Treatment of pancreatic ductal adenocarcinoma with tumor antigen specific-targeted delivery of paclitaxel loaded PLGA nanoparticles

Background Pancreatic ductal adenocarcinoma (PDA) remains the most aggressive cancers with a 5-year survival below 10%. Systemic delivery of chemotherapy drugs has severe side effects in patients with PDA and does not significantly improve overall survival rate. It is highly desirable to advance the therapeutic efficacy of chemotherapeutic drugs by targeting their delivery and increasing accumulation at the tumor site. MUC1 is a membrane-tethered glycoprotein that is aberrantly overexpressed in > 80% of PDA thus making it an attractive antigenic target. Methods Poly lactic-co-glycolic acid nanoparticles (PLGA NPs) conjugated to a tumor specific MUC1 antibody, TAB004, was used as a nanocarrier for targeted delivery into human PDA cell lines in vitro and in PDA tumors in vivo. The PLGA NPs were loaded with fluorescent imaging agents, fluorescein diacetate (FDA) and Nile Red (NR) or isocyanine green (ICG) for in vitro and in vivo imaging respectively or with a chemotherapeutic drug, paclitaxel (PTX) for in vitro cytotoxicity assays. Confocal microscopy was used to visualize internalization of the nanocarrier in vitro in PDA cells with high and low MUC1 expression. The in vivo imaging system (IVIS) was used to visualize in vivo tumor targeting of the nanocarrier. MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide) assay was used to determine in vitro cell survival of cells treated with PTX-loaded nanocarrier. One-sided t-test comparing treatment groups at each concentration and two-way ANOVAs comparing internalization of antibody and PLGA nanoparticles. Results In vitro, TAB004-conjugated ICG-nanocarriers were significantly better at internalizing in PDA cells than its non-conjugated counterpart. Similarly, TAB004-conjugated PTX-nanocarriers were significantly more cytotoxic in vitro against PDA cells than its non-conjugated counterpart. In vivo, TAB004-conjugated ICG-nanocarriers showed increased accumulation in the PDA tumor compared to the non-conjugated nanocarrier while sparing normal organs. Conclusions The study provides promising data for future development of a novel MUC1-targeted nanocarrier for direct delivery of imaging agents or drugs into the tumor microenvironment. Electronic supplementary material The online version of this article (10.1186/s12885-018-4393-7) contains supplementary material, which is available to authorized users.


