Polydopamine-enabled surface functionalization of gold nanorods for cancer cell-targeted imaging and photothermal therapy

Abstract

Aim

A novel biomimetic strategy was employed for presenting antibodies on gold nanorods (NRs) to target growth factor receptors on cancer cells for use in photothermal therapy.

Materials & methods

Polydopamine (PD) was polymerized onto gold NRs, and EGF receptor antibodies (anti-EGFR) were immobilized onto the layer. Cell-binding affinity and light-activated cell death of cancer cells incubated with anti-EGFR-PD-NRs were quantified by optical imaging.

Results

PD was deposited onto gold NRs, and antibodies were bound to PD-coated NRs. Anti-EGFR-PD-NRs were stable in media, and were specifically bound to EGFR-overexpressing cells. Illumination of cells targeted with anti-EGFR-PD-NRs enhanced cell death compared with nonirradiated controls and cells treated with antibody-free NRs.

Conclusion

PD facilitates the surface functionalization of gold NRs with biomolecules, allowing cell targeting and photothermal killing of cancer cells. PD can potentially coat a large variety of nanoparticles with targeting ligands as a strategy for biofunctionalization of diagnostic and therapeutic nanoparticles.

Carcinogenesis is a process whereby cells with upregulated growth and inhibited death characteristics emerge in microenvironments characterized by intermittent hypoxia, leaky vasculature and acidity [1,2]. EGF receptors (EGFRs), also known as ErbB receptors, are often overexpressed on cancer cell surfaces and can lead to activation of multiple pathways linked to malignancy [3]. EGFR overexpression is linked to recurrent and high-risk cases in multiple types of cancer including those of the head and neck [3,4], as well as the breast [5,6]. Although EGFRs are important diagnostic markers and targets for therapeutic intervention [7,8], these treatments still fail in many cases and can give rise to more invasive, drug-resistant tumors [9]. Therefore, new diagnostics and more effective therapies that target growth factor receptors [10] are needed.

Nanoscale platforms are becoming increasingly investigated for cancer diagnosis and therapy [11,12]. A number of nanoparticle systems have been investigated in conjunction with EGFR targeting toward a variety of cancers, including nanoparticles composed of gold [13], iron oxide [14,15], dendrimers [16], high-density lipoprotein [17], L-protein [18], poly(ε-caprolactone) [19], lactic acid–lysine copolymers [20], gelatin [21], human serum albumin [22], heparin [23] and liposomes [24].

Surface plasmon resonant (SPR) gold nanoparticles have broad potential as diagnostic and therapeutic materials due to their low reactivity, sub-100 nm size and shape-controllable dielectric properties, which lead to strong interactions with light throughout the visible and near infrared (NIR) portion of the electromagnetic spectrum [25–32]. Gold nanorods (NRs) [27,29,32–34] have the advantage of possessing strong longitudinal plasmon resonances within the NIR tissue transparency window preferred for in vivo applications.

A limited number of studies have explored the use of EGFR antibody-conjugated gold NRs for cancer diagnostic imaging and photothermal therapy [35,36]. In order to realize the full biomedical potential of gold nanoparticles, nanoparticle aggregation and nonspecific interactions with molecular and cellular constituents of the biological system should be inhibited. These goals can be accomplished through the control of surface chemical composition, for example by the conjugation of stabilizing and bioactive molecules to the nanoparticle surface. Electrostatic interactions [35] and gold–thiolate bonds [36] have been employed to immobilize EGFR-targeting moieties to gold NRs for targeted imaging and photothermal therapy. A further challenge associated with the functionalization of gold NRs is related to the exchange of desired surface ligands with cetyltrimethylammonium bromide (CTAB), a highly toxic surfactant employed during gold NR synthesis [37–40].

Molecular adhesives inspired by catechol-rich mussel adhesive proteins offer many attractive properties for surface modification that can be employed in metal nanoparticle functionalization. Catechols have the ability to form a variety of strong covalent and noncovalent interactions with inorganic and organic material surfaces [41,42], and molecules with catechol anchors are now widely used to functionalize surfaces for a range of biomedical applications [43–50]. Recently, catechol-containing polymers were used to spontaneously form polymer-functionalized gold and silver nanoparticles in aqueous solutions [51], harnessing the multifunctional catechols to reduce metal ions to plasmonic metal crystals, to adhere strongly to the formed crystals and to covalently crosslink the polymer coating surrounding the metal nanoparticles. Catechol-functionalized heparin was bound to gold nanoparticles for diagnostic x-ray imaging of liver cancer [52]. A special example of a mussel mimetic coating is formed through the polymerization of dopamine molecules under alkaline aqueous conditions, resulting in the deposition of thin conformal coatings of polydopamine (PD) on material surfaces [53]. The method is simple to implement, works on most materials, does not require multistep peptide and polymer synthesis and serves as a convenient and versatile ‘primer’ for further deposition of functional inorganic and organic coatings for a variety of uses [54–62].

