IN VITRO DEVELOPMENT OF DISSOCIATED
AND IMMUNOMAGNETICALLY-PURIFIED
EMBRYONIC CHICK OPTIC TECTUM CELLS
By
DENI SCOTT GALILEO
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1988

Copyright 1988
by
Deni Scott Galileo

ACKNOWLEDGEMENTS
I would like to thank Dr. Paul Linser for both
direction and yet ample latitude to explore and satisfy my
own intellectual curiosities. I would also like to thank my
committee comprised of Drs. Francis Davis, Carl Feldherr,
and Chris West for guidance and constructive criticisms
throughout my dissertation research. A special thanks goes
to the University of Florida Division of Pediatric
Hematology and Oncology and especially Dr. Adrian Gee for
making this research possible through his commitment of
continuous gifts of microspheres, antibodies, and
encouragement to me. Also, I thank Dr. John Ugelstad of the
University of Trondheim, Norway, for a generous gift of
microspheres.
I would also like to thank Drs. Gudrun Bennett and
Gerry Shaw of the University of Florida as well as Dr.
Steve Pfeiffer of the University of Connecticut for gifts
of antibodies without which most of this work could not
have been done. Thanks are also in order to Drs. Paul
Begovac and Steve Dworetzsky for many worthwhile
discussions and sharing the common bond of being a usually
helpless guinea pig in the experiment of education that
runs continuously and often without controls in the big
orange and blue apparatus. I whole-heartedly thank anyone
that I have left out that has made my time in graduate
school either intellectually rewarding, pleasurable, or
iii

both. Alternatively, I wish to castigate those that have
made the learning processes of myself and others at the
University of Florida needlessly entangled in red tape,
rules, regulations, grades, prejudices, position, rigidity,
and everything else that runs counter to true learning.
Finally, I wish to thank Dr. Arlene Stecenko for
friendship and encouragement for the past 3 years and for
the ability to see that my quandaries could often easily be
resolved.
xv

TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS iii
LIST OF TABLES vi
LIST OF FIGURES vii
ABSTRACT viii
CHAPTERS
IINTRODUCTION (A NATURAL HISTORY) 1
IISEPARATION OF DAY 7-8 CELLS 4
Introduction 4
Materials and Methods 7
Results 19
Discussion 48
IIISEPARATION OF DAY 12-13 CELLS 69
Introduction 69
Materials and Methods 7 2
Results 76
Discussion 99
IVCONCLUSIONS 115
REFERENCES 118
BIOGRAPHICAL SKETCH 124
v

LIST OF TABLES
Table Page
2-1 Effects of substrata on monolayers 39
2-2 3H-thymidine incorporation: Day 7-8 cells 49
3-1 3H-thymidine incorporation: Day 12-13 cells.... 98
3-2 3H-thymidine incorporation: Monolayers 102
vi

LIST OF FIGURES
Figure Page
2-1 Chamber used for immunomagnetic separations... 13
2-2 Cell isolation in methylcellulose 21
2-3 A2B5 antigen modulation in vitro 25
2-4 In vitro development of A2B5( + ) cells 27
2-5 A2B5 antigen modulation via methylcellulose... 30
2-6 Calculated incomplete separations 33
2-7 Remixed separated cells 35
2-8 Effects of substrata: glass vs polyornithine.. 38
2-9 Neurofilament expression in vitro 43
2-10 GS expression in vitro 47
2-11 3H-thymidine incorporation into monolayers.... 51
2-12 Summary diagram of results 53
3-1 Markers for dissociated cells (graph) 78
3-2 Markers for dissociated cells (photos) 80
3-3 A2B5 antigen modulation in vitro 85
3-4 In vitro development of A2B5( + ) cells 89
3-5 Filament expression in vitro 93
3-6 Oligodendrocyte development in vitro 97
3-7 3H-thymidine incorporation into monolayers .... 101
3-8 Summary diagram of results 104
vii

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
IN VITRO DEVELOPMENT OF DISSOCIATED
AND IMMUNOMAGNETICALLY-PURIFIED
EMBRYONIC CHICK OPTIC TECTUM CELLS
By
Deni Scott Galileo
August 1988
Chairman: Paul J. Linser
Major Department: Anatomy and Cell Biology
Immunomagnetic cell separation techniques were used
to purify embryonic chick optic tectum cells from 2
different developmental ages for in vitro development
studies. This negative cell selection method was based on
reactivity of monoclonal antibody A2B5 with cell surfaces.
Purified A2B5(-) cells obtained initially were >99.99%
pure. Surprisingly, A2B5(+) cells rapidly appeared in the
purified surface A2B5(-) cells in direct response to the
immunomagnetic depletion. After 1 day in culture, levels of
A2B5(+) cells in purified cultures equalled unpurified
levels («50%). Similarly, visual densities of A2B5(+)
neurons were equal in purified and unpurified long-term
monolayer cultures.
vxii

Degeneration of purified cells on the neuron-
selective substratum polyornithine suggested that these
cells contained a paucity of neurons initially after
separation. Immunohistochemistry combined with 3H-thymidine
autoradiography of cells and monolayers demonstrated that
new DNA synthesis was required for neither the acquisition
of surface A2B5-antigen, nor for differentiation into
neurons. These results suggest that in early embryonic
vertebrate brain (days 7 and 8) cells are present which are
capable of replacing depleted neurons in vitro.
With day 12 and 13 cells, nearly all purified A2B5(-)
cells were identifiable as glia by reacting with antibodies
against either glutamine synthetase or galactocerebroside.
Most (=80%) of the purified A2B5(-) cells in culture for
one day became A2B5(+). No increase in the percentage of
A2B5(+) cells from 45% was observed in unpurified cultures.
Long-term monolayer cultures from purified cells contained
many A2B5(+) cells with a flattened glial or round
morphology. These A2B5(+) cells frequently reacted with
antibodies against glial fibrillary acidic protein and
another glial marker, 5A11, which indicated a partial glial
character. Additionally, flattened glial-like cells were
found to contain elaborate networks of anti-neurofilament-M
reactive filaments. The above combinations of markers were
not found in unpurified monolayers and are believed to be a
result of the immunomagnetic removal of neurons. It is
IX

hypothesized that the abnormal phenotypes in purified cell
cultures from day 12 and 13 cells represent unsuccessful
responses of the glia to replenish depleted neurons most
likely due to restricted developmental potentials.
x

CHAPTER I
INTRODUCTION (A NATURAL HISTORY)
The work presented in this dissertation is work that
began as an attempt to study in depth the phenomenon of
glutamine synthetase (GS) production in glia as mediated by
neuronal contact. This enzyme is produced in embryonic
chick retina cultures in glia that are in close apposition
to neurons (Linser and Moscona, 1979). Subsequently, it was
found that this phenomenon also occured in cultures of
embryonic chick optic tectum cells (Linser and Perkins,
1987a) and probably occurs in the vertebrate central
nervous system in general. To study this phenomenon
directly, I wished to obtain purified glia that were not
producing any GS to which I could add back neurons and
trigger the gene expression at will.
I chose to utilize cell purification methods that
employed antibodies as the means of discrimination between
neurons and glia (immunoselection). I became aware of the
immunomagnetic cell separation procedures that use small
paramagnetic microspheres coated with specific antibodies
to remove a "target" cell population from heterogenous
populations. Although this method has been used in
different forms in several different systems, it appears
that it has most seriously been utilized by clinicians to
1

2
remove neuroblastomas and leukemias from human bone marrow
with unparalled efficiency (Treleaven et al., 1984).
Fortunately, one of only two laboratories in the United
States that uses this method clinically is at the
University of Florida in the Department of Pediatrics.
Through the aid of a mutual friend (science is as human an
endeavor as is anything else) I enlisted the aid of Dr.
Adrian Gee, the scientist in charge of such bone marrow
"purging" at U.F. He and his entire laboratory remained
committed to me for supplying microspheres and antibody
with which to coat them from the conception to completion
of my research.
Since then much time was spent on developing a simple
and effective procedure and separation chamber that would
be suitable for the separation of embryonic brain cells. At
first, it seemed that my procedure was not accomplishing
separations since apparent target cells were always in what
was hoped to be purified populations of nontarget cells. It
then became quite clear that these target cells were
appearing from nontarget cells after the separations. This
in itself was a unique and unexpected finding. It was
subsequently found that when different age embryonic cells
were separated that the target cells which appeared in the
purified populations apparently developed different
phenotypes according to the age separated. Early embryonic
cells appeared to be able to compensate for the depletion

3
of target cells (neurons), whereas older embryonic cells
could not. Herein lies the natural division of my work
into the two following chapters according to the results
obtained: separation of day 7-8 cells (Chapter II) and
separation of day 12-13 cells (Chapter III). Although these
results precluded my ability to obtain purified populations
of immature glia as I originally had hoped for, many unique
and interesting experimental phenomena occured in the
purified cultures. These have led to a better (or more
confused, depending upon the point of view) understanding
of the potentials and restrictions of embryonic cells when
their development is perturbed in a controlled fashion.

CHAPTER II
SEPARATION OF DAY 7-8 CELLS
Introduction
During development of the chick optic tectum the
various differentiated cell types emerge in a temporally
stepwise but overlapping manner. The neurons are generally
the first cell type to exit the mitotic cycle (LaVail and
Cowan, 1971b) and to express a differentiation product such
as neurofilaments (Bennett and DiLullo, 1985). The glia are
generally later in becoming post-mitotic and show overt
signs of differentiation (Linser and Perkins, 1987a)
practically coincident with the completion of neurogenesis
(LaVail and Cowan, 1971b; Fujita, 1964). This timing of
overt differentiation, however, does not necessarily
reflect the timing of when different cells are determined
to become one cell type or another. Additionally, this
general pattern does not necessarily imply when a cell is
restricted in its ability to become anything else if its
microenvironment were to change.
The mechanisms that govern cell determination and
differentiation during brain development seem multiple but
are poorly understood. Interactions between cells appear to
influence development at several levels from physical
positioning of neurons (Levitt and Rakic, 1980) to the
4

5
expression of specific glial gene products (Fisher, 1984;
Linser and Moscona, 1979). Glutamine synthetase (GS), for
example, is produced in glia in culture when the glia are
in contact with neurons (Linser and Moscona, 1983; Linser
and Perkins, 1987a; Wu et al., 1988). Obviously, to study a
phenomenon such as this it would be of great advantage to
be able to purify the immature glia so that neurons could
be added back to elicit GS production. Such an ability
could also in itself reveal other phenomena that involve
cell interactions.
A major obstacle to studying interactions that take
place during early development, such as those that lead to
GS production, is that they occur when only few if any
cells are identifiable by commonly recognized
differentiation markers. Also early in development, most
cells do not differ enough from each other physically to
make use of such cell purification techniques as buoyant
density centrifugation (Campbell et al., 1977; Sheffield et
al., 1980). Methods that do not need overt physical
differences for separation are those that utilize
monoclonal antibodies that discriminate between the
surfaces of different types of cells (immunoselection).
Complement-mediated cell lysis has been used successfully
with embryonic neural tissues (Politi and Adler, 1987;
Nagata et al., 1986), but this method does not allow
recovery of both cell types, and not all antibodies fix

6
complement. "Panning" (Wysocki and Sato, 1978) is an
immunoselection method that allows for recovery of both
cell types; however, the purity of cells obtained is
marginally acceptable («95%) for most applications.
One negative separation method that utilizes
monoclonal antibody binding to a target cell population to
remove it from a mixed population is immunomagnetic
purging. This method was developed for removing
neuroblastoma cells from human bone marrow (Treleaven et
al., 1984) and operates by attaching the target cells to
paramagnetic polystyrene microspheres via antibody linkages
and removing them with a magnet. Potentially all types of
monoclonal antibodies against cell surface constituents can
be used, and routine separations result in depletions of
target cells by 4 to 5 orders of magnitude (Philip et al.,
1987 ) .
I have modified this method to work well with
dissociated embryonic chick brain cells. In hopes of
purifying immature glia for in vitro reassociation studies,
I removed the majority of identifiable neurons by using
the monoclonal antibody A2B5 (anti-ganglioside GQi<=;
Eisenbarth et al., 1979). It was found that even though the
initial purification of A2B5(-) cells was complete, by 24
hours in culture approximately 50% of these cells had
become A2B5(+). This modulation of cell surface A2B5
antigen was found to be in direct response to the

7
depletion of A2B5(+) cells. Similarly, visual densities of
A2B5(+) neurons and neurofilaments were equivalent in
purified and unpurified cultures. New DNA synthesis was not
required for either modulation of surface A2B5 antigen or
differentiation of cells into neurons.
Materials and Methods
Animals
White Leghorn chick embryos were used throughout this
study. Fertilized eggs were purchased from the Division of
Poultry Science, University of Florida and stored at 15°C
until initiation of incubation at 37.5°C in a standard
humidified egg incubator. Time in days of incubation was
used as the index of developmental age. For the present
study, 7 and 8 day embryonic optic tecta were dissected at
the tectal commissure and isolated free of non-neural
tissues aseptically in calcium-magnesium free Tyrode's
solution (CMF; Linser and Moscona, 1979).
Cell Culture
Dissociated cells were prepared by incubating tecta
for 30 min. in 0.4% trypsin (Nutritional Biochemicals,
Cleveland, OH) in CMF at 37°C, followed by dissociation
with a Pasteur pipette in Medium 199 (Hank's salts,
Degenstein formula; KC Biological, Lenexa, KS) containing
0.3 mg/ml soybean trypsin-inhibitor (Sigma Chemical Co.,
St. Louis, MO) and 0.03 mg/ml DNase I (Sigma) (SBTI-DNase).

