Redirecting Transport of Nanoparticle Albumin-Bound Paclitaxel to Macrophages Enhances Therapeutic Efficacy against Liver Metastases

Despite improvements in diagnosis, surgical techniques, general patient care, and systemic therapies, most deaths from cancer result from the progressive growth of distant metastases. The liver is one of the dominant metastatic sites for primary gastrointestinal, breast, and lung cancers. Current approaches to treat liver metastasis include hepatic resection (surgery), transcatheter arterial embolization, systemic chemotherapies, molecular targeted therapies, immunotherapies, and various combinations (1). Unfortunately, patients usually do not respond sufficiently to these therapies or develop resistance, limiting 5-year survival rate to approximately 22% (2–4).

Transport of therapeutics from systemic circulation to the tumor cells substantially affects their efficacy (5, 6). Nevertheless, poor transport and tumor uptake have been often neglected as mechanisms of drug resistance in cancer therapy. Major roadblocks for drug transport consist of biophysical barriers, such as tumor-associated blood flow, vascular permeability, and tumor stroma. Majority of liver metastases are hypovascular in comparison with surrounding uninvolved liver. Tumor lesions in the liver frequently appear either hypoattenuating or isoattenuating on clinical CT scans (7–9). In our previous study, vascular permeability in the liver metastases was decreased with respect to primary tumor (10). Thus, systemically injected contrast agent or therapeutics may not be able to accumulate sufficiently in the lesion.

On the other hand, any substance administered intravenously rapidly encounters the phagocytic cells of the liver called Kupffer cells (KC). These tissue associated hepatic macrophages represent approximately 10% to 20% of all the cells in the healthy liver (11), and higher numbers are present in the close proximity of metastatic loci due to the recruitment of monocytes/macrophages from bone marrow in response to tumor-associated inflammation (12). The pronounced presence of these phagocytic cells in the proximity of the liver lesions has been utilized in the clinic for imaging liver tumors with intravenous injection of super paramagnetic iron oxide nanoparticles (SPION; ref. 13). Solid SPION particles are intrinsically taken up by the liver macrophages thus enhancing the T2 contrast around the lesion.

To overcome multiple successive biophysical barriers and carry therapeutics to the tumor microenvironment more efficiently, we have developed multistage nanovector (MSV) delivery systems composed of mesoporous silicon particles (14–16). MSV systems are comprised of several nanocomponents, or stages, which can be designed on the basis of the desired functionality (17, 18). First stage porous silicon particles can target various cell populations, such as tumor-associated endothelial cells or macrophages (17–20). The nanopores of the first stage vector are loaded with the second-stage therapeutic agent/nanoparticles. When MSV associates with the target cells in the tumor microenvironment, the second-stage therapeutic nanoparticles/macromolecules are released into the tumor lesion by time-dependent degradation of the silicon carrier. Several studies by our group have reported an increased efficiency when MSV approach was utilized for the therapy of various primary tumors (21–23).

Nanoparticle albumin-bound paclitaxel (nAb-PTX) is a clinically used nanoformulation of paclitaxel designed to overcome the toxicities associated with the Cremophor-based formulation of PTX. nAb-PTX is currently approved for the therapy of advanced breast, lung, and pancreatic tumors. This albumin-bound therapeutic targets an albumin transporter expressed on vascular endothelial cells (24), thereby, activating transport across the tumor vessel wall and allowing paclitaxel to accumulate in the tumor. Although nAb-PTX has shown better therapeutic efficacy than paclitaxel, its clinical efficiency in the case of tumor mass in the liver lesions is marginal (25).

In this study, we aimed at evaluating the therapeutic efficacy of redirecting nAb-PTX from tumor vessels to the macrophages in the proximity of the metastatic tumors in the liver by means of MSV carrier. MSV-nAb-PTX was designed and characterized in vitro and in vivo studies. In vivo studies in experimental liver metastasis of breast and lung tumor models show that the proposed approach significantly enhanced therapeutic transport to the liver and extended the survival of animals as compared with the clinically used nAb-PTX.

