Modulation of the CD40–CD40 ligand interaction using human anti-CD40 single-chain antibody fragments obtained from the n-CoDeR phage display library

Abstract

CD40 plays a central regulatory role in the immune system and antibodies able to modulate CD40 signalling may consequently have a potential in immunotherapy, in particular for treatment of lymphomas and autoimmune disease like multiple sclerosis. As a first step to achieve this goal, we describe the selection and characterization of a novel set of fully human anti-CD40 antibody fragments (scFv) from a phage display library (n-CoDeR). In order to determine their biological potential, these antibody fragments have been analysed for their ability to promote B-cell activation, rescue from apoptosis and to block the CD40–CD40 ligand (CD40L) interaction. The selected cohort of human scFv could be subcategorized, each expressing a distinct functional signature. Thus scFv were generated that induced B-cell proliferation, rescued B cells from apoptosis and blocked the CD40–CD40L interaction to different extents. In particular, one of the scFv clones (F33) had the ability to abrogate completely this interaction. The epitope recognition patterns as well as individual rate constants were also determined and the affinity was shown to vary from low to high nanomolar range. In conclusion, this panel of human anti-CD40 scFv fragments displays a number of distinct properties, which may constitute a valuable source when evaluating candidates for in vivo trials.

Introduction

CD40 is a 45 000–50 000 MW glycoprotein that belongs to the tumour necrosis factor receptor (TNFR) superfamily. It is expressed on a variety of cells in the immune system, such as B cells, dendritic cells and monocytes. The four extracellular domains of the CD40 molecule consist of several cysteine-rich repeats and each domain is further subdivided into an A- and a B-module.¹ No X-ray structure of CD40 has been reported, but several models have been proposed, using the known X-ray structure of TNFR as a template.^2–4 The role of the CD40 molecule in B-cell development has been extensively studied and has been shown to be of importance for proliferation, differentiation, immunoglobulin production, isotype switching and maturation into memory B cells. CD40 is expressed on B cells during all stages of B-cell differentiation. Ligation of CD40 on antigen-presenting cells (APCs) is of central importance in the immune response, especially for T-cell-dependent B-cell activation. The CD40 ligand (CD40L) is primarily expressed on activated mature T cells.^5–7

The role of CD40 and CD40L in tumour cell proliferation, differentiation and APC function has recently been underlined,⁸ when it was suggested that anti-CD40 antibodies could potentially be used for treatment of lymphomas. Furthermore, anti-CD40 antibodies have also been proposed for treatment of chronic inflammatory clinical conditions.^9,10 It has also been shown that the CD40–CD40L interaction is critical for both the initiation and the progression of experimental autoimmune encephalomyelitis (EAE), a model proposed for multiple sclerosis. Treatment with an anti-CD40L antibody effectively inhibited EAE¹¹ in mice, and it has also been shown that treatment of marmoset monkeys with a monoclonal antibody (mAb) against CD40 (5D12) postponed the onset of EAE.^10,12 Moreover, anti-CD40 antibodies have been shown to have a therapeutic activity in chronic collagen-induced arthritis (CCIA) in mice.⁹

Today only anti-CD40 antibodies of non-human origin are available and the clinical efficacy of these antibodies is limited due to the human anti-mouse response found in most patients.^13,14 In this study, we have selected and characterized a number of anti-CD40 antibody fragments from a fully human phage display library, called n-CoDeR.¹⁵ The kinetic properties, as well as the location of the CD40 epitope recognized by each antibody fragment, were determined. These antibodies were also functionally characterized, in that their ability to stimulate B-cell proliferation, prevent apoptosis and to block the CD40–CD40L interaction was investigated.

Materials and methods

Reagents

The n-CoDeR library was kindly provided by BioInvent Therapeutic AB (Lund, Sweden)¹⁵ and human CD40-Fcγ was kindly provided by Tanox Pharma (Amsterdam, The Netherlands).^16,17 An antibody against the AD2-eptitope of cytomegalovirus, ITC88,¹⁸ was a generous gift from Dr Mats Ohlin (Lund University). M2 mouse anti-FLAG antibody was purchased from Sigma-Aldrich (St Louis, MO). Phycoerythrin (PE)-conjugated rabbit anti-mouse antibody and streptavidin, as well as fluorescein isothiocyanate (FITC)-conjugated rabbit F(ab′)₂ anti-human immunoglobulin G (IgG) were obtained from DAKO A/S (Glostrup, Denmark). Recombinant interleukin-4 (IL-4) was purchased from R & D (Abingdon, UK). Goat anti-human IgM was obtained from Jackson ImmunoResearch (West Grove, PA).

