Colorectal cancer stem cells: a review of targeted drug delivery by gold nanoparticles

The proposed schematic mechanismviawhich 5-fluorouracil-loaded gold nanoparticles conjugated with CD133 antibody target colorectal cancer stem cells.

Her expertise include the identication and functional studies of Regulatory T cells and dendritic cells, allergic airway inammation mouse models, cancer immunology, and development of rBCG for TB vaccines. She is currently a University Lecturer and Principal Investigator at USM. Her current research is focused on nanoparticles that mediate immune response in asthma and cancer research both for human and animal models.

Introduction
Colorectal cancer (CRC) is among the most common cancers and accounts for nearly 9% of all cancers in the world 1,2 with high mortality rate due to advanced stage diagnosis and high rates of recurrence. 3 Studies have underlined the cause of recurrence and metastases as due to the presence of cancer stem cells (CSCs), a small subpopulation of cancer cells which are able to self-renew, differentiate and sustain tumor growth. 2,4,5 Colorectal cancer stem cells (CRCSCs) are characterized by the overexpression of CD133, a ve-transmembrane glycoprotein. 6,7 Currently, common cytotoxic chemotherapies used for the treatment of CRC are 5-uorouracil (5-FU), oxaliplatin and cisplatin. Chemotherapeutic drugs kill cancer cells by interfering with their cell growth and division mechanisms. 8 A few studies both on cellular and animal models have proved that treatment with these chemotherapeutic drugs, either alone or in combination with other chemotherapies, is effective in treating CRC. [9][10][11] Among these chemotherapeutic agents, 5-FU is one of the most prescribed anti-tumor drugs for CRC therapy. 12,13 Owing to its analogous structure to uracil, 5-FU can be incorporated into RNA and DNA and interferes with nucleoside metabolism. However, 5-FU-based chemotherapy also interferes with rapidly dividing healthy cells due to its lack of site specicity, and hence, causes common side effects such as hair loss, nausea and vomiting. 14,15 In addition, rapid clearance from blood circulation and poor distribution limit the therapeutic action of 5-FU. 16 Drug delivery systems in cancer therapeutics provide an alternative to minimize the limitations of conventional cancer chemotherapy by improving specic drug targeting, prolonging the circulation time and controlling drug release. 17,18 In addition, targeted drug delivery increases bioavailability to maintain the drug concentration upon arriving at the cancer site. 19 Within the past few decades, nanoparticles have received great interest Halima Alem is an Associate Professor at Lorraine University in the Department of Chemical Engineering. She has a strong background and more than a decade of experience in the design and elaboration of surfaces and nanomaterials for applications in healthcare and process engineering. She is an Associate Editor of the journal Nanomaterials and Nanotechnology and a member of two other editorial boards. She has co-authored more than 40 peer reviewed publications and participated in more than 70 national and international conferences, some of them as invited speakers. In October 2019, she was awarded by an Institut Universitaire de France delegation as a Junior member.  26 Here, we summarize a proposed system of targeted drug delivery consisting of 5-FU loaded gold nanoparticles (AuNPs) conjugated with CD133 antibody to target CRCSCs, supported by previous studies (Table 1).
