Targeting mesothelin receptors with drug-loaded bacterial nanocells suppresses human mesothelioma tumour growth in mouse xenograft models

Human malignant mesothelioma is a chemoresistant tumour that develops from mesothelial cells, commonly associated with asbestos exposure. Malignant mesothelioma incidence rates in European countries are still rising and Australia has one of the highest burdens of malignant mesothelioma on a population basis in the world. Therapy using systemic delivery of free cytotoxic agents is associated with many undesirable side effects due to non-selectivity, and is thus dose-limited which limits its therapeutic potential. Therefore, increasing the selectivity of anti-cancer agents has the potential to dramatically enhance drug efficacy and reduce toxicity. EnGeneIC Dream Vectors (EDV) are antibody-targeted nanocells which can be loaded with cytotoxic drugs and delivered to specific cancer cells via bispecific antibodies (BsAbs) which target the EDV and a cancer cell-specific receptor, simultaneously. BsAbs were designed to target doxorubicin-loaded EDVs to cancer cells via cell surface mesothelin (MSLN). Flow cytometry was used to investigate cell binding and induction of apoptosis, and confocal microscopy to visualize internalization. Mouse xenograft models were used to assess anti-tumour effects in vivo, followed by immunohistochemistry for ex vivo evaluation of proliferation and necrosis. BsAb-targeted, doxorubicin-loaded EDVs were able to bind to and internalize within mesothelioma cells in vitro via MSLN receptors and induce apoptosis. In mice xenografts, the BsAb-targeted, doxorubicin-loaded EDVs suppressed the tumour growth and also decreased cell proliferation. Thus, the use of MSLN-specific antibodies to deliver encapsulated doxorubicin can provide a novel and alternative modality for treatment of mesothelioma.

Amatuximab and HN1 are both anti-MSLN mAbs with potential to be utilised for treatment of human malignant mesothelioma [38][39][40][41]. Amatuximab was developed as a high affinity chimeric mAb (K D 1.5 nM), that comprises murine anti-MSLN SS1 scFv (K D 11 nM) [39,42] and human constant regions [38]. The preclinical evaluation showed the efficacy of amatuximab in combination with gemcitabine or paclitaxel (Taxol) in a treatment of MSLN-positive xenografts [38]. A phase II clinical trial of amatuximab in combination with pemetrexed and cisplatin resulted in a disease control rate of 90% and improved the median overall survival in comparison to a phase III clinical trial of cisplatin and pemetrexed alone [15,43]. The HN1 scFv was isolated from a naïve, human scFv phage library and had an affinity of K D 100 nM [40]. The scFv was engineered as a human IgGγ1κ mAb, which improved the affinity 30-fold (K D 3 nM) [40].
In this study, we evaluated whether anti-MSLN BsAbs could be used to specifically deliver drug-loaded EDVs to mesothelioma and suppress tumour growth in mesothelioma xenografts (Fig 1). Therefore, we reformatted amatuximab and HN1 scFvs as BsAbs to target EDV to MSLN-positive tumours. The BsAb format utilized a glycine-serine (G 4 S) linker to connect two scFvs, whereby one scFv binds to the lipopolysaccharide antigen (LPS) on the EDV surface and the other binds to the target receptor, as reported by Taylor et al. [21]. The amatuximabderived BsAb (Amatux-BsAb) in conjugation with doxorubicin (Dox)-loaded EDV nanocells ( Amatux EDV Dox ) was capable of binding human mesothelioma cells in vitro, resulting in a release of doxorubicin intracellularly and induction of apoptosis. Additionally, in human malignant mesothelioma xenograft mouse models, treatment with Amatux EDV Dox led to statistically significant suppression of tumour growth, measured by both tumour size and a reduction in nuclear staining of excised tumour by Ki67, an antibody which recognizes a marker of cell cycle proliferation [44][45][46]. These results suggest that the Amatux-BsAb in conjugation with doxorubicin loaded EDV nanocells are effective despite using significantly lower doses of the drug, offering a potential new avenue for malignant mesothelioma-targeted therapy.

