CA2420494A1 - Ca 125 tumor antigen function and therapeutic uses thereof - Google Patents

Description

FIELD OF THE INVENTION
This invention relates to CA 125 tumor antigen. More specifically, it relates to the biological function of CA 125 tumor antigen and to the use of CA 125 tumor antigen and its function as therapeutic targets in the treatment and prevention of diseases wherein CA 125 tumor antigen is overexpressed.
BACKGROUND OF THE INVENTION
l0 Ovarian cancer is one of the leading causes of death in women over 40.
Although most patients respond to initial treatment, the majority relapses partially due to the appearance of chemo-resistant tumor cells. In order to improve therapy, it is essential to understand the underlying mechanisms responsible for the occurrence of ovarian cancer.
CA125 antigen is the most important clinical marker of ovarian cancer CA125 tumor antigen is the most important clinical marker of ovarian cancer as it is used to monitor response to chemotherapy. Rising or falling blood levels of

2 o correlate with progression or regression of the disease. CA125 antigen was first detected in the early 80's using the MAb OC125 which was raised against the human ovarian carcinoma cell line OV433 isolated from a patient with serous papillary cystadenocarcinoma (1). The specific reactivity of the OC125 Mab to a variety of human ovarian carcinoma cell lines and paraffin-embedded ovarian carcinoma tissues has led to the development of a radioimmunoassay to detect the CA125 antigen in serum from ovarian cancer patients (2). Using this assay, rising or falling levels of CA125 were shown to correlate with progression or regression of disease demonstrating that CA 125 levels correlate with clinical course of the disease (2-4).
It is currently employed as a predictor of clinical recurrence in ovarian cancer and to

3 0 monitor response to chemotherapy treatment (5-8).

CA125 biochemical studies Despite the widespread use of CA125 as a clinical marker of ovarian cancer, the biochemical and molecular nature as well as the function of this antigen are poorly understood. Previous biochemical studies demonstrated that the CA125 epitope is carried on a large glycoprotein with a M.W. in the range of 2x105-106 Da, while others reported that CA125 consists of many subunits of 50-200 kDa (9-14). The study of Lloyd et al. showed that CA125 is a high molecular weight glycoprotein having properties of a mucin-type molecule (15). In these studies however, a definite consensus regarding the molecular nature of CA 125 could not be elaborated and no information about its function was provided. A partial cDNA encoding CA125 was recently identified as MUC16. The deduced amino acid sequence proposed an extracellular domain composed of 9 tandem repeats rich in serine, threonine and proline followed by a unique region, a potential transmembrane domain and a short cytoplasmic tail. CA125 is expressed in more than 80% of epithelial ovarian cancer but is not detectable in normal ovary tissues. However its role in the disease is unknown.
There is therefore a crucial need to identify therapeutic targets in order to treat the disease.
?o SUMMARY OF THE INVENTION
An object of the present invention is to provide a therapeutic target that satisfies the above mentioned need.
2 .~
We developed a novel strategy to study the role of proteins that could not be previously studied because the gene was not known or not available. Using this strategy, we derived unique inhibitors of CA125 tumor antigen. We propose a role 3 o for CA125 tumor antigen in the pathogenesis of ovarian cancer as well as other diseases where CA 125 tumor antigen is overexpressed, a non exclusive list of which is endometriosis, cervical cancer, fallopian tube cancer, cancer of the uterus and prostate cancer. Our results have lead to the identification of CA 125 tumor antigen and CA 125 tumor antigen function as novel therapeutic targets for the treatment and prevention of these diseases in mammals.
Accordingly, the present invention provides for the use of CA 125 tumor antigen and CA 125 tumor antigen function as therapeutic targets in the treatment and prevention of any disease, in mammal, wherein CA 125 is overexpressed, such as ovarian cancer, cervical cancer, cancer of the uterus, fallopian tube cancer, prostate cancer and endometriosis.