Discoidal MSV with 1 μm × 0.4 μm (d × h) were fabricated in a microelectronics facility via photolithography and electrochemical etching as previously described (see also Supplementary Information; ref. 17). The particles were modified with 3-aminopropyl-triethoxysilane (APTES; refs. 17, 20, 26) and lyophilized using Freezone Freeze Dry System (Labconco). nAb-PTX (Abraxane, Celgene) particles injectable suspension was loaded to lyophilized MSV in aliquots, followed by drying under low pressure system (27). MSV were fluorescently labeled with DyLight 650 (Abcam) when required (28). Release of FITC-albumin from the MSV was tested in two pHs: 4.5 and 7.4 (for cell organelles and plasma/buffer) at 37°C and 100 rpm for 1 to 96 hours. 1 N NaOH solution (10% v/v per sample) was added to all samples prior to the measurements to avoid the loss of fluorescence of FITC-albumin in acidic pH. The standard curves were plotted on the basis of the FITC-albumin solutions treated in the same manner (See Supplementary Information).

The 4T1 murine breast cancer cells and 3LL murine lung cancer cells were obtained from ATCC and used within 20 passages. The cells were cultured in Minimum Essential Medium (MEM) supplemented with 10% FBS, 1% penicillin–streptomycin, 1% GlutaMAX, 1% NEAA, 1% MEM Vitamin, 1% sodium pyruvate at 37°C in a humidified atmosphere with 5% CO₂.

Brewer thioglycollate (3 mL, Becton Dickinson) was injected into mice peritoneum 72 hours prior to cell harvest. The peritoneal cavity was then injected with 5 mL PBS containing 3 mmol/L EDTA, and the fluid was collected and centrifuged to obtain macrophages. The cells were seeded at approximately 10⁶ cells per well in 6-well plates with MEM and supplements as listed above.

Female Balb/c and C57BL/6 mice, 6 to 8 weeks old, were purchased from Charles River Laboratories. The mice were maintained in animal facilities at Houston Methodist Research Institute approved by the American Association for Accreditation of Laboratory Animal Care and in accordance with current regulations and standards of the United States Department of Agriculture, Department of Health and Human Services, and NIH (Bethesda, MD).

To establish experimental liver metastasis, breast cancer 4T1 cells (1 × 10⁵/100 μL) or lung cancer 3LL cells (3 × 10⁵ cells/100 μL) were injected into the spleen of female Balb/c or C57BL/6 mice, respectively (10). The cells injected in the spleen disseminate to the liver through portal vein, producing experimental liver metastasis (29). To evaluate therapeutic efficacy and survival of the tumor-bearing mice described, the spleen was resected 10 minutes after the injection of the tumor cells to prevent growth of primary tumor in the spleen. For determination of MSV tissue distribution, the spleen was not resected. 4T1 cells were injected into the mammary fat pad of Balb/c mice to establish orthotopic murine breast cancer.

All the surgical procedures are approved by the Institutional Animal Care and Use Committee of Houston Methodist Research Institute under the protocol AUP-0514-0032.

Alexa Fluor 488 40 kDa (green) and Alexa Fluor 538 3 kDa (red) dextrans (Molecular Probes) were injected intravenously into the retro-orbital space of the mice immediately before intravital microscopy (IVM) recording for visualization of blood flow dynamics (30). Mice were anesthetized with isoflurane and mounted on heated microscope stage during IVM imaging under Nikon A1R multiphoton microscope (Nikon Inc.). The liver was exposed by a midline incision through the abdominal wall. The fluorescent images of the tumors and unaffected (control) liver (n = 12–15) were recorded 0–60 minutes after the injection of the tracers. The images were processed for evaluation of fluorescent intensities of the tracers inside and outside the lesions at different times using NIS-Elements Image Processing Software (Nikon). The recorded fluorescent intensities in the site of interest (in arbitrary units) were then normalized to the fluorescent intensity in the unaffected liver.

Seven days after the injection of tumor cells, the mice (n = 5–8) were randomized into three treatment groups: (i) untreated control (PBS); (ii) nAb-PTX (75 mg/kg nAb-PTX, equivalent to 7.5 mg/kg paclitaxel); and (iii) MSV-nAb-PTX (75 mg/kg nAb-PTX). The therapy was administered three times via intravenous injections at 5-day interval. Twenty-four hours after the last treatment, the mice were sacrificed, and the livers were excised and weighed. Livers were either fixed in 4% formalin or embedded in OCT compound for histologic and immunohistochemical analysis. Diameter of liver metastases was assessed by light microscopy analysis of histology slides.