Protein expression of anti-MSLN-LPS BsAbs
Both anti-MSLN amatuximab [IMGT/mAb-DB ID 64] and HN1 [40] were reformatted as BsAbs each with an anti-LPS scFv to allow binding to MSLN and to LPS. The sequences were codon-optimized for expression in Chinese hamster ovary (CHO) cells. The sequence of each BsAb included a mammalian leader sequence from immunoglobulin kappa chain, followed by a 6x Histidine purification tag, then the anti-MSLN scFv sequence, a G 4 S linker, the anti-LPS scFv sequence and a C-terminal myc-tag for detection (Fig 2A) as described in Taylor et al. [21]. Both Amatux-LPS and HN1-LPS BsAbs were cloned into the mammalian expression vector pcDNA3.1(+) (Life Technologies), using HindIII and NotI restriction sites. CHO-S cells (Life Technologies) were cultured in CD-CHO with 8 mM GlutaMAX (Gibco) at 37˚C, 7.5% CO 2 with shaking at 130 rpm. On the day of transfection, cells were prepared at~3.0 × 10 6 cells/mL at 98% viability as measured using Cedex HiRes cell counter. The transient transfection for 200 mL culture was prepared as following: 600 μg of plasmid DNA was mixed with 15 mL OptiPRO serum free media (SFM) (Gibco) (15% total volume) and incubated at room temperature (RT) for 30-60 sec. Concurrently, 2400 μL of PEIpro (Polyplus-transfection) was mixed with 30 mL OptiPRO SFM (15% total volume) and incubated at RT for 30-60 sec. The two solutions were then mixed together by gently pipetting the PEIpro: OptiPRO complex into the plasmid DNA: OptiPRO complex, and incubated static at RT for 15 mins. The transfection complex was then added to the culture in a shake-flask and mixed by gentle swirling. The culture was incubated at 37˚C, 7.5% CO 2 with shaking at 130 rpm for 4 to 6 hrs, after which the culture was diluted 1:2 (v:v) with CD-CHO supplemented with final concentrations of EfficientFeed A and B (Gibco) 7.5% (v/v), 0.4% Anti-Clumping Agent (Gibco), 8 mM Glu-taMAX and then the flasks were kept shaking at 130 rpm at 32˚C, 7.5% CO 2 for 12 days. The supernatant was then harvested and 0.22 μm filtered.

Protein purification
Both His-tagged BsAbs were purified using immobilized metal affinity chromatography with a 5 mL HisTrap excel column (GE Healthcare). The column was equilibrated in 20 mM sodium phosphate pH 7.4, 500 mM NaCl. The harvested supernatant was loaded onto the column and then washed with equilibration buffer containing 20 mM imidazole to remove non-specifically bound contaminants. The protein of interest was then eluted with 20 mM sodium phosphate pH 7.4, 500 mM NaCl, 500 mM imidazole. The final product was desalted into PBS using HiPrep 26/10 column (GE Healthcare) and then filtered through a 0.22 μm syringe filter. The concentration of protein was determined by measuring the UV absorbance at 280 nm using a Nanodrop 1000 spectrophotometer. Protein purity was analysed by SDS-PAGE using 4-12% Bis-Tris gels (Invitrogen).

Binding analysis of BsAbs to human mesothelioma cells by flow cytometry
To evaluate the binding affinity of anti-MSLN BsAbs to MSLN receptors on the human lung mesothelioma cell lines H226 (ATCC, CRL-5826) and MSTO-211H (ATCC, CRL-2081), cells were grown in RPMI-1640 (Sigma) medium supplemented with 10% fetal calf serum (FCS; Bovogen) and Penicillin/Streptomycin (Sigma). Before flow cytometric analysis, the cells were washed twice with PBS and then scraped from T75 flask and centrifuged at 700 x g for 5 mins. The viable cells were counted by trypan blue exclusion method and 8x10 6 cells/mL were washed twice with cold PBS. A centrifugation step at 700 x g was conducted in between the washes. The sample was divided into five vials (2x10 6 cells/sample) prior to the final centrifugation step. The pellets were resuspended in 500 μμL Binding Buffer (PBS containing 0.1% sodium azide and 5% bovine serum albumin (BSA)) and incubated for 10 mins to block nonspecific binding sites. Various primary antibodies (4 μg) were added to each tube as follows: (1) Amatux-BsAb, (2) HN1-BsAb, (3) ABX-EGF-BsAb (anti-EGFR BsAb described in [21]), (4) Anti-CD3-BsAb (as a negative control). The tubes were incubated for 45 mins on ice on a gentle shaker. Then the cells were washed twice with Binding Buffer. The cells were incubated with the secondary antibody, anti-cMyc-FITC (Miltenyi Biotec), diluted 1 in 250 in Binding Buffer for 45 mins on ice on a shaker (moderate movement). Afterwards, the cells were washed three times with Binding Buffer, and re-suspended in 400 μL PBS for flow cytometric analysis on a Beckman Coulter, Cytomics CXP500 flow cytometer using the 488 nm laser and the 530/30 nm filter. The data were analysed using FCS Express Flow version 4 (De Novo Software).

