FEMS Microbiology Letters 189 (2000) 1^8

www.fems-microbiology.org

MiniReview

Antibody engineering and its applications in tumor targeting and intracellular immunization Patrick Chames a , Daniel Baty b

b;

*

a Department of Pathology, Maastricht University, Maastricht, the Netherlands Laboratoire d'Inge¨nierie des Syste©mes Macromole¨culaires, IBSM, CNRS, 31 chemin Joseph Aiguier, 13402 Cedex 20 Marseille, France

Received 2 March 2000; received in revised form 30 May 2000; accepted 30 May 2000

Abstract During the last decade, recombinant antibody engineering has emerged as one of the most promising approaches for the design, selection and production of molecules for basic research, medicine and the pharmaceutical industry. This MiniReview describes the major findings that have led to the development of this powerful technique, with an emphasis on the use of Escherichia coli and filamentous phage as a tool allowing powerful selection procedures from large libraries as well as the use of intracellular expression of antibody fragments as a new class of neutralizing molecules with a potential use in therapy. The future of these rapidly evolving technologies is discussed. ß 2000 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Antibody fragment; Engineering ; Phage display; Intrabody ; Escherichia coli

1. Introduction The idea to use antibodies as magic bullets to target toxins or radioisotopes to de¢ned cell types is now 100 years old [1]. A major step toward this goal was made 25 years ago by Kohler and Milstein who developed hybridoma technology, thereby providing a reliable source of monoclonal antibodies (mAbs) of known speci¢city [2]. However, until very recently, there have been few antibody-based products suitable for therapy available. This delay can be largely explained by the fact that mouse antibodies trigger a human anti-mouse antibody response. In the late 80s, methodologies were developed to produce active antibody fragments in Escherichia coli. A few years later, Greg Winter and colleagues demonstrated that it was possible to select antibody fragments displayed on the surface of ¢lamentous bacteriophage. This pioneering work provided the basis for the development of antibody engineering techniques which have in recent years produced many molecules that have been taken to the clinic. This review focuses on the possibilities o¡ered by these techniques and discusses exciting future applications for

* Corresponding author. Tel. : +33 (4) 91-16-41-17; Fax: +33 (4) 91-71-21-24; E-mail : [email protected]

recombinant antibodies with a special emphasis on their use as intracellular neutralizing molecules or `intrabodies'. 2. Engineering of antibody fragments In 1988, antibody engineering techniques were boosted by the groups of Skerra and Better who demonstrated that active antibody fragments (Fig. 1) could be expressed in E. coli. The smallest such fragment is the Fv fragment, which is obtained by association of the variable domains of the heavy chain (VH) and the light chain (VL) of the antibody [3,4]. However, the hydrophobic interactions between these two domains are not very strong. Thus engineering of a covalent link between the VH and the VL is necessary to obtain a stable molecule. The most common approach is to use a £exible peptide linker of 15^20 residues to join the two domains. The resulting fragment is called single-chain Fv fragment or scFv [5]. Another approach is to engineer a disul¢de bond at the interface between the VH and the VL [6]. This disul¢de-linked Fv is generally more stable to thermal denaturation in serum than the scFv. The second fragment currently in use is the Fab fragment, made by the association of the whole light chain and the Fd chain (Fig. 1) [4]. This fragment possesses several advantages over the scFv fragments : (1) the two chains

0378-1097 / 00 / $20.00 ß 2000 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 0 0 ) 0 0 2 5 1 - 2

FEMSLE 9478 14-7-00

Cyaan Magenta Geel Zwart

2

P. Chames, D. Baty / FEMS Microbiology Letters 189 (2000) 1^8

Fig. 1. Schematic representation of several di¡erent formats of Ig fragments. The IgG is composed of two identical heavy chains (VH+CH1+CH2+CH3) and two identical light chains (VL+CL). Both chains are organized as domains containing about 110 amino acids with each domain possessing an intra-disul¢de bond (not shown). Inter-disul¢de bonds (red lines) link the light chain to the heavy chain and the two heavy chains together. The variable domains (VH and VL) contain the complementary determining regions (CDRs) which bind to the antigen (CDR represented by small dots). The Fv fragment (VH and VL domains) possesses the binding activity. The scFv corresponds to the VH linked to the VL by a £exible peptide linker (black lines). Dia-, tri- and tetrabodies can be obtained by using short linkers (represented by black lines or dots).

