University of Southern California Dissertations and Theses

Studies On Antigens Associated With Wilm'S Tumor (Nephroblastoma)

(USC Thesis Other)

Studies On Antigens Associated With Wilm'S Tumor (Nephroblastoma)

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content STUDIES ON ANTIGENS ASSOCIATED WITH
WILM'S TUMOR (NEPHROBLASTOMA)
by
Kim Stuart W ise
A D issertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In P artial Fulfillm ent of the
Requirem ents for the D egree
DOCTOR OF PHILOSOPHY
(Biochem istry)
June 1973
INFORMATION TO USERS
This material was produced from a microfilm copy of the original document. While
the most advanced technological means to photograph and reproduce this document
have been used, the quality is heavily dependent upon the quality of the original
submitted.
The following explanation of techniques is provided to help you understand
markings or patterns which may appear on this reproduction.
1. The sign or "target" for pages apparently lacking from the document
photographed is "Missing Page(s)". If it was possible to obtain the missing
page(s) or section, they are spliced into the film along with adjacent pages.
This may have necessitated cutting thru an image and duplicating adjacent
pages to insure you complete continuity.
2. When an image on the film is obliterated with a large round black mark, it
is an indication that the photographer suspected that the copy may have
moved during exposure and thus cause a blurred image. You will find a
good image of the page in the adjacent frame.
3. When a map, drawing or chart, etc., was part of the material being
photographed the photographer followed a definite method in
"sectioning" the material. It is customary to begin photoing at the upper
left hand corner of a large sheet and to continue photoing from left to
right in equal sections with a small overlap. If necessary, sectioning is
continued again — beginning below the first row and continuing on until
complete.
4. The majority of users indicate that the textual content is of greatest value,
however, a somewhat higher quality reproduction could be made from
"photographs" if essential to the understanding of die dissertation. Silver
prints of "photographs" may be ordered at additional charge by writing
the Order Department, giving the catalog number, title, author and
specific pages you wish reproduced.
5. PLEASE NOTE: Some pages may have indistinct print. Filmed as
received.
Xerox University Microfilms
300 North Zeeb Road
Ann Arbor, Michigan 48106
il
I
73-31,686
WISE, Kim Stuart, 194-5-
STUDIES ON ANTIGENS ASSOCIATED WITH WILM'S
TUMOR (NEPHROBLASTOMA).
University of Southern California, Ph.D., 1973
Biochemistry
; University Microfilms, A X E R O X Company, Ann Arbor, Michigan
THIS DISSERTATION HAS BEEN MICROFILMED EXACTLY AS RECEIVED.
UNIVERSITY O F SO U TH ER N CALIFORNIA
TH E GRADUATE SCHOOL
UNIVERSITY PARK
LOS ANGELES, C A LIFORN IA 9 0 0 0 7
This dissertation, written by
.............................
under the direction of his..... Dissertation Com
m ittee, and approved by all its members, has
been presented to and accepted by The Graduate
School, in partial fulfillm ent of requirements of
the degree of
D O C T O R O F P H IL O S O P H Y
Dean
June 1973
D ate..........................
DISSERTATION COMMITTEE
......
Chairman
D E D IC A T IO N
To Dr. Lucien Bavetta, for his
guidance at an important point in my life.
To my w ife, Cheri, for her warm
encouragement during these years.
ii
A C K N O W L E D G M E N T S
I appreciate deeply these people who helped make
this work possible:
D rs. Samuel A llerton and John B eierle for
their continuous support in all parts of this work.
D rs. Gary Trump and P. Roy-Burm an for
many useful d iscu ssion s.
D rs. D arleen Pow ars and Robert M cA llister
for biosam ples and clinical-laboratory correlations.
D rs. Kalindi Desmuth and M arcel Nimni,
for the amino acid analyses.
Dr. Jose Miguel Jim enez, for helpful
assista n ce in the Laboratory.
M rs. Marilyn Cheung, for her help and
excellen ce in preparing this m anuscript.
Finally, all those I have known during these
yea rs, whose id eas, ethics and humor have made
life very rich.
iii
[
TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS................................................................................. iii
LIST OF FIG U R E S....................................................................................... v i
Chapter
I. INTRODUCTION.......................................................................... 1
Background
Statement of the Problem
II. MATERIALS AND METHODS............................................... 17
A. IMMUNOLOGICAL PROCEDURES
Immuniz ation
Preparation of A ntisera
Preparation of Immunoglobulins
Absorption of A ntisera
Immunodiffusion
Im m unoelectr ophor e sis
Immunological M aterials
B. CHEMICAL METHODS
Protein Determ ination
Uronic Acid Determ ination
Sensitivity of Antigen to
Enzym atic Degradation
Amino Acid A nalysis
C. PITY SIC AL METHODS
M olecular Sieve Chromatography
DEAE -Sephadex Chromatography
A crylam ide Gel E lectrop h oresis
Iso electric Focusing in A crylam ide G els
A nalytical Ultracentrifugation
iv
C h a p te r P ag e
D. PREPARATION OF EXTRACTS FROM
TISSUES AND CULTURED CELLS
C ell Cultures
T issu e E xtracts
W ater Soluble Antigens
Ammonium Sulfate Fractionation
of Newborn Calf Serum
Nature of the Immunoglobulins Employed
D em onstration of an Abnormal Antigen in
W ilm 's Tumor Extract
Exploration for the Antigen in Other
Human M aterials
D etection of the Antigen on Cultured W ilm 's
Tumor C ells, in Newborn Calf Serum and
in Fetuin
Nature of the Antigen
M olecular Sieve Chromatography
Sensitivity of the Antigen to Enzym atic
Degradation
Amino Acid A nalysis of Fetuin and E lectro -
phoretically Purified Antigen from the
PBS-EDTA Extract of W ilm 's Tum ors
D eterm ination of Uronic Acid and Sialic Acid
III. RESULTS 35
IV. DISCUSSION 111
P roperties of the Antigen Detected
Distribution of Antigen
P resen ce of the Antigen In Vitro
V. SUMMARY 139
APPENDICES 142
BIBLIOGRAPHY 152
LIST OF FIGURES
Figure Page
1. Sedimentation P roperties of Immunoglobulin
F r a ctio n ................................................................................... 36
2. Immunodiffusion Reaction of R 5-19 with Pooled
W ilm 's Tumor Extract and Human S eru m ................. 39
3. Immunodiffusion Reaction of R5 Immunoglobulin
Fraction (Absorbed) with Norm al Human Kidney
Extract, Human Plasm a, and W ilm 's Tumor
E xtract...................................................................................... 42
4. A nalysis for Antigen in Extracts of Human
Fetal T issu e, W ilm 's Tumors and Non-Related
T u m o r s ................................................................................... 47
5. A nalysis for Antigen in Serum of Patients with
W ilm 's and Other Tumors; Human Fetal and
M aternal Serum; and D escription of Patients
Donating A ntigen-P ositive W ilm 's Tumor Pool . . 50
6. Immunodiffusion C ross Reaction of W ilm 's
Tumor Extract and Extract of Cultured W ilm 's
Tumor C e l l s ........................................................................... 53
7. Immunodiffusion C ross Reaction of W ilm 's
Tumor Extract and Extract of Cultured Norm al
Human Fetal Kidney C e l l s .............................................. 56
8. C ross Reaction of Antigens from Calf Serum,
Fetuin, and from Pooled W ilm 's Tumor and
T issu e Culture E xtracts.................................................... 59
9. Im m unoelectrophoresis of R>oled W ilm 's Tumor
Extract and Newborn Calf Serum ............................... 66
10. Im m unoelectrophoresis of Pooled W ilm 's Tumor
Extract and C om m ercial F e t u in .................................. 68
vi
F ig u re P a g e j
11. Im m unoelectrophoresis of Newborn Calf Serum
and Pooled W ilm 's Tumor Extract Against Anti -
Bovine Serum Antiserum ..................................................... 70
12. A crylam ide Gel E lectrophoresis of W ilm 's
Tumor Extract and Com m ercial F e t u in ................. 76
13. Iso electric Focusing Pattern of Pooled W ilm 's
Tumor Extract and C om m ercial F e t u in ................. 80
14. Isoelectric Distribution of Antigen from W ilm 's
Tumor Extract and C om m ercial F e t u in ................... 82
15. Iso electric Focusing of Antigen from Newborn
Calf S e r u m ............................................................................. 86
16. G-75 Sephadex Chromatography of Antigen
M aterial in Pooled W ilm 's Tumor Extract and
in Fetuin ................................................................ 89 |
17. Enzym atic Degradation of Antigen from W ilm 's
Tumor E xtract-D irect A s s a y ........................................ 92
18. Lack of Interfering Enzym atic A ctivity in Antigen
D ig e s ts .................................................................................... 95
19. Enzym atic Degradation of Antigen from W ilm's ;
Tumor Extract-Inhibition A s s a y .................................. 97
20. Amino Acid A nalysis of Fetuin and E lectro-
phoretically-Purified Antigen from Pooled W ilm's
Tumor Extract .................................................................... 102
21. Comparison of Experim ental Amino Acid A nalysis !
of Antigen from W ilm 's Tumor Extract; and
Reported Com positions of F e t u i n ............................... 105
vii
CHAPTER I
INTRODUCTION
Background
N eoplasia has been defined and cla ssified largely on the b asis
of growth ch aracteristics and morphology (1). These characteristic
patterns are distinguishable from those in a norm al organism when
compared at an equivalent tim e in development, and at an equivalent
anatomical position. Though the description of neoplasia varies
I
j
| widely, two unifying features are common to most: an elevated
j
I
| growth rate, and an invasive growth pattern. In neoplasia term ed
malignant, this invasive property is observed not only at the im m e
diate site of neoplastic origin, but at a variety of sites widely d istr i-
I
i
I buted throughout the organism , into which neoplastic lo ci have
t
I
dissem inated and rooted. A fundamental question regarding
neoplasia is: "How do these c ell populations escape the stringent
j controls im posed on norm al c e lls during the development of a com
plex organism ; and what properties do th ese c e lls have that allow
them to survive and flourish outside of these bounds ?" Answering
this question n ecessita tes first, a greater understanding of the
i
j control and differentiation of normal cell populations; and second, a
m ore p recise description of the neoplastic cell, allowing dissection
jof its m orphologic and kinetic properties into experim entally oper-
i
! able parts. 1
2
Of the numerous approaches to th ese problem s, one has
recently gained great impetus: the im m unological response to
neoplasia. This is in part because immunology now provides an
i
effective m eans of studying structurally, som e of the cell com
ponents possibly involved in neoplastic transform ation. A lso,
im m unological m echanism s, capable of controlling erratic cell
j populations in a host, do exist; and are amenable to study.
j
I The relationship between neoplasia and im m unological
i
response may, however, be m ore fundamentally related. Burnett
(2) points out that the appearance of adaptive immunity generally
i
i coincides with the evolution of verteb rates, and that m ost in verte
brate form s lack the ch aracteristics of this com plete system : a
thymus and prim ary lymphoid system ; circulating lym phocytes
and immunoglobulins; a capacity to reject tissu e hom ografts
i (isografts); a capacity for delayed hypersensitivity to antigen; and
I
an ability to form specific antibody to antigenic stim ulus. While
evidence ex ists dem onstrating parts of this com posite system in
invertebrates (3), this generalization is probably sound. Ullman (4)
com piles evidence that neoplastic d isea se, with very rare excep-
i
I
I
i tions, is also restricted to vertebrates. Burnett speculates that
| th ese two phenomena resulted from the em ergence of a genome which
!
| allowed random changes in a portion of its products. Evolutionary
advantages associated with th ese changes are discussed by Burnett J
3
(2). Most germ ane to this d iscu ssion , however; th ese changes
explain the appearance of d iv erse histocom patability differences
within sp ecies and the presen ce of a d iverse immune system able
to discrim inate and react to th ese differences. Furtherm ore, this
immune surveillance system could ea sily evolve into an effective
check for aberrant c ell types arising not from external sou rces,
but from the organism 's own changing cellular phenotype. It may
in this sen se, be a supplem entary control m echanism derived to
make up for the inadequacies of preexisting controls. This may
very w ell have been the m ain evolutionary advantage of the immune
system . The immune resp on se to externally introduced m aterials
m ay have been an incidental side benefit, sin ce prior to the
appearance of an immune system , m echanism s had evolved to deal
very effectively with harm ful foreign m aterial (5, 6).
That neoplastic growth in man may be partially controlled
by immune m echanism s is suggested by a number of clinical
observations: 1) the incidence of neoplastic foci (not detected
clinically) in random b iop sies is much greater than the incidence of
neoplasia presenting clin ically (7, 8); 2) the incidence of neoplasia
is much greater in periods of life where the immune system is
known to be deficient or not fully developed (9, 10); 3) long term
use of im m unosuppressive drugs lead s to an in crease in neoplastic
d isea se (11,12), and when adm inistration of th ese drugs c e a se s.
4
c a se s of tumor rem issio n have been reported; 4) a high incidence
of neoplasia is found in patients having various, sp ecific d eficien cies
in their immune response (13).
In addition to clin ical evidence for immune control of
neoplasia, som e anim al sy stem s have provided a sounder b a sis for
th is conclusion. For instance, neonatal thym ectom y facilitates
som e form s of m alignancy in rodents (14), im plying a norm al
capability for control by the thym us-dependent immune system .
While it is very difficult to rule out other m echanism s controlling
neoplasia such as hormonal sy stem s (15), allogenic inhibition (16)
j or other as yet unknown resp on ses; there is good circum stantial
evidence that the immune system plays a significant role in con-
!
I trolling this p rocess.
The im m unological response of a host to neoplastic growth
can be extrem ely varied and can be categorized into three types:
lack of resp on se, effective anti-tum or response, and tum or-
enhancing response.
Studies by Prehn (17) show that "spontaneous" tum ors
arising in m ouse embryo c e lls which are grown in an im m uno-
lo g ica lly deficient environment (peritoneal diffusion cham bers)
generally lacked im m unogenicity when transferred into syngeneic
m ice which w ere subsequently challenged with tumor c e lls. This
| w as in marked contrast to m ethylcolanthrene (MCA)-induced tum ors
5
which w ere highly im m unogenic. F ir st, this indicates a lack
of im m unogenicity in th ese spontaneous tum ors as opposed to
experim entally induced tum ors, and m ay indicate that etiology of
a given neoplasia is important in determ ining the hosts resp on se.
Second, it argues against the selection , by im m unological d estru c
tion, of non-im m unogenic neoplasia, sin ce all th ese tumor o r ig i
nated without contacting im m unocytes (and thus w ere insulated
from the h ost's immune system ). T his is germ ane to the question
of how neoplastic c e lls escape im m unological control, and suggests
that it is not due to se le ctiv e p ressu res im posed by preferential
| im m unological destruction of certain c e lls.
E ffective anti-tum or resp o n ses have been dem onstrated
I
clin ically by Klein (18); showing com plete rem issio n of basal c e ll
carcinom a of the skin using n on -sp ecific stim ulation of the immune
apparatus with 2, 3, 5 -tri-eth y len e-im in o (1, 4)-benzoquinone. More
extensive work has been pursued by Parr (19) dem onstrating r e s is t
ance to tumor challenge following the injection of B acillu s Colum et-
Guerin (BCG) and irradiated leukem ia c e lls in m ice. Sim ilar
studies by Zbar et al. (20) show this phenomenon in m ethyl -
colanthrene-induced tum ors in guinea pigs. A lso, Lindenmann (21)
has dem onstrated strong im m unogenicity in E hrlich and K rebs-2
a sc ites tum ors following infection of th ese c e lls with influenza
virus.
6
Tum or-enhancing resp on ses have been found in vivo by
Stjernward (22) who showed that m ice from which a MCA-induced
tumor w as rem oved w ere m ore susceptible to autochthonous
tumor c ells than w ere control m ice. This w as a tu m or-sp ecific
susceptibility and w as suggested to resu lt from tem porary immune
p aralysis. Interestingly, m ice having their prim ary tumor sim i
la rly rem oved, and im m unized with irradiated c e lls from the sam e
tum or, showed sp ecific resista n ce to the autochthonous tumor
challenge.
While these studies indicate the d iversity of host resp on ses
I
! to neoplasia, they also suggest a number of questions of paramount
im portance in understanding this immune response in vivo.
1. How do aberrant c e lls contact the immune
m echanism s, and does the route of contact influence the type of
response obtained ?
2. Is the im m unogenicity of neoplastic c e lls in a
host governed only by changes indigenous to the aberrant c e lls,
or is it influenced by a variety of host resp on ses which can
either allow or lim it recognition of the c e lls by the immune
system ?
3. How do growth kinetics of neoplastic c e lls
influence the ability of the immune system to respond effectively ?
7
B ecause th ese questions are com plex, and m ust be dealt
with sim ultaneously in studies of tum or-host relationships, in vitro
system s have evolved with the hope of isolating a part of the h ost's
response which can be m ore ea sily studied.
One resu lt of this approach is the discovery of blocking
I
antibody (23) which inhibits sen sitized lym phocytes from neuro
blastom a patients from in vitro destruction of cultured neuro
blastom a c e lls, hence allowing tumor cell growth. This antibody,
found in som e patients serum , is apparently substituted in others
j by "unblocking antibody" which inhibits the effect of the enhancing
i
j antibody. The provocative resu lts of such in vitro system s w ill
perhaps be very useful in understanding the relationships between a
host's immune system and neoplasia. They are not a p riori sim
plifying, as can be seen, and do present a constant risk of o v er-
interpretation when applied to the actual host-tum or situation.
A highly relevant exam ple of this is the generality (24) that
cell-m ediated immune resp on ses are generally effective in anti-
tumor reactions, w hereas humoral resp on ses are expected to
enhance tumor growth. This concept is not tenable when dealing
with som e in vivo situations. Alexander (25) presents evidence
that rats with prim ary, non-m etastatic tum ors, when drained of
their lymph and reconstituted with the rem oved lymph ce lls,
undergo rapid m etastasis of the tum ors and soon die. In this
8 i
: system , artificial induction of m etastasis can be achieved by j
intra-venous injection of tumor c e lls, w hereas subcutaneous injection ’
of the sam e c e lls fa ils to produce th is effect. This induced
m etastasis is inhibited by p assive tran sfer of the drained lym phatic
fluid containing antibodies. In th is in vivo situation, then there is
strong dependence of the humoral response on control of neoplastic
: ce lls. Such observations m ust be kept in mind when d issectin g the
immune response for apparently m ore critical analysis.
The outcome and m echanism of the encounter between a
h ost's im m unological system and a developing neoplastic c e ll is not
understood at present, nor is enough known about this encounter to
manipulate predictably the h ost's response toward an effective
control of neoplasia; this is probably the ultim ate m edical goal of
studying tumor immunity.
An im plicitly related facet of tumor immunology is the exam
ination of neoplastic c e lls to determ ine changes which might endow
th ese c e lls with abnormal growth ch aracteristics. These changes
hopefully would be m anifest as antigenic alterations which could
be analysed by serological techniques, characterized physically and
chem ically, and shown to have grow th-regulating function. This
approach is not new (26). Much ea rlier research has been d isr e
garded however, because the apparent tumor antigens observed w ere
very probably due to genetically determ ined allogenic differences
9
between donor and recipient of tumor grafts. More recently work
such as that of F oley (27), and Prehn and Main (28), dem onstrating
tumor antigens in syngeneic anim als, has rekindled this field of
i
| ;
! research and has lead to som e understanding of tumor antigens
i
as w ell as to som e new quandaries.
Tumor c e lls experim entally induced by chem ical
carcinogens, DNA, and RNA v iru ses have all been shown to exp ress
new surface antigens accompanying m orphological and growth changes
| characteristic of neoplasia (29). While other internal antigens are
i
I also expressed in som e of th ese c e lls, the surface antigens are
i
j
j particularly interesting in this light since 1) specific surface com -
I ponents of neoplastic c e lls interacting with plant agglutinins (30),
! cytochalasin B (31), and cyclic AMP (32) are known to influence
|
!
! strongly growth ch aracteristics of these cells; and 2) surface
|
structures are generally the only components a ccessib le to immune
defense m echanism s (33). Identification of these antigens is prom
isin g, then, in studying the grow th-affecting surface structures of
c e lls, and in understanding the immune response to these c e lls
in vivo.
A number of questions face this entire approach of studying
neoplastic cells:
1. Do all neoplastic c e lls exp ress new antigenic
structures on their surface ? Are th ese antigens causally related to
10
the neoplastic state, or are they m erely incidental m arkers of
som e m ore fundamental event(s) or change(s) ?
j 2. What is the diversity and specificity of these
antigens ? Should they be expected to share common structural
features, endowing all neoplastic c e lls with the common ch aracter
is tic s of rapid, invasive growth?
3. How do these c e lls acquire the surface antigens
associated with neoplasia ? Is it by cellular synthesis resulting
from selectiv e exp ression of the cell genom e, or can adsorption of
substances from other parts of the host resu lt in altered antigenic
j
! structures on the neoplastic c ell surface ?
4. Are these antigens new, occurring only in the
neoplastic state, or are they norm al developm ental antigens d is
placed in tim e or place which appear abnormal in a neoplastic
situation ?
Some directions of em erging answ ers to these questions
are becom ing apparent:
1. There is a seriou s question as to the presence of
new antigens on all neoplastic c e lls. While many have been found
(34), it is prem ature to suggest that this is a universal occurrence
(35). This applies esp ecially to spontaneous tum ors, which are
relatively little studied, and whose etiology is generally unknown
(36); and particularly to human neoplasia where th ese problem s
11
are compounded with a lim ited capacity for im m unological exp eri
mentation.
An indirect lin e of reasoning suggesting a p ossib le func
tional relationship between tu m or-associated antigens and neoplastic
growth is provied by two p ieces of evidence. 1) Growth of
neoplastic c e lls in vitro can be regulated by the p resen ce of antibody
in the medium directed to c e ll surface components (37). 2) C ell
aggregation factors (potentially influencing the growth ch aracter-
I
i
; is tic s of c ells) have been shown to be antigenic (38), and antibodies
|
| to th ese factors can inhibit aggregation. Furtherm ore, agglutinins
i
■ affecting neoplastic c e ll growth (30, 39) can be inhibited by anti
bodies to th ese substances. If analogs of th ese m aterials are
present on neoplastic c e lls in vivo and are antigenic, this could
I
I
| offer an effective m eans of im m unological control.
2. The sp ecificity of tu m or-associated antigens has
been in a constant state of reevaluation. Evidence suggests that
virus-induced tum ors share antigenic structures indicative of the
viru s group used for transform ation; while chem ically-induced
tum ors have individually unique antigenic structure (29, 34). This
has recently been challenged by the report that certain tum ors pro
duced by c la ss e s of chem ical carcinogens share common antigens
(40).
12
There is a degree of uncertainty about the uniqueness of a
given antigen at a particular stage of experim ental development,
usually due to the sen sitivity lim its of the a ssay m easuring the
antigen. A lso disturbing is the observed d iversity of tum or-
associated antigens; this is difficult to recon cile when attempting
to relate th ese antigens functionally. This feature does, however,
offer a potentially discrim inating m eans of diagnosing and perhaps
treating various neoplastic d isea ses.
i
!
| 3. C ellular synthesis of antigens associated with
i
| the neoplastic state is indicated by in vitro studies w here induced
| neoplastic transform ation is accompanied by the appearance of a
| new antigen (34, 41). Cellular m echanism s m ay also m ediate the
| disappearance of antigens from transform ed c e lls (42). While this
| may represent a significant m eans of obtaining new antigenic
structure in vivo, as seem s to be the ca se with human carcin o-
em bryonic antigen (43); this is not the only possibility. Many types
of norm al c e lls can obtain surface coating antigens from their
environment, including erythrocytes, sperm c e lls and lymphatic
c e lls (44). In addition, environm ental factors such as horm ones
(44), antibodies (37), and shifts in populations of c e lls with a given
antigenic makeup (45) can influence the exp ression of surface
antigens. N eoplastic c e lls would be expected to respond, as w ell,
to the in vivo environment. Substances synthesized by the liv er in
13 |
hepatoma are known to differ antigenically from norm al liv er
proteins, and are known to travel through the humoral system (46).
This m aterial, or other substances produced by the host in response
to neoplastic growth may have the capacity to bind c e lls and alter
their normal antigenic structure. The hormonal environment (47),
the presence of norm al m asking glycoproteins (48) and the abnormal
presence of enzym es (49) have all been im plicated in altering
antigens on neoplastic c e lls.
4. Finally, much attention has been given to the
concept that many antigens associated with neoplasia are m erely
norm al antigens of development which have been tim e-displaced.
Alexander has recently review ed th ese studies (33). Included in
this cla ss of antigen are 1) an a-fetal protein obtained from m ice
with chem ically induced hepatom as, from em bryonic m ouse serum ,
from serum of patients with prim ary hepatoma and from human
fetal liv er (46); 2) an ct-ferro-protein found in human fetus and in
children with a variety of tum ors (50); 3) a fetal sulphoglycoprott-in
present in fetal gut and gastric ju ices of patients with gastric
neoplasia (51); this m aterial disappears after birth and can
reappear in sen escen ce (52); 4) a fetal alkaline phosphatase found
in sera of patients with a variety of neoplasias, and in serum from
pregnant women (53); 5) an antigen from many human tum ors
(with gamma electrophoretic mobility) which cr o ss reacts with
14
fetal serum from many sp ecies including human and bovine (54);
and 6) a carcinoem bryonic antigen of colonic tum ors which c r o s s-
reacts with fetal gut, liver and pancreas up to six months gestation
(55). It apparently reacts in som e system s (56) with sera from
I
patients with non-neoplastic d isea se, and with carcinom a of non
digestive tract origin. R ecently, low le v e ls have been found in
norm al adult colon (57). Sim ilarly trace amounts of 0 1 -fetoprotein
have been reported in adult sera (58).
The value of such "onco-fetal" antigens (33) in diagnosis
! and potential treatm ent of neoplasia is not dim inished by their c r o s s-
| reactivity with fetal or other substances, as long as significant
| differential amounts can be observed in the neoplastic state. These
I
| antigens may indeed answer som e questions about differentiation
! m echanism s connected with neoplastic transform ation. These
1
!
i antigens do, however, challenge the concept of "tum or-specific"
j
i
or "tumor unique" antigens, and warn against prem ature judgement
of an apparently new antigenic structure found in neoplastic
m aterial.
Statement of the Problem
T hese unanswered questions have inspired the present
study. I have chosen to explore the antigenic structure of m aterials
I
associated with W ilm 's tumor, or human nephroblastoma, with hope
of contributing to the further understanding of the changes
indicative, and possibly causative, of this neoplastic state.
W ilm 's tumor is a mixed cell carcinom a (59), usually
detected unilaterally in the kidney of children generally up to three
years of age, but found le s s frequently in older children and in both
kidneys. It readily m eta sta sizes to the lungs and liv e r, and can be
effectively treated by surgery, chemotherapy and radiation, parti
cularly in younger children (60). It is of mixed epithelial and
fibroblastic morphology, has a high degree of vascularization and
can be either encapsuled or soft (61). Abnormal m aterials have
been observed in the serum of patients with W ilm 's tumor (62, 63);
and m aterials have been isolated from tumor extracts (63), sim ilar
on the b asis of electrophoretic m obility and staining with Alcian
Blue. The properties of this m aterial suggest that it is m ainly
acidic m ucopolysaccharide.
The approach taken in this study is to analyse, imm uno-
logically, m aterials in nephroblastoma tumor m a sses, serum from
nephroblastoma patients and extracts from seria lly cultured
nephroblastoma c e lls in vitro. The goal is to determ ine in part
the specificity of these m aterials to W ilm 's tumor; to characterize
them physically and chem ically; and to understand if p ossib le, the
site of their origin. These are hopeful fir st steps toward a
16
diagnostic tool, and a system whereby the function of tumor
associated antigens can be studied.
Imm unological analysis of neoplasia can be approached
with various strategies. The antigenic structure of neoplastic
c e lls shows extrem e d iversity, except possibly in the case of
v ira l etiology. Without knowing the etiology of a given tumor,
no expectations should a rise about the universality of its antigenic
structure. For this reason, a d iverse experim ental approach
focusing on one neoplasia is perhaps as rea listic an approach as
studying a variety of tum ors in hopes of finding common antigens.
i
I
! This has been the reasoning behind the structure of the present
study.
C H A PTER II
MATERIALS AND METHODS
A. IMMUNOLOGICAL PROCEDURES
Immunization
Male New Zealand white rabbits (ABC Caviary, Pomona,
California), approximately one year of age w ere injected initially
in four footpads with 1. 0 m l of tissu e extract (20 mg protein) or
cultured cell pellet (7 x 10 cells) em ulsified with 1. 0 m l Freund's
com plete adjuvant (Calbiochem, Los A ngeles, California).
Subsequent injections of the sam e com position and
volume w ere perform ed at three week intervals intraderm ally
at m ultiple site s along the back. This was continued until a
strong reaction was observed (generally after three to five
injections) between the antisera and the imm unizing antigen, as
judged qualitatively by Ouchterlony double diffusion in agar (see
below). T est bleedings w ere perform ed generally two to three
weeks after an imm unizing injection.
Four to five months after discontinuing this imm unization
schedule, antisera of the sam e potency w ere obtained from these
anim als, following injection into the m arginal ear vein of either
0. 4 m l tissu e extract (8 m g protein) or 0. 4 m l PBS (Appendix 1)
C
containing 7 x 10 c e lls. Two such injections w ere perform ed at
17
a four day interval, and seria l bleedings w ere made starting
three days after the second injection.
Preparation of A ntisera
Pre-im m unization sera and subsequent antisera were
obtained by punture of the central ear artery after vasodilation
with xylene. From 10 to 30 m l of blood w ere collected in g la ss
tubes, allowed to clot at room tem perature for two to three hours,
and refrigerated overnight at 6° C to allow clot concentration.
The supernatant was decanted and centrifuged 15 m inutes at
350 x g to sedim ent residual red blood c e lls. The rem aining
serum (approxim ately 1/2 the original blood volume) was aliquoted,
quick frozen in liquid nitrogen, and stored at -6 0 ° C.
Preparation of Immunoglobulins
A ll step s w ere perform ed at room tem perature. Rabbit
sera w ere diluted with an equal volume of PBS. To this solution
was slow ly added an equal volume of 100% saturated (25° C)
ammonium sulfate (Special Enzym e Grade, Mann R esearch
L aboratories, New York) taking approximately 20 m inutes for
addition during which constant m ixing was maintained by a
m agnetic stirrer. After adjusting the resulting suspension (50%
saturated ammonium sulfate, 25° C) to pH = 7.5, it was allowed
to stir 15 m inutes after which it was decanted into conical cen
trifuge tubes and centrifuged at 1500 x g for 20 m inutes. After
discarding the supernatant, the pellet was rinsed with 40%
saturated ammonium sulfate, and resuspended in a volume of
19
this equal to the original serum volume. It was allowed to stir
10 m inutes and recentrifuged as before. This was repeated one
or two tim es. The final p ellet was dissolved in PBS or BBS
(Appendix 1) to a volume generally one-fourth to one-fifth that of
the original serum . This was stored in aliquots at -6 0 ° C.
Absorption of A ntisera
To elim inate reactions between antibody and antigenic
components of norm al tissu e and cell extracts, the following
procedure was used: antisera (or immunoglobulin fractions of
these) w ere diluted with PBS to a volume tw ice that of the original
antiserum . To this was added the solution or suspension of
absorbing m aterial (see text for details). Incubation with gentle
shaking was carried out at 37° C for two to three hours, then at
4° C for 18 hours. This suspension was then centrifuged at
1500 x g for 20 m inutes at 4° C. The absorbed antisera
(supernatant) was concentrated either by ammonium sulfate pre
cipitation, as above, or by placing it in d ialysis tubing which was
imbedded in dry Sephadex G-100 (Pharm acia, Uppsala, Sweden).
A final volume one-fourth to one-fifth that of the original antiserum
was generally obtained. This was stored at -6 0 ° C in aliquots.
A lternatively, absorption in agar double diffusion exp eri
m ents was achieved by pre-diffusing a solution or suspension of
absorbing m aterial into the antibody-containing w ell prior to the
addition of the antibody. This was, in general, as effective as the
above method, and was used where noted in the text.
To insure com plete rem oval of antibody to norm al
components, a se r ie s of absorptions using increasing amounts of
absorbant was perform ed until antisera preparations no longer
reacted by Ouchterlony double diffusion with the absorbing solution.
Immunodiffusion
Double diffusion in agar was employed, using either of
two agar system s: 1) Prepoured Im m uno-Plates w ere obtained
from Hyland Laboratories (Costa M esa, California). The agar
gel solution contained (w /v in water) Difco Special Noble Agar,
2% ; G lycine, 7.5%; NaCl, 1% ; Sodium Azide 0. 1% (pH = 7. 0 -
7.2 ). W ell diam eter was 2 mm and cen ter-w ell-to-p erip h eral-
w ell distance was 2 mm. W ells contained 10 Pi when filled to a
lev el m eniscus. 2) Ionagar #2 (Consolidated Laboratories, Chicago
H eights, Illinois) 0. 85% (w/v) in BBS (and 0. 1% in sodium azide)
