut do not show directional movement towards the genital ridges and likely undergo apoptosis (14, 15). The exact genital ridge


target of PGC migration is unknown. In zebrafish PGCs initially target pronephric progenitor cells before migrating into the adjacent developing gonad (16). This observation provides a possible link between PGC migration and early kidney development; however, the same connection has not been shown in mice.

The somatic cells of the developing genital ridge likely express factors that act as directional cues for migratory PGCs. In cell culture experiments, PGCs were found to migrate toward explanted E10.5 genital ridge tissue and not tissue from the hindgut mesentery or the limb buds (17). One such factor that has been implicated as a possible

PGC chemo-attractant is Stromal cell-derived factor-1 (Sdf-1) (18, 19). SDF-1 and the associated G-protein-coupled receptor CXCR4 have been shown to be necessary for PGC targeting of the genital ridges in both mice and zebrafish (18, 19). In mice, Sdf-1 is expressed at E9.5 throughout the dorsal body wall, highly enriched in the mesonephric mesenchyme of the genital ridges while migratory PGCs express Cxcr4 (18). PGCs null for Cxcr4 show normal behavior while in the hindgut but are subsequently unable to target the genital ridges (18). However, in a transwell migration assay, cultured PGCs did not migrate towards a high concentration of SDF, indicating that this factor may not be sufficient to attract PGCs (20). Taken together, these results indicate that PGCs are likely attracted towards the genital ridges through a mechanism involving SDF-

1/CXCR4.

Between E10.5 and E11.5 PGCs arrive at the genital ridges and downregulate

Tissue non-specific alkaline phosphatase (Tnap) and upregulate Mouse vasa homologue

(Mvh) and Germ cell-less (Gcl) (21-23). By E11.5 PGCs begin to show gender specific differences in migration as they organize into the sex chords. In males, most anterior


PGCs move in a posterior direction and ultimately group together in stripes (14). In

females, most PGC migration is random, with some posterior directionality, as PGCs

ultimately form small clusters in the developing ovary (14).

PGC migration to the genital ridges is extremely important for the fertility and

health of the animal. Failure of PGCs to reach the genital ridges can result in infertility

or formation of extragonadal germ cell tumors. Extragonadal germ cell tumors are

believed to arise from ectopic migratory PGCs that fail to undergo apoptosis and begin to

differentiate (24).

1.1.3 Sexual differentiation.

In mice, the master sex regulator is Sry found on the Y chromosome, which is sufficient to drive male gonadal development in XX transgenic mice, though insufficient to confer male fertility (25). Sry expression in males specifies the sertoli cell lineage from a population of cells fated to become granulosa cells in females (26, 27). During normal gonadal development in males and females, somatic cells of the coelomic epithelium migrate dorsally into the developing gonad beginning at E11.2 (28). In males, those cells migrating from E11.2 to E11.4 activate expression of Sry and are fated towards the sertoli cell lineage (27-29). All sertoli cells arise from coelomic epithelium cells that have expressed Sry (26). Coelomic epithelium cells continue to migrate into the developing gonad up to E12.5, contributing to the interstitial cells of the gonad (28). At

E12.5 the basement membrane layer of the coelomic epithelium thickens to form the tunica albuginea (28). Additional contributions to the somatic cells of the gonad come from the mesonephros from E11.5 through E16.5, however, none of these cells are


believed to contribute to the sertoli cell lineage (30). Sertoli cells then cluster with germ

cells and become surrounded by Leydig cells, the source of testosterone, and peritubular

myoid cells forming the sex chords by E12.0 (27, 28, 31).

Sexual differentiation hinges on the balance in FGF9 and WNT4 levels. In pre-

sertoli cells, SRY activates expression of Sox9, which upregulates Fgf9 (26, 32).

Expression of Fgf9 then remains high through positive feedback on Sox9, tipping the

FGF9/WNT4 balance in favor of FGF9 (32). In females, Sry is not expressed and there is

no induction of Sox9 within the gonads (25, 33). Without SOX9, WNT4 accumulates

and inhibits Fgf9 and Sox9 expression, repressing the male fate (32). Loss of Wnt4 is

sufficient to result in expression of Fgf9 and Sox9, resulting in a male gonad in the

absence of SRY (32).

Male germ cells continue to proliferate until E16.5 when they undergo mitotic

arrest until shortly after birth. Following reactivation of mitosis, the germ cells begin the

process of spermatogenesis (34). The spermatagonial stem cells of the testis are referred

to as the ASingle (AS) spermatogonia and divide to form either two daughter AS cells or two linked APaired (APr) cells. APr cells remain connected by a cytoplasmic bridge resulting

from incomplete cytokenisis and account for roughly half of the AS divisions. APr cells then continue to divide to form groups of 4, 8, or 16 AAligned (AAl) chains. These AAl spermatogonia then progress through differentiation as A1, A2, A3, A4, intermediate, and

ultimately B spermatogonia, which yield spermatocytes following a final division.

In females, the absence of Sry expression results in shift in the FGF9/WNT4

balance in favor of WNT4 and somatic cells migrating out of the coelomic epithelium

that would become sertoli cells in males, are fated to become granulosa cells in females


(26, 27, 32). During female development, the first structures to form in the ovary are the primordial follicles, comprised of a diplotene stage, prophase I arrested oocyte, surrounded by pre-granulosa support cells (35). Throughout embryonic and juvenile development the follicles grow and mature through various intermediate stages, with all or most being present in an ovary at any given time. Primordial follicles are characterized by squamous granulosa cells surrounding a single diplotene oocyte (36).

Next, the granulosa cells begin to take on a more cuboidal confirmation characteristic of primary follicles (36). In secondary follicles, the granulosa cell layer has expanded to multiple layers of supporting cells (36). Next, small voids within the follicle form, filled with a follicular fluid, or antrum, characteristic of early antral follicles (36). These voids then merge and expand to form a single larger antral space in antral follicles (36). The antral cavity splits the granulosa cell layers into the peripheral mural granulosa cells and the oocyte surrounding cumulus cells (37). Oocytes in late stage follicles reenter meiosis during estrous in response to a surge in luteinizing hormone (LH) produced by the pituitary gland and undergo oocyte maturation (37). Prior to estrous, many follicles progress to the point where they are receptive to LH, yet without this hormone, these follicles are cleared from the ovary through atresia (38).

1.2 The Roles of Bone Morphogenetic Proteins in PGC Development

1.2.1 BMPs and the TGF-
Superfamily.

Members of the transforming growth factor beta (TGF-
) superfamily are involved in countless processes during development, including PGC specification and gonadal development (39). This superfamily includes members of the BMP, TGF-
,


nodal, activin, and growth/differentiation factors (GDF) subfamilies in mammals.

Drosophila homologues include decapentaplegic (DPP) and glass bottom boat (GBB).

BMPs make up a large portion of this superfamily with over 20 members.

BMPs are synthesized and secreted as homo- or hetero-dimers and they interact with their receptors as dimers. BMPs signal through heteromeric receptor complexes comprised of type I and type II serine-threonine kinase receptors (Figure 1-1) (40).

Association of BMPs with the constitutively active type II receptor recruits the inactive type I receptor. The type II receptor then phosphorylates the type I receptor. The now active type I receptor then phosphorylates specific receptor regulated (R) SMAD proteins at two conserved serine residues found in a SS[V/M]S motif within their C termini (41).

BMP activated receptors specifically signal though SMAD1, SMAD5, and SMAD8

(SMAD1/5/8). Phosphorylated SMAD1/5/8 then associate with the common partner

SMAD4 and translocate to the nucleus where they control transcription of target genes through direct DNA binding at AGAC containing sequences or through direct or indirect interaction with other transcription factors (41). Phosphorylated SMADs are a marker of active BMP signaling.

Repression of the BMP signaling pathway can be mediated through several mechanisms. Inhibitor molecules including Noggin, Chordin, and Follistatin can bind

BMPs and block their interaction with their receptors (42-45). Noggin specifically binds

BMP2, BMP4, and BMP7 with high affinity (42, 43). Intracellularly, inhibitory I-

SMADs can disrupt the signal transduction pathway. SMAD6 and SMAD7 can form stable interactions with type I receptors, blocking their interaction with R-SMADs, while

SMAD6 can also block the interaction between SMAD1 and SMAD4 (41). SMAD1/5/8



Figure 1-1: Bone morphogenetic protein (BMP) signaling pathway. BMPs signal as dimers though extracellular association with constitutively active type II BMP receptors. BMP associated type II receptors then recruit and activate type I receptors through phosphorylation. Activated type I receptors phosphorylate SMADs 1, 5, and 8 which then associate with SMAD4 and translocate to the nucleus to regulate gene expression (46).


can also be targeted by SMAD-ubiquitination regulatory factor I (SMURF1) which marks

BMP specific R-SMADs for degradation, capable of controlling the duration of BMP signaling (41, 47).

BMPs and other members of the TGF-
superfamily are capable of signaling through non-SMAD dependent pathways. Cultured human melanocytes treated with

BMP4 showed increased phosphorylation of mitogen activated protein kinase (MAPK) and extracellular-signal regulated kinase (ERK) after 30 minutes, indicating activation of the MAPK/ERK pathway (48). Additionally BMP4 treatment of human umbilical vein endothelial cells induced capillary sprouting through activation of ERK1/2 through a

SMAD4 independent mechanism (49). BMP4 and TGF-
were found to activate the

MAP kinase kinase TGF-
-activated kinase (TAK1) which in turn activates the stress- activated protein kinase/ c-Jun N-terminal kinase (SAPK/JNK) pathway (50-52).

However, it is worth noting that in each of these cases of activation of non-canonical pathways, BMP treatment was done in cell culture and not in vivo.

1.2.2 BMPs are crucial to the stability of various niche environments.

Stem cells provide the body with a replenishing stock of various different cell types as the need arises. Accordingly, these cells must remain in an undifferentiated state until differentiation is needed. The microenviroment that provides the necessary factors for sustaining a stem cell population is referred to as the stem cell niche. BMPs have been shown to be involved in various stem cell niches including the Drosphila ovary and the mammalian intestinal and hematopoietic stem cell niches. The Drosophila ovary is comprised of many smaller ovarioles. Each ovariole consists of a series of egg chambers


at different stages of maturation (53). At the tip of each ovariole is the germarium containing two to three germ line stem cells (GSCs) and various supporting somatic cells constituting a stem cell niche (53, 54). The GSCs reside in the apical tip of the germarium, in direct contact with somatic cap cells and neighboring inner germarial sheath cells (54). Extending from the cap cells outward are additional support cells called terminal filament cells (54). Following division of each GSC, the daughter cell remaining proximal to the cap cells will remain a GSC. The distal cell will separate from the niche and divide to form an egg chamber containing an oocyte and it’s surrounding nurse cells (54). The process of asymmetric GSC division is likely controlled through directed spindle orientation by the GSC spectrosome (55). The position of the differentiating daughter cell relative to the cap cells is believed to be the major factor leading to differentiation. Cap cells and inner germarial sheath cells express BMP homologues Dpp, homologous to Bmp2/4, and Gbb (54). Loss of DPP leads to loss of

GSCs while overexpression leads to an increased number of GSCs at the expense of differentiating daughter cells (56, 57). Ectopic somatic activation of DPP signaling in the cap cells and inner germarial sheath cells has no affect on GSC self renewal indicating that the requirement of DPP signaling is specific to the GSCs (57). GSCs respond to somatic derived DPP and GBB through activation of MAD, the drosophila homologue of

SMAD, and an additional BMP signaling molecule media (MED) (58, 59).

Phosphorylated MAD and MED then bind directly to a silencing element of the GSC differentiation factor Bag of marbles (Bam) repressing its expression in GSCs (59, 60).

BAM is necessary for GSC differentiation in both the male and female gonad, and ectopic expression of Bam in female GSCs is sufficient to induce differentiation and GSC


loss (61, 62). In this manner, BMP signaling maintains GSCs in an undifferentiated state

through repression of Bam.

BMPs also play important roles in stem cell niche environments in mammals.

BMPs control the size of the intestinal stem cell (ISC) population in the adult intestinal

niche. The small intestine of the mouse functions in nutrient digestion and absorption

through a highly regenerative epithelial layer (63). The epithelial cell layer takes on a

villus/crypt structure, whereby functional, differentiated epithelial cells form the

protruding villus structure and undifferentiated, proliferative cells form an invaginated

crypt (64). The vast majority of the villus epithelium is made up of enterocytes

surrounding interior mesenchymal cells (63, 64). These cells are replenished

approximately every three days by intestinal stem cells residing in the neighboring crypts

(64). Near the base of each crypt is a ring of 4-6 ISCs (64). ISCs undergo asymmetric

division to produce transit amplifying (TA) progenitors which fill the remaining space

above the ring of ISCs, where they continue to proliferate in a multipotent state before

being pushed to the surface where they begin to differentiate (64).

The primary regulator of crypt patterning is WNT signaling through stabilized
- catenin, however, BMP signaling has been shown to counteract WNT signaling in controlling niche size (64, 65). Over proliferation of ISCs leads to intestinal polyposis, which is phenotypically similar to juvenile polyposis syndrome (65). WNT signaling activates
-catenin in TA cells and a subset of ISCs to promote cell proliferation (65).

This is countered by BMP4, expressed in the mesenchymal cells surrounding the ISCs

(64, 65). BMP4 leads to activation of PTEN, which inhibits the PI3K/AKT stabilization of
-catenin, effectively blocking WNT induced cell proliferation (64, 65). Loss of BMP


signaling in ISCs leads to expansion of the ISC population and formation of polyps with

multiple crypts (65). In normal crypts, transiently expressed Noggin inhibits the BMP pathway leading to increased activation of PI3K/AKT to allow for ISC self renewal (65).

Through this mechanism, BMP4 controls intestinal niche size through repression of ISC self renewal until such action is needed.

BMPs have also been shown to play a role in regulating niche size in the hematopoietic stem cell (HSC) niche of bone marrow. The HSC niche is characterized by HSCs attached to the endosteal surface of trabecular bone (64). HSCs are tethered directly to spindle-shaped N-cadherin+ osteoblastic (SNO) cells likely through N-

cadherin mediated adhesion (66). These SNO cells secrete factors that promote long term

(LT) HSC quiescence and adhesion to the trabelcular surface (67). Evidence suggests a

direct link between the number of supporting SNO cells and the number of HSCs (64).

One factor that has been found to regulate the number of SNO cells, and in turn the

number of HSCs, is BMP signaling. Bmpr1a is expressed in hematopoietic marrow cells

and surrounding bone cells including most osteoblastic cells, but not in HSCs (66). Loss

of Bmpr1a leads to increased numbers of SNO cells and ultimately more LT-HSCs (66).

In the normal HSC niche, BMP signaling acts to restrict the size of the niche by

repressing the expansion of the supporting SNO cells. By reducing the size of the niche,

BMPs can act to restrict the HSC population.

1.2.3 BMPs and embryonic patterning of the genital ridges.

The region of the embryo referred to as the genital ridge is comprised of several

different cell types derived from lateral plate and intermediate mesoderm (68). The


population of cells within the lateral plate mesoderm will contribute to endothelial,

vascular, and hematopoietic lineages while the intermediate mesoderm will ultimately

contribute to the kidney and its related structures (68, 69). BMPs play an important role

at several stages of mesoderm patterning including the necessary requirement of BMP4

for gastrulation and mesoderm formation (70). Subsequently, evidence from the chick

shows that surface ectoderm derived BMP4 is necessary for maintaining expression of

pronephric marker genes Pax2 and Sim1 leading to nephric duct formation (71). In a similar manner in zebrafish, lateral plate mesoderm derived BMPs restrict endothelial and hematopoietic cell expansion in favor of expansion of Pax2.1 expressing kidney progenitors (69). The requirement of BMP signaling in early pronephric development is also found in Xenopus where BMP signaling is necessary for development of the pronephric tubules and duct and later in the mouse where BMP4 is necessary for ureter development (72, 73).

1.2.4 BMPs play major roles in multiple stages of reproductive development.

Members of the BMP family play crucial roles at several points during reproductive development. In addition to the necessity of BMPs during PGC specification, BMPs have also been shown to be involved in gonadal germ cell development following sexual differentiation. The strongest evidence comes form the study of oocyte follicle maturation in the ovary.

The ovary is comprised of various staged follicles, made up of oocytes surrounded by supporting granulosa cells and theca cells. Theca cell expressed BMP7 has been shown to modulate the sensitivity of granulosa cells to FSH inducing primordial


follicle progression to primary, preantral, and antral follicles (74-77). Oocytes also employ BMP family members to signal back to granulosa and theca cells to control their own environment. Primary follicle oocytes in rats express Gdf9, Bmp15, and Bmp6 while granulosa cells express the corresponding receptors (77). Loss of Gdf9 leads to accumulation of stalled primordial and one cell layer primary follicles, consisting of atypical granulosa cells and lacking theca cells (78). In vivo injection of GDF9 stimulates the progression of primordial and primary follicles to small preantral follicles

(79). Oocyte expressed BMP15 may act synergistically with GDF9 in promoting follicle maturation (80). Additionally, oocyte derived BMP15 promotes expression of the germ cell survival factor Kit ligand (Kitl) in granulosa cells while GDF9 represses Kitl (81).

KITL then represses expression of BMP15 in oocytes in a negative feedback loop (81,

82). BMP15 regulation of KITL likely regulates granulosa cell mitosis (81). Later in more mature follicles, BMP-15, -6, and -7 have been found to reduce cultured bovine cumulus cell apoptosis (83).

The role of BMP proteins in spermatogenesis is less well described. Both Gdf9 and Bmp15 are expressed in the testis (84). Gdf9 is expressed beginning 18 days after birth, reaching a maximum 2 weeks later, and is found in pachytene spermatocytes and round and elongating spermatids (84, 85). Bmp15 is first expressed in gonocytes 3 days after birth, then again later in pachytene spermatocytes (84). Neither protein is found in the sertoli cells, which express the corresponding receptors for GDF9 and BMP15 (84).

Sertoli cells respond to oocyte derived GDF9 through upregulation of target genes including Inhibin B (84, 85).


1.3 The Roles of KITL in PGC Survival and Migration

1.3.1 Kit Ligand and germ cell development.

The cytokine KIT ligand (KITL, SCF, MGF, encoded by the Sl, steel locus) and its receptor KIT (c-KIT, encoded by the W, dominant spotting locus) are involved in melanocyte and hematopoietic development and absolutely required for germ cell development. Null mutations in either Kitl or Kit result in infertile animals with no germ cells, and various hypomorphic mutations of Kitl yield germ cell deficiencies during PGC migration and colonization of the genital ridges (86-88). Analysis of KIT deficient embryos shows PGCs failing to increase in numbers beginning around E9.5, coupled with an increase in the percent of ectopic PGCs (89). KITL has also been implicated to be involved in zebrafish PGC development through PGC culture experiments showing that a

KITL homologue promotes PGC survival (90). However, in vivo analysis of KITL is complicated by duplication of the zebrafish genome leading to a high degree of redundancy (90, 91).

1.3.2 Kit/Kitl embryonic expression.

Beginning at E9.0, Kitl is expressed along the migratory route of PGCs as a gradient with the highest expression at the genital ridges (92). At E10.5, Kitl expression becomes restricted to the developing gonads with high levels on the dorsal rim of the coelomic epithelium and within the pronephric duct (92, 93). Decreased KITL levels outside of the genital ridge during this time may induce apoptosis in ectopic PGCs (15).

The receptor, Kit, is expressed in PGCs starting near the time of their specification at

E7.0 and continuing until E12.5 (94). Following sexual differentiation, Kitl is expressed


in the granulosa cells of the ovary, while Kit is expressed by oocytes (92, 95). In the

testis, embryonic expression of Kit is limited to interstitial Leydig cells until 10 days after

birth when differentiating spermatogonia begin expressing Kit (95, 96). Sertoli cells act

as a source of KITL in the testis (97, 98). KIT signaling is reduced in developing

spermatids and expression of a constitutively active form of KIT results in disrupted

spermatid development, indicating that while KIT signaling may be important early in

spermatogenesis, subsequent downregulation is necessary (99, 100).

1.3.3 The structures of KIT/KITL.

The W locus on mouse chromosome 5, known to be involved in skin

pigmentation, hematopoiesis, and fertility, was found to produce a 5 kb transcript

encoding the KIT receptor (99, 101, 102). The resulting 145 kDa transmembrane protein

tyrosine kinase receptor capable of self phosphorylation was later identified as the

receptor for KITL (101-103).

KITL was first identified as a 30 kDa soluble protein capable of stimulating bone

marrow mast cell proliferation through interaction with KIT (103). Kitl is expressed form

the Sl locus on chromosome 10, containing nine exons (99, 104). The expressed

glycoprotein is comprised of an extracellular domain, including the KIT binding

sequences, a 26 residue hydrophobic transmembrane domain, and a cytoplasmic domain.

In vivo KITL exists as both a membrane bound form and a soluble form.

Alternative splicing of the Kitl mRNA results in the production of two isoforms referred

to as KL1 and KL2 (Figure 1-2) (105). KL1 encodes for a full-length, 248-residue membrane bound KITL that is proteolytically cleaved to produce a soluble, 165-residue


form (99, 105, 106). Two proteases capable of cleaving KL1 have been identified as

mast cell chymase and matrix metalloproteinase-9 (MMP-9) (107, 108). KL2 encodes a truncated protein with an in frame deletion of exon 6 excluding the major proteolytic cleavage sites. As a result, this variant leads to a predominantly membrane bound form.

Although cleavage of KL2 does occur in vivo at a site encoded within exon 7, functional assays suggest that the soluble product of KL2 is less biologically active than the soluble product of KL1 (105, 109). The ratio of expression of the two Kitl splice variants is tissue specific with KL2 being the predominant isoform in the adult testis (105, 106, 109).

Soluble and membrane bound KITL have been shown to have different affects on

PGCs in vivo and in culture. Mice homozygous for a mutation mimicking the predominantly membrane bound KL2 isoform have a significant reduction in soluble

KITL levels yet are fertile and have no coat pigment phenotypes (109). Conversely, the

Steel-dickie (Sld) mutant mouse line expresses only a truncated, soluble from of KITL and is infertile with PGC defects observed as early as E9.5 (87, 110). PGCs in culture are also highly dependent on KITL. Traditionally PGCs are cultured on a feeder cell layer, such as Kitl expressing STO cells. STO cells expressing wild type Kitl are capable of

supporting PGC survival to a greater extent than those expressing the Sld form of KITL

(111). However, addition of soluble KITL to cultured PGCs with or without a feeder cell

layer enhances PGC survival, but not proliferation (112). Culture of later stage germ

cells shows that membrane bound KITL, but not soluble KITL, is able to promote oocyte

growth (113).



Figure 1-2: The structures of KIT and KITL. A) The extracellular region of KIT contains five immunoglobulin (Ig)-like domains (114). The three most N-terminal domains serve in ligand binding while the fourth is necessary for ligand induced receptor dimerization (114, 115). The transmembrane domain is comprised of 23 hydrophobic amino acids (99). Intracellularly, KIT has 433 amino acids forming a proximal ATP binding kinase domain, a non-conserved kinase insert, a phosphotransferase domain, and a C-terminal tail (99). B) KITL exists as either a membrane bound or soluble protein. Full length Kitl mRNA codes for a full-length KITL protein, KL1, which is initially membrane bound. KL1 can then be proteolytically cleaved to produce a soluble isoform, KITL(s). Kitl mRNA can be alternatively spliced to produce a variant lacking exon 6 (red outline). The resulting protein, KL2, lacks the known cleavage sites and is primarily membrane bound. However, KL2 can be proteolytically cleaved to produce a less active soluble form, KL2(s) (105).


1.3.4 KIT/KITL signaling pathways.

KIT and KITL act through various pathways leading to a multitude of cellular outcomes depending on factors such as cell type and developmental stage. In general, binding of KITL to the KIT receptor leads to dimerization of the ligand followed by dimerization of the receptor and activation of the kinase domain (116). KIT then undergoes autophosphorylation at several sites that serve as binding sites for intracellular signaling partners including phosphatidylinositol 3-kinase (PI3K), phospholipase-C gamma (PLCG), JAK2, SRC, and growth factor receptor-bound protein-2 (GRB2) (99,

117, 118). Upon activation of KIT, PI3K associates with Tyr719 and signals through c-

JUN and c-FOS to regulate cell adhesion (99). This is coupled with activation of AKT, which then phosphorylates/deactivates BAD to inhibit apoptosis and associates with p70S6K to promote cell proliferation (99). KIT signaling also promotes cell proliferation through association with PLCG at Tyr728 following activation of KIT by membrane bound KITL (99, 119). KIT signals through the JAK/STAT pathway, activating STAT1 in hematopoietic progenitors, STAT3 in myeloid cells, and STAT5 in mast cells (99).

SRC associates with activated KIT proximal to the interior plasma membrane and signals through LYN/FYN, RAC, GTPases and p38 MAP kinase to regulate cell chemotaxis and proliferation in hematopoietic cells (120, 121). Additionally SCR/LYN activation has affects on multiple other pathways including phosphorylation of STAT3, activation of

JNKs, inhibition of PI3K/AKT, and internalization of the KIT receptor (122). KIT signaling stimulates MAPK3/1 and p38 through association with GRB2 at phosphorylated Tyr703, signaling through son-of-sevenless (SOS), G-proteins RAS and

RAF, MEK1, MAPK3/1, and p38 to regulate gene expression and cell proliferation (99).


In early spermatogonia KIT signaling induces mitosis through activation of

MEK/MAPK3/1 (99, 123).

1.3.5 Affects of Kitl mutations on PGC development.

Several mutations in the Kitl gene have been described in the mouse. The Steel-

grizzle belly (Slgb) allele arises from a deletion of the Kitl coding sequence and leads to

embryonic lethality by E15 and a lack of PGCs by E10.5 (87, 124). Embryos

homozygous for the Slgb allele are observed to have severe PGC deficiencies at E9.5 (87).

