DHA-SBT-1214 Taxoid Nanoemulsion and Anti-PD-L1 Antibody Combination Therapy Enhances Anti-Tumor Efficacy in a Syngeneic Pancreatic Adenocarcinoma Model

Abstract

The goal of this study was to evaluate combination of a novel taxoid, DHA-SBT-1214 chemotherapy in modulating immune checkpoint marker expression and ultimately in improving antibody-based checkpoint blockade therapy in pancreatic adenocarcinoma (PDAC). DHA-SBT-1214 was encapsulated in an oil-in-water nanoemulsion and administered systemically in Panc02 syngeneic PDAC-bearing C57BL/6 mice. Following treatment with DHA-SBT-1214, expression levels of PD-L1 were measured and anti-PD-L1 antibody was administered in combination. The effects of combination therapy on efficacy and the molecular basis of synergistic effects were evaluated. PD-L1 expression was lower on Panc02 pancreatic tumor cells in vitro, which significantly increased after exposure to different chemotherapy drugs. Administration of DHA-SBT-1214, gemcitabine and PD-L1 antibody alone failed to increase CD8⁺ T-cell infiltration inside tumor. However, combination of anti-PD-L1 therapy with a novel chemotherapy drug DHA-SBT-1214 in nanoemulsion (NE-DHA-SBT-1214), significantly enhanced CD8⁺ T-cell infiltration and the therapeutic effects of the anti-PD-L1 antibody. Furthermore, in Panc02 syngeneic model, NE-DHA-SBT-1214 combination therapy group reduced tumor growth to higher extend than paclitaxel, nab-paclitaxel (Abraxane™), gemcitabine, or single anti-PD-L1 antibody therapy groups. Our results indicate that NE-DHA-SBT-1214 stimulated immunogenic potential of PDAC and provided an enhanced therapeutic effect with immune checkpoint blockade therapy, which warrants further evaluation.

INTRODUCTION

In the United States, pancreatic adenocarcinoma (PDAC) is currently ranks fourth leading cause of cancer-related deaths and is expected to rise significantly in the next decade largely due to an increase in type I and II diabetes epidemic ¹. Currently chemotherapy, radiation and surgery are standard treatment options for pancreatic cancer ². The high mortality associated with PDAC is attributed mainly to advanced stage diagnosis and therapeutic resistance to agents, such as gemcitabine ^{2, 3}. During metastasized stage of PDAC, systemic chemotherapeutic drugs like taxanes and gemcitabine, is the most important interventional option. However, these drugs are not tumor specific and can develop multi-drug resistance (MDR) ⁴. Instead of gemcitabine chemotherapy alone, the combination with nab-paclitaxel (Abraxane™), and a cocktail of 5-fluorouracil, oxaliplatin, irinotecan, and leucovorin (FOLFIRINOX) have shown better results in patients’ overall survival (OS) response ⁵. Along with toxicity issues, one of the major limitations of chemotherapy agents in PDAC is their inability to penetrate the highly fibrotic and desmoplastic tumor tissue, thereby limiting the effectiveness to the rapidly proliferating PDAC cells confined to the tumor periphery ². Some of these limitations are responsible for slower rate of drug development against PDAC and some other solid tumors ⁶.

At present, there is dire need for the development of safer and effective drugs that can improve on PDAC’s dismal OS. Over the last several years, we have developed next generation taxoids that show superior antitumor efficacy, especially in refractory tumors by targeting the cancer initiating (or stem) cell population. One of these next generation toxoids is SBT-1214 that has shown superior efficacy in SCID mice against colon tumor xenograft in ⁷ and completely regressed the tumor in another study⁸. In order to decrease its toxicity and further improve its specificity, we have conjugated this next generation taxoid with an an omega-3 polyunsaturated fatty acid, docosahexaenoic acid (DHA)⁹. We have evaluated the efficacy of this conjugated taxoid in different tumor models including non-small cell lung, pancreatic, colon and ovarian tumor xenografts in mouse models and demonstrated improved efficiency as compared all the other control studies ¹⁰.

Because of poor water solubility of this drug and susceptibility of the ester bond in DHA-SBT-1214 to cleave in the presence of esterases in the systemic circulation, we have formulated the drug in omega-3 fish oil-rich nanoemulsion formulation. In a previous studies, we have observed that nanoemulsion encapsulated DHA-SBT-1214 (NE-DHA-SBT-1214) shows improved delivery efficiency, tumor residence, improved therapeutic efficacy, and less systemic toxicity in a subcutaneous stem cell-rich PPT-2 human prostate tumor xenograft model ¹¹.