The present invention provides for the use of CA 125 tumor antigen as a therapeutic target in a method of treatment and prevention of a disease, in a mammal, in which CA 125 tumor antigen is overexpressed, wherein CA 125 tumor antigen is sequestered, knocked-out, inhibited, inactivated or otherwise partially or totally neutralized before, during or after protein synthesis in ways that are known to a person skilled in the art.
The present invention also provides for the use of a CA 125 tumor antigen function as a therapeutic target in a method of treatment and prevention of a disease, in a mammal, in which CA 125 tumor antigen is overexpressed, wherein a CA 125 tumor antigen function is sequestered, knocked-out, inhibited or otherwise partially or totally neutralized in ways that are known to a person skilled in the art.
The present invention also provides for the use of a CA 125 tumor antigen function 2 ~> in a method of diagnosis of a disease, in mammal, in which CA 125 tumor antigen is overexpressed.
The present invention also provides for the use of a CA 125 tumor antigen function in a diagnostic kit for a disease, in a mammal, in which CA 125 tumor antigen is 30 overexpressed.

4 The present invention also provides for the use of a CA 125 tumor antigen function in the identification of an agent useful for the prevention and treatment of a disease, in a mammal, in which CA 125 tumor antigen is overexpressed.
The present invention also provides for the use of a CA 125 tumor antigen function in a screening method for the identification of an agent useful for the prevention and treatment of a disease, in a mammal, in which CA 125 tumor antigen is overexpressed.
The present invention also provides for a pharmaceutical composition targeting CA
125 tumor antigen or CA 125 tumor antigen function and its use in the treatment and prevention of a disease, in a mammal, in which CA 125 tumor antigen is overexpressed .
The present invention also provides for a method of treatment and prevention of a disease, in a mammal, comprising the step of modulating cell surface expression of the CA 125 tumor antigen by administration of an inhibitor of CA 125 tumor antigen expression in mammalian cells. Preferably, the inhibitor is a single-chain antibody (ScFv). More preferably, the modulation consists in a downregulation of the cell 2 o surface expression of the CA 125 tumor antigen in mammalian cells. Even more preferably, the ScFv sequesters CA 125 protein within organelles. Most preferably, organelles are the ER, the trans-golgi or cytoplasm (or any cell compartment).
In another aspect of the present invention the method of treatment and prevention

5 comprises the use of ScFvs that are introduced to the CA 125 tumor antigen by subcloning them into eucaryotic expression vectors specifically targeting them to the ER and the trans-golgi or other specific cell compartments. Preferably, ScFvs are derived from the OC 125 Mab and the VK-8 Mab. More preferably, ScFvs are OC125-3.11 and VK-8-1.9.

r In a further aspect of the present invention, it is provided an inhibitor of CA 125 tumor antigen expression which is OC125-3.11 or VK-8-1.9 for the use in the treatment and prevention of disease in a mammal wherein CA 125 is overexpressed.
The present invention also provides for the use of an inhibitor of CA 125 tumor expression in the making of a medicament for the prevention or treatment of a disease, in a mammal, wherein CA 125 is overexpressed. Preferably, the inhibitor is a scFv selected from the group consisting of ScFv OC125-3.11 and ScFV VK-8-1.9.
rO
The present invention also provides for a method of treatment and prevention of a disease in a mammal comprising the step of identifying a modulator capable of modulating cell surface expression of the CA 125 tumor antigen in cells of the mammal. Preferably, the modulator is an inhibitor. More preferably, the inhibitor is an scFv selected from the group consisting of ScFv OC125-3.11 and ScFV VK-8-1.9. The assay may be automated. The assay may be used for the high through-put screening of a number of modulators.
Other objects, advantages and features of the present invention will become more apparent upon reading of the following non restrictive description of preferred embodiments thereof, given by way of example only with reference to the accompanying drawings.

These and other objects and advantages of the invention will become apparent upon reading the description and upon referring to the drawings in which Figure 1shows the proposed structure of the CA 125 tumor antigen.
30 Figures 2 A, B and C show the construction of an ScFv library.

