University of Southern California Dissertations and Theses

Nude mouse grown human nephroblastomas

(USC Thesis Other)

Nude mouse grown human nephroblastomas

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content NUDE MOUSE GROWN HUMAN NEPHROBLASTOMAS
by
John Michael Deimage
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Cellular and Molecular Biology)
January 1979
UMI Number: DP23649
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
Dissertation PublisWng
UMI DP23649
Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author.
Microform Edition © ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106- 1346
UNIVERSITY OF SOUTHERN CALIFORNIA
T H E G R A D U A T E S C H O O L
U N IV E R S IT Y P A R K
LO S A N G E L E S . C A L IF O R N IA 9 0 0 0 7
This dissertation, w ritten by
John Michael DeImage
under the direction of h.X3... Dissertation C om
mittee, and approved by a ll its members, has
been presented to and accepted by The Graduate
School, in p a rtia l fu lfillm e n t of requirements of
the degree of
D O C T O R O F P H I L O S O P H Y
Dean
DISSERTATION COMMITTEE
Chairman
DEDICATION
To Dr. Samuel Allerton, for his direction and en
thusiasm^ as these have been essential to my graduate
education.
To my wife^ Pat^ for her patience and encourage
ment during my graduate years.
11
ACKNOWLEDGMENTS
I thank the following for their continued support
in niy education^ and their assistance in this work:
Dr. Gregory Mooser^ for his friendship and his
helpful discussions dealing with critical perspectives in
scientific research in general, and of this project in par
ticular.
Dr. Nino Sorgente, for his enthusiastic support of
ray work and for his directions in cell transformation phe
nomena.
Dr. John (Jack) Beierle, for his many useful dis
cussions and for his aid in the immunological aspects of
this work.
Dr. Harold Slavkin, for his encouragement and his
aid in obtaining financial support for my graduate educa
tion .
Drs. Darleen Powers and Roger Terry, for their help
in obtaining clinical specimens and related data.
Dr. Tohru Okigaki, for his generous gift of sup
plies and for his expert instruction in tissue culture
techniques.
Ms. Jeanne Joyce, for her assistance in the devel
opment of the nude mouse tumors and her expert care of the
animals.
iii
TABLE OF CONTENTS
Page
DEDICATION.......................................... 11
ACKNOWT.EDC-MENTS.................................... ill
LIST OF TABLES...................................... vi
LIST OF FIGURES.................................... vil
Chapter
I. INTRODUCTION ................................ 1
Clinical Wilms Tumor
Histopathology
Genetics of Wilms Tumor
The Immunobiology of Wilms Tumor
Biochemistry of Wilms Tumor
In vitro Studies
Wilms Tumor Animal Models
11. MATERIALS AND METHODS....................... 25
Tissue and Cell Culture Methods
Profile of the Nephroblastoma Patient
Care and Inoculation of the Nude Mice
In vitro Techniques for Primary
Culturing
Karyotyping Analysis
Soft Agar Techniques
Biochemical Methods
Tissue Extractions
Colorimetric Assays
Enzymatic Analysis
Gel Chromatography
Isoelectric Focusing
Immunological Methods
Immunization of Rabbits
Antisera Preparation
Antisera Absorption
iv
Page
Immunodiffusion
Immunoelectrophoresis
Absorbent Materials
III. RESULTS...................................... 47
Nude Mouse Tumor Induction
Nude Mouse Grown Tumors
In vitro Culturing
Media and Sera Tests
Growth Characteristics of Cell Cultures
Karyotyping Results
Soft Agar Techniques
Differential Staining Experiments
Biochemical Analysis: Precipitation
by Various Agents
Biochemical Analysis: Measurement of
Enzyme Activity
Biochemical Analysis: Determination of
DNA, Protein, and Carbohydrate
Binding Interactions: Biochemical and
Immunological Studies
Detection of Tumor-Associated Antigens:
The Gamma-Beta Antigen
Immunological Determinations:
EDA and Other Antigens
IV. DISCUSSION.................................. 185
V. SUMMARY....................................... 214
VI. BIBLIOGRAPHY................................ 2l6
V
LIST OF TABLES
Table Page
1. A List of Inoculated Nude Mice, including
Dates of Injections, Tumor Status,
and Termination Dates ....................... 57
2. Results of Soft Agar Growth Analyses......... Ill
3. Precipitation, by Various Agents............... 130
4. Determination of Enzyme Activities:
Phosphatase and Gamma Glutamyl-
transpeptidase .............................. 133
5. Determination of Enzyme Activities:
Hyaluronidase, Protease, and
Glueose-6-phosphatase ....................... 139
6. Determination of DNA, Protein,and Carbo
hydrate Contents of Samples from
Normal and Malignant Tissue Origin ..... l44
7. Tests for Antigen Presence in Samples
from Normal Neo-plastic Sources ............. l80
VI
LIST OP FIGURES
Figure Page
1. Karyotype Analysis: Chromosome morphology
with Method 1 .............................. 34
2. Karyotype Analysis: Chromosome morphology
with Method 2 .............................. 34
3. Karyotype Analysis: Chromosome morphology
with Method 3 .............................. 34
4. Wilms tumor : View of entire t u m o r ........... 49
5. Wilms tumor : View of slice through
midsection.................................. 49
6. Histological paraffin section of the
nephroblastoma .............................. 51
7. Histological paraffin section of the
nephroblastoma .............................. 51
8. A nude mouse with a subcutaneous
nephroblastoma .............................. 55
9. The tumor removed from the mouse in
Figure 8 .................................... 55
10. Histological paraffin section of
nude mouse 194 primary tumor............... 58
11. Histological paraffin section of
nude mouse 194 primary tumor............... 58
12. Cyst formation within the nude mouse
194 primary t u m o r ......................... 6l
13. Cyst formation within the nude mouse
194 primary t u m o r ......................... 6l
14. Histological paraffin section of the
nude mouse 195 primary tumor............... 63
1 5. Histological paraffin section of the
nude mouse 195 primary tumor............... 63
Vll
Figure Page
1 6. Histological paraffin section of the first
passage tumor of mouse 2 3 8 ................. 66
1 7. Histological paraffin section of the
mouse 2 3 8 .................................. 66
1 8. Histological paraffin section of the first
passage tumor of mouse 240................. 69
1 9. Histological paraffin section of mouse 240 . . 69
20. Histological paraffin section of the sec
ond passage tumor 303 72
21. Histological paraffin section of mouse 303 • . 72
22. An early colony growth of EPL cells:
Demonstrates cell morphology, multi-
nucléation, colony morphology, and
early "piling up" of cells................. 77
2 3. A confluent culture of EPL cells:
Demonstrates the "piling up" of
cells at confluency....................... 77
24. A primary culture of normal human embry
onic fibroblasts: Demonstrates
unique colony formation ................... 80
2 5. A primary culture of NUF, nude-mouse
grown nephroblastoma fibroblasts .......... 80
2 6. NUF fibroblasts after three passages ......... 80
2 7. A nearly confluent culture of NEP cells:
Demonstrates typical epithelial cells. . . . 83
2 8. An early colony of MIX cells: Demonstrates
epithelial-like cultures ................... 83
2 9. A nearly confluent culture of TuWi cells:
Demonstrates cell and colony morphology. . . 83
3 0. A growth response curve: Tests of
various sera................................ 88
3 1. Growth Curve: EPL c e l l s ..................... 9I
3 2. Growth Curve: MIX c e l l s ..................... 9I
viii
Figure Page
33- Growth Curve: NEP c e l l s ...................... 94
3 4. Growth Curve : TuWl cells....................... 94
35- Growth Curve : NUF c e l l s ...................... 97
3 6. Karyograrn of nude mouse grown tumor.
Demonstrates "human type" with an
abnormal dicentric and fragment
chromosomes................................ 100
3 7. Karyotype profile : TuWl cells.
Demonstrates bimodal distribution........... 103
3 8. Karyotype profile : MIX cells.
Demonstrates maximum cell bursts ........... 103
39' Karyotype profile: EPL cells. Demonstrates
maximum bursts .............................. IO6
40. Karyotype profile : NEP cells. Demonstrates
between 54-58 chromosomes ................. 106
41. Karyotype profile: NUF cells. Demonstrates
possible bimodal distribution ............. IO9
42. Karyotype profile : NHF cells. Demonstrates
unimodal distribution ..................... 109
4 3. Soft agar colony formation : TuWi cells .... 113
44. Soft agar colony formation : EPL cells......... 113
4 5. Soft agar colony formation: NEP cells......... 115
46. Soft agar colony formation : NUF cells......... 115
4 7. A frozen section from nephroblastoma
patient E.C., stained with
Ruthenium R e d .............................. II8
48. Differential staining of cultured cells :
NEP cells stained with Toluidine Blue . . . 122
4 9. Differential staining of cultured cells:
NEP cells after Varidase digestion ........ 122
5 0. Differential staining of cultured cells:
NEP cells stained with Ruthenium Red .... 125
ix
Figure Page
5 1. Differential staining with Varidase
digestion.................................... 125
5 2. Differential staining: NUF cells with
acetic orcein ................................ 127
53- Differential staining of cells with
Toluidine B l u e .............................. 127
5 4. Demonstration of abnormal Immuno
electrophoresis banding..................... l4g
55. Demonstration of antigen "blockage" by
Immunoelectrophoresis ....................... 149
5 6. Demonstration of abnormal Immunoelectro
phoresis banding in. the presence
of hyaluronic a c i d ......................... 149
57. Chromatographic demonstration of binding
interaction, between hyaluronic acid
and serum protein............................ 153
5 8. Rocket Immunoelectrophoresis .................. 157
59- Demonstration of immunogen!city of
peritoneal fluid ............................ 157
6 0. Demonstration of altered immunogenicity
of peritoneal fluid after
Varidase treatment ......................... 157
6 1. Demonstration of "blockage" of the GB
antigen...................................... I63
6 2. Separation of GB antigen presence from a
peritoneal fluid sample ..................... I66
6 3. Isoelectric focusing of GB Antigen............ I68
64. Demonstration of the heterogeneity
of fetuin.................................... I73
6 5. Demonstration of FLA presence only in
cultures grown in the presence
of fetal calf serum......................... 173
X
Figure Page
66. Examination of antisera to FLA absorbed with
normal human sera............................ I83
6 7. Examination of antisera to FLA absorbed with
normal human plasma .......................... I83
68. Examination of ant is era to fetnin............. I83
XI
CHAPTER I
INTRODUCTION
Human cancer Is often thought of as a disease of
middle and late adulthood. Malignancy, however, is the
second leading cause of death in children under ten years
of age. Leukemia, a relatively common disease, has been
the focus of considerable research interest. Solid tumors
of children, though somewhat rarer, also claim a number of
lives each year.
In youngsters, many of the solid tumors are "embry
onal” growths, being distinct from either the adult-type
epithelial carcinomas or the connective-tissue sarcomas.
They are histologically similar, in some respects, to cer
tain fetal tissues, and are thus termed "embryonic.” In
cluded in this group of malignancies are: neuroblastomas,
retinoblastomas, and nephroblastomas (Wilms tumor). This
thesis is concerned with the study of one of these child
hood malignancies, Wilms tumor.
A number of investigative groups have a potential
interest in the study of model systems for human nephro
blastoma, as well as other "embryonic” tumors. Some of
these are : clinicians, immunologists, geneticists, develop-
1
mental biologists, and molecular biologists.
1. To the clinician, additional research may pro
vide improved diagnostic and therapeutic methods
for nephroblastoma patients. Possibly, the
rarity of the disease and the increasing effec
tiveness of chemotherapy have tended to discour
age in-depth research studies which could
improve the clinical management of this malig
nancy. A Wilms tumor analog could provide a
means for testing new chemotherapeutic agents
and innovative therapeutic regimes.
2. Recently, immunologists have become interested
in tumor-associated antigens of nephroblastoma.
Several reports have now been published concern
ing "marker" antigenic substances in these
tumors. In this case, a model (utilizing human
tissues) could provide a means to monitor the
synthesis of these substances, and a method of
studying their effects.
3. Geneticists could perhaps also profit from a
nephroblastoma model by using it for a genetic
analysis of hereditary and non-hereditary forms
of Wilms tumor. Further, the association of
this disease with certain congenital abnormali
ties could be better investigated.
3
4. Developmental biologists would have a malignant
tumor model, that may have been derived from
abnormal differentiated tissues. The study of
this system may then provide a means for eluci
dating the process of normal development and
maturation of the human kidney.
5. Molecular biologists have been recently inter
ested in studying the production of cell-
membrane-associated components. In the past
few years, our own laboratory has been involved
with studies of cell-surface glycosaminoglycans
(acid mucopolysaccharides) of Wilms tumor. The
development of an analog to nephroblastoma would
aid in the study of these, and many other, sub
stances.
A number of interesting and important questions re
main to be answered about nephroblastoma. A reliable model
system would be useful in attempts to answer a few of these
questions. Essentially, this research describes the estab
lishment and partial characterization of a Wilms tumor
model. Specifically, human malignant tissues have been
grown in athymic mice, and comparisons have been made be
tween these secondary growths and native human tumors. In
this work, histochemical, biochemical, and immunological
methods have been used to compare the animal-grown to the
human-grown nephroblastoma tissues.
: .......... ^ ^ 4 I
i
In addition to the "nude mouse" experiments, stud- '
ies have been completed which reevaluate our previous find-
I
ings on biochemical analyses of Wilms tumor extracts. Of !
j
particular interest has been the examination of binding
interactions between soluble proteins and extractable poly-i
saccharide components. This thesis describes (with the
animal model) the results of biochemical and immunological j
techniques used to study this binding phenomenon. ,
Clinical Wilms Tumor
The first description of human nephroblastomas was
presented by Wilms in I8 8 9. Originally, the disease was |
i
described as an embryonic renal neoplasm, a "mixed embry- j
oma" consisting of a variety of glandular components, con- |
nectlve tissue, and muscular elements. Since the early
pioneering work of Wilms, much has been learned about the |
clinical picture of these tumors. However, the genetic,
biochemical, and immunological features remained largely
unknown. j
i
Clinically, Wilms tumor is the third most common |
malignancy of children. The affected age group is gen- :
erally from two to five years of age, with an incidence from
about 1 in 1 0 0 ,0 0 0 to 1 in 2 0 0 ,0 0 0 live births (Canale),^ ,
although the tumor is rarely found before 6 months of age i
(Powars). The only major symptom of the disease is a i
large abdominal mass. Occasionally, blood pressures and i
5 |
erythrocyte sedimentation rates are also elevated. Usu- '
ally, a preliminary diagnosis of nephroblastoma is made on ,
the basis of abdominal X-rays and intravenous or retrograde |
pyelograms. Definitive diagnosis is based on pathological '
examination of surgical specimens. Histological procedures |
are especially critical for distinguishing between true !
Wilms tumor and mesoblastic nephroma, a benign tumor that id
'morphologically similar to nephroblastoma. Normally, the ;
benign disease is found in infants from birth to about two |
iyears of age, and is simply cured by surgical excision. j
Nephroblastomas are often large tumors, varying in size |
: from 80 to 3000g (with an average of 550g). Externally, |
they are usually surrounded by a pseudocapsule; internally, |
' I
the tissues are soft, with cream to grey colored amorphous |
g 4 !
masses (Potter,^ Rosenstock ). These true nephroblastomas |
I n
: grow very rapidly. For example, Grossman found cases in
I
1 which children had normal retrograms at one period, but had j
; i
massive tumors within nine months or less. :
Nephroblastomas, like other melignant tumors, metas6
tasize readily. Most often, the lung is the site of seconds
ary tumor development, but tumor spread to liver, bone, and |
i
other tissues is not uncommon. ;
The successful management of Wilms tumor is based I
i
on vigorous multimodal treatment plans. The primary weapon I
against the disease is surgical removal of the primary i
6
tumor. Additional therapies to prevent or control non-
resectable tumor tissue include chemotherapy (with Actino-
mycin-D, Vincristin, or Adriamycin) and X-ray irradiation.
A number of problems have been encountered, though, with
current chemotherapy and radiation techniques. There are
substantial risks of toxic overdosage and induction of new
tumors. Adriamycin is effective in some cases. Still,
this drug has caused cardiotoxicity (Prout).^ Actinomycin-i
D and Vincristin, as well as X-irradiation to the spinal
column, have also been associated with developmental abnor
malities, as well as with renal failure (Pochedly).^
I
Histopathology
Histologically, Wilms tumor is characterized by
primative glomeruli and abortive tubule formations, with
spindle cell stroma. These tissues bear a striking resem- I
blance to the developing metanephros (fetal kidney). |
A few pathologists have recognized that the nephro-;
genic tissues of nephroblastoma can produce muscle, fat, or :
cartilage, as well as hair, bone, and other tissues. Still,}
these tumors are considered neither teratomatous nor mixed ;
(Ceschickter and Widenhorm).^ Rather, they are thought to
be a neoplastic exaggeration of a normal developmental
process in the renal cortex of late fetal life (Willis).^
Other investigators have concluded that the metanephric
i
blastema is responsible for the production of renal connec-
7
tlve tissue, as well as for the formation of nephrons. As
such, this primitive tissue is a type of stem cell popula
tion that is required for the development of both tubular
and interstitial elements ( Potter). In Wilms tumors,
both sarcomatous stromal cells and epithelial tubular ele
ments are present simultaneously. In most cases, the char
acteristic picture is one of "islands" of primitive nephro
genic cells surrounded by swirling "sheets" of interstitial
stromal cells.
Under normal circumstances, the ampullae of the
ureteral limb buds act to stimulate an ordered prolifera
tion and differentiation of metanephric blastema cells.
These latter cells, in turn, form nephrogenic and stroma-
genic cells. The nephrogenic cells normally mature into
tubular and glomerular tissues, while stromagenic cells
form the connective elements of the kidney. In nephro
blastomas, the developmental process of ureteral induction
appears to be interrupted. As a result, the metanephrlc
blastema cells are left as "rest" masses with the potential
for continued proliferation. In addition, the tumor cells
maintain the ability to produce primitive tubular and con
nective tissue elements. Histologically, Wilms tumor is
very consi stent with such a developmental pattern
(P ot t e r , Powars^).
8
Genetics of Wilms Tumor
The genetic aspects of human nephroblastoma have
become especially fascinating in recent years. Statistical
'analysis of 97 Wilms patients revealed two categories of
Individuals (Knudson) . They were the familial and bilat
eral cases in one group, and the non-familiai (and non
bilateral cases) in the other. Familial and bilateral tu
mors account for 38^ of all the cases. In this group, di
agnosis was at a relatively early age, which appeared to be
consistent with an autosomal dominant inheritance model.
Spontaneous, non-familiai (and non-bilateral) cases com
prised the majority of nephroblastoma patients surveyed.
These tumors appear at random in the general population at a
rate of 1 per 100,000 live births. A model has been de
signed to explain the two categories of afflicted individ
uals. This model consists of a two-mutation system, in
which familial and bilateral tumors occur in those with a
predisposition for the disease, the first mutation occur
ring in the germ cell line. Such a condition could be at
tributed to the presence of an inherited mutation (or a
virus acting as a mutagen). In these children, only a
single additional mutational event would be required for
induction of the malignancy. In the non-familiai and uni
lateral patients, two somatic mutational events would be
required for tumor formation. This model includes features
9
which help to explain the rarity of the disease in the
general population, as well as its high incidence in some
families. Associations of nephroblastoma with aniridia,
hemihypertrophy, genitourinary anomalies, and trisomy of
chromosome l8 are thought to lend support to this genetic
theory of tumor induction.
Circumstantial evidence supporting a genetic basis
of this tumor has come from several investigative groups.
12
In a review of 440 Wilms patients, Giangicomo found con
genital abnormalities in 67 individuals (15^)- In one case,
a patient had a chromosomal group B-C translocation. Other
17
reports of genetic association have been made by Brown.
This investigator described a familial case in which four
children in three successive generations developed the
disease. In each case, the right kidney was involved, with
one instance of bilateral tumor. A second report of this
kind was written by Strom. In this article, five chil
dren of three generations were afflicted by tumors.
The presence of congenital malformations with Wilms
tumor has been reported often. Rosenstock^ has found
horseshoe shaped kidneys in several nephroblastoma pa
tients. Also, Canale^ has pointed out instances of micro
cephaly, mental and physical retardation, and pseudoher
maphroditism in some patients.
i 1 0 I
The Immunobiology of Wilms Tumor *
A consistent immunological description of Wilms }
tumor has proven highly elusive. Several investigators |
have pursued studies of tumor-associated antigens in human ;
nephroblastoma. Despite intensive efforts, the association
of distinctive "tumor" antigens with this disease remains !
' i
somewhat uncertain. i
ID '
Work in our laboratory by Wise and others has re- ;
i
suited in the discovery of a fetuin-like antigen in at i
least some nephroblastomas. This antigen (FDA) was found |
in a pooled sample of ethylenediaminetotracetate (EDTA)
chelate extracts of tumor tissue from five patients. Anti- !
bodies directed against this material did not react with |
I
fetal tissues (spleen, kidney, thymus, and sers). Immuno- :
i
,logical cross reactivity was also lacking with extracts of ;
: I
'renal clear cell carcinoma, alpha-fetoprotein (AFP, of i
ihepatomas and fetal sera), and carcino-embryonic antifen |
(CEA, of stomach, intestinal, and uterine carcinomas). !
Antibodies to FLA did, however, react with the fetal bovine |
glycoprotein, fetuin. Chemically, the tumor-associated |
antigen seemed to contain both protein and carbohydrate i
components, as pronase treatment abolished recognition of ,
the antigen by directed antibodies, and hyaluronidase i
treatment altered the antigen-antibody precipitin pattern -
observed with an agar double diffusion ( Ouchterlony) tech- I
nique. I
1
At the time Wise ^ and others in our laboratory were'
: n 5 '
Investigating PLA, Buffe found an abnormal alpha-2 (mo
bility) protein in 75^ of Wilms extracts. He also found j
: this protein component in neuroblastomas and in a variety |
of other pediatric tumors. More recently, this antigenic |
- ] y I
material has been identified as an Iso-ferritln (Alpert) '
which seems to be an acidic protein, not found In normal >
liver ferritin preparations. ;
Another report of a Wilms tumor antigen was pub-
2 0 i
: 11 shed by Burtin and Gendron. The material they Iso- |
;lated (the "W" antigen for Wilms) was found In phytic acid
I (insltolhexol-phosphoric acid) extracts of 45^ of the |
; tested nephroblastomas. "W" antigen was found throughout |
the entire alpha and beta mobility zones in immunoelectro-
j phoreses. With aged extracts, two antigenic components
were noted, with crossed lines denoting non-identity. As
^in the work of Wise, no cross reactivity was found with
ID
fetal tissues, nor with normal kidney. Wise, as well as
20
Burtin and Gendron, also found no immunological similari
ties to AFP and CEA. However, unlike the results of our
20
laboratory, Burtin and Gendron found the "W" antigen in
renal clear cell carcinoma and in mammary carcinomas. They
noted that all of the tumors containing this antigen were
comprised partly of glandular elements. Chemically, the
"W" antigen was destroyed by treatment with perchoric acid
12 ;
: i
or pronase; the substance could be stained with alcian blue,
and periodic acid Schiff reagent, indicating the presence
of a carbyhydrate constituent. Unfortunately, the rela- ;
tionship between the FLA and "w" antigens remains uncer- I
tain, as no direct inter-comparisons have been made (be
tween the antigens and antisera preparations).
19
Kumar and Taylor ^ have also been active in the
immunobiology of human nephroblastoma. In two publica
tions, they have described tumor-specific antibodies in a
child with Wilms tumor. This patient had antibodies di
rected against cytoplasmic and cell surface components of
the tumor cells. In these experiments, tumor specimens |
were examined for antibody binding by a fluorescein-conju- |
gated isothiocyanata labeled dual antibody technique. The
chemical nature of the antigenic material that induced i
antibody formation was not, however, determined. These
data, then, are difficult to correlate with the findings ^
o n 2 1 1 6
of Burtin, Wise, and Buffe. There is, though, strong
circumstantial evidence among all of these reports for at |
least one tumor-associated antigen of glycoprotein nature I
in human nephroblastoma. '
Biochemistry of Wilms Tumor
Biochemical manifestations of nephroblastoma have
22
been recently reviewed by Pochedly. Included in chemical
abnormalities associated with this tumor are: abnormal
1 3 :
enzyme profiles, altered serum protein levels, the presence
of acid mucopolysaccharides (glycosaminoglycans) in the
sera and urine, and elevated erythrocyte sedimentation
rates. In particular, there are elevations in serum lac
tate dehydrogenase activity, elevations of alpha serum pro- ;
teins, elevated serum haptoglobin, elevated erythropoietin,
and a decrease in serum albumin (Murphy, Mirand, and Stau-
21 24 2h
bitz, Shalet. Holder, and Walters, Morse and Nussbaum,
26
and Allerton, Beierle, Powars, and Bavetta ).
There have been a number of publications on altered
enzyme levels in neoplastic tissues. Specific patterns for
nephroblastoma have been investigated primarily in experi-
27 !
mental animal nephroblastomas. Jasmin and Riopelle found j
!
that carcinogen-induced rat nephroblastomas have suppressed |
oxidative and hydrolytic enzyme activities. Included in |
these tests were: succinate dehydrogenase, hydroxybutyric |
dehydrogenase, eytochrome oxidase, acid phosphatase 5'-
nucleosidase, and nonspecific esterase. Thomas, Wessel, and|
28 '
Citoler also studied chemically induced Wilms-like tumors
of the rat. They found slightly suppressed levels of acid
and alkaline phosphatase, nonspecific esterase, succinate i
dehydrogenase, and glucose-6-phosphatase. 1
Developmental changes of enzymatic levels in normal i
mammalian kidney tissues have also been well documented. Of
particular interest are the elevation of UDP-pyrophosphatase
- - - - -
and glucose-6-phosphatate levels in the adult proximal
tubule cells of the kidney as compared to those levels of
I the developing tubule cells (Pretlow, Jones, and Dow,
'Cordex^^).
Elevation of serum mucopolysaccharides associated
with Wilms tumor was first examined in detail by Morse and
2D
Nussbaum. These investigators found a hyaluronidase-
sensitive, acid precipitable component in the sera of
nephroblastoma patients. Our laboratory has pursued this
observation further, with the result that large quantities
of EDTA or phytic acid-extractable "mucin” were identified
in Wilms tumor, as well as in some of the patients' sera
(Allerton et al.).^^ This material was found to precipi
tate in acetic acid (3^), ethanol (50^), cetylpyridinium
choride (0.2501.0^), and ammonium sulfate (33-75$). Chemi
cal analyses of tumor extracts have indicated the presence
of hyaluronic acid and chondroitin sulfates as major ex
tract components. Follow-up studies on Wilms patients
showed a disappearance of the "abnormal” mucins from the
sera of these children, upon successful resection of their
tumors.
Several reports have been made concerning the al
teration of glycosaminoglycan (GAG, mucopolysaccharide)
content of transformed cells and neoplastic tissues. For
example, some human lung carcinomas were shown to have
15
markedly increased concentration of GAGs, particularly the
chondroitin sulfates and hyaluronic acid (Hatae and
Mikita).^^ Syrian hamster embryo cells and green monkey i
kidney cells were also found to produce increased quantities
of mucopolysaccharides when transformed by herpes simplex
type-2 or by SV-40 viruses (Satoh, Duff, and Davidson,
Mikita and Shimojo^^). In each of these latter experiments,
both hyaluronic acid and heparan sulfate concentrations
were elevated in the transformed cells, as compared to con
trol cells. Increases of cell surface polysaccharides are
not always associated with viral infection, however. In
qli
two independent studies, Underhill and Keller-^ and Roblin,
ID
Albert, Gelb, and Black-^ found decreased concentration of
GAGs in SV-40 transformed mouse embryo 3T3 cells, as com
pared with control 3^3 cells.
The effects of mucopolysaccharide-rich exogenous
Wilms tumor extracts on in vitro cell cultures have also '
been examined in our laboratory (Beierle, Allerton, and
\ 26 :
Bavetta) . In these studies, EDTA extracts of pooled |
human Wilms tissues were added to several "in vitro" cell i
I
cultures. Examinations of the growth rates for treated
cultures indicated a "selective" effect in increasing the
growth of kidney and lung cell populations. Skin and muscle
cells did not, however, show a significant response to the ;
addition of the EDTA extracts. In another laboratory.
16
27
Ohnishl, Oshlma, and Ohsuka'^' took the opposite approach to
the study of cell surface components. They exposed hepa-
!toma cells to polysaccharide extracts from normal liver |
'tissue. They found a decrease in growth rates and satura- !
tion densities. Together, the above two sets of experiments
suggested a regulatory function for cell surface acid muco
polysaccharides (from normal as well as from neoplastic cell
sources). An obvious difficulty in both studies was the
failure to show that the polysaccharides per se, or indi
vidual polysaccharides in the tissue extracts, were in fact
the regulatory agents. That is, crude extracts were used ;
in both laboratories. As such, the observed effects cannot |
be directly correlated with the presence of particular |
glycosaminoglycans.
It is difficult to correlate the above biochemical ;
findings with a specific cause for Wilms tumor. Still, a
few laboratories have been interested in comparing nephro-
blastoms to other embryonic tumors for possible etiological ;
similarities. Most of this work has involved a "viral ap- |
proach" to tumor induction. For example, Werthein and |
Voute^^ found a 50^ incidence of antibodies to cytomegalo- !
virus in Wilms tumor and neuroblastoma patients. However, :
electron microscopic examinations have failed to reveal the
presence of viral particles in tumor tissues. The connec- I
!
tion between viruses and nephroblastomas remains unsup