Preparation of targeted nanocells
EDVs were manufactured and purified using a cross-flow filtration and lyophilized as previously described [20]. Each lyophilized vial contained 1.1x10 10 EDVs and the lyophilized particles were resuspended in 600 μL of solvent resulting into 1.8x10 10 EDVs/mL. Doxorubicin loading into EDV nanocells to create EDV Dox was carried out as previously described [20]. Additionally, for some experiments EDV nanocells were labelled with Alexa-Fluor 488 (AF-488) (ThermoFisher) as previously described [21]. Drug-loaded or AF-488 labelled EDVs were then incubated with Amatux-BsAb at RT for 30 mins while shaking at 300 x g, to create Amatux-EDV. Samples then underwent 3x PBS spin wash cycles at 9,000 x g for 8 mins each to remove any excess Amatux-BsAb. Amatux EDV nanocells were resuspended to concentrations as required for the experiment.
Internalization analysis of drug-loaded BsAb EDV nanocells into H226 cells by confocal microscopy H226 cells were incubated on coverslips with AF-488 labelled Amtaux EDV Dox , at a ratio of 10,000 EDV nanocells per cell. Cells without any treatment were used as a control. Cells were incubated at 37˚C for 3 hrs and 24 hrs which was followed by 3 washes in sterile PBS. Cells were fixed with 4% paraformaldehyde for 10 mins, followed by 3 times PBS washes. After the final wash, 500 μL PBS was added to each well to cover the coverslips completely. An anti-EGFR mAb (528 mAb) was labelled with x-site fluor 670/755 (Kodak) and used as a cell membrane stain; 4 μg was added directly to each sample. Wells were mixed gently and left for 10 mins to allow cells to stain sufficiently. All coverslips were again washed 3 times with PBS followed by a final wash in Milli-Q water and then mounted on clean slides using Fluka Eukitt (Sigma) quick-hardening mounting medium. Cells were imaged using a Carl Zeiss 710 Inverted Laser Scanning Confocal Microscope housed at the Australian National Fabrication Facility, Queensland Node (ANFF) using a plan-apochromat 63x 1.40 Oil DIC M27 objective and ZEN 2008 software for image formatting. The AF-488 labelled EDV nanocells were detected using ex488/em525 nm. Doxorubicin is auto-fluorescent and detected using ex490/ em560 nm. Cell surface EGFR was detected with ex633/em670 nm.

