An immunologically relevant rodent model demonstrates safety of therapy using a tumour‐specific IgE

Abstract Background Designing biologically informative models for assessing the safety of novel agents, especially for cancer immunotherapy, carries substantial challenges. The choice of an in vivo system for studies on IgE antibodies represents a major impediment to their clinical translation, especially with respect to class‐specific immunological functions and safety. Fcε receptor expression and structure are different in humans and mice, so that the murine system is not informative when studying human IgE biology. By contrast, FcεRI expression and cellular distribution in rats mirror that of humans. Methods We are developing MOv18 IgE, a human chimeric antibody recognizing the tumour‐associated antigen folate receptor alpha. We created an immunologically congruent surrogate rat model likely to recapitulate human IgE‐FcεR interactions and engineered a surrogate rat IgE equivalent to MOv18. Employing this model, we examined in vivo safety and efficacy of antitumour IgE antibodies. Results In immunocompetent rats, rodent IgE restricted growth of syngeneic tumours in the absence of clinical, histopathological or metabolic signs associated with obvious toxicity. No physiological or immunological evidence of a “cytokine storm” or allergic response was seen, even at 50 mg/kg weekly doses. IgE treatment was associated with elevated serum concentrations of TNFα, a mediator previously linked with IgE‐mediated antitumour and antiparasitic functions, alongside evidence of substantially elevated tumoural immune cell infiltration and immunological pathway activation in tumour‐bearing lungs. Conclusion Our findings indicate safety of MOv18 IgE, in conjunction with efficacy and immune activation, supporting the translation of this therapeutic approach to the clinical arena.


| INTRODUCTION
Preclinical safety assessments for monoclonal antibodies (mAbs), especially novel immunomodulatory agents, attracted regulatory attention following the significant unexpected adverse events observed with an anti-CD28 super-agonist mAb targeting regulatory T cells during its first-in-human (FIH) clinical trial in 2006. 1,2 Identification of a pharmacologically relevant animal species for conducting safety studies is now incorporated in the European Medicines Agency guidance on taking novel therapeutics from preclinical studies into FIH clinical trials. 2 Design of biologically and immunologically relevant models represents an even greater challenge for therapeutic agents with diverse, often multimodal functional profiles, or for those representing novel therapeutic classes.
For many therapeutic human IgGs, preclinical evaluation of Fcmediated immune functions is often performed in immunocompetent mice, because human IgG Fc interacts with murine FccRs. 3 For the same reason, preclinical safety evaluation of IgGs is conducted in mice or primates, to study the antibody intended for human administration. In contrast, mouse models for studies of IgE classspecific immunological functions and safety, including for cancer immunotherapy, are not useful in the clinical translation of this class.
Differences between species render the mouse system inadequately representative of human IgE biology. These include the absence of human IgE binding to murine FceRs 4,5 and differential FceRI structure (murine FceRI is tetrameric, while human FceRI exists both as a tetramer and a trimer). 4 Murine FceRIs are expressed solely on mast cells and basophils, 4 whereas in rats, similar to humans, expression is on mast cells, basophils, eosinophils, monocytes and macrophages. 6,7 These features suggest the rat, rather than the mouse, as a model immune system better suited to preclinical evaluations of human IgE.
We previously reported the in vivo efficacy of recombinant chimeric (mouse/human) antibody, hMOv18 IgE, specific for the human tumour-associated antigen folate receptor a (FRa), overexpressed in solid tumours including ovarian carcinomas. The antitumour activity of hMOv18 IgE was superior to the equivalent hMOv18 IgG in a syngeneic rat model, and in two human ovarian carcinoma xenograft mouse models reconstituted with human immune cells. [8][9][10][11][12] In preparation for a clinical trial of hMOv18 IgE, we hypothesized that an immunocompetent rat model would provide the most comprehensive evaluation of safety and efficacy. This model was selected because it provides i) a species in which the native FceR is expressed on effector cell populations similar to human cells, ii) an opportunity to study syngeneic tumours in highly vascularized lungs of immunocompetent animals, iii) a self-replenishing supply of native effector cells and iv) a chance to evaluate antibody safety in the presence of antitumour IgE-mediated responses in tumour-bearing animals. 11 Although human exposure to exogenous IgE has been reported, 13 an IgE antibody recognizing a tumour antigen has not, until now, been introduced as a potential anticancer therapeutic. A conceivable reservation in contemplating a clinical trial is the perceived risk that intravenous IgE administration may trigger systemic or organ-specific toxicities, including blood basophil activation and induction of inflammatory cascades, potentially leading to cytokine storm or allergic reactions. We previously explored the propensity of human antitumour IgE to trigger human blood basophil activation and mast cell degranulation in healthy volunteers' and cancer patients' blood ex vivo. We found no evidence of FceRI-mediated activation of basophils or mast cell degranulation with hMOv18 IgE. 14 However, the safety of hMOv18 IgE in a physiologically and immunologically relevant immunocompetent tumour-bearing in vivo system has not yet been evaluated.
Here, we report the design and implementation of a syngeneic tumour model in immunocompetent rats to examine IgE safety and compare antitumour efficacy with that of the equivalent IgG2b. Surrogate rat MOv18 IgE and IgG2b antibodies (rMOv18) were generated. Treated rats were subjected to clinical, histopathological and metabolic evaluation. Immune cell recruitment and immune mediator release were interrogated to provide preclinical insights into this most appropriate immunocompetent host.