In this report, we describe the preparation and anticancer performance of PD-coated NIR-active gold NRs. Anti-EGFR antibodies were conjugated to PD-coated NRs and their use in targeted photothermal therapy of cancer cells was demonstrated. Antibody-functionalized NRs were significantly more toxic to cancer cells in vitro compared with untargeted NRs when irradiated with a broadband light source. PD-mediated surface modification is suggested to be a useful strategy for the conjugation of cancer-specific ligands to nanoparticle surfaces, enabling the formation of biofunctional diagnostic and therapeutic metal nanoparticles.

Materials

Dopamine hydrochloride, CTAB (99%), NaAuCl₄ 2H₂O (99%), NaBH₄ (98%), ascorbic acid, glycine and AgNO₃ (99%) were obtained from Sigma-Aldrich (MO, USA). The pH value of the glycine solution (0.2 M) was adjusted to 8.0 with 2M sodium hydroxide before use. Methoxy-polyethylene glycol (PEG) -thiol (molecular weight: 5000) was purchased from Laysan Bio (AL, USA). Alexafluor 633 goat anti-mouse IgG antibody was purchased from Invitrogen (CA, USA). Anti-EGFR antibody was obtained as a commercial infusion (erbitux/cetuximab, 2 mg/ml; ImClone LLC, Bristol-Myers Squibb, NY, USA) from the Robert H Lurie Comprehensive Cancer Center pharmacy of Northwestern University (IL, USA). Ultrapure, deionized water (18.2 MΩ cm) was used to prepare all of the aqueous solutions. OSCC15 cells were acquired from the lab of David Crowe at the University of Illinois-Chicago (IL, USA). MCF7 and MDA-MB-231 cells were acquired from Dean Ho at Northwestern University.

Methods

Synthesis of gold NRs

The synthesis of CTAB-coated gold NRs (CTAB-NRs) was performed according to a slightly modified method previously described in the literature [63]. Briefly, CTAB aqueous solution (0.2 M, 5.0 ml, heated to 30°C) was mixed with 0.5 mM NaAuCl₄ (5.0 ml). Ice cold 0.01 M NaBH₄ (0.6 ml) was added to this solution and sonicated for 5 min to form a brownish-yellow seed solution. A total of 50.0 ml of 0.2 M CTAB was then gently mixed with 50.0 ml 1.0 mM NaAuCl₄ and 0.1 ml 0.1 M AgNO₃ to form a growth solution. Ascorbic acid was added to the solution as a mild reductant (78.8 mM, 0.7 ml), followed by the addition of 120 μl of the seed solution. After 45 min, 100 ml of this gold NR solution was mixed with 100 ml 0.2 M glycine (pH 8.0). This solution was allowed to react overnight, without stirring, at an ambient temperature.

Inductively coupled plasma optical emission spectrometry

A total of 1 ml of the NR stock solution was centrifuged, the supernatant was removed and the pellet was resuspended in 500 μl 70% nitric acid. The sample was sonicated in a Branson 2510 sonicator in an ice-water bath for 5 h to dissolve the NRs, after which 154 μl of the resulting solution was added to 4.846 ml pure water and the gold concentration determined in a Varian VISTA-MPX ICP-OES spectrometer (Varian, Inc., CA, USA) using a calibration curve constructed from 0.1, 0.5, 1, 2, 5 and 10 ppm gold standards (dissolved in 2% nitric acid).

Preparation of PD-coated gold NRs

A total of 1 ml of CTAB-NR suspension was centrifuged and the pellet was resuspended in 1.9 ml of a 516 μM dopamine solution buffered to pH 8.5 using 10 mM tris(hydroxymethyl) aminomethane (TR IS) buffer. The NR suspension was sonicated for 30 min, centrifuged at 9000 rpm for 10 min and the supernatant was removed. Pelleted PD-coated NRs (PD-NRs) were resuspended in 2 ml of ultrapure deionized water, and a UV–visible–NIR extinction spectrum was acquired to confirm the presence of longitudinal and transverse SPR peaks. Additional centrifugation and removal of the supernatant was performed to remove unbound dopamine.

