Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
13 41536
Background of the Invention
Field of the Invention
This invention relates to biologically active polypeptides
and their production from natural or synthetic sources, oligopeptides
derived from said polypeptides, and compositions and methods for
usefully applying the biological activity associated with the
polypeptides and oligopeptides derived therefrom in the health
sciences field. More particularly, this invention is directed to a
general polypeptide structure which defines proteins having cell
growth promoting activity and a novel class of transforming growth
factor (TGF) polypeptides which have the property of reversibly
conferring the transformed phenotype on normal cells in vitro and
:
thus appear to be proximate effectors of the malignant phenotype. In
another particular aspect, this invention is directed to antigenic
oligopeptides derived from the TGF polypeptides and to antibodies
raised to the antigenic oligopeptides and the associated TGF
- 1 -
~341536
1 polypeptides which are useful in the detection and treatment of
malignancies. In a further particular aspect, this invention is
directed to a class of oligopeptides which have the ability to bind
to cellular growth factor receptors and thus to potentially interfere
with transformation of certain cell lines into a cancerous state. In
a still further particular aspect, this invention is directed to a
process for isolating TGF polypeptides in homogeneous form from both
transformed human and murine cell lines and body fluids and to the
homogeneous TGF polypeptides so obtained. Other particular aspects
of this invention are directed to compositions and methods for
detection and treatment of cancer and other proliferative diseases
and for cell growth associated treatment, for example, wound healing
and ulcer treatment. Furthermore. another aspect of this invention
is in the treatment and detection of bone-loss diseases, such as
osteoporosis and hypercalcemia-
A number of polypeptide hormone and hormone-like growth
factors have been found in tissue fluids and their relationship in
the control of normal cellular growth or mitosis has been
established. These mitogenic polypeptide growth factors include
insulin, insulin-like growth factors, platelet-derived growth factor.
nerve growth factor, fibroblast growth factor and epidermal growth
factor (EGF). At least some of these known growth factors have an
effect on the growth of transformed cells, however, on the basis of
in vitro tests, it appears that transformed cells require less of
these known growth factors for optimal growth and multiplication than
-2-
13 41536
i do normal cells. In particular, it has been shown, in experiments in
cell culture, that the addition of exogenous growth factors such as
insulin and EGF can cause normal cells to mimic certain changes In
- cellular properties that are analogous to transformation; however,
they are unable to produce all of the changes associated with the
transformed phenotype, e.g., see Sporn et al. (1981) The New Eng. J.
of Med. 15, pp. 878-880.
Recently, new types of polypeptide growth factors
dErsignated as transforming growth factors or TGFs have been found in
certain human and animal carcinoma and sarcoma cells which possess a
g reater complement of the properties apparently essential to
phenotypic transformation (Roberts et al. (1980) Proc. Natl. Acad.
Sc:i. USA 77, pp. 3494-3498 and Todaro et al. (1980) Proc. Nati. Acad.
Sci. USA 77, pp. 5258-5262). The TGF polypeptides as a class are
characterized by the changes which they cause when applied to
untransformed, non-neoplastic indicator cells growing in culture.
These changes include a) loss of density-dependent inhibition of cell
growth in monolayer culture, b) overgrowth of cells in monolayer
culture, c) change in cellular shape, with the result that the
iiidicator cells assume the neoplastic phenotype, and d) acquisition
o-F anchorage-independence, with the resultant ability to grow in soft
agar. The property of anchorage-independent growth of cells in
culture has a particularly high correlation with neoplastic growth in
vivo. At least certain of the TGF polypeptides show some
relationship with EGF in that they are both heat-stable, acid-stable
-3-
i341536
peptides sensitive to reducing agents and proteases and they appear
to specifically interact with, and produce biological effects
through, cellular membrane EGF receptors, TGF competing with EGF for
binding to the cellular EGF receptor. Nowever, TGF is
distinguishable from EGF in several important respects. In
particular, EGF does not induce anchorage-independent growth of cells
in culture nor do antibodies to EGF detect TGF in either
riidioimmunoassay or immunoprecipitation tests. Further, EGF has only
a slight effect on the phenotype of cultured cells, whereas TGF
produces a more pronounced phenotypic alteration in cultured cells
and confers on them the ability to behave as transformed cells.
Interestir:gly, the transformation produced by TGF is not permanent
biit reversible in the absence of TGF and there is no evidence that
Tt;F acts as a complete carcinogen itself. (Todaro et al. (1981)
J., of Supramolecular Structure and Cell Biochem. 15, pp. 287-301).
Thus, TGF polypeptides are a unique class of proteins
d-istinguishable from other growth factors such as EGF from the
standpoint of both biological properties and chemical structure.
These TGFs, in turn, possess a variety of properties of value or
potential value in the health sciences field including potent, but
reversible, cell growth or mitogenic properties which find use in
cell repair including, for example, wound healing and ulcer therapy.
Additionally, the production of TGF polypeptides, or elevated levels
of production, are characteristic of. if not essential to, the
morphologic transformation of certain cell lines in both human and
- 4 -
13 41536
1: murine tissue and/or fluids; therefore. the TGF polypeptides or
antigenic fragments thereof are of value in differentiating normal
cells from tumor cells and antibodies raised thereto have application
in both the diagnosis and treatment of malignancies. Further,
realization that certain TGF polypeptides specifically interact with
and produce their biological effects through cellular membrane EGF
receptors raises the possibility, once the basic TGF polypeptide
structure is determined, of correlating its structure with the
structure of EGF to develop oligopeptides having chemical
characteristics to allow binding to the EGF receptors without
concomitant phenotypic transformation of the celi: Oligopeptides
having this characteristic EGF receptor binding ability find
application in treatment of malignancies, since the oligopeptide will
interfere or compete with TGF for available receptor sites and
thereby interrupt the expression of the transformed properties of the
cell.
With the present invention, a method has been developed to
obtain TGF polypeptides in sufficient quantity and purity to allow
complete structure determination and as a result of this
c;etermination and other observations, including the finding of
substantial homology between human and murine TGFs, a basic peptide
structure has been discovered which has broad application in cell
mitosis and the cell growth related fie:ld of use. Further,
homogeneous TGF polypeptides are obtained having application in both
-5-
1341536
the cell growth field and in the detection and treatment of cancer
and other proliferative diseases.
TGF polypeptides, oligopeptides and antibodies raised to
these polypeptides also have application in the detection and.
treatment of bone-loss diseases, such as osteoporosis, hypercalcemia
and bone resorption.
Additionally, antigenic oligopeptides and oligopeptides
having the ability to bind to cellular growth factor receptors are
derived from the basic peptide structure and the determined TGF
polypeptide structures. Finally, compositions and methods, including
antibodies raised to the TGF polypeptides and antigenic oligopeptides
derived therefrom are provided for use in the health sciences field.
Description of the Prior Art
Marquardt and Todaro (1982) J. of Bio. Chem. Vol. 257,
No. 9, pp. 5220-5225, (published May 10, 1982) describe the isolation
of a low molecular weight human TGF from serum-free medium
conditioned by a human metastatic melanoma tumor line by a sequence
of process steps including extraction in 1M acetic acid and
sequential purification on reversed phase high pressure liquid
chromatography eluting first with acetonitrile solvent followed by
elution with 1-propanol to afford a purified TGF having
characteristic TGF biological activity, e.g., induction of anchorage-
independent cell growth.. Twardzik et al. (1982) Science 216, pp.
894-897 (published May 21, 1982) report the use of the same
purification methodology to purify a TGF from a virus transformed rat
-6-
1341536
1 cell. The biological activity of the purified materials in the cell
culture is also demonstrated. Pike et a1. (1982) J. of Bio. Chem.
Vol. 257, No. 24, pp. 14628-14631 (published December 25, 1982)
- disclose that both partially purified rat and human TGF have the
ability to activate a protein kinase in human tumor cell membranes
iind therefore to stimulate phosphorylation of a synthetic tyrosine-
containing peptide. The only other molecules so far described that
ttave this activity are EGF, insulin and platelet derived growth
factor, all of which are believed to have important physiologic
functions in man and animal. Other references of interest are cited '
in the aforementioned articles.
Summary of the Invention
A basic protein structure has now been found which defines
a new class of biologically active molecules. The finding of this
framework polypeptide affording biological activity, particularly in
the cell growth promotion area, is based on the discovery that a
definite correlation exists between the three dimensional structure
of certain polypeptides, including TGFs,.containing multiple
disulfide bonds and the biological activity attributable to the
polypeptide. Accordingly, in one of its broadest aspects, the
present invention is directed to biologically active polypeptides
containing at least one peptide sequence of the formula 1:
-Cys-(AA)a-Cys-(AA)b-Cys-(AA)c-Cys-AA-Cys-(AA)d-Cys- I
wherein AA is an amino acid residue selected from Val, Ser, His, Phe,
Asn, Lys, Asp, Thr, Gln, Arg, Leu, Glu, Pro, Ala, Gly, Trp and Tyr,
-
- 7
1341536
1 and a is 7. b is 4 or 5, c Is 10, and d is 8. Also contemplated are
compounds according to formula I wherein when b is 4, AA may also be
Ile or Met, in addition to the amino acid residues recited above.
Another aspect of the present invention is directed to
specific classes of polypeptides having transforming growth factor
prDperties which include compounds of the formulas II and IIA or
oligomers thereof:
5 10
Val -Val -Ser-Hi s-Phe-Asn-R-Cys-Pro-Asp-Ser-Hi s-Thr-
20 25
Gln-R'-Cys-Phe-His-Gly-Thr-Cys-Arg-R -Leu-Val-Gln-
30 35 II
10 Glu-Gtu-Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly-R " '-
.40 45 50
Val -Gly-Val -Arg-Cys-Glu-His-Ala-Asp=Leu-Leu-Ala
5 10
Val -Val -Ser-Hi s-Phe-Asn-R-Cys-Pro-Asp-Ser-His-Thr-
15 20 25
Girs-R'-Cys-Phe-His-Gly-Thr-Cys-Arg-R " -Leu-Val -Gl n-
30 35 IIA
Giu-Asp-Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly-R " '-
40 45 50
15 Val-Gly-Ala-Arg-Cys-Glu-His-Ala-Asp-Leu-Leu-Ala
wherein R is Asp or Lys, R' is Phe or Tyr, R'' is Ser or Phe, and
R"' is Phe or Tyr. Also within the scope of the invention are
antigenic oligopeptides derived from the polypeptides of formulas II
and IIA,
An additional aspect of this invention relates to
polypeptide growth factors containing one or more of the following
peptide fragments:
- 8 -
13 4 1 5'36
A. Val-Val-Ser-His-Phe-Asn-Lys-Cys-Pro-Asp-Ser-His-Thr-Gln-Tyr-Cys-
Phe-His-Gly-Thr-Cys
B. Val-Val-Ser-His-Phe-Asn-Asp-Cys-Pro-Asp-Ser-His-Thr-Gln-Phe-Cys-
Phe-His-Gly-Thr-Cys
C. Leu-Val-Gln-Glu-Glu-Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly
0. Cys-His-Ser-Gly-Tyr-Val-Gly-Val-Arg-Cys-Glu-His-Ala-Asp-Leu-Leu-
Ala
F. Leu-Val-G1n-G1u-Asp-Lys-Pro-Ala-Cys-Va1-Cys-His-Ser-Gly and:
F. Cys-His-Ser-Gly-Tyr-Val-Gly-Ala-Arg-Cys-Giu-His-Ala-Asp-Leu-Leu-
Ala.
A still further aspect of the present invention is directed
to a class of oligopeptides which have an ability to bind cellular
growth factor receptors. and therefore, have potential utility in
treatment of malignancy and other disorders of_cell proliferation.
lhis class of oligopeptides is based on the discovery of key
sequences in larger polypeptide molecules which exhibit both
significant amino acid sequence homology in the appropriate three
cimensional structure and have the ability to bind to cellular growth
r-eceptor sites. Accordingly, this aspect of the invention provides
cligopeptides having the formula III:
(AA)x-Cys-(AA)y-Gly-R-(AA)Z-Arg-Cys-(AA)z- III
wherein R is Phe or Tyr and AA is an amino acid residue selected from
Val, Asn, His, Ser, Ile, Gly, Leu, Asp, Cys, Thr, Ala, Tyr, Pro, Glu,
Gin, and Arg, and x is 0 or an integer of from 1 to 6, y is 2, z is
3, and z' is 0 or an integer of from 1 to 6.
-9-
13 415 36
Also contemplated by the invention are biologically active
compositions and methods using the polypeptide and oligopeptide
structures given above including antibodies to the polypeptides of
formulas II and IIA and the antigenic oligopeptides derived
therefrom, said antibodies being optionally labeled with a label
capable of providing a detectable signal for use in diagnostic
methods or labeled with a cytotoxic agent for use in cancer or
proliferative disease therapy. Other compositions and methods
utilizing the cell growth promoting properties of the polypeptides of
the present invention also form part of the present invention. '
Another aspect of the present invention includes a process
for isolating homogeneous transforming growth factor polypeptides
from less pure aqueous solutions containing said polypeptides,
including body fluids and aqueous mediums conditioned with
transforming growth factor-producing cell l-ines, as well as the
homogeneous transforming growth factor polypeptides produced thereby
and antibodies raised to said homogeneous polypeptides. In its
broadest aspects, the process of the invention involves isolation of
a homogeneous transforming growth factor polypeptide from an aqueous
medium containing said transforming growth factor polypeptide in
impure form by the process steps comprising:
1) dialyzing the aqueous medium containing the
transforming growth factor in impure form against
aqueous acetic acid to afford a solvent phase
- 10
-
~34153~
1 containing transforming growth factor polypeptide which
phase is concentrated and optionally clarified,
2) reconstituting the concentrated solvent phase of step
1) with aqueous acetic acid and subjecting the
reconstituted solution to gel permeation chromatography
by applying reconstituted solution to a gel permeation
chromatography column conditioned with aqueous acetic
acid and eluting with aqueous acetic acid to obtain
selected fractions of eluate containing transforming
growth factor polypeptide in an enhanced state of '.
purity, said selected fractions being combined and
concentrated, to afford a partially purified,
transforming growth factor polypeptide-containing
product,
3) subjecting the partially purified, transforming growth
factor polypeptide-containing product of step 2) to
sequential reverse phase high pressure chromatography
by passing said product, after reconstitution in
aqueous trifluoroacetic acid, through one or more
2C hydrocarbon bonded silica matrix columns, which have
been equilibrated with aqueous trifluoroacetic acid,
under high pressure liquid chromatography conditions,
the initial column elution being performed using a
linear acetonitrile gradient in aqueous trifluoroacetic
acid and the subsequent column elution, which is
- 11 -
13 41536
Carried out on the combined, transforming growth
factor polypeptide-containing fractions of the initial
high pressure chromatography step, being performed
using a linear 1-propanol gradient in aqueous
trifluoroacetic acid, said 1-propanol gradient being
increased in sufficiently small 1-propanol
concentration increments to afford the transforming
growth factor polypeptide as a single distinct peak in
the state of a homogenous polypeptide.
In a further aspect, the present invention resides
in a process for preparing a polypeptide of the formula IIA:
10
Val-Val-Ser-His-Phe-Asn-R-Cys-Pro-Asp-Ser-His-Thr-
20 25
Gln-R'-Cys-Phe-His-Gly-Thr-Cys-Arg-R"-Leu-Val-Gln-
30 35
Glu-Asp-Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly-R"'-
40 45 50
Val-Gly-Ala-Arg-Cys-Glu-His-Ala-Asp-Leu-Leu-Ala
(IIA)
wherein R is Asp or Lys, R' is Phe or Tyr, R" is Ser or Phe and
R"' is Phe or Tyr which comprises synthetically preparing a
double stranded DNA or cDNA sequence which codes for the amino
acid sequence of the polypeptide structure, inserting the double
stranded DNA into a suitable site in a cloning vehicle or vector
to form a recombinant DNA molecule and transforming an
appropriate host with said recombinant DNA molecule to obtain
expression of the polypeptide.
-12-
13 41536
In another aspect, the present invention resides in a
process for preparing a polypeptide containing one or more of
the following peptide fragments:
Leu-Val-Gln-Glu-Asp-Lys-Pro-Ala-Cys-Val-Cys-
His-Ser-Gly and
Cys-His-Ser-Gly-Tyr-Val-Gly-Ala-Arg-Cys-Glu-
His-Ala-Asp-Leu-Leu-Ala
which comprises synthetically preparing a double stranded DNA or
cDNA sequence which codes for the amino acid sequence of the
polypeptide structure, inserting the double stranded DNA into a
suitable site in a cloning vehicle or vector to form a
recombinant DNA molecule and transforming an appropriate host
with said recombinant DNA molecule to obtain expression of the
polypeptide.