Many hypomorphic mutations of Kitl have also been characterized in mice including a mutation that leads to the generation of a truncated Kitl transcript. The Steel-Dickie (Sld)

mutation results in a truncated KITL protein that lacks the transmembrane and

cytoplasmic domains (Figure 1-2) (110). As a result mice homozygous for the Sld allele fail to express the membrane bound form of KITL and are viable but sterile. Analysis of

PGC defects during development reveals PGC deficits as early as E9.5 at levels similar to the Slgb mutants (87). The number of PGCs decreases from E9.5 to E10.5 ultimately

resulting to few, if any, PGCs present in the genital ridges at E11.5. Only about half of

the PGCs present in Sld homozygous embryos migrate out of the hindgut, and of those the majority fail to colonize the genital ridges. The PGC phenotype of the Sld homozygous

embryos suggests that membrane bound KITL may be the form required for PGC

survival and migration. However, given the paucity of PGCs at the onset of migration

and the lack of PGCs successfully emerging from the hindgut, it is difficult to make

conclusions about the role of membrane bound KITL during PGC migration. Phenotypes

observed in Sld embryos may be a result of an earlier requirement for membrane bound


KITL and this early requirement must be circumvented in order to observe possible defects occurring during PGC migration.

1.3.6 Role of Kitl in PGC migration.

During PGC migration from the hindgut to the genital ridges beginning at E9.0,

Kitl is expressed by the somatic cells along the PGC migration route as a gradient with the highest levels in the genital ridges (92). This expression pattern could imply a role of

KITL as a chemo-attractant of PGCs. The genital ridges likely express factors that direct the PGCs to their target. This was shown through an under-agarose assay by which cultured PGCs were observed migrating towards a genital ridge explant, suggesting the genital ridges release a soluble attractant (17). However, subsequent experiments have both rejected and supported KITL as a PGC attractant. Initially, in a follow up under- agarose experiment, soluble KITL alone was not capable of attracting cultured PGCs, indicating that the chemo-attractant factor expressed by the genital ridges was likely not

KITL (112). Later work has found that PGCs seeded on the top of a porous membrane in a modified Boyden chamber will migrate across the membrane to a higher concentration of KITL, providing evidence in favor of a model consisting of KITL as a PGC attractant

(20).

While the role of KITL as a chemo-attractant may be debatable, the chemo- kinetic capability of KITL is well described. Addition of exogenous KITL to cultured

PGCs increases the percentage of PGCs displaying a motile phenotype (20, 112).

Activation of PI3K/AKT and Src Kinase was observed in response to addition of KITL and inhibition of these pathways blocked or reduced KITL induced PGC migration (20).


In vivo, Kitl null mutants lack PGCs by the migratory stage making it difficult to study

the role of KITL in PGC migration in vivo. Analysis of apoptosis deficient Bax-/- Kitl-/- mice revealed that surviving PGCs mostly remained in the hindgut at E10.5, a position typical of E9.0 PGCs (15).

1.3.7 PGC-Soma adhesion.

The interaction between migratory PGCs and the somatic environment is important for germ cell development. Adhesion between PGCs and their surroundings has been shown to be necessary for maintaining PGC survival in cell culture and for migration in vivo. PGCs begin to express E-cadherin after they emerge from the hindgut around E9.5 and continue to express it as they colonize the genital ridges (125).

Exogenous treatment of E10.5 tissue slices with an E-Cadherin blocking antibody,

ECCD2, resulted in a failure of PGCs to condense in the genital ridges and an increase in the number of ectopic PGCs remaining in the midline. PGC survival in culture is highly dependent on PGC adhesion to an artificial membrane and removal of this adhesion leads to decreased viability and increased apoptosis (126). Furthermore, membrane bound

KITL has been implicated as being involved in cell adhesion (106). Cultured Kit expressing mast cells require KITL and the extracellular domain of KIT to adhere to Kitl expressing fibroblasts (127). Kitl expressing sertoli cells are able to bind purified Kit expressing spermatogonia in cell culture (98). However, sertoli cells isolated from Sl/Sld mice fail to express membrane bound KITL and are unable to bind purified spermatogonia. This implies that interactions between membrane bound KITL and KIT can act as an adhesion mechanism between PGCs and somatic cells. Additionally,


cultured PGCs adhere more tightly to feeder cells expressing the membrane bound form

of KITL and this adhesive interaction can be reduced or abolished by addition of soluble

KITL or inhibition of the KITL/KIT interaction (128). Taken together, cell culture data

strongly suggests that in addition to other factors, such as E-Cadherin, PGCs engage in

adhesive interactions with their somatic environment through membrane bound KITL.

In this thesis I have investigated how BMP signaling in the E9.5 genital ridges

establishes a PGC niche. I will describe how mesonephric mesenchyme expressed BMPs

support pronephric survival while restricting epithelial expansion and promote expression

of Kitl within the coelomic epithelium. High levels of the membrane bound form of

KITL are necessary for PGC survival and serve to anchor PGCs to the somatic cells of the genital ridge. I propose that the genital ridge serves as a PGC niche, built and maintained by BMP signaling, providing necessary survival factors including KITL, and holding PGCs within the microenvironment through membrane bound KITL mediated adhesion.


Chapter 2: BMP Signaling Regulates PGC Numbers and

Motility in Organ Culture

Brian M Dudley1, Chris Runyan2, Yutaka Takeuchi3, Kyle

Schaible2, Kathleen Molyneaux1

1Department of Genetics, Case Western Reserve University, 10900 Euclid Ave.

Cleveland, OH 44106.

2Division of Developmental Biology, Cincinnati Children’s Hospital Research

Foundation, 3333 Burnet Ave. Cincinnati, OH 45229.

3Department of Aquatic Biosciences, Tokyo University of Fisheries, 4-5-7 Konan,

Minato-ku, Tokyo 108-8477.

This chapter is a modified version of the previously published:

Dudley BM, Runyan C, Takeuchi Y, Schaible K, Molyneaux K. BMP signaling regulates

PGC numbers and motility in organ culture. Mech Dev. 2007 Jan;124(1):68-77.

Note: I personally completed real time PCR experiments represented in figures 2-5 and

2-6, and housekeeping gene experiments referenced in section 2.4.5.


2.1 Abstract

Members of the bone morphogenetic protein (BMP) family play diverse roles in multiple developmental processes. However, in the mouse, mutations in many BMPs,

BMP receptors and signaling components result in early embryonic lethality making it difficult to analyze the role of these factors during organogenesis or tissue homeostasis in the adult. To bypass this early lethality, we used an organ culture system to study the role of BMPs during primordial germ cell (PGC) migration. PGCs are the embryonic precursors of the sperm and eggs. BMPs induce formation of primordial germ cells within the proximal epiblast of embryonic day 7.5 (E7.5) mouse embryos. PGCs then migrate via the gut to arrive at the developing gonads by E10.5. Addition of BMP4 or the BMP-antagonist Noggin to transverse slices dissected from E9.5 embryos elevated

PGC numbers or reduced PGC numbers respectively. Noggin treatment also slowed and randomized PGC movements, resulting in a failure of PGCs to colonize the urogenital ridges (UGRs). Based on p-SMAD1/5/8 staining, migratory PGCs do not respond to endogenous BMPs. Instead, the somatic cells of the urogenital ridges exhibit elevated p-

SMAD1/5/8 staining revealing active BMP-signaling within the UGRs. Noggin treatment abrogated p-SMAD staining within the UGRs and blocked localized expression of Kitl, a cytokine known to regulate the survival and motility of PGCs and Id1, a transcription factor expressed within the UGRs. We propose that BMP signaling regulates PGC migration by controlling gene expression within the somatic cells along the migration route and within the genital ridges.


2.2 Introduction

Bone morphogenetic proteins (BMPs) are members of the transforming growth

factor
(TGF-
) superfamily. Members of this family signal through heteromeric

complexes composed of type I and type II serine-threonine kinase receptors. Binding of

BMPs to their receptors induces phosphorylation of the BMP-specific SMADs (SMAD1,

SMAD5, SMAD8). p-SMAD1/5/8 then associate with SMAD4, translocate into the

nucleus, and activate the transcription of BMP-target genes (129, 130). In the mouse,

targeted or spontaneous mutations in various Bmps, Bmp receptors and Smads have

confirmed the roles of bone morphogentic proteins in the formation of the skeletal system

(131-134) and mesoderm induction (135-138), as well as development of the heart,

nervous system, and urogenital system (139).

BMPs have known roles in germ cell development and function. BMP2 (3),

BMP4 (5) and BMP8b (8) control the formation and early proliferation of primordial

germ cells (PGCs). These factors are thought to be released from extraembryonic tissues

and create a permissive environment for PGC formation in the epiblast (140).

Additionally, members of the BMP family are required for germ cell function in adult

mice. Targeted deletion of Gdf-9 blocks the growth of primary follicles resulting in infertile females (78). Additionally, targeted deletions of Bmp15 and Bmpr1b/Alk6 cause

female infertility due to defects in ovulation and cumulus cell expansion (141) and

mutations in Bmp7, Bmp8a, Bmp8b, and Bmp4 block spermatogenesis (142-144). The requirement for BMPs in fertility is conserved in a number of species. Two Bmp

orthologues, Glass-bottom boat (Gbb) and Decapentaplegic (Dpp), are expressed in the


cap and sheath cells of the Drosophila ovary and regulate differentiation of the germ-line stem cells (59). Gbb and Dpp play similar roles in the fly testis (60).

BMPs are important in controlling PGC formation and gametogenesis, but the

early lethality caused by loss of Bmp4 (70), the BMP type I receptors (Bmpr1a and

Acvr1) (136, 137, 145), the BMP type II receptor (Bmpr2) (135), and SMADs (SMAD1,

SMAD5, SMAD4) (146-148) precludes analysis of BMP signaling during much of germ

cell development. Additionally, redundancy amongst BMP family members and

receptors complicate genetic analysis in this system. To bypass these complications, we

used an organ culture system to study the role of BMP signaling in germ cell migration.

After formation, PGCs proliferate and migrate via the gut to colonize the developing

gonads (14). Migratory PGCs express BMP receptors and SMADs (149) and may be

exposed to BMP signals emanating from the lateral mesoderm or the extraembryonic

mesoderm where Bmp2, Bmp4, Bmp5 and Bmp7 are expressed (70, 150, 151). BMP2 and BMP4 were shown to simulate the migration of osteoblast progenitors (152) and it is possible that BMPs might play a role in germ cell guidance. To test this, BMPs or the

BMP-inhibitor Noggin were added to tissue dissected from E9.5 embryos. BMP treatment increased PGC numbers and Noggin treatment reduced PGC numbers and slowed PGC migration. Additionally, BMPs induced and Noggin repressed expression of

Kitl and Id1. We propose that BMPs regulate the number and position of migratory

PGCs by controlling expression of genes along the migratory route and within the urogenital ridges.


2.3 Materials and Methods

2.3.1 Organ culture.

All animal procedures have been approved by the Case Western Institutional

Animal Care and use Committee. Oct4
PE:GFP +/+ homozygous males (153) established on the FVB background were crossed to CD1 females (Charles River

Laboratories). Embryonic day 0.5 (E0.5) was assumed to be noon on the day that a vaginal plug was detected. E9.5 embryos were harvested and slices were cut and cultured in millicell Organ culture chambers pre-coated with collagen IV as previously described

(18). Recombinant human BMP4 (R&D systems) or mouse Noggin-FC (R&D systems) was added to the medium at the indicated concentrations. Germ cells were counted at

T=0 hours and T=18 hours by optically sectioning the slices as described (18). Slices were filmed using the LSM510 confocal system and germ cell movements were quantitated as described (18). Briefly, one frame was captured every 7 min. for 7.5 hrs.

(63 frames). Individual germ cells (six cells per movie) were manually traced using NIH

Image (http://rsb.info.nih.gov/nih-image/Default.html). The position of a germ cell was determined every 35 min. (5 frames) generating a trace and a velocity measurement.

Velocity measurements (12 per cell) were averaged to yield the average velocity of that cell (Avg. V.). At the end of the trace, the maximum velocity of the cell was recorded

(Max. V) (fastest of the twelve measurements) and the absolute distance that the cell moved from its starting point was recorded (displacement, D.). For an individual slice, data from the six germ cells that were tracked were averaged to yield the Avg. V., Avg.

Max. V. and Avg. D. for that slice.


2.3.2 Immunostaining.

E9.5, E10.5, and E11.5 embryos were dissected from the uterus in PBS/2% FCS.

The heads were removed and the bodies were fixed overnight in 4% PFA/PBS. Embryos were washed in PBS/0.1% TX-100 and a scalpel was used to cut transverse slices from the trunk of the fixed embryos. Slices (~200 mm thick) were permeablized at 4°C o/n in

PBS/0.1% TX-100, and then blocked overnight in 2% donkey serum/PBS (for donkey secondary antibodies) or 2% goat serum/PBS (for goat secondary antibodies). Slices were incubated (overnight at 4oC) with primary antibodies used at a 1:100 dilution in blocking buffer. Rabbit polyclonal antibodies against p-SMAD1/5/8 were generously supplied by Peter ten Dijke. Goat polyclonal antibodies against BMP2/4 (sc-6267) were purchased from Santa Cruz BioTech. Slices were washed 5x 1 hr. in PBS/0.1% TX-100 and incubated (overnight at 4oC) with 1:100 Cy5-conjugated Goat anti-rabbit (Jackson

Immuno Research) or 1:100 Cy5-conjugated Donkey anti-goat secondary antibodies

(Jackson Immuno Research) diluted in blocking buffer.

2.3.3 Semi-quantitative RT-PCR.

Germ cells and somatic cells were isolated by flow cytometry as previously described (149). cDNAs were prepared from 24 ng of mRNA using the Superscript III kit (Invitrogen). Semi-quantitative RT-PCR was performed as described (149). Primers were designed using Primer 3 (http://frodo.wi.mit.edu/cgi- bin/primer3/primer3_www.cgi) and are shown in Table 2-1 of the supplemental material.


2.4 Results

2.4.1 PGCs and somatic cells express BMP receptors and Smads during germ cell

migration.

BMPs signal through a complex containing type I and II serine/threonine kinase

receptors. Upon binding BMPs, type II receptors phosphorylate and activate type I

receptors, which in turn phosphorylate and activate SMADs. Based on data from a

screen of the Affymetrix MGU-74av2 chip (149), both migratory (E10.5) and

postmigratory (E12.5) PGCs express mRNAs for Bmp4, Bmpr1a (Alk3) and Smad4.

Smad1 (based on probe set 102984_g_at) was expressed in migratory and postmigratory

PGCs; however an additional probe set for Smad1 (102983_at) gave weaker and less

reproducible results. The other type I receptors (Acvr1/Alk2 and Bmpr1b/Alk6), the type

II receptors (Bmpr2, Acvr2 and Acvr2b), Smad5 and the inhibitory Smads (Smad6 and

Smad7) were not reproducibly detected in PGCs using the Affymetrix chips.

As a more sensitive method, we used quantitative RT-PCR to examine the

expression of Bmp4, BMP receptors, and Smads in PGCs and somatic tissue at E10.5

(Figure 2-1). PGCs and the adjacent soma expressed Bmp4, Bmp receptors and Smads

with the majority of these transcripts being more highly expressed in the soma than in

PGCs. The germ cell marker, Stella was absent from the somatic fraction (data not shown); whereas Cxcr4, a G-protein coupled receptor implicated in PGC guidance was enriched in germ cells, but present in the soma. The somatic marker Kitl was enriched in the somatic fraction. From this expression data it is likely that both PGCs and the


somatic tissue are capable of responding to BMP-signals at this time point, but the

majority of the BMP signaling components are more highly expressed in the soma.

2.4.2 There is a BMP-signaling center within the mesonephric mesenchyme at E9.5 and

E10.5.

Based on in situ hybridization results, Bmp4 is expressed within the urogenital ridge (154) and BMP4-signaling is thought to regulate the outgrowth of the ureteric bud

(155). Immunostaining using an antibody that cross-reacts with BMP2 and BMP4 revealed that BMP protein is enriched in the mesenchyme and epithelium of the urogenital ridges at E10.5 (Figure 2-2 A and B). To identify cells responding to BMPs, we stained slices with an antibody that recognizes the phosphorylated epitope of

SMAD1, SMAD5 and SMADd8 (p-SMAD1/5/8) (156). At E9.5, cells in the epithelium of the urogenital ridge as well as the adjacent mesenchyme stained positive for p-

SMAD1/5/8 (Figure 2-2C). By E10.5, p-SMAD1/5/8 staining was no longer detected in the epithelium of the genital ridge. Instead, high levels of staining were observed in the condensing mesonephric mesenchyme (Figure 2-2D). By E11.5, p-SMAD1/5/8 staining was absent from the developing kidney, but weak staining was detected within gonadal

PGCs (Figure 2-2E). p-SMAD1/5/8 staining was not detected in migratory PGCs (stages

E9.5 and E10.5) (Figure 2-2C and D), which appeared to accumulate in a region of the genital ridge devoid of p-SMAD1/5/8 staining (Figure 2-2D). These data suggest that germ cells do not respond directly to BMP signals during migration, but the surrounding somatic cells do. In particular, somatic cells within the mesonephric mesenchyme exhibited robust staining for p-SMAD1/5/8.



Figure 2-1: BMP signaling components are enriched in the soma at E10.5. PGCs and somatic cells were purified from E10.5 Oct4
PE:GFP embryos using flow cytometry. The somatic tissue in these sorts represents cells derived from the gut, gut mesentery, ventral body wall (including the urogenital ridges) and notocord. cDNA was prepared from PGCs (GFP+) and somatic (GFP-) samples and used for quantitative RT-PCR. For each marker, the somatic sample was used to generate a standard curve (expression set to 100%). Expression of each marker was normalized to the expression of Odc. Bmp4, Bmp receptors (Bmpr1A, Bmpr1B, Acvr1, Bmpr2, Acvr2, Acvr2b), Smad1, Smad4 and Smad5 were enriched in the soma (compare expression levels with the somatic marker Kitl). Cxcr4, the receptor for the PGC guidance factor Sdf-1 and Smad8 were enriched in the PGCs. Expression of Smad7 was not detected in either the somatic or PGC fractions.



Figure 2-2: Cells of the mesonephric mesenchyme respond to endogenous BMPs. Tissue dissected from E10.5 Oct4
PE:GFP embryos was fixed and stained with naïve goat IgG (A) or anti-BMP2/4 (B). Immune complexes were detected with donkey anti-goat Cy5 and visualized by confocal microscopy. Diffuse staining for BMP2/4 was present throughout the slices, but highly concentrated in both the epithelial and mesenchymal components of the mesonephros. a (aorta) and gr (genital ridge). Slices were dissected from E9.5 (C), E10.5 (D) and E11.5 (E) Oct4
PE:GFP embryos and stained with anti-p- SMAD1/5/8. At E9.5, both epithelial and mesenchymal components of the genital ridge exhibited elevated p-SMAD1/5/8. By E10.5, p-SMAD1/5/8 staining was highly enriched within the condensing mesonephric mesenchyme (mm), but reduced in the ventral portion of the genital ridge (where PGCs accumulate). By E11.5, p-SMAD1/5/8 staining was absent in the mesonephros, but was weakly detected in the PGCs. g (gonad). Scale bars are 40 mm (A) and (B), 30 mm (C) and (D), and 70 mm (E).


2.4.3 Perturbing BMP signaling affects PGC numbers in E9.5 slice culture.

At E9.5, PGCs emerge from the gut and actively migrate towards the genital

ridges. Chemoattractants and survival factors released by the genital ridges are thought

to control this process. To assess whether BMP-signaling affects PGC migration

recombinant human BMP4 or recombinant Noggin were added to slices dissected from

E9.5 embryos (Figure 2-3). PGC numbers were quantified by optically sectioning the

tissue at the start point (T0) and end point (T18 hours) of the culture period (Figure 2-3A

and B), and expressing the number of germ cells after 18 hours of culture as a percentage

of the number at T0, in order to normalize results between slices containing different

numbers of germ cells. After treatment, slices were fixed and stained for p-SMAD1/5/8

to verify the suppression or activation of BMP-signaling (Figure 2-3C-H).

1000 ng/ml Noggin reduced PGC numbers in E9.5 slices (Figure 2-3A). Control

slices had little change in PGC numbers (average of 96.4 + 11.4% s.e.m.) whereas slices

treated with 1000 ng/ml Noggin had reduced PGC numbers (average of 66.7 + 7.7% s.e.m.). This effect was statistically different than controls (ANOVA followed by

Fisher’s LSD post test). High doses of Noggin (2000 ng/ml) appeared to increase PGC numbers (although this effect was not significantly different than controls).

Treating E9.5 slices with BMP4 also resulted in a bi-phasic effect on PGC numbers (Figure 2-3B). Low doses (0.5 ng/ml and 5 ng/ml) increased PGC numbers, whereas higher doses (50 ng/ml and 500 ng/ml) had no effect or actually reduced PGC



Figure 2-3: BMP signaling regulates PGC numbers. The percent change in PGC numbers was determined in slices treated for 18 hrs. with the indicated concentrations of Noggin (A) or BMP4 (B). Moderate doses of Noggin reduced PGC numbers, whereas moderate doses of BMP4 increased PGC numbers in slices dissected from E9.5 Oct4
PE:GFP embryos. n indicates the number of slices in each treatment group. Error bars are s.e.m. Data was analyzed by ANOVA followed by the Fisher’s least significant difference test to determine significant changes. * indicates samples differing from controls (post test p<0.05). ** indicates samples differing from controls (post test p<0.01). Slices treated with 1000 ng/ml Noggin (C) and (D), control slices (E) and (F) or slices treated with 10 ng/ml BMP4 (F) and (G) were fixed and stained with anti-p-SMAD1/5/8. Noggin treatment reduced p-SMAD1/5/8 staining. BMP4 treatment elevated p- SMAD1/5/8 staining over the entire slice (including in PGCs). In control slices, p-SMAD1/5/8 staining was elevated in the mesonephric mesenchyme (mm), but was not observed in PGCs. gr (genital ridge). md (mesonephric duct). Scale bar for (C) (E) and (G) is 115 mm. Scale bar for (D) (F) and (H) is 52 mm. The large arrow points towards the gut and indicates the ventral axis of the slice.


numbers. In the BMP experiments, control slices exhibited an average change in PGC number of 73.1 + 9.1% s.e.m. and the most effective dose of BMP4 (5 ng/ml) had an average change in PGC number of 120.0 + 13.9% s.e.m. This effect was statistically different than controls (ANOVA followed by Fisher’s LSD post test).

p-SMAD1/5/8 staining confirmed the effects of the BMP agonists and antagonists in the slice cultures. Noggin treatment reduced p-SMAD1/5/8 staining (Figure 2-3C and

D), and abolished the high level of response to BMPs in the mesonephros (compare 2-3D and F). Conversely, BMP4 treatment elevated p-SMAD1/5/8 staining in every cell of the slice including PGCs (Figure 2-3G and H).

2.4.4 Noggin treatment suppresses PGC motility.

In addition to affecting PGC numbers, Noggin treatment also affected germ cell position in slice cultures (Figure 2-4A), whereas BMP4-treatment did not (data not shown). In control slices, the majority of PGCs moved away from the midline forming two distinct clusters at the genital ridges. PGC migration is inefficient and a few PGCs remained stranded in the midline region (arrows in Figure 2-4A). However, in slices treated with 1000 ng/ml Noggin there was a dramatic accumulation of PGCs across the midline region of the slice (boxed areas in Figure 2-4A). In the experiment shown, 75%

(3/4) of the Noggin-treated slices were affected. Similar results were seen in two additional experiments (data not shown).

Accumulation of PGCs in midline structures indicates a defect in PGC migration

(18). To determine if Noggin treatment affects the direction and/or speed of PGC migration, we performed time-lapse analysis of PGC behavior in control and Noggin-


treated slices (Figure 2-4B and C). Four control slices and four Noggin-treated slices

were filmed and cell trajectories and velocity measurements were calculated (see

materials and methods). Figure 2-4B shows start- and end-point pictures of a

representative control and Noggin-treated slice. In the control slice, PGC trajectories

were long and straight. In the Noggin-treated slice, PGC trajectories were shorter and

twisted. PGCs starting in the midline were more severely affected than PGCs starting

closer to the genital ridges. Six cells were traced per movie to yield average velocity,

average maximum velocity and average displacement measurements for each slice

(Figure 2-4C). PGCs in control slices had an average velocity of 21.3 + 3 s.d. mm/hr, an average maximum velocity of 46.3 + 10.4s.d. mm/hr. and an average displacement of

111.8 + 27.4 s.d. mm. Noggin treatment reduced the average velocity, maximum velocity and displacement of PGCs. PGCs in Noggin-treated slices had an average velocity of 15.9 + 4.8 s.d. mm/hr., which was statistically different than controls

(Student’s t-test p=0.00009). The average maximum velocity of PGCs in Noggin-treated

slices was 38.7 + 15.2 s.d. mm/hr (t-test p=0.051). Displacement in Noggin-treated slices

was severely reduced with the average displacement being 66.5 + 32.6 s.d. mm (t-test p=0.000004). We conclude that Noggin treatment randomizes and slows PGC migration.

2.4.5 Exogenous BMP4 and BMP5 elevate expression of Kitl and Id1 in organ culture.

Based on the p-SMAD staining and organ culture data, BMP4 is unlikely to act as a chemoattractant for PGCs however BMP signaling to the soma (eg. the mesonephric mesenchyme) may control expression of chemoattractants and survival factors that regulate PGC behavior. To test this, the expression of potential BMP-target genes was



Figure 2-4: Noggin treatment slows and randomizes PGC movements. Slices were dissected from E9.5 Oct4
PE:GFP embryos and cultured for 18 hrs. In control slices (A), the majority of PGCs cleared the midline and formed clusters at the genital ridges. A few PGCs remained near the midline (arrows). In Noggin-treated (1000 ng/ml) slices a large number of PGCs remained scattered across the midline (boxed areas). This effect was observed in 9 out of 12 slices from 3 experiments. g (gut). Four control and four noggin treated slices were filmed for 144 minutes (7.4 hrs.) and PGC movements were traced (6 cells per movie). The start point and endpoint of a representative control slice and a representative Noggin-treated slice are shown in (B). Germ cell trajectories are overlaid in red. The average velocity, average maximum velocity and average displacement of the six traced cells are shown in (C). Movies are available as supplemental material. Each bar represents an individual slice. Error bars are standard deviation in order to display the fact that PGCs, even in control slices, exhibited a wide range of behaviors. Noggin treatment produced a statistically significant decrease in Avg. velocity (t-test p<0.001) and displacement (t-test p<0.001). Again, note that three out of four slices were affected.


examined in organ culture using real time RT-PCR. In initial experiments, results were

normalized using Ornithine decarboxylase (Odc) as a loading control; however, Odc levels were found to drop in culture (relative to input RNA) (data not shown).

