Improved i.p. drug delivery with bioadhesive nanoparticles Yang Denga, Fan Yanga, Emiliano Coccob, Eric Songa, Junwei Zhanga,c, Jiajia Cuia, Muneeb Mohideena, Stefania Belloneb, Alessandro D. Santinb, and W. Mark Saltzmana,c,1 a Department of Biomedical Engineering, Yale University, New Haven, CT 06511; bDepartment of Obstetrics, Gynecology & Reproductive Sciences, School of Medicine, Yale University, New Haven, CT 06511; and cDepartment of Chemical & Environmental Engineering, Yale University, New Haven, CT 06511
Edited by Joseph M. DeSimone, University of North Carolina at Chapel Hill and Carbon, Chapel Hill, NC, and approved August 19, 2016 (received for review November 22, 2015)
| ovarian cancer | intraperitoneal |
H
igh-grade ovarian and uterine serous carcinoma (USC) are biologically aggressive tumors, which commonly spread into the peritoneal cavity, disseminating along the abdominal and pelvic peritoneum and resulting in peritoneal metastases. Although these tumors are often responsive to initial cytoreductive surgery and platinum–paclitaxel chemotherapy, most patients develop recurrence and eventually die from progression of chemotherapy-resistant disease (1). In cases of advanced ovarian cancer, the infusion of chemotherapeutic agents directly into the peritoneum (i.e., i.p. chemotherapy) improves overall and disease-free survival compared with standard i.v. chemotherapy (2). However, despite level 1 evidence from meta-analyses of three large, prospectively randomized, controlled trials, this route of administration is still not widely used in the clinic because of its more severe side effects compared with i.v. chemotherapy and the need for frequent dosing schedules (2, 3). To enhance the retention of chemotherapeutic drugs in the peritoneal cavity and improve their bioavailability, controlled release drug carriers, such as microparticles and hydrogels, have recently been tested in preclinical animal models (4–8). Although these carriers can lead to improvement over conventional delivery methods, both strategies only use the microparticles or hydrogel as reservoirs for drugs. To achieve expected therapeutic efficacy, the concentration of the free drugs in the peritoneal cavity has to be maintained in the effective range. Sustained drug concentrations require a balance between the drug release from carriers and clearance from peritoneal cavity by metabolism. For a specific drug and drug carrier combination, release that is too slow might also compromise the therapeutic efficacy as observed when limited dissolution of paclitaxel from a microparticle/hydrogel system impaired therapeutic outcomes (5). In addition, side effects of some of these carriers do exist (i.e., microparticles accumulate in the lower abdomen and cause peritoneal adhesion) (9–11). Degradable polymer nanoparticles have several potential advantages over microparticles and hydrogels in cancer therapy (12–16). Nanoparticles (∼100 nm) are small compared with microparticles www.pnas.org/cgi/doi/10.1073/pnas.1523141113
Significance Resistance to platinum-based chemotherapies and paclitaxel is common in recurrence of both high-grade ovarian and endometrial cancers. Paclitaxel resistance has been correlated with overexpression of class III β-tubulin, the preferential target of the epothilones, microtubule-stabilizing agents. Epothilone B (EB) is manifold more effective than paclitaxel, but clinical use is limited by side effects. To reduce side effects, we encapsulated EB into bioadhesive nanoparticles (BNPs), reasoning that bioadhesive nanoparticles loaded with epothilone B (EB/BNPs) would interact with abdominal tissues and gradually release EB in proximity of peritoneal cancer implants, thus maintaining EB concentration at the site of action and limiting systemic exposure and toxicity. Our experiments show the higher therapeutic activity and limited toxicity of EB/BNPs compared with nonadhesive nanoparticles loaded with EB or carrier-free EB. Author contributions: Y.D., F.Y., E.C., E.S., J.Z., J.C., M.M., A.D.S., and W.M.S. designed research; Y.D., F.Y., E.C., E.S., J.Z., J.C., M.M., S.B., A.D.S., and W.M.S. performed research; Y.D., F.Y., E.C., E.S., J.Z., J.C., M.M., S.B., A.D.S., and W.M.S. analyzed data; and Y.D., E.C., A.D.S., and W.M.S. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1
To whom correspondence should be addressed. Email: [email protected].
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1523141113/-/DCSupplemental.
PNAS | October 11, 2016 | vol. 113 | no. 41 | 11453–11458
APPLIED BIOLOGICAL SCIENCES
|
drug delivery nanoparticles chemotherapy
(∼10–100 μm) and therefore, should distribute more evenly throughout the peritoneum. Furthermore, nanoparticles are small enough to be internalized by tumor cells (17, 18), allowing them to release the encapsulated drugs near the biological target, which is often in the cell nucleus. Although many studies have shown that nanoparticles are promising vehicles for the delivery of chemotherapeutic agents in preclinical in vivo models, the application of nanoparticles for i.p. delivery remains hampered by their fast clearance from the peritoneal cavity, mainly as a result of lymphatic drainage (19). In a recent study (20), we described a novel form of bioadhesive nanoparticles (BNPs) and showed that their adhesiveness results in prolonged retention on the skin after topical delivery. These new materials are based on polylactic acid block– hyperbranched polyglycerol (PLA-HPG) copolymers, which can be formed into “stealthy,” nonadhesive nanoparticles (NNPs) that circulate for long times after i.v. injection (21). The BNPs were produced by oxidation of the NNPs, which results in conversion of vicinal diols on the surface of NNPs into aldehydes on the BNPs (Fig. 1). Aldehydes spontaneously react with proteins to form a variety of bonds, such as Schiff-base bonds; therefore, the presence of surface aldehydes on BNPs leads to particle adhesion to protein-rich materials. For example, we previously showed the rapid and robust adhesion of the BNPs to poly-Llysine–coated surfaces and pig or mouse skin (20). In this work, we hypothesize that BNPs will interact and adhere strongly with mesothelial cells layering the abdominal cavity after i.p. delivery
ENGINEERING
The i.p. administration of chemotherapy in ovarian and uterine serous carcinoma patients by biodegradable nanoparticles may represent a highly effective way to suppress peritoneal carcinomatosis. However, the efficacy of nanoparticles loaded with chemotherapeutic agents is currently hampered by their fast clearance by lymphatic drainage. Here, we show that a unique formulation of bioadhesive nanoparticles (BNPs) can interact with mesothelial cells in the abdominal cavity and significantly extend the retention of the nanoparticles in the peritoneal space. BNPs loaded with a potent chemotherapeutic agent [epothilone B (EB)] showed significantly lower systemic toxicity and higher therapeutic efficacy against i.p. chemotherapy-resistant uterine serous carcinomaderived xenografts compared with free EB and non-BNPs loaded with EB.
successfully used in prior work to generate consistent i.p. tumors in mice (28).
Fig. 1. Schematic showing the conversion of BNPs from NNPs and the fate of NNPs and BNPs after i.p. delivery. BNPs can be converted from NNPs by NaIO4 treatment. After i.p. delivery, (Left) NNPs are cleared by lymphatic drainage, but (Right) BNPs are retained in the peritoneal cavity because of their bioadhesive property.
(Fig. 1). We also hypothesize that the interaction of the BNPs with abdominal tissue will extend the retention of the BNPs after i.p. injection, therefore significantly improving the bioavailability of their payloads and the efficacy of i.p. chemotherapy. We tested the effectiveness of this approach for i.p. delivery of patupilone [epothilone B (EB)], a microtubule-stabilizing agent that targets class III β-tubulin. Importantly, EB has been shown to be 3- to 20-fold more effective than Paclitaxel (PTX) in vitro against multiple PTX-sensitive and -resistant human tumor cell lines (22–25). Unfortunately, the clinical use of EB in patients with recurrent disease is limited because of high toxicity, including severe diarrhea, vomiting, increased risk of bowel obstruction, and fatigue (26, 27). Here, we use our BNPs loaded with EB to test whether strongly adhesive nanoparticles will improve the efficacy of i.p. chemotherapy to PTX-resistant peritoneal metastasis. In particular, we developed i.p. xenograft models derived from the primary USC cell line USC-ARK-2. This cell line was selected (i) because of its high expression of P-glycoprotein and tubulin-β-III, characteristics that have been recently shown by our group to confer PTX resistance while showing extreme sensitivity to EB (25), and (ii) because this cell line has been 11454 | www.pnas.org/cgi/doi/10.1073/pnas.1523141113
Results In a previous work, we showed that NNPs had an extended blood circulation because of their resistance to nonspecific interaction to proteins (21). In another recent study, we found that BNPs had an enhanced interaction with protein-rich surfaces caused by their aldehyde groups (20). These results suggest that this class of nanoparticles can be converted from a nonadhesive to a strongly adhesive state and then back again by regulation of the density of aldehydes on the nanoparticle surface. To illustrate this flexibility, nanoparticles were converted by a cycle of oxidation and reduction between the hydroxyl-rich NNP state, the aldehyde-rich BNP state, and a reduced nonadhesive nanoparticle (NNP-R) state. We reduced the aldehydes to alcohols by treatment with a common reducing agent (29), NaBH4 (Fig. S1A). The concentration of aldehyde groups on BNPs was monitored as a function of duration of NaBH4 treatment; most aldehyde groups were converted after 2 min of treatment (Fig. 2D). It is important to note that NaBH4 treatment cannot reduce the aldehydes to the original vicinal diols, because one carbon is truncated by the initial conversion of vicinal diols to aldehydes (Fig. S1B). BNP, NNP, and NNP-R were identical in shape by transmission EM (TEM) (Fig. 2 A–C) and hydrodynamic size measured by dynamic light scattering (DLS) (Table 1), suggesting that oxidation and reduction cycles had no detrimental effect to the nanoparticles. To quantitate bioadhesion, we tested nanoparticles at various states of aldehyde activation by examining attachment to poly(L-lysine)–coated plates and USC cells (Fig. 2 E and F). To facilitate quantification of their bioadhesive properties, all nanoparticles were loaded with 0.2% 1,1′-dioctadecyl3,3,3′,3′-tetramethylindodicarbocyanine, 4-chlorobenzenesulfonate salt (DiD) dye. We first evaluated the bioadhesive property of the three nanoparticles on poly(L-lysine)–coated surfaces. After incubation and extensive washing, the BNPs showed much higher retention on the poly(L-lysine)–coated plates (Fig. 2E). The BNPs also showed much higher attachment to USC cells (Fig. 2F). The retention of NNPs or NNP-Rs on both poly(Llysine) surfaces and USC cell monolayers was negligible (Fig. 2 E and F). We further tested the nanoparticles for blood circulation times after i.v. injection in mice. The percentage dose of BNPs dropped to 1.5% after 2 h, whereas significant numbers of NNPs (21.3%) were still circulating after 24 h (Fig. 2G). NNP-Rs circulated with similar duration as the NNPs, reflecting the chemical similarity of their surfaces. We hypothesize that BNPs will adhere avidly to protein-rich tissue surfaces and remain adherent for long periods. To test this hypothesis, we first characterized the adhesive potential of BNPs and NNPs on human umbilical veins. Both DiD/NNPs and DiD/ BNPs were applied individually to the luminal surface of human umbilical veins. After 2 h of incubation within the lumen followed by extensive washing by vessel perfusion with 30 mL PBS, the BNPs showed much higher retention on the vessel endothelium than NNPs (Fig. 3 A and B). Most significantly for this study, we evaluated the retention of nanoparticles within the peritoneum after i.p. administration in mice. In this case, BNPs and NNPs were loaded with 0.5% IR-780 iodide (IR-780) dye to facilitate imaging. At this loading, only a small fraction of the IR780 dye is released from the nanoparticles, and it is released slowly over 10 d of incubation in buffered saline supplemented with physiological levels of protein (Fig. S2A); therefore, the dye molecule serves as a reliable marker for the presence of nanoparticles. As we have shown with other NNP/BNP combinations, the conversion of IR-780/NNPs to IR-780/BNPs (by treatment with NaIO4) had no apparent effect on nanoparticle size and morphology as observed by TEM (Fig. S3) or hydrodynamic size as measured by DLS (Table 1). After i.p. injection of NNPs, most of Deng et al.