In yet another aspect, the present invention resides in
a process for preparing an antigenic oligopeptide
selected from the class consisting of:
Arg-Phe-Leu-Val-Gln-Glu-Asp-Lys-Pro-Ala
Arg-Tyr-Leu-Val-Gln-Glu-Asp-Lys-Pro-Ala
Cys-His-Ser-Gly-Phe-Val-Gly-Ala-Arg-Cys-Glu-His
Ala-Asp-Leu-Leu-Ala
Cys-His-Ser-Gly-Tyr-Val-Gly-Ala-Arg-Cys-Glu-His
Ala-Asp-Leu-Leu-Ala
which comprises synthetically preparing a double stranded DNA or
cDNA sequence which codes for the amino acid sequence of the
oligopeptide structure, inserting the double stranded DNA into a
suitable site in a cloning vehicle or vector to form a
recombinant DNA molecule and transforming an appropriate host
- 12a -
.,~
1341536
12b
with said recombinant DNA molecule to obtain expression of the
oligopeptide.
In yet a further aspect, the present invention resides
in a process for preparing an oligopeptide having an ability to
bind to cellular growth factor receptors and selected from the
class consisting of:
Cys-His-Ser-Gly-Tyr-Val-Gly-Ala-Arg-Cys
Cys-His-Ser-Gly-Phe-Val-Gly-Ala-Arg-Cys and
Cys-Ser-His-Gly-Tyr-Thr-Gly-Ile-Arg-Cys
which comprises synthetically preparing a double stranded DNA or
cDNA sequence which codes for the amino acid sequence of the
oligopeptide structure, inserting the double stranded DNA into a
suitable site in a cloning vehicle or vector to form a
recombinant DNA molecule and transforming an appropriate host
with said recombinant DNA molecule to obtain expression of the
oligopeptide.
In another aspect, the present invention resides in a
polypeptide growth factor containing one or more of the
following peptide fragments:
A. Val-Val-Ser-His-Phe-Asn-Lys-Cys-Pro-Asp-Ser-His-Thr-Gln-
Tyr-Cys-Phe-His-Gly-Thr-Cys
~
13 41536
12c
B. Val-Val-Ser-His-Phe-Asn-Asp-Cys-Pro-Asp-Ser-His-Thr-Gln-
Phe-Cys-Phe-His-Gly-Thr-Cys
C. Leu-Val-Gln-Glu-Glu-Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly
D. Cys-His-Ser-Gly-Tyr-Val-Gly-Val-Arg-Cys-Gly-His-Ala-Asp-
Leu-Leu-Ala
E. Leu-Val-Gln-Glu-Asp-Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly
F. Cys-His-Ser-Gly-Tyr-Val-Gly-Ala-Arg-Cys-Glu-His-Ala-Asp-
Leu-Leu-Ala;
with the proviso that the polypeptide is not a TGFa as found
in nature.
In another aspect, the present invention resides in a
polypeptide of the formula (IIA):
10
Val-Val-Ser-His-Phe-Asn-R-Cys-Pro-Asp-Ser-His-Thr-Gln-
20 25
R'-Cys-Phe-His-Gly-Thr-Cys-Arg-R"-Leu-Val-Gln-Glu-
30 35 40
Asp-Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly-R"'-Val-Gly-
45 50
Ala-Arg-Cys-Glu-His-Ala-Asp-Leu-Leu-Ala
(IIA)
wherein R is Asp, R' is Phe, R" is Phe and R"' is Tyr, whenever
prepared by a process of synthetically preparing a double
stranded DNA sequence which codes for the amino acid sequence of
the polypeptide structure, inserting the double stranded DNA
into a suitable site in a cloning vehicle or vector to form a
recombinant DNA molecule and transforming an appropriate host
with said recombinant DNA molecule to obtain expression of the
polypeptide.
F
13 41536
12d
Description of the Preferred Embodiments
The peptide sequence which characterizes the basic
protein structure according to the invention contains six
cysteine residues positioned at critical positions in the
polypeptide framework. It is speculated that the positioning of
the cysteine residues allows the polypeptide to fold in a
particular fashion as a result of disulfide bridges between
paired cysteines, and therefore to present a three-dimensional
structure which contributes to the biological activity of the
resulting protein. There is some evidence suggesting a
particular disulfide bridging sequence wherein (numbering the
Cys residues (in formula I above) 1 through 6 going from left to
right) Cys-1 is bonded to Cys-3, Cys-2 is bonded to Cys-4 and
Cys-5 is bonded to Cys-6 by disulfide bonds in biologically
active forms of the basic protein structure. The exact chemical
nature of the amino acid residues recited for the amino acid
sequences spaced between the Cys residues in formula I do not
appear to be particularly critical
F
41536
i provided at least 10 different amino acids from the group recited for
formula I are employed and no amino acid is repeated more than four,
preferably three, times as consecutive residues in any given
sequence. Of the amino acid residues listed for formula I,
preference is given to Val, Ser, His, Phe, Lys, Asp, Thr, Gln, Leu.
Glu, Pro, Ala, and Gly. A preferred group of biologically active
polypeptides having at least one peptide sequence of formula I are
polypeptides and oligomers thereof of the formula IV:
(AA)n-Cys-(AA)m-Cys-(AA)o-Cys-(AA)p-Cys-AA-cys-
IV
(AA)q-Cys-(AA)r
wherein AA is an amino acid residue selected from Val, Ser, His, Phe,
Asn, Lys, Asp, Thr, Gin, Arg, Leu, Glu, Pro, Ala, Gly, Trp and Tyr,
and n is an integer of from 4 to 10, m is 7, o is 4 or 5, p is 10, q
is 8 and r is an integer of from 6 to 12. In this preferred group,
even further preference is given to polypeptides or oligomers thereof
wherein o is 4 and n and r are 7. Within this preferred group, the
amino acid residues designated by AA may also be Ile and/or Met, in
addition to the amino acid residues recited previously for formula
IV. Most preferred are polypeptides wherein the amino acid residues
in the sequences spaced between the Cys residues are selected from
Val, Ser, His, Phe, Lys, Asp, Thr, Gin, Leu, Glu, Pro, Ala and Gly.
Typically the polypeptides in accordance with formulas I and IV will
have molecular weights ranging from about 5,000 to about 35,000.
Preferred polypeptides in this respect have molecular weights in the
range of 5,000 to 8,000.
- 13 -
. -~
13 4153s
1 As noted above, the present invention also contemplates a
new class of TGF polypeptides and oligomers thereof of the formulas
(formuias II and IIA above):
10 15
Val-Val-Ser-His-Phe-Asn-R-Cys-Pro-Asp-Ser-His-Thr-Gln-R'-
20 25
5 Cys-Phe-His-Gly-Thr-Cys-Arg-R " -Leu-Val-Gln-Glu-Glu-Lys-
30 35 40 II
Pro-Ala-Cys-Val-Cys-His-Ser-Gly-R " '-Val-Gly-Val-Arg-Cys-
45 50
Giu'His-Ala-Asp-Leu-Leu-Ala
5 10 15
Val-Val-Ser-His-Phe-Asn-R-Cys-Pro-Asp-Ser-His-Thr-Gln-R'-'
20 25
Cys-Phe-His-Gly-Thr-Cys-Arg-R " -Leu-Val-Gln-Giu-Asp-Lys-
30 35 40 IIA
Pro-Ala-Cys-Va1-Cys-His-Ser-Gly-R " '-Val-Gly-Ala-Arg-Cys-
45 50
Giu-His-Ala-Asp-Leu-Leu-Ala
wtierein R is Asp or Lys, R' is Phe or Tyr, R" is Ser or Phe, and
R"' is Phe or Tyr. This novel class of polypeptides is derived from
ttie finding that certain TGFs obtained from a variety of mammalian
species (both murine and human) have substantial homology in the
ar7ino acid make-up of the peptide sequence (greater than 90% of the
sequences be'lng identical) as well as substantially the same
biological properties. In particular, TGF polypeptides in accordance
with the formulas given above cause the loss of density-dependent
inhibition of cell growth in monolayer culture, overgrowth in
monolayer culture, characteristic change in cellular morphology and
acquisition of anchorage-independent growth when applied to
- 14 -
13 4153s
1 untransformed, non-neoplastic, indicator cells grown in culture. In
addition to being extremely potent cell growth promoters and
effectors of cell transformation, the TGF polypeptides in accordance
with the above formulas compete with EGF for binding to the cellular
EGF receptor and also have the ability to activate an enzyme, a
protein kinase, in human tumor cell membranes. Preferred TGF
polypeptides of formula II above include those wherein R is Asp, R'
is Phe, R" is Ser, and R"' is Phe, or where R is Lys. R' is Tyr,
R'' is Phe, and R''' is Tyr. Preferred TGF polypeptides of formula
IIA above include those wherein R is Asp, R' is Phe, R" is Phe, and
R"' is Tyr. As will be discussed in greater detail below, these TGF
polypeptides may be suitably obtained from a variety of transformed
human and m rine cell lines, certain embryonic cell lines and body
fluids of tumor-carrying mammals using the isolation process of the -
invention or by conventional synthetic or recombinant means for
synthesizing polypeptides. Typically the molecular weight of the TGF
polypeptides and oligomers thereof of formulas II and IIA will be in
the range of from about 5,000 to about 35,000. In this regard,
preference is given to TGF polypeptides having a molecular weight of
about 5,000 to 8,000.
Recognition of the substantial peptide sequence homology in
the novel class of TGF polypeptides of formulas II and IIA above and
the commonality of biological properties associated therewith allows
for further definition of a class of polypeptide growth factors which
- 15 -
13 41536
i' are within the scope of the present invention. These polypeptide
growth factors are defined as containing one or more of the following
peptide fragments:
A. Val-Val-Ser-His-Phe-Asn-Lys-Cys-Pro-Asp-Ser-His-Thr-Gln-Tyr-Cys-
Phe-His-Gly-Thr-Cys
E;. Val-Val-Ser-His-Phe-Asn-Asp-Cys-Pro-Asp-Ser-His-Thr-Gln-Phe-Cys-
Phe-His-Gly-Thr-Cys
C. Leu-Val-Gin-Glu-Glu-Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly, and
C. Cys-Hi s-Ser-Gly-Tyr-Val -Gly-Val -Arg-Cys-Gl u-Hi s-Al a-Asp-Leu-Leu-
Ala
F. Leu-Val-Gln-G1u-Asp-Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly, and ~
F. Cys-His-Ser-G1y-Tyr-Val-Gly-Ala-Arg-Cys-Glu-His-Ala-Asp-Leu-Leu-
Ala,
Preferred polypeptides in this respect, include polypeptides
containing a combination of peptide fragments A and C or B and E'and
polypeptides containing a combination of peptide fragments B and C.
Most preferred are polypeptides containing fragments B, C and D or. B
and E. Here again, the polypeptide growth factors containing one or
more of peptide fragments A, B, C, D, E. and F will generally have
rtolecular weight in the range of from about 7,000 to about 35,000,
preferably from 1,000 to 8,000. The lower end of the molecular
weight range would include the above specified peptide fragments
themselves as complete polypeptides having the characteristic growth
factor biological activity.
- 16
-
1341536
Previously it has been noted that the TGF polypeptides of
the present invention are of value in the detection of malignancies
in mammals, since the production andlor elevated levels of production
of the TGF polypeptides are characteristic of morphologic
transformation of certain human and murine cell lines. In this
regard, antibodies to the TGF polypeptides have utility in diagnosis
of malignancy, since they can be used to detect extremely low levels
of TGF polypeptide present in tumor cells or in body fluids. While
the entire TGF polypeptide molecule can be used to generate
1.o antibodies (both polyclonal and monoclonal), it is also possible, and '
advantageous from the standpoints of cost and technical effort, to
determine various regions in the TGF polypeptide sequence which are
likely to be determinant sites and to use these oligopeptides of at
least about eight amino acids, typically at least about 10 and not
more than about 20 amino acids, to define a hapten which can be used
to induce antibody formation. As further discussed below, the
oligopeptide is bound to an appropriate immunogen and introduced into
a vertebrate to produce the desired antibodies. Accordingly, the
present invention also provides a series of oligopeptides
carresponding to antigenic regions in the TGF polypeptides.
E.Kemplary species of the antigenic oligopeptides useful in generating
antibodies in accordance with the present invention are listed below:
- 17 -
13415 36
A. Val-Val-Ser-His-Phe-Asn-Asp-Cys-Pro-Asp-His-Thr
B. Val-Val-Ser-His-Phe-Asn-Lys-Cys-Pro-Asp-His-Thr
C. Arg-Phe-Leu-Val-Gln-Giu-Glu-Lys-Pro-Ala
D. Arg-Tyr-Leu-Val-Gln-Glu-Glu-Lys-Pro-Ala
E. Cys-His-Ser-Gly-Phe-Val-Gly-Va1-Arg-Cys-Gtu-His-Ala-Asp-Leu-Leu-
Ala
F. Cys-His-Ser-Gly-Tyr-Val-Gly-Val-Arg-Cys-Glu-His-Ala-Asp-Leu-Leu-
Ala
G. Arg-Phe-Leu-Val-Gin-Glu-Asp-Lys-Pro-Ala
H. Arg-Tyr-Leu-Va1-Gln-Glu-Asp-Lys-Pro-Ala
1. Cys-His-Ser-Gly-Phe-Val-Gly-Ala-Arg-Cys-Glu-His-Ala-Asp-Leu-Leu-
Ala
J. Cys-Hgs-Ser-Gly-Tyr-Val-Gly-Ala-Arg-Cys-Glu-His-Ala-Asp-Leu-Leu-
Ala
K. Val-Val-Ser-His-Phe-Asn-Asp-Cys-Pro-Asp-Ser-His-Thr and
L. Val-Val-Ser-His-Phe-Asn-Lys-Cys-Pro-Asp-Ser-His-Thr.
Another compositional aspect of the present invention is
directed to a class of oligopeptides which have therapeutic value in
treatment of malignancies. These oligopeptides have the ability to
bind to cellular growth factor receptors without causing phenotypic
transformation of the cell and therefore they can effectively compete
with TGF polypeptides for available receptor sites on the cell and
interrupt or minimize cell transformation which is characteristic of
TGF binding to cell receptors. The oligopeptides according to the
invention which have the abililty to bind to cellular receptors are
of the formula (formula III above):
-
- 18
1341~36
(AA)x-Cys-(AA)y-Gly-R-(AA)z-Arg-Cys-(AA)z~ III
wherein R is Phe or Tyr and AA is an amino acid residue selected from
Val, Asn, His, Ser, Ii e, Gly, Leu, Asp, -C-ts; Thr, Ala, Tyr, Pro, Glu,
Gln, and Arg, and x is 0 or an integer of from 1 to 6, y is 2, z is 3
and z' is 0 or an integer of from 1 to 6. Preferred oligopeptides in
accordance with the above formula include those wherein x and z' are
0 and AA is an amino acid residue selected from Val, His, Ser, Ile,
Gly and Asp. A desirable group of biologically active oligopeptides
related to those of formula III are those containing two glycine
1 o residues in addition to afford a sequence of the following formula:
(AA)x-Cys-(AA)2-Gly-(AA)2-Gly-(AA)2-Cys-(AA)z~
wierein the amino acid residues designated by the AAs are the same as
t)ose mentioned for formula III above but include Phe and subscripts
x and z' are as given for formula III above. These oligopeptides
assume a common three-dimensional structure attributable to the
disulfide bridging between the two cysteines. This disulfide bridge
ciaracterizes the biologically active forms of the oligopeptides.
P-articularly preferred oligopeptides in this regard are selected from
t~le class consisting of:
1) Cys-His-Ser-Gly-Tyr-Val-Gly-Val-Arg-Cys
2) Cys-His-Ser-Gly-Phe-Val-Gly-Val-Arg-Cys
3) Cys-His-Ser-Gly-Tyr-Val-Gly-Ala-Arg-Cys
4) Cys-His-Ser-Gly-Phe-Val-Gly-Ala-Arg-Cys and
5) Cys-Ser-His-G1y-Tyr-Thr-Gly-Ile-Arg-Cys.
The polypeptides and oligopeptides according to the
invention as defined by the structural formulas (formulas I through
- 19 -
13415 36.
1 IV) and peptide sequences given above can be prepared by synthetic
techniques, techniques whereby the peptide is isolated from a
naturally occurring source, e.g., cell lines and body fluids, and by
techniques employing hybrid DNA technology. For those polypeptides
and oligopeptides of the invention containing up to about 50 amino
ac:id residues, conventional solid phase peptide synthesis is suitably
employed. In this general synthetic procedure for making peptides,
wtrich is described, for example, in U.S. Patent No. 4,341,761 to
G<<nfield et al., employs known side-chain protecting groups and
ccinventional polystyrene resins supports - e.g., chloromethylated
resins, hydroxymethyl resins or benzhydrylamine resins - to affect
the amino ac-ici coupling. For polypeptides containing in excess of
at-out 50 amino acid residues, the process according to the invention
fc-r isolating homogeneous TGFs from natural sources (which is
described in detail below) can be suitably employed to obtain pure
fcirms of the desired peptide. In this regard,.particularly suitable
sources of the TGF polypeptides according to the invention include
serum-free medium conditioned by retrovirus-transformed Fischer rat
embryo fibroblasts, in particular fibroblasts transformed with
Sriyder-Theilen feline sarcoma virus, Moloney murine sarcoma virus-
transformed mouse 31-3 cells and human metastatic melanoma cell lines
A2058 and A375. Sources and methods for suitable murine cell lines
are described in DeLarco et al. (1980) J. Biol. Chem. 255, pp. 3685-
3690 and Ozanne et al. (1980) J. Cell. Physiol. 105, pp. 163-180.
