Research ArticleNanomedicine

Preclinical Development and Clinical Translation of a PSMA-Targeted Docetaxel Nanoparticle with a Differentiated Pharmacological Profile

See allHide authors and affiliations

Science Translational Medicine  04 Apr 2012:
Vol. 4, Issue 128, pp. 128ra39
DOI: 10.1126/scitranslmed.3003651


We describe the development and clinical translation of a targeted polymeric nanoparticle (TNP) containing the chemotherapeutic docetaxel (DTXL) for the treatment of patients with solid tumors. DTXL-TNP is targeted to prostate-specific membrane antigen, a clinically validated tumor antigen expressed on prostate cancer cells and on the neovasculature of most nonprostate solid tumors. DTXL-TNP was developed from a combinatorial library of more than 100 TNP formulations varying with respect to particle size, targeting ligand density, surface hydrophilicity, drug loading, and drug release properties. Pharmacokinetic and tissue distribution studies in rats showed that the NPs had a blood circulation half-life of about 20 hours and minimal liver accumulation. In tumor-bearing mice, DTXL-TNP exhibited markedly enhanced tumor accumulation at 12 hours and prolonged tumor growth suppression compared to a solvent-based DTXL formulation (sb-DTXL). In tumor-bearing mice, rats, and nonhuman primates, DTXL-TNP displayed pharmacokinetic characteristics consistent with prolonged circulation of NPs in the vascular compartment and controlled release of DTXL, with total DTXL plasma concentrations remaining at least 100-fold higher than sb-DTXL for more than 24 hours. Finally, initial clinical data in patients with advanced solid tumors indicated that DTXL-TNP displays a pharmacological profile differentiated from sb-DTXL, including pharmacokinetics characteristics consistent with preclinical data and cases of tumor shrinkage at doses below the sb-DTXL dose typically used in the clinic.


Targeting of therapeutic agents to diseased cells and tissues was first suggested more than 100 years ago (1). Since then, disease targeting has been pursued as a strategy to develop safer and more effective drugs for myriad diseases, most notably cancer, where the mainstay of therapy often involves cytotoxic agents (2, 3). The past decade has seen the advent of molecularly targeted agents capable of inhibiting specific pathways that are required for survival, proliferation, and metastasis of diseased cells (4, 5). However, like cytotoxic drugs, these agents are subject to dose-limiting toxicities that affect their overall clinical effectiveness. Targeted nanoparticles (TNPs) have the potential to overcome the toxicity and efficacy limitations associated with traditional cytotoxic agents and newer-generation molecularly targeted drugs by trafficking a greater fraction of the administered drug directly to cancer cells in a controllable and tunable manner (6, 7).

Even though TNPs were first described more than 30 years ago (8), only a handful have entered clinical trials, and none have advanced beyond early-phase testing in humans (7). A key challenge to clinical translation of TNPs has been defining the optimal physicochemical parameters that simultaneously confer molecular targeting, immune evasion, and controlled drug release (9). This is mainly due to the complex interdependence of NP properties (composition, size, shape, rigidity, surface charge, hydrophilicity, and ligand type and density), payload properties (drug type, solubility, loading, and release kinetics), and in vivo physiological barriers to NP trafficking (immune surveillance, particle extravasation, tissue penetration, and cellular uptake) (10). Although a considerable amount of information is available regarding individual factors that affect the in vivo fate of polymeric NPs (1115), a system that achieves optimal characteristics has remained elusive (16).

Recently, combinatorial and high-throughput technologies have been explored as means to address the multifactorial challenge of NP optimization (11, 1719). For example, combinatorial biomaterial libraries have been created by introducing wide chemical diversity in the polymeric components and then screened for a specific readout, such as small interfering RNA silencing in vitro, to identify NPs with the desired in vivo activity (17, 20, 21). Here, we investigated a new approach to NP development and optimization, by which we introduced physicochemical diversity in the NP design, while restricting the particle makeup to a clinically validated set of biomaterials. This approach consisted of evaluating a combinatorial library of TNPs that varied systematically with respect to critical parameters (size, surface hydrophilicity, targeting ligand density, drug loading, and drug release) that affect pharmacokinetics (PK), biodistribution, and efficacy of the encapsulated therapeutic agent.

Here, we present a platform to optimize the design of TNPs and its use in the development of a TNP containing the chemotherapeutic docetaxel (DTXL), which was selected from a combinatorial library spanning more than 100 distinct NP compositions. DTXL-TNP was designed to target the extracellular domain of prostate-specific membrane antigen (PSMA) using a stable small-molecule ligand displayed on the NP surface. PSMA is a clinically validated transmembrane receptor that is overexpressed on the surface of prostate cancer cells and in the neovasculature of nearly all nonprostate solid tumors (22). DTXL is a semisynthetic taxane approved for treatment of several major solid tumor cancers, including breast, prostate, lung, gastric, and head and neck (23). The in vivo performance of DTXL-TNP, including PK, biodistribution, and tolerability, was tested in multiple animal species (rat, mouse, and cynomolgus monkey). In addition, efficacy in mouse xenograft models of prostate, lung, and breast cancer was examined. Finally, we initiated a clinical study to evaluate the tolerability and PK of DTXL-TNP in humans and to obtain a preliminary assessment of its efficacy in patients with advanced and metastatic solid tumor cancers.