Use of immune check point inhibitors and adoptive cell therapies that increase the effectiveness of cytotoxic T-cells have become major breakthroughs in the cancer immunotherapy treatment options of solid tumors^{12, 13}. Immunotherapy with immune check point inhibitors have shown not only preclinical but also clinical efficacy in many different types of tumors, the overall efficacy in PDAC is limited ¹⁴. Among immune check point inhibitors, programmed death-ligand 1 (PD-L1) which is expressed on tumor cells binds with programmed death-1 (PD-1), which is found on activated T-cells and is a member of the CD28 family sends an inhibitory signal to T-cell and stops immune response ^{13, 15}. One of the early immune-oncology agent approved in the United States is nivolumab (Opdivo^®), this is a humanized antibody (IgG4 PD-1), used in the United States for treating patients having melanoma ^{13, 15, 16}, non-small cell lung cancer ¹⁵ and renal cell carcinoma ¹⁷. Along with anti-PD-1 and anti-PD-L1 antibodies, various other immune checkpoint inhibitors, including antibodies against CTLA-4 are approved and being used in combination for many solid tumor types ^18–20. In murine preclinical pancreatic cancer models, anti-PD-1 or anti-PD-L1 inhibitors have performed comparatively better than other check point inhibitors ^{14, 21}. Due to highly variable expression of PD-L1, this immune checkpoint inhibitor as a monotherapy has not demonstrated clinical efficiency in in human pancreatic cancer ²². However, anti-CTLA-4 drug ipilimumab (Yervoy^®) in combination with a cancer vaccine has shown marginal survival benefit in clinical patients with pancreatic cancer ^{23, 24}.

One of the limitations of immunotherapy in PDAC could be due to dense and highly fibrotic tumor stroma that prevents antibodies and T-cell infiltration into the tumor mass. To gain maximum benefits from the immunotherapy, it would need to be combined with antifibrotic agents and other small molecule or immune-based therapies to make the tumors responsive to immunotherapy. In this regard, some studies with anti-PD-1/PD-L1 antibodies have shown contradictory findings ^25–28. In this study, we have investigated how anticancer agents (paclitaxel, gemcitabine, Abraxane™ and DHA-SBT-1214) influence PD-L1 expression in PDAC model. As such, combination of a small molecule chemotherapy along with an antibody against check point inhibitor would have synergistic effect for PDAC treatment. Additionally, we provide evidence that NE-DHA-SBT-1214 in combination with PD-L1 antibody shows improved anticancer effect due to higher penetration of CD8 T-cells inside the tumor.

MATERIALS AND METHODS

Materials

Synthesis of Docosahexaenoic acid conjugate of SBT-1214 (i.e., DHA-SBT-1214) was performed by ChemMaster International, Inc. (Stony Brook, NY) and DHA-SBT-1214 structure is available on company website and has also been reported previously ^{7, 29, 30}. Following reagents were purchased from the respective vendors. Omega-3 rich fish oil from Jedwards International (Quincy, MA), Protease inhibitors, Gemcitabine (GEM), paclitaxel (PTX) and Tween 80 from Sigma Chemicals, Inc. (St. Louis, MO), Lipoid E80 from Lipoid GMBH (Ludwigshafen, Germany), DSPE PEG2000 from Avanti Polar Lipids, Inc. (Alabaster, AL), Dulbecco’s Modified Eagle Medium (DMEM) and LAL endotoxin quantitation kit from Thermo Scientific (Rockford, IL). Trypsin, Penicillin, and streptomycin were purchased from Invitrogen (Grand Island, NY, USA). Amicon Ultra-0.5ml, Centrifugal filters from Millipore (Cork, Ireland). All other analytical grade reagents were purchased through Fisher Scientific. Female C57BL/6 mice (4–6 weeks old) were obtained from Charles River Laboratories Cambridge, MA).

Preparation and Characterization of DHA-SBT-1214 Nanoemulsion Formulations

Blank and DHA-SBT-1214 encapsulated fish oil/water nanoemulsion formulations was prepared as reported previously ^{11, 31–33}. The formulations were characterized for their oil droplet size, surface charge, drug encapsulation efficiency, and the morphology of the oil droplets was evaluated with transmission electron microscopy.

Cell Culture

The mouse pancreatic cancer cell line Panc02, passage number 3 was kindly provided by Professor Michael Hollingsworth from the University of Nebraska Medical Center (Omaha, NE), through material transfer agreement “MTA17314T” on May, 25^th 2017 ³⁴. After receiving cells from our collaborator, passage 3 Panc02 cells were thawed and grown in cell culture flasks and maintained in DMEM medium supplemented with 10% fetal bovine serum (FBS), L-glutamine and penicillin (100 U/ml)/streptomycin (100 μg/ml) (both from Gibco Life Technologies, Carlsbad, CA, USA). Cells were incubated at 37°C in a humidified atmosphere containing 5% CO₂. Cells were passaged once a week for three additional passages before starting animal experiment. Before start of each animal experiment, cells were confirmed to be mycoplasma free with commercially available kit from R&D Systems’ new MycoProbe® Mycoplasma Detection Kit (Catalog # CUL001B), Inc. Minneapolis, MN. We did not perform cell authentication in our laboratory.