Previously, the specic targeting of the CSCs in CRC by CD133 antibody has been studied using methoxy poly(ethylene Table 1 The advantages of the proposed 5-FU loaded AuNPs conjugated with CD133 antibody system targeting CRCSCs

System
Advantages Reference   AuNPs as drug carrier  Facile synthesis  27  Biocompatible  Multi-functionalization  Tunable surface  mPEG-surface coating  Reduce protein adsorption on AuNPs  28 and 29  Avoid macrophage uptake  5-FU  Antimetabolite  30  Inhibit DNA synthesis  Anti-CD133 monoclonal antibody conjugate  Overexpression of CD133 has been associated with the relapse,  metastasis and chemotherapy resistant   7 and 31 Interfere on specic proteins overexpressed CD133 involved in tumorigenesis (targeted cancer therapy) CRCSCs targeting Targeting CRCSCs rather than tumor bulk could be effective in reducing risk of relapse and metastasis 32 Fig. 1 The proposed schematic mechanism of 5-fluorouracil (5-FU) loaded gold nanoparticles (AuNPs) conjugated with CD133 antibodies to target colorectal cancer stem cells (CRCSCs) in a targeted drug delivery system. Methoxy polyethylene glycol (mPEG)-stabilized AuNPs loaded with 5-FU and conjugated with CD133 antibody will target the CRCSCs instead of the bulk of colorectal cancers due to the overexpression of CD133 antigen on the surface of CRCSCs. High affinity binding of the CD133 antibody ligand with the targeted cells will increase the delivery efficiency and hence protect healthy cells. The AuNPs with loaded 5-FU will be internalized by cells via endocytosis. The acidic environment in the tumor may trigger the cleave of 5-FU from the AuNP complex inside cell endosomes to enhance cell toxicity by interfering with DNA synthesis.
glycol)-poly(3-caprolactone) (mPEG-PCL) nanoparticles loaded with anticancer drug 7-ethyl-10-hydroxy-camptothecin (SN-38). 33 The in vitro and in vivo studies showed that CD133targeting nanoparticles have increased cytotoxicity and inhibited tumor growth in HCT116 cells and mouse xenogra models, respectively, compared with the none antibody-targeted nanoparticles. In this case, the therapeutic effects on CRC have been enhanced by the CD133 antibody conjugates because the chemotherapeutic agent SN-38 is efficiently guided to the overexpressed CD133 markers at the tumor site. Fig. 1 shows the proposed schematic mechanism of 5-FU loaded AuNPs conjugated with CD133 antibody to target CRCSCs. In drug delivery systems, nonionic polymer-surface modication of AuNPs by methoxy polyethylene glycol (mPEG) thiol is commonly used to improve blood residence time, reduce reticulo-endothelial system (RES) uptake and avoid non-specic targeting. 28 The AuNP complexes will be internalized by CRCSCs through passive (neovasculature) and active targeting (anti-CD133 ligand-receptor docking) followed by fusion with lysosomes. Owing to the changes of some pathological events in tumor cells, such as pH and concentration of intracellular glutathione (GSH), AuNPs can be engineered to be activated by these endogenous stimuli in order to release the payloads at the tumor site and hence enhance the therapeutic effects and avoid systemic side effects to healthy cells. Regularly, the extracellular pH of solid tumor cells is acidic (5.5 to 6.5) compared to the physiological condition 34 due to the anaerobic glycolysis-produced lactic acid in hypoxia 35 and substantial H + generation because of higher metabolic activity. 36 Besides the acidic tumor environment, 5-FU will be released on exposure to lysosomal enzymes followed by interfering with DNA synthesis and RNA processing and functioning to cause tumor cell death.

Colorectal cancer stem cells
According to the cancer stem cell (CSC) theory, a small subpopulation of cells share embryonic stem cell characteristics, 2 giving rise to metastatic cancer cells and responsible for the high rate of recurrence. 37 Indeed, CSCs share the same major signaling pathways with embryonic stem cells, including the Wnt, Notch, transforming growth factor beta (TGF-b), and Hedgehog signaling pathways. 38 Nevertheless, quiescent CSCs express stemness genes, have the capability to self-renew and differentiate into other non-stem cancer cells and resist traditional chemotherapy and radiotherapy. 39 Another crucial feature of CSCs is possessing tumorigenic potential, which is measured by the ability of the cells to initiate xenogra tumors upon transplantation. 38 Thus, targeting CSCs instead of the tumor bulk population is a therapeutic goal in CRC management, particularly in preventing cancer relapse. The isolation and identication of CSCs from CRC further elucidates several putative stem cell markers. The identication of CRCSC markers will be benecial in identifying the disease progression as well as the risk of recurrence among patients. 40 Table 2 shows several putative stem cell biomarkers of CRC and their prognostic value as previously reported.
