Mesothelin (MSLN) is a differentiation antigen that is highly expressed in many epithelial cancers. MSLN is an important therapeutic target due to its high expression in cancers and limited expression in normal human tissues. Although it has been assumed that shed antigen is a barrier to immunotoxin action, a modeling study predicted that shed MSLN may enhance the action of MSLN-targeting recombinant immunotoxins such as SS1P and similar therapeutics by facilitating their redistribution within tumors. We aimed to determine whether shed MSLN enhances or reduces the antitumor effect of MSLN-targeting immunotoxins SS1P and RG7787. We engineered a cell line, A431/G9 (TACE mutant) that expresses a mutant form of MSLN in which the TNF-converting enzyme protease site is replaced with GGGS. We compared the response of the TACE-mutant cells with immunotoxins SS1P and RG7787 with that of the parental A431/H9 cell line. We show that TACE-mutant cells shed 80% less MSLN than A431/H9 cells, that TACE-mutant cells show a 2- to 3-fold increase in MSLN-targeted immunotoxin uptake, and that they are about 5-fold more sensitive to SS1P killing in cell culture. Tumors with reduced shedding respond significantly better to treatment with SS1P and RG7787. Our data show that MSLN shedding is an impediment to the antitumor activity of SS1P and RG7787. Approaches that decrease MSLN shedding could enhance the efficacy of immunotoxins and immunoconjugates targeting MSLN-expressing tumors. Mol Cancer Ther; 15(7); 1648–55. ©2016 AACR.

Mesothelin (MSLN) is a cell surface protein that is highly expressed in mesothelioma and many other epithelial cancers. Furthermore, small amounts of MSLN have been observed in the blood of patients with mesothelioma and ovarian cancer (1). MSLN has limited expression on normal tissues; its expression is limited to mesothelial cells lining the peritoneum, pericardium, and pleura (2–5). MSLN does not appear to have a critical function in development, because inactivation of both copies of the MSLN gene in mice does not lead to an overt phenotype (6). However, it does appear to have a role in cancer growth and progression making it an attractive therapeutic target for antibody-based therapies of cancer (7, 8).

Antibody-based therapies such as antibody–drug conjugates and recombinant immunotoxins (RIT) are being developed for the treatment of solid tumors (9). However, penetration of these agents into solid tumors is restricted by many factors. Impediments include defective capillaries, slow entry by diffusion because of the lack of functional lymphatics, increased extracellular matrix deposition (10), and high interstitial fluid pressure (11). It has also been proposed that shed antigen within tumors is another barrier to tumor penetration (12). Overcoming these barriers could significantly improve the efficacy of antibody-based therapies.

There are several anticancer agents in various stages of development that target MSLN. These include amatuximab (13, 14), antibody–drug conjugates (15, 16), and RITs (17–20). RITs are chimeric proteins composed of an Fv or Fab fused to a portion of Pseudomonas exotoxin A (PE). SS1P is a RIT in which the disulfide-stabilized Fv of a high-affinity antibody to MSLN is fused to a 38-kDa fragment of PE (17). SS1P kills target cells by inhibiting protein synthesis and activating apoptosis (18, 21). SS1P is very active in killing MSLN-expressing cancer cell lines (17) and has shown antitumor activity in mouse xenograft models (22, 23). In human studies, SS1P has been found to enhance the antitumor activity of chemotherapy agents, gemcitabine, and pemetrexed/cis-platinum (23, 24). When combined with immunosuppressive agents, pentostatin, and cyclophosphamide, it produced major remissions in several patients with advanced chemotherapy-refractory mesothelioma (25). A new anti-MSLN RIT, RG7787 has been recently developed. RG7787 is a clinically optimized anti-MSLN immunotoxin (20). It is designed to be less immunogenic than SS1P. It also contains mutations that remove or suppress many B-cell epitopes and some T-cell epitopes and is very cytotoxic to several MSLN-expressing cancers (19)

MSLN is shed into the extracellular space of tumors and its levels are elevated in the blood of patients with mesothelioma and ovarian cancer (26). Antigen shedding is a well-recognized process associated with malignant cells (27). Previous studies from our group have suggested that shed MSLN within the tumor is a barrier to effective therapy with SS1P, because treatment of MSLN-expressing tumors with chemotherapy lowered MSLN levels in the tumor, which allowed more SS1P to reach the tumor cells and enhanced the antitumor activity of SS1P (10, 11). We have also shown that MSLN is released from cells by the action of the TNF-converting enzyme (TACE) protease (28) and that inhibition of the protease by an inhibitor or knockdown of the TACE protease by siRNA diminishes MSLN shedding and enhances immunotoxin killing of cultured cells (28). These data suggest that shed MSLN should act as a barrier to the killing of cells by SS1P or related immunotoxins.

Pak and colleagues (29) used our data to develop a mathematical model to analyze immunotoxin entry into tumors and cell killing within the tumor to determine which steps in immunotoxin entry are affected by shed MSLN. These calculations predicted that high concentrations of MSLN should not interfere with SS1P action, but instead could promote the redistribution of MSLN within solid tumors and enhance immunotoxin activity.

This study was designed to establish the role of MSLN shedding in the response of tumors to SS1P and RG7787. To examine the effect of MSLN shedding, we constructed a cell line with reduced shedding of MSLN due to a specific deletion of the TACE protease site in MSLN. We find that cells that shed less MSLN are more sensitive to MSLN killing in culture and that tumors derived from these cells shed less MSLN into the blood and are much more sensitive to the antitumor effects of SS1P and RG7787.

Reagents

SS1P was purified in our laboratory as described previously (30). Clinical grade RG7787 was manufactured by Roche Innovation Center. Alexa-labeled RG7787 was made using an Alexa labeling kit from Invitrogen. HB21(Fv)-PE40 immunotoxin was made in our laboratory. Anti-MSLN (mouse) was purchased from Rockland Immunochemicals (code# 200-301-A88), GAPDH (rabbit mAb 14C10; Product #2118) from Cell Signaling Technology, goat anti-mouse IgG-HRP antibody from Santa Cruz Biotechnology (cat. # sc-2004), goat anti-mouse IgG-R-PE from Jackson Immunoresearch (code: 115-116-146), QuantiBrite PE Beads from BD Biosciences, and Cell Counting Kit-8 from Dojindo Laboratories.

