Edinburgh Research Explorer Recombinant canine IgE Fc and an IgE Fc-TRAIL fusion protein bind to neoplastic canine mast cells

18 Screening for expression of the high affinity receptor for IgE by reverse transcriptase PCR, revealed 19 that almost all canine mast cell tumors expressed FcεRIα mRNA, supporting the rationale for 20 developing anti-neoplastic treatments based on molecules that could target this receptor. Use of 21 cytotoxic cytokines to trigger an apoptotic signal is one strategy for inducing cell death in malignant 22 mast cells. The coding sequences for canine IgE and tumor necrosis factor-related apoptosis-inducing 23 ligand (TRAIL) were identified through genome analyses. Selected regions of the coding sequences 24 for these genes were cloned and compared to the predicted genome sequences. The Fc region of 25 canine IgE, death domain of canine TRAIL and an IgE Fc : TRAIL fusion construct were generated and 26 epitope-tagged proteins expressed, using a eukaryotic expression system. Specific binding of 27 recombinant canine IgE Fc-containing proteins to recombinant human FcεRI  and to a canine mast 28 cell tumor line expressing FcεRI  (C2), but not one failing to express FcεRI  (MCLA), was 29 demonstrated. Specific binding of the IgE:TRAIL fusion protein was not abrogated by the TRAIL 30 moiety. These results are proof of principle that canine IgE targeting to FcεRI  can be used as a 31 platform for selective delivery of therapies to FcεRI  -expressing cells, potentially enhancing their 32 therapeutic index and efficacy.