Table 1. Nanoparticle size measurements Nanoparticles EB/NNPs EB/BNPs DiD/NNPs DiD/BNPs DiD/NNP-Rs IR-780/NNPs IR-780/BNPs
Deng et al.
Diameter (nm)
Polydispersity index
127 127 121 117 97 131 128
0.225 0.233 0.215 0.232 0.213 0.331 0.297
EB/BNPs, EB/NNPs, and free EB were separately added to USC cell cultures, which were tested for viability after 3 d of exposure. All EB treatments (EB, EB/NNPs, and EB/BNPs) yielded similar dose–response curves, with the free EB showing the highest toxicity (i.e., lowest IC50) (Fig. 4D). We believe that this difference between the free EB and nanoparticle formulations of EB is caused by the slow release of EB from NNPs and BNPs, resulting in a slightly lower concentration of EB in the culture medium exposed to nanoparticle formulations compared with free EB. To confirm that the cytotoxicity was caused by the EB and not the polymers within the nanoparticles, we also incubated USC cells with unloaded (blank) BNPs and NNPs. Unloaded nanoparticles did not show any toxicity, even when added to the culture medium at high concentrations (1 mg/mL) (Fig. 4E). The tolerable concentration of blank nanoparticles is much higher than the nanoparticle concentration of EB/NNPs or EB/BNPs needed to deliver an effective dose: for example, 10 nM EB is contained within 0.25 μg/mL EB/NNPs or EB/BNPs. To further show the safety of the nanoparticle treatments, we examined the cytotoxicity of blank BNPs and NNPs on human umbilical vein endothelial cells and HeLa cells. Neither BNPs nor NNPs were toxic in these cells up to 1 mg/mL (Fig. S5A). Furthermore, we incubated blank BNPs or NNPs at 1 mg/mL on a variety of immortalized cell lines and observed no cytotoxicity at high concentration (Fig. S5B). To measure retention of EB within the peritoneal tissue, we i.p. administered free EB, EB/BNPs, and EB/NNPs in nude mice. No EB was detected in mice receiving the free EB at 6 h after a single dose (Fig. 4F). In contrast, detectable levels of EB were found at 6 h with both EB/NNPs and EB/BNPs injections: the EB/BNPs-treated mice had substantially higher concentrations of EB than the EB/NNPs-treated mice (21% vs. 5%, P < 0.05) (Fig. 4F). One day after i.p. administration, EB was still significantly higher after EB/BNPs PNAS | October 11, 2016 | vol. 113 | no. 41 | 11455
ENGINEERING
the fluorescent signal, which was initially strong (5 min), was lost by the end of the first day (Fig. 3C). In contrast, after i.p. injection of BNPs, a strong fluorescent signal was observed to last up to 5 d (Fig. 3C); at the end of the 10th day, BNPs were still detectable in the area of the initial injection (Fig. S2B). Remarkably, accumulation of BNPs in the lower abdomen, which is commonly seen after microparticle i.p. delivery (30), was not observed with administration of BNPs; the majority of the BNPs adhered to the parietal peritoneum of the mice (Fig. S4). To determine the efficacy of nanoparticles as vehicles for i.p. delivery, we loaded NNPs and BNPs with EB. Nonadhesive nanoparticles loaded with epothilone B (EB/NNPs) and bioadhesive nanoparticles loaded with epothilone B (EB/BNPs) were of similar hydrodynamic size, ∼130 nm in diameter (Table 1). EB/BNPs and EB/NNPs were spherical, with no differences in morphology as observed by TEM (Fig. 4 A and B). Both NNPs and BNPs had a loading of EB of 2% by weight. When EB/NNPs and EB/BNPs particles were incubated in PBS at 37 °C, the majority (about 80%) of the EB was released during the initial 8 h of incubation (Fig. 4C). To test for in vitro cytotoxicity,
APPLIED BIOLOGICAL SCIENCES
Fig. 2. TEM images of (A) DiD/NNPs, (B) DiD/BNPs, and (C) DiD/NNP-Rs (DiD/NNPs reduced from DiD/BNPs by NaBH4 treatment). (Scale bars: 200 nm.) (D) The concentration of aldehydes on BNPs as a function of incubation time with NaBH4. (E) Retention of BNPs on (poly-L-lysine)–coated plates compared with NNPs and NNP-Rs. *P < 0.05 compared to both other groups. (F) Retention of BNPs on USC cell monolayers compared with NNPs and NNP-Rs. *P < 0.05 compared to both other groups. (G) The percentage of dose of DiD/NNPs, DiD/BNPs, and DiD/NNP-Rs in blood as a function of time after i.v. injection in mice.
(P = 0.03) or EB, with 60% of the animals surviving longer than 110 d (Fig. 6C). As expected, the survival rate for mice treated with blank BNPs was indistinguishable from that of mice treated with PBS (P = 0.83) (Fig. 6C). No significant difference was observed between mice treated with EB/NNPs and mice treated with free EB (P = 0.32), both with ∼10% of the animals surviving to 110 d.
Fig. 3. Imaging of DiD/BNPs and DiD/NNPs ex vivo and in vivo. (A) DiD/BNPs and DiD/NNPs were incubated at 1 mg/mL in Ringer’s buffer within the lumen of umbilical veins for 2 h at 37 °C. The red fluorescence signal represents nanoparticles retained on the luminal surface of the vein after extensive washing. (B) Quantification of fluorescence retained on vein. Data are shown as means ± SD. P < 0.0001 (Student t test). (C ) The retention of the IR-780/NNPs and IR-780/BNPs monitored with live imaging after i.p. administration.