Figures 3 A and B show the selection of soluble ScFvs through a "colony lift assay".
Figure 4 shows selection of soluble ScFvs (periplasmic extracts).
Figure 5 shows expression of ScFvs (ELISA).
Figure 6 shows expression of ScFvs in the pCantab-5E prokaryotic expression system with and without induction.
Figure 7 shows selection of ScFvs binding to CA 125 (ELISA).
Figure 8 shows selection of ScFvs binding to CA 125 (ELISA).
Figure 9 shows cloning of ScFvs binding to CA 125 in eukaryotic expression system.
Figures 10 A, B, C and D show Western blots showing expression of ScFvs directed to the golgi in OVCAR-3 and directed to the ER in PA-1.
Figure 11 shows expression of ScFv OC125 golgi 3.11 compared with expression of CA 125.
Figure 12 show expression of ScFv VK-8 KDEL 1.9 compared with expression of CA
125.
5 Figure 13 shows expression of control linker compared with expression of CA
125.
Figure 14 shows expression of ScFv in golgi and expression of ScFv in ER
compared with expression of proteins native to golgi and ER, respectively.
3o Figure 15 illustrates construction and in vitro validation of anti-CA125 scFvs.
A) CA125 binding activity of anti-CA125 OC125, VK-8-1.9 and VK-8-4.5 scFvs present in periplasmic extracts of bacteria as well as anti-Bcl2 4D7 scFv compared to parental Mabs OC125 and VK-8 and to PBS and periplasmic extract from bacteria, uninduced, IPTG-induced and controls.
B) Expression of scFvs from periplasmic extracts, probed with anti-Etag antibody.
C) Immunoprecipitation and co-immunoprecipitation of Golgi- and ER-targeted OC125-3.11 scFv from transient transfection of pSTCF.GOLGI-OC125-3.11 and pSTCF.KDEL-OC125-3.11 in NIH:OVCAR-3 human ovarian cancer cells using anti-c-myc, anti-CA125 Mabs OC125 and VK-8, western blot probed with anti-c-myc 9E10 antibody.
D) Immunoprecipitation and co-immunoprecipitation of Golgi- and ER-targeted VK-1.9 scFv from transient transfection of pSTCF-GOLGI-VK8-1.9 and pSTCF.KDEL-VK8-1.9 in NIH:OVCAR-3 human ovarian cancer cells using anti-c-myc, anti-CA125 Mabs OC125 and VK-8, Western blot probed with anti-c-myc 9E10 antibody.
E) Immunoprecipitation and co-immunoprecipitation of ER-targeted VK-8-4.5 scFv 1 ~ from transient transfection of pSTCF.KDEL-VK8-4.5 in NIH:OVCAR-3 human ovarian cancer cells using anti-c-myc, anti-CA125 Mabs OC125 and VK-8, western blot prober with anti-c-myc 9E10 antibody.
Figure 16 shows localization of anti-CA125 scFvs and CA125 cell surface down 2 0 regulation.
A) NIH:OVCAR-3 cells were transiently transfected with pSTCF.Golgi-OC125-3.11 or pSTCF.KDEL-VK-8-1.9 constructs and 48hrs later the cells were fixed in ice-cold methanol. Localization of scFvs was detected with the anti-c-myc A14 polyclonal antibody and compared with ER and Golgi residents using anti-calreticulin PA3-900 and anti-ADP ribosylation factor MA3-060 monoclonal antibodies, respectively. Oregon green anti-rabbit and texas red anti-Mouse secondary antibodies were used.
B) NIH:OVCAR-3 cellswere transiently transfected with pSTCF.Golgi-OC125-3.11 j 0 or pSTCF.KDEL-VK-8-1.9 constructs and 48hrs later the cells were fixed in ice-cold methanol. Expression of scFvs and CA125 was detected using the anti-c-myc A14 polyclonal antibody and anti-CA125 M11 monoclonal antibody. Oregon green anti-rabbit and Texas red anti-mouse secondary antibodies were used.
Figure 17 shows Cell surface down modulation of CA125 in stable NIH:OVCAR-3 clones expressing the ER-VK-8-1.9 anti-CA125scFv and relevant control.
A) Stable transfectants expressing the ER-targeted VK-8-1.9 and VK-8-4.5 scFvs and parental cell line NIH:OVCAR-3 were fixed in ice-cold methanol generated and expression of CA125 at the cell surface and scFv was assesed by immunofluorescence using anti-c-myc A14 polyclonal antibody and anti-CA125 1a M11 monoclonal antibody, respectively. Oregon green anti-rabbit and Texas red anti-mouse antibodies were used as secondary antibodies.
8) CA125 expression in the stable trasnfectants was analysed by FACS using anti-CA125 M11 monoclonal antibody and a Phyco-Erythrin-anti-mouse antibody and compared with parental cell line NIH:OVCAR-3; Black, OVCAR-3 levels of CA125 expression at cell surface; grey, CA125 levels in stable transfectants.
Figure 18 shows decreased CA125 cell surface expression influences the proliferation rate, cell-cell interaction and cell migration.
