Research ArticleCancer

Nanodiamond Therapeutic Delivery Agents Mediate Enhanced Chemoresistant Tumor Treatment

See allHide authors and affiliations

Science Translational Medicine  09 Mar 2011:
Vol. 3, Issue 73, pp. 73ra21
DOI: 10.1126/scitranslmed.3001713


Enhancing chemotherapeutic efficiency through improved drug delivery would facilitate treatment of chemoresistant cancers, such as recurrent mammary tumors and liver cancer. One way to improve drug delivery is through the use of nanodiamond (ND) therapies, which are both scalable and biocompatible. Here, we examined the efficacy of an ND-conjugated chemotherapeutic in mouse models of liver and mammary cancer. A complex (NDX) of ND and doxorubicin (Dox) overcame drug efflux and significantly increased apoptosis and tumor growth inhibition beyond conventional Dox treatment in both murine liver tumor and mammary carcinoma models. Unmodified Dox treatment represents the clinical standard for most cancer treatment regimens, and NDX had significantly decreased toxicity in vivo compared to standard Dox treatment. Thus, ND-conjugated chemotherapy represents a promising, biocompatible strategy for overcoming chemoresistance and enhancing chemotherapy efficacy and safety.


Resistance to chemotherapy is a key barrier to cancer treatment. Liver cancer is inherently difficult to treat with chemotherapy because of the innate role of the liver in processing toxins, including chemotherapeutics (1). Furthermore, many chemosensitive cancers acquire chemoresistance after treatment, leading to recurrence and subsequent ineffectiveness of standard chemotherapy. This acquired resistance is particularly troubling because it often leads to cross-resistance against other chemotherapeutics. Overall, intrinsic or acquired drug resistance contributes to treatment failure in more than 90% of metastatic cancers (2). Thus, methods for overcoming drug resistance or improving efficacy of chemotherapy drugs would significantly improve cancer survival rates.

Drug efflux by adenosine triphosphate–binding cassette (ABC) transporter proteins, such as MDR1 and ABCG2, is the most common mechanism of chemoresistance (3). Drug efflux has been demonstrated in cancers in organs, such as the liver and breast, where these proteins have functional roles in normal biology (4, 5), as well as in established cancer cell lines such as the 4T1 mammary carcinoma cell line (6). Although clinical trials of small-molecule inhibitors of these transporter proteins have proven unsuccessful (7, 8), overcoming drug resistance through nanoparticle-mediated drug delivery may be more effective.

Previous studies of nanoparticle drug delivery platforms have identified many of the requisite properties for translation (916). Nanodiamonds (NDs) integrate several of these properties, including surface geometries that mediate high-affinity therapeutic binding/sustained release, diversity of potential conjugates, scalability, and biocompatibility. NDs are carbon nanoparticles with truncated octahedral architecture that are about 2 to 8 nm in diameter and can deliver a wide range of therapeutics, including small molecules, proteins, and nucleic acids (17, 18). A broad range of assays have also revealed in vitro ND biocompatibility (19). Furthermore, the facets present on the ND surface have been shown to possess charge properties that enable potent water binding for dispersability and sustained therapeutic release (20, 21). All of these properties suggest that NDs may serve as a translational platform for the treatment of cancer and other diseases.

NDs have been used primarily as biological imaging agents because of their ability to emit bright fluorescence after the introduction of nitrogen defects (2224). In addition, ND surface–protein interactions via silane linkage and electrostatics have been explored for applications such as extraction in proteomics (25, 26). The aforementioned studies have primarily addressed fundamental mechanisms of ND fluorescence toward imaging, diagnostics, and surface modification. Here, drug binding mediated by ND surfaces was harnessed to develop NDX, a drug delivery platform for doxorubicin (Dox), which was then compared in vivo to drug delivery by unmodified Dox. Although Dox is the standard chemotherapeutic for many cancers, it can be actively effluxed from tumor cells by a wide range of drug transporter proteins, including MDR1, ABCG2, and MRP1 (27). Here, Dox was reversibly bound to the NDs with sodium hydroxide chemical treatment, enhancing sustained release in vitro and in vivo. The resulting ND drug delivery system improved both drug retention in tumor cells and treatment safety and efficacy in murine cancer models when compared with standalone chemotherapy, suggesting that NDs are a promising drug delivery platform for chemoresistant tumors.


Analysis of ND modification, toxicity, and clearance

To evaluate the potential of NDs as viable drug delivery platforms in vivo, we analyzed biological responses to high ND dosages. As seen in Fig. 1A, the NDs (500 μg) failed to induce elevated sera interleukin-6 (IL-6) concentrations, which is indicative of a lack of systemic inflammatory responses. Furthermore, ND treatment did not result in any increase in serum alanine transferase (ALT), suggesting that ND treatment does not adversely affect liver function. Histological analysis revealed no significant changes in multiple tissues that were analyzed after ND treatment (Fig. 1B and fig. S1). These initial biocompatibility results indicate that the NDs may be applicable as drug delivery platforms in vivo. Additionally, ND surfaces enable stable drug sequestering by ND complexes. This resulted in a 10-fold increase in blood circulation halftime by NDX from 0.83 to 8.43 hours while also significantly decreasing myelosuppression by NDX treatment (200 μg of Dox equivalent) compared to Dox treatment (200 μg) (Fig. 1, C and D). Dox extractions from blood and tissue were performed as previously described (28).