The cell lines used were the human B-cell lines, BJAB,¹⁹ and Ramos (ATCC, CRL-1596) and two mouse fibroblast L cell lines expressing CD32 or CD40L respectively, the latter kindly provided by John Pound (Birmingham, UK).

Selection of anti-CD40 antibodies

Selections using biotinylated CD40-Fcγ were performed as described by Söderlind et al.¹⁵ CD40-Fcγ-biotin concentrations between 10⁻⁷ m and 10⁻⁹ m were used. Solid surface selections were performed on Immunotubes coated with 10⁻⁷ m CD40-Fcγ in phosphate-buffered saline (PBS) over night at 4°. The Immunotubes were blocked for 1 hr at room temperature with a buffer containing 5% bovine serum albumin (BSA) and 0·05% Tween-20 in PBS, washed four times and incubated with phage stocks for 1 hr at room temperature. After washing, bound phages were eluted with trypsin (0·5 mg/ml). Cell selections were performed on COS-7 cells transiently transfected with CD40 using lipofectamin, according to the manufacturer's protocol (BRL Life Technology, Täby, Sweden). The cells were incubated with the phages on ice for 1 hr and then washed and the selected phages were eluted with a glycine–HCl buffer, pH 2·2.

Expression and purification of single-chain antibody fragment

The filamentous phage gene3 fragment, which is fused to the single-chain antibody fragment (scFv) in the library vector in order to enable display on the phage surface, was removed and the expression vector was transformed into SF110 cells.²⁰ As positive control we used variable domains from G28-5,²¹ cloned into a scFv expression vector (pFab5c) and produced in SF110. Cultures were grown in TB-medium at 37° to OD₆₀₀=0·5 and then induced with 0·1 mm isopropyl-thio-β-D-galactoside. Cells were grown at 30° and removed 16 hr after induction by centrifugation (15 000 g for 30 min at 4°). Supernatants were concentrated, using an Ultrasette (10 000 MW cut-off Filtron, Northborough, MA). Antibody fragments were purified from the supernatants by batch purification on Ni-NTA agarose (Qiagen GmbH, Hilden, Germany). Endotoxins were removed on prepacked 1 ml Detoxi-Gel Columns (Pierce, Rockford, IL). Higher molecular weight contaminants were removed by Sephadex-75 gel filtration (Pharmacia Biotech, Uppsala, Sweden). The protein concentrations were determined by ultraviolet-spectrophotometry.

Screening of monoclonal phage stocks

High-binding 96-well assay plates (Corning Inc., Corning, NY) were coated with 1 µg/ml antigen in PBS, over night at 4°. The plates were washed five times with PBST (PBS with 0·1% Tween-20) and 25 µl mPBS (4% milk powder in PBS) were added to each well, followed by 75 µl phagestock and the plates were incubated for 2 hr at room temperature. After washing five times with PBST, the plates were incubated with horseradish peroxidase-anti-M13 (Pharmacia Biotech, Uppsala, Sweden) diluted 1 : 2000 in 1% mPBS for 1 hr. The plates were washed and the enzymatic activity was determined, using the chromophore o-phenylenediamine. The optical density was measured at 490 nm.

Domain mapping

The CD40 leader sequence was amplified together with the AD2 epitope sequence²² by overlap extension polymerase chain reaction (PCR) of oligonucleotides (BRL Life Technology) followed by reamplification with flanking primers. The resulting fragment was digested with NheI and NotI and subcloned into the corresponding sites in pcDNA3.1(+) (Invitrogen, Groningen, Germany).

The CD40 domains were PCR amplified, using the primers in Table 1. After amplification the fragments were digested with NotI and XbaI and inserted 3′ to the AD2 epitope sequence in the vector described above. All the CD40 constructs (Fig. 1) were also produced without the AD2 epitope. The CD40 constructs were transfected into COS-7 cells and transiently expressed, using lipofectamin according to the manufacturer's protocol (BRL Life Technology). After 3 days the cells were harvested and incubated with anti-CD40 scFv antibodies or an anti-AD2 antibody (ITC88) on ice for 30 min. The cells were washed and incubated with a secondary antibody, rabbit anti-human F(ab′)₂-FITC or mouse anti-FLAG (M2). The anti-FLAG antibody was detected with rabbit-anti-mouse-PE. The samples were analysed on a FACScan (Becton Dickinson, Mountain View, CA).