Combinatorial synthesis and optimization of targeted DTXL-TNPs

The TNPs were composed of (i) a hydrophobic biodegradable polymeric core that allowed for the encapsulation and controlled release of drugs, (ii) a hydrophilic corona that protects the NP from immune surveillance, and (iii) a targeting ligand that mediated molecular interactions between NPs and PSMA expressed on prostate cancer cells and in nonprostate solid tumor neovasculature (Fig. 1A). The particles consisted primarily of block copolymers of either poly(d,l-lactide) (PLA) or poly(d,l-lactide-co-glycolide) (PLGA) and poly(ethylene glycol) (PEG), in which DTXL was physically entrapped without chemical modification. PLA, PLGA, and PEG are widely used in pharmaceutical products and medical devices and have a long history of safe human use (24). The NP size fell within the range reported to be retained in the vascular compartment after intravenous administration and to access tumors via their leaky vasculature by means of the enhanced permeability and retention (EPR) effect (<400 nm) (25). The NPs were targeted to the extracellular domain of PSMA using S,S-2-[3-[5-amino-1-carboxypentyl]-ureido]-pentanedioic acid (ACUPA), a PSMA substrate analog inhibitor (26). The ACUPA moiety is a constituent of a prostate cancer–imaging agent currently in clinical development (27). In addition, targeting of DTXL-TNPs with ACUPA has been shown to enhance the cytotoxicity of DTXL in a PSMA-expressing prostate cancer cell line (LNCaP) in vitro (28).

Fig. 1

Combinatorial screening and optimization of DTXL-TNPs. (A) Schematic of DTXL-TNP, a PSMA-targeted polymeric NP composed of a hydrophobic PLA polymeric core encapsulating DTXL and a hydrophilic PEG corona decorated with small-molecule (ACUPA) targeting ligands. (B) Generation of a library of DTXL-TNPs prepared by self-assembly of particles from mixtures of DTXL, PLA, PLGA, PLA- or PLGA-PEG (with varying PLA, PLGA, and PEG block lengths), and PLA-PEG-ACUPA. (C) Development and clinical translation of PSMA-targeted DTXL-TNPs. (1) A nanoemulsion process for efficiently encapsulating DTXL in NPs was developed. (2) Small-scale batches of DTXL-TNPs were prepared and evaluated with respect to drug load and encapsulation efficiency, particle size distribution and reproducibility, and in vitro release kinetics. (3) DTXL-TNPs with promising physicochemical properties were evaluated with respect to PK in rats, and tolerability, tumor accumulation, and efficacy in tumor-bearing mouse models. In vivo results informed additional formulation optimization and led to selection of DTXL-TNP composition and process. (4) The DTXL-TNP manufacturing process was scaled up and used to manufacture sterile clinical supplies under cGMPs. (D) Range of formulation parameters and NP physicochemical properties evaluated during development of DTXL-TNPs, with optimized DTXL-TNP parameters and target properties indicated by the red dotted line.

Here, we prepared a combinatorial library of ACUPA-targeted DTXL TNPs with diverse physicochemical properties using single-step self-assembly processes wherein organic solutions containing DTXL and polymeric components were contacted with an external aqueous phase, resulting in encapsulation of DTXL in solid particles. The library was prepared by combining systematically varied proportions of PLA-PEG polymer conjugated to ACUPA (PLA-PEG-ACUPA), DTXL, and PLA, PLGA, PLA-PEG, and PLGA-PEG copolymers of varying PLA, PLGA, and PEG block lengths and the PLGA ratio of glycolic to lactic acid units (Fig. 1B). These mixtures self-assembled into a core-shell structure with PEG and PEG-ACUPA chains oriented toward the NP surface, as demonstrated previously for similar NPs (29, 30). From enzyme-linked immunosorbent assay (ELISA) and surface plasmon resonance (SPR)–based binding studies, we found that NPs decorated with ACUPA bound tightly to PSMA even at mole fractions of PLA-PEG-ACUPA as low as 0.5% relative to nonfunctionalized polymeric components (Fig. 2).

Fig. 2

(A) ELISA of TNP binding to PSMA. NPs containing PLA-PEG-ACUPA mole fractions of 0 to 5% were assessed for binding to recombinant human PSMA in a competitive ELISA using a biotinylated ACUPA analog. Data are means ± SD. (B) SPR of TNP binding to PSMA. NP samples with or without 2.5 mole percent (mol %) PLA-PEG-ACUPA relative to nonfunctionalized polymeric components were compared with respect to PSMA binding.

Our approach for development, optimization, and clinical manufacturing of DTXL-TNP is depicted in Fig. 1C. Initial small-scale studies were carried out to identify a process capable of encapsulating at least 5 weight percent (wt %) DTXL and producing NPs with consistent particle size distribution across batches. Specifically, we evaluated two general processes: nanoprecipitation (in which the organic polymer-drug solution comprised water-miscible solvents, such as acetone or acetonitrile) and nanoemulsion (in which the organic phase consisted of water-immiscible solvents, such as ethyl acetate or dichloromethane). Our studies revealed that a nanoemulsion process using organic phase mixtures of ethyl acetate and benzyl alcohol in ratios between 90:10 and 70:30 (w/w) yielded stable NPs with higher DTXL content compared to nanoprecipitation or nanoemulsion processes using other solvent systems. For example, the DTXL content in NPs produced by nanoemulsion ranged from 2 to 20 wt % depending on formulation and process parameters, but did not exceed 1 wt % in NPs produced by nanoprecipitation.

We then produced a combinatorial library consisting of more than 100 NP compositions varying with respect to formulation and process parameters (Fig. 1D), including the molecular weight and composition of the polymeric components, the theoretical DTXL content, the targeting ligand content, the concentrations of NP components, and processing times, temperatures, flow rates, and mixing speeds. NPs were characterized with respect to particle size, ligand density, DTXL loading, and DTXL release kinetics. Data for 63 of these compositions, which span the range of formulation and process parameters evaluated, are compiled in table S1.

Drug release and PK of DTXL-loaded NPs

DTXL was released in vitro under physiological conditions in a sustained fashion over about 48 hours, with kinetics that depended on the NP composition and physicochemical properties. For example, Fig. 3A shows the in vitro release kinetics of three candidate NPs that varied with respect to their polymeric components and DTXL release rate. The NP composition with the fastest release rate (50% release at 6 hours) contained a mixture of equal amounts of PLA (6.5 kD) and a 28/5 PLGA-PEG copolymer (Table 1). The other two slower-releasing NPs contained either a mixture of equal amounts of PLA (6.5 kD) and a 16/5 PLA-PEG, or 16/5 PLA-PEG only. Among these formulations, the time at which 50% of the encapsulated DTXL was released varied 7.5-fold, ranging from about 6 to 45 hours.

