Int. J. Cancer: 73, 117–124 (1997)
`r 1997 Wiley-Liss, Inc.
`
`Publication of the International Union Against Cancer
`Publication de l’Union Internationale Contre le Cancer
`
`TARGETED THERAPY OF SCHWANNOMA CELLS IN IMMUNOCOMPETENT RATS
`WITH AN erbB2-SPECIFIC ANTIBODY-TOXIN
`Uwe ALTENSCHMIDT, Mathias SCHMIDT, Bernd GRONER and Winfried WELS*
`Institute for Experimental Cancer Research, Tumor Biology Center, Freiburg, Germany
`
`Over-expression of the erbB2-receptor tyrosine kinase is
`frequently observed in many human tumors of epithelial
`origin. Due to its causal involvement in malignant transforma-
`tion and its presence on the tumor cell surface erbB2 is an
`attractive target for directed tumor therapy. We earlier
`described the potent anti-tumoral activity of the recombi-
`nant single-chain antibody toxin scFv(FRP5)-ETA in vitro and
`in nude mouse tumor models in vivo. This molecule consists
`of the variable domains of the heavy and light chains of an
`erbB2-specific antibody genetically fused to a truncated Pseu-
`domonas exotoxin A. Here we have investigated the in vivo
`effects of this immunotoxin on erbB2 expressing NV2Cd
`schwannoma cells growing as s.c. tumors in syngeneic BDIX
`rats. Established tumors were treated either locally by intra-
`tumoral
`injection of scFv(FRP5)-ETA or systemically by
`injection into the tail vein. Both routes of application resulted
`in pronounced inhibition of tumor growth with local treat-
`ment being more effective. Treatment with 25 mg/day of
`scFv(FRP5)-ETA for 10 days suppressed tumor growth al-
`most completely. Antibodies directed mainly against the
`toxin domain of the fusion protein developed in all animals
`treated. Int. J. Cancer 73:117–124, 1997.
`r 1997 Wiley-Liss, Inc.
`
`The erbB-2 gene encodes a 185-kDa transmembrane glycopro-
`tein that is a member of the type I/erbB family of receptor tyrosine
`kinases which also includes epidermal-growth-factor (EGF) recep-
`tor, erbB3 and erbB4 (Peles and Yarden, 1993). Over-expression of
`erbB2 is frequently observed in human tumors arising at many
`sites, including the breast and the ovaries, where it correlates with
`an unfavorable prognosis (Hynes and Stern, 1994). Due to its role
`in cancer development and its accessible location on the cell
`surface erbB2 has been suggested as a promising target for directed
`therapy.
`Pseudomonas aeruginosa exotoxin A is a well-characterized
`protein of 66 kDa which harbors the different activities required for
`cell binding, uptake into cells, and toxic activity in distinct
`functional domains (Allured et al., 1986; Hwang et al., 1987).
`Upon binding of the N-terminal domain I to target cells and
`internalization via receptor-mediated endocytosis the internal do-
`main II becomes activated by proteolytic cleavage and facilitates
`the translocation of a C-terminal fragment containing the enzy-
`matic domain III into the cytosol (Ogata et al., 1992). There the
`toxin ADP-ribosylates elongation factor 2 which results in the
`inhibition of protein synthesis. On the basis of earlier studies
`demonstrating the utility of recombinant ETA as a potential
`therapeutic effector (Pastan and FitzGerald, 1991), we have already
`described chimeric ETA fusion proteins targeted to the different
`members of the erbB receptor family.
`As a cell recognition domain these hybrid molecules employ
`either growth factors like the erbB3/erbB4 ligand Heregulin b1
`(Jeschke et al., 1995; Groner et al., 1997), or recombinant
`single-chain antibody (scFv) fragments directed to erbB2 (Wels et
`al., 1992, 1995) and the EGF receptor (Wels et al., 1995; Schmidt
`et al., 1997). ScFv molecules consist of the variable domains of the
`heavy and light chains of immunoglobulins connected by a flexible
`linker sequence (Winter and Milstein, 1991). The heterologous
`binding domains were genetically fused to a modified form of ETA
`which lacks the natural cell-binding domain of the toxin (Wels et
`al., 1992). The resulting chimeric toxins were expressed in E. coli
`as recombinant molecules and their biological activities were
`characterized. High selectivity for cells expressing the respective
`
`target receptor and potent anti-tumoral activity were observed
`(Wels et al., 1992, 1995; Groner et al., 1997).