It is interesting to note that several discrepancies of prognosis values and tumorigenic potential have been reported among the listed CRCSC biomarkers. Despite the above listed inconsistencies, there is strong evidence from numerous studies suggesting that CD133 is a notable marker to identify CRCSCs. 6,7,53 CD133, also known as prominin-1, is a 97 kDa pentaspan transmembrane glycoprotein 54,55 and has been reported to be mostly localized in membrane protrusions. 56 However, due to glycosylation, CD133 yields a 120 kDa protein. 55 It was rst discovered in 1997 as a hematopoietic stem cell (HSC) marker, which is associated with stem cell maintenance. 55 Recently, CD133 has been widely studied as a putative biological marker for CRCSCs to predict tumor progression, prognosis value and chemoradiotherapy resistance. 7 The enhanced CD133 expression in CRCSCs suggests that CD133 is a suitable biomarker target for the treatment of CRC. Fang et al. demonstrated that CD133 + cells isolated from primary CRC gave rise to spheroid cultures, which have the ability to self-renew and maintain CD133 expression in serum-free media. 57 This nding supports the notion that CRCSCs that are CD133 + are able to initiate tumors, which may increase CRC occurrence.
It was also suggested that CRCSCs escaped conventional chemotherapy targeting rapidly dividing cells due to their quiescent nature, thereby inhibiting the complete eradication of CRC. 58 In addition, the presence of overexpressed multidrug resistance protein 1 (MDR-1), detoxifying enzymes and DNA repair proteins in CSCs increased chemoresistance in malignant tumors. 59,60 A study by Ong et al. reported that CD133 + CRC cells were more resistant towards 5-FU-based chemotherapy. 44 Therefore, CD133-targeted therapeutic strategies in CRC might be valuable to eliminate CSC, which is associated with a high risk of relapse.

5-Fluorouracil
5-FU is an antimetabolite and was discovered in 1957 to display tumor inhibitory activities by Heiderberger et al. 61 Currently, it is used as a rst line chemotherapy agent and is widely used as an antineoplastic drug. 12,62 5-FU is a natural pyrimidine uracil analog with a uorine atom inserted into the C-5 position in place of hydrogen (Fig. 2). 63 Besides CRC, 5-FU is also commonly used to treat other solid malignancies arising from breast, stomach, pancreatic and head and neck cancers. 64,65 The regulation of 5-FU metabolism involves thymidylate synthase (TS), dihydropyrimidine dehydrogenase (DPD) and orotate phosphoribosyl transferase (OPRT) enzymes. 66 It has been reported that only 1-3% of the original dose of 5-FU contributes to the cytotoxicity in rapidly proliferating tumor and normal cells via anabolic actions. 67 Meanwhile, about 80-85% of 5-FU is subjected to biotransformation in the liver and catabolized into inactive metabolites, uorinated b-alanine, by DPD 67,68 and 15-20% is eliminated in urine. 68 Thus, the development of drug delivery agents for 5-FU will enhance the 5-FU cytotoxicity to the targeted tumor cells. As an analog of uracil, 5-FU readily crosses the cell membrane using the same facilitated transport mechanism as uracil and is incorporated into nucleic acids, which eventually contributes to the cytotoxicity. 30,69 5-FU and uracil are in continuous competition with each other due to their identical metabolic pathways (Fig. 3). Instead of uridine triphosphate (UTP), the active metabolite of 5-FU, uorouridine triphosphate (FUTP), is incorporated into ribonucleic acid (RNA) and interferes with its processing and functioning. 67 Fluorodeoxyuridine triphosphate (FdUTP) is another 5-FU metabolite that can be incorporated into deoxyribonucleic acid (DNA) instead of deoxythymidine triphosphate (dTTP). In addition, instead of normal pyrimidines, deoxyuridine monophosphate (dUMP), the metabolite of 5-FU, uorodeoxyuridine monophosphate (FdUMP), forms a covalent ternary complex with TS and 5,10-methylene tetrahydrofolate (CH 2 THF) and inhibits deoxythymidine monophosphate (dTMP) production, which eventually interferes with DNA synthesis. 62,67 In general, 5-FU causes cancer cell death by interfering with DNA synthesis and RNA processing and function.