Additionally, Odc levels were found to decline in vivo between E9.5 and E10.5 (data not shown). This makes Odc an unsuitable loading control for establishing a time course of gene expression at these stages. Instead, RT-PCR results were normalized using the expression level of TATA-binding protein and b-actin, two housekeeping genes whose expression levels remain steady between E9.5 and E10.5 (data not shown).

Both BMP4 and BMP5 induced expression of Kitl (Figure 2-5) in organ culture.

KITL is a known survival factor for PGCs. Kitl expression was induced after 3 hours in culture, but peaked between 6 and 18 hrs. This relatively slow induction suggested to us that BMP-regulation of Kitl might go through an intermediate step. Several transcription factors are expressed in the early genital ridge and have known roles in kidney development and/or development of the somatic components of the gonad. Quantitative

RT-PCR was used to examine the effect of exogenous BMPs on transcription factors expressed within the urogenital ridge (Pax2, Sf1, Lhx1, and Id1). Both BMP4 and BMP5 treatment induced rapid induction of Id1 (Figure 2-5), a helix-loop helix factor and known BMP target (157). BMP treatment slightly reduced expression of Pax2, Sf1 and

Lhx1 (data not shown). Id1 was identified in an in situ screen for factors expressed within the urogenital ridge (158). In addition, Id1 is implicated in controlling cell migration (157) and pluripotency (159).


2.4.6 Noggin treatment reduces Id1 and Kitl expression within the urogenital ridge.

BMP signaling is necessary for PGC formation and early patterning of the embryo. To examine the role of BMP signaling in PGC migration in vivo requires a genetic system with very fine control over the timing at which the BMP signals are ablated in order to bypass this early requirement and yet efficiently and specifically inactivate all BMP signaling by E9.5. Instead, we have again taken an organ culture approach to address whether endogenous BMP signals regulate gene expression within the urogenital ridges. Slices were treated with Noggin for 18 hrs. Following treatment, the UGRs were dissected and expression levels of Kitl and Id1 were quantified in cDNA prepared from UGR and non-UGR tissue samples (see Figure 2-6). Kitl expression increased an average of 3.8x in the UGRs during culture (Figure 2-6C). Kitl expression also increased within the UGRs in vivo (6x comparing E10.5 genital ridges to E9.5 genital ridges). This increase was specific to the genital ridges. Adding Noggin to slices inhibited the increase in Kitl within the UGRs. Id1 expression increased in culture within both the genital ridges and in the rest of the slice albeit this increase (an average of 1.5x) was less dramatic than the increase in Kitl expression (Figure 2-6D). Noggin treatment suppressed this induction and reduced Id1 levels to 0.4-0.6x starting levels.

2.5 Discussion

A founding population of approximately 50 PGCs forms at the posterior end of the mouse embryo at E7.5 (5). From there PGCs migrate and proliferate until ~1000 cells colonize the genital ridges by E10.5. Chemoattractants, repellents, adhesive gradients, and/or localized survival factors have all been proposed to regulate PGC



Figure 2-5: BMP treatment induces expression of Kitl and Id1. E9.5 Slices were treated with the indicated concentrations of BMP4 or BMP5. Slices were harvested at 0, 3, 6 or 18 hrs. after treatment and the expression of Kitl and Id1 were compared using quantitative RT-PCR. The T0 samples were used to generate the standard curve (expression set to 100). A normalization factor was generated by averaging the expression of TATA-binding protein and bACT. Data is an average of five experiments using BMP4 and five experiments using BMP5. Data was analyzed using ANOVA followed by Fisher’s least significant difference test. * indicates samples that differ from T0 values (100) (p<0.1). ** indicates samples that differ from T0 values (p<0.01) Error bars are s.e.m.



Figure 2-6: Noggin treatment represses Kitl and Id1 expression in the genital ridges. E9.5 slices (four slices per treatment group) were treated with 100 ng/ml (0.1 N), 1000 ng/ml (1 N), or 2000 ng/ml (2N) of Noggin. After 18 hrs. in culture the genital ridges were dissected away from the rest of the body wall and assayed for expression of Kitl and Id1 using real time PCR. (A) Schematic of how the dissections are performed. (B) A confocal image of a slice showing dissection of the genital ridge on one side. (C) Real time PCR expression levels for Kitl. (D) Real time PCR expression levels for Id1. Data is an average of three experiments except for the E10.5 data which represents a single comparison. In each experiment the T0 genital ridge sample was used to generate a standard curve (expression level set to 100). A normalization factor was generated by averaging the expression of TATA-binding protein and bAct. Data was analyzed using ANOVA followed by Fisher’s least significant difference test. * indicates samples that differ from T0 genital ridge values (100) (post test p<0.1). ** indicates samples that differ from T0 genital ridge values (post test p<0.01) Error bars are s.e.m.


migration (reviewed in (149)), but molecular evidence for these processes is scant. In a

screen to identify transcripts expressed in migratory PGCs, we found that PGCs express

BMP receptors and hence might be able to respond to endogenous BMPs expressed in the

lateral plate and developing urogenital ridges. BMP signaling is required for PGC

formation and mice lacking Bmp4 (5), Bmp2 (3), Bmp8b (8), Smad1 (147) or Smad5

(160) have a reduced founding population of PGCs as well as defects in mesoderm formation and development of the allantois. In order to address the role of BMPs in germ cell migration, we used an organ culture system to bypass the early role of BMPs in PGC formation and embryonic patterning.

Quantitative RT-PCR revealed that migratory (E10.5) PGCs and the surrounding somatic tissue express Bmp4, Bmp receptors and Smads and that these transcripts were more abundant in the soma than in PGCs. At E9.5, p-SMAD staining was detected in somatic cells along the migratory route and by E10.5 intense p-SMAD staining was observed within the condensing mesonephric mesenchyme within the urogenital ridges.

Surprisingly, PGCs did not exhibit p-SMAD staining until after colonizing the UGRs

(E11.5). A similar expression pattern was observed by Chuva de Sousa Lopes et al.

(161). This suggests that migratory PGCs are not responding to BMP signals or do so using a non-SMAD pathway such as ERK or p38 MAPK. Takeuchi et al. (162) have shown that ~15% of migrating PGCs exhibit pp-ERK, but this signaling is likely due to

FGFs and not BMPs. Migratory PGCs did not stain for phospho-p38 MAPK (data not shown).

Although migratory PGCs did not appear to be responding to endogenous BMPs, perturbing BMP signaling affected both PGC numbers and motility in an organ culture


assay. Addition of 5 ng/ml of BMP4 increased PGC numbers whereas high doses

decreased PGC numbers. Noggin treatment had the converse effect, 1000 ng/ml of

Noggin decreased PGC numbers whereas high doses slightly increased PGC numbers.

1000 ng/ml Noggin treatment also reduced the speed of and partially randomized PGC

movements causing a large number of germ cells to remain on the midline of cultured

slices. BMP4-treatment did not have a significant effect on the direction or speed of

germ cell migration (data not shown). This suggests that BMPs do not act directly as

chemo-attractants for PGCs.

If migratory PGCs are not responding to BMPs, why should Noggin treatment

affect germ cell numbers and motility? Addition of BMP4 and BMP5 induced

expression of Kitl and Id1 in cultured slices. Conversely, Noggin treatment inhibited Id1 expression in slices and blocked expression of Kitl specifically within the urogenital ridges. KITL and its receptor KIT control PGC survival and proliferation during migration (reviewed in (163)). Strong mutations in both the receptor and ligand result in a complete loss of PGCs early in development (164). Additionally, changing expression patterns of Kitl may control germ cell motility by either attractive (via the secreted form of KITL) or adhesive (via the membrane bound form of KITL) mechanisms. Godin et al.

(112) have shown that soluble KITL does not attract PGCs in culture, but certain mutations in Kitl cause mislocalization of PGCs within the gut (89) in addition to affecting PGC numbers. Because of the role of KITL in PGC survival and proliferation it has been difficult to draw conclusions about whether this protein has a specific effect on

PGC motility. However recently, Mahakali Zama et al. (87) used an allelic series of


mutations in the Kitl gene to demonstrate that the role of this protein in PGC survival/proliferation may be separable from its role in migration.

Our data suggest a model whereby BMP signaling regulates PGC behavior by controlling expression of Kitl within the somatic tissue along the migration route. BMPs are required at multiple stages of germ cell development and in many cases mediate their effects at least in part by control of the soma. de Sousa Lopes et al. (140) have shown that BMP signals from the extraembryonic ectoderm regulate PGC formation by controlling the release of secondary signals from the visceral endoderm. Recently,

Kirilly et al. (165) showed that BMPs in the fly ovary not only signal directly to the developing oocyte, but are also required to control proliferation of the somatic cells within the germ cell niche. Likewise in the mammalian ovary, BMP15 and GDF9 released from the oocyte control proliferation of granulosa cells (166). Intriguingly,

BMP15 and GDF9 have opposing effects on Kitl expression. BMP15 was shown to induce Kitl expression within granulosa cells (81) whereas GDF9 was shown to inhibit

Kitl expression both in vivo (82) and in ovary organ culture (167). We have shown that both BMP4 and BMP5 can induce Kitl expression in an organ culture system. It will be intriguing to see if there are additional BMP family members (perhaps GDFs) expressed within the trunk of the E9.5 embryos that are capable of downregulating Kitl expression.

If this is the case, it may explain one of the puzzling aspects of our organ culture data.

We have found that both Noggin and BMP4 have biphasic effects on germ cell numbers.

Moderate doses of BMP4 were found to increase PGC numbers whereas high doses actually reduced numbers. Noggin treatment had the converse effect. Alternatively, it may be possible that different levels of BMP signaling may have different effects on PGC


behavior. For example, it has been shown that low doses of BMP4 can induce hematopoietic stem cell precursors to differentiate whereas high doses promote self renewal (168).

In summary, cells along the migration route and within the developing urogenital ridges respond to BMPs; however signaling was not detected in migratory PGCs. Germ cells exhibited p-SMAD staining after colonizing the gonads. We propose that BMP signaling regulates PGC numbers and may control PGC position by controlling expression of genes along the migration route. Tissue specific targeting will be required to definitively address the role of BMPs in PGC migration in vivo. However, the large number of BMP family members and receptors makes this a challenging task. It may be necessary to target multiple BMP receptors and/or SMADs to reveal the impact of BMP signals on PGCs at this stage.

2.6 Acknowledgements

Funding support was provided by the March of Dimes (#5-FY05-1233). We thank

Peter ten Dijke for the p-SMAD1/5/8 antibody. We thank Greg Matera and Joeseph

Nadeau for critical reading of the manuscript and we acknowledge Patti Conrad for microscopy assistance.


2.7 Supplemental Material

Gene ID Forward Primer (5'-3') Reverse Primer (5'-3') b-actin AGAGGGAAATCGTGCGTGAC CAATAGTGATGACCTGGCCGT Acvr1 TGTACCGTTGGACTCTGCTG CCGTTTCAGAGTTGGGTGTT Acvr2 TGTGCGTTGTCTCTGTACCC AAGAGTTTGAGGGGCCAAAT Acvr2b GATACCCATGGACAGGTTGG CTTGTGGACAACCACCTCCT Bmp4 GAGCCTTTCCAGCAAGTTTGTTC CCATCAGCATTCGGTTACCAGG Bmpr1a GATGGCTGGTTGTGCTCATTTC AATCACGGTTGTAACGACCCC Bmpr1b TGAGTGTCTCAGGCAGATGG GAACTCACTGGGCAGTAGGC Bmpr2 AGGCCCAATTCTCTGGATCT CACTGCCATTGTTGTTGACC Cxcr4 AGCCTGTGGATGGTGGTGTTTC CCTTGCTTGATGACTCCCAAAAG Id1 CAACAGAGCCTCACCCTCTC AGAAATCCGAGAAGCACGAA Kitl CCATGGCATTGCCGGCTCTC CTGCCCTTGTAAGACTTGACTG Lhx-1 CCTACCCTTTGTGCCATCAT ACAAATGGTTCCCGTAGCTG Odc GCCATTGGGACAGGATTTGAC CATCATCTGGACTCCGTTACTGG Pax2 AAAGTTCAGCAGCCTTTCCA CCAGGTAGAGTGGTGCTCGT Sf-1 TACGACGACTACCACCAGGA GCGGAAAGTCCTCACTCTCA Smad1 ATGAGCTTCGTGAAGGGTTG CGGAAGCCACAGGTCTTTTA Smad4 AGCTCCAGCCATCAGTCTGT GAGCTCGGTGAAGGTGAATC Smad5 CCCACCACCGTCTGTAAGAT GACGTCCTGTCGGTGGTACT Smad6 CCACCAACTCCCTCATCACT CTGCCCTGAGGTAGGTCGTA Smad7 CCCGAGTGATTGCTTTTCAT GGGATGCTGTGAGCTGACTTTC Smad8 CTTCACCGACCCTTCCAATA TCTGGACAAAGATGCTGCTG Stella TGAGTTTGAACGGGACAGTG GATTTCCCAGCACCAGAAAA TBP CTTCGTGCAAGAAATGCTGA AGAACTTAGCTGGGAAGCCC

Supplemental Table 2-1: PCR Primers used for Real Time RT-PCR.


Chapter 3: BMP Signaling Controls Formation of a Primordial

Germ Cell Niche within the Early Genital Ridges

Brian Dudley1, Caterina Palumbo2, Jennifer Nalepka3, and

Kathleen Molyneaux1

1Department of Genetics, Case Western Reserve University, 10900 Euclid Ave.

Cleveland, OH 44106.

2College of Medicine, University of Toledo, 2801 Bancroft, Toledo, OH 43606.

3Athersys Inc., 3201 Carnegie Ave, Cleveland, OH 44115.

This chapter is a modified version of the previously published:

Dudley B, Palumbo C, Nalepka J, Molyneaux K. BMP signaling controls formation of a primordial germ cell niche within the early genital ridges. Dev Biol. 2010 Apr 22.

Note: I personally completed all experiments and analysis in this chapter except BMP in situ hybridization (Figure 3-1b and c), Rosa:LacZ/+ Pax2:Cre/+ staining (Supplemental

Figure 3-S2a), and Pax2:Cre/+ PGC counts (Supplemental Figure 3-S2b).


3.1 Abstract

Stem cells are necessary to maintain tissue homeostasis and the microenvironment

(a.k.a. the niche) surrounding these cells controls their ability to self renew or differentiate. For many stem cell populations it remains unclear precisely what cells and signals comprise a niche. Here we identify a possible PGC niche in the mouse genital ridges. Conditional ablation of Bmpr1a was used to demonstrate that BMP signaling is required for PGC survival and migration as these cells colonize the genital ridges.

Reduced BMP signaling within the genital ridges led to increased somatic cell death within the mesonephric mesenchyme. Loss of these supporting cells correlated with decreased levels of the mesonephric marker, Pax2, as well as a reduction in genes expressed in the coelomic epithelium including the putative PGC chemo-attractants Kitl and Sdf1a. We propose that BMP signaling promotes mesonephric cell survival within the genital ridges and that these cells support correct development of the coelomic epithelium, the target of PGC migration. Loss of BMP signaling leads to the loss of the

PGC target resulting in reduced PGC numbers and disrupted PGC migration.


3.2 Introduction

In the adult, the tissue homeostasis and repair of multiple organ systems are controlled by resident stem cells. These stem cell populations divide slowly giving rise to new stem cells or to daughter cells that differentiate to replace tissue lost to aging or injury. The development of stem cell populations and their subsequent ability to renew or differentiate are controlled by an elaborate array of signaling interactions between the stem cells and their surrounding support cells. How these niches arise during embryogenesis is unclear but the process requires careful coordination of both stem cell and surrounding support cell development in order to insure long-term tissue homeostasis in the adult.

Germ cells have emerged as an ideal system to study stem cell biology (169).

Germ cells are the precursor cells of the gametes. In the mouse embryo, primordial germ cells (PGCs) are induced to form within the posterior proximal epiblast between day 6.5 and 7.5 (E6.5-E7.5) of gestation. Once formed, PGCs migrate through the posterior primitive streak and become incorporated into the invaginating hindgut (13). PGCs remain in the hindgut until E9.5 when they actively migrate through the body wall, dorsally, then laterally to colonize the genital ridges (14). At the ridges PGCs associate with the somatic support cells that will either direct their fate towards becoming the self- renewing stem population in the testis, or the non-renewing oocyte pool within the ovary.

Being motile, PGCs occupy many different niche environments prior to colonizing the gonads. For example, the posterior proximal epiblast (PPE) has been dubbed an embryonic germ cell niche (4). PGCs arise in this region in response to multiple BMP family members. First, BMP8b is required to control the anterior-posterior


patterning of the visceral endoderm thereby allowing for establishment of a permissive region within the PPE. Cells in the PPE then respond either directly (6) or indirectly

(140) to additional BMP family members that induce them to become PGCs.

Additional growth factors and adhesive interactions are required to control PGC migration, proliferation and survival as they migrate to colonize the genital ridges (170).

The growth factors KITL (171) and SDF1a (172) are expressed in the hindgut and later in the genital ridges. KITL (15, 86-88) and SDF1a (173) are critical for controlling PGC numbers and may direct PGC migration. However, precisely how the somatic cells along the migration route coordinate expression of these factors remains uncertain and in many cases, even the identity of the supporting cells themselves is unclear.

This study provides evidence that PGC development is coupled to differentiation of the mesonephros and that both systems rely on BMP signaling. In a previous study, an organ culture system was used to demonstrate that BMP signaling regulates expression of

Kitl within the nascent genital ridges (174). In this study, conditional gene targeting was used to confirm the role of BMP signaling during PGC migration in vivo. Additionally, we demonstrate that BMP signaling controls cell survival and possibly cell identity within the genital ridges. We propose that this represents an early and probably transient

PGC niche that is established in the genital ridges prior to differentiation of the sertoli or granulosa cell support lineages.


3.3 Materials and Methods

3.3.1 Mice and tamoxifen injections.

The Case Western Institutional Animal Care and Use Committee approved all experiments involving mice. Bmpr1a-fx mice provided by Y. Mishina were crossed with

Oct4
PE:GFP mice (153) and Rosa:LacZ mice (175), and all three loci were bred to homozygosity. Mice carrying the CAGGCre-ERTM (176) or Pax2-Cre (177) transgene were crossed to mice carrying the Bmpr1a null allele Bmpr-s (137). Male Bmpr1a- fx/Bmpr1a-fx Oct4
PE:GFP/Oct4
PE:GFP Rosa:LacZ/Rosa:LacZ mice were crossed with CAGGCre-ERTM/+ Bmpr-s/+ females resulting in four different genotypes (Figure

3-1H). Noon on the day that a vaginal plug was detected was considered E0.5. E8.5 pregnant females were weighed and treated with tamoxifen via oral gavage (9 mg TM in corn oil per 40 g body weight) (176). Embryos were then harvested 24, 36, or 48 hours later.

3.3.2 PGC counts and immunohistochemistry.

For PGC counts, embryos were harvested at E10.5 and transverse sections between the fore- and hind-limb buds were cut by hand. Optical sections of tissue slices were collected at 5 mm intervals using a Leica TCS AOBS filter-free Confocal Laser

Scanning microscope. Images were processed using Volocity software and the numbers of PGCs were counted. For each slice, the number of PGCs per 100 mm was calculated

(n = number of PGCs per slice / slice thickness * 100 mm). Counts were normalized to the average wild type value for each litter in order to minimize the effect of variations in


age between litters. E9.5 PGC counts were obtained by analyzing images from time-

lapse experiments.

Immunostaining of dissected tissue slices was performed using pSMAD1/5/8

antibodies (Cell Signaling Technology) or cleaved PARP antibodies (Cell Signaling

Technology). For pSMAD1/5/8 staining, embryos were fixed in 8% PFA/PBS overnight

at 4°C. Transverse sections were then cut by hand and slices were dehydrated through a

methanol series. Slices were then rehydrated through the MeOH series and further

permeabilized by treatment in PBS/0.1%TX-100 (PBTX) for at least 2 hours. Slices

were then blocked in 2% goat serum and 2% IgG-free BSA in PBTX (goat block)

overnight at 4°C. Then, slices were incubated with rabbit-anti-pSMAD1/5/8 diluted 100 fold in goat block, overnight at 4°C. The slices were rinsed 5 times, one hour each, in

PBTX at room temperature. Slices were then exposed to goat-anti-rabbit IgG Cy5

secondary antibodies (Jackson Immunoresearch Laboratories, Inc.) diluted 100 fold in

goat block overnight at 4°C. The tissue was then rinsed 5 times, one hour each, in PBTX.

Slices were stored in 1:4 PBS:Glycerol before collecting confocal images. Volocity

software was used to process the images and to measure pixel intensity. Cleaved PARP

antibody staining followed the same protocol as above except that slices were fixed in 4%

PFA/PBS and were not taken into MeOH. The rabbit anti-cPARP antibody was diluted

to 1:250 in goat block, and the secondary goat-anti-rabbit IgG cy5 was diluted to 1:200.

Frozen sections were used for
-gal antibody staining. Briefly whole embryos were fixed

in 4% PFA/PBS overnight at 4°C, rinsed in PBS and PBTX, and then rinsed three times

for 30 minutes each in O.C.T. compound (Sakura Finetek). Embryos in O.C.T. were then

transferred to plastic blocks and frozen. Sections were cut to a thickness of 14 μm and


stored at -20°C. Slides were rinsed twice in 1X PBS and once in PBT (0.25% Triton X-

100, 1X PBS) for 5 minutes each. The tissue was then blocked with 10% donkey serum

in PBT (donkey block) for 1 hour at RT then incubated overnight in 1:250
-gal antibody

(Biogenesis, Inc.) in donkey block at 4°C. Slides were rinsed 5 times for 10 minutes each

in PBT and then exposed to 1:100 donkey-anti-goat IgG (H+L) Alexa Flour 555

(Molecular Probes) in 1% donkey serum in PBT for 1 hour at RT. Slides were rinsed 5

times for 10 minutes each in PBT and mounted in Vectashield with DAPI (Vector

Laboratories, Inc.) overnight at 4°C. Staining was imaged on a Leica DM6000B microscope and images were processed using Volocity software. Whole embryos were stained with X-gal as described (178).

pSMAD1/5/8 staining was measured using Volocity. pSMAD1/5/8 Pixel intensity throughout all z-planes of the image was divided by the volume of the measured image yielding the intensity per volume. Measurements were normalized to the average pixel intensity per volume of the wild type samples.

Cleaved PARP foci were counted in Volocity by selecting the mesonephric mesenchyme, coelomic epithelium, and boundary of the hindgut using the region of interest (ROI) tool. Positive foci were identified by detecting the top 30 – 100% intensity on the Cy5 channel. Foci were restricted to the ROI using the “Clip to the ROI” command. Objects greater than 1000 μm3 were separated (reflecting the fact that clusters bigger than 1000 um3 usually represented multiple touching foci instead of a single foci).

Objects less than 100 μm3 were excluded. Objects matching these parameters were

counted by the software, the volume of the ROI was recorded, and counts were

normalized to the average volume for each region.


3.3.3 Gene expression analysis.

Changes in gene expression were monitored by quantitative real time PCR (q-RT-

PCR). Briefly, E10.5 embryos were harvested and either tissue slices or genital ridges were dissected as described (174). Total RNA was isolated by homogenizing tissue in

300 μl of TriZOL (Invitrogen) in the presence of 5 μg linear polyacrylamide (Sigma-

Aldrich Co.). Isolated RNA was treated with RQ1 DNase (Promega). cDNA was synthesized using the SuperScript III First Strand cDNA Synthesis Kit with oligo dT priming (Invitrogen). A minimum of 15 ng of template RNA was used per 10 ml cDNA reaction. An RT- negative control was run for each experiment to monitor the effectiveness of the DNase treatment by having one sample run without the SuperScript

III reverse transcriptase. Following the RT reaction, samples were diluted 10 fold with nuclease-free water. qPCR was run using the Brilliant II SYBR Green Master Mix

(Promega) and primers for Bmpr1a, Kitl, Pax2, Id1, Sdf1a, Scarb1 and the loading controls bAct and Tbf (Supplemental Table 3-1). Raw data was normalized according to the average of bAct and Tbf for each sample. For each litter, values were then expressed as a percent relative to the average wild type, allowing for comparisons among litters.

3.3.4 Time-lapse.

Tissue dissection and filming was conducted as previously described with some modifications (18). Briefly E8.5 pregnant females were exposed to TM as described above and E9.5 embryos were harvested. Transverse sections were then cultured as described and a single optical section was captured every 7 minutes for approximately


11.8 hours (100 frames). Cell traces were generated using the automated tracking feature

of Volocity software. First, noise was removed from the GFP channel. Cells were

identified based on intensity (top 40% of the GFP channel). Holes were filled in the

identified objects. Touching cells were separated based on a size threshold of 100 μm2.

Objects less than 40 μm2 and larger than 300 μm2 were excluded. Tracking was then performed using the “Shortest Path” algorithm with a maximum distance between nodes of 10 μm. Targeting and average velocity measurements were calculated from 20 traces with the highest trace times, excluding any traces with times less than one hour. PGC targeting was measured by connecting a PGC’s start point and end point with a straight line then extrapolating that through the tissue in the direction the PGC was moving. If the line intersected a genital ridge, it was recorded as targeting (Figure 3-S3).

3.3.5 Statistics.

Gene expression values, pSMAD1/5/8 levels, cleaved PARP staining, PGC counts and tracking data were compared between the four different genotypes using one-way

ANOVA. Following analysis of variance, between group differences were determined using the Fisher’s least significant difference post test. Calculations were performed in

Excel.


3.4 Results

3.4.1 BMP signaling components are expressed in the genital ridges during PGC

migration.