In survival experiments, 4T1 or 3LL liver metastases bearing mice were continuously treated with the above regiments once in every five days (n = 10) until the mice became moribund and were sacrificed.

Formalin-embedded tissues were sectioned and stained with hematoxylin and eosin. Frozen sections were fixed with 4% paraformaldehyde in PBS for 15 minutes. For staining of proliferating cells, or macrophage, the sections were incubated with anti-mouse Ki67 rabbit polyclonal antibody (Abcam), anti-mouse F4/80 or CD68 rat polyclonal antibody (AbD serotec) at 4°C overnight followed by incubation with corresponding secondary antibody, respectively. TUNEL assay was performed using TUNEL System (Promega Biosciences, LLC) according to the manufacturer's protocol. The sections were imaged under laser scanning confocal microscope (Nikon Inc.), and the images were analyzed using NIS Elements software. DAPI nuclear stain was used to define the number of cells and the border of the tumor lesions (based on the density of nuclei). Fluorescent images (n = 10–12) were captured and processed for regions of interest (ROI) using NIS elements software. To evaluate the macrophage distribution in various regions of the liver, the relative fluorescent intensities of F4/80 cells in ROI inside the tumor, tumor periphery (40–50 μm from the tumor border) and in the unaffected tissue were measured and normalized to DAPI signals (number of cells).

MSV biodistribution was assessed as previously described (17, 31). Briefly, silicon levels were quantitatively measured in the liver and major organs using inductively coupled plasma atomic emission spectrometry (ICP-AES). Tissue extracts were processed for ICP-AES analysis spiked with Yttrium as an internal standard and analyzed for silicon contents (250.69, 251.43, 251.61, and 288.158 nm). An aliquot of the MSV suspension injected was dissolved in 0.1 N NaOH overnight and tested by ICP-AES to assess the injected dose of silicon. Silicon levels were calculated as micrograms of silicon per grams of tissue and further normalized to the injected dose to obtain percentage of injected dose. For evaluation of MSV distribution in the liver tissue, 4T1 tumor in the liver bearing Balb/c mice were injected intravenously with fluorescent MSV. The mice were sacrificed 24 hours after injection and the livers were excised, embedded in OCT, and sectioned.

MTT (Sigma) and WST1 (Clonetech) assays were performed according to the manufacturer's protocols to access viability of 4T1/3LL cells and abdominal macrophages, respectively. Cancer cells and abdominal macrophages were seeded into 96-well tissue culture plates at concentrations 1 × 10³ cells/well and 1 × 10⁴ cells/well, respectively, and incubated with increasing concentrations of nAb-PTX and MSV-nAb-PTX for 72 hours. Cell viability was assessed at 570 and 440 nm for MTT and WST-1 assays.

Amount of paclitaxel internalized and released by macrophages was determined in vitro (See Supplementary Information). In brief, macrophages were incubated with MSV, nAb-PTX, or MSV-nAb-PTX for 1 and 2 days and the cells were lysed for analysis of paclitaxel internalization. In addition, the macrophages were incubated with MSV, nAb-PTX or MSV-nAb-PTX, washed and further cultured with fresh medium. After 1 and 2 days, the supernatant was measured for the amount of paclitaxel released. The amount of paclitaxel in the plasma and liver of mice injected with PBS, MSV, nAb-PTX, and MSV-nAb-PTX was also determined after 1 and 3 days of treatment and measured by liquid chromatography-tandem mass spectrometry (LC/MS-MS) analysis (see Supplementary Information).

To evaluate cytotoxic activity of paclitaxel released from macrophages, conditioned medium was added to cancer cells and cell death was evaluated by MTT assay after 3 days.

All quantitative parameters are presented as a mean with SD. Statistical analysis was performed by the Mann–Whitney U test for unpaired samples, and survival analysis was done by the Kaplan–Meier method and compared by the Log rank test using the SPSS software. P value of < 0.05 is accepted as indicative of a significant difference.