A) Growth curve of stable NIH:OVCAR-3 transfectants ER-VK-8-1.9#9 (positive for 2 0 CA125 binding) and ER-VK-8-4.5#12 (negative for CA125 binding) compared to parental cells OVCAR-3. Cells were plated in triplicate in 96-well plate and cell proliferation was measured every day with a XTT assay. Plot represents results from 3 independent experiments B) Cell aggregation assay. Cells were plated onto 0.6% agarose layer in bacterial 25 dishes. Seventy-two hours later photomicrographs were taken (10X
magnification) to visualize the presence of cell aggregates.
C) Wound healing assay. A wound was made using a 13mm-wide razor blade in confluent cell monolayers and 20mM hydroxy-urea was added to block cell proliferation. Forty-eight hours later, the cells were fied in methanol and stained with Giemsa and microphotographs were taken (10X magnification).
D) Tumorigenic assay. Ten millions NIH:OVCAR-3 transfectants ER-VK-8-1.9#9, ER-VK-8-4.5#12 and parental cells OVCAR-3 were inoculated subcutaneously in nude mice Tumors were allowed to grow for 6 weeks after which tumorw were excised and tumor weight was measured and plotted for each transfectant.
Figure 19 A) NEDO cDNA clone FLJ14303 encodes a part of CA125. Cos-7 cells (negative for CA125 by Western blot and ELISA) were transfected with an expression vector encoding the cDNA from the NEDO clone FLJ14303. Reactivity of anti-CA125 OC125 and VK-8 antibodies with the expression product of this cDNA was 1 ~'~ analysed by western blot and compared to CA125 expression in OVCAR-3 cells as well as in mock-transfected Cos-7 cells.
B) Expression of the CA125 cytoplasmic tail fused to Gal4 DNA binding domain.
The CA125 cytoplasmic tail was cloned in the pGBDU and pGAD for the yeast two-hybrid system. The S.Cerevisiae strain PJ69-4a was transformed with the pGBDU empty vector (EV) or with the vector Containing CA125 cytoplasmic tail (Cyto). Three days after growth on appropriate media, proteins were extracted from the 2 transfectants or the wt strain PJ69-4a, ran on 12.5% SDS-PAGE and transfered by western blot on a PVDF membrane. The membrane was probed with anti-Gal4 DNA binding domain antibody Gal-4-DBD RK5C1.
~~ o Figure 20 shows cisplatin sensibility of stable NIH:OVCAR-3 clones expressing the ER-VK-8-1.9 anti-CA125scFv and relevant controls.
Cells were plated in triplicate in 96-well plates and exposed or not to increasing concentrations of cisplatin. Fours days later, cell proliferation was measured with a 2 ~ XTT assay. Percentage of survival was plotted against concentration of cisplatin.
Curves represent results from 3 independent experiments. Red line represents 50%
survival.
3 o Figure 21 shows IC50 of cisplatin for the stable NIH:OVCAR-3 clones expressing the ER-VK-8-1.9 anti-CA125scFv and relevant controls. Inhibitory concentrations of cisplatin resulting in 50% survival of cells were calculated from curves of graph in figure 20 (red line in figure 20). .
Figure 22 shows expression of E-cadherin and av~35 integrin in NIH:OVCAR-3 cells. NIH:OVCAR-3 were fixed in ice-cold methanol generated and expression of E-cadherin and cxv~35 integrin at the cell surface was assesed by immunofluorescence using E-cadherin clone 36 and anti- av~5 integrin clone P1 antibody and Texas red labelled secondary anti-mouse antibody.
Figure 23 shows expression of E-cadherin and scFv in NIH:OVCAR-3 stable 1 c~ transfectant expressing the ER-VK-8-1.9 anti-CA125 scFv without induction with doxycycline. Cells were grown on glass slides for 48hrs and fixed in ice-cold methanol generated and expression of E-cadherin and scFv was assesed by immunofluorescence using E-cadherin clone 36 and anti-c-myc A14 antibody and Texas red or Oregon green-conjugated secondary anti-mouse and anti-rabbit antibodies.
Figure 24 shows expression of E-cadherin and scFv in NIH:OVCAR-3 stable transfectant expressing the ER-VK-8-1.9 anti-CA125 scFv when induced with doxycycline. Cells were grown in presence of doxycycline for 48 hrs and then fixed in ice-cold methanol generated and expression of E-cadherin and scFv was assessed by immunofluorescence using E-cadherin clone 36 and anti-c-myc A14 antibody and Texas red or Oregon green-conjugated secondary anti-mouse and anti-rabbit antibodies.
Figure 25 shows expression of E-cadherin and scFv in NIH:OVCAR-3 stable transfectant expressing the ER-VK-4.5 control scFv without induction with doxycycline. Cells were grown on glass slides for 48hrs and fixed in ice-cold methanol generated and expression of E-cadherin and scFv was assesed by immunofluorescence using E-cadherin clone 36 and anti-c-myc A14 antibody and 3~~ Texas red or Oregon green-conjugated secondary anti-mouse and anti-rabbit antibodies.

