Accepted Manuscript Title: Colloidal Stability of Gold Nanorod solution Upon Exposure to Excised Human Skin: Effect of Surface Chemistry and Protein Adsorption Author: Nouf N. Mahmoud Khaled M. Al-Qaoud Amal G. Al-Bakri Alaaldin M. Alkilany Enam A. Khalil PII: DOI: Reference:

S1357-2725(16)30044-9 http://dx.doi.org/doi:10.1016/j.biocel.2016.02.020 BC 4805

To appear in:

The International Journal of Biochemistry & Cell Biology

Received date: Revised date: Accepted date:

30-1-2016 19-2-2016 23-2-2016

Please cite this article as: Mahmoud, Nouf N., Al-Qaoud, Khaled M., Al-Bakri, Amal G., Alkilany, Alaaldin M., & Khalil, Enam A., Colloidal Stability of Gold Nanorod solution Upon Exposure to Excised Human Skin: Effect of Surface Chemistry and Protein Adsorption.International Journal of Biochemistry and Cell Biology http://dx.doi.org/10.1016/j.biocel.2016.02.020 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Colloidal Stability of Gold Nanorod solution Upon Exposure to Excised Human Skin: Effect of Surface Chemistry and Protein Adsorption Nouf N. Mahmoud1, Khaled M. Al-Qaoud2, Amal G. Al-Bakri1, Alaaldin M. Alkilany1*, Enam A. Khalil1

1

Department of Pharmaceutics & Pharmaceutical Technology, Faculty of Pharmacy, the University

of Jordan, Amman 11942, Jordan 2

Department of biological sciences, Yarmouk University, Irbid, Jordan

*

Corresponding Author: E-mail: [email protected]

1

Abstract In this study, we evaluated the colloidal stability of gold nanorods (with positive, negative and neutral surface charge) in solution upon contact with excised human skin. UV-vis absorption, plasmon peak broadening index (PPBI%) and transmission electron microscope analysis were used to follow nanoparticles aggregation in solution. Our results show that positively charged gold nanorods aggregate extensively upon exposure to excised human skin compared to negatively and neutrally charged gold nanorods. Skin-induced aggregation of cationic gold nanorods was linked to the adsorption of proteins released from the dermis layer to the surface of gold nanorods. Protein adsorption significantly screen nanorod’s effective surface charge and induce their aggregation. Moreover, we demonstrate that the presence of polyethylene glycol polymer on the surface of cationic gold nanorods minimize this aggregation significantly by providing steric repulsion (nonelectrostatic stabilization mechanism). This work highlights the importance of evaluating the colloidal stability of nanoparticles in solution upon contact with skin, which is a “usually overlooked” parameter when studying the nanoparticle-skin interaction.

Keywords Human skin; stratum corneum; epidermis; dermis; nanoparticles; gold nanorods; colloidal stability

2

Introduction Understanding the effect of shape, charge, and surface chemistry of nanoparticles on their penetration into skin and hair follicles is crucial to engineer efficient nanotherapeutics and has been the target of several recent studies. However, the colloidal stability of nanoparticle solution upon exposure to skin is usually ignored in many published reports evaluating the nano-skin interface (Table S1, Supplementary data). This overlooked parameter contributes to the current dilemma and large debate in the reported findings regarding nanoparticles penetration and distribution into skin layers. For example, a recent study conducted by Fernandes et al. (2015) has shown that positively charged gold nanorods (GNR) penetrate the skin in larger quantity in comparison to their negatively charged counterparts. These results are contrary to those drawn by Lee et al. (2013) where they indicated that skin penetration is greater for anionic GNR compared to cationic counterparts. The colloidal stability of nanoparticle solutions exposed to skin in the above two studies, as well as for the majority of published studies, was not investigated. Clearly, a systematic evaluation of the colloidal stability of nanoparticles solution upon contact to human skin is a clear need and will enrich our understanding to the nano-skin interface.