Fig. 1

Nanodiamonds (NDs) are nontoxic and capable of conjugation with a variety of molecules. (A) Serum analysis of FVB/N mice tail vein injected with 500 μg of NDs (n = 3) or PBS (n = 3) for 1 week or lipopolysaccharide (LPS) (2.5 μg/kg)/d-galactosamine (d-GalN) (200 mg/kg) (n = 2) for 6 hours. nc, no change. Data are represented as means ± SD. *P < 0.006; **P < 0.001. (B) 40× hematoxylin and eosin (H&E) histopathological analysis of kidney, liver, and spleen tissue from treated mice. Scale bar, 100 μm. (C) Mean white blood cell (WBC) counts after treatment with PBS (n = 5), doxorubicin (Dox) (400 μg) (n = 5), or ND-conjugated Dox (NDX) (400 μg of Dox equivalent) (n = 5). Data are represented as means ± SD. *P < 0.002. (D) Blood circulation halftime (t1/2) analysis after treatment with Dox (200 μg) (n = 4) or NDX (200 μg of Dox equivalent) (n = 4). Data are represented as means ± SD. (E) FTIR analysis of ND (spectra 1), ND-NH2 (spectra 2), free XenoFluor 750 dye (spectra 3), and XenoFluor 750–ND (spectra 4). Arrows denote the C-N stretch and N-H bend combination at 1261 cm−1, benzene ring stretch at 1508 cm−1, and vibration of aromatic C-H at 926 cm−1. FTIR analysis of reduced ND (spectra 1), ND-NH2 (spectra 2), free Alexa Fluor 488 dye (spectra 5), and Alexa Fluor 488–ND (spectra 6). Arrows show IR features at 1261 cm−1, which represent the C-N stretch and N-H bend combination in –CO–NH–C– groups, suggesting amide formation. Arrows also denote benzene ring stretch at 1619 and 1442 cm−1. (F) Transmission electron microscopy (TEM) images of NDs and NDX. Scale bars, 5 nm. (G) Model of ND, NDX, and XenoFluor 750–ND.

To examine ND distribution and clearance, we labeled NDs with XenoFluor 750 and Alexa Fluor 488. Fourier transform infrared spectroscopy (FTIR) analysis revealed that both XenoFluor 750 and Alexa Fluor 488 covalently bound to NDs by amide formation, denoted by the appearance of benzene ring stretch signals (Fig. 1E, arrowheads). Furthermore, transmission electron microscopy (TEM) was used to verify NDX formation; clear lattice structures were visible in samples of ND alone, and the lattices were markedly less visible after Dox-ND binding (Fig. 1F). Figure 1G shows schematic representations of functionalized ND complexes developed in this study.

Although ND binding lengthens dye retention compared with free dye, ND-dye conjugates are still capable of whole-body clearance [Fig. 2A, days 3 to 10 (120 μg) and days 1 to 10 (40 μg)]. More detailed organ analysis (Fig. 2, B and C) further confirms clearance in lung (day 4), spleen (day 7), kidney (day 7), and liver (day 10). Because the NDs are covalently bound to the fluorescent dyes, the clearance seen can be attributed to labeled NDs and not free dye, which is not visible by whole-body imaging after 24 hours even at concentrations significantly higher than those delivered by ND-dye conjugates (Fig. 2A). Imaging of individual organs in mice confirms that free dye is cleared from the lungs and spleen within 24 hours and the liver and kidneys within 48 hours (Fig. 2B). Because the liver is a primary organ of ND accumulation before clearance, the stability of serum ALT levels (Fig. 1A) provides further evidence of ND biocompatibility.

Fig. 2

NDs are capable of clearance after biodistribution analysis. (A) Whole-body imaging of wild-type FVB/N mice tail vein injected with 40, 120, 400, or 1200 μg of NIR XenoFluor 750 covalently labeled NDs, 3 nmol of free dye, or control NDs. (B) Organ-specific imaging of wild-type FVB/N mice tail vein injected with 120 μg of NIR XenoFluor 750 covalently labeled NDs or 60 pmol of free dye. Data represented as means ± SD. (C) Tissue section analysis of FVB/N mice tail vein injected with Alexa Fluor 488–labeled NDs (400 μg). Sections were analyzed by fluorescent microscopy. Scale bar, 100 μm.