BMP signaling components are present and active in the genital ridges as PGCs

migrate to and colonize this region. E9.5 and E10.5 embryos were harvested and

transverse sections between the fore- and hind-limb buds were isolated (Figure 3-1A). In

situ hybridization data show that at least two BMP ligands are expressed in the genital

ridges at E9.5 and E10.5. Bmp4 expression was detected in the mesonephric

mesenchyme at E9.5 with lower expression in the midline dorsal to the hindgut (Figure 3-

1B). Bmp4 expression was elevated in the posterior end of the embryo encompassing the

tail bud and the remnant of membrane surrounding the umbilical vein. By E10.5 Bmp4

expression remained high in the mesonephros (data not shown) with additional regions

corresponding to known domains of Bmp4 expression including the dorsal margin of the

eye (179), the outflow tract of the heart (180), the limb buds (181) and the

dermomyotome of the developing somites (43). At E12.5 Bmp4 continued to be

expressed in the gonads as well as the developing kidneys (data not shown). Bmp7 was also expressed in the developing urogenital system. At E9.5 Bmp7 was observed in the mesonephric duct of the genital ridges (Figure 3-1C). By E10.5 expression had expanded to include the somites, the limb buds (182), and the myocardium of the heart (151) (data not shown). In the E12.5 urogenital system, Bmp7 expression was found in the kidneys

(183) and in the sex cords of the developing testis (184) (data not shown). Bmp2



Figure 3-1: Tools for targeting BMP-signaling during PGC migration. (A) Transverse tissue sections containing PGCs were isolated as shown. This cartoon illustrates the morphology of a typical slice taken at E9.5. Slices may be larger or smaller based on their position along the AP axis. (B, C) In situ hybridization of E9.5 tissue with probes for Bmp 4 (B) and Bmp7 (C). Bmp4 is expressed in mesenchymal cells along the PGC migration route. In the genital ridge, Bmp7 expression is limited to the mesonephric duct (D, E). Immunostaining for phosphorylated SMAD1/5/8 (pSMAD1/5/8) reveals BMP-responsive cells at (D) E9.5 and (E) E10.5. (F) Cre-LoxP strategy used to target Bmpr1a. (G, G’) Excision efficiency was tested using the Rosa:LacZ reporter. CreER/+ mice were crossed to mice homozygous for the Rosa:LacZ reporter. Tamoxifen was administered at E8.5 and recombination assayed by immunostaining at E10.5. bGal expression (red) is seen in CreER/+, Rosa:LacZ/+ embryos (G) but not in Rosa:LacZ/+ sibling controls (G’). DAPI is shown as blue. (H) Breeding strategy to generate a tamoxifen inducible loss of Bmpr1a. Four genotypes of offspring are possible. In KO and C-Het embryos, tamoxifen will induce


recombination of the Bmpr1a-fx allele. Bmpr-s is a null allele of Bmpr1a. FL = fore-limb bud, NT = neural tube, MM = mesonephric mesenchyme, MD = mesonephric duct, CE = coelomic epithelium, A = aorta, G = gut, PGC = primordial germ cells. Arrows indicate the genital ridges. ______

expression was not detected at E9.5 or E10.5 in the genital ridges but signals were seen in

known domains of Bmp2 expression including the outflow tract of the heart (185) and the

ventral margin of the limb bud (186) (data not shown). At E12.5 Bmp2 expression was

observed in the developing kidneys (183) and weakly in the gonads (data not shown).

To determine which cell populations respond to endogenous BMPs, we performed

immunostaining with an antibody that recognizes phosphorylated SMAD1/5/8. Staining

of E9.5 transverse sections showed activated SMADs at low levels throughout most of

the section (Figure 3-1D) (174). Staining was most intense in the mesonephric

mesenchyme and coelomic epithelium of the genital ridges. Bright staining was also

observed in the dorsal neural tube and in the dermomyotome of the somites. At E10.5,

pSMAD1/5/8 staining was no longer elevated in the coelomic epithelium but remained

strong in the mesonephric mesenchyme and in the dorsal neural tube, with low levels

throughout the slice (174) (Figure 3-1E).

3.4.2 Conditional loss of Bmpr1a leads to decreased BMP signaling.

To target BMP signaling in vivo we used a ubiquitously expressed CreER transgene, CAGGCre-ERTM (176), to knock out a conditional allele of Bmpr1a (187)

(Figure 3-1F). Bmpr1a encodes for one of three type one BMP receptors in the mouse

and is the preferred receptor for BMP4 (129), the ligand that appeared to be the most

abundantly expressed in and along the PGC migration route (Figure 3-1B). First, we

tested excision efficiency of the CAGGCre-ERTM line, hereafter referred to as “CreER.”


CreER/+ mice were crossed to a Rosa:LacZ reporter line. E8.5 pregnant females were injected with tamoxifen and embryos were harvested at E9.5 and either stained for b-gal activity as whole mounts (data not shown) or cryosectioned to examine b-gal expression via immunostaining. Immunostaining with an antibody directed against b-gal showed moderate recombination throughout the entire section (Figure 3-1G). A breeding schema was established to allow for inducible recombination of one conditional allele over a null allele, Bmpr-s (137), to increase efficiency of knockdown (Figure 3-1H). Quantitative real-time PCR was then used to monitor the efficiency of Bmpr1a-fx recombination in tissue slices dissected from between the fore- and hind-limb buds (Figure 3-2). Primers were designed to detect only full length Bmpr1a cDNA (Figure 3-1F, arrows). In the absence of TM, wild type (WT) samples expressed Bmpr1a at an average normalized

level of 95.5 + 4.5%, conditional heterozygotes (C-Het) at 76.8 + 6.1%, Bmpr-s heterozygotes (B-Het) at 50.0 + 6.2%, and conditional knock outs (KO) at 49.5% (Figure

3-2A). Administration of TM to E8.5 pregnant females resulted in near complete loss of

Bmpr1a expression in KO embryos by E10.5 (Figure 3-2B). Average normalized WT

Bmpr1a expression was found to be 104.9 + 8.9%, compared to C-Het expression at 65.8

+ 5.9%, B-Het at 57.2 + 7.6%, and KO at 5.2 + 3.0%. This experiment was repeated while varying the length of time embryos were exposed to TM in utero. Recombination was detected as early as 6 hours after TM exposure (Figure 3-S1). Within 12 hours average expression in KOs was 18% of WT littermates.

To confirm that a reduction in Bmpr1a expression leads to decreased BMP signaling in vivo, TM exposed embryos were fixed and stained with an antibody directed



Figure 3-2: Tamoxifen induced Cre recombination of Bmpr1a-fx leads to decreased expression of Bmpr1a and decreased BMP signaling. q-RT-PCR was used to monitor the level of Bmpr1a mRNA in E10.5 tissue slices taken between the fore- and hind-limb buds. E10.5 KO embryos that had not been exposed to tamoxifen had Bmpr1a mRNA levels similar to their B-Het siblings (A). KO embryos treated with TM at E8.5 and harvested at E10.5 had reduced Bmpr1a levels (B). Immunostaining for pSMAD1/5/8 was used to monitor BMP signaling in WT (C, D) and KO (C’, D’) TM-treated embryos at E9.5 (C, C’) and E10.5 (D, D’). Staining intensity (relative pixel intensity per mm3) was quantified at E9.5 (E) and E10.5 (F) using Volocity software. Expression and pSMAD staining data were compared between the different genotypes using one-way ANOVA followed by the Fisher’s least significant difference post test. Data sets that were deemed equivalent are marked with the same letter (a-c). Arrows mark the genital ridges. TM = tamoxifen, A = aorta, NT = neural tube, -1° = no primary antibody, N = number of embryos. Error bars are standard error of the mean (s.e.m.).


against phosphorylated SMAD1/5/8 (Figure 3-2C-F). Embryos exposed to TM at E8.5 showed no change in pSMAD1/5/8 levels by E9.5 (Figure 3-2C, C’, and E). However, by E10.5, KO embryos had greatly reduced pSMAD1/5/8 staining (Figures 3-2D, D’, and

F). Remaining E10.5 pSMAD1/5/8 activity could be attributed to incomplete turnover of endogenous BMPR1a or the activity of other type I BMP receptors.

The Cre-ER transgene is effective at reducing Bmpr1a mRNA levels and BMP signaling in the entire trunk of the E10.5 embryo. However, to more precisely target

BMP signaling within the genital ridges, we attempted tissue specific targeting utilizing the Pax2-Cre line developed by Ohyama and Groves (177). Pax2 is a transcription factor required for kidney development. It is expressed in the pronephros at E9.0 and in the mesonephric duct and adjacent mesonephric mesenchyme at E10.5 (154). We demonstrate that Pax2-Cre can induce recombination of a LacZ reporter within the mesonephric duct and anterior mesonephric mesenchyme by E9.5 (Figure 3-S2A).

Additionally, Bmpr1a mRNA levels are reduced in genital ridges isolated from Pax2-

Cre/+ Bmpr1a-fx/Bmpr1a-s E10.5 animals (Figure 3-S2B). However, pSMAD1/5/8 levels remained high in genital ridges from these animals (Figure 3-S2C) and there was no significant effect on PGC numbers (Figure 3-S2D). This indicates that this driver line may not recombine early enough or efficiently enough to reduce BMPR1a protein levels and impact signaling during PGC migration. Therefore, we discontinued the use of this line and focused our analysis on the CreER crosses.


3.4.3 Decreased Bmpr1a expression correlates with changes in BMP target gene

expression.

We have previously shown that in culture BMP signaling controls the expression

of target genes within the genital ridges as PGCs migrate to colonize these structures

(174). To test if BMP signals control genital ridge gene expression in vivo, we used real

time PCR to compare the expression of ridge marker genes in induced knock out animals

and their littermates. In the chick, Pax2 expression within the lateral plate is controlled

by BMP signaling (71). The growth factor Kitl is a BMP responsive gene (81, 174)

required for PGC survival and potentially motility. Kitl is expressed along the migratory

route of PGCs and its expression increases within the E10.5 genital ridges as PGCs

colonize these structures (92, 174). To detect tissue specific changes in target gene

expression, transverse tissue pieces were dissected to separate genital ridge tissue from

neural tube (NT) and gut tissue as previously described (174). Bmpr1a expression in KO

tissue was reduced to 23.8 + 3.8% in the genital ridges (Figure 3-3A). Expression of the known BMP target gene Id1 (157) was reduced to 55.0 + 6.1% in KO genital ridges relative to WT littermates (Figure 3-3B). Genital ridge expression of Kitl in KO embryos was reduced to 35.3 + 4.9% that of the wild type (Figure 3-3C). Pax2 expression in KO

embryos was reduced to 46.3 + 7.0% in the genital ridges (Figure 3-3D). Expression of

Sdf1a, a putative PGC chemoattractant expressed in the genital ridges (18), was reduced in C-Het and KO genital ridges relative to WT (Figure 3-3E). While some mesenchymal

(Pax2) and epithelial (Sdf1a and Kitl) marker genes were down regulated in response to

decreased BMP signaling, the expression of Scarb1, an epithelial marker in the ridge, was


slightly up regulated in KO genital ridges relative to wild type and heterozygous

littermates (Figure 3-3F).

3.4.4 Reduced BMP signaling leads to decreased targeting of PGCs to the genital ridges.

Decreased BMP signaling in KO embryos led to reduced expression of Kitl and

Sdf1a (Figure 3-3), two factors that have been shown to be involved in PGC migration at this stage. This suggested that there might be migration defects in the Bmpr1a conditional KO embryos. To monitor PGC migration we used confocal time-lapse microscopy. At E9.5, WT PGCs are located near the midline then migrate laterally to colonize the genital ridges over the next 24 hours (Movie 3-1, 11 hours shown).

However, in KO samples, fewer PGCs migrate towards the genital ridges (Movie 3-2).

To quantify the migration defects in the KO embryos we used Volocity software to track

PGCs (Figure 3-4). PGC tracks were then scored as either targeting the nearest genital ridge or an ectopic region (Figure 3-S3). While only 60% of PGCs in WT embryos were observed targeting the nearest genital ridge, approximately 43% of PGCs in KO embryos were observed targeting the nearest genital ridge (Figure 3-4A-C). There was no significant difference in the average length of time that the software was able to track a

PGC in WT, heterozygous, and KO PGCs (Figure 3-S4). Similar trace times suggest that the observed differences in PGC targeting are not artifacts of the variation in the ability of the program to follow PGCs for their entire migration. In addition to decreased PGC targeting, KO PGCs had a slightly lower average velocity than WT PGCs (Figure 3-4D).



Figure 3-3: Expression of BMP target genes are reduced in the genital ridges of Bmpr1a conditional knock out embryos. TM was administered at E8.5 and embryos were harvested at E10.5. Tissue slices were dissected and the genital ridges were further dissected away from the neural tube and gut. q-RT-PCR was used to measure the expression of Bmpr1a (A), known and putative BMP target genes Id1 (B), Kitl


(C), Pax2 (D), and Sdf1a (E), and the epithelial marker gene Scarb1 (F). Results were significant by ANOVA followed by Fisher’s least significance testing as indicated (lower case letters), F<0.05. WT = wild type, C-Het = conditional heterozygote, B-Het = Bmpr-s heterozygote, KO = knock out, N = number of embryos. Error bars are s.e.m. ______

3.4.5 Conditional Bmpr1a knock out embryos have fewer migratory PGCs.

We have previously shown that migratory PGCs are sensitive to BMP levels in

organ culture. In this study we monitored the effect of decreased BMP signaling on

migratory PGCs in vivo (Figure 3-5). Embryos were exposed to TM at E8.5 and

harvested at E9.5 or E10.5. There was no significant difference in the number of PGCs

in E9.5 KO embryos relative to WT and heterozygous littermates (Figure 3-S5). This is

not surprising considering that 24 hours was insufficient time to allow for a reduction in

pSMAD1/5/8 levels (Figure 3-2E). However, by E10.5, KO embryos exhibited a

significant reduction in PGCs when compared to their littermates (Figure 3-5).

Additionally the few remaining PGCs in KO embryos were often located outside of the

genital ridges in the gut or midline (Figure 3-5D). Volocity software was used to count

the number of PGCs per 100 mm of tissue. To account for stage variability between

litters, PGC counts were normalized to the average number of PGCs per wild type

embryo for each litter. E10.5 KO embryos were found to have significantly fewer PGCs

on average than their WT littermates (38.1 + 9.2% compared to 100 + 4.6%) (Figure 3-

5E). There was no significant difference between the WT embryos and either the C-Hets or B-Hets. However, C-Het embryos had significantly more PGCs than B-Het embryos,

117.6 + 9.9% compared to 92.0 + 7.1 % respectively.



Figure 3-4: Conditional loss of Bmpr1a leads to reduced PGC targeting of genital ridges. Tamoxifen was administered at E8.5 and embryos harvested at E9.5. Transverse sections were then cultured on the stage of a confocal microscope and time-lapse images collected every 7 minutes for 100 frames. Volocity software was used to track the trajectories of the PGCs in wild type (A, black lines, Movie 3-1) and knock out tissue (B, black lines, Movie 3-2). PGCs were then scored as either targeting the genital ridges (A, B, arrowheads) or migrating away from the nearest ridge, revealing decreased targeting in KO tissue significant by ANOVA followed by Fisher’s least significant difference test, F = 0.01 (C). Additionally there was a slight but not significant decrease in PGC velocity in KO tissue relative to the wild type (D). PGCs express the GFP marker and arrows mark genital ridges. NT = neural tube, N = number of embryos. Error bars are s.e.m.



Figure 3-5: Reduced BMP signaling leads to decreased numbers of PGCs in vivo. Representative transverse sections of E10.5 wild type (A), C-Het (B), B-Het (C), and KO (D) embryos show the abundance and location of PGCs (GFP positive). Counts of PGCs in E10.5 WT, C-Het, B-Het, and KO embryos showed that there are significantly fewer PGCs in KO embryos relative to their wild type and heterozygous littermates (E). Results are significant by ANOVA followed by Fisher’s least significant differences test, F = 0.000004. Arrowheads mark genital ridges. A = aorta, NT = neural tube, N = number of embryos. Error bars are s.e.m.


3.4.6 Knock out embryos have increased genital ridge somatic cell death.

The reduction in PGCs in KO embryos could be due to apoptosis, decreased

proliferation, or loss of expression of the Oct4
PE:GFP marker should PGCs loose pluripotency. To determine if the loss of PGCs in KO embryos was due to apoptosis we fixed and stained transverse sections with a cleaved PARP specific antibody. Following

TM exposure at E8.5, there was an increase in the number of apoptotic foci in the somatic tissue of the KO embryos relative to their WT littermates (Figure 3-6). However, few

PGCs were observed undergoing apoptosis. In E9.5 KO embryos there was scattered cell death throughout the slice with an increased concentration of apoptotic cells in the mesonephric mesenchyme of the genital ridges (Figure 3-6A, A’, B, B’ and C). Death in the mesonephric mesenchyme was more dramatic at anterior positions (compare Figure

3-6B and B’). On average, the mesonephric mesenchyme of WT slices had 2.0 + 0.61

apoptotic foci per 600,000 μm3, compared to 6.6 + 0.88 apoptotic foci per 600,000 μm3 in the KO mesonephric mesenchyme (Figure 3-6C). Increased cell death in the mesonephric mesencyhme was also observed in E10.0 genital ridges (Figures 3-6D, E, F,

G and H). Wild type E10.0 slices had moderate levels of cell death throughout the slice

(Figure 3-6D). However, significantly greater cell death was observed in the mesonephric mesenchyme of KO slices than of WT slices, 3.1 + 1.2 compared to 0.72 +

0.41 apoptotic foci per 900,000 μm3 respectively (Figure 3-6H). In two KO embryos out of three litters, massive cell death was observed in the genital ridges in such density that



Figure 3-6: Conditional loss of Bmpr1a leads to increased somatic cell apoptosis in the mesonephric mesenchyme. Immunostaining for cleaved (c) PARP (red) in an anterior slice taken from a WT E9.5 embryo (A) and a more posterior slice taken from the same embryo (A’). cPARP staining in an anterior slice taken from a KO E9.5 embryo (B) and a more posterior slice taken from the same embryo (B’). The mutant has more cell death within the anterior mesonephric mesenchyme than the wildtype. More posterior positions are less effected (B’). Examples of the regions selected for counting cPARP foci are circled in (B). For each slice, Volocity software was used to count the number of apoptotic foci in the mesonephric mesenchyme (mm) (C), coelomic epithelium (ce), and gut (g) (data not shown). Examples of cPARP staining at E10.0 in WT (D), B-Het (E), and KO (F, G) tissue slices. KO embryos have either a modest increase in cell death (F) or in rare cases a dramatic increase in cell death (G) in the mesonephric mesenchyme. cPARP foci were counted as described above and the amount of cell death in the mesonephric mesenchyme is displayed (H). E10.0 samples with extreme cell death (G) were not used for quantitation because it was not possible to count individual foci accurately. There were significantly more apoptotic foci in the KO mesonephric mesenchyme than in the WT at both E9.5 (C) and E10.0 (H). Results are significant by ANOVA followed by Fisher’s least significant difference test, F = 0.000004 (C) and F =


0.039 (H). PGCs express the GFP marker. Scale bars: 100 μm (A – B’), and 160 μm (D-G). Error bars are s.e.m. ______

the number of apoptotic foci could not be counted (Figure 3-6G) and therefore were

omitted from these comparisons (Figure 3-6H). E10.0 B-Het embryos were found to

have slightly increased numbers of apoptotic foci in the mesonephric mesenchyme at 1.7

+ 0.19 per 900,000 μm3, though these counts were not significantly different than the WT

(Figure 3-6E, H). In addition to the mesonephric mesenchyme, apoptotic foci were also counted in the coelomic epithelium and the gut; however, there were no significant differences in these data sets (data not shown).

3.5 Discussion

In this study, the role of BMP signaling in PGC migration was tested by using

conditional targeting to reduce BMPR1a levels. Decreased BMP signaling, as monitored

by pSMAD1/5/8 levels, resulted in fewer PGCs at E10.5 and reduced targeting of

migratory PGCs. This phenotype coincided with an increase in somatic cell death within

the mesonephric mesenchyme of the conditional knock out genital ridges. Consistent

with the high degree of apoptosis in the mesenchyme, the transcription factor Pax2 was

observed to be reduced in the KO genital ridges. Additionally expression of genes

considered to be markers of the coelomic epithelial (a.k.a. pre-gonadal) compartment

were also affected. Expression of the PGC survival factor Kitl was reduced and

expression of the PGC attractant Sdf1a was also reduced. Expression of the epithelial

marker Scarb1 was increased. Our data demonstrates that BMP signaling is required for


survival of the mesenchymal component of the genital ridge and that it may regulate cell

fate decisions within the coelomic epithelial compartment of the ridge. In this manner

BMP signals establish a favorable environment for arriving PGCs.

BMPs are known to be required for PGC development at various stages.

Specification of PGCs in early post-implantation mouse embryos requires

extraembryonic ectoderm derived BMP4 (5) and BMP8b (2) and endoderm derived

BMP2 (3). These various BMP family members may act directly or indirectly to specify

PGCs. In support of an indirect mechanism, de Sousa Lopes et al. (140) have shown that

the PGC deficiency in Bmp4-/- embryos can be rescued by constitutive activation of the type I BMP receptor Activin A Receptor, type 1 (ACVR1) in the visceral endoderm, but not the epiblast where PGC precursors reside. Conversely, there is evidence to suggest that BMPs directly regulate PGC specification. Ohinata et al. (6) showed that extraembryonic ectoderm derived BMP4 can drive expression of Blimp1 in epiblast cells, signaling directly through PGC expressed BMPR1a, as well as SMAD1 and SMAD5.

The effect of BMP4 is dose dependent such that those cells that are closest to the source, and are exposed to the highest concentration, will express Blimp1 and ultimately become

PGCs. BMP4 has also been found to be necessary to differentiate PGCs from embryonic

stem (ES) cells in culture (188), which may indicate a direct requirement of BMP4 in

PGC specification. However, the authors noted that BMP4 can promote the formation of

extraembryonic mesoderm and that PGCs may be developing within a microenvironment

in an indirect way. Our data indicates that BMP signaling also acts within the nascent

genital ridges to establish a supportive microenvironment for arriving PGCs. Loss of

BMP signaling results in death of mesenchymal cells and reduces expression of the


epithelial markers Kitl and Sdf1a. We cannot currently rule out a direct interaction between ridge expressed BMPs and arriving PGCs. Migratory PGCs express multiple

BMP receptors and SMADs and are capable of responding to high levels of exogenously added BMPs (174). However, we were unable to detect pSMAD within PGCs under normal conditions using either immunostaining or a BRE:lacZ reporter mouse (189) (data not shown). Either migratory PGCs use non-SMAD pathways to respond to BMPs or the requirement for BMP signaling during PGC migration is entirely indirect.

Our data indicates that BMP signaling is necessary for the survival of the mesonephric mesenchyme and for PGC migration. This is the first study to couple early kidney development with PGC migration in mammals. There is evidence in zebrafish that pronephric cells of the lateral plate mesoderm are an intermediate target of migrating

PGCs and disruption of this target tissue ultimately leads to decreased PGC accumulation within the genital ridges (16). However, multiple previous studies in mice have suggested that PGC migration is independent of both kidney development and specification of the somatic support cells of the gonads. Mice homozygous for a null mutation in Wilms tumor 1 (Wt1) exhibit increased apoptosis in the E11.0 metanephric blastema and ultimately a complete degeneration of the kidney precursor population by

E12.0 (190). Regardless, PGCs were observed in the E12.0 genital ridge. Additionally, mice mutant for Steroidogenic factor 1 (Sf1) (191) and the LIM homeobox gene Lhx9

(192) fail to form gonads, with defects beginning by E12.0. PGCs were observed in the genital ridges at E10.5 in Sf1 null embryos and at E12.0 in Lhx9 null embryos. Together these data would imply that neither the developing kidney nor the gonad is necessary for proper PGC targeting and colonization of the genital ridges. However, these mutations


resulted in phenotypes later than those observed in conditional targeting of Bmpr1a. In our conditional knock out embryos, we observed increased cell death at E9.5 within the mesonephric mesenchyme. This precedes the mutant phenotypes of the Wt1, Sf1, and

Lhx9 mutants, and results in a rapid loss of mesenchymal cells, as evidenced by increased apoptotic foci at E10.0. This increase in cell death correlates with decreased PGC survival and targeting suggesting that the requirement for mesonephric cells occurs at least from E9.5 to E10.5.

Taken together our results indicate that BMP signaling within the mouse genital ridges during PGC migration acts to establish a permissive environment for arriving

PGCs. This niche is comprised of several cell populations. The Pax2 expressing cells of the pronephric mesenchyme both produce and respond to BMPs, including BMP4, which acts to promote their survival. Conversely, the cells of the coelomic epithelium do not rely on BMPs for survival but require it to maintain Kitl and Sdf1a expression and repress

Scarb1. Others have shown that during early kidney development, BMP signaling in the lateral plate mesoderm drives pronephric fate while restricting hematopoietic and vascular cell fates (69, 70, 72, 73). Our data suggest that BMP signaling may promote similar cell fate choices in the coelomic epithelium, the compartment that will later give rise to the somatic components of the gonads.

In summary, BMP signaling is necessary for PGC migration, cell survival and gene expression within the genital ridges. Future experiments will attempt to dissect the role of BMP signaling in distinct cell populations within the ridge. Ultimately this analysis will provide insight into how the germ cell niche is established. It should also


reveal how cell fate decisions are coordinated to give rise to distinct kidney and gonadal cell populations.

3.6 Acknowledgements

The BMPR-fx and BMPR-s mouse lines were generously provided by Yuji

Mishina. The Pax2-Cre line was provided by Andy Groves. We thank Patti Conrad for microscopy assistance and acknowledge the use of the Leica widefield and confocal microscopes in the Genetics Department Imaging Facility at Case Western Reserve

University, funded by the National Center for Research Resources (NIH-NCRR) Shared

Instrumentation Grant program (1S10RR021228 and 1S10RR017980). Financial support for this work was provided by Case Western Reserve University and the NIH

(R01HD053900 awarded to KM).