Preparation of PEG-grafted gold NRs

For comparison to antibody-coated NRs, PEG-grafted NRs (PEG-PD-NRs) were prepared by suspending PD-NRs overnight in 1.7 ml 10 mM TRIS buffer (pH 8.5) containing 5 μl of 0.4 mM methoxy-polyethylene glycol-thiol at 20°C. PEG-PD-NRs were isolated by centrifugation and resuspended in ultrapure deionized water.

Antibody immobilization on gold NRs

A PD - coated gold N R suspension (9.1 × 10¹² NRs/l) was resuspended in 500 μl 10 mM TRIS buffer (pH 8.5). To immobilize the anti-EGFR antibody, 0–350 nM antibody was added to 500 μl PD-NR suspensions and sonicated (in order to inhibit PD aggregation) for 30 min. To remove unbound antibodies from the suspension of PD-NRs functionalized with anti-EGFR antibodies (anti-EGFR-PD-NRs), the suspension was then centrifuged at 9000 rpm at 23°C for 10 min, supernatant was decanted and the pellet was resuspended in ultrapure water or DMEM. Anti-EGFR-PD-NRs were incubated with 10 nM fluorescent goat anti-mouse IgG secondary antibody, centrifuged, resuspended in 500 μl ultrapure water and tested in a Synergy 4 Hybrid Multi-Mode Microplate Reader (Biotek, VT, USA). To quantify antibody density, the fluorescent signal was normalized to NR background fluorescence, and compared to a standard fluorescence curve from free IgG antibody.

Electron microscopy

Pelleted NRs (5 μl) were dropped on electron microscopy (EM) grids (Ted Pella, CA, USA) and allowed to dry overnight in ambient conditions. Bright-field transmission EM, dark-field Z-contrast EM, EM using secondary electrons and energy-dispersive x-ray spectroscopy spectral imaging were performed on a Hitachi HD-2300 ultra-high resolution field-emission scanning transmission electron microscope (Hitachi City, Japan).

Optical spectroscopy

A Hitachi U-2010 spectrophotometer was used to acquire optical spectra in a two-beam geometry. To match the NR suspension, 10 mM TRIS buffer, ultrapure deionized water or DMEM was used for the reference beam. Spectral scans were performed over the 200–1000 nm range of wavelengths in the UV–visible–NIR region of the spectrum. A deuterium lamp was used for the 200–340 nm UV range illumination and a halogen lamp was used for the visible and NIR illumination. Spectral resolution was 1 nm.

In vitro cell imaging with NRs

A total of 5 × 10⁵ OSCC15, MDA-MB-231 and MCF7 cancer cells were grown in tissue culture plates for 3 days. OSCC15 cells were grown in high-glucose DMEM (10% fetal bovine serum [FBS], 1% gentamycin), MCF7 cells were grown in high-glucose DMEM both with and without insulin (10% FBS, 1% pen/strep) and MDA-MB-231 cells were grown in RPMI-1640 media (10% FBS, 1% pen/strep). Optical imaging was performed on both cells in solution and cells that were adhered to the plates. For the solution experiments, 2 ml of trypsin was added to confluent plates of OSCC15 cells, cells were removed and 400 μl was added to 2 ml of DMEM high-glucose media. Cells were centrifuged at 1500 rpm for 5 min, the supernatant was removed and 2 ml more media was added. NR suspensions were added to the cells and incubated for 30 min. Cells were centrifuged at 1500 rpm for 5 min, the supernatant was removed, 2 ml of media was added and, finally, cells were centrifuged and plated onto glass optical microscopy slides. Bright-field images were acquired in a leica DMRX microscope. To quantify NR binding on cells adhered to plates, 5 × 10⁴ MDA-MB-231 and MCF7 cells were grown in 12-well plates and imaged in an optical coherence tomography (OCT) system after incubation with anti-EGFR-PD-NRs for 1 h (Supplementary Figure 1 & Supplementary material; see online at www.futuremedicine.com/doi/suppl/12.82). OCT intensity between 750 and 850 nm, that was localized to individual cells, was quantified and plotted as a function of antibody concentration.