8
Rotation-mediated suspension cultures of
reaggregating cells were made by placing 2xl07 cells in 3
ml of Medium 199 supplemented with 10% fetal bovine serum
(FBS; Gibco Laboratories, Grand Island, NY), 100U/ml
penicillin + 100 jig/ml streptomycin sulfate (Gibco), and 10
ng/ml gentamycin sulfate (M.A.Bioproducts, Walkersville,
MD). Cells were rotated in capped 25 ml Ehrlenmeyer flasks
at approximately 75 rpm (1 inch radius) at 37°C in a rotary
incubator (New Brunswick Scientific, Edison, NJ) and fed
approximately every other day with the same medium. Two
days previous to assaying cultures for glutamine
synthetase (GS), cultures were fed with the above medium
supplemented with 0.33 ng/ml hydrocortisone (Sigma). GS
levels were assayed by the modified colorimetric method of
Kirk as previously described (Linser and Moscona, 1979).
Adherent monolayer cultures were prepared by
incubating 106 cells in 1 ml of the above medium in 24 well
(200 mm2) tissue culture plates (Corning Glass Works,
Corning, NY). Cells were plated on either 1) the plastic
itself, 2) inserts of 12 mm dia. round glass coverslips, 3)
coverslips coated with 0.1 mg/ml poly-L-ornithine HBr
(m.w.= 100,000; Sigma), 4) coverslips coated with
polyornithine and then 1 mg/ml rat tail collagen (Sigma),
or 5) inserts of 7.5 mil thick Aclar fluorohalocarbon film
(Allied Chemical Corp., Morristown, NJ). Cultures were kept
in a standard tissue culture incubator at 37°C in a 5% CO2/

9
air atmosphere. Monolayer cultures were fed with fresh
medium approximately every other day.
For cell isolation experiments, freshly dissociated
cells were suspended in a semisolid medium of 1.4% methyl
cellulose (Sigma) dissolved in the above culture medium at
a density of 107 cells/ ml. Cultures were kept in loosely
capped sterile Ehrlenmeyer flasks in a standard tissue
culture incubator in a 5% C02 / air atmosphere. After
different lengths of time cells were recovered from the
semisolid medium by diluting the medium with several times
the volume of Tyrode's solution and were collected by
centrifugation.
Immunomaqnetic Separations
Polyclonal sheep anti-mouse antibodies (kindly
provided by Dr. Adrian P. Gee, Div. of Pediatric Hematology
and Oncology, Univ. of Florida) used to coat the
microspheres were prepared by hyperimmunization with
purified mouse IgG (Organon Teknika- Cappell, West Chester,
PA). Useful antiserum was obtained after 5-6 immunizations.
Anti-mouse antibodies were affinity-purified on a column
made from mouse IgG (Cappell) bound to Affi-Gel 10 (Bio-Rad
Laboratories, Richmond, CA) and then mixed with Sepharose
CL-4B (Pharmacia, Inc., Piscataway, NJ).
Cells used for immunomagnetic separation were rinsed
in Tyrode's solution, incubated in 1/25 dilution of A2B5-
conditioned hybridoma medium in Tyrode's with 10% heat

10
inactivated FBS (HI-FBS) for 30 min. at 4°C. Cells were
then rinsed 2x in Tyrode's and resuspended in SBTI-DNase.
The A2B5 hybridoma cell line was obtained from the
American Type Culture Collection, Rockville, MD, through
Dr. Michael F. Marusich, Univ. of Oregon, Eugene, OR.
Cells were mixed for 30 min. at 4°C in SBTI-DNAse + 10% HI-
FBS with paramagnetic polystyrene microspheres (4.5 (J.m;
Dynal, Inc., Great Neck, NY) which were previously coated
with sheep anti-mouse IgG prepared as above. A 15-fold
excess of the number of microspheres/ the number of A2B5(+)
target cells was used which corresponds to a 7.5-fold
excess of microspheres/ total cells, since approximately
50% of dissociated cells were A2B5(+). The microspheres
were ethanol sterilized then coated with 30-40 ug antibody/
mg microspheres in a concentration of at least 0.2 mg/ml
antibody overnight at 4°C with rotation in a microfuge
tube.
The cell-microsphere mix was then poured into the top
of the separation chamber which was prefilled with
Tyrode's, and unbound cells were eluted with Tyrode's by
gravity at a rate of approximately 1-2 ml/ min. in a
sterile hood until no more cells appeared in the eluant. In
a typical experiment, approximately 3xl07 cells were
separated in 30 min. The separation chamber was constructed
from a plastic funnel and tubing (1/4" dia.) held on a ring
stand (Fig.2-1). Two magnet arrays were held against the

11
side of the tubing, one made of ferrite magnets above one
made of samarium-cobalt magnets. A pinch clamp at the
bottom of the tubing was used to control the rate of flow
induced by gravity.
Bound microsphere fraction cells were collected by
first removing the magnet arrays from the side of the
tubing and then washing out the bound cells and
microspheres into a test tube. Cells were released from the
microspheres by trypsinization as above followed by
addition of SBTI-DNAse and vortexing. Freed microspheres
were drawn away from the cells by placing the samarium-
cobalt magnet array against the side of the tube and the
suspended cells were aspirated out of the tube with a
pipette.
Cells to be assessed for cell surface A2B5 following
separation (approximately 2 hrs. after plating) were fixed
in 1% formaldehyde (ACS grade, Fisher Scientific,
Pittsburgh, PA) in phosphate-buffered saline (PBS) for 30
min., rinsed 3x in PBS, then incubated for 30 min. in 1/50
dilution of fluorescein-goat anti-mouse IgM (FITC-GAM IgM;
Boehringer Mannheim Biochemicals, Indianapolis, IN) in PBS
with 5% normal goat serum. Cells assayed for cell surface
A2B5 after one day in culture were incubated live in 1/25
A2B5-conditioned hybridoma medium +10% HI-FBS on ice and
then processed as above. Plating efficiencies of cells were
determined retrospectively from Ektachrome slides by

Fig.2-1. Separation chamber used for the immunomagnetic
separations and Nomarsky micrographs showing cells before
and after the purification. Construction was of plastic
tubing and funnel held on a ring stand. Two magnet arrays
were used to ensure collection of all of the paramagnetic
microspheres. A pinch clamp was used to control the rate of
flow through the chamber. After ethanol sterilization in a
sterile hood the cell-microsphere mix was poured into the
funnel and the purified cells were collected in centrifuge
tubes from the bottom. Micrograph (a) shows smaller
microspheres binding to one cell (arrowhead) but not to
another. Micrograph (b) shows cells that were removed from
the mixture due to their coating of microspheres
(arrowheads). Bottom micrograph (c) shows 6 purified cells
without attached microspheres. Bar, 50nm.

13

14
counting the number of cells that adhered to the
polyornithine coated coverslips initially after the
experiment and after 1 day in culture. No attempt was made
to determine cell numbers or densities in longer-term
cultures.
Miniaggregate cultures were made of both unpurified
and purified A2B5(-) cells in 24 well tissue culture plates
on a substratum of 2% poly(2-hydroxyethyl methacrylate)
(poly(HEMA)) (Interferon Sciences, Inc., New Brunswick, NJ)
(Folkman and Moscona, 1978). This substratum prevented cell
attachment to the plastic so that maximum intercellular
contact and interaction could take place. Cultures were
plated and fed as were adherent monolayers above. Two days
before the GS assay was performed, cultures were fed with
medium containing hydrocortisone as were rotation-mediated
aggregate cultures.
Vital Dye Experiments
Experiments were performed which made use of the vital
fluorescent carbocyanine dye DiO (3,3'-dioctadecyloxa-
carbocyanine perchlorate; Molecular Probes, Eugene, OR)
for cell marking purposes. These experiments were to
determine if A2B5(+) cells would appear if less than all of
the A2B5(+) cells were removed (incomplete separations),
and also to determine if remixing the separated cells
suppressed recruitment of new A2B5(+) cells (remixed
separations). In both of these, cells that were to be

15
labelled were incubated in 200 jig/ml dye solution for 30
min. according to Honig and Hume (1986). DiO stained cells
were viewed on an epifluorescent microscope with
fluorescein optics and DiO fluorescence was not visible
with rhodamine optics. Labelling efficiency with DiO was
nearly 100%.
For calculated incomplete separations, part of the
dissociated cells to be separated were incubated in A2B5 as
in a normal separation followed by incubation in the dye
DiO. These cells were then mixed with dissociated cells
that were not incubated in either of these in some ratio
(Fig. 2-6). This mix was then mixed with microspheres and
separated in the magnetic column as above. Eluted unbound
A2B5(-) cells were plated on polyornithine coverslips as
described above. After one day in culture, these cultures
along with control unseparated cultures were immunostained
live with A2B5 as described above. A rhodamine-goat anti¬
mouse IgM (Fisher) secondary antibody was used to
discriminate the DiO staining. The percent of the DiO(+)
cells that were also A2B5(+) were scored.
In experiments where separated cells were remixed, the
separation was carried out as normal (complete) and the
bead-bound cell fraction was recovered via trypsinization
of the beads (Fig.2-7). Eluted A2B5(-) cells were meanwhile
incubated in DiO, and then the two separated cell fractions
were remixed in even proportions. After one day in culture,

16
remixed cultures along with straight purified and control
unpurified cultures were immunostained live with A2B5 and a
rhodamine-goat anti-mouse secondary antibody as above.
Scored were the percent of the DiO(+) cells that were also
A2B5(+) in remixed cultures and percent A2B5(+) in purified
and unpurified cultures.
Immunohistochemistrv
Immunostaining with monoclonal antibody A2B5 was
performed by incubating coverslips in 1/25 dilution of
hybridoma supernatant in Tyrode's + 10% HI-FBS on ice for
30 min. Coverslips were then rinsed with Tyrode's 3x, fixed
with 1% formaldehyde in phosphate-buffered saline (PBS) for
30 min., and rinsed 3x in PBS. Either fluorescein-goat
anti-mouse IgM or rhodamine-goat anti-mouse IgM diluted
1/50 + 5% NGS in PBS for 30 min. was used as the secondary
antibody. This procedure results in specific labelling of
neurons in monolayer cultures without labelling the
flattened glial cells.
Immunostaining of cells and monolayers with anti-
galactocerebroside (GC) was performed essentially the same
as with A2B5 above. A dilution of 1/50 of a purified
monoclonal antibody against galactocerebroside (Ranscht et
al., 1982; kindly provided by Dr. Steve Pfeiffer, Univ. of
Conn. Health Center) was used. FITC-GAM IgG + IgM
(Boehringer) was used as the secondary antibody.

17
For localization of GS in monolayer cultures,
coverslips were reacted with A2B5 and then fixed in Bouin's
fixative. Followed by rinsing 3x in PBS, cultures were then
incubated in polyclonal antisera specific for GS (Linser
and Moscona, 1979) at a dilution of 1/100 in PBS + 5% NGS
for 30 min. After rinsing 3x in PBS, coverslips were
incubated in a mixture of 1/50 FITC-GAM IgM and Texas Red
goat anti-rabbit IgG (Fisher) + 5% NGS for 30 min.
followed by rinsing 3x in PBS.
For localization of intermediate filaments, coverslip
cultures were first immunostained with A2B5 and
formaldehyde fixed as above. Cultures were then
permeablized with 95% ethanol at -20°C for several minutes
followed by rinsing in PBS. For glial filaments, cultures
were incubated in a 1/50 dilution of polyclonal anti-glial
fibrillary acidic protein (Dakopatts, Denmark) +5% NGS for
30 min. followed by rinsing 3x in PBS and incubation in the
appropriate mixture of secondary antibodies. For
neurofilaments, permeablized cultures were incubated in
either a 1/250 dilution of polyclonal anti-neurofilament-M
antiserum (Bennett et al., 1984) or a 1/100 dilution of a
purified monoclonal antibody NF-1 (Shaw et al., 1986)
followed by the appropriate mixture of secondary
fluorescent antibodies.

18
3H-Thymidine Incorporation
Incorporation of 3H-thymidine was performed on both
cells and monolayer cultures made from both unpurified and
purified cells. Cells were plated on polyornithine coated 8
chamber Lab-Tek tissue culture chamber/slides (Miles
Scientific, Naperville, IL) in culture medium containing 1
HCi/ml [methyl-3H]-thymidine (6.7 Ci/mmol; NEN Research
Products, Boston, MA) for analysis after 24 hours. For
analysis of monolayer cultures, cells were plated on
chamber/slides in the above medium and fed with fresh
medium with label approximately every other day.
Cultures were rinsed in Tyrode's 3x, fixed in 1%
formaldehyde, and immunostained with A2B5 as were coverslip
cultures. Chamber wells and gaskets were removed and slides
were dipped in Kodak NTB2 nuclear track emulsion (Eastman
Kodak Co., Rochester, NY) diluted 1:2 with water. Slides
were developed in Dektol developer (1:1) according to the
manufacturer's directions after 4-5 days exposure at 4°C.
Micrography
All cultures were viewed and photographed using a
Dialux 20 (Leitz, Switzerland) epifluorescent microscope
equipped for mutually exclusive visualization of
fluorescein and rhodamine fluorescence through a 4Ox
Neofluar (Zeiss, West Germany) phase contrast objective
which had a numerical aperture of 0.75.

19
Results
Cell Isolation in Methyl Cellulose
Using the tissue dissociation method above, cells were
obtained that were nearly round and free of cell surface
debris. Approximate yields were 2xl07 cells/ embryo or 107
cells/ lobe using 7 day and slightly higher (2.5xl07/
embryo) for 8 day embryos. This yield was high compared to
dissociations of tissue that is several days older (Chapter
III). When suspended in the methyl cellulose, cells were at
least several diameters apart from each other with no
apparent physical contact.
The results of the effects of cell isolation on the
production of GS after reaggregation are shown in Fig.2-2.
GS was produced by cells isolated for 1 day at levels
equaling those after immediate reaggregation. Thereafter,
levels of GS fell rapidly, however levels were still
measurable after 2 days of isolation. This demonstrated
that these cells could be manipulated for up to 2 days
before reaggregation was necessary if GS production was to
be examined in culture.
Immunomagnetic Separations
Unless otherwise specifically stated, cell reactivity
with A2B5 antibody refers only to cell surface binding and
not to possible intracellular reactivity. The
immunomagnetic cell separation procedure described herein
produced initially extremely pure populations of A2B5(-)

Fig.2-2. Results of GS production after cells were
suspended and isolated in methylcellulose medium for
varying lengths of time. GS levels after one day in
isolation were identical to that of cells which were
reaggregated immeadiately. After longer periods of
isolation the GS levels were reduced. Shown are the means
s.e.m. (n=4).

50
£
>
i—
o
<
o
Lu
O
Ld
Q_
00
00
O
1 2
Days in Methyl Cellulose
0
3

22
cells from dissociated tecta. Purified cells were routinely
>99.99% A2B5(-) as assayed by indirect immunofluorescence.
Also, the vast majority of cells recovered from the
microspheres were low level A2B5(+). Yields of A2B5(-)
cells have ranged from 100% of theoretical yield (50% of
total cells) to less than 50% of theoretical yield,
depending on the batch of microspheres used. M450.40
microspheres gave the highest purified cell yields, while
M450.51 and Dynabeads M450 gave significantly lower yields,
presumably due to increased nonspecific binding to or
trapping of A2B5(-) cells.
For separation of cells via A2B5, I have found it
necessary to use a polyclonal antibody that was
specifically produced for the purpose of microsphere
coating. Several commercial affinity-purified polyclonal
anti-mouse antibodies have been tested for effectiveness
with this monoclonal without satisfactory purifications.
Microspheres precoated with anti-mouse IgG (Dynal) have
also been tested with similar unsuccessful results. The
results presented here represent only experiments where the
purity of A2B5(-) cells exceeded 99%.
A2B5 Antigen Modulation
Although the separations resulted in purified A2B5(-)
cells, it was found that many A2B5(+) cells appeared after
in vitro culture for only several hours. After one day in
culture, approximately 50% of the initially purified cells

23
had become A2B5(+) (Fig.2-3). This is the same level of
A2B5(+) cells in unpurified cultures after one day in
vitro. Plating efficiencies of both unpurified and purified
cells after one day in culture were nearly identical and
ranged from approximately 85-100% as determined by
retrospective counts from slides. Unpurified monolayer
cultures (1-2 weeks in vitro) contained A2B5(+) cells that
exhibited only neuronal morphologies that rested atop a
layer of A2B5(-) flat cells either singly or in
multicellular aggregates (Fig.2-4). Similarly, A2B5(+)
cells with neuronal morphologies with a layer of A2B5(-)
flat cells beneath them were present in monolayers from
purified cells. The clusters of A2B5(+) neurons in both
types of cultures appeared to be the same visual density
under high or low magnification. No difference in the
amount of A2B5(+) neurons in either type of culture was
seen. The cultures made from microsphere cells, however,
always contained visually larger and/or more numerous
aggregates of A2B5(+) neurons than did the unpurified or
purified cultures (not shown).
To assess the requirement of serum (FBS) in purified
cultures where A2B5 antigen was modulated, purified cells
were plated on polyornithine coated coverslips as above.
After one day in culture, cells were immunostained live
with A2B5 as above. To determine whether or not the
modulated A2B5 antigen was trypsin-resistant as was the

Fig.2-3. Quantitation of A2B5(+) cells when unpurified and
purified at the time of separation and after one day in
culture. The unpurified cells show a small but significant
increase in the number of A2B5(+) cells after one day. The
purified cells are initially almost entirely A2B5(-), but
after one day contain the same number of A2B5(+) cells as
unpurified cultures. Shown are means ± s.e.m. for 4
individual experiments.

% A2B5(+) Cells
25
□□ Day 0

Fig.2-4. In vitro development of A2B5(+) cells in both
unpurified (a-c) and purified (d-f) cultures. Initially
after the separation the unpurified cells (a) were =45%
A2B5(+) while the purified cells (d) were devoid of A2B5(+)
cells. After one day in culture both the unpurified (b) and
purified (e) cells were =50% A2B5(+). Similarly, monolayer
cultures after approximately 1-2 weeks contained A2B5(+)
cells with neuronal morphologies. A2B5 did not react with
the flat cells underlying the neurons. Shown are monolayers
after 8 days in culture at high magnification. The fields
of neurons shown in c) and f) were chosen to illustrate the
morphologies of the A2B5(+) neurons and not to show the
equal densities of neuronal clusters that were present.
Bar, 50(im.

