Abstract
Pancreatic cancer has historically proven resistant to anticancer agents. On the one hand, drugs might be more efficient if higher levels could be achieved at the tumor site rather than the normal tissues. On the other hand, the thick stroma and the relative absence of abundant vessels may account at least partially for the failure of successive clinical trials to demonstrate effective treatments in this type of malignancy. In this context, the development and testing in clinical trials of treatment strategies that aim to optimize drug delivery is an important target in improving the prognosis of patients with pancreatic cancer.
Introduction
Pancreatic adenocarcinoma remains one of the most difficult to handle malignancies. Most patients are diagnosed with inoperable disease and their treatment options are limited [Hidalgo, 2010]. Few clinical trials have proven that certain drugs can improve prognosis in this setting: patients treated with gemcitabine survive for 6 months on average whereas erlotinib adds minimal benefit when combined with gemcitabine [Berlin et al. 2002; Senderowicz et al. 2007]. Recently, the combination of 5-fluorouracil (5-FU), leucovorin, irinotecan and oxaliplatin (FOLFIRINOX) proved superior to standard regimens at the cost of substantial toxicity [Conroy et al. 2011]. However, the majority of studies have failed to provide an efficient treatment paradigm.
Systemic regimens use tumor vessels to reach tumor cells and become effective. In this regard, besides their numerous mechanisms to overcome treatment effects, tumors grow to be viable in hypoxic conditions and less dependent on their vascular beds as part of their evolution process [Brahimi-Horn et al. 2007]. At the absence of abundant vessels, administered drugs cannot access their targets easily. Pancreatic adenocarcinoma cells in particular often appear to grow within a fibrotic, abundant, and poorly perfused stroma. Thus, compounds are less likely to accumulate close to such cells, independently of their targeting specificity or their antitumor potential [Neesse et al. 2011].
Conventional chemotherapy halts cell division or even demonstrates toxic activity preferentially on cells that proliferate fast. On the basis that tumor cells divide more rigorously than cells in normal tissues, such regimens have a tumor-specific effect to some extent. Furthermore, novel treatment modalities are designed to target nearly unique features expressed mainly in specific tumors. Nevertheless, both conventional chemotherapy and targeted therapies cause a series of dose-limiting adverse events because of activity on normal tissues. Optimization of drug specificity can therefore render larger accumulation in the tumor area and achieve a more preferable effect to toxicity ratio.
Taking into account the need for effective treatment strategies that spare the normal tissue in patients with pancreatic cancer, tumor-specific delivery of compounds becomes relevant. Advances in nanotechnology have provided potential delivery systems that might add clinical benefit to existing standard treatments. However, manipulation of the tumor microenvironment holds promise for effective treatments to overcome the stromal barrier. This review aims to illustrate the existing evidence and the rationale for incorporating drug delivery optimization systems in the arms of clinical trials in pancreatic cancer.
Rationale for using nanoparticles in pancreatic cancer
Hydrophobicity can limit the effect a drug might manifest against a tumor since efficient antitumor activity requires delivery through aqueous compartments. Addition of various excipients is currently used to provide hydrophilic properties. However, as in the case of paclitaxel, excipients are linked to severe hypersensitivity reactions. An alternative route to improve the efficiency of the delivery systems is the use of nanoparticles.
Nanoparticles are materials less than 100 nm in size. They share size-specific physicochemical properties which are distinct from larger materials. In medicine, both inorganic and polymeric nanoparticles have been used for diagnostic and treatment purposes [Youns et al. 2011]. Modification of their composition provides a broad range of different options, including simple passive vehicles, active compounds and molecule-specific targeting entities. They are biocompatible and allow the optimization of half life, biodistribution and water solubility of anticancer agents. Most of the nanoparticles are ‘PEGylated’ in that they are modified to include poly-ethylene-glycol (PEG) in their surface. PEGylation has proven the best among other nanoparticle modifications in adding stability along with mediating escape from the reticuloendothelial system [Prencipe et al. 2009].