ported, at least for the present. I
17
In vitro Studies
Human Wilms tumor cells have been occasionally used
in experiments involving short term tissue culture. One of
! the most thorough of the reported trials was by Waghe and
Kumar,who described methodology for developing and pre
serving primary cultures from tissue specimens. The major-
;ity of these cultures were fibroblastic in appearance, with
abnormal mitotic figures and multiple nucleoli. The rapid
growth of "in vitro" Wilms tissues has been positively cor-|
related with high thymidine indexes of the cultured cells.
Unfortunately, these cultures were viable for only one to a
I
few population doublings.
A second group of investigators has also estab-
■lished methodology for the short-term cultivation of human ■
I
: nephroblastomas. In the words of Nuss, Duursma, and Pudi- i
I 4o
' fun, cultures were readily formed, with the resulting I
I
cells having "bizarre" shapes and sizes, and lacking an
orderly growth pattern. According to these authors, their
"abnormal" polyhedral cells were presumably of fibroblastic ;
origin. :
Differentiation of short-term Wilms cultures has !
been recently reported by Rousseau-Merck, Lonbard, Nezelof,
and Mouly. In this work, chick and mouse embryonic tis
sues were used to induce kidney tubule formations in culti- :
vated cells. These experiments resulted in a high correla-
: 1 8 1
!
:tion between the presence of tubules in tumor specimens, !
I i
and the induction of tubule formations in cell cultures |
from these tissues. That is, tumors with abundant abortive :
tubules often gave rise to cell cultures that are inducible ;
for tubule formation. Conversely, tumors that lack spe- |
!
cific, differentiated components generated cultures that
lack the capability to form tubules on in vitro induction.!
Cultured human Wilms cells have also been commer- |
cially available for a number of years through the American |
I
Type Culture Collection (ATCC cells). These cells :
'have from 60 to 65 chromosomes and are epithelial-like in I
, i
appearance (the original culture was reported as fibro- 1
\ 42
•blastic in nature). In one report. Wise and Muller found
! that EDTA extracts of TuWi cells were immunologically simi-
!
ilar to comparable extracts of native human tumors.
Some possible difficulties with the use of these
cells have also been reported. Nelson-Rees and Flander-
meyer^^ and Elandemeyer, Hawthorne, and Nelson^^ have re
ported the presence of glueose-6-phosphate dehydrogenase
(G-6-Pd) Type A in TuWi stock cultures of the American Type
Culture Collection. This isozyme has been typically con
nected with HeDa cell strains (malignant uterine carcinoma
cells). Still, it is impossible to establish "bona fide"
cellular contamination on the basis of G-6-Pd activity
alone. In the case of HeLa, cell strain associated anti-
I9i
genic expression is far more reliable. With the TuWi cells |
such an immunological approach has not been reported; none
theless, work with the TuWi cells should be evaluated with
caution. |
Nephroblastoma cells of animal origin have also
been cultivated jji vi tro. Priestly reported the forma- |
tion of an animal system analogous to human Wilms tumor.
In these experiments, spontaneous rat tumors were used to
develop cell cultures. The isolated cells, in turn, served
as the basis for an evaluation of chemotherapeutic methods ^
for the treatment of the human disease. Laboratory tests
included examinations of cell sensitivities to Antinomy- i
45
cin-D, vincristin, and X-irradiation. Priestly found
that his vitro system demonstrated responses similar to
those found clinically with human patients. In particular, j
I
he found that a multimodal therapy regime of chemotherapy ;
plus irradiation was the most effective in eliciting spe- |
cific toxicity (when compared to the sensitivity of "normal'*
human kidney cells vitro). This latter observation is
consistent with the clinical findings of studies involving |
afflicted children who have been treated with multimodal '
therapy.
Wilms Tumor Animal Models I
Within the past two decades, several attempts have
been made to establish Wilms tumor analog models. In one i
20
46
of the first of these, Olcott described the successful
transplantation of rat nephroblastomas (to other rats).
Subsequently, other investigators also reported transplant-
4 7
able renal tumors of the rat (Babcock and Southham) . ' In
both of these early reports, tumors arose spontaneously in
large inbred animal colonies. These models proved to be
unmanageable, though, in that the initial neoplasms gen
erated were random and rare events. This lack of control
over tumor induction has dampened progress with such models.
In fact, after almost twenty years, the initial articles
described above are without follow-up publications.
More recently, investigators have pursued carcino
gen-induced animal tumors. In I9 6 8, Hirono, Laqueur, and
Spatz"^^ were able to induce twenty-three tumors in inbred
Fisher rats. Of these, ten were considered nephroblastomas|
by histological criteria. These neoplasms were transplant- :
!
able through several generations of animals. Interestingly,
I
the transplantation rate for male recipients was 100$, while
!
that for females was only about 50^. These authors have
suggested that estrogen in the female animals may have a
i
constraining influence on the induction of such tumors. |
p y
Another research group (Jasmin and Riopelle) re
ported nephroblastomas in ovariectomized female Sprague-
i
Dawley rats. In this case, the carcinogen 7,12, dimethyl-
benzanthracene (DBMA) caused Wilms tumors in l4$ of the
21
treated animals. The resulting growths were histologically
typical of nephroblastoma. Analysis of enzymatic activi
ties in these tumors indicated reduced levels of both oxi
dative and hydrolytic enzymes of differentiated epithelial
cells (as found in mature tubule cells and in most adult
kidney carcinomas). In these studies. Jasmin and Riopelle
found tumor induction in ovariectomized animals. This
finding is consistent with the suggestion of estrogen in -
48
fluences as reported by Hirono et al.
Animals other than rats have also been used for
Wilms models. Nephroblastomas were produced in l8^ of an
inbred Jackson-Wh rabbit colony by transplacental induction
with 1-ethyl-1-nitrosourea (Fox, Diwan, and Meier).Also
49
in a mouse system, Javadpour and Bush took a novel ap
proach to the problem of developing a Wilms animal model.
They injected fragments of normal mouse renal blastema
(fetal metanephric kidney stem cells) into the testis of
forty adult male mice. Tumors formed in 60^ of the treated
animals. These experiments were interpreted to support the
theory of a "blastema rest" origin of nephroblastomas. Un
fortunately, their attempts to duplicate this work were not
successful (Mount, Thelmo, and Husk).
I, with other members of our laboratory, felt that
a DMEA-rat model would be useful to compare with ongoing
studies of human Wilms tumor. In particular, our interest
I 22 I
was focused on the generation of a reproducible system to
correlate with studies of tumor associated antigens, and
tumor acid mucopolysaccharides (glycosaminoglycans). Under
the direction of Dr. Tohru Okigaki (at that time of the |
Pasadena Foundation for Medical Research), I attempted to
reproduce the work of Jasmin. Six pregnant female Sprague-
Dawley rats were inoculated with intragastric dosages of
DMBA (DOmg each) mixed with mineral oil. Our rationale for j
I
the use of pregnant animals was based on clinical observa
tions and current theories of tumor formation. Human Wilms
tumor is believed by many to originate in the metanephric
blastema tissues of the fetus. If this is the case, then |
these malignancies must begin to develop prenatally. The !
exposure of developing rat fetuses to a carcinogen that is
known to form nephroblastomas may then be a more effective
5 2 I
method of tumor induction. The work of Fox et al., de- |
scribed previously, illustrated the effectiveness of trans- |
placental induction of Wilms-like tumors. In our pilot
study, however, both the mother and offspring animals ap
peared healthy nine months after inoculations. Both :
groups of animals were eventually sacrifices, and various |
tissues (kidney, liver, and spleen) were removed for histo
logical analysis. None of the tissue specimens demonstrated
abnormalities. Due to our lack of success in developing
tumors with DMBA, this approach was deemed unsuitable. Our
; 23
i attention since that time has been redirected to other ap-
! proaches to an analog system.
^ Prior to our DMBA studies. Wise and Muller'^"^ were
able to grow commercially available Wilms cells (TuWi) in
nude mice. They injected solid tumor formations in 70^ of
' 1
their animals by injecting athymic (nu/nu) nude mice sub- !
;cutaneously with a suspension of TuWi cells. In this sys- '
:tem, karyotype analysis indicated a complement of 6O -6 5 ;
' human chromosomes in the neoplastic tissues. This is essen-|
itially identical with the chromosome values for TuWi cells I
: I
: grown vitro. Wise and Muller also found elevated blood j
I gamma-glutamyl transpeptidase activity associated with the |
formation of solid tumors in the mice. This latter finding ;
R Q I
I is consistent with the studies of Ru tenberg, who found ]
I high levels of this enzyme in a number of malignant tissues
i
as well as in the brush borders of kidney proximal tubule
cells.
21
The nude mouse model of Wise and Muller was the
first report of an animal carrier for the growth of Wilms |
cells. There are, however, major disadvantages to their |
system. As described previously, there is now concern aboutj
i
possible HeLa contamination of the TuWi cell cultures. In
I
addition, the commercial Wilms line has undergone "trans- i
formation" due to an extensive time and repeated transfers |
in vitro. Included in this transformation are altered I
chromosomes numbers and altered morphology. As such. It Is
difficult to assess similarities between TuWi tumors and
native human tumors.
Our laboratory has made a second attempt to estab
lish a model for human nephroblastoma. For these studies,
the nude mouse carriers were used, as in the work of Wise
21
and Muller. However, in our experiments, fresh human
tissue has been utilized as a cell source instead of tumor
cells or the TuWi cells. In this way, we have been able to
avoid the disadvantages of possible cellular contamination
(with HeLa) and those of cellular transformation, since our
tissue source was not previously maintained vitro. This
thesis is primarily a report of our success in establishing
human Wilms tumors in nude mice by this approach.
CHAPTER II
MATERIALS AND METHODS
Tissue and Cell Culture Methods
Profile of the Nephro
blastoma Patient
For the induction of Wilms tumor growths (of human
origin) in nude mice, the tumorous tissues from a single
nephroblastoma patient have been used. This white male
patient of Spanish-American descent was first seen as a
newborn infant with bilaterally enlarged kidneys and hyper
tension . At that time a diagnosis of mild bilateral poly
cystic disease was made. At 27 months of age (on routine
clinical evaluation), a large left flank mass was identi
fied. Subsequent tests, including intravenous urograms,
renal angiograms, and chest x-rays, revealed a tumor of the
left kidney. The mass was surgically resected through a
total nephrectomy of the left kidney; the tumor was then
identified as a Wilms tumor (nephroblastoma) by histologi
cal examination. Pathological evaluation also revealed a
cystic deformity of the normal portion of the left kidney.
After postoperative recovery, the child received adjuvant
chemotherapy with Actlnomycin-D for 18 months.
25
26 I
At six years of age, the child was seen at a spe
cial follow-up because of increasing abdominal pain. Clin- '
ical examinations revealed a second Wilms tumor in the |
right kidney. As the child’s entire left kidney had al
ready been, removed, only a partial nephrectomy of the right
kidney was performed. The patient has made an uneventful
postoperative recovery, and is currently receiving chemo
therapeutic agents including Adriamycin, Actinomycin-D, and
Vincristine. For the development of an animal carrier sys
tem, fresh tissue was taken from the second tumorous kid
ney. The malignant tissue was taken immediately from the
operating room (using sterile techniques) and transported
in sera-free culture media to the gynobiotic facility at
the use School of Medicine, where the nude mice are housed.
These tissues were minced and directly implanted subcuta-
neously into the mice.
Care and Inoculation
of the Nude Mice
Athymic nude mice (genetically homozygous nu/nu.
Use NIH [S ] ) were raised and maintained in non-sterile
flexible plastic isolators. The animals were fed Wayne
Lab-Blox for Mice (Allied Mills, Illinois) and sterile
water ad libitum. For inoculation, fresh pieces of human
primary tumor (in sera-free RPMI l620 media. Microbiologi
cal Associates, Maryland) were passed into an aseptic sur-
27
:gery isolator where the tumor tissue was minced and placed
, into 1/2 cc syringes, fitted with l6 gauge needles. The
mice (about two years old) were lightly anesthetized with
methoxyflurane (Pitman-Moore, Inc., New Jersey) and in
jected with tissue fragments subcutaneously in the sub
scapula region (with an inoculations of about 1/2 cc of !
minced tumor for each mouse), All subsequent serial trans- I
plants were accomplished with the methods as described
! above. i
[ Two of three mice in the initial induction were |
inoculated with soft, cream colored tumor tissues; the
'third animal received adipose-like tissues from a different |
: I
: region of the tumor. The two animals with the "soft" tis- |
sue developed tumors approximately two cubic centimeters |
: within a five and one-half month period. The third animal
: did not form a tumor even after nine months of incubation,
i The primary animal-grown tumors were serially passed, with
seven of eight (in the second) group forming tumors in
about three and one-half months. A third group of animals
(the second passage) also developed tumors (of up to 6
grams) in slightly over three months.
In vitro Techniques for
Primary Culturing
Attempts to establish primary Wilms tumor cell cul
tures involved the utilization of several in vitro tech-
; 281
^ niques. A critical first step in these culturing proce
dures was the procurement of fresh sterile tissues. In |
I
I most cases, human nephroblastoma tissue was received from i
j
: the Los Angeles County General Hospital during or shortly |
after surgical resection. Tissue pieces were immediately I
placed into chilled growth media (either MEM, Hams-F-12, or|
Hams F-10 without sera), and were quickly transported to |
our laboratory at the main USC campus. Normally the tumors ;
were at our tissue culture facility within one to two hours j
of surgical removal. i
: !
Initial tissue preparation included carefully dis- |
; secting away necrotic tumor regions, pseudocapsules, and
; i
major ureteral elements. The non-necrotic tumor regions |
were minced with sterile knives (Becton and Dauldison, |
I
: Philadelphia), and dislodged cell aggregates were isolated |
by centrifugation. This last step was accomplished by !
: centrifuging the minced tissue fragments in 15 ml conical |
centrifuge tubes (Corning, sterile) at 3 ,0 0 0 x g for 10 |
minutes at 4^ C. The cell pellets were then washed in |
growth media containing 1 0-20/ sera, then re-centrifuged. j
These finely minced tissues and cell isolates were used in |
a number of culturing methods to establish primary explants. '
An early attempt was made to culture both tumor tisH
sue and peri-tumor "normal kidney tissue," according to i
these methods. The tissue fragments prepared as described |
: 29
were trypslnlzed In O.25/ trypsin (GIBCO, Los Angeles) at
;37° for 30 minutes. Cells isolated by this method were
■removed mechanically from the incubation solution. The re-
imaining tissue fragments were placed into T-3 0 (Corning,
Baton) culture flasks with 10 cc of growth media (MEM with
15/ fetal calf serum). Trypsin-isolated individual cells
and small cell aggregates were washed in growth media and
: centrifuged; they were then placed into T-6 0 flasks (Corn
ing, Baton) with 20 cc of growth media. The plating den
sity in each of the cultures was about 2 0 ,0 0 0 cells per
2
cm . The growth media in each flask were changed at inter
vals of 24 hours for a period of three days. In the early
media changes, the used MEM with loose non-adhering cells
was discarded and decanted.
I Modified techniques for establishing primary cul-
! tures involved the use of alternate enzymatic tissue dis-
iruption techniques, coupled with cell separations based on
density. In these experiments, minced tissues were split
into three groups. One group was trypsinized (as indicated
above), then placed into growth media (Hams P-12 with I5/
fetal calf sera). A second group of cells was layered onto
3 X volumes of Eicoll-Paque (Pharmacia) in conical centri
fuge tubes. For this procedure, only single cells and
small cell aggregates were used. The multi-layered solu
tion was then centrifuged at 750 x g for I5 minutes. Those
: 30
cells remaining at the interface (between the Ficall and
the cell solution) were removed and washed for culturing.
After rinsing in growth media, these cells were plated in
: the usual manner. The third group of cells was treated
with a combination of hyaluronidase (from ovine testis.
Sigma Corp., Los Angeles) and collagenase (from Clostridium
histolyticum, Worthington). In this procedure, cells were
incubated in the presence of 0.2/ of each enzyme for thirty
minutes at 37° C.
After the appropriate incubation, the cells were
washed in growth media and plated in the usual manner. In
one set of experiments, the disaggregation with hyaluroni
dase and collagenase was followed by the cell separation
in Ficoll-Paque.
Karyotyping Analysis
Cells to be analysed were trypsinized twenty-four
hours before beginning karyotyping procedures. This "sub-
culturing" was accomplished by removal of growth media and
addition of 0 . 0 2 5/ trypsin in minimal essential medium
(without serum). The cells were incubated at 37° 0 for
twenty minutes. The loose cells were then transferred to
conical centrifuge tubes and spun at 800 x g for ten min
utes. Supernate solutions containing the trypsin were de
canted, and fresh media were added (l4-15ml each). The
cells were then suspended to "wash" them of residual tryp-
31
sin. Again the cultures were spun as before. The second
supernate was also decanted and discarded. The remaining
cell pellets were suspended in growth media containing 10/
fetal calf serum. These cell suspensions were added to
fresh Corning T-6 0 culture flasks. After incubating for
twenty-four hours (at 37 degrees, with 5^ CO^), ten micro
liters of mitosis-inhibiting agent was added.
In preliminary experiments, colchicine (at 0.8pg/ml
final concentration) was added to inhibit mitosis. After
four hours of incubation (at 3 7° C), the arresting agent
was removed, and the cells were trypsinized as described
previously. This method was in accordance with the pro
cedures described by Moorehead.^^ The dislodged cells were
washed in a hypotonic solution of 0.075 M KCL in distilled
water; they were then centrifuged at 8OO x g for five min
utes. The supernate was decanted and discarded. A second
application of the hypotonic medium followed with incuba
tion for 30 minutes at 3 0° C. After the incubation the
cells were spun at 5OO x g for five minutes. The hypotonic
supernate was then removed and discarded. To the cellular
pellet, three changes of Carnoy's fixative were added (75^
methanol plus 25/ glacial acetic acid). The first applica
tion of fixative involved addition of 2.5 volumes of the
Carnoy's solution to 1.0 volume of the cell pellet. The
cells were suspended in this solution and incubated for ten
32
minutes at 25° C. A second fixation step followed; for
this, the cells were centrifuged and a second 2.5x volume
was added to the cell pellet. Again the cells were incu
bated for ten minutes. A third fixation followed, using
the same conditions as the initial fixations. After this
final step, however, the cells were not centrifuged after
incubating. Rather, a 75x volume of fixative was added to
the suspension. At this point, the cells were aspirated
through a flame-drawn pasteur pipette. The cells were then
ready for spreading.
For this analysis, precleaned microscope slides
were used. They were prepared by rinsing in a detergent
solution, followed by several washes in distilled water.
Before use, the slides were chilled in 40/ methanol at
-10° C. For preparation of the spreads, the excess metha
nol was removed (but not to dryness). Single droplets of
the fixed cells were dropped onto the slides from a dis
tance of about ten inches. In this way, maximum spreading
was achieved. Cell spreads were dried immediately by pla
cing the slides in front of an electric fan. The alternate
procedure for immediate drying was also employed. For this
method, the methanol-acetic acid fixative was ignited with
a flame, and the slide was allowed to burn until dry (ac
cording to Moorehead). Once dry, the preparations were
stained with either Giesma or in 0.1/ acetic orcein (GIBCO,
33 I
Los Angeles), After staining for fifteen minutes at room
temperature, the slides were destained in water (with sev- ;
leral changes) for 30 minutes. Observations of chromosome I
!
spreads were made at 400x and lOOOx magnifications with a |
Zeiss standard microscope. Photographs of preparations
were taken at lOOOx (under oil). For photography, Kodak
high contrast film was utilized. Normally, exposures of 20
to 30 seconds were required for appropriate exposure and
optimal contrast (with a 6 volt filament setting).
A second method of chromosome preparation utilized
0.4pg/ml cholchicine instead of 0.8|Ug/ml. Additionally, thq
' I
'cells were incubated for three hours instead of four hours. |
A third method was also formulated. In this procedure, |
I
colcemid ( 0 . 0 6pg/ml) was used instead of cholchicine, and
incubation with this agent was for only two hours. This
last method proved the best for karyotype analysis of cul
tured nephroblastoma cells. A comparison of these three !
procedures is illustrated in Figures I-3.
Soft Agar Techniques I
!
Soft agar plates were made according to a modifica- ;
d4 !
tion of the methods of Habel and Salzman.^ For these
plates, the base (nutrient) layer and the overlay (cell) :
components were prepared separately. ,
Base layers were made by mixing 20 ml of sterile |
fetal calf sera with 20 ml concentrated (2x) minimum essen- ;
Figure 1. Karyotype Analysis
Chromosome morphology with Method. 1
(Colchicine O.Spg/ml, 4 hr.
10,000 X mag.)
Figure 2. Karyotype Analysis
Chromosome morphology with Method 2
(Colchicine 0.4pg/ml, 3 hr.
10,000 X mag.)
Figure 3. Karyotype Analysis
Chromosome morphology with Method 3
(Colcemid 0.06|ug/ml, 2 hr.
10,000 X mag.)
34
35
Figure 1
Figure 2
Figure 3
« V
9
> * .
5 . 3 '
i
• %
1
I
I
V
-•C'
. . . . . . . . . . . . . . . . . 3 6 ' ;
tlal media (Microbiological Associates, San Diego) and 20
ml triplicate broth (GIBCO, Los Angeles). With these three
solutions added, the mixture was warmed to 46 ° C. A stock
j
solution of 0 .5/ agar (GIBCO, Los Angeles) was melted in !
boiling water, then cooled to 48° C. An aliquot of 80 ml ‘
of the water was then added to the above nutrient broth.
Base layers of approximately 3 mm in height were poured into
p p
either T-6 0 (Corning, 75 cm ) or T-3 0 (Corning, 25 cm ) ;
tissue culture flasks.
For the cell layers, two volumes of sterile agar
(0 .5/)J prepared as indicated above, were added to one
volume of suspended cells ( trypsonized) at 37° C. The cell-|
agar overlay was quickly mixed and injected into flasks with
pre-poured base layers (syringes and 29 gauge needles were
used for innoculations). The final cell layers measured
approximately 1 mm in height. Soft agar cultures were in- j
cubated at 37° C for periods of two weeks to one month. j
Positive plates were determined on the basis of multicellu-
lar colony formation. Examples of growth patterns from !
!
each cell type were photographed for future reference. |
1
Biochemical Methods
Tissue Extractions
Normal and malignant tissues were extracted accord-
26
ing to the methods of Allerton et al. In these proce-
37
dures, a solution of 0.2/ EDTA in CMP-PBS was used to re
move "loosely adhering" cell surface components. Cell cul
tures were extracted with a similar solution, according to
1D
the methods of Wise et al.
Colorimetric Assays
Several "classical" colorimetric assays have been
used to quantitatively analyze extracts and homogenates of
normal and tumorous tissues (as well as extracts and cells
from cultures). Included are the following:
1. Protein content was determined by a modification
ED
of the Lowry method according to Bailey. This
procedure was modified by decreasing the indicated
volumes for both samples and solutions. The assay
was sensitive to 10 pg of protein (with bovine
serum albumin as a reference).
2. Uronic acid was measured by the carbazole methods
e6
of Bitter and Muir. This system was modified
as in the protein determinations. The standard
used was glucuronic acid (Calbiochem, La Jolla,
California), and the sensitivity was approximately
0.5 dS of carbyhydrate.
3. DNA content was estimated through the use of the
diphenylamine reaction as described by modifica
tion of the Dische method according to Burton.
4. Hexose was estimated by the orcinol reaction ac-
38
57
cording to the methods of Winzler. This test
was sensitive to about 1 |og of glucosamine.
5. Hexosamine was determined by a modification of the
Elson-Morgan reaction (with para-dimethylamino-
benzaldehyde) according to Blix.^^ The sensitiv
ity range to glucosamine with these methods was
about 1 pg to 150 pg.
Enzymatic Analysis
A total of seven enzymatic activities was measured
in dialized homogenates and extracts of tumor tissues,
normal "control" tissues, and in preparations from cultured
cells.
1. Alkaline phosphatase activity was measured accord
ing
ing to the methods of Garen and Leventhal. The
substrate for this measurement was para-nitro-
phenolphosphate. This assay was determined at pH
8.0.
2. Acid phosphatase was determined as in the alkaline
phosphatase measurements, except that the reaction
was conducted at pH 5*0 instead of 8.0.
3 . Heat resistant alkaline phosphatase (Placental-
like phosphatase) was measured according to the
procedures described above for alkaline phospha
tase. In this case, though, the samples were
assayed after heat inactivation (at 5 6^ C for
60 minutes). _______ __________
39 I
4. GluCOse-6-phosphatase activity was determined ac-
pQ
cording to the methods of Pretlow et al.
5. Gamma glutamyl transpeptidase activity was deter- ;
mined through the use of an enzyme determination
kit purchased from Calbiochem (La Jolla, Califor
nia) . The results and standards were prepared
according to the recommendations of Calbiochem,
6. Protease activity was determined by both colori
metric and agar tests. The agar tests were based
on a fibrin degradation method of Schill and Schu- '
macher, which causes "clearing zones" around
wells with protease activity; the size of the !
zones is proportional to the concentration of |
protease. A second method of protease test was |
based on the ability of samples to cause dégrada- !
tion of bovine serum albumin. To measure the I
amount of degradation, the albumin (with the sam
ple) was precipitated with 10^ trichloroacetic '
acid. Controls consisted of samples before incu- i
bation (incubation was conducted at 37^ C for one !
hour). i
7. Hyaluronidase activity was determined by assays ,
6l
according to Kass and Seastone. The determina
tion of residual (non-digested) hyaluronic acid,
though, was determined by dialysis of the samples ;
4o
(to retain large molecules of substrate, but re
lease degraded fragments). The residual material
was measured by the carbozole method of Bitter and
56
Muir. Actually, the incubation and dialysis
steps were integrated into one step. That is, the
samples (with exogenous Grade I hyaluronic acid.
Sigma Corporation) were incubated under digestion
conditions within dialysis tubing. In this way,
degraded fragments would be generated then re
leased through the dialysis membranes at the same
time. Control samples consisted of samples (with
endogenous hyaluronic acid) without exogenous sub
strate, which were incubated in parallel with the
substrate containing samples.
Gel Chromatography
Molecular sieve chromatography was used to estimate
molecular weights of tumor associated compounds, as well as
to study binding interactions between hyaluronic acid and
serum proteins. For these experiments, either Sephacryl
S-200 or Sephadex G-200 gel beads (Pharmacia, Uppsala,
Sweden) were prepared according to the recommendations of
the manufacturer. The columns measured 2.5 centimeters in
diameter by 50 cm in length. Before an experimental run,
the columns were equilibrated with 0.5M NaCl or CMF-PBS at
pH 7.2; normally,the column run would utilize these same
41 I
solutions for the chromotography. The flow rate used was
0.5 ml per minute, with fractions collected at every 0.7
ml. The fractions were continuously monitored by a ;
Beckman-DB spectrophotometer using a 1 cm (50 pi) flow cell . |
Isoelectric Focusing !
The methods for focusing have been adapted from the
62
procedures of Wrigley. Specifically, the focusing column
was established in a G-15 Sephadex gel bead slurry (Phar
macia of Sweden) measuring 2.5 cm in diameter by 25 cm in
length. At the bottom of the column, a plug was fabricated
from 10^ aeryamide and this "stopper" was separated from the
G-15 by a 0.46 micron Millipore Filter. The ampholines usedj
for focusing were a mixture of 2.5 ml of LKB pH 2.5-4.0 '
plus 2.4 ml of LKB pH 3.0-9.0 (LKB, Washington, D.C.) di
luted to 50 ml in a solution of 10$ glycerol. In each run, '
a sample of 10 ml was added. Once the gel slurry with the j
ampholines and sample had been poured, a top plug was fab- !
ricated with acrylamide (as with the bottom plus, a 10$ gel
was made; the gel was polermized with the addition of !
85mg/$ ammonium persulfate and 0.5$ TEMED. The top buffer |
in this system consisted of 1$ ethanolamine at pH 11.0,
while the bottom buffer was 0.6 N sulfuric acid. The focus
ing was accomplished by applying a current of 1 5 -2 5 milli
volts to the column for approximately four hours (or until
".................. ’....'...................’..' "L !
the current remains stable). After focusing, the samples
were removed by pumping, and the pH of each sample was
determined. Ampholines were removed from the sample by
chromatography with Sephadex G-75 (Pharmacia, Uppsula,
Sweden).
Immunological Methods
Immunization of Rabbits
i
In this study, female New Zealand white rabbits '
j
were used to develop antisera to tumor tissues, cells, and '
extracts. All of the animals were three months of age at |
the time of initial inoculation. Injections (both primary |
and booster) consisted of 1.5 ml antigen containing samples
(with 1 5 -5 0 mg protein) with 1 .5 ml Freund’s Complete Adju-
vanct (Calbiochem, La Jolla, California). Each inocula- !
i
tion consisted of a simultaneous injection of emulsified j
antigen-adjuvant solution (0.5 ml aliquots) at each of six I
sites along the base of each animal's neck (subscapula re-
I
gion). Secondary or booster injections were administered
I
at intervals of 15-20 days after primary innoculation. The i
booster applications were continued for four or more injec- '
tions (depending on the antisera potency as observed by
Ouchterlony double diffusion analysis).
....................... 43 '
Anti sera Preparation
Within 10-15 days after boosting, the animals were
bled through cardiac puncture. For this procedure, 25 cc
Vacutainer multiple fill needle holders and 1.5 inch 20g |
needles were used (Beetin and Dickinson and Co., Phila
delphia, Pennsylvania). At each bleeding approximately 50 ;
cc of blood would be drawn. This would be allowed to clot
at 4° C for 24-48 hours ; after clotting the serum would be
separated by centrifugation at 8 0 0 g for 10 minutes. The !
serum (supernate fraction) would then be decanted and
frozen at 1 7° C. ;