27

28
antigen in freshly dissociated cells, purified cells were
plated on polyornithine coverslips as above. After one day
in culture, coverslips were subjected to the same
trypsinization regime as was used to dissociate tissue.
Cells were then immunostained for surface A2B5 antigen as
above. In both instances, it was found that there was no
reduction in the percentage of A2B5(+) cells («50%). The
experiments that entailed trypsinization, however, resulted
in a large (unquantitated) release of cells from the
coverslip due to the trypsin treatment.
Subsequent to the finding of the appearance of A2B5(+)
cells in immunomagnetically-purified cell cultures, I
predicted that an analogous increase in A2B5(+) cells may
have occured when unseparated cells were suspended in
methyl cellulose for a day. It was realized that suspension
in methyl cellulose was in effect also a separation that
would not allow the A2B5(+) and (-) cells to interact. This
prediction was found to be true both qualitatively and
quantitatively (Fig.2-5). Over 50% of the cells that were
A2B5(-) when suspended in the methylcellulose converted to
A2B5(+) after one day. Thus by two different techniques the
ability of A2B5(-) cells to become A2B5(+) when isolated
has been demonstrated with similar quantitative results.
Vital Dye Experiments
I made use of the a vital fluorescent dye, DiO, to
label certain cell populations for either controlled

Fig.2-5. Quantitation of A2B5(+) cells initially and after
one day in culture when unpurified, purified, and when
suspended in methylcellulose. Results for unpurified and
purified cells are essentially the same as in Fig.1-3. The
cells isolated in methylcellulose for one day exhibited a
large quantitatively similar increase in the number of
A2B5(+) cells as did the purified cells (h of the cells
that were A2B5(-) on day 0).

CO
80"
70-
60-
o 50-
07 40-
LO
8 30-
<
^ 20-
10-
0
1 JDcy 0
Unpurified
Purified
Unpurified in
Methylcellulose

31
"incomplete" immunomagnetic separations or remixing
experiments as diagrammed in Figs. 2-6 and 2-7. This was
done to determine if new A2B5(+) cells would appear if only
some of the A2B5(+) cells were removed, and also to
determine if appearance of new A2B5(+) cells in purified
cultures could be suppressed by adding back the cells that
were removed. The results of the incomplete separations are
shown in Fig.2-6. The abscissa is the percent cells
incubated in A2B5 before separation which is in effect the
percent of A2B5(+) cells removed, since the remaining cells
had not been exposed to the monoclonal antibody. The
ordinate expresses the percent of the known A2B5(-)/DiO(+)
cells that became A2B5(+) after one day in culture. It can
be clearly seen that the result of removing increasing
numbers of A2B5(+) cells was increasing recruitment of
A2B5(-) cells to become A2B5(+). Thus the increase of
A2B5(+) cells in purified cultures was in direct response
to the depletion of A2B5(+) cells.
Fig.2-7 shows the lack of effect of mixing back those
cells that were removed with purified cells on the
recruitment of newly appeared A2B5(+) cells. Here, complete
separations were performed and the purified A2B5(-) cells
were incubated in dye before remixing with cells recovered
from the microspheres. It is clear that mixing back those
cells that were removed had no effect on decreasing the
amount of newly appeared A2B5(+) cells from those that were

Fig.2-6. Experimental design and results of controlled
incomplete separations. Plain dissociated cells were mixed
in some ratio with those that were incubated in A2B5 and
then the vital fluorescent dye DiO (diagram). After the
immunomagnetic separation cells were plated on
polyornithine. After 1 day in culture cells were
immunostained with A2B5 and scored were the % of DiO(+)
cells that had become A2B5(+), except with the leftmost
points where none of the cells were incubated in dye or
A2B5. Those cells were scored for %A2B5(+) cells initially
and after one day in culture. Plating efficiencies after
one day in culture approached 100%. The results graph shows
that as a greater percentage of the A2B5(+) cells are
removed from dissociated cells, a greater percentage of the
known A2B5(-) cells convert to A2B5(+) cells. If no A2B5(+)
cells are removed then only a slight increase in A2B5(+)
cells occurs. The (2) indicates two identical results at
that point. The results indicate that A2B5(+) cells are
recruited from A2B5(-) cells as a function of the depletion
of A2B5(+) cells.

Dissociated Cells
I
Separate
t
A2B5(- ) Cells
+ 1 Day
1) Immunostain with A2B5
2) Score % of dye(+) cells
that are also A2B5(+)
s
H 1 1 1 1 1 1 1 1 1
0 10 20 30 40 50 60 70 80 90 1
% Cells Incubated in A2B5 Before Separation

Fig.2-7. Design of experiments and results in which
separated cells were remixed. Separation was carried out as
usual. Purified A2B5(-) cells were incubated in the dye DiO
and then remixed with cells that were recovered from the
microspheres after trypsinization. After 1 day in culture
the remixed cultures were immunostained with A2B5 and
scored was the % of Dye(+) cells that had become A2B5(+).
The results graph shows the effect of remixing separated
cells on the recruitment of A2B5(+) cells from A2B5(-)
cells. The results obtained with unpurified (control) and
purified cells initially and after 1 day in culture are
essentially the same as Fig.1-3. Remixing back those
A2B5(+) cells that were removed during the purification had
no effect on reducing the number of A2B5(+) cells that were
recruited from A2B5(-) cells (rightmost graph). Shown are
means ± s.e.m. for three separate experiments.

Dissociated Cells
I
A2B5
I
Separate
X
Eluted A2B5(-)
Cells
t
Dye DiO
Mix 1
X
Beads with Cells
i
Trypsin O
60-
50-
40-
30-
20-
10
+1 Day
1) Imraunostain with A2B5
2) Score % of dye(+) cells
that are also A2B5(+) n
n = 3
Control
I I Doy 0
Purified Dye(+) in
Remixed
60
50
40
30
20
10
0
U>
Ln

36
A2B5(-). Furthermore, when cells were allowed to develop
into monolayers, A2B5(+) neurons that had retained the dye
were found (not shown). Internalization of dye from cell
surfaces as well as a high dye background from labelled
flat cells made it impossible, however, to quantitate the
percentage of A2B5(+) neurons that were dye labelled.
Nonetheless, recruitment of A2B5(+) cells did not seem to
be suppressable with this experimental regime.
Effects of Substrata
The effects of different substrata on the development
of monolayer cultures was investigated to determine whether
purified cells might develop differently than unpurified
cells. Adler et al. (1979) showed previously that different
substrata had profoundly different effects on the
development of day 7 optic tectum cells. Here, when cells
were plated on either plastic, glass, or collagen
multicellular aggregates formed rapidly (<1 day) and then
attached to the substratum after about a week in culture.
Flat cells then migrated out of the aggregates on the
substratum eventually forming a mature monolayer culture
with networks of A2B5(+) neurons and neuronal aggregates
growing on top of the layer of A2B5(-) flat cells (Fig. 2-
8). This was true for both unpurified or purified A2B5(-)
cells. These different substrata appeared to be totally
nonselective for the growth of neurons and glia in culture
(Table 2-1) .

Fig.2-8. The effects of substrata on the development of
monolayer cultures (11 days in vitro) of unpurified (a,e),
purified (b,f), microsphere fraction (c,g), and remixed
separated cells (d,h). Cultures shown were live when
photographed. Monolayer cultures grown on either glass (a-
d), plastic (not shown), Aclar (not shown), or collagen
(not shown) substrata developed into a confluent layer of
A2B5(-) flat cells on top of which was a network of inter¬
connected aggregates (arrowheads) of A2B5(+) neurons.
Cultures from (a) unpurified, (b) purified, and (d) remixed
cultures contained the same visual density of neurons and
clusters of neurons at both high and low magnifications.
Cultures made from microsphere fraction cells (c)
contained larger aggregates of neurons than the others.
Monolayer cultures grown on polyornithine substrata (e-h),
however, resulted in a network of neurons amidst nuclei
from degenerated glia in cultures made from unpurified (e)
and remixed (h) cells. Cultures made from purified A2B5(-)
cells (f) completely degenerated. Cultures made from
microsphere recovered cells (g) contained a denser network
of neurons than from unpurified cells. Bar, 50|im.

38

39
Table 2-1
Effects of substrata on monolayers®
Cells
Substratum
Growth
Unpurified
Glass
All cells
Plastic
r r
Aclar
r r
Collagen
r r
Polyornithine
Neurons only
Purified
Glass
All cells
Plastic
/ /
Aclar
/ f
Collagen
t t
Polyornithine
No growth
Microsphere
Glass
All cells
Plastic
r r
Aclar
/ /
Collagen
/ r
Polyornithine
Neurons only
Mixed
Glass
All cells
Plastic
/ r
Aclar
i r
Collagen
/ f
Polyornithine
Neurons only
® Effects of substrata on the growth of neurons and glia
in unpurified and purified monolayer cultures. The only
substratum out of those tested which was cell type
selective was polyornithine which was selective for
neurons. Practically all cells degenerated on
polyornithine in cultures made from purified A2B5(-)
cells.

40
The substratum polyornithine resulted in much
different development of embryonic tectum cells in vitro.
As observed by Adler et al. (1979), when dissociated day 7
or 8 tectum cells were plated on polyornithine only the
neurons developed and the glia degenerated over a period of
about a week. This resulted in aggregates of neurons with
interconnecting processes attached to the coverslips amidst
the nuclei and debris of dead glia (Fig.2-8). When purified
A2B5(-) cells were plated on polyornithine virtually all
cells degenerated. Conversely, when cells recovered from
the microspheres (A2B5(+)-enriched) were plated on this
substratum many neurons developed. In fact, this fraction
of cells resulted in the most visually dense networks of
neurons. This suggests that the purified A2B5(-) cells were
deficient in neurons and that the microsphere cells were
enriched in them.
Filament Expression In Vitro
The expression in vitro of two neural tissue specific
intermediate filaments was examined by immunohistochemistry
of monolayer cultures. Glial fibrillary acidic protein
(GFAP) positive filaments were found to be present in the
population of A2B5(-) flat glial cells underlying the
neurons (not shown). Immuno-reactivity was seen in most if
not all flat cells but not in cells with a neuronal
morphology. Unlike GFAP, there were cells present in
dissociated 7 or 8 day tectum that reacted with antibodies

41
specific for neurofilaments (see Chapter 3). Twelve percent
of unpurified cells showed immuno-reactivity with a
polyclonal antisera specific for the phosphorylated form of
the middle weight chicken neurofilament triplet protein
(NF-M; Bennett et al., 1984). The pattern of staining was
mostly filamentous surrounding the nucleus. This pattern is
somewhat surprising since neurofilament reactivity was
confined to cellular processes (axons) in tissue sections
from the same age tecta (data not shown). This may suggest
that the dissociation procedure caused a retraction of
processes by neurons. When cells were double labelled with
A2B5, the large majority of NF-M(+) cells were also
A2B5(+). Twelve percent of the total dissociated cells were
NF-M(+). Ten out of 12 of these cells were also A2B5(+), so
83% of the NF( + ) cells were also A2B5(+). Alternatively, 2%
of the total cells were A2B5(-) and NF-M(+), leaving 4% NF-
M(+) cells in the purified A2B5(-) population.
Neurofilament expression was examined in monolayer
cultures with both the polyclonal antisera and a purified
monoclonal antibody (NF1) specific for the phosphorylated
form of the rat heavy (200kD) triplet protein (Shaw et al.,
1986). Many of the A2B5(+) cells with neuronal morphology
in both unpurified or purified cultures contained
neurofilaments as evidenced by both antibodies (Fig.2-9).
Likewise, greater than 90% of neurofilament containing
neurons were A2B5(+). Cell monolayers that were made from

Fig.2-9. Neurofilament expression in monolayer cultures
grown on glass and on polyornithine for approximately 2
weeks. Shown are high magnifications to demonstrate
morphologies of individual cells. Anti-neurofilament
reactivity (NF-M) was found in neurons in both (a)
unpurified and (b) purified cultures to approximately the
same extent on glass. A large number of filaments were
present along with NF-M(+) cell bodies in many A2B5(+)
neurons. Many A2B5(+) neurons (<50%) in both cultures did
not react with antibodies against neurofilaments. A small
number of NF-M(+) neurons were A2B5(-). Cultures on
polyornithine of unpurified (c) and (e) microsphere
fraction cells revealed networks of neurofilaments similar
to those on glass. Purified cells (d) degenerated on
polyornithine and correspondingly contained a paucity of
NF-M reactivity. Unpurified cell cultures (f) on glass that
were double labelled with antibodies against both the heavy
and middle weight triplet proteins (NF1 and NF-M,
respectively) showed similar but not identical patterns of
reactivity in neurons. Anti- (NF-M) appeared to react with
more processes than did NF1. Neither antibody reacted with
filaments in the A2B5(-) flat cells underlying the neurons.
Bar, 5 0)i.m.


44
microsphere fraction cells had visibly more dense
neurofilaments and A2B5(+) neurons (not shown). In
monolayers, more neuronal processes were found that
contained filaments positive for the polyclonal antisera
than for NF1, and processes that contained filaments that
were NF1(+) were almost always NF-M(+) as well. Thus, the
patterns of neuronal and glial intermediate filaments
appeared identical in purified and unpurified cells when
grown on a substratum of glass.
The pattern of neurofilament immunostaining was very
different, however, in purified versus unpurified cultures
when grown on a substratum of polyornithine (Fig.2-9). As
stated above, only a network of neurons survived when
unpurified cells were cultured on polyornithine. As in
neurons cultured on glass, many of the neurons on
polyornithine contained neurofilaments. Purified cultures,
on the other hand, degenerated on polyornithine and
contained a paucity of neurofilaments. The few cells that
survived and produced neurofilaments (or had them at the
start) were presumably those A2B5(-)/NF-M(+) cells (see
above) that eluted in the purified population.
Galactocerebroside Expression In Vitro
A monoclonal antibody specific for the membrane
molecule galactocerebroside (GC) was used to study the
development of oligodenrocytes in cultures of purified and
unpurified tectum cells. No GC(+) cells were present in

45
dissociated 7 or 8 day tectum. Similarly, no GC(+)
oligodendrocytes appeared in cultures of either unpurified
or purified cells (not shown). GC(+) cells appear by day 12
of development (see Chapter 3).
GS Expression In Vitro
The expression of glutamine synthetase was examined in
both aggregate and monolayer cultures to determine whether
or not purified and unpurified cultures were similar in
this respect. A quantitative assay revealed that both
purified and unpurified cells produced the same levels of
GS when allowed to reaggregate on poly(HEMA) (Fig.2-10).
Similarly, immunostaining for GS revealed the same staining
patterns in both purified and unpurified monolayer
cultures; GS was produced in glia under or near aggregates
of neurons (Fig.2-10). Thus, the production of GS in glia
in purified cultures parallels that of unpurified cultures
both quantitatively and with respect to position to
neurons.
DNA Synthesis In Vitro
The synthesis of new DNA was examined by combining
A2B5 immunostaining with 3H-thymidine autoradiography of
both newly plated cells and of monolayer cultures. Newly
plated cells were subjected to continuous labelling with
3H-thymidine for 24 hours, followed by immunostaining with
A2B5 and autoradiography. With both purified and unpurified
cells only a small number of nuclei were labelled (<5%;

Fig.2-10. GS expression in vitro. Graph shows results of GS
assay performed on miniaggregate cultures on poly(HEMA) in
culture for ® 1 week. Unpurified and purified cultures
produced identical levels of GS. Micrographs show
immunohistochemistry of monolayers (9 days in vitro) for
GS. GS was produced only in glia underlying neuronal
aggregates in both (a) unpurified and (b) purified
cultures. Bar, 50nm.