are naturally covalently associated ; (2) the Fab fragment does not have the ability to dimerize (which can be problematic for a¤nity determination as observed with scFvs, see below) ; (3) the CH1 domain can be used as a tag for detection ; (4) the Fab fragment is usually more stable in the puri¢ed form, and (5) its larger size may be advantageous for in vivo applications, due to a slower blood clearance. However, the Fab fragment requires bicistronic constructs to allow simultaneous expression of the light and Fd chains in E. coli. In order to recreate the bivalency of full-size immunoglobulin G (IgG) molecules, some authors have fused the CH3 domain to a scFv or Fab fragment. The resulting dimeric molecules called minibodies (Fig. 1) show decreased dissociation rates. The higher molecular weight of the dimers is also more suitable for in vivo targeting [7]. The genes of scFv and Fab fragments can be easily engineered and this property has been used to humanize murine fragments. This can be achieved either by grafting of murine antigen binding loops onto human antibody framework regions or by replacing surface residues of the murine fragment by their human analogue (for review,

FEMSLE 9478 14-7-00

see [8]). It is also possible to confer additional properties to antibody fragments such as avidity, multispeci¢city as well as/and e¡ector function. 2.1. Engineering multivalency The single-chain format is particularly suited for antibody engineering. It is for example possible to fuse, using another peptide linker, two scFvs in tandem to produce a molecule with two binding sites. As for minibodies, such molecules have been demonstrated to have dramatically decreased dissociation rates with cell-bound antigen [9]. However, peptide linkers are sometimes rather sensitive to proteases. Another more elegant approach to increase the valency of the scFv fragment has been proposed. As mentioned above, the hydrophobic interaction between the VH and VL domains is not very strong. This can lead to dimer formation where the VH of one molecule interacts with the VL of another and vice versa. Consequently, most scFv fragments are found in two forms: monomer and dimer. Holliger and colleagues have exploited this property to create a bivalent scFv fragment [10]. They

Cyaan Magenta Geel Zwart

P. Chames, D. Baty / FEMS Microbiology Letters 189 (2000) 1^8

3

demonstrated that if the peptide linker is shortened from 15 to ¢ve residues, the VH and VL of the scFv are no longer able to bind to one another. The equilibrium is displaced towards production of scFv dimers also known as diabodies (Fig. 1). The dissociation rate of the diabody is signi¢cantly lower than that of the parental scFv [9], and this compact linker-free molecule is thought to be more stable than the tandem scFv. More recently, it has been demonstrated that if the linker is reduced to one residue, the scFvs are preferentially found as tetramers (tetrabodies), whereas no linker at all mainly leads to the formation of trimers (triabodies) (Fig. 1) [9]. However, the ratio of each species is di¤cult to predict and depends on the sequence of each scFv. Another dimerization technique consists in fusing the fragments to another domain capable of self-association. Minibodies, described above, use the IgG CH3 domain. Other domains, such as the leucine zipper, have also been used [11]. 2.2. Engineering multispeci¢city Antibody fragments can be engineered to have several speci¢cities. Bispeci¢c antibodies are a desirable tool, but the ¢rst attempts to create such molecules using chemical modi¢cations of mAbs or hybrids of hybridomas were hampered by the requirement of extensive puri¢cation and as a result were rather ine¤cient. Recombinant antibody technology o¡ers several approaches that may be employed for the production of bispeci¢c antibodies. Two scFvs of di¡erent speci¢cities can be linked via a peptide linker in a tandem construction and expressed in E. coli [12]. In this case, only the molecule of interest is produced. However, the diabody approach has been more widely employed. To produce a bispeci¢c diabody, the VH of scFv `a' is fused to VL of scFv `b' via a ¢ve residue linker, similarly VHb is fused to VHa. The coexpression of these two chains forces interaction of VHa with VLa and VHb with VLb, creating a heterodimeric bispeci¢c molecule. Such constructs have produced interesting results in several studies and may play an important role in therapy, for example in T-cell retargeting against tumor cells (for review, see [13]). Although not demonstrated yet, this approach could be extended to triabodies and tetrabodies, to generate multispeci¢c molecules of modest molecular weights (Fig. 1). 2.3. Fusion with other molecules Other proteins or protein fragments can be fused to antibody fragments to equip them with additional properties. The most studied approach in this ¢eld is the production of immunotoxins (see [14] for an excellent review). These molecules are made by the fusion of a tumor-speci¢c scFv or Fab to a toxin capable of killing the target cell once internalized. Numerous impressive results have

