Keywords

1 Introduction

Cancer is a complex collection of diseases that through deregulation of myriad cellular pathways achieves unchecked growth, metabolism, and migration, thus making it the second leading cause of death in the United States. Solid tumors in particular have complex cyto-architecture and diverse microenvironments, making single-agent therapy largely ineffective due to sub-populations of cells that are resistant to a given agent (Burrell et al. 2013; Hainaut and Plymoth 2013; Hanahan and Weinberg 2000, 2011). Efforts have been taken to better understand the genetic alterations and complex molecular mechanisms in cancer to identify more effective therapies.

Due to the past failure of mono-therapies, most cancer patients now receive some form of combination therapy as the standard of care treatment for their tumor and this approach has improved patient outcomes. However, for the most part, current regimens were not “designed”, rather, they were empirically determined to be combinations of drugs with non-overlapping toxicities, so that each drug could be administered at near-maximal dosage. Since these drugs generally have different mechanisms of action, there tends to be minimal cross-resistance, decreasing the emergence of drug resistant tumors (Mayer and Janoff 2007; Harasym et al. 2007; Ramsey 2005; Zoli et al. 2001). Importantly, however, this does not mean that current combinations are the best cocktail of drugs to achieve lasting remissions in patients. Rather they are the best cocktail of drugs identified to date that avoid unmanageable toxicity. Recent advances in nano-scale drug carriers that can target tumors preferentially and limit systemic toxicity are now revolutionizing the way we approach combination therapy.

2 Drug Combinations for Cancer Treatment

2.1 A Brief History of Combination Cancer Therapy

2.1.1 Combinations of Independently Active Drugs

Following the discovery of cytotoxic chemotherapy by Goodman et al. (1946; Gilman and Philips 1946) in 1943 and Farber et al. (1948) in 1947, researchers and clinicians began focusing on strategies to prolong cancer remissions in patients and to delay drug resistance. Inspired by the observation of synergism between 6-mercaptopurine (6MP) and azaserine (O-diazoacetyl-l-serine), as well as 6MP and antifolates (Skipper et al. 1954) in preclinical mouse models of leukemia, Burchenal and colleagues conducted pilot studies (Burchenal et al. 1954) and later a randomized trial (Heyn et al. 1960) in children with acute leukemia receiving sequential combinations of 6MP, azaserine, and steroids (Fig. 16.1). While widespread improvements in overall survival were not observed, Frei et al. later demonstrated greatly enhanced remission rates in children with acute leukemia treated with 6MP and antifolates in 1961 – noting that “advantage derives from the two drugs acting independently on the patient” (Frei et al. 1961). With the discovery of antineoplastic activity from plant alkaloids from Vinca rosea at Eli Lilly, Freireich et al. (1964) quickly developed a “quadruple combination therapy” known as VAMP (vincristine, amethopterin, 6MP, and prednisone) which augmented remission rates in patients with acute leukemia treated in 1964. VAMP soon gave way to MOMP (Devita et al. 1965) (nitrogen mustard with vincristine, methotrexate, and prednisone) and then to MOPP (Devita et al. 1970) (nitrogen mustard with vincristine, procarbazine, and prednisone), the latter of which demonstrated startling activity in patients with Hodgkin’s lymphoma – nearly 80 % displayed complete remission, and of these, 60 % demonstrated disease free survival. Throughout the advent of MOPP, the rationale for combination chemotherapy relied heavily on two fundamental underlying concepts: independent drug action and non-overlapping toxicity. As stated by Vincent DeVita: “differing mechanisms of action and various dose-limiting toxicities could presumably overcome [low therapeutic index]” (Devita et al. 1970).

Fig. 16.1
figure 1

An early example of sequential combination cancer therapy c.1954. Treatment regimen and corresponding blood counts for a 4 year old male with 6-mercaptopurine-resistent acute leukemia receiving amethopterin (antifolate), 6-mercaptopurine (purine antagonist), cortisone (steroid), and azaserine/P-165 (purine antagonist) at the Sloan-Kettering Institute (Reproduced with permission from Burchenal et al. (1954). Copyright 1954 John Wiley and Sons)

Full size image

2.1.2 Rational Drug Combinations Targeting a Shared Mechanism of Action

In 1982, in vitro studies demonstrated that pretreatment of cells with leucovorin (folinic acid), an innocuous nontoxic agent, to cancer cells could enhance the biochemical effects of the nucleoside analog, 5-fluorouracil (5-FU). Researchers found that leucovorin augmented inhibition of thymidylate synthase, the canonical target of 5-FU, depleting cellular nucleotide levels and inducing apoptosis (Pritchard et al. 2013; Longley et al. 2003). Preclinical studies in tumor xenograft-bearing mice, and later pilot studies in patients (Machover et al. 1982) showed clear benefits in response rates from the combination therapy versus that seen after treatment with 5-FU alone, although the improvements in overall survival were more moderate. The combination, still applied today as part of FOLFOX, FOLFIRI, and FOLFIRINOX treatment regimens, marked a significant departure from the strategy originally established by Frei, Freireich, and Zubrod. These drug combinations came about from “trial and error”/empirical testing in patients and are used not because they are the most effective treatment for tumor killing, but because they are capable of achieving the maximal tumor reduction within the maximal toxicity tolerated by patients. With the advancement of targeted drug delivery, “rational drug combinations” can now be used. Rather than combining drugs with independent activity or dose-limiting toxic effects, rational drug combinations could act cooperatively on overlapping molecular targets to selectively to kill cancer cells.