In Vitro Evaluations of Nanoemulsion Uptake and Cellular Distribution

Panc02 cells (0.5×10⁶) were seeded in 6-well plates onto glass cover-slips for overnight at 37°C containing 5% CO2 in a humidified cell culture incubator. Then cells were incubated with 2 μM of rhodamine administered in nanoemulsion for different time points (0.5 h to 4 h) to allow uptake of nanoparticles by cells. After last incubation time point, the glass cover-slips were washed with PBS before fixing in formalin for 15 minutes. 4’, 6-diamidino-2-phenylindole (DAPI) was used to stain nuclei of the fixed cells. Uptake of rhodamine nanoemulsion was studied by a fluorescence confocal microscope (Zeiss LSM 700) with fixed parameters to have comparable uptake among different time points.

Cell Viability Studies

To evaluate the cytotoxic effect of different drugs and nanoemulsion formulation, 5000 cells were seeded per well of the 96-well plate for overnight at 37°C in a cell culture incubator containing 5% CO₂. Various control and test treatments (i.e., paclitaxel solution, Abraxane™, gemcitabine solution, and solution and nanoemulsion of DHA-SBT-1214) were diluted at different concentrations ranging from 0 nM to 10000 nM and Panc02 cells treated with these concentrations for 96 hours. After treatment, cells were incubated with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT). MTT crystals were dissolved with DMSO and plates read at 570nm absorbance using a BioTek Synergy HTX Multi-Mode Microplate Reader.

PD-L1 expression after Exposure to Different Therapeutic Agents

Cells were seeded at 0.5×10⁶ cells per well in 6-well plates for overnight at 37^oC in a humidified incubator with 5% CO2. After 24 hours, cells were treated with IC₅₀ value of different drug treatments as described in Figure 2 for 48 hours. After that, expression level of PD-L1 was determined using flow cytometry as follows. Briefly, harvested cells were washed twice with 3% BSA/PBS and incubated with either isotype control or rat anti-PD-L1 (mouse, BioXcell, West Lebanon, NH, USA) for 30 minutes at 4^oC. After washing three times, the cells were incubated with anti-rat Alexa Fluor 488 conjugated antibody. The cells washed once more with 3% BSA/PBS and analyzed on FACSCalibur flow cytometer and CellQuest™ Pro version 6.0 software (both from Becton-Dickinson and Co.).

Figure 2.

The activity of different anti-cancer agents against Panc02 cells in vitro. The percentage maximal response as a function of anti-cancer agents when administered to Panc02 cells. The cell viability was measured using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay after 96 h of incubation at 37 ºC. Data represent mean ± standard deviation (n=3). Significant differences are indicated as follows: *p < 0.05, and **p < 0.01.

Immunoblotting

Cells and tumor tissues washed twice with phosphate-buffered saline (PBS) and lysed in ice-cold lysis buffer which contains 2% proteinase inhibitor. Cells were retrieved with a cell scraper, stirred and incubated on ice for 15 min, whereas, mice tumor tissues were sonicated for 10 seconds on ice with an ultrasound probe. After centrifugation of the lysates and protein concentration was determined from supernatants. The supernatants were diluted with lysis buffer to create equal amount of protein. Equal concentration of the protein was separated on 4–12% Bis-Tris gels. By using the iBlot Dry Blotting System, the separated proteins were transferred onto a nitrocellulose membrane. After blocking blots in 1% dry milk in TBS-T for 1 hour, the membranes were incubated at 4°C with different anti-PD-L1 ((ab58810), from Abcam), PD-1 ((D7D5W) XP® Rabbit mAb, 84651S), F4/80 ((D2S9R) XP® Rabbit mAb, 70076S) and Histone 3 ((D1H2) XP® Rabbit mAb, 4499S), (all from Cell Signaling Technology, Inc.) antibodies in TBS-T. After washing, the membranes were incubated with the secondary antibodies either against anti-rabbit and mouse IgG antibodies (Life Technologies) and were detected using western blotting, Bio rad: Gel Doc apparatus (Bio rad, Hercules, Ca).

Quantitative Polymerase Chain Reaction (qPCR)

The expression level of PD-L1 and mRNA for other proteins from the pancreatic cancer cells (Pan02) or tumor tissues was determined using real-time PCR. Total mRNA was extracted using commercially available RNA extraction kit (Thermo Fisher Scientific (Rockford, IL). The extracted RNA either used fresh or stored at −80°C until use. For real-time qPCR, 1 μg of isolated RNA was reverse-transcribed using commercial cDNA synthesis kit (Thermo Fisher Scientific (Rockford, IL). The resulting cDNA was subjected to RT-PCR with Applied Biosystems™ PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific (Rockford, IL), using the following primers for mouse PD-L1: (forward primer, 5’-AAAGTCAATGCCCCATACCG-3’ and reverse primer, 5’-TTCTCTTCCCACTCACGGGT-3’)³⁵; mouse PD-1 (forward primer, 5’- TTCACCTGCAGCTTGTCCAA-3’ and reverse primer, 5’- TGGGCAGCTGTATGATCTGG-3’)³⁵; CD4: (forward primer, 5’- ACACACCTGTGCAAGAAGCA-3’ and reverse primer, 5’- GCT CTTGTTGGTTGGGAATC-3’)³⁶; mouse CD8 (forward primer, 5’- CTCACCT GTGCACCCTACC-3’ and reverse primer, 5’-ATCCGGTCCCCTTCACTG-3’)³⁶; mouse Arginase-1 (forward primer, 5’-GAACACGGCAGTGGCTTTAAC −3’ and reverse primer, 5’-TGCTTAGCTCTGTCTGCTTTGC-3’)³⁷; and mouse β-actin (forward primer, 5’- CTCCTGAGCGCAAGTACTCTGTG-3’ and reverse primer, 5’- TAAAACGCAGCTCAGTAACAGTCC-3’)³⁸. PCR was run using a real-time PCR system (7300; Applied Biosystems, Foster City, CA, US A). Relative quantifications of gene expression were calculated according to mouse β-actin as a housekeeping control.