FEMSLE 9478 14-7-00

Fig. 2. Model for the protection of E. coli from colicin A toxin by scFv. Colicin A is composed of three domains: the central domain (M) that binds to the receptor (OmpF and Btub), the N-terminal domain (N) that crosses the outer membrane (OM) to interact with TolA and the C-terminal domain (C). Killing of the bacterial cell is mediated by the C-terminal domain (C) that inserts into the inner membrane (IM) to form pores. When the scFv is produced in the periplasmic space of bacteria, it binds to the N domain of colicin A, thereby preventing colicin A activity.

been obtained with these molecules and several immunotoxins are currently being tested in clinical trials. Tumor-speci¢c antibody fragments have also been fused to cytokines. In this case, the molecule, called an immunocytokine, is injected into the patient and accumulates on the tumor cell surface, thereby allowing T-cells in the vicinity of the tumor to be activated (for review, [15]). The intrinsic tumor binding activity of these scFv-IL-2 fusions allowed the use of low concentrations and produced impressive results, without the side e¡ects normally associated with systemic IL-2 injections. For biochemical applications, several fusions of antibody fragments with enzymes such as alkaline phosphatase have been created [16]. These reagents are inexpensive and easily produced and do not su¡er from the classical problems presented by chemically labeled reagents such as partial inactivation or heterogeneity. Recently, the activity of the E. coli biotin ligase BirA has been used for the production of biotinylated antibody fragments [17]. The 13 residue sequence recognized by this enzyme was fused with the antibody fragments. Incubation of the puri¢ed fragments with the enzyme allowed addition of a unique biotin to this sequence. The BirA enzyme has also been over-expressed in E. coli to produce in vivo biotinylated antibody fragments [18]. These methods are advantageous over chemical modi¢cations in that the resulting molecules

Cyaan Magenta Geel Zwart

4

P. Chames, D. Baty / FEMS Microbiology Letters 189 (2000) 1^8

Fig. 3. The phage display technique. Libraries of phage antibodies (Phabs) are creating by fusing a repertoire of antibody fragments, obtained by PCR ampli¢cation, to a coat protein of ¢lamentous phage. The Phab repertoire is incubated with the immobilized antigen. Non-binding phages are removed by extensive washing and phages with antigen binding activity are eluted and ampli¢ed by infection of E. coli. Typically 2^5 rounds of selection are required for selection of Phabs with desired speci¢city.

are homogeneous, fully active and can be tetramerized by incubation with streptavidin. Other C-terminal peptide tags such as hexahistidine tags allowing metal-chromatography-based puri¢cation or peptide tags facilitating detection are widely used and are found in most vectors used for antibody fragment production. 3. The phage display system In 1990, MacCa¡erty and colleagues demonstrated in a model experiment that it was possible to use ¢lamentous bacteriophage fd as a tool to select an antibody fragment from a large population of non-binding proteins [19]. This work, based on earlier fundamental ¢ndings concerning peptide phage display by Smith's group [20], was the basis of the development of the technique that has revolutionized the ¢eld of antibody engineering. The principle consists of expressing a protein, in this case an antibody fragment, on the surface of a phage particle by cloning the gene encoding the antibody fragment into the phage genome, as a fusion with a coat protein. The resulting par-

FEMSLE 9478 14-7-00

ticle can be a¤nity-puri¢ed on immobilized antigen. After washings and elution, the selected phages are ampli¢ed by infection of E. coli. This cycle can be repeated several times (Fig. 3). On average, a binder is enriched 103 ^105 fold over non-binding clones. This technique can be used to select antibodies from large repertoires [21]. Two types of repertoires have been built. Immunized repertoires where the V genes used to assemble the antibody fragments are taken from a mouse immunized by antigen, and na|«ve `single pot repertoires' where V genes are taken from peripheral blood lymphocytes of non-immunized individuals (for review, [22]). The main advantage of the ¢rst method is that antibodies selected from immune repertoires are expected to have high a¤nities on the basis that immunization will have induced a large bias toward B-cells speci¢c for the immunogen and having undergone a¤nity maturation. Obviously the main drawback of such repertoires is that a new repertoire is required for each new antigen. In this respect, na|«ve libraries are very convenient since once they are constructed they can then be used almost inde¢nitely against a diverse range of antigens. Most importantly, these libraries can be built using human B-cells, thereby allowing the direct isolation of