Despite the effectiveness of 5-FU as a chemotherapeutic drug, one of its limitations is short plasma half-life (10-20 min) caused by fast degradation. 63 In addition, due to the nonselectivity and poor bioavailability of 5-FU to tumor sites, a higher dose is required in order to reach the effective drug concentration. Nonetheless, high doses of chemotherapeutic agents lead to elevated systemic side effects and increased incidence of drug resistance, thereby limiting the usefulness of 5-FU in CRC chemo-treatment.

Considerations on using nanoparticles in drug delivery
The application of nanoparticles in drug delivery systems has enhanced the conventional anticancer treatments, particularly in terms of bioavailability, specic targeting, and reducing adverse side effects. However, nanoparticle clearance from the circulation by the reticuloendothelial system (RES), including the mononuclear phagocytic system (MNS), is one of the major concerns in drug delivery systems. The MNS consists of the liver, spleen, lymph nodes, bone marrow, skin and macrophages. 26 Meanwhile, scavenger receptors on Kupffer cells are responsible for recognizing, engulng and eliminating the opsonized nanoparticles in the liver. 26,70 Thus, several requirements should be incorporated in designing nanoparticles to evade rapid clearance and enhance effective delivery of chemotherapeutic agents. Generally, the mechanisms of nanoparticles in drug delivery are affected by the physicochemical properties of the nanoparticles, such as size, surface chemistry and particle shape. These characterizations may impact nanoparticle-cell interactions, particularly in cellular and tissue uptake.
First and foremost, it is apparent that nanoparticles offer an ideal system for chemotherapeutic drug delivery due to their tunable sizes. In medical applications, nanoparticles are materials with preferential dimensions of 1-100 nm. 71 In addition, to avoid recognition by the RES, nano-sized particles also contribute to passive targeting into the tumor zone. The idea of nanoparticles as promising carriers for anti-tumor drugs was owing to the study by Matsumura and Maeda, which suggested that nanoparticles could accumulate in tumors. 72 The formation of neovasculature in tumors increases the permeability of the vascular endothelial layer, resulting in leaks and susceptibility for the passage of molecules and/or nanosystems with sizes in the range of hundreds of nm. 73,74 It is reported that the range of size of the vascular pore cut-off exhibited by most solid tumors is between 380 nm and 780 nm. 75 Thus, nano-sized particles within this size range can easily pass through the cancer vascular pores. This phenomenon is known as the enhanced permeability and retention (EPR) effect (Fig. 4); it enables passive targeting of nanoparticles carrying drugs and is reported to enhance drug efficiency by up to 2.3-fold compared with when loaded with the free drugs. 76 Secondly, hydrophobic and highly surface charged nanoparticles are highly recognized by the RES and removed from the circulation. 77 Thus, various current nanoparticle-based drug delivery systems undergo surface modications by graing polymer coatings such as polyethylene glycol (PEG), poly(Nvinyl-2-pyrrolidone) (PVP) and dextran in order to avoid agglomeration and systemic clearance by macrophages, thereby increasing the bioavailability of chemotherapeutic agents. [78][79][80] Hydrophilic PEG shielded the presence of nanoparticles to avoid opsonization by passive diffusion across the lipophilic cell membrane and hence increase the nanoparticle availability in systemic circulation. 79 For instance, doxorubicin-loaded PEGylated liposomes have shown signicant efficacy in breast cancer treatment. 81 In addition, the stability of the formulated nanoparticles in the delivery system inuences the cellular uptake. Monodispersity of nanoparticles in suspension is required to avoid agglomeration, which increases the rate of clearance. Basically, more positively or negatively charged nanoparticles prevent agglomeration in suspension due to their larger degree of repulsion. However, since many plasma membrane surfaces are negatively charged, nanoparticles with cationic surfactants display high affinity for cellular uptake via endocytosis during nanoparticle-cell adhesion. 