!
Antisera Absorption |
To remove antibodies directed against normal human i
plasma and kidney antigens, the antisera prepared as indi- ^
cated were absorbed with combinations of: normal human sera, |
normal human plasma, homogenates or either normal human I
liver or normal kidney, homogenates of native or mouse- '
grown Wilms tumors, and tissue cultured nephroblastoma cells:
(homogenates or extracts). When the antisera were absorbed I
with human sera or plasma, the two solutions would be mixed i
at ratios of from 1 :1 to 1 (part antisera): 3 (parts ab- ■
sorbent sera or plasma). Absorption with tissues (either '
of normal or neoplastic origin) was accomplished by adding
one milliliter (ml) of tissue homogenate to each milliliter
of antiserum. With these latter samples, tissues would be
44
disrupted with teflon and glass homogenizers (with 1 ml of
CMF-PBS for each gram of tissue). The homogenate would
then be added directly to the antisera without prior cen-
I
trifugation; in this way, insoluble tissue components can ;
still act to absorb antibodies. Also, in these homogenates,
epsiIon-amino-caproic acid would be added to a concentra
tion of 0 .5^ to inhibit protease activity. When absorption
was accomplished with purified protein components (as
Spiro's fetuin), the absorbant would be added to a final
^ concentration of at least 1 mg/ml in the antisera.
In all of the above absorptions, the antisera-ab-
O i
sorbant mixtures would be gently mixed at 25 C for one ,
hour, then incubated at 4° C for 24-48 hours. They were I
then centrifuged at 10,000 x g for 10 minutes. The absorbecj
antisera would be decanted and used directly in the pres- '
ence of some soluble absorbent compounds or processed
further with a Protein-A column. The supernate fractions
of absorbed antisera (as described above) would often be
substantially diluted by the procedures. To concentrate :
these preparations, an affinity column was used to isolate i
the Immunoglobulin G fraction from the supernate antisera- |
absorbent mixture. The column used was prepared with
Protein-A/Sephadex (Pharmacia, Sweden). This staphylo
coccal Protein "A" affinity material binds to IgG from both :
human and rabbit sources. In this way, the absorbed anti
sera could be purified and concentrated.
......... 45" i
Immunodiffusion
Ouchterlony double diffusion methods were used to
examine antigen-antibody reactions in several experiments, j
The agar gels used in these studies were purchased from |
Hyland Laboratories (Costa Mesa, California). The gel wells
each held 10 microliters of solution and were separated
from each other by 4 millimeters (the well diameters were
2 millimeters each). These gels were made with 2$ Nobel
Agar, 7.5^ glycine, 1$ NaCl, and 0.1$ sodium azide (at pH
7.2). To develop bands, the wells were filled with the
appropriate antigen or antibody preparation, then incubated |
at room temperature (in a humidified box) for 48 hours. |
The gels were then deprotelnized in CMF-PBS at room tem
perature for 24 hours. Gels were stained in either Buffalo
Black (0 .25$) or in Commassi Brilliant Blue (0.4$) for two i
hours. Destalning was accomplished by soaking the stained
gels in a solution of 7$ glacial acetic acid and 5^ metha
nol. Once destained, the gels were dried in an oven at
j
4 5° C for six hours. |
I
Immunoelectrophore si s :
Immunoelectrophoresis kits were purchased from |
Millipore-Worthington Corporation (Bedford, Mass.). These
kits included prepoured agar plates, electrophoresis buffer
(Barbital buffered to pH 8.6), deproteinizing buffer, and a
Buffalo Black staining solution. This system was utilized
; .. ' 46
las recommended, with the exception that Commassi Brilliant
Blue was substituted for the Buffalo Black stain. The
I sample wells of each well held approximately 2 microliters,
'while the antisera troughs held about 50 microliters. The
isamples were electrophoresed at 100 volts for 35 minutes;
antisera would then be added, and the gels were incubated
:at room temperature for 48 hours (in a humidified box). The
deproteinization and staining procedures were the same as |
i I
those used for the Ouchterlony diffusion gels. I
' Absorbent Materials
Biosamples for these studies have been obtained
; through the courtesy of Dr. Powars (of County/USC Medical
Center, Department of Pediatrics and Hematology). Included
: in these samples have been Wilms tumor tissues, sera, and
Brine samples from nephroblastoma patients. Also, control
,samples have been provided by Dr. Powars (normal human sera
'and normal human plasma samples, as well as normal human
ikidney and liver tissues). An additional human liver
sample was obtained from Dr. Alan Murray (of the UC-Irvine
School of Medicine, Department of Pediatrics).
! CHAPTER III
: RESULTS
Nude Mouse Tumor Induction
Attempts were made on three occasions to grow human
Wilms cells (or tissues) in nude mice. The first two of
these which utilized in vitro cell cultures failed to cause
tumor growth. In the first experiment, TuWi cells (100,000
per animal) were injected into each of four mice. After
four months, no abnormal growths were found and the inocu
lated animals were sacrificed. A second attempt to develop
tumors was made with a human Wilms tumor primary culture.
For this study, Wilms tissue fragments were cultured as
indicated in the Materials and Methods section. After one
week in culture, the primary cells (containing both epi
thelial and fibroblast elements) were injected into five
mice (at 100,000 cells per animal). This second group of
animals was monitored for over three months. Since no
tumors formed, these animals were also sacrificed.
A third attempt to grow Wilms cells (in this case
from tissue) was made using fresh Wilms tumor fragments.
This experiment, unlike the previous ones, was successful
in forming Wilms-like tumors. The original human tumor
47
.............. ' 48°
used for the project is shown at 5/6 original size in
i
Figures 4 and 5. As described previously, this is the
second tumorous kidney from a Wilms patient. Figure 4 |
shows this tumor with the outer surface exposed; note that |
I
a pseudocapsule covers the entire mass. Internally, the |
tumor was nearly homogenous with little indication of
necrosis (Figure 5)• Histological sections of 6 microns !
thickness were prepared and stained with Hematoxylin and
I
Eosin by the Department of Pathology at the County/USC I
iHospital of Los Angeles. Examples of these paraffin sec- i
I
I tions are illustrated in Figures 6 and 7 - Under low power i
I
■magnification (40x), the tissue shows elements typical of ;
human nephroblastoma. Included is a mixture of swirls of I
: interstitial (fibroblastic) cells around "islands" of
, nephtogenic cells. As shown in the Illustration (Figure 6),
: there is an approximately equal distribution of the two
! major tissue populations.
The interstitial tissues in this section are the
lighter stained, spindle shaped cells. Occasionally, vacu
oles are present in these tumors. One such formation is
shown at the bottom of Figure 6 in a fibroblastic region.
!
Also, the presence of blood vessels in this tissue is in- g
dicative of the degree of tumor vascularization. |
I
Metanephric cells are shown as dark staining island^
of rounded cells. Of particular interest are the tubule |
Figure 4. Wilms tumor from patient B.C.
(Used to develop nude mouse grown
tumors)
View of the entire tumor (showing
the pseudocapsule. 0.8 x mag.)
Figure 5- Wilms tumor from patient E.G.
View of a slice through the
midsection
(0.8 X mag,)
4g
50
Figure 4
Figure 5
4
Figure 6. Histological paraffin section of the
nephroblastoma(from patient E.G.,
40 X mag.; Hematoxylin and Eosin)
Figure 7- Histological paraffin section of the
nephroblastoma(from patient E.G.,
2 5 0 X mag.; Hemotoxylin and Eosin)
51
Figure 6
Figure 7
m m :
5 2
%; H'
' f S U
m m M B s
iS
4? 6/,
i * l ? 7 LA
« W ' & % % : ?
%
W9-V:
53
formations found primarily in these clusters of epithelial-
like (nephrogenic) cells. Such abortive tubules are essen
tial to a clinical diagnosis of Wilms tumor. Also in the
"islands" is evidence of vascularization.
A high power magnification of a paraffin section of
this tumor is shown in Figure 7. This illustration depicts
an abortive tubule surrounded by morphologically undifferen
tiated metamephric cells. These epithelial-like cells have
large distinct nuclei, and are deeply stained by hematoxy-
'lin and eosin.
Nude Mouse Grown Tumors
Tissue fragments of the child's tumor were prepared
for inoculation into the nude mice as indicated in the
Materials and Methods section. After injection of 0.5 ml
of minced tissue into each animal, a period of six months
was required for incubation. During the first five months,
no noticeable growth occurred. In fact, an initial "inoc
ulation bulge" in the skin of the animals decreased in size i
during the first few months. During the sixth month, how- |
ever, rapid tumor growth occurred in two of the three |
treated animals. In this study, one animal received
adipose-like tissue from the human tumor. This animal
failed to develop a tumor. The other two animals were
given injections of "white mushy" tissue from the outer
regions of the child's tumor. In each of these latter ani-
54 I
mais, large tumors were found by the end of the sixth monthJ
Figures 8 and 9 illustrate the growths in one animal
(Figure 8 ) as well as the actual tumor removed from this
animal (Figure 9). Both of the primary nude-grown tumors i
were sub-cultured into additional mice. Only small tissue
slices were kept for pathological examination, A complete
list of the tumors and injected animals is shown in Table
1. I
Paraffin sections (of 6 microns, stained with Hema
toxylin and Eosin) were prepared for both of the primary
nude tumors. Figures l4 and 15 show microscopic examina
tions of the two initial tumors. The growth developed in i
mouse # 1 9 4 is shown in these figures. Under low power mag- |
nification (Figure 10), the tumor tissues have many of the ,
i
characteristics of native Wilms tumor. Included are the 1
!
presence of: interstitial and nephrogenic tissues, tubular I
elements, vascular elements, and vacuoles. Also, the over- j
I
all organization of the tumor is characteristic of nephro
blastoma. !
I
Unlike the child's tumor, this particular mouse- ;
. !
grown tumor has an enrichment of the nephrogenic (epithe
lial-like) regions; in this section these epithelial-like
cells comprise 80$ of the tumor, rather than the original
5 0$. Under higher magnification (Figure 11), each of the
two tissue types is clearly visible, with many tubular ele- !
Figure 8. A nude mouse with a subcutaneous
nephroblastoma
(0.7 X mag.)
Figure 9- The tumor removed from the mouse
in Figure 8.
(0.7 X mag.)
55
5 6
iFigure 8
Figure 9
Table 1
A List of Inoculated Nude Mice, including Dates of
Injection, Tumor Status, and Termination Dates
57
Animal # Inj. Tumor/Use Date Term
Primary
195
6-21 No tumor
12-8-77
194 6-21 Trans. 238-9, 242-3
12-8-77
195
6-21 Trans. 236-7, 240-l 12-8-77
1st Passage (Second animal group)
256 12-8 Karyotyping 4-6-78
257
I I
T ran s. ^08-312 3-28-78
258
I I
Trans. 265-268 3-10-78
259
I I
No tumor poss. MHV 4-20-78
240
I I
Frozen Dr. Hashed 4-13-78
241
I I
Biochemistry/Histo. 4-20-78
242
I I
Biochemi st ry/Hi st0. 4-20-78
242
I I
Trans. ^0^-^OJ 3-28-78
2nd Passage (Third animal group)
265 3-10 No tumor: terminated 4-14-78
MHV®*
266
I I
No tumor: term. MHV 3-3-78
267
I I
Biochemistry 6-16-78
268
I I
No tumor; term. MHV
3-3-78
505
3-28 Histology 7-19-78
504
I I
No tumor : term. MHV 4-26-78
505
I I
No tumor: term. MHV 4-26-78
506
I I
No tumor: term. MHV 4-26-78
507
I I
No tumor: term. MHV 6-8-78
508
I I
No tumor: term. MHV 5-12-78
509
I I
No tumor: term. MHV 5-12-78
310
I I
No tumor: term. MHV 6-30-78
311
I I
No tumor: term. MHV 5-2-78
312
I I
No tumor: term. MHV
3-2-78
Mouse Hepatitis Virus.
Figure 10. Histological paraffin section of
nude mouse 194 primary tumor
(40 X mag.; Hematoxylin and Eosin)
Figure 11. Histological paraffin section of
nude mouse 194 primary tumor
(100 X mag.; Hemotoxylin and Eosin)
58 '
5 9
jPigure
Figure 11
« I
& : % - a
60
ment8 being found In the nephrogenic tissues. Throughout
this field, there is also evidence of blood cell components.;
These latter cells are shown as small dark spheres dis- |
persed through the nephrogenic tissues. ^
Figures 12 and I3 show another field from the tumor
of mouse # 1 9 5. In these illustrations, an encapsulated cyst
is shown. Under higher magnification (lOOx), the cells sur
rounding the cyst appear to contain a high proportion of
the presumptive blood cells. It may be that this mouse has
initiated a limited immune response against soluble tumor
substances. These athymic mice do not have thymus-dependent
cell-mediated immune responses; they do have, however, an !
i
intact antibody-mediated system against soluble components. I
It is also interesting to note that the "capsule" is com
posed of fibroblastic cells, which could be of mouse origin.
This would be possible, since the entire tumor in this ani
mal was encapsulated. Histological preparations of the
second nude primary growth (mouse #195) are shown in Figures
l4 and 15. The low power (40x) photograph of this tissue
indicates a tumor that is quite diverse from the other nude |
mouse primary growth. In this second case, all of the I
characteristic elements of Wilms tumor are also present
(including interstitial cells, nephregenic cel 1s, tubules,
vascular elements, and vacuoles). The relative proportions
of the tissues and differentiated structures are quite dif- ;
Figure 12. Cyst formation within the nude mouse
1 9 4 primary tumor
(4o X mag.; Hemotoxylin and Eosin)
Figure 13. Cyst formation within the nude mouse
1 9 4 primary tumor
(100 X mag.; Hemotoxylin and Eosin)
61
im-rnMs
m m # » :
M l ®
Figure 12v^SK|4 ^ *
‘ XV»
a g : '
m
w
Figure 13
i
5*A>«’ b
Figure l4. Histological paraffin section of
the nude mouse 195 primary tumor
(40 X mag.; Hemotoxylln and Eosln)
Figure 1 5. Histological paraffin section of
the nude mouse 1 95 primary tumor
(1 0 0 X mag.; Hemotoxylln and Eosln)
63
Figure
Figure 15
F
a m i #
t oi
: .......................... " ■ ......."... 65
i
:ferent j though. In the second tumor (animal ^195)j the
sections reflect about equal tissue type distribution of
5 0 ^ each (like the native child's tumorbut unlike the
first nude tumor with QOfo eplthe 11 al-like cells). Also,
Figure l4 shows considerably more vascularization than In
either the other nude tumor or the native human tissue.
Higher magnification of this section shows a pattern of
swirling Interstitial cells surrounding nephrogenic ' ’is
land s''j this Is typical of human nephroblastoma. Like the
other nude mouse primary tumor, this second growth has pre-|
:sumptlve blood cells throughout the tissue section. |
As described previously, the primary tumors were !
I
; serially passed to additional animals. Figures I6 and 17 I
Illustrate a first passage growth (in animal # 2 3 8 of the
: second animal group) from the original tumor of mouse #194.
;Thls subcultured animal grown tumor maintained the two major
^tissue types of the first mouse grown tumor (as well as of
the native human tumor). Specifically, the nephrogenic re
gions of this growth accounted for about 750 of the tumor
mass. This Is In contrast to the native human tumor (which
had almost 5 0 ^ ) ; It Is comparable, though, to the first I
nude mouse tumor (194) which had 8 0 ^ eplthellal-like cells, i
Figure 17 shows the major change observed between the first |
and second groups of animals. That Is, the amount of tubule,
formation Is substantially decreased In the second animal I
Figure 1 6. Histological paraffin section of the
first passage tumor of mouse 238
(from the primary growth of mouse 194)
(4o X mag.j Hemotoxylln and Eosln)
Figure 1 7. Histological paraffin section of the
mouse 238
(100 X mag.; Hemotoxylln and Eosln)
66
I
iFigure
Figure 17
68 I
: tumor. Still, some of the differentiated structures are I
still present. Also, like the previous nude mouse tumors,
blood cells were found throughout the tumor section. |
Figures l8 and 19 illustrate a second generation !
nude mouse tumor from the other nude mouse primary (1 9 5) :
growth. In this nephroblastoma, the secondary animal-grown
tumor resembles its "parental” tissue (from mouse # 1 9 5)
more than it does either the original child's tumor or the
other primary growth. Like both the native tumor and the
mouse # 1 9 5 tumor, the interstitial cells are found as !
"swirling sheets" surrounding islands of nephrogenic cells. |
, However, unlike the original human tumor (but like the pri- I
; j
mary mouse tumor #195)^ there are very few abortive tubule |
I
formations. This particular secondary tumor also shows an '
enrichment of large undifferentiated nephrogenic regions. '
Figure 19 illustrates this phenomenon. Accompanying the !
increase in epithelial cells was a distinct local reduction ^
of interstitial cells. Instead of sheets of cells, many of
the fibroblastic regions formed only slender "islands" sur
rounded by a "sea" of undifferentiated nephrogenic cells. |
A second passage of tumors (to the third group of |
mice) was conducted with tumors from each of the two paren- :
I
tal primary tumors. However, the entire nude mouse facility'''
housing these animals became plagued with mouse hepatitis
virus. Most of the second passage animals were terminated ,
Figure l8 . Histological paraffin section of
the first passage tumor of mouse 240
(from the primary tumor of mouse 1 9 5)
(40 X mag.; Hemotoxylln and Eosin)
Figure 19. Histological paraffin section of
mouse 240
(100 X mag.; Hemotoxylln and Eosin)
69
iFigure
I # # #
V '
■>.»-
rf V - i
. e : ' y / , ’ î
X
F P i
> -a, ^ V. V.' _ --
%
Ù -
Figure 19$ *
« s ^ S S
71
by this viral disease. The specific mortality figures are
listed in Table 2. Only two tumors were salvaged from this
third animal group. One of these (mouse # 2 6 7) was used for !
biochemical and immunological determinations. The other |
tumor from mouse # 3 0 3 was used for histological examina
tion. Photographs of paraffin sections of this second pas- 1
sage tumor are shown in Figures 20 and 21. From these 11- '
lustrations, it is clear that both of the major tissue |
types are still present. However, there has been a dis- '
tinct shift toward domination by the epithelial-like tis
sues. Figure 20 shows that the nephrogenic regions are now
massive undifferentiated sheets. The interstitial areas
are also present, but these regions are quite diminished.
As in all the previous nude mouse-grown tumors, this growth
contained numerous blood cells. Unlike the human malig
nancy and the parental mouse tumor (#194), though, this I
second passage growth contained no tubule formations. A |
consistent pattern is then found in the gradual depletion
of these primitive differentiated kidney structures through^
out the completed serial passages. There is also a trend |
toward selective enrichment of epitheliai-like cells. '
In vitro Culturing
Early attempts to maintain Wilms tumor cultures
from surgical specimens were essentially unsuccessful. In ,
Figure 20. Histological paraffin section of the sec
ond passage tumor 3 0 3 (from the secondary
tumor of mouse 24-3; the parental primary
tumor was from mouse 194)
(4o X mag.; Hemotoxylln and Eosin)
Figure 21. Histological paraffin section of mouse 303
(100 X mag.; Hemotoxylln and Eosin)
72
Figure 20
Figure 21
* 1
m m
* ^ T* — 1^ * tm IlM
...............^ 74 '
three such experiments, cells remained alive for six days
or less. Modifications in methodology were therefore made j
to improve establishment of cultures.
Trypsinizations often resulted in the formation of
an insoluble gelatinous material in the tissue extracts.
Such a gel-like substance was also found with Ficoll-paque
(Pharmacia of Sweden) separation. In an attempt to dis
rupt this material, hyaluronidase (ovine testicular source
from Sigma Corp.) and collagenase (from clostridium his-
tolyticum, Worthington Corp.) were added at 0.2^ each in
serum-free MEM media. This combination of enzymes dispersed
the tissues as well as the gel, but the resulting isolated |
cells proved to be non-viable.
Attempts have also been made to culture cells di
rectly from tissue fragments. That is, small tissue pieces I
(less than 1 millimeter in diameter) were placed into cul- :
ture flasks with 15-20^ sera (in Hams F-12 or MEM media). ;
After several days of undisturbed incubation, many of the |
fragments had cells attaching to the culture vessels. From I
these experiments, two conclusions were made. First, the
cells had to be left undisturbed for several days for at- I
tachment to occur. These cultures then exhibited a sub-
I
stantial lag period between seeding and outgrowth. Sec- i
ondly, enzymatic disruptions were generally detrimental to
i 75
ithe primary cultures. Therefore, these methods were
'avoided for primary cells explanation.
The final procedure adopted to grow primary cul
tures from human or mouse-grown Wilms tissues involved two
critical changes in the outlined protocol described previ
ously. Specifically, the initial "feeding" methods were
altered. Previously, the media of primary cultures would
be changed every 24 hours. The used growth media (with any
loose cells) would then be discarded. However, by replat
ing the non-attached cells, it was found that many of these
loose cells were still viable. A preferable approach was
!to change the media for primary cultures only after 48 or
72 hours. In addition, the loose cells from used media
were isolated by centrifugation; they were then resuspended
;in growth médias, and added back to the original culture
vessels. Such a procedure was repeated for several weeks.
‘ This alteration was also adapted for cultivation of tissue
pieces. That is, instead of discarding unattached pieces,
these fragments were added back to the culture vessels. By
this method, a far greater number of cells would attach and
grow in primary culture. A second change dealt with the
media used. Instead of MEM with 10-150 fetal calf sera.
Hams E-12 media with 150 rabbit sera was used. This latter
growth medium provided a better initial attachment and
growth of primary cells.
! 761
!
The outlined protocol was utilized to establish i
cells from the original human nephroblastoma (of patient
E.G.) as well as cultures from nude mouse-grown tumors. In!
‘ addition, an attempt (using these procedures) was made to |
establish growth from a peritoneal fluid sample from a ;
!
terminal patient (J.M.). A total of six cell strains have
been developed to date. Of these, four cell populations
are alive and growing. In each of the cultures, the iso
lated cells have maintained their original morphological
appearance (from primary explantation to the time of this
.writing). A description of the cell types is given below. j
jAlso, photographic illustrations of the cells appear in
Figures 22-29.
1. EPL Cells: These epitheliai-like cells were derived
from a nephroblastoma patient (E.G.).
These were separated from fibroblastic
cells of the primary culture by differ
ential trypsinization (as indicated in
the Methods section). The cells are
cuboidal in shape and form rounded colo
nies; they also tend to "pile-up" on one |
another to depths of 4-5 cells. Many of
the cells are multi-nucleated and most
have 2-3 nucleoli. An example of a small
EPL colony is shown in Figure 22; more
Figure 22. An early colony growth of EPL cells
(epithelial-like cells from the
patient E.C.)
Demonstrates cell morphology, multi
nucléation, colony morphology, and.
early "piling up" of cells.
(100 X mag.; Acetic Orcein)
Figure 23. A confluent culture of EPL cells
Demonstrates the "piling up" of
cells at confluency. (Deeply
stained regions are 3-5 cells
thick.)
(100 X mag.j Acetic Orecin)
77
7 8
-
Figure 22
Figure 23
u
" ' 79"
confluent cells with the characteristic
"piling-up" phenomenon are illustrated
in Figure 2 3.
2. NHF Cells: These are normal human fibroblasts ob
tained from routine amniocentesis
through the generosity of Dr. Barbara
Crandall (of UCLA, Los Angeles). They
served as a control cell population for
comparison with the nephroblastoma cells.
An example of a large colony from a pri
mary culture is shown in Figure 24.
Note the distinct alignment of cells
within the colony, as well as the spindle
shape of the individual cells.
3. NUF Cells: These are fibroblastic cells from a nude
mouse-grown tumor. As shown in Figure
25j a colony made by primary cultures of
these cells is quite similar to that of
the amniotic fibroblasts. The NHF cells
are also spindle shaped, and tend to
align with each other. The "gelatinous
coat" on the NUF colony developed only
after an extensive time in primary cul
ture. These cells were isolated by dif
ferential trypsinization. An example of
Figure 24. A primary culture of normal human
embryonic fibroblasts.
(from normal human amniotic cells,
courtesy of Dr. Barbara Crandall
of U.C.L.A., Los Angeles)
Demonstrates unique colony formation
(100 X mag.; Acetic Orcein)
Figure 25. A primary culture of NUF, nude-mouse
grown nephroblastoma fibroblasts.
(Note the similarity between this colony
and the amniotic fibroblasts in Fig. 24.)
(100 X mag.; Acetic Orcein)
Figure 26. NUF fibroblasts after three passages.
(The cells at this stage show typical
fibroblastic cell morphology; however,
no alignment of cells is found in
colonies of these cells --in contrast
to the same cells in the primary cul
tures, Fig. 26).
(100 X mag.; Acetic Orcein)
80
8l
4
«sure 24
w i m à .
Figure 25
Figure 26 '
82l
NUF cells that have been sub-cultured
appears in Figure 26. From this illus- |
tration, the spindle shape of the cells j
is evident. The NUF cells survived only |
four sub-culturings before they died. !
i
Some biochemical and immunological stud-j
les were completed on these cells, |
though, before they expired. |
t
4. NEF Cells: NEP cultures are nude mouse tumor-derived
epithelial cells. They were isolated by I
I
the same method as the EPL cell popula
tion. Morphologically, these are cu
boidal epithelial cells. They are
slightly irregular in shape, as compared
to the EPL cultures. Also, only a slight)
"piling-up" of the NEP cells is found at
confluency. The cell morphology, colony
formation, and "piling-up" are shown in
Figure 2 7.
5 . MIX Cells: This cell population was originally a
mixture of epithelial and fibroblastic
elements. They were derived from the
nephroblastoma patient (E.C.). After
ten passages, the cells were epithelial-
like in nature, and were indistinguish-
Figure 27. A nearly confluent culture of NEP cells
(an epithelial-like population from a
nude mouse tumor). Demonstrates typ
ically epithelial cells (with some
"piling up").
(100 X mag.; Acetic Orcein)
Figure 28. An early colony of MIX cells; Demon
strates epithelial-like cultures.
(100 X mag.; Acetic Orcein)
Figure 2 9. A nearly confluent culture of TuWi
cells: Demonstrates cell and colony
morphology.
(100 X mag.; Acetic Orcein)
83
841
(Figure
V ' :
i'-pw
% - #
/'Vli
/
Figure 28
Ù^At
1%^,
if.
>
u f t >
\
/
X
ü
Figure 29•
#
... . ....' '851
able (morphologically) from the EPL
cells from the same source. An example I
of these cells and the colonies they |
form is shown in Figure 28.
6. TuWi Cells: A commercially available human nephro
blastoma cell line. These cells appear
fibroblastic when sparsely seeded, but
are epithelial-like near confluency.
Originally, these cells were reported to
be fibroblasts. An example of this line
is given in Figure 29. (They are shown
here with our cultured nephroblastoma
cells [from human and mouse sources].)
7. FBL Cells: These were the fibroblastic counterparts
cultured from the patient (B.C.). Mor
phologically, they were quite similar to
the NUF (and NHF) cells. Unfortunately,
after a slow population doubling, the
cells ceased to divide; they were
Bluffed from the culture dishes and
failed to reattach. A photographic ex
ample of these cells is not available.
Media and Sera Tests
A series of culture media and sera tests was per
formed to determine optimal growth conditions for primary
: 86 I
tumor expiants (as well as for established cultures). In
1
these experiments, primary explants and TuWi cells were ’
used to evaluate the growth promoting capabilities of vari- ,
ou8 nutrient solutions. In all, four commercially available
media were compared. These were: Minimum Essential Media *
(with Earles and Hanks salts). Hams F-10, Hams F-12, and I
RPMI-1620.
With the primary explants, the Hams F-12 provided |
the best plating efficiency and growth, closely followed by
Hams F-10. It should be noted, however, that all of the !
tested media proved adequate for attachment and initial
outgrowth of primary tumor cells. j
Tests of several sera were also conducted. The
choice of sera was especially important for two reasons. ;
First, the serum that provided optimal primary growth was |
I
needed to establish cell cultures. Second, fetal calf |
serum (which is used for most iqi vitro work) is inadequate I
i
for some immunological studies of Wilms tumor. That is, |
calf sera contain the fetal glycoprotein fetuin which is |
i
immunologically cross reactive with a Wilms tumor-associ- j
ated antigen (Wise et al.).^^ As a result, cells to be
examined for production of Wilms tumor cell antigens should,
be cultured in media lacking fetuin (in this case, with i
sera other than fetal calf). I
Five sera were examined for growth promotion. They I
87
were : fetal calf, human, horse, rabbit, and chicken sera.
Each of the tested sera was found to support the attachment
and growth of explants and TuWi cells. Substantial dif-
1
ferences were noted, however, in the population doubling
times with these sera. Figure 30 shows the response of
four cultures to five sera (in MEM). The shaded regions of
the illustrations indicate the seeding density, while the
clear areas show increases in cell number after 48 hours of
incubation. The growth profiles of the tested cells reveal
marked and subtle influences of serum on cellular growth.
The data of Figure 30 indicate, for example, a unique re
sponse to different sera for the various cells. This tends
to argue against cellular contamination among the cell
lines utilized. Rabbit serum had an additional advantage,
: though, in that this serum contained no demonstrable fetuin'
I like protein. In fact, an immunological examination of the
; above sera by the Ouchterlony method showed that only fetal
'calf serum contained fetuin (as measured with anti-fetuin
antisera). For these reasons, rabbit serum was used in the !
establishment of primary cultures from nude mouse-grown |
i
tumors, as well as in the growth of cells in determinations |
of surface antigen presence.
Growth Characteristics of Cell Cultures
Population doubling times and growth curves are |
often considered important criteria in the characterization |
Figure 30. A Growth Response Curve;
Tests of Various Sera
(Seeding populations are indicated by
shaded regions ; clear areas demonstrate
additional cells after an incubation
period of 48 hours. All seras were
tested at 10^ (v/v) in MEM with Hanks
and Earles salts.)
Legend: Ca: Petal Calf Serum
Hu: Human Serum (adult
* Ho: Horse Serum (adult
Ra: Rabbit Serum (adult)
Ch: Chicken Serum (adult)
-X"Shaded area represents seeding, while
clear bar regions demonstrate addi
tional growth after 48 hours.
88
7
6
CD
t
O 5
X
4
CD
_J
_J
3
UJ
o
2
o
EPL (13)
i
i
TuWi (20)
Ca Hu Ho Ra Ch Ca Hu Ho Ra Ch
SERA TYPE
X ^
< 03
ÜJ 2
o
O
NEP (II-3) Mix (10)
1
Ca Hu Ho Ra Ch Ca Hu Ho Ra Ch
SERA TYPE
: 90
'Of newly established in_ vitro cell cultures. These param
eters are especially important in studying specific changes
that cells may have undergone. For example, transformed
and malignant cells generally grow rapidly gjl v^’ tro, while
"normal” cells grow poorly (and slowly) under the same con
ditions. For this reason, these tests have been conducted
with our human and nude mouse-grown Wilms tumor cells.
To investigate the growth characteristics of our
nephroblastoma cells (as well as TuWi cells), eighteen
flasks of each cell type (with six groups of triplicate
vessels) were prepared. With most cultures, approximately
.one million trypsinized cells were seeded into each culture
flask; an exception was with the NUF line (nude mouse-grown
fibroblasts), as only 2 5OO cells were added to each vessel.