Cyaan Magenta Geel Zwart

P. Chames, D. Baty / FEMS Microbiology Letters 189 (2000) 1^8

human antibodies. A major drawback of na|«ve repertoires is that they need to be very large to yield reasonable af¢nities. Typically a na|«ve library of 1010 clones is expected to yield a¤nities in the 1038 M range (for review, [22]). If higher a¤nities are desired, for example in vivo applications, a¤nity maturation techniques can be used (for review, [22]). Alternatively, `synthetic' libraries have been made using PCR techniques and degenerate primers to introduce diversity at precise locations (typically CDR3s) (reviewed in Hoogenboom and Chames, in press). The possibility to control the selection process (by depletion, competition) and to obtain fully human high-af¢nity antibodies against any antigen including toxic, conserved or non-immunogenic targets, possibly in an automated fashion, should ultimately set phage display technology to replace hybridoma technology. Antibody engineering also provides the possibility to tailor these binding fragments for diagnostic or targeting purposes as well as for use as `intrabodies', a rapidly growing ¢eld of application of antibody fragments. 4. The intrabodies 4.1. General concept and mode of action Intracellular antibodies or `intrabodies' represent a new and very promising application of recombinant antibodies. This concept is based on expression of an antibody fragment inside a cell in order to bind and inactivate the protein targeted by this fragment, either by blocking interactions with other proteins, or by diverting the protein from its normal location, thereby preventing its action (reviewed in [23,24]). The utilization of well characterized signal sequences allows scFv fragments to be directed to subcellular compartments of mammalian cells (such as endoplasmic reticulum (ER), nucleus, cytoplasm, mitochondria, secretion pathway), where they can neutralize their target. A possible limitation of this approach is the folding of the fragments. Antibodies are normally folded in the reducing environment of the ER where they interact with speci¢c chaperones. However many interesting targets are cytoplasmic proteins. In 1990, Cattaneo's group showed that some antibodies retain their activity when expressed in the cytoplasm of mammalian cells [25] despite the absence of disul¢de bond formation. Nonetheless, many antibody fragments do not function under such conditions. Several solutions to this problem have been proposed, including engineering antibodies to obtain stable disul¢de-free forms [26], or the use of a stable disul¢de-free non-antibody sca¡old [27]. Alternatively, Visintin and colleagues used the two-hybrid system for direct in vivo selection of antibodies capable of binding antigen under cytoplasmic conditions [28]. However, an interesting recent study demonstrated that good cytoplasmic folding was not necessarily a prerequi-

FEMSLE 9478 14-7-00

5

site for e¤cient blocking activity. Indeed, Lener and colleagues showed that a panel of phage-derived single-chain anti-p21ras antibodies were all able to inactivate Ras in vivo, although they were not interfering with the intrinsic GTPase activity and the Raf binding activity of Ras in vitro [29]. Interestingly, these antibodies were aggregating in mammalian cells, suggesting that they could exist in an intermediate folding state that retained speci¢c binding activity allowing sequestration of antigen by `intracellular coaggregation'. This result is of signi¢cance as it demonstrates that non-blocking binders may be used as intrabodies to divert protein tra¤cking. 4.2. Application: phenotypic knock-out The ability to functionally and speci¢cally block an endogenous protein by expressing an antibody against it has two obvious applications. These are the knock-out of a protein to investigate its role in a pathway and the neutralization of oncogenic molecules. The latter approach has been demonstrated with several proteins, including ErB-2, p21ras and p53. Applicability of this concept was ¢rst demonstrated in 1993 with the inhibition of the insulin-induced meiotic maturation of Xenopus laevis oocytes by cytoplasmic expression of a murine anti-p21 scFv [30]. More recently in vivo studies performed using an anti-p21 scFv cDNA inserted into an adenoviral vector have shown that the scFv dramatically a¡ects tumor growth [31]. Similarly, cytoplasmic expression of an anti-p53 scFv led to formation of p53^scFv complexes and restored the p53 mutant de¢cient transcriptional activity [32]. ErB-2 has also proven to be an interesting target for intrabodies. Curiel and colleagues have demonstrated that intracellular antibody against this protein mediates targeted tumor cell eradication by apoptosis [33]. Similarly, Hynes' group reverted ErB-2 transformation by expressing a speci¢c single-chain antibody in the ER to prevent transit of the ErB2 receptor [34]. These ¢ndings demonstrate that selective oncogene `knock-out' using intracellular antibodies represents a viable novel strategy for anti-cancer gene therapy that potentially o¡ers highly speci¢c, targeted eradication of human tumor cells. The second major application of phenotypic knock-out is to inactivate a protein to understand its biological role. This approach has mainly been used for the prevention of cell surface expression of speci¢c receptors (reviewed in [23]). Although in theory an intrabody can interfere at multiple points within the secretory pathway, the ER is the most strategic location for targeting intrabody expression. Antibody fragments are well produced, show high stability and a long half-life in the ER due to the presence of molecular chaperones enabling e¤cient folding and disul¢de bond formation. Also the signal peptide (KDEL) necessary for the retention of the antibody fragment in the ER is well characterized. Finally, the tubular architecture of this organelle maximizes the chance of interaction be-