Introduction
Canine mast cell tumors (cMCTs) are the most common skin malignancy in dogs (Brodey 1970;Finnie and Bostock, 1979;Bostock 1986;Dobson et al., 2002), likely arising from neoplastic transformation of resident tissue mast cells or their progenitors.Breed, location, stage, gross appearance, size, presence of paraneoplastic syndromes and several proliferation markers have been variably associated with prognosis, however the most consistent predictor of cMCT recurrence, metastasis and survival is histopathological grade (Bostock, 1973;Patnaik et al., 1984;O' Keefe, 1990;Gerritsen et al., 1998;Mullins et al., 2006).Although dogs with a localised cMCT are often cured by local therapy (surgery and/or radiation therapy), those with an inoperable primary mass or confirmed disseminated disease usually die of their disease (reviewed by Welle et al., 2008;Blackwood et al., 2012).The use of both local (e.g.surgery and radiation therapy) and systemic therapies (e.g.chemotherapy, receptor tyrosine kinase inhibitors) is limited by the potential for adverse effects, some of which can be idiosyncratic, cumulative and/or permanent in nature.Thus, there is a need to develop novel therapies for cMCTs with improved efficacy and higher therapeutic indices.
Mast cells avidly and specifically bind to IgE via the high-affinity IgE receptor, FcεRI.Crosslinkage of surface IgE by antigen leads to mast cell degranulation and inflammation, an important defence against parasite infestation.In addition to disease caused by malignant transformation, mast cells are also involved in type I hypersensitivity reactions in allergic disease.Additional beneficial and pathophysiological roles for mast cells are emerging, although many of these can be undertaken by other, more numerous cells (reviewed by Rao & Brown, 2008), and mast cells are not absolutely required for survival, as demonstrated by genetically-modified mouse strains lacking these cells (C57BL/6-KITW sh -/W sh-, Grimbaldeston et al., 2005).However, other genetic mutations associated with mast cell deficiency can lead to significant developmental problems in cells of other lineages.WBB6F 1 -W/W v mice have reduced erythrocyte, granulocyte, platelet and mast cell numbers (Kitamura et al., 1978), and although such mice are more likely than wild type mice to die in a model of acute septic peritonitis involving cecal ligation and puncture, their survival can be enhanced solely by reconstitution with cultured wild type mast cells (Echtenacher et al., 1996).
FcεRI is constitutively expressed by mast cells and basophils during early differentiation, but lower level, inducible expression has also been found on mammalian eosinophils, monocytes, platelets and dendritic cells (Thompson et al., 1990;Rottem et al;1992;Joseph et al., 1997;Kinet 1999;Kita et al., 1999;Seminario et al., 1999).The affinity and avidity of IgE for FcεRI exceeds that of other immunoglobulin/Fc receptor interactions by several orders of magnitude (Ravetch and Kinet, 1991;Maenaka et al., 2001;McDonnell et al., 2001;Wan et al., 2002).Previous work, evaluating the role of each IgE domain in binding to its high affinity receptor, has shown that C H 3 is critical for binding to FcRI, but C H 2 and C H 4 are also required for high affinity and avidity binding (Keown et al., 1997;Garman et al., 2000Garman et al., & 2001;;McDonnell et al., 2001;Wurzburg and Jardetzky, 2002;Hunter et al., 2008).In a previous study, a partially dimerized C H ε2-4 recombinant canine protein was used to raise an anti-canine IgE antibody which detected native canine IgE, suggesting that the quaternary structure of the recombinant C H ε2-4 protein was similar to native canine IgE, despite the lack of a C H 1 domain (Ledin et al., 2006).Hunter et al. (2008) reported an alternative cell-based binding assay using RBL-2H3 cells stably transfected with rcFcεRIα and elegantly evaluated the role of each IgE heavy chain domain in binding to the receptor.IgE has been used as a delivery platform for targeting murine and human mast cells, when linked to Pseudomonas toxin or pro-apoptotic Bcl-2 protein family members.These fusion proteins were capable of binding to and inducing apoptosis in mast cells in vitro and in vivo.(Fishman and Lorberboum-Galski, 1997;Fishman et al., 2000;Belostotsky and Lorberboum-Galski, 2001) A rmIgE 301-473 -Pseudomonas exotoxin A chimeric protein was highly cytotoxic to both malignant murine mast cell lines and bone marrow-derived mast cells, while having no effect on other cell lines of haematopoietic lineage (Fishman and Lorberboum-Galski, 1997), demonstrating its specificity for FcεRI-expressing cells.
Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL, TNFSF10) is a type II membrane protein, an extracellular fragment of which can be cleaved to generate soluble TRAIL (sTRAIL; amino acids 37-281) (Wiley et al., 1995;Pitti et al., 1996;Mariani and Krammer, 1998).In humans and rodents, healthy cells display a repertoire of death and decoy TRAIL receptors, whereas neoplastic/transformed cells preferentially express death receptors over decoy receptors (Sheridan et al., 1997;Simonet et al., 1997;Pan et al., 1998;LeBlanc and Ashkenazi, 2003).Therefore, neoplastic cells are generally more susceptible than healthy cells to TRAIL-mediated apoptosis, making this a tempting therapeutic strategy (LeBlanc and Ashkenazi, 2003).Although the canine TRAIL receptor repertoire is poorly characterized and there are no TRAIL receptor genes identifiable in the syntenic region of the canine genome, compared to human and mouse, we and others have demonstrated that canine neoplastic mast cells are susceptible to the apoptotic effects of TRAIL (Rong et al., 2008;Elders et al., 2009).
The aim of the current study was to generate a recombinant canine IgE Fc protein that was capable of binding to the FcεRI receptor expressed on canine mast cells.The intention was that this recombinant canine IgE Fc molecule could then be modified to allow specific targeting approaches to be evaluated for treatment of canine mast cell tumors.Since we had previously shown that a canine MCT line (C2) was susceptible to TRAIL-mediated apoptosis, a canine IgE-TRAIL fusion protein was planned in the first instance to test the concept that this targeting strategy was feasible.

Case recruitment and sample collection
Biopsy samples were obtained from dogs undergoing surgical resection of a mass, suspected or cytologically confirmed as cMCT.Tissue was stored in RNAlater® (Qiagen, Crawley, UK) at -20 °C prior to molecular analysis.All tumors had histopathology or cytology performed and, where available, representative slides were reviewed by a single pathologist (K.C.S.) using a published grading scheme (Patnaik et al., 1984).In one patient, tumor tissue was disaggregated and cultured to generate a novel cMCT line, designated MCLA, which demonstrated metachromatic granule staining with Toluidine blue and was positive for expression of chymase and tryptase (Elders, 2009).
Residual canine EDTA blood, following completion of processing of a diagnostic sample, was used to provide canine genomic DNA for IgE cloning.Similarly, residual lymph node tissue, following diagnostic processing of a lymphoma sample, was used to provide cDNA for TRAIL cloning.