injection compared with EB/NNPs injection (P < 0.05). No EB was detectable after 3 d for all formulations. Our hypothesis is that EB/BNPs can protect the peritoneum against the attachment of floating cancer cells in the peritoneal fluid; this protection is provided by BNP retention on protein-rich peritoneal surfaces and their ability to deliver EB locally to treat adherent cells (Fig. 1). To simulate the microenvironment in which we expect the nanoparticles to reside after i.p. injection, we evaluated the in vitro efficacy of surface-immobilized EB/BNPs to suppress the growth of USC cells. Here, we used microscope slides coated with poly(L-lysine) and incubated them with free EB, unloaded nanoparticles, or EB-loaded nanoparticles (Fig. 5A). We observed significant suppression of tumor cell growth only in slide regions pretreated with EB/BNPs (Fig. 5 B and C). No suppression of tumor growth was observed on surfaces treated with either blank BNPs or EB/NNPs. Although this is an in vitro model system and lacks many of the features of the i.p. environment, these results suggest that BNPs are retained on the lysine-coated surface of the slides because of their bioadhesive properties and able to deliver EB locally to adjacent cells more effectively than any of the other nanoparticle preparations. We evaluated the safety and efficacy of EB treatments in nude mice harboring class III β-tubulin overexpressing USC xenografts in the peritoneal cavity. One week after tumor inoculation, EB/ BNPs, EB/NNPs, and free EB were administrated weekly at an EB dose of 2.5 (Fig. 6 A and B) or 0.5 mg/kg (Fig. 6 C and D). Our intent was to treat animals in both experiments for 5 wk, but animals receiving the higher dose of free EB (2.5 mg/kg) experienced weight loss (Fig. 6B) and early death caused by EB toxicity (Fig. 6A); therefore, all animals in the high-dose (2.5 mg/kg) experiment were treated for only 3 wk. At both doses, mice treated with EB/BNPs survived, on average, significantly longer than control animals treated with PBS (P = 0.0006, 2.5 mg/kg; P = 0.0005, 0.5 mg/kg), but in mice treated with free EB, only mice treated with the lower dosage showed significance (P = 0.17, 2.5 mg/kg; P = 0.01, 0.5 mg/kg) (Fig. 6 A and C). Especially of note, mice treated with EB/BNPs (0.5 mg/kg) survived, on average, significantly longer than mice treated with EB/NNPs 11456 | www.pnas.org/cgi/doi/10.1073/pnas.1523141113
Discussion In this study, we have synthesized BNPs that interact with peritoneal tissues, dramatically enhancing their duration of retention in the abdominal cavity after i.p. administration. Previous studies have shown that the diameter of the lymphatic ducts in the peritoneal cavity is about 1 μm. Microparticles, which typically have diameters larger than 4 μm, can escape the lymphatic drainage, resulting in long-lasting retention after i.p. administration (4, 31). However, microparticles tend to accumulate in the lower abdomen because of gravitational forces (30), likely leading to regions of high and low drug concentration. This heterogeneous distribution may also cause potential serious side effects, including inflammation and peritoneal adhesion (9–11), limiting their clinical application. Although multiple studies have shown the advantages of nanoparticles in the delivery of drugs and biologic agents, the use of nanoparticles in the abdominal cavity is limited because by their rapid clearance caused by their small size (∼100 nm) (19). Here, we show that the addition of a dense coating of aldehyde groups on the particle surface confers bioadhesive properties to the nanoparticles (converting them into BNPs). This adhesive property promotes nanoparticle interaction with tissues, therefore extending their retention in the abdomen after i.p. injection (Fig. 3). Based on an in vitro model study (Fig. 5), we believe that this prolonged retention of BNPs leads to improved treatment of disseminated tumors in the abdominal cavity. Consistent with this hypothesis, free EB and
Fig. 4. Characterization of EB/BNPs. TEM images of (A) EB/BNPs and (B) EB/ NNPs. (Scale bars: 200 nm.) (C) EB is released from EB/BNPs and EB/NNPs over a period of 24 h of incubation in PBS. Data are shown as means ± SD (n = 4). (D) Viability of USC cells after incubation with free EB, EB/NNPs, and EB/BNPs for 3 d. Data are shown as means ± SD (n = 8). (E) Viability of USC cells after incubation with blank BNPs and NNPs for 3 d. Data are shown as means ± SD (n = 8). (F) EB retention in the peritoneal cavity after i.p. administration of free EB, EB/NNPs, and EB/BNPs. Data are shown as means ± SD (n = 3).
Deng et al.
Deng et al.
APPLIED BIOLOGICAL SCIENCES
EB/NNPs were rapidly cleared from the peritoneal cavity, whereas EB was retained much longer when administered in EB/BNPs (Fig. 4F). Our measurement of release from BNPs in vitro (∼90% released in 24 h) (Fig. 4C) is consistent with the drug retention observed in vivo (∼5% retained at 24 h) (Fig. 4F), despite the fact that the in vitro release was measured in an artificial system in which the nanoparticles were suspended in buffered saline, which is not identical to the conditions experienced by BNPs in the peritoneal space. We believe that slow EB release over 24 h leads to the reduced EB toxicity and enhanced EB effectiveness (Fig. 6A). The prolonged retention time that we observed for BNPs (some particles are retained for 5–10 d) (Fig. 3C and Fig. S2B) may not be needed in this particular case, but this long retention may prove to be useful in other applications: for example, for BNPs loaded with other drugs that are released over longer periods. When not loaded with drug agents, the BNPs are not toxic. In cell cultures, BNPs did not show any cytotoxicity, even at high concentration (i.e., up to 1 mg/mL), and mice treated with BNPs (i.p. injection of 5 mg BNPs once a week for 5 wk) had no evidence of weight loss or behavioral changes. In addition, the repeated i.p. injection of BNPs did not seem to lead to any nonspecific responses that affected tumor growth or animal health (Fig. 6). We attribute the low toxicity of BNPs to several factors. (i) Although free lowmolecular weight aldehydes can be toxic (32), the surface aldehyde groups are covalently attached to BNPs, limiting their dispersion. (ii) High tolerance to aldehydes is expected, because they are widely present in foods, fragrances, and metabolites and can be efficiently detoxified by the enzyme aldehyde dehydrogenase (33). In these studies, we selected EB as an agent for use in i.p. administration. EB has remarkable efficacy against PTX-resistant uterine serous and ovarian cancers (34–36), but prior clinical testing has revealed substantial toxicity when EB is administered as a free drug (26). We hypothesized that encapsulation in nanoparticles would reduce the toxicity of EB and make it
Fig. 6. Therapeutic efficacy of EB/BNPs on mice bearing i.p. USC tumors. (A) Kaplan–Meier survival curves for EB treatments at 2.5 mg/kg. (B) As a measure of toxicity, mice treated as in A were weighed twice a week. (C ) Kaplan–Meier survival curves for EB treatments at 0.5 mg/kg. (D) Mice treated as in C were weighed twice a week. In A and C, the x axis indicates the number of days from the first dose of drug. In B and D, the average weight is displayed as a function of days from the first drug dose. Statistical analysis was performed as described in SI Methods.
PNAS | October 11, 2016 | vol. 113 | no. 41 | 11457
ENGINEERING
Fig. 5. Suppression of USC cell growth on EB/BNPs-treated surfaces. (A) Schematic showing experimental design. A glass slide was divided into five blocks, and each block was incubated individually with EB/BNPs, EB/ NNPs, free EB, BNPs, or PBS for 30 min. After extensive wash, the slide was incubated in RPMI medium with USC cells with density at 2.0 × 105/mL for 24 h. After wash, the USC cells on slides were stained with Hoechst (for nuclei; blue) and live/dead stain (green indicates live cells, and red indicates dead cells) and imaged with a fluorescence microscope. (B) Fluorescence images showing the USC cell growth on poly-(L-lysine)-coated slides pretreated with different EB formulations. (Scale bars: 1 mm.) (C) The surface density of the cells was quantified with ImageJ and normalized to the PBS control. Data are shown as means ± SD (n = 3). *P < 0.05.
suitable for i.p. therapy. We showed the effectiveness of these EB formulations by measuring the survival of mice with PTX-resistant USC xenografts. To further limit EB toxicity in animals, we used a fivefold lower dose of EB compared with previous studies in nude mice with human multiple myeloma xenografts (37). Even at this lower dose of EB, however, significant toxicity (i.e., weight loss, diarrhea, and bowel dilation) was observed in mice treated with free EB. Remarkably, very little toxicity was observed in the mice treated with EB/BNPs (Fig. 6B). We attribute this difference in toxicity to two factors associated with EB/BNPs: the gradual release of EB from the BNPs (Fig. 4C), which likely reduces peak drug levels, and the strong adhesion of BNPs to the peritoneal tissue (Fig. 3C), which likely reduces EB doses to nontarget tissues and vital organs by fast lymphatic drainage, although some local toxicity to adjacent normal tissue is likely unavoidable (although not observed in our studies). We showed that EB was quickly cleared from the abdominal cavity, likely by diffusion through the peritoneal membrane into capillaries and transport through the hepatic portal system (38), which would be expected to lead to side effects and significantly lower bioavailability (39). The main side effect caused by EB in our study was gastrointestinal toxicity, which is in agreement with the known toxicity profile of EB (26, 40, 41). Importantly, in the EB/BNPs-treated mice, we showed significantly improved survival compared with in the controls (P < 0.05), with 60% of the animals alive at the end of the experiment (110 d). We attribute this enhanced effectiveness to EB/BNP bioadhesion and slow release. As evidence for the importance of bioadhesion, EB/NNPs were not as effective as EB/BNPs. Recent studies have shown strong activity of small molecules that selectively target the Her2/PI3K/Akt/mTor pathway against USCs. However, treatments with a single agent were only transiently effective, and tumors rapidly acquired resistance in vivo (36, 42). Importantly, recent preclinical data suggest that dual targeting of HER2/PIK3CA with a Pan-HER inhibitor (Neratinib) and a PIK3CA inhibitor (Taselisib) may be synergistic and able to achieve durable regression of USC xenografts in mice (43). However, this strategy is only applicable for the treatment of patients whose USC tumors harbor the amplification of the c-erbB2 gene and the concomitant amplification or oncogenic mutations in the PI3KCA gene. Therefore, alternate strategies are likely needed. In this study, we propose the use of nanoparticles loaded with EB for the treatment of USC after local/regional (i.p.) administration. This strategy takes advantage of the chemical properties of the nanoparticles and the
extreme sensitivity of USCs to the EB and may, therefore, be used for the treatment of the vast majority of USC patients, regardless of the genetic landscape of their tumors (44). In summary, our results represent a demonstration that i.p. administration of BNPs encapsulating EB can enhance survival of mice with i.p. tumors formed from USC xenografts. Importantly, the unique formulation of EB/BNPs exhibits higher therapeutic efficacy and lower toxicity than free EB and EB/NNPs. Because the bioadhesive property of BNPs is based on the general interaction between aldehydes on nanoparticles and proteins on tissue—which is illustrated in the variety of systems that we studied here (Figs. 2 and 3A)—we expect this approach to be applicable to other strategies for local delivery of active agents.
emulsion–evaporation technique and characterized using TEM and DLS. BNPs were formed using a previously described method (20) of using NaIO4 (Fig. 1) to convert vicinal diols to aldehyde groups and reversing back to NNP-Rs by adding NaBH4 (Fig. S1A). Retention of fluorescent BNPs was tested in vitro using a lysine-coated plate and wells containing USC cells by washing wells after particle incubation and reading the measurements on a fluorescent plate reader. Ex vivo retention was evaluated using a human umbilical cord and visualized with a fluorescent microscope. In vivo retention, distribution, and efficacy of dye- and drugloaded particles were tested in C57BL/6 and nude mice with USC cancers. Characterization was completed with the use of an in vivo imaging system (Xenogen), and efficacy was measured with strict survival guidelines. All animals were handled in accordance with the policies and guidelines of the Yale Institutional Animal Care and Use Committee. Detailed experimental procedures are provided in SI Methods.