NR-mediated photothermal therapy

OSCC15, MDA-MB-231 and MCF7 cells were cultured in 6- and 12-well plates and were allowed to grow to confluence. Cells were incubated with anti-EGFR-PD-NRs (0.1–20 pM) for 30 min–3 h, washed twice with media and irradiated with a 1 mm spot size SuperK^™ Versa (NKT Photonics, Birkerød, Denmark) pulsed laser source (480–850 nm) for 5, 10 or 15 min. The average power of the irradiation was 150 mW, which corresponded to a beam intensity of 60 W/cm², and an energy/area range of 18–54 kJ/cm². At the plasmon peak center wavelength (780 nm), the pulse width was 31 ps and power density was 0.18 W/nm. A total of 1 μl 4 mM calcein and 1 μl 4 mM ethidium bromide was added to each well and cells were incubated for 15 min. Fluorescence microscopy was performed on a Leica DMIRB microscope, with a 250 W maximum Hg arc lamp and a QIClick detector (QImaging, Vancouver, Canada). For each irradiated spot, both calcein-stained cells and ethidium bromide-stained cells were counted and the percentage of viable cells was calculated (n = 1–3 spots, ~1000 cells/spot).

Results

In this study, a general strategy was used to immobilize antibodies onto gold NR surfaces through an intermediate PD layer (Figure 1). First, high yield (>95%) CTAB-NRs were synthesized using established protocols. The apexes of the transverse and more intense longitudinal extinction peaks of CTAB-NRs were located at 520 and 783 nm, respectively, and were stable during storage in excess CTAB at room temperature for at least 6 months. CTAB-NRs were imaged by transmission EM and had an average length of 62 ± 6 nm and a width of 17 ± 3 nm (n = 123) (Supplementary Figure 2). A total of 0.294 ± 0.013 mg/l of gold was detected in the NR suspension treated with nitric acid described in the methods. Taking this concentration, the density of gold and assuming a right circular cylinder shape, the gold NR concentration in the initial stock suspension was determined to be 1.81 × 10¹³ NRs/l (30.1 pM).

Figure 1.

Preparation of EGF receptor antibody-conjugated gold nanorods (anti-EGF receptor-polydopamine-nanorods) using a polydopamine ‘primer’ coating

EGFR: EGF receptor; NR: Nanorod.

PD deposition was accomplished by dispersion of CTAB-NRs in an alkaline dopamine solution, upon which spontaneous deposition of PD onto NRs, accompanied by displacement of the CTAB ligand was observed. The UV-visible extinction spectra acquired during this process revealed a linear redshift over time (Figure 2A), corresponding to a PD coating thickness of 5–10 nm as observed by EM under the secondary electron mode (Figure 2B). To further confirm the presence of PD on the surface of gold NRs, x-ray photoelectron spectroscopy was performed (Supplementary Figure 3). In contrast to the single C1s peak observed from CTAB-NRs, high-resolution C1s spectra of PD-NRs were characterized by a peak centered at 284.5 along with a significant shoulder toward higher energies representative of C–O and C–N bonds in the PD. After this deposition, NRs were nontoxic to cells used in this study compared to CTAB-coated controls (Supplementary Figure 4).

Figure 2. Polydopamine polymerization onto gold nanorods.

(A) Redshift in the longitudinal surface plasmon resonance of the gold nanorods during reaction in 520 mM dopamine in pH 8.5. (B) Electron microscopy of polydopamine-coated nanorods using the secondary electron mode (scale bar = 50 nm).

To provide clinically relevant anticancer functionality, anti-EGFR antibodies were immobilized onto PD-NRs at pH 8.5 for 30 min. Redshifting and broadening of the longitudinal SPR occurred during conjugation (Figure 3A). To further confirm the presence of the antibody on the gold NR, anti-EGFR-PD-NRs were stained with a secondary anti-mouse IgG antibody, and a fluorescent signal representative of an antibody monolayer (~350 antibodies/NR) on the PD coating was detected. The stability of the anti-EGFR-PD-NRs was tested in DMEM media over a 24 h time span, with no statistical loss in SPR intensity observed (Figure 3B).

Figure 3. Anti-EGF receptor antibody immobilization onto polydopamine-coated nanorods.