Cell lines

A431/H9 wild-type cells containing wild-type MSLN (GenBank Accession number NM_005823) were generated in our laboratory by Dr. Mitchell Ho (Laboratory of Molecular Biology, NIH, Bethesda, MD). Cells were tested and authenticated on August 15, 2014 using a short tandem repeat multiplex assay by a molecular detection group (Frederick National Laboratory). Cell line A431/G9 was authenticated in March 2016.

Construction of plasmid-expressing mutant MSLN

The expression plasmid pMSLN del (586–593) was generated by the Protein Expression Laboratory (Frederick Core Facility at NCI) using Gateway cloning technology. To try and maintain the native domain structure, the TACE cleavage site (YLVLDLSV) was replaced by a small flexible linker (GGGGS). The resulting clone was then used to generate the mammalian expression vector with a CMV promoter and a neomycin selectable marker.

Generating A431 cells expressing mutant MSLN

On day one, 5 × 105 A431 cells were added to 2 mL of medium in 6-well plates. The next day, cells were transfected with 5 μg of mutant MSLN DNA and control DNA with Lipofectamine as described in the Life Technologies' transfection protocol. After transfection, mixture was added to cells, cells were placed in the incubator for 5 hours, and then 1 mL of fresh media were added. Cells were cultured for 3 days and then transferred into a 10-cm dish containing media with G418 selection reagent (750 μg/mL). Cells were cultured until confluence and then trypsinized for flow cytometry analysis to identify positive expressers. Positive expressers were sorted and plated in 96-well plates at 1 cell per well to grow single clones. Cells were cultured for 2 weeks after which 15 single clones were selected for further analysis. After single clones were confluent, cells were trypsinized and flow cytometry analysis was used to select a clone with comparable MSLN expression to A431/H9 control cells.

MSLN surface expression

To quantify surface MSLN expression, 5 × 105 cells were trypsinized and harvested. Cells were then centrifuged at 2,000 rpm for 2 minutes, followed by washing once with cold PBS. Next, cells were washed once with cold FACS buffer (PBS with 5% FBS and 0.02% NaN3). After centrifuging, cells were resuspended in 100 μL of FACS buffer containing 5 μg/mL of mouse anti-human MSLN antibody (Rockland Immunochemicals, Inc.) on ice for 1 hour. After 1-hour incubation, cells were washed once with FACS buffer and centrifuged for 2 minutes at 2,000 rpm. Cells were resuspended in 100-μL FACS buffer containing anti-mouse IgG-R-PE (1:250) secondary Ab on ice for 30 minutes. Cells were washed twice with FACS buffer and then resuspended in 500 μL of FACS buffer and processed for flow cytometry (31). Fluorescence intensity was analyzed on a FACSCanto II. QuantiBRITE R-PE beads (Becton Dickinson, cat #340495) were used as a standard to quantitate the number of MSLN sites per cell.

Surface staining using labeled RG7787 was described previously (32). Briefly, 2.5 × 105 cells were plated in 6-well Falcon tissue culture plates in 2-mL media. The next day, cells were incubated in 500 μL of complete media containing 2 μg of Alexa-647–labeled RG7787 for 45 minutes on ice. Cells were washed twice with PBS followed by addition of 500 μL of trypsin and then placed in an incubator for 2 minutes. Afterwards, 1.5 mL of cold media were added and then transferred into tubes and centrifuged for 2 minutes at 2,000 rpm. Cells were washed twice with FACS buffer, and then resuspended in 500 μL of FACS buffer and analyzed on FACSCanto II.

Cell culture and cytotoxicity assay

Cell lines were cultured in RPMI1640 (Bio Whittaker's) supplemented with 10% FBS and 1% penicillin–streptomycin in 96-well plates in a humidified atmosphere of 5% CO2 at 37°C. The cell line, A431/H9, was described previously (33). For cytotoxicity assay, 5 × 103 cells were plated in Costar 96-well flat-bottom plates. The next day, cells were treated with SS1P, RG7787, or HB21(Fv)-PE40 for 72 hours. Cell viability was measured with Cell Counting Kit-8. Cell viability is expressed in reference to control where SS1P was not added.

Western blot analysis

Cells were washed in PBS and lysed in lysis buffer on ice for 30 minutes. After 5-minute centrifugation at 13,000 rpm, 15 μg of supernatants were loaded in a 4%–12% gradient gel and separated by SDS-PAGE under reducing conditions and then transferred to polyvinylidene difluoride membranes. The membranes were incubated with MSLN antibody (1:1,000) overnight in the cold room. The next day, goat-anti-mouse IgG HRP secondary Ab (1:5,000) was applied for 1 hour at room temperature followed by washing, then incubation with ECL, and exposure to X-ray film.

MSLN shedding assay in vivo and in vitro

Soluble MSLN levels in cell line medium were measured in duplicate with a human MSLN ELISA Kit (Novatein Biosciences). Cells (5 × 105) were seeded in 2-mL RPMI cell culture medium in multi-well 6-well polystyrene Falcon tissue culture dishes. The next day, media were removed and replaced with 1 mL of fresh media. Every hour thereafter, 100 μL of medium was collected and stored at −80°C. Other plates remained in the incubator for 4 days and afterwards, media were collected and stored at −80°C. Cells were trypsinized and counted using a cell counting machine. Cell supernatant was diluted 10-fold and added to wells of a 96-well plate previously coated with capture antibody. The amount of MSLN shed by wild-type and TACE-mutant cells was calculated using the MSLN concentration per cell and total cell number.

For in vivo MSLN shedding, 5 athymic nude mice per group were implanted with 2 × 106 A431/H9 or A431/G10 TACE–mutant cells on day 1; two mice were used as controls. When tumors reached 300 to 500 mm3, blood was collected from each mouse and transferred into heparin-coated tubes. Samples were centrifuged for 15 minutes at 1,000 × g, plasma was collected and aliquoted into new heparin-coated tubes, and stored at −80°C. For MSLN ELISA assays, samples were diluted and analyzed.