Dynamic light scattering (DLS) was then performed to determine the average size and ζ potential of NDs, NDX, and XenoFluor 750– and Alexa Fluor 488–labeled NDs. Characterization of ND conjugates revealed that unmodified NDs form average cluster sizes of nearly 50 nm (table S1). NDX complexes possessed an average size of ~80 nm at both high and dilute concentrations and were dispersable in aqueous solutions. In addition, conjugation with XenoFluor 750 or Alexa Fluor 488 increased the average cluster size by two to three times the unmodified cluster size, suggesting that fluorescent dye conjugation makes NDs more prone to agglomeration. However, all of the ND particles exhibited a narrow size distribution, with polydispersity indices ranging from 0.1 to 0.2, calculated at 25°C with a 173° scattering angle via cumulative analysis with Zetasizer Nano ZS software (29, 30). Furthermore, the NDX size indicated minor aggregation and a narrow size distribution. These are favorable properties for drug delivery systems, which should not be prone to excessive aggregation and precipitation. These properties were further observed through successful ND-dye delivery throughout the body via tail vein injection (Fig. 2A).

A potential mechanism for the formation of NDX complexes may be the existence of electrostatic interactions between protonated amines on the Dox molecules and deprotonated carboxylic acid groups on NDs, because oxidized ND surfaces are prone to polar interactions such as hydrogen bonding and electrostatic interactions. DLS data revealed that NDX complexes had higher ζ potentials than unmodified NDs (~17.7 mV) at neutral pH (table S1). ND-NH2 showed a ζ potential of ~48 mV, which was also higher than unmodified NDs and indicative of surface binding. After conjugation with XenoFluor 750 and Alexa Fluor 488 dyes, the labeled ND particles exhibited ζ potentials of −38.8 ± 0.4 mV (SD) and 27.1 ± 2.0 mV (SD), respectively. In coordination with the FTIR spectra, the ζ potential changes after fluorescent agent–ND reaction steps indicate fluorescent dye conjugation to the NDs. UV-vis spectrophotometry analysis was used to confirm ND-mediated pelleting of Dox after centrifugation at 14,000 rpm. From these studies, we conclude that a shift in the 490-nm absorbance peak that is characteristic of Dox was due to drug-ND interactions (figs. S2 and S3). Furthermore, when analyzing the loading efficiency of Dox (from 5 to 50 μg) to 50 μg of NDs, the NDX loading for a 5:1 ND/Dox ratio, the more stable complex of those tested, was 6.54 ± 0.2 μg (SD) of Dox, which corresponds to ~65.4% of the added Dox and ~13.1% loading relative to the added NDs (fig. S4 and table S2). Although higher loading was possible, the 5:1 NDX complex could be stored for at least 3 months at 4°C, remaining unaltered and suitable for sustained circulation and delivery in vivo and facile intravenous injection.

Effect of ND conjugation on Dox efflux and function

Therapeutic efflux is a common mechanism for tumor chemoresistance. A nonspecific method for overcoming efflux may represent a realistic approach to solving chemoresistance. The effect of ND-enhanced drug retention was predicted to overcome efflux, enabling tumor cells to retain chemotherapeutics longer (Fig. 3, A and B). We first analyzed drug retention in MDCK cells overexpressing the human drug transporter MDR1, which has previously been demonstrated to efflux Dox (31). After 1 hour of drug treatment, the cells were washed and allowed to efflux for 4 hours. Effluxed drug was subsequently washed from the media and cells were measured for drug retention. As seen in Fig. 3A, MDCK-MDR cells retained more Dox when treated with NDX (0.26 μg/ml) compared to unmodified Dox (0.02 μg/ml). This drug retention was verified by fluorescent microscopy (Fig. 3B). Efflux assays were repeated with cells from the LT2M cell line, which is derived from liver tumors of Tet-o-MYC/LAP-tTA (LT2-Myc) mice, and the more Dox-resistant 4T1 murine mammary tumor cell line (fig. S5). As seen in Fig. 3, C and D, tumor cells treated with NDX also retained more Dox than those treated with unmodified Dox. Experiments repeated in Huh7 and MDA-MB-231 human cell lines expressing drug transporter pumps (32, 33) also yielded similar results (fig. S6). Thus, it appears that NDX complexes can impair tumor cell efflux of Dox.

Fig. 3

Delayed release of functional Dox from NDs reduces Dox efflux from tumor cells. (A) Dox efflux analysis of MDCK-MDR cells. Data represented as means ± SD. *P < 0.02. (B) Efflux analysis by fluorescent microscopy. Scale bar, 100 μm. (C and D) Dox efflux analysis of LT2M or 4T1 cells. Data represented as means ± SD. *P < 0.001. (E) LT2M tumor cells incubated with Dox (0.5 μg/ml), NDX [Dox equivalent (0.5 μg/ml)], or NDs (2.5 μg/ml) and Dox (0.5 μg/ml). After 4-day incubation, cell growth and viability were analyzed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Data are represented as means ± SD. *P < 0.001; **P < 0.0001; ***P < 0.01. (F) 4T1 tumor cells treated with Dox (1 μg/ml), NDX [Dox equivalent (1 μg/ml)], or NDs (0.5 μg/ml) and Dox (1 μg/ml). After 4-day incubation, cell growth and viability were analyzed by MTT assay. Data are represented as means ± SD. *P < 0.001; **P < 0.001; ***P < 0.01. (G) LT2M tumor cells treated with Dox (5 μg/ml) and NDX [Dox equivalent (5 μg/ml)]. After 1-hour incubation, cells were washed every 24 hours and cell growth was analyzed by MTT assay after 3-day incubation. Data are represented as means ± SD. *P < 0.007; **P < 0.001. (H) 4T1 tumor cells treated with Dox (10 μg/ml) and NDX [Dox equivalent (10 μg/ml)]. After 1-hour incubation, cells were washed every 24 hours and analyzed by MTT assays after 3-day incubation. Data are represented as means ± SD. *P < 0.0006.

