The immune system can recognize cancer-associated antigens and eliminate some cancer cells [1
]. However, cancer cells that evolved into established tumors have outgrown or evaded the immune system. Immunotherapy is designed to shift the cancer-immune balance against tumors and promote tumor regression in patients [3
]. One target being explored for immunotherapy is cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4). It is an immune regulatory protein that is expressed on activated T cells and inhibits immune activation. An antibody that antagonizes CTLA-4 was effective in melanoma patients [4
]. However, its use in other cancers has not yielded positive results so far [5
]. An effort to find combination strategies with anti-CTLA-4 is warranted.
Immunogenic cell death promotes the recognition of dying cell antigens by the immune system. In the past, apoptotic cell death was considered tolerogenic; however, work done in mice shows that some neoplastic agents that kill cells by apoptosis can induce anti-tumor immunity [6
]. Several markers of immunogenic cell death have been described, including the translocation of calreticulin from the endoplasmic reticulum (ER) to the cell surface and extracellular secretion of adenosine triphosphate (ATP). The induction of markers of immunogenic cell death can predict the ability of drugs to induce anti-tumor immunity [8
], and these types of drugs would be good candidates for the potential synergy with anti-CTLA-4.
Immunotoxins are anti-cancer agents composed of a targeting domain genetically fused to a toxin [9
]. In our lab, we develop immunotoxins that use truncated Pseudomonas exotoxin A (PE) as their payload fused to an antibody fragment that targets a tumor antigen. PE is one of the virulent factors of the gram-negative bacterium Pseudomonas aeruginosa
]. After immunotoxins bind to cells and are internalized, the toxin is cleaved from the antibody domain, inhibits protein synthesis, and leads to cell death by apoptosis [11
Mesothelin is a 40 kDa protein that under normal conditions is only expressed on mesothelial cells lining the pleura, peritoneal, and pericardium. The role of mesothelin in mice or humans is not known and mesothelin knocks out mice that do not have a noticeable phenotype [12
]. Importantly, its expression is increased in many epithelial cancers including the lungs, pancreas, ovary, stomach, colon, and mesothelioma [13
]. Mesothelin is being explored as a target for various immunotherapies [15
SS1P is the first generation of anti-mesothelin immunotoxins which has been tested in patients with mesothelin-expressing cancers. As a single drug, it had only a modest effect. Out of 33 patients treated with bolus injections of SS1P, 4 had minor partial responses, and 19 had a stable disease [16
]. One of the obstacles found in this study was the induction of rapidly evolving antibodies which neutralized the drug. To reduce the antibodies forming against SS1P, Hassan et al. combined SS1P with the immune modulating chemotherapies pentostatin and cyclophosphamide. In this study, 3 out of 10 patients had major regressions. All three continue to respond even when the drugs were discontinued and had major tumor regressions lasting up to 5 years [13
]. We suspect that the direct cytotoxic effect of SS1P was accompanied by the induction of anti-tumor immunity.
LMB-100 (also known as RG7787) is a second generation anti-mesothelin PE immunotoxin. It contains a smaller fragment of PE (24 kDa) that is composed of enzymatically active domain III. In addition, it incorporates point mutations designed to reduce B cell recognition [18
]. LMB-100 is currently being tested in clinical trials (NCT02798536, NCT03436732, NCT02810418).
Our previous work showed that injecting immunotoxins directly into tumors in combination with anti-CTLA-4 antibody had a synergistic anti-tumor effect in the 66C14 murine breast tumor model. Disease regressions were accompanied by long-term anti-tumor immunity indicated by the rejection of a second tumor challenge from the same cells [19
In this study, we used a syngeneic AE17M murine mesothelioma tumor model to evaluate immunotoxin efficacy in the mesothelioma. AE17 cells were derived from the peritoneal cavity of C57BL/6 mice treated with asbestos and later modified to express human mesothelin (AE17M) [20
]. We found that immunotoxin treatment promotes markers of immunogenic cell death in culture, and when injected directly into the tumors, immunotoxins enhance the effect of anti-CTLA-4 therapy.