Methods We used previously characterized BNPs to overcome barriers present in i.p. delivery of therapeutics. PLA-HPG was synthesized as previously reported (21). NNPs loaded with DiD, IR-780, and EB were formed using an
ACKNOWLEDGMENTS. We thank Yukun Pan, Tian Xu, Yu Wu, Yao Lu, and Rong Fan of Yale University for access to instruments in their laboratories. This work was supported by NIH Grants CA149128 and CA154460.
1. Greggi S, et al. (2001) Topotecan and gemcitabine in platinum/paclitaxel-resistant ovarian cancer. Oncology 60(1):19–23. 2. Tewari D, et al. (2015) Long-term survival advantage and prognostic factors associated with intraperitoneal chemotherapy treatment in advanced ovarian cancer: A gynecologic oncology group study. J Clin Oncol 33(13):1460–1466. 3. Bartlett DL (2008) HIPEC: The complexities of clinical trials. Ann Surg Oncol 15(5): 1277–1279. 4. Tsai M, et al. (2007) Effects of carrier on disposition and antitumor activity of intraperitoneal Paclitaxel. Pharm Res 24(9):1691–1701. 5. Bajaj G, Kim MR, Mohammed SI, Yeo Y (2012) Hyaluronic acid-based hydrogel for regional delivery of paclitaxel to intraperitoneal tumors. J Control Release 158(3): 386–392. 6. Emoto S, et al. (2014) Intraperitoneal administration of cisplatin via an in situ crosslinkable hyaluronic acid-based hydrogel for peritoneal dissemination of gastric cancer. Surg Today 44(5):919–926. 7. Wang Y, et al. (2010) 5-FU-hydrogel inhibits colorectal peritoneal carcinomatosis and tumor growth in mice. BMC Cancer 10:402. 8. Wu Q, et al. (2015) Thermosensitive hydrogel containing dexamethasone micelles for preventing postsurgical adhesion in a repeated-injury model. Sci Rep 5:13553. 9. Armstrong DK, Fleming GF, Markman M, Bailey HH (2006) A phase I trial of intraperitoneal sustained-release paclitaxel microspheres (Paclimer) in recurrent ovarian cancer: A Gynecologic Oncology Group study. Gynecol Oncol 103(2):391–396. 10. Kohane DS, et al. (2006) Biodegradable polymeric microspheres and nanospheres for drug delivery in the peritoneum. J Biomed Mater Res A 77(2):351–361. 11. Lu Z, et al. (2008) Tumor-penetrating microparticles for intraperitoneal therapy of ovarian cancer. J Pharmacol Exp Ther 327(3):673–682. 12. Zhou J, et al. (2013) Highly penetrative, drug-loaded nanocarriers improve treatment of glioblastoma. Proc Natl Acad Sci USA 110(29):11751–11756. 13. Ediriwickrema A, Zhou J, Deng Y, Saltzman WM (2014) Multi-layered nanoparticles for combination gene and drug delivery to tumors. Biomaterials 35(34):9343–9354. 14. Ediriwickrema A, Saltzman WM (2015) Nanotherapy for cancer: Targeting and multifunctionality in the future of cancer therapies. ACS Biomater Sci Eng 1(2):64–78. 15. Saucier-Sawyer JK, et al. (2015) Systemic delivery of blood-brain barrier-targeted polymeric nanoparticles enhances delivery to brain tissue. J Drug Target 23(7-8): 736–749. 16. Gupta A, Bahal R, Gupta M, Glazer PM, Saltzman WM (January 8, 2016) Nanotechnology for delivery of peptide nucleic acids (PNAs). J Control Release, 10.1016/j. jconrel.2016.01.005. 17. Cheng CJ, Tietjen GT, Saucier-Sawyer JK, Saltzman WM (2015) A holistic approach to targeting disease with polymeric nanoparticles. Nat Rev Drug Discov 14(4): 239–247. 18. McNeer NA, et al. (2015) Nanoparticles that deliver triplex-forming peptide nucleic acid molecules correct F508del CFTR in airway epithelium. Nat Commun 6:6952. 19. Bajaj G, Yeo Y (2010) Drug delivery systems for intraperitoneal therapy. Pharm Res 27(5):735–738. 20. Deng Y, et al. (2015) A sunblock based on bioadhesive nanoparticles. Nat Mater 14(12):1278–1285. 21. Deng Y, et al. (2014) The effect of hyperbranched polyglycerol coatings on drug delivery using degradable polymer nanoparticles. Biomaterials 35(24):6595–6602. 22. Mozzetti S, et al. (2008) Molecular mechanisms of patupilone resistance. Cancer Res 68(24):10197–10204. 23. Nobili S, Landini I, Mazzei T, Mini E (2012) Overcoming tumor multidrug resistance using drugs able to evade P-glycoprotein or to exploit its expression. Med Res Rev 32(6):1220–1262.
24. Paik D, et al. (2010) Higher sensitivity to patupilone versus paclitaxel chemotherapy in primary uterine serous papillary carcinoma cell lines with high versus low HER-2/neu expression in vitro. Gynecol Oncol 119(1):140–145. 25. Roque DM, et al. (2013) Tubulin-β-III overexpression by uterine serous carcinomas is a marker for poor overall survival after platinum/taxane chemotherapy and sensitivity to epothilones. Cancer 119(14):2582–2592. 26. Peereboom DM, et al. (2014) Phase II trial of patupilone in patients with brain metastases from breast cancer. Neuro-oncol 16(4):579–583. 27. de Jonge M, Verweij J (2005) The epothilone dilemma. J Clin Oncol 23(36):9048–9050. 28. Cocco E, et al. (2015) Clostridium perfringens enterotoxin C-terminal domain labeled to fluorescent dyes for in vivo visualization of micrometastatic chemotherapyresistant ovarian cancer. Int J Cancer 137(11):2618–2629. 29. Hermanson GT (2013) Bioconjugate Techniques (Elsevier/AP, London), 3rd Ed. 30. Steinbacher JL, et al. (2010) Gd-labeled microparticles in MRI: In vivo imaging of microparticles after intraperitoneal injection. Small 6(23):2678–2682. 31. Yang M, et al. (2014) Intraperitoneal delivery of paclitaxel by poly(ether-anhydride) microspheres effectively suppresses tumor growth in a murine metastatic ovarian cancer model. Drug Deliv Transl Res 4(2):203–209. 32. O’Brien PJ, Siraki AG, Shangari N (2005) Aldehyde sources, metabolism, molecular toxicity mechanisms, and possible effects on human health. Crit Rev Toxicol 35(7):609–662. 33. Vasiliou V, Thompson DC, Smith C, Fujita M, Chen Y (2013) Aldehyde dehydrogenases: From eye crystallins to metabolic disease and cancer stem cells. Chem Biol Interact 202(1-3):2–10. 34. El-Sahwi KS, Schwartz PE, Santin AD (2012) Development of targeted therapy in uterine serous carcinoma, a biologically aggressive variant of endometrial cancer. Expert Rev Anticancer Ther 12(1):41–49. 35. Black JD, English DP, Roque DM, Santin AD (2014) Targeted therapy in uterine serous carcinoma: An aggressive variant of endometrial cancer. Womens Health (Lond) 10(1):45–57. 36. Roque DM, et al. (2013) Class III β-tubulin overexpression in ovarian clear cell and serous carcinoma as a maker for poor overall survival after platinum/taxane chemotherapy and sensitivity to patupilone. Am J Obstet Gynecol 209(1):62.e1–62.e9. 37. Lin B, et al. (2005) Patupilone (epothilone B) inhibits growth and survival of multiple myeloma cells in vitro and in vivo. Blood 105(1):350–357. 38. De Bree E, Zoetmulder FAN, Romanos J, Witkamp AJ, Tsiftsis DD (2003) Intraperitoneal chemotherapy for prevention and treatment of peritoneal carcinomatosis from colorectal origin. Ann Gastroenterol 16(1):20–33. 39. Goodin S, Kane MP, Rubin EH (2004) Epothilones: Mechanism of action and biologic activity. J Clin Oncol 22(10):2015–2025. 40. Colombo N, et al. (2012) Randomized, open-label, phase III study comparing patupilone (EPO906) with pegylated liposomal doxorubicin in platinum-refractory or -resistant patients with recurrent epithelial ovarian, primary fallopian tube, or primary peritoneal cancer. J Clin Oncol 30(31):3841–3847. 41. Forster M, et al. (2007) A phase Ib and pharmacokinetic trial of patupilone combined with carboplatin in patients with advanced cancer. Clin Cancer Res 13(14):4178–4184. 42. Lopez S, et al. (2014) Taselisib, a selective inhibitor of PIK3CA, is highly effective on PIK3CA-mutated and HER2/neu amplified uterine serous carcinoma in vitro and in vivo. Gynecol Oncol 135(2):312–317. 43. Lopez S, et al. (2015) Dual HER2/PIK3CA targeting overcomes single-agent acquired resistance in HER2-amplified uterine serous carcinoma cell lines in vitro and in vivo. Mol Cancer Ther 14(11):2519–2526. 44. Cross SN, et al. (2010) Differential sensitivity to platinum-based chemotherapy in primary uterine serous papillary carcinoma cell lines with high vs low HER-2/neu expression in vitro. Am J Obstet Gynecol 203(2):162.e1–162.e8.