Sources and methods for human cell lines are similarly described in
-20-
13 41536
Todaro et al. (1980) Proc. Natl. Acad. Sci. USA 77, pp. 5258-5262 and
Giard et al. (1973) J. Natl. Cancer Inst. 51, pp. 1417-1423. The
isolation process of the invention described below can also be used
to obtain TGF polypeptides according to the invention from various
body fluids such as urine, serum, plasma, whole blood or
cerebrospinal fluid of human or murine subjects carrying malignancies
or transformed cells which produce TGF polypeptides. In this regard,
a suitable source of TGF polypeptides according to the invention is
the urine or other body fluids of mice which have been inoculated
with tumor cells (human melanoma or transformed rat) known to produce ~
TGF polypeptides. In all cases the identification and purity of the
TGF polypeptide can be monitored by a radioreceptor assay based on
receptor cross-reactivity with EGF (see experimental examples below).
In techniques utilizing recombinant or hybrid DNA technology, the
oligopeptides according to the invention or segments of the
polypeptides according to the invention containing up to, for
example, 20 amino acids can be used to deduce the codon sequence for
single stranded nucleotide (DNA) probes. These nucleotide probes can
then be synthesized using known synthetic techniques and used as a
probe to obtain messenger RNA (mRNA) coding for growth factor-type
polypeptides in both normal and transformed cells or body fluids
containing said peptides. Once messenger RNA is obtained,
conventional techniques can be used for reverse transcribing of the
mRNA to complementary DNA (cDNA) and subsequent cloning of the cDNA
in a suitable vector to obtain expression of the desired polypeptide.
- 21 -
134 15 36
1 The process according to the invention provides a uniquely
effective means of obtaining TGF polypeptides in homogeneous form
from various aqueous based fluids containing less pure forms of the
TGF polypeptides such as serum-free mediums conditioned by
transformed cell lines which produce TGF polypeptides or body fluids,
e.g., urine, from mammals carrying malignancies or transformed cells
which produce the TGF polypeptides. Important aspects of this unique
isolation or purification process include an initial extraction or
dialysis step using aqueous acetic acid, subsequent gel permeation
chromatography of the acid-soluble TGF-containing activity, and
finally, reverse phase high pressure liquid chromatography using
sequentially acetonitrile and 1-propanol in the presence of aqueous
trifluoroacetic acid. Broadly defined, this process involves
isolation of a homogeneous transforming growth factor polypeptide
from an aqueous medium containing said transforming growth factor
polypeptide in impure form by the process steps comprising:
1) dialyzing the aqueous medium containing the
transforming growth factor in impure form against
aqueous acetic acid to afford a solvent phase
containing transforming growth factor polypeptide which
phase is concentrated and optionally clarified,
2) reconstituting the concentrated solvent phase of step
1) with aqueous acetic acid and subjecting the
reconstituted solution to gel permeation chromatography
by applying reconstituted solution to gel permeation
-22-
13 41536
1 chromatography column conditioned with aqueous acetic
acid and eluting with aqueous acetic acid to obtain
selected fractions of eluate containing transforming
growth factor polypeptide in an enhanced state of
purity, said selected fractions being combined and
concentrated, to afford a partially purified,
transforming growth factor polypeptide-containing
product,
3) subjecting the partially purified, transforming growth
factor polypeptide-containing product of step 2) to
sequential rev-lrse phase high pressure chromatography
by passing said product, after reconstitution in
aqueous trifluoroacetic acid, through one or more
hydrocarbon bonded silica matrix columns. which have
been equilibrated with aqueous trifluoroacetic acid,
under high pressure liquid chromatography conditions,
the initial column elution being performed using a
linear acetonitrile gradient in aqueous trifluoroacetic
acid and the subsequent column elution, which is
carried out on the combined, transforming growth factor.
polypeptide-containing fractions of the initial high
pressure chromatography step, being performed using a
l i near 1-propanol gradient in aqueous tri f l uoroaceti c
acid, said 1-propanol gradient being increased in
sufficiently small 1-propanol concentration increments
-23-
13 4 1 5 3.6-
to afford the transforming growth factor polypeptide as
a single distinct peak in the state of a homogeneous
polypeptide.
In a preferred application, the process according to the
invention is employed to isolate homogeneous TGFs from serum-free
media conditioned by transformed, TGF-producing cell lines. In this
preferred application, the conditioned medium is suitably clarified,
e.g., by centrifugation, and concentrated prior to dialysis and the
TGF-containing solvent phase from dialysis is suitably clarified,
lo e.g., by centrifugation, as well as concentrated prior to gel
permeation chromatography. In any case, the dialysis is suitably
carried out using an aqueous acetic acid solvent having an acetic
acid concentration of from 0.01 to 1 molar, with 0.1 molar acetic
acid being preferred. The gel permeation chromatography may be
carried out using a variety of gels conventionally employed to
separate proteins or polypeptides based on molecular size. Suitable
gels include dextran gels, agarose gels and polyacrylamide gels. In
this regard, preference is given to polyacrylamide gel filtration
resins (Bio-Gels ) such as Bio-Gel P-10, Bio-Gel P-30 and Bio-Gel
P-60, Bio-Ge1 P-10 being especially preferred. The aqueous acetic
acid used to condition the column and to elute the TGF-containing
fractions suitably has an acetic acid concentration of from 0.2 to
2.0 molar, with 1.0 molar acetic acid being preferred. The TGF-
containing fractions which elute from the column can be Identified by
determining their EGF-competing activity and growth promotion
* Trade Mark
-24-
t;;~
41536
1 activity in soft agar (see experimental examples below). After gel
permeation chromatography,, the fractions containing TGF polypeptides
in an enhanced state of purity are pooled together and concentrated
for example. by lyophilization as a preparation step for further
purification by reverse phase high pressure liquid chromatography
(HPLC).
The final stage of the purification process of the
invention involves sequential HPLC with acetonitrile and 1-propanol
in the presence of aqueous trifluoroacetic acid. This sequential
HPLC can be carried out using one or more HPLC columns, but it is
preferred to carry out the sequential HPLC steps using a single HPLC
column. The column packing employed is suitably a porous silica
matrix to which a long chain hydrocarbon, for example, hydrocarbon
containing 16 to 22 carbon atoms, is bound. Preferred packings are
#Bondapak hydrocarbon columns, in particular ABondapak*C18 column
(10-Am particle size, 0.39 x 30 cm, Waters Associates). Typically,
the procedure is carried out under pressure, preferably in the range
of from about 50 to 5,000 pounds per square inch (psi). Prior to
application to the column, the concentrated TGF-containing fractions
are reconstituted in an aqueous 1 to 10% trifluoroacetic acid and
adjusted to a pH in the range of 2 to 5, preferably 3.5 by the
addition of trifluoroacetic acid. The column is suitably
equilibrated with 0.01 to 0.1% aqueous trifluoroacetic acid,
preferably 0.05% aqueous trifluoroacetic acid, before sample
injection. The first elution is carried out with acetonitrile in a
* Trade Mark
-25-
~~
1~41536
1 0.01 to 0.1%, preferably 0.05% trifluoroacetic acid, using a linear
acetonitrile gradient (acetonitrile concentration increased linearly
at a gradient in the range of about 0.1X1min to about 1X/min). The
elution is carried out over a time period of from 0.2 to 3 hours at a
flow rate of about 0.2 to 2 ml/min and at a temperature of from 10 to
500C, preferably about 400C. The pooled fractions containing TGF
activity as determined by EGF competition and soft agar assay are
concentrated. foi- examplef by lyophilization, prior to the second
step of the HPLC using 1-propanol solvent. For the second step of
the sequential HF'LC, the pooled and concentrated fractions from the
first HPLC eluticin are reconstituted in 0.01 to 0.1% trifluoroacetic
acid and rechromatographed on the same column or a second column
equilibrated witth trifluoroacetic acid in a manner identical to that
used for the first column. This second elution is carried out with
1-propanol in a 0.01 to 0.1%, preferably 0.035% trifluoroacetic acid
using a linear 1-propanol gradient. It is important for optimum
results to employ a shallow linear 1-propanol gradient in this step.
In particular, the 1-propanoi concentration should be increased
linearly at a gradient which does not exceed 0.1X/min and preferably
the linear 1-propanol gradient should be maintained between 0.01X/min
and 0.05X/min during the elution. This second elution is suitably
carried out over a time period of from 1 to 5 hours at a flow rate of
about 0.5 to 5 ml/min and at a temperature of from 10 to 600C,
preferably about 400C. By controlling the linear 1-propanol
concentration gradient at the shallow levels given above, it is
possible to elute TGF polypeptides as well-defined peaks of TGF
-26-
13 41536
activity in the form of homogeneous polypeptides.
With the isolation process of the invention, it is possible
to recover up to about 70% of the initial TGF activity in the impure
starting material while achieving degrees of purification in excess
of 200,000-fold. The homogeneous TGF polypeptides obtained by the
isolation process of the invention typically have molecular weights
in the range of about 5r000 to about 35,000 and are of sufficient
purity to permit peptide sequencing. Preferred homogeneous TGF
polypeptides, which are obtained with the process of the invention,
include TGFs h3ving apparent molecular weights of 7,400, 20,000 and =
30,000 to 35,000. These homogeneous TGF polypeptides show
characteristic biological properties of TGF.polypeptides when they
are applied to untransforr::ed, non-neoplastic indicator cells. growing
in culture, including acquisition of anchorage-independence, with the
resultant abiVty to grow in soft agar. Further, the homogeneous TGF
polypeptides can be used to raise antibodies (see below) which have
value in the detection and treatment of cancer and other
proliferative cliseases in accordance with the invention. In this
regard, it is r,ot essential that each of the forms of the TGF be
purified to homogeneity in order to produce antibodies to the
different TGF peptides. These various forms of antibodies to the TGF
polypeptides are also contemplated by this invention.
Antibodies according to the invention include both
monoclonal and polyclonal antibodies raised to the TGF polypeptides
-27-
13 4 ~536
1. of formulas II and IIA given above, antigenic oligopeptides derived
from the TGF polypeptides of formulas II and IIA and the homogeneous
TGF polypeptides obtained using the isolation process of the
_ invention from various transformed cell lines and body fluids of
mammals carrying malignancies or transformed cells. The antibodies
according to the invention can be prepared in a variety of ways known
in the art, depending on whether monoclonal or polyclonal antibodies
are desired. For polyclonal antibodies, a vertebrate, typically a
domestic animal, Is hyperimmunized with antigen and the blood
collected shortly after repeat immunizations and the gamma globulin
isolated. Suitable methods for preparing polyclonal antibodies are
described in the Handbook of Experimental Immunology, 3rd edition,
Weir, Editor, Blackwell Scientific publications, Oxford and London,
1978. For monoclonal antibodies, a small animal, typically a mouse
or rat, is hyperimmunized with antigen, the spleen removed and the
lymphocytes fused with myeloma cells in the presence of a suitable
fusion promoter. The resulting hybrid cells or hybridomas are
screened to isolate individual clones, each of which secrete a single
antibody species to the antigen. The individual antibody species
obtained in this way are each the product of a single B cell from the
immune animal generated in response to a specific antigenic site
recognized on the immunogenic substance. The general process for
obtaining monoclonal antibodies, including those according to the
invention, is described by Kohler and Milstein (1975) Nature 256, pp.
495-497. The polypeptides and antigenic oligopeptides of the
-28-
1341536
1 invention used to produce antibodies (both polyclonal and monoclonal)
may be employed directly in the immunization procedure or they iaay be
bound to a suitable carrier-protein using methods known in the art,
for example, see U.S. Patent No. 4,341,761to Ganfield et al. Use of
a carrier protein is particularly preferred when the immunization is
carried out using the antigenic oligopeptides of the invention.
The antibodies according to the invention may be used in a
variety of ways. In a preferred application, they may be used for
diagnosis of malignancy and other proliferative diseases. In
instances where the antigen may be found in a physiological fluid or
at a concentration differential only when malignancy or other
proliferative disease exists, the physiological fluid, such as serum,
plasma, whole blDod or cerebrospinal fluid may be assayed.
Antibodies employed in assays may be labeled or unlabeled. Unlabeled
antibodies may b~ employed in agglutination; labeled antibodies may
be employed in a wide variety of assays, employing a wide variety of
labels, such as radionuclides, enzymes, fluorescers, enzyme
substrates or cofactors, or the like. These techniques are amply
defined in the l-iterature and exemplary assays may be found in U.S.
Patent Nos. 3,81;7,834, 3,935,074, 4,233,402 and 4,318,980, as
illustrative.
In some techniques it will be useful to label the antigen
or fragment thereof, rather than the antibody, and have a competition
between labeled antigen and antigen in the sample for antibody. In
-29-
~3415361_ this situation, it is common to provide kits which have the
combination of the labeled antigen or labeled fragment and the
antibody in amounts which provide for optimum sensitivity and
accuracy. In other situations, it is desirable to have a solid
support. where either.antigen or antibody is bound. A polyepitopic
antigen can serve as a bridge between antibody bound to a support and
labeled antibody in the assay medium. Alternatively. one may have a
competition betrveen labeled antigen and any antigen in the sample for
a limited amount of antibody.
Where the antigen may not be found in a physiological fluid
or if found there is not diagnostic of malignancy or the target
proliferative d-isease, then cells will have to be isolated and the
cells assayed for the presence of the antigen. For detecting the
antigen, the tissue sample may be lysed by conventional methods,
e.g., base, detergents, or the like, cellular debris separated by
filtration or centrifugation and the filtrate or supernatant isolated
and assayed.
For purposes of therapy, either xenogeneic or allogeneic
antibodies may be employed, depending upon the nature of the
treatment, and whether the foreign antibodies will induce an immune
response. The literature has described a number of ways of making
human antibodies, where it is found that mouse or other mammalian
antibodies are not satisfactory. The antibodies may be used in a
wide variety of ways. By employing the appropriate IgG (other than
IgGl), one may induce lysis through the natural complement process.
- 30 -
13 41536
1 Alternatively, the lysing portion of a toxin may be joined to the
antibodies, particularly a Fab fragment. The antibodies may be bound
to liposomes for directing the liposomes,>to the malignant cells to
- become ingested by the cells by merging of the membranes. Other
labels may also be bound to the antibodies, such as radionuclides,
fluorescers, enzlrmes, and the like. By introducing the antibodies in
vivo, the antibodies will direct the label to the malignant cell,
where the presence of malignancy may be diagnosed or treated.
The foi-rnulation of the antibodies will vary widely,
depending on the nature of the label, the purpose of the antibodies, --
the site to which the antibodies are to be directed, and the like.
Usually, the ant~:bodies will be formulated in a physiologically
acceptable carrie!r, e.g., saline or phosphate buffered saline, and
injected into the host, when possible at the desired site, and when
this is not poss4:ble, into a circulating system, such as blood.
The antibodies obtained in accordance with this invention
can also be used to isolate cells expressing the TGF polypeptides and
to remove cells "n vitro from a heterogeneous cell population
containing cells expressing a TGF polypeptide. Separation can be
achieved with a t'luorescence activated cell sorter (FACS). This same
technique can be used for identifying and isolating cells expressing
a TGF polypeptide. For removing cells expressing a TGF polypeptide
from a mixture of cells, the subject antibodies may be combined with
complement, joined to the lysing fragment (A fragment) of a toxin
- 31 -
1341536
(see E.P.O. Application No. 17,507 and U.S. Patent Application No.
2,034,324) or the cells agglutinated and separated by physical means.
Methods and compositions employing the biologically active
polypeptides and oligopeptides of the invention are also afforded for
treatment of cancer and other proliferative diseases and for
therapies wherein cell growth promotion is beneficial. In
particular, conipositions are provided employing the oligopeptides of
formula III abc+ve for the treatment of malignancies. Further
compositions cantaining biologically active polypeptides of formulas
I, II, and II F,for treatment of cancer and other proliferative
diseases and fcr cell growth promotion applications, e.g., wound
healing and ulcer therapy are also provided. These therapeutic
compositions comprise effective amounts of the indicated
oligopeptides and polypeptides in admixture with pharmaceutically
acceptable carriers. In particular, pharmaceutical compositions that
contain the oligopeptides and/or polypeptides of the invention as an
active ingredient will normally be formulated with an appropriate
solid or liquid carrier depending upon the particular mode of
administration being used. For instance, parenteral formulations are
usually injectable fluids that use pharmaceutically and
physiologically acceptable fluids such as physiological saline,
balanced salt solutions, or the like as a vehicle. Oral
formulations, on the other hand, may be solid, e.g., tablet or
capsule, or liquid solutions or suspensions.