| ME TH ODS
See Appendix S1 for details about cell lines, isolation of rat immune cells from blood, antibody binding to cell surface receptors, quantitative analysis of gene expression in antibody-treated tumourbearing rat lungs, Gene Set Enrichment Analysis and multicytokine bead immunoassays.

| Generation of anti-FRa antibodies with human and rat Fc regions
Chimeric mouse/human antibodies, hMOv18 IgE and IgG1 were prepared as before. 8 Chimeric mouse/rat rMOv18 IgE and IgG2b were designed with rat constant and mouse variable domains specific for human FRa ( Figure 1A). Heavy and light chain region sequences were cloned into single mammalian expression vectors containing a hygromycin resistance gene (pVITRO1) 15 and transfected into Free-Style TM 293-F cells. Selected rMOv18 IgE-and IgG2b-transfected cells were expanded using an AppliFlex Single-Use Bioreactor (Applikon Biotechnology) (50L bioreactor wave bag, 25L working volume).

G R A P H I C A L A B S T R A C T
To study IgE class-specific functions and safety, we designed a surrogate rat tumour model and rat equivalent antibodies of a human/chimeric IgE therapeutic candidate recognising a tumour associated antigen. IgE restricted tumour growth, elevated intratumoural immune cell infiltration and enhanced immune pathway activation, without clinical, histopathological or metabolic toxicity signs, nor immunological evidence of allergic response or cytokine storm. This model may inform translation of anti-tumour IgE antibodies to clinical study, and help to dissect IgE functions in AllergoOncology, allergic inflammation, allergen immunotherapy and anti-parasitic responses.

| Purification of recombinant anti-FRa antibodies with human and rat Fc regions
rMOv18 IgE and IgG2b antibodies were purified using MEP (4-mercapto-ethyl-pyridine) HyperCel TM Mixed-Mode Chromatography Sorbent (Pall International) and affinity chromatography using a HiTrap TM Protein G HP column (GE Healthcare), respectively. 16 The MEP HyperCel TM resin was packed into an XK26/20 column (GE Healthcare) and equilibrated with PBS pH 7.0 (flow rate 5 mL/min).

| Clinical pathology
Blood samples were drawn into EDTA and lithium heparin collection tubes and analysed for haemoglobin concentration, total white blood cell count, neutrophil count, and creatinine, alanine aminotransferase and aspartate aminotransferase concentrations (haematological parameters: ADVIA 120 haematology analyser; biochemistry parameters: Roche Modular Evo (P800) clinical chemistry analyser). Full necropsies were performed at day 30 following tumour challenge. Tissues were preserved in neutral buffered formaldehyde before processing for histological evaluation.

| Histological evaluations
Rat tissues (thymus, injection site (tail), spleen, lymph nodes, liver, kidney, lungs, colon and brain) were submerged in neutral buffered formaldehyde, embedded in paraffin wax, cut at a nominal thickness of 4-5 lm, stained with haematoxylin and eosin and analysed on a Zeiss Axiophot microscope (910, 920 magnification lenses, Carl Zeiss) and NIS Elements imaging software (Nikon).