Background
Pancreatic Cancer is a highly aggressive disease with a 5-year relative survival rate of~9% [1]. Greater than 90% of all pancreatic cancers arise in the epithelial ducts of the pancreas and are designated pancreatic ductal adenocarcinomas (PDA). Only 18-20% of patients diagnosed with PDA are eligible for surgical resection followed by chemo and radiation therapies. For majority of PDA patients, chemo and radiation therapies are the only choices. However, due to chemo-resistance, the overall survival rate with or without surgical resection remains dismal [2]. It is established that one of the reasons for failed therapy is the inefficient delivery of chemotherapy drugs to the tumor site, likely due to the dense stroma and deficient vascular network in the pancreatic tissue microenvironment [3,4]. Therefore, there is a pressing need to develop a novel drug delivery system for PDA that can increase the drug accumulation and uptake in a tumor specific manner [5].
Nanoparticles (NPs) modified to degrade in the tumor microenvironment or target tumor antigens are promising platforms for the targeted delivery of therapeutic drugs to specific cells and tissues [6][7][8][9]. NPs formulated from biodegradable and biocompatible polymers, such as Poly lactic-co-glycolic acid (PLGA), are being utilized increasingly in research due to their excellent systemic characteristics [10]. PLGA NPs allow for the encapsulation of a variety of hydrophobic chemotherapeutics or imaging agents, and can thereby facilitate the systemic delivery of these otherwise insoluble compounds with localization at the tumor site. This localization is the result of the enhanced permeability and retention effect (EPR), which is caused by the vasculature permeability in tumors being greater than in normal tissues, and thus provide a mechanism of selection for the NPs, as they do not penetrate into neighboring normal tissue [11,12]. The unorganized structure of the tumor and lack of lymphatic drainage prolong the retention of NPs after they escape from the leaky vasculature [13]. PLGA NPs with polyethylene glycol (PEG) displayed at the surface have been shown to increase circulatory half-lives of the NPs, while surface modification with targeting agents have been shown to aid in localization of the NPs selectively at targeted tissues [14][15][16][17]. Novel chemotherapeutic agents and combinations like FOLFIRINOX (5-fluorouracil, oxaliplatin, irinotecan, and leucovorin) or Abraxane (nab-paclitaxel, an albumincoated formulation of paclitaxel) have been developed and have seen some success [18,19]. The combination of gemcitabine and nab-paclitaxel has been shown to increase the intratumor concentration of gemcitabine by roughly three-fold in xenograft models [20,21]. Paclitaxel, a taxane agent, (PTX) is one of the most widely used anticancer drugs approved for the treatment of many types of cancer. PTX interferes with cell division by interacting with the polymer form of tubulin and promoting microtubulin assembly. This stabilizes the polymers against depolymerization, which induces M-phase cell cycle arrest and cell death [22,23]. Targeted NPs consisting of PLGA encapsulated PTX will provide a drug delivery system that would increase delivery of PTX to the tumor site, due to the EPR effect [24]. Systemic administration of the drug loaded NPs, however, have many problems associated with it. For instance, if the NP is too large, issues can arise that prevent them from reaching the tumor site, as the NPs have to cross through several biological barriers, such as blood vessels, tissues, organs, and cells. Without any specificity for the tumor site, it may be necessary to use fairly high doses of NPs and drugs to achieve sufficient local concentrations. In PDA, due to poor vascularization and despoplasia, non-targeted NPs may not suitable. Conjugating the NPs with tumor targeting moieties could possibly overcome some of these challenges.
Mucin-1 (MUC1), is a transmembrane protein with an extracellular domain that is heavily glycosylated [25]. It is normally expressed on epithelial cells of the mammary gland, esophagus, stomach, duodenum, uterus, prostate, lung, and pancreas [26]. In healthy tissues, the negatively charged glycosylated extracellular domain of MUC1 creates a physical barrier and an anti-adhesive surface, preventing pathogenic colonization [27]. In over 80% of PDA, MUC1 is hypoglycosylated and overexpressed [28] which in turn is also associated with higher metastasis and poor prognosis [29,30]. These characteristics ranked MUC1 as one of the best tumor antigens for targeted therapy [31]. We have developed a novel monoclonal antibody, TAB004 (OncoTAb, Inc., Charlotte, NC), which specifically targets the hypoglycosylated form MUC1 (tMUC1) [32][33][34].
This study was aimed at investigating the targeting ability of TAB004 conjugated PLGA NPs in vitro and in vivo. NR, FDA, and ICG were used as the imaging agents and PTX as the chemotherapeutic drug. We hypothesized that the conjugation of TAB004 to the surface of PLGA NPs will increase their accumulation and duration at the tumor site and thereby increase the overall therapeutic index of the treatment. For this purpose, PTX, ICG, or FDA&NR were encapsulated in PEGylated PLGA NPs and then conjugated to TAB004. Unconjugated particles were used as controls. Internalization, retention, and therapeutic efficacy were evaluated in vitro in several MUC1 high and low expressing human PDA cell lines.

Determination of NP loading
For the paclitaxel NP (PTX) formulation a 20 mg sample of was dissolved into 600 μl of DMSO-d6 and the concentration of the respective cargo determined using 1

Determination of NP size and polydispersity
Particle size, polydispersity index (PDI), along with zeta potential were determined by dynamic light scattering (Zetasizer Nano, Malvern Instruments) Table 1.

Cargo release profiles
Release profiles of NPs were modeled using FDA NPs. The release characteristics of these particles were characterized in phosphate buffered saline (PBS) at pH 7.4.