- 32 -
~'.
13 41536
1. In the therapeutic methods of the invention, the
oligopeptides and/or polypeptides may be administered to humans in
various manners such as orally, intravenously, intramuscularly,
- intraperitoneally, intranasally, intradermally, and subcutaneously.
The particular mode of administration and dosage regimen will be
selected by the attending physician taking into account the
particulars of the patient, the nature of treatment required, and/or
the disease stat:e involved. For instance, damaged tissue from wounds
is usually treated by daily or twice daily doses over a few days to a
few weeks; whereas tumor or cancer treatment involves daily or
multidaily doses; over months or years. The oligopeptide and/or
polypeptide ther=apy of the invention may be combined with other
treatments and may be combined with or used in association with other
chemotherapeutic: or chemopreventive agents for providing therapy
against proliferative diseases, neoplasms, or other conditions
against which they are effective.
TGF oligopeptides, polypeptides and antibodies raised to
these peptides z.lso are effective in the detection and treatment of
bone-loss diseases. For example, TGF activity is present in the
material responsible for bone-resorbing activity in tumors associated
with hypercalcemia (Ibbotson et al. (1983) Science 221, 1292).
Increased bone resorption may be an endocrine effect of TGFs secreted
by tumor cells.
In addition, elevated levels of TGF stimulate bone
resorption through calcium leaching, which is associated with
-33-
13 4.1536
1,: diseases, such as osteoporosis. Osteoporosis is a disease commonly
seen in elderly individuals, causing bone loss. The use of -
antibodies to TGF may be indicative of the existence of osteoporosis.
Therefore, TGF antibodies are useful in the analysis and therapy of
this important disease..
This invention provides a method for detecting bone loss in
a human host wliich comprises contacting cells or body fluids with an
antibody of TGl= and determining the level of binding of said antibody
to said cells or cellular products in the body as diagnostic of a-
1 0 host with bone loss. Also provided are compositions for treatment of
bone-loss dise<<se, employing effective amounts of the oligopeptides,
polypeptides, cr antibodies of TGF together with a pharmaceutically
acceptable carrier therefor.
The follcwing examples are offered by way of illustration and
not by way of limitation:
Example I
Production, Purification and Characterization of low Molecular
Weight Human Transforming Growth Factors (hTGFs)
A. Experimental Procedures
Source of hTGFs
hTGFs were purified from the serum-free medium conditioned
by a human metastatic melanoma line A2058 (Todaro et al. (1980) Proc.
Natl. Acad. Sci. USA 77. pp. 5258-5262) derived from a brain
metastasis in a 43-year-old man. Cells were grown to 90% confluency
in roller bottles containing Dulbecco's modified Eagle's medium
- 34 -
~3 41536
1 (Grand Island Biological Co., 430-2100), supplemented with 10% calf
serum (Colorado Serum Co..) at 370C. The cells were washed for 1 h
with 50 ml of serum-free Waymouth's medium (Grand Island Biological
- Co.. MD 705/1). This and a second collection of supernatant fluid,
24 h later, wer=e discarded. Subsequent collections were made every
other day, or every 3rd day, for a 2-week period.
The medium was collected by decantation, stored for up to
24 h at 40C in the presence of the protease inhibitor phenylmethane-
sulfonyl fluoride (ip g/ml), and clarified by continuous flow
centrifugation at 32,000 rpm at 40C. Flow rates of 5 liters/h in the
CF-32 continuous flow rotor (Beckman) in the model L5-50
ultracentrifuge (Beckman) were used. The supernatant, after high
speed centrifugation, will be referred to as A2058-conditioned
medium.
The A2058-conditioned medium was immediately concentrated
in the hollow fiber Dialyzer/Concentrator (model DC10, type H1OP5-20
cartridge, Amicon Corp.) at 10OC. The'concentrate was drained after
a 150-fold reduction in volume. The cartridge was washed with 1000
ml of Waymouth's niedium. The ultrafiltrate was discarded.
Purification of hTGFs
Dialysis and Centrifugation
The combined retentate and cartridge wash after
ultrafiltration of A2058-conditioned medium was dialyzed for 60 h
against 0.1 M acetic acid in Spectrapor 3 dialysis tubing (Spectrum
Medical Industries). The retentate was centrifuged at 100,000 x g
-35-
;3 41 536
for 1 h at 40C. The pellet was discarded. The supernatant was
concentrated by lyophilization and reconstituted in 0.5 ml of I M
acetic acid/liter of original A2058-conditioned medium.
Chromatography on Bio-Gel P-10
Following concentration, dialysis. and centrifugation, the
supernatant con'taining hTGF activity was further purified by gel
permeation chroinatography on a column (2.5 x 85 cm) (420 ml bed
volume) of Bio-Ge1 P-10 (200-400 mesh, Bio-Rad Laboratories). The
column was equilibrated with iM acetic acid at 220C. Samples of
protein (65-115 mg) in 1 M acetic acid (5 ml) were applied to the
column. To ensure a constant flow rate, the column effluent was
regulated at 12 mi/h with a peristaltic pump. 4.8-mi fractions were
collected. Aliquots were lyophilized for subsequent determinations
of EGF-competin(l activity and growth-promoting activity in soft agar.
1 5 Fractions repre>enting the major portions of a given peak were pooled
and concentrateii by lyophilization.
Reverse Phase High Pressure Liquid Chromatography
The final purification of hTGF was achieved by reverse
phase HPLC, usiiig the general.procedure described in Marquardt et al.
(1980) J. of Biul. Chem. 256, pp. 6859-6865. All separations were
performed on a #Bondapak C18 column (10-pm particle size, 0.39 x 30
cm. Waters Associates) at a flow rate of 1 ml/min at 400C.'
Lyophilized samples were reconstituted in 0.05% (v/v) trifluoroacetic
acid in water, adjusted to pH 2 with 10% (v/v) trifluoroacetic acid,
and applied through the sample injector to the column which was
-36-
~3 4 5 6
equilibrated with 0.05% trifluoroacetic acid. -The column was then
eluted with a linear acetonitrile gradient in 0.045% trifluoroacetic
acid. The column effluent was collected in 1.5-m1 fractions.
Aliquots were lyophilized for subsequent EGF competition and growth
stimulation assays. Pools of fractions comprising the major hTGFs
activity were concentrated by iyophilization.
hTGFs-containing pools were reconstituted in 0.05%
trifluoroacetic acid and rechromatographed on the same column,
previously equilibrated with 0.05% trifluoroacetic acid in water.
The column was then eluted with a linear 1-propanol gradient in
0.035% trifluoroacetic acid. The column effluent was collected in
1.5-mi fractions. Aliquots were lyophilized for EGF competition and
growth stimulati:)n assays.
SDS-Polyacrylamile Gel Electrophoresis
Sodium dodecyl sulfate (SDS) -polyacrylamide gel
electrophoresis 'was performed as described in Laemmli (1980) Nature
(Lond) 227, pp. 580-685. A 15-30% acrylamide gradient slab (140 x
120 x 0.75 mm) was prepared with a 4% stacking gel. The gels were
run at 30 V with an electrode buffer containing Tris (0.05 M),
glycine (0.38 M), and SDS (0.1%, w/v) until the tracking dye
(bromphenol b'iue) had run off the end of the gel. After
electrophoresis, gels were fixed in 50% methanol, 10% acetic acid for
2 h. washed in 5% methanol, 7% acetic acid overnight, and stained
with silver (Oakley et al. (1980) Anal. Biochem. 105, pp. 361-363).
-37-
1341536
1. Protein Determination -
Total protein was determined using bovine serum albumin as
a standard. Prior to protein determination, the starting material
was dialyzed against phosphate-buffered saline to remove components
of the culture medium interfering with the color reaction or
lyophilized, if samples had been dissolved in volatile acids.
Protein was also determined by amino acid analysis. Lyophilized
samples were hydrolyzed at 110oC for 24 h in evacuated Pyrex*tubes
with 0.1 ml of 6 N HC1 containing 0.1% liquid phenol, and analyzed
with a Durrum D-500 analyzer equipped with a PDP 8/A computing
integrator using o-phthalaldehyde for the fluorogenic detection of
primary amines (Bensen et al. (1975) Proc. Natl. Acad. Sci. USA 72,
pp. 619-622).
Radioreceptor Assay
Purified EGF was labeled with Na1251 by a modification of
the chloramine-T method as described in DeLarco et al. (1978) Proc.
Natl. Acad. Sci. USA 75, pp. 4001-4005. The 125I-EGF binding assay
was performed on subconfluent monolayers of formalin-fixed A431 human
carcinoma cells as previously described (in DeLarco et al. (1980)
J. of Biol. Chem. 255, pp. 3685-3690). The fixed cells were washed
twice with 0.5 ml of binding buffer (Dulbecco's modified Eagle's
medium containing 1 mg/ml of bovine serum albumin and 50 inM
2-[bis(2-hydroxyethyl)amino]ethanesulfonic acid. pH 6.8).
Competitions were initiated by the addition of 0.2 ml of binding
buffer containing 0.4 ng of 1251-EGF with or without potential
* Trade Mark
-38-
p~t
13 41536
.1 Inhibitor. After incubation for 1 h at 220C. the specifically bound
125I-EGF was detemined. The TGF content was expressed by its degree
of inhibition of the binding of 1251-EGF to the EGF receptor. One
EGF-competing activity unit is defined as the amount of protein that
inhibits the binding of 125I-EGF to its receptor by 50%.
Soft Agar Growth Assay
The assay for colony -growth in soft agar, using normal rat
kidney fibroblasts, clone 49F, was performed as reported in Todaro et
al. (1980) Proc. Natl. Acad. Sci. USA 77, pp. 5258-5262. Lyophilized
samples to be tefted were reconstituted in 0.5 ml of Dulbecco's
modified Eagle's medium, supplemented with 10% calf serum. 1.5 ml of
0.5% (w/v) agar (Difco) in the supplemented medium and 0.5 ml of
supplemented medium contaiziing 2.3 x 104 cells were added. 2.3 ml of
the resultant mixture were pipetted on a 2-ml base layer (0-5% agar
in supplemented inedium) in 60-mm Petri dishes (Falcon). The cells
were incubated a-t 370C in a humidified 5% C02/95% air atmosphere.
The assay was re,id unfixed and unstained at 5 days and at 10-14 days.'
B. Resul ts
Source, Concentr,jtion, and Initial Fractionation of hTGFs
hTGFs was isolated from serum-free conditioned medium of
the highly trans-formed human metastatic melanoma cell line. A2058.
The quantitation of hTGFs was based on two of its properties: the
- 39 -
1341536
1 capacity to induce anchorage-independent growth of normal rat kidney
fibroblasts in soft agar, and the ability to compete with 125I-EGF
for the EGF receptor sites on A431 human carcinoma cells. A summary
of the steps leading to the isolation of hTGFs and its recovery is
s presented in Table I.
-40-
TAOLE I
Purification of hTGFs from conditioned medium of human melanoma cells, A2058
Ul.
Purification step Proteina recovered competing Relative specific Degree of
purification Recovery
activity activity
rccovered
mg units units/mg -fold
1. A2058-conditioned medium 1,02G 4,525 4.4 1 100
2. Acid-soluble supernatant 837 4,299 5.1 1 95
3. 0to-Gel P-10
Pool P-10-A 29.7 = 2,077 70 16 (1) 45.9 (100)
Pool P-10-8 14.5 2,033 140 32 (1) 44.9 (100)
4, uBondapak C18(acetonitrite) 0.202 1,628 8,059 1,832 (S7) 36.0 (80.1)
5, u8ondapak Ct0(1-propanoi) 0.0015 1,476 984,000 223,636 (6,988) 32.6 (72.6)
aTotal protein was determined using bovine serum albumin as a standard. The
quantitation of step65 hTGFs was based on amino -''
acid analysis. The absolute specific activity of a companion aliquot was found
to be 1-1.5 x 10 units/mg. W
bOne EGF-competing activity unit is defined as the amount of protein that
inhibits the binding of 12SI-EGF to 1ts receptor
by 50%. tJt
~
, = =
13 4 5 3 6
1 To remove serum proteins, A2058 cells were extensively
washed with Waymouth's medium prior to their culture in serum-free
medium. The supernatant fluids were collected every other day for a
2-week period. Culture conditions were such that at the end of the
culture period more than 90% of the cells were still viable and
attached as monolayers. The intitial clarified A2058-conditioned
medium of 136 liters, containing 1.02 g of total protein and 4525
units of EGF-competing activity, was concentrated to about 900 ml
using a hollow fiber concentrator with cartridges of 5000 molecular
weight cutoff. The total EGF-competing activity was retained and a
recovery above 95% was obtained.
Dialysis of the concentrated A2058-conditioned medium
against acetic acid and s=jbsequent centrifugation resulted in 95%
recovery of the initial total EGF-competing activity. 18% of the
protein was aci-j-insoluble and was discarded. The acid-soluble,
partially purified hTGF was subjected to gel permeation
chromatography i~n..Bio-Gel P-10. The column was eluted with 1 M
acetic acid. The bulk of the contaminating protein was eluted in the
exclusion voluml~ of the column and was well separated from the EGF-
competing activity and growth-stimulating activity. Two peaks of
activity were found to be well resolved from each other. Fractions
with both EGF-competing and growth stimulating activity (P-10-A and
P-10-B) had apparent molecular weights of 10,500 and 6,800,
respectively. Fractions iiaving only one of the two activities were
not observed. hTGF-containing fractions were pooled as indicated,
-42-
13 4 5 3 6
1 lyophilized, and further purified. The larger molecular weight TGF
eluted from the column in a broad peak (P-10-A) and appeared to be
associated with polypeptides of different sizes. P-10-A contained
46% of the initial EGF-competing activity. The small molecular
weight TGF eluted from the column in a sharp peak (P-10-B) and
represented 45% cif the initial total EGF-competing activity. The
cumulative yield of total input EGF-competing activity from step 1
through the gel permeation chromatography step was 91% (Table 1).
hTGF was eluted as two distinct major peaks that varied
quantitatively from one preparation to another. In some preparations ~
of A2058-conditicned medium essentially all the growth-promoting
activity was in the hTGFs region.
Purification of hTGFs
hTGFs was further purified by reverse phase HPLC. Pool
P-10-B, after gel permeation chromatography of the acid-soluble EGF-
competing activity of concentrated A2068-conditioned medium on Bio-
Gel P-10, was reconstituted in 0.05% trifluoroacetic acid in water,
and then chromatographed on a pBondapak C18 column. EGF-competing
and growth-stimulating activities in soft agar of individual
fractions were determined. hTGFs was well separated from the bulk of
contaminating protein which eluted at higher concentrations of
organic solvent. Fractions containing hTGFs were pooled,
lyophilized, and taken for rechromatography. A 57-fold purification
of hTGFs after gel permeation chromatography was obtained. 80% of
-43-
1341536
1 the initial EGF-competing activity in pool P-10-B was recovered
(Table I).
Rechromatography of the hTGFs-containing fractions on
pBondapak C18 support was chosen for the final purification step,
since only relatively small losses of EGF-competing activity were
observed on these columns. In order to obtain a distinct separation
of hTGFs from inipurities, it was necessary to use a shallow linear 1-
propanol gradierit in 0.035% trifluoroacetic acid. A bulk of
contaminating peptide material was separated from a well-defined peak
of activity. ECF-competing and growth-stimulating activities
copurified with a distinct absorbance peak at 13% 1-propanol.
Fractions containing hTGFs were pooled and further analyzed. The
purification of hTGFs was approximately 7000-fold after gel
permeation chroniatography with a yield of 33% of the initial total
EGF-competing activity. The overall recovery of hTGFs from step 3
through the final reverse phase HPLC step was 73%, and the recovery
range per step was 80-100% (Table I).
Characterization of hTGFs
The purity of the final hTGFs preparation was determined by
analytical SDS-polyacrylamide gel electrophoresis. The gel was
stained with silver. One major polypeptide band, with an apparent
Mr = 7400, was observed. The same pattern was obtained when samples
were electrophoresed under nonreducing conditions indicating that TGF
is a single chain molecule.
- 44 -
13 41536
1. The receptor reactivity of hTGFs was compared with EGF in
the radioreceptor assay. The quantitation of hTGFs was based on
amino acid analysis of a companion aliquot. Both hTGFs and EGF
competed with 125I-EGF for the EGF receptor sites of A431 human
carcinoma cells. The specific EGF-competing activity of hTGFs was
found to 1-1.5 x 106 units/mg; 1.1 ng of hTGFs or EGF were required
to inhibit EGF binding by 50%.
hTGFs enabled normal anchorage-dependent rat kidney cells,
clone 49F, to gn)w in soft agar. The half-maximal response of hTGFs
in soft agar was reached with 1 EGF-competing unit, or 1.1 ng of
hTGFs, whereas EGF does not stimulate growth of these cells in soft
agar even when tested with up to lO#g.