ll Figure 26 shows expression of E-cadherin and scFv in NIH:OVCAR-3 stable transfectant expressing the ER-VK-4.5 control scFv when induced with doxycycline. Cells were grown on glass slides in the presence of doxycycline for 48hrs and fixed in ice-cold methanol generated and expression of E-cadherin and scFv was assesed by immunofluorescence using E-cadherin clone 36 and anti-c-myc A14 antibody and Texas red or Oregon green-conjugated secondary anti-mouse and anti-rabbit antibodies.
Figure 27 shows expression of av~5 integrin and scFv in NIH:OVCAR-3 stable transfectant expressing the ER-VK-8-1.9 anti-CA125 scFv when induced or not with doxycycline. Cells were grown in the absence or presence of doxycycline for 48hrs and subsequently fixed in ice-cold methanol generated and expression of av~35 integrin and scFv was assesed by immunofluorescence using anti- av~5 integrin clone P1 F6, anti-c-myc 9E10 antibody and Texas re or Oregron green 15 labelled secondary antibodies.
Figure 28 shows expression of av~5 integrin and scFv in NIH:OVCAR-3 stable transfectant expressing the ER-VK-8-4.5 anti-CA125 scFv when induced or not with doxycycline. Cells were grown in the absence or presence of doxycycline for 48hrs and subsequently fixed in ice-cold methanol generated and expression of av~5 integrin and scFv was assesed by immunofluorescence using anti- av(35 integrin clone P1 F6, anti-c-myc 9E10 antibody and Texas re or Oregron green labelled secondary antibodies.
Figure 29 shows alignment of deduced amino acid sequence for anti-CA125 scFvs VK-8-1.9 and OC125-3.11 as well as control scFv VK-8-4.5 Nucleotidic sequences encoding the anti-CA125 scFvs and their control were determined from pCANTABSE/scFv constructs using the scFv specific primers S1 and S6 from the pCANTAB5 sequencing primer set (Amersham Pharmacia Biotech, Piscataway, NJ).
3G Sequences were determined using the LI-COR automatic sequencing system (Bio S&T Inc., Lachine, QUE). Amino acid sequence was deduced from the nucleotidic sequences and aligned using the Alibee multiple alignment software available at www.genebee.msu.su/services/malign reduced.html. The area in boxes represent consensus sequences of frameworks 1-4 of heavy and light chain, asterisks correspond to differences between VK-8-1.9 and OC125-3.11 anti-CA125 scFvs whereas arrows identify differences between VK-8-1.9 and VK-8-4.5 scFvs.
While the invention will be described in conjunction with example embodiment, it will be understood that it is not intended to limit the scope of the invention to such embodiment. On the contrary, it is intended to cover all alternatives, modifications and equivalents as may be included as defined by the appended claims.
DESCRIPTION
The CA125 tumor antigen is a protein associated with the majority of human epithelial ovarian cancer, the most common form of the disease. It is also known to be overexpressed in other diseases such as endometriosis, cervical cancer, cancer of the uterus, fallopian tube cancer, prostate cancer, etc. It was recently proposed that CA125 is part of the mucin family of proteins. Mucins are known to play an important role in cell adhesion in cancer cells as well as in normal cells. In some cancer cells, mucins promote the metastatic process.
To elucidate the function of CA125, we developed 2 anti-CA125 single-chain antibodies (scFvs) and show that they act as CA125 specific inhibitors. When expressed intracellularly and retained to the ER or Golgi, the anti-CA125 scFvs presumably entrap CA125 within the secretion pathway and therefore prevent its proper cell surface localization in the human ovarian cancer cell line OVCAR-3. In addition, we have shown that stable inhibition of CA125 in OVCAR-3 cells results in an increased cell proliferation and reduced cell adhesion and migration and prevent tumor growth in nude mice. Our hypothesis is that CA125 modulates cell proliferation, adhesion and migration through molecular mechanisms. We have 3o confirmed our observations using alternative knockout approaches such as RNA

