Gemcitabine and doxorubicin in immunostimulatory monophosphoryl lipid A liposomes for treating breast cancer

Abstract Cancer therapy is increasingly shifting toward targeting the tumor immune microenvironment and influencing populations of tumor infiltrating lymphocytes. Breast cancer presents a unique challenge as tumors of the triple‐negative breast cancer subtype employ a multitude of immunosilencing mechanisms that promote immune evasion and rapid growth. Treatment of breast cancer with chemotherapeutics has been shown to induce underlying immunostimulatory responses that can be further amplified with the addition of immune‐modulating agents. Here, we investigate the effects of combining doxorubicin (DOX) and gemcitabine (GEM), two commonly used chemotherapeutics, with monophosphoryl lipid A (MPLA), a clinically used TLR4 adjuvant derived from liposaccharides. MPLA was incorporated into the lipid bilayer of liposomes loaded with a 1:1 molar ratio of DOX and GEM to create an intravenously administered treatment. In vivo studies indicated excellent efficacy of both GEM‐DOX liposomes and GEM‐DOX‐MPLA liposomes against 4T1 tumors. In vitro and in vivo results showed increased dendritic cell expression of CD86 in the presence of liposomes containing chemotherapeutics and MPLA. Despite this, a tumor rechallenge study indicated little effect on tumor growth upon rechallenge, indicating the lack of a long‐term immune response. GEM/DOX/MPLA‐L displayed remarkable control of the primary tumor growth and can be further explored for the treatment of triple‐negative breast cancer with other forms of immunotherapy.


INTRODUCTION
The engineering of the tumor immune response has rapidly become an integral part of cancer therapies. Treatments such as checkpoint inhibitors have significantly improved patient prognosis in late-stage non-small cell lung cancer 1 and melanoma. 2 Studies have shown that breast cancer, while traditionally considered immunologically cold, 3 may also manifest host antitumor immune responses that may be amplified through use of immunotherapy. 4,5 However, few clinical trials of checkpoint inhibitor monotherapy in the treatment of triple negative breast cancer have demonstrated substantial efficacy. 6 The mechanisms by which breast cancer cells escape immune recognition are still not fully recognized, but include recruitment of suppressive immune cells such as regulatory T cells and tumor-associated macrophages, as well as the secretion of immune inhibitory cytokines. 7 Breast cancer subtypes also express relatively low levels of tumor antigens, which makes recognition difficult for activated cytotoxic Tcells. 8 The use of immune adjuvants to boost recognition of otherwise poorly immunogenic antigens can potentially improve the immune microenvironment of breast cancer. Clinically approved immune adjuvants include oil/water emulsions, aluminum salts, and agents that activate innate immunity by binding to "Toll"-like receptors (TLRs) that recognize pathogen-associated molecular patterns. 9 One such adjuvant, monophosphoryl lipid A (MPLA), is a detoxified derivative of lipopolysaccharide (LPS) from Salmonella minnesota R595. MPLA was the first TLR adjuvant approved for clinical use and is currently licensed for use in Ceravix (human papilloma virus- 16 and -18 vaccine) and Fendrix (Hepatitis B vaccine). 10 MPLA has also been incorporated in liposomes in the malaria vaccine AS01E (or AS01B) and was shown to induce stronger cytotoxic T cell reactions than formulations that had similar composition but smaller particle size. 11 Recent work has shown MPLA to be effective in altering the tumor immune environment when used in liposomes containing immune stimulating cytokines. 12 MPLA may also sensitize breast cancer tumors to doxorubicin (DOX) treatment. 13 However, the effect of MPLA in combination with different drug pairs has not been extensively explored. The immune effects of chemotherapy have long been disregarded, as drug cocktails were administered to the point of patient myelosuppression. 14 Also, human-derived tumor cell lines are typically implanted in immunodeficient mouse models to ensure tumor growth, resulting in the development of most chemotherapy combinations without consideration of immune effects. However, in the past decade focus has shifted to understanding the immune interactions of low-dose chemotherapy with immunotherapy, and the identification of immunogenic chemotherapy combinations that can enhance immune responses. [15][16][17][18] We have recently shown very effective tumor control with gemcitabine (GEM) and DOX liposomes in the orthotopic 4T1 murine breast cancer tumor model. 19 GEM and DOX, both commonly used chemotherapeutics, were co-loaded into liposomes with lipid content representative of clinically used formulations. DOX has been reported to stimulate immunogenic cell death of tumor cells, prompting immune recognition and activation, 20 and GEM has been shown to restrict myeloid-derived suppressor cells while promoting antigen cross-presentation in dendritic cells. 21