3.7 Supplemental Material

Supplemental Figure 3-S1: A time course of Bmpr1a-fx recombination. Tamoxifen was administered to pregnant females at E8.5 (48hrs), E9.5 (24hrs), E10.0 (12hrs), or E10.25 (6hrs) and embryos were harvested at E10.5. q-RT-PCR was used to compare the expression level of full length Bmpr1a mRNA in wild type (WT), heterozygous (B-Het), and knock out embryos exposed to TM. Bmpr1a mRNA levels in KO embryos that were not exposed to TM (-TM) resemble expression levels in heterozygous animals (B- Het). N = number of embryos. Error bars are s.e.m.



Supplemental Figure 3-S2: BMP signaling was not reduced in the E10.5 genital ridge of Pax2-Cre/+ Bmpr1a-fx/Bmpr-s embryos. Pax2-Cre drives expression of Cre recombinase in the mesonephric mesenchyme and adjacent mesonephric duct by E9.5 as evidenced by recombination of the Rosa:LacZ reporter (A). By E10.5, tissue specific knock out genital ridges have reduced expression of full length Bmpr1a mRNA relative to wild type controls (B). However, at E10.5 there is no difference in the GR staining intensity of anti-pSMAD1/5/8 as measured using Volocity software (C). Additionally there is not a significant difference in the number of PGCs in E10.5 tissue specific knock out tissue relative to wild type littermates (D). Arrowheads mark the genital ridges. NT = neural tube, A = aorta, WT = wild type (Bmpr1a-fx/+), BH = heterozygote (Bmpr1a-fx/Bmpr-s), TSH = tissue specific heterozygote (Bmpr1a-fx/+ Pax2-Cre/+), TSKO = tissue specific knock out (Bmpr1a-fx/Bmpr-s, Pax2-Cre/+), N = number of embryos. Error bars are s.e.m.



Supplemental Figure 3-S3: Examples of PGCs exhibiting targeting and non-targeting migration. Two PGCs (green) have been traced using Volocity. The highlighted PGC is migrating toward the genital ridge (GR) (A). The superimposed dashed arrow projects the direction of the PGCs path and intersects with the nearest genital ridge, implying that the PGC is targeting the genital ridge during this stage of migration. The highlighted PGC in (B) is migrating towards a position not within the genital ridge but dorsal to the mesonephric mesenchyme. The superimposed arrow does not intersect with the nearest genital ridge and therefore this PGC is not targeting the genital ridge. GR = genital ridge, NT = neural tube.



Supplemental Figure 3-S4: PGC trace times are similar in movies generated from wild type, heterozygous, and conditional KO embryos. Volocity is only capable of tracing a germ cell through a portion of each film. The program typically fails to follow cells that contact other cells during migration. The trace times of movies taken from each genotype were compared to insure that differences in trace times are not responsible for the differences observed in PGC tracking. Results were not significant by ANOVA followed by Fisher’s least significant difference test. N = number of embryos. Error bars are s.e.m.

Supplemental Figure 3-S5: Injection with TM at E8.5 does not affect PGC numbers at E9.5. To determine the number of PGCs in E9.5 slices, images from time-lapse experiments were analyzed using Volocity software. The T = 0 time point correlates with E9.5, and PGCs from each T = 0 optical section were counted for each genotype. PGC counts were compared by ANOVA followed by Fisher’s least significance testing. There was no significant difference among the four different genotypes. F = 0.39. N = number of embryos. Error bars are s.e.m.


Gene ID Forward Primer (5'-3') Reverse Primer (5'-3') bAct AGA GGG AAA TCG TGC GTG AC CAA TAG TGA TGA CCT GGC CGT Bmpr1a CAG CAG GAC CAG TCA TTC AA CTG GCT TCT TCT GGT CCA AG Id1 CAA CAG AGC CTC ACC CTC TC AGA AAT CCG AGA AGC ACG AA CTG CCC TTG TAA GAC TTG ACT Kitl CCA TGG CAT TGC CGG CTC TC G Pax2 AAA GTT CAG CAG CCT TTC CA CCA GGT AGA GTG GTG CTC GT TGA CAT CAG GGA CTC AGA GTA Scarb1 TTT GGA GTG GTA GTA AAA AGG GC G Sdf1a CTT CAT CCC CAT TCT CCT CA GAC TCT GCT CTG GTG GAA GG Tbp CTT CGT GCA AGA AAT GCT GA AGA ACT TAG CTG GGA AGC CC

Supplemental Table 3-1: RT-PCR primer sequences.

Movie 3-1: Time-lapse confocal microscopy of E9.5 wild type PGC migration. TM was administered to pregnant females at E8.5 and embryos harvested at E9.5. Transverse sections between the fore- and hind- limb buds were cultured overnight on the stage of a confocal microscope. Images were collected of each slice at 7-minute intervals for 100 frames and processed using Volocity software. Images were compiled into movies at 10 frames per second.

Movie 3-2: Time-lapse confocal microscopy of E9.5 Bmpr1a conditional knock out PGC migration. Images were collected at 7-minute intervals for 100 frames and processed using Volocity software. Images were compiled into movies at 10 frames per second. The gamma settings of the bright field channel were adjusted to reduce brightness in order to see the PGCs in the exported video. No gamma adjustments were made to the GFP channel.


Chapter 4: PGC Colonization of the Genital Ridge Involves

Membrane Bound KITL Mediated PGC/Soma Adhesion

Brian Dudley1, Nia Bhadra1, Kathleen Molyneaux1

1Department of Genetics Case Western Reserve University, 10900 Euclid Ave. Cleveland,

OH 44106

Note: I personally completed all experiments and analysis in this work except the PGC counts displayed in the ACK2 death curve figure (Figure 4-4).


4.1 Abstract

Primordial germ cells (PGCs) rely heavily on their somatic environment during their migration to and colonization of the genital ridges. Here we provide evidence of a mechanism of PGC migration involving membrane bound KIT Ligand, KITL(mb), mediated PGC/soma adhesion. As PGCs migrate from the hindgut to the genital ridges between E9.5 and E10.5 they depend on somatic cell expressed KITL for survival.

Inhibition of KITL activity by ACK2 led to a reduction in the number of E9.5 PGCs after

18 hours in culture. Addition of exogenous soluble (s) KITL led to a similar, though weaker, reduction in PGC numbers while increasing the accumulation of midline PGCs.

Time-lapse analysis revealed that KITL(s) treatment reduced PGC targeting of the E9.5 genital ridge while increasing PGC velocity. This decrease in targeting resulted in a reduced ability of PGCs to clear the midline, leading to midline accumulation. Treatment of E10.5 slices, when many PGCs have arrived in the genital ridges, led to an increase in the efflux of PGCs from the ridge. PGC/soma adhesion assays point to a role of

KITL(mb) in adhering PGC to their somatic environment. We propose that PGCs crawl along a trail of KITL(mb) towards the genital ridges. Once there, high levels of

KITL(mb) tether PGCs to the ridge.

4.2 Introduction

Mammalian reproductive development requires a complex sequence of events. In the mouse, the progenitors of the germ cells, the primordial germ cells (PGCs), are first


specified at approximately E6.25 – E6.5 on the proximal layer of the epiblast, near the base of the allantois (4). These cells then activate various PGC markers including Kit, encoding the receptor for KIT ligand (KITL), and become incorporated into the invaginating hindgut (13, 94). The PGCs then migrate actively and passively to fill out the extending hindgut until approximately E9.5 when they exit the hindgut, migrate dorsally through the gut mesentery, then laterally to colonize the genital ridges between

E10.5 and E11.5 (14).

Fertility hinges on the successful colonization of the genital ridges by the PGCs.

Factors that regulate PGC survival and migration during this process are crucial for reproductive development. We have previously shown that bone morphogenetic protein

(BMP) signaling within kidney precursors of the genital ridge promotes PGC survival and migration and positively regulates expression of the somatic cell expressed cytokine

Kitl (174, 193). KITL, and its receptor KIT, are known to absolutely crucial for PGC survival. Mutation or deletion of either Kit or Kitl leads to germ cell deficiencies in mice by E9.5, and near complete losses by E11.5 (86, 87). KITL is present as two isoforms in vivo resulting from two splice variants. A full length KL1 mRNA leads to production of a full length membrane bound KITL that is cleaved to produce a soluble form of KITL

(105). A truncated KL2 results from the in frame splicing out of exon 6, leading to a membrane bound KITL that lacks the major proteolytic cleavage site, resulting in a membrane bound isoform of KITL (105). These two isoforms show tissue specific expression patterns with KL2 being the predominant form in the mouse testis (105).

Previous experiments have shown that the membrane bound form of KITL, KITL(mb), is


capable of interacting with PGC expressed KIT to form an adhesive interaction between cultured somatic cells and PGCs (128).

During PGC migration, Kitl is expressed as a gradient along the migration route of PGCs, with the highest expression localized to the coelomic epithelium of the genital ridges (92). In a previous study, addition of exogenous soluble mouse KITL, or KITL(s), to cultured E9.5 transverse sections was found to counteract an observed decrease in midline Kitl expression, leading to an increase in midline PGCs (15). However, we had previously found that Kitl levels outside of the genital ridges remain constant from E9.5 to E10.5, indicating that other mechanisms may be involved in midline PGC death (174).

Here we show that treatment of E9.5 tissue with exogenous KITL(s) leads to a reduction in the total number of PGCs in organ culture. While an increase in midline

PGCs is observed as previously described by Runyan et al (15), we propose that this is a result of decreased PGC targeting of the genital ridges, leading to a reduction in the percent of PGCs clearing the midline. Additionally, exogenous KITL(s) treatment of

E10.5 tissue leads to an increase in PGC emigration from the genital ridges, in effect loosening the hold of PGCs within the target tissue. We propose that endogenous

KITL(mb) tethers PGCs to somatic cells along the PGC migration route and within the genital ridges. Exogenous KITL(s) competitively inhibits this interaction reducing PGC adhesion. This leads to increased PGC velocity, decreased targeting of the genital ridges, and ultimately decreased tethering to the somatic cells of the genital ridges. Preliminary cell culture adhesion assays support previous data by Pesce et al (128) that somatic cells are capable of forming this interaction, however, we were unable to repeat results showing that KITL(s) is capable of blocking this interaction in cell culture.


4.3 Materials and Methods

4.3.1 Organ culture.

The Case Western Institutional Animal Care and Use Committee approved all

animal protocols. Oct4
PE:GFP (153) male mice were crossed with CD1 females

(Charles River Laboratories). Noon on the day a vaginal plug was detected was

considered day E0.5. Embryos were harvested on E9.5 or E10.5 as indicated. Transverse

sections were cut by hand between the fore and hind limb buds and cultured as previously

described with some modifications (18, 174). Briefly, transverse sections were cultured

on millicell organ culture chambers in DMEM/F12 with 2% BSA in the presence of

mouse recombinant KITL(s) (R&D Systems), IgG2b or ACK-2 (eBiosciences). Optical

sections were taken of each slice at 5 μm intervals using a Leica TCS AOBS filter-free

Confocal Laser Scanning microscope at T=0 and T=18 hours. Counts are expressed as percent of PGCs present at T=18 relative to T=0.

4.3.2 Time lapse.

Time lapse imaging of KITL(s) and ACK2 treated E9.5 and E10.5 slices was performed as previously described (18, 193). Volocity was used to analyze time lapse data using previously described parameters with some modifications (193). Targeting and velocity data was collected for the 20 traces with the highest trace times excluding those traces with trace times less than 45 minutes. PGC targeting was measured as previously described (193). PGC survival curves were created by counting the number of

PGCs visible in each slice every ten frames (70 minutes). These counts were divided by


the total starting number of PGCs to generate a percent survival for each time point. In

E10.5 experiments, PGCs were scored as exiting the genital ridge if they migrated

outward from a position within the ridge.

4.3.3 Cell culture and adhesion assays.

Sl/Sl4, hSCF-220, and hSCF-248 cells were grown and maintained according to

the providers recommendations (ATCC). Briefly, all cells were grown in DMEM/High

Glucose supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 U/ml

streptomycin (Thermo Scientific).

To isolate PGCs for unsorted culture experiments, E11.5 genital ridges were

dissected from a full litter and treated with accutase (Sigma-Aldrich) for 15 minutes at

37°C. The tissue was triturated by repeated pipetting with a 200 μL pipette. The mixture

was filtered through a cell strainer to remove large clumps. Unsorted PGCs were seeded

onto SL/SL4, hSCF-220, and hSCF-248 cells and cultured in 6 well culture treated plates

with 3 ml of media.

PGCs were further sorted using the MACS kit (Miltenyi Biotec) according to the

manufacture’s protocol with some adjustments. Following accutase treatment, unsorted

PGCs were blocked in PBS with 2% BSA and 2% normal goat serum (Goat Block)

(Jackson Immunoresearch Laboratories) for 20 minutes. Cells were then incubated in

PBS with 1:100 MC-480/SSEA1 antibody (David Solter, Developmental Studies

Hybridoma Bank) for 15 minutes at 4°C. The cells were then washed two times in goat block. IgM microbeads were then added at 20 μl per 107 cells and the mixture was incubated with rocking at 4°C for a minimum of 45 minutes. The cells were then washed


two times and resuspended in PBS+2% IgG free BSA. The cells were then applied to three MACS Separation MS columns (Miltenyi Biotec) in succession. The final elution was collected and reconstituted in PBS with 2% BSA, and 2% goat serum. Purity and yield were calculated by counting total cells and total GFP+ cells with a hemocytometer.

Average purity following three passes was 56.6%. Sorted PGCs were then seeded onto cultures of hSCF-220 and hSCF-cells and cultured for 1 hour at 37°C. Sorted PGCs were cultured in 24 well culture treated plates with 622 μl of media and 1 μg/ml mouse

KITL(s) as indicated.

Following the initial 1 hour incubation, PGC images and counts were collected using a Leica DMI6000 inverted microscope. PGCs were counted at 10 different locations within the culture well for each condition/cell type. The cultures were then agitated by rotating the plate for 5 minutes on a bench-top nutator then rinsing the cells with one volume of PBS. Fresh medium was then added to the cultures and additional

PGC counts were collected. The percent of PGCs remaining after agitation and rinsing was calculated for each condition.

4.3.4 Statistics.

PGC counts, targeting, velocity, and trace time were compared using one-way

ANOVA followed by the Fisher’s least significant difference post-test. PGC adhesion data was compared using the Students T-Test. All calculations were performed in Excel.


4.4 Results

4.4.1 PGCs require KIT/KITL during migration to the genital ridges.

To determine if PGCs rely on KITL as they migrate to colonize the genital ridges

from E9.5 to E10.5, transverse slices between the fore and hind limb buds were treated

with the KIT inhibitor and blocking antibody, ACK2 (Figure 4-1) (194). PGCs in control

slices migrated normally, with PGCs arriving at the genital ridges and few PGCs

remaining in the midline (Figure 4-1A). During the 18-hour culture period, the number

of PGCs in control slices fell slightly to 86.9 + 6.0% (Figure 4-1C). ACK2 treatment resulted in a considerable decrease in the number of PGCs (Figure 4-1B and C).

Remaining PGCs in ACK2 treated slices were often observed remaining in the midline, with few PGCs colonizing the genital ridges (Figure 4-1B). ACK2 treatment at 9
g/ml or 18 μg/ml significantly reduced the number of PGCs to 16.2 + 4.9% and 28.5 + 8.3%

respectively (Figure 4-1C). Treatment of slices with a rat IgG2b control at 9
g/ml or 18

μg/ml yielded average PGC survivals of 85.4 + 15.1% and 98.7 + 16.7% respectively,

which were not significantly different than the untreated controls (Figure 4-1C).

4.4.2 Exogenous soluble KITL reduces PGC numbers in tissue culture.

In previous studies we observed an increase in Kitl expression coupled with an

increase in PGC numbers when slices were treated with exogenous BMP4 (174).

Conversely, reduction of BMP signaling in vivo led to decreased Kitl expression and

fewer PGCs by E10.5 (193). Together these data suggest that BMPs may act through



Figure 4-1: Inhibition of KIT by ACK2 reduces PGC numbers in cultured E9.5 slices. Confocal images of optical sections of slices were collected after 18 hours in culture under control conditions (A) and in the presence of 18 μg/ml ACK2 (B). PGCs were counted at T=0 and T=18. PGC survival is expressed as the average percent of PGCs remaining at T=18 in controls and following treatment with 9μg/ml and 18μg/ml

IgG2b and 9 μg/ml and 18 μg/ml ACK2 (C). Lower case letters indicate equivalent groups based on

ANOVA analysis, F < 9x10-8 (C). NT = neural tube, N = number of slices. Arrows mark the genital ridges. Error bars indicate standard error.


KITL to regulate PGC survival. To test the effect of artificially increasing KITL levels

during PGC migration, transverse tissue slices were treated with KITL(s) and PGCs were

counted (Figure 4-2). In untreated controls, PGCs have mostly cleared the midline and

begun to colonize the genital ridges after 18 hours in culture (Figure 4-2A). Addition of

soluble 40 ng/ml KITL(s) led to a decrease in PGC numbers coupled with increased PGC

accumulation within the midline (Figure 4-2B). This decrease was observed with varying

concentrations of exogenous KITL(s). In control slices, on average 99.4 + 10.1% of

PGCs remained after the 18-hour time course (Figure 4-2C). Treatment with 4, 40, and

400 ng/ml KITL(s) resulted in a significant decrease in PGC numbers to 67.8 + 9.3%,

63.4 + 8.6%, and 68.0 + 10.3% respectively (Figure 4-2C).

4.4.3 PGC migration is disrupted following treatment with soluble KITL.

Midline accumulation of PGCs in KITL(s) treated slices suggests that exogenous

KITL(s) might be having a negative effect on PGC migration. Time-lapse confocal

microscopy was used to observe PGC migration in KITL(s) treated and untreated

controls. In untreated slices, E9.5 PGCs begin in the gut and the dorsal gut mesentery

near the midline and migrate dorsally then laterally to form two clusters in the genital

ridges (Movie 4-1). PGCs remaining in the gut and midline eventually are lost through

apoptosis (15). In soluble KITL treated slices, PGCs do not form clusters within the

genital ridges as many PGCs fail to migrate laterally from the midline (Movie 4-2).

Volocity was used to trace migratory PGCs, and PGCs were scored as either targeting the

nearest genital ridge or an ectopic location. In control slices, 66.5 + 2.4% of the PGCs



Figure 4-2: Exogenous soluble KITL reduces the number of PGCs in organ culture. Confocal images of optical sections were collected of untreated (A) and 40 ng/ml KITL(s) treated (B) slices after 18 hours in culture. PGCs were counted at T=0 and T=18. PGC survival is expressed as the average percent of PGCs remaining at T=18 in controls and following treatment with 4 ng/ml, 40 ng/ml, and 400 ng/ml KITL(s) (C). Lower case letters indicate equivalent groups based on ANOVA analysis, F < 0.05 (C). NT = neural tube, N = number of slices. Arrows mark the genital ridges. Error bars indicate standard error.


traced targeted the nearest genital ridge (Figure 4-3). Treatment with 0.4 ng/ml and 40

ng/ml soluble KITL reduced PGC targeting to 51.1 + 3.3% and 42.2 + 2.9% respectively

(Figure 4-3A). KITL(s) treatment also had an effect on PGC velocity. Untreated PGCs

migrated at an average velocity of 20.3 + 0.2 μm/hr (Figure 4-3B). While treatment with

0.4ng/ml KITL(s) had no affect on PGC velocity, addition of 40 ng/ml KITL(s)

significantly increased PGC velocity to 21.5 + 0.3 μm/hr (Figure 4-3B). Inhibition of

KIT/KITL with ACK2 treatment reduced both PGC targeting and velocity to 27.1 + 3.0%

and 18.4 + 0.3 μm/hr respectively (Figure 4-3A-B). Time lapse analysis of ACK2 treated

slices revealed that PGCs halt migration prior to gradually disappearing from the slice

(Movie 4-3; Figure 4-4).

4.4.4 Soluble KITL treatment increases PGC emigration from E10.5 genital ridges.

At E10.5, PGCs are completing their lateral migration from the midline and

beginning to form tight clusters within the genital ridges (Movie 4-4). In untreated

controls, only 1.6 + 0.7% of PGCs were observed migrating away from the genital ridges

(Figure 4-5A). Treatment with 40 ng/ml KITL(s) led to a slight increase to 5.7 + 3.18% of total PGCs exiting the genital ridges while treatment with 0.4 ng/ml KITL(s) resulted in a slight reduction to 0.8 + 0.5% (Movie 4-5 and 4-6, Figure 4-5A). PGC emigration

was observed in 18% of untreated control slices and 17% of 0.4ng/ml KITL(s) treated

slices, whereas treatment with 40 ng/ml increased the occurrence to 42% (Figure 4-5B).

In addition to PGCs exiting the genital ridges, KITL(s) treatment also led to midline-

crossing PGCs. It is very rare for a PGC to commit to one genital ridge then cross the

midline and target the opposite genital ridge. No midline crossing PGCs were observed



Figure 4-3: Manipulation of KIT/KITL signaling with KITL(s) or ACK2 disrupts PGC migration. Confocal time-lapse microscopy was used to follow PGC migration in E9.5 cultured slices following treatment with 0.4 ng/ml and 40 ng/ml KITL(s) and 9 μg/ml ACK2. PGCs were scored as targeting the nearest genital ridge or an ectopic location (A). Volocity was used to measure average PGC velocity in slices cultured in KITL(s) or ACK2 (B). Lower case letters indicate significantly equivalent groups based on ANOVA analysis, F < 3x10-21 (A) and F < 5x10 (B). N = number of PGCs. Error bars indicate standard error.



Figure 4-4: PGC survival curves in control and ACK2 treated slices. Transverse slices were cultured in the presence or absence of 9 μg/ml ACK2 on the stage of a confocal microscope. A single optical section was captured every 7 minutes. PGCs were counted every 10 frames (70 minutes) and counts were expressed as the percent of PGCs remaining relative to T=0. Error bars indicate standard error. N = 4 slices. ______

in 31 control slices whereas 3 were observed in 56 KITL(s) treated slices. An example of

this was observed following treatment of an E10.5 slice with 0.4 ng/ml KITL(s) (Movie

4-6).

4.4.5 Membrane bound KITL adheres cultured PGCs more tightly than soluble KITL.

KITL(s) treatment reduces PGC targeting of the E9.5 genital ridge, increases PGC

velocity, leads to increased PGC emigration from the E10.5 ridge, and results in the

occurrence of midline-crossing PGCs. Taken together, these results suggest a possible

decrease in PGC adhesion due to the competitive interaction between exogenous KITL(s)

and endogenous PGC expressed KIT at the expense of endogenous KITL(mb).

KITL(mb) mediated PGC adhesion has been previously described in cell culture (128).



Figure 4-5: Exogenous KITL(s) leads to increased PGC emigration from the E10.5 genital ridge. Confocal time-lapse microscopy was used to monitor PGC behavior in E10.5 control, and 0.4 ng/ml and 40 ng/ml KITL(s) treated slices. Slices in which more than one PGC was observed migrating away from a cluster of genital ridge PGCs were counted and expressed as a present of the total number of slices (A). In each slice, PGCs migrating away from genital ridge PGC clusters were counted and expressed as a percent of total cells present (B). N = number of slices. Error bars indicate standard error.



Figure 4-6: Cultured PGCs adhere more tightly to cells expressing KITL(mb) than cells expressing KITL(s) or no KITL. E11.5 genital ridges were dissociated and seeded onto Sl/Sl4, human KITL(mb) expressing hSCF-220, or human KITL(s) expressing hSCF-248 cells and assayed for adhesion, expressed as the percent of PGCs remaining following agitation (A). E11.5 PGCs were purified using the MACS kit and seeded onto hSCF-220 and hSCF248 cells with or without 1 μg/ml mouse KITL(s) (B). N = number of independent experiments. Error bars indicate standard error. ______

These experiments were repeated using Sl/Sl4 cells expressing either a predominantly

membrane bound isoform of KITL (hSCF-220), a predominantly soluble isoform of

KITL (hSCF248), or no KITL (Sl/Sl4) (Figure 4-6). In experiments using unsorted

PGCs, consisting of a mix of PGCs and somatic cells, 26.1 + 10.1% of PGCs seeded onto

Sl/Sl4 cells remained following agitation (Figure 4-6A). A larger percent (41.9 + 5.9%)

of unsorted PGCs adhered to cells expressing membrane bound KITL. In contrast few

PGCs (18.3 + 3.3%) remained adhering to cells expressing soluble KITL (Figure 4-6A).

To test the model that soluble KITL can interfere with PGC adhesion to

membrane bound KITL, unsorted PGCs were cultured on feeder layers in the presence or

absence of KITL(s). In 3 experiments, KITL(s) treatment did not affect adhesion of

PGCs to either hSCF-220 cells or hSCF-218 cells. Since we were not purifying PGCs,

cell behavior in these assays might have been affected by the presence of contaminating


somatic cells. To test this, MACS sorting was used to purify PGCs prior to seeding onto hSCF-220 or hSCF-248 cells. Following treatment with 1 μg/ml KITL(s) there was no significant change in PGC adhesion to either cell line (Figure 4-6B). Additionally, unlike unsorted PGCs, purified PGCs did not adhere more firmly to cells expressing membrane bound KITL (compare Figure 4-6A and B). These results are in contrast to a previously published study in which purified PGCs were able to adhere more firmly to cells expressing KITl(mb) (128). Pesce et al. (128) were also able to show that exogenous

KITL(s) could reduce PGC adhesion to cells expressing KITL(mb). We believe that the difference between our results and the Pesce results are procedural. For example, our

MACS purification procedure is fairly long and it is possible that cell lysis during the procedure might release DNA resulting in aberrantly high baseline adhesion. Baseline adhesion of unpurified PGCs (to cells expressing KITL(s)) was only 18.3 percent (Figure

4-6A); however MACS purified PGCs exhibited 40% binding to cells expressing

KITL(s) (Figure 4-6B).

4.5 Discussion

We have shown that migratory PGCs continue to rely on somatic expressed KITL during their migration and colonization of the genital ridges. Blocking of the KIT receptor by ACK2 leads to a gradual loss of E9.5 PGCs in organ culture. ACK2 treatment interferes with PGC migration leading to decreased PGC velocity and targeting of the genital ridges. While PGCs require KITL, they are sensitive to the isoform of

KITL. Addition of exogenous KITL(s) was found to decrease PGC survival and


targeting. This outcome is counterintuitive of a model where KITL promotes both PGC survival and migration.