Binding analysis of Amatux EDV Dox on human mesothelioma cell by flow cytometry
To evaluate binding of Amatux EDV Dox to MSLN receptors on H226 and MSTO-211H cells, lyophilised EDV Dox nanocells were reconstituted, loaded with doxorubicin, and then labelled with Amatux-BsAb or with anti-Murine Epithelial Cell Adhesion Molecule (mEpCAM)-BsAb as a negative control. The BsAb EDv Dox complexes were incubated rotating for 1 hr at RT, before washing three times and resuspending in 500 μL of PBS. H226 and MSTO cells were grown as previously described. Cells were washed twice with PBS and then detached using Accutase (Sigma) for 10 mins at 37˚C and centrifuged at 300 x g for 5 mins. Viable cells were counted by trypan blue exclusion method on a hemocytometer and 1x10 7 cells/mL were resuspended in serum free media. The cell suspension was divided into three vials to be treated with: (1) Non-targeted EDV Dox , (2) mEpCAM EDV Dox , and (3) Amtaux EDV Dox . EDVs (5x10 9 EDV/sample) were added to 5x10 5 cells (1:10,000 cell:EDV ratio) and incubated rotating at 4˚C. After 2 hrs, unbound EDVs were removed by washing the cells with PBS+1%BSA at 300 x g for 5 mins. Cells were fixed with 4% paraformaldehyde for 10 mins at RT, followed by a PBS+1%BSA wash. Bound EDVs were stained with 1 μg of anti-LPS mAb (Biodesign International) labelled with AF-488 (Life Technologies) for 45 mins at RT in the dark. The cells were washed three times with PBS+1%BSA at 300 x g for 5 mins, before analysis on a Beckman Coulter, Gallios flow cytometer using the 488 nm laser and the 530/30 nm filter. The data were analysed using FCS Express Flow version 4 (De Novo Software).
A flow cytometry dose-response of Amatux EDV binding to the aforementioned mesothelioma cell lines, and the murine colon carcinoma CT26 cells (ATCC, CRL-2638), was performed in similar manner, but without loading the EDVs with doxorubicin. The cell suspensions were treated with different amounts of Amatux EDV (1:1000, 1:2500, 1:5000 and 1:10,000 cell: Amatux EDV ratios). EDVs charged with anti-CD3 BsAb were also included as negative controls when we incubated the cell suspensions with constant concentration of CD3 EDV (1:10,000 cell: CD3 EDV). The data were analysed using FCS Express Flow version 4 (De Novo Software).

Cytotoxicity of Amatux EDV Dox nanocells
Induction of apoptosis was demonstrated by staining cells with Annexin V-FITC after incubating MSLN-positive H226 cells with Amatux EDV, Non-targeted EDV Dox and Amatux EDV Dox . H226 cells were grown in DMEM (Sigma) with 10% FCS (Bovogen) and 1% Penicillin/Streptomycin at a seeding density of 3x10 5 cells per well in 6 well plates. The cells were incubated with Amatux EDV, Non-targeted EDV Dox and Amatux EDV Dox at a ratio equivalent to 10,000 EDVs per cell, for 2 h at 37˚C to allow EDVs to bind to cells. Then the cells were washed 3 times with PBS to remove unbound EDVs, before adding fresh media and incubating at 37˚C for 72 hrs to allow EDV internalisation and doxorubicin release. Washing the unbound particles is necessary, because if left in a well more than 24 hrs, they will start to leak out the doxorubicin into media [20,22]. After 72 hrs, the media were collected and the cells were detached with trypsin-EDTA and collected in media. The cells were washed twice in PBS, then resuspend in 1x Binding Buffer at 1x10 6 cells in 1 mL. Each sample was divided equally (500 μL) into different tubes, with one treated with 5 μL of Annexin V-FITC while the other was treated with 5 μL of Binding buffer as a control for doxorubicin interference by autofluorescence. The tubes were incubated at RT for 10 min before analysing the fluorescence of the cells on a Beckman Coulter, Cytomics CXP500 flow cytometer, using the 488 nm laser and the 530/30 nm filter. Data were analysed using FCS Express 4 Flow cytometry (De Novo Software).

Effect of Amatux EDV Dox nanocells on tumour regression
The experiments associated with the effect of Amatux-targeted nanocells on tumour regression were performed essentially as described by Taylor et al. [21]. H226 mesothelioma cells were grown in RPMI with 10% FCS and penicillin/streptomycin. Six week-old female Balb/c mice were injected subcutaneously on the left flank with 5×10 6 cells in 200 μL of media together with 200 μL of growth factor reduced matrigel (BD Biosciences). Once the tumours reached 110 mm 3 (Day 11) the mice were randomized into treatment groups which included 5-6 mice per group, as follows: Group 1 -Saline, Group 2 − Amatux EDV, Group 3 − Non-targeted EDV Dox , and Group 4 − Amatux EDV Dox . Each mouse was intravenously injected with 1×10 9 of the relevant EDVs administered 3 times per week for 3 weeks, and tumour size was monitored throughout the treatment period up to Day 39 post xenograft. The mice were weighed twice a week for 3 weeks, as per previous methodology [20,27]. A two-sample t-test with equal variance was performed to directly compare the mean tumour volume from each individual treatment to determine statistical significance between treatments and control. Statistical significance was determined as p 0.05 [47]. The mice were euthanased after a total of 12 injections, tumours were harvested, paraformaldehyde fixed and embedded in paraffin. Note; Experiments were conducted in compliance with the Australian National Health and Medical Research Council guidelines for the care and use of laboratory animals and with the approval of the EnGeneIC Animal Ethics Committee.

