Download Distinct High-Performance Liquid

[CANCER RESEARCH 44, 3613-3619,
August 1984]
Distinct High-Performance
Liquid Chromatography Pattern of Transforming
Growth Factor Activity in Urine of Cancer Patients as Compared with
That of Normal Individuals
Edward S. Kimball,1 William H. Bohn, Karen D. Cockley, Thomas C. Warren, and Stephen A. Sherwin
Laboratory ol Molecular Immunoregulation,
Facility, Frederick, Maryland 21701
Biological Response Modifiers Program, Division of Cancer Treatment, National Cancer Institute-Frederick
ABSTRACT
Reverse-phase
high-performance
liquid Chromatography
(HPLC) performed on urine from cancer patients and normal
controls revealed the presence of seven chromatographically
distinct peaks of transforming growth factor (TGF) activity, as
measured by colony formation of normal rat kidney cells in soft
agar. Comparison of urines from normal donors and cancer
patients showed no differences in EGF (epidermal growth factor)dependent 0-TGF-like activity but did reveal distinct patterns of
EGF-related, EGF-independent o-TGF-like activity. All urine sam
Cancer Research
absence of EGF; and 0-TGFs, which do not bind to EGF recep
tors and must act in synergy with either EGF or a-TGFs to
promote anchorage-independent growth (1, 3, 20). Like EGF, aTGFs will catalyze phosphorylation of tyrosine moieties in the
EGF receptor (17, 28).
TGF activity has been identified in tissue culture supernatants
of virally and chemically transformed rodent cells (6,15,16) and
from human tumor cell lines (13,22,26). In addition, acid extracts
of both neoplastic and normal tissues have been shown to
contain TGF activity (3,19, 21, 24). Since TGFs are acid stable,
low-molecular-weight peptides, urine has been examined for the
presence of TGF activity. Sherwin ef a/. (23) have recently
reported elevated levels of M, 30,000 to 35,000 TGF activity in
the urine of most cancer patients, whereas this molecular weight
form was detected only infrequently, and at low levels, in normal
controls. Normal controls as well as cancer patients express a
M, 6,000 to 8,000 TGF.
In this report, we further characterize by using HPLC the EGFcompeting activity and soft-agar growth-promoting activity found
in acidified, concentrated human urine extracts. Reverse-phase
HPLC revealed the existence of at least 5 species of EGF-related
a-TGF-like activities and a smaller number of non-EGF related
0-TGF-like activities when cancer patients and normal controls
were examined, whereas gel filtration earlier had revealed only 2
different urinary TGFs (23, 27). In addition, we noted distinct
differences between normal donors and cancer patients when
compared with regard to the relative levels of the EGF-related
TGF activities, one of which was found to correlate with the high
M, 30,000 TGF observed previously (23, 27) in the urine of
cancer patients. Furthermore, we observed that the M, 30,000
tumor-associated TGF gradually underwent conversion to a M,
ples contained at least two chromatographically distinguishable
forms of EGF-dependent TGF activity, eluting from HPLC as
broad peaks with 30 and 43% acetonitrile. The remaining five
TGFs eluted as sharp peaks with 32, 34, 35, 37, and 38%
acetonitrile, demonstrated EGF-competing activity, and thus
were functionally related to EGF.,Two of the five EGF-related
TGFs were consistently elevated only in the urine of cancer
patients and eluted with 32% (TGFA) and 37% (TGFD)acetonitrile.
Two of the other EGF-related TGFs, eluting with 34% (TGFB)
and 35% (TGFC) acetonitrile, were commonly found in both
normals and cancer patients. The fifth EGF-related TGF, TGFE,
eluting with 38% acetonitrile, was found only in normal donor
specimens. TGFA corresponded to the unique M, 30,000 TGF
activity previously identified only in the urine of cancer patients.
These observations demonstrate that cancer patients produce
high levels of EGF-related TGF activities which can be readily
distinguished, using reverse-phase HPLC, from EGF-related
TGFs produced by normal individuals. Using a solid-phase com
petitive radioreceptor binding assay for EGF, we demonstrated
that quantitation of EGF-competing activity is as sensitive and
effective as the soft-agar colony formation assay for distinguish
ing HPLC profiles of urinary TGF from cancer patients versus 6,000 TGF.
that from normal individuals.
MATERIALS AND METHODS
INTRODUCTION
TGFs2 are peptides which reversibly promote anchorage-in
dependent growth and colony formation in semisolid medium of
nontransformed cells (6, 21). Two functional classes of TGFs
have been described (18, 20): EGF-related a-TGFs, which com
pete with EGF for binding to EGF receptors (6,18,26) and which
can promote growth of normal fibroblasts in soft agar in the
1To whom requests for reprints should be addressed, at Connective Tissue
Research, Division of Immunology, Ayerest Laboratories Research, Inc., CN 8000,
Princeton, NJ 08540.