In vitro immunotoxin internalization and FACS analysis

Wild-type and TACE-mutant cells (2.5 × 105) were seeded in a multi-well 6-well tissue culture dishes. After 16 to 20 hours, cells were exposed to 10 ng/mL of RG7787-Alexa Fluor 647 in 1 mL of media for 90 minutes and for 4, 8, 16, and 24 hours. Surface-bound RG7787-Alexa Fluor 647 was stripped for 5 minutes using a low pH glycine buffer (0.2 mol/L glycine-HCl, pH 2.5) and 1% BSA to remove surface-bound RG7787. Cells were then trypsinized, washed with cold internalization-FACS buffer (PBS with 1% FBS and 0.1% NaN3) twice, and then resuspended in 500 μL cold FACS buffer and analyzed using FACSCanto II. Quantification of internalized RG7787 molecules was performed as described previously (32). For surface staining, cells were either detached using trypsin or citric saline for comparison. Cells were treated with 2 μg/mL of labeled immunotoxin for 45 minutes on ice.

Xenograft tumor model

On day 0, 2 × 106 wild-type and TACE-mutant cells in Matrigel (4 mg/mL) were implanted subcutaneously into the rear legs of 4- to 6-week-old athymic nude mice. When tumors reached 100 mm3 (day 5–7), mice were randomized and 8 to 10 mice in each group were injected intravenously with the maximum safe dose of SS1P (8 μg) or with HB21(Fv)-PE40 (2 μg) prepared in 200-μL PBS with 0.2% human serum albumin (HSA). To be consistent with the clinical practice, mice were injected with SS1P injections every other day for a total of 3 days (23). Mice were monitored for tumor size and tumor volumes were calculated and measured as (l × w2) × 0.4 (34, 35). All animal experiments were stopped when tumors reached 600–700 mm3 in volume. All animal experiments were approved by the National Cancer Institute Animal Care and Use Committee.

Statistical analysis

Data were analyzed and statistical analyses were performed using GraphPad Prism 6 software and Excel. Each experiment was performed at least twice and we used duplicates or triplicates for each point in every assay. Data are presented either as mean ± SD or ± SEM. Unpaired student test and one-way ANOVA with the Mann–Whitney test for comparisons was performed. A P value of <0.05 was considered statistically significant.

Generation and characterization of A431 cells expressing mutant MSLN

A431/H9 cells were engineered previously by transfecting A431 cells, which do not express MSLN, with a plasmid encoding a full-length human MSLN cDNA and isolating a single clone with high MSLN expression (36). To make a cell line with diminished MSLN shedding, but equal expression, we constructed a plasmid with a deletion of the TACE protease site (Fig. 1A) and stably transfected it into A431 cells. After antibiotic selection and cell sorting, we isolated several clones with high MSLN expression. We chose clone A431/G9 for further analysis, although several other clones have the same expression level. Flow cytometry analysis shows that clone A431/G9 has high MSLN expression that is comparable with that of A431/H9 cells (Fig. 1B and Supplementary Fig. S1). We measured the level of surface MSLN expression by two different methods. First, we used RG7787 directly labeled with Alex-647 (Fig. 1B). Next we used an anti-MSLN antibody followed by a fluorescently labeled secondary antibody (Supplementary Fig. S1). Both analyses showed that the level of MSLN expression was very similar in A431/H9 and TACE-mutant cells. MSLN expression was also analyzed by Western blots, and the results revealed a slight but statistically insignificant increase in total MSLN protein expression in TACE-mutant cells (Fig. 1C and D).

Figure 1.

Generation and characterization of mutant MSLN cell line. A, schematic representation of wild-type and mutant MSLN protein structure. The diagram shows the TACE sheddase site in MSLN. B, flow cytometry analysis using Alexa-647–labeled immunotoxin, RG7787, shows similar surface MSLN protein expression on wild-type (black solid line) and TACE-mutant cells (dotted line). C, Western blot analysis comparing MSLN expression of total cell lysates from TACE-mutant and wild-type cells. Four separate samples were run for each cell line. D, quantification of pooled Western blot analysis results using image J. Error bars from pooled Western blot analysis are represented as mean ± SD.

Figure 1.

Generation and characterization of mutant MSLN cell line. A, schematic representation of wild-type and mutant MSLN protein structure. The diagram shows the TACE sheddase site in MSLN. B, flow cytometry analysis using Alexa-647–labeled immunotoxin, RG7787, shows similar surface MSLN protein expression on wild-type (black solid line) and TACE-mutant cells (dotted line). C, Western blot analysis comparing MSLN expression of total cell lysates from TACE-mutant and wild-type cells. Four separate samples were run for each cell line. D, quantification of pooled Western blot analysis results using image J. Error bars from pooled Western blot analysis are represented as mean ± SD.

Close modal

Comparison of MSLN expression in A431/H9 and TACE-mutant cells

Figure 1B shows that wild-type and TACE-mutant cells have similar MSLN surface expression. To determine whether both cell lines saturate at the same concentrations of RIT, equal numbers of wild-type and mutant cells were plated overnight and the next day cells were detached using trypsin reagent. Cells were washed and exposed to increasing concentrations of labeled RG7787 on ice for 45 minutes and then flow cytometry was performed. As shown in Fig. 2A, both cell lines saturate at the same concentrations.

Figure 2.

Comparison of cell lines. A and B, flow cytometry analysis showing that wild-type and TACE-mutant cells saturate at similar concentrations of labeled RG7787 RIT. C, ELISA assay comparing shed MSLN rate by wild-type (opened squares) and TACE mutants (opened circles). Error bars represent variability from duplicate wells. D, in vitro cell growth assay comparing the growth of TACE-mutant cells (opened circle) and wild-type cells (opened square). Error bars are calculated from data from three independent experiments and values pooled. All data are presented as mean ± SD.

Figure 2.

Comparison of cell lines. A and B, flow cytometry analysis showing that wild-type and TACE-mutant cells saturate at similar concentrations of labeled RG7787 RIT. C, ELISA assay comparing shed MSLN rate by wild-type (opened squares) and TACE mutants (opened circles). Error bars represent variability from duplicate wells. D, in vitro cell growth assay comparing the growth of TACE-mutant cells (opened circle) and wild-type cells (opened square). Error bars are calculated from data from three independent experiments and values pooled. All data are presented as mean ± SD.

Close modal

Trypsin is known to remove surface proteins and might affect our data due to variability of trypsinization time. To be certain that there are similar amounts of MSLN molecules of both cell lines, we repeated our FACS experiments using citric saline to remove cells because it acts nonenzymatically to detach cells. We then treated cells with increasing concentrations of labeled RG7787. Our results show that both cell lines express similar amounts of MSLN and also saturate at the same concentration (Fig. 2B).