: In every case, the cells were grown in MEM with 10 percent
fetal calf serum (with incubation under the usual condi
tions) . All of the cultures were fed only once, forty-
eight hours after inoculation.
The results of this study are illustrated in Fig
ures 3 1-3 5. In the tested cultures, a relatively high
growth potential was found with all of the epithelial-like
cells. The single fibroblastic strain, though, had a poor
growth response, with an extended population doubling time.
A specific example of one of the growth curves is shown in
Figure 3I. This graph of the EPL cell response indicates
Figure 31. Growth Curve; EPL cells
(Seeding population is indicated as
time zero. Standard error indicated
by perpendicular bars. Cells were
fed at 48 hours post-seeding. Cell
number determined by Coulter Counter
and hemocytometer techniques.)
Figure 32. Growth Curve: MIX cells*
(*Note the differences between this
graph and that of figure 3 1--both
populations were derived from the
same tissue, and now have the same
morphology.)
91
92
CD
I
o
cn
o
EPL ( 13 )
18
16
14
12
10
6
4
2
O
10 20 30 40 50 60 70 80 90 100 110 120
CD
I
o
o
Mix (10)
18
16
14
12
10
8
6
4
2
O
10 20 30 40 50 60 70 80 90 100 MO 120
HOURS POST-IN MOGUL ATI ON
: . . . . . . . . . . . . . . . . . . . . . . . ' 9 3 I
several growth-related aspects of these cultures. Included!
are: a brief lag period (approximately 15 hours)^ a short
population doubling time (approximately l8 hours)^ and a
rapid depletion of growth medium (as indicated by a de- '
;creased number of cells at 120 hours). ;
i
Another cell population is shown in Figure 32. In
this graph j the MIX cell growth profile is indicated. With
these cells,, rapid growth is also found. Specifically^ the
cells show a short lag period (approximately 15 hours)^ a |
population doubling time of 26 hours^ and continued growth !
throughout the entire 120-hour test period. This pattern
is similar to that of the EPL cells (from the same tissue
.source)^ except that the MIX cells have a slightly lower j
growth rate.
Epithelial-like cells from a nude mouse-grown tumor |
were also examined. The results of this test appear in Fig-j
!
ure 33* The indicated data show a pattern distinct from thej
previous two cell strains. In the case of the nude mouse- j
grown cellS; a short lag period is found (approximately 15 I
hours)j with a short population doubling time of about 28 |
hours. Unlike the other two cultures^ though,, the mouse- i
passaged cells appear to reach a plateau (or stationary |
phase) at 120 hours post-innoculation. i
The commercial TuWi cell line was also investigated.!
The results of tests with these cultures are shown in Figure|
Figure 13. Growth Curve: NEP cells
Figure 3 4. Growth Curve: TuWi cells
94
95
NEP ( II-3 )
14
CD
I
o
X
_J
_J
LlI
o
10 20 30 4 0 50 60 70 80 90 100 110 120
HOURS POST-INNOCULATION
CD
I
O
X
CO
LlI
CD
TuWi (20)
16
14
12
10
8
6
4
2
0
10 20 30 40 50 60 70 80 90 100 110 120
HOURS POST-INNOCULATION
;...... ’ ’ ..... ' ' ’ ........ 96
3^. A striking difference from the first three profiles
' I
was noted with the lag phase. With the TuWl c e l l s a lag !
I I
period of nearly 72 h ours was found (as compared to approxi-|
; j
'mately 15 hours for the other^ epithelial-likej, cells) . j
; StillJ the population doubling times (which are determined |
during the log phase of growth) were similar to those of I
: the NEP cells (with 28 hours) and the MIX cells (with 26
hours).
A final growth study was made with the NÜP fibro-
;blastic cell population from a nude mouse-grown tumor.
These cells exhibited an unusually slow growth in culture.
The actual results are shown in Figure 35- As indicated in
the figure J , only two data points were determined. This was
due to the availability of only a few cultures of these
cells. Because of the lack of information in the figure
(35), it is impossible to determine the "lag phase" for
these cells. It is only possible to extrapolate an approx
imate population doubling time of I50 hours (assuming that
the entire 120 hours is not a lag period).
From the above results^ two major groups of cells
are found on the basis of growth characteristics. The
first group contains the epithelial-like (EPL^ NEPj, TuWi)
cells of nephroblastoma origin. In culture^ these cells
have characteristics of malignant or transformed popula
tions. Howeverj the second group contains only the NUF
Figure 35* Growth Curve: NUF cells*
(*Note that the scale for
cell number Is altered
from the previous graphs.)
97
98
I
O
in
UJ
o
NUF ( II- 3 ) 10
8
6
2
0
10 20 30 40 50 60 70 80 90 100 110 120
HOURS POST-INNOCULATION
Figure 35
! 99
fibroblastic cells. This latter group has a growth profile
of "normal" (or non-transformed) human cells. This study^
then,, tends to support the hypothesis that the epithelial-
like cultures (from patient E.G.) are the malignant (or at
least the more malignant) cells.
Karyotyping Results
Analyses of chromosome numbers and patterns have
been made from nude mouse-grown tumors, as well as from
each of the established cell strains. Methodology for this
study has been described in the Materials and Methods sec
tion. The results of these experiments are shown in Fig
ures 3 6-4 2.
A single karyogram was completed on tissue frag
ments from a mouse-passaged tumor (#2 3 6). This profile is
illustrated in Figure 3^* (The actual preparation was com
pleted at the County/ÜSC Hospital of Los Angeles.) There
are two important facets of the pattern shown. One is that
the tumor chromosomes are of the human type with metacen-
tric form (unlike the mouse chromosomes which are primarily
telocentric). Second, two abnormal chromosomes were found,
as a dicentric with a smaller fragment. In all, 44 chromo
somes were isolated and categorized in internationally es
tablished groups. Some of these groups had appropriate
chromosome numbers ; exceptions were group "A" (missing one).
Figure 36. Karyogram of nude mouse grown tumor
(mouse 236 -- a first passage growth
from partental tumor 195). Demon
strates "human type" with an abnormal
dicentric and fragment chromosomes.
100
101
ill! illAl4»AAé«»
« • * A # A ■■ *
■4
1
Figure 36
102
group "E" (missing a #l8), group "E" (missing a #20), and
group "G" (missing the "y"). It is likely that the above |
profile is incomplete, though, as individual chromosomes
are often lost in preparation of the spreads. Such a loss '
is especially applicable to this case, as the illustrated |
karyogram was made with the only clear spread from the
mouse tumor.
Tissue-cultured cells were also used for chromosome
analysis. In each of the cultures to be karyotyped, the
cells were stimulated to an active (exponential) growth
phase by trypsinization twenty-four hours before initiation|
of the tests. After such stimulation, the cells were
treated with a mi tosi s-inhibiting agent, then processed as |
described previously. The resulting counts from these '
studies of vitro cultures are shown in Figures 37-42. i
The first of these illustrations (Figure 37) depicts the :
pattern for the TuWi cells (commercial line) at greater than
twenty passages in vitro. The graph shows a bimodal chromo
some distribution with maxima at 54-60 and 6 1 - 6 2 chromosomesj.
This is in contrast to the values of 55-60 reported for the |
TuWi cells by the American Type Culture Collection. Pos- i
sible explanations for the observed differences include a ,
cellular contamination of our cells (by Hela?), or a coex
istence of two sub-populations of the TuWi cells in our
laboratory.
Figure 37. Karyotype profile: TuWi cells
Demonstrates bimodal distribution
(with peaks near 54-56 and 61-62
chromosomes) .
Figure 3 8. Karyotype profile: MIX cells
(from patient E.C.)
Demonstrates maximum cell bursts
(at 4 9 - 5 6 chromosomes).
103
104
TuWi (20)
CO
q:
-7
ÛÛ
ÜJ
o
40 4 4 52 48 56 60 64 68 OVER
CHROMOSOME NUMBER 70
Figure 37
Mix (10)
CO
h-
co
□c
ZD
-7
ÛÛ
m ÜJ
o
40 48 52 60 44 56 68 OVER 64
CHROMOSOME NUMBER
Figure 38
105!
The second karyogram (Figure 3 8) Illustrates the
chromosome analysis for the MIX cells In their tenth pas
sage. The results indicate a maximum number of cells hav
ing either 55 or 58 chromosomes; stilly most of the spreads ^
fall within a range of 49-58 chromosomes. These data are |
j
In contrast to the EPL (eplthe11al-like cells) karyogram |
which appears In Figure 39. This latter graph shows a uni-■
modal profile with an even distribution between 5 2 -5 8
chromosomes. It Is Interesting to note that even though
the EPL and MIX cell populations were Isolated at the same
time from the same source^ substantial differences are
found In comparison of their chromosome numbers.
I
Nude mouse tumor-derived cells have also been ana- |
lyzed. Figure 40 depicts the results for the NEP (epl- I
!thellal-llke) cells In their third passage. As with the j
other eplthe1 1al-like cell strains established In our lab- |
oratoryj these cells have a unlmodal distribution. In this |
case,, most of the counts fall between 54-58 chromosomes.
These c e l l s then^ have a profile that resembles the EPL ■
I
cells more than It does either the TuWl or the MIX cells.
Fibroblasts from nude mouse nephroblastomas have |
been characterized as well. The NUF cells were karyotyped |
In their third passage. Unlike the previous cell cultures^ 1
some difficulty was encountered In finding an appropriate
number of "clean cell bursts." This condition resulted In i
Figure 39- Karyotype profile: EPL cells
(from patient E.C.)
Demonstrates maximum cell bursts
(at 5 2 - 5 8 chromosomes).
Figure 4o. Karyotype profile: NEP cells
(from nude tumor 24-1)
Demonstrates between 54-58
chromosomes.
106
107'
i f )
I-
co
cr
=)
CD
LU
O
EPL (13)
-7
H
I • • • •
» • • • •
• • • • •
# e e #
n u j i J L
4 0
4 4 48 52 56 60
CHROMOSOME NUMBER
a
64 ' 68 OVER
70
Figure 39
NEP (11-3)
o -
68 OVER 56 52 60 64 40 48 4 4
CHROMOSOME NUMBER
70
Figure 40
- - ÏÔ8'
a profile composed of relatively few counts. The actual
data are shown In Figure 4l. From the Illustration It ap
pears that the tested cells may have a bimodal distribution
with maxima at 46 and 50 chromosomes. These data contrast
with previous results for the epltheliai-like cells (which
all had unlmodal curves and maxima In the range of 5 2 -5 8
chromosomes).
A final "control" profile was determined for the
NHP cells from amniocentesis. The results of this test are
shown In Figure 42. A single peak of counts was found In
44-46 chromosomes. However^ when these same cells were
tested at the Medical Genetics facility at UCLA (Los Ange-
les) they were considered normal with a karyotype of 46
chromosomes. It appears from this discrepancy that the
technique used In the present study may generate profiles
that are slightly lower In chromosome number than the actual
counts for the cells.
Soft Agar Techniques
Cells have been grown on soft agar plates on two
occasions^ with triplicate plates prepared each time. The
results of the tests are listed In the table below.
As Indicated in the growth data^ TuWl cells grew
well In soft agar. In all of the tests,, large colonies
formed within 10 to l4 days of plating. In some colonies,,
Figure 4l. Karyotype profile: NUP cells
(from nude tumor 241)
Demonstrates a possible bi
modal distribution (at 46 and
50 chromosomes)
Figure 42. Karyotype profile : NHP cells
(from normal human amnio
centesis )
Demonstrates a unlmodal dis
tribution (between 44-46 chrom
osomes)
109
110!
C/5
h-
(/)
(T
ÛÛ
NUF (II-3)
44 48 52
CHROMOSOME
56 60
NUMBER
—I 1 -----1 -----r
64 68 OVER
70
Figure 4l
-7
h-
ÛÛ
44 60 64 68 OVER 52 56 48
o
CHROMOSOME NUMBER 70
Figure 42
Ill
Table 2
Results of Soft Agar Growth Analyses
Cell Type Growth
Ave. Col. Size
l4 day
Ave. Percent
Cells Growing
TuWi ++++ 50 cells hoio
EPL ++ 50 cells 25^»
MIX ++ 25 cells
%
KEP + 10 cells 2 .$
MJF 0
-- --
PEK 0
--
hundreds of cells were present. The EPL (epithelial-like) i
cells,, from a human nephroblastoma,, also grew in soft agar.
The growth of these cells was noticeably slower, though,
than the TuWi cells. After about four weeks in culture, |
colonies averaged about 25 to 6o cells each. Cell numbers |
were determined by direct counting and colony size measure- :
ments. The colonies were measured in terms of cell diam- |
eter, and cell number was calculated on the basis of
"perfect" spherical volume.
i
MIX cells also grew in soft agar. These cultures, i
however, did not show the vigorous growth of either TuWi or |
EPL cells. Only after 3 to 4 weeks did noticeable growth I
occur. The resulting colonies were distinct but smaller !
masses than found with the previous two cultures. In one
instance, the cells failed to form any colonies. NEP cells ,
from a nude mouse-grown tumor grew poorly in soft agar. A
few small colonies were formed only after approximately one |
1121
month In culture. Two cell strains which failed to grow in
the soft agar were the NUE cells (fibroblastic cells de
rived from mouse-passaged tumor) and NHP cells (normal i
human fibroblasts from amniotic cultures). |
Prom these studies, two major observations were j
made. Pirst, only the epithelial cells grew in soft agar, i
Second, there was a positive correlation between the time '
these cells have been in culture, and their ability to form
colonies. Specifically, TuWi cells (with greater than
twenty passages) formed the largest colonies in the short
est period of time, whereas the NEP cells (three passages)
showed a limited ability to grow. The EPL cells (with
i
thirteen passages) were intermediate in capability between |
TuWi cells and the MIX cells (with eight passages). |
Examples of colonies formed by cell cultures are
illustrated in Pigures 43-46. A "giant" colony formed by
TuWi cells (approximately 1000 cells) is shown in Pigure 43.:
Pigure 44 is an example of EPL cell growth, but the pattern
is also similar for MIX cells. NEP cells having limited |
capability in agar are shown in Pigure 45. Also, Pigure 46 |
I
shows NHP cells which failed to grow; these cells have i
growth characteristics similar to normal human fibroblast |
i
(amniotic) cells. Ï
Figure 4 3. Soft agar colony formation: TuWl cells
(The commercial cell line developed
massive colonies as Illustrated
here [with 500-1000 cells];
400 X mag.)
Figure 44. Soft agar colony formation: FPL cells
(This line [from patient E.G.] shows
moderate growth^ with colonies of 2 5-
30 cells. The pattern Is also typ
ical for MIX cells; 400 x mag.)
113
114
[Figure 43
»
À
Figure 44 J ^
S2
%
o
#
0
9 V
G
V
w
(h,
- ? > -
Figure 45. Soft agar colony formation: NEP cells
(This line [from a nude mouse grown
tumor] demonstrates slight growth
in agar_, with colonies of 5-25 cells
each; 400 x mag.)
Figure 46. Soft agar colony formation: NUF cells
(This line [from a nude mouse tumor]
shows a complete lack of growth
[negative plate]; the pattern is
also typical for NHF amniotic cells;
400 X mag.)
115
116
Figure 45
* »
?
•
#
#
> m*
• «
c
*
%
' n
»
. # . 3 # : , # ' #
r i d i l
Figure 46
117 I
Differential Staining Experiments
As described previously^ our laboratory is inter
ested in various aspects of cell surface anomalies of |
nephroblastoma. Of particular interest are cell surface- '
associated "tumor antigens" and glycosaminoglycans (acid
mucopolysaccharides). Until now^ we have had no indication
as to the tissue source of either the mucopolysaccharides
or the "abnormal" antigens. This set of experiments was
designed to probe the individual nephroblastoma cell types
for their association with glycosaminoglycans (GAGs). For
this purpose,, both frozen tissue sections from a nephro
blastoma patient (E.G.) and vitro nephroblastoma cell
cultures were examined by differential staining techniques.'
Included in these studies have been the treatment of the
malignant tissues and cells with polysaccharide-staining
chemicals (Alcian Blue^ Toluidine Blue^ and Ruthenium Red).
An example of a frozen section (8 microns, fixed
with 10^ neutral formalin for 48 hours) from Wilms tumor
patient E.G. is shown in Figure 47. This particular sec
tion was stained with a 1$ solution of Ruthenium Red in
GMF-PBS. As illustrated^ the stained regions are almost
exclusively in clusters of undifferentiated epithelial
cells (at the top and bottom center of the photograph).
Interstitial tissues shown at the bottom left of the figure
remained essentially unstained. (it should be noted that
Figure 47* A frozen section from nephroblastoma
patient E.C.^ stained with Ruthenium
Red. (The two deeply stained "patches"
are nephrogenic [epithelial-like]
regions; 400 x mag.)
118
1 1 9
Figure 47
120
:Ruthenium stains a faint red in color; this is particularly
difficult to photograph. In addition^ the Ruthenium Red
will not cross cell membranes; as a result^ stained regions
are typically cell surface components. From this initial
observationj it was suspected that the epithelial-like
nephrogenic cells may be responsible for most of the gly
cosaminoglycans associated with this malignancy. In an
effort to pursue this further^ additional staining was con
ducted on isolated cell populations from both of the major
tumor cell types.
A total of six cell types was used in conjunction
with four stains in these experiments. In each test^
nearly confluent cultures were fixed in 5 percent glutarde-
hyde (in serum-free complete media) for two hours or more
before staining. Some cultures were also treated with
Varidase (O.l percent in serum-free media) for twenty-four
hours before staining. In these latter experiments^ the
cells were fixed and washed prior to the addition of en
zyme. Alsoj control cultures (without Varidase treatment)
were incubated in parallel with the treated cells.
The data from these studies are shown in specific
examples in Figures 48-53. Generally^ the epithelial-like
cells (EPL^ MIX;, NEP;, and TuWi) all reacted similarly with
the stains. As a result^, the NEP cells were chosen to
represent the phenomena exhibited by this cell type. Fig
121
ure 48 Illustrates the staining of NEP cells with Toliudine
Blue. The photograph depicts four distinct regions of
stain uptake. They are: the nucleoli;, the cytoplasm (dif
fusely stained);, the plasma membrane S;, and scattered extra -
jcellular aggregates. A duplicate of the culture of Figure
'48 was Varidase digested; this culture (stained in the same
manner) is shown in Figure 49. From this illustration;, it
I
is evident that radical changes have occurred in the stain-
I
able substances of the Varidased cells. The clarity of the
non-digested cells is in contrast with the diffuse indis
tinct staining of the digested culture. In addition^ the
quantity of stain picked up by the treated cells is sub
stantially less than the untreated control cells' stain.
A second stain that was used for these determina
tions was Alcian Blue. In this case^ all of the cells ex
amined were stained equally; they were all diffusely tinted
a faint blue color. Both epithelial and fibroblastic cells
appeared equally stained by this procedure^ and staining by
this agent was not pursued further. (illustrations are not
shown of cells stained with this agent.)
Ruthenium Red was used with more success in the cell
systems of the cultures. As indicated previously;, this
stain was found to bind primarily to the epithelial elements
of Wilms tumor tissue. When used in the same manner (as
previously outlined);, this chemical also stained epithelial-
Figure 48. Differential staining of cultured colls
NEP cells stained with Toluidine Blue.
(Demonstrates clearly defined cells
[diffusely stained] with aggregates
of extra-cellular substances [deeply
stained]; 1000 x mag.)
Figure 49* Differential staining of cultured
cells: NEP cells after Varidase
digestion. (1000 x mag.)
122
1 2 3
Figure 48
Figure 49
.... 124 :
like nephroblastoma cells. Figure 50 shows a culture of
NEP cells stained with Ruthenium Red. In contrast to the
cells stained with Toluidine Blue;, Ruthenium Red gave a
I
diffuse rather than distinct staining pattern. As in the |
previous experiments^ Varidase-treated cultures were also
stained (Figure 51). After enzyme treatment^ a substantial
decrease was found in the stability of the cells. This re
sult i 8 ; , then,, consistent with the studies using Toluidine
Blue. ^
Differential staining techniques were also applied
to fibroblastic cell cultures. Included were cells from a ;
nude mouse-passaged tumor (NUF) and normal human embryonic
fibroblasts (NHF) . Since these cells stained similarly;, |
only the NUF cells are illustrated (Figures 52-53)* The
tested fibroblastic cells showed staining properties analo- ;
gous to the interstitial (fibroblastic) regions of the
!
frozen tumor sections ( Figure 47). That is ; , these cells
failed to be stained by Toluidine Blue or Ruthenium Red.
Figure 52 demonstrates the morphology of the mouse derived |
I
fibroblasts when stained with Acetic Orcein (a general his- |
i
tological stain) . Also, , Figure 53 shows a duplicate cul- -
ture stained with Ruthenium Red. A similar pattern was
found when these cells were stained with Toluidine Blue.
It appears from these results that the nephrogenic elements
of the tumor and the epithelial-like cultured cells may
Figure 50. Differential staining of cultured cells
(NEP cells stained with Ruthenium Red.
Demonstrates diffuse staining^ propor
tional to cell density; 1000 x mag.)
Figure 51. Differential staining with Varidase
digestion. (NEP cells stained as in
Figure 50^ but after Varidase diges
tion. Demonstrates a substantial
loss of stainable material; 1000 x
mag. )
125
126
M ÿ M JCKiX
Figure 50
m
»
Figure 51
Figure 52. Differential staining: NUF cells with
acetic orcein.
(Demonstrates fibroblastic appearance
in both sparse and dense colony re
gions; 1000 X mag.)
Figure 53. Differential staining of cells with
Toluidine Blue.
(Demonstrates an inability to pick
up stain [cells appear as "clear
cell ghosts"]; typical pattern also
for NUF cells stained with Ruthenium
Red; 1000 x mag.)
127
128
'T V.
"; .1 / ' •
''r <
T f r i
V- •
*t.
PC-
/
# Ê W
V i
Figure 52
m
. ■ >>
/j
V
Figure 53
129 !
synthesize more ce11-associated GAGs than the interstitial i
tissues or the fibroblastic cells.
Biochemical Analysis: Precipitation I
by Various Agents
Several reagents designed for precipitation experi
ments were used to chemically characterize extracts of
mouse-passaged tumors^ pooled extracts of Wilms tumor tis-
suesj and vi tro cell cultures. The specific test re
agents utilized in this section were chosen to coincide ,
with previous studies from our laboratory (Allerton et
al.).^^ In all^ six chemical precipitation analyses were
completed. The results from these studies are compared to i
the findings of Allerton et al. in Table 3*
From this table^ it is clear that the findings of
26 ' i
Allerton et al. are reproducible with EDTA tumor extracts{
of Wilms tumor tissues. In the tests conducted here^ pre
cipitations were found in every case with my EDTA pool [
(from five patients); this is consistent with the earlier
26
reported work. Some deviation has been founds though,, ,
when the EDTA extracts of nude mouse-grown tumors were com- !
pared to the pooled extracts of human tumors. Specifically,,^
no precipitation was found with acetone at 75^ final volume. ;
The other five test results were the same for either the '
nude mouse-grown or the human tumors. In contrast,, numer- :
ous differences were found between extracts of tumor tis-
130
Table 5
Precipitation by Various Agents
Acetic
Acid
1 o
CPC^
Ifo
Ethanol
50^
Trichloro
acetic
Acid 5^
Acetone
75/0
Ammon.
Sulfate
50^
EDTA ext.^ + + + + + +
b
Trypsin ext. + + + + + +
EDTA ext. pool + + + + + +
EDTA ext. nude
mouse-grown + +
+ + 0 +
tumor
EDTA ext.
EPL cells
O O
0 + 0 +
EDTA ext.
I\I EP cells
O o 0 + 0 +
EDTA ext.
HUE cells
o o 0 + 0 +
EDTA ext.
MIX cells
o o 0 + 0 +
EDTA ext.
TuWi cells
o o 0 + 0 +
a
Cetylpyridinium Chloride.
b
Allerton et al.
1 3 b
sues and those of vl tro cell cultures. As Indicated in
the tablej only two of the solvents would generally pre
cipitate components of these samples . TCA (5^) and ammo- i
niun sulfate (50^) caused precipitates to form. In all^ i
five different cell types were examined; in each case^ the |
extracts of in vi tro cultures were precipitable only with ^
the two indicated protein-precipitating reagents. I
i
The data from these experiments indicate that the ;
EDTA extracts of the nude mouse-grown tumors are somewhat
similar to the extracts of human nephroblastoma tissue. In I
particular^ the precipitation with 3$ acetic acid and CPC ’
may be correlated with the presence of significant amounts
of glycosaminoglycans in the mouse-grown tumors. However^ |
most of the test reagents failed to precipitate extracts of
the cell cultures. Exceptions were found with two chemi
cals which generally precipitate proteins (nonspecifically).
Unfortunately^ no reaction was found with either acetic
acid or CPC. It is likely^ then^ that these cultures do
not produce large amounts of EDTA-extractable polysaccha- [
rides. However^ the presence of hyaluronidase activity in |
the EDTA extract of these cultures could account for the |
low chemical precipitation profiles (see enzyme analysis
section).
132 :
Biochemical Analysis: Measurement
of Enzyme Activity
The activity of several enzymes has been determined
as a part of the general characterization of nude mouse-
grown nephroblastomas and cell cultures from these tumors.
Specific procedures are given in the Methods section. This
particular work has focused on the hydrolytic enzymes asso
ciated with either neoplastic formations or with normal
kidney tissues. For examplej comparatively high levels of
gamma glutamyltranspeptidase activity have been found in
both normal human kidney tissues and in a number of neo
plastic tissuesj including one nude mouse tumor developing
from TuWi commercial nephroblastoma cells (Wise and Mul-
42
1er). Also j elevated hyaluronidase and protease activi
ties have been associated with various malignant tumors
(Cameron).An elevated level of glucose-6-phosphatase
activity has also been found in proximal kidney tubule
cells (Pretlow et al.).^^ In all^ eight enzymatic activi
ties have been tested here. The results of these studies
are indicated in Table 4 and Table 5-
As shown in Table 4^ phosphatase activities were
determined (acid^ alkaline^ and heat-resistant alkaline).
The findings indicate that extracts (EDTA or KCL) from
either the human Wilms tissues or the nude mouse-grown
tumors lacked all three of the tested phosphatase activi-
133
I
E -t
0
1
E
- O ■
ill
(U
! I C C
M
o Q O o o o o
8 g 8 8 3 8 8
d d d d d d d
C - j
O
+ 1
si i 5
o o
? o . « T ,
? 3
O O
+ 1
CJ O 'ê
d
4-1
6 o o o d o
8
d
-t-i
r j 1
|l
8§
d d
o
d d o o o
§ 8
i . i
d d d d d
d c>
-n -n
r n O (
SSSi
8. 8
d d
+ 1 4-1
o o o o o o o o
g liilp S ||S ilg
o o o o o o o o o o o o o o
O - d - O vO O - CO 'O t— o
( —1 I — .~J o
o o o o rvj
4 : ' ^ -
L Z \ L T \
O r-i ■
o o o o o o
O O - ! O CM O rH
88 S8
O r - io 3 o o
Or 4o d d d
^ O ! o
2^. s ' I
g !è l
■ - i 'd -d S'»
C - I X
z ! * ! ± i 4
-P
K
g
X
r . - q H U • j
i 1 i 1 33
U -V 1
3" Cr Tj -a
3 3 r - i
1? S - 4
| 3 i
c. o d
ill
I
I
3
U
I
5
~ A
i l l
* 3
= :
g o
E
0) +j
5 - 1 O
oT §
I g ^
f - i a i T 3
G : - 1 C'
; t : ^
-P P
d i . - i )
SI
%
dO Cl
a M
d i
Is
a j
. . D
d â - C i
> O o O --. PO
3
- « - 4
. m
3 3
r iJ D
O
M
%
>5 :
2O j a O
S j
134
ties. Some low but detactable activity was founds, however^
I in the post-extraction tissue residues from human and
;mouse-derived tumors^ as well as in a homogenate of normal
I human kidney tissue. Extracts and "residues” (cells after
IEDTA extraction) of nephroblastoma cultures grown in our
I laboratory were also found to have demonstrable phosphatase
I activity.
I In all of the positive samples^ the activity was
'generally confined to the alkaline and heat resistant alka-
I line phosphatase; in most cases acid phosphatase levels
were only a few percent of the alkaline phosphatase activi
ties. An exception to this was noted with the normal human
!
jkidney tissue homogenate^ where approximately equal amounts
; of acid and alkaline phosphatase were found. Heat-re-
Isistant phosphatase activity was also confined to a few
j samples. In this case^ only the vitro cell culture
I residues were positive. Specifically^ the TuWi^ MIX^ and
EPL cells (after EDTA extraction) all had a ratio of alka-
I line phosphatase to heat-resistant alkaline phosphatase of
! about 0.3; the NEP cells had a ratio of about 0.03; in this
testj no heat-resistant activity was found in either the
'normal human kidney homogenate or in the NUF culture homoge-
;nate.
! Unlike the EDTA extracts of the human and nude
i
I
I mouse tumorsj the extracts and cell culture residues were
135
positive for alkaline phosphatase. In general^ the EDTA
extracts contained 10^ or less of the activity found in the
cell residues from the same culture. The highest activity
was found in the MIX cultures. In comparison^, the NEP
cells had 75^ of the MIX activity^ while the EPL and TuWi
cells had ^0% and the NUF cells had 3^. A number of con
clusions can be made from the examination of phosphatase
activity. One is that intact tissues do not appear to be a
rich source of phosphatase. In vi tro systems however ^
have easily detectable activity^ particularly in the cell
residue fractions. This latter finding is important since
phosphatases are typically cytoplasmic in nature; it ap
pears from this study that these enzymes are not removed by
EDTA extraction. As a result^ the cell membranes (and thus
the cells) appear to remain intact or at least are not
grossly damaged so as to leak out enzymes. The EDTA^ then^
appears to act primarily on extracellular material of the
cultures. Since very little phosphatase activity was iso
lated from the tumor tissues^ a conclusion could not easily
be reached concerning the effect of EDTA on intact tissues.
Gamma glutamyl transpeptidase (GGTP) activity was
also monitored for the same samples as the phosphatase
activity. The GGTP enzyme may be particularly applicable
to malignant kidney tissue. In the first place,, the GGTP
activity is the highest (with respect to mammalian organs)
: 136 i
'In normal kidney tissues (liver tissue has the second high-;
'est activity). Second^ a number of malignant tissues have
I
;been reported to have elevated GGTP activity. In one re-
j i 2 I
port by Wise and Mullerj human Wilms tumor cells (TuWi) i
I
grown in nude mice had elevated kidney "type” gamma gluta- I
' my1transpeptidase. Unfortunately^ these investigators gave i
their data only as "units GGTP"; as such it is impossible |
to directly compare with the present work which has been
calculated as units per milligram protein. |
It is evident from the data in Table 4 that the |
I
normal human kidney tissue had by far the greatest activity |
with about 5100 units. Extracts of human or mouse-derived j
nephroblastoma^ though^ had little if any activity (0-20 j
units). In comparison^ some GGTP activity was found in the |
extracted cell residues^ but not the EDTA extracts^ of |
cultured Wilms cells. Specifically^ the EPL and MIX cells
had essentially identical activity levels of 29O units. In
this study^ the NEP cells had 1 2 5 units of activity^ the |
TuWi cells had 50 units^ and the NEP cells had the least |
activity with 3O units. The above GGTP assays verify that i
this enzymatic activity is present at high concentrations |
in the kidney. They further suggest that the activity may !
not be associated with nephroblastoma tissuesas little ;
activity was found with either human or mouse-grown tumors. :
With in vitro culture activity was founds as in the case of ;
137
the phosphatase,, only In the cell residue fractions; again
this evidence supports the contention that the enzyme is
intracellular and the residual cells are intact.
The specific data from the cell cultures are also
interesting. The EPL and MIX cells (from the same tissue
source) have the same GGTP activity. In contrast, the NEP
cells from a mouse-grown tumor and the TuWi commercial cell
line have substantially less activity. It appears from
this information that there is an inverse correlation be
tween the amount of activity and the "temporal distance"
from the human tumor condition. That is, the EPL and MIX
cells were derived directly from a Wilms patient. The NEP
cells, with less GGTP activity, though, were passed through
two groups of animals before being cultured; in this case,
the cells have been environmentally removed from the child