GEM/DOX liposome fabrication
Liposomes (40 μmol, molar ratio 56.4% DSPC, 5.3% DSPE-mPEG2000, 38.3% cholesterol) were made by the conventional thin-film hydration method. When making MPLA liposomes, 0.5 mg of MPLA was incorporated as well. The lipids were dissolved in chloroform and added to a dry round-bottom flask. The lipids were dried under reduced pressure and heating to produce a thin lipid film. The lipids were then resuspended using 75 mg/ml GEM in 1.1 ml of ammonium sulfate buffer (250 mM, pH 5.5) and hydrated at 70 C for 30 min, followed by extrusion through a 50 nm polycarbonate membrane to create liposomes of similar size. Then, a pH gradient was created through the removal of extra-liposomal ammonium sulfate salts and unencapsulated GEM by PD-10 size exclusion columns from GE Healthcare (Chicago, IL). The pH gradient served to actively load DOX (20 mg/ml, 50 μl) at 65 C for 30 min. During this step, 100 μl of 95 mg/ml GEM was also added to reduce GEM loss from diffusion. Then, unencapsulated drugs were removed once more by size exclusion chromatography.

Liposome characterization
Samples were diluted 10-fold in 9:1 methanol: water with 0.05% trifluoroacetic acid. After a brief sonication, MPLA was detected by reverse phase HPLC. A Zorbax 300Extend C18 3. Liposomal size and zeta potential were measured by dynamic light scattering using a Malvern Zetasizer. Size was obtained from the number distribution. In order to detect drug content, samples were diluted 10-fold in 1:1 methanol: acetonitrile with 0.05% trifluoroacetic acid (n = 3). Samples were then sonicated in a water bath for 30 min and centrifuged for 5 min. Sample supernatant was then analyzed for drug concentration by reverse phase HPLC. The Zorbax column used previously in the detection of MPLA was equilibrated with 0.5 ml/min 99% mobile phase A (water with 0.1% trifluoroacetic acid) and 1% mobile phase B (acetonitrile with 0.1% trifluoroacetic acid). Sample (10 μl) was injected at this composition. After injection, the gradient changed to 60% mobile phase B at 10 min. The solvent composition reverted to 1% mobile phase B at 15 min and was maintained until the end of the run at 20 min.
Liposomal release was measured using Amicon Ultra mini dialysis filters purchased from Millipore Sigma. A 10-fold dilution (100 μl) of the liposomes was placed into the mini dialysis filter (n = 5), which was installed over a reservoir of PBS. Samples were kept under constant shaking at 37 C. At each timepoint, the PBS reservoirs were replaced to maintain sink conditions. Released drug was quantified using the drug detection HPLC method described previously. Treatment efficacy was evaluated with the same tumor implantation procedure. Liposomal formulations were administered when tumors were~15 mm 3 . Treatment was administered on day 5, 9, and 16 after tumor inoculation. Tumor volume and mice weight were monitored every other day until the control group tumors reached the endpoint criteria of 1000 mm 3 , at which point the study was terminated and tumors were extracted for mass measurements. Mice body weight loss greater than 15% was also a criterion for euthanasia.

In vitro cellular assays
In performing the tumor rechallenge, tumors were established with the same implantation procedure. When tumors were~15 mm 3 in size, two injections of liposomal formulations were administered 4 days apart. Tumors were observed for~20 days, at which point 10 5 4T1 cells in PBS were injected in the opposite mammary fat pad. Mice were monitored for tumor growth and weight loss.