To determine the effect of the TACE deletion on MSLN shedding, we collected media from growing cells and performed an ELISA assay to measure MSLN levels. The MSLN concentration was 831 ng/mL in the A431/H9 medium and 160 ng/mL in the medium from TACE-mutant cells indicating a major decrease in MSLN shedding (Table 1). These data indicate shedding was greatly decreased but not completely abrogated. Next, we investigated and compared the amount of MSLN shed per cell per hour over a 6-hour period by wild-type and mutant cells. The data are presented in Fig. 2C and Table 2. The wild-type cells have shed more than 100% of the total on the surface at 1 hour and continue to shed at this high rate for 6 hours.

Table 1.

Summary of important data for wild-type A431/H9 and TACE mutant A431/G9 cells

Cell lineMSLN sites/cellMSLN in medium (ng/mL)MSLN in blood (ng/mL)SS1P IC50 (ng/mL)RG7787 IC50 (ng/mL)HB21 IC50 (ng/mL)
A431/H9 1.1 × 106 ± 3.7 831 ± 20 464 ± 70 0.05 ± 0.01 0.62 ± 0.10 0.001 ± 0.01 
A431/G9 1.1 × 106 ± 3.5 169 ± 1.4 24 ± 2.5 0.01 ± 0.01 0.16 ± 0.05 0.001 ± 0.01 
Cell lineMSLN sites/cellMSLN in medium (ng/mL)MSLN in blood (ng/mL)SS1P IC50 (ng/mL)RG7787 IC50 (ng/mL)HB21 IC50 (ng/mL)
A431/H9 1.1 × 106 ± 3.7 831 ± 20 464 ± 70 0.05 ± 0.01 0.62 ± 0.10 0.001 ± 0.01 
A431/G9 1.1 × 106 ± 3.5 169 ± 1.4 24 ± 2.5 0.01 ± 0.01 0.16 ± 0.05 0.001 ± 0.01 

NOTE: MSLN site per cell, MSLN shedding, and average IC50 values of immunotoxins on A431/H9, and A431/G9 cells are shown. To determine MSLN site/cell, 250,000 A431/H9 and A431/G9 cells were plated on day 1. On day 2, cells were detached using trypsin or citric saline and processed for flow cytometry analysis. To compare levels of shed MSLN in vitro by A431/H9 and TACE-mutant cells, 5 × 105 were plated in a T75 flask on day 1 and allowed to grow for 4 days. After 4 days, media were collected and used for MSLN-specific ELISA assay. To determine shed MSLN in serum of mice with A431/H9 and TACE-mutant tumors, blood was collected from 5 mice per group after tumors reached 300–500 mm3 (A431/H9 mouse 1 = 424 mm3; Mouse 2 = 424 mm3; Mouse 3 = 335 mm3; Mouse 4 = 448 mm3; Mouse 5 = 768 mm; TACE-mutant mouse 1 = 389 mm3; Mouse 2 = 389 mm3; Mouse 3 = 3,457 mm3; Mouse 4 = 389 mm3; Mouse 5 = 561 mm3) Plasma was isolated and used for MSLN-specific ELISA assay. Shed MSLN concentrations (A431/H9 mouse 1 = 413 ng/mL; Mouse 2 = 520 ng/mL; Mouse 3 = 355 ng/mL; Mouse 4 = 567 ng/mL; Mouse 5 = 469.3 ng/mL; TACE-mutant mouse 1 = 31 ng/mL; Mouse 2 = 27.2 ng/mL; Mouse 3 = 25 ng/mL; Mouse 4 = 8.3 ng/mL; Mouse 5 = 30 ng/mL) were pooled between groups. Average data from duplicate wells are shown. IC50 values were derived from 72-hour WST8 cell viability assays. The IC50 values represent averages from four independent experiments in quadruplicate. The data are represented as ± SD.

Table 2.

MSLN shedding by wild-type A431/H9 and TACE-mutant A431/G9 cells

Cell lineCell numberTotal MSLN on cell surface at time zero (ng)Total MSLN shed after 1 hour (ng)Total MSLN shed after 2 hours (ng)Total MSLN shed after 3 hours (ng)Total MSLN shed after 5 hours (ng)Total MSLN shed after 6 hours (ng)
A431/H9 9.8 × 105 71.1 100 ± 0.02 270 ± 0.6 270 ± 0.4 440 ± 0.6 660 ± 0.1 
A431/G9 7.8 × 105 56.6 16 ± 0.01 28 ± 0.02 50 ± 0.03 77 ± 0.2 92 ± 0.07 
Cell lineCell numberTotal MSLN on cell surface at time zero (ng)Total MSLN shed after 1 hour (ng)Total MSLN shed after 2 hours (ng)Total MSLN shed after 3 hours (ng)Total MSLN shed after 5 hours (ng)Total MSLN shed after 6 hours (ng)
A431/H9 9.8 × 105 71.1 100 ± 0.02 270 ± 0.6 270 ± 0.4 440 ± 0.6 660 ± 0.1 
A431/G9 7.8 × 105 56.6 16 ± 0.01 28 ± 0.02 50 ± 0.03 77 ± 0.2 92 ± 0.07 

NOTE: MSLN sites per cell, total MSLN per well containing cells, and total MSLN shed per hour values by A431/H9, and A431/G9 cells are shown. To determine total shed MSLN/hour, 5 × 105 cells were plated on day 1. On day 2, media were removed and replaced with 1 mL fresh media. Every hour thereafter, 100 μL of media were collected for ELISA assay to determine total shed MSLN. At the end of the experiment, cells were trypsinized and counted. Shed MSLN values were measured by a MSLN-specific ELISA assay. Total shed MSLN per hour values were calculated using predetermined MSLN sites/cell numbers (1.1 × 106 sites/cell), number of cells counted the day after plating.

In contrast, the cells shed much less MSLN with the TACE site deleted. After 1 hour, wild-type cells had shed all their surface MSLN, whereas mutant cells shed only 27% (Table 2).