interference (RNAi) to ablate CA125 expression in OVCAR-3 cells and extend our observations to a panel of human cell lines expressing various levels of CA125 such as OV-90 and SKOV3ip1 (expressing high to low levels of CA125) and primary ovarian cancer cells (result not shown). We also observe that loss of CA125 at the cell surface affects the in vivo behaviour of ovarian cancer cells in xenograft mouse models.
EXAMPLE I : Construction and in vitro validation of anti-CA125 scFvs We constructed single-chain antibody libraries derived from the OC125 and VK-8 l0 hybridoma cell lines specific for CA125. The 2 scFv libraries were screened for CA125 binding activity by ELISA using commercially purified human CA125. ScFvs that bound to CA125 by ELISA, OC125-3.11 and VK-8-1.9, and one that did not bind, VK-8-4.5 were selected (figure 15 A-B). We hypothesized that if those scFvs (CA125 binders) once expressed intracellularly were localized to and retained within the ER
or the Golgi then CA125 antigen would be entrapped during synthesis and thus be unable to localize at cell surface and interact with other intracellular and/or extracellular proteins to achieve its function(s). The scFvs were targeted to the ER
or trans-median Golgi by sequence fusion with an IgK secretion leader and a KDEL
signal or fusion with the N-terminal 81 amino acids of human beta 1,4-2 o galactosyltransferase, a protein resident of the trans-medial Golgi (47-49) in addition of a c-myc tag at the C-terminus. Immunoprecipitation experiments showed that the anti-CA125 scFvs were immunoprecipitated with anti-c-myc antibody whereas only OC125-3.11 and VK-8-1.9 (both positive for CA125 binding) were co-immunoprecipitated using anti-CA125 OC125 and VK-8 MAbs (figure15 C-E). These a 5 results demonstrate that the OC125-3.11 and VK-8-1.9 anti-CA125 scFvs bind to CA125 in vitro.
EXAMPLE II : Localization of anti-CA125 scFvs and cell surface down-3 o regulation of CA125 Proper localization of our anti-CA125 scFvs in transient transfection of human ovarian cancer cells OVCAR-3 was demonstrated by immunofluorescence. Results obtained with ER-VK-8-1.9 and GOLGI-OC125-3.11 are shown in figure 16A.
Immunofluorescence studies showed that cells expressing the Golgi-targeted OC125-3.11 or ER-targeted VK-8-1.9 lost expression of CA125 at the cell surface.
However surrounding cells that did not express the scFvs (not transfected) were positive for CA125 at the cell surface. In addition, the presence of the ER-targeted VK-8-4.5, which did not bind CA125 by ELISA and immunoprecipitation experiments, did not affect expression of CA125 in the cells expressing this scFv. These results show that the expression and retention of ER- or GOLGI-targeted anti-CA125 scFvs results in CA125 down-regulation at the cell surface. Anti-CA125 scFvs act therefore 1 o act as potent inhibitors of CA125.
EXAMPLE III : Consequences of CA125 cell surface down-regulation in human ovarian cancer cell line NIH:OVCAR-3 To determine the effects of down-modulating CA125 expression at the cell surface, 15 we derived stable clones encoding the ER-targeted VK-8-1.9 and Golgi-OC125-3.11 (both positive for CA125 binding) and ER-VK-8-4.5 (negative for CA125 binding) scFvs in human ovarian cancer cell lines OVCAR-3 (high expresser of CA125), OV-90 (moderate expresser) and SKOV3ip1 cells (low expresser). Some of the OVCAR-3 clones have been already characterized for scFv and CA125 expression and all of Go the other clones (including in SKOV3ip1) have also been evaluated.
Characterization of stable clones ER-VK-8-1.9#9 and ER-VK-8-4.5#12 is shown in figure 17. A
dramatic decrease in CA125 expression at the cell surface was observed in stable clone ER-VK-8-1.9#9 (positive for CA125 binding) while CA125 expression was not affected in the clone ER-VK-8-4.5#12 (negative for CA125 binding) although the scFv in this clone was expressed at adequate levels (figure 17A). Similar results were obtained from FACS analysis (figure 17B). These results are consistent with results obtained previously from transient transfection experiments. Taken together these results demonstrate that our scFvs act as specific inhibitors of CA125 and that our stable clones behave as unique tools to study CA125.
i Expression of E-cadherin and av~3v integirin at the cell surface To further characterize the stable transfectants expressing the anti-CA125 scFvs we LJ
evaluated the expression of E-cadherin and ~xvw integrin at the cell surface for each stable transfectant. Figures 23 through 26 show that E-cadherin expression at the cell surface is not affected by the presence of the ER-VK-8-1.9 anti-CA125 scFv or the control ER-VK-8-4.5 demonstrating that E-cadherin expression is not modulated by CA125 levels. Expression of cxv~v integrin at the cell surface of ER-VK-8-4.5 transfectant is also not affected by the expression of the scFv (figure 28).
However, the stable transfectant expressing the ER-VK-8-1.9 anti-CA125 scFv shows a reduced level of av~3v integrin at the cell surface demonstrating that CA125 influences levels of cxv~3v integrin expression at the cell surface.
~~o Cell proliferation on adhesive support In vitro growth kinetics of ER-VK-8-1.9 and ER-VK-8-4.5 clones was evaluated compared with that of the parental cell line using a XTT cell proliferation assay (50).