pH = 8„ 4, dissolved by heating, was readjusted to original volume,
filtered and poured into empty Im m uno-Plates. W ells w ere
punched with a reem ed 20-gauge needle using a pentagonal die
pattern resulting in a 2 mm cen ter-w ell-to-p erip h eral-w ell
distance. W ells contained 10 1 4 when filled to a lev el m eniscus.
After addition of antigen and antibody to w ells, g els w ere
incubated at room tem perature from 24 to 48 hours in a humidified,
sealed chamber. P recipitin reactions could be observed at this
tim e and photographed. A lternatively, reacted gels w ere washed
with saline and dried as described by Clausen (64). Dried agar
slid es w ere stained by 10 minute im m ersion in a 0. 25% (w/v)
21 ;
solution of Buffulo Black NBR (Allied Chem ical, New York, New
York) dissolved in m ethanol-glacial acetic acid-w ater solution
(v /v ratio 5:1:4). Stained slid es w ere rinsed in m ethanol-glacial
acetic acid-w ater solution and allowed to dry. These could be
stored and subsequently photographed.
To in crease the amount of antigen reacting in som e
immunodiffusion experim ents (as noted in text) w ells w ere
enlarged to accomodate 15 nl of solution, but the w ell-to -w ell
distance was not altered.
Im m unoelectrophoresis
Im m unoelectrophoresis was performed in gels of 0. 85%
(w/v) Ionagar #2 dissolved in sodium barbital im m unoelectro-
p horesis (IEP) buffer (Appendix 1) pH = 8. 3, r /2 = 0. 05. G els
w ere poured in a LKB immunophor (LKB Instrum ents, Washington,
D, C .) slide fram e, and punched with a LKB gel punch to give a
2 mm trough width, with 4 mm w ell-to-trough distance. W ells
contained 2 n l of sam ple.
After the antigen was applied to the w ells, a voltage of
135 V (m easure w ick-to-w ick) was applied for 75 m inutes. The
distance between wicks was 22 cm resulting in a field strength
of 6. 1 volts /cm .
Agar was rem oved from the cut troughs, antisera added
and incubation allowed for 24 hours at room tem perature.
Developed slid es w ere rinsed, dried and stained as described for
immunodiffusion plates, after which they w ere photographed and
stored.
22
Im m unological M aterials
Rabbit antibody to bovine serum was obtained from
Hyland Laboratories and stored at 4° C. Sodium azide was added
(0. 1% w /v) for preservative.
Fetal calf serum and newborn calf serum w ere obtained
from Gibco (sterile, filtered, Grand Island B iologicals C o .,
Grand Island, New York) and stored at -6 0 ° C.
A human plasm a pool was obtained from the A m erican
Red C ross (distributed by Courtland Laboratories, Los A ngeles,
California) and from Dr. Robert Nakamura (Los A ngeles County-
U niversity of Southern California M edical Center, Los A ngeles,
California; sixteen individual sera w ere pooled). These w ere
m ixed in equal volum es and stored at -6 0 ° C in aliquots.
Fetuin was obtained from Gibco (lyophilized, "99. 9%
pure," Spiro m ethod).
A ntisera to ABO c ells w ere obtained from Ortho Pharm a
ceutical Corp. , Raritan, New Jersey.
Anti Forssm an (hemolysin) was obtained from Hyland
Laboratories, Costa M esa, California.
Hyaluronic acid (human um bilical cord) was obtained
from Calbiochem, Los A ngeles, California.
a -feta l protein assay kit was obtained from Behringwerke,
H oechst Pharm aceutical, I n c ., Som erville, New Jersey.
Human urinary erythropoietin was procured by the
Department of Physiology, U niversity of the Northeast, C orrientes,
I Argentina, and p rocessed by the Hematology R esearch Laboratories,
| Childrens H ospital of Los A ngeles, for distribution by the National j
Heart and Lung Institute under R esearch Grant HE-10880. It was
authorized for distribution by the Com m ittee on Erythropoietin of j
the National Heart and Lung Institute. j
B. CHEMICAL METHODS !
: j
Protein Determ ination !
A nalysis of sam ples for protein was perform ed using the j
: i
m odified Lowry method as described by Eastoe (65). A ll test j
sam ples w ere in water, PBS or BBS. These sa lts w ere shown not ;
1
to effect the colorim etric reaction. Human IgG (M iles Labora
tories, Kankakee, Illinois) and human serum albumin (recry sta l
lized, supplied by Dr. Samuel E. Allerton) w ere used as standards
!
and showed no significant difference in sen sitivity to this assay, j
which was 3 to 5 yg protein (detectable).
Uronic Acid Determ ination
Uronic acid was analysed by the m odified carbazole
I
I
i method of B itter and Muir (66). A lpha-D -glucurono-lactone
(Eastman Organic C hem icals, R ochester, New York) was used as
i
a standard and a sen sitivity of about 5 vg was achieved.
Sensitivity of Antigen to Enzym atic Degradation
Enzym es w ere dissolved to a concentration of 0. 2% (w/v).
Pronase (Calbiochem, Los A ngeles, California), trypsin (9500 u /g ,
2x crystallized , General B iochem icals, Chagrin F a lls, Ohio),
ribonuclease (bovine pancreatic, 5x crystallized , Calbiochem,
Los A ngeles, California), and neuram inidase (influenza virus
neuram inidase, obtained in solution of 500 U /m l, and used directly I
without dilution; G eneral B iochem icals, Chagrin F a lls, Ohio)
w ere dissolved in PBS (pH = 7. 2).
H yaluronidase (ovine, grade 1A, M iles-Seravac L td ., j
t
Maidenhead, B erkshire, England) was dissolved in 0. 1 M m ono- j
sodium phosphate buffer (pH = 5. 3). T oO .lO m l of each of th ese |
enzym e solutions was added 0. 10 m l of antigen solution which had
j
been dialysed extensively against PBS. As a control, 0. 10 m l of \
this sam e antigen solution was added to 0. 10 m l of each buffer used I
to d issolve the enzym es. In the case of hyaluronidase, the antigen
solution was adjusted to pH = 5. 3 before dilution. These solutions j
w ere m ixed and allowed to incubate at 37° C for 24 hours in sealed
vials. Im m ediately after incubation, these various solutions w ere j
placed in imm unodiffusion w ells to evaluate the rem aining antigenic
activity. To insure against p ossib le degradation of immuno- j
globulins by residual trypsin during immunodiffusion, soybean j
trypsin inhibitor, 2% w /v in PBS (General B iochem icals, Chagrin j
F a lls, Ohio) was prediffused into the antibody-containing w ell prior j
to adding the antibody. j
i
To test for the presence of residual enzym e activity which i
!
m ight have interfered with the precipitin reaction, a 0. 1% w /v |
solution of each enzym e (prepared by diluting 0.2% enzym e stock
with PBS) was incubated for 24 hours, 37° C, and prediffused into j
i 25
| peripheral antibody w ells. After this, antibody w as added to these
! w ells and test antigen (incubated alone in PBS for 24 hours, 37° C)
; |
was placed in the center w ell to determ ine whether residual
enzym es had degraded either antigen or antibody.
Absorption of antibody with enzym e hydrolysates of
antigen was perform ed by prediffusing the hydrolysates into p eri- i
pheral w ells, after which antibody was placed in these w ells. The
center w ell was filled with antigen (incubated alone in PBS, 24
hours, 37° C). The inhibition of the precipitin line was observed
; for each enzym e hydrolysate.
Amino Acid A nalysis
|
Amino acid analysis was perform ed using a JEOL
Automatic Amino Acid A nalyser. The sam ple was hydrolysed
24 hours at 110° C under nitrogen in sealed g la ss vials in 6N HC1.
After this, sam ples w ere air dried, and taken into acetate buffer, ■
pH = 2 .2 for application to the chromatographic column. j
i
: j
C. PHYSICAL METHODS j
’ |
’ i
i
; M olecular Sieve Chromatography j
: i
G -75 Sephadex (Pharm acia, Uppsala, Sweden) was allowed j
1 I
to sw ell in Water over a three day period at 4° C. The gel was |
then equilibrated to 0. 10 M NaCl pH = 6 .5 , by repeatedly suspending
the settled beads in this solution. At room tem perature, after
evacuating the suspension to rem ove air bubles, a g la ss column,
; 0. 9 x 30 cm, was filled with a slurry of the beads, and allowed to
26
settle over the scintered g la ss filter in the bottom of the column.
The final height of the column bed was 27 cm. This bed was
j
thoroughly rinsed with the equilibrating solution (approxim ately j
500 m l over three days) with a flow rate of 0 .0 5 m l per minute
(using a 50 cm p ressu re head). ;
Sam ples w ere added to the column (0. 50 m l), allowed to
enter and rinsed in with the equilibrating solution. Fractions j
i
w ere collected (0. 26 m l per fraction), diluted by adding 1. 0 m l
water, and analysed for their absorption at 280 nm in a cuvette
of 1. 0 cm path length, using a Z eiss PMQ II spectrophotom eter.
A ll sam ples w ere run on the sam e column under identical
j
conditions. F erric chloride was obtained from Matheson Chem ical j
C orp ., E ast Rutherford, New Jersey. Ovalbumin, m olecular j
|
weight 45, 000 daltons, was obtained from Pharm acia, Uppsala, j
Sweden. Serum albumin (Fraction V) was obtained from Santa \
Monica Chem ical Co. , Santa Monica, California.
DEAE-Sephadex Chromatography
Maximal chromatographic resolution of antigen was
achieved using DEAE-Sephadex A -50 medium (Pharm acia, Uppsala,
Sweden). A column of 2. 5 x 31 cm was poured at 4° C after
equilibrating the DEAE-Sephadex beads in a buffer of 0. 10 M
form ic acid, adjusting to pH = 8. 3 with ammonium hydroxide. The j
sam ple was applied in 2. 5 m l, adjusted to pH = 8. 3 with ammonium
hydroxide and was eluted with 1, 000 m l of a linear gradient
ranging from 0. IN form ic acid to 1. O N form ic acid (adjusted to
27
pH = 8. 3 with ammonium hydroxide). Fractions of 5. 2 m l w ere
collected until the gradient was depleated, at an average flow rate
of 0. 4 - 0 .5 m l per m inute. Following this, IN form ic acid buffer,
adjusted to pH = 10. 3 with ammonium hydroxide, was flushed
through the column at the sam e flow rate. Fractions w ere analysed
for absorbance at 240 and 280 nm on a Beckman DB-G spectro
photometer. Pooled fractions w ere lyophilized, red issolved in 1. 5
m l of water and analysed for antigenic activity by immunodiffusion.
A crylam ide Gel E lectrophoresis
E lectrophoresis was perform ed in 7% acrylam ide gels as
described by D avis (67). Sam ples w ere incorporated into the
spacer gel before polym erization. G els of 7. 5 cm length and 0. 4
cm diam eter w ere run at a constant current of 4 mA per tube for
60 to 80 m inutes. During electrop h oresis, gels w ere maintained
at 8° C by im m ersion in the lower buffer chamber of a Blichler
P oly-prep (Buchler Instrum ents, Ft. Lee, New Jersey) electro
phoresis apparatus, cooled by a circulating ice bath. Some gels
w ere sectioned manually, and the sections eluted in water over
night at 4° C with shaking. Unsectioned gels w ere stained by
im m ersion for one hour at room tem perature in B rilliant Blue R
(C oom m assie Blue, Sigma C hem icals, St. Louis, M issouri), which
was initially dissolved in water and adjusted to a final concentration
of 0. 1% (w/v) in 12. 5% (w/v) trichloroacetic acid (TCA). G els
w ere subsequently destained by rinsing in 12. 5% TCA and photo
graphed.
28
Iso electric Focusing in A crylam ide G els
Iso electric focusing methodology was adapted from that of
W rigley (68). G els w ere prepared by m ixing 6. 0 m l of 4%
Ampholine (LKB Instrum ents, Washington, D. C .; diluted from
40% stock in water, pi range 3 - 10 or 3 - 6); 6. 0 m l aqueous
solution of 28% (w/v) acrylam ide and 0. 74% (w/v) m eth yl-(b is)-
acrylam ide (Bio Rad Laboratories, Richmond, California); 5. 6 m l
of 0. 004% (w/v) aqueous riboflavin; and 4. 4 m l water. The
resulting solution (22. 0 ml) was taken to 24. 0 m l with an aqueous
solution of the sam ple to be run, or proportionately partitioned
when running m ultiple sam ples. This final solution was evacuated
to rem ove dissolved air, poured into g la ss tubes prepared identi
cally as for electrop h oresis, and allowed to photopolym erize for
20 m inutes with a fluorescent lamp (G. E. 140W Daylight lamp,
2 inches from gels).
A fter polym erization, tubes w ere placed in a Buchler
apparatus as before (maintained at 8° C), the cathode solution con
taining 1% (v/v) ethanolamine and the anode solution 1% (v/v)
phosphoric acid. A current was m anually maintained at 2 mA per
tube until a voltage of 400 volts was reached (in 90 minutes) after
which the current was allowed to drop to a stable, low value
(typically 0. 5 mA per tube). At this point the run was discontinued.
Standards including hemoglobulin and cytochrom e C showed sharp,
dilineated bands after this procedure (about 120 m inutes after run
was begun).
; 29
Tubes w ere rem oved, and gels sliced and eluted as
i
described for electrop h oresis procedures. The pH of each section
j
eluate was determ ined. j
i
A nalytical Ultracentrifugation J
o !
Sam ples w ere placed in a standard 4 section, plastic j
centerpiece cell, and centrifuged at 59, 780 rpm in an AN-D rotor j
i
at 20. 0° C using a Beckman Model E Ultracentrifuge (Spinco j
i
D ivision, Beckman Instrum ents, Palo Alto, California). Sedi
mentation velocity studies w ere perform ed using sch lieren optics
i
( 0 = 70°). Photographs w ere taken at 8 minute intervals, and
; peak positions m onitored by a Nikon two-dim ensional comparator. j
|
Sedimentation coefficients w ere calculated from the slope of a In j
i
r vs. t plot, where r is the distance of the boundary from the j
center of rotation, and t the tim e of centrifugation. j
A reas under the sch lieren patterns w ere determ ined by
1 weighing traced enlargem ents of the pattern cut out of uniform
i plastic sheet. j
; j
j i
i D. PREPARATION OF EXTRACTS FROM TISSUES AND j
CULTURED CELLS j
' C ell Cultures
Serial cultures of W ilm 's tumor cells (CCL 31 TuWi, |
■ A m erican Type Culture Collection, R ockville, Maryland) or human
fetal kidney c e lls (MA-147, M icrobiological A ssociates, Bethesda,
; Maryland) w ere grown in a humidified incubator at 37° C to nearly
30
confluent m onolayer in Blake flasks (B ellco B iological G lassw are,
Vineland, New Jersey), using E agles, MEM (Appendix 1) supple
mented with 10% newborn calf serum . After decanting the medium, j
|
c ells w ere vigorously washed four tim es with BSS, using 100 m l
per w ash per flask. I
i
I
C ells used for imm unization w ere m echanically rem oved j
from the flasks with a rubber policem an, suspended in PBS and j
i
centrifuged at 350 x g for 20 m inutes at room tem perature. These !
p ellets w ere frozen and thawed three tim es to elim inate c ell j
viability, which was tested by reseed in g a portion of the pellet in
a culture tube using the growth medium above. C ell counts w ere
perform ed using a haem ocytom eter prior to centrifugation. P e llets j
w ere stored at -6 0 ° C until used. Injection schedules are j
i
indicated under Immunization. I
C ells used for ethylenediam inetetracetic acid (EDTA) |
I
extraction of antigen w ere incubated (following the BSS wash) with j
i
20 m l PBS-EDTA (Appendix 1) for ten m inutes at 37° C, at which |
|
tim e the c e lls released from the g la ss. C ells w ere triturated into |
suspension, counted for ce ll number (as above) and for c e ll j
viability using trypan blue exclusion (69). Viability as m easured
by this criterion was greater than 95%. C ells w ere pelleted
im m ediately after suspension, as above. The supernatant (PBS- |
i
EDTA extract) was lyophilized to dryness, resuspended in water
to an appropriate volume (about 5 m l), and dialysed extensively
against PBS. This was repeated to obtain a very sm all volume
without accumulating salts from the extract.
31
An alternative procedure of preparing concentrated
solution of antigen from tissu e culture was to hom ogenize p ellets
of m echanically rem oved c ells (following the BSS wash) in PBS-
EDTA using a P otter-E lvajem hom ogenizer at 4° C.
I
| T issu e Extracts
| Human fetal kidney, human adult kidney, and tumor
tissu e was frozen (-6 0 ° C or quick frozen in liquid nitrogen)
1-24 hours after rem oval from the patient, and stored at -6 0 ° C
until use. In two c a ses, tissu e was extracted fresh within two
hours of rem oval. Tumor tissu e was typically necrotic in the
! center, vascularized to varying d egrees, and ranged in texture
I from encapsuled, rigid to amorphous, flaccid.
j Kidney and tumor tissu e s w ere extracted with PBS-EDTA
i by m incing thawed tissu e into approxim ately 2 mm cubes. Care
was taken in all c a ses to use representative tissu e when large
specim ens w ere used. M inces w ere added to a volume (in ml) of
PBS-EDTA equal to the weight (in gram s) of the wet tissu e. These
! w ere incubated three hours at room tem perature, while stirring
| with a m agnetic stirrer. Prom this point, all steps w ere carried
out at 4° C. The extract was centrifuged at 20, 000 x g for 15
m inutes in a Servall (Ivan Sorvall, I n c ., Norwalk, Conn.) centrifuge.
A portion of the resulting supernatant (PBS-EDTA extract) was
further centrifuged at 115, 000 x g for 120 m inutes on a Beckman
i Model L preparative ultracentrifuge (42, 000 rpm, #50 rotor). The
i supernatant was decanted and the p ellet resuspended in PBS
32
j (one-fifth the original volume). T hese preparations w ere stored
i at -6 0 ° C.
i I
: i
A second tissu e preparation utilized hom ogenization in \
PBS. R epresentative tissu e p ieces w ere added to a volume of
PBS (in ml) equal to the weight (in gram s) of the wet tissu e. This
was hom ogenized in a Sorvall Om nim ixer (5 m l cup, Ivan Sorvall,
I
In c ., Norwalk, Conn.) for 30 - 45 seconds, #6; centrifuged 15
m inutes at 20, 000 x g, 4° C, as above, and part of the super
natant (PBS-homogenate) recentrifuged at 115, 000 x g for 120
m inutes. This supernatant was decanted, the p ellet suspended as
before, and all stored at -6 0 ° C.
A final extraction procedure utilizing 3M KC1 (70) was
applied to these tissu e s. Wet tissu e was lyophilized to dryness j
and pulverized in a m ortar and p estle. A PBS-KC1 solution i
(Appendix 1) was added and the suspension allowed to shake gently '
at 4° C for 20 hours. The resulting viscous solution was cen tri- |
j
fuged at 115, 000 x g for 120 m inutes. The supernatant was
i
dialysed extensively first, against PBS, and then against distilled
i
w ater, at 4° C. The w ater-dialysed m aterial was centrifuged at
i
! 1500 x g for 15 m inutes at 4° C to rem ove particulate m aterial; and |
the supernatant of this (KC1 extract) was frozen and stored at
j -6 0 ° C.
A procedure identical to this was used for extracting
cultured W ilm 's tumor c e lls with the exception that p ellets of
m echanically rem oved c e lls (see above) w ere employed for
extraction rather than lyophilized powder.
33
W ater Soluble Antigens
To obtain large amounts of water soluble antigen from
calf serum , and PBS-EDTA extract of tumor tissue; these
!
| m aterials w ere extensively dialysed against d istilled w ater (pH = 5)
I at 4° C. The resulting suspension was centrifuged at 1500 x g for
| 30 m inutes to rem ove particulates, and the resulting supernatant
l
| was lyophilized to dryness. This was then reconstituted in an
I appropriate volume of water for further purification.
Sim ilarly prepared solutions w ere obtained by dialysing
I tissu e and c e ll extracts and hom ogenates (PBS, PBS-EDTA or
|
! KC1) against distlled w ater. These w ere reconstituted in the
j sm a llest volume p ossib le for analysis of antigenic activity. This
| procedure was also applied to serum of tumor patients which was
j
collected at different tim es during diagnosis and treatm ent, and
I stored at -6 0 ° C prior to use.
I
| Ammonium Sulfate Fractionation of Newborn Calf Serum
!
| Fractionation of newborn calf serum to obtain a crude
purification of antigenic m aterial was perform ed at room tem
perature by slow ly adding 100% saturated ammonium sulfate (25° C)
to a solution of w ater-dialysed calf serum (pH = 5) diluted to a
protein concentration of 20 m g per m l, with constant stirring. At
the appropriate concentration of ammonium sulfate, the suspension
i was centrifuged at 1500 x g for 30 m inutes, the supernatant decanted
and the p ellets stored. To a known volume of supernatant was
added 100% saturated ammonium sulfate to obtain the next
34
appropriate concentration; and the procedure repeated. P e llets
w ere obtained from 0 - 30% saturated ammonium sulfate, 30 - 45%
saturated ammonium sulfate, 45 - 60% saturated ammonium sulfate,
60 - 75% saturated ammonium sulfate, and 70 - 90% saturated
ammonium sulfate. They w ere dissolved in, and extensively
dialysed against disttiled water, and stored at -6 0 ° C.
CHAPTER III
RESULTS
Nature of the Immunoglobulins Employed
A nalytical ultracentrifugation was utilized to exam ine the
quality of the immunoglobulins obtained by ammonium sulfate frac -
tionation of rabbit antisera, described in MATERIALS AND
j METHODS. Figure la illu stra tes the pattern observed with one anti
serum , R 5-19, obtained after a boost injection of the PBS-EDTA
extract of four pooled W ilm 's tum ors. The ammonium sulfate p re-
jcipitate of this anti serum was dissolved in PBS and passed through a
column of G -25 Sephadex, equilibrated with the sam e buffer, to
rem ove e x cess ammonium sulfate. A portion of the m aterial
j chromatographing in the void volume was diluted with PBS to a final
i
concentration of 5. 0 m g protein per m l (based on the Lowry method
described), and centrifuged. The sch lieren pattern of this sed i-
menting m aterial (Figure la) shows a peak with a sedim entation
I
coefficient (20. 0° C, PBS) of s = 7. 0 Svedbergs. No other peaks
w ere apparent, indicating uniform ity by this technique, and
suggesting Immunoglobulin (IgG) type G as the major immunoglobulin
c la ss present in this preparation.
As an additional check on purity, com m ercially obtained
rabbit IgG w as subjected to the sam e centrifugation conditions. This
m aterial showed a sch lieren pattern (Figure lb) indicating a peak
35
Fig. 1. Sedimentation P roperties of
Immunoglobulin Fraction.
A. Schlieren optical pattern of
R5 immunoglobulin preparation,
5. 0 mg protein per m l in PBS
tim e of run = 39 m inutes
9 = 70°
rpm = 59, 780
T = 20. 0° C
direction of sedimentation: right to left
B. Schlieren optical pattern of
com m ercial rabbit IgG 10 m l
protein per m l PBS
tim e of run = 44 m inutes
9 = 70°
rpm = 59, 780
T = 20. 0° C
direction of sedimentation: right to left
36
37
f t i <
!?CV< ' -
% $:$? U / ‘i " \ 1
T y - A ' T ' ' ” v ^
i ^ , -> t ^
38
with a sedim entation coefficient of s = 6. 8 Svedbergs, with a trace
contaminant peak sedim enting faster. The ratio of protein concen
tration to the m en iscu s-co rrected area under the sch lieren pattern
was determ ined for this preparation, and for the preparation of
R 5-19 above. This ratio was identical for the two preparations,
within 5% experim ental error.
The protein concentration subsequently referred to in this
antiserum fraction is 7 s protein, sin ce no other sp ecies w ere
apparent. T his preparation was used in m ost of the studies where
antibody to tumor extract was em ployed, and is designated R5.
!Identical preparations from bleedings im m ediately prior to R 5-19
i
| w ere ocassion ally used and are also designated R5. No differences
i
w ere apparent in the quality of these preparations.
|D em onstration of an Abnormal Antigen in W ilm 's Tumor Extract
l
j
| Antibody preparation R5 was em ployed in Ouchterlony
| double diffusion in agar to study its reaction with the im m unizing
antigen, a PBS-EDTA extract of four pooled W ilm 's tum ors. Figure
2 shows the reaction observed between R5 and this extract. A
reaction between this antibody and norm al human serum was seen as
w ell. Pre-im m unization serum from the sam e rabbit showed no
reaction when tested against the sam e extract, or against human
serum (not shown). That R5 reacted with human serum would be
i
|
expected, sin ce serum was undoubtedly present in the imm unizing
m aterial. To elim inate the reaction of R5 with th ese substances and
with norm al components of human kidney tissu e , this antibody
F ig . 2. Immunodiffusion R eaction of R 5-19 with
Pooled W ilm 's Tumor Extract and Human
Serum.
Center well: R5 immunoglobulin fraction,
750 pg 7 s protein in 10 pi
PBS
W ell 1, 2, 5: PBS-EDTA extract of four
pooled W ilm 's tumors
W ell 3, 4: pooled human serum
40
41
preparation was absorbed as described in MATERIALS AND
METHODS. One m l R5 (containing 75 m g 7 s protein) was incubated
with one m l pooled human plasm a and one m l PBS-EDTA extract of
norm al adult human kidney, after which incubation and centrifugation
was perform ed as described. The supernatant from this was placed
in d ialysis tubing, and im m ersed in dry Sephadex G -200 until a final
volume of one m l (the original volume of R5) was obtained.
The reaction of this absorbed antibody with the sam e
PBS-EDTA extract of four W ilm 's tum ors is illustrated in Figure 3.
Adjacent w ells contain the PBS-EDTA extract of norm al adult human
kidney used for absorption, and pooled human plasm a. A marked
i precipitin band p er sists with the tumor extract which is not apparent
i
with the two norm al m aterials. Absorption of this antiserum p re
paration with increasing amounts of these norm al m aterials failed
to alter the intensity or position of this band (the absorbed prepara
tion was always concentrated to its original volum e, initially con
taining 75 m g of rabbit 7s protein per m l).
By these criteria, an antigen was dem onstrated in this
tumor extract which was not present at the sam e concentration in the
norm al human kidney and plasm a tested.
Some substances found norm ally in humans are known to be
abnormally distributed, or abnormally abundant during neoplastic
d isea se. Some of these w ere examined for p ossib le im m unological
relationship to the antigen detected in the W ilm 's tumor extract.
Hyaluronic acid (71) and human urinary erythropoeitin (72) w ere
| dissolved at a concentration of 1. 0% w /v in PBS, placed in w ells
F ig. 3. Immunodiffusion R eaction of R 5 Immunol
globulin Fraction (Absorbed) with Norm al
Human Kidney Extract, Human Plasm a,
and W ilm 's Tumor Extract.
Center Well: R5, absorbed with pooled
human plasm a and norm al
adult human kidney PBS-
EDTA extract
W ell 1, 3, 5: PBS-EDTA extract of
pooled W ilm 's tum ors
(diluted 1:5 with PBS)
W ell 2: norm al adult human kidney
PBS-EDTA extract
W ell 4: pooled human plasm a
42
43
' I > ~ ?
44
adjacent to the W ilm 's tumor extract described, and allowed to
react against R5, absorbed as before. There was no evidence of
any reaction of these substances with R5; and prediffusion of these
m aterials into the center w ell, prior to the addition of R5, failed to
inhibit the reaction of this antiserum and the extract of the four
pooled W ilm 's tum ors. T his indicates that there is no apparent
im m unological relationship between these substances and the antigen
observed in the tumor extract.
Antibody to F orssm an antigen, and to A group and B group
red blood c e lls (Rht-) failed to show any reaction with the pooled
tumor extract by double diffusion. Furtherm ore, incubation of the
extract with th ese antisera, prior to reacting the extract against R5
in double diffusion, failed to inhibit the observed reaction with R5.
This suggests that the antigen observed in the W ilm 's tumor extract
is not related to th ese blood group antigens.
The W ilm 's tumor PBS-EDTA extract containing the
antigen observed was assayed for the p resen ce of a-fetal protein
described by A belev in human hepatoma (46). A radical diffusion
assay em ploying sp ecific antibody to this protein was utilized, and
known a -feta l protein standards w ere included. There was no
evidence of this substance in the PBS-EDTA extract of the pooled
W ilm 's tum ors, insuring a concentration of le s s than 50 g per m l.
Additionally, “ -fe ta l protein was prediffused into the antibody con
taining w ell (2. 2 g total) prior to addition of R5, to test its ability
to alter the immunodiffusion pattern obtained with this antibody and
the extract of W ilm 's tumor. No effect was observed, indicating
45
lack of absorption with this a -feta l protein, and the apparent
im m unological unrelatedness between this substance and the antigen
observed in the PBS-EDTA extract of pooled W ilm 's tum ors.
Exploration for the Antigen in Other Human M aterials
A. F etal T issu es
Kidneys w ere obtained from six aborted fetu ses, rep re
senting gestation periods throughout the fir st two tr im e ste r s. These
w ere hom ogenized in PBS-EDTA using a ground g la ss P o tter-
Elvajem hom ogenizer, at room tem perature. In double diffusion
experim ents these hom ogenates failed to show any reaction with
antibody preparation R5 (absorbed as above with human plasm a and
norm al adult human kidney extract); they w ere also unable to
inhibit the reaction of this antibody with the PBS-EDTA extract of
pooled W ilm 's tum ors by prediffusion into the antibody containing
w ell.
Spleen and thymus tissu e from three of th ese fetu ses (all
in the second trim ester of gestation) w ere prepared and tested in
the sam e manner. These also failed, by the sam e criteria, to
dem onstrate the antigen detected in the W ilm 's tumor extract.
B. Other W ilm 's Tum ors and N on-R elated Tum ors
T issue from three individual W ilm 's tum ors and one clear
c e ll carcinom a of the kidney w ere extracted with PBS-EDTA as
described. Each of th ese was tested for its ability to react
directly with R5 (absorbed as above), and to inhibit the reaction of
R5 with the PBS-EDTA extract of the pooled W ilm 's tum ors. By
46
th ese criteria, none of the preparations from th ese tum ors showed
antigenic related n ess to the antigen observed in the original
im m unizing extract of pooled W ilm 's tum ors.
Portions of one of these W ilm 's tum ors (R. S. ) w ere
alternatively prepared as a PBS homogenate and as a KC1 extract,
as described in MATERIALS AND METHODS. T hese preparations
failed also to dem onstrate a precipitin reaction related to that of
the antigen observed in the PBS-EDTA extract of the pooled W ilm 's
tum ors. Since the antigen of in terest was subsequently found to be
non-dializable and w ater soluble, concentration of the potentially
undetected antigen in th ese additional tumor extracts was attempted.
The extracts w ere dialysed against water, centrifuged at 1500 x g
for 30 m inutes, and the supernatant lyophilized to dryness. This
was then reconstituted with w ater and tested as above by imm uno
diffusion, after an eight- to ten-fold concentration of water soluble,
non-dializable m aterial. T hese concentrates also failed to demon
strate cro ss-rea ctio n with the antigen from the pooled W ilm 's
tumor extract. The observed reaction with the pooled tumor
extract is detectable at a 1:20 dilution of the original extract,
using the R5 preparation described. The lev el of antigen detect
able, then, in these eight- to ten-fold concentrates, would be
160- to 200-fold le s s than that in the original im m unizing extract
of the pooled W ilm 's tum ors. This would n ecessita te a rather
large quantitative difference in th ese various extracts to escape
detection.
The resu lts of section A and B are sum m arized in Figure 4;
F ig . 4. A nalysis for Antigen in Extracts of
Human F etal T issu e, W ilm 's Tumors
and N on-R elated Tum ors.
T issue Patient Sex Age Preparation Antigen'1 '
Human fetal kidney Y. S. M 18 weeks PBS-EDTA homogenate
-
D. A. M 18 weeks PBS-EDTA homogenate -
M. N. F 19 weeks PBS-EDTA homogenate —
J. M. F 24 weeks PBS-EDTA homogenate
-
---- ----
15 weeks PBS-EDTA homogenate -
---- ----
22 weeks PBS-EDTA homogenate
-
Human fetal thymus Y. S. M 18 weeks PBS-EDTA homogenate
-
D. A. M 18 weeks PBS-EDTA homogenate -
M. N. F 19 weeks PBS-EDTA homogenate
-
Human fetal spleen Y. S. M 18 weeks PBS-EDTA homogenate
-
D. A. M 18 weeks PBS-EDTA homogenate
-
M. N. F 19 weeks PBS-EDTA homogenate
-
W ilm's tumors M. C. F
--- -
PBS-EDTA extract
-
F. G. M
----
PBS-EDTA extract
-
R. S. F 7 months PBS-EDTA extract
-
PB S -homogenate
-
PBS-KC1 extract
“
Clear cell carcinoma I. C. F. PBS-EDTA extract
_
vl,
^assayed by direct immunodiffusion with R5 (absorbed) and by ability to absorb R5
against test antigen ^
00
49
C. Serum from Tum or-Bearing and Norm al Patients
Eighteen individual serum sam ples from W ilm 's tum or-
bearing patients w ere obtained at various stages of tumor develop
ment and treatm ent. These w ere subjected to the sam e immuno
diffusion assay as above to evaluate the presence of antigen related
to that observed in the PBS-EDTA extract of pooled W ilm 's
tum ors. These w ere assayed im m ediately upon thawing (when
received in dry ice), or fresh when available. They w ere then
w ater-dialyzed, centrifuged, lyophilized and reconstituted as
above, resulting in a fiv e- to ten-fold concentration of water
soluble m aterials over the initial serum concentration, and this
concentrate assayed. None of these sera dem onstrated an antigen
related to that observed in the pooled W ilm 's tumor extract.
Serum from four patients with non-related tum ors (two
with osteogenic sarcom a, one with rhabdom yosarcom a and one
with clear c ell carcinom a) w ere sim ilarly prepared and tested.
These did not show a reaction with R5 in immunodiffusion. T issu e
from the tumor of one of these patients (clear c e ll carcinom a) was
tested for the presen ce of antigen, described above, but no tissu e
from the other c a se s was available for analysis. The resu lts of
this section are sum m arized in Figure 5.
D etection of the Antigen on Cultured W ilm 's Tumor C ells,
in Newborn Calf Serum and in Fetuin
An analysis of the PBS-EDTA extract obtained from an
established culture of W ilm 's tumor c e lls (line TuWi CCL 31,
Fig. 5. A nalysis for Antigen in Serum of Patients
with W ilm 's and Other Tumors; Human
F etal and M aternal Serum; and D escrip
tion of Patients Donating A ntigen-P ositive
W ilm 's Tumor Pool.
Tumor or Sample Type
Number of
Patients
Sera Tested
Number of
Sera
Dem onstrating
Antigen
W ilm 's tumor 18 0
O steogenic sarcom a 2 0
Clear cell carcinom a 1 0
Rhabdomy o sar coma 1 0
Human fetus serum (18 w eeks) 1 0
Mother of W ilm 's tumor patient 1 0
Human serum 6 y ears old m ale 1 0
D escription of patients donating pool of four W ilm 's tumor
dem onstrating antigen:
Patient Sex Age at Surgery Blood Group Chemotherapy
at Surgery
L. S. F 14 months A + none
K. I. M 24 months O + none
M. C. F 42 months A + Actinom ycin D
K. S. F 45 months AB + none
52
A m erican Type T issu e C ollection, R ockville, M d.) revealed the
p resen ce of an antigen im m unologic ally related to the antigen
detected in the PBS-EDTA extract of pooled W ilm 's tumor.
C ells w ere grown, washed and extracted with PBS-EDTA
as described in MATERIALS AND METHODS; and the extract
O
concentrated to 0. 1 m l, representing 10 extracted c e lls. This
concentrate was tested by imm unodiffusion techniques described
above, against R5 (absorbed as before with norm al pooled human
plasm a and PBS-EDTA extract of norm al adult human kidney). A s
seen in Figure 6, a single precipitin band occurs with this PBS-
EDTA extract of cultured W ilm 's tumor c e lls, which shows a
reaction of identity with the antigen from the PBS-EDTA extract
of pooled W ilm 's tum ors. A control w ell containing pooled human
plasm a is included in this figure. Identical experim ents w ere
perform ed in which the PBS-EDTA extract of norm al adult human
kidney was placed in the control w ell. Sim ilarly, no reaction was
apparent between this absorbing extract and absorbed R5.
That these c e lls w ere over 95% viable by trypan blue
exclusion, after the PBS-EDTA extraction, indirectly suggests
that the antigen m ay resid e on the outside surface of these c e lls,
sin ce g ro ss c e ll breakage can be ruled out as a source of released
antigen. It is p ossib le, how ever, that the antigen is internal, and
relea sed by this extraction procedure without cellular destruction.
An identical extraction procedure was carried out with
norm al human fetal kidney c e lls (M icrobiological A sso cia tes,
Bethesda, M d.) grown in the sam e medium; the final
F ig . 6. Immunodiffusion C ross Reaction of
W ilm 's Tumor Extract and Extract of
Cultured W ilm 's Tumor C ells.
Center w ell: R5 (absorbed) 15 pi
W ell 1, 3, 5: PBS-EDTA extract of
pooled W ilm 's tum ors
(diluted 1:5 with PBS)
W ell 2: PBS-EDTA extract of cultured
W ilm 's tum or c e lls (10^ c e lls
represented per m l of extract)
W ell 4: pooled human plasm a
53