In some cell lines, MSLN has been shown to regulate cancer cell growth and proliferation (37, 38). To determine whether the decrease in MSLN shedding by mutant cells affected the growth of the cells, we measured cell number over several days and found no change in growth rate between wild-type A431/H9 and TACE-mutant cells (Fig. 2D). On the basis of these data, we conclude the cell lines were suitable for additional comparative studies.

Sensitivity of TACE-mutant cells to SS1P, RG7787, and HB21(Fv)-PE40

To determine whether decreased MSLN shedding had an impact on the efficacy of MSLN-targeting immunotoxins in vitro, we treated wild-type and TACE-mutant cells with SS1P and RG7787. As shown in Fig. 3A and Table 1, TACE-mutant cells were 5-fold more sensitive to SS1P than A431/H9 cells (0.01 ng/mL compared with 0.05 ng/mL), respectively. The TACE mutant was also about 3.5-fold more sensitive to RG7787 than A431/H9 cells (0.16 ng/mL compared with 0.62 ng/mL), respectively (Fig. 3B and Table 1). To show that the difference in the cytotoxic effect was MSLN-specific, we treated the cell lines with an immunotoxin targeted to the human transferrin receptor, HB21(Fv)-PE40 (20) and found no difference in its effect on the two cell lines indicating the difference is MSLN-specific as shown in Fig. 3C.

Figure 3.

Reduced MSLN shedding renders cultured cells more sensitive to MSLN-targeting RITs. Cells were incubated for 72 hours with SS1P (A), RG7787 (B), or a transferrin receptor–targeting immunotoxin, HB21(Fv)-PE40 (C). Growth inhibition was assessed with a WST8 cell assay. Results compare sensitivity to immunotoxins between wild-type cells (opened squares) and TACE-mutant cells (opened circles). The error bars from all data represent the variability of triplicate wells from at least four experiments. Data are presented as mean ± SD.

Figure 3.

Reduced MSLN shedding renders cultured cells more sensitive to MSLN-targeting RITs. Cells were incubated for 72 hours with SS1P (A), RG7787 (B), or a transferrin receptor–targeting immunotoxin, HB21(Fv)-PE40 (C). Growth inhibition was assessed with a WST8 cell assay. Results compare sensitivity to immunotoxins between wild-type cells (opened squares) and TACE-mutant cells (opened circles). The error bars from all data represent the variability of triplicate wells from at least four experiments. Data are presented as mean ± SD.

Close modal

Uptake of RG7787 by TACE-mutant and A431/H9 cells

Because TACE-mutant cells were more sensitive to SS1P and RG7787 than A431/H9 cells, we examined their rate of uptake of immunotoxn assuming that shedding would reduce uptake. To do this, cells were seeded in 6-well dishes and 16 hours later 10 ng/mL of RG7787-Alexa Fluor 647 was added and cells harvested after 90 minutes, 4, 8, 12, and 16 hours. As shown in Fig. 4, there was a 2-fold increase in uptake at all time points by TACE-mutant cells compared with wild-type cells.

Figure 4.

Reduced MSLN shedding leads to increased internalization of Alexa-647–labeled RG7787 immunotoxin. Cells were incubated with 10 μg/mL of Alexa-647-RG7787 and uptake measured. TACE-mutant cells (open circles), wild-type cells (open squares). Error bars represent variability from three separate experiments. Data are presented as mean ± SD.

Figure 4.

Reduced MSLN shedding leads to increased internalization of Alexa-647–labeled RG7787 immunotoxin. Cells were incubated with 10 μg/mL of Alexa-647-RG7787 and uptake measured. TACE-mutant cells (open circles), wild-type cells (open squares). Error bars represent variability from three separate experiments. Data are presented as mean ± SD.

Close modal

Sensitivity of TACE-mutant tumors to immunotoxins SS1P, RG7787, and HB21-(Fv)-PE40

The data in Table 1 show that that TACE-mutant cells shed 5-fold less MSLN than wild-type cells. Next, we investigated the levels of circulating shed MSLN in wild-type and TACE-mutant mouse xenograft models. To do this, we collected blood from mice with wild-type and TACE-mutant tumors and analyzed the plasma using a MSLN-specific sandwich ELISA kit. Our results show greater than 20-fold less MSLN in the plasma of mice with TACE-mutant tumors compared with wild-type tumors (Fig. 5A). Next, we carried out antitumor experiments in mice to test the mathematical model. We implanted 2 × 106 A431/H9 or TACE-mutant cells subcutaneously into athymic nude mice (18). When tumors reached about 100 mm3, SS1P, RG7787, or HB21(Fv)-PE40 was intravenously administered every other day for a total of three doses. As shown in Fig. 5B, there was a remarkable difference in the response to SS1P. The TACE-mutant tumors were much more susceptible to SS1P than the A431/H9 tumors. There was a statistically significant difference in the size of TACE-mutant tumors compared with A431/H9 tumors beginning on day 9 (P < 0.0001), 2 days after treatment was initiated, and the difference in tumor size increased over time. The TACE-mutant tumors continued to decrease in size and were undetectable on day 15, 6 days after the last SS1P treatment. In contrast, A431/H9 tumors remained about the same size during treatment and resumed growth when treatment with SS1P was completed.

Figure 5.

Reduced MSLN shedding increases response of tumors to MSLN-targeting immunotoxins. A, MSLN levels in plasma of mice with cells expressing wild-type MSLN (open squares) and mutant MSLN (closed symbols). Data are presented as mean ± SD. **, P < 0.005. B, tumor-bearing mice were treated with D-PBS HSA 0.2% i.v. (control, n = 15/group) and 1 cycle of 3 × 0.4 mg/kg SS1P every other day for a total of 3 days (QOD) i.v. (n = 18/group). TACE-mutant tumors were significantly more sensitive to SS1P antitumor activity than wild-type tumors. C, tumor-bearing mice were treated with D-PBS HSA 0.2% i.v. (control, n = 15/group) and 1 cycle of 3 × 2.5 mg/kg RG7787 QOD i.v. (n = 20/group). D, tumor-bearing mice were treated with D-PBS HSA 0.2% i.v. (control, n = 12/group) and 1 cycle of 3 × 0.1 mg/kg HB21(Fv)-PE40 QOD i.v. (n = 12/group). The error bars represent the deviation of tumor measurements combined from two independent experiments. Data are presented as mean ± SEM.

Figure 5.