Cyaan Magenta Geel Zwart

6

P. Chames, D. Baty / FEMS Microbiology Letters 189 (2000) 1^8

tween a resident antibody and its target. A clear example of this approach was given by Richardson and colleagues who demonstrated complete abrogation of cell surface expression of IL-2RK in stimulated Jurkat cells by intracellular expression of an anti-IL-2RK scFv with a C-terminal ER retention signal, thereby creating a powerful tool for the study of IL-2-mediated signal transduction. Similarly, the activity of epidermal growth factor receptor (EGFR) has been inhibited by directing an anti-EGFR scFv to the secretory pathway [35]. Vanhove and colleagues also used an intrabody to inhibit the activity of K-1,3-galactosyltransferase in pig cells and thus generation of epitopes recognized by human anti-Gal antibodies that take part in later stages of xenograft rejection. This led to more than a 90% reduction in cytotoxicity involving anti-Gal antibodies and complement [36]. Recently, we have extended the application of intrabodies to the ¢eld of fundamental microbiology, by studying the mode of action of bacterial toxin colicin A (Fig. 2) [37]. E. coli was protected from the lethal action of colicin A toxin by periplasmic expression of a scFv directed against the N-terminal region of the colicin A molecule. This allowed us to establish that the N-terminal domain of colicin A is accessible in the periplasm when its central domain is bound at the cell surface and that the C-terminal domain forms the lethal pore in the inner membrane. 4.3. Virus resistance by intracellular immunization One very interesting application of intrabodies is to confer new resistance to cells. This has been demonstrated in plant cells by constitutive expression in transgenic plants of a scFv antibody, directed against the plant icosahedral tombusvirus artichoke mottled crinkle virus, which results in reduction of infection incidence and delay in symptom development [38]. More importantly, intracellular antibodies have been shown to have a potential use in human gene therapy. Several intrabodies directed against structural, regulatory or enzymatic proteins of HIV-1 have been studied (see [39] for review). Encouraging results have been obtained with an antibody directed against the envelope glycoprotein gp120 and its precursor gp160. Interestingly, production of this antibody as a Fab fragment in the ER blocked processing and incorporation of gp120 into virion particles. In addition, the secreted Fab neutralized cell-free HIV-1 virions, thus exhibiting a dual e¡ect [40]. In a clinical trial, transduction of human lymphocytes with this Fab intrabody gene using adeno-associated virus vector resulted in blocking of infection by several primary isolates [40]. Tat is an essential regulatory protein that activates transcription of viral genes, as well as cellular genes that in turn can result in HIV-1 gene transcription. This protein is thus an interesting target. Amongst several alternative formats studied, it appeared that the cytoplasmic expression of an anti-Tat scFv fused to a constant kappa

FEMSLE 9478 14-7-00

Ig domain (for multimerization) resulted in the strongest resistance to HIV-1 infection in stably transfected lymphocytes [41], seemingly by sequestration of Tat in the cytoplasm. Several studies have reported targeting of the regulatory protein Rev. Cytoplasmic anti-Rev intrabodies were able to inhibit transport of Rev to the nucleus, thereby preventing viral mRNA processing and HIV-1 production by the cell. Human lymphocytes transduced with retroviral vectors expressing these intrabodies were resistant to HIV-1 infection [42]. Other HIV-1 proteins that have been chosen as intrabody targets include HIV-1 reverse transcriptase, integrase as well as matrix and nucleocapsid proteins (for review, [39]). All these studies resulted in signi¢cant inhibition of HIV-1 replication, demonstrating the versatility of intrabodies in their ability to inhibit different stages of the viral life cycle. More recently, Steinberger and colleagues proposed an alternative approach for intrabody-based HIV therapy. They described an intrabody capable of blocking the expression of the chemokine receptor CCR5, an essential coreceptor required for entry of HIV-1 into macrophages, the ¢rst cells targeted by HIV-1 during infection. Transduced cells were highly refractory to challenge with virus or infected cells [43]. This approach aiming at preventing viral entry may have advantages over strategies targeting later stages of the HIV-1 life cycle. Clearly intrabodies have great potential for anti-HIV therapy, possibly providing a treatment strategy for the immune reconstitution of infected patients with advanced disease, experiencing recurrence of plasma viremia after an initial response to highly active anti-retroviral therapies [44]. An alternative future application for intrabodies, predicted to be of great signi¢cance, is in the ¢eld of functional genomics. Combined with phage display to rapidly select binders, the phenotypic knock-out may in the near future become an important tool to ascertain the function of the plethora of new proteins that will require characterization. 5. Conclusion and perspective Since the development of recombinant antibody technology combined with the power of selection by phage display, exciting results have been obtained in an impressive variety of ¢elds. The selection of speci¢c high-a¤nity human antibody fragments has never appeared so feasible. From these fragments, tailor-made reagents can be constructed by choosing valency, (multi)speci¢cities as well as e¡ector functions. These developments have meant that several antibodies have already been accepted as therapeutic products and hundreds of antibody-based products (representing over 30% of all biological proteins) are currently undergoing clinical trials for diagnosis and therapy (for review, see [45]). The concept of the magic bullet is beginning to be practically realized. We anticipate that in