(A) UV-visible extinction of PD-NRs (30-min PD deposition time, 5-nm-thick PD coating) before and after anti-EGFR antibody conjugation (maximum peak intensity normalized to one; nanorod = 0.015 nM). (B) Stability of longitudinal plasmons of PD-NRs functionalized with anti-EGFR antibodies in serum-containing DMEM (normalized to initial intensity). CTAB: Cetyltrimethylammonium bromide; EGFR: EGF receptor; NR: Nanorod; PD: Polydopamine.

To explore EGFR targeting, anti-EGFR-PD-NRs were incubated with two cancer cell lines characterized by high EGFR expression, OSCC15 oral cancer cells [64] and MDA-MB-231 breast cancer cells [65]. Optical microscopy revealed that anti-EGFR-PD-NRs visibly interacted with cells, whereas NRs were not visible on cells treated with PEG-PD-NRs in identical conditions (Figure 4A & B). Furthermore, a concentration-dependent increase in NIR light intensity (750–850 nm) localized to cellular structures was observed in MDA-MB-231 cells incubated with anti-EGFR-PD-NRs (Figure 4C), with a 4.1-fold increase in light intensity observed at an antibody concentration of 5.7 nM ([NR] = 14 pM). No increase in intensity was observed in low-EGFR-expressing MCF7 breast cancer cells incubated with the same concentration of anti-EGFR-PD-NRs (Supplementary Figure 5A).

Figure 4. Optical imaging of cells incubated with nanorods.

Bright-field light microscopy images of OSCC15 oral cancer cells incubated with (A) polyethylene glycol-PD-NRs and (B) anti-EGFR-PD-NRs. Scale bars in (A) and (B) = 20 μm. (C) Optical coherence tomography intensity from individual MDA-MB-231 breast cancer cells as a function of anti-EGFR-PD-NR concentration.

EGFR: EGF receptor; NR: Nanorod; PD: Polydopamine.

Next, we demonstrated photothermal therapy of OSCC15 cells targeted with anti-EGFR-PD-NRs (Figures 5 & 6). In control experiments, no toxicity of cells was observed immediately after irradiation in the absence of NRs (Figure 5A) or in cells incubated with anti-EGFR-PD-NRs without irradiation (Figure 5B). Whereas cells that were incubated with 9.92 pM PEG-PD-NRs and then irradiated maintained 91% viability (Figure 5C), viability of cells incubated with the same concentration of anti-EGFR-PD-NRs and then irradiated decreased to 24% (Figure 5D). Thus, cells targeted with anti-EGFR-PD-NRs and irradiated with light were killed more efficiently than any other treatment (Figure 6A).

Figure 5. Photothermal therapy of OSCC15 cells.

Images of cells treated as follows: (A) cells irradiated for 5 min in the absence of nanorods (NRs); (B) cells incubated for 1.5 h with 9.92 pM anti-EGF receptor-polydopamine (PD)-NRs with no irradiation; (C) cells incubated for 1.5 h with 9.92 pM polyethylene glycol-PD-NRs and irradiated for 5 min; and (D) cells treated with 9.92 pM anti-EGF receptor-PD-NRs and irradiated for 5 min.

Figure 6. Viability of OSCC15 and MDA-MD231 cells treated with nanorods followed by photoirradiation.

(A) Viability of OSCC15 cells after incubation with anti-EGFR-PD-NRs ([antibody] = 4.0 nM; [NR] = 9.92 pM), irradiation for 5 min in the absence of NRs, irradiation for 5 min after incubation with PEG NRs ([NR] = 9.92 pM) and irradiation for 5 min after incubation with anti-EGFR-PD-NRs ([antibody] = 4.0 nM; [NR] = 9.92 pM). (B) Dependence of MDA-MB231 cell viability on anti-EGFR-PD-NR concentration after incubation with anti-EGFR-PD-NRs for 1 h followed by 5 min of irradiation. (C) Dependence of MDA-MB231 cell viability on incubation time with anti-EGFR-PD-NRs ([NR] = 3.0 pM). Viability was measured (A) 30 min or (B & C) 19 h after a 5-min irradiation treatment. EGFR: EGF receptor; NR: Nanorod; PD: Polydopamine; PEG: Polyethylene glycol.