GS Specific Activity
47
GS ASSAY

48
Table 2-2) and all of these cells in both cultures were
A2B5(-). See Fig. 2-11 for examples of labelled nuclei.
Similarly, long-term (1-2 weeks) monolayer cultures
were subjected to continuous labelling with 3H-thymidine
followed by A2B5 immunostaining and autoradiography. The
A2B5(+) neurons that developed in both purified and
unpurified cultures did not contain labelled nuclei (Fig.2-
11). Out of several hundred neurons examined where cell
bodies could be clearly seen not one contained a labelled
nucleus. The majority of A2B5(-) flat cells, however, had
densely labelled nuclei. These nuclei were rather large in
diameter, quite flat, and oval shaped.
Discussion
A diagrammatic summary of the results is presented in
Fig.2-12. I have developed a method for the purification of
dissociated embryonic brain cells based on the removal of a
target cell population by specific antibody linkages to
paramagnetic microspheres. This method was developed to
purify embryonic glia so that phenomena such as the
induction of GS by neurons could be studied in vitro in a
controlled fashion. The cell isolation experiments in
methylcellulose demonstrated that embryonic tectum cells
could be manipulated for a significant length of time («2
days) before they lost competence for GS production when
reaggregated. Thus it seemed entirely possible to develop a
method of separation that would be useful within this time

49
Table 2-2
3H-thymidine incorporation:
Day 7-8 cells®
Cells
Score
% Labelled Nuclei
Unpurified
21/519
4.0%
Purified
28/566
4.9%
® 3H-thymidine incorporation into cells during 24 hours in
culture. Both unpurified and purified cultures contained a
small percentage of cells with labelled nuclei. All of
these labelled cells were A2B5(-).

Fig.2-11. 3H-thymidine autoradiography combined with
immunohistochemistry of monolayer cultures after 13 days in
vitro. Cultures were continually exposed to label in the
culture medium and dipped slides were exposed 4 days before
developing. Each micrograph set (a and b) are comprised of
(from left to right) phase contrast, brightfield, and A2B5
immunostaining. The photographic emulsion appears as phase-
dark wrinkles in the left micrographs. All black in the
middle unstained brightfield micrographs represents label.
Individual nuclei appear as the dense dark areas. In both
(a) unpurified and (b) purified cultures 3H-labelled nuclei
were found only in the flat A2B5(-) glia. Most glia were
labelled. No labelled nuclei were found in A2B5(+) neurons
either singly or when in aggregates (arrowheads). Bar,
50jim.


Fig.2-12. Diagrammatic summary of the results, taking into
account immuno-phenotype, morphology, and 3H-thymidine
incorporation of unpurified (left) and purified A2B5(-)
(right) cells. Time in culture is denoted on the left. "A"
on a cell indicates cell surface A2B5 antigen, "5"
indicates binding of antibody 5A11, "N" indicates NF-M
immunoreactivity, "G" indicates anti-glial fibrillary
acidic protein immunoreactivity, and "GS" indicates anti¬
glutamine synthetase immunoreactivity. The numbers below
cells indicate percentages of either unpurified (left) or
purified (right) cells. Labelled nuclei are shown as
intracellular stiplings.

Day O
<3 ¿X O ®
35 10 53 2
Day 7-8
Purify
A2B5 (-) cells
o ®
96 4

54
frame. This approach would not be possible if the tissue
were retina, however, since the cells lose their
competence for reaggregation and GS production in a matter
of hours (Linser, 1987 and unpublished).
Separation Technique
The immunomagnetic separation technique that was used
was ideally suited for my purpose. Populations of extremely
pure A2B5(-) cells were obtained in a relatively short time
period. The ability to separate the tectum cells based
solely on surface antibody binding seemed necessary since
most 7 or 8 day cells are not sufficiently differentiated
to make use of other purification strategies.
Immunomagnetic cell separation has been applied to neural
tissues previously to purify oligodendrocyte precursors
(Meier et al., 1982; Meier and Schachner, 1982) with
success. This was a positive selection of the target cells
using large polyacrylamide-coated beads, however, in
contrast to the negative selection employed here. They
reported an enrichment of oligodendrocytes from 1.5% to 91%
purity and an average yield of 19%. The method described
here resulted in much higher purity of cells as well as a
higher yield. Also, cells could be recovered from the
microspheres with a purity comparable to the
oligodendrocyte selection above with much better yield.
Several major modifications of the technique used to
purge bone marrow (Treleaven et al., 1984) had to be made

55
for optimum separations of dissociated embryonic brain
cells. It was necessary to mix cells and microspheres in a
medium containing DNase. This was presumably due to the
release of DNA due to cell lysis. Various types of
separation chambers have been used for clinical
applications (Gee et al., 1987; Treleaven et al., 1984).
The chamber that was used for separation was much simpler
than those used for bone marrow purging. This open system
proved to be entirely satisfactory and contamination due
to this never occured. Lastly, satisfactory separations
were obtained with a much lower microsphere/target cell
ratio than is used for marrow purging. This is believed to
be due to increased collisions between target cells and
microspheres as a result of a higher percentage of target
cells (50%) than in infected bone marrow (=1%).
A2B5 Antigen and its Modulation
A2B5 (Eisenbarth et al., 1979) was chosen as the
target cell antibody for the immunomagnetic separations.
This was because of its high specificity for neurons in
long-term monolayer cultures of dissociated differentiated
tectum cells. Other factors that led to its choice were
that it has been reported to be neuron specific in humans
(Kim, 1985; Kim et al., 1986) and to bind to most or all
neurons in chick brain (Schnitzer and Schachner, 1982). The
antigen it recognizes is also a protease resistant
ganglioside (Eisenbarth et al., 1979; Kasai and Yu, 1983).

56
When compared with immunostaining for neurofilaments of
freshly dissociated cells it was found that a small
percentage of NF(+) cells were A2B5(-). The majority of the
NF(+) cells, however, were A2B5(+) (83%) and were removed.
Thus the specificity of A2B5 for neurons initially is not
complete and all-inclusive but encompases the majority of
identifiable neurons. No other markers that react with
differentiated neurons are known that react with 7 or 8 day
cells. Neuron specific enolase in chick brain does not
appear until later in development and apparently is not
produced in brain cell cultures (Ledig et al., 1985). Aside
from markers, culturing cells on neuron-selective
polyornithine suggests that at least most of the A2B5(+)
cells were neurons because the microsphere fraction was
enriched for neurons and the purified A2B5(-) cultures did
not contain them.
The appearance of A2B5 antigen on the surfaces of the
purified A2B5(-) cells was curious. This phenomenon was
first observed when the purified cells were incubated in
A2B5 and then the fluorescent secondary antibody. This was
done for fear that some A2B5(+) cells may have quickly
cleared the monoclonal from their surfaces after the
separation. Subsequent experiments revealed that no
reduction in the number of A2B5(+) cells occured after a
day in culture (data not shown). Therefore, the correct
assay to assess the degree of depletion of A2B5(+) cells

57
was to incubate the purified cells in only the secondary
antibody since rapid turnover of the antigen did not seem
to occur.
As was stated earlier, the antigen recognized by A2B5
is a ganglioside (Eisenbarth et al., 1979; Kasai and Yu,
1983). These sialoglycosphingolipids are believed to be
synthesized in an progressive fashion from individual
sugars transfered from nucleotide conjugates (Ledeen,
1985). This apparently takes place in the Golgi apparatus
and probably the smooth endoplasmic reticulum through
membrane bound multienzyme complexes specific for synthesis
of each ganglioside. This occurs only in the cell soma of
neurons in the chick visual system (Landa et al., 1979)
after which the gangliosides translocate to nerve endings
via fast axonal transport (Ledeen, 1985). In fractionated
cells gangliosides are found predominantly in the
synaptosomal and microsomal fractions (Hamberger and
Svennerholm, 1971). Gangliosides exhibit turnover and are
degraded primarily in lysosomes in an ordered fashion and
evidence suggests that postsynthetic processing of them
does not occur (on the cell surface) between synthesis and
degradation (Ledeen, 1985).
The mechanisms that control the synthesis and export
of gangliosides to the cell surface not at nerve endings
are not well understood. From what is known about
ganglioside biosynthesis (Ledeen, 1985) it is probable that

58
the appearance of A2B5 antigen on purified cell surfaces is
not due to the modification of existing gangliosides on the
surface or even intracellularly. This cannot be ruled out,
however, because increases in the complex gangliosides
(including GQic) during development of the chick optic
tectum are correlated with decreases in simpler precursors
(Gd3) (Rosner, 1980). Other possibilities that may explain
the appearance of A2B5 antigen are new synthesis and export
to the surface, or export of an intracellular pool. Since
it is known that chick brain cells have intracellular pools
of A2B5 antigen (see below; Schnitzer and Schachner, 1982)
this possibility is quite real. Export to the surface
could be via exocytotic vesicles or via a ganglioside
transfer protein found in brain (Gammon and Ledeen, 1985).
These types of export mechanisms could account for the
appearance of A2B5 antigen on the surfaces of purified
cells within the several hours that it has been seen to
occur (see Results).
The number of "recruited” A2B5(+) cells was a function
of the number of original A2B5(+) cells that were removed
as demonstrated by the calculated incomplete separations.
The question remains as to whether this triggered
modulation of A2B5 antigen was a result of the removal of
neurons or of A2B5(+) cells. Some (20%) of the A2B5(+)
cells that were removed are known to be neurons because
they contained neurofilaments. The remaining 80% were of

59
unknown type. They may, however, have been neurons that
either did not contain neurofilaments because of lack of
synthesis in the tissue or because of severing of axons
during dissociations. The results of culturing cells
recovered from the microspheres on glass and polyornithine
suggest that these are neurons because cultures from these
A2B5(+)-enriched cells appeared to be neuron-enriched.
Later on, there was a tight correlation between a cell
being A2B5(+) and having a neuronal morphology in long-term
monolayer cultures. Therefore, modulation of A2B5 antigen
on purified A2B5(-) cell surfaces may be a response to
neuronal depletion as effected by the removal of A2B5(+)
cells. Consistent with this hypothesis, surface A2B5
antigen modulation has been shown to occur with mouse
cerebellar astrocytes in culture in response to complement-
mediated depletion of neurons using an independent neuronal
marker (Nagata et al., 1986).
The inability to prevent recruitment of new A2B5(+)
cells by mixing back the cells that were removed implies
that the recruitment was very rapid and irreversible. There
exists the possibility that this was due to damage of the
cells that were removed from the microspheres by the second
trypsinization. This seems unlikely, however, since the
cells were trypsinized before dissociation. These recovered
cells also grew well in culture. It appears that the
events that led to the expression of A2B5 antigen on the

60
recruited cells were irreversible within the time frame in
which they were remixed (an hour).
A main question that remains is the significance of
A2B5 antigen and its appearance on the surfaces of purified
cells. Gangliosides are major constituents of the
glycocalyx of neural cells (Ledeen, 1985). Their general
stability in the membrane makes them ideal candidates for
roles such as adhesion and recognition. Gangliosides are
known to be the neural cell surface receptors for tetanus
toxin (Gdiid and G-rib) and for cholera toxin (Gmi) (Ledeen,
1985). Gangliosides are also thought to influence in some
way the formation of synapses (Grunwald et al., 1985) and
the process of myelination (Ledeen, 1985). In culture,
purified gangliosides have been shown to mediate adhesion
of embryonic chick retina cells (Blackburn et al., 1986)
and to alter the morphology and growth of astrocytes from
fetal rat brain (Hefti et al., 1985). It is possible that
the ganglioside recognized by A2B5 on cell surfaces
similarly may function as a recognition molecule. It is
unlikely that it serves as an adhesion molecule since cells
that had bound antibody on their surfaces did not appear
to have diminished ability to form either heterotypic or
homotypic contacts with other cells. It may even be
proposed that A2B5 antigen is a molecule that is involved
in the communication between neurons and other cells since
A2B5 is neuron-specific in monolayer cultures, here, and

61
since its modulated appearance on cell surfaces depends on
the removal of cells already expressing it at their
surface.
Development of Purified Cells
It is clear that A2B5(+) cells were recruited in
purified cultures. It is also clear that neurons appeared
in purified cultures when grown on a nonselective
substratum. Initially, the small proportion of identifiable
neurons present before separation were depleted by 83% as
evident by NF immunoreactivity. Degeneration of purified
cells on polyornithine also suggests that neurons were
depleted because no cells survived. Presumably, they would
have grown if they were present in the purified cells. This
degeneration occured even though the cells had become
A2B5(+). Thus, the presence of A2B5 antigen on a cell
surface does not in itself correlate with survival on
polyornithine. Microsphere fraction cells (A2B5(+)-
enriched), on the other hand, resulted in the visually most
dense network of neurons, which suggests that the majority
of neurons were A2B5(+).
When grown on a nonselective substratum such as glass
for « 1-2 weeks there appeared to be no decrease in the
density of A2B5(+) neurons or of neurofilaments in purified
cultures as compared to unpurified monolayers. If the
purified A2B5(-) cells were initially devoid of the
majority of neurons as is suggested above, then, neurons

62
must have come from preexisting nonneuronal cells to
exhibit the same density as in unpurified cultures (see 3H-
thymidine discussion below). The simplest explanation for
equal densities of neurons is that A2B5 expression on the
surfaces of cells is irrelevant to the development of
neurons in long-term monolayer cultures. A2B5 antigen may
have been modulated up and down on cell surfaces in both
unpurified and purified cultures. Then, the A2B5(+) neurons
at the culture endpoints may not have developed from the
A2B5(+) cells seen initially or after a day in culture. The
phenomenon of recruitment of A2B5(+) cells in purified
cultures may be separate from the appearance of neurons in
these cultures. No experiments were performed that could
conclusively demonstrate which of the possibilities had
occured.
On the other hand, neurons as defined by morphology
and/ or neurofilament content were almost always surface
A2B5(+) (>90%) in long-term monolayer cultures. Conversely,
cells that expressed A2B5 antigen on their surfaces always
had a neuronal morphology in monolayer cultures.
Additionally, the presence of cells that were surface
A2B5(+) always preceded the development of neurons in long¬
term cultures. A2B5(+) cells were present at the outset in
unpurified cultures, and were present after one day in
purified cultures. These correlations raise the possibility
that A2B5(+) neurons in long-term cultures developed from

63
the A2B5(+) cells that were seen after one day in culture.
The A2B5(+) neurons that developed in unpurified cultures
may have originated from the A2B5(+) cells that were
present when the tissue was dissociated (which were
presumably were the same cells that were A2B5(+) after a
day in culture). The fact that either no change or a slight
increase in the percentage of A2B5(+) cells occured in
unpurified cells after one day in culture is consistent
with A2B5 antigen not being modulated in unpurified
cultures. Similarly, the A2B5(+) neurons that developed in
purified cultures may have originated from the recruited
A2B5(+) cells seen after a day in culture. If this
hypothesis were true, then, this would imply the existence
of a previously unidentified intermediate cell type in the
optic tectum neuronal lineage. This cell would have the
characteristics of being A2B5(+)/ NF-M(-) and would be
susceptible to degeneration on polyornithine.
The analysis of GS in culture revealed another manner
in which purified cultures were identical to unpurified
cultures. GS was produced in purified monolayer cultures in
an indistinguishable pattern from unpurified cultures.
Quantitatively, the expression of GS in the two types of
cultures was also the same. These results suggest several
possible explanations. One is that the purification did
not result in separation of neurons and glia. This has been
discussed above. Another is that a small number of neurons