Reduced MSLN shedding increases response of tumors to MSLN-targeting immunotoxins. A, MSLN levels in plasma of mice with cells expressing wild-type MSLN (open squares) and mutant MSLN (closed symbols). Data are presented as mean ± SD. **, P < 0.005. B, tumor-bearing mice were treated with D-PBS HSA 0.2% i.v. (control, n = 15/group) and 1 cycle of 3 × 0.4 mg/kg SS1P every other day for a total of 3 days (QOD) i.v. (n = 18/group). TACE-mutant tumors were significantly more sensitive to SS1P antitumor activity than wild-type tumors. C, tumor-bearing mice were treated with D-PBS HSA 0.2% i.v. (control, n = 15/group) and 1 cycle of 3 × 2.5 mg/kg RG7787 QOD i.v. (n = 20/group). D, tumor-bearing mice were treated with D-PBS HSA 0.2% i.v. (control, n = 12/group) and 1 cycle of 3 × 0.1 mg/kg HB21(Fv)-PE40 QOD i.v. (n = 12/group). The error bars represent the deviation of tumor measurements combined from two independent experiments. Data are presented as mean ± SEM.

Close modal

To confirm our finding with SS1P, we examined the antitumor efficacy of RG7787 on A431/H9 and TACE-mutant tumors. Although RG7787 is more active than SS1P on many epithelial cancers (20), it is less active on A431/H9 cells as shown in Table 1, and therefore larger doses were used in mice. TACE-mutant tumors treated with 50 μg given every other day for a total of 3 days were also much more responsive to RG7787 than A431/H9 tumors (Fig. 5C). Similar to the SS1P antitumor study, there was a statistically significant difference in the size of the TACE-mutant tumors after RG7787 treatment compared with A431/H9 tumors beginning on day 9 (P < 0.0001), and the difference in tumor size continued to be significant for the remainder of the experiment.

To show that the difference in response to immunotoxin treatment was related to MSLN shedding and not to some other difference in the tumors, we also treated the A431/H9 and TACE-mutant tumors with HB21(Fv)-PE40 that targets the human transferrin receptor and observed no difference in response (Fig. 5D). Both tumor types responded similarly, consistent with our in vitro results. We conclude that the difference in the antitumor effect is specific for the anti-MSLN immunotoxins and directly correlate with MSLN shedding.

It is well known that many proteins that are targeted by antibody-based therapies are shed from the cell surface. This shed antigen has the potential to act as a decoy receptor and inhibits the action of these agents. However, because solid tumors are not in direct communication with the blood, and because immunotoxin SS1P took many hours after its entry into tumors to reach the maximum number of cells in the tumor, we carried out mathematical analysis of our data, which unexpectedly predicted that antigen shedding should enhance and not block the efficacy of RITs that target MSLN-expressing tumors (29). This report suggested that shed antigen acts as a protective reservoir, which carries RITs into spaces far from the blood vessel.

In this study, our goal was to use both in vitro and animal studies to test this mathematical model. To do this, we generated and characterized a cell line, A431/G9 (TACE mutant), which expresses a mutant form of MSLN that lacks the TACE sheddase site. We performed several experiments to make sure the mutant cells were characteristically similar to wild-type cells. Flow cytometry and cell culture experiments showed both wild-type and mutant cells displayed similar numbers of MSLN sites per cell and similar growth rates. We also show that both wild-type and mutant cells saturate at the same antibody concentration. It was previously shown that knockdown of TACE led to greater than 50% inhibition of MSLN shedding (28). These results suggested that TACE was a major MSLN-cleaving protease; however, there are other enzymes that cleave MSLN because knockdown of TACE did not completely abrogate MSLN shedding. We tested for MSLN shedding into cultured media and in serum from tumor-bearing mice. Our results demonstrated greatly reduced, but not 100%, abrogation of MSLN shedding by mutant cells. This finding indicates there must be a second shedding mechanism. Because MSLN is anchored to the cell by phosphatidyl inositol, and is released from cells by treatment with phospholipase C (2), it is likely a phospholipase also contributes to shedding.

In contrast with the mathematical model, our animal studies with tumors that shed MSLN at vastly different rates show that shedding greatly reduces the efficacy of two immunotoxins (SS1P and RG7787) that target MSLN but not an immunotoxin that targets the transferrin receptor. We have not yet determined the reason for the failure of the model to predict this result but future studies will address this issue.

Our results show that wild-type cells shed 100% of their surface MSLN every hour and it is replenished by fresh MSLN, whereas the TACE mutant sheds very little MSLN per hour. If this is the case, then cells expressing TACE mutant should have much more uptake of RIT than wild-type A431/H9 cells; however, our results show only a 2-fold increase of labeled RG7787 by the TACE-mutant cells. This difference could be due to the fact that we used 10 μg/mL to measure uptake. We could not use lower concentrations closer to the IC50 for cell killing because the uptake assay is not sensitive enough.

Because there are equal numbers of MSLN molecules on the surface of the mutant cells and the uptake rate is higher, one could reason that the binding site barrier effect is larger in tumors derived from these mutant cells. Therefore, without large amounts of shed MSLN molecules, the immunotoxin would be unable to travel far from the blood vessel into the tumor tissue and its efficacy would be reduced. Contrary to this expectation, the animal studies reported here with tumors that are derived from the H9 and G9 cells show that reduced shedding greatly enhances the efficacy of two different immunotoxins (SS1P and RG7787) that target MSLN.

The findings presented here indicate that there are factors that are more important in regulating immunotoxin action than the binding site barrier effect. One possibility is that the immunotoxin action is hindered more by a low endocytosis rate than by possible binding site barrier effect. The endocytosis rate used in the mathematical model, we reported previously (29, 39) was 22% per hour, which was derived by analyzing earlier in vitro data (40). The binding site barrier effect is the dominant factor for this model and the efficacy of immunotoxins increases monotonically as the shedding rate increases when the number of receptor molecules is one million or more. However, when the endocytosis rate is reduced to 8%, the immunotoxin is less effective in the tumor model with 20% shedding rate than those with 1.5% or 5% shedding rate. Thus, one possibility is that the in vivo endocytosis rate is much less than that measured in in vitro experiment in which case increasing the uptake rate by reducing shedding becomes beneficial rather than a barrier.

In conclusion, our results suggest that strategies that prevent MSLN shedding may significantly enhance immunotoxin efficacy in the treatment of MSLN-expressing cancers.