We propose an adapted model where endogenous KITL(mb) supports PGC survival and modulates migration. Addition of KITL(s) effectively blocks the interaction between KITL(mb) and KIT, inhibiting the action of endogenous KITL(mb). Exogenous

KITL(s) is insufficient to promote PGC survival and causes a decrease in PGC numbers.

This decrease is not as severe as observed with ACK2 treatment indicating some KITL(s) is better than no KITL.

In a previous study, addition of KITL(s) was proposed to increase midline PGC numbers by decreasing midline PGC apoptosis (15). Our data support an alternative model. While we observed that exogenous KITL(s) does increase the number of midline

PGCs, this effect is unlikely to be due to increased survival since the overall numbers of

PGCs actually decrease in response to KITL(s) treatment. We propose that the mechanism leading to midline accumulation is defective PGC migration. Time-lapse analysis of PGC migration from E9.5 to E10.0 reveals that KITL(s) treatment leads to a reduced ability of PGCs to target the nearest genital ridge. This results in a failure of

PGCs to clear the midline, leading to an accumulation of midline PGCs. In control midline tissue, PGC numbers are low following culture as PGCs migrate out of this tissue or undergo apoptosis. In treated slices, PGCs fail to migrate out of the midline leading to an increase in midline PGC numbers but a decrease in the total number of PGCs per slice.

During the course of our 18-hour assay many PGCs in control slices have effectively cleared the midline and reached the genital ridges, where Kitl expression is high (15, 92).

Based on our results we propose that the form of KITL present in the genital ridges is


KITL(mb). KITL(s) addition leads to inhibition of KITL(mb) and leads to two outcomes.

First, PGCs located outside of the genital ridges accumulate in the midline. There, numbers decline somewhat as exogenous KITL(s) provides minimal support for their survival. Second, PGCs exposed to high levels of endogenous KITL(mb) within the genital ridges drop in numbers due to inhibition of KITL(mb) by KITL(s). The net outcome is an increase in midline PGC accumulation and a decrease in total PGC numbers.

PGC migration is also disrupted in response to exogenous KITL(s) as evidenced by decreased targeting of the E9.5 genital ridges and increased PGC emigration from the

E10.5 genital ridges. We propose that KITL(mb)/KIT mediated PGC/soma adhesion regulates PGC migration to the genital ridges and tethers PGCs to their target tissue once they arrive in the ridge. As PGCs migrate they interact with somatic cells expressing a gradient of KITL(mb), with the highest levels in the genital ridges. PGCs crawl along this somatic trail as they migrate. Upon arriving in the ridge, the trail of KITL(mb) in the midline and gut is removed and the high level of genital ridge KITL(mb) holds them in the target tissue. The tethering of PGCs within the ridge may reduce the likelihood of

PGCs migrating to ectopic locations and forming germ cell tumors. Treatment with

KITL(s) inhibits this adhesive interaction leading to a loss of this tethering. As a result

E9.5 PGCs fall off of the somatic KITL(mb) track and migrate with less directionality towards the genital ridges. In E10.5 tissue, exogenous KITL(s) loosens the adhesion holding them within the ridge leading to PGCs migrating away.

Our adhesion model is consistent with the previous observation that

KITL(mb)/KIT interact to form an adhesive bond between PGCs and somatic cells (128).


In cell culture, PGCs were found to adhere more tightly to a cell feeder layer expressing

KITL(mb) than expressing KITL(s). In addition, they showed that exogenous KITL(s) was capable of reducing PGC adhesion to KITL(mb) expressing cell layers. In our study we were able to repeat only those findings that cells expressing KITL(mb) adhere PGCs more tightly than those expressing KITL(s). Unfortunately, we were unable to repeat the competition experiments. A possible flaw in the experimental design of using cells expressing the human form of KITL rather than the mouse form may be at the root of our discrepancy. In future experiments, PGCs should be cultured on cells expressing mouse

KITL(mb) or KITL(s), complementing the PGC expressed mouse KIT and the exogenous mouse recombinant KITL(s). Alternatively, exogenous human KITL(s) could be used in adhesion assays using mouse PGCs on human KITL expressing feeder layers.

4.6 Acknowledgements

We thank Patti Conrad for microscopy assistance and acknowledge the use of the

Leica widefield and confocal microscopes in the Genetics Department Imaging Facility at

Case Western Reserve University, funded by the National Center for Research Resources

(NIH-NCRR) Shared Instrumentation Grant program (1S10RR021228 and

1S10RR017980). Financial support for this work was provided by Case Western Reserve

University and the NIH (R01HD053900 awarded to KM).


4.7 Supplemental Material

Movie 4-1: Time-lapse confocal microscopy of untreated E9.5 PGC migration. Oct4
PE:GFP/ Oct4
PE:GFP males were crossed with CD1 females and embryos harvested at E9.5. Transverse sections between the fore- and hind-limb buds were cultured overnight on the stage of a confocal microscope. Images were collected of each slice at 7-minute intervals for 100 frames and processed using Volocity software. Images were compiled into movies at 10 frames per second.

Movie 4-2: Time-lapse confocal microscopy of KITL(s) treated E9.5 PGC migration. Oct4
PE:GFP/ Oct4
PE:GFP males were crossed with CD1 females and embryos harvested at E9.5. Transverse sections between the fore- and hind-limb buds were cultured overnight in the presence of 40 ng/ml KITL(s) on the stage of a confocal microscope. Images were collected of each slice at 7-minute intervals for 100 frames and processed using Volocity software. Images were compiled into movies at 10 frames per second.

Movie 4-3: Time-lapse confocal microscopy of ACK2 treated E9.5 PGC migration. Oct4
PE:GFP/ Oct4
PE:GFP males were crossed with CD1 females and embryos harvested at E9.5. Transverse sections between the fore- and hind-limb buds were cultured overnight in the presence of 9 g/ml ACK2 on the stage of a confocal microscope. Images were collected of each slice at 7-minute intervals for 100 frames and processed using Volocity software. Images were compiled into movies at 10 frames per second.

Movie 4-4: Time-lapse confocal microscopy of untreated E10.5 PGC migration. Oct4
PE:GFP/ Oct4
PE:GFP males were crossed with CD1 females and embryos harvested at E10.5. Transverse sections between the fore- and hind-limb buds were cultured overnight on the stage of a confocal microscope. Images were collected of each slice at 7-minute intervals for 100 frames and processed using Volocity software. Images were compiled into movies at 10 frames per second.

Movie 4-5: Time-lapse confocal microscopy of KITL(s) treated E10.5 PGC migration. Oct4
PE:GFP/ Oct4
PE:GFP males were crossed with CD1 females and embryos harvested at E10.5. Transverse sections between the fore- and hind-limb buds were cultured overnight in the presence of 40 ng/ml KITL(s) on the stage of a confocal microscope. Images were collected of each slice at 7-minute intervals for 100 frames and processed using Volocity software. Images were compiled into movies at 10 frames per second.

Movie 4-6: Time-lapse confocal microscopy of KITL(s) treated E10.5 PGC migration displaying an example of a PGC crossing the midline. Oct4
PE:GFP/ Oct4
PE:GFP males were crossed with CD1 females and embryos harvested at E10.5. Transverse sections between the fore- and hind-limb buds were


cultured overnight in the presence of 0.4 ng/ml KITL(s) on the stage of a confocal microscope. A midline crossing PGC is first observed 7 seconds into the movie near the top genital ridge. It then migrates towards the gut then down into the lower genital ridge. Images were collected of each slice at 7-minute intervals for 100 frames and processed using Volocity software. Images were compiled into movies at 10 frames per second.


Chapter 5: Discussion and Future Directions

By

Brian Mason Dudley


I have investigated the role of bone morphogenetic protein (BMP) signaling

during primordial germ cell (PGC) development. Through in vitro and in vivo studies, I

have found that BMPs expressed in the E9.5 genital ridges promote PGC survival and

migration through the establishment and maintenance of a PGC niche. Within this niche

environment, BMPs support the survival of pronephric cells while suppressing the

expression of the epithelial marker gene Scarb1. In the underlying coelomic epithelium,

mesonephric derived BMPs promote the expression of the PGC required cytokine Kit

ligand (Kitl). As PGCs colonize this tissue, they rely on high KITL levels for survival.

Here I propose that the membrane bound (mb) isoform of KITL is the predominant

regulator of PGCs in the genital ridges, and KITL(mb) interacts with PGC expressed KIT

receptor to tether PGCs to the somatic cells of the ridge. Loss of BMP signaling or the

adhesive interaction between KITL(mb) and the KIT leads to decreased PGC survival,

decreased PGC targeting of the ridge, and decreased PGC tethering within the ridge.

Taken together, BMPs act through KITL(mb) to establish a PGC niche within the genital

ridges complete with survival factors, adhesive interactions, and regulated size.

5.1 BMP signaling during PGC colonization of the genital ridges

The migration of PGCs to the genital ridges is concurrent with the development of

various genital ridge structures. At E9.0, cells within the developing pronephros initiate

expression of the kidney progenitor transcription factor Pax2. By E9.5, Bmp4 is expressed within the mesonephric mesenchyme, midline and dorsal gut mesentery, and

Bmp7 is expressed within the adjacent mesonephric duct. Cells within the E9.5


mesonephric mesenchyme and duct, the underlying coelomic epithelium, and to lesser extent the midline, dorsal gut mesentery, and coelomic epithelium are all responding to these BMP signals. By E10.5 midline and gut mesentery cells show reduced BMP signaling, while the mesonephric mesenchyme and duct. At this stage PGCs are located near the coelomic epithelium and mesonephric mesenchyme.

It is likely that PGCs do not respond directly to endogenous BMPs. pSMAD1/5/8 positive PGCs are only observed when exogenous BMP4 is added to cultured slices.

While BMPs can act through other SMAD independent pathways, there is little evidence of this occurring in PGCs in vivo. PGCs do show activated MAPK signaling, a possible non-SMAD dependent pathway member, however, this has been attributed to FGF signaling (162). There is prior evidence of BMPs and FGF acting synergistically. In the mouse ovary, cumulus cells upregulate factors necessary for glycolysis only in response to both BMP15 and FGF8b (195). However, cumulus cells do have active SMAD signaling, suggesting the cooperation may still require SMAD signaling (195). In granulosa cells, FGF8 has been shown to enhance SMAD dependent BMP signaling through activation of the JNK signaling pathway (196). In both of these cases, BMPs continue to act through SMADs. However, to conclude that BMPs are not acting directly on PGCs in vivo, SMAD dependent BMP signaling should be knocked out in PGCs. One way to do this would be to create a Stella:CreER line of mice that specifically expresses an inducible Cre in migratory PGCs. Using the Bmpr1a-fx allele, we could then target

BMP signaling in migratory PGCs. If the reduced capacity to respond to BMPs has no affect on PGCs, then it is unlikely that BMPs are acting directly on PGCs in vivo.


If PGCs are not responding directly to BMPs, then it is likely BMPs are acting on

PGCs through regulation of their somatic environment. Within the mesonephric

mesenchyme, BMPs maintain the survival of pronephric cells, loss of which results in

disrupted PGC migration and survival. In coelomic epithelium, BMP signaling

suppresses the epithelial fate while driving expression of Kitl.

PGC survival and migration depends on the successful patterning of the genital ridge by BMPs. Within 24 hours of Bmpr1a conditional knock out there is no measurable decrease in pSMAD1/5/8 levels within the E9.5 ridge, yet cells within the mesonephric mesenchyme begin to undergo apoptosis. At this stage, PGC numbers in conditional knock out embryos do not differ from those in wild type embryos. Twelve hours later the cell death persists and, in extreme examples, most cells within the mesonephric mesenchyme are undergoing apoptosis, with no increase observed in the coelomic epithelium. During this time, PGCs in conditional knock out slices begin to show defects in targeting and a decline in numbers. The occurrence of mesonephric mesenchyme cell death prior to a detectable decrease in pSMAD1/5/8 indicates that these cells are highly sensitive to the level of BMP signaling. A slight, immeasurable drop is sufficient to change the landscape of the genital ridge. However, PGC loss follows the onset of somatic cell death, suggesting that their decreased survival is secondary to the collapse of the mesonephric portion of the genital ridge.

Mesonephric cell death is only half of the story. In addition to the loss of mesonephric mesenchyme cells, the expression profile of the ridge begins to change in response to decreased BMP signaling. Kidney progenitor markers such as Pax2 are down

regulated while epithelial markers such as Scarb1 are up regulated. The decrease in Pax2


expression could be evidence of BMP regulation of Pax2, as previously described (71).

However, Pax2 expression was not significantly increased in BMP4 or BMP5 treated

E9.5 slices in tissue culture experiments. Alternatively, the drop in Pax2 levels may be secondary to the death of Pax2 expressing mesonephric cells. Analysis of apoptosis in

Bmpr1a knock out genital ridges revealed the greatest cell death within the mesonephric mesenchyme, the same region of high Pax2 expression. At this stage in this region of the embryo, Pax2 is predominantly expressed within the mesonephric mesenchyme, with little to no expression detected elsewhere in the slice. Loss of this population of cells in conditional knock out embryos disproportionately reduces the level of Pax2 transcript relative to ubiquitously expressed housekeeping genes. If Pax2 were widely expressed, and only a subset of Pax2 expressing cells were dying, normalization to housekeeping genes would be sufficient to account for the decreased levels of Pax2. However, if a high percentage of Pax2 cells were lost, then the relative decrease in Pax2 levels would be greater than the decrease in housekeeping gene levels, resulting in an apparent down regulation of Pax2. In these experiments, RNA was collected only from genital ridges, enriching the contribution of mesonephric mesenchyme cells. Under these conditions, we would assume that cell death would reduce the expression of housekeeping genes to a similar extent as Pax2. However, we see a more substantial decrease in Pax2 expression suggesting that cell death may not be the only factor at play. Barring Pax2 in situ data combined with apoptosis staining, I cannot conclude whether BMPs regulate the expression of Pax2 in vivo or are necessary for the survival of Pax2 expressing cells.

Regardless of the mechanism, loss of BMP signaling within the genital ridge leads to a decline in pronephric cell fate.


Cell death in the mesonephric mesenchyme of Bmpr1a conditional knock out embryos is not mirrored in the coelomic epithelium, the source of KITL and SDF1a. In this region, data suggest that BMPs control expression of the epithelial marker Scarb1.

Loss of BMP signaling leads to increased expression and, preliminary data suggests,

BMP treatment leads to decreased expression of Scarb1. This is in line with evidence in the zebrafish ridge where BMP signaling has been shown to promote the pronephric fate while repressing endothelial and hematopoietic fates (69). In future experiments, a panel of epithelial markers, such as LIM homeobox protein 9 (Lhx9) (192) or steroidogenic factor 1 (Sf1, a.k.a Nr5a1) (197), should be checked to conclude that BMPs are repressing the epithelial fate in the mouse genital ridge. Interestingly, BMP treatment may slightly repress Sf1 expression in tissue culture (174)

I propose that BMP signaling within the E9.5 genital ridge establishes and maintains a PGC niche (Figure 5-1). Mesonephric mesenchyme expressed BMPs act on mesonephric mesenchyme cells to promote survival while repressing expansion of the epithelial cell lineage. Within the coelomic epithelium, BMP signaling represses Scarb1 expression while driving the expression of Kitl. BMP activated pSMAD1/5/8, or a downstream target, is both negatively regulating Scarb1 and positively regulating Kitl in the same population of cells. This creates a high KITL concentration within the ridge supporting PGC survival and targeting of the ridge.

5.2 BMP regulation of Kitl in the genital ridges

As PGCs migrate from the hindgut to the genital ridges from E9.0 through E10.5 their environment is undergoing substantial change. One constant is the expression of



Figure 5-1: Bone morphogenetic protein (BMP) signaling in the E9.5 genital ridge establishes a permissive environment for arriving primordial germ cells. Mesonephric mesenchyme (MM) and duct (MD) derived BMPs lead to the phosphorylation of SMAD1/5/8 (p-smad) in MM and coelomic epithelium (CE) cells. Within the MM, active BMP signaling promotes cell survival and a pronephric fate. In the CE, BMP signaling represses the expression of the epithelial marker Scarb1, while promoting the expression of Kitl. This leads to the increase in membrane bound (mb) KITL levels within the coelomic epithelium. PGCs are represented by green ovals.


Kitl by the somatic cells they encounter. Upon reaching the ridge, the coelomic

epithelium has the highest level of KITL. PGCs residing in the ridges rely on KITL for

survival while ectopic PGCs likely undergo apoptosis. Maintaining high levels of KITL

in the genital ridges is absolutely crucial for the successful colonization of these

structures by PGCs.

In vitro and in vivo evidence suggests BMP signaling within the E9.5 genital

ridge indirectly regulates the expression of Kitl. At E9.5, cells within the coelomic epithelium, midline, and dorsal gut mesentery are all actively responding to BMP signaling through activation of SMAD1/5/8. BMP signaling during this stage parallels

Kitl expression. By E10.5, BMP signaling becomes restricted to the genital ridges. At this stage KITL levels are highest in the coelomic epithelium. To test if BMPs regulate

Kitl expression within the genital ridges, cultured E9.5 slices were treated with exogenous BMP4 or BMP5. Following treatment, Kitl levels were increased after 18 hours. Inhibition of BMP signaling pharmacologically, using Noggin, or in vivo through conditional knock out of Bmpr1a, led to down regulation of Kitl. Despite this apparent

control, BMPs most likely do not directly regulate the expression of Kitl. First, Kitl

expression only changes after 18 hours in culture with BMPs or Noggin, while expression

of the known target gene Id1 responds in less than 3 hours. Second, in silico analysis of

the proximal Kitl enhancer region and promoter failed to reveal any conserved BMP

response elements (BRE), characteristic SMAD binding sites suggestive of direct BMP

targets.

I propose that BMPs indirectly regulate the expression of Kitl either through

control of cell fate decisions within the genital ridge or through the expression of an


intermediate transcription factor. BMPs appear to promote a pronephric fate within the

genital ridges while suppressing an epithelial fate. Kitl expression could be controlled via

cell fate decisions or survival, in a similar fashion to Pax2 in the mesonephric

mesenchyme. However, this mechanism is not well supported by the Bmpr1a knock out

experiments. In these experiments, increased cell death was only observed in the

mesonephric mesenchyme of conditional knock out embryos, and not in the coelomic

epithelium, the source of Kitl. Instead, Kitl expression is decreasing in cells not

undergoing apoptosis.

More likely, Kitl expression is being affected at the level of transcription. Given

the lack of evidence for direct regulation of Kitl by BMP signaling, BMPs likely regulate

the expression of an unidentified transcription factor within the coelomic epithelium,

such as Pax2, Lhx9, or Sf1. Initial in silico analysis of the proximal Kitl enhancer and

promoter revealed several conserved transcription factor binding sites including Pax2.

While Pax2 may be regulated by BMPs, it is not expressed by coelomic epithelium cells,

and therefore is unlikely to be regulating Kitl expression. Further analysis of the Kitl

enhancer may yield more candidate intermediate factors downstream of BMP signaling

and upstream of Kitl.

5.3 The role of BMP driven Kitl expression in PGC survival and migration

Throughout this research a common theme has emerged: increased Kitl correlates

with increased PGC survival and decreased Kitl correlates with reduced PGC survival and disrupted PGC migration. KITL has long been known to be necessary for PGC


survival, and in this study I was able to confirm the requirement persists through PGC colonization of the genital ridges. Additionally, I showed that loss of KITL/KIT signaling nearly halts directional PGC migration by reducing PGC velocity and targeting.

What is less well known is the specific role of the membrane bound isoform of KITL during PGC migration.

While previous work has shown that the two isoforms of KITL are differentially expressed, it remained unclear which isoform was present in the genital ridges during

PGC migration. Various clues have pointed to KITL(mb) including apparent cell surface antibody staining, the drastic PGC deficit in Steel-dickie mice lacking membrane bound

KITL, and the predominance of KITL(mb) in the adult testis (15, 87, 105, 106). Here I provide novel evidence of KITL(mb) within the genital ridges and propose that this isoform acts with KIT to adhere PGCs to their somatic neighbors. I propose a model whereby PGCs migrate along a track of KITL(mb) directing them towards the genital ridges, where increased KITL(mb) tethers them to the developing gonad.

In Bmpr1a conditional knock out experiments, Kitl expression is significantly reduced within the E10.5 genital ridges. As PGCs attempt to colonize these structures, the somatic environment is in decay. The mesonephric mesenchyme has increased cell death and the coelomic epithelium has reduced Kitl expression. Arriving PGCs are no longer exposed to high KITL levels necessary for their survival, and as a result their numbers drop between E9.5 and E10.5. PGCs in conditional knock out embryos also show a defect in their ability to target the genital ridges. During PGC migration Kitl is expressed in a gradient with the highest levels localized to the coelomic epithelium of the genital ridges (92). In conditional knock out embryos this gradient is disrupted as KITL


levels drop in the target tissue. PGCs within the dorsal gut mesentery and midline are no

longer exposed to a KITL gradient and fall off track. PGC targeting is not completely

abolished indicating that other directional cues continue to have an influence on PGCs.

Yet more PGCs can be observed migrating towards ectopic positions, migrating in non-

directional small circles, or passing through the target tissue without stopping.

It is also possible that the migration defects observed in conditional knock out

embryos are not related to KITL. Other directional cues emanating from the genital ridge

could be disrupted or lost following reduced BMP signaling. One such factor is the

putative PGC chemo-attractant Sdf1a. At E10.5, conditional knock out genital ridge

Sdf1a expression is reduced to less than 60% that of the wild type. This reduction could account for the drop in PGC targeting. However, Sdf1a genital ridge expression is reduced to an equal level in conditional heterozygous (C-Het) embryos. Targeting in C-

Het embryos is only slightly and insignificantly reduced relative to wild type embryos, indicating that Sdf1a reduction alone is not sufficient enough to account for the migration defects in Bmpr1a conditional knock out embryos. Additionally, SDF was not able to attract migratory PGCs in transwell migration assays, leaving room for a role for KITL

(20).

Tissue culture experiments provide additional evidence that KITL(mb) plays a role in PGC migration to the genital ridges. A similar migration defect as observed in conditional knock out embryos is also observed in KITL(s) treated slices. In control slices, PGCs begin in the midline and dorsal gut mesentery at E9.5 and migrate laterally to colonize the genital ridges by the approximately E10.0-E10.5, with few PGCs remaining in the midline. From E10.5 to E11.5, PGCs cluster within the ridge as they


begin to associate with gonad precursors. When KITL(s) is added to cultured E9.5 slices, migratory PGCs show significantly reduced targeting and increased ectopic accumulation in the midline. Following treatment of E10.5 tissue, some PGCs are observed exiting the ridge. KITL(s) treatment effectively loosens the grip of the somatic cells of the genital ridge on PGCs.

KITL(mb) has previously been shown to interact with PGC expressed KIT to adhere PGCs to somatic cells in culture (128). Addition of KITL(s) was found to inhibit this adhesive interaction. In this work, I have shown evidence of this interaction occurring in the mouse genital ridge. From E9.5 through E11.5, KITL(mb) levels are high in the genital ridges. As PGCs colonize this structure, they adhere to their somatic environment through the interaction between KITL(mb) and KIT (Figure 5-2A). This tethers them to the genital ridges preventing them from migrating away from this niche environment and may provide a stopping mechanism for arriving PGCs. As PGCs colonize the ridge, it may be necessary for the somatic cells to tether them in order to prevent them from passing through. In Bmpr1a conditional knock out embryos the level of KITL(mb) within the genital ridges drops, leading to a loss of this adhesive interaction.

As a result PGCs fall off of their track, show defects in migration, and die (Figure 5-2B).

If KIT is blocked using ACK2, the interaction between KITL(mb) and KIT is completely lost and PGCs fail to target the ridge and die (Figure 5-2C). If KITL(s) is added, excess of the soluble form of KITL competitively inhibits the interaction between endogenous

KITL(mb) and KIT leading to PGCs no longer adhering to their somatic track and failing to migrate correctly to the genital ridges (Figure 5-2D). In E10.5 slices, loss of adhesion causes PGCs to migrate away from the developing gonad either subtly through the



Figure 5-2: Membrane bound (mb) KITL tethers primordial germ cells (PGC) to their somatic neighbors in the genital ridges. A) In controls, PGCs migrate along a gradient of KITL(mb) towards the genital ridges where high levels of KITL(mb) tether them to the somatic cells of the ridge. B) If BMP signaling is reduced in vivo through conditional knock out of Bmpr1a, KITL(mb) levels drop within the ridge and PGCs loose their adhesion to the somatic cells of the ridge. C) Treatment with the KIT blocking antibody ACK2 blocks the interaction between KITL(mb) and KIT leading to loss of adhesion. D) If soluble (s) KITL is added it competitively inhibits the adhesive interaction between KIT and endogenous KITL(mb). GR = genital ridge.


loosening of the PGC cluster or more apparently through PGCs migrating to ectopic locations. The chemo-kinetic properties of KITL(s) could reactivate migration in PGCs no longer tethered to the genital ridges, increasing migration away from the ridge.

The tethering of stem cells within a niche is not a unique observation. In the

Drosophila ovary the germ line stem cells are anchored to cap cells, which provide necessary signals for maintaining the stem cells. In the mammalian hematopoietic stem cell (HSC) niche, HSCs are attached spindle-shaped N-cadherin+ osteoblastic cells, which secrete factors promoting HSC survival and pluripotency. KITL(mb) likely provides the same function within the mouse genital ridge PGC niche, holding PGCs in proximity to necessary regulatory factors.

KITL(mb) serves multiple functions in the E9.5 to E10.5 PGC niche. PGCs rely on a KITL(mb) gradient to target the genital ridges, and depend on KITL(mb) for survival. Once within the ridge, KITL(mb) promotes PGC survival and tethers them to the PGC niche. By holding PGCs within the ridge, they remain close to survival cues as well as factor preventing their early differentiation.