Stable clone ER-VK-8-1.9#9 (positive for CA125 binding) grew faster than the ER-VK-8-4.5#12 (negative for CA125 binding) which grew at a rate similar to that of the parental OVCAR-3 cells (figure 18A). Stable clone ER-VK-8-1.9#9 seems to adhere faster to the plastic than OVCAR-3 cells or ER-VK-8-4.5#12 clone (not shown).
These results show that loss of CA125 at the cell surface affects cell proliferation on adhesive support.
a c;
Sensitivity to cisplatin Consequently, sensitivity to cisplatin of the various stable transfectants was determined. Stable clone ER-VK-8-1.9#9, the ER-VK-8-4.5#12 and the parental cell lines were plated in triplicate in 96-well plates and exposed or not to increasing concentrations of cisplatin. Fours days later, cell proliferation was measured with a XTT assay. Percentage of survival was plotted against concentration of cisplatin.
Curves represent results from 3 independent experiments (figure 20). Results showed that stable clone ER-VK-8-1.9#9 was more sensitive to cispaltin than control cells. IC50 were calculated and figure 21 shows that the stable clone ER-VK-8-1.9#9 is approximately 10-fold more sensitive to cisplatin than ER-VK-8-4.5#12 and the parental cell lines. Similar experiments were performed using taxol and results showed no difference between sensitivity of stable transfectant ER-VK-8-1.9#9, the 1. 6 ER-VK-8-4.5#12 and the parental cell lines confirming that the increased sensitivity of transfectant ER-VK-8-1.9#9 is linked to the increase in cell proliferation, These results demonstrate that CA125 influence cell proliferation and thereby controls the sensitivity to therapeutics drugs such as cisplatin.
Cell-cell interactions and anchorage independent We also assessed the effect of CA125 cell surface down-regulation on the ability of the cells to mediate cell-cell interaction using a cell aggregation assay (51). Cell-cell interactions are measured by the ability of the cells to aggregate to each other and 1 c grow in clumps. Transfectant ER-VK-8-4.5#12 formed small aggregates and grew in small clumps similarly to the parental OVCAR-3 cells (figure 18B). In contrast, transfectant ER-VK-8-1.9#9 (positive for CA125 binding) did not form aggregate and only isolated single cells were observed. In addition, the single cells observed in this clone did not grow and looked as if they were dead or dying. These results show that 1 ~ loss of CA125 at the cell surface impairs the cells ability to mediate cell-cell interactions and to survive in anchorage-independent conditions.
Cell migration We also determined the consequence of reducing CA125 expression levels at the ? o cell surface on cell migration. We evaluated the cell motility of ER-VK-8-1.9#9 and ER-VK-8-4.5#12 stable transfectants and compared to the parental cell line using the wound or scratch assay (52). Cells were plated in 6-well plates and when confluent a wound was made in the monolayer using a razor blade. To distinguish between cell proliferation and cell migration, cell proliferation was inhibited with 20mM
2. 5 hydroxyurea (53). Cells were incubated in the presence or the absence of FBS. In the absence of FBS none of the cells, neither the parental cells, were able to migrate and fill in the wound (not shown). This suggests that some factors present in the serum may be required for stimulating cell motility as showed by others in different tumor cell lines (54). However, in the presence of FBS, the only cells that did not 30 migrate and fill in the wound were the cells from clone ER-VK-8-1.9#9 (positive for CA125 binding) (figure 18C). Cells from clone ER-VK-8-4.5#12 (negative for binding) migrated in a similar manner as the parental cells. These results show that 1 ~7 CA125 affects cell migration of the OVCAR~-3 cell line.
Tumor growth We also determined whether the loss of CA125 expression at the cell surface affects the in vivo behaviour of human ovarian cancer cells in tumor-bearing mice subcutaneously or intraperitoneally. This was achieved by evaluating tumor growth, tumor burden, formation of ascites, presence of tumor cells in ascites, pattern of metastases spread and survival of mice. The tumorigenicity of each stable transfectants was also determined in nude mice. Stable clone ER-VK-8-1.9#9, the 1 o ER-VK-8-4.5#12 and the parental cell lines were inoculated subcutaneously in nude mice and tumors were allowed to grow for 6 weeks after which tumor were excised and tumor weight was measured. Figure 18D shows that tumor derived from stable clone ER-VK-8-1.9#9 were significantly much smaller (if existant) than those from the ER-VK-8-4.5#12 and the parental cell lines demonstrating that CA125 influence the tumorigenic potential of ovarian cancer cells. When injected intraperitoneally, stable clone ER-VK-8-1.9#9 showed a significant slower growth, reduced volume of ascites, a decrease in the total number of viable tumor cells in suspension (in ascites) and therefore an overall increased in survival of mice (not shown).
These results taken together point to a role for CA125 in the pathogenesis of ovarian cancer by influencing tumor cell proliferation, tumor cell adhesion and migration, and in tumorigenesis.
EXPERIMENTAL PROCEDURES