2The abbreviations used are: TGF, transforming growth factor; EGF, epidermal
growth factor; HEGF, human epidermal growth factor from urogastrone; HPLC,
high-performance liquid Chromatography; MeCN, acetonitrile; NRK, normal rat
kidney.
Received October 17,1983; accepted May 1, 1984.
Urine Preparation. Morning void urines were collected from patients
with a variety of disseminated cancers and stored at 4°.The patients
ranged in age from 18 to 63 years, had received no anticancer therapy
for a period of at least 3 weeks, and had normal renal function, as judged
by serum creatinine and the absence of proteinuria. Urines were also
collected from normal laboratory workers ranging in age from 22 to 49
years and also consisted of morning voids. Normal female donors in this
group were nonpregnant. Urine specimens were acidified with glacial
acetic acid to a final concentration of 1 M, clarified by centrifugation at
20,000 x g for 30 min, and then filtered through a 0.45-,«mfilter (Nalge
Corp., Rochester, NY). Between 20 and 50 ml of acidified urine, contain
ing 10 mg of protein, were concentrated to less than 1 ml by ultrafiltration
on an Amicon XM-50 membrane (Amicon Corp., Danvers, MA) (M, 50,000
cutoff). The concentrate was dialyzed against 1 M acetic acid in the
ultrafiltration cell. This procedure generally yielded 200 to 500 fig of
protein in the retained portion and contained almost all of the EGF-
AUGUST1984
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1984 American Association for Cancer Research.
3613
E. S. Kimball et al.
competing activity and soft-agar growth-promoting
activity when com
pared to dilution curves for the unconcentrated urine and the Amicon
filtrate. The retained portion was then applied to a reverse-phase HPLC
column.
HPLC. A Waters Associates
(Milford, MA) CurpBondapak
reverse-
phase column (3.9 x 300 mm) was used on an Altex Model 324 HPLC
(Beckman Instruments, Palo Alto, CA). Gradient elution over a period of
2 hr at a flow rate of 1 ml/min at room temperature, from 0 to 60%
acetonitnle (Burdick and Jackson Laboratories, Inc., Hoffman-LaRoche,
Inc.) in 0.05% tnfluoroacetic acid (Pierce Chemical Co.), adjusted to pH
2.5 with ammonium hybroxide, was used to develop the chromatogram
after an initial isocratic elution for 5 min with 0.05% trifluoroacetic acid.
One-mi fractions were collected. Duplicate aliquots (100 ^l) were taken
for EGF competition assays, and SOO-plaliquots were used for soft-agar
colony assays, as described below.
Soft-Agar Growth Assay. Column fractions were tested singly for the
presence of factors capable of stimulating nontransformed NRK (clone
49F) fibroWasts, taken at passages 17 to 19, to grow as colonies in soft
agar as described previously (5, 6). A single-cell suspension of approxi
mately 10* NRK cells was mixed with the lyophiiized sample to be tested
and seeded in 2.0 ml of Dulbecco's modified Eagle's medium containing
forming activity that is also present in normal control urine were
detected. In order to further characterize such urinary TGF
activity and permit comparisons to TGFs secreted by human
tumor cell lines in vitro (13, 22,26), acid extracts of human urine
from normal controls and patients with disseminated neoplastic
disease were examined by reverse-phase HPLC. Representative
HPLC profiles of acidified urine extracts, from a normal 38-yearold male control and a 50-year-old male patient with renal cell
carcinoma, are shown in Chart 1. Four peaks of soft-agar growthpromoting activity (Chart 1, fop) which coeluted with EGF-competing activity (Chart 1, bottom) were seen in the cancer patient,
whereas the normal control contained only 3 such peaks. There
were noteworthy quantitative and qualitative differences be
tween the normal control and the cancer patient. The normal
control showed very little, if any, of the TGF activity (TGFA) which
250 H
10% calf serum (Grand Island Biological Co., Grand Island, NY) plus
0.3% agar (Difco Laboratories, Detroit, Ml) into 60-mm Costar Pétri
40
dishes containing a base layer of the same medium plus 0.5% agar.
Plates were refed with 2.0 ml of the same medium plus 0.3% agar at 7
days. Colonies which typically contained 8 to 20 cells (23) were scored
in 4 random tow-power fields at both 7 and 14 days. Two independent
readings were performed for each assay. NRK cells do not grow as
colonies in 0.3% agar except in the presence of soft-agar growthpromoting factors, nor do they form colonies in the presence of EGF at
concentrations as high as 10 ng/ml (6,23).
Solid-Phase EGF Competition Assay. To quanti tate EGF competition
in the HPLC eluates, a rapid solid-phase competitive binding assay suited
for screening a large number of samples was developed and is discussed
in detail elsewhere (12). In brief, receptor-rich A-431 (7) cell membrane
fragments in 100 nI Dulbecco's phosphate-buffered saline, pH 7.2, were
dried overnight at 37°onto 96-well polyvinyl chloride plates (Dynatech
Laboratories, Inc., Arlington, VA), 2.5 /¿gmembrane protein per well.