11458 | www.pnas.org/cgi/doi/10.1073/pnas.1523141113
Deng et al.
Edited by Joseph M. DeSimone, University of North Carolina at Chapel Hill and Carbon, Chapel Hill, NC, and approved August 19, 2016 (received for review November 22, 2015)
| ovarian cancer | intraperitoneal |
H
igh-grade ovarian and uterine serous carcinoma (USC) are biologically aggressive tumors, which commonly spread into the peritoneal cavity, disseminating along the abdominal and pelvic peritoneum and resulting in peritoneal metastases. Although these tumors are often responsive to initial cytoreductive surgery and platinum–paclitaxel chemotherapy, most patients develop recurrence and eventually die from progression of chemotherapy-resistant disease (1). In cases of advanced ovarian cancer, the infusion of chemotherapeutic agents directly into the peritoneum (i.e., i.p. chemotherapy) improves overall and disease-free survival compared with standard i.v. chemotherapy (2). However, despite level 1 evidence from meta-analyses of three large, prospectively randomized, controlled trials, this route of administration is still not widely used in the clinic because of its more severe side effects compared with i.v. chemotherapy and the need for frequent dosing schedules (2, 3). To enhance the retention of chemotherapeutic drugs in the peritoneal cavity and improve their bioavailability, controlled release drug carriers, such as microparticles and hydrogels, have recently been tested in preclinical animal models (4–8). Although these carriers can lead to improvement over conventional delivery methods, both strategies only use the microparticles or hydrogel as reservoirs for drugs. To achieve expected therapeutic efficacy, the concentration of the free drugs in the peritoneal cavity has to be maintained in the effective range. Sustained drug concentrations require a balance between the drug release from carriers and clearance from peritoneal cavity by metabolism. For a specific drug and drug carrier combination, release that is too slow might also compromise the therapeutic efficacy as observed when limited dissolution of paclitaxel from a microparticle/hydrogel system impaired therapeutic outcomes (5). In addition, side effects of some of these carriers do exist (i.e., microparticles accumulate in the lower abdomen and cause peritoneal adhesion) (9–11). Degradable polymer nanoparticles have several potential advantages over microparticles and hydrogels in cancer therapy (12–16). Nanoparticles (∼100 nm) are small compared with microparticles www.pnas.org/cgi/doi/10.1073/pnas.1523141113
Significance Resistance to platinum-based chemotherapies and paclitaxel is common in recurrence of both high-grade ovarian and endometrial cancers. Paclitaxel resistance has been correlated with overexpression of class III β-tubulin, the preferential target of the epothilones, microtubule-stabilizing agents. Epothilone B (EB) is manifold more effective than paclitaxel, but clinical use is limited by side effects. To reduce side effects, we encapsulated EB into bioadhesive nanoparticles (BNPs), reasoning that bioadhesive nanoparticles loaded with epothilone B (EB/BNPs) would interact with abdominal tissues and gradually release EB in proximity of peritoneal cancer implants, thus maintaining EB concentration at the site of action and limiting systemic exposure and toxicity. Our experiments show the higher therapeutic activity and limited toxicity of EB/BNPs compared with nonadhesive nanoparticles loaded with EB or carrier-free EB. Author contributions: Y.D., F.Y., E.C., E.S., J.Z., J.C., M.M., A.D.S., and W.M.S. designed research; Y.D., F.Y., E.C., E.S., J.Z., J.C., M.M., S.B., A.D.S., and W.M.S. performed research; Y.D., F.Y., E.C., E.S., J.Z., J.C., M.M., S.B., A.D.S., and W.M.S. analyzed data; and Y.D., E.C., A.D.S., and W.M.S. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1
To whom correspondence should be addressed. Email: [email protected].
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1523141113/-/DCSupplemental.
PNAS | October 11, 2016 | vol. 113 | no. 41 | 11453–11458
APPLIED BIOLOGICAL SCIENCES
|
drug delivery nanoparticles chemotherapy
(∼10–100 μm) and therefore, should distribute more evenly throughout the peritoneum. Furthermore, nanoparticles are small enough to be internalized by tumor cells (17, 18), allowing them to release the encapsulated drugs near the biological target, which is often in the cell nucleus. Although many studies have shown that nanoparticles are promising vehicles for the delivery of chemotherapeutic agents in preclinical in vivo models, the application of nanoparticles for i.p. delivery remains hampered by their fast clearance from the peritoneal cavity, mainly as a result of lymphatic drainage (19). In a recent study (20), we described a novel form of bioadhesive nanoparticles (BNPs) and showed that their adhesiveness results in prolonged retention on the skin after topical delivery. These new materials are based on polylactic acid block– hyperbranched polyglycerol (PLA-HPG) copolymers, which can be formed into “stealthy,” nonadhesive nanoparticles (NNPs) that circulate for long times after i.v. injection (21). The BNPs were produced by oxidation of the NNPs, which results in conversion of vicinal diols on the surface of NNPs into aldehydes on the BNPs (Fig. 1). Aldehydes spontaneously react with proteins to form a variety of bonds, such as Schiff-base bonds; therefore, the presence of surface aldehydes on BNPs leads to particle adhesion to protein-rich materials. For example, we previously showed the rapid and robust adhesion of the BNPs to poly-Llysine–coated surfaces and pig or mouse skin (20). In this work, we hypothesize that BNPs will interact and adhere strongly with mesothelial cells layering the abdominal cavity after i.p. delivery
ENGINEERING
The i.p. administration of chemotherapy in ovarian and uterine serous carcinoma patients by biodegradable nanoparticles may represent a highly effective way to suppress peritoneal carcinomatosis. However, the efficacy of nanoparticles loaded with chemotherapeutic agents is currently hampered by their fast clearance by lymphatic drainage. Here, we show that a unique formulation of bioadhesive nanoparticles (BNPs) can interact with mesothelial cells in the abdominal cavity and significantly extend the retention of the nanoparticles in the peritoneal space. BNPs loaded with a potent chemotherapeutic agent [epothilone B (EB)] showed significantly lower systemic toxicity and higher therapeutic efficacy against i.p. chemotherapy-resistant uterine serous carcinomaderived xenografts compared with free EB and non-BNPs loaded with EB.
successfully used in prior work to generate consistent i.p. tumors in mice (28).
Fig. 1. Schematic showing the conversion of BNPs from NNPs and the fate of NNPs and BNPs after i.p. delivery. BNPs can be converted from NNPs by NaIO4 treatment. After i.p. delivery, (Left) NNPs are cleared by lymphatic drainage, but (Right) BNPs are retained in the peritoneal cavity because of their bioadhesive property.
(Fig. 1). We also hypothesize that the interaction of the BNPs with abdominal tissue will extend the retention of the BNPs after i.p. injection, therefore significantly improving the bioavailability of their payloads and the efficacy of i.p. chemotherapy. We tested the effectiveness of this approach for i.p. delivery of patupilone [epothilone B (EB)], a microtubule-stabilizing agent that targets class III β-tubulin. Importantly, EB has been shown to be 3- to 20-fold more effective than Paclitaxel (PTX) in vitro against multiple PTX-sensitive and -resistant human tumor cell lines (22–25). Unfortunately, the clinical use of EB in patients with recurrent disease is limited because of high toxicity, including severe diarrhea, vomiting, increased risk of bowel obstruction, and fatigue (26, 27). Here, we use our BNPs loaded with EB to test whether strongly adhesive nanoparticles will improve the efficacy of i.p. chemotherapy to PTX-resistant peritoneal metastasis. In particular, we developed i.p. xenograft models derived from the primary USC cell line USC-ARK-2. This cell line was selected (i) because of its high expression of P-glycoprotein and tubulin-β-III, characteristics that have been recently shown by our group to confer PTX resistance while showing extreme sensitivity to EB (25), and (ii) because this cell line has been 11454 | www.pnas.org/cgi/doi/10.1073/pnas.1523141113
Results In a previous work, we showed that NNPs had an extended blood circulation because of their resistance to nonspecific interaction to proteins (21). In another recent study, we found that BNPs had an enhanced interaction with protein-rich surfaces caused by their aldehyde groups (20). These results suggest that this class of nanoparticles can be converted from a nonadhesive to a strongly adhesive state and then back again by regulation of the density of aldehydes on the nanoparticle surface. To illustrate this flexibility, nanoparticles were converted by a cycle of oxidation and reduction between the hydroxyl-rich NNP state, the aldehyde-rich BNP state, and a reduced nonadhesive nanoparticle (NNP-R) state. We reduced the aldehydes to alcohols by treatment with a common reducing agent (29), NaBH4 (Fig. S1A). The concentration of aldehyde groups on BNPs was monitored as a function of duration of NaBH4 treatment; most aldehyde groups were converted after 2 min of treatment (Fig. 2D). It is important to note that NaBH4 treatment cannot reduce the aldehydes to the original vicinal diols, because one carbon is truncated by the initial conversion of vicinal diols to aldehydes (Fig. S1B). BNP, NNP, and NNP-R were identical in shape by transmission EM (TEM) (Fig. 2 A–C) and hydrodynamic size measured by dynamic light scattering (DLS) (Table 1), suggesting that oxidation and reduction cycles had no detrimental effect to the nanoparticles. To quantitate bioadhesion, we tested nanoparticles at various states of aldehyde activation by examining attachment to poly(L-lysine)–coated plates and USC cells (Fig. 2 E and F). To facilitate quantification of their bioadhesive properties, all nanoparticles were loaded with 0.2% 1,1′-dioctadecyl3,3,3′,3′-tetramethylindodicarbocyanine, 4-chlorobenzenesulfonate salt (DiD) dye. We first evaluated the bioadhesive property of the three nanoparticles on poly(L-lysine)–coated surfaces. After incubation and extensive washing, the BNPs showed much higher retention on the poly(L-lysine)–coated plates (Fig. 2E). The BNPs also showed much higher attachment to USC cells (Fig. 2F). The retention of NNPs or NNP-Rs on both poly(Llysine) surfaces and USC cell monolayers was negligible (Fig. 2 E and F). We further tested the nanoparticles for blood circulation times after i.v. injection in mice. The percentage dose of BNPs dropped to 1.5% after 2 h, whereas significant numbers of NNPs (21.3%) were still circulating after 24 h (Fig. 2G). NNP-Rs circulated with similar duration as the NNPs, reflecting the chemical similarity of their surfaces. We hypothesize that BNPs will adhere avidly to protein-rich tissue surfaces and remain adherent for long periods. To test this hypothesis, we first characterized the adhesive potential of BNPs and NNPs on human umbilical veins. Both DiD/NNPs and DiD/ BNPs were applied individually to the luminal surface of human umbilical veins. After 2 h of incubation within the lumen followed by extensive washing by vessel perfusion with 30 mL PBS, the BNPs showed much higher retention on the vessel endothelium than NNPs (Fig. 3 A and B). Most significantly for this study, we evaluated the retention of nanoparticles within the peritoneum after i.p. administration in mice. In this case, BNPs and NNPs were loaded with 0.5% IR-780 iodide (IR-780) dye to facilitate imaging. At this loading, only a small fraction of the IR780 dye is released from the nanoparticles, and it is released slowly over 10 d of incubation in buffered saline supplemented with physiological levels of protein (Fig. S2A); therefore, the dye molecule serves as a reliable marker for the presence of nanoparticles. As we have shown with other NNP/BNP combinations, the conversion of IR-780/NNPs to IR-780/BNPs (by treatment with NaIO4) had no apparent effect on nanoparticle size and morphology as observed by TEM (Fig. S3) or hydrodynamic size as measured by DLS (Table 1). After i.p. injection of NNPs, most of Deng et al.