Because Dox must release from NDs and intercalate DNA to inhibit DNA replication and cell growth, we analyzed whether NDX complexes could release functional Dox. Although ND conjugation of Dox resulted in less apoptosis than either unmodified Dox or an unconjugated mix of ND and Dox in both LT2M and 4T1 cells, NDX complexes can still elicit a significant amount of apoptosis in both cell lines (Fig. 3, E and F). Thus, ND conjugation of Dox results in sustained functional drug release (figs. S7 and S8). Gradual Dox release was also confirmed in vitro through UV-vis spectrophotometry, revealing a gradual accumulation of Dox released from NDX complexes in phosphate-buffered saline (PBS) at 37°C (figs. S7 and S8).

We subsequently investigated if increased retention by NDX would result in continued tumor cell killing after clearance of drug treatment. As seen in Fig. 3G, Dox and NDX apoptotic effects are equivalent in tumor cells that were washed after 1 hour of LT2M cell treatment. In the more Dox-resistant cell line 4T1, only NDX treatment resulted in statistically significant killing (Fig. 3H). Additionally, for both cell lines, NDX killing is not affected by the universal drug transporter inhibitor verapamil (Fig. 3, G and H). Comparatively, Dox has greatly increased apoptotic effects when drug transporters are inhibited by verapamil. Thus, NDX can increase tumor cell drug retention and impair Dox efflux by drug transporters, allowing for continued killing of tumor cells via sustained drug release from NDX.

Increased acute apoptotic response mediated by NDX in tumor models in vivo

Upon demonstrating that the NDX platform is capable of impairing drug efflux from cancer cells and maintaining functional drug for a longer period of time, we asked if NDX was more effective at promoting acute apoptosis than Dox in tumor models in vivo. Mice bearing tumors were treated with NDX or Dox, and tumors were analyzed 7 days after treatment for apoptotic response. As seen in Fig. 4, A and B, NDX treatment resulted in increased apoptosis in murine LT2-Myc liver tumors compared to treatments with Dox. Murine 4T1 mammary tumors were more resistant to Dox, exhibiting minimal apoptosis even at 200 μg of Dox equivalent treatments (Fig. 4, C and D). NDX treatments, however, resulted in notably increased apoptosis as measured by terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) staining. Additionally, the increased efficacy of NDX was not due to a general systemic increase in apoptosis, because adjacent normal tissue was unaffected by NDX treatment (Fig. 4, A and B). Thus, delivery of chemotherapeutics by NDs appears to improve acute apoptotic response in vivo specifically in tumors.

Fig. 4

ND delivery of Dox increases apoptosis in murine liver and mammary tumor models. (A and B) Apoptosis analysis of liver tumor–bearing LT2-Myc mice treated with PBS (n = 3), Dox (n = 3), or NDX (n = 3) by tail vein injection 7 days after treatment. (A) Representative images of TUNEL stains. Scale bar, 100 μm. (B) TUNEL stain quantification by ImageJ. Data are represented as means ± SD. *P < 0.02; **P < 0.02. (C and D) Apoptosis analysis of 4T1 mammary tumor–bearing BALB/c mice treated with Dox (n = 3) or NDX (n = 3) by tail vein injection 7 days after treatment. (C) Representative images of TUNEL stains. Scale bar, 100 μm. (D) Quantification of TUNEL stains by ImageJ. Data are represented as means ± SD. *P < 0.03.