55
concentrated PBS-EDTA extract represented 5 x 10 c e lls in
0. 1 m l (20 tim es le s s c e lls represented per volume than in the
extract of the W ilm's tumor cell line). When this extract was
tested in an analogous experim ent for the presence of an antigen
related to that in the PBS-EDTA extract of pooled W ilm 's tum ors,
no precipitin line was observed (Figure 7).
This experim ent may not p recisely answer the question
as to whether or not this antigen was present on norm al human
fetal kidney c e lls in vitro for the following reasons: 1) the antigen
O
concentration obtained from 10 c e lls per 0. 1 m l of W ilm 's tumor
cultured c e lls is approximately the sam e as the antigen concentra
tion in a four-fold dilution of the PBS-EDTA extract of pooled
W ilm 's tum ors, as judged by the position of the precipitin lin es in
Figure 6; 2) the c e ll number represented in the extract of
norm al kidney c e lls is 20-fold le s s than that in the tumor cell
extract; 3) even if the sam e amount of extractable antigen w ere
present on the norm al c ells as on the tumor c e lls, this extract of
norm al c e lls would contain 80 tim es le s s antigen per volume than
the extract of pooled W ilm 's tumors; and 4) this dilution of
antigen is undetectable by the method employed.
Insufficient cultured norm al kidney c e lls w ere obtainable
to extract the quantity of antigen n ecessary for detection, nor was
enough m aterial available to attempt absorption of R5 with the
extract of norm al kidney c ells.
As an alternative approach, a sim ilar PBS-EDTA
O
homogenate (representing 10 c e lls per 0 .1 ml) was made of
Fig. 7. Immunodiffusion C ross Reaction of
W ilm 's Tumor Extract and Extract
of Cultured Norm al Human Fetal
Kidney C ells.
Center well: R5 (absorbed)
W ell 1, 3, 5: PBS-EDTA extract of
pooled W ilm 's tumors
(diluted 1:5 with PBS)
W ell 2: PBS-EDTA extract of cultured
human embryonic kidney c e lls
(5 x 1 C )8 c e lls represented per
m l extract)
W ell 4: PBS-EDTA extract of norm al
adult kidney
56
57
’ l i}
L .
58
polyoma virus-transform ed baby ham ster kidney c e lls (PY-BHK
C l3 line) grown in the sam e medium and washed as above, but
m echanically rem oved and hom ogenized after a p ellet was obtained
by centrifugation. This homogenate, representing the sam e
number of c e lls per volume as did the extract of cultured W ilm 's
tumor c e lls, failed to dem onstrate a reaction in immunodiffusion
against R5, nor did it inhibit the reaction between R5 and the
extract of pooled W ilm 's tum ors. Thus, the extract of these c e lls
(which are m alignantly transform ed kidney ce lls, though of another
species) failed to dem onstrate the common antigen found in the
PBS-EDTA extract of pooled W ilm 's tumor tissu e, and in the
analogous extract of cultured W ilm 's tumor c e lls.
As a further control experim ent, newborn calf serum ,
com plem ent-deactivated at 56° C for one hour (the sam e as that
used for the tissu e culture medium) was tested as a source of the
antigen found in cultured W ilm 's tumor c e lls. This calf serum
did contain antigenic m aterial dem onstrating a line of identity
with the antigen from the PBS-EDTA extract of pooled W ilm 's
tum ors, and with the antigen from the PBS-EDTA extract of
cultured W ilm 's tumor c e lls. In addition, a second precipitin
band was observed in this calf serum , not seen in the other two
extracts. Both of these calf serum antigens w ere present in an
ammonium sulfate fraction (30-45% saturated ammonium sulfate,
25° C) of the sam e calf serum , w ater-dialysed and prepared as
described in MATERIALS AND METHODS. T hese reactions are
shown in Figure 8.
Fig. 8. C ross Reaction of Antigens from Calf
Serum, Fetuin, and from Pooled W ilm 's
Tumor and T issu e Culture E xtracts.
Center
W ell 1,
W ell 2:
W ell 3:
W ell 5:
well: R5 (absorbed)
4: PBS-EDTA extract of
pooled W ilm 's tum ors
(diluted 1:5 with PBS)
Fetuin (15 yg per m l in PBS)
PBS-EDTA homogenate of
cultured W ilm 's tumor ce lls
(approx. 5 x 10? ce lls per ml)
30 - 40% saturated ammonium
sulfate (25° C) fraction of w ater-
dialysed newborn calf serum
(2. 2 m g protein per ml)
59
60
I
61
The unexpected finding of this antigen in newborn calf
serum , and in this particular ammonium sulfate fraction of calf
serum , lead to the further pursuit of its identity. C om m ercial
fetuin, prepared from fetal calf serum by the method of Spiro (73)
showed by immunodiffusion a reaction of identity with the common
antigens in the PBS-EDTA extract of pooled W ilm 's tum ors, in the
PBS-EDTA extract of cultured W ilm 's tumor c e lls, and in newborn
calf serum (Figure 8).
To sum m arize, an antigen was detected by immunodiffu
sion using antiserum prepared against a PBS-EDTA extract of
four pooled W ilm 's tum ors, and absorb ed with norm al human
kidney and plasm a. T his antigen was dem onstrated in 1) the
imm unizing extract (of four W ilm 's tumors); 2) a sim ilar PBS-
EDTA extracts of cultured W ilm 's tumor cells; 3) a partially
defined fraction of the newborn calf serum used to culture these
cells; and 4) a com m ercial preparation of fetuin. It was not
observable in a PBS-EDTA extract of norm al human kidney c e lls
grown in the sam e medium, nor in a PBS-EDTA homogenate of
polyoma virus-transform ed baby ham ster kidney c e lls also grown
in the sam e medium.
An attempt w as made to dem onstrate this sam e antigen
using antisera prepared against cultured W ilm 's tumor c ells,
since this was an apparent source of the common antigen detected.
Preparation of these c e lls for imm unization is described in
MATERIALS AND METHODS. B riefly, cultures of W ilm 's tumor
c e lls w ere grown as before, rinsed and harvested by m echanical
rem oval without extraction with PBS-EDTA. After centrifugation,
the resulting p ellet was suspended in Freund's com plete adjuvant
and injected as described. A ntisera w ere obtained, the immuno
globulin fraction prepared as before, and absorption carried out
with norm al human plasm a and extract of adult human kidney.
After absorption, this antibody preparation against
cultured tumor c e lls failed to dem onstrate any reaction by immuno
diffusion with 1) the PBS-EDTA extract of the four pooled W ilm 's
tumors; 2) the extract of cultured W ilm 's tumor cells; 3) new
born calf serum; or 4) com m ercially obtained fetuin. A ll of these
had dem onstrated a common antigen when using the R5 antibody pre
paration. E xtracts of other individual W ilm's tumors (tested as
before with R5) also failed to show any reaction with the absorbed
antibody preparation to cultured W ilm 's tumor cells.
The failure of this antibody preparation to dem onstrate
the previously detected antigen is not fully understood. The anti -
body preparation before absorption reacted strongly with both
kidney and tumor extracts, indicating the rabbit's ability to
respond to a number of antigens. That a much lower dose of
antigen was injected, and that the form of the antigen upon injection
was different from that used to prepare R5 may be partial explana
tions.
D etection of these common antigens was also attempted
using rabbit antiserum to bovine serum , and is d iscu ssed under
Im m unoelectrophoresis below.
63
Nature of the Antigen
To describe m ore p recisely, and possibly to differentiate
the common antigens observed in the extracts of W ilm 's tumor
tissu e and cultured tumor c ells, in newborn calf serum and in
com m ercially obtained fetuin; a physical and chem ical charac
terization of these antigens was pursued.
A. G eneral P roperties
Antigenic activity was determ ined qualitatively as above
by double diffusion in agar employing R5 absorbed with human
plasm a and norm al human kidney extract. PBS-EDTA extract of
the four pooled W ilm 's tum ors, newborn calf serum and com m er
cial fetuin w ere subjected to a number of procedures; these pro
cedures w ere tested for their ability to alter the intensity, sharp
n ess or the position (i. e . , the antigen concentration of the p re
cipitin band compared to the precipitin band obtained with these
m aterials prior to the given procedure. The following resu lts
w ere obtained:
1. Upon dialysing these antigen preparations
against PBS, at 4° C, all antigenic activity rem ained inside the
d ialysis tubing.
2. When these w ere dialysed against water, a p re
cipitate com monly formed inside the tubing (except in the case of
com m ercial fetuin). Antigen activity was found in the supernatant
after rem oving the precipitate by centrifugation at 1000 x g for
15 m inutes. This was found when the pH of the water was
64 |
maintained at 4. 5, 5. 2, 7 .5 or 8. 5. These aqueous solutions of
antigens w ere stable for w eeks at 4° C.
3. Repeated freezing and thawing did not effect the
antigenic activity of any of these antigen preparations, or their
w ater-dialysed counterparts.
I
4. Antigenic activity was not precipitated in 3%
v /v aqueous acetic acid (pH = 2. 7).
5. M aintaining th ese antigen preparations at 56° C
for one hour did not effect the antigenic activity.
6. Antigenic activity was not sedim ented at
100, 000 x g after two hours.
7. The diffusion coefficient in agar was greater
for these antigens than that of the 7 s globulin in R5 (judged
qualitatively by the curvature of the precipitin band in im m uno
diffusion experim ents).
Following any of th ese procedures, these antigen prepara- ;
tions showed a reaction of identity when placed in adjacent w ells
in immunodiffusion experim ents. The additional precipitin band
observed in newborn calf serum was not affected by these pro
cedures.
Insufficient m aterial w as available to test the properties
of the antigen obtained from cultured W ilm 's tumor c e lls by these
criteria.
B. Im m unoelectrophoresis Studies 1
Im m unoelectrophoresis was perform ed as described in
MATERIALS AND METHODS to determ ine the heterogeneity and
65 j
relative m ob ilities of the common antigens detected by imm uno
diffusion. Initially, whole newborn calf serum and the PBS-EDTA
extract of pooled W ilm 's tum ors w ere co-electroph oresed , and
allowed to react against R5, absorbed as before with human plasm a
and norm al adult human kidney extract. Figure 9 illu stra tes the
resulting pattern, indicating a single precipitin arc in the case of
the tumor extract. Two, n o n -cro ss-rea ctin g antigens appear in
the newborn calf serum: one having a m obility identical to that of
the tumor extract antigen, and the other slightly slow er.
Figure 10 rep resen ts an identical experim ent com paring
the sam e tumor extract with com m ercial fetuin. Fetuin shows a
single, broad arc with a m obility slightly faster than that of the
tumor extract antigen. By this technique, the heterogeneity is the
sam e as observed in immunodiffusion studies: the PBS-EDTA
extract of pooled W ilm 's tum ors and fetuin show a single antigenic
substance, w hereas newborn calf serum contains two antigenic
m aterials reacting with R5, only one of which c r o ss-r e a c ts with
the tumor extract antigen.
The m obility of these antigens was estim ated by com
parison with the m obility of bovine albumin under the sam e condi- i
tions. Figure 11 shows the reaction of electrophoresed newborn
!
calf serum and rabbit antiserum to bovine serum . The distance ‘
from the origin to the prominent albumin arc was used as a
standard referen ce of m obility. Since the g els represented in
F igures 9, 10 and 11 w ere all part of the sam e com posite gel, and
w ere therefore electrophoresed sim ultaneously under identical
F ig. 9. Im m unoelectrophoresis of Pooled
W ilm 's Tumor Extract and Newborn
Calf Serum.
Trough: R5 (absorbed)
Top w ell (CS): Newborn calf serum
(diluted 1:10 with
IEP buffer)
Bottom w ell (WT): PBS-EDTA extract
of pooled W ilm 's
tum ors (diluted 1:3
with IEP buffer)
66