NR-mediated photothermal therapy was also performed on MDA-MB-231 breast cancer cells, which exhibit overexpression of EGFRs [65]. Like OSCC15 cells, MDA-MB-231 cells treated with anti-EGFR-PD-NRs but not irradiated maintained high viability at all time points tested in this study, whereas irradiation significantly decreased viability in a NR concentration-dependent manner (Figure 6B). After 5 min of irradiation, cells not treated with NRs were 79% viable; in the same conditions, cells incubated with anti-EGFR-PD-NRs were between 37 and 70% viable depending on the NR concentration. Furthermore, longer incubation times with anti-EGFR-PD-NRs provided enhanced toxicity when coupled with irradiation (Figure 6C). By contrast, no significant toxicity difference was observed in MCF7 breast cancer cells incubated with 0.1–20 pM anti-EGFR-PD-NRs and irradiated with light compared with cells that were irradiated without NR incubation (Supplementary Figure 5B–D).

Discussion

Gold NRs have shown promise as contrast agents and photothermal therapeutic agents in vitro and in vivo [66,67]. However, a versatile chemical approach is needed to couple the optically functional metal nanoparticles with a broad range of biologically specific molecules in physiological environments to fully realize their biomedical potential. Catecholamine molecular adhesives, inspired by the protein glues of mussels that adhere to wet, chemically heterogeneous surfaces in diverse aqueous environments, can be exploited to functionalize metal nanoparticle surfaces for biomedical applications. PD is perhaps the simplest form of a mussel-mimetic coating in terms of ease of use and versatility, spontaneously depositing as a thin conformal coating on surfaces by taking advantage of a rich repertoire of chemical interactions with surfaces [53–62]. The formation of PD from dopamine is driven by the auto-oxidation of dopamine catechols to quinones at alkaline pH, leading to the formation of dihydroxyindole, which subsequently polymerizes to form heterogeneous oligomers and macromolecules resembling natural eumelanin [68]. These highly conjugated species are further capable of aggregation into larger structures via π-stacking [68], ultimately leading to the formation of nanometer-thick layers of melanin-mimetic PD [69]. The thickness, permeability and deposition rate of PD coatings can be easily tailored through deposition conditions such as pH, dopamine concentration and the presence of oxygen [70–72].

PD deposition on the gold NR surface was indicated by a redshift of the longitudinal SPR (Figure 2A), x-ray photoelectron spectroscopy analysis showing a C1s shoulder representative of C–O and C–N bonds in the PD coating (Supplementary Figure 3) and EM images using the secondary electron mode revealing a 5-nm thick PD layer surrounding gold NRs (Figure 2B). The redshift of the plasmon can be related to the thickness of the coating through the equation:

$$\mathrm{\Delta }\lambda =\text{m}\mathrm{\Delta }\text{n}[1-exp(-2\text{d}/{\text{I}}_{\text{d}})]$$

with d representing the coating thickness [28]. Using the observed shift of 3.3 nm for a 30 min PD reaction, m = 224 nm/refractive index units, Δn = n_PD − n_water = 1.6 – 1.33 and l_d = 200 nm [28,69,73], the calculated thickness of the PD coating was 4.5 nm, generally matching the coating thickness detected by EM. A curve correlating the PD coating thickness to reaction time can be seen in Supplementary Figure 6. After PD deposition, no toxicity related to the presence of CTAB was observed (Supplementary Figure 4).

A variety of interactions are possible between PD and the metal NR surface, including π-electron interactions [74] and metal coordination [75] from the catechols as well as electrostatic interactions between positively charged amines and the negatively charged noble metal surface. However, the interaction between the conformal PD layer and the metal NR surface does not just rely on individual metal–polymer interactions, but arises additionally from the robust, cohesively crosslinked coating surrounding the entire NR. This cooperative strategy mimics the molecular interactions employed by mussel adhesives to survive in the turbulent and high ionic strength sea environment.

PD coatings further serve as a platform or ‘primer’ onto which further surface modifications can be performed. In this study, PD was used in this way to provide a versatile chemical interface for the conjugation of antibodies and other biomolecules onto the surface of NIR-active gold NRs for targeting cancer cells. When coupled to metal nanoparticles, antibodies offer complementary biological functions such as specific binding to human cell surface receptors associated with cancer phenotypes such as EGFR, HER2 [76], glucose transporters like GLUT1 that are upregulated under glycolysis [77] and mucin receptors such as MUC1 [78]. To quantitatively characterize the general immobilization of antibodies onto PD-NRs, fluorescently tagged IgG antibodies were incubated with PD-NRs in alkaline conditions. A 22-fold increase in immobilized antibody was observed compared with CTAB-NRs (Supplementary Figure 7A), illustrating the advantage of using PD for biomolecule conjugation to surfaces. The number of antibodies per NR could be controlled between 8 and 350 (Supplementary Figure 7B), the latter number generally corresponding to an antibody monolayer immobilized onto a 5-nm thick PD layer surrounding a 62 × 17 nm gold NR. Control experiments and geometrical calculations ruled out significant effects from fluorescence enhancement or quenching from metal NRs. We hypothesize that amines from the antibody covalently react with quinones in the PD layer through Michael addition or Schiff base reactions under the mildly basic conditions of the functionalization step, although electrostatic or other noncovalent interactions between the antibody and the PD cannot be entirely ruled out [79].