82 Lastly, nanoparticle shape is another crucial factor that can affect the cellular uptake efficiency and actions of nanoparticle delivery. In a study conducted by Huang et al., mesoporous silica nanoparticles (MSNs) of long-rod shape showed an increased number of internalized particles compared with short-rod shaped and sphere-shaped MSNs. 83 However, when compared to spherical shape MSNs, the long-rod shaped MSNs showed higher cytotoxicity than the short-rod shaped and sphere-shaped particles. Another study found that the sharpness of the nanoparticles inuenced the intracellular translocation and excretion of nanoparticles. 84 In that study, it was reported that the sharp-shaped nanodiamonds translocated from the endosome to the cytoplasm quickly, with difficult cellular excretion compared with the round-shaped nanodiamonds, suggesting that the sharp-shaped nanoparticles ruptured the endosomal membrane, thereby reduced the Fig. 4 The enhanced permeability and retention (EPR) effect, which allows passive targeting of nanoparticles. The neovasculature phenomenon in the tumor causes disordered vascular endothelial layers and permits the passage of nanoparticles. excretion rate, and remained in the cytoplasm for an extended time. These studies suggest the interplay between cell interactions and nanoparticle design as one of the important aspects in developing drug delivery systems. In addition, it is important to note that in cellular systems, particles enter the cells via the endocytosis route, which consists of phagocytosis or pinocytosis as summarized in Fig. 5. Large particles are taken up by phagocytosis/micropinocytosis, while nanoparticles with sizes lower than 200 nm are engulfed by micropinocytosis, either by a clathrin-dependent, caveolae-dependent or clathrin/caveolae-independent route. 78,85 The clathrin-mediated pathway transfers the nanoparticles into lysosomes and a portion of the nanoparticles are recycled back to the extracellular space. Meanwhile, nanoparticles that follow the caveolae pathway are transported to caveosomes to be either translocated to the endoplasmic reticulum/Golgi body or enter the endosomal pathway. 86 Engineered physicochemical properties as well as elemental compositions of the nanoparticles are crucial, since they are responsible for the fate of the nanoparticles generally in blood circulation and hence determine the effectiveness of the carried chemotherapeutic agents. 87 Table 3 summarizes the common physicochemical properties of nanoparticles and the pathways for their internalization as described by Kou et al. 78 Apart from the physicochemical properties of the particles, it is also important to note that the initial materials for nanoparticles should be non-toxic materials or biocompatible to avoid toxic effects. 88

The applications of gold nanoparticles in drug delivery systems
Gold (Au) has been applied in the biomedical eld for many years. It has been used in bioimaging 89 as well as in the treatment of arthritis/inammation since the beginning of the 20 th century. 90 Since gold salts show anti-inammatory activities, they have been incorporated into disease-modifying antirheumatic drugs (DMARDs), one of the drug classes to treat rheumatic arthritis. 91 However, gold salts as DMARDs have been discontinued and replaced by others, probably due to the affinity of gold for DNA and hence its interference in cell function. 80 The toxic effects of AuNPs have been supported by a study conducted by Qiu et al. which reported that acute and chronic exposure to AuNPs interrupts gene expression. 92 To date, no gold-based nanomedicines have been approved by the US Food and Drug Administration (FDA) yet. 80 Besides as a direct therapeutic agent, AuNPs have been intensively studied in other biomedical applications, including in plasmonic photothermal therapy, photodynamic therapy and targeted drug delivery. 93 To date, various studies have proven the overwhelming potential of AuNPs with some in clinical trial phases. A recent phase II clinical study into the photothermal effect (AuroLase Therapy with identication number: NCT02680535) utilized the unique optical tunability of AuNPs that can convert light into heat to thermally kill prostate cancer. 94 A near-infrared laser that did not destroy healthy tissues was specically designed to excite the accumulated AuNPs in cancerous cells and hence, the removal of the prostate tumor was precise enough without signicant damage to surrounding healthy tissues.