In Vivo Studies – Subcutaneous Tumor Inoculation and Growth

The experiments in which animals were involved, performed as per recommendations in the guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH). All experiments were carried out under strict accordance of the protocol for animal experiments was approved by Northeastern University’s Institutional Animal Care and Use Committee (IACUC). For inoculation of tumor cells, after propagation twice, Panc02 murine pancreatic cancer cells were mixed in 1:1 PBS/Matrigel and 2 × 10⁵ cells injected to the right flanks of a 6 weeks old C57Bl/6 mice subcutaneously. Development of the tumor was monitored twice weekly and the tumor size was measured with a caliper twice a week along with animal body weight and tumor volumes determined using the well-established formula 0.5ab², where b represents the smaller of the two perpendicular diameters. The mice were sacrificed according to the guidelines of IACUC protocol when the tumor volume reached ≥ 1,500 mm³ in diameter.

In Vivo Single and Combination Therapies

Mouse antibody against PD-L1 (10F.9G2) and relevant isotype IgG control was purchased from Bio X Cell. Two hundred micrograms of antibody against PD-L1 and relevant isotype IgG control was injected through i.p. per mice twice a week for 3 weeks. Gemcitabine solution and Abraxane™ at 120mg/kg was injected through i.p. once a week. Paclitaxel 120mg/kg and NE-DHA-SBT-1214 either 10mg/kg or 25mg/kg was injected once a week through i.v. All chemotherapy drugs were either injected in combination to isotype IgG control or anti PD-L1 antibody. In total, three treatments were given per experiment.

Histology and Immunohistochemistry (IHC) Analysis of Tumor Tissues

Tumor tissues from mice were fixed in formaldehyde and embedded in paraffin to do histological analysis by hematoxylin and eosin (H&E) staining. Fixed and paraffin-embedded tissue sections were also used for IHC against PD-1 ((D7D5W) XP® Rabbit mAb, 84651T), CD4 ((D7D2Z) Rabbit mAb, 25229T) and CD8 ((D4W2Z) XP® Rabbit mAb, 98941T) antibodies (Cell signaling technology, Danvers, MA). IHC was processed according to the protocol and recommended dilution from Cell Signaling Technology.

Data Analysis and Statistics

For in vivo therapeutic efficacy experiments, each treatment group contains three mice. GraphPad Prism 6 software was used for performing statistical data analysis by using ANOVA and unpaired Student’s t test. The differences were considered to be statistically significant at p < 0.05. The results are expressed as the means ± SD here.

RESULTS

DHA-SBT-1214 Nanoemulsion Formulation and Characterization

Nanoemulsion mode of drug delivery has shown better therapeutic efficacy previously ^{11, 33}. For current study, an oil-in-water nanoemulsion, a new-generation taxoid (DHA-SBT-1214) was formulated using fish oil. We used a high pressure homogenization technique to formulate this uniform, milky-white and stable nanoemulsion ³⁹. As shown in Figure 1a and b, the nanoemulsion droplets were near spherical in morphology with an average diameter of approximately 220 nm, as observed by light scattering and transmission electron microscopy (TEM). The bioavailability of the nanoemulsion is predicted by its uniformity, charge and particle size. Uniformity of droplet size in a sample is shown by polydispersity index (PDI) and the value of PDI lower than 0.2 dictates uniform and stable nature of nanoemulsions. In our study, values of PDI of the nanoemulsions were <0.1, indicating their best uniformity. The oil droplets in the nanoemulsions have an average surface charge of −28.9 mV (Figure 1c). The free fatty acids of the fish oil which was used to make these nanoemulsions might be responsible for their negative charge.

Figure 1.

Characterization of DHA-SBT-1214 nanoemulsion formulations. (A) – Transmission electron microscopy image of DHA-SBT-1214 encapsulated nanoemulsion. (B) – The oil droplet particle size determination in nm. (C) – The measurement of zeta potential or surface charge on the oil droplets in mV; and (D) – The uptake of rhodamine-encapsulated nanoemulsion formulation in Panc02 cells. Fluorescence microscopy images showing the blue (nucleus), red (rhodamine encapsulated nanoemulsion) and overlay images in purple color. The images were acquired at 63x magnification. The image scale bar is 100 μm.