Table 1. Primers used for PCR-based amplification of CD40 leader sequence, AD2 epitope and CD40 domains.

Leader sequence primer:

ACTGCTTGCGGCCGCTCCGTACTTGAGGGTAGTGTTG- ″TAGATAGTCTCGTTGGCTGGATGGACAGCGGTCAG- ″CAA

AD2 epitope:

GTGCTAGCTAGCATGGTTCGTCTGCCTCTGCAGTGCGT- ″CCTCTGGGGCTGCTTGCTGACCGCTGTCCATCCAGCC

5′ reamplification Leader sequence:

GTGCTAGCTAGCATGGTT

Reamplification 3′ AD2 epitope:

ACTGCTTGCGGCCGCTCC

5′ D1 CD40 (NotI):

ACTGCTTGCGGCCGCGAACCACCCACTGCATGCAG

5′ D1/B2 CD40 (NotI):

ACTGCTTGCGGCCGCCTGTTCTTTGTGCCAGCCA

5′ D2 CD40 (NotI):

ACTGCTTGCGGCCGCTTGCGGTGAAAGCGAATTCCTA

5′ D2/B1 CD40(NotI):

ACTGCTTGCGGCCGCCCACCAGCACAAATACTGC

5′ D3 CD40 (NotI):

ACTGCTTGCGGCCGCCTGTGAAGAAGGCTGGCACTGT

5′ D4 CD40 (NotI):

ACTGCTTGCGGCCGCCTGCCCAGTCGGCTTCTTCTCC

3′ CD40 (XbaI):

GAATTCTGATCTAGATTATCACTGTCTCTCCTGCACTG-″AGAT

Figure 1.

(a) Models of the extracellular part of the different CD40 constructs. Each domain (D) can be subdivided into one A (A1 or A2) and one B (B1 or B2) module. The constructs are named after the outermost domain or module they contain. To detect surface expression a short (13 amino acids) tag, AD2, was used. The position of the AD2 epitope on the CD40 constructs is indicated by a filled circle. All constructs, except D4 have also been expressed without the AD2 epitope. (b) The localization of each antibody's epitope on CD40 was analysed using a domain mapping approach. The ability of the anti-CD40 scFv antibodies to bind to the different CD40 constructs on COS-7, as measured by FACScan, is indicated in the Table. *These clones only bind this construct in the absence of the AD2 epitope. §ITC88 is an anti-AD2 antibody used to detect expression of the CD40 constructs that contained the AD2 epitope.

Isolation of B cells

Human peripheral blood lymphocytes were obtained from Lund University Hospital (Lund, Sweden). B cells were positively selected on anti-CD19-coated magnetic beads (Dynal A/S, Oslo, Norway). The beads were removed from the cells with Detachabeads (Dynal A/S). The selected B cells were routinely >98% pure, as determined on FACScan (Becton Dickinson).

B-cell proliferation assay

The ability of the selected anti-CD40 scFv fragments to stimulate B-cell proliferation was tested in a system similar to that described by Banchereau et al.²³ Briefly, 10⁵ B cells were cultured in 200 µl R1-medium (RPMI-1640, containing 1% fetal calf serum, 2 mm l-glutamine, 1% non-essential amino acids and 50 µg/ml gentamycin) (BRL Life Technologies), supplemented with 100 U/ml recombinant human IL-4. In order to cross-link the scFv anti-CD40 antibodies, M2 anti-FLAG antibody was used (the anti-FLAG antibody to scFv molar ratio was 4 : 1). To further cross-link the anti-FLAG/scFv complex 2×10⁴ irradiated L cells expressing CD32 were added to the culture. Anti-CD40 scFv were added at concentrations varying from 1 to 10 µg/ml. B-cell proliferation was measured after 3 days by a 16-hr [methyl-³H]thymidine pulse (0·5 µCi/well, Pharmacia Biotech).