In vivo inhibition or regression of long-term tumor growth by NDX

To determine whether the increased apoptotic response in tumor models seen after NDX treatment results in more effective therapy against tumor growth, we analyzed long-term tumor burden in LT2-Myc liver and 4T1 mammary tumor models. Tumors were first allowed to develop in both models based on previous studies (34, 35), after which tumor-bearing mice were treated weekly with NDX, Dox, NDs, or PBS (Figs. 5A and 6A). ND treatment alone did not significantly affect liver/tumor weight (Fig. 5B). As seen in Fig. 5, B to D, treatment with 100 μg of Dox equivalent of either free Dox or NDX effectively inhibited liver tumor growth, but NDX resulted in significant improvement in tumor treatment compared to Dox. This was determined by lower liver/tumor weight, fewer mice with visible nodules, and fewer nodules in mice with visible tumors. Additionally, NDX treatment translated into a limited but more significant improvement of survival over Dox treatment and control mice (Fig. 5E). Tumor retention studies indicate that although Dox exhibits greater penetration into liver tumors than NDX at 30 min after injection, NDX was retained in tumors longer, with more NDX (1.5 μg of Dox per gram of tumor) present at day 7 than unmodified Dox (Fig. 5F). Analysis in LT2-Myc and 4T1 tumor models suggests that, like near-IR (NIR)–labeled NDs, NDX is retained longer in multiple tissues including tumors, with the liver and kidneys retaining the highest amounts of NDX (fig. S9). Increased retention in normal tissue may not result in overall increased systemic apoptosis as evidenced by lack of apoptosis in adjacent normal tissue after NDX treatment (Fig. 4, A and B).

Fig. 5

ND delivery of Dox inhibits tumor growth in murine liver tumor models. (A) Model of long-term drug treatment of tumor-bearing LT2-Myc mice. Upon weaning, doxycycline (Doxy) was removed from diet of LT2-Myc mice. Five weeks after doxycycline withdrawal from diet, mice were treated with PBS (n = 5), Dox (100 μg) (n = 12), ND (400 μg) (n = 5), or NDX (100 μg of Dox equivalent) (n = 12) by tail vein injection every 7 days. (B) Total liver and liver/tumor weight analysis after treatment. Data are represented as means ± SD. *P < 0.005; **P < 0.0001; ***P < 0.02. (C) Percentage of mice exhibiting macroscopic tumor nodules (defined as >1 mm). *P < 0.03. (D) Images of livers/tumors from treated mice. (E) Kaplan-Meier survival plot for LT2-Myc mice treated with PBS (n = 5), Dox (100 μg) (n = 8), or NDX (100 μg of Dox equivalent) (n = 7) by tail vein injection every 7 days. *P < 0.03; **P < 0.06. (F) Tumor retention analysis of LT2-Myc mice treated with Dox (200 μg) (n = 4) or NDX (200 μg of Dox equivalent) (n = 4) by tail vein injection. Data are represented as means ± SD. *P < 0.002.

Fig. 6

ND delivery of Dox inhibits tumor growth in a murine mammary carcinoma model. (A) 4T1 cells (5 × 106) injected into the left and right number 4 mammary gland of BALB/c virgin female mice. Mice were treated with PBS (n = 18), Dox (100 μg) (n = 18), NDX (100 μg of Dox equivalent) (n = 18), Dox (200 μg) (n = 10), or NDX (200 μg of Dox equivalent) (n = 10) by tail vein injection. Black arrows denote injection days. Asterisks indicate statistically significant differences between Dox- and NDX-treated mice. Data are represented as means ± SD. *P < 0.02; **P < 0.0005. (B) Kaplan-Meier survival plot for 4T1 mice treated with PBS (n = 7), Dox (100 μg) (n = 10), NDX (100 μg of Dox equivalent) (n = 10), Dox (200 μg) (n = 5), or NDX (200 μg of Dox equivalent) (n = 5) by tail vein injection every 6 days. *P < 0.003. (C) Whole mouse weight analysis of 4T1 mice treated with PBS (n = 10), Dox (100 μg) (n = 5), NDX (100 μg of Dox equivalent) (n = 5), Dox (200 μg) (n = 5), or NDX (200 μg of Dox equivalent) (n = 5) by tail vein injection every 6 days. *P < 0.003. (D) Tumor retention analysis of 4T1 mice treated with Dox (200 μg) (n = 4) or NDX (200 μg of Dox equivalent) (n = 4) by tail vein injection. Data are represented as means ± SD. (E) Representative images of excised tumors from treated mice.

Although both Dox and NDX treatment appear to have at least some effect on LT2-Myc liver tumor burden, the increased efficacy of NDX was even more evident in the more Dox-resistant 4T1 tumor model. Lower unmodified Dox dosages did not affect tumor growth when compared to untreated mice. Higher Dox dosages elicited some effect on tumor growth (Fig. 6A), but were also accompanied by extreme toxicity, with none of these mice surviving beyond day 15 (Fig. 6, A and B). In contrast, NDX treatment markedly inhibited tumor growth (Fig. 6, A and E). Additionally, although treatment with unmodified Dox results in similar survival curves compared to PBS-treated mice, NDX treatment significantly improved survival probability compared to PBS- or Dox-treated 4T1 mammary carcinoma models (Fig. 6B). Furthermore, although maximal unmodified Dox dosages (200 μg) were highly toxic and resulted in early mortality, equivalent NDX dosages demonstrated the greatest efficacy in tumor treatment and survival among all treatment conditions (Fig. 6, A and B). The decreased toxicity of Dox delivery by NDs was further demonstrated by no significant change in weight after treatment by NDX compared to high Dox dosages (Fig. 6C). NDX was retained longer in 4T1 tumors compared to Dox and cleared from tumors by day 7 (Fig. 6D). Thus, NDs serve as an effective drug delivery option to overcome the limitations of unmodified chemotherapy.