2.1.3 Molecularly Targeted Therapies

The discovery that multidrug transporter proteins could play a key role in the development of adaptive resistance to chemotherapy (Rothenberg and Ling 1989), led to the initiation of clinical trials in the late 1980s to concurrently inhibit promiscuous drug efflux pumps during the administration of cytotoxic anti-tumor agents. Here, competitive substrates of the multidrug transporters – already approved to treat other conditions (e.g. cyclosporine and verapamil) – were co-administered in patients with refractory disease. Although these trials, and those that followed over multiple generations of inhibitors, were largely disappointing (Fletcher et al. 2010; Kaye 1993), the general strategy of identifying and preemptively blocking the biological mechanisms responsible for adaptive resistance continued to guide a significant portion of subsequent work.

During this same period in the late 1980s, researchers were also beginning to understand structure-activity relationships for a novel small molecule contraceptive agent that showed unusually selective antitumor activity in breast cancer patients – tamoxifen (Jordan 2003). The drug was found to selectively inhibit the estrogen receptor, its ‘target’ protein, in breast tumors. These findings led to a shift in the focus of commercial anti-cancer drug discovery away from enhanced non-specific cell killing towards rational target inhibition. So-called ‘molecularly targeted therapies’ against the BCR-ABL fusion protein (i.e. imatinib/Gleevec) (Capdeville et al. 2002), monoclonal antibody therapies (e.g. trastuzumab/Herceptin, and related molecules targeting EGFR family members) (Hudis 2007), and recombinant proteins (e.g. interleukin-2) (Dranoff 2004) followed soon afterward. Rational drug combinations incorporating these molecularly-targeted drugs added further complexity to prior combination therapeutic approaches.

2.1.4 Large-Scale Screens, Nucleic Acid Therapies, and Beyond

Later, with the discovery of RNA interference in mammalian cells in 2001 (Elbashir et al. 2001), large scale loss-of-function screens were used to identify a subset of new molecular targets for cancer therapy (Ngo et al. 2006; Luo et al. 2008), as well as new gene combinations whose pairwise loss blocked cancer cell survival, resulting in ‘synthetic lethality’ (Kaelin 2005; Luo et al. 2009). A number of these early discoveries were complicated by poor reproducibility, despite improvements in small interfering RNA (siRNA) and messenger RNA (mRNA) delivery (Kormann et al. 2011). Improvements in CRISPR-Cas9 technology have overcome many of these difficulties (Platt et al. 2014), further expanding this toolkit to include amplified or synthetic protein expression, in addition to genetic loss of function. Although it is currently unclear to what extent epigenetic modifiers (e.g. chromatin regulators) (Floyd et al. 2013; Keung et al. 2014), immune checkpoint antagonists (Mahoney et al. 2015), or chimeric proteins/receptors (Kalos et al. 2011; Morsut et al. 2016) will contribute to future multiplexed combination therapies, the number of possible pairwise combinations of the above drug classes alone provides ample opportunities for the creation of new and highly potent therapies with durable treatment responses.

2.2 Challenges in Delivering Drug Combinations to Tumors

2.2.1 Co-delivery

Implementation of drug combinations generally requires co-localization of each agent within the malignant cells for efficient tumor cell killing. This presents a variety of complex intrinsic challenges due to the unique physiochemical properties of each drug such as size, charge, hydrophobicity, and stability, among others. For example, current frontline two-agent therapy for ovarian cancer requires co-delivery of cisplatin and paclitaxel. Although both drugs are roughly neutral at physiological pH, paclitaxel is more than double the molecular weight of cisplatin and its relative hydrophobicity (octanol:water partition coefficient) is more than five logs higher than cisplatin. Cellular colocalization in the complex tumor microenvironment is thus severely constrained. Further challenges include colocalization of either drug with bevacizumab – a monoclonal antibody directed against the pro-angiogenic cytokine VEGF-A, which is 1000 times larger in molecular weight, as part of platinum-sensitive disease therapy, or liposomal doxorubicin, a topoisomerase inhibitor whose nanoparticle is almost 10,000 times larger than the free small molecule paclitaxel – as part of platinum-insensitive disease therapy. Interestingly, Wittrup and coworkers (Schmidt and Wittrup 2009) have modeled tumor uptake data for biomolecules of varying size and affinity and found that intermediate-sized targeting agents (ca. 25 kDa, 5.3 nm dia) exhibit the lowest tumor uptake, whereas higher tumor uptake levels are observed for either smaller agents (e.g. peptides and small molecules) and larger agents (e.g. IgG, liposomes). Bawendi, Jain, and Fukumura (Stylianopoulos et al. 2012) have similarly examined size-dependent penetration of nanoscale particles into the interstitium of tumor xenografts and found that drug size is inversely correlated with tumor penetration (Fig. 16.2a). Because intratumoral distribution profiles of drugs are often heterogeneous, enhanced or diminished cell killing can thus occur in a spatially dependent manner. Using inorganic colloids as model drug carriers, Chan and coworkers also found that 20 nm particles efficiently penetrate and are retained in the tumor interstitium at significantly greater depths than comparable 40–100 nm particles (Perrault et al. 2009).