| Statistical methods
Parametrically distributed data were compared with an unpaired two-tailed Student's t test. Nonparametric data were compared using the Mann-Whitney U test. All statistical analyses were performed using GraphPad TM Prism software (version 5.03, GraphPad TM ). P values were represented as follows: *P < .05, **P < .01, ***P < .001, ****P < .0001. Error bars represent SD and SEM in in vitro figures, and SEM in in vivo figures and histological analyses. All results represent at least 2 independent experiments.

| Generation and receptor-binding characteristics of rat MOv18 IgE and IgG2b
We generated anti-human FRa mouse/rat chimeric (mouse variable/ rat constant domains) rMOv18 IgE and IgG2b, surrogates of the original mouse/human chimeric hMOv18 IgE and IgG1, respectively. Rat IgG2b isotype was chosen as equivalent to mouse IgG2a/b and human IgG1, based on complement activation and antibody-dependent cell-mediated cytotoxicity (ADCC) functions ( Figure 1A). 19

| Establishment of a tumour-bearing surrogate rat model to interrogate IgE functions
Folate receptor a expression is reported to be absent or lowly expressed in normal tissues, 23

| Safety of rat MOv18 IgE and IgG2b in vivo
We interrogated the rat model with respect to safety using scoring criteria previously described to evaluate general toxicities in rodents (Table S1A) and in vivo anaphylactic responses (Table S1B). 17,18 We also conducted histopathological evaluations of several organs to detect tissue damage, especially from immunological infiltration or tissue remodelling.
Safety evaluations of rMOv18 IgE were conducted at weekly doses of 5, 10 and 50 mg/kg and compared with rMOv18 IgG2b and PBS. Buffer alone-treated rats demonstrated a single toxicity of subdued but responsive behaviour (69%, mild score; Table S2).
Of the possible general and anaphylactic toxicities, only 3 categories were observed in rats treated with rMOv18 IgE or IgG2b.
These were as follows: subdued but responsive behaviour; piloerection; and hunching. All toxicities were scored as mild only, occurred immediately upon waking from anaesthetic, and lasted for 10-15 minutes thereafter. Severity was similar for both IgEand IgG2b-treated rats (Table S2, Figure 3A). Additionally, the number of toxicities experienced by each animal in each treatment group was similar between the IgE-and IgG-treated groups at each dose ( Figure 3B). None of the rats in any treatment group experienced any symptom that would constitute an anaphylactic response, according to a published anaphylaxis scoring system (Table S1B). 18 Histopathological evaluations of several tissues were conducted to determine potential off-target toxicities. No obvious pathological lesions associated with 5 or 50 mg/kg IgE, or IgG2b administration, were detected in major physiological systems (spleen, lymph nodes, liver, kidney, colon, brain) ( Figure 3C). IgE administration had no effect on body weight ( Figure 4A), serum blood chemistry or haematological parameters (creatinine, liver transaminases, haemoglobin and white blood cell counts) compared with PBS treatment ( Figure 4B-E). As we previously reported in sera of patients with FRa-expressing carcinomas, 14 in this model, we found circulating antibodies recognizing FRa, irrespective of treatment or antibody dose. We detected antidrug antibodies (ADA) in the serum of one of F I G U R E 1 Rat MOv18 IgE and IgG2b antibody engineering, purification and characterization. A, Schematic representations of the cloned rat rMOv18 IgE and IgG2b chimeric antibodies with mouse variable domains specific for the human folate receptor a (FRa) and rat constant domains (light green: human Ce, dark green: rat Ce, dark orange: human Cc, light orange: rat Cc, light grey: human Cj, dark grey: rat C L , purple: mouse V H , brown: mouse V L ). B, The heavy and light chains of each rMOv18 antibody were cloned into a single mammalian expression vector per antibody (pVITRO1). This resulted in the production of pVITRO-Rat MOv18 IgE and IgG2b. C, Purification of rMOv18 IgE with MEP HyperCel TM resin and 25% isopropanol elution buffer (MW: molecular weight marker, F1-F8: fractions 1-8). D, E, non-reducing SDS-PAGE (D) and HPLC-SEC elution profiles (E) of HPLC-purified hMOv18 IgE and IgG1, and rMOv18 IgE and IgG2b demonstrate intact monomeric antibodies of purities >85%. F, rMOv18 IgE and IgG2b bind to rat primary monocytes, RBL-2H3 rat mast cells, FRa-expressing rat colon adenocarcinoma CC531tFR cells and human ovarian carcinoma IGROV1 cells, but not to non-FRa-expressing human melanoma A375 cells. G, hMOv18 IgE shows weak binding, while rMOv18 IgE demonstrates significant binding to rat FceRI-expressing RBL-2H3 cells. Human and rat MOv18 IgE display similar binding to human FRa-expressing CC531tFR cells [Colour figure can be viewed at wileyonlinelibrary.com] ten rats treated with 10 mg/kg and in the serum of one of ten rats treated with 50 mg/kg IgE ( Figure 4F-G).
We concluded that administration of rMOv18 IgE to tumour-bearing rats induced no clinical, histopathological or metabolic signs of a type I hypersensitivity reaction, even at weekly 50 mg/kg doses.