Receptor grade mouse EGF (Collaborative Research, Inc., Waltham,
MA), was further purified to homogeneity over HPLC as described (2,
12,14), radtolabeted with 125I(Amersham/Searie Corp., Chicago, IL) as
described (8), and used as the '"l-labeled binding component. HPLCpurified unlabeied mouse EGF was also used for deriving standard
inhibition curves. Binding was initiated by the addition of 100 ¡Abinding
buffer (6) [Dulbecco's minimum essential medium containing bovine
serum albumin (1 mg/ml) and 50 mw 2|bis(2-hydroxyethyl)amino|ethanesulfonic acid adjusted to pH 6.8], containing 125l-labeled EGF (4 ng/
ml) with or without an aliquot of the sample to be tested. After a 1-hr
incubation at 23°,the contents of the wells were aspirated, and the wells
were washed 4 to 5 times with binding buffer (150 ^l). Finally, the wells
were cut out of the plate with a hot wire, deposited into test tubes, and
the amount of 125Mabeled EGF specifically bound was determined. EGFcompeting activity was quantitated as ng equivalents of EGF by com
parison to a standard inhibition curve.
Other Growth Factors. Sarcoma growth factor from Moioney sarcoma
virus-transformed 3T3 cells (clone 3B11) was prepared and partially
purified according to the methods of DeLarco and Todaro (6). Chemically
homogeneous, recombinant HEGF was a generous gift from Chiron
Research Laboratories (Emeryville, CA).
RESULTS
Previous studies revealed that urine from cancer patients
contained a unique soft-agar colony-promoting activity that had
a molecular weight of approximately 30,000 (23,27). In addition,
elevated levels of a M, 6,000 to 8,000 fraction with colony-
3614
50
60
70
FRACTION
80
NUMBER
Chart 1. Comparison of soft-agar growth-promoting activity and EGF-competing activity in HPLC fractions. Portions retained by ultratiltration of 10 mg of
acidified urines were chromatographed as described in "Materials and Methods."
Top, soft-agar colonies containing 8 to 20 cells/colony in 4 low-power microscope
fields; bottom, EGF-competition represented as ng equivalents of EGF activity. •,
50-year-old male patient with renal cell carcinoma; O, 38-year-old male;
,
MeCN gradient. The tetters A to E designate each of the individual EGF-related
TGFs resolved by HPLC. Gradient elution over a period of 2 hr at a flow rate of 1
ml/min, at room temperature, from 0 to 60% MeCN (Burdick and Jackson Labo
ratories, Inc.) in 0.05% trifluoroacetic acid (Pierce Chemical Co.), adjusted to pH
2.5 with ammonium hydroxide, was used to develop the chromatogram after an
initial isocratic elution for 5 min with 0.05% trifluoroacetic acid. One-mi fractions
were collected.
CANCER RESEARCH
VOL. 44
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1984 American Association for Cancer Research.
HPLC of TGF in Human Urine
eluted at 32% MeCN, or of the activity (TGFD) which eluted at
37% MeCN, whereas these activities were observed at high
levels in the cancer patient urine. Furthermore, the activity des
ignated TGFE (38% MeCN) in the control urine appeared to be
below the limits of detection in the patient urine. The levels of
TGFB and TGFC had similar ranges in the urine of patients and
normal controls. The observation that all of the above soft-agar
colony-promoting activities were associated with EGF-competing
activity and formed small-to-moderate-sized colonies in the ab
sence of EGF conforms to the definition of a-TGFs, as previously
described (18, 20,26).
The foregoing suggested a potential difference between TGF
activity profiles of cancer patients and normal individuals with
respect to the relative amounts of TGFA, TGFD, and TGFE. To
determine whether such differences were consistent, we per
formed similar HPLC analyses on a series of urines from cancer
patients and normal controls. Table 1 compares levels of colonyforming and EGF-competing activity for the above-mentioned
TGFs in urine from 2 pairs of age- and sex-matched normal
donors and cancer patients. Soft-agar growth-promoting and
EGF-competing activity were concomitantly elevated for TGFA
and TGFD in cancer patients but depressed in normal controls,
whereas the converse was observed for TGFE. It may also be
noteworthy that although there were similar levels of EGFcompeting activity for TGFC from normal male and female con
trols, the levels of soft-agar growth-promoting activities showed
a wider distribution. Also, the male urines demonstrated more
TGFE than did female urines when tested by the radioreceptor
assay but less TGFE when tested in the soft-agar assay. This
may be due in part to additional factors which are present in
female urine, as observed by Twardzik et al. (27), which could
contribute to the colony-forming activity present in those frac
tions.