The drug concentration inside the nanoemulsion formulations was determined by an HPLC assay. The concentration of DHA-SBT-1214 was 20 mg/ml with 97%. The drug encapsulated nanoemulsion were filtered through 0.2-micron filter to ensure their sterility which was confirmed through Limulus Amebocyte Lysate (LAL) assay throughout the study.

In Vitro Evaluations of nanoemulsion Formulations in Panc02 cells

To study internationalization of nanoemulsions in Panc02 cells, confocal microscopy studies were done on rhodamine encapsulated nanoemulsions. The spheroid and cell uptake of 2 μM rhodamine containing nanoemulsions was observed at different time points (Figure 1d). As shown in Figure 1d, the images clearly depict that the dye encapsulated nanoemulsions internalize in spheroids and in the Panc02 cells and has higher intracellular signal at later time points. After confirmation of internalization of rhodamine encapsulated nanoemulsions, we replaced DHA-SBT-1214 with rhodamine in the nanoemulsion formulation and examined its effect on viability of cells with different anti-cancer drugs.

The cell-kill efficiency of different anti-cancer drugs was studied in Panc02 cells using the MTT assay. In addition to blank nanoemulsion or vehicle control, the concentrations of DHA-SBT-1214 nanoemulsion for these studies ranges from 0.01 nM to 10,000 nM based on previous studies of SBT-1214 ⁴. The concentration-response of DHA-SBT-1214 nanoemulsions and other anticancer agents in Panc02 cells are shown in Figure 2. After 96 hours of treatment at 37ºC, the results are presented as percent viability of cells from each treatment. Greater cell death was observed at 10 and 100nM concentrations of nanoemulsion compared to its aqueous solutions. However, under in vitro conditions, gemcitabine showed highest potency with average IC₅₀ value of 154 nM, followed by 215 nM for DHA-SBT-1214 nanoemulsion and 262 nM for DHA-SBT-1214 in solution. In contrast, the average IC₅₀ values of paclitaxel and Abraxane™ were significantly higher than DHA-SBT-1214, at 443 nM and 428 nM, respectively.

Evaluation of PD-L1 Expression Following Drug Therapy in Panc02 Cells

Panc02 cells were treated with gemcitabine, nab-paclitaxel (Abraxane™), paclitaxel and DHA-SBT-1214 both in solution and nanoemulsion for 48 hours to examine the induction of PD-L1 protein expression. PD-L1 expression levels on tumor cells were determined by flow cytometry. The flow cytometry data is shown as the Δ mean fluorescence intensity after subtracting MFI of anti-PD-L1 subtracted from the isotype control. Figure S1a, shows induction of PD-L1 expression after treatment with different anticancer drugs at their IC₅₀ values in Panc02 cells. PD-L1 upregulation in response to the anticancer agents tested was significantly increased compared to the untreated control. As reported previously, PD-L1 level enhanced in pancreatic tumor tissues compared to in vitro growing cells as shown in Figure S1b^{40, 41}.

In Vivo Evaluation of Combination Drug and anti-PD-L1 Antibody Therapy

We have examined the effect of different anticancer agents either alone or in combination to blocking antibody against PD-L1 in syngeneic mouse pancreatic tumor model. Panc02 tumors were established subcutaneously and animal body weight and tumor volumes were measured twice a week till the end of the experiment. After tumor size reaches approximately 100 mm³, we randomized the mice to have approximately equal tumor volume among all treatment groups. Then, mice were treated with either anticancer agents alone or in combination with immune check point inhibitor for three weeks. Figures 3a, b, and c, show the tumor growth inhibitory effects of each treatment group at the end of experiment. Our results show that compared to the untreated control group, each of the treatment group had better inhibitory effect which was most prominent in 25mg/kg NE-DHA-SBT-1214 in both IgG and PD-L1 antibody combination treated groups. These results indicate that blocking of only PD-L1 was not efficient in reducing tumor growth but in combination with 25mg/kg NE-DHA-SBT-1214 significantly inhibited tumor growth. NE-DHA-SBT-1214 treatment even at 10mg/kg in combination to PD-L1 antibody was more effective in suppressing tumor growth compared to standard chemotherapy drug, gemcitabine. Treatment with 10mg/kg DHA-SBT-1214 was superior to Abraxane™ treatment at 120 mg/kg. Overall, a combinational treatment of NE-DHA-SBT-1214 with anti-PD-L1 antibody showed a synergistic effect compared with single treatment, and this was particular noticeable for in particular for NE-DHA-SBT-1214 10 mg/kg plus PD-L1, compared to NE-DHA-SBT-1214 plus IgG. As a crude proxy for acute toxicity of the treatment protocols, we evaluated the changes in body weight loss. There was no significant body weight loss among the treatment groups (data not shown).

Figure 3.