Ki67 cell staining and quantification by immunohistochemistry
Paraffin embedded human mesothelioma tumour tissue section (5 μm), prepared from tumours excised from the four xenograft treatment groups at Day 39 (Saline, Amatux EDV, Non-targeted EDV Dox , Amatux EDV Dox ) were dewaxed in xylene, rehydrated in graded alcohols and deionized in distilled water. Heat-induced antigen retrieval (HIER) was performed by immersing the slides for 20 mins in 0.01 M Citric Acid-Based Buffer, pH 6.0 (Vector), at 100˚C in a microwave, followed by cooling for 30 mins at RT. The endogenous peroxidase activity was quenched by incubating the slides with methanol containing 1% hydrogen peroxide for 30 min at RT in the dark. Then the slides were washed using PBS with 1% Tween20 (PBS-T) for 5 min and then blocked with 5% goat serum in PBS-T for 30 min at RT. Rabbit anti-human pKi67 antibody (Sino Biological, US) was added to the slides at 20 μg/μL. PBS substituted the primary antibody for the control slide. The slides were incubated overnight at 4˚C in a covered humidity chamber. On the next day, the slides were washed in PBS-T and the HRP-conjugated secondary antibody goat anti-rabbit IgG (Life Technologies) was diluted in 5% goat serum in PBS-T at 1:200 before it was incubated with the slides for 2 hrs at RT. Then the slides were washed 3 times in PBS-T, then incubated with AEC (3-amino-9-ethylcarbazole) substrate solution (Vector) until the desired colour developed (10 min). The slides were counterstained with hematoxylin (Blue) for 30 sec and mounted with VectaMount AQ (Vector). The percentage of Ki67 positive cells was estimated by blind counting of Ki67 positive cells (red nuclei) in two different Ki67 expressing regions consisting of 1000 cells in total using 20x magnification. The percentage was estimated from three different mice representing each group. Images were captured using Aperio digital pathology slide scanner and analysed using Aperio ImageScope software (Leica Biosystems). The mean values for the percentage of Ki67 positive cells for all treatments were analysed for variance using one-way ANOVA with Tukey post-hoc test was performed to directly compare the means from each individual treatment to determine if there is any significant difference between treatments and control. Statistical significance was determined as p 0.05 [47].

Hematoxylin and Eosin (H&E) staining for mesothelioma paraffin embedded tumour samples
Paraffin embedded human mesothelioma tumour tissue sections (5 μm), prepared from tumours excised from the four xenograft treatment groups at Day 39 (Saline, Amatux EDV, Non-targeted EDV Dox , Amatux EDV Dox ) were dewaxed in xylene, rehydrated in graded ethanol solutions. Tissue samples were then stained for 7 min in Mayers Haematoxylin and 20 sec in 1% Eosin. Samples were then dehydrated through graded ethanol solutions and cleared with xylene. Finally, slides were coverslipped using DPX solution (Sigma). Samples were visualised using an upright bright-field slide scanner (Zeiss Axio Imager Z2). The percentage of necrotic areas was estimated by blind measuring of multiple scattered necrotic regions (absence of nuclear staining) using 20x magnification. The percentage was estimated from three different mice representing each treatment group. Images were captured using a Zeiss Axio Imager Z2 slide scanner and analysed using ImageJ software v1.50i (National Institutes of Health, Bethesda, MD). The mean values for the percentage of necrotic regions for all treatments were analysed for variance using one-way ANOVA with Tukey post-hoc test. Statistical significance was determined as p 0.05 [47].

Production of anti-MSLN BsAbs
The BsAbs derived from amatuximab and HN1 were expressed in CHO cells followed by IMAC purification using HisTrap excel affinity chromatography column. SDS-PAGE analysis showed molecular weights of 56 kDa for both BsAbs (Fig 2B). Binding specificity between purified MSLN and purified LPS and the BsAbs was confirmed by conducting indirect ELISA on immobilised MSLN and LPS and detection with HRP-conjugated anti-myc antibody (data not shown).