Fig. 16.2
figure 2

Challenges in delivering drug combinations. (a) Drugs with differing targets often display disparate physiochemical properties that, here, affect intratumoral distribution, but also pharmacokinetics, pharmacodynamics, and tissue disposition. Model drug A is 10-fold smaller than B and penetrates more deeply into HT1080 tumor xenografts at 6 h as measured by fluorescence microscopy of hitological tissue sections. (b) In many canses, multiplexed combinations exhibit optimally synergistic drug stoichiometry that is difficult to recapitulate using free drug compounds. Here, small molecule inhibitors of Mek and PI3K are synergistic toxic only over a narrow range of drug ratios in vitro as measured by CellTiter Glo. (c) The sequence and timing with which drugs modulate (rewire) complex cell signaling networks also determines capacity for cell killing. Here, combinations of the cytotoxic chemotherapeutic, doxorubicin, and erlotinib, an inhibitor of EGFR, optimally induce apoptosis (double positive flow cytometry) in a sequence-dependent manner in vitro. (d) Dose-limiting toxic effects can also act synergistically, here increasing the relative risk of gastrointestinal bleeding 19 % beyond additivity when selective serotonin reuptake inhibitors (SSRIs) and nonsteroidal anti-inflammatory drugs (NSAIDSs) are combined in patients (n = 223,336) (Reproduced with permission from (a) Stylianopoulos et al. (2012), (b) Dreaden et al. (2015), (c) Lee et al. (2012), (d) Anglin et al. (2014). Copyright (a) 2012 Elsevier, (b) 2015 American Association of Cancer Research, (c) 2012 Cell Press, (d) 2014 Nature Publishing Group)

Full size image

2.2.2 Stoichiometry/Ratiometric Dosing

Historically, combinations of free drugs are administered at their respective maximum-tolerated-doses (MTDs); however, it is now widely acknowledged that drug combinations can act synergistically at specific drug ratios, as well as additively or even antagonistically at other drug stoichiometries. For example, Dreaden et al. found that MEK and PI3K inhibitors, when co-administered, exhibit stoichiometry-dependent drug synergy in vitro (Fig. 16.2b). Likewise, optimally synergistic pairwise combinations of irinotecan/floxuridine (Batist et al. 2009), cytarabine/daunorubicin (Tardi et al. 2009a), irinotecan/cisplatin (Tardi et al. 2009b), paclitaxel/tanespimycin (17-AAG) (Katragadda et al. 2013), and quercetin/vincristine (Wong et al. 2010) have also been identified.

2.2.3 Drug Sequence and Timing

Cellular responses to perturbations occur in a time-dependent manner and drug combinations that exploit these response networks can often maximize therapeutic potential through sequence- and time-staggered delivery. Although conventional delivery methods such as intravenous or intraperitoneal infusion can be staged manually, poor drug colocalization and unfavorable drug stoichiometry within tumors can abrogate the therapeutic potential of even the most potent rational combination therapies. Lee et al. (2012), for example, recently employed a systems-based approach to understand how targeted cancer therapies rewire oncogenic cell signaling networks, and methods by which this ‘dynamic rewiring’ can be exploited to improve tumor cell killing (Fig. 16.2c). Interesting, they found that time-staggered inhibition of epidermal growth factor receptor (EGFR), but not simultaneous co-administration, could sensitize a subset of triple negative breast cancer cells to DNA damaging chemotherapy. Not only was sequence a critical determinant for enhanced cell death, but a time delay of as little as 4 h could dramatically alter apoptotic response in vitro. In a related approach, Seino et al. (2016) recently found that JNK pathway may play a key role in adaptive resistance of ovarian cancer cells towards frontline platinum/taxane therapy. Although concurrent targeted inhibition of JNK induced contrasting effects on cisplatin (enhanced cytotoxicity) and paclitaxel (decreased cytotoxicity), time-staggered inhibition of JNK prior to chemotherapy greatly augmented in vitro cell killing in a time-dependent manner. These findings also suggest that basal JNK activity may correlate with drug resistance in ovarian cancer. Sengupta and coworkers (Goldman et al. 2015) have also examined drug-induced tumor cell perturbations and found that taxane therapy induces a transient cell state characterized by Src family kinase (SFK)/Hck pathway activation and suppression of apoptosis. By studying the in vivo induction kinetics of this transient state, the authors could preemptively block anti-apoptotic responses in vivo through targeted inhibition of SFK/Hck staggered approximately 6 days after taxane therapy.