Nucleic acid extraction and polymerase chain reaction
Tumor tissue or cultured cells were homogenized in Lysis Solution for total RNA (Sigma, Poole, UK) and RNA extracted using the GenElute TM Mammalian Total RNA Miniprep Kit (Sigma).On-column (RNase-free DNase set, Qiagen) and in-solution (Turbo DNase TM , Ambion, Austin, USA) digestion steps were performed to remove contaminant genomic DNA.Reverse transcription of mRNA into complementary DNA (cDNA) was performed using oligo(dT) 15 primer and ImProm-II reverse transcriptase (Promega, Southampton, UK).
Polymerase chain reaction was used to amplify genes of interest using specific primers (Table 1).Each 25 µl reaction contained 1 µl both sense and anti-sense primers (final concentration 10 pmol/µl) and cDNA or gDNA (1 µl) as template.Reactions also contained 1× Hi-Spec Additive, 1× NH 4 buffer, MgCl 2 (final concentration 2.0-3.0 mM), dNTP (final concentration 10 mM) and 0.5 IU Immolase™ DNA polymerase per reaction (all from Bioline, London, UK), made up to 25 µl with water.Reactions were heated to 95 °C for 10 mins, followed by 20 -35 cycles at 94 °C for 40 s, 55 -62.5 °C for 30 s, and 72 °C for 1 min; with a final extension step at 72 °C for 7 mins using a G Storm thermocycler (Gene Technologies Ltd, Essex, UK).For PCR products that were to be cloned, a proofreading DNA polymerase (Easy-A® High-Fidelity PCR Cloning Kit, Stratagene, Amsterdam, The Netherlands) was used in place of Immolase.

Cloning of inserts into plasmid DNA vectors
PCR amplicons were extracted (GenElute TM gel extraction kit; Sigma), cloned into the pSC-A® vector (Stratagene) and transformed into E. coli (Solopack TM , Stratagene).Plasmid DNA was then isolated using the GenElute™ Plasmid Miniprep Kit (Sigma) and sequencing to confirm the integrity of the inserts (Geneservice, Cambridge UK).Donor (recombinant pSC-A) and recipient (pSecTagA®; Invitrogen) plasmid DNA were digested using the indicated restriction enzymes to allow directional subcloning.Following purification, insert DNA was ligated into digested recipient vector (Quick-Stick TM DNA Ligase, Bioline) and E. coli.transformed.Endotoxin-free plasmid DNA was prepared from E. coli cultures using GenElute™ Endotoxin-free Plasmid Midiprep Kit (Sigma), according to the manufacturer's instructions.

Transfection of Chinese hamster ovary cells
Cells for transfection were seeded at 1×10 5 cells per well into 24-well plates in a 500 μl volume of culture medium containing 5% FBS.When 90-95% confluent, the medium was replaced with FBS and antibiotic-free medium.Cells were transfected with 800 ng endotoxin-free recombinant plasmid DNA using Lipofectamine TM 2000 (Invitrogen) according to the manufacturer's instructions.Plates were incubated for 6 h then the plasmid-containing medium was replaced with medium containing 10% FBS but lacking antibiotics, and after a further 18 h this medium was replaced with culture medium.After 72 h incubation, the supernatant, containing recombinant protein, was recovered, centrifuged (13.8 ×g, 5 mins), filtered (0.22 μm filters, Millex ® GV syringe filter unit, Sigma), and used for experiments.