64
may be able to induce production of GS in a certain number
of glia as well as a large number of neurons could. This
possibility is compounded by the fact that there are many
different types of neurons in the optic tectum (LaVail and
Cowan, 1971a) and it is not known which types are capable
of GS induction. Maybe the small percentage of A2B5(-)
neurons that were in the purified cells were the neurons
that induced GS. Another possible explanation is that
neurons were recruited in the neuron-depleted purified
cells in a rapid manner so that they could interact with
the glia to produce GS. This would have to have been within
about a day as was determined by the isolation of cells in
methyl cellulose. If the appearance of A2B5 antigen on cell
surfaces was an indication of commitment to becoming a
neuron as is suggested then the ability to induce GS may
also have occured rapidly as did the expression of A2B5
antigen. However, this hypothetical change was not
sufficient to ensure survival of cells on polyornithine.
It is also worth mentioning that levels of GS produced
in aggregate cultures made from cells that were isolated in
methylcellulose were identical to levels produced by
immeadiately reaggregated dissociated cells. This is
surprising since it was shown that a much greater number
of A2B5(+) cells existed initially in cells from the
methylcellulose («75%) than in immeadiately reaggregated
cells («50%). So either the number of A2B5(+) cells has

65
nothing to do with the amount of GS produced or the proper
ratio of neurons to glia is somehow obtained. Although it
was shown that GS is produced in glia under clusters of
A2B5(+) neurons it could not have been determined whether
more GS activity was induced under large clusters than
under small clusters. The proper ratio of neurons to glia
could have been accomplished by division of the glia since
these cells have been shown to incorporate 3H-thymidine in
culture.
There appears to be an absolute correlation between
having cell surface A2B5 antigen and not synthesizing new
DNA in cultures of tectum cells. The lack of 3H-thymidine
labelled nuclei in neurons in cultures from unpurified day
7 or 8 cells was expected since other workers have found
that the majority of neurogenesis in the tissue has already
occured by this time (LaVail and Cowan, 1971b; Puelles and
Bendala, 1978). The fact that no labelled neurons were
found suggests that all of the neurons that survived in
culture had completed their terminal S-phase by day 7.
Similarly, the lack of any labelled nuclei in recruited
A2B5(+) cells that appeared in purified cultures
demonstrated that new DNA synthesis was not required for
cell surface expression of A2B5 antigen. The antigen may
have been expressed via a mechanism as discussed in the
section above. The neurons that developed in these
cultures, likewise, did not require new DNA synthesis for

66
differentiation (i.e. the synthesis of neurofilaments).
This finding suggests that these neurons originated from
cells that were already born. This may have been from a
resting blast cell that was limited to the choices of
either differentiating into a neuron or dying.
Alternatively, these neurons may have originated from cells
that would have become glia had the need for more neurons
not occured. This possibility would be exciting, since it
suggests that a change in phenotype or even cell type may
occur in post-mitotic cells if the neurons that developed
in purified cultures were recruited as is suggested.
Glial Development and Plasticity
Perhaps the most interesting question that the results
here pose is: From what population of cells were the
recruited neurons taken? Two theoretical possibilities
exist. The recruited cells either would have become
something else had there been no depletion, or they were
resting blast cells that normally would degenerate if not
needed. And since these cells were of neuroectodermal
origin, if they were not to become neurons then they were
to become glia. The results presented here do not support
one or the other possibility. Analysis of cell lineage in
the rat retina by using recombinant retroviral vectors has
shown that both neurons and glia are produced from common
progenitors throughtout development (Turner and Cepko,
1987). However, results with purified cells from day 12 or

67
13 optic tectum suggest that the recruited cells in
cultures stem from the astrocyte lineage (Chapter 3). The
majority of the recruited A2B5(+) cells in these purified
cultures showed immunoreactivity for GS thus identifying
them as glia. There is also evidence in the chick
peripheral nervous system that certain glial precursors are
capable of being diverted to a neuronal lineage under
certain transplantation conditions (Le Lievre et al.,
1980) .
The mechanism of recruitment is also unclear. One
possibility is that there exists either a negative feedback
system between A2B5(+) and A2B5(-) cells or a positive
feedback system between A2B5(-) cells only. With the
negative feedback system the A2B5(-) cells would have
recognized the loss of the A2B5(+) cells by some mechanism
and then reacted as a result of this. With the positive
feedback system the A2B5(-) cells would have sensed an
increase in the density of A2B5(-) cells in the purified
cultures and reacted to replenish A2B5(+) cells. The cell
isolation experiments, however, clearly rule out the latter
possibility since recruitment of A2B5(+) cells occured as a
result of a loss of contact between all cells. What remains
in question then is whether the communication between
A2B5(+) and A2B5(-) cells is via cell contact or soluble
factors.

68
An interesting finding concerning glial development in
culture is the lack of appearance of oligodendrocytes in
culture. These results using an antibody for the marker
galactocerebroside (GC) essentially confirm the results of
Linser and Perkins (1987a) who failed to find cells
positive for the oligodendrocyte markers myelin basic
protein and S-100 (Linser, 1985). This is in contrast to
cultures made from day 12 or 13 tectum cells where GC(+)
cells are present and develop in vitro (Chapter III). Thus
it seems that either future cellular interactions were
disrupted that were required for oligodendrocytes to
develop, or that the culture conditions did not contain
some growth factor(s) that was required earlier in
development. It should be noted that these results are in
contrast to those with rat brain cells where
oligodendrocytes appear in cultures that do not contain
them initially (embryonic day 10) on time with those that
appear in vivo 13-14 days later (Abney, Bartlett, and Raff,
1981).

CHAPTER III
SEPARATION OF DAY 12-13 CELLS
Introduction
The previous chapter described an immunomagnetic
separation method to separate cells from early (day 7-8)
embryonic chick optic tectum. This method resulted in
extremely pure populations of cells that were negative for
the cell surface marker A2B5 (Eisenbarth et al.f 1979). It
was found, however, that the A2B5 antigen was modulated on
the surfaces of about half of the purified cells in direct
response to the depletion of A2B5(+) cells. Neurons were
also apparently recruited in these cultures. Thus, day 7
and 8 optic tectum cells showed a remarkable ability to
maintain the correct number of A2B5(+) cells and neurons.
It was not clear, though, from what population of cells the
recruitment was occuring. This was largely due to the fact
that no glial differentiation markers which occur later in
development were present at days 7 and 8 to identify
definitive glia.
During development of the chick optic tectum, commonly
recognized glial differentiation markers do not appear
until relatively late in development (Linser and Perkins,
1987a). Glutamine synthetase (GS) is detectable by
immunohistochemistry in some astrocytes at day 9 and is
69

70
produced eventually in most if not all astrocytes. Glial
fibrillary acidic protein (GFAP) appears in a small
population of astrocytes beginning at day 16 of
development. The oligodendrocyte specific marker myelin
basic protein (MBP) appears on day 12 followed by S-100 on
day 16. This is in striking contrast to the neuronal
marker, neurofilaments, which appears beginning on day 3 of
development when neurons begin to exit the mitotic cycle
(Bennett and DiLullo, 1985). Unfortunately, however,
another widely used neuronal marker, neuron-specific
enolase, does not appear in chick brain until much later
(day 17) (Ledig et al., 1985).
The monoclonal antibody A2B5 which was used to
tentatively identify and remove neurons immunomagnetically
seemed at first to be a relatively stable marker with
freshly dissociated cells. It reacted with approximately
50% of dissociated cells from days 7-13. If it was assumed
that A2B5 was reacting with the same population of cells
throughout this range of ages, it would be interesting to
examine whether induced deficiencies of A2B5(+) cells and/
or neurons would be compensated for if day 12 or 13 A2B5(-)
cells were purified as they were with day 7 and 8 cells.
Since two known glial differentiation markers have appeared
in the tissue by this time (GS and GC), it might be
hypothesized that the capacity for neuronal "recruitment"
from other cells would be very limited. In several other

71
respects, as stated by LaVail and Cowan (1971a), age 12
tissue differs from earlier ages as follows: By this time
all mitosis in the neuroepithelium has ceased (Cowan et
al., 1968). The major 6 laminations of the tectum have been
arranged, and many cells have obtained their relative final
positions. Retinal axons by now have penetrated all parts
of the tectum, and by this age the superficial tectal
laminae are dependent upon retinal contact for survival
(Kelly and Cowan, 1972).
Therefore, I have immunomagnetically purified A2B5(-)
cells from day 12 and 13 cells for purposes of studying
their development in vitro as compared to unpurified cells.
A2B5 was compared to the differentiation markers present at
this time (NF and GS) as well as to the oligodendrocyte
marker galactocerebroside (GC) in freshly dissociated
cells. Monolayer cultures of purified and unpurified cells
were analyzed by immunohistochemistry for these markers and
for glial filaments (GFAP). 3H-thymidine autoradiography
was also performed on cells and monolayers to examine the
role of new DNA synthesis in recruitment and
differentiation. Reaggregation of cells to elicit GS
production as an indicator of neuronal-glial interaction
was not performed, since this age tissue is beyond the age
at which this can be done successfully (unpublished).
Similarly, the effects of polyornithine substrata on glial
degeneration (Adler et al., 1979) could not be utilized,

72
because glia can survive on this substratum by this age
(see Results).
I have found that glia which have been deprived of
neuronal contact can alter their immuno-phenotype
drastically. Day 12 and 13 tectum cells have a very limited
ability, if any, to replenish depleted neurons in vitro.
Instead, phenotypes were found in purified cultures that
were intermediate between, or showed characteristics of
both, neurons and glia. These were not found in unpurified
cultures and apparently represented reactions to depletion
of neurons. Taken with previous findings in Chapter II, the
inability to replenish neurons coincident with the
appearance of abnormal phenotypes suggest that the non¬
neuronal cells in day 12 and 13 tissue (glia) are
restricted in their potential to become neurons in vitro.
Materials and Methods
White Leghorn chick embryos were used throughout this
investigation. Fertile eggs were purchased from the
Division of Poultry Science, University of Florida, and
stored at 15°C until initiation of incubation at 37.5°C in
a humidified egg incubator. For this series of
investigations, 12 and 13 day embryonic optic tecta were
dissected as described in Chapter II.
To obtain sufficient quantities of dissociated cells
treatment of tissue with two different proteases was
required. Tissue was minced finely and then incubated in

73
5mg/ml neutral protease (Dispase; Boehringer Mannheim
Biochemicals, Indiannapolis, IN) for 1 hour with aggitation
followed by incubation in 0.4% Trypsin (Nutritional
Biochemicals, Cleveland, Ohio) for 30 min., both at 37°C.
Dissociation of tissue into single cells was as described
in Chapter II.
Adherent monolayer cultures were prepared on glass or
polyornithine as described in the previous chapter.
Briefly, 10s cells/ well were plated on coverslips in 24
well tissue culture plates in Medium 199 supplemented with
10% fetal bovine serum. Monolayer cultures were kept in a
standard tissue culture incubator in a 5% C02/ air
atmosphere. Cultures were fed with fresh medium
approximately every other day.
Cell "isolation" in methylcellulose was accomplished
by suspending dissociated cells in a semisolid medium of
1.3% methylcellulose in Medium 199 according to the
previous chapter.
Immunomaqnetic Separations
Immunomagnetic separations were carried out exactly as
described previously and so will not be described further
here. Cells assessed for cell surface A2B5
immunofluorescently following a separation were fixed with
1% formaldehyde in phosphate-buffered saline (PBS) for 30
min., rinsed, then incubated for 30 min. in a 1/50 dilution
of fluorescein-goat anti-mouse IgM (FITC-GAM; Boehringer

74
Mannheim) in PBS with 5% normal goat serum added. Cells
assayed after one day in culture were incubated live in
1/25 A2B5-conditioned hybridoma medium + 10% heat-
inactivated fetal bovine serum for 30 min. on ice and then
processed as above. Plating efficiencies of cells were
determined retrospectively from Ektachrome slides by
counting the number of cells that adhered to the
polyornithine coated coverslips initially after the
experiment and after 1 day in culture. No attempt was made
to determine cell numbers or densities in longer-term
cultures.
Immunoradiometric Assay for A2B5
An immunoradiometric assay (Hunter, 1978) was used to
quantitate cell surface binding of A2B5 both immediately
after the separation and after 1 day in culture. For this,
12sI-labelled antibodies were prepared by the Iodo-Gen
method (Pierce Chemical Co., Rockford, IL) according to the
manufacturer's directions using polyclonal goat anti-mouse
IgG +IgM + IgA (Organon Teknika- Cappell, West Chester,
PA) . Cells assayed for cell surface A2B5 were processed
exactly as those for the immunofluorescent assay, with the
iodinated secondary antibody in place of the fluorescent
secondary antibody. Processed coverslips were placed in
vials and bound radioactivity was measured with a Beckman
Gamma 7000 (Beckman Instruments, Norcross, GA). Specific
DPMs were obtained by subtracting mean nonspecific DPMs

75
from total mean DPMs of triplicate or quadruplicate
coverslips. Nonspecific background DPMs were obtained by-
incubating unpurified cells in an irrelevant monoclonal
antibody specific for pipefish vitelline envelope (provided
by Dr. P.C.Begovac, Whitney Laboratory) for 30 min. on ice
and then processing them identically as those incubated in
A2B5.
Immunohistochemistrv
All immunohistochemistry of cells and monolayer
cultures was performed as described in the previous chapter
except as noted below. 5A11 (Linser and Perkins, 1987b)
immunohistochemistry of monolayers was carried out
identically to that for A2B5. Immunostaining of cells for
glutamine synthetase was accomplished by permeablization
after formaldehyde fixing with 95% ethanol for 2 min. at
-20°C, rinsing, and incubation in a 1/100 dilution of
polyclonal anti-GS (Linser and Moscona, 1979) for 30 min.
Other NF antibodies rasied against rat neurofilaments and a
monoclonal antibody specific for GFAP (DA3, NN18, anti-MSH,
A5; kindly provided by Dr. G. Shaw, Univ. of Florida) were
also used to immunostain monolayers.
3H-Thymidine Incorporation
New DNA synthesis in cell cultures was investigated by
3H-thymidine autoradiography combined with immunohisto¬
chemistry as described in the previous chapter.

76
Results
The dissociation protocol used for younger tissue
(Chapter II) did not produce satisfactory numbers of single
cells when using older tissue (day 12 and 13). Therefore,
the double protease treatment was adopted. This
still resulted in only approximately 5xl06 cells/ embryo.
The double protease treatment was used only because a
greater number of viable cells could be obtained for
experiments and the results herein are not believed to
reflect an effect caused by the tandem proteases (see
Discussion).
Markers for Dissociated Cells
Unless otherwise specifically stated, the reference as
to whether a cell is A2B5(+) or (-) refers to cell surface
binding of this antibody only. With dissociated day 12 and
13 optic tectum cells, discrete populations were labelled
by antibodies against two recognized glial antigens and one
neuronal antigen, as well as with A2B5 (Figs.3-1,3-2).
Approximately 10% of dissociated cells were reactive with
the polyclonal antibody against neurofilaments. The
majority of these were A2B5(+). It was more difficult to
attempt to quantitate this because, unlike younger cells,
there was an absence of reactive filaments in the cell
cytoplasm around the nucleus. NF-M(+) (Bennett et al.,
1984) cells were less distinct and mostly showed
immunoreactivity in what was apparently membrane blebs of

Fig.3-1. Graph comparing the cell surface reactivity of
marker A2B5 to reactivity against NF, GS, and GC. The
percent of total dissociated cells that react with each
antibody is quantitated on the abscissa. The percent that
react with purified cells is obtained by multiplying the
percents on the A2B5(-) side by 2. A2B5 bound to
approximately 50% of dissociated cells as shown on the top.
Anti-NF reactivity was present in =12% of the cells and 10
out of 12 of these were A2B5(+) (this data only is from
days 7 and 8 cells; see text). Anti-GS antibodies reacted
with =40% of total cells and all of these were A2B5(-).
Similarly, anti-GC antibodies reacted with only A2B5(-)
cells (6%) and therefore were enriched in the purified
fraction. Approximately 95% of the purified cells could be
identified as reactive with the markers NF, GS, and GC.
More than 90% of the purified cells could be identified as
either GS(+) astrocytes or GC(+) oligodendrocytes.