2.2.4 Compounding and Overlapping Toxicity

As the focus of combination cancer therapies shifts towards synergy and network rewiring, compounded dose-limiting toxic effects again present a significant challenge to multi-drug treatment regimens. Drug interactions are often complex to predict, and dose-limiting toxic effects – like therapeutic effects – can occur synergistically. For example, combined administration of common therapeutics such as nonsteroidal anti-inflammatory drugs (NSAIDSs) with selective serotonin reuptake inhibitors (SSRIs) increases the relative risk of gastrointestinal bleeding 19 % beyond the additive expectation (Anglin et al. 2014) (Fig. 16.2d). Similarly, doxocubicin and trastuzumab (anti-HER2/ErbB2; Herceptin) – commonly co-administered in breast cancer patients – are known to induce Type I and Type II cardiotoxicity, respectively, effects which compound the risk of dose-limiting cardiotoxic events in patients (Cardinale et al. 2010). More recently, combined blockade of MAPK and PI3K pathway signaling has been shown to synergistically kill a variety of solid tumors and pre-emptively block resistance-associated signaling in preclinical animal models (Engelman et al. 2008). A retrospective Phase I clinical study of patients with advanced solid tumors receiving small molecule inhibitors of MAPK or PI3K pathway signaling, alone or in combination, found that simultaneous blockade significantly decreased average tumor burden relative to monotherapy (Shimizu et al. 2012); however, these favorable outcomes were accompanied by a 2.0-fold increase in the prevalence of dose-limiting toxicity and a 3.0-fold increase in the prevalence of drug-related high grade (>III) adverse events, primarily hepatic-related.

3 Nanoparticle Formulations to Optimize Anti-cancer Combination Therapies

Nanomaterials can be used to co-deliver multimodal cancer therapeutic agents to achieve maximum therapeutic effects (Hu and Zhang 2012). Some of the nanomaterials used to study combination therapy include liposomes, polymer-drug conjugates, dendrimers, and polymeric micelles (Peer et al. 2007). We will discuss a few notable examples in detail below.

3.1 Mesoporous Silica Nanoparticles

Mesoporous silica nanoparticles (MSNs) exhibit a range of features that are amenable to drug delivery (Baek et al. 2015). They have high surface area, which allows for large amounts of drug loading, tunable porosity and size, structural diversity, easily modifiable chemistry and suitability for functionalization, and are biocompatible. MSNs have been used extensively as multifunctional nanocarrier systems by combination or hybridization with biomolecules, drugs, and other nanoparticles, and can be stimulated by signals such as pH, optical signal, redox reaction, òr electric and magnetic fields (Baek et al. 2015).

Lipid-coated MSN have recently been used to delivery synergistic gemcitabine and paclitaxel to both subcutaneous and orthotopic pancreatic tumors in mice. Mice with both subcutaneous and orthotopic tumors receiving systemic gemcitabine and paclitaxel loaded MSN achieved more effective tumor shrinkage than those receiving individual or free drug. The authors also observed elimination of metastatic foci without evidence of local or systemic toxicity (Meng et al. 2015).

Another way to use MSNs is for the combination of photodynamic and chemotherapies. One such example is the use of MSNs loaded with aluminum chloride phthalocyanine (AlClPc) and cisplatin for cancer treatment. Vivero-Escoto and colleagues showed that these MSNs are taken up by HeLa cells, and upon light exposure, the AlClPc-cisplatin-MSN combination was more cytotoxic then the AlClPc-MSN and cisplatin-MSN controls. These data suggest that there is great potential for the use of MSN platforms as nanocarriers for combination photodynamic and chemotherapies to treat cancer (Vivero-Escoto and Elnagheeb 2015). In another example, Zhang and colleagues synthesized a polymeric prodrug (doxorubicin)-coated hollow mesoporous silica nanoparticles (HMSNs) with an NIR absorbing dye IR825 loaded into the hollow cavity of the HMSN to form a multifunctional hybrid HMSNs-DOX/IR825 (Zhang et al. 2016). Cancer cells efficiently took up the hybrid nanoparticle, and the conjugated doxorubcin was successfully released in the cellular environment. In vitro cytotoxicity study showed that anticancer activity of HMSNs-DOX/IR825 was significantly improved by the NIR irradiation, suggesting that the hybrid nanoparticle could potentially be used for combined photothermal-chemotherapy of cancer (Zhang et al. 2016).

3.2 Self-Assembly Copolymer Carriers – Micelles

Many self-assembling molecules are amphiphilic, comprising of both hydrophobic and hydrophilic domains. Amphiphilic copolymers can self-assemble into micelles, vesicles, and molecular gels composed of tubules, fibrils, and fibers (Giddi et al. 2007; Nishiyama and Kataoka 2006; Rösler et al. 2001).

Micelles are amphiphilic molecules that self-assemble into a spherical structure with a hydrophobic core and hydrophilic exterior making it suitable for encapsulating hydrophobic cancer drugs (Jhaveri and Torchilin 2014). It is estimated that about 40 % of marketed drugs and up to 75 % compounds under development are poorly water soluble (Jhaveri and Torchilin 2014; Di et al. 2009; Williams et al. 2013). Polymeric micelles can be used for combination therapy by loading multiple anticancer agents in polymeric micelles in a one-step drug-loading process without chemical modification of drugs (Shin et al. 2011). Multi-drug release may occur by simple hydrolysis or be triggered by an acidic pH and/or lysosomal enzymes, and can be tuned by chemical linkage for concurrent or sequential delivery (Duncan 2006; Greco and Vicent 2009).