Cyaan Magenta Geel Zwart

P. Chames, D. Baty / FEMS Microbiology Letters 189 (2000) 1^8

the future, with the increase in the number and quality of phage antibody libraries, and development of automation, desired speci¢cities will be isolated in matter of days rather than weeks or months. One may also consider functionbased selections as opposed to selections based on binding alone, such as phage antibodies selected for internalization, receptor triggering or catalytic activity. Finally, the ¢eld of functional genomics and the human genome project are producing copious amounts of sequences that need to be characterized. Automated selection of phage antibodies against sequence-predicted peptides followed by expression studies, as well as intracellular antibody approaches should provide valuable information eventually leading to a panel of pharmaceutically useful reagents for use as targeting and/or neutralizing molecules. Acknowledgements The authors would like to thank Rajinder Bhoday and Stephen Slatin for critical reading of the manuscript. Work in the authors' laboratory was supported by the Centre National de la Recherche Scienti¢que and the Association de Recherche contre le Cancer.

References [1] Erlich, P. (1957) In: Immunology and Cancer Research, Vol. 2 (Himmelweit, F., Ed.), Permagon Press, Oxford. [2] Kohler, G. and Milstein, C. (1975) Continuous cultures of fused cells secreting antibody of prede¢ned speci¢city. Nature 256, 495^497. [3] Skerra, A. and Pluckthun, A. (1988) Assembly of a functional immunoglobulin Fv fragment in Escherichia coli. Science 240, 1038^1041. [4] Better, M., Chang, C.P., Robinson, R.R. and Horwitz, A.H. (1988) Escherichia coli secretion of an active chimeric antibody fragment. Science 240, 1041^1043. [5] Bird, R.E., Hardman, K.D., Jacobson, J.W., Johnson, S., Kaufman, B.M., Lee, S.M., Lee, T., Pope, S.H., Riordan, G.S. and Whitlow, M. (1988) Single-chain antigen-binding proteins [published erratum appears in Science, 1989, April 28; 244(4903):409]. Science 242, 423^ 426. [6] Reiter, Y., Brinkmann, U., Webber, K.O., Jung, S.H., Lee, B. and Pastan, I. (1994) Engineering interchain disul¢de bonds into conserved framework regions of Fv fragments: improved biochemical characteristics of recombinant immunotoxins containing disul¢de-stabilized Fv. Protein Eng. 7, 697^704. [7] Hu, S., Shively, L., Raubitschek, A., Sherman, M., Williams, L.E., Wong, J.Y., Shively, J.E. and Wu, A.M. (1996) Minibody : A novel engineered anti-carcinoembryonic antigen antibody fragment (singlechain Fv-CH3) which exhibits rapid, high-level targeting of xenografts. Cancer Res. 56, 3055^3061. [8] Hurle, M.R. and Gross, M. (1994) Protein engineering techniques for antibody humanization. Curr. Opin. Biotechnol. 5, 428^433. [9] Le Gall, F., Kipriyanov, S.M., Moldenhauer, G. and Little, M. (1999) Di-, tri- and tetrameric single chain Fv antibody fragments against human CD19: e¡ect of valency on cell binding. FEBS Lett. 453, 164^168. [10] Holliger, P., Prospero, T. and Winter, G. (1993) `Diabodies' : small bivalent and bispeci¢c antibody fragments. Proc. Natl. Acad. Sci. USA 90, 6444^6448.