Fig. 3

In vitro release, PK, and biodistribution of DTXL-TNPs. (A) In vitro release and PK of DTXL-TNP formulations. For PK, rats (n = 6 per group) were administered a single bolus dose of DTXL-TNPs or sb-DTXL (5 mg/kg) through a lateral tail vein. Serial blood samples were collected from each animal at various times, and total DTXL plasma concentrations were analyzed. NPs contained a mixture of equal amounts of 6.5-kD DL-PLA and a PLGA-PEG copolymer comprising 28-kD PLGA (50:50 lactide/glycolide monomer ratio) and 5-kD PEG (28/5 PLGA-PEG), a mixture of equal amounts of 6.5-kD DL-PLA and a PLA-PEG copolymer comprising 16- and 5-kD PEG (16/5 PLA-PEG), or 16/5 PLA-PEG. (B) Biodistribution in rats (n = 3 per group) of sb-DTXL and NPs in plasma, liver, bone marrow, and spleen after a single intravenous dose of 14C-labeled sb-DTXL, 14C-labeled NPs (labeled on the targeting polymer), or 14C-DTXL encapsulated in NPs. In vitro release data are means (n = 2); PK and biodistribution data are means ± SD.

Table 1

In vitro release and in vivo PK parameters of DTXL-TNPs. Time to 50% in vitro release was determined on the basis of measurements performed in duplicate. PK parameters are average values.

View this table:

DTXL-TNPs with promising physicochemical and release properties were characterized with respect to PK. Figure 3A shows a typical study in which the same three particle compositions were evaluated in healthy Sprague-Dawley rats (n = 6). A bolus dose of DTXL-TNPs (5 mg/kg) or the commercially available solvent-based DTXL formulation (sb-DTXL, or Taxotere) was injected intravenously, and blood samples were collected at different time points over the course of 12 hours to measure the plasma concentration of DTXL (Fig. 3A). After administration of sb-DTXL, plasma DTXL concentrations rapidly declined as the drug distributed to the peripheral compartment, which was consistent with the reported 0.02-hour distribution half-life (t1/2α) of DTXL in rats (31). In contrast, DTXL plasma concentrations in animals receiving the same dose of DTXL-TNP remained up to three orders of magnitude higher for the duration of the experiment, yielding areas under the concentration-time curve (AUCs) between 46- and 670-fold higher than sb-DTXL (Table 1). We observed that the rate of decay of DTXL plasma concentration in vivo depended on the NP formulation and correlated with the rate of DTXL release in vitro (for instance, slower DTXL release from NPs in vitro corresponded to slower DTXL clearance in vivo). This indicated that the DTXL PK profile could be controlled by manipulating DTXL-TNP formulation and process parameters.

Overall, DTXL release kinetics and PK were dependent on the type and molecular weight of the polymeric components, with PLA-PEG NPs displaying slower release kinetics than PLGA-PEG, as exemplified in Fig. 3. Inclusion of lower–molecular weight polymers, such as 4.5- or 6.5-kD PLA, resulted in faster release kinetics, and higher PLA/PEG ratios resulted in more rapid clearance of DTXL in vivo, presumably owing to faster particle clearance rates (Fig. 3). Finally, levels of PLA-PEG-ACUPA within the range of 0.5 to 2.5 wt % relative to nonfunctionalized polymeric components imparted increasing PSMA targeting with no additional benefit observed at higher levels (Fig. 2A).

On the basis of these findings, we selected and manufactured for clinical evaluation a DTXL-TNP composition containing 10 wt % DTXL encapsulated in 100-nm NPs (polydispersity index <0.2), with a surface charge (ζ potential) of about −6 mV (Fig. 1D). The optimal DTXL-TNP was composed of a PLA-PEG copolymer comprising 16-kD PLA and 5-kD PEG (16/5 PLA-PEG) and a PLA-PEG-ACUPA targeting polymer also comprising 16-kD PLA and 5-kD PEG, with the PLA-PEG and PLA-PEG-ACUPA copolymers representing 97.5 and 2.5% of the polymer mass, respectively (Fig. 1D).

Tissue distribution of DTXL-TNPs

To characterize the biodistribution of DTXL TNPs in vivo in rats, we examined the plasma and tissue concentrations in the major clearing organs—the liver, spleen, and bone marrow—after intravenous administration of (i) PSMA-targeted NPs with 14C-labeled polymer, (ii) PSMA-targeted NPs with 14C-labeled DTXL, or (iii) 14C-labeled sb-DTXL (Fig. 3B). Circulating NPs with the 14C-labeled polymer exhibited a decrease in plasma concentration of about twofold over the 24-hour period studied. This prolonged circulation time is indicative of limited or delayed particle clearance via the mononuclear phagocyte system, which mediates the rapid clearance of most intravenously administered polymeric NPs (25). The plasma profiles for 14C-DTXL encapsulated in NPs and 14C-sb-DTXL were consistent with the PK data presented in Fig. 3A, indicating that encapsulation of DTXL prevents its rapid distribution from the plasma compartment and results in substantially higher plasma DTXL concentrations relative to sb-DTXL. With respect to tissue distribution, lower levels of 14C-labeled NPs were detected in liver and bone marrow relative to plasma, whereas in the spleen, NP concentrations were slightly higher than plasma at 12 and 24 hours (Fig. 3B).