F ig. 10. Im m unoelectrophoresis of Pooled
W ilm 's Tumor Extract and Com m er
cial Fetuin.
Trough: R5 (absorbed)
Top well: C om m ercial fetuin
(1% w /v in IEP buffer)
Bottom w ell (WT): PBS-EDTA extract
of pooled W ilm 's
tum ors (diluted 1:3
with IEP buffer)
I
I
!
68

Fig. 11. Im m unoelectrophoresis of Newborn
Calf Serum and Pooled W ilm 's Tumor
Extract Against A nti-Bovine Serum
Antiserum .
Trough: Rabbit antiserum to bovine
serum (absorbed with pooled
human plasma)
Top w ell (CS): Newborn calf serum
(diluted 1:10 with
IEP buffer)
Bottom w ell (WT): PBS-EDTA extract
of pooled W ilm 's
tum ors (diluted 1:3
with IEP buffer)
70

72
conditions, the distance from the origin to a given antigen arc can
be directly com pared to obtain this relative m obility to albumin.
The values observed w ere as follows:
Antigen Source
D istance from origin to arc cen ter/
D istance from origin to albumin arc
PBS-EDTA extract of
pooled W ilm 's tumor 0. 61
C alf Serum
(fast component) 0. 61
C alf Serum
(slow component) 0. 53
C om m ercial fetuin 0. 69
Albumin in newborn
calf serum O
o
T hese values w ere con sisten t to within 2% variation in other
experim ents of the sam e design.
Sim ilar values w ere calculated from established serum
patterns (74) to determ ine the relative m obility range, based on
serum albumin, for a and 3^ m ob ilities. These w ere
as follows:
M obility
D istance from origin to arc cen ter/
D istance from origin to albumin arc
“ 1
0. 66 - 0. 92
“ 2
0. 36 - 0. 75
*1
0. 02 - 0. 48
73
The antigens observed in these studies could be placed in the
following categories:
PBS-EDTA extract of pooled W ilm 's tumors: “2
Calf serum (fast component): “2
Calf serum (slow component): < * 2
C om m ercial fetuin: a2 “ al
Identical resu lts w ere obtained using human albumin as
a referen ce m obility (not shown). In this ca se, goat antiserum to
human serum was reacted against electrophoresed PBS-EDTA
extract of pooled W ilm 's tum ors. Since this contained human
albumin, this prominent arc w as used for a referen ce point to
determ ine the m ob ilities of the other antigens.
By these studies, then, antigens in newborn calf serum
and PBS-EDTA extract of pooled W ilm 's tum ors showed identical
electrophoretic m ob ilities in the “2 range. C om m ercial fetuin
dem onstrated an antigen of faster m obility in the “2 or “1 range.
The second antigen in newborn calf serum showed a slow er
m obility, s till in the range. No additional sp ecies of antigen
w ere detected by this method, so the heterogeneity in th ese studies
was the sam e as detected in immunodiffusion. M aterial w as not
available in sufficient quantity from the PBS-EDTA extract of
cultured W ilm 's tumor c e lls to study by im m unoelectrophoresis.
The presence of a second antigen in newborn calf serum ,
detected by R5, an antibody to a tumor extract, r a ise s a puzzling
question, sin ce the antigen was not detected in the im m unizing
extract. An attempt was made to detect this antigen, or others,
in the im m unizing PBS-EDTA extract of pooled W ilm 's tum ors,
using rabbit antiserum prepared against bovine serum in immuno-
electrop h oresis. C om m ercially obtained antiserum to bovine
serum was shown to react against electrophoresed human serum
with a number of precipitin arcs. Since human serum was present
in the PBS-EDTA extract of pooled W ilm 's tum ors, this rabbit
antiserum was absorbed with human plasm a (as described in
MATERIALS AND METHODS) prior to reacting it against the tumor
extract. Figure 11 shows the reaction of this absorbed antiserum
with newborn calf serum and PBS-EDTA extract of pooled W ilm 's
tum ors, following electrop h oresis as before. While the antiserum
dem onstrates numerous reactions with the calf serum , no apparent
reaction occurs with the extract of the tumor. A very faint stain
is observed in the region of albumin m obility, which may be due to
incom plete absorption, since there was a strong reaction of this
unabsorbed antiserum with human serum albumin; this does not
appear as a w ell-defined arc, and may not even be an antigen-
antibody precipitate. In a sim ilar experim ent, this absorbed anti
serum was allowed to react against fetuin following electrop h oresis.
No precipitin arc was observed (not shown).
The antiserum to bovine serum did not, then, dem onstrate
either the cro ss-rea ctin g antigens of the PBS-EDTA extract of
pooled W ilm 's tum ors and of fetuin; nor did it dem onstrate, in the
tumor extract, the second antigen observed in newborn calf serum ,
using R5. The bovine serum used to r a ise this com m ercial
antiserum is not described, and m ay have been exclu sively of adult
origin. If the antigens detected in newborn calf serum using R5 are
restricted to fetal stages, as may be the case with fetuin (75), then i
i
this antiserum might not detect them. This study does corroborate, |
|
however, that the PBS-EDTA extract of W ilm 's tum ors used for j
im m unization to obtain R5 was not generally contaminated with !
j
bovine serum antigens, sin ce these would be detected with the j
present antiserum . This point is also supported by the fact that
only two antigens w ere detected in newborn calf serum using R5. i
The source and identity of this second calf serum antigen is not
|
clear at this tim e.
C. A crylam ide Gel E lectrop h oresis Studies
A crylam ide gel electrop h oresis was employed to determ ine
the relative m ob ilities, in this medium, of antigenic m aterial from j
the PBS-EDTA extract of pooled W ilm 's tum ors, newborn calf j
serum and com m ercial fetuin. G els loaded with these antigen p re- j
i
parations and a control gel with only water added w ere run sim u l- j
!
taneously, rem oved, and sectioned (or stained) as described in j
MATERIALS AND METHODS. The eluates of these sections w ere
concentrated by lyophilization, and reconstituted with water to a
uniform volum e. These w ere analysed for antigenic activity by
double diffusion against R5 absorbed as before.
Figure 12 sum m arizes the resu lts obtained. Sections are
aligned with C oom m assie Blue stained gel patterns, and those
showing antigenic activity are indicated by a (+). When a high
concentration of antigen was evident by the position of the
F ig. 12. Acrylam ide Gel E lectrophoresis of
W ilm 's Tumor Extract and C om m ercial
Fetuin.
Protein staining patterns and antigen ^
distribution in acrylam ide gels:
left: PBS-EDTA extract of pooled
W ilm 's tum ors (water dialysed,
representing water soluble
m aterial from 25 y l original
extract per gel)
right: C om m ercial fetuin (100 y g per gel)
76
JfsX '.
I precipitin band in immunodiffusion, this is indicated by (++).
; Antigen activity in the gel containing fetuin re sid e s in a m ore
anodic region than the activity in the gel containing the extract of
pooled W ilm 's tum ors. This p arallels the finding in im m uno-
; electrop h oresis, that the m obility of the antigen in com m ercially
obtained fetuin is slightly greater than that in the tumor extract.
The gel containing ca lf serum failed to electrop h orese properly,
as judged by the staining pattern (not shown) and w as not able to
be com pared. Duplicate g els w ere obtained, however, within one
c la ss of sam ple. Unlike electrop h oresis in agar, the antigenic
activity m igrated with or ahead of albumin, in the ca se of fetuin;
and only partially behind albumin in the extract of W ilm 's tum ors.
No distinct staining bands w ere apparent in the region of antigenic
activity, and much activity seem ed to be in regions of the gel not
staining at all with C oom m assie Blue. L o ss of the antigen during
' the staining p ro cess can be ruled out, sin ce 12. 5% trich loroacetic
acid (the staining solvent) w as shown independently to precipitate
fetuin. Some caution should be ex ercised in interpreting th ese
electrophoretic stu d ies. M obilities of a given m aterial can change
depending upon the com position of the m edium in which they
1 m igrate. Though th ese studies w ere alw ays perform ed sim u l
taneously in identical g els, the amount of protein in the extract of
the tumor w as generally greater than that in the rela tiv ely uniform
fetuin preparation.
79
D. Iso electric Focusing Studies
To determ ine the approximate iso e le c tr ic point of the
antigens observed, and as an additional criterion of heterogeneity,
iso e le c tr ic focusing in acrylam ide gels was utilized. The su p er
natant (1500 x g, 30 m inutes, 4° C) of w ater-d ialysed newborn calf
serum , of the PBS-EDTA extract of pooled W ilm 's tum ors, and of
fetuin w ere incorporated into g els as described in MATERIALS
AND METHODS, in the p resen ce of ca rrier am pholytes ranging in
iso e le c tr ic point from pi = 3 to pi =6. These w ere run sim u l
taneously, and the g els w ere either stained or sectioned, as
described. Sections from four gels w ere pooled and eluted with
w ater to determ ine the pH gradient in the gel and to obtain enough
m aterial to analyse by immunodiffusion as before.
Figure 13 and 14 sum m arize the resu lts of one experim ent,
providing inform ation on the distribution of a n ti g e n i c activity,
staining pattern and pH gradient in the g els. The pH gradient in
th ese gels was reproducible and linear in the range from pH = 3. 3
to pH = 5. 2. No variations w ere seen between sam p les when
duplicate g e ls w ere tested. Antigenic activity was detected in the
regions shown (+), dem onstrating a range of pi = 3. 2 to pi = 4. 2
for the antigen in fetuin and a range of pi = 3. 4 to pi = 4. 2 for the
antigen from the extract of W ilm 's tum ors. A ntigenic activity was
not recovered from the gel containing newborn calf serum , though
am ple m aterial was loaded to allow detection. The pH gradient in
the g els containing calf serum did not vary from that of the other
g els, indicating that failure to recover antigen w as not due to an
Fig. 13. Iso electric Focusing Pattern of Pooled
W ilm 's Tumor E xtract and C om m ercial
Fetuin.
P rotein staining pattern and antigen
distribution in acrylam ide g els after
iso e le c tr ic focusing:
G els contained:
BI: water
WT EXT: PBS-EDTA extract of pooled
W ilm 's tum ors (water dialysed,
representing w ater-solu b le
m aterial from 33 Ul of original
extract per gel)
Fet: C om m ercial fetuin 80 yg per gel
CS: Newborn calf serum (water dialysed,
20 y l per gel)
80
C O
1III 111 1 1 1 » 1 1 1 1 1 »11 11111 1 1 111 1111 11
C O
LLi
+ + + + + + + +
++++++++++++++
i i i i i i i i i i i i i i i ............. i i i i i i i n
gas?*- ' - - f
1
1 1 1
u.
+++++++++++ +
I m i l L i-i.-i_i-.ij I - I -
................ i i i i i i i i i n i i i _ i
F ig. 14. Iso electric Distribution of Antigen from
W ilm 's Tumor Extract and C om m ercial
Fetuin.
pH gradient in acrylam ide g els after
focusing antigen (+) from PBS-EDTA
extract of pooled W ilm 's tumor ( • -----•);
com m ercial fetuin (o o); and
newborn calf serum (*-------- *). G els
are identical to those represented in
Figure 13.
82
5j 0-
4.5-
3.5-
AG FET
AG WT EXT
GEL SECTION
im proper establishm ent of a focusing gradient. By this technique,
the antigens from the fetuin and tumor extract appear to have
virtually identical iso elec tric ranges. The stained gel patterns
(Figure 14) suggest that the breadth of this pH range is not due to
incom plete or discontinous focusing, since sharp bands appear in
the region of antigenic activity in the fetuin gel. A lso, bands
appearing in the other g els have a fairly distinct structure. It is
interesting that while antigen activity is apparent in the region
shown of the gel containing the tumor extract m aterial, no staining
is observed. The amount of antigenic m aterial applied to (and
recovered from) the g els containing fetuin was greater than that of
the gels containing the tumor extract m aterial. This m ay account
for the observed difference in staining, sin ce the m aterial in the
tumor extract could be below the lev el of m aterial n ecessa ry to
stain. The presen ce of numerous bands in the antigen region of
the fetuin gel would suggest, however, that a fam ily of sp ecies
ex ists sharing common antigenic m o ieties, but differing in the
ionizable groups determ ining the iso elec tr ic point of each sp ecies.
This may be true of the antigen region of the gel containing tumor
extract as w ell, though it is not observable in this case. This
technique has dem onstrated no apparent difference in the is o
electric properties of the antigen from the PBS-EDTA extract of
pooled W ilm 's tum ors, and that from the com m ercial preparation
of fetuin.
In a separate experim ent, em ploying a 2% ampholyte con
centration and a total run tim e of five hours, gels containing
85
w ater-d ialysed , newborn calf serum w ere electrofocused and
p rocessed as above. In this ca se, shown in Figure 15, antigenic
m aterial from this calf serum was distributed in the sam e range
of the pH gradient as was antigenic m aterial from both the tumor
extract and com m ercial fetuin in the previous experim ent. Again,
this m aterial dem onstrated a fam ily of antigenic ally related, but
iso elec trica lly d iscreet and d isp erse sp e cie s, as did the antigens
from the above sou rces.
From this study two points em erge: 1) no difference w as
observed in the iso elec tric points determ ined for antigens from the
pooled tumor extract, newborn calf serum or com m ercial fetuin;
and 2) all of th ese antigens w ere apparently represented by a
c la ss of sp ecies showing distinct, but clo se ly related iso e le c tric
points. No conclusions w ere made about the relative distribution
of antigenic m aterial from th ese various sources in a given is o
e lectric c la ss, sin ce a quantitative evaluation of antigen activity
or of protein staining m aterial was not obtained.
In the gel containing newborn calf serum , a sin gle section
(representing an iso e le c tr ic point near 3. 0) was shown to contain
an antigen not c r o ss reacting with antigens eluted from the other
section s of the sam e gel. This m ay rep resen t the second antigenic
component observed in ca lf serum using imm unodiffusion and
im m unoelectrophoretic techniques. This m aterial was not subjected
to further experim ents to confirm th is, sin ce it was not available
in sufficient quantity. It cannot, at this point, be equated to the
second calf serum component with certainty.
F ig. 15. Iso elec tr ic Focusing of Antigen from
Newborn C alf Serum.
pH gradient in acrylam ide g e ls after
focusing antigen (+) of newborn calf
serum (w ater-dialysed 20 y l per gel).
P rotein staining pattern shown in
in set with antigen activity (+).
86