for some time before tissue culturing. In addition, the
TuWi cells (with the lowest GGTP activity) have undergone
an vitro transformation due to an extended period in
culture. These results appear to reflect a "de-differenti
ation" of the tumor cells after removal from the "human
tissue environment" of a nephroblastoma patient.
Another hydrolytic enzyme that was studied was
gluCOse-6-phosphatase (GôPtase). In this set of experi
ments, an attempt was made to compare the "normal human
kidney" to nephroblastoma. This GôPtase activity has been
138 i
associated with the proximal tubule cells of mammalian kid-
p Q
neys (Pretlow et al.) as discussed previouslythese same
tubule cells are Implicated as a possible origin for nephro
blastoma (through the primitive metanephrlc blastema). The ^
samples used for these enzyme tests were the same as those i
specimens of the phosphatase and the GGTP assays. The re
sulting data are Illustrated In Table 5* Of particular
Interest are the low values found with normal human kidney
(no activity)j cell residues of NEP and EPL cells (no
activity) extracted nude mouse tumor residues (O.O96
units)j and pooled Wilms tumor EDTA extract (0.022 units).
These values are In contrast to the high enzyme activities :
found with Wilms tumor tissue residue (O.6 5 units)^ EDTA
extract of mouse tumor (1.20 units)^ EPL EDTA extract (O.37
units) MIX cell EDTA extract (O.68 units) ^ MIX cell resi
due (0.4l units) j NEP cell EDTA extract (0.4l units) TuWl I
I
cell extract (0.74 units) and TuWl cell residue (.71 I
units). These results of the glucose-6-phosphatase assay
are difficult to explain. In general^ most of the EDTA ;
extracts had more enzyme activity than the cell (or tissue)
residues of the same source. This may Indicate that much
of the enzyme Is located on the external surface of the
nephroblastoma cells_, and Is also EDTA extractable. This
observation Is consistent with the cultured cells and the
nude mouse tumorsbut not for the pooled Wilms tumor j
!
specimens.
139
.
' ( y , > ^ 3 e /
Defcer.'aiiiatlou o f Eas.^'ins A c ^ i'fitlc iT . : . Pi oreaoi;, a:ut ülucot:a-6-phcs:j.ba,.aae
îlyaluron idaoe A o ' j a . y
' O t e a ' i C
A g a z - Gcrtase
kCl (GK) 0.000 0 0.000
V-lCi (GK) — o.oco c 0.000
' • pyy'
> . . . . J
0 . 92? e - 1 )
ferlboneai 1.
—
o.cooo 0 0.000
: » l ' ! x iiesi vue C'.lie + C. 008 0. 1.1!) + 0,017 0
0. 09' i
• r0.011
VI'. Residue 0,^10 ■t O.IPC 0. OdO T 0,01; 0 0,090 0.084
îutK .1.070 + 0.190 2.:;60 3 0.220 Temsi 0.000
Ti’ G'i Scrape l . w + 0.l!:0 i.9.'0 r u. .120 i i î l i U 0 . 24c O.Gpl
I Ï Ï J 5 - 6c rope 2.2^0 I
0. W O.Cbo c. 009 0.159
0.010
Il'L- E.et. 0.29.0 1. IcO -r C-, J.IO 0
0.370 Ü.G32
EPL- C i.CpO + O.lSO o.bCo T 0.0;/ 0 O.Ocrfi
M IX - Ext. 6.7.10 +
0.790 0.490
■*■ o.o_ ' = 6 0 0.690 + 0,078
ÎIIX- C j.oyo 'O.990 0,270 T
:'.C 5-
0 0.610 - + 0.049
IIEP- ixb. --- 0,690 - 0.085 0 0.045 T
0.057
?JEP- C ^.380 + U.410 0.290 + 0.026 0 0.000
TuWi-Ext, i.oyo + O.'lOO 0,096 + 0.093 0 0.740 2 0. 060
ruV/i-C 5.190 + 0.290 0.!)95
+ 0.0(18 0 0,710 +
0.091
N a ce-Ext, 1.0 70 4-
0.067
0 .S20 -> 0.071 0 1.200 + 0.100
Wude-C 1.950 £ 0.026 O.ptIO f O.QlO 0 0,009 - + C .002
E P . ' . lie rape 2.ISO + 0,290 1.040 + 0.210 c
Nirp 0.000 0.033
LaAT^.id: t X . l : ^ K C l •b .ic ra c t o f u a V f li n o cm nor
iëtC H e s ià u e ;
yîuiie mouse tu jn o r t is s u e s , r e in a in lr ^
a f t e r EDTA e :c tr a z t io n
V -K C l: V a rid a s e t r e a t e d K E l e x tra W o f W lim o cu.nior
V f R e s id u e :
C e ll b is o u e rcoroii. d u g a io é i’ SCI'A e x t r u c t io r i
o f human W i3 arvs tu m o rs
ïffîK : Nont’a l human k id n e y (Hoziogexiaoe}
S c ra iJ e : C e lls scraped, fi'o m c u lt u r e ' e s s e Is
(n o b extrftv^bec.)
Exb . : E;yfA e x tx -a c c s o f caJ. ra re
C : C e llu la r " p r e c jp io a t e " a f t e r EBfTi e x t l'a c tio n o f
Î 2 L P : l i O i i n e J h ' m ï u r ! p i a s i c v i
c e ils
l4o I
Hyalunonidase was also investigated. This panticu- '
Ian activity was studied because of its association with
neoplasia (Cameron)^ and because of the presence of EDTA-
extractable polysaccharides (containing substantial hya- ‘
luronic acid) in nephroblastoma tissues. The connection^
then,, between the presence of hyaluronic acid and detectable
levels of hyaluronidase was examined. The results of this
set of experiments are shown in Table 5» Unlike the previ
ous enzyme assays^ each of the samples tested had enzymatic
activity (even though control samples of Grade 1 hyaluronic
acid remained undigested under parallel conditions). It
should be noted that all of the samples used in this assay
were dialyzed for 48 hours prior to testing in order to
minimize soluble chemicals having hyaluronidase-like ac
tivity such as ascorbic acid. Prom the results of this as
say ^ as with the glucose-6-phosphatase determination^ cor- |
relations between samples are difficult to make. However,,
some trends are noted. For example,, low values have been
found with pooled EDTA extracts of primary human tumor
tissues as well as with mouse-passaged tumor extracts. It I
must be stated,, though^ that high activity was found with
MIX and EPL cell extracts.
These two examples of maxima and minima for the
hyaluronidase do correlate positively with the results of
the gluCOse-6-phosphatase assay. Another interesting ;
l4ll
parallel between these two determinations is that,, gen- i
erally^ the tissue cultured cell extracts have higher
activities than the cell residues. In both cases,, the ac
tivity appears to be extractable from these cells (presum- ’
ably from the cell surface). A "sideline note" to the
above tests is that the cell residues from EDTA extraction
remain viable (in some cases the cells were replated after
such EDTA treatment) . This information,, together with the
results of the GGTP and hyaluronidase assays,, tends to sug
gest that the enzymes are either associated with the exter
nal surface of the plasma membranes,, are secreted into the
EDTA extraction solution during these procedures. Alter- I
natively,, they may reflect the presence of substantially
more protein in the cell residues than in the extracts,,
since the above values are shown as ng/Pi released/min/mg j
protein.
A final enzymatic activity examined was nonspecific i
protease. This activity^ like the GGTP and hyaluronidase^ |
has been generally linked to neoplastic development. To j
test for pro tease levels,, two methods were employed. One j
was a fibrin agar technique (according to the methods of i
60
Schill and Schumacher ; plates were donated for this study
through the courtesy of Dr. Sorgente of the USC School of
Dentistry). A second procedure (described in the Methods
section) was based on the digestion of acetic acid-precipi-
142
table bovine serum albumin. The results of these tests are
outlined in Table 5- In the fibrin-agar method,, only three
samples (Wilms EDTA poolnormal human kidney,, and scraped
TuWi cells) were positive^, as indicated by clearing zones
around sample walls. In each of these,, the activity found
was equivalent to a solution of about 0.5WM trypsin. The
results from the BSA digestion test were more complex. As
found with the hyaluronidase assay,, all of the samples gave
rise to some degradation of the test sub stance. High lev
els of protease activity were found in pooled EDTA extracts
of both EPL and TuWi cells. However,, low values were found
with HUP cellSj Wilms tissue residues,, and nude mouse-grown
tumor residues. As with the hyaluronidase and glucose-6-
phosphatase assays,, more activity was noted with the EDTA
extracts than with corresponding cell residues. With the
results from the protease determination of cultured cells
alone^ there is a strong correlation between growth poten
tial and the presence of protease activity. That iSj, the
HUP have a low growth rate as well as a low protease level.
The EPL and TuWi cells, though, have high growth potentials
and high protease levels. In addition, HEP cells have an
intermediate growth rate and an intermediate protease level.
This finding correlates well with the findings of other in
vestigators who have noted high protease activity in rap
143
idly growing cell cultures and tissues (especially those of
neoplastic origin),
Biochemical Analysis: Determination of
DNA, Protein, and Carbohydrate
A series of biochemical assays was completed on
human and nude mouse-passaged tumor samples (as well as in
vitro specimens) in order to further define the chemical
nature of these materials. In this set of tests, classical
colorimetric assays were used to determine the protein,
DMA, uronic acid, hexose, and hexosamine contents of both
EDTA extracts and residual cells and tissues. The specific
procedures utilized are described in the Methods section.
The data from these experiments are presented in Table 6.
For the expression of protein values, an absolute term
(mg/ml) has been used. In all of the other tests, the
indicated values are expressed in relative terms of "mg" or|
i
"pg" per "mg" of protein.
The protein concentrations of the tested samples
are shown in the first column of Table 6. As expected, the I
Wilms tumor EDTA extract pool and the normal human kidney
homogenate were relatively rich in protein. With the other
samples, unequal dilutions account for some unexpected
values. For example, the NEP EDTA extract has more protein
per volume than the NEP cell residues. This is due to a
higher dilution of the cell fraction than of the EDTA frac-;
tion.
144
c a
h \
'a
o
u
D
a
S^i
It!
“ 1
< 1 ; d
II
f i
&
§ *
? i i g
-- O a.p4
x
K
b C > <
0 1 CÛ
Q S
5 0 •
^'11
X g
Ü C L ,
3 50
■5
r 4 - d - \0 i f \ i r v - d - m 4 tOv\o i r \ K\ X " " » O b —
rHCOOOOOOOOOOOOJOO
t r —co c—r-iC\oOr-iaoü\oo
O LTN r - i C — C — C - —L O
I —I r -4 : — 1
-H r - H CU L P . .
04 '~ t '~ i
CO O O O C— CO O'. cr\ r - H rH CO r - H r - H
O CM I - H CJ < —1 -d" CM r * \00 -d" O CM " —I MO
?, S c! S' S S ^ i l Ï ? 3 ^ ci ^ s! a b
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
CJ CJ CT\ .H CO IXA O CM CJ C-\0 ^ CQ O
rA LPv r - t ^ ,H CO CJ -3- CM -1 C‘
C M O rH O O O O
Cl L T \ - : 4 -
(6 o o o o’ o* d o
I'
5 c
èà
. .
0
u : X 0
1 ) X V X p V 1 - 4
r - * V J - i ‘ V
< -H
r —(
d
r - H
1 - 4 g
r —i
g
0 1; 0 3 .0
f
; j CM V
%
4 Oy
5 3 Z 1 4 1
t: 5
U
A
3 3
p
Î - .
t3
a
« d o
'ta
3
2 3
3
-p
V i V4 0| (r i
0
0
0
a
■ 3 1» 3 ■3
i-l V
C' A
1
0 ■ j 5
2=
b Vi 3
a X X 2
•Ü 1 )
y 4
4
d
, —1
S r-H
■L) A u 3
0 3 0
K il S o l
: ^ 145;
The determination of uronic acid concentrations^ '
listed in column two of Table 6, shows approximately four
times the uronic acid in the Wilms EDTA extract pool (and ;
the nude mouse tumor EDTA extract) than in the normal human ■
kidney or any of the cell culture samples. These data sup
port the visual observations that the EDTA extracts of the
tumors (either human or nude mouse) were quite viscous;
none of the other samples demonstrated significant vis
cosity. From these findings^ it appears that the viscosity
may be due to a high level of hyaluronic acid. One sur
prising result of this test was that the EDTA extracts of
the cultured cells had generally less uronic acid than the i
!
cell residuesj in contrast to the findings with the tumor
tissues. Howeverwith the in vitro specimens^ there is a
correlation with the presence of hyaluronidase. That is^
the EDTA extracts were found to have higher hyaluronidase
activity but a lower uronic level than the cell residues. :
A notable exception was the normal human kidney in which
low levels of both uronic acid and hyaluronidase were found.;
In contrasts the EDTA extracts of human primary tumors or |
i
nude mouse-grown tumors had high levels of uronic acid with i
low levels of hyaluronidase.
Of particular interest was the presence of detect
able DNA in all of the EDTA extracts. From this finding,
it appears that EDTA may cause damage to some cells. Al- ;
I 146
jternately, these results may reflect a necrotic or degen-
lerative state of the tissues or cells at the time of ex-
I traction.
; One area of analysis that is of special interest to
I our laboratory is the carbohydrate determination (including
hexose and hexosamine) also shown in Table 6. Relatively
low levels of hexose and hexosamine were found in the EDTA
extracts of both the Wilms tumor tissues and the nude mouse-
; grown tumors; the normal human kidney extract also had low
levels of these carbohydrates'. The residues of the ex
tracted Wilms tissues and all of the cultured cells had
substantially higher levels of these components. A compa-
'rable interest in the uronic levels of these latter samples
was not found. It seems, then, that the cultures have an
I enhanced ratio of hexose (and hexosamine) to uronic acid.
Further, in the study of the cell cultures, a consistent
pattern of hexose and hexosamine presence (in either EDTA
extract of the cell residues) was not found. The EPL and
MIX cultures had more carbohydrate in the EDTA extracts,
while the TuWi and NEP cells showed more sugars in the cell
residues. From the indicated data, the ratio of hexose to
hexosamine was calculated; this is shown in the last column
of Table 6. Most of the ratios are relatively consistent,
ranging from 0.3 to 0.6. In contrast, both the EDTA ex
tracts of the human and nude mouse-grown tumors indicated
147 :
ratios of greater than 1.0 (actually 1.1-2.0). Another
interesting correlation has been made with previous reports
26
of carbohydrate contents. Specifically, Allerton et al.
reported a ratio of hexoamines to uronic acid of 1.45. In
the above studies, a ratio of 1.40 was found with a pooled
Wilms tumor EDTA extract (other than the original Wilms
extract used by Allerton et al.). In addition, the nude
mouse tumor EDTA extracts had a ratio of 2.68. In com
parison, the average ratio for the cultured cells was about
j
175-200. From these data, it appears clear that the nephro
blastoma tissues favor uronic acid production (as hyaluronic
acid) and the cultured cells favor hexose and hexosamine |
production (most likely as glycoproteins).
Binding Interactions: Biochemical I
and Immunological Studies !
For some time, our laboratory has been investigat
ing Wilms tumor-associated antigens and cell surface muco-
i
polysaccharides. In the course of my studies, evidence has ;
been found for binding interactions between polysaccharide
components of the tumor and antigenic proteins. To evalu- ,
ate and characterize this phenomenon, a combination of bio- '
chemical and immunological approaches was used. Included
have been chromatographic techniques, enzymatic modification
tests, and determinations of immunogenicity. ,
Preliminary evidence for binding interactions ap-
: 148 i
peared in an Immunoelectrophoresis study of a peritoneal ;
fluid sample from a Wilms patient (J.M.). Specifically, |
immunoprécipitation bands in these experiments appeared I
abnormally elongated and somewhat distorted. This profile, '
shown in Figure 54, illustrates band elongation particu
larly with serum albumin (to the right of the photograph),
with the trailing edge of this protein band extending back
to the loading well. However, when this same peritoneal
fluid was pretreated with a hyaluronate-specific hydrolyic
enzyme (Varidase-streptokinase/streptodornase, American
Cyanamid Co.) at 0.1 percent in CMF-PBS, for 24 hours at
37° C, the observed banding abnormalities disappeared. This;
[ I
latter profile was essentially identical to that formed by
using normal human serum as an antigen source. An example ;
of an immunoelectrophoresis pattern of a Varidase-treated
peritoneal fluid is shown in Figure 60. Note that, in this
photograph, each of the bands has a typical symmetrical arc
without elongation.
After the initial observation with the peritoneal ,
fluid sample, studies of the possible binding interaction ;
were expanded to include a number of EDTA extracts of Wilms ;
tumor specimens and ijn vitro cell cultures. In each case,
abnormal banding was found before, but not after, Varidase
treatment. Of particular interest was a second observation
that, in some samples, additional bands were disclosed
Figure 54.
Demonstration of abnormal immunoelec
trophoresis banding
(Well: Peritoneal Fluid from patient (j.M.)
Troughs: Anti-normal human sera)
Figure 55.
Demonstration of antigen "blockage” by
immunoelectrophoresis
EDTA extract ”A”
2: Varidase treated EDTA extract
Varidase treated EDTA extract
EDTA extract ”B”
(Well 1:
Well
Well 3:
Well 4:
"A”
”B"
All troughs: Antisera to the GB antigen)
Figure 5 6.
Demonstration of abnormal immunoelectro
phoresis banding in the presence of
hyaluronic acid
(Well 1: BSA and Hyaluronic Acid
Trough: Anti-BSA
Well 2: Sample as in well 1, after
Varidase treatment.)
149
150
Figure 54
îmÊmÊÊÊÊiÊÊÊMÊÊm
Figure 55
Figure 56
151
after hyaluronidase (Varidase) treatment. Two examples of
such an inhibition of antigenic expression are illustrated
in Figure 55* In both of the patterns shown, a particular
band (in this case, the Gamma Beta [GB] antigen) appears
only after Varidase treatment.
Subsequent to the above findings, attempts were
made to reproduce the apparent binding phenomenon with
purified reagents. Since serum albumin and hyaluronic acid
were both implicated as being involved in binding interac
tions, these substances were utilized. In one such study,
human mercaptalbumin (prepared by S. E. Allerton) and human
umbilical cord hyaluronic acid (Grade I, Sigma Corp.) were
mixed at room temperature for 3O minutes before immunoelec
trophoresis with anti-normal human sera (GIBCO) as antibody
source. At certain concentration ratios, electrophoretic
banding abnormalities analogous to those described above
were found. Figure 56 depicts the profile achieved with a
test solution containing 25 mg/ml and 0 .5 mg/ml hyaluronic
acid (in CMF-PBS). The band distributions shown are quite
similar to those found in the peritoneal fluid; in addi
tion, the concentrations of the purified components used
were similar to those found in the peritoneal fluid (hya
luronic acid was measured in the peritoneal fluid sample by
the carbazole reaction of Bitter and Muir and the albumin
content was determined by radial immunodiffusion tech-
152
niques). As a control, the albumin-hyaluronic acid mixture
was treated with Varidase. For these enzyme-hydrolyzed
samples, profiles identical to native albumin alone were
generated, as illustrated in Figure 5 6.
The binding phenomenon between proteins and hya
luronic acid was pursued with gel chromatographic tech
niques. In this set of experiments, purified proteins
mixed with hyaluronic acid (as in the immunoelectrophoresis
studies) were chromatographed using either Sephacryl S-200
or Sephadex 6b (Pharmacia, Sweden) gel columns. Details of
the procedure are described in the Methods section. The
results of a typical chromatogram are shown in Figure 57.
In this case, Sephacryl S-200 was used to resolve a mixture
of human serum proteins (igO, IgM, and albumin at 50 mg/ml
each, in CMF-PBS at pH 7.2), indicated by the dotted lines.
A duplicate experiment was performed with the above three
proteins mixed with human hyaluronic acid (umbilical cord.
Grade I from Sigma Corp., at 10 mg/ml final concentration).
The elution profile for this protein-hyaluronate mixture is
also shown in Figure 57 by the solid line. Two features of
this second analysis are evident. First, none of the three
proteins used was re solved by the column (even after a 3 0-
hour run on a large column) in the presence of hyaluronic
acid. Subsequently, the presence of the three proteins was
verified by double-diffusion technic on Ouchterlony plates
Figure 57. Chromatographic demonstration of
binding interaction between hya
luronic acid and serum protein
(*The dotted line represents an
initial chromatography run with
human IgM, IgG, and albumin.)
(**The solid line represents a du
plicate chromatography run with
the addition of Hyaluronic Acid
[Grade I from Sigma Corp.])
153
154
IgG
ALb IgM
0 .9 -
0.8-
0 .7 -
0.6-
c v j 0 .5 -
Q
0 .4-
0 .3 -
0.2-
20 25 3C
HOURS
Figure 57
155
i after Varidase digestion. In the Initial run^ the peaks
Identified by these methods are appropriately labeled. In
the second chromatogram^ though^ all of the proteins were
found equally distributed In the single broad peak. That
Is^ Its position (displaced to the right at higher elution
volume) normally corresponds to lower molecular weights of
the eluted components. However^ the test proteins are al
most certainly still of very high molecular weighty since
they are still Immunelogically reactive^ and no protease
activity was found In hyaluronic acid preparations tested
60
by the fibrin agar technique of Schill and Schumacher.
If the proteins and hyaluronic acid are essentially Intact^
then an aggregation phenomenon between these two Is likely;
this would produce a single broad peak. Also^ very large
aggregates could be physically Inhibited from migrating |
through the column. Alternatelythe combination of poly- I
saccharide and protein may chemically Interact with the gel I
beads. Since hyaluronic acid alone migrates to the left of
the IgM peak,, a massive aggregation of the proteins with :
hyaluronate Is strongly suggested. ;
i
A final set of experiments In the Interactions be- I
tween purified serum proteins and hyaluronic acid dealt ;
with the quantitative aspects of binding. For these tests, ,
rocket Immunoelectophoresls was used In conjunction with >
I
bovine serum albumin (BSA) and antl-BSA sera. The rocket
156
plates were made according to the description given In the
Materials section. Standard curves were made by serial
dilution of a BSA standard solution. In addition^ various
concentrations of hyaluronic acid were added to different
concentrations of BSA for a combination of samples. In j
ithese experiments,, all of the test mixtures were loaded IntJ
the wells of the gel (at 5 pi each),, and the samples were j
Ielectrophoresed for one hour at 100 volts (4° C). The re-
'suiting gels were deprotelnlzed and stained as usual. One
j such electrophoresis pattern Is shown In Figure 5 8. From
'the first four wells^ a nearly linear Increase In migration
I
jdlstance Is found with Increasing BSA. In the Illustrated
! rocket pattern^ the migration was 5-7 mm/mg BSA. However,,
IIn the presence of 0.1 mg/ml hyaluronic acld^ the migration
I
dropped to 4.3 mm/mg BSA. With 1 mg/ml hyaluronic acid the
mobility dropped even further to 2.7 mm/mg BSA. A final
sample with 10 mg/ml of the polysaccharide completely In-
ihlblted the formation of rocket bands for either 0.5 pg
1(0.1 mg/ml) or 1.0 pg (0.2 mg/ml) BSA. From these experl-
mentSj, a quantitative approximation for the binding capa-
!blllty was made. Specifically^ 1 pg of hyaluronic acid was
found to block the antigenic expression (as Immunopreclpl-
i
;tatlon with antibody of about 0.4 pg of BSA). From this
I set of experiments,, It was concluded that hyaluronic acid
I could physically bind to proteins and block their antigenic
Figure 58. Rocket Immunoelectrophoresis
(Well 1
0.125 pg
BSA
Well 2
0.25
pg
BSA
Wei]
3 0 . 5 pg
BSA
Well 4 1.0
pg
BSA
Well
5 0 .5 pg
BSA + 0.1
pg
hyaluronic acid :
Well 6 1.0
pg
BSA + 0.1
pg
hyaluronic acid
Well
7 0 . 5 pg
BSA + 1.0
pg
hyaluronic acid
Well 8 1.0
pg
BSA + 1.0
pg
hyaluronic acid
Well
9 0.5 pg
BSA + 10.0
pg
hyaluronic acid
Well 10 1.0
pg
BSA + 10.0
pg
hyaluronic acid)
Figure 5 9. Demonstration of Immunogenicity of
peritoneal fluid
(Well: Peritoneal fluid from patient J.M,
Trough: Antisera to the peritoneal fluid)
Figure 6 0. Demonstration of altered Immunogenicity of
peritoneal fluid after Varidase treatment
(Well: Peritoneal fluid from patient J.M.
Trough : AntIsera to Varidase treated
peritoneal fluid)
157
158
\ / . /
\
Figure 58
Figure 59
Figure 60
159
j(immunogenic) sites. To correlate the findings of the
rocket Immunoelectrophoresis study with BSA to the speci
mens from Wilms patients^ a test of Immunogenicity was
de signed.
This last approach toward Investigating the Inter
actions of serum proteins with tumor-associated polysac
charides Involved tests In rabbits. In these experiments
aliquots of the peritoneal fluid (from patient J.M.) were
Injected Into animals from Freunds Complete Adjuvant. A
I second group of animals was Inoculated with an Identical
I sample pretreated with Varidase (as described previously).
For purposes of a control, the sample without enzyme was
Incubated In parallel with the Varidase sample during the
enzyme digestion period. After 4-6 weeks, both of the
! animal groups were bled as usual, and the serum was col
lected. Immunoelectrophoresis experiments with these anti-
sera Indicated radically different responses to the antigen
preparations. The animals treated with Intact peritoneal
fluid generated antibodies only against five minor beta
globulin proteins In the sample (totally lacking were the
major bands of albumin and IgG). With the digested samples,
however, a complete spectrum of bands was found. These
latter profiles were comparable to patterns obtained with
commercial anti sera to normal human serum proteins. Most
noticeable was the presence of large IgG and albumin arcs.
These results are shown In Figures 59 and 6o.
160 I
In all of the above experiments dealing with bind
ing interactions, certain consistencies have been found.
I
One Is that hyaluronic acid can bind to certain serum pro- :
telns (and tumor-associated antigens, as found with the GB
antigen). Further, such binding alters the migration of
proteins In gel chromatography, Immunoelectrophoresis,
rocket electrophoresis, and Ouchterlony double diffusion.
More Importantly, once binding occurs, specific antigenic
sites appear to be blocked. This has been demonstrated by '
the differential response of rabbits to control and Varl-
■dase-treated specimens from a Wilms patient. Additional
; circumstantial evidence comes from the antisera prepared j
against EDTA (or phytic acid) extracts of tumor tissues by i
previous Investigators In our laboratory (K. Wise and A.
Kasparlan). In these previous studies, only a portion of
the total variety of proteins present In the extracts
elicited an Immunological response In rabbits. This has
been demonstrated by Immunoelectrophoresis studies Iden
tical to those shown above. Considering these findings. It
Is not surprising that the Immunological nature of Wilms i
tumor has been difficult to determine. Not only does
hyaluronic acid appear to bind and alter the physlo-chem-
Ical nature of some proteins, but, under certain conditions,:
this polysaccharide can act to generate selective Immune- ■
suppression through the reduction of antigenicity. i
I6l
Detection of Tumor-Associated Antigens:
The Gamma-Beta Antigen
In the course of searching for tumor-associated
j
antigens of human nephroblastoma, several primary Wilms
tumor specimens and nude mouse-passaged Lumoi's were used to
: develop antisera (for antigen detection). The specific
methodology Is described In the Methods section. Of these
antlseras, most were unreactlve with tumor cell components
I after appropriate absorption. Included In the "negative"
antisera were antl-TuWl cell extract, anti-nude mouse tumor,
antl-EPL cell extract, and antl-Wllms tumor (tissue frag
ments) . In one preparation against a peritoneal fluid
sample (from patient J.M.), however, antibodies were made
which were capable of detecting an "abnormal" protein pres
ent. This component migrated between the gamma and beta
regions of Immunoelectrophoresis, and was thus termed the
Gamma-Beta (GB) antigen. An extensive examination of EDTA |
extracts of Wilms tumors Indicated a positive reaction (by I
either Ouchterlony double diffusion or Immunoelectrophore- j
sis) In about 70^ of the extracts. Interestingly, the stud-)
les of a binding phenomenon between hyaluronic acid and j
serum proteins became useful at this point. After the "GB
negative" extracts were treated with Varidase (as Indicated
previously), all of these samples became positive for this
component. A preparation of Burtln’s "W" antigen was also
162 I
positive. An example of the GB antigen, as shown by immuno
electrophoresis, Is Illustrated In Figure 6l. '
Tests for GB antigen were also conducted on ex- i
tracts of normal human tissues (liver and kidney) as well
as a number of normal human sera from children (5 5) and
human plasma samples from children and adults (6l). These
experiments made It clear that the GB antigen as such Is
not abnormal, as all of the control plasmas and tissue ex
tracts were positive while the sera were negative. It ap
peared, then, that the GB antigen was a plasma protein In
corporated In the formation of a blood clot. The most
likely candidate for this protein appeared to be fibrinogen
(as this Is the major protein lost during formation of a
plasma clot). To Identify the GB antigen further, molecular
weight determinations were made, and the component was char
acterized by Isoelectric focusing; In addition, the sensi
tivity of this material to enzyme degradation was Investi
gated. The enzyme studies confirmed that the material (or
at least the antigenic determinants) was protein In nature.
Both pronase and chymotrypsin were found to destroy the
antigen, while hyaluronidase and nonspecific glycosldases
did not alter Its expression. Gel chromatography (with
Sephacryl S-200, Pharmacia, Sweden) was used to resolve two
peaks of antigenically Identical protein. One peak was
found at 300,000-500,000 daltons MW, while the other was
Demonstration of "blockage" of the
GB antigen
Figure 61. (Well 1: EDTA extract of Wilms tumor
Trough: Antisera to peritoneal fluid
absorbed with normal human
sera
Well 2: Sample In well 1 after Varl-
dase treatment)
163
164
r^iTànrria
F i g u r e 6 l
165
noted at about 75^000-100^000 MW. This is shown in Figure
62. It was this gel pattern that distinguished the
"normal” from the "malignant" patients' sera and tissue
extracts. In the samples from Wilms patients^ both the
high and the low-molecular weight components were found (in
plasma and peritoneal fluid samples as well as In tissue
extracts). With control normal plasmas and tissue ex
tracts, only the high-molecular weight component was found.
In addition, normal human sera lacked any detectable GB
antigen; the sera from the Wilms patients, however, con
tained both components. Antigenic material Isolated by
Sephacryl gel chromotography (both high- and low-mw frac
tions) was further characterized by Isoelectric focusing.
An example of such a test Is shown In Figure 6 3. This pro
file depicts the pH gradient (dotted line), the amphollnes
(dashed lines), and the sample material (continuous line).
From the graph. It appeared that only a single peak of
activity (measured precipitation by directed antisera) was
detected at pH 5.0-5.6. It was also noted from the figure
that the antigenic material was not a major component of
the focused sample from the gel run. From a combination of
the results by Immunoelectrophoresis, gel chromotography,
and Isoelectric focusing. It appeared that the "abnormal"
component was probably a fragment of a normal plasma pro
tein. The exact Identification of the GB antigen remains
Figure 62. Separation of GB antigen presence from
a peritoneal fluid sample (from patient
E.G. Antigen presence Is shown by [x];
IgM presence Is Indicated by closed
circles; and albumin Is shown by closed
triangles [determined by Ouchterlony
analysis]).
166
167
0.9n
0.8-
0.7-
0.6-
00 0.5-
CVJ
Q _
o 0.4-
0.3-
0.2-
o oo oo
50 40 30 20
FRACTION NUMBER
Figure 62
Figure 63. Isoelectric focusing of GB Antigen
(fractions from chromatographic sepa
ration were pooled. The dashed line
Indicates absorption of araphollne^
while the solid line Indicates sample
absorption^ crossed bars [+] indicate
antigen presence as determined by
Ouchterlony analysis).
168
169
0.8
0.7
0.6
0.5 o
o
ro
00
0.4 o
Q.
0.3 PH
5.0 6.0 7.0
0.2
30 40
FRACTION
20 50 60
Figure 63
170 I
lundetermined. Stilly there are three plasma proteins which:
have characteristics in common with the unknown antigens.
They are: ferritin (4^0000 mw_, pi 5.8 with an acidic com- :
ponent often associated with neoplasia) fibrinogen (340 j-
000 daltons MW with a neoantigenic fragment of 8 5^000 dal-
tons MW_, pi 5 . 8 and electrophoretic mobility in the gamma-
beta range) and finally,, plasminogen (8 7 ^ 0 0 0 daltons M¥^
pi 5 .8, , also associated with neoplasia through "plasminogen
activator"). Of the above plasma proteins,, fibrinogen of