Tolerability of TNPs

To obtain a preliminary characterization of the toxicological properties of NPs exclusive of the known toxicities of the DTXL active ingredient, we performed a single-dose tolerability study in rats of PSMA-targeted and identical nontargeted NPs, both containing no DTXL. Rats received single intravenous doses of NPs at 1000, 1500, or 2000 mg/kg [about 6000, 9000, and 12,000 mg/m2, respectively, corresponding to up to 20 g of NP in an adult human of average body surface area (32)] and were monitored for clinical observations, changes in body weight, and blood chemistry. No abnormal clinical observations or appreciable body weight loss were observed (fig. S1), and blood chemistry parameters determined at study termination (tables S2 and S3) were within normal ranges (33). These results indicated that NPs on which the DTXL-TNP formulation was based, including the ACUPA targeting ligand, were well tolerated without clinical observations suggestive of hypersensitivity or anaphylactic reactions, which have been reported after infusion of high doses of particles (34).

Drug efficacy and tumor accumulation

To evaluate the antitumor activity of DTXL-TNPs, we performed efficacy studies in mice bearing PSMA-expressing human LNCaP prostate cancer xenografts. Twenty-nine days after inoculation with LNCaP cells, mice were treated every 4 days for a total of four doses with equivalent doses of DTXL-TNP, nontargeted DTXL-NPs (lacking the ACUPA targeting ligand), sb-DTXL, or a 10% sucrose vehicle. Treatment with DTXL-TNPs resulted in tumor regression over the 50-day period of study, outperforming nontargeted DTXL-NPs and sb-DTXL (Fig. 4A). For DTXL-TNP, tumor mass decreased by an average of 26% compared to that of nontargeted DTXL-NPs, which saw a 75% average increase. sb-DTXL–treated tumors increased by 100% over the 50-day study period. DTXL-TNPs appeared to be better tolerated on the basis of minimal body weight loss compared to nontargeted DTXL-NPs and sb-DTXL (Fig. 4B), but this was not statistically significant.

Fig. 4

Therapeutic effect of DTXL-NPs and accumulation of DTXL in mouse tumor xenograft models. (A) Efficacy of DTXL-TNP, nontargeted DTXL-NPs [DTXL-NP (NT)], and sb-DTXL in athymic nude mice with LNCaP tumor xenografts. Beginning on day 29 after inoculation, mice were administered 5 mg/kg every 4 days for a total of four doses. Data are mean tumor weights ± SEM (n = 6 per treatment group). *P < 0.05 (DTXL-TNP versus sb-DTXL and DTXL-TNP versus nontargeted DTXL-NPs at day 80), Wilcoxon test. (B) Body weight in the LNCaP efficacy study was monitored as an indicator of tolerability and is presented as mean body weight change ± SD. (C) Efficacy of DTXL NPs in MX-1 breast cancer xenografts. Beginning on day 14 after inoculation, mice (n = 10 per treatment group) were administered 10 mg/kg every 4 days for a total of three doses. Treatment groups were the same as in (A). **P < 0.01 (DTXL-TNP versus sb-DTXL on day 46; DTXL-TNP versus nontargeted DTXL-NPs; P > 0.05, not significant), Wilcoxon test. (D) Efficacy of DTXL NPs in NCI-H460 NSCLC xenografts. Beginning on day 8 after inoculation, mice (n = 10 per treatment group) were administered two doses (30 mg/kg) 7 days apart. Treatment groups were the same as in (A). Comparison of tumor weights for all DTXL treatment groups at day 46 was not significant (P > 0.05). (E) Accumulation of DTXL-TNP or sb-DTXL in LNCaP prostate xenografts. Mice (n = 6 to 10 per treatment group) were administered a single intravenous bolus dose of sb-DTXL or DTXL-TNP (50 mg/kg), tumors were excised 2 or 12 hours after dose administration, and DTXL content was quantified. Data are means ± SEM. **P < 0.01. NS, not significant.

These results suggest that, for long-circulating NPs, active targeting improves tumor response compared to passively targeted NPs. This hypothesis was further tested in tumor models that do not express PSMA. Specifically, we compared the efficacy of DTXL-TNP, nontargeted DTXL-NPs, and sb-DTXL in mouse xenograft models of breast cancer (MX-1) and non–small cell lung cancer (NSCLC) (NCI-H460). Mice were implanted with fragments of MX-1 or NCI-H460 tumors from an in vivo passage. Fourteen days after inoculation, mice bearing MX-1 tumor xenografts were administered 10% sucrose vehicle or DTXL-TNP, sb-DTXL, or DTXL (10 mg/kg) encapsulated in nontargeted NPs, every 4 days for a total of three doses. Mice bearing NCI-H460 tumors received two doses (30 mg/kg) of the same test articles 7 days apart beginning 8 days after inoculation. (These treatments varied across different tumor models owing to differences in sensitivity to DTXL.) DTXL-TNP and the corresponding nontargeted DTXL NPs were equally effective to one another in both models and, in the MX-1 breast cancer model, exhibited significantly enhanced efficacy over sb-DTXL by drastically and durably shrinking tumors, whereas sb-DTXL resulted in only temporary cessation of tumor growth (Fig. 4, C and D). Although not statistically significant, a similar trend was observed in the NCI-H460 lung cancer model. The similar effects of targeted and nontargeted NPs in non–PSMA-expressing tumors were presumably due to increased tumor drug accumulation by means of the EPR effect.

To verify the tumor-targeting capabilities of DTXL-TNPs, we measured intratumoral DTXL concentrations in mice bearing LNCaP prostate cancer xenografts at different time points after intravenous administration of DTXL-TNPs and compared to the same dose of sb-DTXL (Fig. 4E). Tumors were excised 2 and 12 hours after dosing, and the concentration of DTXL was determined by liquid chromatography–mass spectrometry (LC-MS). At 2 hours, similar levels of DTXL were found for both DTXL-TNP and sb-DTXL. However, at 12 hours, the DTXL concentration in tumors had decreased appreciably in the sb-DTXL group, whereas the DTXL concentration continued to increase in the DTXL-TNP group such that the concentration of DTXL in the tumor was about sevenfold higher than that in the sb-DTXL group (21.2 ± 5.2 versus 3.30 ± 0.76 ng of DTXL per milligram of tumor mass ± SEM). Overall, these results demonstrate that a combination of long NP circulation time, NP targeting, and controlled release of DTXL markedly increases the amount of DTXL delivered to the tumor.