M olecular Sieve Chromatography
As a m eans of comparing the m olecular size of the
antigens detected in the PBS-EDTA extract of pooled W ilm 's
tum ors and in com m ercially obtained fetuin, chromatography on
G -75 Sephadex was em ployed. Figure 16 illu stra tes the resu lts
obtained, indicating the elution profile (absorbance at 280 nm, and
antigenic activity) of th ese chromatographed sam ples. To obtain
the void volume and volume of com plete retention, a m ixture of
Blue Dextran 2000 and ferric chloride was chromatographed on
the sam e column. The volume of all sam ples applied to this
column was uniform (0. 50 m l).
Antigenic activity is observed in the sam e elution volume
upon chromatography of either fetuin or the supernatant (1500 x g,
30 m inutes, 4° C) of the w ater-dialysed PBS-EDTA extract of
pooled W ilm 's tum ors. Rechrom atography of the antigen containing
fractions of the tumor extract profile resulted in an identical
elution position for this antigen, indicating an independence from
the effect of other m aterials in the crude tumor extract on the
antigen's elution behavior. The antigenic activity elutes at a peak
of absorbance both in the case of fetuin and the extract of W ilm 's
tumor. If contaminating protein sp ecie s are present in either case,
this technique would probably not be useful in separating them from
the antigen.
The exclusion lim it for globular proteins is approxim ately
70,000 daltons for Sephadex G -75 (76). Since the antigen elutes
after void volume, this places its maximum m olecular weight
F ig. 16. G -75 Sephadex Chromatography of
Antigen M aterial in Pooled W ilm 's
Tumor Extract and in Fetuin.
Absorbance (280 nm) of eluted fractions.
Sample volume applied:
0. 50 m l in 0. 05 M NaCl pH = 6. 5
Arrow indicates elution position.
(+) indicates antigen activity.
A: F erric chloride/B lue dextran m ixture
B: Serum albumin MW approxim ately
70, 000 daltons
C: Ovalbumin MW approxim ately
45, 000 daltons
D: C om m ercial fetuin
E: PBS-EDTA extract of pooled W ilm 's
tum ors (w ater-dialysed to rem ove
water insoluble fraction) (•— •);
rechrom atography of three antigenic
peaks (o o).
89
90
|------- ]------p----1 ----- <
•M
M
'~ * y —1»— -j»-»-j-*-*
M l
I I
m
■ •« g C o o o o & OOP—o c - c - olS-0-5=ft=8=S«»*
FRACTION NUMMR
91
somewhat below this value, assum ing it does not show aberrant
behavior. This is a fair assum ption for fetuin, since its structure
is known to be a prolate ellipsoid (77).
No difference was observed, then, in the m olecular size
of the antigens from com m ercial fetuin, and from the PBS-EDTA
extract of pooled W ilm 's tum ors. The elution behavior is not
inconsistent with the known m olecular weight of fetuin (48, 000
daltons).
Sensitivity of the Antigen to Enzym atic Degradation
To explore the chem ical nature of the antigenic m aterial
observed, its sen sitivity to enzym atic degradation was determined.
The PBS-EDTA extract of pooled W ilm 's tumor, which demon
strated this antigen, was extensively dialysed against PBS, and
enzym atically treated as described in MATERIALS AND METHODS.
Figure 17 indicates the effect of enzym atic digestion on
this antigenic m aterial. Pronase and trypsin digestion (24 hours,
37° C) of the extract prevented formation of the precipitin line
between the treated extract and antibody R5, absorbed as before.
In contrast, incubation with ribonuclease and neuram inidase, for the
sam e period and tem perature, did not affect the reaction between
this extract and R5. In the case of hyaluronidase, a slight shift in
the position and clarity of the precipitin line indicated a possible
alteration of antigen following incubation of the extract with this
enzyme.
Fig. 17. Enzym atic Degradation of Antigen
from W ilm 's Tumor E xtract-D irect
A ssay.
Center well: R5 (absorbed)
P eriph eral w ells:
PBS-EDTA extract of pooled
W ilm 's tum ors (PBS-dialysed
and diluted 1:4 with PBS) after
treatm ent with (1) pronase,
(2) hyaluronidase, (3) trypsin*
(4) neuram inidase, and (5)
ribonuclease.
93
1 ^ '
94
To exam ine the p ossib ility that residual enzym e activity
was present in these treated extracts, which in som e way could
have affected the precipitin reactions observed, enzym e solutions
w ere incubated alone under identical conditions and diffused into
the peripheral w ells shown in Figure 18. Following this diffusion,
antibody R5 was added to these w ells, and the center w ell was
filled with the PBS-EDTA extract of the pooled tum ors. This
extract had also been incubated alone under identical conditions of
the digestion procedure. The immunodiffusion pattern shown
indicates a precipitin pattern unaffected by any of the prediffused
enzym es, and confirm s that residu al enzym atic degradation of
antigen or antibody is not occurring. In addition, this shows that
antigen incubated alone under the conditions of enzym atic digestion
does not lo se its activity, since such antigen (in the m iddle well)
dem onstrates a typical precipitin reaction. The figure shows
antigen incubated at pH = 7. 5 in the center w ell, though when
incubated at pH = 5. 3 (as a control for the hyaluronidase treatment)
the sam e resu lts w ere obtained.
The digestion products resulting from the incubation of
each of th ese enzym es with the tumor extract w ere tested for their
ability to inhibit the reaction between R5, absorbed as above, and
enzym atically untreated tumor extracts (again incubated without
antigen under the condition identical to enzym e digestion). The
enzym e d igests of the extract w ere prediffused into the peripheral
w ells shown in Figure 19. Following this, R5 was added to each
of these w ells and the control tumor extract was placed in the center
Fig. 18. Lack of Interfering Enzym atic A ctivity
in Antigen D igests.
Center well: PBS-EDTA extract of
pooled W ilm 's tum ors
(PBS-dialysed, diluted
1:4 with PBS); incubated
alone under conditions
of enzym e digestion
P eripheral w ells:
Enzym es, incubated without antigen,
w ere prediffused into w ells (1) pronase,
(2) hyaluronidase, (3) trypsin, (4)
neuram inidase, and (5) ribonuclease.
R5 (absorbed) was then added to all
peripheral w ells (1-5).
95
96
fj? v
♦ ' * fk U V & iJP in ,
F ig. 19. Enzym atic Degradation of Antigen from
W ilm 's Tumor Extract-Inhibition A ssay.
Center well: PBS-EDTA extract of pooled
W ilm 's tum ors (PBS-dialysed
and diluted 1:4 with PBS)
incubated alone under condi
tions of enzym e digestion
P eripheral w ells:
D igestion products from enzyme
treatm ent of PBS-EDTA extract
of pooled W ilm 's tum ors (PBS-
dialysed and diluted 1:4 with PBS)
w ere prediffused into peripheral
w ells (1) pronase digest, (2)
hyaluronidase digest, (3) trypsin
digest, (4) neuram inidase digest
and (5) ribonuclease digest. R5
(absorbed) was then added to all
peripheral w ells.
97
98
99
w ell. Only the d igests of trypsin and pronase fail to inhibit the
reaction of R5 with the control tumor extract, w hereas the d igests
of the other enzym es com pletely inhibit this reaction. This
indicates that absorption of the antibody occurs in these other ca se s
and shows that antigen was not altered by these enzym es in a way
which would elim inate their binding with antibody. Trypsin and
pronase apparently did alter the antigenic structure to render it
unable to interact with antibody.
It should be noted that the reaction products of hyaluroni
dase incubation failed to react fully with the ex cess antibody in
the w ells containing pronase and trypsin products (1 and 3 in the
figure). This would suggest again that partial alteration of the
antigen had occurred as a resu lt of incubation with this enzym e.
The hyaluronidase preparation was not analysed for the p resen ce of
proteolytic enzym e contaminants, so no conclusions w ere drawn
regarding the effect of this enzym e on the antigen, other than
g ro ss alteration of the antigen was not observed using this enzym e
preparation.
Since pronase and trypsin w ere shown to inhibit the
reaction of antigen with R5, both by a d irect immunodiffusion
assay and by an inhibition technique; it is suggested that the protein
m oiety of the antigen is n ecessa ry for the integrity of the antigenic
site(s) interacting with antibody. R ibonuclease, neuram inidase and
hyaluronidase had little or no effect on the antigen, which suggests
that the antigenic structure recognized by the antibody is not
dependent on portions of the m olecule susceptible to these enzym es.
100
Whether the actual antigenic determ inant recognized com p rises a
protein structure cannot be answ ered by this experim ent, though
it is a possibility.
Amino Acid A nalysis of Fetuin and E lectrophoretically Purified
Antigen from the PBS-EDTA Extract of W ilm 's Tum ors
D eterm ination of the amino acid com position of a purified
preparation of antigen from the W ilm 's tumor extract was pursued
in order to ch aracterize the protein m oiety of this m aterial, and
to provide a m eans of com paring chem ically this antigen to the
antigen in com m ercial fetuin.
DEAE-Sephadex chromatography was in itially employed
to fractionate the crude tumor extract, having been found superior
to m olecular siev e chromatography for this purpose. W hile a
single fraction (of eight) containing the antigenic m aterial was
obtained from the eluate of this chromatographic column (see
Appendix 2); this fraction appeared heterogeneous by acrylam ide
gel electrop h oresis. For this reason, d irect elution of antigenic
m aterial from electrophoresed g els containing w ater-dialysed
tumor extract (prepared as described under A crylam ide Gel
E lectrophoresis) was em ployed to obtain m aterial for amino acid
an alysis. M ultiple g els w ere electrophoresed as before, sim u l
taneously. After electrop h oresis, these w ere sectioned and
corresponding section s of th ese g els w ere pooled and eluted with
water. These eluates w ere analysed for antigenic activity, which
w as found to be slightly diffuse over a region of the g el (see
101
Figure 12). The three pooled section s showing the highest antigen
concentration (judged by imm unodiffusion patterns) w ere con
centrated and hydrolysed as described in MATERIALS AND
METHODS. This pool represented approxim ately 0. 5 m g protein
by a Lowry determ ination.
S im ilarly, an aliquot (4. 0 m g protein) of com m ercial fetuin
was hydrolysed. Both sam ples w ere subjected to amino acid
analysis as described, and a standard of amino acids was run to
quantitate the amino acids in the hydrolysed product. Figure 20
lis ts the resu lts of this analysis. In the case of the e lectro
phoretically purified antigen of the tumor extract, glycine, present
in the electrop h oresis buffer, was in e x c e ss and not quantitatable.
Due to the breadth of the glycine peak, valine was not separated,
and therefore could not be quantitated either. No evidence of
m ethionine was apparent, though this peak, if in sm all quantity,
would have been hidden by an abnorm ally large buffer-change peak.
The rem aining amino acids w ere used for com parison of the
electrophoretically purified antigen with whole, com m ercial fetuin.
This was accom plished by obtaining values of each amino acid
relative to aspartic acid, a relatively stable amino acid under these
hydrolysis conditions. This ratio w as com pared for the purified
tumor extract antigen and for com m ercial fetuin. Of the amino
acids listed for both antigen hydrolysates; lysin e, histidine,
threonine, tyrosine and phenylalanine showed different com positional
ratios, outside the experim ental erro rs. Notably, the absence of
tyrosine and the low value of proline distinguished the tumor
Fig. 20. Amino Acid A nalysis of Fetuin and
E lectrophoretically-P urified Antigen
from Pooled W ilm 's Tumor Extract.
103
A. A. ytfm oles (+ 2. 5%) ^ m o le s / m ole asp (+ 5 . 0%)
WT FET WT FET
L ys 0. 058 0. 155 1. 798 + 0. 090 0. 696 + 0. 035
His 0. 011 0. 097 0.344 + 0. 017 0.435 + 0. 022
n h 3 1. 374 0. 426 42.750 + 2. 137 1. 916 + 0.096
Arg 0. 020 0. 132 0. 614 + 0. 031 0. 592 + 0. 030
Asp 0. 032 0. 222 1. 000 + 0. 000 1. 000 + 0. 000
Thr 0. 017 0. 136 0. 520 + 0. 026 0. 610 + 0. 031
Ser 0. 021 0. 156 0. 658 + 0. 033 0. 702 + 0. 035
Glu 0. 035 0. 222 1. 076 + 0. 054 0. 998 + 0. 050
Pro 0. 018 0. 253 0.552 + 0. 028 1. 136 + 0. 057
Gly 0. 165
o,.
i '
0. 741 + 0. 037
Ala
1/2
V
0. 255 1. 145 + 0. 057
Cyst 0. 014 0. 096 0.442 + 0. 022 0. 430 + 0. 025
Val 0. 029 0. 246 0. 903 + 0. 045 1. 106 + 0. 055
Met
***
0. 024
« A . O * s V
'1 ' *V ' l Z
0. 109 + 0. 005
lieu 0.014 0. 090 0.439 + 0. 022 0. 404 + 0.020
Leu 0.027 0. 201 0.839 + 0. 042 9.906 + 0. 045
Tyr trace 0. 093 trace 0.419 + 0. 021
Phe 0. 013 0. 109 0.413 + 0. 021 0. 490 + 0. 025
‘ 'Glycine buffer present; reading off scale.
Hidden under shoulder of glycine peak.
'‘ ' ‘ " ‘ Hidden under shoulder of buffer change peak, if present at all.
104
extract antigen from com m ercial fetuin. W hile lysin e also differs
m arkedly, this chromatographic peak eluted abnorm ally, in the
case of the tumor extract antigen, and should not be included in
quantitative com parisons.
Additional verification of the com position of fetuin was
obtainable from the literature. The ratio of amino acids to
aspartic acid was calculated from the data of Spiro and Spiro (78)
and of F ish er et al. (79). T hese are presented in Figure 21,
accompanying the sam e ratios for the experim ental values in
Figure 20 and for serum albumin, calculated from the data of
Shultze (80). In com paring the experim entally obtained amino
acid com position of com m ercial fetuin (isolated by the method of
Spiro) with the fetuin com position reported by Spiro and Spiro;
the amino acids tyrosine, phenylalanine, lysin e, histidine and
arginine appear elevated in the com m ercially prepared fetuin
studied here.
In com paring the com position of fetuin reported by Spiro
and Spiro with that reported by F ish er et a l .; the amino acids
threonine, tyrosine, 1 /2 -cystin e, histidine and arginine differ
significantly. Since the antigen isolation from the electrop h oresis
g els cannot be directly com pared to any of th ese chem ical m ethods
of purification, a com positional com parison was made between the
tumor extract antigen and fetuin, including com posite data from
the two reported amino acid analyses of fetuin, and from the
experim entally determ ined analysis of com m ercially obtained fetuin.
The notable lack of tyrosine and the low value of proline still
Pig. 21. Com parison of Experim ental Amino
Acid A nalysis of Antigen from W ilm 's
Tumor Extract; and Reported Com
positions of Fetuin.
106
A. A. m oles of amino a cid /m o les of aspartic acid
i F e t1 F et2 F e t3 WTE4 Albumin5
!
Asp 1. 000 1. 000 1. 000 1. 000 1. 000
Thr 0. 755 0. 597 0. 610 0. 520 0. 538
Ser 0. 794 0. 769 0. 702 0. 658 0. 461
Glu 1. 027 1. 149 0. 998 1. 076 1. 558
Pro 1. 021 1. 302 1. 136 0. 552 0.481
Gly 0. 733 0. 683 0. 741
----
0. 231
Ala
1/2
1. 012 1. 034 1. 145
— — —
1. 192
Cyst 0. 367 0. 668 0. 430 0. 442 0. 635
Val 1.221 1. 216 1. 106 0. 903 0. 788
Met none 0. 020 0. 109
----
0. 096
lie 0.452 0. 451 0. 404 0. 439 0. 154
Leu 0. 812 1. 034 0. 906 0. 839 1. 153
Tyr 0. 206 0. 284 0. 419 trace 0. 308
Phe 0. 327 0. 350 0. 490 0. 413 0. 577
Lys 0. 500 0. 582 0. 696 1. 798 1. 076
His 0. 309 0. 399 0. 435 0. 344 0. 288
Arg 0. 358 0. 582 0. 592 0. 614 0. 442
^Calculated from the data of Spiro and Spiro (78).
^Calculated from the data of F ish er et al . (79).
O
^Calculated from experim ental values of Figure 20 for
com m ercial fetuin (Spiro method).
4
Calculated from experim ental values of F igure 20 for
antigenic m aterial eluted from acrylam ide g el electrop h oresis
of the PBS-EDTA extract (w ater-dialysed) of pooled W ilm 's
tum ors.
^Calculated from the data of Schultze (80).
107
distinguish the com position of the tumor extract antigen from fetuin.
A lso, low values of threonine, serine and valine w ere apparent,
though it was noted by Spiro and Spiro (78) that these amino acids
w ere dependent strongly upon the conditions and tim e of hydrolysis.
Since only one hydrolysis tim e was used in the present study (24
hours) due to lack of m aterial, com parison of these three amino
acids should be made cautiously.
Two points em erge from this data. F irst, fetuin isolated
by different chem ical m ethods, and even isolated by the sam e
method on different occasions, dem onstrate variable amino acid
com positions. This is very probably due to the criteria of purity
applied to a fetuin preparation prior to amino acid analysis. It
has been shown in this study to be heterogeneous by iso elec tr ic
focusing. Second, the antigen isolated from the W ilm 's tumor
extract by electrophoretic purification apparently differs in amino
acid com position from fetuin as defined by chem ical purification
procedures. However, the apparent discrepancy in these com posi
tions is not definitive evidence for the non-identity of antigenic
m aterials from the tumor extract and from fetuin. It is possible,
for instance, that the electrophoretically purified antigen is a
su b -sp ecies of fetuin which would have been only part of the
population of chem ically isolated fetuin.
Fetuin, as defined by Spiro and Spiro (78), was com pletely
devoid of methionine. Unfortunately, the m ethionine peak of the
electrophoretically isolated antigen was not quantitatable. No
peak or shoulder was apparent under the buffer-change peak which
108
obscured the region of m ethionine, indicating that it w as not present
in appreciable quantity. A low level, however, would have gone
undetected. If the antigen w ere shown to contain m ethionine, this
would differentiate it from fetuin as defined by Spiro and Spiro.
However, fetuin isolated by F ish er et al. (79), and fetuin obtained
com m ercially for this study, both dem onstrated low lev els of
methionine.
A point is reached where the definition of fetuin becom es
important in a com parative study of this nature. The definition
of this m aterial widely used (73, 7 5, 81), is based on free boundary
electrop h oresis, analytical ultracentrifugation and its behavior in
chem ical fractionation sch em es. These are also used to monitor
its purity. This appears to be insufficient criteria for detailed
com parison, since obvious differences are apparent in the nature
of fetuin isolated by various techniques (73, 79, 81). By m ore
sen sitive techniques, such as iso elec tric focusing, fetuin has been
shown to be heterogeneous; this should be considered in further
characterization of this m aterial.
D eterm ination of Uronic Acid and Sialic Acid
A s a further chem ical com parison of the antigens from
the PBS-EDTA extract of W ilm 's tum ors and from com m ercially
obtained fetuin, the content of n -acetyl neuram inic acid (NANA) and
uronic acid (UA) was determ ined as described in MATERIALS AND
METHODS. The solutions used for these determ inations w ere
those described for amino acid analysis, and w ere obtained just
109
prior to hydrolysis. These represented, then, electrophoretically
purified antigen from the tumor extract, and whole com m ercial
fetuin described above. The values obtained are represented as
gram s of carbohydrate m easured per 100 gram s of protein, which
was m easured by the Lowry method described in MATERIALS AND
METHODS. The values are as follows:
g NANA/100 g protein g UA/100 g protein
Purified
WT Extract 0 .3 3.2
Fetuin 1 .6 1 .4
Fetuin.1 8.2
Fetuin2 (5.7)
■'■From the data of Spiro and Spiro (78)
2From the data of F ish er et al. (79) Calculated.
The content of NANA is m arkedly low er in the electro
phoretically purified antigen from tumor extract than in the
com m ercially obtained fetuin. Both of th ese values are consider
ably below the lev els reported for fetuin by the w orkers cited. The
lev el of uronic acid appears greater in the antigen from the tumor
extract than in the com m ercially obtained fetuin.
While differences in the sia lic acid content of fetuin have
been noted as the resu lt of various isolation procedures (73, 78, 79),
values as low as the present are generally not found u n less
rem oval of the sia lic acid is attempted by acid or enzym e hydro
ly sis . The low value of the electrophoretically isolated antigen
110
and com m ercial fetuin is not as marked, and the content of this
carbohydrate in fetuin could not be found in the analyses of others.
It does not, then, dem onstrate a difference between the tum or-
derived antigen and com m ercial fetuin in this study. The sam e
considerations may be given to these data as w ere given to the
amino acid composition: if the electrophoretically isolated antigen
w ere a su b -cla ss of cla ssica lly defined fetuin sp ecies, it could
very w ell have a different chem ical com position than com posite
sp ecies com prising fetuin.
CHAPTER IV
DISCUSSION
P roperties of the Antigen Detected
The antigen detected in these studies was common to a
pooled PBS-EDTA extract of four W ilm 's tumors; a com m ercial
preparation of fetuin; both fetal and newborn calf serum; and a
PBS-EDTA extract of cultured W ilm 's tumor c e lls. These w ere
identical by immunodiffusion studies, indicated by a line of fusion.
In com paring the physico-chem ical properties of these antigens,
som e points should be kept in mind: 1) the antigen was detected
in a ll ca ses by its ability to react with an antibody which presum ably
recogn izes a sp ecific group of antigenic determ inants on a m olecule;
for a precipitin reaction to occur, there m ust be at lea st two such
determ inants on the m olecule in question, w hile absorption of the
anti serum with antigen requires only one such determ inant per
m olecule; 2) the antigenic determ inants may be attached to a
m olecule in a number of ways, and are not n ecessa rily covalently
linked, or an integral part of the m olecule; 3) the physical and
chem ical properties of a m olecule having these determ inants may
be dictated by portions of the m olecule not structurally affecting
or not part of the antigenic determinant recognized. Thus, while
antigen-containing m olecules may differ in siz e , charge, shape,
e t c .; they may very w ell have determ inants indistinguishable by a
given antibody preparation.
I l l
112
In physically and chem ically com paring the antigens
detected in the W ilm 's tumor extract and in com m ercially prepared
fetuin (or in calf serum) the actual question asked is: "How do the
substances in these preparations associated with these common
antigenic determ inants com pare as to their physical and chem ical
nature?" This question is to be distinguished from the question:
"What is the chem ical and physical structure of the antigenic site s
recognized by a given antibody preparation, and how do these
determ inants on two m olecules com pare?" It is the form er question
which was pursued in m ost of this study, while the latter question
was only partially dealt with.
A further consideration in the present study is the se n si
tivity of detection of antigen. Quantitation of antigen w as difficult
to obtain for two reasons: 1) the double diffusion technique
employed is only sem i-quantitative, since the position of a p re
cipitin line under given conditions of diffusion is an in sen sitive
m easure of antigen concentration; and 2) no standard "purified"
antigen was available for estim ating antigen concentration. Low
lev els of antigen would go undetected by this double diffusion method.
While this is a drawback to quantitative analysis of antigen, the
studies presented w ere perform ed in the detectable range of antigen,
so the relative distribution of antigen could be com pared in any of
the techniques em ployed. Fetuin was detectable at a concentration
of eight pg per m l (80 ng added to a w ell in 0. 010 m l). This was
based on dilutions of com m ercially prepared fetuin, prepared on the
b asis of lyophilized dry weight.
113
The heterogeneity and description of fetuin is s till an
unsettled and debated problem (75). The preparation of com m ercial
fetuin used in this study showed an analytical centrifuge pattern
which indicated no heterogeneity by this criterion (see Appendix 3).
The sedim entation coefficient (20. 0° C, 0. 12 M NaCl, 1% w /v,
pH = 6.5) was s = 2. 8 Svedbergs, which is in agreem ent with the
value determ ined by Spiro (73) for a sim ila rly isolated fetuin p re
paration. This can probably be considered, then, a preparation of
fetuin free from g ross contamination.
By m olecular siev e chromatography, antigen-associated
components from com m ercial fetuin, and from the tumor extract,
dem onstrated indistinguishable m olecular size p roperties, in
agreem ent with the previously reported m olecular weight of fetuin,
45, 000 - 50, 000 daltons (73, 79). That antigenic m aterial was
associated with the absorption maximum of a single broad peak
suggests that the antigen resid e s on one m ajor s iz e -c la s s of
m olecule. In iso elec tric focusing, antigen from the tumor extract,
from calf serum and from com m ercial fetuin showed virtually
identical iso e le c tr ic ranges. By this technique, a marked degree of
heterogeneity was dem onstrated by all of these antigen-containing
preparations, indicating a fam ily of m olecu les of different is o
e lectric points, all sharing common antigenic determ inants. That
com m ercially prepared fetuin, uniform by the method of analytical
ultracentrifugation, showed such heterogeneity in this technique
illu stra tes the m acroscopic nature of sedim entation m easurem ents
as averaged values over a population of potentially slightly different
114
sp e c ie s. The difference between sp ecies in this case is due to
ionizable groups which m ight not be expected to influence greatly
the sedim entation properties of a m olecule, but which n everth eless
can be detected by electrofocusing. The differences in ionizable
groups apparently did not influence the nature of the antigenic
determ inants, sin ce antigen was detected over the entire protein
staining range of com m ercial fetuin. This range, from pH = 3. 2
to pH = 4. 2 is consistent with previously reported iso electric
points of fetuin, pi = 3. 3 (73) and pi = 4. 1 (82), determ ined by free-
boundary electrop h oresis at different pH values. This technique is
lim ited in that it m easu res the overall behavior of a population of
m olecu les. H eterogeneity of the population with resp ect to charge
w ill be m easured as a weighted average, depending upon the
relative amounts of subspecies within the population. Splitting of
the population m ay be observed at interm ediate values of iso elec tr ic
points but does not offer a sen sitive m eans of detecting these sub
sp e cies. The technique of iso ele c tric focusing has dem onstrated a
graded s e r ie s of sp ecies between these two reported iso elec tric