`To evaluate the in vivo effects of such reagents, pre-clinical
`animal model systems are of fundamental importance. Since most
`of these molecules are specific for human tumor antigens, experi-
`mental settings chosen are often based on human tumor xenografts
`growing in immunocompromised animals. However, a major
`limitation of such animal models is their greatly impaired immuno-
`logical response to the effector which is applied for therapy. Since
`recombinant toxins contain a large protein domain of bacterial
`origin, it is predictable that in an immunocompetent organism a
`humoral immune response directed against this type of therapeutics
`will develop which might neutralize the anti-tumoral activity.
`Here we have evaluated the in vivo effects of the erbB2-specific
`scFv-Pseudomonas exotoxin-A fusion protein scFv(FRP5)-ETA in
`immunocompetent rats. Our tumor model is based on NV2Cd rat
`schwannoma cells stably transfected with a human erbB-2 cDNA
`construct (Nikitin et al., 1996). Upon s.c. injection NV2Cd-erbB2
`cells formed rapidly growing tumors in syngeneic BDIX rats.
`Treatment of established tumors with the erbB2-specific antibody
`toxin by intratumoral application resulted in pronounced inhibition
`of tumor growth. Systemic treatment also led to growth inhibition,
`but to a lesser extent. During the course of the treatment all animals
`developed a strong antibody response against the fusion protein.
`
`MATERIAL AND METHODS
`Cells and culture conditions
`The SKBR3 and MDA-MB468 human breast-tumor cell lines
`and the A431 human epidermoid-tumor cell line were maintained
`in DMEM (GIBCO, Eggenstein, Germany) supplemented with
`10% heat-inactivated FBS. The established rat schwannoma cell
`line NV2Cd and its derivative NV2Cd-erbB2 stably transfected
`with a human erbB-2 cDNA construct (kindly provided by Dr. M.
`Rajewsky) were generated by Nikitin et al. (1996). The cells were
`maintained in DMEM supplemented with 8% FBS, 2 mM L-
`glutamine, 100 µg/ml streptomycin, 100 U/ml penicillin, and 0.5
`mg/ml G418 (GIBCO) for NV2Cd-erbB2.
`
`Bacterial expression of scFv(FRP5)-ETA and preparation
`of inclusion bodies
`From the plasmid pSW202-5 (Wels et al., 1995), which encodes
`under the control of the IPTG-inducible tac promoter, the erbB2-
`
`Abbreviations: ETA, Pseudomonas exotoxin A; DMEM, Dulbecco’s
`modified Eagle’s medium; FBS, fetal bovine serum; PBS, phosphate-
`buffered saline; IC50, 50% inhibitory concentration; IPTG, isopropyl-b-D-
`thiogalactopyranoside; scFv, single chain Fv antibody fragment; PMSF,
`phenylmethylsulfonyl fluoride; PVDF, polyvinylidenedifluoride.
`
`Contract grant sponsor: Deutsche Forschungsgemeinschaft; Contract
`grant number: SFB 364 C1.
`
`*Correspondence to: Dr. Winfried Wels, Institute for Experimental
`Cancer Research, Tumor Biology Center, Breisacher Strasse 117, D-79106
`Freiburg, Germany. Fax: 149 761 206 1599. E-mail: wels@tumorbio.uni-
`freiburg.de
`
`Received 2 March 1997; Revised 6 June 1997
`
`IMMUNOGEN 2085, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`118
`
`ALTENSCHMIDT ET AL.
`
`specific scFv(FRP5)-ETA immunotoxin fused at the N-terminus to
`the E. coli ompA signal peptide, a synthetic FLAG epitope, and a
`polyhistidine tag, the ompA signal sequence was removed, result-
`ing in plasmid pSW220-5 which upon induction of expression
`allows the accumulation of large amounts of recombinant protein
`as inclusion bodies. Plasmid pSW220-5 was transformed into E.
`coli HB101 (Bolivar and Backman, 1979). A single colony was
`grown overnight at 37°C in terrific broth (12 g/l tryptone, 24 g/l
`yeast extract, 4 ml/l glycerol, 1.7 mM KH2PO4, 7.2 mM K2HPO4)
`containing 0.6% glucose and 100 µg/ml ampicillin. The culture was
`diluted 30-fold in the same medium and grown at 37°C to an OD550
`of 10; expression of the recombinant protein was induced with 1%
`lactose for 16 hr at 37°C. Cells were harvested by centrifugation at
`10,000g for 10 min at 4°C and the cell pellet from 1 l of culture was
`re-suspended in 2 3 v/w of ice-cold 40 mM Tris, pH 8.0, at 0°C.
