Marriage of cell biology (the concept of ‘lysosomotropic drug delivery’) and the realization that water-soluble synthetic polymers might provide an ideal platform for targeted drug delivery led to the first synthetic polymer–drug conjugates that entered clinical trials as anticancer agents. Conceptually, polymer conjugates share many features with other macromolecular drugs, but they have the added advantage of the versatility of synthetic chemistry that allows tailoring of molecular mass and addition of biomimetic features. Conjugate characteristics must be optimized carefully to ensure that the polymeric carrier is biocompatible and that the polymer molecular mass enables tumour-selective targeting followed by endocytic internalization. The polymer–drug linker must be stable in transit, but be degraded at an optimal rate intracellularly to liberate active drug. Our early studies designed two HPMA [N-(2-hydroxypropyl)methacrylamide] copolymer conjugates containing doxorubicin that became the first synthetic polymer–drug conjugates to be tested in phase I/II clinical trials. Since, a further four HPMA copolymer–anticancer drug conjugates (most recently polymer platinates) and the first polymer-based γ-camera imaging agents followed. Polymer–drug linkers cleaved by lysosomal thiol-dependent proteases and the reduced pH of endosomes and lysosomes have been used widely to facilitate drug liberation. It is becoming clear that inappropriate trafficking and/or malfunction of enzymatic activation can lead to new mechanisms of clinical resistance. Recent studies have described HPMA copolymer conjugates carrying a combination of both endocrine and chemotherapy that are markedly more active than individual conjugates carrying a single drug. Moreover, current research is investigating novel dendritic polymer architectures and novel biodegradable polymers as drug carriers that will provide improved drug delivery and imaging probes in the future. The present paper reviews the clinical status of polymeric anticancer agents, the rationale for the design of polymer therapeutics and discusses the benefits and challenges of lysosomotropic delivery.
- N-(2-hydroxypropyl)methacrylamide copolymer (HPMA copolymer)
- lysosomotropic delivery
- polymer therapeutics
More than a century ago, Paul Ehrlich foresaw the potential of perfectly targeted immuno- and chemo-therapy, but effective targeting of drugs and macromolecular therapeutics (including antibodies, proteins, peptides and nucleotides) to diseased cells and, moreover, to specific intracellular compartments has proved to be difficult to achieve in practice . Although parenteral administration can bypass external biological barriers (such as the gastrointestinal tract and skin), limited extravasation (particularly blood–brain barrier), immune-surveillance mechanisms and the lack of unique cell-specific receptors still present major obstacles to disease-specific targeting. At the cellular level, the cell membrane and the inherent compartmentalization of organelles are additional obstacles. Nevertheless, the last two decades have witnessed the first nanosized drug-delivery systems to enter clinical trial and in some cases routine clinical use (reviewed in ). They include antibody conjugates (e.g. Mylotarg®, Tositumomab® or Zevalin®) [2–4], liposomes (e.g. DaunoXome™ or Doxil®/Caelyx®) , the first anticancer nanoparticle, Abraxane® , and polymer therapeutics. The latter include polymeric drugs, polymer–protein and polymer–drug conjugates [7,8]. Moreover, there has been growing realization that nanotechnology applied to medicine (‘nanomedicine’) has considerable potential to improve diagnosis and treatment of cancer [NIH (National Institutes of Health)/NCI (National Cancer Institute) Cancer Nanotechnology Plan: http://nano.cancer.gov/alliance_cancer_nanotechnology_plan.pdf]. Nanomedicines or ‘nanopharmaceuticals’ have been defined as complex nanosized drug-delivery systems or biologically active drug products consisting of at least two components, one of which is the biologically active ingredient (European Science Foundation Forward Look on Nanomedicine http://www.esf.org/publication/214/Nanomedicine.pdf). Thus, in this context, the above technologies are first-generation nanomedicines, and it is important to note that greatest success has been achieved via interdisciplinary research bringing together cell biology, polymer chemistry, medicine and pharmaceutical science.