fers the most likely (and most intriguing) candidate for
the identification of the GB antigen. Even though any of
the three proteins (as well as others) could account for I
the observed data,, only the fibrinogen has been previously
found to generate immunelogically unique fragments (frag
ment neoantigens "E" and "D" from the work of Plout et al.).^
In addition,, one of the reported fibrinogen fragments (d) |
of molecular weight 7 8 ,,0 0 0 daltons would coincide with the 1
smaller protein in the GB analysis. ;
The GB protein is less likely to be either ferritin j
or plasminogen. An acidic iso-ferritin has been associated,
with neoplasia including nephroblastoma (as a "beta onco
fetal antigen" or BOFA [Laurence and Neville]).However,,
the GB antigen failed to react with liver tissue homoge- ;
nates j even though the BOFA has been reported to react with|
such extracts. Plasminogen,, though,, is similar to the
1711
"smaller GB protein" (since both the "smaller" GB antigen ^
and plasminogen are about 80^000-90,000 daltons MW); un- ^
; I
fortunately, the larger molecular weight component of the ;
normal plasmas (300,000-500,000 daltons MW) could not be !
accounted for. Despite the chemical "fit" (or lack of it) i
between the GB antigen and the three Indicated plasma pro- |
teins, conclusive identification of the GB protein remains
to be determined. ■
I
Immunological Determinations: I
FLA and Other Antigens |
Studies of tumor-associated antigens of human and |
nude mouse-grown nephroblastomas have been performed in ;
order to help clarify the immunological nature of Wilms |
tumor. Unfortunately, the search for "tumor antigens" has
proved to be the most frustrating and the least rewarding
segment of this project. After several hundreds of hours’
work, the presence of such antigens remains somewhat un
clear. Still, some information concerning the "fetuin-
like antigen" (FLA, as described by Wise et al.)^^ and
other potential tumor antigens has been found.
Preliminary searches for FLA in extracts of tumor
tissues and cell cultures were based on analyses with
Spiro’s fetuin (GIBCO) and antisera to this bovine glyco
protein. In this early work, hundreds of experiments were
conducted with Ouchterlony and immunoelectrophoretic meth-
172
ods (as outlined In the Methods section) . The results of
these studies are as follows.
1. Spiro’s fetuin preparations contain at least three
distinct antigenic components (including bovine
serum albumin). Both rabbit and guinea pig anti
sera to commercial fetuin show reactivity toward
multiple antigens. The guinea pig serum used here
was obtained through the courtesy of Dr. Belerle.
An example of these findings Is shown In Figure 64.
A critical question arises from this result. That
is, if there is an antigen associated with Wilms
tumor that reacts with antisera to Spiro’s fetuin,
then which of the antigenic components (fetuin,
BSA, or others) is responsible for the cross-reac
tivity? This question remains unanswered, however,
since the anti-fetuin did not show a general reac
tivity with Wilms tumor tissues or extracts. An
exception was found with the Wilms tumor EDTA ex
tract pool of Wise (formerly of our laboratory).
This exception is described further below.
2. The production of FLA by cell cultures was studied
with EDTA extracts of both TuWi cells and a primary
cell culture of nephroblastoma (EPL), established
in our laboratory. In a preliminary experiment, an
FLA was found in extracts from both cell popula-
Figure 6 4 .
Demonstration of the heterogeneity
of fetuin
(Well 1: Fetal Calf Sera
Trough: Anti-Spiro’s Fetuin
Well 2: Spiro’s Fetuin)
Demonstration of FLA presence only
In cultures grown In the presence
of fetal calf serum
Figure 65. (Well 1.
Trough :
Well 2:
EDTA extracts of TuWl cells
grown In media with human
serum.
Anti-Spiro’s Fetuin
EDTA extract of TuWl cells
grown In media with fetal
calf sera .)
173
174
Flgure 64
6a'
Figure 65
175
tions. This finding verified the reports of Wise
15
et al. These experiments were then taken one
step further. The above cells were grown in media ,
with rabbit (or human) serum supplements, instead
of with fetal calf serum which contains high con
centrations of fetuin. In this way, cells were
grown in the absence of fetuin (as verified by
Ouchterlony double-diffusion analysis with anti-
fetuin). These latter cells were extracted in the
same manner as the cultures grown with fetal calf
serum. On examination (by Ouchterlony technique or
Immunoelectrophoresis), the extracts from the rab- ;
bit (or human) serum-grown cells failed to react
with anti-fetuin. From these studies, it seemed
that the FLA found in extracts of cells grown in
fetal calf serum was actually bovine fetuin which
had been bound to the cell surfaces. An example of >
the Immunoelectrophoresis patterns found with the
culture extracts is shown in Figure 6 5. '
!
3 . Rabbit and guinea pig antisera raised to Spiro’s !
fetuin were used to examine a number of Wilms tumor ;
tissues and sera samples. In one set of experi
ments, twelve EDTA extracts of Wilms tumor tissues
and one pool of five extracts were examined. In
addition, KCL extract and guanidine hydrochloride ;
176 I
extract from the tumor on one patient were studied,
as well as extracts and tissue homogenates from
nude mouse-grown tumors. With each of these tested
samples, no fetuin or fetuin-like material was de- i
tected by the Ouchterlony method (or alternatively
by Immunoelectrophoresis), even though bovine fetuin
was easily detectable. A second set of studies
with additional samples gave more promising results.|
Some positive reactions were found in this second
group between the anti-fetuin and the antigen prepa-,
ration of Kim Wise (this was a sample from the
original preparation of pooled EDTA tumor extracts |
that demonstrated the EDA); normal human kidney
(NHK) homogenates were also positive. Other posi
tive reactions were also found between the anti - i
fetuin and both an individual EDTA tumor extract
I
(patient H.P.) and a sample of normal human sera.
It should be noted, though, that these latter two
reactions were found only with antisera from guinea i
pigs; the reactions could also be inhibited by ab- |
i
sorbing the antisera with normal human sera before i
testing. The results of studies with the antigenic
material in the preparations of Wise and NHK ho
mogenates was, however, unaffected by this absorp
tion of the antisera. Prom these studies, it ap-
177
pears that some FLA component(s) are present In
both the extract pool of Wise and In the normal
human kidney homogenate.
A number of anti sera to nephroblastoma extracts and
tissues were prepared (according to the procedures In the
Materials and Methods section) in the course of these stud
ies. With each of these rabbit antisera, tests with numer
ous combinations of samples (from tissues, extracts, and
cultured cells) were conducted in search of "tumor-associ- ,
ated" antigens. In all of these tests, the antisera failed
to detect such specific antigens. The antisera and their j
!
reactivities are as follows: |
A. Anti-TuWi cell extract (EDTA): reactive only with i
normal human sera components (extracted cells had
been cultures in the presence of human sera). No !
reactions were found between this anti serum and j
pooled Wilms tumor extracts, the preparation of
Wise, Spiro’s fetuin, nude mouse-grown tumor ex
tracts, normal human kidney homogenates, or cul
tured cell extracts and residues (from EPL, MIX,
NUF, and TuWi cultures).
B. Anti-Wilms tumor EDTA pool (5 patients): reactive
only with normal human plasma components; included
reactivity with normal human sera proteins and with ,
the "GB" human plasma antigen. :
178
G. Anti-nude mouse tumor (a combination of homogenate
and EDTA extract): reactive with normal mouse
plasma proteins only. Unreactive with extracts and
cells from both human and nude mouse tumors, as
well as with cultured nephroblastoma cell prepara
tions. The lack of antibodies against normal human
plasma proteins suggests that the sera protein and
GB antigen are not generated by the tumor tissues
(probably these proteins are adsorbed to the tumor
cells, perhaps in conjunction with the cell surface
mucopolysaccharides).
D. Anti-EPL cell extract (EDTA): reactive only with
rabbit sera proteins (cultures were grown in media
supplemented with rabbit sera).
E. Anti-normal human kidney (homogenate): reactive
with normal human sera proteins and one human kid
ney-specific protein (by Ouchterlony analysis).
These antisera, once absorbed with normal human
sera, still failed to react with samples from
either the tumors or the cultured cells. This re
sult is surprising, since some reactivity against
tumor tissues might be expected. However, the po
tency of these antisera cannot be directly tested
(as the antigenic components are present in unknown
179 I
concentrations). As a result, the negative results
here may reflect "weak" antisera.
A final set of experiments was conducted in con
junction with the original antigen and anti sera prepara
tions of Wise. (it was found that after five years in
storage at -17*^ C, the antigens and antisera of Wise were
still quite reactive. For these studies, minute samples of
anti sera that remained frozen, including the WTpex, the
anti-WTpex, and the anti-V23 preparations, were utilized |
[and exhausted].) The results of these studies are shown
in Table 7. From the data, a number of interesting find
ings have been noted. For example, a "FLA" activity was ^
found (in the EDTA extracts of human and nude mouse tumors
as well as in homogenate of normal human kidney tissue) with
the antisera of Wise before, but not after, absorption with !
homogenates of normal human kidney. Also, absorption of ;
his antisera with fetuin caused some reduction of the reac-j
tivities with both normal human kidney and with the WTpex ;
tumor-associated antigen preparation. This latter finding |
suggests that the antigen of Wise was actually a component i
of normal human kidney tissue; this component appears to
have some, but not all, of its antigenic sites in common
with a protein in the Spiro's fetuin. i
It was also apparent that the preparation termed
"W" antigen in a Wilms tumor extract (a gift from Dr.
i8o
Table 7
Tests for Antigen Presence in Samples from
Normal Neo-plastic Sources
(Using the antisera of Wise)
3. NE2
h. n-Homo
5. Pool
6. Pf
7. #3
8. #1
9. EDTA
(H.P.)
1 0. KCl
11. KW
1 2. Burt.
1 3-
l4.
EPL
TuWi-
PCS
Antisera
Well Sample Oi Pet.
1. NHK
2. NEl
a GB
Pet.
15. Pel.
CX WTpex
ABS
NHP
Q( V23
ABS
NES
d V 2 3
ABS
NHP
d V 2 3
ABS
PET.
d V 2 3
ABS
NHK
+ +
(+)
+ + + +
+
+
+
+ + + + + +
+ + + +
+ + +
(+)
+ +
+
+ +
(+) (+)
Legend:
I8l
Sample s
NHK: Normal human kidney homog
enate
NEl: First EDTA extract of a
nude mouse tumor
NE2: Second EDTA extract of a
nude mouse tumor
n-Homo: Homogenate of a nude mouse
tumor
Pool: Pool of ( 3)EDTA extracts
of Wilms tumor
Pf: Peritoneal fLuid from a
Wilms patient
Fraction 3 of an equilib
rium
# 1: Fraction 1 of an equilib
rium
EDTA : EDTA extract of tumor from
(H.P.) patient E.G.
KCl: 3M KCl extract of a Wilms
tumor
KW : FLA preparation of Wise
Burt.: "W" antigen preparation
of Burtin
EPL: EDTA extract of EPL ceils
grown in rabbit sera
TuWi-: EDTA extract of pf TuWi cells
PCS grown in fetal calf sera.
Antisera
a Pet.: Rabbit anti-fetuin
d GP-Pet.: Guinea pig anti-
fetuin
d WTpex: Antisera to Wilms
tumor extracts
d V23: Antisera to FLA
(of Wise)
NHS: Normal human sera
NHP: Normal human plasma
PET : Spiro's fetuin
NHK: Normal human kidney
homogenate
182
i
Burtln) contained detectlble quantities of the GB antigen
(in immunoelectrophoresis) as did the WTpex antigenic prep
aration, the peritoneal fluid of a Wilms patient (J.M.),
the nude tumor EDTA extract, and the KCL extract of a Wilms I
tumor (prepared by Dr. Allerton). Further, human kidney-
specific antigens were noted in the following samples:
normal human kidney homogenate, nude mouse tumor EDTA ex- i
tract, a Wilms tumor EDTA extract (from patient H.P.), a j
I
KCL Wilms tumor extract, and in the WTpex extract pool of '
Wise. All of these reactions are illustrated in Figures
66-68. I
From the experiments with the antisera of Wise, it !
I
is clear that his preparations demonstrated more antigenic |
components than did any of my antisera. As I used a sim
ilar method, it is unlikely that the "methodology" was re
sponsible for the discrepancy found. Rather, it is more
i
probable that the original Wtpex (wilms tumor pooled ex
tract) preparation contained more of the antigenic compo
nents, and his inoculated animals thus gave a more potent
immunological response. The overall results with the above|
I
antisera of Wise in addition to the previous tests strongly'
suggest that all of the observed antigens are of either
human plasma or human kidney origin.
Figure 66 Examination of antisera to FLA absorbed with
normal human sera
(Center wells: antisera to FLA [Vgg]
absorbed with normal human sera)
Figure 6 7 Examination of antisera to FLA absorbed with
normal human plasma
(Center wells : antisera to FLA [Vgo]
absorbed with normal human plasma)
Figure 68
Legend :
Examination of antisera to fetuin
(Peripheral Wells [from top center, clockwise]
1. Normal human kidney homogenate
2. Nude mouse tumor, extract 1
3. Nude mouse tumor, extract 2
4. Minced nude mouse tumor tissue
5. Pooled EDTA extract of Wilms tumor
6. Peritoneal fluid from a Wilms patient
7. Fraction 3 from an equilibrium run [Kimmel
8. Fraction 1 from an equilibrium run [Kimmel
9. EDTA extract from patient [H.P.]
10. KCl extract [3M] from a Wilms tumor
11. FLA preparation of Wise
12. "W" antigen preparation of Burtin
1 3. EPL: EDTA extract [grown in rabbit sera]
14. TuWi: EDTA extract [grown in fetal calf
sera]
1 5. Fetuin [Spiro's])
183
Figure 66
Figure 68
184
2 ) / . , \
: O : O
O
Figure 67
CHAPTER IV
DISCUSSION j
This study has Involved the development and partial
characterization of nude mouse-grown nephroblastomas (and |
cultured cells from these tumors). Comparisons of these
tissues and cells with "native" Wilms tumor tissues and
!
commercial putative nephroblastoma cells have been com
pleted. The specific results of this work are reviewed '
j
here in the light of pertinent publications. Included are |
I
discussions of histological and morphological studies, en- I
zymatic analyses, immunological studies, and biochemical |
analysis. |
The histological patterns of nude mouse-grown tumorb
i
have been presented in the Results section. One key obser
vation was the tendency toward "de-differentiation" of the
tissues with multiple transfers from one animal to another.
This change consisted of a loss of kidney tubule formation
with a simultaneous increase in the ratio of epithelial-
like cells to fibroblastic interstitial cells. These
changes in morphology have interesting implications when
the mouse-passaged tumors are categorized according to the
classification of Lawler, Marsden, and Palmer^^ (for Wilms
185
186
tumor tissues). Lawler et al. compared the occurrence of
Wilms tumors in each of four histological categories (based
on the presence of kidney tubules In the tumors) with the
survival rate for the 75 patients Involved In the study.
Their findings Indicated a strong correlation between the
presence of tubules and the survival rate of the afflicted
children. Specifically, patients with no tubules In the
tumor tissues had a poor prognosis (with a 17^ three-year
survival), while those with many such formations had a good
prognosis (with an 8 3^ survival rate).
If the original Wllms tumor used here (patient
E.G.) is compared with the various nude mouse-grown deriva
tive tumors by the system of Lawler et al., a specific pat
tern is found. That is, the tumorous tissues tend to "de
differentiate" (as indicated by a loss of tubules) with
successive passages through the animals. In this case, the
original tumor would be classified as in Lawler et al.'s
system as a "++" tumor with moderate extent of tubule for
mation ; patients with tumors of this type had a fair prog
nosis (with a 46^ 3-year survival rate). The primary nude
mouse tumor #194 would also be classified in this category;
the tumor from mouse # 1 9 5: » however, would be considered
"+" with few tubules. Also, after the first transplanta
tion, the progeny from the tumor of mouse #194 had de
creased numbers of tubules (and would now be considered
187 I
By the time that the third group of animals (second
passage) developed tumors, no tubules were found. At this
point, the mouse tumors would be classified with tumors of
patients having a poor prognosis.
One could speculate from the above observations
that the malignant cells In the mouse-passaged tumors were
becoming more "potent" or "malignant" than the original
patient's tumor. Speculations could be carried even
further to suggest that the increase in the ratio of epi
thelial-like cells to interstitial cells (in the second
passage animal tumors) represented an expression of this
increased "malignancy." Of course. It can be argued that j
the nude mice provided an inadequate anatomical and molecu- ;
lar environment for the continued presence of the tubules
(due to the lack of essential human hormones or other spe- .
cific "inducers"). It could be argued also that the inter- |
stitial cells may be critical for tubule formation, and |
!
that with the loss of such cells, fewer tubules would be
formed. Still, there is additional circumstantial evidence |
for the theory of enhanced "malignant potential" (with in- j
creases in epithelial tissue regions and decreases in '
tubular and glomerular formations). According to Canale
and Mueeke,^ Wilms tumor métastasés have relatively few
tubules in addition to increased regions of undifferenti
ated metanephric cells. This has also been noticed in our '
I 188
2
Iown studies of metastatic Wilms tumor (Powars et al.).
I Currently^ the formation of métastasés Is viewed by many as
a process of selection In which only the most "potent" of
'the malignant cells survive In "ectopic sites" (Nlchol-
82
I son). It has also been found by a number of Investlga-
! tors that tumor cells become more "virulent" with passage
: either vltro or ln_ vivo. That Is^ fewer of the malig
nant cells are required for tumor development after each
I passage In a tissue culture or an animal system.
: The selection theory Is compatible with the finding
i
'of this study. Specifically^ a definite progression of
I tumor adaptation can be found In which the cell type which
I becomes Increasingly dominant Is also the one which grows
most rapidly and survives longest on vitro culture. For i
I I
j Wilms tumor j the key malignant cell appears to be the epl-
' thellal-llke cell. As such_, the observed human and mouse |
I tumors seem to be more like a carcinoma (having epithelial |
neoplastic cells) than like a sarcoma.
I Along these same lines,, Shlmosato^ Kameya,, and
I gh
Nagal established l4 serially transplantable human tumors
In a nude mouse system. Of these tumors^ only two showed
: alterations In tissue morphology upon repeated transplanta
tion. The tumor types Involved were both carcinomas (a
! gastric carcinoma and a breast carcinoma). These two spe-
:clflc tumors grown In the mice showed certain similarities
1891
with our animal-grown Wilms tumor. Included were the oh- |
servatlons that the initial human tumors contained glandu
lar elements and mixed tissue types. Also, In both the ;
studies of Shlmosato et al. and the present work, a de- I
crease In stromal elements was found with repeated passage !
: of the tumors. This evidence may be considered further
support for the contention that the Wilms tumor described
here was actually carcinomatous In nature.
Despite the arguments presented above, there are
alternate explanations for the observed morphological |
changes In the nude mouse tumors. One such alternate pos
sibility could Involve morphological reversion. Such a i
condition (normally found In vitro) Is denoted by an altera-!
tlon In the appearance of a cell line. An example of this j
6 5 I
activity has been Illustrated by Oklgakl with rat liver j
I
cells; In his experiments, he could reverslbly change the |
I
morphological appearance of the cells from an epithelial to I
a fibroblastic shape by alterations In culture conditions, j
: In the case of our mouse-passaged tumors, the Interstitial j
' I
and nephrogenic elements could have regressed to a common j
morphology. In this way, the cells could not be dlstln- j
gulshed from one another. Under such conditions. It would |
be easy to (erroneously) draw the conclusion that only one j
of the Initial populations remains (in a morphologically
homogeneous tissue or culture). This alternate explanation
I îî
190
could be supported by the contention that human nephroblas
tomas arise from metanephrlc blastema rests. These latter |
cells are believed to be stem cells which give rise to both I
the Interstitial and nephrogenic elements of the kidney.
On the basis of this theory, some Investigators have as
sumed that both the epithelial and the fibroblastic tissues
found In Wilms tumor are malignant In nature (since the
stem cells are thought to be malignant).
Ordinarily, the above alternate theory. Involving
regression" to a uniform morphology, would be Impossible
: to discount. However, the results obtained In this study
I
with ÔU1 vltro cultures of nephroblastoma cells tend to re-
|fute the Idea of reversion to a common morphology. As out-
; lined In the Results section, both the eplthellal-llke
(nephrogenic) and the fibroblastic Interstitial elements of
human Wilms tumor (and mouse tumors) have been cultured and
jIsolated. These cells have been characterized by morpho-
I logical studies, biochemical analysis, differential staln-
,Ing, karyotyping, growth soft agar, and studies of enzy
matic patterns.
In the course of this work, the various cell types
were maintained In tissue culture for a number of months.
I During that time, each of the cell populations retained Its
original cell morphology. This was particularly notable
:wlth the fibroblastic cells of both the original tumor and
191
of subsequent nude mouse-passaged tumors. These inter
stitial cells maintained a "typical" fibroblastic appear
ance (as compared to the normal human amniotic cell fibro
blasts) . In addition, differential staining techniques
revealed distinct similarities between the Isolated fibro
blastic cells and the Interstitial regions of the original
human tumor tissue. Also, the fibroblasts maintained a
very low growth rate over a period of several months ; they
even failed to grow In soft agar (even though malignant
cells generally grew under these conditions). In contrast
to the fibroblasts, the eplthellal-llke cells demonstrated
a morphology and differential staining patterns equivalent
to those found In the nephrogenic regions of the original
tissue. These latter cells also had a short population
doubling time (28 hours) and had the ability to grow well
In soft agar. However, like the fibroblasts, the epithelial,
cells maintained their morphology throughout this study.
The results of these experiments with the nephroblastoma
cultures appear to be In conflict with the theory of morpho-i
I
logical reversion to a common appearance. However, the
findings with the cell cultures do support the Idea that
the eplthellal-llke cells are the malignant cells of this
particular tumor (from E.G.). In addition, the data also
support the belief that the enrichment of undifferentiated
nephrogenic regions In the later animal-passaged tumors
192
(and in human metastatic disease) is due to a selection and
subsequent enrichment of the malignant epithelial cells.
The above conclusions are consistent with the pres-;
ent data from the nude mouse-passaged tumors and vltro |
cell cultures. However, It Is Impossible to generalize the i
present findings to show a widespread "Wllms tumor phenom- :
enon." The present study has been based primarily on the
tissues from one nephroblastoma patient. As such, no di
rect extrapolations can be made to encompass all Wilms
tumors; still, valuable Information Is generated by a com- :
pari son of the present work with the findings of other lab-
I
oratories. In this regard, the nude mouse system has fared |
well, with many Wilms tumor characteristics being shared |
between the primary tumors and mouse-passaged derivations.
i
Other Interesting aspects described here Include a compari
son of the nude mouse system of Wise and Muller with the !
present mouse system.
The utilization of nude mice to develop tumors as
described In this work Is, In most respects, distinct from j
i
the methods and results of nude mouse use developed by Wise |
i i 2 !
and Muller with TuWl cells. The one common element found !
In these systems was the use of congenitally athymlc mice
for heterotransplantation. The differences between these
systems Include the type and form of Innoculum, the "lag
time" for tumor growth, and the histological pattern of
193 I
tissue (tumorous) development. With the studies described '
here, fresh human tumor tissue was used for the primary ;
tumor Induction In the mice,* the resulting growths required '
a five-month "lag period" before Initiation of rapid growth,!
and the resulting animal tumors contained both nephrogenic j
and Interstitial tissues In addition to abortive tubule
formations. These animal-grown tumors closely resembled ;
the native tumors. The tumors from the TuWl cells, how- !
ever, did not require a substantial "lag period"; these |
I
latter growths, furthermore, contained only a single tissue !
morphology of "uniform undifferentiated cells with frequent |
mitoses." The general growth and histological characterls- {
tics reported by Wise and Muller are, however, similar to |
the "tertiary tumors" of this study (the tumors of the thlrc^
group of animals). It seems, then, that there Is a trend |
i
In the animal tumors generated from the tissue (mixed cell) I
source to become less differentiated, with fewer Inter- |
stltlal elements and an enrichment of "sheets" of eplthe- |
llal-llke cells. Both of these nude mouse systems do have
I
one critical element In common: the resulting tumor mass Is I
composed of cells which are more eplthellal-llke than flbro-|
blastlc In appearance. I
Another element of the work of Wise and Muller dealt
with gamma glutamyltranspeptldase (G-GTP) activity. It was
the contention of these authors (along with Alpert et al.) ■
194
that this enzyme may be a logical marker for nephroblastoma;
there are some difficulties with this viewpoint, though.
Specifically, the GGTP activity is not confined to kidney
tissue. It Is rather widely distributed In normal tissue
(especially In liver and kidney) and In neoplastic diseases
(Nodogrondsky, Tate, and Melster)For the determination
of tissues of origin, specific Isozymes must be Isolated
and quantitated by electrophoretic methods. The study of
GGTP could, however, be a valuable marker for cell extrac
tion and membrane purification methods. From our studies
with kidney cultures. It appears that this cell surface
enzyme remains with the "cell residue" fraction after EDTA
extraction. As a result, GGTP activity could be monitored
to establish a criterion of cell Integrity.
According to the data presented here, glucose-6-
phosphatase activity may be a more reliable "tumor marker"
I than GGTP. From the work of Pretlow et al. , GÔPtase
'levels have been found particularly elevated In proximal
I tubule cells of the kidney. Interestingly, this enzyme
,activity has been associated with EDTA extracts of both
nude mouse-passaged tumors and eplthellal-llke cells from
these tumors or the original human primary growth. As with
'the GGTP activity, the GôPtase could serve as a cell
I marker. In this case, though, the enzyme Is associated
: with the soluble EDTA fraction, rather than with the cell
195
residue. This is consistent with the location of this en
zyme In tubular cells having brush borders (Ugolev and De
j g Y
jLaey). As with GGTP, glucose-6-phosphatase Is a common
! enzyme ; but In the case of GôPtase, normal human kidney
'homogenates (or extracts) have little activity. As a re
sult, an Increase of this enzyme (perhaps In the sera of
' Wilms patients) could be correlated with progress of the
jdisease. This particular approach has not been pursued
'here, but such experiments would be relatively easy to
! conduct. I
I The findings reported here also suggest that nephro-i
jblastoma cells have EDTA extractable hyaluronldase and |
; protease activities. The hyaluronldase activity Is most '
! !
prominent In the cell cultures, but some hyaluronldase Is I
I
I also found associated with EDTA extracts of human and nude
I
'mouse tumors. This observation could explain the nature of
jpolysaccharide removal from the tumor tissue and subse-
Iquent transport of the glucosamlnoglycans In the sera of
patients. That Is, cell surface polysaccharides could be
partially digested In such a manner to release them from
the cells. The "free" polysaccharide could then find Its
way Into the circulation. Such a theory fits well with
previous reports of circulating "mucins" In three Wilms
patients (Powars et al., from our laboratory), with the
tumor EDTA extracts having a higher sedimentation coeffl-
196 ^
oient than their serum "counterparts." Such a presence of
"extractable" hyaluronldase could also help to explain the
elevated serum polysaccharides found In various neoplastic |
conditions (Cameron).It Is also Interesting to note |
that the "ground substance" of basal lamina that separates
epithelial from vascular mesenchymal tissues Is essentially
composed of a "fabric" made up of elastln, collagen, and
mucopolysaccharides. Thus, In order for a tumor to express
malignant potential (spread), this material must be solu
bilized. This process probably Involves the action of both
:proteases and mucopolysaccharldases (i.e., hyaluronldase).
Even for unregulated growth In a confined region, cells re
quire the ability to "disassociate" themselves from neigh
boring cells. Again, a combination of polysaccharides and
proteins must be solubilized. And again, there Is a need
for extracellular mucopolysaccharldases and proteases.
As Indicated above, proteases may be associated
with various cellular types and conditions. Including neo
plastic development. In the case of nephroblastoma, this
thesis work suggests that extracellular proteases are pres
ent during tumor development. Specifically, cultured
nephroblastoma cells have significant protease activity In
EDTA extracts of these cells. A high protease activity was
also found In pooled EDTA extracts from five Wilms tumors.
In comparison, tissue homogenates, but not EDTA extracts.
197 :
of normal human kidney demonstrate relatively high protease
activity.
The extracellular protease activity revealed by
I these studies provides a possible explanation for the pres- |
ence of "abnormal" fragments of plasma proteins such as the
GB antigen, for which the parent protein may be fibrinogen.
If the GB protein Is related to clot-forming protein, then
the cell-associated protease could be of the group of tumor-
associated proteases known as "plasminogen activator" (Teng
and C h e n , Chen and Buchanan) . Additional work Is neces-i
sary to confirm such a contention. However, the results Î
reported here are consistent with a "plasminogen activator"
hypothesis. The significance of the presence of hydrolytic
enzymes, especially of protease and hyaluronldase. Is not ;
merely of academic Interest. One area of particular con
cern would be the possible modification of antigens and '
their Immunogenlclty.
As outlined In the Introduction, the Immunobiology
of Wilms tumor has been difficult to establish. The pres
ent work has attempted to clarify some of the tumor Immu
nology through the use of nude mouse-grown human nephro
blastomas. The results found In this study have not
demonstrated the presence of tumor-associated antigens In
preparations from either human or mouse-grown tumors. In
addition, the experiments described here have failed to
198 ;
show "tumor antigens" In cultured nephroblastoma cells.
Still, It Is not assumed that these "negative" results In
dicate a lack of such antigens In human nephroblastoma. On
the contrary, the data reviewed here demonstrate some of '
the difficulties encountered In the study of Wilms tumor
antigens.