Clinically, liver lesions frequently appear hypoattenuated, pointing towards the low perfusion, extravasation, and diffusion of the contrast reagents in the tumors as compared with the uninvolved organ. To confirm that our animal models of liver metastasis are characterized by similar features, we have performed IVM imaging of intravenously injected 3 kDa and 40 kDa dextrans in the liver lesions. Figure 1A and B shows that the transport of both molecules into the loci of liver metastasis was significantly impaired. Quantitatively (Fig. 1B), the intensity of 3 kDa probe in the lesion ranged from 35% to 70% of that of the unaffected tissue, while for 40 kDa dextran, the concentration in the unaffected tissue was four times higher than inside the lesion. nAb-PTX is an albumin-bound drug and albumin has a similar hydrodynamic diameter as the high molecular weight dextran (∼10 nm). We have also confirmed that similar to the clinical findings (32), both models show increased number of macrophages in the proximity of the tumor lesions in the liver (Fig. 1C). The concentration of macrophages on the periphery of the tumor lesions in the liver was twice higher than in the uninvolved organ (39.8% vs. 21.7%, P < 0.01, respectively).

As we have previously shown that MSV particles have high affinity to macrophages in vitro and in vivo (17, 18, 20), we have further designed the MSV-nAb-PTX carrier to shift the intrinsic transport characteristics from albumin transporters involved in nAb-PTX and located on the blood vessels towards association with macrophages in the proximity of the lesions in the secondary liver tumors. The pores of MSV carrier were loaded with nAb-PTX and FITC-albumin as a model fluorescent molecule with the same physicochemical characteristics and the release profiles in aqueous media of two different pHs (7.4 and 4.5) were assessed by fluorimetry. The two pHs were chosen to represent the physiologic pH of the injection solution and the plasma and the acidic pH of cell endosomes. MSV were efficiently loaded with FITC-albumin and nAb-PTX (Fig. 2). Zeta potential (ζ) of MSV was affected by the loading. APTES-MSV are positively charged due to primary amines and ζ changes due to the loading of albumin-bound drug/probe as albumin has a net negative charge in the neutral environment. FITC-albumin loading efficiency was in the range of 80% to 94%, and the MSV were uniformly loaded with the macromolecule as can be seen from the confocal micrographs. It is known that nAb-PTX in the powder have particle diameter of 130 nm (33). However, when immersed into aqueous solution, it disintegrates into albumin aggregates that have a hydrodynamic diameter of 12.4 nm (Supplementary Fig. S1), comparable with the hydrodynamic diameter of the albumin. Thus, the particles can be efficiently loaded into the pores of MSV with diameters ranging from 20 to 50 nm. The optimized loading efficiency of nAb-PTX was 64%, which depended on the method of loading, with the most efficient one being vacuum assisted loading (as described in Materials and Methods section).

In terms of release, albumin was more prompt to be released in the normal pH and was not released from the MSV in acidic environment (Fig. 2C). This points toward the ability of MSV to protect the drug in the endosomal compartment. Our previous studies have shown that MSV can be efficiently released from the macrophages and endothelial cells by exo- and transcytosis (34).

The organ distribution of MSV and the drug in the models of experimental liver metastasis was evaluated (Fig. 3). In both mice bearing breast tumors in the liver and healthy mice, the highest concentration of MSV among the organs tested was in the liver. In the animals with liver metastasis, significantly higher concentrations of MSV in the liver were assessed as compared with the healthy mice (∼36.9% vs. 23.4%, P < 0.01). This can be related to the increased quantities of macrophages in the affected livers (Fig. 1C).

Furthermore, to assess the levels of nAb-PTX delivered intravenously in the blood plasma and livers, the mice bearing liver metastases were injected either with MSV-nAb-PTX, nAb-PTX, or PBS (untreated control) (Fig. 3B and C). One day after the administration, paclitaxel concentration in the plasma was not significantly different for nAb-PTX versus MSV-nAb-PTX. On the other hand, the amount of paclitaxel accumulated in the liver was three times higher for MSV-nAb-PTX than for nAb-PTX (P < 0.05). To confirm the localization of the MSV in the liver tissue, we have immunostained macrophages with CD68 antibody. As shown in Fig. 3D, colocalized confocal images of MSV (red) and macrophages (green) confirmed the homing of MSV into the liver macrophages.