Our early research [11,12] was born from a combination of Ringsdorf's vision of the polymeric drug carriers , and Trouet and de Duve's realization that the endocytic pathway might be useful for ‘lysosomotropic drug delivery’ . Use of a water-soluble polymer as a platform for drug-targeting limits cellular uptake to the endocytic route, and with careful choice of the carrier can produce long-circulating conjugates [t1/2α (initial half-life)≥1 h] that display passive tumour targeting due to the leakiness of angiogenic tumour blood vessels by the EPR (enhanced permeability and retention) effect . Conjugates can also be synthesized to contain targeting ligands to further promote enhanced uptake by receptor-mediated endocytosis (reviewed in [7,16–18]). So far, only linear polymers such as HPMA [N-(2-hydroxypropyl)methacrylamide] copolymers, PGA [poly(glutamic acid)], PEG [poly(ethylene glycol)] and polysaccharides (e.g. dextran) have been explored clinically as anticancer conjugates . However, recent advances in polymer chemistry are now creating many novel polymers that are being developed as drug carriers and imaging agents. These include new biodegradable polymers  and more complex polymeric architectures including block copolymer micelles, dendrimers and dendronized polymers (reviewed in ).
Several features are essential for effective design [7,12]. The polymer must be non-toxic and non-immunogenic, and polymer molecular mass must be high enough to ensure long circulation (for non-biodegradable polymeric carriers this must be <40000 g/mol to ensure eventual renal elimination of the carrier), but low enough to ensure endocytic internalization (typically <100000 g/mol ). It must be able to carry an adequate payload in relation to drug potency and the polymer–drug linker must be stable during transport, but able to release drug at an optimum rate on arrival. Agents with an intracellular pharmacological target must have access to the correct intracellular compartment, but it should be noted that lysosomotopic delivery with subsequent transfer of drug out of the endosomal/lysosomal compartment also provides the opportunity to bypass mechanisms of drug resistance associated with membrane efflux pumps such as P-glycoprotein.
Design of HPMA copolymer conjugates
Early studies showed HPMA copolymers to be non-toxic and non-immunogenic . Endocytic internalization with subsequent trafficking into lysosomes was verified using a variety of cell lines, together with 125I-labelled probes, HPLC assay of bound drug, epifluorescence and confocal microscopy and subcellular fractionation [23,24]. Conjugates lacking targeting ligands are usually internalized slowly as a solute by fluid-phase pinocytosis, but, in some cell types, conjugate-bearing hydrophobic drugs interact with the plasma membrane leading additionally to non-specific adsorptive uptake. Whereas low-molecular-mass chemotherapeutic agents usually enter cells rapidly (within minutes) by passage across the plasma membrane, HPMA copolymer conjugates enter cells slowly by endocytosis and active drug is liberated slowly thereafter (∼100% in 24–48 h). This makes comparative in vitro screening of cytotoxic activity meaningless. The first two HPMA copolymer conjugates to be tested clinically [25,26] were optimized to a molecular mass of ∼30000 g/mol  and they contained doxorubicin linked to the polymer via a tetrapeptidyl Gly–Phe–Leu–Gly linker that is stable in the circulation, but cleaved intracellularly by lysosomal thiol-dependent proteases, particularly cathepsin B . Many drug-delivery systems of that era were designed to deliver anthracyclines as they were new and showed a broad spectrum of activity.
Pre-clinical and clinical studies have begun to elaborate the multifactorial mechanism of action of HPMA copolymer conjugates (reviewed in ). Drug conjugation alters biodistribution dramatically, and pre-clinical and clinical pharmacokinetics have a good correlation. After intravenous administration, conjugates without a receptor-targeting moiety are initially retained in the vascular compartment (t1/2α is typically 1–6 h), and, if the drug–polymer linker is stable, the levels of free drug detected in plasma are very low, >100–1000 times lower than seen for the conjugated drug. Conjugate elimination occurs predominantly via the kidney with >50% of conjugated drug excreted within 24 h. Limitation of drug access to potential sites of toxicity (such as the heart and bone marrow) leads to a significant reduction in non-specific toxicity. Concurrently, EPR-mediated targeting increases tumour targeting [AUC (area under the curve) increases >70-fold in animal models]. This, coupled with an appropriate rate of drug liberation (activation) at the target site, are the two most important factors governing improved anti-tumour activity of polymer conjugates compared with free drug in vivo .