Synthesis of PCL 14K -PEG 1000
PCL 14K -PEG 1000 was prepared according to the following procedure. Polycaprolactone (2 g, M w~1 4,000) was added to a 50 ml oven dried round-bottom flask fitted with a claisen adapter and equipped with a magnetic stir bar, a rubber septum, and a reflux condenser with attached drying tube. To this was added 20 ml of thionyl chloride via syringe, and the rubber septum replaced with a ground-glass stopper, and the resulting solution heated to reflux for 3 h. The thionyl chloride was then removed under reduced pressure using a rotary evaporator. The resulting residue was placed under a nitrogen atmosphere and 50 ml of freshly distilled tetrahydrofuran (THF) was added by cannula followed by PEG 1000 -diol (2.9 g, 20 equivalent) and triethylamine (

Synthesis of PCL 14K -PEG 1000 -NH2
PCL 14K -PEG 1000 -NH2 was prepared by the according to the following procedure. PCL 14K -PEG 1000 (1 g) was added to a 50 ml oven dried 2-neck round-bottom flask equipped with a magnet stir bar and a rubber septum with nitrogen inlet. To this was added 20 ml of dry methylene chloride (DCM) followed by 1,1′-carbonyldiimidazole (100 mg, .62 mmol) and the resulting solution left to stir for 6 h at room temperature. To this was added 1,3-diaminopropane (1 ml, 12.19 mmol) and the resulting solution left to stir for 12 h at room temperature. The DCM was then removed under reduced pressure using a rotary evaporator. The resulting viscous yellow liquid was dissolved into THF (20 ml) and precipitated by pouring the solution into 250 ml of vigorously stirred DI water. The precipitate was isolated by filtration, re-dissolved into THF (20 ml), and precipitated as before. This process was repeated three times. Finally, the isolated product was dried under vacuum at 25°C for 72 h. The desired product was isolated as a yellow solid (.5 g, 50%). Although the resonances for the end-group -C(O)NHCH 2 CH 2 CH 2 NH 2 -are Nanoparticle preparation: General method Nanoparticles (NPs) were prepared by the nanoprecipitation according to the method of Langer et al. [37]. Briefly; 100 mg of PLGA (50:50, M w~2 0 K), 5 mg of PCL-PEG 1000 , 1 mg PCL-PEG 1000 -NH2, and 1 -5 mg of cargo was dissolved into 10 ml of acetone. This solution was then added dropwise via syringe into a stirred solution of 1% PVA (20 ml) at a rate of 90 ml/hr. controlled using a syringe pump. The resulting colloidal suspension was then transferred to a 100 ml round-bottom flask, and the acetone removed under reduced pressure using a rotary evaporator. NPs were then purified by centrifugation (25 min, 30,000×g) using three successive washes of sterile filtered 18 Ω water at 4°C. The resulting NP pellet was then resuspended into sterile filtered 18 Ω water (10 ml), whereupon dextrose (10 mg) was added as a lyoprotectant. This colloidal suspension was then flash frozen in liquid nitrogen then lyophilized at 25°C and 50 mTorr for 24 -48 h resulting in a flocculent solid. Paclitaxel (PTX), Fluorescein Diacetate (FDA), and Nile Red (NR) were all prepared according to the general method described above. See Table 2 for the amount of cargo used in the preparation of the respective nanomaterials.

ICG preparation
ICG NP's were prepared similarly using a modified nanoprecipitation method according to procedure reported by Cai [38]. Briefly; 100 mg of PLGA (50:50, M w~2 0 K), 5 mg of PCL-PEG 1000 , 1 mg PCL-PEG 1000 -NH2 was dissolved into 9 ml of acetonitrile. Meanwhile 1 mg of ICG was dissolved into 1 ml of sterile filtered 18 Ω water. The two solutions were then mixed, and vortexed rapidly for 2 min. The resulting solution was then added dropwise via syringe into a stirred solution of 1% PVA (20 ml) at a rate of 90 ml/hr. controlled using a syringe pump. The resulting colloidal suspension was then transferred to a 100 ml round-bottom flask, and the acetonitrile removed under reduced pressure using a rotary evaporator. NPs were then purified by centrifugation (25 min, 30,000×g) using three successive washes of sterile filtered 18 Ω water at 4°C. The resulting NP pellet was then resuspended into sterile filtered 18 Ω water (10 ml), whereupon dextrose (10 mg) was added as a lyoprotectant. This colloidal suspension was then flash frozen in liquid nitrogen, and lyophilized at 25°C and 50 mTorr for 24 -48 h resulting in a flocculent green solid.
FDA release profiles