`Cells were disrupted in a French press and the lysate was diluted
`1:4 in ice-cold 40 mM Tris, pH 8.0 and centrifuged for 20 min at
`9,000g. The supernatant was discarded, and the pellet was washed
`with 200 ml of ice-cold 40 mM Tris, pH 8.0, and centrifuged as
`described above. The washed pellet containing the inclusion bodies
`was stored at 270°C.
`
`Purification of scFv(FRP5)-ETA
`Inclusion bodies were re-suspended in 10 3 v/w of buffer A (8 M
`urea, 50 mM Tris, pH 9.0, 0.5 M NaCl) for 30 min at room
`temperature. The suspension was clarified by centrifugation at
`16,000g for 30 min at 4°C and solubilized scFv(FRP5)-ETA
`protein was purified via binding of the polyhistidine tag included in
`the molecule to Ni21 saturated chelating sepharose (Pharmacia
`Biotech, Freiburg, Germany) equilibrated with buffer A. The
`solution was loaded onto the column at a flow rate of 1.5 ml/min.
`Unspecifically bound proteins were removed by washing with
`buffer A containing 20 mM and 40 mM imidazole. Specifically
`bound scFv(FRP5)-ETA protein was eluted with buffer A contain-
`ing 200 mM imidazole. The protein content of the eluate was
`determined using a protein quantification kit (Biorad). The typical
`yield of purified protein was 30 mg/l of original bacterial culture
`with a purity of more than 90%, as determined by SDS-PAGE and
`Coomassie-brilliant-blue staining. ScFv(FRP5)-ETA was re-
`natured by rapid 1:10 dilution of the purified protein into a buffer
`containing 400 mM L-arginine in PBS at 4°C and subsequent
`dialysis against PBS.
`
`Cell-viability assay
`The in vitro cell-killing activity of scFv(FRP5)-ETA was mea-
`sured basically as described (Wels et al., 1992). Human tumor cells
`or rat NV2Cd-erbB2 cells were seeded in 96-well plates at a
`density of 1 3 104 cells/well in normal growth medium. Various
`concentrations of scFv(FRP5)-ETA protein were added to triplicate
`samples and the cells were incubated for 40 hr (human tumor cells)
`or 96 hr (NV2Cd-erbB2). We added 10 µl of 10 mg/ml MTT
`(3-(4,5-dimethylthiazole-2-yl)-2,5 diphenyltetrazolium bromide)
`(Sigma, Deisenhofen, Germany) in PBS to each well and incubated
`the cells for another 3 hr. Cells were lysed for 3 hr by the addition
`of 90 µl of 20% SDS in 50% dimethyl formamide, pH 4.7. The
`absorption at 590 nm was determined in a microplate reader
`(Dynatech, Denkendorf, Germany) as a measure of the number of
`viable cells in comparison with cells grown in the absence of
`recombinant protein.
`
`In vivo anti-tumor activity of scFv(FRP5)-ETA
`in immunocompetent rats
`Female BDIX rats (weight approximately 350 g) were purchased
`from Charles River (Sulzfeld, Germany). NV2Cd or NV2Cd-erbB2
`cells (1 3 107) were re-suspended in 100 µl of PBS and injected
`s.c. into the flanks of the animals (3 rats/group). After 7 days, when
`the tumors were palpable, the animals were treated either by
`intratumoral or by intravenous injection of purified scFv(FRP5)-
`ETA or control proteins for 10 consecutive days. For local
`treatment, 50 µl each of the antibody-toxins were injected directly
`into 2 opposite sides of the tumor at total daily doses of 12.5 and 25
`µg. Systemic treatment was performed by injection of immunotox-
`
`ins into the tail vein at daily doses of 125 and 250 µg/kg,
`corresponding to approximately 44 and 88 µg per injection for a
`350-g rat. Tumor diameters were measured once or twice weekly
`with a calliper, and tumor volumes were calculated according to the
`formula: length 3 (width)2 3 0.5 (Arteaga et al., 1993).
`
`Analysis of rat sera and preparation of cyst fluid
`Blood samples were taken from the tail vein of rats. After
`clotting at 4°C for 16 hr and centrifugation at 8,000g for 20 min at
`4°C serum was collected and stored at 220°C. After 15 µg of
`scFv(FRP5)-ETA or the irrelevant control protein scFv(gE10)-ETA
`were separated by SDS-PAGE on a 10% gel, they were blotted onto
`PVDF membranes (Millipore, Eschborn, Germany). The filters
`were probed with sera diluted 1:100 from tumor-bearing rats
`treated either with scFv(FRP5)-ETA or with PBS. As a positive
`control, a rabbit anti-ETA serum was used at 1:4,000 dilution.