No conflicts of interest were disclosed.

Conception and design: P. Awuah, T.K. Bera, M. Folivi, I. Pastan

Development of methodology: P. Awuah

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P. Awuah, O. Chertov

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P. Awuah, T.K. Bera, O. Chertov, I. Pastan

Writing, review, and/or revision of the manuscript: P. Awuah, T.K. Bera, O. Chertov, I. Pastan

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P. Awuah

Study supervision: I. Pastan

The authors thank Drs. David Fitzgerald, Fabian Mueller, and Christine Alewine for providing feedback on the article. The authors also thank Devorah Gallardo for assisting with animal care.

This work was supported in part by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research and by the National Cancer Institute, NIH, Contract HHSN261200800001E (to O. Chertov).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Scholler
N
,
Fu
N
,
Yang
Y
,
Ye
Z
,
Goodman
GE
,
Hellstrom
KE
, et al
Soluble member(s) of the mesothelin/megakaryocyte potentiating factor family are detectable in sera from patients with ovarian carcinoma
.
Proc Natl Acad Sci U S A
1999
;
96
:
11531
6
.
2.
Chang
K
,
Pastan
I
,
Willingham
MC
. 
Isolation and characterization of a monoclonal antibody, K1, reactive with ovarian cancers and normal mesothelium
.
Int J Cancer
1992
;
50
:
373
81
.
3.
Chang
K
,
Pai
LH
,
Pass
H
,
Pogrebniak
HW
,
Tsao
MS
,
Pastan
I
, et al
Monoclonal antibody K1 reacts with epithelial mesothelioma but not with lung adenocarcinoma
.
Am J Surg Pathol
1992
;
16
:
259
68
.
4.
Argani
P
,
Iacobuzio-Donahue
C
,
Ryu
B
,
Rosty
C
,
Goggins
M
,
Wilentz
RE
, et al
Mesothelin is overexpressed in the vast majority of ductal adenocarcinomas of the pancreas: identification of a new pancreatic cancer marker by serial analysis of gene expression (SAGE)
.
Clin Cancer Res
2001
;
7
:
3862
8
.
5.
Chang
K
,
Pastan
I
,
Willingham
MC
. 
Frequent expression of the tumor antigen CAK1 in squamous-cell carcinomas
.
Int J Cancer
1992
;
51
:
548
54
.
6.
Bera
TK
,
Pastan
I
. 
Mesothelin is not required for normal mouse development or reproduction
.
Mol Cell Biol
2000
;
20
:
2902
6
.
7.
Wang
Y
,
Wang
L
,
Li
D
,
Wang
HB
,
Chen
QF
. 
Mesothelin promotes invasion and metastasis in breast cancer cells
.
J Int Med Res
2012
;
40
:
2109
16
.
8.
Tang
Z
,
Qian
M
,
Ho
M
. 
The role of mesothelin in tumor progression and targeted therapy
.
Anticancer Agents Med Chem
2013
;
13
:
276
80
.
9.
Govindan
SV
,
Goldenberg
DM
. 
Designing immunoconjugates for cancer therapy
.
Expert Opin Biol Ther
2012
;
12
:
873
90
.
10.
Choi
IK
,
Strauss
R
,
Richter
M
,
Yun
CO
,
Lieber
A
. 
Strategies to increase drug penetration in solid tumors
.
Front Oncol
2013
;
3
:
193
.
11.
Heine
M
,
Freund
B
,
Nielsen
P
,
Jung
C
,
Reimer
R
,
Hohenberg
H
, et al
High interstitial fluid pressure is associated with low tumour penetration of diagnostic monoclonal antibodies applied for molecular imaging purposes
.
PLoS ONE
2012
;
7
:
e36258
.
12.
Zhang
Y
,
Pastan
I
. 
High shed antigen levels within tumors: an additional barrier to immunoconjugate therapy
.
Clin Cancer Res
2008
;
14
:
7981
6
.
13.
Lee
JH
,
Kim
H
,
Yao
Z
,
Lee
SJ
,
Szajek
LP
,
Grasso
L
, et al
Tumor and organ uptake of Cu-labeled MORAb-009 (amatuximab), an anti-mesothelin antibody, by PET imaging and biodistribution studies
.
Nucl Med Biol
2015
;
42
:
880
6
.
14.
Hassan
R
,
Kindler
HL
,
Jahan
T
,
Bazhenova
L
,
Reck
M
,
Thomas
A
, et al
Phase II clinical trial of amatuximab, a chimeric antimesothelin antibody with pemetrexed and cisplatin in advanced unresectable pleural mesothelioma
.
Clin Cancer Res
2014
;
20
:
5927
36
.
15.
Golfier
S
,
Kopitz
C
,
Kahnert
A
,
Heisler
I
,
Schatz
CA
,
Stelte-Ludwig
B
, et al
Anetumab ravtansine: a novel mesothelin-targeting antibody-drug conjugate cures tumors with heterogeneous target expression favored by bystander effect
.
Mol Cancer Ther
2014
;
13
:
1537
48
.
16.
Scales
SJ
,
Gupta
N
,
Pacheco
G
,
Firestein
R
,
French
DM
,
Koeppen
H
, et al
An antimesothelin-monomethyl auristatin e conjugate with potent antitumor activity in ovarian, pancreatic, and mesothelioma models
.
Mol Cancer Ther
2014
;
13
:
2630
40
.
17.
Weldon
JE
,
Xiang
L
,
Zhang
J
,
Beers
R
,
Walker
DA
,
Onda
M
, et al
A recombinant immunotoxin against the tumor-associated antigen mesothelin reengineered for high activity, low off-target toxicity, and reduced antigenicity
.
Mol Cancer Ther
2013
;
12
:
48
57
.
18.
Liu
XF
,
Xiang
L
,
FitzGerald
DJ
,
Pastan
I
. 
Antitumor effects of immunotoxins are enhanced by lowering HCK or treatment with SRC kinase inhibitors
.
Mol Cancer Ther
2014
;
13
:
82
9
.
19.