Rescue from apoptosis

Goat anti-human IgM (10 µg/ml) was used to induce apoptosis in the Ramos cell line. Anti-CD40 scFv (1 µg/ml) was added to the cells, with or without the cross-linking M2 anti-FLAG antibody (4 µg/ml). After a 24-hr incubation at 37° the cells were treated with annexinV-FITC (BD Biosciences) and propidium iodide and the level of viable cells was analysed using flow cytometry. Figure 3 shows the percentage of viable cells after treatment with scFv compared to cells treated with PBS (used as control). Solid bars indicate that the scFv have been cross-linked with anti-FLAG, while the hatched bars show viability without cross-linking.

Figure 3.

Ramos cells were treated with anti-IgM antibodies (10 µg/ml) and anti-CD40 scFv (1 µg/ml), or PBS (as control), with or without the M2 anti-FLAG cross-linking antibody. After 24 hr, the cells were treated with annexin-FITC and propidium iodide for 15 min at room temperature and the level of viable cells was determined by flow cytometry. The figure shows the ratio of rescued cells treated with anti-IgM and scFv antibodies, compared to cells treated only with anti-IgM. Solid bars indicate that the scFv have been cross-linked with anti-FLAG, while hatched bars show viability without cross-linking. Data represent mean values ± SD of three independent experiments.

Blocking the CD40–CD40L interaction

Biotinylated CD40-Fcγ (4 µg/ml) was incubated with anti-CD40 scFv antibodies (100 µg/ml) in 100 µl PBS (BRL, Life Technology) for 1 hr on ice. Then, 4×10⁵ CD40L-transfected L cells in 100 µl PBS, supplemented with 1% BSA, were added and the mixture was incubated on ice for 30 min. PE-conjugated streptavidin was used to detect the amount of biotinylated CD40-Fcγ bound to the CD40L-expressing L cells, using FACScan analysis.

Determination of kinetic parameters and affinity constants

CD40-Fcγ molecules were immobilized on a BIAcore sensorchip (CM5) at 1200 and 1400 RU, using conventional amine coupling. Monomeric scFv antibodies were analysed for binding at different concentrations between 100 and 1000 nm in HEPES complete buffer (10 mm HEPES, 3·4 mm ethylenediaminetetraacetic acid, 0·15 m NaCl, 0·05% BIAcore surfactant P20, pH 7·4) at a flow rate of 3 µl/min. The association was followed for 11 min and the dissociation for 10 min. Regeneration was performed using 100 mm HCl for 3 min. The kinetic parameters and the affinity constants were calculated, using BIAEvaluation software version 3.0.

Results

Selection of CD40-specific scFv fragments

A panel of human anti-CD40 scFv were selected from the n-CoDeR phage display library. Four rounds of selections were performed using biotinylated recombinant CD40-Fcγ fusion protein as bait, followed by subsequent capture of bound phages on streptavidin-coated magnetic beads. In the fourth round, additional selection strategies were applied using CD40-coated immunotubes as well as CD40-expressing cells. This latter approach avoided enrichment of streptavidin or biotin-specific binders and ensured that we obtained binders recognizing naturally CD40 expressed on a cell surface. More than 800 clones were screened in a CD40-specific enzyme-linked immunosorbent assay (ELISA) and 79 clones were selected based on their ability to bind to CD40-Fcγ. These clones were further evaluated in ELISA, using either CD40-Fcγ, biotinylated BSA, human Fcγ, or streptavidin. Fifty-one scFv that specifically bound to CD40 were recovered. These antibody fragments were subsequently analysed for binding to a B-cell line (BJAB) using a FACScan. Forty clones bound strongly to the BJAB cells and nine of these clones expressed a unique nucleotide sequence (Table 2). Only one unique clone (B44) was recovered from the selection on CD40-expressing cells, whereas all unique clones were retrieved at least once from the selection in immunotubes. Four of the nine clones (denoted A[clone number]) shared the same CDRH3 sequence (Table 2), while the CDRH3 of the other five clones (B44, C27, D13, E30 and F33) was shown to vary in length (8–15 amino acids) as well as in sequence (Table 2).