`Binding of antibodies was detected by incubation of the filters with
`species-specific horseradish-peroxidase-labeled secondary antibod-
`ies and chemiluminescent detection using the ECL system (Amer-
`sham, Braunschweig, Germany). Cyst fluid was collected from
`established tumors by aspiration, centrifuged at 10,000g for 30 min
`and processed as described (Altenschmidt et al., 1997).
`
`RESULTS
`Expression and purification of scFv(FRP5)-ETA
`The 67-kDa antibody toxin scFv(FRP5)-ETA consists of an
`erbB2-specific scFv domain genetically fused to truncated Pseudo-
`monas exotoxin A (Wels et al., 1992). The ETA fragment represents
`the toxin’s translocation domain II and the enzymatic domain III
`(amino acids 252 to 613) which, upon binding to target cells,
`facilitate uptake into the cytosol and inhibition of cellular protein
`synthesis (Hwang et al., 1987; Ogata et al., 1992). At
`the
`N-terminus, the recombinant protein carries a synthetic FLAG
`epitope followed by a cluster of 6 histidine residues for purification
`via Ni21 affinity chromatography. In an earlier study, the scFv-
`(FRP5)-ETA protein was expressed in E. coli and directed to the
`periplasmic space via the ompA signal peptide included in the
`expression plasmid pSW202-5 (Wels et al., 1995). In order to allow
`the accumulation of large amounts of the recombinant protein as
`inclusion bodies in the cytoplasm of E. coli here, the ompA signal
`sequence was removed from the plasmid pSW202-5, resulting in
`the construct pSW220-5. This plasmid facilitates expression of a
`scFv(FRP5)-ETA molecule, which is identical to the previously
`used protein except for the lack of the secretion signal.
`The modified expression plasmid was transformed into E. coli
`strain HB101, and expression of the recombinant protein was
`induced by the addition of 1% lactose. Protein expression peaked
`16 to 20 hr after induction, in contrast to 3 hr when the synthetic
`IPTG was used for induction (data not shown). Cells were
`harvested, inclusion bodies were isolated and solubilized in a buffer
`containing 8 M urea, 50 mM Tris, pH 9.0, and 250 mM NaCl.
`Decreasing the pH of the buffer resulted in a lower yield of
`solubilized recombinant protein, whereas a lower NaCl concentra-
`tion negatively influenced the binding of denatured scFv(FRP5)-
`ETA to Ni21-saturated chelating sepharose in the subsequent
`purification step (data not shown). ScFv(FRP5)-ETA was purified
`from solubilized inclusion bodies via a single round of Ni21 affinity
`chromatography and re-natured by rapid dilution and subsequent
`dialysis as, described in ‘‘Material and Methods’’. The yield of
`purified protein was typically 30 mg/l of original bacterial culture,
`with a purity of more than 90% as determined by SDS-PAGE
`analysis (data not shown).
`
`Cytotoxic activity and specificity of scFv(FRP5)-ETA
`In order to determine the cytotoxic activity and specificity of
`scFv(FRP5)-ETA purified from inclusion bodies, in vitro cell-
`killing experiments were performed. SKBR3 and MDA-MB468
`human breast-carcinoma cells and A431 human epidermoid-tumor
`cells were incubated for 40 hr with the antibody toxin at concentra-
`tions ranging from 1 ng/ml to 1 µg/ml and the relative number of
`
`IMMUNOGEN 2085, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`TREATMENT OF SCHWANNOMA CELLS WITH ANTI-erbB2 IMMUNOTOXIN
`
`119
`
`viable cells in comparison with cells grown in the absence of toxin
`was determined. The results are summarized in Table I. ScFv(FRP5)-
`ETA purified from inclusion bodies was cytotoxic for SKBR3 cells,
`which highly over-express erbB2 on their surface (IC50 of 50
`ng/ml). The antibody toxin did not affect the growth of MDA-
`MB468 cells, which express high levels of the EGF receptor but
`only very low levels of erbB2, indicating that it is highly specific
`for erbB2-expressing cells. ScFv(FRP5)-ETA was also active on
`A431 cells, which express approximately 5 3 104 erbB2 molecules
`per cell. We have shown that A431 cells, despite moderate erbB2
`levels, are highly sensitive to scFv(FRP5)-ETA, probably due to
`increased receptor turnover caused by heterodimerization of erbB2
`with EGF receptor, which is highly over-expressed in A431 cells
`and activated via autocrine stimulation by TGF-a (Wels et al.,
`1995; Schmidt et al., 1996). As shown in Table I, the in vitro
`cell-killing activity of scFv(FRP5)-ETA expressed from a construct
`lacking the ompA signal sequence and purified from inclusion
`bodies was very similar to that of the recombinant protein used in earlier
`studies. This indicates that the properties of the antibody toxin were not
`altered by the different way of expression and purification.
`
`In vitro cell-killing activity of scFv(FRP5)-ETA
`Implantation of NV2Cd rat schwannoma cells in syngeneic
`BDIX rats initially results in a strong tumor-specific cytotoxic-T-
`cell response (Altenschmidt et al., 1997). However, the tumor cells
`eventually escape elimination by the host immune system, through
`the secretion of immunosuppressive factors such as isoforms of
`TGF-b. In this respect our model resembles a possible clinical
`situation, since in certain cases anti-tumor immune responses
`observed in cancer patients are not sufficient to accomplish tumor
`rejection. In contrast, a tumor-specific antibody toxin should not be
`affected in its activity by immunosuppressive agents. NV2Cd-
`erbB2 cells have been generated by Nikitin et al. (1996) via stable
`transfection of NV2Cd cells with a human erbB2 cDNA construct.
`The resulting cells display in vitro and in vivo growth characteris-
`tics indistinguishable from the parental cell line (data not shown).
`The expression of human erbB2 by NV2Cd-ErbB2 cells was
`confirmed by immunoblot analysis with a monoclonal antibody
`(MAb) specific for human erbB2 (Fig. 1a). The binding of the
`scFv(FRP5)-ETA immunotoxin to the surface of NV2Cd-erbB2
`cells was verified by FACS analysis (data not shown). The in vitro
`cytotoxic activity of
`scFv(FRP5)-ETA on NV2Cd-erbB2
`schwannoma cells was analyzed in cell-killing experiments. The
`cells were incubated for 96 hr with the antibody toxin at concentra-
`tions ranging from 1 ng/ml to 10 µg/ml, and the relative number of
`viable cells was determined. The results are shown in Figure 1b. In
`contrast to the erbB2-over-expressing tumor cell lines shown in
`Table I, NV2Cd-erbB2 cells were only moderately sensitive to
`scFv(FRP5)-ETA in vitro, which might reflect their lower level of
`erbB2 expression of approximately 1.5 3 105 receptors/cell.
`However, parental erbB2-negative NV2Cd cells remained unaf-
`fected by the immunotoxin, indicating that the growth inhibitory
`effect was specific (Fig. 1b). As shown for human erbB2-expressing
`tumor cells (Harwerth et al., 1992) the parental erbB2-specific antibody
`FRP5 in the absence of the cytotoxic ETA effector domain had no effect
`on the growth of NV2Cd-erbB2 cells (data not shown).
`
`In vivo anti-tumoral activity of scFv(FRP5)-ETA
`BDIX rats carrying established s.c. NV2Cd-erbB2 tumors (3
`animals/group) were treated with the erbB2-specific scFv(FRP5)-
`ETA toxin. NV2Cd-ErbB2 schwannoma cells (1 3 107) were
`injected s.c. into the flanks of syngenic immunocompetent rats.
`After 7 days, when the tumors were palpable, the animals were
`treated with scFv(FRP5)-ETA for 10 days and the effects of local
`and systemic treatment on tumor growth were investigated. The
`results are shown in Figure 2. Intratumoral application of 25 µg/day
`resulted in significant growth inhibition of NV2Cd-erbB2 tumors in
`comparison with PBS-treated controls, whereas treatment with a
`lower daily dose of scFv(FRP5)-ETA (12.5 µg/day) was not
`effective (Fig. 2a). In a different set of experiments, scFv(FRP5)-
`ETA was applied systemically by injection into the tail vein at daily
`
`TABLE I – IN VITRO CELL-KILLING ACTIVITY OF scFv(FRP5)-ETA PURIFIED
`FROM DIFFERENT SUB-FRACTIONS
`
`Cell line
`
`Number of
`erbB2 receptors1
`
`IC50 (ng/ml)2
`Total lysates3
`Inclusion bodies
`
`A431
`MDA-MB468
`SKBR3
`
`5 3 104
`,5 3 103
`1–2 3 106
`
`33
`.1000
`34
`
`18
`.1000
`50
`
`1ErbB2-receptor numbers were reported by Jannot et al. (1996).–
`2IC50 values were determined in a cell viability assay, as described in
`‘‘Material and Methods’’.–3The cell-killing activity of scFv(FRP5)-
`ETA purified from total cell lysates were determined earlier (Wels et
`al., 1995).
`
`doses of 125 or 250 µg/kg for 10 consecutive days (Fig. 2b).
`Systemic treatment with 250 µg/kg/day inhibited NV2Cd-erbB2
`tumor growth, but was less effective than local treatment with 25
`µg/day. Treatment with 125 µg scFv(FRP5)-ETA/kg/day had no
`effect on in vivo tumor growth. Upon termination of the treatment,
`the tumors regrew in all animals. The antibody toxin was well
`tolerated by the animals, and no weight loss or other signs of
`systemic toxicity were observed (data not shown).
`Specificity of scFv(FRP5)-ETA in vivo
`In order to confirm that the anti-tumoral effects of scFv(FRP5)-
`ETA were mediated via specific binding of the toxin to erbB2, a
`similar experiment was performed with animals carrying tumors of
`the parental NV2Cd cells, which do not express human erbB2.
`Established s.c. NV2Cd tumors in BDIX rats were treated by
`intratumoral injection of 25 µg/day of scFv(FRP5)-ETA for 10
`consecutive days. As controls, animals carrying NV2Cd-erbB2
`tumors were treated either with the same dose of scFv(FRP5)-ETA
`or with PBS (3 animals/group). The results are shown in Figure 3a.
`As observed before, treatment with 25 µg/day of the antibody toxin
`inhibited the growth of NV2Cd-erbB2 tumors. In contrast, the
`growth of erbB2-negative NV2Cd tumors was not affected by scFv-
`(FRP5)-ETA, indicating that the antibody toxin is strictly dependent on
`erbB2 on the tumor cell surface to elicit a therapeutic effect.
`To further exclude the possibility that unspecific toxic effects
`contribute to the observed inhibition of tumor growth, established
`NV2Cd-erbB2 tumors were treated as described above by intratu-
`moral injection of 25 µg/day of the antibody toxin scFv(225)-ETA
`for 10 consecutive days. ScFv(225)-ETA is very similar in its
`structure to scFv(FRP5)-ETA, but
`is selectively cytotoxic for
`human tumor cells over-expressing the EGF receptor (Wels et al.,
`1995). As shown in Figure 3b, at an identical dose the EGF-receptor-
`specific antibody toxin, in contrast to scFv(FRP5)-ETA, had no
`effect on the in vivo growth of NV2Cd-erbB2 tumors. These results
`clearly show that the growth inhibitory effects of scFv(FRP5)-ETA
`are highly specific.
`
`Effect of treatment schedule on the anti-tumoral activity
`of scFv(FRP5)-ETA
`To analyze the dependency of scFv(FRP5)-ETA anti-tumoral
`activity on the period of tumor establishment before treatment,
`BDIX rats carrying s.c. NV2Cd-erbB2 tumors were treated by
`intratumoral injection of 25 µg/day of scFv(FRP5)-ETA either from
`day 7 to day 16 or from day 13 to day 22 after implantation of the
`tumor cells. Control animals were treated with PBS from day 13 to
`day 22. The results are shown in Figure 4. Tumor growth was
`suppressed by scFv(FRP5)-ETA during the treatment period when
`the treatment was begun on day 7. In contrast, when the tumors
`were larger at the onset of treatment on day 13, no significant
`inhibition of tumor growth at the same dose of scFv(FRP5)-ETA
`was observed in comparison with PBS-treated controls.
`We have shown that NV2Cd tumors develop necrotic areas
`inside the tumor approximately 15 days after s.c. implantation of
`cells and form fluid-filled cysts (Altenschmidt et al., 1997).
`Proteolytic activity in the cyst fluid could be responsible for rapid
`degradation of intratumorally injected scFv(FRP5)-ETA toxin. In
`order to test this possibility, scFv(FRP5)-ETA protein was incu-
`
`IMMUNOGEN 2085, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`120
`
`A
`
`ALTENSCHMIDT ET AL.
`
`FIGURE 1 – (a) Immunoblot analysis of cell lysates from NV2Cd transfectants. Equal amounts of cell lysates from NV2Cd-erbB2 rat
`schwannoma cells stably transfected with human erbB2 cDNA, parental NV2Cd cells, and erbB2-over-expressing SKBR3 human mammary-
`carcinoma cells were separated by electrophoresis on 7.5% SDS-polyacrylamide gels and transferred to PVDF membranes as indicated.
`Immunodetection was performed with a MAb specific for human erbB2, followed by incubation with an anti-mouse horseradish-peroxidase-
`labeled antibody and chemiluminescent detection. The position of the erbB2 protein is indicated. (b) Inhibition of the in vitro growth of
`NV2Cd-erbB2 schwannoma cells. NV2Cd-erbB2 cells and erbB2-negative parental NV2Cd cells were incubated for 96 hr with the indicated
`concentrations of the erbB2-specific scFv(FRP5)-ETA, and the number of viable cells in comparison with PBS-treated cells was determined by an
`enzymatic assay, as described in ‘‘Material and Methods’’.
`
`bated in vitro for different time intervals with cyst fluid removed
`from 23-day-old tumors at a ratio of 1 µl of cyst fluid per µg of
`protein. The integrity of the fusion protein was subsequently
`analyzed by SDS-PAGE and immunoblotting with a specific
`antibody. As shown in Figure 5, the immunotoxin was partially
`degraded after 2 hr of incubation with cyst fluid at 37°C (lane B)
`and completely degraded after 4 hr of incubation (lane C). These
`results show that proteolytic activity is present in the necrotic areas of
`NV2Cd-erbB2 tumors, and rapidly destroys the scFv(FRP5)-ETA
`immunotoxin. This explains the failure of the antibody toxin to inhibit
`the growth of tumors at time points when cysts have already formed.
`Specific antibody response in scFv(FRP5)-ETA-treated animals
`After termination of the in vivo experiments, serum was obtained
`from the tumor-bearing rats, in order to investigate the generation
`of immunotoxin-specific antibodies. Purified scFv(FRP5)-ETA was
`separated by SDS-PAGE, blotted onto PVDF membranes, and
`incubated with rat sera. The results are shown in Figure 6a. Both,
`sera from rats treated systemically with different doses of scFv-
`(FRP5)-ETA (lanes 1 and 2) and sera from animals treated by
`intratumoral injection (lanes 3 and 4) contained anti-immunotoxin
`antibodies with higher titers in the sera of systemically treated rats.
`As expected, no anti-scFv(FRP5)-ETA antibodies developed in
`PBS-treated animals (lane 5). The antibodies were directed mainly
`against the toxin domain, since the sera reacted equally with an
`irrelevant scFv(9E10)-ETA control protein containing a different
`antibody domain specific for a myc epitope (data not shown) (Fig.
`6b, lanes 8–11). Neither PBS-treated animals nor animals treated
`with the immunotoxin developed antibodies directed against the
`human erbB2 molecule (data not shown).
`The capacity of the anti-immunotoxin antibodies to neutralize
`the anti-tumoral activity of scFv(FRP5)-ETA was tested. Rats
`
`carrying s.c. NV2Cd-erbB2 tumors were treated for 10 days as
`described above by intratumoral injection of 25 µg/day of scFv-
`(FRP5)-ETA pre-treated with the serum obtained from toxin-
`treated animals, at a concentration of 20 µl serum/mg recombinant
`antibody toxin. As shown in Figure 6c, no significant reduction in
`the anti-tumoral activity of scFv(FRP5)-ETA was observed upon
`pre-treatment with the rat serum as compared with the results of
`treatment with the native antibody toxin. This is consistent with the
`results of
`in vitro cell-killing experiments with erbB2-over-
`expressing human tumor cells and scFv(FRP5)-ETA in the pres-
`ence of rat sera. In these experiments, no difference in the effects of
`sera from toxin-treated or PBS-treated animals on the cytotoxic
`activity of the immunotoxin was observed (data not shown),
`indicating that upon treatment with the recombinant toxin a humoral
`anti-toxin response develops. However, this response did not result in
`neutralization of the anti-tumoral activity of the fusion protein.
`
`DISCUSSION
`
`Members of the erbB family of receptor tyrosine kinases have
`been shown to play an important role in tumor development and
`progression. In particular, over-expression of erbB2 and the EGF
`receptor has been correlated with poor clinical outcome in sub-sets
`of human malignancies (Hynes and Stern, 1994; Gullick, 1991).
`Due to the differential expression of such receptors, high on the
`tumor cell surface and low on most normal tissues, therapeutic
`strategies specifically targeted to erbB2 and the EGF receptor hold
`promise to improve the clinical situation. MAbs specific for the
`extracellular domains of such receptors have been shown to
`interfere with signal transduction, and are being evaluated in
`clinical studies for their therapeutic effects (Baselga and Mendel-
`
`IMMUNOGEN 2085, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`TREATMENT OF SCHWANNOMA CELLS WITH ANTI-erbB2 IMMUNOTOXIN
`
`121
`
`FIGURE 2 – Effect of scFv(FRP5)-ETA on the in vivo growth of NV2Cd-erbB2 schwannoma cells. Tumor cells (1 3 107) were injected s.c. in
`syngeneic BDIX rats on day 0. From day 7 to day 16 the animals received daily injections of scFv(FRP5)-ETA directly into the tumor (a) of 25
`µg/day (filled triangles) or of 12.5 µg/day (open triangles), or systemically into the tail vein (b) of 250 µg/kg/day (closed circles) or of 125
`µg/kg/day (open circles). Control animals were treated with PBS (open squares). Tumor size was measured at the indicated time points and tumor
`volumes were calculated. The mean values for each group are shown and the standard deviation is represented by error bars.
`
`FIGURE 3 – Specificity of scFv(FRP5)-ETA in vivo. (a) NV2Cd-erbB2 (closed circles) or erbB2-negative parental NV2Cd schwannoma cells
`(open circles) (1 3 107 in each case) were injected s.c. in syngeneic BDIX rats on day 0. From day 7 to day 16 the animals received daily
`intratumoral injections of 25 µg/day of scFv(FRP5)-ETA. Control animals were treated with PBS (open squares). (b) NV2Cd-erbB2 tumor cells
`(1 3 107) were injected s.c. in syngeneic BDIX rats on day 0. From day 7 to day 16 the animals received daily intratumoral injections of 25 µg/day
`of the erbB2-specific scFv(FRP5)-ETA (closed circles) or the EGF-receptor-specific control protein scFv(225)-ETA (closed squares). Control
`animals were treated with PBS (open squares). Tumor size was measured at the indicated time points and tumor volumes were calculated. The
`mean values for each group are shown and the standard deviation is represented by error bars.
`
`sohn, 1994; Baselga et al., 1996). The growth-inhibitory potential
`of such antibodies can be potentiated by connecting them to
`cytotoxic effector functions such as bacterial toxins. We have
`described the potent anti-tumoral activity of recombinant single-
`chain immunotoxins directed to erbB2 and the EGF receptor (Wels et
`al., 1992, 1995; Schmidt and Wels, 1996; Schmidt et al., 1996, 1997).
`
`Here we have studied the anti-tumoral activity of the erbB2-
`specific scFv-Pseudomonas exotoxin-A fusion protein scFv(FRP5)-
`ETA in immunocompetent rats. The in vivo effects of molecules
`such as immunotoxins containing species-specific binding domains
`are generally evaluated in nude mice or rats carrying human tumor
`xenografts which would be rejected in animals capable of develop-
`
`IMMUNOGEN 2085, pg. 5
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`122
`
`ALTENSCHMIDT ET AL.
`
`to play a role in the development of human Schwann-cell tumors.
`However, erbB2/neu-transformed rat
`tumors might serve as a
`model for human brain tumors such as astrocytomas and meningio-
`mas, in which the expression of erbB2 has been documented
`(Schwechheimer et al., 1994; Schlegel et al., 1993).
`Upon s.c. injection, NV2Cd schwannoma cells form rapidly
`growing tumors in BDIX rats (Nikitin et al., 1996; Altenschmidt et
`al., 1997). NV2Cd-erbB2 cells which are transfected with a human
`erbB2 cDNA construct displayed in vitro and in vivo growth
`characteristics indistinguishable from the parental cell line. Estab-
`lished s.c. NV2Cd-erbB2 tumors were treated either locally by
`intratumoral
`injection of scFv(FRP5)-ETA or systemically by
`injection into the tail vein. Both routes of application resulted in the
`inhibition of tumor growth during the treatment, local treatment
`being more effective. The anti-tumoral activity of scFv(FRP5)-ETA
`was mediated via specific binding to erbB2, since the toxin did not
`affect the growth of the erbB2-negative parental tumor cells, nor
`did a similar fusion protein directed to the human EGF receptor
`affect the growth of NV2Cd-erbB2 tumors. During local treatment
`with 25 µg/day of scFv(FRP5)-ETA,
`tumor growth could be
`suppressed almost completely. However, upon termination of the
`treatment
`tumors regrew in all animals. In contrast
`to many
`established human tumor cell lines, NV2Cd-erbB2 cells express
`only