Upon contact with biological media/compartments, nanoparticles adopt new surface properties due to the formation of “protein corona” resulting in a dramatic change in their colloidal stability and biological behavior (Mahmoudi et al., 2011; Zanganeh et al., 2016). Colloidal stability of nanoparticles in various biological media is reported including: various types of cell culture media, alveolar and lysosomal fluid, synthetic lung fluid, blood, serum and plasma (Urban et al., 2016). Similarly, the physiochemical properties of nanoparticles are expected to change significantly upon exposure to skin considering its anatomical complexity and the possible interaction of nanoparticles with the diverse population of released proteins and other compounds from the exposed skin (Labouta et al., 2011; Labouta and Schneider, 2013). Recently, surfactant proteins were found in epidermis, dermis, hair follicles, sweat and sebum (Kankavi, 2006). The transcriptome analysis shows that 63% of all human proteins are expressed in the skin (Waldera-Lupa et al., 2014).

In addition to their unique optical properties and promising applications, gold nanoparticle (GNPs) are frequently used as a “model nanoparticle” to understand the nano-bio interface due to the ease of their quantification, visualization in complex biological compartments and following their

3

aggregation (Alkilany et al., 2013; Jain et al., 2008; Dykman and Khlebtsov, 2012). For example, we had reported a significant change in GNR’s physiochemical properties upon exposure to cell culture media due to adsorption of serum albumin protein (Alkilany et al., 2009). Herein, we employ GNR with different effective surface charge (cationic, anionic and neutral) to evaluate the colloidal stability of nanoparticle solutions upon exposure to human skin. We also show that biomolecules (like proteins) that are released from the deeper skin layer (dermis) adsorb onto the surface of cationic GNR and play a critical role in defining the colloidal stability of nanoparticles upon contact with human skin. Finally we show that pegylation of cationic GNR prevent the nanoparticle aggregation by providing steric repulsion (non-electrostatic stabilization mechanism).

Materials and methods

1.

Materials and Instrumentation:

Chloroauric acid (HAuCl4.3H2O, 99.9%), Sodium borohydride (NaBH4, 99%), Silver nitrate (AgNO3, 99%), Ascorbic acid (99%), Poly(allylamine hydrochloride), (PAH, MW ~15000 g/mole), Poly (acrylic acid, sodium salt) (PAA, MW ~15000 g/mole), methoxypolyethylene glycol thiol (mPEG-SH, MW~2000 Da), Cetyltrimethylammonium bromide (CTAB, 99%), Bovine serum albumin, (BSA), Trypsin from porcine pancreas (lyophilized powder, 1000-2000 BAEE units/mg solid), and

Bradford reagent were obtained from Sigma Aldrich Chemicals. Cystamine

dihydrochloride (97%) was obtained from Acros, UK. Sodium dodecyl sulfate (SDS) was obtained from (S.D.Fine-Chem Ltd) and 1,4-dithiothreitol (DTT) was obtained from Biochemika. Filter membrane (Spectra/Pro cellulose membrane, MWCO 6-8,000 Da, Spectrum laboratories Inc, USA) was used to filter the conditioned solution. Franz diffusion cells (FDC) with 1.0 cm² exposed area and flat ground joint clear glass (PermeGear Inc., Hellertown, PA, USA) was used to perform stability experiments conducted on skin. All solutions were prepared with purified 18 MΩ water and all glassware were cleaned with aqua regia and rinsed with purified 18 MΩ water before use. UV-vis spectra of GNR were collected using UV-vis spectrophotometer (Spectrascan 80D, Biotech Eng., UK) over the range from 400 to 1100 nm. Transmission electron microscopy (TEM) images were obtained using (Versa 3D, FEI, Holland) operating at 30 kV. TEM grids were prepared for imaging by drying ~10 μL of diluted purified GNR solution on Formvar coated copper TEM grids (300 mesh, Ted Pella Inc., Redding, CA). Uranyl acetate stain (Fluka AG, Chem.) was used to stain 4

PAH-GNR-protein sample for TEM imaging. Dynamic light scattering (DLS) and zeta potential analysis was performed on Microtraczetatrac (USA).

2.