Nanoparticle preparation and characterization
We evaluated the size and release profile of PLGA NPs to determine an optimal size for use (Fig. 1a). As shown, PCL 14K -PEG 1K and PCL 14K -PEG 1K -NH 2 partitions into the aqueous environment during self-assembly of the nanoparticles, thereby generating a nanoparticle having a pegylated surface with a small percentage of nucleophilic amines available for chemical modification. During selfassembly the cargo is encapsulated in the hydrophobic core. The functionalization of the NP surface was performed using the NuLink bis-electrophile (Fig. 1b). Fluorescein Diacetate (FDA) PLGA NPs were used as a model system to investigate the size and release profile of the NP platform described (Fig. 2a, b). In vitro cargo release of the NPs was evaluated in PBS at pH 7.4. FDA was steadily released over the course of 120 h. The percent of FDA released at 24, 48, 72, and 96 h was 24%, 37%, 50%, 59%, and 70% respectively (Fig. 2b).

PLGA NPs internalize into BxPC3.MUC1 and BxPC3.Neo human PDA cell lines
We determined whether the PLGA NPs internalizes into a human PDA cell line. Wild-type BxPC3 cells have minimal expression of endogenous MUC1. We generated BxPC3. MUC1 cells that stably express full-length MUC1. As control, we generated BxPC3.Neo that expresses the empty vector. BxPC3.MUC1 cells express high levels of MUC1 while BxPC3.Neo cells express minimal levels of MUC1 [30]. In the later experiments, this will enable us to assess the specificity of the TAB004 antibody in an otherwise genetically identical PDA cell line. BxPC3. MUC1 and BxPC3.Neo cell lines were treated for 1.5 h with FDA and Nile Red loaded NPs. This matched the total treatment time of PDA cell lines in the cell viability assay. After 1.5 h, FDA and NR loaded PLGA NPs internalized through endocytosis into both BxPC3.MUC1 and BxPC3.Neo cell lines equally (Fig. 3) suggesting that internalization is independent of MUC1 expression levels. Punctate green fluorescence, from the hydrolysis of FDA in the NPs, can be seen within the cytoplasm of the cells, indicating internalization of the NPs (Fig. 3). Fluorescence is only observed if FDA is hydrolyzed within the cells [39]. Although we detect a slight trend of increased endocytosis in the BxPC3.MUC1 versus BxPC3.Neo cells, the difference is not statistically significant.