Tumor vessels have bigger fenestration pores than normal vasculature [Yuan et al. 1995]. Furthermore, they are wider and they lack normal smooth muscle layers. Thus, they respond to various vasoconstriction stimuli with a hypertension-mediated passive dilatation instead of vasoconstriction [Nagamitsu et al. 2009]. In addition, tumors’ lymphatic drainage network is inadequate to clear macromolecules that enter in the tumor area [Ji, 2005]. The unique features of tumor angiogenesis and lymphangiogenesis contribute to accumulation of systemically administered macromolecular compounds, a phenomenon better known as enhanced permeability and retention effect (EPR; Figure 1) [Maeda, 2001]. When the size of such a drug exceeds a certain threshold, the drug gathers in the tumor area through the abnormally enlarged tumor vessel openings. However, they spare normal tissues since they cannot go through the narrow openings of normal vessels. Scant and defective lymphatics cannot transfer the drug from its initial deposition in the tumor to a different compartment. Besides, increased size prevents rigorous renal clearance and results in prolonged circulation. Taken together these features lead to accumulation of macromolecular compounds inside tumors. The size threshold for the EPR effect to occur has been defined as 40 kDa [Seki et al. 2009]. The entrapment of an active drug in the tumor can be an attractive treatment strategy as it allows higher and potentially more effective dosage schemes while sparing the normal tissues.

Illustration of the enhanced permeability and retention effect. Nanoparticles can easily go through the pores of the leaky tumor vasculature, while they do not have access to healthy tissues with normal vessels (top panel). In contrast, conventional drug delivery results in accumulation of the drug in both the tumor and the healthy tissues, giving rise to systemic adverse events (bottom panel).
Nanoparticles can optimize delivery of conventional cytotoxics
Figures 2(a) and (b) show a liposomic nanoparticle basic structure and a copolymer-based nanoparticle.

(a) Liposomic nanoparticle basic structure. A lipid layer forms a micelle which exerts hydrophilic properties and its size is within the nanotechnology spectrum. Antibodies or other peptides can boost the particle’s specificity to tumor epitopes like epidermal growth factor receptor. (b) Copolymer-based nanoparticle. In this system the drug, shown here in red, is covalently bound to a polymer. The resulting nanoparticle is water soluble and has the appropriate size to accumulate in the tumor area. Tumor epitope specific peptides can be added to this basic scheme.
Paclitaxel is a hydrophobic compound and is traditionally administered with castor oil so that it becomes solvent in aqueous media [Rischin et al. 1996]. A different approach is based on linking the active compound to albumin and forming albumin-based, hydrophilic nanoparticles [Di Costanzo et al. 2009]. Pretreatment is not required to avoid hypersensitivity reactions with nab-paclitaxel. It has been tried in a phase I trial in variable tumors, including pancreatic cancer, and it has demonstrated an acceptable toxicity profile [Chien et al. 2009]. Data for paclitaxel efficacy in pancreatic cancer are limited; however, nab-paclitaxel might be superior to paclitaxel due to better pharmacokinetics and more favorable accumulation in the tumor. Data from a phase I trial [Drengler et al. 2008] of nab-paclitaxel plus gemcitabine and a phase II trial [Hosein et al. 2010] of nab-paclitaxel after gemcitabine failure in advanced pancreatic cancer are promising. In particular, in the phase II trial patients although refractory to gemcitabine had an overall median survival of 7.3 months and progressed after a median time of 1.3 months. In addition, a great number of phase I and II clinical trials are testing nab-paclitaxel in various combinations and settings of pancreatic cancer. A recently published phase I/II study tested the safety and efficacy of nab-paclitaxel in combination with gemcitabine in the first-line setting of advanced pancreatic cancer [Von Hoff et al. 2011]. Neutropenia was the dose-limiting toxicity and median survival was 12.2 months. Interestingly, expression of secreted protein acidic and rich in cysteine in the stroma was associated with higher response rate. In the same study the combination of gemcitabine with nab-paclitaxel achieved stromal depletion and higher gemcitabine concentration in the tumor area compared with gemcitabine alone in xenograft models in mice.