5.4 Future work

Many additional questions remain regarding the PGC niche. While BMPs are necessary for the survival of the pronephric compartment of the genital ridge and for Kitl expression in the coelomic epithelium it is unclear if both are necessary for PGC survival and migration. I have proposed that the mesonephric mesenchyme may act as an intermediate target for migratory PGCs, however, it is possible that defects in PGC


migration are completely due to the loss of factors expressed in the coelomic epithelium,

such as Kitl and Sdf1a. To determine the contribution of each compartment, KITL(mb) and SDF1a could be added back to Bmpr1a knock out genital ridges. This would attempt to restore coelomic epithelium KITL(mb) and SDF1a levels while allowing for the degradation of the mesonephric mesenchyme. Conversely, Bmpr1a conditional knock

out experiments could be done on a Bax-/- background, deficient in a key apoptotic

pathway member. This could block mesonephric mesenchyme cell death while

presumably allowing for down regulation of Kitl and Sdf1a. If one or both of these targets no longer shows reduced expression, this may also shine light on the mechanism of BMP regulation of Kitl.

It remains unclear how BMP signaling in the mesonephric mesenchyme and coelomic epithelium regulates the expression of Kitl. Due to the delayed response of Kitl expression and the lack of apparent BMP response elements within the proximal Kitl enhancer, it is unlikely that BMPs directly regulate the expression of Kitl. To characterize this regulation, I have begun examining the Kitl enhancer using Chromatin

IP followed by sequencing. This should provide insight into the regulatory regions of

Kitl.

I have proposed a model where KITL(mb) acts as a track that PGCs are tethered to as the migrate towards the genital ridges, following this KITL(mb) gradient. While I have provided evidence that KITL(mb) is the predominant form of KITL in the genital ridges, I have not provided direct proof. Additional experiments looking at the expression of the membrane bound and soluble forms of KITL within the ridge are necessary to conclusively show that one form predominates. If KITL(mb) is the


predominant form, and PGCs follow a track of increasing KITL(mb) levels, it would be interesting to create an artificial track of KITL(mb) in cell culture or ideally in tissue culture to see if PGCs would accumulate at an ectopic source of KITL(mb). Previous experiments have provided evidence of a soluble attractant directing PGC migration, suggesting that PGCs may not follow an ectopic KITL(mb) track (17, 20). However, these experiments only looked at soluble attractants, overlooking the possible contributions of membrane bound directional cues. PGC migration is likely affected by many cues, as is evidenced by some experiments showing that KITL(s) or SDF are attractants while others show the opposite. I propose that a gradient of KITL(mb) provides additional support for PGC targeting of the genital ridges.

Lastly, while I have shown that BMP signaling is necessary to maintain the PGC niche and KITL(mb) acts to support PGC survival while tethering them to the niche, it is unclear what other factors are necessary within the ridge during this time. It is likely that the niche provides cues supporting PGC proliferation and blocking differentiation.

Additional work should be done to further characterize this environment.

In this work I have identified a niche environment within the pregonadal genital ridge. This microenvironment provides a permissive environment for arriving PGCs and holds them in place until sexual differentiation and subsequent development begins. I have shown that this niche is established and maintained through BMP signaling, loss of which leads to the collapse of the niche environment. A primary somatic factor supporting PGC survival and retention within the ridge is KITL(mb). PGCs interact with their somatic neighbors through the adhesive interaction between PGC expressed KIT and somatic cell expressed KITL(mb).


References

>8
#$,0*$)
8
#
$ )/$!$/$*)5
*-$"$)5
)
($"-/$*)
*!
/#
+-$(*- $'
" -(
 ''.
$)
/#
(*0.
(-4*8
)/
 8
>FBA
 6>>E:?;7>@B9AC8

?8
$)"
5
$
5
#*
8
) 0/$*)
*!
+-$(*- $'
" -(
 ''.
!-*(
(0-$)
+$'./.
4
.4) -"$./$
/$*)
*!
A
)
E
.$")'$)"
+/#24.8
-*
/'

$


8
?==>
0'
@6FE:>A;7DEBE9C?8

@8
$)"
5
#*
8
**+ -/$*)
*!
) * -(9 -$1
?
)
3/- (-4*)$
/* -(9 -$1
A
$)
+-$(*- $'
" -(
 ''
" ) -/$*)
$)
/#
(*0. 8
 1
$*'8
?==>
+-
>B6?@?:?;7AEA9F?8

A8
$/*0
5
-/*)
5
0-)$
8

(*' 0'-
+-*"-((
!*-
/#
.+ $!$/$*)
*!
" -(
 ''
!/
$)
($ 8
/0- 8
?==?
0'
>E6A>E:CEFB;7?F@9@==8

B8
2.*)
5
0))
5
* ' )
5
 $)./-
5
1$.
5
-$"#/
5
/
'8
(+A
$.
- ,0$-
!*-
/#
" ) -/$*)
*!
+-$(*- $'
" -(
 ''.
$)
/#
(*0.
(-4*8
 ) .
 18
>FFF
 
>B6>@:A;7A?A9@C8

C8
#$)/
5
#/
5
#$" /
5
()&
5
&4(
5
$/*0
8

.$")'$)"
+-$)$+'
!*-
/#
.+ $!$/$*)
*!
/#
" -(
 ''
'$) "
$)
($ 8
 ''8
?==F
4
>6>@D:@;7BD>9EA8

D8
#$)/
5
4 -
5
<--*''
5
) '$)
5
)*
5
)*
5
/
'8
'$(+>
$.

-$/$'
/ -($))/
*!
/#
" -(
 ''
'$) "
$)
($ 8
/0- 8
?==B
0'
>A6A@C:D=AE;7?=D9>@8

E8
$)"
5
$0
5
-'
5
2.*)
5
#*
8
 ,0$- ( )/
*!
(+E
!*-
/#
" ) -/$*)
*!
+-$(*- $'
" -(
 ''.
$)
/#
(*0. 8
*'
) *-$)*'8
?===
0'6>A:D;7>=B@9C@8

F8
0/
5
0-$(*/*
5
#$)/
5
 &$
5
$/*0
8
 )
3+- ..$*)
4)($.
0-$)"
" -('$)
.+ $!$/$*)
$)
($
$ )/$!$
4
,0)/$//$1
.$)"' 9 ''
" )
3+- ..$*)
+-*!$'$)"8
$*'
 +-* 8
?==C
*16DB:B;7D=B9>C8

>=8
0-$(*/*
5
0/
5
#$)/
5
#$" /
5
()&
5
$/*0
8
*(+' 3
" )*( 92$
/-).-$+/$*)
4)($.
*-# ./-/
4
'$(+>
!*-
/#
.+ $!$/$*)
*!
/#
" -(
 ''
'$) "
$)
($ 8
 ) .
 18
?==E
0)
>B6??:>?;7>C>D9@B8

>>8
0-$(*/*
5
(%$
5
 &$
5
$/*0
8
+ $!$/$*)
*!
/#
" -(
 ''
'$) "
$)
($ 7

+-* ..
*-# ./-/
4
/#
9 *($)
+-*/ $).5
'$(+>
)
- (>A8
 ''

4' 8
?==E
*1
>B6D:??;7@B>A9E8

>@>
<=6
 $#
3
1+"#
3
,("
3
#2-!$#
3
#,(-
3
,+-#
6
0, '+#.
'
(* *%1
* )*(!*&&#'!
(
! '(& 7/#
"*(&,#'
&(# #,#('+
++(#, 
/#,"
+) # #,#('
'
*%1
 . %()& ',
(
! *&
 %%+
#'
&# 6
 .
#(%6
=;;@
 
<@4=BC9=:5??;7@C6

<>6
' *+('
3
() %'
 3
"(% *
3
+&'
3
1%#
6
"
('+ ,
(
! *&
 %%
&#!*,#('
#'
,"
&(-+
&*1(6
 "
 .6
=;;;
*
<4D<9<7=:5A<7C6

>9=?:5?CA<7D6

A6

@B7A>6


 )4<>;9;A9=:5@B=7C>6

=<6
(1(($
3
+-' $/
3
$"+"#
3
,+-#
3
,("
3
(
6
0)* ++#('
'
#',* %%-%*
%(%#2,#('
(
&(-+
+7"(&(%(!-
)*(, #'
-*#'!
! *&
 %%
 . %()& ',6
 "
 .6
=;;;
14D>9<7=:5<>D7?D6

==6
#&-*
3
(&(!#
3
/#
3
,(
3
$'(
6
(% -%*
%('#'!
'
! '(&#
(*!'#2,#('
(
&(-+
"(&(%(!-
(
*(+()"#%
! *&
 %%7% ++
'
#,+
0)* ++#('
#'
! *&
%#' !
 %%+6
#(" &
#()"1+
 +
(&&-'6
7>;6

=>6
 * '
6
*#&(*#%
! *&
 %%+
#'
,"
&(-+ 6
 .
#(%6
=;;>
,
<4=A=9<:5<7<@6

<>=
<>6
 &!*
3
 -+,*
3
*!,+
 3
-
3
$+'&
3
-*
3
,
$6
-$,!('!&,
!%(*!&,!&
&$1+!+
!&!,+

'%%'&
(*-*+'*
$$
'*
'&$
&
&'& '&$
(!,*!
*%
$$
,-%'*+6
&*
+6
<::;
,
;4@;8;C95A<@B7A@6


,
;?48<95:8;95C<7;:A6


,
+
&, '&!+,!
+! &$+
,'
* -$,
%%%$!&
+0
,*%!&,!'&6
 '
!'$6
<::@
-&4>8@95;BA6

==6
&,
3
 ,$1
3
&*/+
3
!&$!*
 3
''(%&
6

%$7+(!!
*'$
'*
C
!&
.*,*,
+0
,*%!&,!'&6
.$'(%&,6
;CC@
(4;<<8C956

''!"
3
*'', '
6
(*%,' '&!$
+,%
$$+6
-**
(!&
$$
!'$6
;CCB
4;:8@95@C>7A:;6

=?6
 !#*!
3
!-
6
'$-$*
% &!+%+
-&*$1!&
, 
,!.,!'&
'
%%%$!&
(*!%'*!$
'$$!$+6
&'*
.6
<::C
- 4=:8?95>=B7@>6

=@6
1*+
3
*!,,
3
*'*
3
$!&
 3
**
6
, '+
'*
)-&,!1!&
'$$!-$*
&-%*+
/!, !&
, 
%'-+
'.*16
(*'-,!'&6
<::>
14;

;==
>B7
 "&'((
7
.)*-
(
-.,.-
#(
'''&#(
))2. -6
,  (.
0( -
#(
/( ,-.(#(!
."
, !/&.#)(
)
' #).#
,, -.
(
))2.
'./,.#)(7
 *,)/.#)(7
=;;@
 5<>;9A:6BD<8D7

>C7
 . ,-
7
"
 0 &)*' (.
)
."
')/-
)0,2
,)'
#,."
.)
'./,#.27
.
(),#()&
9)* (":7
D7
--!/
7
8 .
-#!(&
.,(-/.#)(7
((/
 0
#)" '7
8 D<7

?;7
.'(
4
 (#-
4
,%.
4
)0 &(
7
&&
#(
."
'#&26
8 .
'#&2
.#)(
#(
. -.#-
 0 &)*' (.7
 *,)/.#)(7
=;;A
/!5<>=9=:6=>>8?A7

?<7
. (
#$%
4
#23)()
4
&#(
 7
#!(&#(!
#(*/.-
)(0 ,!
)(
(/& ,
.),-
#(
8 .
-#!(&#(!7
, (-
#)" '
#7
=;;;
 5=@9=:6A?8B;7

?=7
#'' ,'(
4

-/-8-),
4
,&(
7
"
* '((
),!(#3 ,
-#!(&
()!!#(
#(-
(
#(.#0. -
)(
'),*")! ( .#
*,). #(
?7
 &&7
5CA9?:6@DD8A;A7

?>7
")(
4
%
4
#'' ,'(
4
(
4
,&(
4
")(
7
)!!#(8' #. 
(.!)(#-'
)

-#!(&#(!
#-
, +/#, 
),
!,)1."
(
*.. ,(#(!
)
."
( /,&
./
(
-)'#. 7
 ( -
 07
C8@=7

??7
#)&)
4
-#
4
/
4

) ,.#-
7
),-)0 (.,&
*.. ,(#(!
#(
 ()*/-6
#("##.#)(
)
0 (.,&
-#!(&-
2
#, .
#(#(!
)
"),#(
.)
8?7
 &&7
5CA9?:6@CD8DC7

?@7
'-"#.
4
. (
#$%
4
/2& ,) %
4
'*."
 4
(,# -
4
'#."
4
.
&7
-. )! (#
*,). #(8<
#(-
.)
.#0#(
.2*

,  *.),-
(
#(/ -
 ,.#(
.#0#(8&#%
.-7

 &&
#)&7
;9<:6=>8?<7

?B7
"/
4
0-%
4
)&&"
4
,(
4
")'- (
 7


/#+/#.#(
&#!-
.,! .-
."

*."12
(
 .-
',2)(#
*.. ,(
),'.#)(7
./, 7
7

?C7
,%
4
/
4
,
4
."/,
4
)-'%#
4
#&", -.
7
)&
)
8 ?
(
.-
#!(&#(!
."12-
#(
/&./, 
/'(
 &()2. -7
(.

 &&
#)&7
=;;D5=;;D6B@;?C=7

?D7
")/
4
#(%
4
,!-
4
#((#%
4
,/--
4
)
4
.
&7

-#!(&#(!
#-

 (.,&
, !/&.),
),
8?
 * ( (.
*#&&,2
-*,)/.#(!7
,#)0-
 -7
=;;B
 
<5BA9>:6>D;8D7

<>?
>94
+ "
1
,"*"&
1
0)*
1
*(!
4
&,+"'&
'
('(+'*"*
/
=1

+)&*)"(+"'&$
+')
) ,$+
/
+)&* ')%"&
)'.+!
+')5+
+!)', !
*+)**5 +"-+
()'+"&
#"&*65 ,&
5+)%"&$
#"&*
7 6  8
*" &$"&
(+!./4

"'$
!%4
:BB@
+
<2;@;7=983;=@<:5=4

>:4
&
1
!',
1
,
1
'
1
&
 4
+"-+"'&
'
+!
!%+'('"+"
()' &"+')
#"&*5:
7  :85(&&+1
*+)**5+"-+
5 ,&
5+)%"&$
#"&*
7  8
(+!./
/
+)&* ')%"&
)'.+!
+')
+
75+85+"-+
#"&*
7 :81

#"&*
%"+')
'

+
*" &$
+)&*,+"'&4

"'$
!%4
:BB@
(
>2;@;74

>;4
% ,!"
1
!")#
1
!",/
1
)"
1
"*!"
1
&'
1
+
$4
&+" "+"'&
'

%%)
'
+!

%"$/
*

('+&+"$
%"+')
'
5 +
*" &$
+)&*,+"'&4
"&4
:BB>

;;2;@97>;==83;99A5::4

><4
'&0$05/*
4
+%
$$*1
&"!*
&
!)"&*3

-".
)'%
)'*'(!"$4

$$
"4
;99<
)
:>2::?7+
?83B=B5>=4

>=4
"
1
()$"&
4

&"!
%"&+"&"&
)%
$"&
*+%
$$*
"&
+!
)'*'(!"$
'-)/4
"&4
;999
+
:<2;B97>=B983<;A5<94

>>4
&
1
"&
4
(+)'*'%*
&
,*'%*
&!')
%"+'+"
*("&$*
,)"&
*/%%+)"
)%
$$
"-"*"'&*
&
"$"++
+!
')%+"'&
'

('$)"0
%")'+,,$
))/
')
''/+
*(" "+"'&
"&
)'*'(!"$4
-
"'$4
:BB@
(
:2:AB7:83@B5B=4

>?4
"
1
()$"&
4
(&+($ "
"*
**&+"$
')
+!
%"&+&&
&
"-"*"'&
'
)%$"&
*+%
$$*
"&
+!
)'*'(!"$
'-)/4
$$4
:BBA
,$
;=2B=7;83;>:5 ?94

>@4
'$"-)
1
)*'&
1
'(05&"-
1
'&0$05/*
4
&+"
"**+"'&
'

*+%
$$
&"!3
+!
*
'
+!
)'*'(!"$
'-)/4
-
/&4
;99?
'-2;<>7::83;B?B5@B4

>A4
"
1
()$"&
4
&
%(+/
)'*'(!"$
*+%
$$
&"!
)+"-+*
+!
()'$" )+"'&
'
+'("
$$*4
)'
+$

"


4
;99<
()
:>2:997A83=?<<5A4

>B4
'&
1
'&
1
.*
1
"
1
"&
1
)+!/
1
+
$4
%(
*" &$*
)'%
&"!
$$*
")+$/
)()**
+)&*)"(+"'&
'

" )&+"+"'&5()'%'+"&
&1

'
%)$*1
"&
)%$"&
*+%
$$*
"&
+!
)'*'(!"$
'-)/4
-$'(%&+4
;99=
)2:<:7?83:<><5?=4

?94
.*
1
'&
1
"&
1
"
4
6%(
*" &$"&
"*
**&+"$
')
%"&+"&"&
)%$"&
*+%
$$*
&
')
)()**"&
%
+)&*)"(+"'&
"&
+!
)'*'(!"$
+*+"*4
-$'(%&+4
;99=
)2:<:7?83:5@>4

:<>
A<7
 +"'
 4
)+%"'
7
 8( 8&+%,6

+(,()!"%
'
+*."+
-(
"'"-"-
(-!
&%
'
&%
&-( ',",7
',
/7
A@<8A=7

A>7

'-
++
4
/'
'
+"'$
4
(+-,
 7
/%()&'-
'
" +'-"-"('
(
-!
"'-,-"'%
)"-!%".&7
%%
(%
" 
"7
=;;>
.%5A;9B:6<>==8>=7

A?7
((+
4
&",!$
7
-&
%%,
'
-!"+
'"!,7
"'7
=;;A
+
><5><<9@BAD:6A9<;:6<<
-
=>5?=@9ADA;:6C>A8?<7

AB7
%",,('
4
(',,('
7
!
!&-()("-"
,-&
%%
'"!6
%(0
"'
(12 '
.-

'"
)%
-(
7

%%
!2,"(%7
=;<;
'5===9<:6>9<<:6=8C7

B=7
+$'
 4
"3+$
4
 . !%"'
7
--+'"'
-!
&+2('"
$"'26
 
," '%"'
&"-,
-!
" +'-"-"('
(
-!
)+(')!+"
-..%,
'
.-
"'
'().,
%/",7
/
2'7
=;;C
'5=>B9<:6<>=8??7

B>7
'
 4
+''+8''-!+&
4
. !'
4
+3%"' +
7
'- ('",&
(
 ?
," '%"'
",+.)-,
,&((-!
&.,%
"'/,-&'-
(
-!
.+-+
'
.+-+()%/"
#.'-"('7

+(%7
=;;D
'5

<>A
@=5

2
''&
 2
''&
 2
!
2

 2
!"%+#"
2
,
$5
 ,+
'
'&
%'*(!' &,"
(*',"&6@
76@8
'&
(*"%'*"$
'$$"-$*
*'/,!
"&
,!
%'-+
'.*05
'$
(*'
.5
;99=
,3?B7;84:>B6?<5

@>5
!"%+#"
2
!'/
 2
"
2
"%
2
%-*
2
&'
2
,
$5

-&,"'&$
'&
%'*(!' &,"
(*',"&
+0+,%
"&
,!
'.*05
*'
,$

"


5
:BBB
-&
;;3B?7:<84@;A;6@5

@?5

2
,+-#
2
''*
 2
!"%+#"
5
 ,
'
'&
%'*(!' &,"
(*',"&6@
'&
'$$"-$' &+"+
&
'.-$,"'&
"&
,!
*,5
"'$
(*'5
;99:
,3?>7=84BB=6B5

@@5
&" !,
2
$"+,*
5
6,
+-(* %"$0
%%*+
&
'.*"&
'$$"$
.$'(%&,5
(*'-,"'&5
;99?
- 3:<;7;84:B:6;9?5

@A5
'&
2
$*,"&"
2
"+!"%'*"
2
-%*
2
-
2
,1-#
5
*'/,!
" *&,","'&
,'*6B
"+
*)-"*
-*"&
*$0
'.*"&
'$$"-$' &+"+5
,-*5
:BB?
,
:93<:6>5

@B5
",,
2

2
0+!"
2
+-!
 5
&
.".'
,*,%&,
/",!
6B
+,"%-$,+
(*"%'*"$
&
(*"%*0
'$$"$
(*' *++"'&
&
,!
$$
%*#*
:@
"&
'.*"+
'
"%%,-*
*,+5
&'*"&'$' 05
;999
,3:=:7:984
&
*'/,!
" *&,","'&
,'*
B
"&
'.*"&
-&,"'&5
'$
&'*"&'$5
;99:
-&3:>7?84A>=6??5

A:5
,+-#
2
!"%+#"
5

& ,".
#
+0+,%
,/&
''0,
'&
%'*(!' &,"
(*',"&
:>
&
*&-$'+
$$
#",
$" &4
",+
*'$
"&
* -$,"&
*&-$'+
$$
%",'+"+5
*'
,$

"


5
;99;
-&
::3BB7:;84A9?96>5

A;5
$."&
2
&
2
&
2
"+!"%'*"
2
,1-#
5
'$-$*
!*,*"1,"'&
'
,!
'$$"$
 ,+
"&
,!
*'/,!
" *&,","'&
,'*
B6 ""&,
'.*05
'$
&'*"&'$5
:BBB
-&3:<7?84:9:A6<=5

A<5
-++"&
2
*'"$&
2
%,'
2
!'%(+'&
2
"$!*"+,
5
'0,+
(*.&,
-%-$-+
$$
('(,'+"+
0
%"&,"&"&

%'*(!' &"
(**"&
*"&,
'
'&
%'*(!' &,"
(*',"&+5

$$
"5
;99>
'.
:>3::A7,
;;84>;>@6?A5

A=5
"!'$$+
 2
**"+'&
2
"$!*"+,
2
*&/'*,!
2
,&,'&
5
*'/,!
" *&,","'&
,'*
B
"+

*%
$$
* -$,'*
'
*,'$"
$$
-&,"'&5
&'*"&'$' 05
;99B
03:>97>84;=A:6B95

:<@
DA6
&
3
! '$$+
 3
,&,'&
3
**!+'&
3
**"
3
!$ *!+,
3
,
$6
0,*7'.*!&
0(*++!'&
&
,!.!,1
'
*'/,
!*&,!,!'&
,'*
E6

&'*!&'$6
><<>9?:5@=E7?<6

DB6
&&,,
6
.$'(%&,$
&$1+!+
'

%-,,!'&
/!,
($!',*'(!
,+
!&
, 
%'-+6
'-*&$
'
'*( '$' 16
=EAB4ED9>:5=EE7>??6

DC6
 #$!
%
3
-+'&
3
?*3
$$
6
&$1+!+
'
1('%'*( !
!,$$
%-,&,+
+- +,+
!*&,
*)-!*%&,+
'*

!&
(*'$!*,!'&
&
%! *,!'&
'
%'-+
(*!%'*!$
*%
$$+6
!'$
(*'6
><<5?AC7@AE6

DE6
- *
3
 *&
3
*,$1
3
*$!&
6
*'$!*,!'&
&
%! *,!'&
'
(*!%'*!$
*%
$$+
!&
8
%'-+
%*1'+6
.
1&6
=EE?
'.4=ED9?:5=D>7E6

E<6
&
3
''&
3
'&
3
*'!&
3
'$$'!
6
*!+
(*!%'*!$
*%
$$
-$,-*+
*!.
*'%
.+55
,*&+ &!
%*1'+6
,%
$$+
.6
><<B4?9=:5=C6

E>6
+ ,
3
1%&
3
!$$!%+
3
&*+'&
3
&#!&+
3
'($&
3
,
$6
%*1'&!