| Immune cell infiltration and immune activation signatures associated with MOv18 IgE in vivo
IgE antitumour efficacy is likely to function by recruitment of host immunity. We therefore investigated whether inhibition of lung metastases in IgE-treated rats was associated with immune cell infiltration and activation. We measured growth inhibition of lung metastases with 4 (weekly) doses of rMOv18 IgE between 1, 10 and 50 mg/kg ( Figure 5A-C). Histological evaluation of rat lungs showed dose-dependent reduction in tumour islet density, alongside immune and stromal cell infiltration appearing progressively greater with increasing doses of IgE compared with lesions from PBS-treated rats ( Figure 5D).
We generated mRNA expression profiles from tumour-bearing rat lungs excised from animals treated with PBS or rMOv18 IgE. We employed Gene Set Enrichment Analysis (GSEA) to identify gene sets correlated with treatment. This yielded evidence of significant enrichment of gene expression associated with immune activation in IgE-treated tumours, compared with enrichment of cell cycle-associated pathways with vehicle treatment ( Figure 5E). The IgEenhanced, activated genes included those associated with IL-12 and NK cell immune activation pathways ( Figure 5F).  hunching and reduced activity). B, Percentage of animals within each treatment group to experience 1, 2 or 3 toxicities. C, Histopathological evaluations of tissues 30 days after administration of two doses (every 14 days) of rMOv18 IgE or IgG2b at 5 mg/kg and 50 mg/kg or PBS rats (n = 10 per group). Sections were stained with H&E and observed at brain = 25 9 magnification; colon, kidney, liver, mesenteric lymph nodes (LN), tail and thymus = 50 9 magnification; spleen = 100 9 magnification [Colour figure can be viewed at wileyonlinelibrary.com] F I G U R E 4 Absence of rMOv18 IgE-associated changes in body weight or blood haematological or biochemical parameters. A, Weekly weights of rats following weekly treatment with rMOv18 IgE at 1 mg/kg (n = 10), 10 mg/kg (n = 10) and 50 mg/kg (n = 10) or PBS (n = 10). B-E, Blood biochemical and haematological parameters were measured immediately before necropsy on day 30 following 4 weekly doses of rMOv18 IgE at 1 mg/kg (n = 10), 10 mg/kg (n = 10) and 50 mg/kg (n = 10) or PBS (n = 10). Dotted lines indicate normal range values for each parameter determined in healthy Wistar Albino Glaxo (WAG)/RijCrl. 