Curcumin has demonstrated anticancer activity in a panel of different tumors in vitro [Altenburg et al. 2011; Lu et al. 2011; Zhang et al. 2010]. Curcumin has been shown to increase apoptosis and block nuclear factor κB in pancreatic cancer cells [Jutooru et al. 2010]. Nonetheless, there is a gap between the in vitro promising data and the efficiency of the compound in vivo due to its hydrophobicity and subsequent impractical pharmacokinetic profile [Garcea et al. 2004]. Recently, a polymeric nanoparticle containing curcumin has been developed [Bisht et al. 2007]. This has been shown to be a promising ‘nanocurcumin’ that demonstrates similar antitumor activity against pancreatic cancer cells and subcutaneous and orthotopic xenografts. In addition, its effect is additive to the effect of gemcitabine. Furthermore, another study has shown that curcumin forms nanoparticles in the presence of rubusoside that are active against breast colon and pancreatic cancer cell lines [Zhang et al. 2001]. The effect of nanocurcumin has not been evaluated in clinical trials as yet but might be a reasonable option given the anticancer activity of curcumin combined with the favorable nanotechnology pharmacokinetics.
CPT-11 has demonstrated significant efficacy against pancreatic cancer as part of the FOLFIRINOX scheme. SN38 is the active metabolite of CPT11 responsible for the antitumor activity and the adverse events. Saito and colleagues have tested the efficacy of NK012, a polymeric micelle that contains SN38, in the setting of xenograft models generated by hypervascular and hypovascular pancreatic tumors [Saito et al. 2008]. Interestingly, they reported that NK012 achieved a better and more durable release of SN38 in the tumor area compared with CPT-11 in both xenograft models. In their experiments, NK012 but not CPT-11 resulted in eradication of all the tumors treated. Based on this study, nanoparticle-based SN38 might lead to specific accumulation of the drug in the tumor area even in relatively avascular tumors surrounded by a thick stroma. Another possible modality of more effective irinotecan delivery is the liposomic preparation of CPT-11 or PEP02. In the interim analysis of a phase II study PEP02 monotherapy in metastatic gemcitabine refractory pancreatic cancer achieved 3-month survival of 74%. The most common grade 3–4 side effects were fatigue, neutropenia, nausea, vomiting and diarrhea [Ko et al. 2011].
Nanoparticle-based delivery of small interfering RNAs
Gene therapy has been a promising field in the therapeutics of human tumors. Small interfering RNA (siRNA) refers to small RNA molecules that can effectively target and suppress specific messenger RNAs (mRNAs). This technology can be employed to modify gene expression at the mRNA level once it is effectively delivered in the target cells and avoids endosomal entrapment and degradation. Nanotechnology provides a range of candidate carriers in this regard. Calcium precipitates have been used in the past for nucleic acid delivery inside the cells [Fasbender et al. 1998]. Pittella and colleagues have created a complex of PEG, charge conversional polymer (CCP), and calcium precipitates as stable nanoparticles to bind polyanions and siRNAs [Pittella et al. 2011]. PEG provides colloidal stability whereas CCP inhibits endosomal entrapment. The investigators proved that this system can effectively deliver siRNA against vascular endothelial growth factor in pancreatic cancer cells in vitro. Although primitive, this work has created an option for siRNA delivery which needs to be validated in higher organized systems.
Other groups have employed nanotechnology to effectively deliver siRNAs in pancreatic cancer models. In one of these studies, mesoporous nanoparticles (MSNPs) were synthesized from tetraethyl orthosilicate and a micellar template. Their large surface allows delivery of drugs in larger amounts [Xia et al. 2009]. The addition of polyethyleneimine polymer made the surface of MSNPs cationic and favorable for attaching nucleic acids like siRNAs. A second group successfully used poly lactic- co -glycolic acid (PLGA) poloxamer nanoparticles for delivering siRNA against methyl-CpG binding domain protein 1 in pancreatic cancer cells [Luo et al. 2009]. Cell proliferation was then inhibited and apoptosis was activated. Finally, in a third study antihuman epidermal growth factor receptor 2 siRNA was effectively delivered and inhibited tumor growth with an immunoliposome-based nanoparticle in pancreatic cancer models in vitro and in vivo [Hogrefe et al. 2006]. This system specifically targeted tumor cells with the addition of a single chain transferrin specific antibody. Again, these interesting approaches require further experimentation before any conclusions regarding their efficacy in pancreatic cancer can be drawn.