points. This pattern is probably not a resu lt of overloading the gel,
sin ce at low er gel loads, sim ilar staining patterns w ere observed
(in this case, however, antigen was only detectable at the position
of h eaviest staining, in the region of pH = 4). These subspecies
apparently do not rep resen t structural subunits of antigen-
associated m oieties, sin ce their precipitin bands dem onstrated
curvature sim ilar to the unfractionated antigen. This curvature
would be expected to in crease m arkedly, due to the higher diffusion
115
coefficient of a sm aller structural unit. They probably rep resen t
heterogeneity in the ionogenic groups associated with a core
structure.
After rem oval of sia lic acid by m ild acid hydrolysis, fetuin
has been shown to shift in its iso elec tric point from pi = 3. 3 to
pi = 5. 2 (73). Whether or not the sia lic acid resid u es w ere respon
sib le for the observed variation in iso elec tr ic point in the study
presented here was not determ ined. If this w ere the case, perhaps
a n on -sp ecific lo ss of sia lic acid resid u es due to hydrolysis or
enzym atic degradation in situ or during preparation was responsible
for the observed heterogeneity. The distribution of m aterial within
these subspecies of fetuin was not obtainable in quantitative term s.
This was due to lack of a quantitative antigen assay, and the
difficulty of accuratedly quantitating a protein staining pattern. In
addition, the amounts of antigen obtainable from focused acrylam ide
g els was sm all. The staining pattern observed with the com m ercial
preparation of fetuin showed denser regions in the range of higher
pH, perhaps suggesting that the bulk of the m aterial was lacking
the m ore acidic ionogenic groups. This is consistent with the
p ossib le lo ss of sia lic acid in this fetuin preparation, since it
dem onstrated only 1. 4% sia lic acid, compared to 8% reported in
other analyses of fetuin (78). Whether or not the uronic acid
content of fetuin and tumor extract contributed to the iso elec tric
heterogeneity of these substances was not determ ined.
An alternative explanation of the spread in iso elec tric
points of the antigen-associated substances in fetuin is the binding
116
of acidic ionizable groups to a core m olecule containing the antigenic
determ inants. If this binding w ere random, m olecules of fetuin
could contain m ultiple charges in a graded manner, and the resu lt
ing iso elec tric points could vary as observed. That fetuin binds
anions is strongly suggested by the observation of Spiro (73) that
the isoion ic point (2% w /v fetuin in distilled water) of fetuin,
pH = 4. 03, is much higher than the observed iso electric point of
this preparation, pi = 3. 3. The "isoionic point" was raised to
pH = 4. 21 in the presence of 0. 15 M KC1, dem onstrating a p referen
tial anion binding. This type of binding, then, m ust be considered
as a possible source of iso elec tric heterogeneity in the present
study. It may also be important with resp ect to the biological
activity of fetuin, reported by F ish er e t a l. (79). The activity of
fetuin, in inducing c e ll spreading in tissu e culture, was rem oved
by passing a preparation of fetuin through D E A E -cellulose, an
anion exchanger. The fetuin fraction (as defined by electrophoretic
and sedim entation properties) was, however, recovered alm ost
com pletely. Lieberm an et al. (83) reported that the biologically
active fraction of crude fetuin was retained on D E A E -cellulose,
and eluted after the fetuin fraction. The ionogenic groups which
may bind to or be part of fetuin m ust be considered in a ssessin g its
m olecular properties and its biological function.
Studies on the electrophoretic behavior of antigen-
associated m aterial from com m ercial fetuin and W ilm 's tumor
extract indicate an apparent difference in the properties of these
substances. Both in agar and acrylam ide gel electrop h oresis, the
117
antigen on fetuin showed a higher m obility (anodic) than the antigen
in the tumor extract. While this may be due to an increased charge
density in the case of the fetuin antigen, other explanations are
feasib le. Since the m olecular size and the iso elec tric properties
of these antigens are indistinguishable, no great difference in the
electrophoretic m obility would be anticipated, even in a technique
utilizing m olecular sieving such as in acrylam ide g els. D ifferential
binding to the support medium or differences in m olecular shape
could explain the observed difference in m igration in an electric
field of these two antigens. While the nature of this interaction
could not be determ ined here, it could serve as a b asis of distin
guishing the antigen-associated m aterials from the two sources.
Another explanation of the observed m igration of these m aterials
is the dependence of electrophoretic m obility on accompanying
substances in the electrophoresed preparation. Fetuin, relatively
uncontaminated in its electrophoretic pattern, differs from the
tumor extract which dem onstrates numerous components accom
panying the antigen-associated m aterial. These components could
alter the relative electrophoretic m obility of the antigen by
increasing the ionic strength of the medium, or by changing the
resista n ce of the medium leading to an altered voltage gradient,
both of which affect directly the m obility of a m aterial in an electric
field. That this type of phenomenon m ay be occurring is suggested
by the behavior of antigen in whole calf serum , which m igrates with
a m obility identical to that of the antigen in the tumor extract. Both
of these antigen-containing preparations (calf serum and tumor
118
extract), show many contaminating m aterials (serum proteins). If
the antigen in calf serum , showing im m unological identity with
fetuin, is indeed fetuin in its native state, then its electrophoretic
m obility is identical to that of the tumor extract antigen. The
com m ercially prepared fetuin, perhaps because of its lack of con
taminating substances, or because of alteration during purification,
m ight dem onstrate the altered electrophoretic m obility observed.
B ecause of these considerations, it would be unfounded to conclude
at this tim e that the antigens from com m ercial fetuin and from the
tumor extract w ere different due to their observed electrophoretic
properties.
Regarding the chem ical makeup of the antigenic d eter
m in an ts) recognized by the antibody preparation used in these
studies, som e information can be presented. The determinant(s)
are apparently not affected by heat (56° C, one hour), repeated
freeze-thaw ing, or pH ranging from 2. 7 (3% v /v glacial acetic
acid) to 8. 9 (acrylam ide gel electrop h oresis). These common
properties w ere shared among the antigens from com m ercial fetuin,
W ilm 's tumor extract and newborn calf serum . That the d eter
m inants w ere affected by pronase and trypsin digestion was
apparent in the antigen from the tumor extract. Whether this was
a resu lt of p roteolysis of the actual determinant or alteration of
other portions of the m olecule affecting the determinant was not
determ ined. The lo ss of reactivity of the determinant(s) was not,
however, due m erely to the cleavage of the antigen m olecule into
119
non-precipitating segm ents, since the protease digestion products
did not inhibit the reaction of antibody with intact antigen.
The role of uronic acid-containing m oieties in the makeup
of the antigenic determ inant, or in the antigen-associated m aterial
is not clear. Hyaluronidase showed a slight degradative effect on
the antigen in the crude tumor extract. This may have been due
to proteolytic enzym e contamination of this enzym e preparation;
this was not explored. The p ossib ility that antigen-containing
m aterial can a sso cia te with uronic acid-containing substances
is indirectly suggested by a number of observations: 1) Fetuin,
as w ell as other acid glycoproteins, preferentially bind anions.
2) In DEAE-Sephadex chromatography, the antigen-containing
fraction dem onstrated an uronic acid to protein weight ratio of
0 .1 6 , four tim es higher than the ratio of the crude extract. This
was a fraction shown to be extensively contaminated with other
proteins in addition to the antigen. 3) T his sam e fraction demon
strated a ratio of absorbance at 240 nm to absorbance at 280 nm of
approxim ately 5, much higher than any of the other fractions. This
could resu lt from the association of the antigen with a high
m olecular weight polym er such as hyaluronic acid, giving increased
light scattering at low er w avelengths. 4) After acrylam ide gel
electrop h oresis, the antigen-containing fraction contained 3%
uronic acid on a protein weight b asis.
If the antigen is associated with hyaluronic acid, or other
uronic acid-containing m ucopolysaccharides, it can apparently be
ea sily separated from m ost of it, since 3% aqueous acetic acid
120
precipitates 90% of the uronic acid in the W ilm 's tumor extract
and leaves the antigen in the supernatant. Furtherm ore, under
condition of immunodiffusion, the antigen dem onstrates a diffusion
coefficient in agar greater than that for IgG, inconsistent with its
association with a large m olecule. The uronic acid associated
with the electrophoretic ally isolated antigen may be residual
impurity, or m ay be part of the antigen-containing m olecule. The
1. 5% uronic acid content of com m ercial fetuin is lower than the
value observed for the tumor extract antigen; but since it had been
isolated chem ically, this may only rep resen t differences in
residu al uronic acid contaminants, left by these procedures of
isolation.
The potential im portance of hyaluronic acid and other
m ucopolysaccharides in m alignant d isea se has been d iscu ssed at
length by Cameron (71). That these substances can interfere with
cellular antigen recognition (84) and c ell invasion and interaction
(71, 85, 86) has been shown; and indicates the need to exam ine the
association of these substances with antigens such as described
in the present study. The antigenic determinant(s) recognized on
the m aterials studied here w ere not shared by hyaluronic acid,
since hyaluronic acid neither reacted with antibody preparation R5,
nor inhibited the reaction of this antibody with antigen.
The presence of an abnormal uronic acid-containing m uco
polysaccharide protein com plex in the serum of W ilm 's tumor
patients has been reported (63, 87). The relationship between this
m aterial and the present antigen was not extensively studied; since
121
sera from th ese patients, containing this m aterial, w ere not
available to pursue antigen detection and isolation. The abnormal
component in th ese W ilm 's tumor sera was precipitable in 3%
aqueous acetic acid, as was the uronic acid-containing m aterial
in the W ilm 's tumor extract. Since antigen was not precipitable
by this technique, it is unlikely that the antigen could be strongly
associated with the m ajor uronic acid-containing sp ecies, though
it is p ossib le that they associate in untreated serum .
The amino acid com position of antigen-associated m aterial
was pursued because, from enzyme hydrolysis studies, the protein
portion of the m olecule appeared essen tia l to the integrity of the
antigenic determ inant(s). Though the relative amino acid com
positions of antigenic m aterial from com m ercial fetuin and from
the W ilm 's tumor extract differed (esp ecially with resp ect to
proline and tyrosine) som e reservations ex ist about the meaning
of this difference:
1) Preparations of fetuin, hom ogeneous by physico
chem ical m ethods showed variable amino acid com positions, as
reported in the literature (78, 79). The preparation of com m ercial
fetuin used in th ese studies, homogeneous by analytical ultracentri
fugation showed an amino acid com position in disagreem ent with
both the fetuin preparations cited.
2) The tumor extract m aterial eluted from
acrylam ide gel is not exactly analogous to a preparation of fetuin
obtained by chem ical m ethods, and may represent components
differing from its chem ically purified counterpart. This was,
however, the m ost nearly homogeneous preparation of tumor
extract antigen available in quantities sufficient for amino acid
analysis.
3) Low lev e ls of tyrosine and proline in the
isolated tumor extract antigen may reflect the com position of a
subpopulation, different from the overall com position of unfrac
tionated fetuin.
4) Although acrylam ide gel electrop h oresis was
the m ost suitable method for obtaining purified antigenic m aterial
from the tumor extract, the m aterial so obtained was not r e -
electrophoresed to confirm its homogeneity, and thus could be
contaminated with accompanying m aterials. These would, how
ever, n ecessa rily lack tyrosine, and have a low proline level.
Serum albumin, a likely contaminant, does not m eet this criterion.
For these reason s, amino acid analysis of the sam ples in
this study is not a definitive m easure of chem ical difference
between the antigen detected in fetuin and that in the W ilm 's tumor
extract, though it is suggestive that the two antigens may differ
chem ically. A nalysis of electrofocused m aterials would be much
m ore useful in this com parison, since this technique offers the
m ost discrim inating m eans of separating antigen-associated
sp ecies. This would be p ossib le by obtaining pooled eluates of
sectioned g els, or by using a preparative electrofocusing pro
cedures to obtain suitable amounts of m aterial.
In sum m ary, many physico-chem ical sim ila rities w ere
dem onstrated between the antigens from com m ercial fetuin, from
newborn calf serum and from the PBS-EDTA extract of four pooled
W ilm 's tum ors. The differences found suggest that the antigens
m ight be of different chem ical com positions, but did not constitute
unequivocal grounds for distinguishing these antigens. These
differences could be of significance if studied under m ore appro
priate conditions. An inherent problem in making this com parison
between antigens, was that fetuin is only g ro ssly defined, is not
reproducibly isolated, and shows extrem e charge heterogeneity.
D issection of cla ssica lly defined fetuin is n ecessary before it can
be compared to the antigen in question. The fact that m aterials
from the sources mentioned share antigenic determ inants cannot
be questioned, and this im m unological property com plem ents the
physico-ch em ical suggestion that they are sim ilar. Under these
circum stances the question as to whether the antigen in W ilm 's
tumor extract is fetuin, is perhaps better postponed or replaced
by an inquiry into the detailed structure, and the function of fetuin.
D istribution of Antigen
A lm ost no conclusions can be made regarding the d istr i
bution of the antigen detected in these studies. This stem s from
two lim itations: first, the number of tumor, tissu e and serum
sam ples studied does not provide sufficient b asis for meaningful
distribution data; and second, the sen sitivity of the assay employed
may sim ply be inadequate to detect low amounts of antigen. For
th ese reason s, the specificity of this antigen is uncertain. It has
been shown however that antigens of the ABO(H) system , the
124 ;
m ajor Rh antigens, and plasm a antigens of a large pool fail to
cro ss react with the antigen detected. In addition, whole blood
(type A, Rh+) failed to cr o ss react with the antigen, indicating that
the antigen was probably not sim ilar to antigens common to human
blood ce lls. The lack of cro ss reaction with Forssm an antigen
determ inants was also dem onstrated, indicating that this antigen,
found in som e viral transform ed c e ll cultures (49, 88), was not
the antigen detected in this study. Human urinary erythropoeitin,
which can be synthesized by W ilm 's tumor ce lls (72), also failed
to dem onstrate cro ss reactivity with the antigen detected. The
human histocom patibility antigens of the HL-A system w ere not
thoroughly tested for cro ss reaction with the antigen detected.
Spleen, thymus and kidney from human fetu ses failed to cro ss
react with the antigen, though the number of sam ples tested would
not cover the entire spectrum of HL-A antigens. Two other
arguments favor the idea that the antigen detected is not of the
HL-A group. F irst, HL-A antigens have not been found in calf
serum , in contrast to the antigen found in this study; and second,
the antigens of the HL-A system are not as ea sily solubilized as
the antigen studied here. The solubilization procedure employed
for preparing HL-A antigens n ecessita tes detergents, sonication
or hypertonic salt solutions. These w ere not n ecessa ry to prepare ;
the soluble antigen studied here. j
Two features of the antigen detected in this study, common
to W ilm 's tumor extract and to calf serum fetuin, warrant further
discussion. F irst, the antigen is sim ilar to fetuin, a substance
(or group of substances) prominent during fetal development and j
possibly involved with growth promotion during this tim e. Second,
the antigen detected here c r o sse s sp ecies lin es, in that it was
observed in human tumor extract and in bovine serum .
It has been known for som e tim e that the globulin fraction
of serum w ill undergo qualitative and quantitative changes during
and around the tim e of transition from fetal development to p ost
natal growth. These changes have been generally detected by
im m unological and physico-chem ical criteria, and are common to
a number of m amm alian sp ecies, including humans (75, 89). Of
particular in terest are the findings that certain malignant and ;
d isease states in adults bring about a reversion of the serum
globulin patterns to a fetal-lik e distribution. Perhaps the two m ost
notable exam ples in humans are the a -fetoprotein, associated with
em bryonic serum and serum from adult patients with hepatoma;
and the carcinoem bryonic antigen found by Gold (90), associated
with em bryonic gut and adult neoplasia of the gastrointestinal tract.
While these have been m ost thoroughly studied, other antigens,
common to fetal development, have been reported in adult tum or-
bearing subjects. These have been recently review ed by Alexander j
(33), who suggests the name "oncofetal" antigens for this group of
substances. Included in this group is a sulfoglycoprotein present
in fetal gut and gastric juices which appears in gastric juices of
patients with gastric neoplasia, and in sen escen ce (51, 52); and an
antigen having y-electrophoretic m obility, common to many human i
tum ors and cro ss reacting with fetal serum from many m am m alian :
126
sp ecies, including human and bovine (54). A fetal glycoprotein
has been described by Alpert et aL (91), which resem b les fetuin,
and which appears in the plasm a of patients with prim ary liver
hepatoma. It is not found in secondary liver neoplasia, nor in
other hepatic d isea ses.
A s m ore sen sitive a ssa y s for these antigens have
developed, it has becom e apparent they are not qualitatively
indicative of a neoplastic state. Low lev els of carcinoem bryonic
antigen, for exam ple, have been found in norm al adult serum;
and marked elevation of these le v e ls noted in non-neoplastic
d isease (56, 57). While the "tumor specificity" of this group of
oncofetal antigens is poorly established, they have been useful in
diagnosing neoplastic disorders in many ca ses.
Very little is known about the function of these substances
associated with fetal development, or their role in neoplastic
d isea se. It has been hypothesized that these oncofetal antigens
represent a reversion of the c e lls producing these antigens to a
m ore "primitive" state, perhaps resem bling the undifferentiated
em bryonic c e ll (33). Whether th ese substances represent incidental
side products of neoplastic transform ation, or are m ore intim ately
related to the growth pattern of neoplastic c ells is not known. In
m ost ca ses, it is not known which c e lls produce the oncofetal
antigen. Apparently, the carcinoem bryonic antigen of the gut can
be synthesized in vitro by cultured tumor c e lls (43), w hile the
fetoprotein of A belev et al. (46) is synthesized by the liv er.
127 |
i
The p ossib ility that these fetal glycoproteins antigens
i
are instrum ental in the growth properties of em bryonic (and
possibly tumor) c e lls is im plied by studies with fetuin., demon
strating the n ecessity of this m aterial in the growth of c e lls in
the absence of serum . As mentioned, the sia lic acid residues of
this glycoprotein m ay play an important role in its biological role,
though this point is not settled. Other glycoproteins, notably
human chorionic gonadotropin and follicle-stim u latin g hormone,
are deactivated by rem oval of their sia lic acid residues (92, 93).
Control of fetal development could conceivably be under strong
control by a population of glycoproteins able to effect such changes,
either by direct interaction with developing c e lls, or by tran s
porting sm aller m olecu les such as horm ones, or m etabolites to
these c ells. The carbohydrate m oieties of these m olecules may
be particularly instrum ental in determ ining their functional role.
These p o ssib ilities are conceptual, and little concrete evidence is
available regarding the actual function of these fetal glycoproteins.
The a-globulin cla ss of fetal serum proteins is w ell
known for its ability to cro ss react with a -globulins of other
sp ecies. Notably, the a-fetoprotein associated with human
prim ary hepatoma and fetal serum has been shown to cro ss react
with analogous “ -globulins from the fetal serum of six m am m alian 1
sp ecies (89). This cro ss reaction was extended by A belev (46) to ;
include the a-fetoprotein of m ouse, found both in embryonic
serum , and in prim ary m ouse hepatoma serum . Other serum
proteins, such as albumins and Y -globulins are known to cro ss !
react between sp ecies (46, 94), suggesting that these proteins
share a common core, which can be altered to create additional
antigenic determ inants, perhaps sp ecies sp ecific. Sim ilar con
siderations have been given to the a -globulin fraction of serum
(89). An interesting exam ple of this phenomenon may be the a-
fetoprotein of the sheep. This cro ss reacts with human “ -feto
protein (89), and also cro ss reacts with calf fetuin (95). The
m olecular weight of human “ -fetoprotein, approximately 70, 000
daltons (89), differs from that of fetuin (45, 000 daltons), indicating
that they are associated with m olecules differing in size. Antigenic
determ inants of each of these “ -fetoproteins are apparently
present on the sheep “ -fetoprotein, however, and suggest a
p ossib le relationship between fetuin and human “ -fetoprotein.
D irectly related to the present study is the finding by
Bodman (96) that antiserum prepared against calf serum fetuin
was able to precipitate an ^ -glob u lin from human fetal serum
(and serum from the m other). This globulin was apparently not
“ -fetoprotein, which dem onstrates a m obility between “j and
albumin. This was indirectly confirm ed by Muralt and Roulet (97)
who showed that antiserum against human a -fetoprotein did not
precipitate fetuin. The argument was raised by these w orkers
that the human “2 -globulin precipitated by antiserum to fetuin,
was probably due to the presence of a contaminating a2"gl°kulin
in the immunizing fetuin preparation. Whether or not this is the
case, it clearly indicates that an human “ -globulin was p re
cipitated by antiserum (rabbit) prepared against bovine serum
129 j
j
components, which dem onstrates the existence of common antigenic j
determ inants on a-globulins of both sp ecies. It is particularly
interesting that in the present study, the antiserum to the W ilm 's
tumor extract reacted with an antigen common to the extract and
to calf serum , both of which dem onstrated m obility. The
common antigen in com m ercially prepared fetuin dem onstrated,
also, a m obility between ag and c t ^ unlike the fast a ^ m obility
reported by others (73).
It is entirely possible that among the antigenic d eter
minants present on a given ot-globulin, som e of these are shared
by globulins of other sp ecies, and of a different electrophoretic
globulin cla ss. Since these c la sse s are largely defined by the
ionogenic groups of a population of m olecu les, there is no a priori
reason to expect an obligatory difference in antigenic determinants
among these groups. It has been shown in the present study that
the antigenic determ inants recognized on fetuin are common to
su b classes of this substance differing in ionic makeup. It also
seem s apparent that the antigenic determ inants detected in the
W ilm 's tumor extract and fetuin resid e on m olecules of sim ilar
size , though the detailed chem ical com position may be sign ifi
cantly different. While this is of in terest in comparing two antigens,
it is again not to be expected a priori since, as in the case of sheep ;
a -fetoprotein, antigens common to m ultiple sp ecies may be shared
by m olecules of different size . Until m ore information about the
detailed structure of these glycoproteins is obtained, and until the |
details of their synthesis and m etabolism are elucidated, it w ill
be difficult to understand the distribution of antigenic determ inants
in serum glycoproteins, and the significance of this distribution.
Whether or not the antigen detected in the W ilm 's tumor
extract and in newborn and fetal calf serum is an "onco-fetal"
antigen cannot be firm ly established at this tim e. It was not
found in any of three human fetal kidneys, spleens or thym uses,
nor in one serum sam ple of human fetus. The negative findings
m ay be due to lack of sen sitivity in the assay employed or to the
rea l absence of antigen in these preparations. It could also be due
to the site of synthesis of the antigen. If, for instance, the antigen
w ere produced in the liver, it may not circulate to the other organs
effectively, or in quantities sufficient for detection. Its serum
lev el could depend also on the stage of fetal development. In the
case of the carcinoem bryonic antigen of the digestive tract, the
production dim inishes after the fir st two trim esters of gestation.
Other fetal proteins appear at variable tim es throughout the
development of the fetus (75). B ecause only one serum sam ple
and three fetal organ sam ples (all from the late first or second
trim ester) w ere used in this study, no indication of the tim e-
distribution of antigen was obtained.
The inability to detect this antigen in the serum of W ilm 's
tumor patients, or in three other W ilm 's tum ors would suggest
lim itations to the use of this antigen for diagnostic purposes.