Table 1. Nanoparticle size measurements Nanoparticles EB/NNPs EB/BNPs DiD/NNPs DiD/BNPs DiD/NNP-Rs IR-780/NNPs IR-780/BNPs
Deng et al.
Diameter (nm)
Polydispersity index
127 127 121 117 97 131 128
0.225 0.233 0.215 0.232 0.213 0.331 0.297
EB/BNPs, EB/NNPs, and free EB were separately added to USC cell cultures, which were tested for viability after 3 d of exposure. All EB treatments (EB, EB/NNPs, and EB/BNPs) yielded similar dose–response curves, with the free EB showing the highest toxicity (i.e., lowest IC50) (Fig. 4D). We believe that this difference between the free EB and nanoparticle formulations of EB is caused by the slow release of EB from NNPs and BNPs, resulting in a slightly lower concentration of EB in the culture medium exposed to nanoparticle formulations compared with free EB. To confirm that the cytotoxicity was caused by the EB and not the polymers within the nanoparticles, we also incubated USC cells with unloaded (blank) BNPs and NNPs. Unloaded nanoparticles did not show any toxicity, even when added to the culture medium at high concentrations (1 mg/mL) (Fig. 4E). The tolerable concentration of blank nanoparticles is much higher than the nanoparticle concentration of EB/NNPs or EB/BNPs needed to deliver an effective dose: for example, 10 nM EB is contained within 0.25 μg/mL EB/NNPs or EB/BNPs. To further show the safety of the nanoparticle treatments, we examined the cytotoxicity of blank BNPs and NNPs on human umbilical vein endothelial cells and HeLa cells. Neither BNPs nor NNPs were toxic in these cells up to 1 mg/mL (Fig. S5A). Furthermore, we incubated blank BNPs or NNPs at 1 mg/mL on a variety of immortalized cell lines and observed no cytotoxicity at high concentration (Fig. S5B). To measure retention of EB within the peritoneal tissue, we i.p. administered free EB, EB/BNPs, and EB/NNPs in nude mice. No EB was detected in mice receiving the free EB at 6 h after a single dose (Fig. 4F). In contrast, detectable levels of EB were found at 6 h with both EB/NNPs and EB/BNPs injections: the EB/BNPs-treated mice had substantially higher concentrations of EB than the EB/NNPs-treated mice (21% vs. 5%, P < 0.05) (Fig. 4F). One day after i.p. administration, EB was still significantly higher after EB/BNPs PNAS | October 11, 2016 | vol. 113 | no. 41 | 11455
ENGINEERING
the fluorescent signal, which was initially strong (5 min), was lost by the end of the first day (Fig. 3C). In contrast, after i.p. injection of BNPs, a strong fluorescent signal was observed to last up to 5 d (Fig. 3C); at the end of the 10th day, BNPs were still detectable in the area of the initial injection (Fig. S2B). Remarkably, accumulation of BNPs in the lower abdomen, which is commonly seen after microparticle i.p. delivery (30), was not observed with administration of BNPs; the majority of the BNPs adhered to the parietal peritoneum of the mice (Fig. S4). To determine the efficacy of nanoparticles as vehicles for i.p. delivery, we loaded NNPs and BNPs with EB. Nonadhesive nanoparticles loaded with epothilone B (EB/NNPs) and bioadhesive nanoparticles loaded with epothilone B (EB/BNPs) were of similar hydrodynamic size, ∼130 nm in diameter (Table 1). EB/BNPs and EB/NNPs were spherical, with no differences in morphology as observed by TEM (Fig. 4 A and B). Both NNPs and BNPs had a loading of EB of 2% by weight. When EB/NNPs and EB/BNPs particles were incubated in PBS at 37 °C, the majority (about 80%) of the EB was released during the initial 8 h of incubation (Fig. 4C). To test for in vitro cytotoxicity,
APPLIED BIOLOGICAL SCIENCES
Fig. 2. TEM images of (A) DiD/NNPs, (B) DiD/BNPs, and (C) DiD/NNP-Rs (DiD/NNPs reduced from DiD/BNPs by NaBH4 treatment). (Scale bars: 200 nm.) (D) The concentration of aldehydes on BNPs as a function of incubation time with NaBH4. (E) Retention of BNPs on (poly-L-lysine)–coated plates compared with NNPs and NNP-Rs. *P < 0.05 compared to both other groups. (F) Retention of BNPs on USC cell monolayers compared with NNPs and NNP-Rs. *P < 0.05 compared to both other groups. (G) The percentage of dose of DiD/NNPs, DiD/BNPs, and DiD/NNP-Rs in blood as a function of time after i.v. injection in mice.
(P = 0.03) or EB, with 60% of the animals surviving longer than 110 d (Fig. 6C). As expected, the survival rate for mice treated with blank BNPs was indistinguishable from that of mice treated with PBS (P = 0.83) (Fig. 6C). No significant difference was observed between mice treated with EB/NNPs and mice treated with free EB (P = 0.32), both with ∼10% of the animals surviving to 110 d.
Fig. 3. Imaging of DiD/BNPs and DiD/NNPs ex vivo and in vivo. (A) DiD/BNPs and DiD/NNPs were incubated at 1 mg/mL in Ringer’s buffer within the lumen of umbilical veins for 2 h at 37 °C. The red fluorescence signal represents nanoparticles retained on the luminal surface of the vein after extensive washing. (B) Quantification of fluorescence retained on vein. Data are shown as means ± SD. P < 0.0001 (Student t test). (C ) The retention of the IR-780/NNPs and IR-780/BNPs monitored with live imaging after i.p. administration.