Nanoparticles with attached epitope-specific targeting molecules
Pancreatic cancer specificity of nanotechnology-based treatment platforms can be optimized by the attachment of antibodies to nanoparticle surface, able to recognize and bind epitopes uniquely expressed in pancreatic adenocarcinoma. In this context, epidermal growth factor receptor (EGFR) is often present and abundant in pancreatic cancer cells [Dancer et al. 2007]. Cetuximab is an EGFR-specific antibody which inhibits the receptor dimerization and activation. Cetuximab can be conjugated to nanoparticles carrying chemotherapy drugs and then exert receptor internalization upon binding to EGFR in a distinct pharmacodynamic fashion compared with cetuximab alone [Bhattacharyya et al. 2010]. In a study that illustrates this rationale, gemcitabine was effectively delivered with cetuximab-coated gold nanoparticles to pancreatic cancer cell lines in vitro and to pancreatic cancer xenografts in vivo [Patra et al. 2008]. Uptake of the nanoconjugates is then proportional to EGFR presence on the cancer cells. A different approach that combines radiofrequency thermal effects with cetuximab-coated gold nanoparticles has been described [Glazer et al. 2010]. In this strategy, gold nanoparticles accumulate in the tumor cells that express EGFR and mediate the radiofrequency thermal effects in vitro and in vivo. Cell viability was reduced only in cell lines that expressed EGFR rather than cocultured adjacent cell lines that did not express the receptor. In addition, this approach was able to reduce tumor growth in xenograft models using cetuximab or an antibody against mucin 1 (MUC1) as a targeting molecule.
Besides EGFR and MUC1, other pancreatic-cancer-specific epitopes have been targeted with antibodies and other molecules conjugated with nanoparticles in preclinical models: a half antibody that recognizes carcino-embryonic antigen (CEA) as part of a lipid-polymeric hybrid nanoparticle has been described and shown to specifically target CEA-overexpressing pancreatic cancer cells [Hu et al. 2010]. Pentagastrin and decagastrin have been conjugated to pegylated calcium phosphosilicate nanoparticles and target orthotopic mouse pancreatic cancer models that express gastrin receptors [Barth et al. 2010]. In another study, a nanoparticle containing the chemotherapeutic agent doxorubicin was targeted against integrin avβ3, which is preferentially expressed in some areas of the tumor vasculature [Murphy et al. 2008]. In these experiments, a 15-fold increase in apoptosis was reported in the areas of the tumor where the targeted integrin was expressed in the adjacent vessels. Last but not least, Luo and colleagues described LyP-1, which is a nine-amino-acid peptide that homes to lymphatic vessels of the tumors and lymph node metastases, conjugated nanoparticles and their ability to accumulate in the lymph node metastases of pancreatic cancer mouse models [Luo et al. 2010].
A slightly different and interesting approach has been proposed by Weissleder and colleagues [Weissleder et al. 2005]. Specifically, this group has created a library of over a hundred fluorescent magnetic nanoparticles conjugated with small molecules like amines and alcohols. Then they screened pancreatic cancer cells against this library and isolated the small molecules that would preferentially target these cells but not macrophages or endothelial cells. In this paradigm, the specificity of the nanoparticles is based on nonspecific small molecules rather than antibodies or other entities specific to tumor antigens.