One of the advantages of AuNPs compared to other metal nanoparticles is that the gold core is relatively inert, and so it is considered to be biocompatible and non-toxic. 95 Previously, it has been reported that ionic AuNPs show obvious toxicity at 25 mM. 96 Nevertheless, another study found that plain AuNPs did not show any cytotoxicity to human cells up to 250 mM. 97 Thus, it is suggested that AuNPs themselves are non-toxic, but then toxicity could be derived from their functionalization, such as with surface coatings. In spite of this, AuNPs have shown promising potential to deliver chemotherapeutic drugs targeting cancer cells either via passive targeting, active targeting or both. As previously described, passive targeting is due to the neovasculature formed by tumor cells. Meanwhile, the active targeting approach is via the surface functionalization of AuNPs using immobilized ligands such as proteins, antibodies or small molecules for specic cell targeting on the targeted surface of the membrane. 80 Thus, targeted cancer therapies can be benecial for the controlled delivery of chemotherapeutic agents stemming from the fact that cancer cells overexpress specic antigens. 98 Owing to the expression of specic antigens, drug carriers are oen conjugated with specic antibodies to bind with the particularly expressed antigen, which will be discussed later. The incorporation of a biocompatible coating seems to enhance the usefulness of AuNPs in drug delivery systems. A recent clinical trial on an AuNP-based product, Aurimune (CYT-6091), recombinant human tumor necrosis factor (rhTNF) bound to colloidal gold via a PEG-linker, showed that AuNPloaded rhTNF was three times as effective in treating advanced stage cancer patients compared to the native rhTNF without introducing any signicant toxic effects. 99 This nding supports the notion that AuNPs with some modications could be a promising targeted drug carrier in mediating conventional chemotherapeutic agents to eradicate tumors.
Undoubtedly, in order to justify further development, in vitro studies using human cell line models are very valuable for initial screening prior to clinical validation and translation. However, preclinical in vitro studies are benecial to study specic tumor biology and treatment responses only. Thus, the in vitro effects may not fully represent the effects in medical application, since more complex interactions exist among the diverse organs and physiological systems in the human body. In spite of this, there are some AuNP studies that have reached clinical trials as previously discussed, and this could be a promising area for researchers to further explore their applications, particularly in drug delivery systems for cancer management. Table 4 shows the application of functionalized-AuNPs in drug delivery systems targeting various types of cancers that have been reported in vivo and in vitro.

Design of gold nanoparticles for cellular uptake
As previously discussed, physicochemical properties in terms of size, shape and surface properties play a major role in determining the competency of AuNPs as an optimal drug carrier, particularly in order to overcome certain biological barriers, such as macrophage clearance. For drug delivery purposes, the interaction between membrane receptors and nanoparticles is one of the most important aspects that regulates the rate of cellular uptake (endocytosis) and hence increases the accumulation of drug-loaded nanoparticles at the tumor site. In this regard, the nanoparticle size plays an important role to avoid early clearance by the MPS organs. Several previous studies have reported the rate of cellular uptake and AuNP accumulation among different sizes of AuNPs (Table 5). In AuNP chemical synthesis using the Turkevish method, the size of the synthesized AuNPs can be easily adjusted by varying the ratio of gold salt and citrate. 105 Thus, tunability of the AuNP size during chemical synthesis can maximize the efficient delivery of therapeutic drugs to targeted cells.