Shin and colleagues showed that they could encapsulate paclitaxel, rapamycin, and 17-allylamino-17-demethoxygeldanamycin (17-AAG) in PEG-b-PLA micelles without changing the pharmacokinetics of each drug at low doses. The pharmacokinetic profiles were however altered when drugs are delivered at higher doses (Shin et al. 2012). Rapamycin and 17-AAG acts concurrently to inhibit the PI3K/Akt/mTOR and Ras/Raf/MEK/ERK signaling pathways, enhancing cancer cell killing by paclitaxel. Bae and colleagues directly conjugated doxorubicin and 17-hydroxyethylamino-17-demethoxygeldanamycin (GDM-OH) to a poly(ethylene glycol)-poly(aspartate hydrazide) block copolymers through acid-labile hydrazone bonds. The pH-sensitive micelles were combined and appeared to minimize a schedule-dependent change in combined drug efficacy when compared to the free drug combination (Bae et al. 2010). Karaca and colleagues, reported the use of methoxy poly(ethyleneglycol)-block-ploy(2-methyl-2-carboxyl-propylenecarbonate)-graft-dodecanol (mPEG-b-PCC-g-DC) copolymer to conjugate gemcitabine and encapsulate a Hedgehog inhibitor, Vismodegib (GDC-0449) into its hydrophobic core for the treatment of pancreatic ductal adenocarcinoma (PDAC). The in vivo stability of gemcitabine increased significantly after conjugation, and the drug combination, when administered to athymic nude mice bearing subcutaneous tumors generated using MIA PaCa-2 cells, efficiently inhibited tumor growth (Karaca et al. 2016).

3.3 Nanotechnology Approaches to Enhance Co-delivery

By physically confining drug combinations within a single carrier, pharmacokinetics for multiple drugs can be unified, ensuring all particle-treated cells receive a pairwise combination of drugs – maximizing therapeutic potential. One key challenge in this area involves combining three general categories of drugs, all of which display differing combinations of physiochemical properties: (i) large and hydrophilic proteins, (ii) small and hydrophobic small molecules, and (iii) moderately sized, hydrophilic, and highly anionic nucleic acids. Targeted, nanoscale delivery of therapeutic proteins remains an area currently underexplored in cancer therapy – particularly immunotherapy. Fahmy and coworkers recently addressed this challenge in the development of a combination cancer immunotherapy that reverts immunosuppressive tumor microenvironments (Fig. 16.3a) (Park et al. 2012). Using the cytokine, IL-2, which stimulates NK cell and cytotoxic T lymphocyte activity, in combination with TGF-β signal blockade, which does the same while additionally depleting tumor-promoting regulatory T lymphocytes (Tregs), the authors were able to develop a systemically administerable drug carrier that decreased tumor burden and improved overall survival when compared to single drug-loaded particles or to intratumoral injection of free drugs in immunocompetent, subcutaneous tumor xenograft mouse models (B16-F10). To achieve this, IL-2 was entrapped in a photo-polymerizable hydrogel matrix composed of poly(ethylene glycol) (PEG), polylactide, and a small molecule/β-cyclodextrin inclusion complex. The hydrogel core is encapsulated by a liposomal shell and photocured under ultraviolet light to obtain PEG-stabilized ‘nanolipogels’. A variety of other novel co-delivered drug combinations are listed in Table 16.1.

Fig. 16.3
figure 3

Nanomedicines Overcome Challenges in Combination Drug Delivery. (a) Bulky and hydrophilic proteins are notably difficult to co-deliver with small, hydrophobic molecular inhibitors. Here, IL-2 is entrapped in a photo-polymerizable hydrogel matrix composed of poly(ethylene glycol), polylactide, and a small molecule/β-cyclodextrin inclusion complex. The hydrogel core is encapsulated by a liposomal shell and photocured under ultraviolet light. Systemically administered ‘nanolipogels’ decreased tumor burden and improved overall survival better than single drug-loaded particles or intratumoral injection of free drugs in immunocompetent, subcutaneous tumor xenograft mouse models (B16-F10). IL-2 promotes a hostile tumor microenvironment by stimulating NK cell and cytotoxic T lymphocyte activity, while a TGF-β signal blockade (SB505124) does the same while additionally depleting tumor-promoting regulatory T lymphocytes (Tregs). (b) The delivery of synergistic drug ratios can also be ‘pre-programmed’ into nanoscale drug carriers. Here, a hydrophobic inhibitor of MAPK pathway signaling is co-encapsulated with a hydrophilic inhibitor of the PI3K pathway at a pre-defined synergistic drug ratio. Layer-by-Layer (LbL) polymer assembly around the drug-containing liposomal core facilitated both affinity-based and microenvironment-responsive tumor targeting, while simultaneously blocking both pathways and enhancing drug synergy beyond that of the free drug combination in vitro and in vivo. (c) Time-staggared delivery can be achieved through structural partitioning of drugs within nanoscale drug carriers. Here, a hydrophobic inhibitor of EGFR (erlotinib) partitions into the lipid bilayer of a nanoscale lipid vesicle, while a hydrophilic genotoxic agent, doxorubicin, is loaded into the aqueous particle core. Initial release of erlotinib ‘dynamically rewires’ cell signaling in lung and breast tumors in a manner that enhances subsequent cell killing by doxorubicin in a time-dependent manner. Systemically administered, folate receptor-targeting constructs induce a partial response in lung (A549) and breast (BT-20) tumor xenograft-bearing mice, while mice receiving single drug-loaded particles (doxorubicin) exhibit progressive disease. (d) Nanomedicines can also mitigate dose-limiting side effects of drugs with overlapping toxicity profiles. Although, doxocubicin and Fig. 16.3 (continued) trastuzumab (anti-HER2/ErbB2; Herceptin) – alone – induce Type I and Type II cardiotoxicity, respectively, the two are often co-administered in breast cancer patients. Engineering of an trastuzumab-doxorubicin antibody drug conjugate (ADC) rescues doxorubicin toxicity towards human primary cardiomyocytes by limiting cellular uptake to HER2-expressing cells. Here, a heterobifunctional linker (SMCC) containing an amine-reactive NHS ester and a thiol-reactive malimide crosslinks doxorubicin and IgG, respectively (Reprinted with permission from (a) Park et al. 2012, (b) Dreaden et al. 2015, (c) Morton et al. 2014, and (d) Zhang et al. 2013). Copyright (a) 2012 Nature Publishing Group, (b) 2015 American Association of Cancer Research, (c) 2014 American Association for the Advancement of Science, and (d) 2013 original authors under the Creative Commons Attribution License)