Table 2. Summary of the CD40-specific clones obtained after screening. The scFv named A[clone number], which share the same CDRH3, is the most common clone type obtained from the selections.

scFv clone	No. identical sequences	CDRH3
A24	1	ARAPVDYSNPSGMDV
A43	20	ARAPVDYSNPSGMDV
A49	5	ARAPVDYSNPSGMDV
A54	1	ARAPVDYSNPSGMDV
B44	7	ARILRGGS...GMDL
C27	2	ARADWEYYYY.GMDV
D13	2	ARHIYPW....GMDV
E30	1	ARMTPWYY...GMDV
F33	1	ARGWLL.......DY

Epitope mapping of the different anti-CD40 scFv fragments

In order to analyse the fine specificity of the selected scFv, the location of each scFv epitope was determined by epitope mapping. The ability of the scFv fragments to bind to truncated CD40-constructs, expressed on the surface of transfected COS-7 cells (Fig. 1a), were measured using FACScan analysis (Fig. 1b).

The four type A scFv fragments and D13 bound only to the intact CD40 molecule. Even the addition of a small peptide tag (AD2) N-terminal to the intact CD40 molecule abolished the binding, which indicated that the type A antibodies and D13 shared an epitope in the N-terminal part of the molecule. B44, C27 and E30 scFv mapped to the D1/B2 domain, which appears to be the same epitope that many murine mAbs bind to (Malmborg Hager et al. manuscript in preparation). Only one scFv clone, F33, was able to bind to D2/B1, whereas no clone bound to only the D3 or D4 domain. In conclusion, the selected scFv fragments represented different families of anti-CD40 specificities that all bound to different epitopes located in the two distal domains of CD40.

Determination of the kinetic parameters of scFv fragments

The affinity of the different clones was determined by surface plasmon resonance. The results are presented in Table 3, showing a wide span of affinities ranging from low nanomolar (3·1 nm for F33) to high nanomolar (600 nm for D13), i.e. approximately a 200-fold difference. The major contribution to the difference in affinity could be found in the association rate constants (k_a) of the different scFv fragments, which ranged from 0·1 to 160×10⁴ (Ms⁻¹). Furthermore, E30 has a similar affinity compared to F33, although the relative contribution of the association rate constant and the dissociation rate constant differ significantly, in that F33 has a more than 10 times lower off-rate than E30.

Table 3. Kinetic constants for the interaction between immobilized CD40 and anti-CD40 scFv antibodies as determined by surface plasmon resonance.

scFv clone	ka (×104/Ms)	kd (×10−3/s)	KD (nm)
A24	14·1	0·4	10
A43	1·8	0·6	35
A49	2·5	1·1	44
A54	2·3	0·9	40
B44	1·5	0·9	60
C27	41	6·6	16
D13	0·1	0·6	600
E30	160	5·8	3·6
F33	13	0·4	3·1

k_a, association rate constant; k_d, dissociation rate constant; KD, dissociation constant.

Effect of anti-CD40 scFv fragments on B-cell proliferation

In order to investigate the functional potential of the selected scFv fragments, a B-cell proliferation assay was used. The ability to induce proliferation of human B cells was assayed by cross-linking cell-surface-bound scFv fragments, using an anti-FLAG antibody (M2). The FLAG epitope was present and detectable on all the anti-CD40 scFv fragments, as determined by FACScan and ELISA. The different scFv varied over a wide range in their ability to activate and support B-cell proliferation (Fig. 2). ScFv B44 and E30 induced proliferation in the same range as G28-5,²⁴ while D13 and F33 induced significantly lower proliferation. G28-5, which we also have produced as a monomeric scFv, is a murine anti-CD40 antibody that efficiently promotes B-cell proliferation. ScFv C27 induced a rather weak proliferation (<20% compared to G28-5), whereas the type A scFv fragments (represented by A24 in Fig. 2) did not induce any proliferation at all. Even at concentrations as high as 10 µg/ml only a modest B-cell proliferation was detected.

Figure 2.

Ability of the selected anti-CD40 scFv to induce proliferation of B cells. Human B cells were cultured with irradiated CD32⁺ L cells, anti-CD40 scFv (1 µg/ml) and IL-4 (100 U/ml), with and without M2 anti-FLAG. Anti-FLAG cross-links the scFv on the L cells, thereby enabling the anti-CD40 scFv to oligomerize CD40. The ability to induce proliferation was measured with a thymidine incorporation assay after 3 days. A24 is representative for all the type A scFv. Data represent mean values ± SD, from three independent experiments. G28-5 is a well-characterized activating murine anti-CD40 antibody²⁷ that we produced as scFv fragment. Solid bars indicate that the scFv have been cross-linked with anti-FLAG, while hatched bars show proliferation without cross-linking.