In vivo efficacy of the PD-L1 antibody in combination to different therapeutic drugs including gemcitabine solution, Abraxane™, paclitaxel in solution, DHA-SBT-1214 in solution and NE-DHA-SBT-1214 against Panc02 induced syngeneic mice tumors. (A)– Graph summarizing all treatment modalities. The values are means ± SD (n=3). Significant differences are indicated as follows: *p < 0.05, and **p< 0.01. (B)– Tumor images taken at the time of harvest from different treatment modalities. (3b-A)– Tumors from mice treated with vehicle; (3b-B)– Three tumors each from PD-L1 (200μg) treated mice; (3b-C, D)–Tumors from Abraxane™ plus IgG or PD-L1 (200μg) treated mice respectively; (3b-E)– Tumors from NE-DHA-SBT-1214 (10mg/kg) plus IgG (200μg) treated mice; (3b-F, G)– Tumors from gemcitabine plus IgG or PD-L1 (200μg) treated mice respectively; (3b-H)– Tumors from NE-DHA-SBT-1214 (10mg/kg) plus PD-L1 (200μg) treated mice; (3b-I, J)– Tumors from NE-DHA-SBT-1214 (25mg/kg) plus IgG or PD-L1 (200μg) treated mice respectively. (C)– Graph for all the tumors from (3B) to show their progression over time.

Anticancer Drugs Upregulate PD-L1 Expression In Vivo in Panc02 Tumor Model

To examine whether the anticancer agents increase the PD-L1, PD-1, CD4, CD8 and Arginase-1 mRNA expression in pancreatic tumor tissues. We determined the mRNA level of CD4, CD8, PD1, PD-L1 and Arginase-1either alone or in combination of immune checkpoint inhibitor treatment groups by RT-PCR. PD-L1 mRNA level was upregulated in combination therapy among all the anticancer agents compared to their respective IgG control groups as shown in figure 4a. However, CD4 and PD-1 mRNA level was lower in anti- PD-L1 plus anticancer agents except gemcitabine which was not significantly higher compared to its IgG treated group as shown in Figure S2a and S2c respectively. CD8 mRNA level was upregulated in response to combination treatment of all anticancer agents when combined with immune check point inhibitor compared to their IgG treated groups as shown in Figure S2b respectively. However, IgG treatment group has higher level of Arginase-1 in comparison to the immune check point inhibitor treatment group as described in Figure S2d. In addition to induction of PD-L1 mRNA expression level, treatment of anticancer agents in combination with immune check point inhibitor also enhance PD-L1 protein expression levels as shown in Figure 4b and c. Similar to PD-L1, PD-1 protein expression was also up-regulated compared to its IgG treatment group, except by higher dose of NE-DHA-SBT-1214 as shown in Figure 4b and Supplementary Figure S4. Higher PD-L1 protein level might be attributed to presence of macrophages in this higher dose NE-DHA-SBT-1214 treated group which is evident due to higher protein level of F4/80 in Figure 4b.

Figure 4.

In vivo PD-L1 surface protein expression in response to different therapeutic modalities. (A)– mRNA expression of PD-L1 from different mouse tumor treatment groups analyzed using RT-PCR. Relative gene expression for RT-PCR data was calculated relative to murine β-actin. (B)– Tumor tissue lysate from different treated groups was prepared and protein level of different proteins was analyzed using western blotting. (C)– The bands corresponding to PD-L1 were quantified using Image J software and was normalized relative to band intensities for the corresponding Histone 3 loading controls. The bar represents the mean ± standard deviation of data from at least 3 independent experiments; *p<0.05, **p<0.01.

Infiltration of PD-1, CD4+, and CD8+ Cells in Panc02 Tumor

We investigated the infiltration of different cell types including CD4+, CD8+, and PD-1 cells in tumor tissues at the end of experiment by histology (Figure 5) and by immunohistochemistry (Figure 6a and Supplementary Figures 3Sa, b and c). The tumor tissue histology from different treatment groups showed that tumor from NE-DHA-SBT-1214 treated group has less dense stroma compared to solid tumor mass from other treatment groups (Figure 5).

Figure 5.

Histopathological evaluation of the Panc02-induced tumor tissues collected from control and different combination treated mice (hematoxylin & eosin staining). Significant reduction in tumor stroma observed with combination of NE-DHA-SBT-1214 and anti-PD-L1 treated groups. The images were taken at 63x magnification.

Figure 6.

Immunohistochemical analysis of infiltrating CD4 or CD8 cells by immunohistochemistry and their quantification. (A & Supplementary Figure S3a)–Tumor tissue from all treatment groups were fixed in PFA and stained with anti-CD4 antibody, (B & Supplementary Figure S3b)– anti-CD8 antibody, and (C & Supplementary Figure S3c)– anti-PD1 antibody according to vendors protocol. The images were taken at 63x magnification.

Untreated control tumor tissues have a relatively small number of CD4+ cells whereas, infiltration of CD4+ cells was significantly increased by treatment of anti-PD-L1 antibody in combination with different anticancer agents (Figure 6a and Supplementary Figure 3a).