Detection of recombinant proteins
Pre-blocked, nickel coated ELISA strips (Ni-NTA HiSorb®, Qiagen) were used to capture polyhistidine-tagged recombinant proteins from transfected CHO supernatants (100 μl/well).After 2 h incubation, wells were washed with PBS supplemented with 0.1% Tween 20 (Sigma).A caprine anti-c-myc:HRP conjugate (A190-104P; Universal Biologicals, Cambridge, UK), diluted 1:5,000 in phenol red-free MEM/10% FBS was added for 2 h, followed by six washes.One hundred microliters of substrate (Supersensitive TMB Liquid Substrate for ELISA; Sigma) was added per well and incubated for 20 min.Twenty five microliters of 0.5 M sulphuric acid (SLS, Nottingham, UK) was added to stop the reaction, and the optical density (O.D.) at 540 nm subtracted from that at 450 nm to generate absorbance values per well.The ELISA was modified by substituting the anti-c-myc antibody with either a polyclonal caprine anti-canine IgE:HRP conjugate (1:10,000 dilution of AHP946P, Abd Serotec Ltd., Oxford, UK), or a polyclonal caprine anti-human TRAIL antibody (1:100 dilution of AF375, AbD Serotec) which was in turn detected using a polyclonal anti-caprine IgG:HRP conjugate (1:10,000 dilution of OBT1500P, Abd Serotec).Samples of transfected CHO supernatants, containing recombinant protein, were also submitted to HESKA (Fribourg, Switzerland) for analysis using their validated Allercept® canine IgE ELISA (Foster et al., 2003).

Western blotting
Polyhistidine-tagged proteins were purified from CHO supernatants using the MagneHis TM Protein Purification System (Promega).Purified recombinant proteins (32.5 μl) were added to 5 μl of 500 mM dithiothreitol (DTT) reducing agent and 12.5 μl NuPAGE ® LDS sample buffer (both Invitrogen) and incubated at 70 °C for 10 mins.Proteins were separated by PAGE using 4 -20% gels and sodium dodecyl sulphate (SDS) running buffer (both PAGEgel.com,California, USA), in the X-cell SureLock TM Mini-cell (Invitrogen) at 200 V constant for 35 mins.Proteins were then transferred to nitrocellulose membranes under reducing conditions in the X-Cell II TM Blot Module (Invitrogen) and transfer buffer (PAGEgel.com)at 30 V constant for 1 h.Membranes were rinsed, blocked overnight at 4 °C in 50 ml PBS/5% Marvel, then incubated with murine anti-His:HRP conjugate (MCA1396P, Abd Serotec, diluted 1:2500 in 25 ml PBS/5% Marvel™/0.1% Tween 20) at room temperature for 1 h.
After washing, immunoreactivity was detected by chemiluminescence using the ECL TM Western blotting analysis system (GE Healthcare, Chalfont St Giles, UK) and film (Kodak X-Omat AR TM , Kodak, Harrow, UK).

Analysis of FcεRIα mRNA expression in canine mast cell tumors and selected mastocytoma cell lines
The canine FcεRIα coding sequence (762 bp), consisting of 5 exons, is located on chromosome 38 (GenBank Accession# NM_00110766.1,Goitsuka et al., 1999).Screening primers, designed to amplify a 273bp fragment, were used to demonstrate that 23 of the 25 canine MCT biopsy samples were positive for FcεRIα mRNA expression (Fig. 1A), suggesting that this represents a suitable target for potential immunotherapeutic strategies.The C2 mastocytoma cell line but not the newly established MCLA mast cell tumor line, expressed FcεRIα mRNA (Fig. 1B).

Generation of recombinant canine IgE Fc and IgE Fc-TRAIL fusion constructs
The canine IgE heavy chain gene, encoding V H ε and C H ε1-4, is located on chromosome 8 and consists of 5 exons (GenBank Accession# XM_548007.1).Using genomic DNA as template, primers were designed at the start of exon 3 (encoding C H ε2) and the end of exon 5 (encoding C H ε4), to generate a PCR amplicon spanning the C H ε2-4 coding region, but which contained additional intronic sequence.This was cloned into the pSecTagA vector and following sequence verification, plasmid DNA was transfected into Chinese hamster ovary (CHO) cells.Isolation of cDNA from the transfected cells and further PCR analysis revealed three amplicons (Fig. 2), the smallest of which was consistent with splicing out of both introns.This was confirmed by sequencing, and this canine C H ε2-4 coding region was subsequently cloned into the pSecTagA vector between the HindIII and EcoRI sites (pSecTagA/rcIgE 100-427 ).
The canine TRAIL gene, consisting of 5 exons is located on chromosome 34 (GenBank Accession# NM_001130836.1).Using cDNA prepared from a canine lymphoma biopsy as template, canine TRAIL-specific primers were used to generate an amplicon encoding the predicted C-terminal TNF-like death domain (amino acids 114-282), which was cloned into pSecTagA (pSecTagA/rcTRAIL 114-282 ).In addition, a fusion construct was generated by subcloning the TRAIL sequence, between EcoRI and XhoI sites, downstream of and in-frame with the IgE Fc fragment to generate a fusion construct (pSecTagA/rcIgE 100-427 : rcTRAIL 114-282 ).