Á2B5 M««:i
A2B5(+)
NF 1
GS
1 1 1 1
GC
1 1 1 1
50 40 30 20 10 0 0 10 20 30 40 50
% CELLS

Fig.3-2. Double-label immunostaining of dissociated
unpurified (a-c) and purified (d-f) cells with A2B5 and
for either GS (a,d), NF (b,e), or GC (c,f). These
micrographs correspond to the graph shown in Fig.2-1. Refer
to that figure for quantitative analysis of double
labelling. Bar, 50|im.

80

Fig. 3-2 Cont'd

82
retracted or severed axons on the surfaces of the cells.
Therefore, the most accurate count of NF-M(+) cells was
obtained with the day 7 and 8 cells described in the
previous chapter. Even this data may be an underestimate of
the number of NF-M(+) cells for the above reasons.
Polyclonal antibodies specific for the definitive
glial marker GS reacted with approximately 40% of
dissociated day 12 and 13 cells. These cells were A2B5(-)
and so the purified population of A2B5(-) cells (see
analysis of purification below) was approximately 80% GS(+)
(Figs.3-1,3-2). Thus, the astrocytes that were identifiable
as such by immunoreactivity with anti-GS were concentrated
in the purified population.
Dissociated cells at this time also were reactive with
a monoclonal antibody specific for the oligodendrocyte
marker galactocerebroside (GC; Ranscht et al., 1982).
Approximately 6% of dissociated cells were GC(+). Similar
to the analysis of GS, all of these cells were A2B5(-) and
so the purified A2B5(-) cells were 12% GC(+) (Figs.3-1,3-
2). Taken with the results for GS, >90% of the purified
cells could be identified as glia by these two
differentiation markers. A small population of the purified
cells (~4%) could be identified as neurons that contained
neurofilaments.

83
Effects of Substrata
As in the preceeding chapter, both unpurified and
purified A2B5(-) cells were cultured on substrata of either
glass or polyornithine. It was found that although there
were slight differences in the development of monolayers
with regards to the extent of cell aggregation before
spreading, polyornithine was not found to be selective for
neurons as it was with day 7 and 8 cells. No cell phenotype
that was observed to develop on glass was absent from
monolayers on polyornithine. Thus, the degeneration of
glial precursors on polyornithine was not manifested in
day 12 and 13 cells.
A2B5 Antigen Modulation
By the immunofluorescent assay described here and in
the preceedind chapter, the immunomagnetic separations
resulted in extremely purified populations of A2B5(-)
cells. Unpurified cells were approximately 45% A2B5(+)
(Fig.3-3). Purified cells were virtually free of A2B5(+)
cells, with a typical purity of >99.99% A2B5(-). After one
day in vitro, however, large numbers of A2B5(+) cells
(=80%) were present in the cultures of initially purified
cells (Fig.3-3). No increase in A2B5(+) cells was observed
in cultures made from unpurified cells. Thus, A2B5 antigen
is apparently modulated on the surfaces of the majority of
A2B5(-) cells in purified cultures, but not on the surfaces
of A2B5(-) cells in unpurified cultures. Plating

Fig.3-3. Analysis of initial purification of A2B5(-) cells
and A2B5 antigen modulation on purified cell surfaces by
immunofluorescence (a) and immunoradiometric (b) assays.
The immunofluorescent assay (a) indicated that the purified
population of cells was initially devoid of A2B5(+) cells.
After 24 hours in culture, however, the purified population
was «80% A2B5(+). The unpurified population did not show
any increase in A2B5(+) cells. The immunoradiometric assays
(b) confirmed that the purified cells were devoid of cell
surface A2B5 antigen initially (left graph). After one day
in culture the purified cells expressed levels of A2B5
antigen approaching those of unpurified cells (right
graph). Shown in all graphs are means ± s.e.m. DPMs are
higher in the +1 day immunoradiometric assay due to use of
125I-labelled antibody with 3x the specific activity.

85
—
"q3
cj
+
irT
GQ
OJ
<
a)
Unpurified Purified
DAY 0
IRA
CL
O
o
*o
<0
Cl
CO
b)
3000
2500
2000
1500
1000
500
0
Unpurified
2
CL
o
o
'o
<u
a.
n=3 ^
i
Purified
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
+ 1 DAY
IRA
Unpurified Purified

86
efficiencies of unpurified and purified cells after a day
in culture were nearly identical and ranged from
approximately 60-90% as determined by retrospective counts
from slides.
To be certain that the observation of the modulation
of A2B5 antigen by immunofluorescence was accurate,
immunoradiometric assays were performed to quantitate cell
surface A2B5 antigen. Assays identical to the
immunofluorescent ones were performed using an iodinated
secondary antibody. The results of this assay are shown in
Fig.3-3 and confirm the immunofluorescent results of
depletion of A2B5(+) cells in purified populations. This
assay also established that the immunofluorescent assay was
a sensitive and satisfactory method for assessing the
presence or absence of A2B5(+) cells. Similarly, the
immunoradiometric assay confirmed the presence of near
control levels of A2B5 antigen on the surfaces of purified
cells after one day in vitro (Fig.3-3). There was no
quantitative increase in the levels of cell surface A2B5
antigen in unpurified cultures after one day.
Similar to the results with day 7 and 8 cells, an
increase in the number of A2B5(+) cells («75%) was observed
after dissociated cells were kept suspended in isolalation
in a semisolid medium containing methylcellulose for 24
hours.

87
Development of A2B5Í+1 Cells In Vitro
Although up to this point the phenomenon of appearance
of A2B5(+) cells in purified cultures of day 12 and 13
cells appeared similar to recruitment in purified day 7 and
8 cells, the appearance of A2B5(+) cells in long-term
monolayers differed markedly. A2B5(+) cells that appeared
in purified cultures for the most part did not have a
neuronal morphology. They exhibited a flattened glial
morphology (Fig.3-4) after about a week in culture and
rested atop the A2B5(-) flat cells and were then round
after several more days or when the cultures had become
confluent. The appearance of these A2B5(+) cells with
nonneuronal morhpology in purified cultures was not
prevented by adding back the cells that were removed by the
microspheres (data not shown). A2B5(+) cells with
nonneuronal morphology also appeared in cultures made from
dissociated cells that were kept in methylcellulose for a
day (Fig.3-4). This was in contrast to the appearance of
A2B5(+) cells in unpurified monolayer cultures. These cells
exhibited a purely neuronal morphology in vitro and not the
nonneuronal morphologies observed in the purified cultures.
Nonneuronal A2B5(+) cells were also absent from cultures
made from cells recovered from the microspheres and this
fraction appeared to contain the largest and most numerous
aggregates of neurons (data not shown). Some A2B5(+) cells
with neuronal morphology were observed in purified

Fig.3-4. In vitro development of A2B5(+) cells in monolayer
cultures made from unpurified (a-c) and purified (d-f)
cells. Shown are phase/ fluorescent pairs with the
antibody staining shown on the right. A2B5(+) cells in
unpurified cultures appeared solely as neurons after about
a week (a) or 2 weeks (b) in vitro. In these cultures, the
antibody 5A11 reacted only with the flat A2B5(-) glia (c).
In purified cultures, the A2B5(+) cells appeared
predominantly with a flat glial morphology after about a
week in culture (d) and were rounded up after several more
days or when the monolayers approached confluency (e). Some
A2B5(+) neurons with processes also developed in these
cultures. The A2B5(+) nonneuronal cells reacted with 5A11
(f, arrowheads) revealing a partial glial phenotype. The
micrograph pair in (g) shows several A2B5(+) nonneuronal
cells in a monolayer culture that was made from unpurified
cells that were suspended in methyl cellulose medium for 24
hours prior to plating. Thus these cells can be generated
in culture by 2 different isolation procedures. Bar, 504m.

89

90
cultures. These may have been the small number of
NF(+)/A2B5(-) neurons that were detected in the initially
purified cells (see above). The A2B5(+) cells in purified
(long-term) monolayer cultures with nonneuronal morphology
reacted with the monoclonal antibody 5A11. This antibody is
known to be specific for glia in retina (Linser and
Perkins, 1987b) and similarly reacted only with the
surfaces of A2B5(-) glia in the cultures of unpurified
optic tectum cells (Fig.3-4). Purified monolayer cultures
contained many 5A11(+) round cells. These were presumably
the A2B5(+) round cells seen in duplicate cultures, but
double-label analysis could not be performed because both
A2B5 and 5A11 are an IgM. Thus the A2B5(+) cells in
purified cultures not only developed a nonneuronal
morphology in culture but expressed a surface antigen
(5A11) that normally seems to be restricted to glia.
Intermediate Filament Expression In Vitro
Two types of intermediate filaments were investigated
in cultures of unpurified and purified optic tectum cells.
Glial filament expression was examined with both a
commercial polyclonal antibody against bovine glial
fibrillary acidic protein (GFAP) and monoclonal A5 (Debus
et al., 1983). A5 did not react with anything in our
cultures and so the following results are from using the
commercial polyclonal antibody. In unpurified monolayer
cultures GFAP reactivity was confined to the A2B5(-) flat

91
cells (Fig.3-5). Immunoreactivity was present in most of
these cells. No reactivity was observed in the A2B5(+)
neuronal cells. In purified cultures GFAP reactivity was
also present in most of the A2B5(-) flat cells. Reactivity
was also present as filaments, however, in many of the
A2B5(+) flat and round cells that appeared in these
cultures (Fig.3-5). So, not only do the A2B5(+) cells with
nonneuronal morphology express the surface antigen
recognized by 5A11 but some also appear to express glial
filaments as well.
Neurofilament expression in cultures was examined with
both a polyclonal antibody (NF-M) specific for the
phosphorylated form of the middle weight neurofilament
triplet protein and a monoclonal antibody (NF1) specific
for the phosphorylated heavy weight triplet protein. In
paraffin sections of tissue from 12 day tectum, NF-M
reactivity was confined to neuronal processes and
predominantly the 2 axonal layers, the stratum opticum and
stratum album céntrale (data not shown). In monolayer
cultures made from unpurified cells, neurofilaments were
observed in the processes and many cell bodies of A2B5(+)
neurons. Not all NF-M(+) neurons were A2B5(+), however.
Small numbers of neurofilament containing neurons were
A2B5(-). An almost identical pattern of reactivity was
exhibited by NF-M and NF1 in neuronal processes, although
sometimes filaments that were NF-M(+) were not NF1(+). In

Fig.3-5. Expression of neurofilament reactivity (b,d,e) and
glial filament reactivity (a,c) in 1-2 week old monolayer
cultures of unpurified (a,b) and purified (c-e) cells. GFAP
reactivity in unpurified cultures (a) was confined to the
majority of flat A2B5(-) glia that underlie the networks of
single and aggregated (labelled "a") neurons. In purified
cultures (c) GFAP reactivity was also found in the A2B5(+)
nonneuronal cells (arrowheads). An A2B5(+) cell with
extended processes (white arrow) also appears to react with
GFAP antibodies. Neurofilament reactivity in unpurified
cultures (b) were present in neurons (mainly A2B5(+)) and
not in the A2B5(-) glia. In purified cultures (d,e) NF
reactivity also was found in some flat cells that were
A2B5(-) or low level A2B5(+). Distinct filaments could be
seen in these cells as shown in a higher magnification of
several cells from a different field in (e). Bar (a-d),
50p.m. Bar (e), 25pm.


94
purified cultures, NF-M(+) filaments were found in many of
the A2B5(+) cells that had neuronal morphology (neurons).
Surprisingly, many NF-M(+) flat and round cells were also
observed dispersed throughout the culture (Fig.3-5). The
NF-M reactivity was in the form of arrays of filaments.
These cells were generally either low level A2B5(+) or
A2B5(-). Immunoreactivity was faint or negative with NF1
for filaments in these cells. The other monoclonal
antibodies specific for neurofilaments (DA3 and NN18) did
not react with the monolayers, and so no double-labelling
for GFAP and NF in these cells could be performed, because
NF-M and the commercial anti-GFAP antiserum were both from
rabbit. Anti-MSH which also reacts with the middle weight
neurofilament triplet protein (Shaw et al., 1985) reacted
with some neurons in monolayers similarly to NF-M, and also
reacted with the A2B5(+) non-neuronal cells in purified
cultures (not shown). The flat cells that contained
neurofilament reactivity were not found in unpurified
cultures.
Galactocerebroside Expression In Vitro
The development of oligodendrocytes in vitro was
explored using a monoclonal antibody specific for the
membrane marker galactocerebroside (GC; Ranscht et al.,
1982). The GC(+) cells that initially were detected in
dissociated and purified cells apparently developed into
oligodendrocytes in culture. Numerous GC(+)

95
oligodendrocytes with multiple processes were found in both
unpurified and purified long-term monolayers (Fig.3-6). As
would correlate with the initial enrichment of GC(+) cells
in purified cells, GC(+) oligodendrocytes were more
numerous in purified monolayers. The oligodendrocytes were
found primarily near clusters of neurons in both cultures,
although there were many exceptions to this. All GC(+)
cells that were observed were A2B5(-).
DNA Synthesis In Vitro
New DNA synthesis in the various cell types was
investigated by 3H-thymidine autoradiography combined with
immunohistochemistry. This was done in order to determine
if any patterns of new DNA synthesis could be discerned and
correlated with particular cell phenotypes that appeared in
culture. When freshly prepared cells were grown in medium
containing 3H-thymidine for 24 hours, a small number of
unpurified cells contained labelled nuclei (Table 3-1). In
2 out of 3 experiments, purified cells also contained
approximately equal small numbers of cells with labelled
nuclei. In one experiment, however, the percent labelled
nuclei of purified cells was several-fold that of the
unpurified cells. In all instances, the cells that
contained labelled nuclei were A2B5(-). Thus, the purified
cells that became A2B5(+) by 24 hours did not synthesize
new DNA in culture.

Fig.3-6. In vitro development of oligodendrocytes as
defined by immunostaining with antibodies against GC. Shown
are micrograph sets of cultures double-labelled for GC and
A2B5. GC(+) oligodendrocytes developed in cultures from
both unpurified (a) and purified (b) cells and were always
A2B5(-). Oligodendrocytes frequently extended processes
that were in close contact with A2B5(+) neurons (a). A
large aggregate of neurons is demarcated by the "a". In
purified cultures GC(+) oligodendrocytes were more numerous
than in unpurified cultures. Shown in (b) is an
oligodendrocyte in the same field as several A2B5(+) cells.
Oligodendrocytes tended not to be in close proximity to
A2B5( + ) nonneuronal cells. Bar, 50jim.

97

98
Table 3-1
3H-thyxnidine incorporation:
Day 12-13 cells®
Cells
Score
% Labelled Nuclei
Unpurified
1
11/315
3.5%
2
21/790
2.7%
mean=3.0+0.4
3
9/326
2.8%
Purified
1
33/722
4.6%
2
94/814
11.5%
mean=6.7+3.4
3
15/383
3.9%
“ 3H-thymidine incorporation into cells during 24 hours in
culture. The results of 3 experiments are shown. In 2 of
the 3 experiments (#1 and 3) both unpurified and purified
cells contained a small percentage (<5%) of labelled
nuclei. In 1 experiment (#2) the purified cells contained
a much higher percentage of labelled nuclei (4x) than did
unpurified cells. In all cases, labelled nuclei were in
A2B5(-) cells. Newly recruited A2B5(+) cells in purified
cultures therefore did not require new DNA synthesis for
this change in phenotype.