The focus of this study was to develop a biocompatible, ND-based drug delivery agent (NDX) to treat chemoresistant tumors with enhanced efficacy. The NDX complex enabled prolonged drug retention compared to unmodified Dox in vitro. After intravenous ND treatment, there did not appear to be any increases in myelosuppression, systemic immune response, or serum ALT secretion, a liver toxicity indicator. Furthermore NDX-mediated efficacy improvements to chemoresistant tumor therapy over Dox alone were observed in vivo.

Acquired and intrinsic chemoresistance are common problems in cancer therapy, particularly for metastatic tumors. Drug transporters that efflux chemotherapeutics, such as MDR1 and ABCG2, are commonly found in cancer. The contribution of these and other pumps likely varies among tumor types and among individual patients suffering from similar cancers. Chemical inhibitors that block drug efflux have been identified for transporter pumps that play a key role in chemoresistance, and clinical use of these inhibitors remains controversial, resulting in either high toxicity or low efficacy depending on the inhibitor (7, 8). A key barrier in overcoming chemoresistance that is not met by small-molecule inhibitors is their lack of specificity for particular efflux pumps. Thus, a passively targeted approach to overcoming chemoresistance may be more potent. ND-mediated drug delivery does not target specific drug transporters, and the in vivo studies presented here demonstrate that NDs can mediate chemotherapeutic delivery that results in drug response by LT2M liver tumors and 4T1 mammary tumors, which are otherwise less sensitive (LT2M) or virtually insensitive (4T1) to unmodified drugs. This effect also makes ND drug delivery potentially applicable to multiple tumor types with varying mechanisms for chemoresistance.

Coupled with previous work on ND toxicity (19), the studies presented here suggest that NDs can potentially serve as clinically relevant drug delivery systems. ND-mediated drug delivery has the advantage of increased circulation time and increased tumor retention in both LT2M and 4T2 tumors compared to free Dox. Although unmodified Dox elicited higher apoptosis in vitro when efflux was chemically impaired, the gradual release of Dox from NDX complexes allowed for enhanced Dox retention, and response is unaffected by chemical impairment of efflux in tumor cells. Yet, acute imaging studies with fluorescently labeled NDs indicate that they can clear in about 7 to 10 days, whereas 60% of the previously explored larger particles (about one order of magnitude increase in diameter) are still retained in the liver at ~28 days (36). Therefore, the application of smaller-diameter NDs is expected to result in shorter clearance times and improved biocompatibility. NDs are also able to bind various therapeutics and other reagents by both permanent covalent bonds and reversible ionic bonds. These include covalently linked imaging agents and small interfering RNA (siRNA) release, making NDs a versatile drug delivery platform with the potential to be applied to a wide range of diseases (37).

With regard to developing nanoparticles as therapeutic carriers, it is important that the nanoparticle is safe for use in vivo. Previous nanoparticle studies have revealed adverse responses indicative of material toxicity. For example, carbon nanotubes have been used for drug delivery and hyperthermia treatment (38) but nanotube treatment of epidermal keratinocytes resulted in increased IL-8 gene expression or pulmonary injury in mice (39). Iron nanoparticle interactions with neurons resulted in the suppression of neural growth factor–induced neurite formation (40). Studies with pulmonary titanium dioxide (TiO2) nanoparticles demonstrated that the degree of toxicity was dependent upon particle size and surface properties (41).

Regarding NDX safety over Dox alone, myelosuppression, which may result in neutropenic fever, infections, hemorrhage, or even death, was examined because it is often the dose-limiting toxic side effect of unmodified Dox (42). NDX administration resulted in significantly decreased myelosuppression compared to free Dox. Furthermore, early mortality and toxicity from maximal Dox dosages (200 μg) were not observed during long-term treatment with equivalent NDX doses. These NDX doses resulted in enhanced treatment efficacy characterized by lack of early mortality and even smaller tumor sizes compared to 100 μg of NDX treatment. These results were seen even though NDX resulted in enhanced retention of Dox in both normal tissue and tumor tissue. Because normal tissue cells are slower dividing than tumor cells, this suggests that the eventual clearance of NDX prevents systemic apoptosis.

These studies indicate that NDs unite several important drug delivery properties into one system. They include biocompatibility, the ability to deliver several types of therapeutic (hydrophilic/hydrophobic small molecules, proteins, nucleic acids, etc.), high-throughput ND-drug binding, and scalable processing. Facile ultrasonication, ultracentrifugation, and ball milling yield particles with 2- to 8-nm diameters. Furthermore, acid washing produces ND facets coated with carboxyl groups, enabling drug conjugation, dispersion in water, and unique surface properties. Alternating electrostatic potentials on the ND facets result in attraction toward surrounding water molecules, creating a hydration shell at the ND-solvent interface. These electrostatic properties that mediate the binding of ions and water also enable applications toward potent and reversible drug binding, resulting in sustained drug release as molecules are exchanged on the ND surface, and magnetic resonance imaging with unprecedented increases in relaxivity, for example (21, 4345). This makes ND surfaces distinct from other nanomaterials ranging from quantum dots to polymers and fullerenes (4648).