The mice bearing experimental liver metastasis of breast and lung tumors were treated with MSV-nAb-PTX, nAb-PTX, or PBS (untreated control), three to four times once every 5 days. The mice were sacrificed and the livers were excised and processed for evaluation of the size/number of the lesions macroscopically and microscopically, weighted, and processed for evaluation of cancer cell apoptosis and proliferation. Macroscopically, we have confirmed the presence of numerous large tumor (light) lesions on the liver surface in the untreated control group (Supplementary Fig. S2). The tumor lesions appeared smaller in the nAb-PTX group, but the least number of lesions and the smallest size were observed in the tumor-bearing mice treated with MSV-nAb-PTX. Liver weight was lower in the both models treated with MSV-nAb-PTX as compared with nAb-PTX (Fig. 4A, P < 0.05). It is noteworthy that in the healthy animals, the liver weights ranged between 0.9 and 1.2 grams, similarly to what was measured for liver metastasis bearing mice following the therapy with MSV-nAb-PTX.

Furthermore, Fig. 4B shows MSV-nAb-PTX enabled a significant extension of survival of the tumor-bearing mice as compared with the other two therapies (nAb-PTX vs. MSV- nAb-PTX: P < 0.05 for 4T1 and P < 0.01 for 3LL, n = 10/each group). No changes in the number of macrophages in the unaffected liver (Fig. 4C) and body weight of the animals (data not shown) were observed, indicating both nAb-PTX and MSV-nAb-PTX were not toxic.

Histologic evaluation consequently revealed that the tumor diameters at microscopic level in breast (4T1 cells) or lung (3LL cells) liver metastases were significantly smaller in the mice treated with MSV-nAb-PTX as compared with other therapy groups (Fig. 5A and Supplementary Fig. S3). We further performed an immunohistochemical evaluation of the lesions. Percentages of Ki67-positive cancer cells in the lesions as a marker for the cell proliferation and TUNEL-positive cells as a marker for apoptosis were imaged (Fig. 5B). The number of Ki67-positive cancer cells (pink) was reduced in MSV-nAb-PTX group as compared with two other groups. The proliferation index (% proliferated cells) after MSV-nAb-PTX treatment was also significantly lower than those mice were treated with nAb-PTX (P < 0.01 in 4T1 model, P < 0.05 in 3LL model; Fig. 5C). On the other hand, as shown in Fig. 5D, the number of TUNEL-positive cancer cells (green cells) and the apoptosis index were significantly higher in MSV-nAb-PTX–treated group in both models as compared with the mice treated with nAb-PTX (P < 0.05 for both models).

The cytotoxicity of MSV-nAb-PTX and nAb-PTX to the cancer cells and primary abdominal murine macrophages was assessed using MTT and WST-1 assays, respectively (Supplementary Fig. S4). Cancer cells were less sensitive to MSV-nAb-PTX than to nAb-PTX, pointing towards slow release of nAb-PTX from the carrier. The IC₅₀ of nAb-PTX versus MSV-nAb-PTX for inhibition of the 4T1 breast and 3LL lung cancer cells growth were in the range of 0.1 to 0.5 μg/mL. In contrast, macrophages were not sensitive to the agents and their viability was slightly affected only in the 50 μg/mL concentration range. These in vitro results support the in vivo findings that macrophages in the uninvolved liver were not affected by the therapeutics (Fig. 4C).

We have then examined the capability of macrophages to serve as a depot for paclitaxel. The amount of paclitaxel internalized into the macrophages was examined 1 and 2 days after the incubation of abdominal macrophages with MSV-nAb-PTX or nAb-PTX in vitro, and determined by LC/MS-MS. The quantities of paclitaxel accumulated in macrophages treated with MSV-nAb-PTX were significantly higher than those treated with nAb-PTX (Fig. 6A).

Next, we have evaluated whether macrophages can release the internalized drug extracellularly. In this experiment, macrophages were incubated with media containing control PBS, nAb-PTX, or MSV-nAb-PTX for one or two days, then media was discarded, macrophages were washed, and fresh media without therapeutics was added to culture wells. The amount of paclitaxel released into the culture wells 24 hours after the incubation was quantified by LC/MS-MS. Higher amount of paclitaxel was released from macrophages into culture media when macrophages were preincubated with MSV-nAb-PTX as compared with nAB-PTX (Fig. 6B). Interestingly, expression of p-glycoprotein, which is associated with efflux pumps for various chemotherapeutics including paclitaxel, was increased in macrophages when incubated with MSV-nAb-PTX and nAb-PTX as compared with control media (Fig. 6C and D).