The addition of targeting ligands, such as galactose, transferrin, MSH (melanocyte-stimulating hormone), GRGD (Gly-Arg-Gly-Asp) and antibodies, can be used to promote receptor-mediated internalization (reviewed in [7,18]), but subsequent intracellular trafficking will be governed by the specific receptor/ligand used. Inappropriate trafficking can impact on anti-tumour activity. For example, galactose-containing HPMA copolymer conjugates designed to target the ASGR (asialoglycoprotein receptor) are directed to lysosomes. Subcellular fractionation and HPLC analysis has confirmed time-dependant lysosomal drug release from HPMA copolymer–Gly–Phe–Leu–Gly–doxorubicin  and HPMA copolymer–Gly–Phe–Leu–Gly–[3H]daunomycin–galactose conjugates. In contrast, addition of transferrin promotes rapid exocytosis via the transferrin receptor recycling pathway. Cathepsin B-mediated drug release is thus compromised and consequently anti-tumour activity is reduced. Importantly, endocytic uptake of HPMA copolymer conjugates has never provoked lysosomal dysfunction (enzyme inhibition or increased lysosomal membrane permeability), but, as more novel polymer conjugates are proposed (even for chronic use), careful experimentation must continue to investigate this possibility.
HPMA copolymer anticancer conjugates kill cells via necrosis and/or apoptotic pathways . Although non-degradable conjugates sometimes demonstrate in vitro cytotoxicity, this is due to contaminating free drug not the conjugate itself. In vivo experiments have shown that drug liberation is a prerequisite for activity, as non-degradable conjugates such as HPMA copolymer–Gly–Gly–doxorubicin and HPMA copolymer–Gly–Gly–en-Pt (where en is ethylenediamine and Pt is platinum) were all inactive. Corresponding conjugates containing degradable polymer–drug linkers demonstrate greater anti-tumour activity in many in vivo models than seen for the parent drug [12,16,18]. Growing evidence supports the immunostimulatory role of HPMA copolymer–doxorubicin conjugates. They can increase circulating levels of NK (natural killer) cells and anticancer antibodies in animals, and LAK (lymphokine-activated killer) cells in breast cancer patients treated with HPMA copolymer–Dox–IgG . Current evidence suggests that HPMA anticancer conjugates act, after EPR-mediated targeting, at the level of the tumour cell and not via effects on the tumour vascular endothelium. However, recent studies have described the first HPMA copolymer anti-angiogenic conjugates synthesized to contain O-(chloractyl–carbamoyl) fumagillol (TNP470) . This agent failed clinically because of unacceptable clinical neurotoxicity; however, TNP470 conjugation enhanced and prolonged the angiogenic activity in several in vivo models and abolished neurotoxicity. Drug release was again essential for activity.
Clinical status of polymer–drug conjugates: lessons learnt
HPMA copolymer–doxorubicin (PK1, FCE28068; ∼8 wt% doxorubicin) was the first synthetic polymer conjugate to enter phase I/II trials in 1994 . The MTD (maximum tolerated dose) was 320 mg/m2 (doxorubicin-equivalent), which is ∼4–5-fold higher than for doxorubicin (∼80 mg/m2). No polymer-related toxicity or immunogenicity was observed, and the DLTs (dose-limiting toxicities) were typical of the anthracyclines, for example febrile neutropenia and mucositis. Anthracycline cardiotoxicity was absent, even though individual cumulative levels of up to 1680 mg/m2 (doxorubicin-equivalent) were used; this is much greater than the maximum allowable for doxorubicin of 550 mg/m2. Activity was seen in chemotherapy refractory patients and, although γ-camera imaging with a 131I-labelled FCE28068 analogue generally showed poor resolution, uptake was seen in six of the 21 patients studied . The fact that cumulative doses of HPMA copolymer (never before administered to humans) of >20 g/m2 were given safely underlined the suitability of this polymer for further clinical development.