Derivation of anti-CA 725 scFv constructs - The hybridoma cell line VK-8 which express a monoclonal antibody against CA 125 tumor antigen has been described previously and was kindly provided by K.O. L_loyd (Sloan-Kettering Memorial Cancer Center, New York, NY) (18). Total mRNA was extracted from VK-8 hybridoma using 3 o the PolyA-track kit from Promega (Cie, city, state). Total mRNA extracted from OC125 hybridoma cell line was kindly provided by R.C. Bast (MD Anderson Cancer Center, Houston, TX). ScFvs constructs were generated using the Recombinant 1 ~5 Phage Antibody System (Amersham Pharmacia Biotech, Piscataway, NJ) according to the manufacturer's instructions. Briefly, the variable heavy and light chains (VH and V~) were amplified from the cDNA by PCR using mouse variable region primers.
The VH and V~ DNA fragments were linked together by overlap extension PCR using a (GIy4Ser)3 linker to generate 750bp scFv constructs with flanking Sfil and Notl sites.
The scFv DNA fragments were inserted into Sfil/Notl sites of the prokaryotic expression vector pCANTABSE from the Mouse ScFv Module (Amersham Pharmacia Biotech, Piscataway, NJ). Screening of recombinant clones expressing a soluble scFv was accomplished by a colony lift assay as described previously (26).

Claims