Full size image
Table 16.1 Selected examples of nanotechnology-enabled combination cancer therapies
Full size table

3.4 Nanotechnology Solutions: Stoichiometry/Ratiometric Dosing

Through ratiometric drug loading into nanoscale carriers, intracellular drug concentrations can be ‘pre-programmed’ for drug synergy, thus maximizing therapeutic potential and avoiding possible antagonistic interactions resulting from spatially and temporally heterogeneous delivery of free drug compounds. For example, Dreaden et al 2015 . co-encapsulated a hydrophobic inhibitor of MAPK pathway signaling with a hydrophilic inhibitor of the PI3K pathway at a pre-defined synergistic drug ratio (Fig. 16.3b) (Dreaden et al. 2015). Layer-by-Layer (LbL) polymer assembly around the drug-containing liposomal core facilitated both affinity-based and microenvironment-responsive tumor targeting, while simultaneously blocking both pathways and enhancing drug synergy beyond that of the free drug combination in vitro and in vivo. A related therapeutic in clinical development, liposomal irinotecan/floxuridine (CPX-1, 1:1 mol; Celator), co-encapsulates drugs at a previously identified synergistic drug ratio and maintains this drug stoichiometry both in plasma and in the tumor bed (Batist et al. 2009). Interestingly, efficacy of the liposomal formulation in tumor-bearing mice was superior to both free drugs dosed at their respective MTDs and the additive expectation from both singly loaded liposomes. Strikingly, a liposomal formulation encapsulating a previously identified antagonistic drug ratio (10:1 mol) was less effective than its singly loaded counterpart, suggesting a putative role for heterogeneous combination drug delivery in promoting resistance to therapy. Similar approaches employing cytarabine/daunorubicin (CPX-351, 5:1 mol) (Tardi et al. 2009a), irinotecan/cisplatin (CPX-571, 7:1 mol) (Tardi et al. 2009b), and paclitaxel/tanespimycin (17-AAG) (Katragadda et al. 2013), quercetin/vincristine (1:2 mol), and doxorubicin/camptothecin/Pt (1:2.5:3.6 mol) (Liao et al. 2014) are currently under investigation.

3.5 Nanotechnology Approaches to Tailor Drug Combination Timing and Sequence

Another powerful property of nanoscale drug carriers is their capacity to not only spatially regulate drug release in the body, but also to temporally control the sequence and kinetics of therapeutics released. To exploit the observation by Lee et al. that time-staggered inhibition of EGFR could sensitize breast cancer cells to DNA damaging chemotherapy, Yaffe, Hammond, and coworkers engineered a nanoscale drug carrier which achieved staged drug delivery through structural partitioning of drugs within a liposomal vesicle (Fig. 16.3c) (Morton et al. 2014). Here, a hydrophobic inhibitor of EGFR, erlotinib, partitions into the lipid bilayer of the vesicle, while a hydrophilic genotoxic agent, doxorubicin, is loaded into the aqueous particle core. The initial release of erlotinib from these nanoparticles ‘dynamically rewired’ cell signaling in lung and breast tumors in a manner that recapitulates optimally staggered delivery kinetics seen with free drug administration, enhancing subsequent cell killing by doxorubicin. Systemically administered, folate receptor-targeted Erlotinib/doxorubicin nanoparticles were shown to induce a partial response in both lung (A549) and breast (BT-20) tumor xenografts in nude mice, while mice receiving only single drug-loaded nanoparticles (doxorubicin) exhibited progressive disease.

Gnanasammandhan and colleagues described a noninvasive method to deliver drugs that allow for a high degree of spatial and temporal control. Upconversion nanoparticles (UCNs) were used to convert deeply penetrating near-infrared (NIR) light to UV-visible wavelengths that match the absorption spectrum of photosensitive therapeutics. This allowed for the use of deep-penetrating and biologically friendly NIR light for photoactivation (Gnanasammandhan et al. 2016). The UCNs were used for photodynamic therapy (PDT) and photoactivated control of gene expression. For PDT, the UCNs are coated with polyethylene glycol (PEG) for stabilization and folic acid for tumor targeting and then loaded with photosensitizers that would be expected to kill cells by singlet oxygen production, whereas for the photoactivated control of gene expression, knockdown of essential tumor genes is achieved using UCNs loaded with caged nucleic acid (Gnanasammandhan et al. 2016).