TAB004 antibody internalizes into PDA cell lines that express tMUC1
Next we determined the specificity of TAB004 to MUC1 and quantified the internalization of TAB004 antibody by fluorescent microscopy (Fig. 4). The presence and uptake of TAB004 was visualized by conjugating the antibody to pHrodo Red, which is non-fluorescent outside the cell, but fluoresces red only post endocytosis (Fig. 4a). The green fluorescence is wheat germ agglutinin that stains the cell membrane. The fluorescent signal from TAB004 is significantly increased in BxPC3.MUC1 when compared  (Fig. 4b). There is some internalization observed in BxPC3 Neo, which can be caused by the very low level of endogenous MUC1 that is present, or by non-specific endocytosis as PDA cells have been shown to actively swallow their surroundings through macropinocytosis [40,41].
TAB004 conjugated PTX loaded PLGA NPs are specific for and inhibit growth of tMUC1 expressing cells The successful conjugation of TAB004 to the surface of the NPs (T-NPs) was determined by flow cytometry (Additional file 1: Figure S1). The linking reagent, a thioester, was tested and was successful in linking TAB004 to the NPs (Additional file 1: Figure S1B). Flow cytometry data shows a shift in fluorescence when NPs are conjugated to TAB004 and labeled with FITC conjugated anti-mouse IgG1. Unconjugated NPs did not display any shift in fluorescence (Additional file 1: Figure  S1A). NPs without the linking reagent but incubated with TAB004, or FITC anti-mouse IgG1, or both also served as controls and as expected did not show any shift in fluorescence signal. This suggests that the thioester linker was successful in conjugating TAB004 to the NPs.
We compared the internalization of unconjugated NPs to TAB004 conjugated NPs (T-NPs) in MUC1 high BxPC3.MUC1 versus MUC1 low BxPC3.Neo cells. BxPC3.MUC1 and BxPC3.Neo cells were treated with FDA loaded NPs or T-NPs and fluorescence signal quantified over time (Fig. 5). There was no significant increase in fluorescence between T-NPs and NPs in the MUC1-low BxPC3.Neo cells (Fig. 5c). However, a significant increase in fluorescence was detected in BxPC3. MUC1 cells treated with T-NPs compared to when treated with NPs. This significant increase in fluorescence was observed at 60 and 90 min post treatment. The data indicates that linking TAB004 to the NPs was highly effective in longer term retention of the NPs within the MUC1-high cells compared to NPs alone and that this retention was antigen specific.
Therefore, we next determined the cytotoxicity of PTX loaded T-NPs compared to PTX loaded NPs in the same cells (BxPC3.Neo and BxPC3.MUC1) as well as in a panel of other human PDA cell lines with varying levels of MUC1 expression and sensitivity to PTX (Fig. 6). The comparison we were interested in was between the treatment groups (NP and T-NPs) and not necessarily between the various cell lines. We selected a single dose of PTX for each cell line where at least 90% of cells remained viable post PTX treatment (~IC 10 ). We determined if there was any added cytotoxic effect of NPs or T-NPs loaded with PTX at the same concentration as the PTX alone. There was no difference in viability between PTX, NP-PTX or T-NP-PTX treated BxPC3. Neo cells (Fig. 6). This was expected based on Fig. 5c where no difference in internalization and retention was observed in BxPC3.Neo cells between NPs and T-NPs. On the other hand, in MUC1-expressing BxPC3 MUC1, MiaPaca2, and HPAC cells, we observed a significant decrease in cell viability between NP-PTX versus T-NP-PTX at a single dose (Fig. 6). This decrease was not noted in HPAF-II cell even though these cells express high levels of MUC1 (Fig. 6). The reasons for the lack of responsiveness to T-NPs in HPAF-II cells are not currently known. Although the effect of T-NPs versus NPs is modest, it is highly significant because of PDA's high resistance to chemotherapy. shown is mean ± SEM (n = 3) and determine by two-way ANOVA and Bonferroni's post-hoc test, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 TAB004 accumulates at the tumor site and its conjugation to PLGA NPs appears to increase their accumulation in an Orthotopic PDA tumor model We demonstrated the specificity of TAB004 in vitro, but the same needed to be determined in vivo. C57BL/6 immune competent mice bearing murine syngeneic orthotopic KCM tumors [35] were injected intraperitoneally with TAB004 conjugated with ICG and imaged 24 h post injection. The KCM cells stably expressed the luciferase gene and thus bioluminescent tumors could be visualized by IVIS post luciferin injection. TAB004 localizes and persists specifically at the tumor site 24 h later (Fig. 7a-d). Images of 4 representative mice are shown. It is clear that the TAB004-ICG localizes only to the bioluminescent pancreatic tumors. The fluorescent radiant efficiency values for region of interests (ROIs) around the tumor site were acquired for TAB004-ICG injected mice and displayed significant increase over tumor bearing control mice that were not injected with TAB004-ICG (Fig. 7e).
Next, we tested ICG loaded NPs and ICG loaded T-NPs in mice bearing the same KCM bioluminescent orthotopic tumors to determine if TAB004 can increase the accumulation of NPs at the tumor site (Additional file 2: Figure S2). ICG loaded NPs appear to clear from the mouse between  Figure S2), similar to the biodistribution profile of ICG loaded NPs injected in non-tumor bearing mice (data not shown). However, ICG loaded T-NPs appeared to accumulate and persist at the tumor site 24 and 48 h post injection (Additional file 2: Figure S2B). Ex vivo images of the tumor and liver of the mice were taken 48 h post injection and ICG loaded T-NPs seem to accumulate and persist in the tumor while ICG load NPs cannot be detected in the tumor post 48 h. We noted that the fluorescence in the liver was identical for ICG loaded T-NPs and ICG loaded NPs (Additional file 2: Figure S2C) suggesting that the tumor localization is extremely high for T-NPs versus NPs. Thus, TAB004 conjugated NPs may be developed as a potential platform for targeted delivery of not only PTX, but other drugs and imaging agents directly to the Fig. 6 Cell viability of PDA cell lines treated with PTX, PTX loaded NPs and PTX loaded TAB004 conjugated NPs. Concentration of PTX is 3.05 × 10 − 3 μg/ml. Data shown is mean ± SEM (n = 3) and determine by a one-sided t-test comparing treatment groups at each concentration, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.000 pancreatic tumor while reducing toxicity to other major organs. Future studies will evaluate the in vivo anti-tumor efficacy in several models of PDA.