PK across different animal species

A key challenge commonly encountered in drug discovery is that results obtained with one animal species do not necessarily translate across different species (32). Therefore, we compared the PK of DTXL-TNP and sb-DTXL in tumor-bearing mice and in healthy rats and nonhuman primates (Fig. 5A). The PK of DTXL in the TNPs was consistent across species, maintaining substantially higher plasma concentrations compared to sb-DTXL for at least 48 hours. Plasma concentration profiles in all species were consistent with retention of DTXL-TNPs in the plasma compartment and the controlled release of DTXL.

Fig. 5

PK of DTXL-TNP and sb-DTXL in tumor-bearing mice, rats, and cynomolgus monkeys. (A) Single-dose intravenous PK studies were performed in tumor-bearing mice (n = 4 per group) administered DTXL or DTXL-TNP (10 mg/kg), Sprague-Dawley rats (n = 4 per group) administered sb-DTXL or DTXL-TNP (5 mg/kg), and cynomolgus monkeys (n = 2 per group) administered sb-DTXL or DTXL-TNP (2.1 mg/kg). (B) Dose proportionality of DTXL-TNP PK in cynomolgus monkeys (n = 2 to 4 per dose level) administered doses ranging from 5 to 50 mg/m2. r2 > 0.99 for Cmax and AUC versus dose. (C) PK of DTXL-TNP in cynomolgus monkeys after repeat dosing (n = 4 per treatment group). Animals were administered DTXL-TNP (13.5 mg/m2) once a week for 3 weeks. Total DTXL plasma concentrations were quantified after the first and third doses. (D) Total and encapsulated DTXL in cynomolgus monkeys receiving DTXL-TNP. Animals (n = 4) were administered a single intravenous dose (13.5 mg/m2) of DTXL-TNP, and total and encapsulated DTXL concentrations were quantified. For all graphs, data are means ± SD.

In addition to assessing interspecies PK profiles, we evaluated dose proportionality and the effect of repeat dosing on the PK of DTXL-TNP in cynomolgus monkeys. Both the maximum DTXL plasma concentration (Cmax) and AUC of DTXL-TNP showed excellent dose proportionality over the range of doses tested (r2 > 0.99) (Fig. 5B). With respect to repeat dosing, DTXL-TNP was administered to cynomolgus monkeys on a weekly schedule and the PK profiles after administration of the first (day 1) and third (day 15) doses were evaluated (Fig. 5C). There was no appreciable change in the PK profile from repeat exposure to DTXL-TNP, indicating that at the dose tested and within the time frame of the study, the PK of DTXL-TNP was not affected by saturation of particle clearance mechanisms and that repeated exposure to DTXL-TNP did not result in any detectable acceleration or inhibition of particle clearance.

The measurement of the plasma DTXL levels in these studies reflects the total DTXL concentration (released DTXL plus DTXL remaining encapsulated in NPs). To characterize the PK of DTXL-TNP, we developed an analytical method to measure the concentration of encapsulated DTXL and then analyzed plasma samples collected from cynomolgus monkeys during the PK study described above. The concentration profiles of total and encapsulated DTXL were nearly superimposable (Fig. 5D), indicating that essentially all of the measured DTXL was encapsulated in NPs. This is consistent with the low toxicity of DTXL-TNP observed in animal models, despite plasma exposure levels of total DTXL two to three orders of magnitude higher than sb-DTXL (Fig. 5A).

PK and initial efficacy in humans

Given the promising and favorable preclinical results obtained with DTXL-TNP, we initiated a phase 1 human clinical trial (NCT01300533) where DTXL-TNP was given by intravenous infusion every 3 weeks to patients with advanced or metastatic cancer. This study is currently in progress and full results will be reported in a subsequent publication. The primary objective of the phase 1 study is to assess the dose-limiting toxicities and to determine the maximum tolerated dose of DTXL-TNP. Secondary goals of the study are to characterize the PK of DTXL-TNP administered on day 1 and to assess preliminary evidence of antitumor activity through imaging evaluation based on Response Evaluation Criteria in Solid Tumors (RECIST) (35). In view of the widespread expression of PSMA in solid tumor neovasculature, the activity of sb-DTXL in multiple important cancer indications, and the differentiated efficacy of DTXL-TNP in preclinical models of both prostate and nonprostate solid tumors, we chose to conduct the trial in all solid tumors instead of restricting the population to prostate cancer patients. The clinical trial uses a standard dose-escalation study design in which patients are assigned to cohorts receiving progressively higher doses until a dose is reached at which dose-limiting toxicities are observed. For these studies, DTXL-TNPs were prepared under current Good Manufacturing Practices (cGMPs) (36) using a multi-kilogram manufacturing process that was capable of producing clinical and commercial quantities of sterile drug product while maintaining the same physicochemical properties as the NPs produced at the bench scale (fig. S2).

Interim data for patients receiving doses up to 75 mg/m2 indicate that DTXL-TNP displayed pharmacological properties differentiated from sb-DTXL and consistent with the preclinical data we described for the animal models. Figure 6A shows a comparison of the mean PK profile for DTXL-TNP in three patients versus published data for sb-DTXL in solid tumor patients at a dose of 30 mg/m2 (37). The PK of DTXL-TNP in humans closely resembled the PK observed in preclinical species, in that the plasma levels of DTXL administered as DTXL-TNP were at least two orders of magnitude higher than those of an equivalent dose of sb-DTXL for all time points >1 hour after administration. Furthermore, the high plasma concentrations of total DTXL persisted for at least 48 hours, potentially enabling a larger fraction of the administered DTXL to traffic to tumor sites. Comparison of the DTXL plasma concentration profiles at DTXL-TNP doses ranging from 3.5 to 75 mg/m2 indicated that PK was essentially dose proportional (correlation coefficients for Cmax and AUC versus dose were 0.87 and 0.79, respectively; Fig. 6B).