In biomedical applications, it is important to take note that the size of AuNPs should not be too small to avoid the interactions with biological macromolecules leading to cytotoxicity, but not too large to reduce the chance of macrophage clearance. Based on previous studies, it was highly suggested that the size of endocytosis-susceptible guest particles should be around 40-50 nm. 106,112,114 This is in agreement with the 100 nm size limit of the plasma membrane to form endocytic vesicles. 119 In addition, AuNPs with a size around 40 to 60 nm were reported to have optimal values of membrane bending rigidity and ligandreceptor binding interaction for endocytosis. 114 However, it is another important consideration to note that the hydrodynamic diameter of AuNPs will increase upon coating or surface modi-cation and hence affect the rate of cellular uptake. Thus, it is crucial to consistently monitor the increase of AuNP size before and aer modications to avoid clearance due to larger size.
In addition to size, the shape of AuNPs is another crucial physical factor to affect endocytic uptake by cells. The particle shape can enhance the process of cellular membrane wrapping during endocytosis. 120 It was reported that spherical AuNPs increased the cellular uptake when compared to rod-shaped Table 4 The current applications of gold nanoparticle-loaded anticancer drugs in drug delivery systems AuNPs 106,115 probably due to the elongated particles requiring longer time for membrane wrapping. 121 Despite the fact there were some previous studies reported that rod-shaped AuNPs enhanced the therapeutic outcomes in cancer drug delivery, including prolonged circulation times 122 and enhanced drug loading, 123 it is also important to note that the synthesis of gold nanorods requires the use of a growth-directing surfactant (cetyltrimethylammonium bromide, CTAB). Free CTAB without nanorods was shown to induce cytotoxicity in the human colon cancer HT-29 cell line. 124 However, it was suggested that a polyacrylic acid (PAA) coating was able to reduce the exposure of the cells to the CTAB and hence attenuated the CTAB-capped gold nanorod toxicity. Thus, instead of focusing on the rate of cellular uptake alone, the toxic effects of drug carriers on the cellular system need to be prioritized as well. In a separate study by Xie et al., mPEGylation of triangular AuNPs resulted in the highest cellular uptake by RAW264.7 cells, followed by rods and star-like structures. 85 The highest uptake of the triangular-like shape was highly associated with the cytoskeleton arrangement and mediated by the dynamin-dependent endocytosis pathway. These studies indicate that the geometry of AuNPs is an additional feature that can inuence the biological interactions, particularly in cellular uptake. However, more repetitive studies need to be conducted to gain a better understanding of the cellular uptake and toxicity implications of drug carriers with different shapes on the cellular system.

Effect of surface properties of AuNPs on cellular uptake
In addition to size and shape, the interactions of nanoparticles with lipid bilayer cell membranes mainly rely upon the chemical functionalities coated on the nanoparticle surfaces. Generally, modication of the AuNP surface is an important feature that inuences the effective use of the particles for drug delivery systems. According to Pissuwan et al., surface modications for AuNPs are needed in drug delivery systems to prolong residence of the AuNP conjugates in circulation, avoid RES clearance, ensure effective attachment of the desired targeting or therapeutic molecules, improve AuNP stability by preventing agglomeration and nally, neutralize the possible cytotoxicity caused by stabilizing surfactants during AuNP synthesis. 125 For instance, the surface of AuNPs is coated by either neutral charged groups such as PEG or zwitterionic ligands, or charged functional groups that are anionic or cationic to avoid rapid clearance via non-specic uptake by the RES and provide active nanoparticle interactions with cells, respectively. 126 Moreover, a previous preclinical trial of CYT-6091 showed that PEGylated AuNPs carrying rhTNF can escape immunogenicity by avoiding RES phagocytic clearance, thereby allowing the nanotherapeutic to be circulated longer in the blood circulation. 99 Another study using gold nanorods reported that surface-modied gold nanorods with PEG showed no agglomeration due to the nearly complete neutral charge on the nanorod surface, as measured by zeta-potential, while increasing the circulation of the AuNP conjugates in in vivo systems. 127 An example of drug delivery using charged functional groups is provided by Hauck et al. 128 In this study, gold nanorods of 18 Â 40 nm were coated with various layer-by-layer polyelectrolyte (PE) coatings to produce positively and negatively charged nanorods by electrostatic interactions. Nanorods coated with negatively charged poly(4-styrenesulfonic acid) (PSS) followed by a layer of positively charged poly(diallyldimethylammonium chloride) (PDADMAC) exhibited a higher cellular uptake via passive targeting compared to the negatively charged PSS-coated nanorods. Electrostatic interactions between the negatively charged cellular membrane and positively charged nanoparticles may explain this higher uptake.