So far the only targeted polymer conjugate to enter clinical testing is an HPMA copolymer–Gly–Phe–Leu–Gly–doxorubicin conjugate also containing galactosamine (PK2, FCE28069) . Designed to target primary liver cancer, phase I/II trials showed an MTD of 160 mg/m2 (doxorubicin-equivalent), and several hepatocellular patients displayed partial responses and/or stable disease. γ-Camera imaging revealed conjugate liver levels of 15–20% dose (at 24 h), and hepatoma-associated drug was 12–50-fold higher than would have been achieved with administration of free doxorubicin .
Phase I evaluation of HPMA copolymer conjugates containing paclitaxel (PNU166945)  and camptothecin (MAG-CPT) was disappointing [32–35]. HPMA copolymer–camptothecin had no anti-tumour activity and severe cystitis was the DLT, and HPMA copolymer–paclitaxel showed paclitaxel-like neurotoxicity, even though some anti-tumour activity was seen. Both conjugates contain an ester linkage between the drug and the polymer that can degrade in the bloodstream and/or during renal elimination, unlike the peptide linkages designed for lysosomotropic drug delivery. Moreover, in colorectal cancer patients given a single dose of MAG-CPT before surgery, it was found that while the conjugate achieved similar levels in tumour and normal tissue at 24 h, released drug levels were lower in tumour than in normal tissue, indicating a lack of preferential accumulation by the EPR effect . These observations underline the need for careful conjugate design in relation to linker stability, drug loading and pharmacokinetics. More recently, two HPMA copolymer platinates have had greater success [36–38]. One contains a malonate ligand (carboplatinum analogue) (AP5280) and the other is a DACH (1,2-diaminocyclohexane) platinate (AP5346). In phase I clinical trials, both displayed reduced platinum-related toxicity and the DACH platinate also exhibited anti-tumour activity.
Although more than 11 polymer–anticancer drug conjugates have entered clinical development (reviewed in [7,17]) a PGA–paclitaxel conjugate (CT-2103, XYOTAX™) shows particular promise [39–41]. This contains drug linked through the 2′ position to PGA (molecular mass ∼17000 g/mol) to give a highly water-soluble product with an overall molecular mass of 49000 g/mol with a high drug loading (∼37 wt%). Unlike HPMA copolymers, the PGA polymer backbone is itself degraded by cathepsin B to liberate paclitaxel . Experiments in cathepsin B homozygous knockout mice have confirmed the importance of enzymatic degradation for drug release/activity. A recent randomized phase III clinical trial in NSCLC (non-small-cell lung cancer) patients showed the PGA–paclitaxel to have significantly reduced severe side effects and a statistically significant improvement in survival (40%) compared with that using vinorelbine [43,44]. Interestingly, there was a greater increase in survival in women (compared with men), and it appears that this correlates with oestrogen levels, which in turn correlate with an increased expression of cathepsin B. A pivotal trial is now ongoing to compare PGA–paclitaxel and paclitaxel (175 mg/m2) as a first-line therapy for women with NSCLC. In our research, pre-clinical studies in a variety of murine and xenograft tumour models have suggested that variations in cathepsin B levels (rate of drug release) may be more variable than the degree of EPR-mediated targeting (R. Duncan, Y.N. Sat, A.M. Burger, M. Bibby, H.H. Fiebig and E.A. Sausville, unpublished work).
To the future
Conjugates as a platform for combination therapy
Clinical proof of the concept has paved the way for synthesis of second-generation conjugates containing experimental chemotherapy and the use of water-soluble polymers as platforms for delivery of a cocktail of pendant drugs. Of particular interest are the first conjugates to combine endocrine therapy (an aromatase inhibitor, aminogluthimide) and chemotherapy doxorubicin . HPMA copolymer–aminogluthimide–doxorubicin showed greater cytotoxicity towards MCF-7 breast cancer cells in vitro than with either of the individual conjugates alone, or a simple mixture of them. HPMA copolymer–AGM (aminoglutethimide) inhibits intracellular aromatase after drug liberation , and mechanistic studies suggest that the enhanced activity is due to the kinetics of lysosomotropic drug liberation, leading to enhanced apoptosis . Given that acquired drug resistance is a particular problem for many of the new molecular targets, this new concept provides an interesting opportunity to provide a platform for delivery of two or more drugs (potentially at different rates) that might act synergistically to block multiple cellular pathways simultaneously.