Fig. 6

PK and efficacy of DTXL-TNP in humans. (A) PK in patients with advanced solid tumors of DTXL-TNP at a dose of 30 mg/m2 (n = 3) compared to published sb-DTXL data at the same dose (n = 3). Data are means ± SD. (B) PK profiles over the first 8 hours after single-dose administration of DTXL-TNP (3.5 to 75 mg/m2). n = 1 patient per dose level at 3.5 and 7 mg/m2, n = 2 at 15 and 75 mg/m2, and n = 3 at 30 and 60 mg/m2. Cmax versus dose r2 = 0.87; AUC versus dose r2 = 0.79. (C) Axial images from contrast-enhanced CT scans obtained from a 51-year-old male cholangiocarcinoma patient with lung metastases at baseline and at day 42 after two treatment cycles of DTXL-TNP (15 mg/m2). Red circles indicate locations of metastatic lesions observed in the baseline scan. (D) Coronal images from contrast-enhanced CT scans obtained from a 63-year-old male patient with tonsillar cancer at baseline and at day 42 after two treatment cycles of DTXL-TNP (30 mg/m2). Target tonsillar lesion is outlined in red.

Finally, early signs of DTXL-TNP activity have been observed in patients receiving doses below the dose of 75 mg/m2 at which sb-DTXL is typically administered (23), with two patients exhibiting stable disease at doses of ≤30 mg/m2. For example, computed tomography (CT) images of a 51-year-old male patient with metastatic cholangiocarcinoma receiving DTXL-TNP (15 mg/m2) revealed disappearance or shrinkage of multiple lung metastases after two treatment cycles (Fig. 6C). This result is unexpected because sb-DTXL has only minimal activity in cholangiocarcinoma (38), and the dose of DTXL-TNP was only 20% of the typical dose for sb-DTXL (75 mg/m2). Similarly, after two dosing cycles at 30 mg/m2, we observed 25% shrinkage of a target tonsillar lesion in a 63-year-old patient with tonsillar cancer (Fig. 6D).


It is widely believed that nanotechnology will have a revolutionary impact on the treatment of cancer (39). A particularly promising application of nanotechnology in cancer therapy is the use of TNPs to enhance the accumulation and uptake of anticancer agents at specific sites, thus limiting exposure to off-target tissues and improving the therapeutic index of conventional chemotherapeutic drugs and molecularly targeted therapeutics (40). Despite this promise, the translation of these technologies from basic research into improving human health has remained elusive (41). A major obstacle preventing this translation is the inability to achieve the optimal interplay of physicochemical parameters that confer tumor targeting, evasion of particle clearance mechanisms, and controlled drug release (42). Our approach has been to first select a biomaterial with favorable pharmaceutical properties and a long history of safe use in humans and then introduce physicochemical diversity in the formulation by producing NPs of distinct size, targeting ligand density, surface hydrophilicity, drug loading, and drug release kinetics. To form the NPs, we used a reproducible, scalable, one-step self-assembly process that favors the orientation of hydrophilic PEG chains and targeting ligands toward the particle surface (11).

The optimal DTXL-TNP composition displayed properties previously unobserved in polymeric NPs, such as a blood circulation half-life approaching 20 hours. Although long-circulating liposomes have been described, they generally lack the particle stability, high drug loading, and controlled-release properties of polymeric NPs (43). Moreover, minimal liver accumulation was observed, which was unexpected considering the predominant role of the liver in clearance of previously described polymeric NPs (25, 4447). The in vitro release and in vivo PK results suggest that DTXL-TNPs largely remained in circulation and gradually released DTXL at a rate determined by the NP composition. These characteristics resulted in high initial levels and a gradual decay in the DTXL plasma concentration over time, yielding an increase of up to 1000-fold in the concentration of DTXL in the plasma that was available to access tumor sites compared to a conventional solution formulation of DTXL. These differentiated PK properties are the product of systematic parallel screening of multiple NP parameters to arrive at the optimal combination of precisely controllable physicochemical characteristics.

Among nanotechnology therapeutic products currently approved for clinical use, liposomal drugs and polymer-drug conjugates are two dominant classes (48). The first-generation NP therapeutic products for cancer therapy function by accumulating in tumor tissue through the EPR effect and releasing their payload in the extravascular tumor tissue. Tumor tissue accumulation is a passive process that requires a long circulating half-life to facilitate time-dependent extravasation through the leaky tumor microvasculature. This process is largely mediated by the physicochemical properties of the NPs and not by active targeting. However, once particles enter the tumor tissue, their retention and specific uptake by cancer cells could potentially be facilitated by active targeting and receptor-mediated endocytosis. This process can result in higher intracellular drug concentration and increased cellular cytotoxicity. Accordingly, our studies showed that long-circulating DTXL NPs targeting PSMA significantly enhanced the antitumor efficacy when compared to equivalent nontargeted NPs in the PSMA-expressing LNCaP prostate cancer xenograft model.

Here, we show in mice, rats, and nonhuman primates that DTXL administered as DTXL-TNP remains in the plasma at concentrations at least an order of magnitude higher than equal doses of sb-DTXL over periods of at least 24 hours, and that in mouse xenograft models, DTXL-TNP enhances the trafficking of DTXL to solid tumors, thereby improving its efficacy. The interim clinical data for DTXL-TNP are consistent with these observations. The preclinical results suggest that DTXL-TNP may potentiate the current treatment of solid tumor cancers by enhancing the therapeutic index of DTXL through PSMA targeting and controlled drug release of DTXL in the tumor microenvironment. Because PSMA is expressed on prostate tumor cells and on the neovasculature of nonprostate solid tumors, the potential exists for DTXL-TNP to have utility in the treatment of many solid tumors in which DTXL has activity but confers modest clinical benefit.