Stromal targeting therapies
Since pancreatic cancer cells are surrounded by a thick, poorly perfused stroma which halts penetration of otherwise effective treatments, it is reasonable to incorporate a stroma depleting agent into the therapeutic plan of patients with pancreatic adenocarcinoma. The basic idea that reducing the size of the stroma will expose the tumor to chemotherapy is further supported by studies that illustrate the active role of the tumor microenvironment in the biology of pancreatic cancer. In this regard it has been shown that hedgehog signaling from the pancreatic cancer cells activates the hedgehog pathway in the adjacent stroma rather than the cancer cells themselves [Yauch et al. 2008]. The hedgehog pathway has been linked to the biology of the cancer stem cells and mediates important aspects of the molecular conversation between the pancreatic cancer cells and the stromal cells. While this interaction is only partially understood and research on this topic is ongoing, it has been shown that inhibition of the hedgehog pathway in pancreatic cancer mouse models enhances the formation of blood vessels in poorly vascularized tumors and decreases their stromal component [Olive et al. 2009]. Taken together these effects result in better distribution of active anticancer compounds in the tumor area which would not have otherwise access to the tumor cells. These experiments support the notion of combining hedgehog inhibitors with other anticancer drugs. Indeed, a number of clinical trials based on this rationale are ongoing in pancreatic cancer. For example, IPI-926 an inhibitor of Smoothened, which is a molecule in the hedgehog pathway, was recently reported to show some activity and acceptable toxicity in a phase 1b trial [Richards et al. 2011] in pancreatic cancer whereas the hedgehog inhibitor GDC-0449 was safe in a phase I trial [Palmer et al. 2011]. More mature data from these trials and larger trials will clarify further the role of this strategy in the treatment of pancreatic cancer.
Likewise, with hedgehog signaling, other stromal-depleting strategies have been reported in mouse models. Specifically, iRGD is a tumor-penetrating peptide which has been conjugated to several anticancer drugs in the past. Sugahara and colleagues have shown that this peptide can increase the permeability and the vascularity of pancreatic cancer models when administered without prior conjugation to any other agent [Sugahara et al. 2010]. This is important since conjugation may alter the physicochemical properties of a drug and therefore its efficacy. In their study, Sugahara and colleagues proved that iRGD optimized the penetration of several drugs in the pancreatic cancer area, including small molecules, nanocarriers, and monoclonic antibodies. In another study, Kano and colleagues examined the effect of inhibiting transforming growth factor β (TGFβ) signaling with a low-dose TGFβ type 1 receptor inhibitor in pancreatic cancer mouse models [Kano et al. 2007]. With this strategy they achieved reduction in the pericyte coverage of tumor vessels and therefore the enhancement of the accumulation of macromolecular nanomedicines in the area of the tumor. However, this strategy did not have any effect on the abundance of the fibrous elements of the pancreatic cancer models. More studies are needed to provide insight into the possible effect of such approaches in patients with pancreatic cancer. Finally, CD40 agonists activate the CD40 superfamily of tumor necrosis factor receptors and are associated with stroma shrinkage in pancreatic adenocarcinoma models in genetically engineered mice [Beatty et al. 2011]. This effect was mediated by activation of the tumor-associated macrophages. Interestingly, in a primitive trial, CD40 agonist in combination with gemcitabine achieved a partial response to 4 out of 21 evaluable patients, further supporting the effectiveness of tumor immunity surveillance in the treatment of such tumors [Beatty et al. 2011].
Concluding remarks
There is a growing body of literature dealing with the optimization of drug delivery in pancreatic cancer. Table 1 summarizes relevant strategies currently in clinical trials in several pancreatic cancer settings. Given the special characteristics of this tumor type, it becomes plausible that the failure of clinical trials to show benefit from drug interventions might be a matter of poor drug accumulation at the sites of interest rather than endogenous antitumor drug activity. Incorporation of research biopsies and ‘mapping’ of the distribution of such anticancer drugs with novel imaging modalities is needed to establish the efficiency of drug delivery in relevant studies. We conclude that clinical trials that employ drug delivery optimization strategies like nanotechnology-based systems or stroma-depleting regimens are reasonable and promising.
Strategies for optimizing drug delivery in pancreatic cancer in clinical trials.
siRNA, small interfering RNA.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement
The authors declare no conflict of interest in preparing this article.