injection compared with EB/NNPs injection (P < 0.05). No EB was detectable after 3 d for all formulations. Our hypothesis is that EB/BNPs can protect the peritoneum against the attachment of floating cancer cells in the peritoneal fluid; this protection is provided by BNP retention on protein-rich peritoneal surfaces and their ability to deliver EB locally to treat adherent cells (Fig. 1). To simulate the microenvironment in which we expect the nanoparticles to reside after i.p. injection, we evaluated the in vitro efficacy of surface-immobilized EB/BNPs to suppress the growth of USC cells. Here, we used microscope slides coated with poly(L-lysine) and incubated them with free EB, unloaded nanoparticles, or EB-loaded nanoparticles (Fig. 5A). We observed significant suppression of tumor cell growth only in slide regions pretreated with EB/BNPs (Fig. 5 B and C). No suppression of tumor growth was observed on surfaces treated with either blank BNPs or EB/NNPs. Although this is an in vitro model system and lacks many of the features of the i.p. environment, these results suggest that BNPs are retained on the lysine-coated surface of the slides because of their bioadhesive properties and able to deliver EB locally to adjacent cells more effectively than any of the other nanoparticle preparations. We evaluated the safety and efficacy of EB treatments in nude mice harboring class III β-tubulin overexpressing USC xenografts in the peritoneal cavity. One week after tumor inoculation, EB/ BNPs, EB/NNPs, and free EB were administrated weekly at an EB dose of 2.5 (Fig. 6 A and B) or 0.5 mg/kg (Fig. 6 C and D). Our intent was to treat animals in both experiments for 5 wk, but animals receiving the higher dose of free EB (2.5 mg/kg) experienced weight loss (Fig. 6B) and early death caused by EB toxicity (Fig. 6A); therefore, all animals in the high-dose (2.5 mg/kg) experiment were treated for only 3 wk. At both doses, mice treated with EB/BNPs survived, on average, significantly longer than control animals treated with PBS (P = 0.0006, 2.5 mg/kg; P = 0.0005, 0.5 mg/kg), but in mice treated with free EB, only mice treated with the lower dosage showed significance (P = 0.17, 2.5 mg/kg; P = 0.01, 0.5 mg/kg) (Fig. 6 A and C). Especially of note, mice treated with EB/BNPs (0.5 mg/kg) survived, on average, significantly longer than mice treated with EB/NNPs 11456 | www.pnas.org/cgi/doi/10.1073/pnas.1523141113
Discussion In this study, we have synthesized BNPs that interact with peritoneal tissues, dramatically enhancing their duration of retention in the abdominal cavity after i.p. administration. Previous studies have shown that the diameter of the lymphatic ducts in the peritoneal cavity is about 1 μm. Microparticles, which typically have diameters larger than 4 μm, can escape the lymphatic drainage, resulting in long-lasting retention after i.p. administration (4, 31). However, microparticles tend to accumulate in the lower abdomen because of gravitational forces (30), likely leading to regions of high and low drug concentration. This heterogeneous distribution may also cause potential serious side effects, including inflammation and peritoneal adhesion (9–11), limiting their clinical application. Although multiple studies have shown the advantages of nanoparticles in the delivery of drugs and biologic agents, the use of nanoparticles in the abdominal cavity is limited because by their rapid clearance caused by their small size (∼100 nm) (19). Here, we show that the addition of a dense coating of aldehyde groups on the particle surface confers bioadhesive properties to the nanoparticles (converting them into BNPs). This adhesive property promotes nanoparticle interaction with tissues, therefore extending their retention in the abdomen after i.p. injection (Fig. 3). Based on an in vitro model study (Fig. 5), we believe that this prolonged retention of BNPs leads to improved treatment of disseminated tumors in the abdominal cavity. Consistent with this hypothesis, free EB and
Fig. 4. Characterization of EB/BNPs. TEM images of (A) EB/BNPs and (B) EB/ NNPs. (Scale bars: 200 nm.) (C) EB is released from EB/BNPs and EB/NNPs over a period of 24 h of incubation in PBS. Data are shown as means ± SD (n = 4). (D) Viability of USC cells after incubation with free EB, EB/NNPs, and EB/BNPs for 3 d. Data are shown as means ± SD (n = 8). (E) Viability of USC cells after incubation with blank BNPs and NNPs for 3 d. Data are shown as means ± SD (n = 8). (F) EB retention in the peritoneal cavity after i.p. administration of free EB, EB/NNPs, and EB/BNPs. Data are shown as means ± SD (n = 3).
Deng et al.
Deng et al.
APPLIED BIOLOGICAL SCIENCES
EB/NNPs were rapidly cleared from the peritoneal cavity, whereas EB was retained much longer when administered in EB/BNPs (Fig. 4F). Our measurement of release from BNPs in vitro (∼90% released in 24 h) (Fig. 4C) is consistent with the drug retention observed in vivo (∼5% retained at 24 h) (Fig. 4F), despite the fact that the in vitro release was measured in an artificial system in which the nanoparticles were suspended in buffered saline, which is not identical to the conditions experienced by BNPs in the peritoneal space. We believe that slow EB release over 24 h leads to the reduced EB toxicity and enhanced EB effectiveness (Fig. 6A). The prolonged retention time that we observed for BNPs (some particles are retained for 5–10 d) (Fig. 3C and Fig. S2B) may not be needed in this particular case, but this long retention may prove to be useful in other applications: for example, for BNPs loaded with other drugs that are released over longer periods. When not loaded with drug agents, the BNPs are not toxic. In cell cultures, BNPs did not show any cytotoxicity, even at high concentration (i.e., up to 1 mg/mL), and mice treated with BNPs (i.p. injection of 5 mg BNPs once a week for 5 wk) had no evidence of weight loss or behavioral changes. In addition, the repeated i.p. injection of BNPs did not seem to lead to any nonspecific responses that affected tumor growth or animal health (Fig. 6). We attribute the low toxicity of BNPs to several factors. (i) Although free lowmolecular weight aldehydes can be toxic (32), the surface aldehyde groups are covalently attached to BNPs, limiting their dispersion. (ii) High tolerance to aldehydes is expected, because they are widely present in foods, fragrances, and metabolites and can be efficiently detoxified by the enzyme aldehyde dehydrogenase (33). In these studies, we selected EB as an agent for use in i.p. administration. EB has remarkable efficacy against PTX-resistant uterine serous and ovarian cancers (34–36), but prior clinical testing has revealed substantial toxicity when EB is administered as a free drug (26). We hypothesized that encapsulation in nanoparticles would reduce the toxicity of EB and make it
Fig. 6. Therapeutic efficacy of EB/BNPs on mice bearing i.p. USC tumors. (A) Kaplan–Meier survival curves for EB treatments at 2.5 mg/kg. (B) As a measure of toxicity, mice treated as in A were weighed twice a week. (C ) Kaplan–Meier survival curves for EB treatments at 0.5 mg/kg. (D) Mice treated as in C were weighed twice a week. In A and C, the x axis indicates the number of days from the first dose of drug. In B and D, the average weight is displayed as a function of days from the first drug dose. Statistical analysis was performed as described in SI Methods.
PNAS | October 11, 2016 | vol. 113 | no. 41 | 11457
ENGINEERING
Fig. 5. Suppression of USC cell growth on EB/BNPs-treated surfaces. (A) Schematic showing experimental design. A glass slide was divided into five blocks, and each block was incubated individually with EB/BNPs, EB/ NNPs, free EB, BNPs, or PBS for 30 min. After extensive wash, the slide was incubated in RPMI medium with USC cells with density at 2.0 × 105/mL for 24 h. After wash, the USC cells on slides were stained with Hoechst (for nuclei; blue) and live/dead stain (green indicates live cells, and red indicates dead cells) and imaged with a fluorescence microscope. (B) Fluorescence images showing the USC cell growth on poly-(L-lysine)-coated slides pretreated with different EB formulations. (Scale bars: 1 mm.) (C) The surface density of the cells was quantified with ImageJ and normalized to the PBS control. Data are shown as means ± SD (n = 3). *P < 0.05.
suitable for i.p. therapy. We showed the effectiveness of these EB formulations by measuring the survival of mice with PTX-resistant USC xenografts. To further limit EB toxicity in animals, we used a fivefold lower dose of EB compared with previous studies in nude mice with human multiple myeloma xenografts (37). Even at this lower dose of EB, however, significant toxicity (i.e., weight loss, diarrhea, and bowel dilation) was observed in mice treated with free EB. Remarkably, very little toxicity was observed in the mice treated with EB/BNPs (Fig. 6B). We attribute this difference in toxicity to two factors associated with EB/BNPs: the gradual release of EB from the BNPs (Fig. 4C), which likely reduces peak drug levels, and the strong adhesion of BNPs to the peritoneal tissue (Fig. 3C), which likely reduces EB doses to nontarget tissues and vital organs by fast lymphatic drainage, although some local toxicity to adjacent normal tissue is likely unavoidable (although not observed in our studies). We showed that EB was quickly cleared from the abdominal cavity, likely by diffusion through the peritoneal membrane into capillaries and transport through the hepatic portal system (38), which would be expected to lead to side effects and significantly lower bioavailability (39). The main side effect caused by EB in our study was gastrointestinal toxicity, which is in agreement with the known toxicity profile of EB (26, 40, 41). Importantly, in the EB/BNPs-treated mice, we showed significantly improved survival compared with in the controls (P < 0.05), with 60% of the animals alive at the end of the experiment (110 d). We attribute this enhanced effectiveness to EB/BNP bioadhesion and slow release. As evidence for the importance of bioadhesion, EB/NNPs were not as effective as EB/BNPs. Recent studies have shown strong activity of small molecules that selectively target the Her2/PI3K/Akt/mTor pathway against USCs. However, treatments with a single agent were only transiently effective, and tumors rapidly acquired resistance in vivo (36, 42). Importantly, recent preclinical data suggest that dual targeting of HER2/PIK3CA with a Pan-HER inhibitor (Neratinib) and a PIK3CA inhibitor (Taselisib) may be synergistic and able to achieve durable regression of USC xenografts in mice (43). However, this strategy is only applicable for the treatment of patients whose USC tumors harbor the amplification of the c-erbB2 gene and the concomitant amplification or oncogenic mutations in the PI3KCA gene. Therefore, alternate strategies are likely needed. In this study, we propose the use of nanoparticles loaded with EB for the treatment of USC after local/regional (i.p.) administration. This strategy takes advantage of the chemical properties of the nanoparticles and the
extreme sensitivity of USCs to the EB and may, therefore, be used for the treatment of the vast majority of USC patients, regardless of the genetic landscape of their tumors (44). In summary, our results represent a demonstration that i.p. administration of BNPs encapsulating EB can enhance survival of mice with i.p. tumors formed from USC xenografts. Importantly, the unique formulation of EB/BNPs exhibits higher therapeutic efficacy and lower toxicity than free EB and EB/NNPs. Because the bioadhesive property of BNPs is based on the general interaction between aldehydes on nanoparticles and proteins on tissue—which is illustrated in the variety of systems that we studied here (Figs. 2 and 3A)—we expect this approach to be applicable to other strategies for local delivery of active agents.
emulsion–evaporation technique and characterized using TEM and DLS. BNPs were formed using a previously described method (20) of using NaIO4 (Fig. 1) to convert vicinal diols to aldehyde groups and reversing back to NNP-Rs by adding NaBH4 (Fig. S1A). Retention of fluorescent BNPs was tested in vitro using a lysine-coated plate and wells containing USC cells by washing wells after particle incubation and reading the measurements on a fluorescent plate reader. Ex vivo retention was evaluated using a human umbilical cord and visualized with a fluorescent microscope. In vivo retention, distribution, and efficacy of dye- and drugloaded particles were tested in C57BL/6 and nude mice with USC cancers. Characterization was completed with the use of an in vivo imaging system (Xenogen), and efficacy was measured with strict survival guidelines. All animals were handled in accordance with the policies and guidelines of the Yale Institutional Animal Care and Use Committee. Detailed experimental procedures are provided in SI Methods.