99
Monolayer cultures were exposed to medium that
contained 3H-thymidine for the life of the culture (»1^
weeks). In both unpurified and purified monolayers, the
majority of the A2B5(-) flat cells contained labelled
nuclei (Fig.3-7). In unpurified monolayers, A2B5(+) neurons
were not labelled. Similarly, GC( + ) oligodendrocytes did
not contain labelled nuclei. In purified monolayers,
neither GC(+) oligodendrocytes nor the A2B5(+) flat cells
contained labelled nuclei. Thus, the only cell phenotype
that synthesized new DNA in monolayer cultures was the
A2B5(-) flat cell. These results are presented in tabular
form in Table 3-2.
Discussion
A summary diagram of the results that includes cell
phenotypes, approximate percentages of cells, and new DNA
synthesis is presented in Fig. 3-8. The work presented
here is an investigation of the in vitro development of
dissociated and immunomagnetically purified 12 and 13 day
embryonic chick optic tectum cells. The cell purification
method that was used resulted in extremely purified
populations of cells that were negative for the target cell
monoclonal antibody A2B5. It was necessary to utilize a
double protease treatment prior to the disruption of tissue
into single cells with the day 12 and 13 tissue that was
used. This was not necessary when less developed day 7 or 8
tissue was used (Chapter 2) and is believed to reflect the

Fig.3-7. 3H-thymidine autoradiography combined with
immunohistochemistry of monolayer cultures after 8 days in
vitro. Cultures were continuously incubated in culture
medium containing label, and dipped slides were exposed for
5 days before developing. Micrograph sets are comprised of
(from left to right) phase contrast, brightfield, A2B5
immunostaining, and (a only) GC immunostaining. All black
is autoradiographic label in the unstained brightfield
micrographs, and nuclei appeared as large densely labelled
areas. In both unpurified and purified cultures labelled
nuclei were present only in A2B5(-) flat glia. A2B5(+)
neurons (a), A2B5(+) nonneuronal cells (b), and GC(+)
oligodendrocytes (a) did not contain labelled nuclei. Bar,
50nm.


102
Table 3-2
3H-thymidine incorporation:
Monolayers^
Cell Type
Marker
Labelled Nuclei
Flat
GFA/5A11
+
Neuron
A2B5/NF
-
Oligodendrocyte
GC
-
Abnormal
A2B5/GFA/
5A11/NF
-
a 3H-thymidine incorporation into monolayer cultures
exposed to continuous labelling. Of the cell phenotypes
described below, the only one that contained labelled
nuclei was the A2B5(-) flat cell. This demonstrates that
new DNA synthesis is not required for either recruitment
of A2B5(+) cells from A2B5(-) cells or their
differentiation. Similarly, cells that developed into
neurons or oligodendrocytes did not need new DNA
synthesis in culture.

Fig.3-8. Diagrammatic summary of the results which shows
immuno-phenotype, morphology, and 3H-thymidine
incorporation of cells in unpurified (left) and purified
(right) cultures. Time in culture is denoted on the left.
The numbers below the cells indicate percentages of either
unpurified (left) or purified (right) cells. "A" on a cell
indicates cell surface A2B5 binding, "5" indicates binding
of antibody 5A11, "GC" indicates galactocerebroside
reactivity, "N" indicates neurofilament reactivity, "G"
indicates glial fibrillary acidic protein reactivity, and
"GS" indicates glutamine synthetase reactivity. Labelled
nuclei are shown as intracellular stiplings.

Day 12-13
104

105
more elaborate processes and apparently tighter adhesion
between cells in the older tissue. This procedure resulted
in smooth surfaced round cells which were required for
optimum separations. An incomplete dissociation resulted in
considerable trapping of nontarget cells that otherwise
would have eluted.
Markers for Dissociated Cells
Day 12 marks the beginning of the growth and
maturation phase in chick optic tectum (LaVail and Cowan,
1971a). With day 12 and 13 optic tectum cells, two widely
used glial markers reacted with discrete populations of
cells. Glutamine synthetase antibodies (Linser and Moscona,
1979) reacted with a large percentage of dissociated cells
(=40%) and the majority of purified cells (=80%)
identifying them as astroglia. The fact that these cells
were A2B5(-) further supports the belief that A2B5 is
neuron-specific in chick optic tectum at this
developmental age. The oligodendrocyte marker
galactocerebroside (Ranscht et al., 1982) was also found to
react with a discrete population of dissociated cells (=6%)
and similarly these cells were all A2B5(-) and enriched for
in the purified population (=12%). Thus, the purified cells
could be positively identified as either astrocytes or
oligodendrocytes to a level of >90%.
Although more cells could be identified with glial
markers, antibodies against neurofilaments were less useful

106
as a marker for neurons than at day 7 or 8 (see Chapter 2).
Day 12 and 13 dissociated cells did not contain the easily
discernable rings of neurofilaments that were
characteristic of the day 7 and 8 cells. Most NF-M
reactivity was in the form of small blebs on one side of
the cell. These were presumably remnants of axons that were
sheared off during tissue dissociation. This difference in
immunoreactivity encountered in the older dissociated cells
possibly reflected more stable attachments between neurons
and thus an inability to retract their axons. Day 7 and 8
cells are in a phase of migration (LaVail and Cowan, 1971a)
and therefore might not have had the stable attachments of
the older tissue.
A2B5 Antigen and its Modulation
Purified cells were initially devoid of A2B5(+) cells.
This was demonstrated by both the immunofluorescent as well
as the immunoradiometric assays immeadiately following the
separations. The appearance of A2B5(+) cells in the
purified cell population in vitro was similar to that
reported with day 7 and 8 cells and was similarly confirmed
by the immunoradiometric assay. This suggests that the same
mechanism may be operating for the recruitment of
additional A2B5(+) cells with day 12 and 13 cells as was
operating with day 7 and 8 cells. A major difference,
though, is the extent to which this occurs. Here, the
percentage of A2B5(+) cells that appears after 24 hrs. in

107
culture (80%) far surpasses the levels in unpurified
cultures (45%). This suggests that the population of
reactive cells is not taken from a resting blast
population, as it is unlikely that more resting blast cells
would be present after the phases of proliferation and
migration within the tectum have ceased (LaVail and Cowan,
1971a,b) than during the proliferative phase.
Alternatively, blast cells may have been released more
easily from the tissue during dissociation than were
differentiating cells. More importantly, since >90% of the
purified cells were identified as definitive glia already
expressing differentiation markers the reactive population
must have come largely from definitive glia. This also
demonstrates that, in the chick system, A2B5 antigen on the
surfaces of glia was induced experimentally. And since the
GC(+) cells in the purified population did not become
A2B5(+) it appears that the reactive cells are wholly of
the astrocyte lineage. Lastly, it is worth noting that none
of the recruited A2B5(+) cells incorporated 3H-thymidine.
This clearly demonstrated that new DNA synthesis, and hence
mitosis, was not required for A2B5 antigen modulation.
Possible explanations of the mechanisms that may be
operating to effect this change in phenotype are given in
the Discussion section in Chapter II.

108
Development of Purified Cells In Vitro
The appearance of the A2B5(+) cells in purified long¬
term monolayer cultures made from day 12 and 13 cells
differed markedly from that of A2B5(+) cells from day 7 and
8 (Chapter 2). Instead of appearing to be neurons, the
majority of A2B5(+) cells exhibited a flattened glial and
then round morphology. They also expressed the glial
antigens GFAP and 5A11 in monolayer culture. These
combinations of antigen expression (A2B5 with 5A11 or GFAP)
were not found in unpurified cultures and are believed to
be abnormally induced by the purifications. Likewise, in
mouse cerebellar cultures other workers have experimentally
induced the expression of A2B5 antigen on the surfaces of
GFAP containing astrocytes as a result of neuronal
depletion by complement mediated lysis using an independent
neuronal marker (Nagata et al., 1986). Thus, the presence
of A2B5 antigen on vertebrate astrocytes which have been in
culture for even short periods of time (hours) should not
be considered characteristic of a normal phenotype without
reservation due to the apparent ease with which the antigen
can be modulated.
Another phenotype that was presumably artifactually
induced as a result of the purifications was flat cells
containing dense networks anti-neurofilament-M reactive
filaments. Some of these cells appeared to exhibit low
level surface A2B5 reactivity, but the majority appeared to

109
be A2B5(-). The polyclonal antibody that was used (anti-NF-
M) was well characterized (Bennett et al., 1984; Bennett
and Dilulo, 1985) and was found not to react with other
types of intermediate filaments on Western blots. In
addition, most of the other flat cells that contained glial
filaments as evidenced by reactivity with antibodies
against GFAP did not exhibit reactivity with NF-M. These
results indicate that the NF-M(+) reactivity in these cells
represents the presence of neurofilament reactivity and not
cross-reactivity with, say, GFAP filaments. The fact that
the NF-M(+) flat cells did not show reactivity with the
monoclonal antibody NF1 which is specific for the heavy
weight triplet protein was curious. This may be explained
by the results of others who have demonstrated that the
expression of the 3 triplet proteins in chicks is not
necessarily coordinate (Bennett et al., 1984; Dahl and
Bignami, 1986) and that the last triplet protein to be
expressed in phosphorylated form is often the heavy weight
one (Dahl and Bignami, 1986; Dahl et al., 1986). In any
case, the presence of neurofilament reactivity in cells
that did not acquire a neuronal morphology is unique.
The lack of new DNA synthesis in all cells that were
A2B5(+) was absolute. In purified cultures this negative
correlation indicates several things. The first is that the
recruited A2B5(+) cells at 1 day in culture did not require
new DNA synthesis for the acquisition of cell surface A2B5

110
antigen. Also, the A2B5(+) cells did not require new DNA
for their differentiation, whether it was into apparently
normal neurons or into cells that also expressed GFAP
and/or 5A11. In this respect these cells were like neurons
(post-mitotic) and not the majority of A2B5(-) flat cells
that were apparently glia and proliferated in culture. The
lack of labelled nuclei in oligodendrocytes also indicates
that the only cell population that may have proliferated in
culture was the astrocytes.
The possibility of the existence of "type 2"
astrocytes (Raff et al., 1983a) in the cultures must be
discussed. These cells express both GFAP and A2B5 antigen
and are thought to be a normal cell phenotype in vivo. To
my knoweldge, these cells have never been documented in the
chick system. Even in the rat system where they have been
reported they are peculiar to the white matter fiber
tracts, optic nerve and corpus callosum, and were not found
in the brain (Raff et al., 1983b). Furthermore, when
A2B5(+) cells were lysed via complement in that system, no
new A2B5(+) cells appeared in culture. GC(+) cells were
found to develop on time in cultures made from dissociated
rat brain cells 13-14 days before they normally appeared in
vivo (Abney et al., 1981). But, GC(+) oliodendrocytes did
not develop in cultures of purified A2B5(-) cells from day
13-15 cells (Abney et al., 1983). Thus, the rat system
appears to be different from chick in several respects.

Ill
However, in view of our results and others (Nagata et al.,
1986) which indicate that A2B5 antigen may be modulated by
loss of neuronal contact it seems possible that its
expression on the surface of type 2 astrocytes may also be
in response to neuronal deprivation. This would occur when
the optic nerve was dissected as only axons are present
there. By the time that the nerve is dissociated and the
cells are plated several hours would have passed which in
my system is ample time for many A2B5(-) cells to express
the antigen on their surfaces (data not shown). This
possibility would not be easy to rule out since I and
others (Schnitzer and Schachner, 1982; Dr. M.F.Marusich,
personal communication) have found that all neural cells,
including glia, contain intracellular epitopes for A2B5.
This last point precludes the usefulness of A2B5 as a
neuronal marker in tissue that was first fixed and
permeablized before incubation in A2B5, such as in tissue
sections.
Explanation of Abnormal Phenotypes
The phenotypes that appeared in purified cultures such
as A2B5(+)/5A11(+) round cells, NF-M(+) flat cells, and
A2B5(+)/GFAP(+) flat cells were clearly abnormal. These
markers that were co-localized in purified cultures were
always segregated into different cells in unpurified
cultures. The possibility that these abnormal phenotypes
were generated as a result of the procedure itself seems

112
unlikely. Cells trypsinized from the microspheres did not
produce them when placed in culture (not shown). The
peculiar cells were also absent from eluted cultures that
were not made from highly purified cells due to ineffective
antibody coating of the microspheres (not shown).
Separations of day 12-13 cells were also performed after
dissociation of tissue with papain (not shown). This also
resulted in A2B5(+) flat cells in the purified but not the
unpurified cultures. Similarly, day 7-8 cells were
separated after tissue dissociation with either papain or
dispase. This resulted in A2B5(+) cells in purified
cultures with neuronal morphologies, as was the result with
dissociation after trypsin alone (see Chapter II). Thus,
the presence of A2B5(+) flat cells in purified cultures is
not believed to be a result of using a different
dissociation protocol than with day 7-8 cells. However, the
different age embryonic cells may react differently to the
different proteases and, so, this possibility cannot be
ruled out completely.
The fact that adding back microsphere fraction cells
to the purified cells did not prevent the appearance of the
A2B5(+) cells with nonneuronal morphology may have been due
to their inability to communicate after separation. This
may have been due to an alteration of the cell surfaces
caused by the binding and/or removal of the cells from the
microspheres. Alternatively, it may have been a reflection

113
of a mechanism that was simply rapid and irreversible. It
is not known which of these or other possibilities had
occured.
Just exactly what these abnormal phenotypes that
appear in purified cultures represent is unclear. Several
possibilities exist. One possibility is that they represent
changes in astrocyte phenotypic characteristics which have
been shown to be modulated by neuronal contact. Many
instances of this type of phenomena have been documented in
the past several years such as modulation of marker
profiles (Fischer et al., 1986; Fisher, 1984; Holton and
Weston, 1982; Linser and Perkins, 1987a; Nagata et al.,
1986) and proliferation (Fischer et al., 1986; Hatten,
1987; Sobue and Pleasure, 1984).
Another possibility, however, is that these cells were
reacting to neuronal depletion by attempting to switch cell
types to become neurons. The range of mixed phenotypes
would represent a continuum of transition states from glia
to neurons. The phenotypes that exhibited more neuronal
character represented cells that were able to more fully,
but not totally, complete the attempted conversion from
glia to neurons. The cells that exhibited more glial
character represented those cells that had a more
restricted potential and less ability to convert to
neurons. This hypothesis is supported by the results of the
previous chapter in which A2B5(+) cells in purified

114
monolayer cultures from younger day 7 and 8 tissue appeared
to be normal neurons. While none of these possibilities can
be settled upon, the drastic change in phenotype in vitro
of purified A2B5(-) cells is further evidence of the
plasticity of embryonic glial cells and their capacity to
change in response to changes in their relationships with
neurons.

CHAPTER IV
CONCLUSIONS
A diagrammatic summary of the results of separations
of day 7-8 tectum cells is shown in Fig. 2-12. From the
results presented here it is concluded that A2B5 is an
accurate marker for neurons in long-term monolayer cultures
made from unpurified embryonic chick OT cells. The
immunomagnetic separation procedure used is an extremely
effective technique for the separation of embryonic neural
cell types based on cell surface antibody binding.
Immunomagnetic depletion of A2B5(+) cells, and possibly
neurons, from dissociated embryonic tectum cells resulted
in recruitment of new A2B5(+) cells from cells that
otherwise would not have been recruited.
Recruited cells from day 7 or 8 tissue compensated for
this depletion by presumably developing a purely neuronal
phenotype in vitro. The deleterious effect of polyornithine
■>
substrata on recruited cells suggests that the recruited
cells were not initially neuronal and that the acquisition
of the neuronal phenotype was not a single step process.
New DNA synthesis was not required for A2B5 antigen
modulation, nor for neuronal development. It is believed
that it is not possible to purify day 7 or 8 glioblasts
that remain as such unless the A2B5(-) cells that are
115

116
recruited to become neurons are a set population and a cell
surface marker can be found to remove them as well. Future
experiments should be aimed at this question.
One such experiment would be to perform a double
tandem separation. Between separations enough time would be
allowed so that recruitment of the A2B5(-) cells to A2B5(+)
ones could occur (hours). The A2B5(+) cells would then be
removed from this population and the purified A2B5(-) cells
would be analyzed for further recruitment. If no second
wave of recruitment occured, then it would seem that the
reactive population of cells was a discrete one. Then this
may very well allow the purification of immature glia
devoid of neurons that presumably would not produce any GS.
Neurons could then be added back to the glia to elicit GS
production on cue. This would obviously lead to a whole
range of further experiments aimed at the mechanism
controlling contact-mediated GS production.
A diagrammatic summary of the results obtained from
the separation of day 12-13 tectum cells is presented in
Fig.3-8. Purified A2B5(-) cells from this developmental age
are comprised almost entirely of identifiable glia with a
small number of identifiable neurons. A2B5(+) cells
appeared in purified cultures, but long-term monolayers
contained A2B5(+) cells with a nonneuronal morphology that
also express glial antigens GFAP and 5A11. Flat cells that
contained abundant neurofilament reactivity were also

117
present in purified cultures. Galactocerebroside(+)
oligodendrocytes were present in dissociated cells from
this age, were all A2B5(-), and apparently continued to
grow in culture. Only a small percentage of cells, all
A2B5(-), incorporated 3H-thymidine during the first 24
hours in culture of either unpurified or purified cells.
Continuous labelling of monolayer cultures resulted in
labelling of the majority of A2B5(-) flat cells, and of no
A2B5(+) neurons, no A2B5(+) nonneuronal cells, and no
oligodenrrocytes. Thus, new DNA synthesis was not required
for the recruitment of A2B5(+) cells or presumably their
differentiation in culture.
It is concluded that the phenotypic instability of
the recruited cells from day 12-13 tissue reflected an
unsuccessful attempt at developing a neuronal phenotype in
vitro. This most likely was due to a restricted
developmental potential of the reactive cells.