Methods

2.1. Synthesis of CTAB capped GNR (CTAB-GNR) CTAB-GNR were synthesized according to the seed-mediated surfactant-assisted wet-chemical method (Sau and Murphy, 2004). For seed synthesis, a solution of 0.25 mM HAuCl4 was prepared in 0.1M CTAB. NaBH4 (0.6 mL, 10 mM) was added to 10 ml of gold-CTAB solution with stirring. For GNR synthesis, the following were added to the CTAB aqueous solution (95 mL, 0.1 M) with gentle mixing: silver nitrate solution (1.1 mL, 10 mM) to get GNR with aspect ratio of around 4, HAuCl4 solution (5 mL, 10 mM), ascorbic acid solution (0.55 mL, 0.1 M) and finally seed solution (0.12 mL). GNR mixture left undisturbed at 25°C overnight. To get rid of the excess CTAB, the GNR solution was centrifuged twice for 15 min at 12 000 rpm. The pellet was re-suspended in purified 18 MΩ water.

2.2. PAA-coating of CTAB-GNR To each 1.0 mL of twice purified CTAB-GNR solution, 0.2 mL of PAA solution (10 mg/mL prepared in 10 mM NaCl solution) and 0.1 mL NaCl solution (10 mM) were added simultaneously. The solution was mixed and left for 30 minutes (Gole and Murphy, 2005). To get rid of the excess PAA polymer after coating, the coated GNR solution was centrifuged for 10 min at 10 000 rpm. The pellet was re-suspended in purified 18 MΩ water.

2.3. PAH-coating of PAA-GNR To each 1.0 mL of purified PAA-GNR solution, 0.2 mL of PAH solution (10 mg/mL prepared in 10 mM NaCl solution) and 0.1 mL NaCl solution (10 mM) were added simultaneously. The solution was mixed and left for 30 minutes (Gole and Murphy, 2005). To get rid of the excess PAH polymer, the coated GNR solution was centrifuged for 10 min at 10 000 rpm. The pellet was resuspended in purified 18 MΩ water.

5

2.4. Pegylation of GNR (PEG-GNR) Self-assembly of thiolated PEG was employed to displace CTAB with PEG-SH. To each 1.0 mL of twice cleaned CTAB-GNR solution, 0.1 mL of PEG-SH solution (10 mg/mL) was added and mixed for 12 hours. To get rid of the excess PEG-SH polymer after coating, the coated GNR solution was centrifuged twice for 10 min at 10 000 rpm. The pellet was re-suspended in purified 18 MΩ water.

2.5. Preparation of PEG-Cystamine-GNR (PEG-Cys-GNR) Assembly of mixed thiol monolayers on GNR using two ligands (thiolated PEG and cystamine) was employed to displace CTAB and thus to prepare PEG-Cys-GNR. To each 1.0 mL of twice cleaned CTAB-GNR solution, 0.1 mL of PEG-SH solution (5mg/mL) were added and mixed for 6 hours. To get rid of the excess PEG polymer after coating, the coated GNR solution was centrifuged for 10 min at 10 000 rpm. The pellet was re-suspended in DI water. To each 10 mL of purified PEG-GNR solutions, 1.0 mL of cystamine dihydrochloride (30 mM) was added and mixed overnight. To get rid of the excess cystamine dihydrochloride after coating, the coated GNR solution was centrifuged for 10 min at 10 000 rpm. The pellet was re-suspended in purified 18 MΩ water.

2.6. Human skin preparation Human skin was obtained from female patients aged 34-40 years, who had undergone abdominal plastic surgery. The approval of the donors was granted after we explained the purposes of experiment to them. The skin was cut into discs and the subcutaneous fatty tissue was removed, then the skin cleaned by tapping with dry wipes and placed on paperboard that was wrapped with aluminum foil so that the skin surface is facing upward. Then the specimens were stored in polyethylene bags at -80C and used within 6 months. For the experiments, discs of proper diameters were punched out from the frozen skin sheets and left for few minutes to thaw at room temperature and treated according to the experimental details below.