In order to fully understand the significance of
the present work, the data must be analyzed In the light of
previous reports. A consistent conclusion of several re
cent publications has been that Wilms tissues have at least
one tumor-associated antigen (Wise et al. , Burtln and
20 QQ 71 72
Gendron, Waghe and Kumar,Numata, Linder, Kumar, :
Waghe, and Taylor,Edynak, Old, Vrana, and Lurdla^^). !
Another consistent finding has been that the presence of
these antigens Is difficult to demonstrate, with not all of
the tumors having detectable abnormal antigens. I believe ;
that a highly variable expression of antigenic material In
the tumors could well have contributed significantly to my
own findings. One obvious disadvantage to the present sys
tem was the utilization of a single tissue source. That
Is, only tissues from one Wilms patient (E.G.) were used
for the development of both nude mouse tumors and vitro
cell cultures. As a result, the productions of potent
antisera against the tumor tissues were dependent on ma
terial from a single Wilms patient. It Is possible that
199 ;
this particular tumor simply did not have any (or at least
very little) of the tumor antigens demonstrated by other
Investigators. In this case, the tumor would fit Into the
"negative tumors" category In each of the previous publica
tions. In order to avoid this disadvantage, a large number
of human tumors would have had to be grown In nude mice,
thereby allowing a more Intensive search for Wilms tumor
antigens. Unfortunately, the establishment of multiple
nude mouse colonies for the growth of Wilms tumors would
not have been realistic under the condition wherein this
work was completed. Specifically, only two fresh samples ;
of Wilms tissues were available during the research period.
Also, human nephroblastoma tissues have not been grown (to j
our knowledge) In nude mice prior to these studies. In ef
fect, then, this work has been a successful pilot study.
As for the "tumor antigen story," further work will be
necessary to conclusively define the presence and the occur
rence of specific antigenic substances. Still, It Is rea
sonable to compare specific findings In this work to prevl- ■
ous studies In order to speculate on the nature of Wilms I
tumor-associated "antigens." !
One of the most Interesting reports on the Immunol
ogy of Wilms tumor was that of Wise et al. , who found and :
described the "fetuln-llke antigen" (PLA). This work by
Wise and his associates (from our laboratory) Is Important '
I 200
I for two reasons. First, It describes a chemical similarity
{between a human neoplastic component and a fetal bovine
serum glycoprotein; from this one could speculate on the
developmental processes which these two systems could have
In common. Second, the demonstrated presence of a tumor-
associated protein Is Itself significant.
In the course of these present studies with the
mouse-grown nephroblastomas, some aspects of the PLA story
have been further Investigated. The finding of PLA In
nephroblastoma cell cultures (as reported by Wise et al.)
was partially verified. That Is, PLA was found In cultures
grown In the presence of fetal calf serum containing fetuln,
I
but It was not found In cultures grown In the absence of
fetuln (human or rabbit sera). The Implications here are
that the observed PLA was In fact bovine fetuln which had
become attached to the cell surfaces. This explanation,
I though, could not be Invoked for the presence of PLA In
I
i tissue extracts, since the tumors were not exposed to bovine
;fetuln or fetal calf serum.
Experiments with the original tumor-derived antigen
preparation of Wise et al. demonstrated a definite reaction
with rabbit antl-fetuln. Under these same conditions, a
pool of five EDTA extracts of native tumor tissue as well
as extracts and homogenates of nude mouse-grown tumors (and
; cell cultures) were negative; eight Individual EDTA tumor
201
(human) extracts were negative, and one Individual (H.P.)
EDTA extract was positive. More detailed Investigations
revealed several Interesting findings. One Is that anti-
sera to fetuln must be absorbed with normal human plasma,
as some of the antisera react with an alpha-2-proteln of
human plasma. Also, when anti-tumor serum Is absorbed with
fetuln, there Is a risk of nonspecific removal of directed
antibodies. This Is based on the finding that affinity
columns prepared with Spiro’s fetuln (GIBCO) linked by
cyanogen bromide groups to Sephadex beads (Pharmacia,
Sweden) will bind to at least three types of protein found
In rabbit sera; they Include albumin, gamma globulins, and
an unidentified protein of beta mobility.
Another finding from these experiments was that
Spiro's fetuln would only partially remove antibodies to
the tumor ELA (as noted by decreased Ouchterlony band In
tensity after fetuln absorption). In contrast, fetal calf
serum did not show this marked cross-reactlvlty with the
tumor component. Also, absorption with normal human kidney
removed all ELA reactivity. From this. It appears that
only some of the sites of the "tumor antigen" are shared
with bovine fetuln. There Is some correlation between the
results presented here and other previous reports. In the
studies of Waghe and Kumar,the absorption of their anti-
sera with fetal calf sera did not Influence their results.
202 :
Prom this, the authors speculated that their alpha-moblllty
protein was probably not related to PLA. However, It Is
possible (based on the present work) that the use of fetal
calf sera (Instead of purified Spiro's fetuln) could ac
count for the Ineffectiveness of the absorbant. The possi
bility that the antigens found by Wise were the same as
those described by Waghe and Kumar cannot be ruled out. !
20
Burtln and Gendron also described a "nephroblas- I
I
toma antigen" (called the "W" antigen). In their work, the i
"W" antigen was found In Wilms tumors, as well as In other
neoplastic diseases (including carcinomas of the breast, :
stomach, and kidney). Through the courtesy of Dr. Burtln, i
a sample of the "W" antigen preparation was sent to our '
laboratory. This material has since been studied by both
Dr. Wise and myself, with the same results. Burtln's prep
aration contained trace amounts of material reacting with |
antisera to PLA. In addition, experiments described here '
show that the "W" antigen preparations also contain the
"GB" antigen discussed previously. However, neither of the
rabbit antisera from our laboratory (the antl-PLA or the
anti-GB) generates Immune-electrophoretic patterns that are
In agreement with those of Burtln. As a result, both the
"GB" and "PLA" antigens may be present In Burtln’s prepara
tions as trace (contaminate) components. The "W" antigen
does, however, have some marked similarities to the Beta-
203 i
oncofetal antigen (BOFA, a type of "abnormal acidic ferri
tin" as reported by Buffe et al. ) . One of the similari
ties Is that both "W" and "BOBA" are found associated with
nephroblastoma (with 8 9^ of 27 tumors tested being positive
for BOBA). Another striking similarity Is that both of
these "tumor antigens" are also found associated with other i
neoplastic conditions (Including breast and stomach carci
nomas) . In the work presented here, an attempt has been
made to verify (or to deny) the presence of this ferritin- !
like antigen In human and nude mouse-transferred tumors, as i
i
well as to examine the possibility that BLA and BOBA (as ;
well as the "W" antigen) are all actually "abnormal ferrl- i
tin components." Unfortunately, I have been unable to show I
!
cross-reactlvlty between the BLA and ferrltin-contalnlng |
I
human liver preparations. Still, In order to conclusively j
show an association (or lack of It) between these antigens, |
specific antisera (directed against each of the antigens) |
I
must be tested with combinations of the tumor antigens. |
There are additional difficulties that have been |
found In the study of the Immunobiology of Wilms tumors. '
One Is that the polysaccharides seem to have the ability to ;
bind extracellular serum proteins. This has been exten- ;
slvely demonstrated for hyaluronic acid (a major component !
of nephroblastoma polysaccharides) and serum proteins. It ,
becomes difficult, then, to distinguish between components i
204
which are synthesized by the tumors and those which may
have become bound to them. Other problems can arise from
the high enzymatic activities found associated with the
external surface of the tumor cells. According to the pres
ent study, high levels of both protease and hyaluronldase
have been found. Prom these results. It becomes evident
that cell surface components (including tumor-associated
antigens) may become susceptible to enzymatic degradation.
It even becomes possible to form "neo-antigens" (as with
fibrinogen) through the Interactions of normal cell surface
(or serum) proteins with the proteases. New antigenic
sites could also be formed through Interactions of the pro
teins with substances such as surface polysaccharides.
Another finding of this study has been the Inhibi
tory effect of polysaccharides on the antigenicity (and
Immunogenlclty) of certain proteins. In this way, the bio
chemical and Immunological aspects of nephroblastoma are
found to be Interrelated In an unexpected manner. Prom the
74
work of Plgman, Grambling, and Holley, It has become
clear that hyaluronic acid can bind to serum albumin (and
possibly other serum proteins) to form electrophoretlcally
stable complexes. In those studies, bovine synovial fluid
(containing hyaluronic acid and serum albumin) was Ini
tially used; but Plgman and his associates found that puri
fied hyaluronic acid could form these complexes when added
I
I to bovine serum. A similar situation was found In the pres-i
|ent work with human peritoneal fluid (from patient E.G.),
|as well as with EDTA extracts of human Wilms tumor tissues. |
j
; Although the demonstrated electrophoretic abnormalities
jwere not totally surprising (in the light of Plgman's work),
they were certainly more dramatic with the zonal Immuno-
: electrophone tic methods used; they also Implicated clearly
the Involvement of several protein classes (in addition to
albumin) In the "binding phenomenon."
87
A second publication by Deutsch reported an ab-
' normal agglutination of white blood cells In certain pa-
!
tlents having reticulum cell sarcoma and neuroblastoma,
j This abnormality was tentatively traced to high levels of
serum hyaluronic acid (up to 1.5 g$). Deutsch further de
scribed serum components with unusual sedimentation pat-
I terns; these were attributed to the formation of complexes
I
'of hyaluronic acid with serum proteins. The findings of
'the present studies are quite similar to the work of
8 7 74
Deutsch and Plgman et al. Specifically, abnormal blnd-
I
Ing Interactions and complex formations were found between
polysaccharide components of tumor extracts or peritoneal
fluid and certain (serum) proteins. These Interactions
were easily demonstrated by Immunoelectrophoresis, Ouch-
1terlony double-diffusion, and column chromatography; they
were also eliminated by Varldase (a hyaluronldase) pre-
206
treatment. By themselves, these findings are Interesting,
but not particularly exciting to Investigators of Wilms |
tumor. However, an additional discovery was made that |
(once complexed) hyaluronic acid (and perhaps other poly
saccharides) could block or mask the antigenic sites of
certain proteins. Such a masking process was found In the
peritoneal fluid studied, as well as In EDTA extracts of
Wilms tumors. A specific case In point was with the "GB"
antigen; several EDTA extracts did not demonstrate the
presence of this antigen (by Ouchterlony or 1mmunoelectro-
phoretlc methods) until after hyaluronldase treatment.
I
Even more surprising was the discovery that, under appropri
ate conditions, this "masking" effect of hyaluronic acid
could even prevent Immunogenic responses (in rabbits)
against the proteins Involved In complex formations with
It. Such a condition has apparently not been previously
noted In the literature. From these experiments. It seems
that It would be relatively easy for hyaluronic acid to
block "tumor-associated antigens" as well as other proteins.
In this way, the Immune system of a tumor-bearing host
could be seriously Impaired In Its ability to respond to
foreign or "neo" antigens. This could explain the low
Incidence of antibodies to tumor cell surfaces In nephro
blastoma (Kumar et al. reported antibodies to tumor cell
components In only one of 45 Individuals tested). Of
207 '
course, this interpretation assumes that there are "tumor
antigens" for an Immune response to be generated against In
the first place. The above "masking phenomenon" could also
help to explain some of the difficulties found In determln-■
Ing the qualitative and quantitative nature of antigens
associated with Wilms tumor.
Although the "masking phenomenon" of hyaluronic
acid with proteins appears to be a new discovery, the asso
ciation between this mucopolysaccharide component and Wilms
tumor has been known for some time. From the preliminary
2g
work of Morse and Nussbaum, It was noted that hyaluronic
acid was present at elevated levels In the serum of a |
nephroblastoma patient. This material was preclpltable In !
% acetic acid, and was categorized as an acid mucopolysac
charide by moving boundary electrophoresis. A later publi
cation from members of our laboratory (Allerton et al.)
described the Isolation and characterization of polysac
charide substances (similar to that of Morse and Nussbaum)
from EDTA or trypsin extracts of Wilms tumor tissues.
Other evidence for an association between nephroblastoma
and cell surface glycosamlnoglycans has come from several
Investigative groups ( Toma si and Robert son, McManus,
Powars et al.,^ and Belerle et al.^^). The general con
clusion reached by these authors Is that nephroblastomas
produce large quantities of cell surface mucopolysaccharide
substances (primarily composed of hyaluronic acid).
, 208
Since the above associations have been established
for native human Wilms tumors. It was only reasonable to
'Investigate the possibility of similar findings with the
nude mouse-grown nephroblastomas. As Indicated In the Re
sults section, EDTA extracts and cell homogenates (after
EDTA extraction) were examined for the presence of muco
polysaccharides by specific precipitation, biochemical
analysis, and differential staining techniques.
The results of precipitation of tumor extract sam
ples by specific solvents Indicate striking similarities
between the native human tumors and the nude mouse-grown
; tumors. The results which were most supportive of previous
studies were the precipitation of mucopolysaccharide sub
stances from EDTA extracts (of native and mouse-transformed
tumors) by both acetic acid (3^) &nd cetylpyrldlnlum chlo
ride ( 0 .2 5-1 . 0 0 0^). The data obtained from other precipi
tation analyses (with ethanol, trichloracetic acid, and
ammonium sulfate) also suggested that the polysaccharide
sub stances found In EDTA extracts of the human tumors are
quite similar to those extracts of the nude mouse tumors.
Analyses of hexose, hexosamlne, and uronlc acid
contents were also consistent for the human and nude mouse-
transferred tumors. Unfortunately, the cultured nephro
blastoma cells did not have the same precipitation charac
teristics or the same biochemical profiles as the tumor
209 !
i
!
tissues. Specifically, the extracts and homogenates of the
cultured cells did not form precipitates with either acetic
acid or with cetylpyrldlnlum chloride. In addition, the :
!
Icultures contained a significant Increase In the ratio of ■
hexose to hexosamlne, as well as an Increase of hexose com
pared to uronlc acid. These findings seem to suggest that ;
the cell cultures are not synthesizing large quantities of
extractable cell surface hyaluronic acid; alternately, the '
cultured cells may have a higher turnover of the polysac- i
: charldes (resulting In less material being present for
extraction). j
Some results In conflict with the above data have
also been found In experiments with differentially stained
cell cultures. As described previously, both epithelial
I
and fibroblastic cell cultures have been stained with |
Toluldlne Blue and Ruthenium Red. Comparisons have also j
been made between these stained specimens and comparably |
stained frozen sections from the original Wilms tumor tis
sue utilized In these experiments (from patient E.G.). The |
j
results of the staining experiments suggest that mucopoly- {
saccharide substances (sensitive to removal by Varldase |
treatment) are present on the surfaces of epithelial cells I
from both tissue culture and frozen tissue sections. These ;
7 6 '
findings are consistent with the studies of McManus, who '
demonstrated the presence of polysaccharides In the tubular '
210
regions of Wilms tumor tissues. The above results are also
consistent with hlstochemlcal studies of other epithelial
cell systems (as In the work of SI sea, Langkamp, and Thon-
77
ard with human amniotic cells, and the work of Thonard
and Scherp^^^ with human gingival epithelium). This ap
parent conflict between results of staining and biochemical
tests may reflect differences of detectability. That Is,
for precipitation (or for colorimetric determination) rela
tively high concentrations of polysaccharide are necessary.
However, histological techniques are more sensitive, and
less material Is required. As a result, the nephroblastoma
cells may be producing some polysaccharide material, but at
Insufficient levels for detection In biochemical assays.
From these findings, one can speculate on the possible
roles that the polysaccharide substances may play In the
development and survival of the tumors.
There are actually several reports which lend sup
port to the contention that the nephroblastoma cell surface
polysaccharides could play a role In cell proliferation,
pathogenesis, and avoidance of an Immune response. A clas
sical study on cell surface polysaccharides was conducted
78
with pathogenic bacteria. Specifically, Krause found
that the most virulent strains of Streptococcus pneumoniae
and Streptococcus hemolytlca (Group A) were covered by a
"capsule" composed primarily of hyaluronic acid. It was
211
: I
I
noted In these studies that the encapsulated bacteria did :
not elicit a strong Immune response, as did the non-encap-
sulated bacteria. Perhaps the Wilms tumor cells represent !
a malignant cell parallel to the bacterial condition. That {
Is, the tumor cells are able to "protect themselves" from i
an Immune response by making a mucopolysaccharide "coat." |
This could easily also be tied to the "masking" of :
antigens as discussed previously. In this way, the tumor -
: I
: cells might be "insulated" from cell-cell contacts with j
each other as well as with other tissues, possibly leading |
to continued proliferation. Also, tumor-associated anti- I
gens released by necrotic (or secretory) tissues could be i
i
"masked" by Interactions with solubilized hyaluronic acid.
Such Interactions would be consistent with the findings of |
high levels of hydrolytic enzymes In the EDTA extracts of |
I
both human and nude mouse-grown tumors. A further consist- |
ency with the bacterial analogy Is the presence of both
extracellular hyaluronldase and protease. In the bacterial |
systems discussed above, the cells produce high levels of I
I
I
both these two enzymes during rapid growth. Many mlcrobl- |
ologlsts have speculated that these enzymes are used to I
disaggregate host tissues during a bacterial Infection. As I
discussed previously, these two enzyme activities are also :
associated with human or mouse-grown nephroblastoma tissues.!
The presence of these enzymes could play a dual |
212
role. First Is that protease digestion of cell surface com
ponents (and. In some cases, hyaluronldase action) Is lmpll-4
cated as a necessary precondition for cellular division.
In fact, such enzymes are used to "subculture" cells In
: vitro. Second, such enzymes are also necessary for the
"invasion" of tumor cells Into foreign tissues (for the
solubilization of tissues and extracellular matrices or
: I
basal lamlnas). In this way, cells with high levels of
I
Iextracellular hydrolytic enzymes could be stimulated to con-
i
tlnual growth and simultaneous Invasion. An experiment by
Belerle et al.^ (in our laboratory) may have Indirectly
i
{supported these contentions. In experiments with human em-
!
Ibryonlc cells, the addition of EDTA extracts of Wilms tumor
tissues caused marked growth stimulation of these cells.
Thls effect was suspected to be caused by the presence of
Ithe tumor polysaccharides. However, the present studies
have shown that similar EDTA extracts also contain high
ilevels of protease and lyaluronldase. As these enzymes
ihave been Implicated In the processes leading to cell
division, perhaps the enhancement of cell growth was due to
the enzymes as well as to the presence of polysaccharides.
In summary, the nude mouse tumors appear to have
EDTA-extractable polysaccharides that are similar to the
icomponents found associated with the native human tumors.
;ln addition, the production of these polysaccharides (espe-
213
dally hyaluronic acid) appears to be primarily by the
epithelial-like tumor cells. The physiological role of
these substances remains uncertain. Still, there are some
Interesting similarities between the tumor cell surface
polysaccharides and surface carbohydrates of other patho
genic cells. Included In the role of cell surface-associ
ated components may be uncontrolled proliferation and
avoidance of Immune "detection."
CHAPTER V
SUMMARY
Nude mice have been used to develop a new Wilms
tumor (human nephroblastoma) model system. This method
utilizes congenitally athymic mice as human tumor carriers.
Tissues for this project were derived from the second tu
morous kidney of a seven-year-old bilateral Wilms patient.
The Induction of tumor growth required subcutaneous Innocu-
latlons of freshly minced tissues, followed by Incubation
periods of over six months. Once established In the mice,
the primary outgrowths were removed and serially trans
planted to other animals. To date, the tumors have been
successfully transplanted twice. Tissue specimens removed |
!
from each growth have been compared with the original human |
j
neoplasm, as well as with other nephroblastomas. Methods :
of comparison Include: histology, karyotyping, Immunologl- |
cal screening. In vitro tissue culturing, and biochemical i
analyses on chemical extracts. I
I
Histological studies have Indicated marked slmllarl-î
ties with native Wilms tumors. Especially striking Is the |
presence of both nephrogenic and stromatogenlc tissues,
with Interspersed abortive kidney tubule formations. Karyo-:
214
215
typing analyses have shown human chromosomes In the animal-
grown tumors, with an abnormal marker chromosome In some of
these tissues. Also, two morphologically distinct cell
I lines have been established from the nude mouse growths,
as well as from the native tumors. One of the lines Is
epithelial-like, while the other Is fibroblastic. Our
studies Indicate that the growth of human nephroblastomas
In nude mice Is feasible, and may be useful In the study of
the tumor development, as well as In the testing of chemo
therapeutic agents.
Analyses for growth characteristics, karyotyping
profiles, enzymatic activities, and staining patterns for
tissues and cell cultures have been completed. These data
taken collectively suggest that the epithelial-like cells
(presumptive nephrogenic cells) of both the native and the
nude mouse tumors are actually the neoplastic cells of
these malignancies. Additional studies have shown that
carbohydrate components (especially hyaluronic acid) of the
tumors can bind and subsequently block (mask) certain pro
tein antigens. The Implications of these studies are that
hyaluronic acid and hyaluronldase are both Involved In Im
portant regulatory functions of cell proliferation as well
as In possible suppression of host Immunological responses.
BIBLIOGRAPHY
1. Canale, V. and Mueeke, E. (I97l) Journal of Clinical
Pediatrics 23^ 66
2. Powars, D., Allerton, S. E., Beierle, J. W. and
Butler, B. (1972) Cancer 29, 6
3. Potter, E. L. (1972) Normal and Abnormal Development
of the Kidney, p. 53^ Yearbook Medical Publishers,
Inc., New York
4. Rosenstock, G. and Bishop, H. (1974) Seminars In
Oncology, p. 97^ Academic Press, New York and
London.
5. Grossman, H. (1976) Radiology 121, 697
6. Prout, M. (1 9 7 7) Cancer 39, 62
7 . Pochedly, C. (1971) New York State Journal of Medi
cine 5j 196
8. Cescheckter, C. E. and Wldenhorn, H. (1934) American
Journal of Cancer 22, 620
9 . Willis, K. A. (1 9 6 8) The Borderland of Embryology and
Pathology, p. 921, Lutterworth and Co., London
10. Potter, E. L. (1972) Normal and Abnormal Development
of the Kidney, p. 96I Yearbook Medical Publishers^
Inc., New York
11. Knudson, A. (1976) Pediatric Research 10, 513
12. Glanglcomo, J., Penchansky, L. and Monteleone, A.
(1 9 7 5) The Journal of Pediatrics 86, 98
1 3. Brown, W., Subhash, P., Altman, D. and Hardin, H.
(1 9 7 2) Surgery 72, 756
14. Strom, T. A. (1957) Acta Pediatrics 46, 6 0 1
1 5. Wise, K., Allerton, S. E., Trump, G., Powars, D. and
Beierle, J. (1972) International Journal of Cancer,
16, 199
216
217;
16. Buffe, D., Rirnbaut, C. and Lenierle, J. ( 1970) Inter
national Journal of Cancer 5^ 85
1 7. Alpert, M. E., Uriel, J. and Nechaud, B. (1 9 6 8) New
England Journal of Medicine 2 7 8, 984
1 8. Burton, K. (1956) Biochemistry Journal 62, 315 '
1 9. Kumar, S. and Taylor, G. (1977) International Journalj
of Cancer 19; 351 |
20. BurtIn, P. and Gendron, M. (1973) Proceedings of the ;
National Academy of Sciences (USÂT) 70, 2051 i
21. Wise, K .a n d Muller, E. (1976) Expe r l m e n t1a 32, 294 ,
22. Pochedly, C. and Miller, D. (1976) Wilms Tumor, p. !
2 6 2, Wiley Biomedical Publishers, New York ;
2 3. Murphy, G., M lr a n d , E. and Staubltz, W. (1976)
Oncology 33; 154 ;
24. Shalet, M., Holder, T. and Walters, T. (1 9 6 7) Journal|
of Pediatrics 70, 6 1 5 |
2 5. Morse, B. and Nussbaum, M. (1 9 6 7) American Journal of|
Medicine 42, 996 |
2 6. Allerton, S., B e i e r l e , J., Powars, D. and Bavetta, L.;
(1 9 7 0) Cancer Research 30, 679 i
2 7. Jasmin, G. and Rlopelle, J. (1970) Cancer Research |
3 0, 3 2 1 !
2 8. Thomas, C., Wessel, W. and Cltoler, P. (1970) B e l t r.
P a t h . Bd. 145; 6 8 I
2 9. Pretlow. T., Jones, J. and Dow, S. (1974) American
Journal of Pathology 74, 275 Ï
1
3 0. Cordex, C. (I9 6 7) Journal of Histochemistry and Cyto-j
chemistry 22, 103?
3 1. H a t a e , Y. and Maklta, A. (1977) Gann 68, 59
3 2. S a to h , C., Duff, R. and Davidson, E. (1975) Proceed
ings of the National Academy of Sciences 7ÔI 5?
33. M l k l t a , A. and S h lm o jo , H. (1973) B lo c h e m lc a et Bio- '
physica Acta 304, 5 7 1
2181
3 4. Underhill, C. and Keller, J. (1976) Journal of Cellu-'
lar Physiology 8 9, 53 '
35. Roblln, R., Albert, S., Gelb, N. and Black, P. (1975)
Biochemistry l4, 347
3 6. Beierle, J., Allerton, S. and Bavetta, L. (1971) '
Expcrlencla 27; 435 :
3 7. Ohnlshl, T., Oshlma, E. and Ohsuka, M. (1975) Experi
mental Cellular Research 93; 136
3 8. Wertheln, P. and Voute, P. (1976) Journal of the Na- i
tlonal Cancer Institute 57; 701
39. Waghe, M. and Kumar, S. (1977) European Journal of
Cancer 3, 993
40. Nuss, D., Duursma, J. and Pudlfun, D. (1977) Journal
of Pediatric Surgery 12, 4
41. Rosseau-Merck, M., Lonbard, M., Nezelof, G. and
Mouly, H. (1 9 6 9) European Journal of Cancer 13; 163
42. Wise, K. and Muller, E. (1976) Experlentla 32, 294
4 3. Nelson-Rees, W. and Elandermeyer, R. (1975) Science
I9I; 96 I
44. Elandermeyer, R. Hawthorne, P. and Nelson, W. (1975) |
International Journal of Cancer 16, 74 1
4 5. Priestley, R. (1972) The Journal of Urology, Balti
more 107; 6 9 3 I
46. Olcott, C. (1 9 5 0) Cancer Research 10, 626 |
4 7. Babcock, V. and Southern, C. (1 9 6I) Cancer Research ;
21, 1 3 0 i
48. Hlrono, I., Laqueur, G. and Spatz, M. (I9 6 8) Journal :
of the National Cancer Institute 40, lOII
4 9. Javadpour, N. and Bush, I. (1972) The Journal of
Urology 107; 931
5 0. Bush, I. (1 9 7 2) The Journal of Urology 107; 934 ■
5 1. Mount, B., Thelmo, W. and Husk, M. (1974) The Journal,
of Urology 111, 738 '
219
52. Pox, R., Dlwan, B. and Meier, H. (1975) Journal of
the National Cancer Institute 54, 1439
53. Rutenherg, A. (1 9 6 9) Journal of Histochemistry and
Cytochemistry 17, 5 1 7
5 4. Habel, K, and Salzman, N. (I9 6 9) Fundamental Tech
niques In Virology, p. 223; Academic Press, New
York
5 5. Bailey, J. (I9 6 7) Techniques In Protein Chemistry,
p. 372; Elsinore, London
5 6. Bitter, T. and Muir, H. (1 9 6 2) Analytical Biochemis
try 4, 3 3 0
57. Wlnzler, R. (1955) Analysis of Biochemistry, p. 279;
Academic Press, New York
5 8. Blltx, R. (1 9 4 8) Acta Chemlca, 2, 467
59- Garen, A. and Leventhal, C. (1 9 6O) Blocheml et Blo-
physlca Acta 3 8, 470
6 0. Schlll, ¥. and Schumacher, G. (1973) Biology of the
Cervix, p. 179; Blandau Publishers, University of
Chicago Press
6 1. Kass, E. and Seastone, C. (1976) Journal of Experi
mental Medicine 79; 319
6 2. Wrlgley, C. (1971) Methods In. Enzymelogy 22, 197
6 3. Lawlen, ¥., Marsden, H. and Palmer, M. (1975) Cancer
3 6, 1 1 2 2
64. Shlmosato, Y ., Kameya, T. and Nagal, K. (1976) Jour
nal of the National Cancer Institute 5 6, I2 5I
6 5. Oklgakl, T. (1 9 7 7) Personal Communication
66. Novogrodsky, T., Tate, S. and Melster, A. (1976)
Proceedings of the National Academy of Sciences
73; 2414
6 7. Ugolev, A. and De Laey, P. (1973) Blochemlca et Blo-
physlca Acta, 300, 105
68. Cameron, H. (I9 6 6) Hyaluronldase and Cancer, p. 2 9I;
Pergamon Press, Los Angeles
220
69- Teng, N. and Chen, L. (1975) Proceedings of the Na
tional Academy of Sciences 72, 413
70. Chen, L. and Buchanan, J. (1974) Proceedings of the
National Academy of Sciences 72, 1132
71. Numata, T. (1974) Journal of the American Medical
AssociaiIon 229; 123
72. Binder, E. (1 9 6 9) International Journal of Cancer
4, 2 3 2
7 3. Edynak, E., Old, L., Vrana, M. and Lardla, M. (1972)
New England Journal of Medicine 286, II78
7 4. Plgman, W., Grambllng, E. and Holley, H. (1 9 6I)
Blochemlca et Blophyslca Acta 46, 100
7 5. Tomas1, T. and Robertson, W. (1 9 6 6) Journal of Clin
ical Investigation 45; IO80
7 6. McManus, J. (1 9 6 7) "Histochemistry of Renal Cell
Carcinoma," p. 297; Renal Dysplasia, King, Little
Brown Co.
7 7. Slsca, R., Langkamp, H. and Thonard, J. (1 9 7 1) Histo-
chetnle 2 7; 123
7 8. Krause, P. (1 9 6 3) Bacteriology Review 27; 369
7 9. Preut, M. (1 9 7 7) Cancer 39; 62 I
8 0. Lawrence, D. and Neville, H. (1972) British Journal
of Cancer 26, 335
8 1. Moorehead, P. (1973) "Karyology of Cells In Cul
ture," p. 723; Tissue Culture, Academic Press,
London
8 2. Nicholson, G. (1978) Personal Communication
8 3. Deutsch, H. (I9 6 7) Journal of Biological Chemistry
224, 7 6 7
84. Beierle, J. (1976) Protides, Biological Fluids, p.
211, at Brugge Belgium Press, Brugge
8 5. Bellanl, E., Gasparlnl, M. and Bonadonna, G. (1975)
European Journal of Cancer 11, 593
8 6. Bodman, J. (1959) Clinical Chemlca Acta 1, 103 i
8 7. CatsImpoolas, N. (1 9 7 6) New York Academy of Sciences |
Annals 209, 145 |
88. Chlpowskl, S., Lee, Y. and Roseman, S. (1973) Pro- i
ceedlngs of the National Academy of Sciences 70, 8 |
8 9. Gahmberg, C. and Hakamorl, S. (1974) Biochemical and ;
Biophysical Research Communications 59; 1 I
9 0. Hakomorl, S. (1970) Proceedings of the National Ac ad-{
erny of Sciences 673 1741 I
9 1. Klthler, K. and Poullk, M. ( 1972) Archives of Bio- |
chemistry and Biophysics 2 7 8, 505 |
9 2. Knudson, A. (1975) Cancer 35; 1022 |
93- Knudson, A. and Strong, L. (1975) Proceedings of the
National Academy of Sciences 72, 5116
9 4. Kumar, S., Waghe, M. and Steward, J. (1973) Journal |
of Pathology 111, II7
9 5. Nash, M. and Edelmann, C. (1973) Nephron. 11, 71 |
9 6. Pritchard, D. and Todd, C. (1976) Cancer Research
3 6, 4699
97. Satoh, C., Krelder, K. and Banks, J. (1975) Cancer
Research 35; 2532
9 8. Sullt, H. (1 9 7 6) International Journal of Cancer 17;
461
99. Zoff, D. and Ginsberg, W. (1977) Blochemla Pharmo-
loglca 2 6, 6 5 9
100. Murata, K. (1975) Clinical Chemla Acta 63; 157
101. Nooter, K. and Bentvlezen, P. (1975) Cancer Research
35; 11 7
102. Lemerle, Jean (1976) Cancer 37; 2557