Although translating NDs to human use remains to be established, the drug-binding potency evident in NDX and its resultant chemoresistant tumor treatment enhancement show that NDs integrate many of the characteristics needed for translationally relevant drug delivery. This serves as a promising foundation for continued NDX development and potential clinical application.

Materials and Methods

Mouse tumor models and cell lines

The Tet-o-MYC/LAP-tTA (LT2-Myc) hepatoblastoma tumor model has previously been described (34, 49). Doxycycline (200 mg/kg) repression of MYC transgene expression was performed. Induction of transgene expression and tumor formation were performed by doxycycline removal from the diet. The 4T1 mammary tumor model has been previously described (35). LT2M cells were derived from a primary MYC-driven liver tumor. All animal studies were approved by the Committee for Animal Research at the University of California, San Francisco. See Supplementary Material for further information.

ND-OH modification for XenoFluor 750 and ND–Alexa Fluor 488 synthesis

ND-OH complexes were prepared according to previous protocols (17). After ultrasonication in dry tetrahydrofuran (THF) for a minimum of 2 hours, 2.5 g of pristine ND material in dry THF was placed in a two-necked flask under argon. Under stirring, 25 ml of a 1 M BH3·THF solution in dry THF was added dropwise. The reaction mixture was stirred under reflux for 24 hours. After cooling the mixture to room temperature, hydrolysis was performed with 2 N hydrochloric acid until the cessation of hydrogen gas evolution. The solid product was isolated by centrifugation. After washing was performed with water and acetone in consecutive washing/centrifugation cycles until the supernatant liquid reached a pH of 7, the sample was freeze-dried, resulting in a gray powder.

ND-O-Si(OMe)2(CH2)3-NH2 modification for XenoFluor 750 and ND–Alexa Fluor 488 synthesis

ND-OH (1 g) was sonicated in dry acetone for at least 2 hours and subsequently added to 140 ml of a 5% solution of (3-aminopropyl)trimethoxysilane in dry acetone and stirred for 48 hours under argon treatment. After centrifugation, the precipitate was washed with acetone with consecutive washing and centrifugation cycles (five cycles), yielding light gray powder after freeze drying. Elemental analysis was as follows: N, 3.54; C, 81.85; H, 1.58; O, 6.64%. Elemental analysis revealed that ~0.94 mmol of amino groups was present per gram of synthesized ND-NH2 (17).

XenoFluor 750 and ND–Alexa Fluor 488 conjugation

XenoFluor 750– or Alexa Fluor 488–labeled ND particles were prepared by incubating ND-NH2 particles with excess XenoFluor 750 succinimidyl ester or Alexa Fluor 488 succinimidyl ester, respectively, in 0.1 M sodium bicarbonate buffer (pH 8.3) at room temperature for 2 hours in the dark according to the manufacturers’ instructions. To remove excess reaction agents, we washed the particle products with Millipore water at least five times, yielding a colorless supernatant. Fluorescent agent–labeled ND particles were resuspended in water at a stock of 2 mg/ml by ultrasonication for several minutes. Both fluorescent agent–labeled ND particles were stored at 4°C before use.

NDX functionalization

NDs were dispersed in pure water, underwent ultrasonication overnight, and were diluted to 10 mg/ml. A solution (2 mg/ml) of aqueous Dox was prepared. After sterilization of NDs via liquid autoclave, the NDs were mixed with Dox to produce an NDX conjugate solution (5 mg/ml–1 mg/ml). The solution was adjusted with 2.5 mM NaOH to promote drug complexing, resulting in a pH of ~7.74 followed by vortex for 1 min.

Spectroscopic analysis of NDX functionalization and loading

Solutions of ND, Dox, NDX, and NDX with NaOH were prepared. The concentrations of ND and Dox were 250 and 50 μg/ml for all solutions, respectively. The absorbance was read from 300 to 700 nm on 100 μl of each solution with a UV-vis spectrophotometer. Each solution was then centrifuged for 2 hours at 14,000 rpm and room temperature. To assess loading efficiency, we converted absorbance measurements to Dox concentration via a standard curve obtained by serial dilution from the specified Dox concentration. Samples used for standard curve generation were kept at 2.5 mM NaOH, whereas Dox was diluted accordingly. Absorbance values of both experimental and standard curve samples were analyzed on the same plate to avoid plate-to-plate variations.