Because there are often large variations in the absolute level
of growth-promoting activity in individual urines, and because
the soft-agar bioassay often demonstrates day-to-day variation,
comparisons between specimens were facilitated by determining
the TGFA/TGFE ratio for the number of soft agar colonies and
for ng equivalents of EGF-competing activity recovered from
reverse-phase HPLC (Chart 2). The TGFA/TGFE ratio was <0.5
in all normal controls and >0.5 in all cancer patients for both
soft-agar colony promotion and EGF competition.
In order to determine the relationship of the above EGF-related
TGFs to the tumor-associated M, 30,000 TGF described previ
ously (23), high-molecular-weight TGF (M, 30,000) was isolated
by gel filtration from the urine of a patient with breast carcinoma
and then applied to reverse-phase HPLC. Chart 3 demonstrates
that the Mr 30,000 TGF activity eluted with the same retention
as TGFA (32% MeCN). Since normal control urine had been
shown previously by gel filtration to contain little or no M, 30,000
TGF (23) and also lacked significant levels of TGFA activity by
reverse-phase HPLC, we tentatively concluded that TGFA is
probably equivalent to high-molecular-weight (approximately, M,
30,000) tumor-associated growth factor in urine. However, in
order to further correlate the M, 30,000 TGF with TGFA, the
latter was isolated from reverse-phase HPLC and was promptly
rechromatographed over Bio-Gel P-30. The result of that exper
iment, shown in Chart 4, reveals the presence of 2 molecular
weight species, one of approximately M, 25,000 to 30,000 and
one of approximately M, 6,000. Both demonstrated EGF-com
peting activity and soft-agar colony-forming activity. It was there
fore of further interest to determine whether the M, 6,000 TGF
had arisen from the 30K TGF. Accordingly, a M, 30,000 TGF
pool was isolated by gel filtration and, following storage for 4
days at 4°in the column eluant (1 M acetic acid), was rechro
matographed
by gel filtration over Bio-Gel P-30. The results
Cancer Patients
7
*
Normal Controls O
u,
6
u_
p
H
5
LL
CD
4
i"
=
2-
EGr
COMPETITION
COLONY
NUMBER
Chart 2. Comparisons of ratios of TGFA/TGFE for normal controls and cancer
patients. Values are for numbers of soft-agar colonies and for ng equivalents of
EGF-competing activity.
Table1
Comparison of growth factor activity in urine from patients and normal subjects
Urines containing 10 mg of protein were concentrated to 500 n\ and applied to reverse-phase HPLC. Each
fraction was analyzed for stimulation and growth of NRK cells in soft agar and for EGF competition on
membranes of receptor-rich A-431 cells (see "Materials and Methods" for details). Equivalents of EGF in ng
were calculated by means of an EGF competition curve using HPLC-purified mouse submaxillary EGF as
standard. Values represent activity in peak fractions of the respective TGFs identified in Chart 1.
No. of colonies of NRK cells in soft agar/starting
protein analyzed
10 mg of urinary
Donor31-yr-old
mate with colon car
(88)580
(144)0
(11)388
cinoma36-yr-old
normal male
(28)
(66)
(0)
(50)
41-yr-old female with breast
930
(230)16
240
(40)380(80)TGFc270
220
(280)185
1100(1150)0
(10)244(220)
0
carcinoma
49-yr-old normal femaleTGFA300(155)"5
(25)TGFB110(22)154(22)
(56)TGFo275
(0)TGFE0
" Numbers in parentheses, ng equivalents of EGF-competing activity per liter of urine
AUGUST
1984
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1984 American Association for Cancer Research.
3615
E. S. Kimball et al.
40
R
o
LU
20 O
o
o
60
50
70
FRACTION
rately chromatographed under the same conditions as was the
urinary TGF (Chart 6). The tissue culture-derived sarcoma growth
factor activity eluted earlier in the gradient, approximately 26%
MeCN, and hence is less hydrophobic than is urine-derived TGF.
Also, as shown, the HEGF eluted with the same retention time
as did TGFc (see Chart 1) and we tentatively conclude that some,
if not all, of the EGF-competing activity in that region of the
chromatogram may be due to EGF.
Since substantial soft-agar colony-forming activity was con
sistently observed with TGFC>and because in our experience, at
doses «10 mg/ml, neither mouse EGF nor HEGF intrinsically
caused the formation of colonies by NRK cells, it seemed likely
that some of the activity might be due to the presence of nonEGF-related TGF. Therefore, in order to further characterize the
80
NUMBER
150-1
Chart 3. Reverse-phase HPLC of tumor-associated high-molecular-weight TGF.
Reverse-phase HPLC of M, 30,000 activity isolated by Bio-Gel P-30 chromatography, as described by Sherwin et al. (23). from an acid extract of urine from a female
patient with breast carcinoma. Soft-agar growth-promoting activity (•)
represented
as number of colonies containing 8 to 20 cells in 4 low-power microscope fields.