Finally, we determined paclitaxel released from macrophages preserves the cytotoxic effect on cancer cells. The supernatants containing the drug released from the macrophages were added to the cancer cells and assessed for their effect on the tumor cell proliferation. Figure 6C and D show that the macrophage condition media following the preincubation with MSV-nAb-PTX were the most efficient to inhibit the growth of 4T1 (Fig. 6E) and 3LL cells (Fig. 6F) respectively, demonstrating that the released drug is therapeutically active.

In this study, we describe a novel approach to increase the efficacy of currently used large molecular weight/nanoparticle–based agents for the therapy for liver metastases of breast and lung cancer. Currently, efficient therapies for tumors metastasizing to the liver are not available and the condition is frequently considered terminal. nAb-PTX is approved by the FDA for therapy against advanced breast, non–small cell lung and pancreatic cancers; however, its clinical efficiency against liver lesions has not been reported. It is the first of a new class of anticancer agents that incorporate albumin-nanoparticle technology and exploit the unique properties of albumin, a natural carrier of lipophilic molecules in the mammalian serum (35). The albumin-bound nanoparticle was designed to target an albumin-specific receptor (gp60) expressed on vascular endothelial cells, thereby activating transport across the tumor vessel wall, allowing paclitaxel to penetrate the endothelium and reach the tumor cells in higher concentration (35). In addition, nanotherapeutics including nAb-PTX have an advantage over free chemotherapeutics because they preferentially accumulate into tumor sites utilizing enhanced permeation and retention (EPR) effect on tumor (36). Thus, efficacy of nAb-PTX can be correlated with the levels of angiogenesis in the tumors. In contrast to the primary tumors, which generally possess angiogenesis and enhanced blood flow and are characterized by the pronounced EPR, secondary tumors growing in the liver are not well perfused with blood and their vascular permeability is limited as confirmed by our IVM imaging of the tumor-bearing mice (Fig. 1). These findings are in agreement with the strong clinical evidence of CT imaging data of patients with secondary liver tumors, which appear less perfused when contrast agents are injected intravenously. These biophysical barriers for transport of therapeutics in the liver metastases diminish the transport advantage of nAb-PTX, thus alternative strategies to increase the therapeutic efficacy of this agent is being pursued.

To overcome the barriers and produce higher drug concentrations at the tumor site in the liver, hepatic arterial infusion (HAI) through catheter has been tested. Phase I clinical study of HAI-nAb-PTX has shown safety, clinical efficacy, and higher drug concentration in the liver as compared with nAb-PTX given intravenously (37). However, the drawbacks of the HAI include the invasiveness of the procedure, need for an experienced team of surgeons, radiologists, and medical oncologists as well as possible complication of implanting the catheter into the major artery (38). Our results show that by utilizing the interactions of MSV with the liver macrophages surrounding the tumor lesions, we could enhance the transport of nAb-PTX into the liver by intravenous injection of therapeutic, which is more feasible, does not require specifically trained surgical personnel, and safer as compared with HAI-nAb-PTX. Our approach is sustained by the observation that higher numbers of macrophages were found to accumulate in the close proximity of the tumor due to the recruitment of monocytes/macrophages from the bone marrow, responding to tumor-related inflammatory signals (12). It is well known that any solid particulate administered intravenously immediately encounters and internalized by the liver macrophages. On the other hand, albumin-based conjugates do not possess similar characteristics. Thus, to increase the localization of the nAb-PTX into the liver metastasis loci by enhancing its association with macrophages, we utilized MSV systems composed of mesoporous silicon particles (14–16).

The MSV system is comprised of several stages: first stage, mesoporous silicon particles (S1MP), target elements in the tumor microenvironment, while the second stage nanoparticles (S2NP, e.g., nAb-PTX) encapsulated in the pores of S1MP are slowly released in the proximity of the tumor cells and affect these cells. Redirecting nAb-PTX transport from tumor vessels, which represent its original target and are not functional in the secondary liver tumors, to macrophages, which are abundant in the proximity of breast and lung liver metastasis should increase transport of therapeutics to tumors in the liver. Previous studies have shown that micrometastasis accumulate higher quantities of macrophages than more mature tumor lesions (39, 40), indicating that MSV-nAb-PTX could also be beneficial to prevent the development of micrometastases into macrometastasis.