Cross-linking of the scFv with the M2 anti-FLAG was necessary to induce proliferation of the selected clones, indicating that oligomerization is necessary to achieve an optimal B-cell activation via CD40 (Fig. 2). This also showed that the observed effect was not caused by any lipopolysaccharide contamination derived from the Escherichia coli expression system.

Anti-CD40 scFv fragments can rescue B cells from apoptosis

It has previously been shown that the ability to rescue B cells from apoptosis, in contrast to the ability to promote B-cell proliferation, is less dependent on epitope.²⁴ Consequently, we investigated whether the type A anti-CD40 scFv were able to rescue Ramos cells, a human Burkitt lymphoma cell line, from IgM-induced apoptosis.

The ability of the type A anti-CD40 scFv (with or without cross-linking with anti-FLAG) to abrogate induction of the apoptotic progamme was measured by annexin-FITC staining (Fig. 3). A24 scFv was unable to rescue Ramos cells from the anti-IgM-induced apoptosis, whereas scFv fragments that were able to promote proliferation, i.e. F33 and B44, almost completely protected the B cells from apoptosis after cross-linking with anti-FLAG. The A24 could not rescue the B-cell line from apoptosis even at 10 times higher concentration (data not shown).

Anti-CD40 scFv fragments can block the CD40–CD40L interaction

The different anti-CD40 scFv fragments were compared for their ability to inhibit the CD40–CD40L interaction. Biotinylated CD40-Fcγ was preincubated with scFv antibodies and any subsequent binding of this complex to CD40L-expressing cells was measured by FACS analysis (Fig. 4). The data demonstrated that the scFv fragments could inhibit the CD40–CD40L interaction to different degrees. F33 inhibited the CD40–CD40L interaction by almost 100%, suggesting that the epitope recognized by F33 is within or very close to the CD40L binding site. F33 recognized an epitope in the D2/B1 module (Fig. 1b), which is in agreement with findings that residues important for the CD40–CD40L interaction are located in the D2 and D3 domains.^4,25,26 The type A scFv fragments (represented by A24), B44, C27 and E30 were partial inhibitors (60–70%), whereas D13 scFv exhibited the lowest inhibitory effect (<45%).

Figure 4.

Ability of CD40-specific scFv to block the binding of soluble CD40Fc-biotin to L cells, expressing CD40L, as measured by flow cytometry. Data represent mean values ±SD of three independent experiments.

Discussion

Antibodies able to modulate the signalling via CD40 are potentially important in several therapeutic applications. However, all previously reported anti-CD40 antibodies are of non-human origin, which is a limiting factor for clinical success, since these antibodies are known to be associated with a number of side-effects. Analysis of the fully human phage display library called n-CoDeR, suggested that clones from this library had the lowest reported number of T-cell epitopes reported today²² and scFv fragments selected from this library could consequently constitute a preferred source for antibody-based clinical trials.

Therefore, we set out to select a variety of human anti-CD40 scFv fragments displaying different functional activities. The results demonstrate that we have managed to obtain scFv fragments with different abilities to induce B-cell proliferation, block the CD40–CD40 interaction and to prevent apoptosis. Some of the scFv fragments had an activating potential comparable to that of G28-5, a well-characterized murine mAb, whereas others did not induce proliferation at all (at 1 µg/ml). As reported by Ledbetter et al.,²⁷ G28-5 (in scFv format) was also able to induce a low level of proliferation without cross-linking on the cell surface. All our activating scFv fragments required cross-linking by the M2 anti-FLAG antibody to induce B-cell proliferation, an observation which is consistent with previous studies, using mouse antibodies.^24,28,29 Hence, when using our activating CD40 binders in vivo or in vitro it is necessary to use a multimeric format, which can be obtained by converting the scFv into a whole IgG or by using other recombinant techniques.^30–33

The reason for the inability of type A anti-CD40 scFv to promote B-cell proliferation is not currently known, although it might be a result of the properties of the specific epitope recognized by this set of scFv fragments. It has previously been shown that CD40-mediated rescue from apoptosis is epitope independent whereas the ability to induce cell cycle progression is not.²⁴ Interestingly, clone A24 was unable both to rescue Ramos cells³⁴ from anti-IgM induced apoptosis, and to induce proliferation. The strongly activating anti-CD40 scFv, F33 and B44 were, on the other hand, able to rescue cells from apoptosis, indicating that the epitope specificity did play a role in defining the functional properties of our scFv fragments.