In this study, we found that anti-mesothelin immunotoxins increased the expression of surface calreticulin and the extracellular secretion of ATP. Both are markers of immunogenic cell death. When SS1P is injected directly into AE17M tumors, it significantly enhances the effect of anti-CTLA-4 therapy. This result is in accordance with our previous report of a synergy between anti-CTLA-4 and locally injected anti-mesothelin immunotoxins found in the 66C14-M murine tumor model [19
We found a significant increase in the number of cells that present calreticulin on their membrane and an increase in ATP secretion after exposure to cytotoxic levels of SS1P. Calreticulin is usually found in the ER but can translocate to the cell surface in response to ER stress and the phosphorylation of eukaryotic initiation factor 2 (eIF2). Once it reaches the cell surface, it acts as an “eat me” signal to phagocytic cells [26
]. The ability of PE immunotoxins to induce ER stress and phosphorylation of eIF2 was previously shown in several cell lines [28
]. PE’s mechanism of action is protein synthesis inhibition; thus, our result indicates that the surface calreticulin pathway can be activated without the production of new proteins.
We found that in response to SS1P or LMB-100, ATP is released from dying AE17M cells. Extracellular ATP is another marker of immunogenic cell death. It can bind to the P2X7 receptor on dendritic cells and prime them to promote anti-tumor immunity [22
]. Much of the characterization of immunogenic cell death was done by vaccinating tumor-naïve mice with dying tumor cells that were pretreated in vitro with different chemotherapies [7
]. The model of vaccination with dying cells is different from the clinical use of chemotherapies in which the therapy is given to cancer patients directly. Besides inducing immunogenic cell death in cancer cells, immunogenic chemotherapies, such as doxorubicin and oxaliplatin, are myelosuppressive and can induce leukopenia [30
]. In patients, their ability to induce immunogenic cell death is countered by a decrease in circulating immune cells. Unlike chemotherapies, leukopenia is not one of the side effects of anti-mesothelin immunotoxins [16
Our findings that localized injections to tumor site are a useful method of treatment has also been shown by Luther et al., who evaluated the use of the locally delivered 8H9scFv-PE38 immunotoxin into athymic rats bearing brain tumors. They found that locally delivered 8H9scFv-PE38 reduced the size of tumors, caused tumor necrosis and prolonged the survival of rats from 24 to 43 days [33
Very few tumor models have been developed to evaluate the effect of PE immunotoxins on anti-tumor immunity. Ochiai et al. showed that when the MR1-1 immunotoxin was injected intratumorally immediately after inoculation of SMA560 EGFRvIII cells, it only prevented the formation of tumors in mice with an intact immune system. Moreover, the mice rejected a second challenge with the same cells, indicating that the MR1-1 immunotoxin mediates anti-tumor immunity [34
]. Kawakami et al. showed that injecting the IL13-PE38 immunotoxin into D5 IL13α2 tumors slowed the growth rate of both injected and un-injected tumors growing in the same mice. This effect was abolished when the mice were depleted of CD4 and CD8 expressing cells, indicating that the effect depends on the adaptive immune system [35
The AE17M mesothelioma tumor model is uniquely suitable to support clinical development of anti-mesothelin immunotoxins. Up to 80% of patients with mesothelioma have high mesothelin expression and thus are candidates for anti-mesothelin immunotoxin therapy [36
]. In this model, injecting anti-mesothelin immunotoxins into the tumors had a transient anti-tumor effect and treatment with anti-CTLA-4 and i.t. PBS affected only a minority of tumors. However, the combination of i.t. anti-mesothelin immunotoxins with systemic anti-CTLA-4 induced tumor regression in the majority of the mice, supporting the idea that the combination of the two outperforms each treatment alone.
This work has a few limitations. One is that the AE17M tumor model did not respond to intravenous (i.v.) immunotoxins; thus, we were unable to evaluate whether i.v. immunotoxin sensitizes tumors to the therapeutic effect of anti-CTLA-4. Minor responses reported in clinical trials with i.v. SS1P suggests that i.v. immunotoxin can kill some tumor cells [16
]. Nevertheless, the combination of locally injected immunotoxins with anti-CTLA-4 might be useful in unresectable local disease. As a single drug, TP-38 and NBI-3001 immunotoxins have occasionally resulted in the complete or durable partial responses in brain tumors [37
]. The use of i.t. VB4-845 in cutaneous metastases resulted in the partial or complete regression in many of the injected tumors. Some patients also experienced regressions in un-injected tumor sites [41
]. These effects might be further potentiated by a combination with anti-CTLA-4.