REFERENCES
Abney, E. R., Bartlett, P. P., and Raff, M. C. (1981).
Astrocytes, ependymal cells, and oligodendrocytes
develop on schedule in dissociated cell cultures of
embryonic rat brain. Dev. Biol■, 83, 301-310.
Abney, E. R., Williams, B. P., and Raff, M. C. (1983).
Tracing the development of oligodendrocytes from
precursor cells using monoclonal antibodies,
fluorescence-activated cell sorting, and cell culture.
Dev. Biol.. 100. 166-171.
Adler, R., Manthorpe, M., and Varón, S. (1979). Separation
of neuronal and nonneuronal cells in monolayer
cultures from chick embryo optic lobe. Dev. Biol., 69,
424-435.
Bennett, G. S., and DiLullo, C. (1985). Transient
expression of a neurofilament protein by replicating
neuroepithelial cells of the embryonic chick brain.
Dev. Biol. 107. 107-127.
Bennett, G. S., Tapscott, S. J., DiLullo, C., and Holtzer,
H. (1984). Differential binding of antibodies against
the neurofilament triplet proteins in different avian
neurons. Brain Res. 304, 291-302.
Blackburn, C. C., Swank-Hill, P., and Schnarr, R. L.
(1986). Gangliosides support neural retina cell
adhesion. J. Biol. Chem., 261, 2873-2881.
Campbell, G. LeM., Schachner, M., and Sharrow, S. 0.
(1977). Isolation of glial cell-enriched and-depleted
populations from mouse cerebellum by density gradient
centrifugation and electronic cell sorting. Brain Res.
127. 69-86.
Cowan, W. M., Martin, A. H., and Wenger, E. (1968). Mitotic
patterns in the optic tectum of the chick during
normal development and after early removal of the
optic vesicle. J. Exper. Zool., 169. 71-92.
Dahl, D., and Bignami, A. (1986). Neurofilament
phosphorylation in development. A sign of axonal
maturation? Exper. Cell Res. 162. 220-230.
Dahl, D., Crosby, C. J., Gardner, E. E., and Bignami, A.
(1986). Delayed phosphorylation of the largest neuro¬
filament protein in rat optic nerve development. J.
Neurosci. Res. 15, 513-519.
118

119
Debus, E., Weber, K., and Osborn, M. (1983). Monoclonal
antibodies specific for glial fibrillary acidic (GFA)
protein and for each of the neurofilament triplet
proteins. Differ. 25, 193-203.
Eisenbarth, G. S., Walsh, F. S., and Nirenberg, M. (1979).
Monoclonal antibody to a plasma membrane antigen of
neurons. Proc. Natl. Acad. Sci. USA, 76. 4913-4917.
Fischer, G., Keilhauer, G., and Schachner, M. (1986).
Comparison of astrocytic morphology, proliferation,
marker profile, and response to neurons in wild-type
and weaver mutant mouse cerebella in culture. Develop.
Brain Res., 28, 247-257.
Fisher, M. (1984). Neuron-glia interactions and glial
enzyme expression in mouse cerebellum. Adv. Exper.
Biol. Med.. 181, 203-213.
Folkman, J., and Moscona, A. (1978). Role of cell shape in
growth control. Nature (London), 273, 345-349.
Fujita, S. (1964). Analysis of neuron differentiation in
the central nervous system by tritiated thymidine
autoradiography. J. Compar. Neurol., 122, 311-327.
Gammon, C. M., and Ledeen, R. W. (1985). Evidence for the
presence of a ganglioside transfer protein in brain.
J. Neurochem., 44, 979-982.
Gee, A.P., Lee, C., Bruce, K., Janssen, W., Graham-Poole,
J., Pawlina, T., Pick, T., Harvey, W., Koch, P., and
Gross, S. (1987). Immunomagnetic purging and
autologous transplantation in stage D neuroblastoma.
Bone Marrow Transplant. . _2 (Suppl. 2), 89-93.
Grunwald, G. B., Fredman, P., Magnani, J. L., Ginsburg, V.,
and Nirenberg, M. (1985). Monoclonal antibody 18B8
detects gangliosides associated with neuronal
differentiation and synapse formation. Proc. Natl.
Acad. Sci. USA. 82. 4008-4012.
Hamberger, A., and Svennerholm, L. (1971). Composition of
gangliosides and phospholipids of neuronal and glial
cell enriched fractions. J. Neurochem., 18. 1821-1829.
Hatten, M. E. (1987). Neuronal inhibition of astroglial
proliferation is membrane mediated. J. Cell Biol.,
104. 1353-1360.

120
Hefti, F., Hartikka, J., and Frick, W. (1985). Gangliosides
alter morphology and growth of astrocytes and increase
the activity of choline acetyltransferase in cultures
of dissociated septal cells. J. Neurosci.. 5., 2086-
2094.
Holton, B., and Weston, J. A. (1982). Analysis of glial
cell differentiation in peripheral nervous tissue: II.
neurons promote S100 synthesis by purified glial
precursor cell populations. Develop. Biol.. 89. 72-81.
Honig, M. G., and Hume, R. I. (1986). Fluorescent
carbocyanine dyes allow living neurons of identified
origin to be studied in long-term cultures. J. Cell
Biol., 103, 171-187.
Hunter, W. M. (1978). Radioimmunoassay. In "Handbook of
Experimental Immunology. Volume 1: Immunochemistry"
(D. M. Weir, Ed.), pp. 14.1-14.40. Blackwell
Scientific Publications, Oxford, England.
Kasai, N., and Yu, R. (1983). The monoclonal antibody A2B5
is specific to cianglioside GQ1C. Brain Res. , 277, 155-
158.
Kelly, J. P., and Cowan, W. M. (1972). Studies on the
development of the chick optic tectum. III. Effects of
early eye removal. Brain Res.. 42, 263-288.
Kim, S. U. (1985). Antigen expression by glial cells grown
in culture. J. Neuroimmunol. , 8_, 255-282.
Kim, S. U., Moretto, G., Lee, V., and Yu, R. K. (1986).
Neuroimmunology of gangliosides in human neurons and
glial cells in culture. J. Neurosci. Res., 15, 303-
321.
Landa, C. A., Maccioni, H. J. F., and Caputto, R. (1979).
The site of synthesis of gangliosides in the chick
optic system. J. Neurochem. , 33., 825-838.
LaVail, J. H., and Cowan, W. M. (1971a). The development of
the chick optic tectum. I. Normal morphology and cyto-
architectonic development. Brain Res., 28, 391-419.
LaVail, J. H., and Cowan, W. M. (1971b). The development of
the chick optic tectum. II. Autoradiographic studies.
Brain Res., 28, 421-441.
Ledeen, R. (1985). Gangliosides of the neuron. Trends in
Neurosci. , 8., 169-174.

121
Ledig, M., Tholey, G., and Mandel, P. (1985). Factors
involved in expression of neuron-specific and non¬
neuronal enolase activity in developing chick brain
and in primary cultures of chick neurons. Develop.
Brain Res.. 21, 107-113.
Le Lievre, C. S., Schweizer, G. G., Ziller, C. M., and Le
Douarin, N. M. (1980). Restrictions of developmental
capabilities in neural crest cell derivatives as
tested by in vivo transplantation experiments. Dev.
Biol.. 77. 362-378.
Levitt, P., and Rakic, P. (1980). Immunoperoxidase
localization of glial fibrillary acidic protein in
radial glial cells and astrocytes of the developing
rhesus monkey brain. J. Comp. Neurol.. 193. 417-448.
Linser, P. J. (1985). Multiple marker analysis in the avian
optic tectum reveals three classes of neuroglia and
carbonic anhydrase-containing neurons. J. Neurosci.,
5(9) . 2388-2396).
Linser, P. J. (1987). Neuronal-glial interactions in retina
development. Amer. Zool., 27, 161-169.
Linser, P. J., and Moscona, A. A. (1979). Induction of
glutamine synthetase in embryonic neural retina:
Localization in Muller fibers and dependence on cell
interactions. Proc. Natl. Acad. Sci. USA, 76, 6476-
6480.
Linser, P. J., and Moscona, A. A. (1983). Hormonal
induction of glutamine synthetase in cultures of
embryonic retina cells: Requirement for neuron-glia
interactions. Dev. Biol. , 96., 529-534.
Linser, P. J., and Perkins, M. (1987a). Gliogenesis in the
embryonic avian optic tectum: Neuronal-glial
interactions influence astroglial phenotypic
characteristics. Dev. Brain Res., 31, 277-290.
Linser, P. J., and Perkins, M. (1987b). Regulatory aspects
of the in vitro development of retinal Muller glial
cells. Cell Differ., 20, 189-196.
Meier, D. H., Lagenaur, C., and Schachner, M. (1982).
Immunoselection of oligodendrocytes by magnetic beads.
I. Determination of antibody coupling parameters and
cell binding conditions. J. Neurosci. Res., 1_, 119-
134.

122
Meier, D., and Schachner, M. (1982). Immunoselection of
oligodendrocytes by magnetic beads. II. In vitro
maintenance of immunoselected oligodendrocytes. J.
Neurosci. Res., 1_, 135-145.
Nagata, I., Keilhauer, G., and Schachner, M. (1986).
Neuronal influence on antigenic marker profile, cell
shape, and proliferation of cultured astrocytes
obtained by microdissection of distinct layers from
early postnatal mouse cerebellum. Dev. Brain Res., 24,
217-232.
Philip, T., Favrot, M., and Philip, I. (1987). Complement
and magnetic purging: A comparison in the Burkitt
lymphoma model. Bone Marrow Transplant.. 2 (Suppl. 2),
78-80.
Politi, L., and Adler, R. (1987). Selective destruction of
photoreceptor cells by anti-opsin antibodies. Invest.
Ophthalmol. Vis. Sci. 28. 118-125.
Puelles, L., and Bendala, M. C. (1978). Differentiation of
neuroblasts in the chick optic tectum up to eight days
of incubation: A Golgi study. Neurosci. . 3., 307-325.
Raff, M. C., Abney, E. R., Cohen, J., Lindsay, R., and
Noble, M. (1983b). Two types of astrocytes in cultures
of developing rat white matter: Differences in
morphology, surface gangliosides, and growth
characteristics. J. Neurosci., 3., 1289-1300.
Raff, M. C., Miller, R. H., and Noble, M. (1983a). A glial
progenitor cell that develops in vitro into an
astrocyte or an oligodendrocyte depending on the
culture medium. Nature (London), 303, 390-396.
Ranscht, B., Clapshaw, P. A., Price, J., Noble, M., and
Seifert, W. (1982). Development of oligodendrocytes
and Schwann cells studied with a monoclonal antibody
against galactocerebroside. Proc. Natl. Acad. Sci.
USA. 79, 2709-2713.
Rosner, H. (1980). Ganglioside changes in the chicken optic
lobes and cerebrum during embryonic development.
Transient occurrence of "novel" multisialo
-gangliosides. Wilhelm Roux's Archives, 188, 205-213.
Schnitzer, J., and Schachner, M. (1982). Cell type
specificity of a neural cell surface antigen
recognized by the monoclonal antibody A2B5. Cell Tiss.
Res., 224, 625-636.

123
Shaw, G., Fischer, S., and Weber, K. (1985). Anti-MSH and
neurofilament M-protein share a continuous epitope but
not extended sequences. FEBS Letters, 181, 343-346.
Shaw, G., Osborn, M., and Weber, K. (1986). Reactivity of a
panel of neurofilament antibodies on phosphorylated
and dephosphorylated neurofilaments. Eur. J. Cell
Biol. 42, 1-9.
Sheffield, J. B., Pressman, D., and Lynch, M. (1980). Cells
isolated from the embryonic neural retina differ in
behavior in vitro and membrane structure. Science,
209. 1043-1045.
Sobue, G., and Pleasure, D. (1984). Astroglial
proliferation and phenotype are modulated by neuronal
plasma membrane. Brain Res.. 324, 175-179.
Treleaven, J. G., Ugelstad, J., Philip, T, Gibson, F. M.,
Rembaum, A., Caine, C. D., and Kemshead, J. T. (1984).
Removal of neuroblastoma cells from bone marrow with
monoclonal antibodies conjugated to magnetic
microspheres. Lancet, JL, 70-73.
Turner, D. L., and Cepko, C. L. (1987). A common progenitor
for neurons and glia persists in rat retina late in
development. Nature, 328. 131-136.
Wu, D. K., Scully, S., and de Vellis, J. (1988). Induction
of glutamine synthetase in rat astrocytes by co¬
cultivation with embryonic chick neurons. J.
Neurochem., 50. 929-935.
Wysocki, L. J., and Sato, V. L. (1978). "Panning" for
lymphocytes: A method for cell selection. Proc. Natl.
Acad. Sci. USA. 75. 2844-2848.

BIOGRAPHICAL SKETCH
Deni S. Galileo was born in Pittsburgh, Pennsylvania,
on November 23, 1961. He attended elementary school there
until moving to Meadville, Pennsylvania in 1968. He
attended the public schools there and graduated from
Meadville Area Senior High School in 1979. Deni attended
and finally graduated from New College, the honors college
of the University of South Florida in Sarasota, in 1983
after being academically dismissed after his first term.
Deni obtained a Bachelor of Arts degree in cell and
developmental biology with his bachelor's thesis entitled
"Investigations on primary mesenchyme cell migration in the
sea urchin Lvtechinus varieqatus; a model system for
studying selective cellular adhesion in vivo" under Dr.
John B. Morrill. He entered graduate school the following
year at the Department of Anatomy and Cell Biology,
University of Florida to further his education in
developmental and cell biology with Dr. Paul J. Linser at
the Whitney Laboratory. He obtained the degree of Doctor of
Philosophy in basic medical sciences from the College of
Medicine in August, 1988. Deni has accepted a postdoctoral
research position at Washington University School of
Medicine in St. Louis, Missouri, to continue his studies of
brain development in chicks using recombinant retroviral
vectors with Dr. Joshua Sanes.
124

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Professor of Anatomy
and Cell Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Carl M. Feldherr
Professor of Anatomy and Cell
Biology
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Francis C. Davis, Jp
Associate Professor of
Microbiology and Cell Science
I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Associate Professor of Anatomy
and Cell Biology

This dissertation was submitted to the Graduate Faculty of
the College of Medicine and to the Graduate School and was
accepted as partial fulfillment of the requirements for the

UNIVERSITY OF FLORIDA
3 1262 08554 4020