When the affinity of the selected anti-CD40 scFv fragments was analysed in relation to B-cell activation no clear correlation could be seen, which also agrees well with previous work on anti-CD40 antibodies.³⁵ This can be illustrated by D13, one of the best activating scFv, which had approximately 200 times lower affinity compared to F33. Despite this difference in affinity F33 and D13 induced B-cell proliferation to the same degree. Furthermore, the scFv that induced the most vigorous B-cell proliferation, B44, had a moderate affinity.

Epitope mapping revealed that most of the scFv fragments recognize an epitope close to the N-terminus of CD40 and only clone F33 recognizes an epitope that contains D2/B2. The epitope of F33 is, consequently, located in the vicinity of the amino acid residues that have been reported to be critical for CD40L binding (located in D2 and D3).^3,4,25 This correlates well with the results from the blocking experiments where F33, in contrast to the other anti-CD40 scFv fragments, completely abrogates the CD40–CD40L interaction. However, the CD40L does not bind to the same CD40 construct as F33, since it requires D1/B2 for binding (Malmborg Hager et al., manuscript in preparation). It is possible that the CD40L require that CD40 preassociate, via its N-terminal domain in order to bind. This would indicate that the CD40–CD40L interaction depends on preassociation mediated by domains outside the CD40L binding site. This has recently been shown to be the case for other members of the TNFR superfamily.^36,37

Sequence analysis of all the selected scFv fragments demonstrated large differences in the CDRH3 sequences, with the exception of the type A family. This indicates that the scFv fragments most probably bind to different epitopes³⁸ and the domain mapping of the scFv also confirmed that they recognized at least three different epitopes on CD40. It is interesting to note that all the selected scFv fragments contained a combination of CDRs derived from several different germline loci. These combinations of CDR 1–3 do not exist in rearranged human variable genes, which illustrates that antibody fragments from the n-CoDeR library have access to a specificity space that is inaccessible for the human antibody repertoire in vivo. Consequently, specificities not accessible to Nature might be found in the n-CoDeR library, which may indicate that scFv fragments with novel functions can be selected.

Activating anti-CD40 antibodies have the potential to be used in treatment of haematological malignancies, since it has been shown that CD40 stimulation of some lymphomas leads to up-regulation of major histocompatibility class I, co-stimulatory/adhesion molecules as well as to apoptosis.⁸ Furthermore, this type of anti-CD40 antibody has, despite their immunostimulatory effect, been shown to have a protective effect in the evolution of CCIA.⁹ Another application for activating anti-CD40 antibodies has been suggested by Rolph et al.³⁹ They reported that a non-immunogenic, non-viable Listeria monocytogenes could be converted to an immunogenic vaccine by simultaneous delivery of an activating anti-CD40 mAb. This indicates that co-administration of anti-CD40 antibodies might be an approach for the development of novel vaccine formulations. In this study we have identified four strongly activating human anti-CD40 scFv fragments (B44, D13, E30 and F33). These clones are presently converted to whole IgG format and are further evaluated in cellular assays.

It might also be possible to use anti-CD40 antibodies that block the CD40–CD40L interaction in multiple sclerosis.¹⁰ Disrupting the CD40–CD40L interaction has proven to be successful in several animal models of autoimmune diseases,^40–43 and particularly the treatment of marmoset monkeys with a murine anti-CD40 mAb (5D12) had a distinct therapeutic effect, in that it delayed the onset of EAE.¹² One of the scFv families from our panel of anti-CD40 scFv, the type A scFv fragments, significantly blocks the CD40–CD40L interaction and yet induced a very low level of proliferation. The functional properties of these clones will also be evaluated as whole IgG molecules.

In this investigation, we have described a successful strategy for selection of a diverse set of fully human anti-CD40 antibody fragments. These scFv fragments exhibit a variety of different functional effects on human B cells, which will be further evaluated in preclinical models in an attempt to identify potential therapeutic candidates.

Abbreviations

CCIA

chronic collagen-induced arthritis

CDR

complementarity-determining region

EAE

experimental autoimmune encephalomyelitis

MS

multiple sclerosis

scFv

single-chain antibody fragment

TNF

tumour necrosis factor