Optimizing polymer structure and architecture
The first-generation polymers used as polymeric carriers have been inert linear architectures, but more sophisticated bio-responsive endosomolytic polymers are now being designed to aid delivery of macromolecular therapeutics into the cytosol . In addition, second-generation polymeric carriers that are pH-sensitive in the main chain (PEG–polyacetals ) can be advantageous as they undergo degradation at the lower pH encountered in endosomes and lysosomes following endocytosis. Unusual patterns of cellular trafficking and whole-body pharmacokinetics have been observed using highly branched polymers called dendrimers. These polymers have particular advantages for use as drug carriers as they have uniformity of structure and controlled three-dimensional molecular architecture with multiple termini for attachment of drugs, targeting groups and imaging agents. Our early studies synthesized the first dendrimer–chemotherapeutic agent conjugates in the form of a PAMAM (polyamidoamine) generation 3.5 Pt conjugate which contained ∼25 wt% Pt . The conjugate showed improved in vivo anticancer activity compared with that using cisplatin, and also displayed tumour targeting by the EPR effect .
Interestingly, in these studies, we noticed that the dendrimer–Pt showed significantly faster tumour efflux. This suggested an effect of molecular architecture on local pharmacokinetics. Subsequent experiments showed that PAMAM dendrimers showed very high rates of transcellular transport across an intestinal model system . This effect was dependent on the surface functionality (particularly rapid transfer for dendrimers with a carboxy group surface) and the generation. Subsequent experiments using endothelial cell lines and tumour cell lines confirmed that dendritic architecture promoted higher rates of cellular uptake, and, in some cases, also a high rate of exocytosis . These observations suggest that particular dendritic architecture may provide an interesting opportunity for the transport of drugs across biological barriers and/or to specific intracellular compartments. Ongoing experiments are trying to better understand the structure–activity relationships of such a transport phenomenon. On a cautionary note, it is essential that dendrimers chosen for further development towards clinical use are ‘biocompatible’ with respect to the proposed route and frequency of administration. While examining the cytotoxicity, haemolytic activity and biodistribution of several families of dendrimer, it became clear that many would not be suitable for parenteral use . The long-term success of these fascinating new architectures will be dependent on careful design of all features on a case-by-case basis with respect to proposed use and the future potential is obvious .
Endocytic pathways are still most often studied using the biological macromolecules that are physiological substrates (e.g. peptides, proteins and antibodies) and those intracellular pathogens that hijack these pathways to gain access to the cell. Nevertheless, collaborative efforts involving synthetic polymer chemistry and cell biology have, over the last three decades, successfully developed the first polymer-based probes and therapeutics that can be used both as tools and realistic drug-delivery systems. Polymers can be chosen that are not antigenic and decrease the immunogenicity of proteins bound to them. Moreover, the structure can be carefully tailored. Increased awareness of the opportunities of nanotechnology applied to medical challenges, and most importantly market approval of the first polymer therapeutics for routine clinical use leading to an explosion of interest in this field. Further understanding of both intracellular trafficking pathways and optimized design of novel polymers to harness such pathways without perturbing normal cell function will surely be a prerequisite for success!
Cellular Delivery of Therapeutic Macromolecules: Biochemical Society Focused Meeting in association with the Royal Society of Chemistry, the Royal Pharmaceutical Society of Great Britain and the Academy of Pharmaceutical Sciences held at Cardiff University, U.K., 29–31 August 2006. Organized and Edited by S. Akhtar (Cardiff, U.K.), M. Gait (MRC - LMB, U.K.), M. Gumbleton (Cardiff, U.K.) and A. Jones (Cardiff, U.K.).
Abbreviations: DACH, 1,2-diaminocyclohexane; DLT, dose-limiting toxicity; EPR, enhanced permeability and retention; HPMA, N-(2-hydroxypropyl)methacrylamide; MTD, maximum tolerated dose; NSCLC, non-small-cell lung cancer; PAMAM, polyamidoamine; PEG, poly(ethylene glycol); PGA, poly(glutamic acid)
- © 2007 The Biochemical Society