Binding activity of anti-MSLN BsAbs toward human mesothelioma cells
Flow cytometry was used to assess the binding of Amatux-BsAb and HN1-BsAb to mesothelioma cell lines H226 and MSTO-211H. ABX-EGF-BsAb specific for EGFR [21] was used as a positive control, since malignant mesothelioma frequently shows over-expression of EGFR in either cell lines or the tumour itself [29]. We observed that Amatux-BsAb and ABX-EGFR-B-sAb bound strongly to the H226 cells, whereas HN1-BsAb showed no binding (Fig 2C). Both Amatux and HN1-BsAbs showed no binding to MSTO-211H cells (Fig 2D). Amatux EDV Dox nanocells cellular uptake BsAb EDV Drug nanocells are capable of being internalized by cancer cells and releasing cytotoxic drug intracellularly [22]. Here, we prepared AF-488 labelled Amatux EDV Dox nanocells to monitor and visualize binding and internalisation of the targeted EDV system to H226 cells at the cellular level by confocal microscopy. Within a 3 hr time-frame, the Amatux EDV Dox nanocells were co-localized on the cell surface, demonstrated by green fluorescence staining on the H226 cell membrane (Fig 3A). After 24 hrs of incubation, the AF-488 labelled Amatux EDV Dox nanocells migrated and accumulated within the H226 cells and released doxorubicin intracellularly, demonstrated by purple staining representing doxorubicin autofluorescence (Fig 3B). To confirm that the Amatux EDV nanocells binding to H226 cells was Amatux-BsAb-dependent, we used flow cytometry to test binding of Non-targeted EDV Dox , mEpCAM EDV Dox and Amatux EDV Dox to H226 and MSTO-211H cell lines . Only Amatux EDV Dox nanocells were able to bind to the MSLN on H226 cells, but did not bind to the low MSLN expressing cell line MSTO-211H (Fig 3C and 3D).
A dose-response shift was generated when H226 mesothelioma cells were treated with different concentrations of Amatux EDV (Fig 3E). The intensity of the binding shift was concentration-dependent, wherein the strongest shift was observed when the H226 cells were treated with 10,000 Amatux EDV per cell, while a gradual reduction of the binding intensity was observed each time we reduced the Amatux EDV:cell ratio by half (Fig 3E). Under the same conditions, Amatux EDV showed no binding to both MSTO-211H and murine colon carcinoma CT26 cells (MSLN negative cell line) (Fig 3F and 3G). Also, as a negative control we included 1:10,000 cell: CD3 EDV, which showed no binding to the three cancer cell lines.

Cytotoxicity of Amatux EDV Dox nanocells
Apoptosis causes externalization of phosphatidylserine from the inner leaflet of the plasma membrane phospholipid bilayer to the cell surface, which can be probed with Annexin V for detection, while living cells remain Annexin V negative [48]. Therefore, we stained Amatux EDV, Non-targeted EDV Dox and Amtaux EDV Dox treated cells with Annexin V-FITC and performed flow cytometry to evaluate the induction of apoptosis. Approximately 37.7% of Amatux EDV Dox targeted H226 cells showed Annexin V positive staining, which was around 9-fold greater than that of Amatux EDV and Non-targeted EDV Dox samples (Fig 4). The same samples were also analysed in the absence of Annexin V; the Amatux EDV Dox sample showed a small increase in fluorescence attributed to the autofluorescence of doxorubicin, which has a comparable emission wavelength to FITC, yet the shift remained within the Annexin V negative gate (data not shown). These results suggest that Amatux EDV Dox nanocells are capable of targeted killing of MSLN-expressing cells. Consequently, tumour regression studies were performed using Amatux EDV Dox nanocells in a H226 cell line xenograft model.