Spectroscopic analysis of drug release from NDX

Dox adsorbed to NDs at an initial concentration of 0.1 μg/μl was desorbed in PBS or media diluted 1:1 and 1:10 in PBS. The percent of Dox desorption is based on the amount of Dox initially adsorbed onto the NDs and on the fluorescent signal of Dox in the sample. Experimental samples were performed in triplicate, and represented as an average, with error bars representing SD. Centrifugation at 2500 rpm for 5 min to pellet the ND-drug complex was performed. The adsorption solvent was replaced with 500 μl of Media 199 (Mediatech) supplemented with 10% fetal bovine serum (Sigma-Aldrich) and diluted 1:1 or 1:10 in PBS (Sigma-Aldrich). The ND-drug pellet was suspended by gentle inversion and subsequently incubated at 37°C. Dox fluorescence (emission, 480 nm; excitation, 550 nm) in the supernatant was measured on hours 2, 3, 4, and 5 on day 1, as well as every day until day 16. After each measurement, the supernatant was discarded and the ND-drug pellet was replaced with fresh media, subsequently inverted, and then incubated as explained above. NDX desorption in PBS was tracked at 4, 6, 24, 48, and 72 hours. Individual solutions of NDX were prepared for each time point and kept at 37°C. At each time point, the sample was centrifuged for 1 hour and 100 μl of the supernatant was subsequently extracted for spectroscopy. The experiment was performed in triplicate.

Acute and long-term in vivo tumor progression studies

Liver tumor–bearing LT2-Myc mice were treated with PBS, ND, Dox, or NDX by tail vein injection once a week. For acute response studies, mice were sacrificed 7 days after initial treatment. For long-term tumor progression studies, mice were treated once a week. For liver/tumor weight and nodule analysis, mice were analyzed after 21 days of treatment. BALB/c mice bearing 4T1 mammary tumors were treated with PBS, Dox, or NDX by tail vein injection every 6 days. For acute response studies, mice were sacrificed 7 days after initial treatment. For long-term tumor progression studies, mice were subsequently treated every 6 days and tumor volumes and weight were measured every 3 days. See Supplementary Material for further information.

Supplementary Material

Materials and Methods

Fig. S1. 100× H&E histopathological analysis.

Fig. S2. NDX adsorption spectrophotometry analysis.

Fig. S3. NDX adsorption comparison.

Fig. S4. Dox loading analysis.

Fig. S5. Cancer cell line Dox resistance.

Fig. S6. Dox efflux analysis in human tumor cells.

Fig. S7. Dox desorption from ND agglomerates.

Fig. S8. Dox release spectrophotometry analysis.

Fig. S9. NDX and Dox tissue retention.

Table S1. Size and ζ potential of functionalized NDs.

Table S2. Dox loading efficiency.


  • Citation: E. K. Chow, X.-Q. Zhang, M. Chen, R. Lam, E. Robinson, H. Huang, D. Schaffer, E. Osawa, A. Goga, D. Ho, Nanodiamond Therapeutic Delivery Agents Mediate Enhanced Chemoresistant Tumor Treatment. Sci. Transl. Med. 3, 73ra21 (2011).

References and Notes

  1. Acknowledgments: We thank J. Michael Bishop (J.M.B.) for financial support of this project and critical manuscript review, L. Moore for ND–XenoFluor 750 preparation assistance, and J. Wu for TEM assistance. Funding: E.K.C. was supported by an American Cancer Society Postdoctoral Fellowship (PF-08-196-01-MGO). J.M.B. was supported by the University of California, San Francisco G.W. Hooper Research Foundation. A.G. was supported by NIH grant R01 CA136717. D.H. was supported by the National Science Foundation CAREER Award (CMMI-0846323), Mechanics of Materials grant CMMI-0856492, Center for Scalable and Integrated NanoManufacturing (DMI-0327077), National Center for Learning and Teaching, V Foundation for Cancer Research Scholars Award, Wallace H. Coulter Foundation Translational Research Award, and NIH (U54-A1065359). R.L. was supported by a Northwestern University Ryan Fellowship. M.C. was supported by the Weinberg College of Arts and Sciences. E.O. was supported by NEDO, Japan. Author contributions: E.K.C. performed the in vitro and in vivo efficacy and biocompatibility/distribution/imaging experiments. X.-Q.Z. synthesized/characterized the ND-imaging agents. M.C., R.L., E.R., H.H., and D.S. synthesized/characterized NDX. A.G. developed the LT2M cell line and provided manuscript preparation guidance. E.O. synthesized/characterized the ND material. E.K.C., X.-Q.Z., A.G., E.O., and D.H. analyzed the data. E.K.C., X.-Q.Z., M.C., A.G., and D.H. wrote the manuscript. E.K.C., X.-Q.Z., M.C., and D.H. wrote the Supplementary Material. Competing interests: X.-Q.Z., M.C., R.L., E.R., H.H., and D.H. are inventors on filed United States Patent No. 20100305309 entitled “Nanodiamond particle complexes” regarding this work. E.O. is an inventor on United States Patent No. 7300958 entitled “Ultra-dispersed nanocarbon and method for preparing the same.” The other authors declare that they have no competing interests.
View Abstract

Navigate This Article