The key factors that contributed to our ability to translate this targeted nanotherapeutic platform from the bench to human clinical trials included the use of polymeric materials already used in pharmaceutical products and the development of a robust, scalable process for NP formation that enabled optimization to yield a clinical candidate NP with differentiated preclinical pharmacological properties and subsequent cGMP manufacturing in kilogram quantities. To maximize the potential of DTXL-TNP and other similar therapies, a number of areas are worthy of further exploration, including optimization of the dosing schedule, which in the phase 1 trial is based on the 3-week treatment cycle widely used for sb-DTXL, identification of patient populations most likely to benefit from treatment with DTXL-TNP based on, for example, PSMA expression levels, or assessment of the impact of PSMA targeting on activity in DTXL-resistant tumors or in androgen-resistant prostate cancer cells. Looking beyond DTXL-TNP, the presented approach of screening and optimizing physicochemically diverse, targeted NPs comprising clinically validated biomaterials are potentially applicable to other agents, including cytotoxic drugs and molecularly targeted therapeutics, for which preferential tumor accumulation can increase efficacy and decrease toxicity.

Materials and Methods

Targeting polymer preparation

The PLA-PEG-ACUPA targeting polymer was synthesized from the allyl-protected ACUPA precursor (S)-diallyl 2-(3-((S)-1-(allyloxy)-6-amino-1-oxohexan-2-yl)-ureido) pentanedioate, trifluoroacetate salt (Organix Inc.), as described in the Supplementary Materials.

NP preparation

NPs were prepared by a single-emulsion method (otherwise called nanoemulsification) (24). Briefly, polymers and drugs were dissolved in an organic phase (typically a mixture of ethyl acetate and benzyl alcohol) and combined with a water phase undergoing high-energy emulsification with a microfluidizer (Microfluidics). The resulting NP suspension was diluted with an aqueous polysorbate 80 solution, purified and concentrated with tangential flow ultrafiltration/diafiltration (Millipore), and then stored as a frozen suspension in a 10% aqueous sucrose solution. NP characterization methods are described in the Supplementary Materials.

In vivo studies

All procedures were approved by the appropriate Institutional Animal Care and Use Committee before initiation. All animals used in PK, biodistribution, and efficacy studies were allowed to acclimate for at least 72 hours in the respective animal facilities before experimentation. Animals were exposed to a 12-hour light/dark cycle and received food and water ad libitum throughout the studies. Dosages of DTXL NPs are expressed as the quantity of DTXL administered. For quantitation of DTXL plasma concentration, blood samples were collected into lithium heparin tubes and plasma was generated. Total DTXL was extracted from plasma with a supported liquid extraction with methyl tert-butyl ether and analyzed by LC-MS.

PK studies

Mice. Female athymic nude mice (6 to 8 weeks of age) were implanted with 1 × 106 MX-1 breast cancer cells [National Cancer Institute (NCI)] in 50% growth medium and 50% Matrigel (BD Biosciences) by subcutaneous injection in the right flank. When tumors were about 200 mg in weight, mice were randomized (n = 4 per group) such that the mean and median tumor weights were similar between groups. Mice were then administered a single intravenous dose of sb-DTXL or DTXL-TNP (10 mg/kg) diluted in 0.9% sterile saline. At various times, a group of four mice was euthanized and blood collected by cardiac puncture was analyzed for DTXL plasma concentration.

Rats. Male Sprague-Dawley rats (about 6 weeks; n = 4 to 6 per group) with indwelling jugular vein cannulae were administered DTXL-TNP or sb-DTXL (5 mg/kg) diluted as appropriate in 0.9% saline. At various times, serial bleeds were collected from the jugular vein cannulae and analyzed for DTXL plasma concentration. NP tissue distribution studies in rats are described in the Supplementary Materials.

Nonhuman primates. Cynomolgus monkeys were administered DTXL-TNP or sb-DTXL by 30-min infusion through a saphenous vein. Doses ranged from 5 to 50 mg/m2. At various times after dosing, serial bleeds were collected from a femoral vein and analyzed for total DTXL plasma concentration. To evaluate the PK after multiple-dose administrations, animals were administered three weekly doses of DTXL-TNP (13.5 mg/m2) and serial blood samples were collected after the first and third doses and analyzed for total DTXL concentration. To evaluate the amount of DTXL remaining encapsulated in NPs, plasma samples collected from cynomolgus monkeys administered a single dose (13.5 mg/m2) were used. Encapsulated DTXL was separated from released DTXL with solid-phase extraction. Plasma samples were applied to hydrophilic-lipophilic balance (HLB) 96-well plates (Waters Corp.) to remove free DTXL before LC-MS analysis.

Efficacy and DTXL tumor accumulation studies

LNCaP prostate tumor xenografts. Male athymic nude mice (about 6 weeks) were subcutaneously inoculated in the right flank with 2 × 107 human LNCaP prostate cancer cells (American Type Culture Collection) resuspended in 50% growth medium and 50% Matrigel. Twenty-nine days after inoculation, the mean tumor weight was about 200 mg. Animals were randomized into groups of six mice such that the mean tumor weights were similar between groups. Mice were administered 10% sucrose (vehicle control) or DTXL-TNP, sb-DTXL, or DTXL (5 mg/kg) encapsulated in nontargeted NPs every 4 days for a total of four doses. Tumor measurements and body weights were monitored twice weekly beginning on the first day of treatment. Tumor volume was determined by caliper measurements with the formula for an ellipsoid sphere (L × W2/2 = mm3), where L and W refer to the larger and smaller perpendicular dimensions collected at each measurement. Tumor volume was converted to tumor weight assuming unit density (1 mm3 = 1 mg).