2.7. Assessment of colloidal stability of GNR solution upon exposure to excised human skin. Four excised human skin discs were mounted onto four Franz diffusion cells (FDC, exposed area =1.0 cm²), the exposed area with the top layer of the skin facing the donor chamber. FDC was used to define the exposed area. The donor champers of the FDC received (300 µL, 3.0 nM) of each

6

PAA-GNR, PAH-GNR, PEG-GNR and PEG-Cys-GNR solutions followed by incubation at 37°C under occlusive conditions. Samples of GNR solutions were taken after 3, 6 and 24 hours of exposure UV-vis spectra were measured and plasmon peak broadening index, PPBI% (Alkilany et al., 2014), was calculated. Zeta potentials for GNR solutions were measured.

2.8. Assessment of colloidal stability of PAH-GNR and PEG-Cys-GNR solution upon exposure to conditioned solution. Three skin discs were mounted on FDC (exposed area =1.0 cm²), 500 µL of purified 18 MΩ water were added to each and left for 6, 12 and 24 hours. At each time point, 300 µL of the incubated “conditioned” solution were mixed with PAH-GNR or PEG-Cys-GNR solutions (150 µL, 3.0 nM) for 6 hours. Purified 18 MΩ water was incubated at 37C in the absence of skin and used as negative control. Aggregation was followed by obtained UV-vis spectra, calculated PPBI%, particle size analysis and zeta potentials measurements.

2.9. Assessment of colloidal stability of PAH-GNR and PEG-Cys-GNR solution upon exposure to isolated human skin layers. Stratum corneum (SC) was separated from skin by incubating whole skin disc (diameter = 8.0 mm) with 1% trypsin for 48 hours. The isolated SC was washed thoroughly with purified 18 MΩ water to remove excess trypsin. The epidermis including the SC was separated from the skin by the heat separation method (Kassis and Søndergaard, 1982). The whole skin was immersed for one minute in water bath at 60°C, followed by peeling the epidermal layer from the dermis. The isolated skin layers and the whole skin were incubated with PAH-GNR or PEG-Cys-GNR solutions (500 µL, 3.0 nM) in glass vials at 37 °C for 24 hours. UV-vis spectra were measured after 1, 3, and 24 hours of exposure and PPBI% was calculated.

2.10. Quantification of proteins in the conditioned solution and adsorbed proteins on PAH-GNR. To three skin discs mounted on FDC (exposed area =1.0 cm²), 500 µL of purified 18 MΩ water were added to each and incubated for 6, 12 and 24 hours. At each time point, 1.5 mL of Bradford reagent was added to 50 µL of the conditioned solution. UV-vis absorbance at 595 nm was measured and protein concentrations were calculated from a fresh BSA calibration curve using the reagent manufacturer’s instructions. The amount of protein adsorbed on PAH-GNR was quantified

7

using Bradford assay by calculating the difference between amount of protein in the 24 hoursconditioned solutions and amount of unbound protein in the supernatant after centrifuging the mixed conditioned solution with PAH-GNR.

2.11. Analysis of proteins in conditioned solution and adsorbed proteins on PAH-GNR using SDSpolyacrylamide gel electrophoresis (SDS-PAGE). PAH-GNR were mixed with 24 hour conditioned solution for 6 hours and then purified from unbound protein by repeated centrifugation. Bound proteins were striped by incubating PAH-GNR pellets in 2% (w/v) SDS and 50 mM dithiothreitol for 5 minutes at 95°C (Docter et al., 2014). 20 μL of the supernatant containing striped protein and 20 μL of the conditioned solution before mixing with PAH-GNR were loaded on the SDS-PAGE. For preparing resolving gels; 4 mL distilled water, 2.5 mL running buffer (pH 8.8), 3.3 mL of 30% acrylamide-bisacrylamide solution, 100 μL of 10% ammonium persulfate (APS) and 10 μL TEMED were mixed. Stacking gels (4%) were prepared by adding 6 ml of distilled water, 2.5 ml staking buffer (pH 6.6), 1.3 mL of 30% acrylamide-bisacrylamide solution, 100 μl of 10% APS and 10 μl TEMED. Samples were mixed with an equal volume of sample buffer have Bmercaptoethanol (pH 6.8). For band size determination, prestained molecular weight protein standard (Wide Range BLUeye prestained protein Ladder, GeneDireX, Taiwan) was used. Electrophoresis was carried out using running buffer with pH 8.3 at 100 volts for 30min. then at 130 volts for about 100 minutes. The gel was stained using silver stain according to manufacturer protocol (Biorad, CA. USA)

2.12. Statistics: Statistical analysis was performed by applying the unpaired t-test using GraphPad Prism version 6.0 (San Diego, CA). Results were considered significant when p < 0.01.