The biodistribution of MSV into the various major organs of healthy and the tumor-bearing mice after intravenous injection determined that significantly higher quantities of MSV were localized in the livers of tumor bearing mice as compared with control healthy animals (36.9 vs. 23.4% of injected dose, respectively; Fig. 3A). This can be related to the tumor associated inflammation and recruitment of the monocytes from bone marrow to the area surrounding the liver metastasis. We found preferential accumulation of MSVs in the liver as compared with other organs (e.g., lungs) that also contain resident macrophages. The difference can be explained by the abundance of macrophages in the liver; macrophages represent approximately 10% to 20% of all the cells in healthy liver and 80% to 90% of whole tissue body macrophages (11). Liver macrophages comprise liver capillary wall along with endothelial cells. In contrast, resident macrophages in the other organs exist outside blood vessels or inside vessels with lesser extent (41, 42), so injected MSVs have less of a chance to encounter macrophages in the lungs and spleen as compared with liver macrophages.

To validate that macrophages serve as a depot for MSVs, we have previously tested the elimination of macrophages in the liver using clodronate liposomes (43) to reduce the accumulation of MSV in the organ. For this purpose, we used mice with primary breast 4T1 tumor grown in the mammary fat pad with the hypothesis that elimination of tissue macrophages (including liver macrophages) could significantly shift the biodistribution of MSVs towards the tumor. However, we observed that clodronate liposomes induced severe and acute toxicity in tumor-bearing animals even at the lower dose of the drug. Close to 50% of the animals were unable to overcome this toxicity and succumbed due to the drug administration. Thus, we concluded that in cancer-bearing sick animals, macrophages play a protective role for the overall animal health. Total elimination of the tissue macrophages can significantly jeopardize the well-being of the animals. The number of survivors was not sufficient for the statistical analysis, but the results showed a clear trend that elimination of liver macrophages increased the accumulation in the primary breast tumor (Supplementary Fig. S5). In addition, we have performed studies in 3D tumor spheroids using a coculture of tumor cells and macrophages. These studies show the lack of macrophages significantly hampers the therapeutic efficacy of MSV-nAb-PTX (manuscript submitted).

We have constructed MSV-nAb-PTX by loading the S1MP nanopores with nAb-PTX and evaluated the amount of paclitaxel in the systemic circulation and in the livers of the tumor-bearing mice using LC/MS-MS. Although the amount of paclitaxel was significantly higher in the liver of mice treated with MSV-nAb-PTX as compared with that in the mice treated with nAb-PTX, there was no significant difference in the paclitaxel concentration in the systemic circulation of the two groups of mice. These results indicate the potential clinical advantage using MSV-nAb-PTX to increase therapeutic efficacy for the tumors in the liver without increasing the incidence of systemic side effects. It is important to note that there was no body weight reduction in the mice treated with MSV-nAb-PTX as compared with the other groups. This is in accordance with a clinical study using HAI-nAb-PTX, which has shown increased concentration of pacliatxel in the liver without increasing peak concentration in the systemic circulation and the chance for systemic adverse events (37, 38). The 3-fold higher concentration of the drug in the affected liver delivered by MSV also suggests that treatment interval can be extended for MSV-nAb-PTX as compared with nAb-PTX to possibly enable similar therapeutic benefits.

Interestingly, macrophages in the liver were not killed by the chemotherapeutic agent as confirmed by in vivo and in vitro studies. The number of macrophages in the uninvolved liver counted by immunofluorescence imaging of macrophages in the mice treated with control, nAb-PTX, or MSV-nAb-PTX was unchanged (Fig. 4C). Macrophages maintained their viability at high concentrations of paclitaxel, which were more than hundred times higher than those required to kill more than 90% of 4T1 and 3LL cells (Supplementary Fig. S4). One of the reasons for resistance of macrophages to paclitaxel can be attributed to increased expression of p-glycoprotein, which is associated with efflux pumps on macrophages when incubated with paclitaxel (Fig. 6C and D). This is also an important factor in the ability of macrophages to excrete the accumulated drug to the tumor microenvironment. Paclitaxel enhances the polymerization of tubulin to stable microtubules and blocks cells in the G₂–M phase of the cell cycle and such cells are unable to form a normal mitotic apparatus (44). On the other hand, macrophages are terminally differentiated monocytes, a nondividing cell population (45); therefore, they may not be affected by paclitaxel.