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Casein receptors on human polymorphonuclear neutrophil leukocytes

PDF

Purification and characterization of leptinotarsin-d, a neurotoxin which stimulates the release of neurotransmitters from synaptosomes

PDF

Induction, extracellular matrix vesicles, and RNA: A study of epitheliai-mesenchymal interaction during late embryonic rabbit incisor development

PDF

Studies of two naturally occurring compounds which effect release of acetylcholine from synaptosomes

PDF

Chemical and immunological studies on a heterophile antigen extracted from bovine erythrocyte stroma

PDF

Mechanism of action of the glucocorticoid triamcinolone acetonide in the induction of cleft palate in A/Jax mice

PDF

Glucocorticoid action and properties of the glucocorticoid receptor in cultured melanoma cells

PDF

Brain and pituitary glucose-6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase: Effects of age and regulatory hormones

PDF

In vitro transcription of chromatin and reconstituted nucleoprotein complexes

PDF

Studies on the aerobic oxidation of fatty acids by bacteria

PDF

Opossum cell synchrony, chromosome sorting, and methylation analysis of an x-linked gene, phosphoglycerate kinase

PDF

Interactions of tightly-bound nonhistone chromosomal proteins with DNA in chromatin

PDF

Characteristics of the enzyme phosphoenolpyruvate carboxylase isolated from 'Chlorella vulgaris'

PDF

Assembly of DNA with histone and non-histone chromosomal proteins in vitro

PDF

Mediation of murine odontoblast differentiation: The role of homotypic epithelium, heterotypic epithelium, and the basal lamina

PDF

Effects of Streptozotocin-diabetes on collagen metabolism

PDF

Biology and gametogenesis of 'Listriolobus pelodes' (Echiura)

PDF

The role of microbes in the formation of the tubeworm fouling community in Fish Harbor, Los Angeles Harbor, California

PDF

Numerical analysis of ecological survey data

PDF

Construction of human linkage maps by sperm typing

Asset Metadata

Creator Delmage, John Michael (author)

Core Title Nude mouse grown human nephroblastomas

School Graduate School

Degree Doctor of Philosophy

Degree Program Cellular and Molecular Biology

Degree Conferral Date 1979-01

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

Tag health and environmental sciences,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c30-217759

Unique identifier UC11226223

Identifier DP23649.pdf (filename),usctheses-c30-217759 (legacy record id)

Legacy Identifier DP23649.pdf

Dmrecord 217759

Document Type Dissertation

Rights Delmage, John Michael

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