24 F-G, Rat anti-human folate receptor a (FRa) antibodies and rat antidrug antibodies (ADA) measured in rat serum following weekly treatment with PBS, or rMOv18 IgE at 1 mg/kg, 10 mg/kg and 50 mg/kg (n = 10 per group) [Colour figure can be viewed at wileyonlinelibrary.com] F I G U R E 5 Rat MOv18 IgE induces immune cell infiltration and immune activating signatures. A, B, Percentage (%) tumour occupancy and number of metastases/cm 2 quantified following four (weekly) doses of PBS, or rMOv18 IgE at 1 mg/kg, 10 mg/kg and 50 mg/kg (n = 10 per group). C, Representative images of Indian ink-stained lungs from PBS-and rMOv18 IgE-treated rats (n = 10 per group). D, Representative images of paraffin-embedded lung sections from PBS-and rMOv18 IgE-treated rats stained with H&E (n = 10 per group). Magnification 200x. Black arrow: glandular tumour islet; white arrow: immune and stromal cells. E, GSEA analyses showing enrichment of immune-associated signatures (red) in rMOv18 IgE-treated (two doses, 2 weeks apart) tumour-bearing rat lungs and cell cycle-associated signatures (yellow) in PBS-treated tumour-bearing rat lungs. Other signatures are shown in blue. The y-axis corresponds to the normalized enrichment score (NES). F, Heat maps showing the differential expression of all genes that constitute the IL-12 and NK cells signature between rMOv18 IgE-and PBStreated rat lungs [Colour figure can be viewed at wileyonlinelibrary.com] We previously reported that TNFa-expressing mature macrophage subsets infiltrated tumour lesions and that elevated levels of TNFa were detected in lungs of rMOv18 IgE-treated rats compared with controls. 11 Here, we also observed significantly elevated circulating TNFa in IgE-treated compared with PBS-treated rats (P = .0044) ( Figure 6A-B). 11 Previously, we found elevated macrophage chemoattractant protein-1 (MCP-1) in the lung environment of IgE-treated animals potentiated by local upregulation of TNFa following IgE cross-linking on tumour-associated macrophages. Here, we measured detectable serum MCP-1 in IgE-treated rats, but these were not significantly elevated compared with controls ( Figure 6C).
On the other hand, concentrations of the classical "cytokine storm"related cytokines IL-6 and IFN-c were undetectable in rat sera from all groups. Furthermore, concentrations of IL-4, a key cytokine in the development of allergic inflammation, were undetectable in sera of IgE-treated and PBS-treated rats ( Figure 6D). These data are consistent with more prominent activation of immune cascades locally in tumours rather than systemically.
Together these findings point to immune cell recruitment, and activation of immunological pathways and mediators associated with F I G U R E 6 Rat MOv18 IgE triggers TNFa but no evidence of cytokine storm or allergic response. A, Cytokine/chemokine production measured in rat serum following weekly treatment with PBS, or rMOv18 IgE at 1 mg/kg, 10 mg/kg and 50 mg/kg (n = 10 per group). B-D, Concentrations of selected analytes in rat sera IgE treatment, in the absence of any evidence of overt toxicities, or of any cytokine profiles that would indicate signs of cytokine storm or allergic response activation.