To measure tumor accumulation of DTXL in prostate cancer xenografts, male severe combined immunodeficient (SCID) mice bearing LNCaP prostate tumors (n = 6 to 10 per treatment group) were administered a single intravenous dose of DTXL (50 mg/kg) as either DTXL-TNP or sb-DTXL. Mice were euthanized 2 or 12 hours later. The tumors from each group were excised and homogenized. DTXL was extracted with supported liquid extraction and quantified by LC-MS, as described for the PK studies.

MX-1 breast and NCI-H460 NSCLC tumor xenografts. Female athymic nude mice (about 6 weeks) were implanted subcutaneously near the right flank with a 30- to 40-mg fragment of MX-1 human mammary or NCI-H460 NSCLC (American Type Culture Collection) tumors from an in vivo passage. Mean tumor weights were about 300 mg 14 days after MX-1 inoculation and 130 mg 8 days after NCI-H460 inoculation. Animals were randomized into groups of 10 mice such that the mean tumor weight was similar between groups. Mice were administered 10% sucrose (vehicle control) or DTXL-TNP, sb-DTXL, or DTXL encapsulated in nontargeted NPs. Mice bearing MX-1 tumors received doses of 10 mg/kg every 4 days for a total of three doses. NCI-H460 mice received two doses (30 mg/kg) 7 days apart. Tumor measurements and body weights were monitored twice weekly beginning on the first day of treatment. Tumor weight was determined as described for LNCaP xenograft studies.

Human clinical study

DTXL-TNP (designated BIND-014) is currently undergoing evaluation in a phase 1 clinical trial (NCT01300533), where DTXL-TNP is given by intravenous infusion every 3 weeks to cancer patients. The main eligibility criteria are ≥18 years old; advanced or metastatic cancer for which no standard or curative therapy exists; measurable or evaluable disease per RECIST version 1.1 (35); Eastern Cooperative Oncology Group performance status of 0 or 1; and life expectancy of >12 weeks. The clinical trial uses a standard dose-escalation design in which patients are assigned to cohorts receiving progressively higher doses until a dose is reached at which dose-limiting toxicities are observed. Single-patient cohorts are enrolled at low-dose levels; subsequent dose levels are enrolling three-patient cohorts. Blood samples for PK analysis are analyzed for total DTXL concentration with LC-MS. All patients provide written informed consent before participation in the study.

PK and statistical analyses

PK parameters were assessed with Phoenix WinNonlin (version 6.0). Statistical comparison of efficacy and tumor accumulation data used the Wilcoxon/Kruskal-Wallis test (JMP 9, SAS Institute Inc.).

Supplementary Materials

Materials and Methods

Fig. S1. Tolerability of targeted and nontargeted NPs in rats.

Fig. S2. Clinical-scale manufacturing of DTXL-TNPs.

Table S1. Combinatorial synthesis of DTXL-TNPs.

Table S2. Clinical chemistry measurements in rats (n = 3 per group) treated with nontargeted placebo NPs.

Table S3. Clinical chemistry measurements in rats (n = 2 to 3 per group) treated with PSMA-targeted placebo NPs.

References and Notes

  1. Acknowledgments: We thank P. M. Valencia for assistance with drafting and critical review of this manuscript, and R. Korn for assistance in reviewing CT scans. Funding: BIND received funding from NCI Small Business Innovation Research Contract No. HHSN261200700060C and National Institute of Standards and Technology Advanced Technology Program Award No. 70NANB7H7021. Author contributions: J.H. led the DTXL-TNP development program and performed study designs and interpretation of the data; D.V.H. is the principal investigator on the DTXL-TNP clinical study; J. Summa was the project leader and oversaw design and conduct of mouse xenograft studies; E.S. designed the clinical study; M.M.A., J.A., and T.C. synthesized the polymers; B.R., D.D.W., M. Figueiredo, and Y.H.S. performed the combinatorial optimization of DTXL-TNPs; A.S., G.T., M. Figa, and T.V.G.H. developed and manufactured the DTXL-TNPs; K.M. and A.H. developed and performed SPR analyses; J.J.S. and E.A. developed and performed PSMA binding ELISA analyses; E.P. and D.T. performed analytical characterization of DTXL-TNPs; S.L. designed and oversaw animal studies and development of the encapsulated DTXL bioanalytical method, and performed PK analyses and writing of the manuscript; J.W. performed interpretation of data and statistical analysis; P.L. is an investigator on the clinical study; P.W.K., N.H.B., and C.S. were scientific advisors on the preclinical animal models and clinical development strategy; O.C.F. and R.L. developed the combinatorial NP optimization platform and performed data interpretation and writing of the manuscript; S.Z. performed study design, data interpretation, and writing of the manuscript. Competing interests: E.S., M.M.A., J.A., B.R., D.D.W., M. Figueiredo, Y.H.S., J. Summa, A.S., G.T., M. Figa, T.V.G.H., K.M., A.H., J.J.S., E.A., E.P., D.T., S.L., J.W., S.Z., and J.H. are employees of BIND Biosciences, a biotechnology company developing nanoparticle therapeutics. P.W.K., N.H.B., C.S., O.C.F., and R.L. disclose financial interest in BIND Biosciences: P.W.K., N.H.B., and C.S. serve as members of its Scientific and Clinical Advisory Boards. O.C.F. and R.L. cofounded BIND and serve as members of its Board of Directors and Scientific Advisory Board. No aspect of this research was carried out in the laboratories of O.C.F. and R.L., and no academic research grants were used to support this research and development effort. Data and materials availability: All reasonable requests for collaboration involving materials used in the research will be fulfilled provided that a written agreement is executed in advance between BIND Biosciences Inc. and the requester (and his or her affiliated institution). Such inquiries or requests for additional data should be directed to the corresponding authors.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article