Expression of recombinant canine IgE Fc and IgE Fc-TRAIL fusion proteins
Chinese hamster ovary cells were transfected with pSecTagA containing, rcIgE 100-427 , rcTRAIL 114-282 , or rcIgE 100-427 : rcTRAIL 114-282 fusion constructs.Polyhistidine and c-myc epitope-tagged recombinant proteins were detected by ELISA in the supernatants of CHO cells transfected with the three constructs (Fig. 3A), while none was found in cell lysates (data not shown).This ELISA was modified by substitution of the anti-c-myc detection antibody with either an anti-canine IgE conjugate (Figure 3B), or an anti-human TRAIL antibody (Figure 3C).This confirmed expression of the relevant epitope-tagged recombinant proteins containing canine IgE Fc and/or TRAIL protein domains.

Binding of recombinant canine IgE Fc and IgE Fc-TRAIL fusion proteins to FcεRIα
Using the Allercept system, it was demonstrated that rcIgE 100-427 and rcIgE 100-427 :TRAIL 114-282 were able bind to recombinant FcεRIα, whereas rcTRAIL 114-282 was unable to do so (Fig. 5).By flow cytometric analysis, C2 cells (FcεRIα positive) bound monoclonal canine IgE as well as rcIgE 100-427 (Figure 6A), whereas MCLA cells (FcεRIα negative) failed to demonstrate any binding to monoclonal IgE or rcIgE 100-427 (Figure 6B).Competitive inhibition of binding of rcIgE 100-427 to C2 cells (detected with anti-myc antibody) was demonstrated by co-incubation with monoclonal IgE, with MFI values falling from 64.4 in the absence of the monoclonal antibody to 27.5 in the presence of the monoclonal antibody.In addition to binding to rcIgE 100-427 , C2 cells were able to bind to the rcIgE 100- 427 :TRAIL 114-282 fusion protein but not rcTRAIL 114-282 (Figure 6C).