Example II
Larger Scale Pro;luction, Purification and Amino Acid Sequencing
of Human (hTGF), Rat rTGF), and Mouse (mTGF) Transforming
Growth Factors
A. Expeririental Procedures
Source of TGF
rTGF, niTGF and hTGF were purified from the serum-free
medium conditioned by Fisher rat embryo fibroblasts, FRE CL10, a
subclone of FRE sA (Sacks et al. (1979) Virology 97, pp. 231-240),
nonproductively transformed by Snyder-Theilen feline sarcoma virus
(Snyder et al. (1969) Nature (Lond), 221, pp. 1074-1075), a Moloney
murine sarcoma virus-transformed 3T3 cell line, 3B11-IC (Bassin et
al. (1970) Int. d. Cancer 6, pp. 95-107), and two human metastatic
-45-
~341536
1. melanoma lines, A2058 (see Example I), and A375-(Girad et al. (1973)
J. Natl. Cancer Inst. 51, pp. 1417-1423), respectively. Cells were
grown in 2-liter plastic roller bottles containing Dulbecco's
modified Eagle's medium supplemented with 10% calf serum and
subsequently maintained in serum-free Waymouth's medium as described
in DeLarco et al. (1978) Proc. Natl. Acad. Sci. USA 75, pp. 4001-
4005. Serum-free conditioned medium was collected every 24 h, for a
3-day period, clarified by continuous flow centrifugation, and the
supernatant concentrated (Marquardt et al. (1980) J. of Biol. Chem.
255, pp. 9177-9181). The concentrate of conditioned medium was the
starting material for the purification of TGFs.
Purification of TGF
The TGFs were prepared essentially as previously described
in Example I for the purification of the melanoma-derived hTGF. The
retentate after ultrafiltration of conditioned medium was dialyzed
against 0.1 M acetic acid, and the supernatant, after centrifugation,
concentrated by lyophilization and reconstituted in 1 M acetic acid
for subsequent gel permeation chromatography on a column (2.5 x 85
cm) of Bio-Gel P-10 (200-400 mesh, Bio-Rad Laboratories). The column
was equilibrated with 1 M acetic acid. Fractions comprising the
major EGF-competing activity with an apparent molecular weight of
approximately 7,000 were pooled and lyophilized.
The final purification of rTGF, mTGF, and hTGF was achieved
by reverse phase HPLC using the chromatography system described in
Example I. The separations were performed on a#Bondapak C18 column
-46-
13 49536
1 (l0 m particle size, 0.39 x 30 cm, Waters Associates). The mobile
phase was 0.05% trifluoroacetic acid and the mobile phase modifier
was acetonitrile containing 0.045% trifluoroacetic acid. The
concentration of acetonitrile was increased linearly' (0.083X/Min)
during 2 h at a flowrate of 1 ml/min,at 400C for elution of
peptides. TGF-containing pools were lyophilized and reconstituted in
0.05% trifiuoroacetic acid and rechromatographed on the same column,
using as the mobile phase modifier 1-propanol containing 0.035%
trifluoroacetic acid. The 1-propanol concentration was increased
linearly (0.05X/min) during 2 h at a flow rate of 1 ml/min at 400C.
Pools of fractions comprising the major EGF-competing activity were
lyophilized.
Assay for TGF
TGF was quantitated in a radioreceptor assay based on
receptor cross-reactivity with mouse submaxillary gland epidermal
growth factor (mEGF). Purified mEGF was labeled with Na 125I.by a
modification of the chloramine-T method as described in Example I.
The 1251-EGF binding assay was performed on formalin-fixed A431 human
carcinoma cells, 8 x 103, in Micro Test II plates (Falcon). The
concentration of TGF was expressed in mEGF ng equivalents/ml and was
based on the amount of TGF required to produce equal inhibition of
1251-EGF binding to A431 cells as a known amount of unlabeled mEGF.
Amino Acid Sequence Determination of TGF
For amino acid sequence analysis, rTGF (3pg) was reduced
with dithiothreitol (20 mM) in 100,CL of Tris-HCI buffer (0.4 M)
-47-
1s~-1~36
1 containing guanidine-HC1 (6 M) and Na2-EDTA (0.1%), pH 8.5, for 2 h
at 500C, and subsequently S-carboxamidomethylated with lodoacetamide
(45 mM) for 30 min at 220C. The S=carboxamidomethylated rTGF was
desalted on a pBondapak C18 column. Peptide was eluted with a
gradient of aqueous acetonitrile containing 0.045% trifluoroacetic
acid. The concentration of acetonitrile was increased linearly
(1%/min) during l h at a flow rate of 1 mi/min at 400C.
Automated sequence analyses (Edman et al. (1967) Eur. J.
Biochem. 1, pp. EI0-91) of S-carboxamidomethylated rTGF and unmodified
mTGF and hTGF were performed with a gas-liquid solid phase
microsequenator (Hewick et al. (1981) J. of Biol. Chem. 256, pp.
7990-7997). Sequenator fractions were analyzed by reverse phase HPLC
(Hunkapiller et al. (1983) Science 219, pp. 650-659).
B. Results
Purification of TGF
Purified preparations of a small molecular weight rTGF,
mTGF and hTGF were obtained from the conditioned medium of
retrovirus-transl'ormed rat and mouse fibroblasts and two human
melanoma cell liiies, respectively. The purification was achieved by
gel permeation chromatography of the acid-soluble EGF-competing
activity on Bio-Gel P-10 in I M acetic acid, followed by reverse
phase HPLC on Bondapak C18 support using sequentially a linear
gradient of aqueous acetonitrile and subsequently I-propanol
containing 0.035% trifluoroacetic acid. The elution patterns of the
final purification step of rTGF, mTGF and hTGF show that EGF-
- 48 -
1341536
1 competing activity co-purified with a distinct -absorbance peak, and
was effectively separated from contaminating UY-absorbing material.
The major protein peak of rTGF, mTGF and hTGF preparations eluted
from a#8ondapak C18 column under standard conditions between 48 and
55 min.
Gel permeation chromatography on Bio-Gel P-10 provided a
separation of the small molecular weight TGFs from larger molecular
weight TGFs and reduced the load of protein applied to a4Bondapak
C18 column in tr7e following purification step. The small molecular
weight TGFs repr!sented 45 to 80% of the initial total EGF-competing
activity. Reverse phase HPLC of TGFs on pBondapak C18 support in the
following two purification steps was very efficient. each giving in a
typical preparation a recovery range of 80 to 100% per step. The
final recovery of the small molecular weight TGFs was approximately
70%, based on th~ maximal total EGF-competing activity detected
during the course of the purification. The average yield of purified
rTGF was 90 ng/liter, of mTGF 50 ng/liter and of hTGF 10 ng/liter of
conditioned mediim. This calculation is based on the specific
activity determiied for isolated TGFs and on the assumption that the
EGF-competing activity measured in the radioreceptor assay reflects
levels of total large and small molecular weight TGFs only. No
immunoreactive mEGF was detected in conditioned medium.
Purity of TGF
The purity of rTGF, mTGF and hTGF, suggested by the
chromatographic elution profiles, was assessed in the EGF
-49-
1341536
1 radioreceptor assay and by amino acid sequence analysis. rTGF, mTGF
and hTGF competed with 125I-EGF for the EGF receptor sites on A431
human carcinoma cells and were qualitatively and quantitatively
nearly indistinguishable from mEGF. Hence, the final TGF
preparations were believed to be highly purified and essentially at
homogeneity. A single amino-terminal sequence was determined by
automated Edman degradation for rTGF, mTGF and hTGF. Any unblocked
minor peptide sequence present at >5% could have been detected by the
methods used. The homogeneity of hTGF was confirmed in addition by
analytical SOS-polyacrylamide gel electrophoresis. The purified
preparation gave one major polypeptide band.
Amino Acid Sequeicing of TGF
The cojnplete sequencing of the rTGF, mTGF and hTGF was
accomplished and the amino acid sequences for the three polypeptides
are given below. It will be noted from the sequences reported that
rTGF and mTGF ar-e identical in chemical make-up and further that
substantial homology exists between the murine TGFs and hTGF with
different amino acid residues occurring at only limited positions in
the sequences.
1) rTaF
5 10
Val-Val-Ser-His-Phe-Asn-Lys-Cys-Pro-Asp-Ser-His-Thr-Gln-
15 20 25
Tyr-Cys-Phe-His-Gly-Thr-Cys-Arg-Phe-Leu-Val-Gln-Glu-Glu-
35 40
Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly-Tyr-Val-Gly-Val-Arg-
45 50
Cys-Glu-His-Ala-Asp-Leu-Leu-Aia
-50-
13 4 1 36
1 - 2) mTGF -
10
Vai-Val-Ser-His-Phe-Asn-Lys-Cys-Pra-Asp-Ser-His-Thr-Gin-
20 25
Tyr-Cys-Phe-Hts-Gly-Thr-Cys-Arg-Phe-Leu-Val-Gln-Glu-Glu-
30 35 40
Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly-Tyr-Val-Gly-Val-Arg-
45 50
5 Cys-Giu-His-Ala-Asp-Leu-Leu-Ala
3) hTGF:
5 10
Val-Val--Ser-His-Phe-Asn-Asp-Cys-Pro-Asp-Ser-His-Thr-Gln-
15 20 25
Phe-Cys--Phe-His-Gly-Thr-Cys-Arg-Phe-Leu-Val-Gln-Giu-Asp-.-
30 35 40 -
Lys-Pro--Ala-Cys-Val -Cys-Hi s-Ser-GIyYTye-Val-G1y-Ala~+Arg-
45 50
10 Cys-Glu--His-Ala-Asp-Leu-Leu-Ala
Example III
Production of TGF in Vivo and its Isolation
Tumor cell lines (1 x 106) known to produce TGFs (human
melanoma and transformed rat) were inoculated into athymic "nude"
15 mice and tumors were allowed to develop. The urine from the tumor-
carrying mice was collected and analyzed for the presence of TGF
using the isolation procedure and analytical techniques given in
Example I above. TGF was detected in the urine of the tumor carrying
mice which has the same size and elution properties on HPLC as does
the cell culture rierived TGF which is described in Examples I and II
above. Further, using the procedures described in Example I above,
the TGF present in the mouse urine was found to have the
characteristic TGF biological properties, in that it stimulates
anchorage-independent growth of cells and binds to the EGF receptor.
Subsequently, the tumors were removed from the tumor-carrying mice
- 51 -
13 41536
1 and the urine of the mice after tumor removal were tested for the
presence of TGF using the above mentioned procedures. In this case.
no TGF was found having the characteristic elution properties on HPLC
or TGF-like biological activity. These results demonstrate that
tumor cells produce TGF in whole animals as well as cell cultures and
that TGF can be detected in and isolated from body fluids using the
process of the invention. Similar experiments were also performed
with rats having chemical carcinogen-induced tumors and they were
found to have TGF in their urine, based on the biological and
biochemical properties listed above, while untreated rats did not.
Example IV
Inhibition of Retroviral Transformed Cell Growth In Vitro
with Antibodies to Antiqenic TGF Oligopeptide
An oligopeptide having the following amino acid sequence
(which corresporids to amino acid sequences 34 through 50 of rat TGF):
Cys-His-Ser-Gly-Tyr-Val-Gly-Val-Arg-Cys-Glu-His-Ala-Asp-
Leu-Leu-Ala
was synthesized using the solid-phase technique of Ohgak et al.
(1983) Journal cf Immunol. Meth. 57, pp. 171-184. This oligopeptide
was then coupled to keyhole limpet hemocyanin in accordance with the
procedure of Baron et al. (1982) Cell 28, pp. 395-404, and used to
immunize rabbits (Baron et al. (1982) Cell 28, pp. 395-404) and sheep
(Lerner (1982) Nature 299, pp. 592-596). Antisera were assayed
against peptide by a peroxidase-iinked immunoassay (Kirkegaard and
Perry Laboratories, Gaithersberg, MD) and against homogeneous rat TGF
-52-
. 13 41536
1. (purified according to Example II above), by immunoprecipitation
(Bister et al. (1980) J. Virol. 36, pp. 617-621) and Western blotting
techniques (Burnett (1981) Analyt. Biochem. 112, pp. 195-203).
Binding of 125I-labeled rat TGF and mouse EGF (Bethesda Research
Labs, Bethesda, MD) to A431 cells grown in 96-well microtiter plates
was as described in Pross et al. (1977) Proc. Natl. Acad. Sci. USA
74, pp. 3918-3921.
Antisera prepared in one rabbit and two sheep reacted with
the corresponding peptide in peroxidase-linked immunoassays in titers
of at least 104. Reactions of the rabbit antiserum with rat TGF were
documented by immunoprecipitation and confirmed.by Western blotting.
The antipeptide antisera did not immunoprecipitate iodinated mouse
EGF in this study.
Human epidermoid carcinoma A431 cells have in excess of 106
EGF receptors per cell (Fabricant et al. (1977) Proc. Natl. Acad.
Sci. USA 74, pp. 565-569). TGF competes with EGF for binding to
these receptors. Binding of 125I-labeled rat TGF to these ceils was
blocked by an excess of unlabeled EGF and by antiserum to peptide. A
blocking effect of antiserum on TGF binding was observed even if the
antibody-TGF complex was not removed from the medium surrounding the
A431 cells by S.aureus protein A-facilitated immunoprecipitation.
Blocking by the antibody was a result of interaction with the TGF
molecule rather than with the receptor, since antisera did not react
in immunoprecipitation assays with purified iodinated EGF receptor.
- 53 -
1341536
As anticipated from immunoprecipitation data, antiserum to peptide
did not Interfere with binding of murine EGF to A431 cells.
With the availability of antisera that blocked cellular
binding of TGF but not EGF, it was possible to test whether TGF
functions by an autocrine mechanism, stimulating the growth of
malignant cells, and whether antibody can thus inhibit such growth in
vitro. Consequently, untransformed normal rat kidney cells (NRK) and
a variety of retroviral cell lines were plated at low densities (500
to 2,000 cells per dish) in serum-containing medium and after a brief
period of adherence were switched to serum-free medium. Sheep or
rabbit antibodies to TGF peptide or to various irrelevant non-~cross-
reacting peptides were added to the medium, and the effect on cell
growth, on microscopic and macroscopic colony formation, and on
colonial morphology was observed. All antibodies were affinity
purified on peptide columns, extensively dialyzed, concentrated, and
reconstituted to an antipeptide titer of 103 to 104. The results are
provided in Table II below.
-54-
. " =
1341536
1 . TABLE II _
EFFECT OF ANTIPEPTIDE ANTIBODIES ON GROWTH OF CELLS IN VITRO
% Inhibition of Colony
Antibody Source Volume Other Formation by
(#i) Addition
NRK CL10 Ki-NRK src-3T3
cells cells
Anti-TGF
oligopeptide Rabbit 5 --- 0 85 NT 53
5 TGF peptide NT 36 NT 6
Anti-TGF
oligopeptide Sheep 2 0 50 49 NT
5 0 68 85 77
10 0 79 73 100
5 TGF peptide NT 9 20 43
5 Irrelevant
peptide NT 85 NT NT
Anti-hetero Sheep &
peptides rabbit 0 <15 S5 0
- 55 -
1341536
As seen in Table II above, neither rabbit nor sheep anti-
TGF oligopeptide inhibited colony formation by NRK cells, but within
48 hours inhibited growth of Kirsten-transformed NRK cells, feline-
sarcoma-virus-transformed rat embryo fibroblasts (CL10 cells). and
Rous-sarcoma-virus-transformed 3T3 cells. (CL10 cells are known to
be prolific producers of TGF (Marquardt et al. (1983) Proc. Natl.
Acad. Sci. USA 80, pp. 4684-4688)). The few surviving colonies in
anti-TGF antibody-treated cells tended to be smaller and lacked the
robust appearancr of normal colonies. Non-adherent unlysed cells
1 o floated free in the medium. This inhibition was partially reversed
when TGF peptide, but not non-cross-reacting irrelevant peptide, was
added concomitantly with antibody to TGF. Rabbit and sheep
antibodies to foLir diffes-er~t non-cross-reacting peptides had no
effect on growth of colonies of NRK or retroviral transformed cell
lines. Replacement of antibody-containing medium by fresh antibody-
free medium after- 72 hours failed to reverse the inhibition of colony
formation, but tfe surviving colonies grew vigorously.
Example V
Synthesis and Characterization of rat TGF
The chemical synthesis of rat TGF, having the chemical
formula given in Example II, was performed manually by the stepwise
solid-phase approach according to the general principles described by
Merrifield (1963) J. Amer. Chem. Soc. 85, pp. 2149-2156. The
differential acid-labile protecting group strategy was adopted for
this synthesis with the conventional combination of
-56-
1341536
I tertbutyloxycarbonyl for N-amino terminus and benzyl alcohol
derivatives for the side chains. A more acid stable benzyl ester
linkage that anchored protected amino acids to the polymeric support
was used to minimize loss of peptides during the repetitive acid
treatments (Mitchell et al. (1976) J. Amer. Chem. Soc. 92, pp. 7357-
7362). Complete deprotection and removal of peptide from the resin
was by the low-high HF method of Tam et al. (1983) Tetrahedron Lett.