Discussion
The ability to target drug-loaded nanoparticles to the tumor site would greatly enhance efficacy of the drug and reduce toxicity. PLGA is one of the most effective biodegradable polymers used to construct polymeric nanoparticles (NPs). It has been approved by the US FDA for use in drug delivery systems due to controlled and sustained-release properties, low toxicity, and biocompatibility with tissue and cells [42][43][44]. PEG-functionalized PLGA NPs are especially desirable, as pegylated-NP platforms have demonstrated significantly reduced systemic clearance compared with similar particles without PEG. This design parameter is important for the passive targeting of nanocarrier to tumor by the EPR effects [45]. To enhance tumor-specific targeting, in this study, we aimed to investigate PEG-functionalized PLGA NPs conjugated to monoclonal antibody TAB004. TAB004 specifically recognizes the hypoglycosylated tumor form of MUC1 [32,46,47] while sparing recognition of MUC1 on normal epithelial cells. Over 80% of PDA expresses this tumor form of MUC1 and is an established target for immunotherapy [48,49]. In Fig. 4, we show that TAB004 specifically internalizes in the BxPC3.MUC1 ells but not in BxPC3.Neo cells. Further, we showed that compared to the unconjugated NPs, TAB004 conjugated NPs had significantly enhanced and prolonged cellular accumulation in the BxPC3.MUC1 versus BxPC3.Neo cells confirming antigen specific targeted internalization (Fig. 5). This enhanced cellular internalization and accumulation of T-NPs over NPs is most likely due to the specific binding of TAB004 to tumor form of MUC1 expressed on BxPC3. MUC1 cells thus enabling the NPs to readily internalize through a process of macropinocytosis [50]. Although modest, PTX-loaded T-NPs showed significantly enhanced cytotoxicity (Fig. 6) in an antigen specific manner. The modest enhancement of cytotoxicity may be attributed to the limited time (1.5 h) of exposure of cells to the drug. Longer incubation with the nanoparticles caused degradation of the NPs, which then interfered with the OD values in the survival assay. It is well established that the antitumor effect of PTX results from its intracellular accumulation over time [51]. In vivo in an immune compromised mouse model, we observed specific localization and accumulation of TAB004 only to the orthotropic BxPC3.MUC1 tumors generated in the pancreas (Fig. 7) but not in MUC1-negative tumors or in normal epithelial organs [34]. In a pilot in vivo experiment using immune competent mice, we showed that compared to ICG-NPs, TAB004 conjugated ICG-NPs accumulated in the KCM tumor while unconjugated ICG-NPs failed to accumulate in the tumor (Additional file 2: Figure S2). Thus, we believe that the modest cytotoxic advantage observed in vitro will be significantly enhanced in vivo. Future studies will evaluate the in vivo efficacy of various drug loaded TAB004-NPs in several PDA models. Taken together the data validates the tumor specificity of TAB004 and that loaded NPs conjugated to TAB004 may be a promising nanocarrier for targeted therapy and imaging of PDA.

Conclusion
Conjugation of NPs to TAB004 greatly enhanced the internalization, retention, and targeting ability of NPs in vitro and in vivo in orthotopic models of human and mouse PDA. TAB004 conjugated PTX loaded NPs showed modest but significant increase in cytotoxicity against PDA cells in vitro. The anti-tumor efficacy of chemotherapeutic drugs in vivo will need to be investigated using this delivery platform.