Results and discussion

Synthesis, characterization and colloidal stability of PAA-GNR, PAH-GNR and PEG-GNR solutions upon exposure to excised human skin

8

Adopting available protocols, CATB-GNR with aspect ratios (length/width) ~ 4 were prepared by wet chemical method where the cationic surfactant (CTAB) was used as a shape directing agent. UV-vis spectra of CTAB-GNR show typical longitudinal plasmon peak, which indicates no aggregation (Figure 1A). TEM images further confirm the shape and size uniformity of the prepared nanorods (Figure 1B). The use of CTAB-GNR was avoided in this study due to their welldocumented toxicity. Layer-by-layer (LBL) coating was employed to wrap CTAB-GNR with anionic polyelectrolytes resulting in PAA-GNR, which were further, coated with the cationic polyelectrolytes resulting in PAH-GNR. Pegylated GNR were prepared by CTAB displacement on the surface of CTAB-GNR with PEG-SH as per available protocols (Alkilany et al., 2014). Typical UV-vis spectra and constant solution color confirm the absence of aggregation associated with the surface modification reactions (Figure 1A). Moreover, surface effective charge of prepared nanorods was confirmed by zeta potential analysis indicating successful surface functionalization (cationic CTAB-GNR; anionic PAA-GNR; cationic PAH-GNR; neutral PEG-GNR; Figure 1C). GNR provide an excellent “nanoparticle model” due to the ease and sensitivity in following their aggregation by simple color change and spectrophotometric analysis (UV-vis spectra). In this work, the broadening of plasmon peaks was quantified using the plasmon peak broadening index, PPBI%, as we reported recently (Alkilany et al., 2014). UV-vis spectra of GNR were fitted (absorption range from 600 to 1100 nm) to Gaussian distribution curves followed by obtaining the standard deviation (SD) of the fitted curves as an indicator of broadening. PPBI% was calculated using the following equation: PPBI% = SD0/SDn ×100%, where (SD0) and (SDn) are the standard deviation values of the fitted Gaussian distribution curves of the UV-vis spectra for GNR before (0) or after exposure for (n) hours to skin.

The colloidal stability of GNR upon exposure to human skin discs was evaluated using Franz diffusion cell (FDC), which ensures a fixed and reproducible surface area of exposed skin as presented in (Figure 1D). FDC is considered the “gold standard” setup in skin research for the determination of the parameters of drug release and penetration and to understand the interaction of various types of nanomaterials with skin (Franz, 1975; Labouta et al., 2011; Labouta and Schneider, 2013). Simply, excised human skin discs were mounted onto FDC (exposed area =1.0 cm²) with the top layer of the skin (SC) facing the donor chamber containing the GNR solution

9

(Figure 1D). We would like to highlight that our experimental setup, measures the aggregation of gold nanorods in their solution upon exposure to human skin discs and does not measure aggregation on the skin itself. Anionic-GNR (PAA-GNR) showed excellent colloidal stability upon exposure to skin over 24 hours as evident from typical absorption spectrum and low PPBI% (Figures 2A & 2E, respectively). However, absorption spectra of cationic PAH-GNR show peak broadening after 6 hours of skin exposure (PPBI%~160) indicating nanorods aggregation. The peak broadening of the cationic PAH-GNR increased significantly after 24 hours (PPBI%~1290) of exposure and the aggregated GNR precipitated out from the aqueous solution on the surface of the skin (Figures 2B & 2E). TEM images of PAH-GNR exposed to skin confirmed the aggregation of PAH-GNR (Figure 3A & 3B).