Discussion
The features of the canine patients and cMCT recruited for this study are similar to previous reports (Sfiligoi et al., 2005;Mullins et al., 2006;Newman et al., 2007;Thamm and Vail, 2007).
Although a retrospective study of archived tumor tissue would have produced a much larger cohort, we have found RNA extracted from formalin-fixed material to be of relatively poor quality, compared to tissue specifically preserved for molecular analysis (Stell, 2008).FcεRIα mRNA expression was consistent throughout the biopsy samples assessed in the study, which contained both well-differentiated and poorly-differentiated tumors.Although mRNA expression does not necessarily indicate expression of functional protein, consistent FcεRIα mRNA expression suggests that development of IgE-based therapies might be worthwhile.The logical next step would be to analyse a sample cohort composed primarily of high grade, poorly-differentiated or metastatic MCTs where IgE-based immunotherapeutics could be particularly valuable, given that the majority of high grade and disseminated tumors are refractory to current anti-cancer treatments.
Established from a cross-breed dog with tumor recurrence, C2 mastocytoma cells are highly differentiated, expressing surface FcεRI capable of binding canine IgE (Lazarus et al., 1986, Brazis et al., 2002;Hunter et al., 2009).The MCLA mastocytoma cell line was established from a Labrador retriever with a cytologically diagnosed, metastatic MCT.Cells demonstrated typical morphology, although they failed to express FcεRIα.Expression of FcRI might have been absent at the time of biopsy, consistent with a neoplasm derived from an early mast cell precursor cell (Thompson et al., 1990;Rottem et al., 1992), de-differentiation and loss of FcRI expression in a tumor arising from a later precursor or a mature mast cell, or might have been downregulated during establishment of the cell line, in the absence of IgE which stimulates expression of the receptor (Furuichi et al., 1985;Hsu and MacGlashan, 1996;Yamaguchi et al., 1997;Kubo et al., 2001).There were no detectable KIT exon 11 mutations in this cell line, which also lacked KIT mRNA expression (data not shown).This might also be evidence of an early mast cell precursor lineage, although canine mast cell lines lacking KIT expression and KIT mutation have been reported which are also independent of stem cell factor for their propagation (Ohmori et al., 2008).However, it is difficult to reconcile the lack of FcRI and KIT mRNA expression with the characteristic morphological appearance of this cell line and the expression of chymase and tryptase mRNA (data not shown), which would suggest more advanced differentiation.
Native IgE (with dimerized heavy chains) binds to the FcεRIα chain on the cell surface with a prolonged half-life, compared to other types of Fc receptor, whereas IgE heavy chain monomers are rapidly internalized (Menon et al., 1986).In the current study, constructs were designed to allow dimerisation of the heavy chain fragments of rcIgE 100-427 as this might facilitate long-term cell surface binding for detection and in the case of rcIgE 100-427 :TRAIL 114-282 would also increase the local concentration of TRAIL death domain moieties on the cell surface for signalling through relevant TRAIL receptors.However, formation of inappropriate intra-molecular disulphide bonds or intermolecular disulphide bonds between rcIgE 100-427 proteins, might have resulted in an altered tertiary and/or quaternary structure.There are two unpaired cysteine residues within the fusion protein (IgE 100CYS and TRAIL 230CYS ) that could potentially lead to misfolding.Cysteine 230 in human TRAIL is important for trimerisation, with mutation of this amino acid abrogating its apoptosisinducing effect (Seol and Billiar, 2000;Trabzuni et al., 2000;Kelley et al., 2001).An IgE construct lacking the N-terminal cysteine was generated (pSecTagA/IgE 99-427 ) although this failed to express recombinant protein when transfected into CHO cells (data not shown).
An ELISA exploiting the polyhistidine epitope tag for capture and c-myc tag for detection consistently demonstrated recombinant proteins secreted into the supernatant of transfected CHO cells, although the relative immunoreactivity varied between experiments, likely reflecting differences in transfection efficiencies in different wells.Substituting the c-myc antibody with anticanine IgE or anti-TRAIL antibodies revealed the presence of IgE and/or TRAIL epitopes in the recombinant proteins, suggesting adequate levels of protein folding.Furthermore, a modified Allercept system demonstrated binding of epitope-tagged rcIgE 100-427 to a recombinant FcRI conjugate, usually employed to detect serum IgE in dogs suffering from allergic disease.Similar binding of the rcIgE 100-427 :TRAIL 114-282 fusion protein was also demonstrated using this assay, confirming that addition of the TRAIL component did not interfere with the ability of the IgE Fc to bind to its cognate receptor.
Binding of recombinant IgE Fc constructs to canine FcRI was further investigated by flow cytometry, taking advantage of the availability of the novel FcRI negative MCLA mast cell line as a control for the FcRI positive C2 mastocytoma line.Specific binding of recombinant IgE 100-427 to C2 cells was demonstrated, which could be inhibited by monoclonal canine IgE.These results provide strong supportive evidence for specific binding of rcIgE 100-427 to canine FcεRI.The C2 cellular binding experiments were conducted at 4 °C and it might be the case that at 37 °C rcIgE 100-427 proteins could possibly be endocytosed and internalized.This could have been detected through incubating cells for varying periods at 4 °C or 37 °C and comparing the binding of recombinant proteins on the cell surface by anti-c-myc/anti-cIgE staining.Alternatives could have included assessing binding at 37 °C in the presence or absence of a metabolic inhibitor (e.g.sodium azide) to prevent internalisation, or the use of confocal microscopy.
The authors have previously demonstrated the preferential efficacy of rhTRAIL 114-281 in inducing apoptosis in C2 cells compared to non-neoplastic MDCK cells (Elders et al., 2009).Herein, rcTRAIL 114-282 was selected as a prototype cytotoxic agent that was added to the IgE delivery system, as this would potentially combine selective targeting and selective killing of malignant cells of this lineage.This particular TRAIL fragment is similar to the one shown to induce apoptosis in several canine neoplastic cell lines (Rong et al., 2008), although the rcTRAIL protein used by Rong et al., (2008) is slightly bigger than the current protein (by 3 amino acids at the N-terminus) and the current construct expressed C-terminal c-myc and polyhistidine tags.
Although receptor-binding studies demonstrated that rcIgE 100-427 :TRAIL 114-282 fusion protein was able to bind to C2 cells, subsequent cell viability assays failed to demonstrate any increased levels of apoptosis when C2 cells were cultured in the presence of the recombinant fusion protein (data not shown).Reasons for the lack of biological activity of the fusion protein include inadequate concentration, interference from other components in the cell supernatant or mis-folding of the recombinant protein.Furthermore, whereas IgE monomers form dimers, TRAIL functions exclusively as a trimer (Trabzuni et al., 2000).Thus, the fusion protein as designed would not have an intrinsic apoptosis-inducing effect.Further experiments combining IgE Fc-TRAIL fusion protein with TRAIL monomers, attempting to enhance C-terminal trimerisation also failed to demonstrate any biological effect (data not shown).