23, pp. 4435-4438, which differed from the conventional HF
deprotection method and removed benzyl protecting groups by the SH2
mechanism in dilut:e HF solution to minimize serious side reactions
due to carbocatioias generated in the conventional SN1 deprotection
method. Furthermore, it is also designed to reduce many cysteinyl
side reactions that often hamper the synthesis of proteins containing
multiple disulfide linkages.
After H(- treatment and prior to any purification, the crude
and reduced synthetic rTGF was oxidized and regenerated by the mixed
disulfide method =in the presence of a combination of reduced and
oxidized glutathione (Ahmed et al. (1975) J. Biol. Chem. 250, pp.
8477-8482). This avoided the formation of polymeric materials during
purification. The regenerated, crude rTGF contained 40-50% of EGF-
radioreceptor and tyrosine-specific protein kinase activities when
compared to the natural rTGF. Crude synthetic rTGF was purified to
homogeneity in three steps: (1) gel filtration on a Bio-Gel P-10
column; (2) ion-exchange chromatography on a CM-Sephadex column: and
(3) preparative high pressure liquid chromatography on a C18 reverse
-57-
1341536
1 phase column. An overall yield, based on starting loading of Ala to
resin, was 31%.
Under reducing or nonreducing conditions, the purified
synthetic rTGF was found to give a single band with an apparent
molecular weight of 7000 on SDS-PAGE electrophoresis. Amino acid
analysis by 6N HC1 and enzymatic hydrolysis provided the expected
theoretical molar ratio of the proposed sequence. No free thiol was
detected by Eliman's method of sulhydryl determination on synthetic
rTGF, but upon thiolytic reduction, the expected theoretical value of
six cysteines was obtained. These findings support the conclusion
that synthetic rTGF is_a single chain polypeptide containing six
cysteines in disulfide linkages, which is in agreement with the
expected chemical properties of the natural rTGF. Additionally,
synthetic rTGF coeluted with the natural rTGF as a single symmetrical
peak in C18 reverse phase HPLC.
Synthetic rTGF prepared in ac:.ordance with this Example was
compared with natural rTGF in three assays for biological properties
of the putative transforming growth factor. In the mitogen assay,
the stimulation of growth of serum-deprived normal rat kidney cells
by rTGF was measured by the incorporation of 125I-Iododeoxyuridine.
In the soft agar assay in the presence of fetal bovine serum and a
second TGF, TGF-beta, the morphological and phenotypic alterations by
rTGF could be quantitated by colony formation in soft agar. The
latter transforming assay has been shown to correlate well with
tumorigenicity (Stoker et al. (1968) Int. J. Cancer 3, pp. 683-693).
-58-
1341536
1 Fetal bovine serum or TGF-beta alone does not i.nduce transformation
of NRK cells in culture. Similarly, TGF, natural or synthetic, does
not produce such an effect in the absence of TGF-beta. Both
synthetic and natural rTGF displayed si.milar dose response curves*and
half maximal activities in these two assays.
Since rTGF competes with mEGF for the binding of EGF
receptors on cellular membranes, synthetic rTGF was compared with the
natural rTGF for binding on A431 human carcinoma cells. Again, the
response and activities of the tiatural and synthetic rTGF were found
to be indistinguishable from eac:h other. The concentration required
for 50% inhibition of 1251-EGF binding was found to be 3.5 and 4.1 nM
for the natural and synthetic rl-GF respectively. A consequence of
TGF or EGF binding to the EGF membrane receptors is the stimulation
of phosphorylation of tyrosine t=esidues of synthetic peptides or
endogenous substrates (Pike et <<l. (1982) J. Biol. Chem. 275, pp.
14628-14631). Synthetic rTGF w<<s found to stimulate the
phosphorylation of the synthetic: angiotensinyl peptide substrate with
a half maximal activity of 0.3 riM, an activity comparable to the
value for natural rTGF, reportecf by Reynold et al. (1981) Nature 292,
pp. 259-261.
Example VI
Wound Healing Using TGFs
Human EGF (hEGF), rat TGF, analog of human TGF, natural
vaccinia virus growth factor (VGF) and recombinant VGF were used in a
wound healing test, according to the procedure below, to determine
-59-
134 1536
1 the healing effects of each factor on second degree burns. The rat
TGF was synthesized as described above in Example V and had the
chemical formula given in Example II. The analog of human TGF, which
was prepared using standard recombinant techniques, had the following
amino acid sequence:
5 10
Val -Val -Ser-Hi s-Phe-Asn-Asp-Cys-Pro-Asp-Ser-Hi s-Thr-
20 25
Gin-Phe-Cys-Phe-His-Gly-Thr-Cys-Arg-Ser-Leu-Val-Gln-
30 35
Giu-Glu-Lys-Pro-Ala-Cys-Val-Cys-His-Ser-Gly-Phe-Val-
40 45 50
Gly-Val -Arg-Cys-Glu-His-Ala-Asp-Leu--Leu-Ala
10 Natural VGF was purifiecl as described below in Example XV.
Recombinant VGF was also produced using standard recombinant
techniques. The natural VGF is a 25 kd protein containing the
following amino acid sequence wHch falls within the scope of formula
I above:
5 10
15 Leu-Cys-Pro-Glu-Gly-Asp-Gly-Tyr-Cys-Leu-His-Gly-Asp-
15 20 25
Cys-Ile-His-Ala-Arg-Asp-Ile-Asp-Gly-Met-Thr-Cys-Arg-
30 35
Cys-Ser-His-Gly-Tyr-Thr-Gly-Ile-Arg-Cys-Gln-His-Val-
Val-Leu-Val
The full amino acid sequence for the VGF molecule purified as
20 described below in Example XV (iind that expressed via recombinant
technique) is as follows:
5 10
Asp-Ser-Gly-Asn-Ala-Ile-Glu-Thr-Thr-Ser-Pro-Glu-Ile-
15 20 25 -
Thr-Asn-Ala-Thr-Thr-Asp-Ile-Pro-Ala-Ile-Arg-Leu-Cys-
30 35
Gly-Pro-Glu-Gly-Asp-Gly-Tyr-Cys-Leu-His-Gly-Asp-Cys-
40 45 50
25 Iie-His-Ala-Arg-Asp-I1e-Asp-Gly-Met-Tyr-Cys-Arg-Cys-
-60-
115 41536
55 60 " 65
1 Ser-His-Gly-Tyr-Thr-Gly-Ile-Arg-Cys-Gin-His-Val-Val-
70 75
Leu-Va1;~-Asp-Tyr-Gln-Arg-Ser-Glu-Asn-Pro-Asn-Thr-Thr-
80~ 85 90
Thr-Ser-Tyr-Ile-Pro-Ser-Pro-G1y-Ile-Met-Leu-Val-Leu-
95 100
Val-Gly-Ile-Ile-Ile-Ile-Thr-Cys-Cys-Leu-Leu-Ser-Val-
105 110
Tyr-Arg-Phe-Thr-Arg-Arg-Thr
Three female piglets of approximately 10 pounds each were
anesthetized with Ketamine and Rompum and their backs were shaved and
the remaining hair was totally removed with a commercial depilatory
cream. A brass template (3x3 cin, 147 gm) was equilibrated in a 700C
water bath and then placed in firm contact with the bare skin for
exactly 10 seconds. Five wounds were placed on each side of the
spine and were separated from each other= by approximately 1 inch.
The top of each resulting blister was totally removed and the wounds
were treated twice a day with Silvadene alone, Silvadene containing
one of the growth factors, or l-eft untreated. The piglets were
allowed to eat and drink at will.
After 9 or 10 days, the piglets were anesthetized and the
eschar from each burn was removed. All burns were photographed and a
punch biopsy was taken within each burn.
The following table indicates the approximate percentage of
each burn that was epithelialized using visual judgment.
;, - 61 -
134153~
TABLE III
Natural VGF pg/ml
Right Side Silvadene Untreated 0.1 0.1
15% 0% 70% 65%
Pig 1
9 Days Post-Burn hEGF ic /mg 1
Left Side Silvadene Untreated 0.1 0.1 0.1
75~ 55X 60% 70% 30%
Recombinant VGF yg/m1
Right Side Silvadene Untreated 0.1 0.5 1.0
501 0% 95% 95% 60%
Pig 2
10 Days Post-Burn
Left Side Silvaiiene Untreated 0.1 0.5 1.0
50't 0% 60% 75% 40%
Rat TGF p/mg 1
Right Side Silvadene Untreated 0.1 0.5 10
20% 15% 0% 65% 90%
Pig 3
9 Days Post-Burn Analog of Human TGF ~/mg 1
Left Side Silvadene Untreated 0.1 -0.5 10
25% 5% 90% 85% 65%
-62-
13 41536
1. As shown in Table III, the TGFs were-very effective in
healing wounds, when compared to the untreated controls or even when
compared to Si]vadene alone.
Example VTI
Wound Healing
Another experiment was performed to measure the effect of
the analog of human TGF described above on the healing of second
degree burns. The experimental conditions were similar to those
described above. However, the following changes were made in the
procedures. One Yorkshire piglet was anesthetized with Ketamine and
each of 12 burns was made using the template for 11 seconds per burn.
After each blister was removed, the wounds were treated once a day
with one of the following:
a. 1pg/ml of TGF in Silvadene;
b. 0.1 pg/ml of TGF in Silvadene;
c. untreated;
d. Silvadene alone;
e. 1pg/ml of hEGF in Silvadene; and
f. 0.1 pg/mT of hEGF in Silvadene.
After 7 days, the esc.har was removed. The percentage of
wound healing, calculated from planimetry results, is shown below.
TGF was very effective as a wound healer.
-63-
13 41536
TABLE IV
TREATMENT PERCENTAGE OF WOUND HEALED
Burn A Burn B
TGF 1 pg/ml 9 26
TGF 0.1 /cg/ml 34 57
Untreated 3 6
Silvadene 14 21
hEGF I g/m) 14 17
hEGF 0.1 /ig/ml 17 14
Example VIII
Corneal Wound Healing of rTGF
A. Preparation of rTGF polypeptide
The TGF polypeptide was synthesized based on the amino acid
sequence reported in Example l.I above; for TGF purified from the
conditioned medium of Fisher r=at embryo fibroblasts transformed by
feline sarcoma virus. The chemical synthesis of the rTGF was
performed as described in Exaniple V above.
,
B. Preparation of treatment formulation
The rTGF-a polypeptide was combined with isotonic (285 m
osmoles) sterile phosphate buffered saline (pH 7.4) at a
concentration of 50 ng/ml.
C. Corneal Stromal Incisions
Totally penetrating incisions 5 mm in length, which
extended into the anterior chamber along their entire length, were
made in the center corneas of adult female Macaca fasicularis
primates. The right eyes served as controls and were treated three
times each day with two drops of isotonic (285 m osmoles) sterile
-64-
1341536
1 phosphate buffered saline (pH 7.4) without rTGF: The left eyes were
treated on the same schedule with the treatment formulation described
above. After three days of treatment, the strength of the wounds was
quantitatively measured by inserting a small bore needle (25 gauge)
into the anterior chamber through the limbus of the cornea. The
needle was connected to an aneroid manometer, and the pressure was
slowly and steadily increased until the wounds first began to leak,
and then burst. This procedure is described in detail by Weene
(1983) Anal. Ophthalmol. 15, 438,
D. Results
The results shown in Tiible V demonstrate that the bursting
strength of TGF treated corneas 'is significantly stronger than the
saline treated control corneas.
TABLE V
mm of Hg
Control Right Eye TGF-treated Left Eye
Leak Burst Leak Burst
Monkey P-21 25 30* 210 225*
Monkey P-20 35 90* 155 >300*
*t = 40.0, p 0.025, paired T-test.
-65-
1.3 41536
I Example IX -
In vivo Studies with anti-TGF Antibody
An anti-TGF serum was prepared by immunizing a rabbit with
a glutaradehyde conjugate of the 17-amino acid sequence of rTGF
described in Example IV above, and KLH (keyhole limpet hemocyanin),
then the Ig fraction was prepared by precipitating the serum twice
with a 45% saturated solution of ammonium sulfate, redissolving it to
the original volume, and dialyzing against PBS. -
The anti-TGF serum and a control serum were diluted from
their original volume of 0.5 ml to 3 ml. The resulting volumes were '
used to inject mice intraperitoneally at a dosage'of 0.1 ml per
mouse. Ten mice, about 3 mo!iths old, were.each previously
subcutaneously transplanted with two small pieces of a rat tumor
derived from the Snyder-Thei'!en feline sarcoma virus. (Each
transplant was counted as on<! site, which results in 20 total sites
with 10 sites for the control group and 10 sites for the treated
group.) The mice were injected starting the day after the transplant
and also were injected on days 4, 7 and 11.
Starting on the seventh day, tumor diameters were measured
in two directions with calipers. The tumors were measured at regular
intervals on days 9, 11, 14, 16, 18, and 24.
At the time of the first measurement, 7 of the 10 sites for
the control group had evidence of tumor growth, while only 4 of the
10 sites for the treated group had tumor growth. In addition, 6 of
the 10 sites in the control group had tumors that were at least 3 mm
-66-
13 41536
1 in diameter, while only 3 sites in the treated group had tumors of
that size. Four sites in the control group were at least 5 mm in
diameter on day 9, while only 2 sites in the treated group were that
size. By day 11, the control group and the experimental group had
similar tumors.
Example IX
Detection of TGF Activity in Urine of Cancer Patients
Various samples of urine were pretreated at 900C for one
minute after adding 1/9 volume of 20% SDS. 0.4 M DTT, 0.4 M Hepes,
.08 M sodium chloride, 1 mM EDTA, to a pH of 7.4. A 20 pL sample was ~
added to an incubation mixture (50 uL to al ) modified to contain 1%
NP-40. Incubation of sample, a,itibody raised to the 17-amino acid
sequence of rat TGF in Example [V and 125a-labelled synthetic 17-
amino acid sequence was performed in 96 well plates for 60 minutes,
followed by 30 minutes with Pan:;orbin. Antibody-bound peptide was
pelleted on a cushion of dibuty'l phthalate and pellets were counted
in a gamma counter using 1/8 inch thick lead sleeves to shield the
unpelleted isotope.
TGF levels were standardized using the synthetic sequence
and human melanoma (A375 cell l'ine) TGF from concentrated serum-free
conditioned culture medium. TGF levels were calculated as ng of TGF
per mg of creatinine. In each experiment, one or more normal samples
were assayed and the average normal value was assigned a value of one
Relative TGF Equivalent.
- 67 -
1341536
1 A. Sample Preparation
Fresh, unclarified urine samples were stored in 25 ml
aliquots at -700C for up to six months. In most cases, the samples
were rapidly thawed and immediately treated with protease inhibitors
( 0.5 mM phenylmethylsulfonyl fluoride, 0.05 mM pepstatin A, Sigma
Chemical Co.) for 30 minutes at 40C. Ten ml samples were then
dialyzed at 40C against either 0.1 M acetic acid or 0.1 M ammonium
bicarbonate. Dialysis tubing with three different pore sizes was
used: 3,500; 6,000-8,000; or 12,000-14,000. (In some cases, the urine
was initially lyophilized and extracted, but the neutralization and
ethanol/ether precipitation steps were omitted.)
After 2 to 4 days of dialysis, the urines were clarified
(10,000 xg for 10 minutes), lyaphilized, redissolved at 50 times the
original concentration in 5 mM formic acid, and 50 mM sodium
chloride, then clarified briefly (12,000 xg, one minute) prior to
pretreatment.
B. Creatinine Assays
An untreated aliquot of each sample was tested in duplicate
for creatinine, using a manual colorimetric test.
C. Results
The samples were standardized for creatinine levels. For
each of 6 experiments, statistical analysis (T-test) indicated
significant differences between cancer patients and normals (P< .001,
.001, .001, .001, .01, .05). In the six experiments, TGF
concentration in urines from normal individuals averaged .18 ng of
-68-
1341536
TGF per mg of creatine, while the urines from all cancer patients
averaged .64 ng of TGF. Elevated levels of TGF was observed in most
cancer patients. Results in each experiment were also normalized
relative to the average normal value in that experiment. Data are
expressed numerically in Table VI. Using an arbitrary cutoff of
twice the average normal level of TGF, in 81% of the various cancers
tested, higher than normal TGF levels were detected.