Other factors m ay be involved, however. Many of the patients
w ere receiving chemotherapy, radiation and surgical treatm ent
at the tim e of obtaining the serum sam ple. A lso these serum
sam ples w ere often received in frozen state, or after incubating
for unknown periods of tim e. These factors may have influenced
the amounts of detectable antigen present. In seven ca ses, how
ever, serum sam ples w ere shipped quick-frozen from patients
with no treatm ent for W ilm 's tumor. The lack of detectable
antigen in these sera could resu lt from at lea st three conditions:
1) the detection system is not sen sitive enough; 2) the antigen
is present in a cryptic form; or 3) the antigen is not in the
serum due to som e localizing factor. This could include, for
instance, restriction of the antigen to the tumor tissu e due to
lack of vascular drainage. This last possib ility does not explain
the lack of antigen in other W ilm 's tum ors, individually extracted.
It is possible, how;ever, that the production of antigen is restricted
to a particular stage of tumor development, analogous to the
production of surface glycoprotein at specific stages of m onolayer
confluency in cultured c ells (42). A close analogy to this ex ists
with the a -fetoprotein associated with hepatoma. This hepatoma,
grown in vitro can lo se its ability to synthesize the a -fetoprotein
when transplanted back into anim als (46). It does, however,
m aintain its malignant properties, and its ability to produce
albumin, transferrin and an organospecific liver antigen. This
was found in one of three cultures studied. That sim ilar events
m ay occur in vivo is a possibility. D ifferences from tumor to
tumor may be compounded in the case of W ilm 's tumor by the
fact that its morphology is of a m ixed cell type (59). With a non-
uniform cell population representing a given tumor, it is possible
that antigen production w ill vary with the c ell type predom i
nating.
As presently developed, the study of this antigen is m ost
lacking in a m ore sen sitive assay. In the case of carcinoem bryonic ’
antigen, significantly high lev els of serum antigen are 5 - 25 ng
per m l (57). The sen sitivity of the assay employed here is
about 8 yg per m l, in sen sitive to this sam e range of antigen even
after a ten-fold concentration of serum . Even the lev els in tumor
extracts may go undetected by the present method. With the m eans
of obtaining relatively uniform antigen, it m ay be p ossib le to
develop a m ore sen sitive radioim m unoassay for detection of this
antigen.
P resen ce of the Antigen In Vitro
Perhaps the m ost prom ising facet of this study is the
dem onstration of this antigen in cultures of W ilm 's tumor c ells.
This offers a m eans by which the interactions of the antigen with
c e lls can be m onitored. The possible growth effects of this fetuin- ‘
like substance can be m easured; the ability of this antigen to
m ediate cellular or com plement-dependent humoral immune r e - i
sponses to these c e lls can be determined; and the biosynthesis of j
this antigen can possibly be studied. This inform ation would lend
functional significance to the antigen which has been partially
described physically and chem ically. ■
A fundamental question is whether the antigen is synthe
sized by these cultured c e lls, or derived from the newborn calf
serum in which the c e lls w ere grown. P recedents exist for
either case. Apparently, carcinoem bryonic antigen of the colon
is synthesized by cultured c e lls from a colonic tumor (43); this
antigen is secreted into the culture medium. Tumor specific
transplantation antigens (TSTA) are observed in cultures of many
neoplastic c e lls, including those of chem ical-induced and viral-
induced tum ors (98); these antigens are apparently synthesized
by the cultured c e lls. On the other hand, evidence suggests that
a number of serum proteins can be adsorbed to ce lls in vitro as
w ell as in vivo (44, 48), and that these adsorbed substances can
act as surface antigens. A sim ilar case is the adsorbed antigens
from sem inal fluid coating sperm c e lls (99). The sialoglyco-
proteins of serum have been particularly im plicated as agents in
this coating p rocess. Apffel and P eters (48) have suggested the
name "sybodies" for this c la ss of m aterials which are largely
liver-d erived , but occur on tumor c e lls in vivo. Evidence for
this phenomenon m vitro is also review ed by these w orkers, and
suggests that such agents m ay be involved in blocking or m asking
the recognition of cellular surface antigens. Other mucinous
m aterials have been reported with a sim ilar capacity for blocking
antigen recognition (84).
The question regarding the source of the antigen detected
in this study cannot be answered by the m ethods em ployed. The
antigen from cultured c e lls was identical by immunodiffusion with
both the antigen from calf serum and from the extract of W ilm 's
tumor tissu e. Since these antigens w ere indistinguishable by
134 !
|
im m unoelectrophoresis, and iso elec tr ic focusing, these criteria |
would not have elucidated the source of antigen from the cultured
W ilm 's tumor c e lls. Insufficient m aterial was available to
determ ine the chem ical com position of the antigen from cultured
c e lls. This is the only criterion strongly suggesting a difference
between the W ilm 's tumor extract antigen and that of calf serum
fetuin.
If the antigen w ere adsorbed from the newborn calf
serum in which the c e lls w ere grown, a number of considerations
would apply:
1) Only the antigen from calf serum cro ss reacting
with the tumor extract would be absorbed; the second antigen
observed in calf serum was not detected in the extract of cultured
tumor cells.
2) Other calf serum proteins w ere not absorbed
after the rinsing procedure. This is suggested by the fact that
antiserum prepared against m echanically rem oved W ilm's tumor
c e lls, and absorbed with norm al human plasm a and human kidney
extract, showed no reaction against newborn calf serum in immuno-j
j
diffusion. This would not be expected if a general adsorption of
calf serum proteins had occurred, since these would be present
j
in the imm unizing m aterial. This point is partially weakened
by the fact that the sam e antibody preparation was unable to
dem onstrate the antigen found in the extract of cultured W ilm 's
tumor c e lls, using R5 antibody prepared against W ilm 's tumor i
tissu e extract. The failure of the antibody against the m echanically :
135
rem oved cells to dem onstrate the antigen m ay be due to the much
reduced amount of antigen used per injection; or the form of the
antigen injected (attached to c e lls instead of in a PBS-EDTA
extract). It was apparently not due to the general lack of immune
response by this rabbit, since unabosrbed antiserum reacted
strongly with the cultured W ilm 's tumor c e ll extract, and with
extracts of tumor and norm al kidney tissu e.
3) The presence of antigen in the PBS-EDTA
extract of cultured W ilm 's tumor c e lls was probably not due to the
rem oval by this agent of antigen adsorbed to the g la ss of the
cultured flask. Two points suggest this: F irst, antigen was
extractable from BSS-washed W ilm 's tumor c ells m echanically
rem oved from the flask as w ell as from m onolayered cells.
Second, the antigen was not found in extracts of norm al human
kidney cells grown in identical flask and extracted while in mono
layer. Antigen adsorbed to g la ss would have been present in this
extract as w ell.
4) Qualitatively there was an apparent, preferential
binding of the antigen to W ilm 's tumor c e lls, compared to polyoma
transform ed baby ham ster kidney ce lls, sim ilarly cultured. The
relative amount of antigen binding to norm al human kidney c e lls
could not be determ ined since the number of c ells extracted was
not comparable.
That the cultured W ilm 's tumor ce lls specifically
synthesize the antigen is an alternative to the above possibility,
136 ;
l
though no evidence suggests this explanation over that of adsorp- j
tion. Both of these p o ssib ilities warrant further investigation,
since either cellular synthesis or specific adsorption of a fetuin-
like antigen is germ ane to the problem of determ ining this
antigen's function in vivo. !
A number of approaches to these problem s are available
using the tissu e culture system studied here:
1) W ilm 's tumor c e lls could be adapted to growth
in human serum , elim inating the problem of absorption of antigen.
These c e lls could be extracted; if antigen is found, it im plies
cellular synthesis.
2) Fetuin, iodinated with radioactive iodine, could
be added to culture medium of W ilm 's tumor c e lls. If adsorption
occurs, this label should be found in the extracts of these cells.
This technique could be used also to study the preferential binding
of fetuin to various c e ll lin es, by sim ply counting the amount of
label rem aining with a given number of c e lls after washing. It
m ight be useful in determ ining the surface properties of these
c e lls in a way analogous to the use of specific plant agglutinins ;
for elucidating the surface structure of various tumor c e lls (30). j
3) C ells labeled with amino acid and amino sugar
precursors and chased with cold m etabolites, could be extracted.
j
The presence of label in extracted antigen would suggest cellular j
synthesis.
These are som e approaches to the elucidation of the
source of antigen on the W ilm 's tumor c ells. Additional
137 !
inform ation could be obtained on the ability of this antigen to effect I
an im m unological response to these c e lls m vitro. This could
shed light on the possible in vivo action of this antigen in provoking
(or blocking) an effective immune response to the tumor. D eter
mining the location of this antigen in the c ell would also be p er
tinent to its role as an antigen. There is evidence based on
im m unofluorescence (100, 101) that fetuin can enter living cells
infected with influenza virus. While the antigen detected in the
present study was extracted without cellular damage (judged by
trypan blue exclusion), it is possible that it was internal and
extracted by this procedure. The presence of virus in the tumor
extracts and cell cultures was not determ ined. The location of
an antigen within a cell m ay w ell effect its ability to elicit an
immune response m vivo (33), since this property is generally
restricted to surface antigens.
The tissu e culture system explored in this study, then,
offers a com plem entary m eans of determ ining the function of the
antigen detected and described by im m unological and physico
chem ical techniques; and allows the structure of this antigen to
be studied in term s of its biological properties. At present, the
function of the antigen detected in this study is com pletely unknown,
its distribution in human subjects has not been determ ined and its
relationship to W ilm 's tumor is established only by its presence
in a PBS-EDTA extract of four pooled W ilm 's tum ors, and of
cultured W ilm 's tumor c ells. Its presence in fetal and newborn
calf serum , and in the fetuin fraction of this serum has been
dem onstrated, but is not understood functionally. The use of the
|
im m unological assay developed here has apparent lim itations in
its use as a diagnostic tool for the detection of antigen in patients
with W ilm 's tumor.
The partial characterization of this antigen; its detection |
in W ilm 's tumor and in W ilm 's tumor in tissu e culture; the
dem onstration of its heterogeneity and its structural relationship :
to fetuin; and the relationship between its antigenic structure and
its physico-chem ical properties has been partially determined in
this study. This offers a first step in understanding the function of
the antigen, and in developing a sen sitive assay for this antigen of
possible diagnostic use. During the course of this study, insight
was gained into the nature of serum glycoproteins, their antigenic
makeup and chem ical heterogeneity. These features are important
in pursuing the isolation and biological ro les of these compounds.
CHAPTER V
SUMMARY
Using a rabbit immunoglobulin fraction directed against
an ethylenediam inetetraacetate (EDTA) extract of human nephro
blastom a tissu e and absorbed with various norm al human m aterials,
an antigen was detected by immunodiffusion techniques in nephro
blastom a extract. This antigen w as also present in a sim ilar
extract of cultured human nephroblastoma c e lls, in newborn and
fetal calf serum and in fetuin. The antigens from th ese sources
w ere sim ilar in solubility, heat sen sitivity, m olecular siz e , and
iso electric properties. Some differences w ere observed in
electrophorectic m obility and chem ical com position of antigen
isolated from nephroblastoma extract and fetuin.
This antigen, detectable at a concentration (fetuin) of
approximately ten m icro g ra m s/m illiliter, w as not found in
extracts of normal adult kidney nor in adult plasma; nor was it
found in kidney, spleen, thymus, or serum of second trim ester
human fetu ses. Serum from nephroblastoma patients did not
dem onstrate the antigen, nor did extracts of other nephroblastoma
tissu e. These sam ples w ere obtained at various stages of tumor
growth and treatm ent. The antigen did not cro ss react with
139
140
blood group antigens of the ABH, Rh or F orssm an type.
Antigen from nephroblastom a c e lls cultured in m onolayer
was extractable using EDTA in salin e, with c e ll viability over 95%.
This suggests that the antigen may resid e on the c e ll surface. The
antigen w as also extractable by hom ogenization of m echanically
rem oved c e lls in EDTA saline solution. Sim ilar preparations of
polyoma transform ed baby ham ster kidney c e lls did not reveal
the antigen, nor did EDTA extracts of m onolayer cultured c e lls
derived from fetal human kidney.
Antigen in nephroblastom a extract, calf serum , and
fetuin was nondialysable, water soluble over a pH range of 2 - 9,
and iso electric over the pH range 3 .4 - 4. 2. The protein staining
pattern of th ese g els suggested a population of sp ecies having
charge m icroheterogeneity.
Antigen from calf serum and nephroblastoma extract
showed identical electrophoretic m obility (alpha2> in agar
im m unoelectrophoresis. Antigen in fetuin dem onstrated a slightly
faster m obility by this technique. Sim ilarly, fetuin-associated
antigen was of faster m obility in acrylam ide gel electrophoresis.
Chromatography on G-75 Sephadex dem onstrated antigen from
both nephroblastoma extract and fetuin as having a m olecular
weight of 40, 000 to 50, 000 daltons.
141
The antigen from nephroblastoma extract was sen sitive
to pronase and trypsin, in sen sitive to ribonuclease and neuram ini
dase. Hyaluronidase produced a slight alteration, though the
p ossib ility of protease contamination of this enzyme preparation
was not checked. These studies suggested that the protein portion
of the antigen was n ecessary for its integrity.
Amino acid analysis of antigen from nephroblastoma,
isolated by acrylam ide gel electrop h oresis, revealed a low proline
and tyrosine content, compared to the amino acid com position of
fetuin. A lso, the uronic acid and neuram inic acid content of
isolated tumor extract antigen was greater than the corresponding
content in com m ercial fetuin.
The tissu e and sp ecies distribution of the antigen was
not fully established, nor was its absolute relationship to nephro
blastom a. However, the approach used and resu lts obtained w ill
aid in developing m ore sen sitive im m unological a ssa y s for d etec
tion of nephroblastom a-associated antigens. Dem onstration of
heterogeneity of the antigenic m aterial by iso elec tric focusing
illu stra tes the difficulties involved in com plete characterization
of such m aterials. Finally, the identification of the antigen in
tissu e culture provides a convenient system whereby the synthesis
and function of this antigen may be studied.
A P P E N D I C E S
142
A PPE N D IX 1. B U FFE R S AND MEDIA
143
144
BUFFERS AND MEDIA
A. PBS
NaCl 8. 0 gm
KC1 0. 3 gm
Na2H P04 0. 073 gm
KH2P 0 4 0. 020 gm
W ater to 1000 m l
B. PBS-EDTA
NaCl 8. 0 gm
KC1 0. 2 gm
Na2H P04 1. 15 gm
KH^PO^ 0. 2 gm
EDTA (tetrasodium ) 0. 2 gm
W ater to 1000 m l
C. PBS-KC1
D issolve KC1 to 3 M in PBS
145
D. BSS
NaCl 8. 00 gm
KC1 0. 40 gm
Na2H P04 ' 2H20 0. 060 gm
k h 2p o 4 0. 060 gm
N gS04 * 7H20 0. 10 gm
CaCl2 0. 140 gm
MgCl2 • 6H20 0. 10 gm
NaHCOg 0. 350 gm
Glucose 1. 00 gm
Phenol red . 010 gm
Water to 1000 m l
146
E. MEM (Eagle)
A ll components of BSS
L Amino Acids:
Arg 0. 105 gm
Cys 0. 024 gm
' v -
Glu 0. 292 gm
His 0. 031 gm
lie 0. 053 gm
Leu 0. 053 gm
Lys 0. 058 gm
Met 0. 015 gm
Phe 0. 032 gm
Thre 0. 048 gm
Try 0. 010 gm
Tyr 0. 036 gm
Val 0. 046 gm
Choline Cl 0. 001 gm
F olic acid 0. 001 gm
i-In ositol 0. 002 gm
Nicotinam ide 0. 001 gm
D-C a
pantothenate 0. 001 gm
Pyridoxal HC1 0. 001 gm
Riboflavin 0. 0001 gm
Thiamine HC1 0. 001 gm
Water to 1000 m l
147
F. BBS
B oric acid 6. 184 gm
Na2B^Oi7 * IOH2O 9. 536 gm
NaCl 4. 384 gm
W ater to 1000 m l
G. IEP Buffer
pH = 8. 2 C/2 = 0. 05
1540 m l 0. 1 N sodium barbital
+ 460 m l 0. 1 N HC1
2000 m l Solution A
1300 m l Solution A
+700 m l water
2000 m l working buffer
APPENDIX 2. DEAE-SEPHADEX CHROMATOGRAPHY
OF POOLED WILM'S TUMOR EXTRACT
Absorbance 280 nm (---------- );
i
Absorbance 240 nm ( • • • • )
148
i
ABSORBANCE
**«*
«* •••• • •• *
a .
»*** OB • • I H f ) a y i i i i
• • • • • • • ! • • • • • • • •
• • • •
N
■ • . * • • • •
FORMATE (M)
APPENDIX 3. SEDIMENTATION PATTERN OF
COMMERCIAL FETUIN
Schlieren pattern of 1. 0% w /v
fetuin in 0. 12 M NaCl pH = 6 .5
6 = 60°
tim e of run = 84 m inutes
rpm = 59, 780
T = 20. 0° C
director of sedimentation: right to left
150
151
B IB L IO G R A P H Y
152
153
BIBLIOGRAPHY
1. Leighton, J . , The Spread of Cancer, Academ ic P r e ss, New
York, 1967, p. 1.
2. Burnett, F. M ., P ro g ress Exp. Tumor R e s ., 13:1, 1970.
3. Good, R. A ., "D isorders of the Immune System , " In
Immunobiology (R. A. Good and D. F ish er, E d s .),
Sinauer, Stanford, 1971, p. 1.
4. TJllman, S. B ., Oncologia, 16:1, 1963.
5. Salt, G ., Federation Proceedings, 26:1671, 1967.
6. M archalonis, J. and Edelman, G. M ., J. Exp. M ed.,
122:601, 1965.
7. Beckwith, J. B. and Perrin, E. V ., Am er. J. P a th .,
43:1089, 1965.
8. A shley, D. J. B ., J. Path. B a c t., 90:217, 1965.
9. Levin, A. G .; McDonough, E. F . , J r .; M iller, D. G.
and Southam, C. M ., Ann. N. Y. Acad. S c i.,
120:400, 1964.
10. Pollock, W. F . ; H astings, N. and Snyder, W. H .,
Surgery, 48:606, 1960.
11. Skinner, M. D. and Schwartz, R. S ., New Eng. J. M ed.,
287:281, 1972.
12. Woodruff, M. F. A ., Annotation: Lancet i :1299, 1968.
13. Comments by R. A. Good In Immune Surveillance (R. T.
Smith and M. Landy, E d s .), A cadem ic P r e ss, New
York, 1970, p. 439.
154
14. Law, L. W .; Ting, R. C. and Leckband, E ., In Ciba
Foundation Symposium, J. and A. Churchill, L td .,
London, 1966, p. 214.
15. Hertz, R .; R oss, G. T. and L ipsett, M. B ., Ann. N. Y.
Acad. S c i., H4:881, 1964.
16. KLein, G ., Ann. Rev. M icrob iol., 20:223, 1966.
17. Comments by R. T. Prehn In Immune Surveillance (R. T.
Smith and M. Landy, Eds. ), Academ ic P r e ss, New
York, 1970, p. 451.
18. Klein, E ., N. Y. State J. of M ed., 68:900, 1968.
19. P arr, I ., Br. J. Cancer, 26:174, 1972.
20. Zbar, B .; B ernstein, I. D .; B artlett, G. L .; Hanna,
M. G. and Rapp, H. J . , J. Natl. Cancer In st.,
49:119, 1972.
21. Lindenmann, J. and Klein, P . , R ec. R esults in Cancer
R es. , 9:1, 1967.
22. Stjernward, J . , J. Natl. Cancer Inst. , 37:505, 1966.
23. H ellstrom , K. E. and H ellstrom , I . , Adv. Cancer R es. ,
12:167, 1969.
24. For exam ple, in Martin, D. S ., Hospital P ractice, 8:129, 1973.
25. Comments by P. Alexander In Immune Surveillance (R. T.
Smith and M. Landy, Eds. ), A cadem ic P r e ss, New
York, 1970, p. 488.
26. Old, L. J. and B oyse, E ., Ann. Rev. M ed., 15:167, 1964.
27. F oley, E. J . , Cancer R es. , 13:835, 1953.
28. Prehn, R. T. and Main, J. M ., J. Natl. Cancer Inst. ,
18:769, 1957.
29. Z ilber, L. A. and A belev, G. I . , The V irology and
Immunology of Cancer, Pergam on, Oxford, 1968, p. 351.
155
30. Burger, M M. and Noonan, K. D ., Nature, 228:512, 1970.
31. Plagem ann, P G. and E stensen, R. D ., J. C ell B io l.,
55:179, 1972.
32. Johnson, G. S .; Friedm an, R. M. and Pastan, I . ,
Proc. Natl. Acad. Sci. (USA), 68:425, 1971.
33. Alexander, P ., Nature, 235:137, 1972.
34. Old, L. J . ; B oyse, E. A .; G eering, G. and Oettgen,
H. F . , Cancer R e s ., 28:1288, 1968.
35. W eiss, D. In Immunological P aram eters of Host-Tum or
R elationship (D. W eiss, E d .), Academ ic P r e ss,
New York, 1971, p. 1.
36. Klein, G ., In Im m unological P aram eters of H ost-Tum or
Relationship (D. W eiss, Ed. ), A cadem ic P r e ss,
New York, 1971, p. 113.
37. B oyse, E. A .; Old, L. J . ; Stockert, E. and Shigeno, N .,
Cancer R e s ., 28:1280, 1968.
38. M oscona, A. A ., Developm ental B iol. , 18:250, 1968.
39. Cline, M. J. and Livingston, D. C ., Nature New Biology,
232:155, 1971.
40. R enier, J. and Southam, C. M ., Cancer R e s ., 29:1814, 1969.
41. R obertson, H. T. and Black, P. H ., Proc. Soc. Exp.
B iol. M ed., 130:363, 1969.
42. Hakomori, S .; Kijim oto, S. and Siddiqui, B ., Federation
P roceedings, 30:1043, 1971.
43. Goldenberg, D. M .; Pavia, R. A .; Hansen, H. J. and
Vandenvoorde, J. P . , Nature New B io l., 239:189, 1972.
44. Scblesinger, M ., Prog. Exp. Tumor R es. , 13:28, 1970.
45. C ikes, M ., J. Natl. Cancer In st., 45:979, 1970.
46. A belev, G . I . , Cancer R e s ., 23:1344, 1968.
156
47. B oyse, E. A .; Old, L. J. and Stockert, E ., In IV Intl.
Symposium Immunopathology, Schwabe, B asel, 1965,
p. 23.
48. Apffel, C. A. and P eters, J. H ., J. Theor. B io l., 26:
47, 1970.
49. Kijim oto, S. and Hakomori, S ., F ebs L etters, 25:38, 1972.
50. Buffe, D .; Rimbaut, C .; L am erle, J . ; Schweiguth, O.
and Burton, P ., Intl. J. Cancer, 5:85, 1970.
51. Hakkinen, I ., Ann. Surgery, 169:277, 1969.
52. deB oer, W. R. G .; Forsyth, A. and Nairn, R. C .,
B rit. Med. J . , 3:93, 1969.
53. Stolbach, L. L .; Krant, M. J. and Fishm an, W. H .,
New Eng. J. M ed., 281:757, 1969.
54. Edynak, E. M .; Old, L. J . ; Vrana, M. and L ardis, M .,
Proc. A m er. A ssoc. Cancer R e s ., 11:22, 1970.
55. Thompson, D. M. P . ; Krupey, J . ; Freedm an, S. O. and
Gold, P ., Proc. Natl. Acad. Sci. (USA), 64:161, 1969.
56. M oore, T. L . ; Kupchik, H. Z . ; Marcon, N. and Zamcheck,
N ., Am er. J. D igestive D ise a se s, 16:1, 1971.
57. Gerfo, P. L .; Krupey, J. and Hansne, H. J . , New Eng.
J. Med. , 285:138, 1971.
58. Rouslahti, E. and Seppala, M ., Nature, 235:161, 1972.
59. Pochedly, C ., N. Y. State J. M ed., 71:1089, 1971.
60. P ollock, W. F . ; H astings, N .; Snyder, W. H .,
Surgery, 48:606, 1960.
61. Ragab, A. H .; V ietti, T. J . ; C rist, W .; P erez, C.
and M cA llister, W ., Cancer, 30:983, 1973.
62. M orse, B. S. and Nussbaum, M ., A m er. J. M ed.,
42:996, 1967.
157
63. A llerton, S. E . ; B eierle, J. W .; Pow ars, D. R. and
j Bavetta, L. A ., Cancer R es . , 30:679, 1970.
64. Clausen, J ., Immunological Techniques for the Identifi
cation and E stim ation of M acrom olecules, A m erican
A lsev ier, New York, 1969, p. 529.
65. E astoe, J. E. and Courts, A ., P ractical A nalytical
Methods for Connective T issu e P roteins, E. and
P. N. Spon, London, 1963, p. 42.
66. B itter, P. and Muir, H. M ., A nalytical B io ch em ., 4:
330, 1962.
67. D avis, B. J . , Ann. N. Y. Acad. S c i., 121:404, 1964.
68. W rigley, C. W. , "Gel E lectrofocusing, " In Methods in
Enzym ology, Vol. 22 (W. B. Jakoby, Ed. ), A cadem ic
P r e ss, New York, 1971, p. 559.
69. Merchant, D. J . ; Kahn, R. H. and Murphy, W. H .,
Handbook of C ell and Organ Culture, B u rgess,
M inneapolis, 1964, p. 157.
70. Adopted from: M eltzer, M. S .; Leonard, E. J . ; Rapp,
H. J. and B o rso s, T ., J. Natl. Cancer Inst. , 47:
703, 1971.
71. Cameron, E ., Hyaluronidase and C ancer, Pergam on,
Oxford, 1966.
72. Erythropoietin has been detected in cultures of W ilm 's
tumor c e lls. (Dr. Eugene G oldw asser, personal
communication)
7£. Spiro, R. G ., J. Biol. Chem. , 235:2860, 1960.
74. Arquembourg, P. C .; Salvaggio, J. E. and B ick ers,
J. N ., P rim er of Im m unoelectrophoresis, Ann
Arbor-H um phreys, Ann Arbor, Michigan, 1970, p. 30.
75. Shultze, H. E. and Heremans, J. P ., M olecular B iology
of Human P roteins, Vol. 1, E lsev ier, New York,
1966, p. 518.
158
76. Sephadex Gel Filtration in Theory and P ractice, Pharm acia
Fine C hem icals Appelbergs B octryckeri AB 1970.
77. Turner, K. J ., Biochim . Biophys. Acta, 69:518, 1963.
78. Spiro, M. J. and Spiro, R. G ., J. Biol. C hem ., 237:
1507, 1962.
79. F ish er, H. W .; O'Brien, D. and Puck, T. T ., Archiv.
Bioc. B iop h ys., 99:241, 1962.
80. Shultze, H. E. and H erem ans, J. F . , M olecular Biology
of Human Proteins, Vol. 1, E lsev ier, New York, i
1966, p. 182. |
81. D eutsch, H. F . , J. Biol. Chem. , 208:669, 1954. j
i
82. Shultze, H. E. and Herem ans, J. F . , M olecular Biology
of Human Proteins, Vol. 1, E lsev ier, New York,
1966, p. 524.
83. Lieberm ann, I.; Lamy, F. and Ove, P ., Science, 129:
43, 1959.
84. Lippman, M ., Nature, 219:33, 1968.
i
85. Taylor, H. E. and Culling, C. F. A ., Lab. In v est.,
15:196, 1966.
86. C urrie, G. A ., Nature, 215:164, 1967.
87. A llerton, S. E . ; Bavetta, L .; Pow ars, D. and B eierle, J. 9
"An Abnormal M ucopolysaccharide in the Serum of |
W ilm 's Tumor Patients, " In Tenth Intl. Cancer C ongress. I
Houston, T exas, Abstr. 1257, 1970. j
i
88. Burger, M. M ., Nature New Biology, 231:125, 1971.
i
j
89. Gitlin, D. and Boesm an, M ., Comp. Biochem . Phvsiol» * j
21:327, 1967. j
|
90. Gold, P. and Freedman, S. O ., J. Exp. Med. , 122:467, 1965.
91. A lpert, M. E . ; U riel, J. and deNechaud, B. , N e w Eng.
J. M ed., 278:984, 1968.
159
92. Gottschalk, A .; Whitten, W. K. and Graham, E. R. G .,
Biochim . Biophys. Acta, 38:183, 1960.
93. Goverde, B. C .; Veenkamp, F. J. N. and Himan, J. D. A .,
Acta Endocr. (Kobenhavn), 59:105, 1968.
94. Boesm an, M. and Gitlin, D ., J. Clin. Invest. , 45:
1826, 1966.
95. Bergmann, F. H .; Levine, L. and Spiro, R. G .,
Biochim . Biophys. Acta, 58:41, 1962.
96. Bodman, J ., C linica Chim ica Acta, 4:103, 1959.
97. Muralt, P. G. and Roulet, D. L. A ., Helv. paediat.
Acta, 16:517, 1961.
98. Z ilber, L. A. and Abelev, G. I ., The Virology and
Immunology of Cancer, Pergamon, Oxford, 1968,
p. 277.
99. W eil, A. J. and Rodenburg, J. M ., Proc. Soc. Exp.
Biol. N. Y ., 109:567, 1962.
100. Blough, H. A .; Ballow, M. and K reider, J. W .,
J. Im m un., 100:209, 1968.
101. Blough, H. A ., V irology, 31:514, 1967.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Isolation, Separation, And Structural Identification Of Streptomyces Quinones