Similar trend was observed for CD8+ cell infiltrations, showing less number in control tumor tissue and highest in combination of anti-PD-L1 antibody with NE-DHA-SBT-1214 treated group (Figure 6b and Supplementary Figure 3b). The infiltration of CD8+ cells in the core of pancreatic tumor might be responsible for the suppression of tumor growth in each of the different treatment groups. However, induction of PD-1 was comparable among all treatment groups (Figure 6c and Supplementary Figure 3Sc).

DISCUSSION

Pancreatic cancer remains an intractable disease due to development of resistance to conventional anticancer agents. Currently, there is a great enthusiasm for immunotherapy in many treatment regimens due to better response of immune checkpoint inhibitors, also adoptive cell transfer therapy, and chimeric antigen receptor (CAR) T-cell therapy are tending towards better efficacy in clinic ⁴². However, immune checkpoint inhibitors as a single treatment has not shown promising results in many tumor types, especially in certain solid tumors, such as PDAC. As such, there are extensive efforts toward improving response rate and efficacy by combining chemotherapy and immune check point inhibitors.

For PDAC patients, gemcitabine is currently used as a frontline treatment in combination with Abraxane™; however, the survival benefit is minimal. For different solid tumors, Paclitaxel is still a first choice of treatment ^{43, 44} because of its apoptosis and cell cycle arrest causing properties^{45, 46}. But most solid tumors including colon and prostate express higher level of P-glycoprotein (Pgp), and actively pumps out paclitaxel from the cell, the drug is not effective ⁴. To overcome the Pgp efflux issues, paclitaxel has been conjugated with DHA. This conjugated Paclitaxel has increased affinity for human serum albumin and acts as a carrier for PUFAs in the bloodstream ¹⁰. However, when paclitaxel is cleaved by esterases from the DHA conjugate in the systemic circulation, the free drug is still susceptible to efflux by Pgp and other ABC transporters in tumors ¹⁰.

In comparison to paclitaxel, our new-generation taxoid, has shown better efficacy against efflux transport over expressing cancer cells^{4, 8, 10}. In previous studies, DHA-conjugated SBT-1214 improved therapeutic efficacy by increased presence of drug inside the tumor microenvironment because of the enhancement permeability and retention effect ³⁹. To further improve the efficacy of DHA-SBT-1214, we formulated nanoemulsion of DHA-SBT-1214 in fish oil. This oil works as a drug reservoir. This colloidal system has desired size and charge to preserve the stability of this nanoemulsion formulation in vitro and enhance its permeability in vivo ⁴⁷. The surface morphology DHA-SBT-1214 nanoemulsion formulation was spherical and uniform. The efficient cellular uptake of the nanoemulsion formulations suggests that the nanoemulsions successfully deliver the drug to the subcellular sites in the Panc02 cell and has better efficacy than its drug solution. In our recent study ¹¹, we observed that DHA-SBT-1214 suppressed tumor growth to a higher extent when delivered in nanoemulsion formulations emphasizing its higher therapeutic efficacy when used as stand-alone therapy. Our data from the previous study demonstrated that nanoemulsion of the DHA-SBT-1214 has better tumor regression efficacy and will be a novel anti-cancer drug candidate ¹¹. In this study, we have explored the efficacy of the combination therapy containing immune check point inhibitors and anticancer agents in pancreatic cancer. As reported previously, paclitaxel, AbraxaneTM, DHA-SBT-1214 and gemcitabine upregulates expression of PD-L1 in Panc02 cells.

This study is one of the few that address the effect of chemo and immunotherapy combination in a syngeneic pancreatic cancer mouse model. In literature, the effect of anticancer agents on PD-L1 induction has been conflicting ^25–28, with some studies, reporting upregulation PD-L1 expression, and one study reporting the opposite in response to chemotherapy agents ²⁵. For instance, Gong et al showed that paclitaxel induces PD-L1 expression, at both mRNA and protein levels, in two different cancer cell models ²⁷. Similarly, PD-L1 expression was increased by paclitaxel, gemcitabine or carboplatin in ovarian cancer cell lines ²⁸. In contrast, doxorubicin upregulated expression of PD-L1 inside nucleus and its downregulated on the surface of breast cancer cells ^26,48. One possible explanation for the conflicting findings among these previous studies, might be the differences of cancer cell lines and/or anticancer agents being used.