| DISCUSSION
Identification of an appropriate animal model for studies involving IgE functions, including cancer immunotherapy, especially with regard to class-specific immunological efficacy and safety, represents a major hurdle in clinical translation of this antibody class. Previous studies have reported substantial evidence of comparable or superior antitumour efficacy of hMOv18 IgE compared to its IgG counterpart. [8][9][10]25,26 To translate this antitumour antibody safely into the clinic, studies conducted in a pharmacologically relevant species are required.
Here, we report findings in support of the in vivo safety of a tumour-targeting IgE in an immunologically relevant immunocompetent syngeneic rat tumour model, selected because of similarities in the structure and distribution of FceRI between humans and rats.
We describe upregulation of immune-associated pathways in lung metastases of IgE-treated rats, associated with antitumour efficacy, and evidence of elevated serum concentrations of the pro-inflammatory mediator TNFa. By contrast, neither the classical "cytokine storm"-related cytokine IL-6 nor the key allergic inflammatory cytokine, IL-4, was upregulated. Despite the activation of immunological pathways associated with IgE treatment, no evidence of systemic or organ-specific adverse events was detected.
We completed the rat model by generating surrogate rat IgE and IgG. We present a simple process for cloning heavy and light chains of any antibody clone into a single mammalian expression vector and report serum-free production by mammalian cells by scale-up in bioreactors. These processes are easily adaptable for generating antibodies of any isotype or specificity. Furthermore, the only currently reported purification method for monoclonal rat IgE is for a dinitrophenyl-specific rat IgE (DNP IgE), purified by an affinity column based on antigen specificity. 20 By contrast, our purification process for rat IgE regardless of antigen specificity using HCIC is also applicable for human IgE purification. Engineering, cloning and purification of intact, functional IgE of sufficient purity and yield for in vivo studies reported in the design of this rat model represents a methodology applicable to the generation of IgE and IgG antibodies of any specificity and species.
Administration of rMOv18 IgE and IgG2b to tumour-bearing rats demonstrated only three toxicities, observed with both isotypes.
Accordingly, none of the anaphylaxis-specific toxicities detailed in previously published criteria 18 were detected with any treatment.
Moreover, immunohistochemical toxicity studies on several tissues from IgE-, IgG2b-and PBS-treated rats revealed no evidence of pathological lesions, even at weekly 50 mg/kg antibody doses.
Together with our previously published data, these findings support clinical systemic MOv18 IgE administration. 14 We previously reported that human IgE cross-linked on the surface of human monocytes triggered monocyte activation and TNFa upregulation; furthermore, human IgE antitumour ADCC was associated with enhanced TNFa. In parallel, we detected TNFaexpressing macrophage subsets and elevated concentrations of TNFa in lung metastases of IgE-treated rats. 11 Consistent with these, here IgE treatment enhanced immune cell recruitment in rat lung tumours, and activation of immunological pathways in tumour-bearing lungs, alongside elevated TNFa in the sera of IgE-treated rats compared with controls. Although previously we measured elevated concentrations of MCP-1 alongside enhanced TNFa in rat bronchoalveolar lavage fluid of IgE-treated animals, here, we did not observe statistically significant elevation of serum MCP-1 in IgE-treated rats. This may be because IgE Fc-mediated cross-linking on tumour-associated macrophages can trigger elevated TNFa, stimulating local tumour cells to produce MCP-1 in the tumour, and not systemically. 11 It is possible that tissue-resident mast cells may also contribute to IgE-mediated enhanced TNFa expression, and this requires further investigation. 27 These suggest principally local induction of tumour-associated immune mediator release, immune cell activation and recruitment by IgE immunotherapy.
The immunocompetent syngeneic tumour model may also pose certain limitations, including lack of human FRa expression by normal rat tissues, and unlikely cross-reactivity of rat FRa homologue with rat MOv18 IgE. Although observed in the circulation of a proportion of patients with FRa tumours such as ovarian carcinomas and mesotheliomas, 28 we could not detect circulating shed human FRa (from CC531tFR cells) in rat sera in any of our studies (data not shown). Furthermore, FceRI-expressing cells in rats remain insufficiently characterized compared with human cells. Finally, laboratory rats, similar to mice, are not exposed to environmental allergens and thus may not display the array of allergen-reactive antibodies potentially found in human patients. In the light of these, other species such as canine patients who spontaneously develop FRa-expressing carcinomas may be interrogated. Considerable immunological homology between humans and dogs, alongside potential to better recapitulate the complexities of the microenvironment in spontaneously developed tumours in an immunocompetent species, may further benefit evaluations of antitumour IgE antibodies. 29 In conclusion, we designed and implemented an immunologically relevant and physiologically representative tumour model which supports the preclinical safety of an antitumour IgE antibody, MOv18 IgE.
Simultaneously, using this model, we report efficacy of the therapy in inducing tumour immune cell infiltration and immune pathway activation in the absence of physiological or immunological evidence of cytokine storm or allergic reaction. Our findings support the translation of IgE into the clinic and contribute to the rapidly growing field of AllergoOncology. Design of this and other immunologically congruent models may also help delineate IgE class-specific functions in allergic inflammation, allergen immunotherapy and antiparasitic responses.

ACKNOWLEDG MENTS
We acknowledge the Biomedical Research Centre (BRC) Immune Monitoring Core and Genomics Core Facility team at Guy's and St Thomas' NHS Foundation Trust for assistance.

S U P P O R T I N G I N F O R M A T I O N
Additional Supporting Information may be found online in the supporting information tab for this article.