To achieve controlled release of drugs, Liao and colleagues recently used two novel macromonomers (MMs) and a novel cross-linker as building blocks for the construction of a multi-drug-loaded nanoparticle. CPT-MM and DOX-MM are branched MMs that release unmodified CPT and DOX in response to cell culture media and long-wavelength ultraviolet (UV) light, respectively (Liao et al. 2014).

3.6 Nanotechnology Approaches to Limit Compounding and Overlapping Toxicity

Combination drug carriers can overcome challenges from overlapping toxicity profiles by biasing tissue disposition away from off-target tissues or by decelerating bolus drug release in vital organs. To address cardiotoxicity from doxorubicin and trastuzumab (anti-HER2/ErbB2; Herceptin) combination therapy, Zhang et al. engineered an antibody-drug conjugate from the pair, thereby limiting cytotoxic doxorubicin delivery to cells expressing high levels of HER2, while blocking the compounded toxicity towards human primary cardiomyocytes (Zhang et al. 2013) (Fig. 16.3d). To address dose-limiting hepatotoxic effects from combined MAPK and PI3K pathway inhibition, Dreaden et al. engineered LbL nanoparticles that biased tissue disposition towards solid tumors and rescued both hepatic and renal tissue damage while improving antitumor efficacy in vivo. Similarly, Farokhzad and coworkers have found that aptamer-targeted PLGA nanoparticles can rescue the nephrotoxic effects of platinum chemotherapeutics while maintaining equivalent antitumor activity in vivo (Dhar et al. 2011; Kolishetti et al. 2010) and also accommodating the chemotherapeutic, docetaxel, in polylactide containing particles (Xu et al. 2013).

3.7 Combining Nucleic Acid Therapies with Other Drug Combinations

Since the discovery of RNA interference (RNAi) in 1997, there has been great interest in harnessing RNAi for the treatment of disease. RNAi is activated by double-stranded RNA (dsRNA), which includes short interfering RNA (siRNA) and microRNA (miRNA) and utilizes the endogenous RNAi pathway for the post-transcriptional silencing of gene expression. MicroRNAs form central nodal points in cancer development pathways and exert their effects by targeting various oncogenes and tumor suppressors (Kong et al. 2012; Zhang et al. 2007), while siRNAs can be used to efficiently silence the expression of any gene with high specificity. These include targets that are considered to be difficult to drug. Here we describe some platforms used to deliver RNAi-drug combinations.

One particular tumor type, for example, that could greatly benefit from RNAi therapy is triple-negative breast cancer (TNBC), which is characterized by the lack of progesterone, estrogen and HER2 receptors. It is non-responsive to conventional hormonal therapy (such as tamoxifen or aromatase inhibitors) or therapies that target HER2 receptors, such as Herceptin (trastuzumab) (Foulkes et al. 2010). RNAi-based approaches can therefore be beneficial for the treatment of TNBC. One way to accomplish this is by modulating endogenous miRNA levels in TNBC. miR-221 and miR-205 have been shown to be up and down regulated in TNBC respectively (Nassirpour et al. 2013; Piovan et al. 2012). Conde and colleagues recently showed that they could deliver a miR-205 mimic and an antagomiRNA (miRNA inhibitor) using a self-assembled RNA-triple-helix structure which is conjugated to dendrimers to form stable triplex nanoparticles that can achieve nearly 90 % tumor shrinkage 2 weeks post-gel implantation in a triple-negative breast cancer mouse model (Conde et al. 2016).

A promising approach made available by nanotechnology is to combine siRNA and chemotherapeutics in a single platform. Deng and colleagues used a controlled layer-by-layer process to co-deliver siRNA against a drug-resistance pathway (multidrug resistance protein 1) and a chemotherapy drug (doxorubicin) to challenge a highly aggressive form of triple-negative breast cancer resulting in an 8-fold decrease in tumor volume compared to control treatments with no toxicity observed (Deng et al. 2013). Xu and colleagues combined siRNA against targets involved in error-prone translesion DNA synthesis pathway (REV1 and REV3L) with conventional DNA-damaging chemotherapy (cisplatin prodrug) through self-assembly of a biodegradable poly(lactide-coglycolide)-b-poly(ethylene glycol) diblock copolymer and a self-synthesized cationic lipid. This nanoparticle formulation had a synergistic effect on tumor inhibition in a xenograft mouse model of human lymph node carcinoma of the prostate that was noticeably more effective than platinum monotherapy (Xu et al. 2013).

Other carriers that have been used to deliver the siRNA based combinations include liposomes (Gabizon et al. 1994; Chen et al. 2009a, b, 2010a, b; Li et al. 2008), micelles (Zheng et al. 2013; Shim et al. 2011; Zhu et al. 2010), polymers poly (d,l-lactide-co-glycolide) (PLGA) (Li et al. 2001; Fonseca et al. 2002), poly lactic acid (PLA) (Tobío et al. 1998; Dong and Feng 2004), polycapro lactone (PCL) (Yang et al. 2006), dendrimers (Biswas et al. 2013; Kaneshiro and Lu 2009; Kulhari et al. 2011), natural chitosan polymeric nanoparticles (Wei et al. 2013; Nagpal et al. 2010), silica (Santra et al. 2001; Qhobosheane et al. 2001; Kneuer et al. 2000) and other inorganic nanoparticles e.g calcium, gold, quantum dots, etc. (Sokolova and Epple 2008).