Amatux EDV Dox nanocells mediate anti-tumour effect in human mesothelioma xenograft models
Eleven days after tumour subcutaneous-implantation when the volume of the tumours reached~110 mm 3 , mice were randomly sorted into four different groups (n = 6 per group): saline control group, Amatux EDV, Non-targeted EDV Dox and Amatux EDV Dox . By the end of the treatment period (Day 39), all the mice in saline, Amatux EDV and Non-targeted EDV Dox treatment groups had comparable tumour volumes (p > 0.1), while the Amatux EDV Dox treated group showed significant tumour regression (p < 0.001) with the tumour volume being 3-fold less than all other treatment groups (Fig 5A). Within the Amatux EDV Dox treated group, one mouse showed a partial response (!30% tumour regression) during the whole study period. Three mice showed transient partial response followed by slight tumour growth and then tumour stasis (no tumour growth or progressive disease). One mouse demonstrated stable tumour size during the whole treatment course (data not shown). The Non-targeted EDV Dox did not evoke any significant anti-tumour effect through passive targeting. Thus, active targeting through BsAbs is essential for the efficient delivery of EDV nanocells to tumours.
All mice survived throughout the treatment period, except one mouse in the Amatux EDV Dox group that died of unknown cause(s) on the first day of the experiment, thus n = 5 in subsequent analysis of data for this group. Physical observation of all the mice showed no abnormalities in behaviour nor physical condition, and there was no difference in weight throughout Targeting mesothelin receptors with drug-loaded bacterial nanocells the treatment period between the groups of mice (Fig 5B). This suggests the doxorubicinloaded EDV nanocells are well tolerated despite the high and repeated dose regimen. Targeting mesothelin receptors with drug-loaded bacterial nanocells Amatux EDV Dox nanocells mediate anti-proliferative effect and reduce necrosis in human mesothelioma xenograft models The nuclear protein Ki67 is a proliferation marker and an independent prognostic factor in malignant mesothelioma that can be utilised to evaluate the effectiveness of a chemotherapy regimen [49,50]. The Ki67 biomarker was therefore utilised to detect the proliferation rate differences between the four treatment groups. The Ki67 immunohistochemical staining of tissue sections from excised xenograft tumours from the Amatux EDV Dox treated group showed a significant decrease in the number of proliferating cells, compared to the saline (p < 0.01), Amatux-EDV (p < 0.01) and Non-targeted EDV Dox (p < 0.05) treated groups (Fig 6A and 6B). Using H&E staining, tissue sections from saline treated mice revealed larger areas of necrosis (12.52%) compared to Amatux EDV (10.9%), Non-targeted EDV Dox (9.5%) and Amatux EDV Dox (5.65%) treatment groups (Fig 6C and 6D). Statistically, the Amatux EDV Dox treated mice had significantly Targeting mesothelin receptors with drug-loaded bacterial nanocells smaller necrotic areas compared to the saline group, Amatux EDV and Non-targeted EDV Dox (p 0.0005, p < 0.01 and p < 0.05, respectively).