EGF-competing activity (O) represented as percentage of inhibition of binding of
18SI-EGF.
, MeCN gradient used, as described in "Materials and Methods."
120-
90-
36-1
30K
36
25K
14K
8
32-
60-1
o
28 Z
28-
30-
O
— 24-
I
•»u.
oUJ
V)
Ul 20-
20 -
8
16 o
g
12-
20
8
20
24
28
32
36
FRACTION
40
44
52
Chart 4. Rechromatography of TGFA by gel filtration. TGFA, isolated after HPLC,
was applied to a Bio-Gel P-30 column (2 x 40 cm) equilibrated in 1 M acetic acid
and run at a flow rate of 0.2 ml/min. Two-mi fractions were taken, lyophilized, and
then tested for colony-promoting activity (•)
and EGF-competing activity (O).
revealed that the Mr 30,000 TGF had been converted exclusively
to a M, 6,000 TGF (Chart 5). These results suggest that there is
a correlation between the M, 30,000 tumor-associated TGF and
TGFA and further suggests that the M, 6,000 urinary TGF is at
least in part derived from the M, 30,000 urinary TGF.
Two other issues which were of concern were how the elution
from HPLC of urinary TGF differed from that of TGF secreted in
vitro by tumor cells and which of the various urinary EGF-related
TGF reverse-phase fractions (TGFA-E) might actually contain
urogastrone (HEGF). Accordingly, a sarcoma growth factor prep
aration (6) and homogeneous recombinant HEGF were sepa-
3616
36
40
44
48
52
100-1
^
NUMBER
32
Chart 5. Rechromatography of M, 30,000 TGF by gel filtration. M, 30,000 TGF,
isolated by Bio-Gel P-30 chromatography, was rechromatographed using the same
gel filtration column and conditions as in Chart 4. Samples were tested only for
colony-promoting activity in soft agar.
90-
fc?
-4
28
FRACTIONNUMBER
12 t
O
I
8-
24
é
80-
m
u.
70-
7
60
in
0
son
O
Z
40-
Z
20-
o
10-ootÃ-50
60
70
80
FRACTION NUMBER
Charte. Reverse-phase HPLC of sarcoma growth factor (SGF) and HEGF.
Partially purified sarcoma growth factor and highly purified HEGF were individually
applied to C-18 reverse phase and chromatographed using the same conditions as
those used
, MeCN
for gradient
urine concentrates
used; |, elution
(seeposition
"Materials
of TGFA.
and Methods" and Chart 1).
CANCER
RESEARCH
VOL. 44
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1984 American Association for Cancer Research.
HPLC of TGF in Human Urine
urinary TGF activities separated by HPLC and to look for other
possible differences between cancer patient and normal urines,
assays were performed for non-EGF-related 0-TGF-like activity.
Accordingly, suboptimal amounts (2 ng) of HPLC-purified mouse
EGF was added to urinary HPLC fractions in a soft-agar colony
profiles between normal controls and cancer patients (not shown)
could be discerned.
assay. Results of a representative HPLC experiment with a urine
sample taken from a normal female control are in Chart 7. In the
absence of exogenous EGF, most colonies contained only 8 to
20 cells and coeluted with EGF-competing activity (a-TGF activ
ity). The addition of purified mouse EGF (2 ng/ml) to the cultures
enhanced the numbers of colonies for all fractions that contained
EGF-competing activity and increased the size of those colonies
to greater than 40 cells. Mouse EGF alone caused the formation
of only low numbers of small colonies (<6 cells). In addition,
when in the presence of added mouse EGF large colonies were
found to be promoted by certain HPLC fractions (approximately,
Fractions 65 and 91), eluting as broad, heterogeneous peaks
with 30 and 43% MeCN. These fractions had shown no EGF
receptor-binding activity and no colony-forming activity in the
absence of EGF supplements. Hence, these fractions probably
contain /3-TGF-like activity, as described previously (1, 18, 20).
In contrast to the distinctions in a-TGF-like profiles between
normal and cancer urine, no differences in /3-TGF-like activity
Human urine has previously been shown to be a source of
both HEGF (4, 9), which has limited soft-agar colony-promoting
activity (6,23) and EGF-related transforming growth factor activ
ity (23, 27). Gel filtration of acid extracts of urine showed that
normal donors expressed low-molecular-weight
EGF-related
o
LU
<
OC
2
C3
O
50
60
70
80
FRACTION NUMBER
90
;co
Chart 7. HPLC profilo of EGF-dependent and EGF-independent activity in urine
from a normal female. Top, soft-agar colonies per 3 mg urine protein. •,colonies
formed in the absence of EGF (EGF independent). Colony size was 8 to 20 cells/
colony. O, augmentation of EGF-independent colony formation activity and dem
onstration of EGF-dependent colony-forming activity. Soft-agar colonies per 3 mg
urine protein formed in the presence of 2 ng of supplemental mouse EGF. Colony
size was typically »40cells/colony. Bottom, EGF-competing activity (O) per 3 mg
of urine protein expressed as ng equivalents of EGF.
DISCUSSION
TGF activity (M, 6,000 to 8,000) and that the majority of cancer
patients demonstrated an additional high-molecular-weight EGFrelated TGF activity (M, 30,000 to 35,000) (23, 27). The present
study was undertaken to further analyze urinary TGF activity.
We now report that reverse-phase HPLC analyses of acid
extracts of urine from normal donors and cancer patients re
vealed the presence of a total of 5 EGF-related growth factors
with soft-agar colony-promoting activity and 2 non-EGF related
factors which had colony stimulating activity only in the presence
of added EGF. Of the 5 EGF-related activities observed, one
(TGFA) was elevated in the urine from cancer patients and
correlated with previously reported M, 30,000 to 35,000 tumorassociated TGF. A second (TGFD) occuring mainly in the urine
from cancer patients, and a third, (TGFE), found at high levels
only in normal control urines, were also identified.
The relative proportions of the various a-TGF-like activities
observed within each specimen tested were in general the same
when tested for either soft-agar growth-promoting activity or
EGF-competing activity. However, the EGF competition assay is
a rapid, 1-day radioreceptor assay, vis-Ã -vis the soft-agar bioassay which requires 7 to 14 days. Moreover, the solid-phase EGF
assay is as effective and sensitive as the soft-agar bioassay for
distinguishing the HPLC-separated TGFs. Thus, we were able
to demonstrate that Ultrafiltration, reverse-phase HPLC, and
solid-phase EGF-competition assay techniques could be com
bined to provide a rapid, convient procedure for monitoring EGFrelated growth factor activity expression in the urine. Using this
technology, we were able to discriminate between urinary EGFrelated TGFs produced by cancer patients and normal controls.
The soft-agar colony formation assay warrants further com
ment. The colonies formed by the NRK cells were reported
originally to consist of greater than 20 cells (5, 6) when sarcoma
growth factor was used as a colony promoter. Other investiga
tors report that purified EGF-related TGF (a-TGF) promotes
formation of only very small colonies (18, 20). Similarly, there
have been reports that EGF does not promote colonies (5, 6) or
promotes large colonies (11,18). These differences may, in part,
be due to differences in growth factor purity but may also reflect
qualitative differences in the colonies which resulted from the
different growth factors examined. Also, the NRK cell line has
an increased sensitivity to growth factors and heightened sus
ceptibility to spontaneous transformation when cells from higher
passages are used (11).
In the studies with human urine-derived growth factor activity,
NRK cells were taken at passages 17 to 19 for soft-agar assays.
Under these conditions, most of the colonies formed consisted
of approximately 20 cells after a 7- to 14-day incubation period.
Purified mouse EGF and HEGF caused the formation of few
colonies, if any, containing no greater than 6 cells (23) when the
EGF alone was used at concentrations as high as 10 ng/ml. In
AUGUST 1984
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1984 American Association for Cancer Research.
3617
E. S. Kimball et al.
the present study, an EGF supplement (2 ng/ml) was able to
enhance the size and number of colonies promoted by some
HPLC fractions that had EGF-competing activity and, in addition,
was able to promote colony formation by HPLC fractions which
otherwise had shown neither colony-forming nor EGF-competing
activity. This is a novel discovery, since EGF-dependent growthpromoting activity in urine has not been described. Because
there are reports that a-TGF colony-forming activity is not en
hanced by EGF supplements (18, 20), this result suggests that
suboptimal amounts of non-EGF-related TGF (/3-TGF) may be a
contaminant in those a-TGF fractions examined and is consistent
with the demonstration that 3 orders of magnitude less EGF than
a-TGF is required to augment the promotion of /3-TGF colony
formation (18,20). Further purification of the components in urine
with growth promoting activity is required to further resolve this
issue. Accordingly, we have purified recently a M, 6,000 EGFrelated TGF from melanoma patient urine which retained EGFcompeting activity and EGF-independent soft-agar growth-pro
moting activity, and which upon isoelectric focusing revealed a
single peak that was distinct from the isoelectric point of urogastrone.3
The presence of high levels of tumor-associated TGFA after
HPLC purification of urine from cancer patients is in agreement
with earlier observations by Sherwin ef a/. (23) of a unique M,
30,000 to 35,000 TGF activity found after gel filtration to be
predominantly in urine from cancer patients. Our results indicate
that the M, 30,000 TGF and TGFA are equivalent and that the M,
30,000 TGF wiill also convert to an active M, 6,000 TGF under
conditions which have been reported not to promote conversion
of high-molecular-weight HEGF (10). In addition, tumor-associ
ated TGFA did not elute on reverse phase at the same retention
as did HEGF. Thus, TGFA and M, 30,000 TGF are not HEGF,
although the possibility exists that they may be either precursors
or consist of altered forms of EGF.
Finally, we observed that these urinary TGFs appear to be
chemically distinct from tissue culture-derived TGF by displaying
increased hydrophobic character. If the urinary tumor-associated
TGF is, in fact, a tumor-derived substance and not the result of
a host response, its increased hydrophobicity over that of a
tissue-culture-derived TGF suggests that it may have undergone
chemical alteration in vivo. This was shown to be the case in
another study involving urine from tumor-bearing mice.4
The dramatically elevated levels of TGFA and TGF0, which
were found exclusively in the urine from cancer patients could
be the result of changes in normal TGF metabolism. Thus, the
high-molecular-weight tumor-associated TGF, corresponding to
TGFA, may be present in the urine from cancer patients because
it is not as efficiently metabolized as it is in normal controls. This
interpretation is based on analogy to the conversion of highmolecular-weight EGF to low-molecular-weight EGF (10, 25) and
is consistent with the demonstration that partially purified highmolecular-weight TGF (M, 30,000) will gradually convert to lowmolecular-weight TGF (Mr 6,000). We are examining currently
the chemistry of Mr 30,000 TGF conversion to M, 6,000 TGF.
Immunological and biochemical comparisons to HEGF have also
been undertaken to further resolve this aspect of our studies,
3M. K. Kim, T. C. Warren, and E. S. Kimball, manuscript in preparation.
4D R. Twardzik, S. Sherwin, E. S. Kimball, J. R. Ranchalis,and G. J. Todaro.
A comparison of growth factors functionally related to EGF in the urine of normal
and tumor bearing mice, submitted for publication.
3618
and we are in the process of completely characterizing these
various TGFs in order to understand the biochemical relation
ships which exist between them, to compare them to TGFs
produced in vitro, and to ultimately understand their biological
function. Finally, the observation that urine from cancer patients
contains TGF activities which, via reverse-phase HPLC and a
solid-phase EGF-competitive binding assay, can be distinguished
from TGF activity found in normal controls further supports the
possibility that TGF activity may be a clinically useful biology
marker for certain types of cancer.
REFERENCES
1. Anzano, M. A., Roberts, A. B., Meyers, C. A., Komoriya, A., Lamb, L. C.,
Smith, J. M., and Sporn, M. B. Synergistic interaction of two classes of
transforming growth factors from murine sarcoma cells. Cancer Res., 42:
4776-4778,1982.
2. Burgess, A. W.. Knesel.J., Sparrow, L. G , Nicola, N. A., and Nice, E. C. Two
forms of murine epidermal growth factor: rapid separation by using reversephase HPLC. Proc. Nati. Acad. Sci. USA, 79: 5753-5757,1982.
3. ChikJs,C. B., Proper, J. A., Tucker, R. F., and Moses, H. L. Serum contains a
platelet-derived transforming growth factor. Proc. Nati. Acad. Sci. USA, 79:
5312-5316,1982.
4. Cohen, S., and Carpenter, S. Human epidermal growth factor: isolation and
chemicaland biological properties. Proc. Nati. Acad. Sci. USA, 72:1317-1321,
1975.
5. Delarco, J. E.. and Todaro, G. J. Epitheloid and Iibroblastic rat kidney cell
clones: epidermal growth factor (EGF) receptors and the effect of mouse
sarcoma virus transformation. J. Cell. Phystol.,94:335-342,1978.
6. De Lateo. J. E., and Todaro, G. J. Growth factors from murine sarcoma virustransformed cells. Proc. Nati. Acad. Sci. USA, 75: 4001-4005,1978.
7. Fabricant, B. N., DeLarco, J. E., and Todaro, G. J. Nerve growth factor
receptors on human melanomacells in culture. Proc. Nati. Acad. Sci. USA, 74:
565-569,1977.
8. Greenwood, F. C., Hunter, W. M., and Glover, J. S. The preparation of 131Ilabeledhuman growth hormone at high specific radioactivity. Biochem. J., 69:
114-123,1963.
9. Gregory, H. Isolation and structure of urogastrone and its relationship to
epidermalgrowth factor. Nature (Lond.), 257: 325-327,1975.
10. Mirata, Y., and Orth, D. N. Conversion of high molecular weight human
epidermal growth factor (hEGF)/urogastrone (UG) to small molecular weight
hEGF/UG by mouse EGF-associated arginine esterase. J. Clin. Endocrinol.
Metab.,49. 481-483,1979.
11. Jetten, A. M., and Goldfarb, R. H. Action of epidermal growth factor and
retinoids on anchorage-dependent and -independent growth of nontransformed rat kidney cells. Cancer Res., 43:2094-2099,1983.
12. Kimball. E. S., and Warren, T. C. Epidemial growth of factor binding to
membranes and immobilized in microtiter wells and estimation of EGF-related
transforming growth factor activity. Biochem. Biophys Acta, 777: 82-88,
1984.
13. Marquardt, H., and Todaro, G. J. Human transforming growth factor. Produc
tion by a melanoma cell line, purification, and initial characterization. J. Biol.
Chem., 257: 5220-5225,1982.
14. Matrisian, L. M., Larsen, B. R., Finch,J. S., and Magun, B. Further purification
of epidermal growth factor by high-performanceliquid chromatography. Anal.
Biochem., 725: 339-351,1982.
15. Moses, H. L., Branum, E. L., Proper, J. A., and Robinson, R. A. Transforming
growth factor production by chemically transformed cells. Cancer Res., 47:
2842-2848,1981.
16. Ozanne, B., Fulton, R. J., and Kaplan, P. L. Kirsten murine sarcoma virus
transformed cell lines and a spontaneously transformed rat cell-line produce
transforming factors. J. Cell. Phystol., 705:163-180,1980.
17. Reynolds, F. H., Jr., Todaro, G. J., Fryling, C., and Stephenson, J. R. Human
transforming growth factors induce tyrosine phosphorylationof EGFreceptors.
Nature (Lond.), 292: 259-262,1981.
18. Roberts, A. B., Anzano, M. A., Lamb, L. C., Smith, J. M., Frolik, C. A.,
Marquardt, H., Todaro, G. J., and Sporn, M. B. Isolation from murine sarcoma
cells of novel transforming growth factors potentiated by EGF. Nature (Lond.),
295:417-419,1982.
19. Roberts, A. B., Anzano, M. A., Lamb, L. C., Smith, J. M., and Sporn, M. B.
New class of transforming growth factors potentiated by epidermal growth
factor: isolation from non-neoplastic tissues. Proc. Nati. Acad. Sci. USA 78:
5339-5343,1981.
20. Roberts, A. B., Frolik, C. A., Amzano, M. A., and Sporn, M. B. Transforming
growth factors from neoplastic and nonneoplastic tissues. Fed. Proc., 42:
2621-2626,1983.
21. Roberts, A. B., Lamb, L. C., Newton, D. L., Sporn, M. B., De Larco, J. E., and
Todaro, G. J. Transforming growth factors: isolation of polypeptides from
CANCER
RESEARCH
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1984 American Association for Cancer Research.
VOL. 44
HPLC of TGF in Human Urine
virally and chemically transformed cells by acid/ethanol extraction. Proc. Nat).
Acad. Sci. USA, 77:3494-3498,1980.
22. Sherwin, S. A., Minna, J. P., Gazdar, A., and Todaro, G. J. Expression of
epidermal and nerve growth factor receptors and soft agar growth factor
production by human lung cancer cells. Cancer Res., 41:3538-3542,1981.
23. Sherwin, S. A., Twardzik, D. R., Bonn, W. H., Cocktey, K. p., and Todaro, G.
J. High-molecular-weight transforming growth factor activity in the urine of
patients with disseminated cancer. Cancer Res., 43: 403-407,1983.
24. Sporn, M. B., Roberts, A. B., Shu«,J. H., Smith, J. M., Ward, J. M., and
Sodek, J. Polypeptide transforming growth factors isolated from bovine
sources and used for wound healing in vivo. Science (Wash. DC), 279:13291331,1983.
25. Taylor, J. M . Mitchell, W. M., and Cohen, S. Characterization of the high
molecular weight form of epidermal growth factor. J. Biol. Chem., 249: 31983203,1974.
26. Todaro, G. J., Fryling, C., and DeLarco, J. F. Transforming growth factors
produced by certain human tumor cells: polypeptides that interact with epidermal growth factor. Proc. Nati. Acad. Sci. USA, 77: 5258-5262,1980.
27. Twardzik, D. R., Sherwin, S. A., Ranchalis,J., and Todaro, G. J. Transforming
growth factors in the urine of normal, pregnant, and tumor-bearing humans. J.
Nati. Cancer Inst., 69: 793-798,1982.
28. Ushiro, H., and Cohen, S. Identification of phosphotyrosine as a product of
epidermal growth factor-activated protein kinase in A-431 cell membranes. J.
Bid. Chem., 255: 8363-8365,1980.
AUGUST 1984
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1984 American Association for Cancer Research.
3619
Distinct High-Performance Liquid Chromatography Pattern of
Transforming Growth Factor Activity in Urine of Cancer Patients
as Compared with That of Normal Individuals
Edward S. Kimball, William H. Bohn, Karen D. Cockley, et al.
Cancer Res 1984;44:3613-3619.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/44/8/3613
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 18, 2017. © 1984 American Association for Cancer Research.