Active targeting or tumors by antibody conjugates
To date, no commercial antibody-conjugated drug-loaded nanoparticles have been applied in cancer therapy. However, antibody-conjugated drugs, such as Mylortag®, are commercially available on the market to treat acute myeloid leukaemia. 75 The application of antibodies in cancer therapy has received great attention owing to the expression of specic antigens by cancerous cells. Thus, the development of antibody conjugates as the active targeting ligands in nanoparticletargeted drug delivery will improve the specicity of chemotherapeutic drug delivery without inducing signicant damage to healthy cells.
Antibody graing via covalent linkages at the surface of AuNPs is preferable compared to physical adsorption to reduce competitive displacement of the antibodies in the blood. 75 To demonstrate the efficacy of functionalized antibodies in drugloaded nanoparticles targeting tumors, Kou et al. synthesized cationic SM5-1 single chain antibody (scFv) and polylysine (SMFv-polylys)-coated poly(lactide-co-glycolide) (PLGA) nanoparticles loaded with paclitaxel to induce cytotoxicity in human hepatocellular carcinoma cell lines (Ch-hep-3). 129 It was found that the paclitaxel-loaded PLGA in the presence of scFv antibody showed higher cell death compared to the paclitaxel-loaded PLGA only. In a recent AuNP study, conjugation of EpCAM or TARP antibodies with paclitaxel-loaded AuNPs via an N-(3dimethylaminopropyl)-N 0 -ethylcarbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) (EDC/NHS) coupling reaction showed signicant T47D breast cancer cell reduction compared to the paclitaxel-loaded AuNPs without antibody conjugates. 130 Both studies support the notion that antibody conjugation provides specic drug delivery in order to enhance the chemotherapeutic action at the targeted tumor site, and hence could be further developed to hinder interactions with healthy cells.

Concluding remarks
The treatment of CRC has emerged as a topic of interest, since CRC displays high rates of recurrence as well as poor prognosis. The discovery of subpopulations in CRC known as stem cells had led to advanced studies targeting CRCSCs, owing to the stem cell theory that they are able to self-renew and sustain tumor growth. The traditional CRC chemotherapeutic agent, 5-FU, targeting RNA and DNA synthesis and function causes cytotoxicity by interfering with nucleoside metabolism. However, due to the short half-life, non-selectivity and poor biodistribution, the usefulness of 5-FU has been limited. Thus, there is an urgent need to construct an effective targeted drug delivery system to maximize the efficacy of chemotherapeutic agents. Recently, AuNPs have been shown to resolve the limitations of traditional chemotherapy via drug delivery systems. The noble AuNPs have various properties that enable them to be designed and functionalized to act as an effective drug carrier, such as biostability, non-toxicity and feasibility for surface modication. Owing to the overexpression of CD133 in CRCSCs, the specic targeting of AuNPs loaded with 5-FU towards CRCSCs can be enhanced via bioconjugation of anti-CD133 to the nanoparticles. Therefore, it will be possible for AuNPs to reach and release 5-FU at the targeted site (CRCSCs), thus reducing the interactions with the healthy cells and hence eliminating the systemic side effects previously reported for naked 5-FU without nanodrug carriers.

Conflicts of interest
There are no conicts of interest to declare.