Conclusions
The specific binding of recombinant canine IgE Fc and a TRAIL fusion protein to canine mast cells expressing FcεRI is proof-of-principle that an IgE-targeted approach is a feasible strategy for mast cell-directed immunotherapeutics in dogs.TRAIL targeted to mast cells via IgE offered the possibility of induction of apoptosis specifically in neoplastic mast cells; however, this proved inefficacious in vitro.Future work aims to generate alternative IgE-Fc fusion proteins that can be utilized for antibody-directed enzyme prodrug therapy (ADEPT).As other cells seem capable of performing the majority of the beneficial functions of mast cells in vivo, the therapeutic index for such IgE-based immunotherapeutics is likely to be high for cMCT-bearing patients, provided that adequate endo-and ecto-parasite control is implemented during therapy, with likely repopulation of the tissues with mast cells from bone-marrow derived precursors after treatment is complete.

Fig. 1 .
Fig. 1.Screening of canine MCT biopsies and mastocytoma cell lines for expression of FcεRIα mRNA.PCR was used to assess FcεRIα mRNA expression (273 bp amplicon) in (A) cMCT biopsies (B-Z) and (B) mastocytoma cell lines C2 and MCLA (MC).PCR products were separated by agarose gel electrophoresis and visualized under UV illumination.La: 100 bp molecular weight ladder.H2O: water negative control.

Fig. 2 .Fig. 3 .
Fig. 2. PCR screening for canine IgE mRNA in control and transfected CHO cells.PCR was used to amplify the C H ε2-4 region of canine IgE from cDNA prepared from CHO cells transfected with the pSecTagA vector containing this region of the genomic sequence.PCR products were separated by agarose gel electrophoresis and visualized under UV illumination.L = 100 bp molecular weight ladder.H = water negative control; G = genomic DNA positive control; UC = untransfected CHO cDNA; UN = untransfected CHO NRT; TN = transfected CHO NRT; TC = transfected CHO cDNA, which produced amplicons with neither (largest fragment), one or both (smallest fragment) introns excised.

Fig. 6 .
Fig. 6.Binding of recombinant proteins to C2 and MCLA mastocytoma cells.C2 (A and C) or MCLA cells (B) were incubated with mock-transfected CHO supernatant (medium) or supernatant containing rcIgE 100-427 (rcIgE), rcIgE 100-427 :TRAIL 114-282 (rcIgE/TRAIL) or rcTRAIL 114-282 (rcTRAIL), followed by labelling with polyclonal anti-canine IgE:FITC (panel B) or polyclonal anti-c-myc:FITC (panel C) or in the absence of a secondary antibody (panel A) and analysed by flow cytometry.The graphs show histogram overlays of cells incubated with the various recombinant proteins.The tables show mean fluorescence intensity (MFI) data for each incubation condition.

Table 2 .
Details of MCT-bearing patients.