FEMSLE 9478 14-7-00

7

[11] Kostelny, S.A., Cole, M.S. and Tso, J.Y. (1992) Formation of a bispeci¢c antibody by the use of leucine zippers. J. Immunol. 148, 1547^1553. [12] De Jonge, J., Brissinck, J., Heirman, C., Demanet, C., Leo, O., Moser, M. and Thielemans, K. (1995) Production and characterization of bispeci¢c single-chain antibody fragments. Mol. Immunol. 32, 1405^ 1412. [13] Cao, Y. and Suresh, M.R. (1998) Bispeci¢c antibodies as novel bioconjugates. Bioconjug. Chem. 9, 635^644. [14] Kreitman, R.J. (1999) Immunotoxins in cancer therapy. Curr. Opin. Immunol. 11, 570^578. [15] Lode, H.N., Xiang, R., Becker, J.C., Gillies, S.D. and Reisfeld, R.A. (1998) Immunocytokines: a promising approach to cancer immunotherapy. Pharmacol. Ther. 80, 277^292. [16] Muller, B.H., Chevrier, D., Boulain, J.C. and Guesdon, J.L. (1999) Recombinant single-chain Fv antibody fragment-alkaline phosphatase conjugate for one-step immunodetection in molecular hybridization. J. Immunol. Methods 227, 177^185. [17] Saviranta, P., Haavisto, T., Rappu, P., Karp, M. and Lovgren, T. (1998) In vitro enzymatic biotinylation of recombinant fab fragments through a peptide acceptor tail. Bioconjug. Chem. 9, 725^735. [18] Sibler, A.P., Kempf, E., Glacet, A., Orfanoudakis, G., Bourel, D. and Weiss, E. (1999) In vivo biotinylated recombinant antibodies : high e¤ciency of labelling and application to the cloning of active anti-human IgG1 Fab fragments. J. Immunol. Methods 224, 129^ 140. [19] McCa¡erty, J., Gri¤ths, A.D., Winter, G. and Chiswell, D.J. (1990) Phage antibodies : ¢lamentous phage displaying antibody variable domains. Nature 348, 552^554. [20] Scott, J.K. and Smith, G.P. (1990) Searching for peptide ligands with an epitope library. Science 249, 386^390. [21] Barbas, C., Kang, A.S., Lerner, R.A. and Benkovic, S.J. (1991) Assembly of combinatorial antibody libraries on phage surfaces : the gene III site. Proc. Natl. Acad. Sci. USA 88, 7978^7982. [22] Hoogenboom, H.R., de ruine, A.P., Hufton, S.E., Hoet, R.M., Arends, J.W. and Roovers, R.C. (1998) Antibody phage display technology and its applications. Immunotechnology 4, 1^20. [23] Biocca, S., Pierandrei-Amaldi, P., Campioni, N. and Cattaneo, A. (1994) Intracellular immunization with cytosolic recombinant antibodies. Biotechnology 12, 396^399. [24] Richardson, J.H. and Marasco, W.A. (1995) Intracellular antibodies: development and therapeutic potential. Trends Biotechnol. 13, 306^ 310. [25] Biocca, S., Neuberger, M.S. and Cattaneo, A. (1990) Expression and targeting of intracellular antibodies in mammalian cells. EMBO J. 9, 101^108. [26] Proba, K., Worn, A., Honegger, A. and Pluckthun, A. (1998) Antibody scFv fragments without disul¢de bonds made by molecular evolution. J. Mol. Biol. 275, 245^253. [27] Beste, G., Schmidt, F.S., Stibora, T. and Skerra, A. (1999) Small antibody-like proteins with prescribed ligand speci¢cities derived from the lipocalin fold. Proc. Natl. Acad. Sci. USA 96, 1898^1903. [28] Visintin, M., Tse, E., Axelson, H., Rabbitts, T.H. and Cattaneo, A. (1999) Selection of antibodies for intracellular function using a two-hybrid in vivo system. Proc. Natl. Acad. Sci. USA 96, 11723^ 11728. [29] Lener, M., Horn, I.R., Cardinale, A., Messina, S., Nielsen, U.B., Rybak, S.M., Hoogenboom, H.R., Cattaneo, A. and Biocca, S. (2000) Diverting a protein from its cellular location by intracellular antibodies. The case of p21Ras. Eur. J. Biochem. 267, 1196^1205. [30] Biocca, S., Pierandrei-Amaldi, P. and Cattaneo, A. (1993) Intracellular expression of anti-p21ras single chain Fv fragments inhibits meiotic maturation of Xenopus oocytes. Biochem. Biophys. Res. Commun. 197, 422^427. [31] Cochet, O., Kenigsberg, M., Delumeau, I., Virone-Oddos, A., Multon, M.C., Fridman, W.H., Schweigho¡er, F., Teillaud, J.L. and Tocque, B. (1998) Intracellular expression of an antibody fragment-