-69-
.~
13 41536
1- TABLE VI: DETECTION OF ALPHA TGF ANTIGEN IN URINE
PATIENT GROUP OVERALL SAMPLE PROCESSING PH
POSITIVE AND RIA CONCENTRATION FACTOR
pH 2 pN 8
9x 18x 9x 18x
Apparently healthy
controls 1f18 (6%) 0/6 1/12 1/15 0/15
Patients with benign
conditions 1/3 0/3 1/3 0/1 0/1
colon (villous adenoma) 0/1 0/1 0/1
breast (fibrocystic) 0/1 0/1 0/1 0/1 0/1
pregnancy (normal, 38 wks) 1/1 0/1 1/1
Patients with cancer 26/32 17/28 22/31 9/16 12/16
(81%) (61%) (71%) (56%) (75%)
Lung 11/13 11/13 12/13 3/4 4/4
(85%) (85%) (92%) (75%) (100%)
Gastrointestinal 4/7 4/7 4/7 3/7 3/7
(57%) (57%) (57%) (43%) (43%)
Urogenital 3/3 1/3 1/3 2/3 3/3
(100%) (33%) (33%) (67%) (100%)
Breast 5/5 1/1 5/5
(100%) (100%) (100%)
Lymphoid 3/4 0/4 1/4 2/2 2/2
(75%) (0%) (25%) (100%) (100%)
*Cut off 2 x average normal value
1341536
1 Example X
Detection of TGF Activity in Urine of Cancer Patients
Human urine from 50 people (25 normals and 25 individuals
having advanced cancer) was subject to immunoassay using antibody to
TGF using the procedure described above in Example IX. In this
study, 100 ml of urine from the individuals under test was collected,
dialyzed against 1.0 molar acetic acid, and concentrated about 100-
fold. This concentrated urine was then subject to an immunoassay
using rabbit polyclonal antibody raised to the 17 amino acid
oligopeptide as in Example IV above, and the result$ are reported in
the table below.
TABLE VII
DETECTION OF TGF ACTIVITY IN URINE
OF CAPlCER PATIENTS
DIAGNOSIS NO. POSITIVE / NO. TESTED
Lung Cancer 4 / 5
Breast Cancer 4 / 5
Colon Cancer 3 / 5
Melanomas 5 / 5
Leukemias 0 / 5
Normal Conditions 1 /25
In this test, the level of TGF in urine of the majority of cancer
patients was at least 5-fold higher than that present in the normal
individuals under test.
- 71 -
13 41536
Example XI
Detection of TGF with a Monospecific Antiserum Directed Against a
Synthetic Peptide
Peptide Synthesis
Peptides, corresponding to amino acids from portions of
rTGF amino acid sequence, described above in Example II, were
synthesized commercially (Peninsula Labs) by the standard solid phase
technique of Ohgak et al. (1983) Journal of Immunol. Meth. 57, pp.
171-184. If necessary, peptides were purified by reverse phase high
performance liquid chromatography (HPLC) prior to use.
The sequences used were:
Peptide I - Val-Val-Ser-His--Phe-Asn-Lys-Cys-
Pro-Asp-Ser-Hi s--Thr
Peptide II - Cys-His-Ser-Gly-Tyr-Val-Gly-Va1-
Arg-Cys
Peptide III - Cys-His-Ser-Gly-Tyr-Val-Gly-Val-
Arg-Cys-Glu-His-Ala-Asp-Leu-Leu-Ala
Peptide IV ~ Val-Gly-Val-Arg-Cys-Glu-His-Ala-
Asp-Leu-Leu-Ala
Peptide Conjugation and Immunization
A sample of peptide (10 mg) was mixed with keyhole limpet
hemocyanin (10 mg; Calbiochem) in 3.5 ml of 0.1 M sodium phosphate at
pH 7.4. Four milliliters of 25 mM glutaraldehyde was added, and the
mixture was incubated, with shaking, for 1 hr at 230C. Glycine was
then added to a final concentration of 0.1 M and the mixture was
-72-
1341536
-1 shaken overnight at 230C. The resulting conjugate was mixed with an
equal volume of Freund's complete adjuvant prior to injection.
Rabbits were irtmunized by subcutaneous Injection of 1 mg of
conjugate Injected subcutaneously at four separate sites. Two
booster injections were administered at bi-weekly intervals following
the initial injection. The serum used in this study was collected
eighty days after the initial injection. An immunoglobulin fraction
of this serum was prepared by ammonium sulfate precipitation. The
product of antibodies specific for immunizing peptide was monitored
1 0 with a solid phase enzyme-linked immunoabsorbent assay.
Radioiodination of Peptides
Peptides were labeled with Na125I using the chloramine-T
procedure, essentially as described in Das et al. (1977) Proc. Natl.
Acad. Sci. 74 pp. 2790-2794. Solutions containing mouse submaxillary
gland EGF (mEGF) (170 p moles), synthetic rTGF (4 p moles) or peptide
III (400 p moles) were mixed with 1-2 mCi Na 1251 (Amersham, 2 mCi/n
mole) in 2 M potassium phosphate at pH 7.5. Chloramine T (50-100 ug)
was added and incubation at 40 was continued for 1 minute (mEGF and
rTGF) or 7 minutes (peptide III). Reactions were terminated by the
addition of Na2S205 (10-20 ug) and labeled proteins were separated
from unreacted Na1251 by chromatography on Sephadex G-10 (Pharmacia).
Specific activities of peptides labeled in this fashion were: 1 x
1010 cpm/n mole (EGF); 6 x 108 cpm/n mole (rTGF); and 1 x 109 cpm/n
mole (peptide III).
* Trade Mark
- 73 -:
~. ..,..,
13 41536
1. Radioreceptor Assay -
The binding of 1251-EGF to its receptor on monolayers of
formalin-fixed A431 cells was measured as described above in Example
1. 1251-EGF was added at a final concentration of 0.33 nM in the
presence or absence of competing substances. Addition of unlabeled
EGF at 0.3-0.5 nM resulted in half-maximal inhibition of 125I-EGF
binding. TGF concentrations were expressed as the amount required to
produce an inhibition of 125I-EGF binding, equivalent to a known
amount of TGF.
Radioimmunoassay
Reactants were mixed in a final volume of 50 microliters of
a solution containing: 20 mM sodium phosphate, at pH 7.4; 200 mM
NaC1; 40 mM dithiothreitol; 0.17,(w:v) BSAs 0.1% (w:v) NaN3;.125I-
peptide III; (104 cpm); antiserum at a final dilution of 1/15,000;
and other additions, as specified. The reaction was initiated by the
addition of antiserum and continued at 230C for 90 minutes. An equal
volume of 10% formalin-fixed Sti.phalococcus A (Pansorbin, Calbiochem)
was then added and incubation w<<s continued for an additional 30
minutes at 230C. The immunoadscirbant was removed by sedimentation
through a cushion of 10% (w:v) s,ucrose and the amount of bound 125I-
peptide III was determined using a gamma counter. Under these
conditions, approximately 25% of 125I-peptide was bound by antibody
in the absence of competitor. The amount of bound peptide III was
corrected for non-specific binding measured in the absence of
antibody (less than 5% of the total) and expressed as a percentage of
-74-
~3415~6
maximal binding. Concentrations of competitors-were expressed as the
amount of peptide III required to give an equivalent inhibition of
precipitation.
Chromatography
Gel filtration chromatography was performed according to
the manufacturer's instructions on columns of Bio Gel*P-10 (BioRad)
*
equilibrated in 1 N acetic acid. HPLC was performed on a Novapak C18
column (0.39 x 10 cm; Waters Associates) using a flow rate of 1
ml/min at 230C.
1 0 Preparation of Conditioned Medium
Serum-free conditioned medium from Snyder Theilen-
transformed Fischer rat embryo cells (ST-FrSV-FRE clone-10) was
collected as previously described in Twardzik et al. (1983) Virology
124, pp 201-207. The medium was clarified by low speed
centrifugation and lyophilized. The residue was then resuspended in
1 N acetic acid and dialyzed extensively against 0.1 N acetic acid.
Insoluble protein was removed by centrifugation and the supernatant
was lyophilized. Finally, the residue was resuspended in one-
hundredth the original volume of 1 N acetic acid and stored at 40C.
Immunoblotting Analysis
Samples to be analyzed were first subjected to sodium
dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) on a
15-20% acrylamide gradient, and then run under reducing conditions,
using the system described by Laemmli (1970) Nature 227, pp. 680-682.
Following separation, proteins were electrophoretically transferred
* Trade Mark
- 75 -
A~
1341536
1. to nitrocellulose (Schleicher and Schuell, BA85, 0.45 u) as described
by Burnette (1981) Anal. Biochem. 112, pp. 195-203. Transfer was
accomplished at approximately 5V/cm for a period of 3-4 hours at
230C. The resultant protein blot was incubated overnight in a milk-
based cocktail, BLOTTO (Johnson et al. Gene Anal. Techn. 1. pp. 3-8).
Antiserum was diluted in BLOTTO and added to the blot in the presence
or absence of excess peptide III. Incubation with antibody was
continued with frequent agitation for 2-3 hours at 230C. The blot
was then washed and incubated with 1251-protein A (4 x 106 cpm/ml;
4 x 107 cpm/#g) for one hour at 230C, and washed with BLOTTO.
Antibody binding sites were visualized following autoradiography on
Kodak* XAR-5 film at -700C using intensifying screens.
Results
Using the radioimmunoassay described above, input 125I_
1 5 peptide III was nearly quantitatively precipitated at low dilution of
antisera; the antiserum showed a titer of >10,000. Antibody affinity
was measured by incubating the antiserum at a final dilution of
1/5,000 with varying concentrations of 125I-peptide III. Analysis of
these data by the method of Scatchard revealed a single class of
binding component(s) having an affinity constant (Ka) of 8.3 x 108
M-1 for 125I-peptide III.
Specificity of Radioimmunoassay
To determine the specificity of the radioimmunoassay,
various peptides were tested for their ability to inhibit
precipitation of 125I-peptide III. Unlabeled peptide III inhibited
* Trade Mark
-76-
~~~
1341535
1. the precipitation of 125I-peptide-III in an approximately linear
fashion in the range of 0.13 to 11 nM; the concentration of unlabeled
peptide III, which produced half-maximal inhibition of precipitation.
was 0.7 nM. Peptide IV was only slightly less effective an inhibitor
of precipitation than was peptide III, while peptide II was totally
ineffective at concentrations up to 1.1 uM. Peptide I, corresponding
to the amino terminal 11 residries, also was ineffective as a
competitor. These results indicate that the epitope detected under
standard assay conditions was '{ocalized to the carboxy-terminal 11
amino acids of peptide III. A summary of the relative inhibitory
concentrations of various pepti'des is presented in'Table VIII.
TE-SLE VIII
Addition Relative Inhibitory Activity
PeptideIII 1
Peptide IV 3.5
Peptide I >1500
Peptide II >1500
rTGF + dithiothreitol 1
rTGF - dithiothreitol 20
mEGF >1500
The indicated additions were tested for their ability to inhibit the
standard RIA as described above. The concentration of each addition
necessary to achieve half-maximal inhibition of precipitation was
determined and is expressed relative to the concentration of peptide
III required to give the same level of inhibition. The symbol ">"
indicates the highest concentration tested. (The inhibitory
activities of peptides I-IV were unaffected by the inclusion of
diethiothreitol in the assay.)
rTGF was also tested for its ability to inhibit
precipitation of 1251 labeled peptide III. Synthetic rTGF, which was
indistinguishable from the native molecule in biologic activity, was
-77-
1341 5
36
1 equally effective (on a molar basis) an inhibitor as was peptide III.
The relative inhibitory capacity of rTGF was considerably diminished
when a reducing agent was omitted from the reaction mixture (Table).
mEGF was totally ineffective as a competitor, at concentrations
ranging up to greater than 1 M. This indicates that the RIA
described here differs from other available assays for TGF in that it
is specific for TGF, but not EGF.
In order to show directly that the antiserum binds rTGF,
125I-labeled rTGF was substituted for 125I--peptide III in the
standard RIA. A dose dependent precipitation of 1251-TGF was
observed, which was inhibited substantially by the addition of excess
unlabeled peptide III. Scatchard analysis of the binding data
obtained in this fashion indicated that the antiserum exhibited a Ka
of 2.3 x 108 M-1 for 1251-TGF. This value was in reasonable
agreement with the Ka determined for 1251-peptide III (8.3 x 108 M-1)
and indicated that the antiserLm binds to rTGF.
Detection of rTGF in Conditioned Medium of Transformed Cells
One source of TGF is the culture medium of cells
transformed by RNA tumor viruses. In order to detect rTGF in
conditioned medium from culturcd transformed cells, the following
procedure was used. A Fischer rat embryo cell line transformed by
the Snyder-Theilen strain of feline sarcoma virus (ST-FeSV FRE clone-
10), has previously been shown to produce elevated levels of rTGF
(Gray et al. (1983) Nature 303, pp. 722-725). Serum-free conditioned
medium was collected, processed and concentrated as described above.
-78-
1 536
1 An aliquot of medium was then subjected to gel _filtration
chromatography on 6io-Gel P-10 under acidic conditions. Fractions
were analyzed for rTGF by the EGF receptor competition assay or by
RIA.
Two size classes of EGF-competing activity were detected
under these conditions, one of approximate Mr = 10,000 and one of
Mr = 20,000; both species eluted well behind the bulk of protein in
the sample. Both size classes of EGF-competing activity also showed
imnunologic activity. Additional immunologic activity was found in
the excluded volume of the column where no EGF-competing activity was ~
detected. These findings indicate that previously described size
classes of rTGF are active in the RIA. The ratio of EGF-competing to
immunologic activity was greatly reduced in the high molecular weight
TGF fractions, indicating that the biologic activities of these
TGF(s) species are less than that of the low molecular weight
fractions.
To confirm that the immunologic activity was carried by the
same molecular species as EGF-competition activity, pooled fractions
containing the Mr = 10,000 and Mr = 20,000 size classes of rTGF from
a preparative scale version of this experiment were subjected to
reverse phase chromatography on HPLC; conditions employed were
similar to those used in the purification of rTGF as described in
Marquardt et al. (1983) Proc. Natl. Acad. Sci., 80, pp. 4684-4688.
For the Mr = 10,000 TGF pooled fractions, both EGF-competing and
immunologic activity were shown to co-purify with essentially
- 79
-
.
16 4 1 5 6
quantitative yields. The co-purification of both activities during
both gel filtration chromatography and HPLC strongly suggested that
both activities are carried by the same molecular species. When the
Mr = 20,000 EGF-competition activity was subjected to HPLC, a similar
co-purification of EGF-competing and immunologic activities was
observed. These experiments indicate that the bulk of immunologic
activity found in conditioned medium of ST-FeSV-FRE-clone 10 was due
to rTGF.
Immunoblotting analysis of different size classes of rTGF
The larger size classes of TGFs found in conditioned medium
from retroviral transformed rat cells could represent either
aggregated states or distinct molecular forms of the TGF molecule.
To distinguish between these alternatives, synthetic rTGF and both
the Mr = 10,000 and Mr = 20,00.) size classes of native TGF were
subjected to immunoblotting an.alysis, following separation of
component polypeptides by SDS-PAGE under reducing conditions.
Immunoreactive synthetic rTGF and low molecular weight native rTGF
co-migrated as single polypept-ide chains of Mr = 6,000, immunologic
reactivity of these molecules ivas blocked by incubation of antiserum
with excess peptide III. In contrast, large molecular weight size
class contained three immunoreacted peptides of Mr = 24,000, Mr =
40,000, and Mr = 42,000; immunologic reactivity of these peptides was
also blocked by addition of excess peptide III. Another radioactive
band having a migration of greater than Mr = 43,500 was noted in all
lanes. This material was incompletely removed by inclusion of excess
-80-
13 41536
1 peptide III and was seen consistently in all experiments, regardless
of the sample analyzed; therefore, it seems to represent an
experimental artifact. It is noteworthy that no Mr = 6,000 TGF was
detected in the large molecular weight size class. These results
demonstrate that the higher molecular weight size classes of rTGF
represent distinct forms of TGF.
Example XII
Synthetic Fragment of rTGF with Receptor Binding and Antigenic
Properties
Peptides
EGF was purified from mouse submaxillary glands by
extraction with 1 M HC1 containing 1% trifluoroacetic acid (TFA), 5%
formic acid, and 1% NaCI, concentration on Sep-Pak*columns (Waters),
and reversed-phase high performance liquid chromatography on
pBondapak C-18 columns (Waters) as in Elson et al. (1984)
Biochemistry Int. 8, pp. 427-435. The complete synthesis of native
rat TGF has been described above in Example V.
The synthetic peptide fragments were prepared on
chloromethyl-polystyrene-1X divinylbenzene resin (Bio-Rad) using
Na-t-butoxycarbonyl protection. Peptides were deprotected and
cleaved from the resin using HF or, for the analogs with blocked C-
termini, the peptide was first removed by ammonolysis (NH3/MeOH).
The crude deprotected peptides were diluted and cyclized to the
disulfide form by oxidation with 0.01 M K3Fe(CN)6. The peptide
solution was loaded on a Bio-Rex 70 cation exchange column, washed
* Trade Mark
- 81 -
A'
13 41 536
1 with 300 mL of H20, and eluted with a gradient to 50% AcOH. The
peptide was purified by prep-HPLC on a 2.5 x 100 cm column (Altex) of
Vydac 218TP C-18 packing using approximately 20% CH3CN eluent, 0.03 M
in NH4OAc at pH 4.5 (10). Peptides 1 and 2 represent the amino acid
sequences 34 through 43 of rTGF with free or blocked (N-Ac; C-amide)
ends, respectively. Peptide 3 is the methyl ester of the
corresponding loop of hTGF (Ac-Cys-His-Ser-Gly-Tyr-Val-Gly-Ala-Arg-
Cys-OMe) prepared by transesterification from the resin (base/MeOH).