PDF

The Human 3 Beta-Hydroxysteroid Dehydrogenase Type Ii Gene And Genetic Predisposition To Prostate Cancer

PDF

Protein Kinase Activity In Human Lymphocyte Nuclei: Effect Of Cyclic Nucleotides And Phytohemagglutinin

PDF

Biochemical factors determining tumor response to 5,10-dideazatetrahydrofolate, a folate antimetabolite inhibitory to de novo purine biosynthesis

PDF

Physicochemical Studies Of High-Density Lipoproteins Of Bovine Plasma

PDF

The Effect Of Polyols On Placental Estradiol Dehydrogenase

PDF

Studies of human ɑ₂ macroglobin: physical and enzyme binding properties

PDF

The ferric complexes of citrate, ETDA and NTA: their nature and the ligand exchange reaction with human transferrin

PDF

A POU gene expressed in the early sea urchin embryo: Structure, evolution and requirement for embryonic development

PDF

Steric effects in the acid, amine, and glyceraldehyde-3-phosphate dehydrogenase catalyzed hydrolysis of acyl phosphates

PDF

Changes of DNA methylation patterns during human development and transformation.

PDF

Crosslink Precursors In Fetal Calf Skin Soluble Collagens

PDF

The inability of collagen cross-linking to influence the insolubilization of proteodermatan sulfate and a sialo-glycopeptide

PDF

Kinetic Studies Of Ultracentrifugal Demulsification

PDF

Studies on the pulmonary surface-active lipoprotein: isolation and characterization of its peptide moiety

PDF

Analysis of multistep mammary tumorigenesis in Wnt1 transgenic mice: The role of fibroblast growth factor-8 in oncogenesis and apoptosis

PDF

Studies On The Quantitative Precipitin Reaction

PDF

FGF8, isoform b, is an etiological factor in prostate tumorigenesis

PDF

Kinetic investigation of the hydrolysis of aryl β-D-glucopyranosiduronic acids by β-glucuronidase

PDF

Development of replication -competent retroviral vectors for efficient, targeted gene therapy of cancer

Asset Metadata

Creator Wise, Kim Stuart (author)

Core Title Studies On Antigens Associated With Wilm'S Tumor (Nephroblastoma)

Contributor Digitized by ProQuest (provenance)

Degree Doctor of Philosophy

Degree Program Biochemistry

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag chemistry, biochemistry,health sciences, immunology,OAI-PMH Harvest

Language English

Advisor Allerton, Samuel E. (committee chair), Roy-Burman, Pradip (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-869885

Unique identifier UC11364305

Identifier 7331686.pdf (filename),usctheses-c18-869885 (legacy record id)

Legacy Identifier 7331686

Dmrecord 869885

Document Type Dissertation

Rights Wise, Kim Stuart

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA