lnt J Clin Lab Res 21: 159-164, 1991

REs dI 9 Springer-Verlag 1991

Retargeting of human lymphocytes against human ovarian carcinoma cells by bispecific antibodies: from laboratory to clinic Delia Mezzanzanica, Silvana Canevari, and Maria I. Colnaghi Division of Experimental Oncology E, Istituto Nazionale Tumori, Via Venezian 1, 1-20133 Milan, Italy

Summary. We have selected a monoclonal antibody (MOvl8) reactive with ovarian carcinoma, which exhibits a restricted tumor specificity, a high affinity constant and which recognizes a 38-kDa glycoprotein homogeneously expressed on the cell surface of 90% of human ovarian carcinomas. Localization studies with radiolabelled MOv18 showed that MOv18 could specifically reach ovarian carcinoma cells growing in the peritoneal cavity of nu/nu mice. MOvl 8 did not mediate antibodydependent cellular cytotoxicity via Fr and could not be used as a carrier for toxins due to poor internalization of the antigen-antibody complex. Bispecific F(ab')2 antibodies made with MOvl 8 and an antibody reactive with CD3 were able to redirect the cytotoxicity of peripheral blood lymphocytes (PBL) against ovarian carcinoma cells both in vitro and in vivo in an animal model. The treatment of athymic mice bearing a human ovarian carcinoma intraperitoneally, with human PBL coated with the bispecific F (ab')2, significantly prolonged survival of the animals compared with tumour-bearing untreated and control mice treated with PBL alone. Four ovarian cancer patients have been treated with autologous lymphocytes coated with this bispecific F(ab')2 in a phase I clinical trial. No serious side-effects were observed but patients developed human anti-murine antibodies mainly directed against the idiotype of MOvl8. We have now begun phase II clinical trials in ovarian cancer patients. Key words: Ovarian carcinoma - Target cellular cytotoxicity - Bispecific antibody - Athymic mice Choice of the anti-tumor reagent One of the major aims of our research group is the production and characterization of monoclonal antibodies (mAbs) suitable for use in clinical oncology [6]. Theoretically for such applications the mAbs should be reactive with tumour cells only, even though some reagents withOffprint requests to: M. I. Colnaghi

out such restricted reactivity have been successfully applied in immunohistochemistry and immunocytology [21]. Moreover, mAbs selected on the basis of an operational tumor specificity [7] were used ex vivo for bone marrow purging in autologous bone marrow transplantation after high doses ofchemo- or radiotherapy [41]. In the case of the in vivo use of anti-tumor mAbs, in addition to the specificity which is essential to avoid the systemic side-effects of conventional therapy, many other factors such as binding kinetics and mAb affinity, heterogeneous expression of the recognized antigen, presence of the relevant circulating antigen and fate of the antigenantibody complex, have to be carefully considered [11]. We selected a mAb (MOvl8) against ovarian carcinoma which exhibits a restricted tumor specificity, a high affinity constant (109 M - t ) and which recognizes a 38-kDa glycoprotein homogeneously expressed on the cell surface of 90% of human ovarian carcinomas [27]. Due to these characteristics MOvl 8 seemed to be a good candidate for in vivo diagnostic and therapeutic applications [8]. Ovarian carcinoma is one of the most severe human malignancies as far as early diagnosis and 5-year survival are concerned. Until the disease reaches stages II b and III metastatic spreading is confined to the abdominal cavity [42]. This could allow a locoregional treatment which would facilitate the mAbs to reach the antigenbearing cells. This has been confirmed in preliminary experiments in which radiolabelled MOvl 8 has been used for the localization of tumor masses in the peritoneal cavity of ovarian cancer patients [9]. The availability of a reliable animal model which mim. ics disease progression in the human would allow the evaluation of different therapeutic approaches in preclinical studies. Two human ovarian carcinoma lines, IGROV1 and OVCAR-3 [1, 16], which grow both in vitro and in vivo in the peritoneal cavity of athymic nu/nu mice and whose characteristics are summarized in Table 1, were found to be suitable for pre-clinical studies. Due to its high expression of the relevant antigen, IGROV1 has been selected for localization studies with radiolabelled

160

D. Mezzanzanica et al. : Retargeting of human lymphocytes by bispecific antibodies

Table I. Characteristics of the human ovarian carcinoma lines used

Table 2. Bifunctional antibodies used

Cell line Reactivity with MOvl8 a

In vivo tumor development b

Type

No. cells No. cells Affected injected recovered ~ organs day 0 day 15

HeteroChemicalconconjugate jugation of parental mAbs

% Posi- Intentive cells sity IGROV1 96

+++

1.2•

4•

OVCAR3 78

++

107

2 x 10s

Pancreas, diaphragm liver, mesenteric lymph nodes Pancreas, mesenteric lymph nodes

a Evaluated by immunofluorescence on live cells b Tumor cells were injected i.p., for both cell lines the tumoral uptake was 90% Tumor cells contained in the ascites

Heteroisotypic hybrid mAb Homoisotypic hybrid mAb Recombinant F (ab') 2

Production

Fusion of hybridomas producing mAbs of different isotypes Fusion of hybridomas producing mAbs of the same isotype Chemicalcrosslinking of parental Fab'

Advantages

Disadvantages

Easy preparation

Preparation not homogeneous Instability of the chemical linkage Possibleeasy Hybrid hybridelimination oma instability of parental mAb Hybrid Possible difficult hybridoma elimination of stability parental mAbs Stability of chemical linkage

Low yield excessive chemical manipulation

mAb, Monoclonal antibody MOvl 8 [12]. These studies indicated that M O v l 8 could specifically reach the IGROV1 tumor cells in the peritoneal cavity of the nu/nu mice.

mice [14, 39, 40]. In our studies, the CD3 activating molecule [35] was selected for human lymphocyte retargeting against human ovarian carcinoma cells.

Selection of the therapeutic approach

Production of bispecific antibodies

MOvl 8 was tested in a s 1Cr-release assay for its ability to mediate antibody-dependent cellular cytotoxicity (ADCC) against ovarian carcinoma lines. Human peripheral blood mononuclear cells (PBMC), activated with gamma-interferon or with recombinant interleukin-2 (rlL-2), were used as effector cells. In both cases the cytotoxicity mediated by MOvl 8 was comparable with that obtained with the effector cells alone, suggesting that MOvl 8 (IgG1) is unable to mediate ADCC, which is in agreement with the very low affinity o f murine IgG1 for human Fc receptors (FcRs) [10]. The binding o f MOv18 to its antigen induces only a poor internalization o f the antigen-antibody complex which could prevent the use of this mAb as a carrier for toxins which require to be internalized to exert their enzymatic activity [23, 29]. In accordance with this, immunoconjugates of MOv18 with the ribosome-inactivating protein restrictocin which lacks the B-like chain (which mediates internalization) and with blocked ricin were unable to inhibit protein synthesis o f target cells [4, 5]. Recently, a method has been developed by which the natural recognition system of the effector cells can be artificially manipulated, in order to retarget cytotoxic cells of any specificity to react specifically with selected target cells [18, 30, 38]. The method employs bifunctional antibodies (bi-mAbs) in which one specificity is directed against tumor cells, whereas the second is directed against an activating receptor on effector cells [36]. The bi-mAbs, by promoting the formation of effector-target conjugates, induce the cytotoxic cell to deliver a lethal hit [35]. Several studies have demonstrated the possibility o f specific retargeting of human effector cells against h u m a n tumor cells both in vitro [17, 20, 31] and in vivo in nu/nu

Three different methods have been used for the production of bi-mAbs. One method involves the chemical cross-linking of the parental mAbs to produce randomly cross-linked heteroconjugates which are mutlivalent [30]. Alternatively, it is possible to produce monovalent bimAbs using two different methods. By fusion of the parental mAb-producing hybridomas it is possible to obtain hybrid hybridomas which secrete mAbs derived from all the possible associations of heavy and light chains (parental mAbs, bi-mAbs and different mismatched antibodies) [26]. The other possibility is the production of recombinant bi-F(ab')2 obtained either by disulfide exchange using the native sulfhydryl groups [2], or by introduction of a maleimidie group between the two parental Fab' [15]. Table 2 summarizes the characteristics of the bi-mAbs used in our studies.

Pre-clinical in vitro activity of bi-mAbs The heteroconjugate and two hybrid antibodies (homoand heteroisotypic) were compared for their ability to mediate ovarian carcinoma cell lysis in the presence of cytotoxic lymphocytes o f different origins: alloreactive T cytotoxic clones and in vitro-activated P B M C obtained either from healthy donors or from ovarian carcinoma patients [3, 19, 23, 32]. As shown in Table 3, the two parental mAbs did not increase the cytotoxicity o f the effector cells, whereas the percentage of specific lysis was increased in the presence of the different bi-mAbs. The best results were obtained by using the homoisotypic hybrid mAb which was also able to retarget the activity o f T ceils from cancer pa-

D. Mezzanzanica et al.: Retargeting of human lymphocytes by bispecific antibodies Table 3. Bispecific mAb-dependent lysis of an ovarian carcinoma line by different cytotoxic T cells" mAb ~

None MOvl8 Anti-CD3 Heteroconjugate Hybrid (heteroisotypic) Hybrid (homoisotypic)

% of specific lysis induced by Cytotoxic T-cell clones c

PBMC from healthy donors ~

PBMC from ovarian cancer patients d

1 2 2 78 40 80

2 6 12 50 22 70

14 17 15 ND ND 46

ND. Not determined; PBMC, peripheral blood mononuclear ceils All numbers represent the mean of 2-5 experiments, PBMC were activated with phytohemagglutinin and recombinant interleukin-2 b Final antibody concentration 100 ng/ml; mAbs alone did not give tumor cell lysis r Effector to target ratio 10 : 1 d Effector to target ratio 100:l tients. In this last case, however, a higher effector to target ratio was necessary because of the lower efficiency of these effectors in comparison with those from healthy donors. The lowest activity was obtained with the heteroisotypic hybrid mAb, probably because the non-homologous reassociation has a low efficiency [26]. Although the heteroconjugates are easy to prepare and are able to mediate good target cell lysis, they cannot be considered as candidates for in vivo use. In fact, each preparation varies in relation to the degree of cross-linking (with consequent formation of species having different molecular weights) and activity. Moreover, their multivalent nature might cause the modulation of the relevant receptors and prevent the retargeting of the effector cells.

Optimization of bi-mAbs to be used in in vivo studies Since the IgG purified from hybrid hybridoma ascitic fluids theoretically contained all the possible combinations of heavy and light chains, including bivalent parental mAbs [26], the purification of the bispecific component of the homoisotypic hybrid mAb was carried out. The presence of parental mAbs in bivalent form may affect the activity of the bi-mAbs due to competition for the relevant binding sites; effector cell auto-killing may also be mediated by the bivalent anti-CD3 [34]. A two-step purification protocol was carried out. The first step was the purification of the ascitic fluid by antiM O v l 8 idiotype affinity chromatography [33]. This allowed the elimination of the anti-CD3 bivalent component: the unbound material was reactive on activated PBMC, but not on ovarian cancer ceils. After elution, the bound material was further fractionated using a hydroxyapatite high-performance liquid chromatography column; this allowed the separation of the bi-mAb and the bivalent MOvl 8. Purification of the bi-mAb improved its ability to mediate target cell lysis in the presence of PBMC indi-

161

cating that the presence of parental or mismatched mAbs in unpurified preparations may affect the system [33]. Furthermore, in order to avoid FcR-mediated lysis which may occur on cells negative for the relevant antigen but bearing FcRs, we prepared and tested bi-F(ab')2 fragments [23]. The homoisotypic hybrid mAb was digested with pepsin and fractionated by ion exchange chromatography to eliminate the parental F(ab') 2 contained in the mAb mixture from the hybrid hybridoma. The activity of this fractionated bi-F (ab')z (named antiOC/TR) in retargeting human T cells was compared with that of a chemically produced bi-F(ab')2 obtained by using MOvl8 as the anti-tumor component and TR66 [19] as the anti-CD3 component. The two parental mAbs were digested with pepsin and then reduced to Fab'. The bi-F (ab')2 was then prepared by introducing a thioether bond between the two fragments (MOvld-TR66). As shown in Fig. 1, the elimination of the Fc portion of the bi-mAb did not affect the lysis of MOvl 8 antigen-bearing cells, but eliminated the non-specific lysis of FcR-positive, MOvl 8-negative cells. Both bi-F (ab')2 mediated a comparable lysis o f the target cells and were used for further studies.

Pre-clinical in vivo activity of bi-F (ab')2 The chemically obtained MOv18-TR66 bi-F(ab% and the natural hybrid anti-OC/TR bi-F(ab')2 were used to investigate whether a bi-mAb could retarget the activity of human activated peripheral blood lymphocytes (PBL) given to nu/nu mice bearing an established i.p. human ovarian carcinoma line [22]. For this study, OVCAR-3 was selected because of its ability to grow as ascite in the peritoneal cavity ofathymic mice. Human PBL were activated with anti-CD3 and rIL-2 [13, 25]. Figure 2 shows the protocols used for the pre-clinical studies in the nu/nu mice. The animals were injected i. p. with 107 OVCAR-3 cells on day 0. On day 4, when the tumor was already established, the mice received a single treatment of two doses, over a 4-h period, of bi-F(ab')zcoated PBL. The total amount o f bi-F (ab% was 15 I~g and the total number of PBL 3 x l0 T, with an estimated effector to target ratio of 1 : 1. Control mice were treated with PBL alone or PBL coated with parental mAbs. Tumor growth was evaluated 15 days after tumor injection by counting the tumor cells recovered after peritoneal wash. In the long-term survival experiments the mice were followed for 5 months after treatment; histological analysis was carried out on some o f the surviving mice, others were rechallenged with OVCAR-3. As shown in Fig. 3 the mice treated with PBL alone or PBL coated with parental mAbs (controls) were only slightly protected from tumor growth, whereas mice treated with PBL coated with either bi-F(ab% were protected in a highly significant way. The two b i - F ( a b % were not only able to protect the mice from ascites formation over a 15-day period, but they were also able to prolong the survival time of the tumor-bearing mice. In fact, by 150 days after treatment all untreated mice were dead; control mice had a slightly prolonged survival time.

162

D. Mezzanzanica et al.: Retargeting of human lymphocytes by bispecific antibodies 100

8O I

r

--O

60

'~

40

i

I

20 1] PBMC mAb

0

+ -

+

+ MOvl 8

TR66 F(ab')2

+ + + ~xOC/TR ctOC/TR MOv18-TR66 F(ab')2 F(ab')2

Fig. 1. Bispecific F(ab')z-mediated lysis of ovarian carcinoma cells (MOvl 8 positive, l-l) and Fc receptor-positive cells (MOvI8 negative, []) in the presence of human peripheral blood mononuclear cells (PBMC). Intact monoclonal antibodies (mAbs) or bi-F(ab') z (100 ng/ml) were used to retarget phytohemagglutinin-/recombinant interleukin-2-(rlL-2) activated PBMC against target cells at effector to target ratio of 10:1 in a standard 5~Cr-release assay

mAb- coated PBL 7.5 IJg mAb / 1.5 x 107 PBL per dose

107 OVCAR-3

:~* ~J7

I

I

Day 0

Day 4

Tumour growth evaluation 15 days

I

Fig. 2. Protocols used for in vivo retargeting of human peripheral blood lymphocytes (PBL) against established OVCAR-3 in nu/nu mice by bi-F(ab') 2. PBL were activated using immobilized anti-CD3 and rlL-2. Both the tumor injection and treatment were given intraperitoneally

>•-

Survival evaluation 150 clays

1

/ 100

(42/52) 8O

(16123) /

60 0

E

J/ 20

(11/64) ~

(4/20)

~

~ PBL mAb

0 v"

--

-

+

-

+

MOv18

+

TR66 F(ab')2

7

+ + ctOC/TR MOv18-TR66 F(ab')2 F(ab')2

Fig. 3. Tumor growth evaluation in nu/nu mice treated with retargeted human PBL. Tumor growth evaluation was performed by peritoneal washing. The numbers in parentheses represent the number of protected mice over the number of treated mice

D. Mezzanzanica et al.: Retargeting of human lymphocytes by bispecific antibodies In contrast, mice treated with bi-F (ab')z-Coated PBL had a mean survival time of 100 days, twice as long as controls (54 days) and 3 times as long as untreated animals (30 days). Fifty percent of the animals treated with anti-OC/ T R F (ab')z-coated PBL were still alive after 150 days of observation [22]. The histopathological analysis carried out on some of the surviving mice showed that some animals were tumor free and some had a necrotic tumor. The surviving mice rechallenged with OVCAR-3 developed tumors as expected, thus excluding the possibility that the observed protection was due to a natural immunity against the t u m o r [25].

Clinical trial The anti-OC/TR bi-F(ab') z is now being employed in a phase I clinical trial in ovarian cancer patients [22]. The bi-F (ab') 2 was purified under good manufacturing practice conditions for clinical application. The clinical protocol is summarized in Fig. 4. The patients had two cycles of treatment with activated autologous P B M C coated with a n t i - O C / T R bi-F(ab')2. No severe side-effects were observed in the four treated patients. Particular attention was given to the development of human anti-murine antibodies ( H A M A ) [37]. The H A M A which were found were mainly directed against the idiotype of the M O v l 8 component of the bi-F(ab') 2, whereas anti-idiotypic responses against the CD3 component were less evident. This could be explained by the fact that the patients received P B M C coated with the bi-F(ab')2; the anti-CD3 binding site was bound to the effector cells, whereas the M O v l 8 c o m p o nent was exposed during the time lapse needed to reach the tumor. M O v l 8 was not only available for immune

Clinical trial: phase I Second look ~ Patient selection Leukaphaeresis and lymphocyte activation PBMC-coating: 1 ~tg bi-F(ab')2/106 cells First cycle of treatment Day 1-4: excalating doses from 106 cells/1 ~tg bi-F(ab')z to 109 cells/1 mg bi-F (ab')z Day 5-12: rest Second cycle of treatment Day I3-17: maximal dose 109 cells/1 mg bi-F(ab') 2 Day 18, 19:1 mg bi-F(ab')2 Toxicity evaluation Fig. 4. Protocol used for the phase I clinical trial in ovarian cancer patients, with macroscopic residual disease after chemotherapy, using bi-F(ab')2-retargeted autologous PBMC

163

system recognition, but was also presented by the effector cells in the context of the m a j o r histocompatibility complex. We are still analyzing the nature of this anti-idiotypic response.

Conclusions The beginning of a phase I clinical trial should be the result of very accurate laboratory research work which allows the selection of the m o s t appropriate reagent to be used for a particular therapeutic application [24]. In our case the first step was the selection of a m A b against ovarian carcinoma with all the properties required for in vivo injection. It was then necessary to identify the most suitable therapeutic approach. Other important steps included the development of adequate methods for the purification of the bispecific reagent and the activation of the effector cells. A complex series of laboratory studies has allowed us to carry out phase I and, more recently, phase II clinical trials in ovarian cancer patients. Promising results have already been obtained using this therapeutic approach in malignant glioma patients [28] in whom a locoregional treatment was carried out.

References 1. Brnard J, De Silva J, De Blois MC, Boyer P, Duvillard P, Chiric E, Riou G, Characterization of a human ovarian adenocarcinoma line, IGROVI, in tissue culture and in nude mice. Cancer Res 45: 4970, 1985 2. Brennan M, Davison PF, Paulus H, Preparation of bispecific antibodies by chemical recombination of monoclonal immunoglobulin GI fragments. Science 229:81, 1985 3. Canevari S, Mrnard S, Mezzanzanica D, Miotti S, Pupa SM, Lanzavecchia A, Colnaghi M I, Anti-ovarian carcinoma anti-T3 heteroconjugates or hybrid antibodies induce tumor cell lysis by cytotoxic T-cells. Int J Cancer [Suppl] 2: 18, 1988 4. Canevari S, Cogliati T, Mezzanzanica D, Orlandi R, Colnaghi MI, Anticorpi monoclonali anti carcinomi umani quali vettori di tossine. J Chemother [Suppl] 2: 370, 1990 5. Cogliati T, Brusa P, Canevari S, Caldera M, Mezzanzanica D, Dosio F, Cattel L, Colnaghi MI, Preparation and biological characterization of conjugates consisting of ricin and a tumorspecific non-internalizing Mab. Anticancer Res 11:417, 1991 6. Colnaghi MI, Canevari S, Conde E Leoni F, Mrnard S, Orlandi R, Pizzetti P, Porro G, Ripamonti M, Monoclonal antibodies: perspectives for tumor therapy. In: GaUo RC, Della Porta G, Albertini A (eds) Monoclonals and DNA probes in diagnostic and preventive medicine. Raven, New York, pp 137-146, 1987 7. Colnaghi MI, Canevari S, Mrnard S, Monoclonal antibodies in immunodiagnosis and immunotherapy of human cancer. Eur J Surg Oncol 14: 1, 1988 8. Colnaghi MI, Buraggi GL, Canevari S, Di Re F, Leoni F, Mangioni C, M~nard S, Orlandi R, Evaluation of the suitability of a monoclonal antibody raised against human ovarian carcinoma for therapeutic approaches. Nucl Med Biol 16:633, 1989 9. Crippa F, Buraggi GL, Di Re E, Gasparini M, Seregni E, Canevari S, Gadina M, Presti M, Marini A, Seccamani E, Radioimmunoscintigraphy with t31I-MOv18 administrated by intraperitoneal and intravenous routes in ovarian cancer patients. Eur J Cancer 27: 724, 1991 10. Fanger MW, Graziano RF, Shen L, Guyre PM, FcyR cytotoxicity exerted by mononuclear cells. Chem Immunol 47:214, 1989

164

D. Mezzanzanica et al.: Retargeting of human lymphocytes by bispecific antibodies

11. Foon KA, Biological response modifiers: the new immunotherapy. Cancer Res 49: 1621, 1989 12. Gadina M, Canevari S, Ripamonti M, Mariani M, Colnaghi MI. Preclinical pharmacokinetics and localization studies of the radioiodinated anti-ovarian carcinoma MOvl8. Nucl Med Biol 18: 403, 199t 13. Garrido MA, Perez P, Titus JA, Valdayo MJ, Winkler DF, Barbieri SA, Wunderlich JR, Segal DM, Targeted cytotoxic cells in human peripheral blood lymphocytes. J Immunol 144: 2891, 1990 I4. Garrido MA, Valdayo MJ, Winkler DF, Titus JA, Heeht TT, Perez P, Segal DM, Wunderlich JR, Targeting human Tlymphocytes with bispecific antibodies to react against human ovarian carcinoma cells growing in nu/nu mice. Cancer Res 50: 4227, 1990 15. Glennie MJ, McBride HM, Worth AT, Stevenson GT, Preparation and performance of bispecific F (ab') z antibody containing thioether-linked Fab' fragments. J Immunol 139:2367, 1987 16. Hamilton TC, Young RC, Louie KG, Behrens BC, McKoy W M, Grotzinger KR, Ozol RF, Characterization of a xenograft model of human ovarian carcinoma which produces ascites and intraabdominal carcinomatosis in mice. Cancer Res 44:5286, 1984 17. Jung G, Honsik CJ, Reisfeld RA, Muller-Eberhard HJ, Activation of human peripheral blood mononuclear cells by anti-T3: killing of tumor target cells coated with anti-target anti-T3 conjugates. Proc Natl Acad Sci USA 83:4479, 1986 18. Karpovsky B, Titus JA, Stephany DA, Segal DM, Production of target-specific effector cells using hetero-cross-linked aggregates containing anti-target cell and anti-Fcy receptor antibodies. J Exp Med 160: 1686, 1984 19. Lanzavecchia A, Scheidegger D, The use of hybrid hybridomas to target human cytotoxic T lymphocytes. Eur J lmmunol 17: 105, 1987 20. Liu MA, Kranz DM, Kurnick JT, Boyle LA, Levy R, Eisen HN, Heteroantibody duplexes target cells for lysis by cytotoxic T lymphocytes. Proc Natl Acad Sci USA 82: 8648, 1985 21. Menard S, Rilke F, Della Porta G, Mariani-Costantini R, Regazzoni M, Tagliabue E, Alasio L, Colnaghi MI, Sensitivity enhancement of the cytologic detection of cancer cells in effusion by monoclonal antibodies. Am J Clin Pathol 83: 571, 1985 22. Mezzanzanica D, On behalf of the International Group for Ovarian Carcinoma Therapy, Human PBL retargeting by bispecific antibodies: preclinical data and phase l clinical trial. J Cell Biochem [Suppl] 15E: 137, 1991 23. Mezzanzanica D, Canevari S, M~nard S, Pupa SM, Tagliabue E, Lanzavecchia A, Colnaghi MI, Human ovarian carcinoma lysis by cytotoxic T cells targeted by bispecific monoclonal antibodies: analysis of the antibody components. Int J Cancer 41:609, 1988 24. Mezzanzanica D, Canevari S, Casalini P, Della Torre G, Colnaghi MI, Monoclonal antibodies in cancer therapy: selection of the most appropriate reagent for different approaches. In: Metzgar R, Mitchell M (eds) Human tumor antigens and specific tumor therapy. Liss, New York, p 181-190, 1989 25. Mezzanzanica D, Garrido MA, Neblock DS, Daddona PE, Andrew SM, Zurawski VR Jr, Segal DM, Wunderlich JR, Human T lymphocytes targeted against human ovarian

26. 27.