Immunogen Preparation and Immunoassay
TGF peptide 1 was coupled to keyhole limpet hemocyanin
(KLH) or bovine serum albumin (BSA) by thiol/maleimide linkage (King
et al. (1979) J. Immunol. Methods 28, pp. 201-206). Carrier-peptide
complexes were purified by gel filtration. An average of 60 moles
peptide/mole KLH and 20 moles peptide/mole BSA was achieved. Rabbits
were immunized with the KLH-peptide 1 conjugate by multiple
subcutaneous (s.c.) and intramuscular (i.m.) injections of 2 mg
protein in complete Freund's adjuvant. Rabbits were boosted s.c.
every two weeks with the immunogen in incomplete Freund's adjuvant.
The IgG fraction of the antiserum after 5 boosts was used for
immunoassays and radiolabeled using Na125I (NEN) and Enzymobeads
(Bio-Rad) to 4 x 105 cpm/ g protein. Polyvinyl chloride 96-well
plates (Costar) were coated with 200 pg/mL BSA-peptide 1 conjugate
and countercoated with 10% normal rabbit serum in phos.phate-buffered
saline. Plates were incubated for 4 hours at 370C with [125I]anti-
* Trade Mark
-82- a
~~~ap
1-v 415 36.
KLH-peptide I IgG (50,000 cpm/well) in the presence or absence of
inhibitors, washed and individual wells were counted.
Radioreceptor Assay
A-431 human epidermoid carcinoma cells or human foreskin
fibroblasts (HFF), established from primary cultures, were grown to
confluence in 24-well cluster dishes (Costar) in Dulbecco's minimal
essential medium (DMEM, Gibco) containing 10% fetal calf serum (FCS,
Hyclone). EGF was radio-iodinated as above to between 4 x 104 and 8
x 104 cpm/ng protein. Cells were incubated at 40C for 60 minutes
with 1 nM [125I]EGF in the presence of inhibitor peptides in DMEM (pH
7.4; 20 mM Hepes) containing 0.1% BSA. Cells were washed 3x with
cold buffer, lysed with 0.1 N PIaOH, and cell-associated radioactivity
was measured (ry-counter). Nons;pecific binding assessed in the
presence of a 100x excess of ccild EGF was less than 5% and 10% of the
specific binding for A-431 and HFF cells, respectively.
Cell Prol i feration Assay
HFF cells were grown to confluence in 48-well cluster
dishes in DMEM-10% FCS and brotight to quiescence by starvation for 2
days in DMEM-0.5% FCS. Mitogens and peptides were incubated with
cells at 370C for 18 hours priar to a 4 hour pulse with 1 Ci
[3H]methyl-thymidine (NEN) and trichloroacetic acid precipitable
radioactivity was determined.
Results
Rabbit antibodies against the KLH-peptide 1 conjugate
reacted with BSA-TGF peptide by solid phase RIA. This reaction was
-83-
i3 41536
1 inhibited in a concentration-dependent fashion-by peptide I and
slightly less by native TGF. In contrast, EGF did not cause
significant inhibition.
Rat TGF competitively displaces the binding of EGF to its
receptors. The ability of TGF-peptides to compete with [125I]EGF
binding to either A-431 cells or HFF was evaluated. Peptide 1
partially inhibited [125I]EGF binding to A-431 cells at >10-6 M.
Peptides 2 and 3, however, exhibited an improved binding inhibition,
with IC50's of 4 x 10-6 M and 4 x 10-7 M, respectively (i.e.,
approximately 0.02 and 0.2% of 'the binding potency of EGF or TGF-a).
Similar observations were made for the inhibition of EGF on HFF
(Table IX). The receptor specificity of these interactions is
illustrated by the inability of these peptides to inhibit either the
binding or mitogenic effect of Endothelial Cell Growth Factor on HFF
(not shown).
None of the TGF peptides possessed intrinsic mitogenic
properties up to 10'5 M when incubated with quiescent HFF (data not
shown). However, the TGF peptides inhibited, in a concentration-
dependent fashion, the induction of DNA synthesis in quiescent HFF by
EGF. These antagonists were equally potent when TGF was used as
mitogen (Table IX). As seen for the binding potency, the
antagonistic potency of the TGF peptides was improved by capping of
the amino- and carboxy-termini.
-84-
13 4 1 5 36
TABLE IX Biological Activities of Synthetic TGF Fragments
Inhibition of Mitogenesisb
Binding a IC50-k1 IC50 (M)
A-431 tiFF EGF TGFa
Peptide 1 8 2 x 10-5 6 2' x 10'5 >10'5 >10'5
Peptide 2 4 2 x 10-6 2+2' x 10-6 5 2 x 10-6 3 1 x 10-6
Peptide 3 4 I x 10-7 3 I x 10-7 6 2 x 10-7 5 + 3 x 10'7
1 0 aIC50's derived from inhibition curves for the binding of
1 nM [125I]EGF to cells. t'IC50's are the concentrations that
decrease by half the enhancement over control of [3H]methyl-
thymidine incorporation in quiescent HFF induced by 1 nM of
EGF or TGFa.
Example XIII
TGFa Stimulates Bone Resorption in Vitro
Transforming growth fiictor preparations were purified from
human melanoma cell conditioned medium and human platelets as
described in Example I and synthetic rat TGF was prepared as
previously described in Example II. Biological activity of the TGF
preparations and synthetic TGF tias monitored using an EGF
radioreceptor assay or using the soft agar colony formation, also as
described in Example I.
Synthetic rat TGF (Mr = 5600)) resorbed bone in a
concentration dependent manner. Concentrations greater than 2 ng EGF
equivalents/ml stimulated bone resorption in three separate
experiments. The synthetic TGF required at least 72 hours to
-85-
134 1536
stimulate bone resorption. In this respect the_synthetic form
appears similar to EGF which resorbs bone over a similar time course
in this bioassay. (Tashjian et al. (1978) Biochem. Biophys. Res.
Commun. 85, 966.)
Partially purified preparations of high and low molecular
weight TGFs prepared from a huinan melanoma cell line were also tested
in the bone resorption assay. Both high (27,000 daltons) and low
(6,000 daltons) molecular weight forms stimulated bone resorption.
The high molecular weight form stimulated resorption within 48 hours,
while the low molecular weight form required 72-96 hours to stimulate
resorption, a similar time cou rse to that of synthetic rat TGF-I and
EGF (See Tashjian, supra.)
The results show thac: TGF can stimulate bone resorption in
vitro. Synthetic rat TGF and low molecular weight human TGF
preparations behaved in a manner similar to that of EGF, since they
required prolonged incubation E-eriods to stimulate bone resorption.
They were effective at concentr-ations of about one order of magnitude
less than EGF. The high molecular weight human melanoma TGF
stimulated bone resorption within 48 hours.
Example XIV
Stimulation of Bone Resorption in Vitro by Synthetic Transforming
Growth Factor
Synthetic rat TGF was prepared as described above in
Example V. Purity of the protein was confirmed by sodium dodecyl
sulfate-polyacrylamide gel electrophoresis, amino acid analysis, and
-86-
~34 15 3fi
1_ reverse phase high-performance liquid chromatography. The biological
activity of TGF preparations and synthetic TGF was monitored by means
of an EGF radioreceptor assay. (Todaro et al. (1976) Nature 264,
26.) Bone resorption was assessed by measuring the release of 45Ca
from previously labeled fetal rat long bones. Pregnant rats at the
18th day of gestation were irijected with 200 #Ci of 45Ca. (Raisz
(1975) J. Clin. Invest. 44, 1.03.) The mothers were killed on the
19th day of gestation, and ttie fetuses were removed. The mineralized
shafts of the radii and ulnaE! were dissected free of surrounding
tissue and cartilage and placed in organ culture. The bones were
incubated in BGJb medium (Irvine Scientific) for 24 hours at 37o.C in*
a humidified atmosphere of 5 percent of C02 and 95 percent air to
allow for the exchange of loosely complexed 45Ca. The bones.were
then cultured for 48 to 120 kours in BGJb medium supplemented with.5
percent fetal calf serum (KC Biologicals) containing control or test
substances. Bone-resorbing activity was measured as the percentage
of total 45Ca released into the medium and was expressed as a
treated-to-control ratio. Statistical significance was determined
with Student's t test for unraired data.
Synthetic rat TGF (molecular weight, 5600) in
concentrations greater than 2 ng of EGF equivalents per milliliter
stimulated bone resorption in a concentration-dependent manner in
three separate experiments. Synthetic TGF caused no significant bone
resorption during the first 48 hours of bone culture, but clearly
stimulated resorption over the following 3 days. In this respect the
-87-
13 41536
synthetic form of TGF appears to be similar to fGF, which resorbs
bone over a similar time course in this bioassay. (Raisz et al.
(1980) Endocrinology 107, 270.) The effect of TGF on bone.resorption
appeared to be independent of prostaglandin synthesis.
Since rat TGF resorbed bone in vitro, human preparations
containing TGF activity were also tested for their effects on bone.
Partially purified preparations of high and low molecular weight
TGF were prepared from a human melanoma cell line. (Marquardt et al.
(1982) J. Biol. Chem. 275, 522(1.) These preparations were partially
purified by acid extraction ancl gel filtration chromatography. Both
high (13,000) and low (6,000) nolecular weight forms stimulated bone
resorption. The high molecularr weight fcnn stimulated resorption
within 48 hours, whereas the lc,w molecular weight form required 72 to
96 hours to stimulate resorpticn--a time course similar to that of
synthetic rat TGF and EGF.
Example XV
Vaccinia Virus Infected Cells F.elease a Novel Polypeptide
Functionally Related to Transfcrming and Epidermal Growth Factors
Cell Culture and Virus
Cercopithecus monkey kidney (BSC-1.) cell monolayers were
maintained in Eagles basal medium supplemented with 10% fetal calf
serum. Vaccinia virus (VV) (strain WR) was grown in Hela cells and
purified by sucrose density gradient sedimentation. (Moss (1981) In
Gene Amplification and Analysis. Vol. 2, eds., Chirickjian and Papis
(Elsevier/North Holland, New York pp. 253-266.)
- 88 -
13 4
1 Preparation of Conditioned Medium -
BSC-1 cell monolayers were incubated at 370C with purified
virus in Eagles basal medium supplemented with 2% fetal calf serum.
Cell culture supernatants were clarified by low speed centrifugation
and lyophilized. The residue was then resuspended in I M acetic acid
and dialyzed extensively against 0.2 M acetic acid. Insoluble
material was removed by centrifugation and the supernatant was
lyophilized and resuspended in one-hundredth of the original volume
of 1 M acetic acid and stored at 40C.
Chromatography
*
Gel filtration was performed on columns of Bio-Gel P-10
(BioRad, Richmond, CA) equilibrated in 1 M acetic acid. Sizing on
high pressure liquid chromatography system utilized two Bio-Sil* TSK-
250 columns (BioRad) in series.
The material isolated, designated as Vaccinia virus growth
factor (VGF), possessed the amino acid sequence given in Example VI
above.
Radioreceptor Assay
The Radioreceptor assay used was similar to the one
described above in Example I. TGF and VGF concentrations were
* Trade Mark
-89-
~
13 41536
1 expressed as the amount required to produce an inhibition of 125I-EGF
binding equivalent to a known amount of EGF.
Radioimmunoassay
The Radioimmunoassay used was similar to the one described
above in Example XI.
Results
Presence of EGF-Competing Activity in the Culture Medium of VV
Infected Cells.
The supernatant, derived from BSC-1 cell 24 hr after
infection with VV, was tested fol- the presence of material that could
compete with 125I-labeled EGF fo~~ binding to EGF receptor-rich human
epidermoid carcinoma cells (A431;). VV infected cells released a
potent EGF competing activity which essentially saturates (>10 ng)
the assay with the equivalent of 10 ng of material resuspended from
0.5 ml of culture fluid. The activity was designated VV growth
factor (VGF). (In contrast, mock infected BSC-1 control cultures
contained minimal EGF competing activity even at the lowest dilution
tested.)
The next experiments were designed to examine the kinetics
of VGF production. At the earliest time examined, 2 hr after
infection, enhanced levels of EGF competing activity already were
present in the culture medium suggesting rapid synthesis and release
of VGF. By 12 hr, maximal amounts of this activity were found in
culture supernatants; only a slight increase was noted at 24 hr.
Since the VV encoded polypeptide with structural homology to EGF and
-90-
134153s
TGF is an early gene product. VGF expression was'monitored in BSC-1
cell cultures treated with cytosine arabinoside (AraCj starting
immediately after the 1 hr virus adsorption period. Inhibition of
DNA synthesis by AraC blocks the expression of late vaccinia genes.
whereas early gene products are not similarly affected. VGF
production was not inhibited in AraC treated cultures but, relative
to infected control cultures, was enhanced more than two-fold at 13
and 24 hr post infection. The ;level of VGF production was also a
function of the virus inoculum (Table XI). With a plaque forming
unit (PFU) to cell ratio of 20:1, approximately 3 ng EGF equivalents
of VGF per ml were detected in culture supernatants at 24 hr post
infection. VGF production was proportional to multiplicity of
infection, with about 6 and 10 ng equivalents of EGF detected in BSC-
I cultures infected at ratios o-f 40 and 80 PFU/cell respectively.
- 91 -
TABLE XI: EFFECT OF MULTIPLICITY OF INFECTION ON VGF RELEASE
Virus Multiplicity V'UF Released
PFU/cell ng Eq. of EGF/ml
0 -
10 2.7
20 4.5
40 6.1
80 10.0
Partial Purification of VGF
The EGF-competing activity found in VV infected BAS-1 cells
was partially purified from acid extracted culture supernatants at 24
hr post infection. Acid solub,ilized polypeptides (10.5 mg) from VV
infected cell culture supernatants were applied to a Bio-Gel P-10
column equilibrated in 1 M acetic acid and samples of each fraction
were tested for EGF competing activity. The major peak of EGF
competition (fraction 42) eluted, slightly after the Mr = 29,000
carbonic anhydrase marker, with an apparent molecular weight of
25,000. No significant activity was detected in fractions
-92-
13 41~~6
1 corresponding to the known elution position of EGF (fraction 100) or
rat TGF (fraction 78) on this column. Peak VGF activity (fraction
42-44) was used for subsequent biological_studies. The molecular
weight was confirmed utilizing tandemly-linked Bio-Sil TSK 250 HPLC
sizing columns. All of the EGF competing activity eluted as a major
peak in the region of the Mr = 25,000 protein marker.
Immunological Comparison of EGF and TGF
To further compare TGF with VGF, the latter was tested in
competitive radioimmunoassay for TGF as described above. The assay
can distinguish EGF from TGF and recognizes both low and high
molecular weight forms of TGF af rat and human origin. (Linsley et
al. (1985) Proc. Natl. Acad. Sci. USA 82, pp. 365-369.) Rat TGF
effectively competed with 125I-labeled lGF peptide for binding to
antibody with a slope, similar to that af unlabeled TGF peptide. A
50% displacement of antigen fr-om antibody was observed at an antigen
concentration of approximately 0.2-0.3 ng equivalents of EGF. When
VGF was tested at equivalent c:oncentraticns, no competition was
observed, suggesting that VGF is not a member of the TGF family of
peptides, insofar as immunoloclical properties are concerned. In a
competitive radioimmunoassay 'For native EGF, VGF preparations
exhibited a minimal displacement (<10%) of 125I-labeled EGF from a
polyclonal antibody to native EGF.
Biological Activity of VGF
At least an order of magnitude more VGF (>2000 ngll) was
found relative to the highest TGF producer, Fisher rat embryo cells
- 93 -
1341535
1 (FRE) non-productively transformed by Snyder-Theilen Feline sarcoma
virus.
In a TGF dependent soft agar assay for anchorage
independent cell growth, VGF stimulated normal rat kidney cells to
form progressively growing colonies in soft agar. On a ng equivalent
basis, partially purified VGF preparations produced 102 colonies
whereas EGF and rat TGF produced 120 and 154 colonies respectively.
(It is possible that contaminating activities in partially purified
VGF preparations may influence quantitat-ion in this type of
biological assay.)
VGF also exhibited potent mitogenicity when tested on a
variety of cultured fibroblast>. When tested at equivalent EGF
receptor binding levels with serum starved (48 hours) mink
fibroblasts (which have relati,eely high numbers of EGF membrane
receptor sites), VGF elicited a 78% increase in DNA synthesis
relative to unstimulated serum deprived cells, whereas a 59%
stimulation in deoxyuridine in,-orporation was seen with mouse
submaxillary gland EGF. Thus the mitogenic effect of VGF is at least
as strong as that of EGF.
-94-