To create an anticancer NR using the biomimetic PD approach, anti-EGFR antibodies were immobilized onto the PD layer. Immobilization of anti-EGFR antibodies onto gold NRs produced a redshift and broadening of the longitudinal and transverse plasmon bands of the gold NR (Figure 3A). Anti-EGFR-PD-NRs were stable in the serum-containing medium for at least 25 h (Figure 3B), thus avoiding NR aggregation and facilitating targeting to EGFR-expressing cells (Figure 4) where they provided a light-activated therapeutic response (Figure 5 & 6). The use of amine-containing TRIS buffer during the reactions did not affect the cell-targeting ability of the antibody-functionalized NRs. Furthermore, while sonication is not generally used with functional biomolecules, its use here was confined to inhibiting aggregation during nanoparticle preparation and did not affect the functionality of the antibody. It is also noted that anti-fouling PEG polymers can be integrated onto the PD layer alongside antibodies to further decrease the potential nonspecific interactions that can occur between functionalized NRs in suspension and components of the biological milieu.

Anti-EGFR-PD-NRs added to cell culture media strongly interacted with EGFR-expressing cells (Figure 4), producing a fourfold enhancement of the OCT signal in the NIR range (Figure 4C) due to the backscattering from gold NRs bound to cells. OCT signal increases were detected at antibody concentrations as low as 0.3 nM, significantly lower than the dissociation constant values for unbound antibody of 4.54 nM reported in the literature [80], which could be caused by multivalent enhanced binding [81] through multiple antibodies on a single NR. By contrast, OCT intensity from MCF7 cells lacking high EGFR expression was not enhanced by incubation with anti-EGFR-PD-NRs at any of the concentrations tested (Supplementary Figure 5A), implying that the increase in intensity from the MDA-MB-231 cells upon incubation with anti-EGFR-PD-NRs was due to binding of NRs to cells through a specific antibody–receptor interaction.

In the future, quantitative EGFR-targeted spectroscopic imaging with anti-EGFR-PD-NRs could be used to help identify high risk, EGFR-expressing phenotypes with bright-field microscopy of ex vivo biopsies or OCT of tumors in vivo. Due to their sub-100 nm particle size, targeting specificity, and the PD interface that decreases in vivo toxicity and immunological response [60], anti-EGFR-PD-NRs should be capable of intravenous in vivo administration. These properties may increase antibody delivery efficiency to tumor sites compared with unbound antibodies through the enhanced permeability and retention effect that occurs in leaky tumor vasculature [12].

Once bound to cancer cells, irradiation with light transforms NRs into potent therapeutic agents, providing a second therapeutic mechanism in addition to the current antibody-based EGFR inhibition. Our results indicate that light irradiation of anti-EGFR-PD-NRs targeted to oral or breast cancer cells provided enhanced therapeutic efficacy compared with control treatments. The observed toxicity arising from NIR irradiation of anti-EGFR-PD-NRs bound to OSCC15 and MDA-MB-231 cells is likely due to direct thermal damage caused by cavitation, as irradiation of gold NRs can produce extreme heating at the NR surface that can increase bulk temperatures by 10 –50°C [66,67,82]. Indeed, in an experiment where a PD-NR suspension was irradiated with a 60 W/cm² beam for 5 min (18 kJ/cm²), bulk temperature rises of over 20°C were observed, which correspond to much higher temperature increases at the surface of the NRs (Supplementary material). Alternatively, toxicity may also arise from secondary effects associated with hyperthermia such as calcium influx-induced membrane blebbing [83]. While the substantial increase in NR suspension temperature upon irradiation is compelling evidence that photoinduced heating causes direct cellular damage, further research should be conducted to fully elucidate the mechanism of cell death produced by light irradiation of anti-EGFR-PD-NRs. However, taken together, this study demonstrates that the PD-based strategy can target NRs to cells with high EGFR expression and can specifically kill them with light.