0(*++!'&
(,,*&+
'
, 
7#!,
*(,'*
&
!,+
' &,
$! &
+- +,
%-$,!($
-&,!'&$
*'$+
!&
%'-+
.$'(%&,6

6
=EE=
(4=<9E:5>@>A7?A6

E?6
 *$7 $$*
3
+,'&
6
$,*
$$7+-*
,* ,!&
'
+,%
$$
,'*
-++
$'++
'
%$&'1,
(*-*+'*+
!&
,$=C
%-,&,
%!6
.
!'$6
=EEE
-&
=4>=<9=:5C=7DB6

E@6
'.$&
3
 $,,
6
,%
$$
,'*
&
7#!,
!&
, 
%%%$!&
,+,!+5
$++'&+
'*! !&,!&
*'%
', *
,-*;+
&
#&'#'-,+6

&'*!&'$6
=EEC
-&4=A?9?:5??C7@@6

EA6
**7*,* *
3
.!.!
3
!%%*
3
!.'$
3
*&
3
'&!
6
.$'(%&,$
0(*++!'&
'
7#!,3

(*','7'&' &
&'
1
, 

$'-+6
.$'(%&,6
=EE<
- 4=?6

EB6
* -
3
!+,*!
 3
 - $!&
3
'%&
3
*&
3
&!+
3
,
$6
0(*++!'&
'
7 !,
*(,'*
%
&
(*',!&
!&
, 
.$'(!& 3
-$,
&
!**!,
*'&,
,+,!+6
(*'-,!'&6
><

=?D
DB6
#*
 3
+%*
3
!.*'.
6
",
$" &
%",+
+-*.".$
'
,1(

+(*%,' '&"
&
".""&
+(*%,'1,+
"&
('+,&,$
%'-+
,+,+6
'$
(*'
.6
:5>;>7<;6

DC6
*2"$"
3
22*'
3
'**&,"&'
6
"&"&
'
*%
$$+
,'
%-,&,
$
*,'$"
$$+
"+
 ,".
&
"+
*+-
1
0(*++"'&
'
,!
,*&+%%*&
'*%
'
,!
7#",
$" &6
.
"'$6

14<@B9<:5C9@:5B?>7@B6

<;;6
!&$
3
%"*2
3
*,+&+,"&
3
 1
3
'%$"
6
,'("
0(*++"'&
'
",C>9<:5=D7?;6

<;<6
!',
3
,(!&+'&
3
!(%&
3
+%*
3
*&+,"&
6
!
(*','7 '&' &
7#",
&'"&

,*&+%%*&
,1*'+"&
#"&+
*(,'*
%(+
,'
,!
%'-+

$'-+6
,-*6
>@9A>?<7@<6

<;>6
'#
3
-#
3
."
3
+%*
6
&",
$" &
'*
,!
7#",
,*&+%%*&
#"&+
*(,'*5
3

"*'$+,
*".
*'/,!
,'*
+,"%-$,+
%+,
$$+
&
*1,!*'"
(*' &",'*+6

6
=CB7D?6

<;?6
+'
3
"$$"%+
3
"++$*
3
*'-1
3
*,"&
 3
,#"&+
3
,
$6
,%
$$
,'*
"+
&'
,
,!
$
$'-+
'
,!
%'-+
&
"+
,!
$" &
'*
,!
7#",
,1*'+"&
#"&+
*(,'*6
$$6
9<:5=<>7=?6

<;@6
-&
 3
'#
3
-#
3
+%*
6
" *&,"$
0(*++"'&
&
(*'++"&
'
,/'
$$
++'",
'*%+
'
,!
#",7$" &5
7<
&
7=6
'$
"'$
$$6
9>:5>?D7A=6

<;A6
$& &
3
!&
3
*
6
*&+%%*&
'*%
'
,!
#",
$" &
*'/,!
,'*
"+
,*%"&
1
$,*&,".
+($""&
&
"+
%"++"&
"&
,!
$
%-,&,6
$$6
@6

<;B6
'& $1
 3
1**$$
3

3
"$$"%+
3
$&
3
& $1
3
,
$6
!1%+
$. 
'
+,%
$$
,'*
1"$+

"',".3
+'$-$
(*'-,6
*'
,$

"


6
<4<;D9@:5A=@7>B6

<>D
=E5EA9><:6==EE5?A>9B??D:6D7
("'
4

4
(($
4
,(
4
1-+
4
2%"
7
 -,
(
-!
,-%
'
)+(.-
('
&(.,
)+"&(+"%
+&
%%,
"'
.%-.+7
-.+7
=EE=
.
>E5?A>9B??D:6D
)+(&(-,
&.+"'
((2-
+(0-!
"'
/"-+(7
"(%
)+(7
><6=>8>=7

==A7
%!&'
4
/
4
+
4
",',-"'
4
$,
4
( %
4
-
%7
!
(.+-!
"&&.'( %(.%"'
(&"'
(
-!
,-&
%%
-(+
+)-(+
(.)%,
%" '
"'"'
-(
," '%
-+',.-"('7
%%7
=EEA
'
=?5D<9=:6=
.
A5>BC9>>:6=AEC<8C7

==C7
"-!
4
%%",
4
2&'
4
'+,('
4
"%%"&,
4
+',-"'
4
-
%7
" '%
-+',.-"('
2
'(+&%
",( (+&,
'

&.-'-
/+"'-,
(
-!
"-
+)-(+
-2+(,"'
$"',7

7
=EE=
)5=<9E:6>@A=8E7

==D7
(,$(,$"
4
+7
" '%"'
2
"-
)+(-"'8-2+(,"'
$"',88-!
,-&
%%
-(+
+)-(+7
"(!&
"()!2,
,
(&&.'7
><<<<

=5EB9=>:6?C?@8@>7

=><7
; . !%"'8.''+
4
(,/"
4
2%(+
 4
!"/$+.)4
++2
4
-% 
4
-
%7
2'
",
+*."+
(+
'(+&%
,-&
%%
-(+8"'.
)+(%" +-"('
'
!&(-1",
(
)+"&+2
!&-()("-"
%%,7
%((7
><<=
.%
=A5ED9>:6?@?8A<7

=@<
=>=6
%1/* &
3
(-*
3
*!
6
&.'$.%&,
'
1&
#!&+
!&
!,
&
!&, *!&7%!,

,!.,!'&3
1,'+#$,$
*'* &!2,!'&3
&
 %',0!+
'
%+,
$$+6
$''6
><>6
 !.#*-(
3
!&&#!&
6
1&
'&,*!-,+
,'
* -$,!'&
'
%-$,!($
!,7 (&&,
+! &$!&
(, /1+
!&
%-*!&
'&
%**'/
%+,
$$+6
$$
! &$6
><?6
'$!
3
$$ *!&!
3
!
 '+,!&'
3
*%!
3
'++!
6
! &$!&
, *'-
0,*$$-$*
+! &$7* -$,
#!&+
!+
*)-!*
'*
+(*%,' '&!$
(*'$!*,!.
*+('&+
,'
+,%
$$
,'*6

!'$
 %6
><<=
,
>B4>CB9@?:5@<>>A7??6

=>@6
$$
3
$.$&
3
;-$$!.&
3
'($&
3
&#!&+
6
$,!'&
&
!&,*$$$!
'%($%&,,!'&
&$1+!+
'
,$
%-,&,
%!6
&,!+6
=EEB
*4=@>9?:5E?A7@@6

=>A6
&$7,&2$
3
'%(*,+
3
&*+'&
3
+%&
3
1$!
6
 
*'$
'
 *!&+
-*!&
(*!%'*!$
*%
$$
%! *,!'&
&
*$1
'&
'*%,!'&
!&
, 
%'-+6

.6
><<<
*
=4E=9=7>:5=@?7A>6

=>B6
*!&!
3
$**!
 3
'&
3

$
3

$!!
6
*'/,
,'*+
+-+,!&
(*!%'*!$
*%
$$
+-*.!.$3
(*'$!*,!'&
&
&,*!&
!&,'
%!'+!+
!&
, 
+&
'
+'%,!
$$+6
.
!'$6
><DA9=:5@E7AB6

=>C6
 !
3
!
3
!+ !#/
3
1+ !
3
%2#!
3
+- !
3
,
$6
++!,1
'
0,*$$-$*
'%!&
'

97#!,:
*(,'*+
'*
,, %&,
'
%-*!&
-$,-*
%+,
$$+
,'
!*'$+,+6
$''6
=EE>

=4CE9?:5BA<7B6

=>D6
+
3
!
*$'
3

$!!
6
 
7#!,
*(,'*
!+
!&.'$.
!&
, 
 +!'&
'
%'-+
(*!%'*!$
*%
$$+
,'
+'%,!
$$+
!&
-$,-*6

.6
=EEC
'.4BD9=7>:5?C7@@6

=>E6
,&
!"#
3
'* 1&+#1!
3
$!%*+',,!*
3
'-%&+
 6
'&,*'$$!&
$$
,
1
'&
%'*( ' &,!
(*',!&
*(,'*+6
'$
$$
&'*!&'$6
><==9=7>:5=<
,
?<4C=9?:5?EE7@=<6

=@=
>@?8
/)
5
)!'((
5
,)(% ,-
 5
)#)%$
5
, & 3
5
,- (.3
8
9D
$-
(
$( / ,
)!
( *#,)" ( -$-5
(
$-
&-)
, +/$,
!),
3
0 &)*' (.
(
-% & .&
*.. ,($("8
 ( -
 08
>FFB
)0
>B6F;??<7?E=E9?=8

>@@8
.),'
5
/3(#
5
)* &(
5
(%$(-
5
$("-& 3
5

 8
$'
&. ,.$)(-
$(
,#3*) $-'
'$
/
.)
'/..$)(-
$(

( 1
' ' ,
)!
.#

 .9-/* ,!'$&38
./, 8
>FFA
*,
>A6@CE;CAD?<7C@F9A@8

>@A8
/ & 3
5
3)(-
5
) ,.-)(
 8

, +/$, ' (.
!),
)(
'),*#)" ( .$
*,). $(9D
/,$("
0 &)*' (.
)!
.#
'''&$(
%$ ( 3
(
3 8
 ( -
 08
>FFB
)0
>B6F;??<7?DFB9E=D8

>@B8
 **/
5
1.
5
'').)
5
#3.$&
5
$()1
5
) 
5
.
&8

.3*

,  *.),
$-
, +/$,
!),
"-.,/&.$)(
(
,&3
0 &)*' (.
)!
')/-
',3)-8
 0
$)&8
?===
3
>6??>;><7?AF9BE8

>@C8
$-#$(
5
,)'$
5
, & 3
5
 #,$(" ,
8
/&.$*&
,)& -
!),
.$0$(9&$%
%$(- 9?
-$"(&$("
/,$("
')/-
',3)" ( -$-8
 0
$)&8
>FFF
 *
>B6?>@;?<7@>A9 ?C8

>@D8
$-#$(
5
/4/%$
5
 ()
5
 #,$(" ,
8
'*,
() -

.3*

)(
'),*#)" ( .$
*,). $(
,  *.),
.#.
$-
-- (.$&
!),
"-.,/&.$)(
/,$("
')/-
',3)" ( -$-8
 ( -
 08
>FFB
 
>B6F;?A<7@=?D9@D8

>@E8
$,,
5

&
)'*
5
&$
5
.$
5
$,.-)-
5
# /("
5
.
&8
#
./'),
-/**, --),
" (
' A:*A
$-
, +/$,
!),
"-.,/&.$)(
(
&. ,
!),
(. ,$),
0 &)*' (.
)!
.#
')/-
',3)8
 ( -
 08
>FFE
(
>6>?;><7>=D9>F8

>@F8
#)
8
)(- +/ ( -
)!
%()%$("
)/.

-$"(&$("
$(
.#
')/- 8
 ( -$-8
?==@
(6@B;><7A@9BC8

>A=8

)/-
)* -
5
) & (
5
)(. $,)
5
'' (-
5
$(
5
$
5
.
&8

-$"(&$("
' $.
3
 ?
$(
.#
0$- ,&
( ) ,'
$-
(  --,3
!),
.#
" ( ,.$)(
)!
*,$'), $&
" ,'
 &&-
$(
.#
')/-
',3)8
 ( -
 08
?==A
/"
>6>E;>B<7>E@E9AF8

>A>8
#$'-%$
5
)),
 5
.-/%
5
,$%-)(
8
#
)(
'),*#)" ( .$
*,). $(
-3-. '
$(
'''&$(
, *,) /.$)(8
( ),
 08
?==A
 6?B;><7D?9>=>8

>A?8
/
5
# (
5
("
5
$
5
$
5
)
5
.
&8
 0 &)*' (.&
2*, --$)(
(
!/(.$)(
)!
'*A
$(
-* ,'.)" ( -$-
(
$(
'$(.$($("
*$ $ 3'&
$(. ",$.38
 0
$)&8
?==A
 
>6?DC;><7>BE9D>8

>A@8
#)
5
 ("
5
)-%3
5
$1
5
)"(
 8
#
" (
() $("
)(
'),*#)" ( .$
*,). $(
E
$-
, +/$,
!),
.#
$($.$.$)(
(
'$(. ((
)!
-* ,'.)" ( -$-
$(
.#
')/- 8
 ( -
 08
>FFC
/&
>6>=;>@<7>CBD9CF8

>A?
?BB9
#*
6
$2
6
*")
 9
*)
(*-+#*" ) /$
+-*/ $)
F
+'4.

-*'
$)
/#
($)/ ))
*!
.+ -(/*" ) .$.
)
/#
$)/ "-$/4
*!
/#
+$ $ 4($.9
 1 '*+( )/9
?GGF
-7?@CA:?@9

?BC9
0
6
 4)*' .
6
*)"
6
$
6
 $% )
6
0
6
/
'9
#
/4+

. -$) ;/#- *)$)
&$).
-  +/*-
/ 
< @=
$.
- ,0$-
!*-
"./-0'/$*)
*!
/#
(*0.
(-4*9
 1 '*+( )/9
?GGG
0)7?@D>?
 +7?@FG:@?9

?BF9
)"
6
$
6
0
6
 )"
9
#
/0(*-
.0++- ..*-
B;B
$.
.. )/$'
!*-
+$'./
+-*'$! -/$*)
)
( .* -(
$) 0/$*)
$)
($ 9
-*
/'

$


9
?GGF
-
A?7GC>B
-7B<@=8?DE:F?9

?C>9
0%$2-
6
 #-/
6
0'$&
6
*")
 9
$./$)/
- ,0$- ( )/.
!*-
3/-: (-4*)$
)
(-4*)$
*)
(*-+#*" ) /$
+-*/ $)
B
$)
/#
!*-(/$*)
*!
/#
)*
)
+-$($/$1
./- &
)
**- $)/$*)
*!
' !/:-$"#/
.4(( /-4
$)
/#
(*0. 9
 1 '*+( )/9
@>>@
/7?@G<@>=8BDFC:GD9

?C?9
*''*24
 6
* -/.*)
 9
-'4
(-4*)$
' /#'$/4
$)
(+C7(+E
*0'
(0/)/
($
.0"" ./.
!0)/$*)'
- 0) )4
2$/#$)
/#
D>
.0"-*0+9
 1 '*+( )/9
?GGG
+-7?@D>@7FEC:?@9

?CA9
) -.*)
6
..' -
6
 *-" .: *0 ..
6
4) .
6
 -
 6
- $  -"
6
/
'9
*0.
+-$(*- $'
" -(
 ''.
'&$)"
 /?
$)/ "-$).
)/ -
/#
" -('$)
0/
!$'
/*
($"-/
)*-(''4
/*
/#
"*) .9
 1 '*+( )/9
?GGG
+-7?@D>A
0'7?A>FC: GB9

?BA
>BB8
#34%#
5
-"#'
5
)!)
5
)!(
 5
"#%1
8
)(
'),*")! ( .#
*,). #(
A
, !/&. -
."
/#(!
-#.
(
&)(!.#)(
)
."
')/-
/, . ,8

&#(
(0 -.8
?===
*,6>=B;D<7EC@9D@8

>BC8
"(!
5
. (
#$%
5
/
 8

*."13-
,
#(0)&0 
#(
).#
*-/&

),'.#)(
(
*#." &#&9' - ("3'&
-#!(&#(!
#(
."
 0 &)*#(!
"#% (
#(( ,
,8
 0
#)&8
?==?
)0
>B6?B>;?<7@E=9FA8

>BD8
&#',-)..#,
5
)/'(-
 5
)- ("&
5
,/!'(
5
.)"
5
,#(
5
.
&8
.#'/&.#)(
)
>
2*, --#)(
3
)(
'),*")! ( .#
*,). #(
#-
-/ ## (.
(
(  --,3
),
)(
'),*")! ( .#
*,). #(9#(/ 
.#0.#)(
)
()." &#&
 &&-8
#,/&.#)(8
?==?
.
??6>=C;>D<7??C@9D=8

>BE8
 ,.4
5
,,'((
8
,! 9-&
-, (
),
! ( -
#(0)&0 
#(
!)(
 0 &)*' (.8
 "
 08
?===
)06FE;>9?<7B>9D=8

>BF8
#(!
 5
#")&-
5
"' ,-
5
'#."
8

#(/.#)(
)

*,). #(-
-/**, -- -
# , (.#.#)(
(
-/-.#(-
',3)(#
-. '
 &&
- & 9, ( 1&
#(
)&&),.#)(
1#."
@8
 &&8
?==@
.
@>6>>B;@<7?E>9F?8

>C=8
"(!
5
.4/%
8
'B
#-
, +/#, 
),
')/-
*,#'),#&
! ,'
 &&
 0 &)*' (.8
 "
 08
?==>
/(6>=A;>9?<7C>9D8

>C>8
"/0

)/-
)* -
5
0(
 (
,# -"
5
,0&")
 5
,--)(
5
!! (
5
/,(#
5
.
&8
&. , 
*,#'),#&
! ,'
 &&
'#!,.#)(
#(
."
- (
)
.,(- ),'#(!
!,)1."
.),
 .
-#!(&#(!
0#
 B8
 0
#)&8
?==B
/!
>6?EA;><7>FA9?=@8

>C?8
% /"#
5
)&3( /2
5
/(3(
5
"#&
5
3&#
8
"
,)& -
)

-#!(&#(!
#(
! ,'
 &&
'#!,.#)(
#(
."
')/- 8
 0 &)*' (.8
?==B
 6>@?;?A<7B@FF9A=F8

>C@8
3&#
8
 ,'
 &&-8
 &&8
>FFF
(
??6FC;?<7>CB9DA8

>CA8
 -' ,
5
()0
5
/..&#(! ,
5
/(!
 5
% ,
5
3--& ,
5
.
&8
"
%#.9&#!(
;-. &
.),<
(
#.-
,  *.),
9%#.:7
*& #).,)*#
,)& -
#(
!' .)! ( -#-
(
' &()! ( -#-8
 0
/**&8
>FF@7>?B9@D8

>CB8
#,#&&3
5
*(
5
 ,,#')(
5
! ..
5
#
8

-#!(&#(!
#-
, +/#, 
),
)(.,)&&#(!
-)'.#
-. '
 &&
- & 9, ( 1&
#(
."
,)-)*"#&
)0,38
 0
 &&8
?==B
)06F;B<7CB>9C?8

>CC8
.-/%
5
)
5

5
'').)
5
,#%-)(
5
"#'-%#
8
)(
'),*")! ( .#
*,). #(9>B8
 (.# #.#)(
)
.,! .
 &&-
(
#)&)!#&
/(.#)(-8


#)&
" '8
?===
 
>B6?DB;B=<7@FB?@9E8

>AA
=BC7
)2
4
&+%
4
 ()&
 4
**#!
7
)'*+#,)(
)
+ )'#((-
!+)0-"
# + (-#-#)(
-)+8E
(
))2-
+ !.&-#)(
)

&#!(
' ,, (! +
+#)(.& #
#
1*+ ,,#)(
#(
').,
)/+#(
)&&#& ,7
#)&
 *+)7
><<<
 5B?:B;6=BBE8CA7

=BD7
"-#
4
)(( -
4
.
4
.+)"
4
+(
4
&&" +
4
-
&7
)(
')+*")! ( -#
*+)- #(,
+ !.&-
-"
 / &)*' (-&
*+)!+'
)
".'(
" '-)*)# -#
,- '
 &&,7

1*
 7
=EEE
*+
A5=DE:C;6==?E8@D7

=BE7
)()/(
 4

#!. &
7
.+(#(!
! +'
 &&,
#(-)
,- '
 &&,7
.++
*#(
 ( -
 /7
><<<@5@D:A8B;6A?C8@@7

=C=7
 "+& 8 && +
4
 ,-)(
7
)&.&
(
 &&8).(
)+',
)
,- &
-)+
-#/#-2
*&2
#,-#(-
+)& ,
#(
' &()2-
*+ .+,)+
#,* +,&
(
,.+/#/&
)(
-"
&- +&
( .+&
+ ,-
'#!+-#)(
*-"027
 / &)*' (-7
=EEA
+5=>=:?;6C?=8@>7

=C>7
- & +
4
*# & +
4
&(" /
4
)&2( .1
4
#"- +
4
)$)+.
4
-
&7
+#')+#&
! +'
 &&
'#!+-#)(
#(
-"
"#%
(
').,
'+2)6
-"
+)&
)
-"
" ')%#(
8=9 =>7
 /
#)&7
><<@
.!
=A5>C>:>;6?A=8B=7

=C?7
+
4
%'.+
4
!0
4
.!#2'
4

4
#,"#')-)
4
-
&7
'*#+ 
)&)(#3-#)(
)
-"
!)(,
2
*+#')+#&
! +'
 &&,
#(
'#
&%#(!

" ')%#( 4
,-+)'&
 &&8 +#/ 
-)+8=
:8=;7
+)
-&

#


7
>?7

=C@7
.& 2
4
.(2(
4
% ."#
4
"#&
4
)&2( .1
7

,#!(&#(!
+ !.&- ,

(.' +,
(
')-#&#-2
#(
)+!(
.&-.+ 7
 "
 /7
><@:=;6BD8CC7

=CA7
)+#()
7
 ( + 
&
1*+ ,,#)(
0#-"
-"
>B
+
+ *)+- +
,-+#(7
-
 ( -7
=EEE
(5>=:=;6C<8=7

=CB7
2,"#
4
")(
7
 ## (-
+ )'#(-#)(
#(
#/ +,
-#,,. ,
2

-')1# (8#(.#&
)+'
)
+ 6

-))&
)+
- '*)+&&2
+ !.&- 
! (
-#/-#)(9#(-#/-#)(
#(
-"
')., 7
 /
#)&7
><<>
*+
=A5>@@:>;6?8+
'#
2
')# #-#)(
)

1>
- +#&
+-# ##&
"+)'),)' 7
 ( ,#,7
><<@
*+5?D:@;6=EA8E7

=CD7
)!(
7
(#*.&-#(!
-"
').,
'+2)
6

&)+-)+2
'(.&7
>(
7
&#(/# 04
776
)&
*+#(!
+)+
)+-)+2
+ ,,5
=EE@7

=CE7
.+.-
4
)!(
 7
@
#,
,, (-#&
)+
& (,
#(.-#)(
#(
-"
').,
'+2)7
 ( ,
 /7
=EED
 
=5=>:>?;6?CB@8CA7

=@A
=D<7
#(
4
.% ,,
4
(&)$#',
4
"(.
4
--,
4
%0#'
4
-
%7
'
,, '-#%
+(%
(
&)@
#'
-"
-+#(/ '-+#.%+
, )--#('
(
-"
&(.,
" +-7
 ' ,
 /7
>8C7

=D=7
(!'
 4
% ,,#'!
4
#''# +
4
.3.$#
4
(' ,
7
+(0-"
-(+,
#'
 / %()& '-6
-"
+(%
(
8 -
+ %- 
)(%2) )-#
,#!'%%#'!
&(% .% ,
#'
&+2(! ' ,#,7
 /
.))%7
=EE@6A?8B<7

=D>7
$#'
4
 / +,
4
.+(
4
2(',
4

( +-#,
7
#+ '(& %#
#'
&)C
'
,!
(&)(.'
&.-'-
&# 6
+ *.#+ & '-
(+
&)
,#!'%#'!
#'
-"
 / %()& '-
(
/ '-+%
)(,- +#(+
& ,( +&7
 / %()& '-7
><:=<;6>@DE8EE7

=D?7
.% 2
4
( +-,('
 7
/ +%))#'!
1)+ ,,#('
(&#',
(
('
&(+)"(! ' -#
)+(- #'
&#%2
& & +,
)(- '-#%%2
(.'-
(+
%#&#- 
-#,,.
 -,
#'
C
 ## '-
&+2(,7
 /
2'7
=EEC
+5>7

=D@7
(,,
4
.'! +
4
) %
7
&)C
+ !.%- ,
! +&
 %%
)+(%# +-#('
#'
&(.,
-%
!(',7
 1
 /7
><;6=>C8?C7

=DA7
"+#,-( %,
4
((!+,
4
 ,,+#
4
%(.-
4
((+&'
4
&)#('
7
8(1
-+',+#)-#('
-(+
1>
+ )+ ,, ,
# + '-#-#('
'
(+&-#('
(
-"
+#
"& +,7
 /
2'7
><<@
)+5>>E:@;6CB?8C<7

=DB7
(('
4
(.% -
4
) "#
7
(+&%
%#&
 / %()& '-
#'
('#-#('%
&.-'-,
(
! @7
 / %()& '-7
><<<
+5=>C:A;6EDE8EB7

=DC7
#,"#'
4
'$,
4
#.+
4
%%*.#,-
4
 "+#'! +
7
 ' +-#('
(
&)+9%$?
('#-#('%
$'($(.-
&# 7
 ' ,#,7
><<>
 5?>:>;6BE8C>7

=DD7
 #
4
#'!
4

4
#.
4
"'!
4
'!
4
-
%7
+#&(+#%
! +&
 %%
,) # #-#('
+(&
&+2('#
,- &
 %%,7
 (
' 7
><;6 @<=?7

=DE7
('- #+(
4
 --(
7
%.(,
+ )+ ,,#('
(
=
1)+ ,,#('
#,
& #- 
2
(+=)
'
,>)
-"+(.!"
#'"##-#('
(
,'>9@)
#'
"+(&2 ,
 + /#,# 7

#+(#(%
--7
><<@
 
=A5>@=:>;6>>=8D7

=E<7
+ # +!
4
+#(%
4
(+#'!
4
 
4
 %% -# +
4
(.,&'
4
-
%7
8=
#,
+ *.#+ 
(+
+%2
$#' 2
 / %()& '-7
 %%7
=EE?
.!
>C5C@:@;6BCE8E=7

=E=7
.(
4
$ 
4
+$ +
7

 %%8,) # #
'.% +
+  )-(+
#,
,, '-#%
(+
+ '%
'
!('%
 / %()& '-
'
, 1.%
# + '-#-#('7
 %%7
=EE@
2

><5CC:@;6@D=8E<7

=@B
:B;4
!("
1
)!%&
1
))!
1
& #!*!
1
 &
1
 &
1
*
#4
 

&$&&.
%
.B
!)
))%*!#
&(
$&+)
&%
&($*!&%4
*+(4
;999

;=2=9<7?@@;83B9B5:<4

:B<4
+#/
1
#+$&
1
#'"
1
&#/%+.
4

)! %#!%
&%*(&#)
&($*!&%
&

'(!$&(!#
($
##
%! 
-!* !%
* 
(#/
%!*#
(! )4
,
!
;9:9
'(
;;4

:B=4
 -
1
*)+0"!
1
!) !"-
1
/) !
1
+%!)
1
+&
1
*
#4
.'())!&%
%
+%*!&%
&
5"!*
!%
$&'&!*!
'(& %!*&(
##)4

.'
4
:BB:
+#
:2:@=7:83?<5@:4

:B>4
+ !+(
1
+
1
!0
 1
% )
1
 ($
1
! #)-&(*
1
*
#4
&/*5 (!,
:>
%
)
&&'(*
*&
'(&$&*
#/&#/)!)
!%
+$+#+)
##)4
,#&'$%*4
;99@
+#2:<=7:=83;>B<5?9<4

:B?4
!/&) !
1
*)+"
1
$) !*
1
% "!
1
"$+(
1
)+"$&*&
1
*
#4
+%*!&%#
(#*!&%) !'
*-%
!(&#)*
(&-*
*&(5A
%
&%
$&(' & %*!
'(&*!%)
!%
( +#*!%
)*(&!& %)!)
/
(*
(%+#&)
##)4
&#
##
%&(!%
;9:9
'(
;B4

:B@4
!% $
1
($5 +(,(!
1
(
1
("(
4
,#&'$%*
&

)*(&!& %!
*&(
:6(
*(%) %!
$&+)
#!%4
%)!)4
;99?
'2==7B83=:B5;=4

:=@