Discussion
Human malignant mesothelioma is an intractable tumour that develops from the mesothelial cells mainly as a result of asbestos exposure. Challenges exist for current mesothelioma treatment protocols due to the lack of specificity of cytotoxic drugs and their low therapeutic index. To address these challenges, we targeted drug-loaded bacterial EDV nanocells to MSLN receptors, which are overexpressed in mesothelioma, to deliver doxorubicin directly to the mesothelioma cells. This was achieved by engineering the MSLN-specific mAb amatuximab as a bispecific antibody, whereby one arm binds the EDV and the other arm binds to tumour cells.
Here, we showed the ability of Amatux EDV Dox to suppress the growth of human malignant mesothelioma in xenograft models. EDV nanocells are derived from Salmonella typhimurium, and have proven to be effective, safe and well-tolerated in different species [20, 23-25, 27, 51]. This study builds on a body of published research outcomes including clinical research, associated with a nonliving, bacterially-derived drug nanodelivery system [20,23,25,27]. The efficiency of BsAb EDV Drug nanocells as a system that is able to suppress tumours expressing MSLN, is demonstrated.
We assessed the binding of Amatux-LPS and HN1-LPS BsAbs (anti-MSLN-LPS BsAbs) by performing flow cytometry analyses using two mesothelioma cell lines (H226 and MSTO-211H). Amatux-BsAb was able to bind to H226 cells but not to MSTO-211H while HN1-BsAb was unable to bind either cell line (Fig 2C and 2D). This may be a reflection of the lower affinity of the parent HN1 scFv (K D 100 nM) compared with the amatuximab parent scFv (K D 11 nM), and therefore we excluded HN1 scFv from the remainder of experiments. Both BsAbs showed no binding to MSTO-211H; this is most likely due to the low abundance of MSLN on MSTO-211H cells compared with H226 cells as previously reported [52].
scFvs typically demonstrate lower affinity compared to the original bivalent antibody due to a loss of avidity [21,[53][54][55], but binding of BsAbs to multiple LPS sites on EDVs, forms an antibody corona, creating a highly avid BsAb EDV nanocarrier system [20]. This is evident by the greater binding of the Amatux EDV Dox to H226 cells compared with the Amatux-BsAb ( Fig  3C compared with Fig 2C).
In this study, we showed that BsAb EDV Drug nanocells are able to specifically target tumour cells, internalise, release drugs intracellularly and induce apoptosis in vitro. Additionally, we showed that treatment of mice containing mesothelioma cell xenografts with BsAb EDV Drug results in suppression of tumour growth. Previously, the BsAb EDV Drug nanocells have been shown to effect significant tumour growth regression in various in vivo models [20,21,27,56,57]. Various chemotherapeutic agents, siRNA and miRNA can be efficiently packaged within the EDV nanocells and delivered to cancer cells [20,27,51]. These nanocells have a greater capacity to package drugs (<10 million drug molecules) [20], when compared with other systems such as liposomes and antibody-drug conjugates, which are only able to be packaged with~10,000 drug molecules per liposome and <10 drugs molecules per antibody, respectively [58]. Packaging of drug molecules within EDV nanocells allows a significantly lower overall dose compared to systemic drug delivery and as a consequence less potential side effects. For example, the doxorubicin dose was reduced around 1764-fold with CD3 EDV Dox compared to that of conventional chemotherapy treatment [20].
Rapidly growing malignant tumours are characterised by poor and fragile vascularisation that results in low oxygen levels which leads to necrosis; thus large, malignant tumours have greater levels of necrosis [59][60][61]. In mesothelioma, necrosis is a sign of tumour growth [62,63]. We were able to identify necrotic areas in xenograft tumour sections, and observed that tumours from mice treated with Amatux EDV Dox showed significantly smaller areas of necrosis compared to Non-targeted EDV Dox (p < 0.05). This likely reflects the difference between the two treatment groups in terms of tumour size, with the larger tumours showing more necrosis. Thus, targeted delivery of encapsulated doxorubicin to mesothelioma cells caused tumour growth retardation leading to smaller necrotic areas compared to the other treatment groups. Combined with the results obtained from staining the proliferation marker Ki67, which showed that cells within the tumour sections from Amatux EDV Dox -treated mice had significantly lower proliferation, we have demonstrated that Amatux EDVs can be used to target mesothelioma cells and deliver therapeutic doses of doxorubicin, resulting in reduced cell proliferation, lower cellular necrosis and suppression of tumour growth.

Conclusions
We showed that the full-length antibody amatuximab can be engineered into a BsAb format, which is capable of binding doxorubicin-loaded nanocells, forming Amatux EDV Dox . These targeted nanocells are functional and able to bind MSLN receptors on the surface of H226 cells in vitro and deliver doxorubicin intracellularly, resulting in the induction of apoptosis. We also showed that the new drug-delivery system Amatux EDV Dox is able to deliver and release doxorubicin in mesothelioma xenografts and suppress the growth of malignant cells. These results indicate that MSLN expression is retained in implanted xenografts and the protein is accessible by the BsAb EDV. MSLN abundance on tumour cells is sufficient for efficient antibody targeting of doxorubicin-loaded EDVs to xenografts, resulting in suppression of tumour growth as well as reducing cell proliferation as shown by Ki67 staining of excised tumours ex vivo. This proofof-concept animal study provides essential data that the anti-MSLN antibody and the EDV drug delivery system are functional in vivo and warrants further translational research of this treatment regime. This is the first investigation associated with usage of MSLN as a target molecule for mesothelioma treatments using doxorubicin-loaded EDV nanocells. Further studies investigating dose optimization, biodistribution and survival rates will be required to obtain a better understanding of this drug delivery system and compare its efficacy to existing malignant mesothelioma treatments. Targeting mesothelin receptors with drug-loaded bacterial nanocells Collaborative Research Infrastructure Strategy to provide biologics, nano-and micro-fabrication facilities for Australian researchers. This project also utilised EnGeneIC Ltd facilities for binding assays and xenografts studies.