Alewine
C
,
Xiang
L
,
Yamori
T
,
Niederfellner
G
,
Bosslet
K
,
Pastan
I
. 
Efficacy of RG7787, a next-generation mesothelin-targeted immunotoxin, against triple-negative breast and gastric cancers
.
Mol Cancer Ther
2014
;
13
:
2653
61
.
20.
Hollevoet
K
,
Mason-Osann
E
,
Liu
XF
,
Imhof-Jung
S
,
Niederfellner
G
,
Pastan
I
. 
In vitro and in vivo activity of the low-immunogenic antimesothelin immunotoxin RG7787 in pancreatic cancer
.
Mol Cancer Ther
2014
;
13
:
2040
9
.
21.
Weldon
JE
,
Pastan
I
. 
A guide to taming a toxin–recombinant immunotoxins constructed from Pseudomonas exotoxin A for the treatment of cancer
.
FEBS J
2011
;
278
:
4683
4700
.
22.
Hassan
R
,
Williams-Gould
J
,
Steinberg
SM
,
Liewehr
DJ
,
Yokokawa
J
,
Tsang
KY
, et al
Tumor-directed radiation and the immunotoxin SS1P in the treatment of mesothelin-expressing tumor xenografts
.
Clin Cancer Res
2006
;
12
:
4983
8
.
23.
Hassan
R
,
Broaddus
VC
,
Wilson
S
,
Liewehr
DJ
,
Zhang
J
. 
Anti-mesothelin immunotoxin SS1P in combination with gemcitabine results in increased activity against mesothelin-expressing tumor xenografts
.
Clin Cancer Res
2007
;
13
:
7166
71
.
24.
Hassan
R
,
Sharon
E
,
Thomas
A
,
Zhang
J
,
Ling
A
,
Miettinen
M
, et al
Phase 1 study of the antimesothelin immunotoxin SS1P in combination with pemetrexed and cisplatin for front-line therapy of pleural mesothelioma and correlation of tumor response with serum mesothelin, megakaryocyte potentiating factor, and cancer antigen 125
.
Cancer
2014
;
120
:
3311
9
.
25.
Hassan
R
,
Miller
AC
,
Sharon
E
,
Thomas
A
,
Reynolds
JC
,
Ling
A
, et al
Major cancer regressions in mesothelioma after treatment with an anti-mesothelin immunotoxin and immune suppression
.
Sci Transl Med
2013
;
5
:
208ra147
.
26.
Pastan
I
,
Zhang
Y
. 
Modulating mesothelin shedding to improve therapy
.
Oncotarget
2012
;
3
:
114
5
.
27.
Law
SK
. 
Antigen shedding and metastasis of tumour cells
.
Clin Exp Immunol
1991
;
85
:
1
2
.
28.
Zhang
Y
,
Chertov
O
,
Zhang
J
,
Hassan
R
,
Pastan
I
. 
Cytotoxic activity of immunotoxin SS1P is modulated by TACE-dependent mesothelin shedding
.
Cancer Res
2011
;
71
:
5915
22
.
29.
Pak
Y
,
Zhang
Y
,
Pastan
I
,
Lee
B
. 
Antigen shedding may improve efficiencies for delivery of antibody-based anticancer agents in solid tumors
.
Cancer Res
2012
;
72
:
3143
52
.
30.
Pastan
I
,
Beers
R
,
Bera
TK
. 
Recombinant immunotoxins in the treatment of cancer
.
Methods Mol Biol
2004
;
248
:
503
18
.
31.
Onda
M
,
Willingham
M
,
Nagata
S
,
Bera
TK
,
Beers
R
,
Ho
M
, et al
New monoclonal antibodies to mesothelin useful for immunohistochemistry, fluorescence-activated cell sorting, Western blotting, and ELISA
.
Clin Cancer Res
2005
;
11
:
5840
6
.
32.
Mason-Osann
E
,
Hollevoet
K
,
Niederfellner
G
,
Pastan
I
. 
Quantification of recombinant immunotoxin delivery to solid tumors allows for direct comparison of in vivo and in vitro results
.
Sci Rep
2015
;
5
:
10832
.
33.
Liu
XF
,
FitzGerald
DJ
,
Pastan
I
. 
The insulin receptor negatively regulates the action of Pseudomonas toxin-based immunotoxins and native Pseudomonas toxin
.
Cancer Res
2013
;
73
:
2281
8
.
34.
Liu
XF
,
Xiang
L
,
Zhang
Y
,
Becker
KG
,
Bera
TK
,
Pastan
I
. 
CAPC negatively regulates NF-kappaB activation and suppresses tumor growth and metastasis
.
Oncogene
2012
;
31
:
1673
82
.
35.
Onda
M
,
Nagata
S
,
Tsutsumi
Y
,
Vincent
JJ
,
Wang
Q
,
Kreitman
RJ
, et al
Lowering the isoelectric point of the Fv portion of recombinant immunotoxins leads to decreased nonspecific animal toxicity without affecting antitumor activity
.
Cancer Res
2001
;
61
:
5070
7
.
36.
Ho
M
,
Hassan
R
,
Zhang
J
,
Wang
QC
,
Onda
M
,
Bera
T
, et al
Humoral immune response to mesothelin in mesothelioma and ovarian cancer patients
.
Clin Cancer Res
2005
;
11
:
3814
20
.
37.
Bharadwaj
U
,
Li
M
,
Chen
C
,
Yao
Q
. 
Mesothelin-induced pancreatic cancer cell proliferation involves alteration of cyclin E via activation of signal transducer and activator of transcription protein 3
.
Mol Cancer Res
2008
;
6
:
1755
65
.
38.
Zheng
C
,
Jia
W
,
Tang
Y
,
Zhao
H
,
Jiang
Y
,
Sun
S
. 
Mesothelin regulates growth and apoptosis in pancreatic cancer cells through p53-dependent and -independent signal pathway
.
J Exp Clin Cancer Res
2012
;
31
:
84
.
39.
Pak
Y
,
Pastan
I
,
Kreitman
RJ
,
Lee
B
. 
Effect of antigen shedding on targeted delivery of immunotoxins in solid tumors from a mathematical model
.
PLOS ONE
2014
;
9
:
e110716
.
40.
Zhang
Y
,
Hansen
JK
,
Xiang
L
,
Kawa
S
,
Onda
M
,
Ho
M
, et al
(2010) A flow cytometry method to quantitate internalized immunotoxins shows that taxol synergistically increases cellular immunotoxins uptake
.
Cancer Res
2010
;
70
:
1082
9
.