Cyaan Magenta Geel Zwart

8

[32]

[33]

[34]

[35]

[36]

[37]

[38]

P. Chames, D. Baty / FEMS Microbiology Letters 189 (2000) 1^8 neutralizing p21 ras promotes tumor regression. Cancer Res. 58, 1170^1176. Caron de Fromentel, C., Gruel, N., Venot, C., Debussche, L., Conseiller, E., Dureuil, C., Teillaud, J.L., Tocque, B. and Bracco, L. (1999) Restoration of transcriptional activity of p53 mutants in human tumour cells by intracellular expression of anti-p53 single chain Fv fragments. Oncogene 18, 551^557. Deshane, J., Grim, J., Loechel, S., Siegal, G.P., Alvarez, R.D. and Curiel, D.T. (1996) Intracellular antibody against erbB-2 mediates targeted tumor cell eradication by apoptosis. Cancer Gene Ther. 3, 89^98. Beerli, R.R., Wels, W. and Hynes, N.E. (1994) Intracellular expression of single chain antibodies reverts ErbB-2 transformation. J. Biol. Chem. 269, 23931^23936. Richardson, J.H., Hofmann, W., Sodroski, J.G. and Marasco, W.A. (1998) Intrabody-mediated knockout of the high-a¤nity IL-2 receptor in primary human T cells using a bicistronic lentivirus vector. Gene Ther. 5, 635^644. Vanhove, B., Charreau, B., Cassard, A., Pourcel, C. and Soulillou, J.P. (1998) Intracellular expression in pig cells of anti-K-1,3-galactosyltransferase single-chain FV antibodies reduces Gal K1, 3Gal expression and inhibits cytotoxicity mediated by anti-Gal xenoantibodies. Transplantation 66, 1477^1485. Chames, P., Fieschi, J., Baty, D. and Duche, D. (1998) Intracellular immunization of prokaryotic cells against a bacteriotoxin. J. Bacteriol. 180, 514^518. Tavladoraki, P., Benvenuto, E., Trinca, S., De Martinis, D., Cattaneo, A. and Gale¤, P. (1993) Transgenic plants expressing a func-

FEMSLE 9478 14-7-00

[39]

[40]

[41]

[42]

[43]

[44]

[45]

tional single-chain Fv antibody are speci¢cally protected from virus attack. Nature 366, 469^472. Rondon, I.J. and Marasco, W.A. (1997) Intracellular antibodies (intrabodies) for gene therapy of infectious diseases. Annu. Rev. Microbiol. 51, 257^283. Chen, J.D., Yang, Q., Yang, A.G., Marasco, W.A. and Chen, S.Y. (1996) Intra- and extracellular immunization against HIV-1 infection with lymphocytes transduced with an AAV vector expressing a human anti-gp120 antibody. Hum. Gene Ther. 7, 1515^1525. Mhashilkar, A.M., Bagley, J., Chen, S.Y., Szilvay, A.M., Helland, D.G. and Marasco, W.A. (1995) Inhibition of HIV-1 Tat-mediated LTR transactivation and HIV-1 infection by anti-Tat single chain intrabodies. EMBO J. 14, 1542^1551. Duan, L., Zhu, M., Bagasra, O. and Pomerantz, R.J. (1995) Intracellular immunization against HIV-1 infection of human T lymphocytes: utility of anti-rev single-chain variable fragments [published erratum appears in Hum. Gene Ther., 1998, March 20; 9765]. Hum. Gene Ther. 6, 1561^1573. Steinberger, P., Andris-Widhopf, J., Buhler, B., Torbett, B.E. and Barbas III, C.F. (2000) Functional deletion of the CCR5 receptor by intracellular immunization produces cells that are refractory to CCR5-dependent HIV-1 infection and cell fusion. Proc. Natl. Acad. Sci. USA 97, 805^810. Collier, A.C., Coombs, R.W., Schoenfeld, D.A., Bassett, R., Baruch, A. and Corey, L. (1996) Combination therapy with zidovudine, didanosine and saquinavir. Antiviral Res. 29, 99. Hudson, P.J. (1999) Recombinant antibody constructs in cancer therapy. Curr. Opin. Immunol. 11, 548^557.

Cyaan Magenta Geel Zwart