Tumor dissociation and immune profiling
Two days after the second administration of treatment, 4T1 tumors were extracted and weighed. Each tumor was cut into small pieces and enzymatically digested using Collagenase Type I (5 mg/ml) and DNAse I (50 U/ml) in 5 ml of HBSS buffer at 37 C for 60 min. Afterwards, the cells were passed through 70 μm cell strainers with trituration and then centrifuged and resuspended in ACK red cell lysis buffer for 2 min at room temperature. The cells were then resuspended in PBS with 50 U/ml DNAse with volume adjusted to obtain 10 6 cells/ml. One hundred microliters of the cell suspension for each tumor was pelleted and treated with blocking buffer for 30 min at room temperature in a round-bottom 96 cell plate. Blocking buffer was made by supplementing cell staining buffer (1× PBS, 3% FBS, 30 μM EDTA) with 1% CD16/32. After washing the cells once with cell staining buffer, the tumors were treated with cell marker staining antibodies (Table S1)

Statistical analysis
Statistical comparison of groups was done using a one-way analysis of variance with Tukey's multiple comparison test and Student's t-test in GraphPad Prism v5. Statistical significance was defined as *p < 0.05, **p < 0.01, ***p < 0.001.  (Table 1).

In vitro cellular activation
MPLA has been shown to increase dendritic cell activation. 23,24 Both blank liposomes and liposomes with~5 μg/ml MPLA were administered to JAWSII immature murine dendritic cells. 1 μg/ml of liposaccharides (LPS) was used as a positive control for dendritic cell activation. The amount of LPS used was lower than the amount of MPLA because LPS is highly stimulating and a potential cause of decreased cellular viability. 25  had no significant differences between the two treatments ( Table 2).
The IC 50

In vivo efficacy and immune profiling
The liposomal formulations were next evaluated in vivo for immunogenicity and tumor response in the highly aggressive orthotopic 4T1 model. The 4T1 model is also regarded as immunologically cold, making it representative of human breast cancers. 29 The liposomal formu- Tumors after extraction are shown in Figure S10, and a direct comparison between the tumor masses of GEM/DOX-L and GEM/DOX/ MPLA-L is given in Figure S11.
To further investigate and evaluate the relevance of MPLA addition into GEM/DOX liposomes, we proceeded with a tumor rechallenge in the opposite mammary fat pad using the 4T1 model in BALB/c mice. As before, treatment occurred when tumors were~15 mm 3 in size. However, one notable difference in this study was that two injections of treatment were given to remain consistent with tumor immune profiling conditions. The MPLA content in this experiment was slightly lower at has not been studied extensively in combination with chemotherapy.
We used MPLA to enhance the natural immunogenicity of chemotherapeutics in a novel and translatable dual-loaded liposome with MPLA in the lipid bilayer.
Other studies have confirmed that MPLA can increase the efficacy of DOX liposomes. 13 However, the separate administration of MPLA microparticles and DOX liposomes inevitably leads to differences in pharmacokinetic profiles, which can reduce the impact of the combination. Additionally, we incorporated a second chemotherapeutic, GEM, as the combination of GEM and DOX has been tested extensively in clinical trials for breast cancer 45,46 and GEM has been shown to stimulate different anti-tumor immune responses than DOX. 21 We pursued a co-loaded DOX, GEM, and MPLA liposomal formulation to ensure controlled drug ratios and consistent MPLA concentration throughout the circulation time of the formulation. We  51 We hypothesize that the combination with chemotherapy may cause overlapping toxicity profiles, and dosing will need adjustment presumably upon translation to different animal models.
Despite initial immune activation in treated tumors, a tumor rechallenge study with 4T1 cells in the opposite mammary fat pad was not able to produce significant differences in subsequent tumor growth. Based on this, the addition of MPLA was unable to create sustained immune responses. Other treatments involving immuno-