Altogether, this study shows that the local administration of anti-mesothelin immunotoxins potentiates the effect of anti-CTLA-4 in a murine mesothelioma model and induces markers of immunogenic cell death, providing preclinical support to pursue this combination strategy in the clinic.
4. Materials and Methods
4.1. Cell Culture and Reagents
SS1P was manufactured by ABL (Rockville, MD, USA). LMB-100 and anti-CTLA-4 (clone 9D9, isotype IgG2a) were manufactured by Roche. The AE17M cell line was kindly provided by Dr. Steve Albelda from the University of Pennsylvania. Cells were cultured in RPMI 1640 with 10% heat-inactivated FBS (incubated in a 56 °C water bath for 30 min) and supplemented with 100 U/mL penicillin, and 100 U/mL streptomycin. Cultures were maintained at 37 °C with 5% CO2.
4.2. Cytotoxicity Assays
Cells were plated at 2500 cells/well in 96-well flat-bottom plates and incubated overnight. Cells were treated with various concentrations of immunotoxins and incubated for 72 h. Cell viability was determined by a WST-8 cell counting kit (Dojindo Molecular Technologies, Inc, Kumamoto, Japan) per the manufacturer’s instructions. Absorbance was measured at 450 nm and normalized to 0% viability (cycloheximide treatment) and 100% viability (media).
4.3. Mouse Experiments
Female, wild-type C57BL/6 mice lot 027 at 6-9 weeks of age were purchased from Charles River. All mouse experiments followed NIH guidelines approved by the Animal Care and Use Committee of the National Cancer Institute (Animal Protocol LMB-014, date of approval: 3 March 2015). AE17M cells (2 × 106) in PBS were inoculated subcutaneously on the flank. Tumor volumes were measured two to three times per week using a caliper. The tumor volume was calculated using the formula 0.4 × length × width2. Mice were euthanized if the tumor volume exceeded 400 mm3, if they became hypoactive, or lost more than 10% of their weight. The day of euthanasia was used to calculate survival. Immunotoxin was injected intravenously in a 100 µL volume, or directly into the tumor (i.t.) in a 30 µL volume. Prior to the i.t. injection, the injection site was sterilized with povidone-iodine and alcohol pads. Anti-CTLA-4 diluted with PBS was injected into the peritoneum (i.p.) at a volume of 200 µL. All mice were followed for 90 days or more. The cured mice were re-challenged with AE17M cells (2 × 106) on the contralateral flank 40 days from the first tumor inoculation.
4.4. Flow Cytometry
AE17M cells were plated at 1 × 106 cells in 10-cm tissue culture dishes and allowed to adhere overnight. Cells were treated with either 65 µM tunicamycin (Sigma, St. Louis, MO, USA), 5 µg/mL oxaliplatin (Teva) or 100 ng/mL SS1P. All cells were harvested (including cells floating in the media) and stained with a rabbit polyclonal anti-calreticulin antibody (ab2907, Abcam, Cambridge, UK) at 1:200 for 60 min. Secondary anti-Rabbit FITC antibody (ab6717, Abcam) was added at 1:1000 for 30 min. Dead cells were labeled using 7AAD (BD Pharmigen, San Diego, CA, USA) and excluded from the analysis. Data were acquired on a FACSCanto II flow cytometer (BD Bioscience, Franklin Lakes, NJ, USA) and analyzed using FlowJo.
4.5. ATP Assay
AE17M cells were plated at 8 × 104 in a 24 well plate and allowed to adhere overnight. Cells were treated with either Doxorubicin (Pfizer, New York, NY, USA), Docetaxel (Winthrop, Bridgewater, NJ, USA) or anti-mesothelin immunotoxins in various concentrations and time durations. Next, the plates were spun at 1000 rpm for three minutes to reduce cell contamination and the supernatant from each well was transferred to a different 96 plate and ATP measured using the ENLITEN (Promega, Madison, WI, USA) kit according to the manufacturer’s instructions. Each 24-well plate was processed and evaluated separately to reduce the processing time until ATP was measured. Bioluminescence was analyzed using Victor3 (PerkinElmer, Waltham, MA, USA).
4.6. Statistical Analyses
Statistical analyses and graphing were performed with the GraphPad Prism software. The log-rank (Mantel-Cox) test was used to compare survival of mice. The Mann–Whitney test was used to compare the percentage of calreticulin positive cells and ATP values between groups. Error bars represent SEM.