28. 29. 30. 31.

32.

33.

34. 35. 36. 37. 38. 39.

40.

41.

42.

carcinoma with bispecific F(ab')2 antibody prolong host survival in a murine xenograft model. Cancer Res, in press, 1991 Milstein C, Cuello AC, Hybrid hybridomas and the production of bi-specific monoclonal antibodies. Immunol Today 5:299, 1984 Miotti S, Canevari S, Menard S, Mezzanzanica D, Porro G, Pupa SM, Regazzoni M, Tagliabue E, Colnaghi MI, Characterization of human ovarian carcinoma-associated antigens defined by novel monoclonal antibodies with tumor-restricted specificity. Int J Cancer 39:297, 1987 Nitta T, Sato K, Yagita H, Okumura K, Ishii S, Preliminary trial of specific targeting therapy against malignant glioma. Lancet 335: 368, 1990 Pastan I, Willingham MC, FitzGerald DJ, Immunotoxins. Cell 47:641, 1986 Perez P, Hoffman RW, Show S, Bluestone JA, Segal DM, Specific targeting of cytotoxic T cells by anti-T3 linked to antitarget cell antibody. Nature 316: 354, 1985 Perez P. Hoffman RW, Titus JA, Segal DM, Specific targeting of human peripheral blood T cells by heteroconjugates containing anti-T3 crosslinked to anti-target cell antibodies. J Exp Med 163: 166, 1986 Pupa SM. Canevari S, Fontanelli R, Menard S, Mezzanzanica D, Lanzavecchia A, Colnaghi MI, Activation of mononuelear cells to be used for hybrid monoclonal antibody-induced lysis of human ovarian carcinoma ceils. Int J Cancer 42:455, 1988 Pupa SM, Canevari S, Colzani E, Damgard EM, Menard S, Miotti S, Colnaghi MI, Purification of bispecific monoclonal antibodies produced by a hybrid hybridoma. J lmmunol Res 3: 16, 1991 Roosnek E, Lanzavecchia A, Triggering T cells by otherwise inert hybrid anti-CD3/antitumor antibodies requires encounter with the specific target cell. J Exp Med 170:297, 1989 Segal DM, Snider DP, Targeting and activation of cytotoxic lymphocytes. Chem Immunol 47: 179, 1989 Segal DM, Wunderlich JR, Targeting of cytotoxic cells with heterocrosslinked antibodies. Cancer Invest 6: 83, 1988 Shawler DL, Bartholomew RM, Smith LM, Dillman RO, Human immune response to multiple injection of murine monoclonal IgG. J Immunol 135: 1530, 1985 Staerz UD, Kanagawa O, Bevan MJ, Hybrid antibodies can target sites for attack by T cells. Nature 314:628, 1985 Titus JA, Garrido MA, Hecht TT, Winkler DF, Wunderlich JR, Segal DM, Human T cells targeted with anti-T3 crosslinked to anti-tumor antibody prevent tumor growth in nude mice. J Immunol 138:4018, 1987 Titus JA. Perez P, Kaubish A, Garrido MA, Segal DM, Human K/natural killer ceils targeted with heterocrosslinked antibodies specifically lyse tumor cells in vitro and prevent tumor growth in vivo. J Immunol 139:3153, 1987 Tondini C, Pap SA, Hayes DF, Elias AD, Kufe DW, Evaluation of monoclonal antibody DF3 conjugated with ricin as a specific immunotoxin for in vitro purging of human bone marrow. Cancer Res 50: 1170, 1990 Young RC, Knapp RC, Perez CA, Cancer of the ovary. In: De Vita VTJ, Hellman S, Rosenberg SA (eds) Cancer principles and practice of oncology. Lippincott, Philadelphia, p 884, 1982