Methods We used previously characterized BNPs to overcome barriers present in i.p. delivery of therapeutics. PLA-HPG was synthesized as previously reported (21). NNPs loaded with DiD, IR-780, and EB were formed using an
ACKNOWLEDGMENTS. We thank Yukun Pan, Tian Xu, Yu Wu, Yao Lu, and Rong Fan of Yale University for access to instruments in their laboratories. This work was supported by NIH Grants CA149128 and CA154460.
1. Greggi S, et al. (2001) Topotecan and gemcitabine in platinum/paclitaxel-resistant ovarian cancer. Oncology 60(1):19–23. 2. Tewari D, et al. (2015) Long-term survival advantage and prognostic factors associated with intraperitoneal chemotherapy treatment in advanced ovarian cancer: A gynecologic oncology group study. J Clin Oncol 33(13):1460–1466. 3. Bartlett DL (2008) HIPEC: The complexities of clinical trials. Ann Surg Oncol 15(5): 1277–1279. 4. Tsai M, et al. (2007) Effects of carrier on disposition and antitumor activity of intraperitoneal Paclitaxel. Pharm Res 24(9):1691–1701. 5. Bajaj G, Kim MR, Mohammed SI, Yeo Y (2012) Hyaluronic acid-based hydrogel for regional delivery of paclitaxel to intraperitoneal tumors. J Control Release 158(3): 386–392. 6. Emoto S, et al. (2014) Intraperitoneal administration of cisplatin via an in situ crosslinkable hyaluronic acid-based hydrogel for peritoneal dissemination of gastric cancer. Surg Today 44(5):919–926. 7. Wang Y, et al. (2010) 5-FU-hydrogel inhibits colorectal peritoneal carcinomatosis and tumor growth in mice. BMC Cancer 10:402. 8. Wu Q, et al. (2015) Thermosensitive hydrogel containing dexamethasone micelles for preventing postsurgical adhesion in a repeated-injury model. Sci Rep 5:13553. 9. Armstrong DK, Fleming GF, Markman M, Bailey HH (2006) A phase I trial of intraperitoneal sustained-release paclitaxel microspheres (Paclimer) in recurrent ovarian cancer: A Gynecologic Oncology Group study. Gynecol Oncol 103(2):391–396. 10. Kohane DS, et al. (2006) Biodegradable polymeric microspheres and nanospheres for drug delivery in the peritoneum. J Biomed Mater Res A 77(2):351–361. 11. Lu Z, et al. (2008) Tumor-penetrating microparticles for intraperitoneal therapy of ovarian cancer. J Pharmacol Exp Ther 327(3):673–682. 12. Zhou J, et al. (2013) Highly penetrative, drug-loaded nanocarriers improve treatment of glioblastoma. Proc Natl Acad Sci USA 110(29):11751–11756. 13. Ediriwickrema A, Zhou J, Deng Y, Saltzman WM (2014) Multi-layered nanoparticles for combination gene and drug delivery to tumors. Biomaterials 35(34):9343–9354. 14. Ediriwickrema A, Saltzman WM (2015) Nanotherapy for cancer: Targeting and multifunctionality in the future of cancer therapies. ACS Biomater Sci Eng 1(2):64–78. 15. Saucier-Sawyer JK, et al. (2015) Systemic delivery of blood-brain barrier-targeted polymeric nanoparticles enhances delivery to brain tissue. J Drug Target 23(7-8): 736–749. 16. Gupta A, Bahal R, Gupta M, Glazer PM, Saltzman WM (January 8, 2016) Nanotechnology for delivery of peptide nucleic acids (PNAs). J Control Release, 10.1016/j. jconrel.2016.01.005. 17. Cheng CJ, Tietjen GT, Saucier-Sawyer JK, Saltzman WM (2015) A holistic approach to targeting disease with polymeric nanoparticles. Nat Rev Drug Discov 14(4): 239–247. 18. McNeer NA, et al. (2015) Nanoparticles that deliver triplex-forming peptide nucleic acid molecules correct F508del CFTR in airway epithelium. Nat Commun 6:6952. 19. Bajaj G, Yeo Y (2010) Drug delivery systems for intraperitoneal therapy. Pharm Res 27(5):735–738. 20. Deng Y, et al. (2015) A sunblock based on bioadhesive nanoparticles. Nat Mater 14(12):1278–1285. 21. Deng Y, et al. (2014) The effect of hyperbranched polyglycerol coatings on drug delivery using degradable polymer nanoparticles. Biomaterials 35(24):6595–6602. 22. Mozzetti S, et al. (2008) Molecular mechanisms of patupilone resistance. Cancer Res 68(24):10197–10204. 23. Nobili S, Landini I, Mazzei T, Mini E (2012) Overcoming tumor multidrug resistance using drugs able to evade P-glycoprotein or to exploit its expression. Med Res Rev 32(6):1220–1262.
24. Paik D, et al. (2010) Higher sensitivity to patupilone versus paclitaxel chemotherapy in primary uterine serous papillary carcinoma cell lines with high versus low HER-2/neu expression in vitro. Gynecol Oncol 119(1):140–145. 25. Roque DM, et al. (2013) Tubulin-β-III overexpression by uterine serous carcinomas is a marker for poor overall survival after platinum/taxane chemotherapy and sensitivity to epothilones. Cancer 119(14):2582–2592. 26. Peereboom DM, et al. (2014) Phase II trial of patupilone in patients with brain metastases from breast cancer. Neuro-oncol 16(4):579–583. 27. de Jonge M, Verweij J (2005) The epothilone dilemma. J Clin Oncol 23(36):9048–9050. 28. Cocco E, et al. (2015) Clostridium perfringens enterotoxin C-terminal domain labeled to fluorescent dyes for in vivo visualization of micrometastatic chemotherapyresistant ovarian cancer. Int J Cancer 137(11):2618–2629. 29. Hermanson GT (2013) Bioconjugate Techniques (Elsevier/AP, London), 3rd Ed. 30. Steinbacher JL, et al. (2010) Gd-labeled microparticles in MRI: In vivo imaging of microparticles after intraperitoneal injection. Small 6(23):2678–2682. 31. Yang M, et al. (2014) Intraperitoneal delivery of paclitaxel by poly(ether-anhydride) microspheres effectively suppresses tumor growth in a murine metastatic ovarian cancer model. Drug Deliv Transl Res 4(2):203–209. 32. O’Brien PJ, Siraki AG, Shangari N (2005) Aldehyde sources, metabolism, molecular toxicity mechanisms, and possible effects on human health. Crit Rev Toxicol 35(7):609–662. 33. Vasiliou V, Thompson DC, Smith C, Fujita M, Chen Y (2013) Aldehyde dehydrogenases: From eye crystallins to metabolic disease and cancer stem cells. Chem Biol Interact 202(1-3):2–10. 34. El-Sahwi KS, Schwartz PE, Santin AD (2012) Development of targeted therapy in uterine serous carcinoma, a biologically aggressive variant of endometrial cancer. Expert Rev Anticancer Ther 12(1):41–49. 35. Black JD, English DP, Roque DM, Santin AD (2014) Targeted therapy in uterine serous carcinoma: An aggressive variant of endometrial cancer. Womens Health (Lond) 10(1):45–57. 36. Roque DM, et al. (2013) Class III β-tubulin overexpression in ovarian clear cell and serous carcinoma as a maker for poor overall survival after platinum/taxane chemotherapy and sensitivity to patupilone. Am J Obstet Gynecol 209(1):62.e1–62.e9. 37. Lin B, et al. (2005) Patupilone (epothilone B) inhibits growth and survival of multiple myeloma cells in vitro and in vivo. Blood 105(1):350–357. 38. De Bree E, Zoetmulder FAN, Romanos J, Witkamp AJ, Tsiftsis DD (2003) Intraperitoneal chemotherapy for prevention and treatment of peritoneal carcinomatosis from colorectal origin. Ann Gastroenterol 16(1):20–33. 39. Goodin S, Kane MP, Rubin EH (2004) Epothilones: Mechanism of action and biologic activity. J Clin Oncol 22(10):2015–2025. 40. Colombo N, et al. (2012) Randomized, open-label, phase III study comparing patupilone (EPO906) with pegylated liposomal doxorubicin in platinum-refractory or -resistant patients with recurrent epithelial ovarian, primary fallopian tube, or primary peritoneal cancer. J Clin Oncol 30(31):3841–3847. 41. Forster M, et al. (2007) A phase Ib and pharmacokinetic trial of patupilone combined with carboplatin in patients with advanced cancer. Clin Cancer Res 13(14):4178–4184. 42. Lopez S, et al. (2014) Taselisib, a selective inhibitor of PIK3CA, is highly effective on PIK3CA-mutated and HER2/neu amplified uterine serous carcinoma in vitro and in vivo. Gynecol Oncol 135(2):312–317. 43. Lopez S, et al. (2015) Dual HER2/PIK3CA targeting overcomes single-agent acquired resistance in HER2-amplified uterine serous carcinoma cell lines in vitro and in vivo. Mol Cancer Ther 14(11):2519–2526. 44. Cross SN, et al. (2010) Differential sensitivity to platinum-based chemotherapy in primary uterine serous papillary carcinoma cell lines with high vs low HER-2/neu expression in vitro. Am J Obstet Gynecol 203(2):162.e1–162.e8.
11458 | www.pnas.org/cgi/doi/10.1073/pnas.1523141113
Deng et al.