Conclusion

PD offers a versatile chemical interface for the coupling of biological molecules with metal nanoparticles, as illustrated by the synthesis of PD-functionalized gold NRs conjugated with an anti-EGFR antibody. In this study, the anti-EGFR-PD-NRs were stable in serum-containing media and when selectively bound to cells overexpressing EGFR, where they were detected by optical imaging. They further provided a potent light-activated photothermal therapeutic response, producing significant photoinduced toxicity of breast and oral cancer cells upon exposure to NIR light. These nanoparticles have both diagnostic and therapeutic potential against cancer.

Future perspective

While the responses observed in both oral and breast cancer cells in this study demonstrate the clinical potential of anti-EGFR-PD-NRs toward EGFR-specific pathologies, the biomimetic surface-modification strategy is broadly applicable in the formation of biofunctionalized metal nanoparticles. More specifically, any antibody could be immobilized onto the PD layer to target NRs to other cancer-related cell-surface receptors such as HER2 [76], GLUT1 [77] and MUC1 [78]. The optical signals from nanoparticles that specifically target disease-related cell-surface receptors such as EGFR can be employed for clinically relevant diagnostic imaging, wherein the signal is quantitatively related to cell-surface receptor density and is used to identify specific cellular pathologies. Furthermore, the versatility of the PD surface coating facilitates a great variety of highly tailored surface modifications that could include multiple biological targeting ligands, passivating polymers such as PEG and other functional molecules such as peptides [36,84] and DNA aptamers [85], providing an avenue to form nanoparticles that target and treat cells through multiple mechanisms of action.

We also note that our approach could prove to be a useful adjunct to existing surgical oncology practices. For example, we envision that NIR irradiation of NRs could be used as a post-surgical complement to current tumor-excision techniques to decrease the likelihood of recurrence. Furthermore, employing NRs as opposed to spherical gold nanoparticles shifts the relevant wavelengths into the NIR range, allowing light to penetrate far deeper into tissue compared with visible wavelengths, which in principle can provide noninvasive activation of NRs for treatment of currently inoperable tumors. Since therapeutic efficacy was observed even at the lowest concentration tested (62.5 pM antibody, 0.21 pM NR), we therefore hypothesize that systemic in vivo administration of NRs could provide a sufficient dose to tumor sites to observe a light-induced therapeutic response. Alternatively, employing direct administration to tumor sites immediately postsurgery would increase the NR concentration and therefore the photothermal potential of the treatment. Future in vivo studies should help to identify the optimum protocol for these NR treatments, which may be different for pathologically distinct situations.

Finally, we note that catecholamine-based coatings can be integrated with many other organic and inorganic functional nanomaterials, and can then be functionalized with a broad range of biotargeting and therapeutic organic molecules to form an array of agents for diagnosis and therapy of other diseases.

Executive summary.

Nanoparticle surface modification

• A conformal 5–10 nm-thick ‘primer’ layer of polydopamine (PD) was deposited onto surface plasmon resonant gold nanorods (NRs) to form a versatile interface for biofunctionalization.

• Polyethylene glycol polymers were covalently reacted to the biomimetic PD layer to passivate the surface.

• Antibodies were immobilized onto the PD-coated NRs to provide bioactivity, and the number of antibodies per nanoparticle was tuned between 8 and 350.

• To form anticancer nanoparticles, anti-EGF receptor (EGFR) antibodies were immobilized onto the PD-coated NR surface, and the functionalized nanoparticles were stable in serum-containing medium for 24 h.

Nanoparticle-targeted imaging & photothermal therapy

• Antibody-functionalized NRs were bound specifically to EGFR-overexpressing oral and breast cancer cells, which were detected in both bright-field microscopy and optical coherence tomography.

• Targeted NRs provided a strong synergistic therapeutic response with broadband light irradiation, causing significant death to EGFR-expressing cancer cells in vitro.

Future perspective

• The versatile PD surface modification offers attractive properties to functionalize nanoparticles composed of many materials with a broad range of bioactive molecules such as antibodies, peptides and DNA aptamers for multifunctional cancer treatment.

• PD-enabled nanoparticles have the adaptable chemical repertoire to form versatile, multifunctional diagnostic and therapeutic agents for a broad range of other diseases.