Since, as earlier discussed, macrophages are engulfing MSV, it was important to determine whether macrophages can internalize and release the therapeutic agent from MSV in vitro (Fig. 6). Macrophages internalized and released significantly higher quantities of paclitaxel when the cells were cultured with MSV-nAb-PTX as compared with nAb-PTX. These results indicate advantage of using MSV as a carrier of nAb-PTX to enhance the transport of the paclitaxel into the liver tumor microenvironment. In the next set of experiments, following the incubation of macrophages with nAb-PTX or MSV-nAb-PTX, the cells were washed and further incubated with fresh media without therapeutics. The concentration of paclitaxel released from the cells was three times higher when the cells were preincubated with MSV-nAb-PTX as compared with nAb-PTX. Finally, we confirmed that the released drug from macrophages maintained its cytotoxicity to the breast and lung tumor cells in vitro.

To summarize the proposed mechanism of action, the efficacy of MSV-nAb-PTX as compared with nAb-PTX for the therapy of liver metastasis of breast and liver cancer can be related to redirecting the transport of nAb-PTX from tumor vessels towards the macrophages in the tumor microenvironment, which are abundant in the proximity of the lesions in the liver. We have shown that higher quantities of nAb-PTX can be transported to the liver when nAb-PTX is packed in MSV. MSV-nAb-PTX is internalized by the liver macrophages accumulating especially in the close proximity of the metastatic lesions in the liver. Macrophages can retain nAb-PTX, keep viability and release the drug, possibly through efflux pumps. The increased concentration of the therapeutic agent for the prolonged period of time in the proximity of the liver lesions governs the diffusion of the agent into the microenvironment, eventually affects the viability of cancer cells nearby, inhibits tumor cell proliferation, and induces apoptosis. MSV-nAb-PTX significantly enhanced therapeutic efficacy of nAb-PTX in terms of inhibition of tumor growth and extension of survival.

The therapeutic efficacy of agents in battling tumors can be hampered not only by the intrinsic molecular resistance mechanisms developed in cancer cells but also by the insufficient quantity of the therapeutics transported to cancer cells from the systemic circulation (5, 6). Transport of therapeutics in the primary tumors and secondary locations significantly differs and, thus, various elements have to be considered. Understanding transport differentials in oncology, termed “transport oncophysics,” can provide a novel framework to improve cancer therapy (46). We have shown that by designing a solid particle system that will retain nAb-PTX and redirect transport from tumor vessels to the relevant and abundant elements in the liver tumor microenvironment (macrophages), therapeutic efficacy of an agent can be significantly improved without increasing chance of developing systemic side effects. Our study provides strong rationale for clinical translation of MSV-nAb-PTX as a novel therapy for liver metastases. Relevant elements in the tumor microenvironment as a target for efficient transport of therapeutics can be different by the location of the secondary lesions (i.e, brain, lung, bone, lymph nodes, etc.). Although most of the tumors in the liver are characterized by the increased concentration of the macrophages, there are cases when macrophages are underrepresented in these tumor lesions. In these instances, delivering the drug with MSV is not expected to provide therapeutic benefit and the deciding therapeutic modality should be rendered on the basis of the hemodynamics in the tumors evaluated by the imaging modalities (CT/MRI with contrast reagent; refs. 6, 47). Ultimately, personalized cancer therapy should consider the transport information in the tumor lesions of the specific patient.

No potential conflicts of interest were disclosed.

Conception and design: T. Tanei, M. Ferrari, B. Godin, K. Yokoi

Development of methodology: T. Tanei, F. Leonard, X. Liu, B. Godin, K. Yokoi

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T. Tanei, F. Leonard, X. Liu, J.F. Alexander, Y. Saito, B. Godin, K. Yokoi

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T. Tanei, F. Leonard, M. Ferrari, B. Godin, K. Yokoi

Writing, review, and/or revision of the manuscript: T. Tanei, F. Leonard, M. Ferrari, B. Godin, K. Yokoi

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): T. Tanei, J.F. Alexander, Y. Saito, B. Godin

Study supervision: B. Godin, K. Yokoi

The authors thank David A. Engler (Houston Methodist Research Institute, Department of Proteomics) for LC/MS-MS analysis. M. Ferrari and K. Yokoi would like to acknowledge Ernest Cockrell Jr. Presidential Distinguished Chair. The authors also thank Ms. Hannah Kim for assistance with editorial review.

The authors gratefully acknowledge funding from the NCI 1-U54-CA143837 and NIH 1-U54CA151668-01.

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