4 Limitations to Developing Combination Chemotherapeutics, Tumor-Specific Targeting, and Enabling Approaches

Like Frei, Freireich, and Zubrod, drug discovery has historically focused on the development of compounds with independent antitumor activity – those intended for use as monotherapies. Modern approaches to combination development have, in contrast, been largely limited to off-patent cytotoxic drugs. To address the challenge of integrating patent-protected targeted therapies in combination approaches, Merck and AstraZeneca initiated a seminal partnership in 2009 to share the costs of developing combination candidates, for example, AstraZeneca’s MEK inhibitor (AZD6244) and Merck’s protein kinase B inhibitor (MK-22060). Merck and Sanofi later followed with a similar agreement to investigate Merck’s MEK inhibitor, MSC1936369B, in combination with Sanofi’s PI3K/mTOR inhibitor, SAR245409, and class I PI3K inhibitor, SAR245408 . Bristol-Myers Squibb and Roche have likewise established agreements to combine Roche’s vemurafenib (Zelboraf) with BMS’s ipilimumab (Yervoy) for BRAF mutant metastatic melanoma. While highly promising, dose-limiting hepatotoxicity from the latter two combination approaches in Phase I clinical trials (Xu et al. 2013) highlights a key weakness of this approach: by neglecting combination effects during the discovery phase, more safe and/or effective combination candidates may be discarded simply as a result of exhibiting less potent independent antitumor activity. Although small molecule targeted therapies, in the past, have provided less incentive for early combination development (due to their single agent efficacy), the recent resurgence of cancer immunotherapy development will likely accelerate the integration of combination approaches earlier in the development pipeline, providing opportunities for the investigation of combinations with weak independent activity, but potent and safe combined therapeutic effects in the future.

A major but as-yet incompletely realized opportunity for nanoparticle therapeutics is the potential for tumor-specific targeting. Compounded organ-specific toxic effects such as those described above highlight a potential intrinsic advantage of – and challenge to – nanoscale drug delivery, whereby dose-limiting toxicities could be mitigated through nanoparticle-altered pharmacokinetics combined with favorable tumor tissue targeting profiles. The latter phenomenon can occur through size-dependent ‘passive’ tumor targeting or the ‘active’ targeting of tissues via stimuli-responsive behavior or affinity directed accumulation. While a number of recent publications seek to revisit the importance and prevalence of passive targeting in tumor delivery (Prabhakar et al. 2013; Park 2013) – largely attributed to the so-called enhanced permeability and retention (EPR) effect (Matsumura and Maeda 1986; Matsumoto et al. 2016) – nanoparticles are well known to preferentially accumulate in organs of the mononuclear phagocyte system (MPS), namely the liver and spleen, and to augment the accumulation of renally excreted drugs (i.e. <10 nm) in solid tumors. This property can be advantageous when designing treatments for hepatocellular carcinoma or immunotherapies, respectively; however, affinity directed targeting of tumor tissues remains an integral and underexplored challenge to the field. Chan and coworkers (Wilhelm et al. 2016), highlight this disparity in a recent retrospective literature analysis, noting only modest (0.5-fold) improvements in median tumor accumulation afforded by active targeting across multiple studies. In contrast, relatively smaller antibody-drug conjugates (ADCs) (Vaklavas and Forero-Torres 2012) and molecular polyconjugates (Rozema et al. 2007), while less prevalent, have demonstrated notably reproducible in vivo targeting capabilities. Future research investigating specific targeting ligands including small molecules (e.g. folic acid, bisphosphonate, carbohydrates), peptides (e.g. GE11, RGD, knottin), and proteins (e.g. IgG, Fab fragments, Centyrins), is expected to improve both the effectiveness and reproducibility of affinity targeting strategies employed by the field, as well as subsequent treatment outcomes from rational drug combinations.

5 Outlook and Conclusions

The intersection between cancer biology and nanotechnology is an exciting and emerging area of current academic and industrial research. With ongoing efforts in both fields, combination anticancer treatments are continuing to evolve, raising hopes for unprecedented antitumor responses and reduced toxicity. The emergence of newly engineered combination drug delivery platforms should allow us to combine different classes of drugs into a single nanoparticle with tunable functionality over local or temporal control of drug delivery increasing safety, efficacy, and durability. The co-delivery of different cancer therapeutic agents provides promising options to overcome chemoresistance. Recent reports provide strong evidence that combining different drugs using nanoparticles improves tumor killing compared to single agent therapy. While these approaches hold great promise, there still remain key limitations in their proof of concept. Most nanocarrier studies are currently performed in preclinical models, and desperately need to be translated into human clinical trials, particularly since the biodistribution, localization, and release profiles of these drugs may differ in humans. It is also pertinent for the safety profiles of the various carriers used for the delivery of these therapeutic agents to be further studied, with special focus on their toxicity and immune response. Given the progress that has been made in the field during the past 5 years, the future of rationally designed and personalized combination therapy using customizable nanoparticle delivery platforms looks promising.