In this study, we tested Abraxane™, gemcitabine, paclitaxel as well as both solution and nanoemulsion of DHA-SBT-1214, either used alone or in combination with immune check point inhibitors against pancreatic cancer. The concentration of anticancer agent used in this in vivo study was based on IC₅₀ value of Panc02 cells ^{25, 49, 50}. The difference in concentration of drug being used has not significantly influenced the expression level of PD-L1 and the PD-L1 surface protein was upregulated in response to all anticancer agents as determined by flow cytometry ²⁶. In tumor cells, expression of PD-L1 is regulated by both innate and adaptive immune resistance mechanisms as reported previously ^51,48. In general, chemotherapy drugs along with cytotoxicity, can also cause alteration in tumor immune response both of which can induce tumor immune escape. It has been well reported that PD-1/PD-L1 interactions is important for immune homeostasis because of their negative regulation of T-cell activation ^{52, 53}. This negative regulation from PD-1/PD-L1 interaction makes cancer cells to escape from tumor-specific T-cell immunity ^{54, 55}. Blockade of PD-L1 in pancreatic cancer has reported to regress tumor, although these studies have not used anticancer agents along with immune check point inhibitor ¹⁴. In our syngeneic pancreas cancer model study, blocking of PD-L1 reduced rate of tumor growth in when used as a single treatment option or when used in combination with commonly used anticancer agents (Paclitaxel, Abraxane and Gemcitabine) for pancreatic cancer. However, combination of NE-DHA-SBT-1214 with PD-L1 blockade showed significant tumor suppression and kept tumor regressed even after treatment, suggesting that PD-L1 might be a novel target for treating pancreatic cancer.

Freeman et al, showed that PD-L1 reduced T-cell proliferation, however our results reported an increase in number of tumor-infiltrating cells after treatment with anti-PD-L1 antibody. Previously, an increase in level of IFN-gamma has been reported after blocking of the PD-1/PD-L1 in different studies⁵⁶. Increase in expression of PD-L1 after PD-L1 antibody treatment might be due to the increased infiltration of IFN-gamma producing CD8+ T cells to tumor tissue. Besides CD8+ T cells ⁵⁷, Natural killer (NK) cells ⁵⁸ have also been reported to upregulate PD-L1 expression inside tumor microenvironment that might be due secretion of a cytokine (IFNγ) both by NK and CD8+ T cells. Another possible explanation for the upregulation of PD-L1 mRNA and protein after anti-PD-L1 antibody treatment is the recruitment of Myeloid derived suppressor cells (MDSC) and macrophages, which also express PD-L1. The increased IFN-gamma from infiltrating CD8+ T cells play an important role in antitumor activity because IFN-gamma expression in the tumor microenvironment recruits inflammatory cells such as M1 macrophages^{54, 59}. Macrophages in the tumor microenvironment overexpress Arginase-1 indicating that these macrophages are M1 in addition to possible presence of MDSC. In this study, tumor regression effect of anti-PD-L1 antibody in NE-DHA-SBT-1214 combination treatment group might be due to the increased number of tumor-infiltrating effector cells. In other words, in untreated group, PD-L1 might decrease the infiltration of IFN-gamma-producing T-cells and M1 macrophages. As same cells that were injected into mice to form a pancreatic tumor expressed very high level of PD-L1 after IFN-gamma treatment in vitro. In our study, anti-PD-L1 antibody treatment did not decrease the number of tumor-infiltrating CD4+ T-cells. Overall, our results suggested that PD-L1 blockade can decrease the pancreatic tumor burden through additive effect of NE-DHA-SBT-1214. Furthermore, histology of tumor tissues from different treatment groups showed that tumor from NE-DHA-SBT-1214 treated group has less dense stroma compared to solid tumor mass from other treatment groups. However, neither single nor the combination therapy of most commonly used anticancer agents show an additive anti-tumor effect except NE-DHA-SBT-1214. One possible explanation for better efficacy of NE-DHA-SBT-1214 is its role in treating cancer stem cells as compared to other anti-cancer agents.

CONCLUSIONS

The results of this study indicate a significant tumor growth suppression by blocking PD-L1 in combination to NE-DHA-SBT-1214. Blockade of PD-L1 increased intra-tumoral IFN-gamma producing T-cells and inflammatory macrophages, that play role in tumor regression. However, the level of both PD-1 and PD-L1 were high in combination of commonly used anti-cancer agents emphasizing increased tumor infiltration of Treg cells, which might be culprit for increased tumor burden. These differential roles of different anticancer agent combinations may be a good start to explore other clinical treatments options for pancreatic cancers. Overall, the results shown here, may aid in the design of more effective anticancer treatments by combining chemotherapy and immunotherapy.

ABBREVIATIONS

MDR

Multidrug Resistance

DHA

Docosahexaenoic Acid

PEG

Poly(Ethylene Glycol

DMEM

Dulbecco’s Modified Eagle’s medium

DSPE-PEG2000

1, 2-Distearoyl-Sn-glycero-3-Phospho Ethanolamine-N-[amino [Poly(Ethylene Glycol)-2000]

TEM

Transmission Electron Microscopy

LAL

Limulus Amebocyte Lysate

HPLC

High Performance Liquid Chromatography

IACUC

Institutional Animal Care and Use Committee

PTX

Paclitaxel

AUC

Area Under the Curve

PDI

Polydispersity Index

OS

Overall Patient Survival

PD-L1

Programmed Death Ligand-1

PD-1

Programmed Death-1

Panc02

Mouse Pancreatic Adenocarcinoma Cell line

GEM

Gemcitabine

MDSC

Myeloid Derived Suppressor Cells