Biomolecules responsible for PAH-GNR aggregation upon exposure to excised human skin Upon contact with biological compartments, nanoparticles adapt new surfaces (corona) due to the adsorption of proteins and other moieties to their surfaces changing the nanoparticle’s physiochemical properties and may induce nanoparticle stabilization or aggregation (Maiolo et al., 2015; Monopoli et al., 2011). Protein adsorption to nanoparticle depends on several factors such as type, size, shape and surface chemistry of nanoparticles (Mahmoudi and Serpooshan, 2011; Nel et al., 2009; Saptarshi et al., 2013). Studies investigated the adsorption of proteins or other secreted molecules from skin on nanoparticles are extremely rare.

Interestingly, zeta potential of PAH-GNR significantly decreased from ~+60 to ~+5.5 mV after 6 hours of exposure (Figure 2F). We may postulate that negatively charged components or biomolecules, which are most likely released proteins from the skin, adsorb on the surface of cationic GNR resulting in a significant decrease in their effective surface charge and induce nanoparticle aggregation. However, the PAH-GNR surface charge did not flip all the way to a negative value suggesting that amount of negative charged skin biomolecules is not enough to form a complete corona around PAH-GNR and thus to flip their surface charge. To confirm the release of proteins from skin, we injected a blank solution (water) to the donor compartment over a clean skin discs. This solution (referred as conditioned solution afterward in this manuscript) was assayed for protein content using Bradford test. The protein concentration had slightly increased after 12 hours (~0.07 mg/cm²) and maintained constant after 24 hours of exposure (Figure S1,

10

Supplementary data). Interestingly, we found only one report measuring the amount of protein released from skin (including albumin, beta globulin and gamma globulin) as 0.06 mg/cm², which is in excellent agreement with our quantitation (Rosenthal et al., 1957). To further confirm protein release from skin, we analyzed the conditioned solution using silver-stained SDS-PAGE, which confirmed the presence of proteins and showed a visible abundance of high molecular weight (~68 kDa) band which is related to albumin in addition to other less intense bands (Figure 3C).

In order to test our hypothesis that the released proteins from skin are responsible for the observed aggregation of PAH-GNR, we mixed PAH-GNR with conditioned solution. Interestingly, addition of the conditioned solution resulted in PAH-GNR aggregation but to a lower extent compared to the aggregation upon their direct exposure to skin on the FDC setup (PPBI~ 200% vs ~1290% after 24 hours, respectively; Figure 4A & 4B). The observed aggregation upon addition of condition solution was associated with a significant decrease in effective surface charge of the PAH-GNR (from ~+60mV to ~+10Mv, Figure 4C). In addition, the protein-PAH-GNR interaction increased the hydrodynamic size of the PAH-GNR which is most likely related to aggregate particles and protein adsorbed on the PAH-GNR surface (Figure S2, Supplementary data). The amounts of adsorbed protein on the PAH-GNR surface (or associated with the nanoparticles) were quantified using Bradford assay (0.000094 picogram/nanorod). Moreover, protein adsorption on PAH-GNR was further tested by SDS-PAGE analysis of the striped proteins. Only one intense high molecular weight (~68kDa) band related to albumin protein was observed (other proteins separated in conditioned media were either not adsorbed or not striped from PAH-GNR). Assuming only albumin, we calculate that quantity of protein that adsorb to PAH-GNR as ∼860 albumin protein/ nanorod (Figure, Supplementary data). A recent work investigating the adsorption process of bovine serum albumin (BSA) on GNR reported that ∼1000−3000 BSA molecule bind to each gold nanorod (Boulos et al., 2013).

While these results support our hypothesis that proteins are released from skin and screen the effective surface charge on cationic GNR and thus induce nanoparticle aggregation upon exposure to human skin, it further highlight that direct skin contact increases the rate and extent of aggregation significantly. In agreement to our results, Rancan et al., (2012) reported that coating of silica nanoparticle with positively charged groups enhanced the tendency of nanoparticles to

11

aggregate, resulting in lower internalization by in vitro cultured primary skin cell line. Albanese et al. (2014) reported that secreted metabolites and proteins by cultured cells in vitro induce nanoparticle aggregation in culture media. However, we cannot exclude other hypothesis stating that low-molecular weight components (such as salt, amino acids, sugars and lipids) may contribute to the observed aggregation of PAH-GNR. To test the later hypothesis, we filtered the conditioned solution (24 hours incubated) thorough 8000 Da membrane to separate low-molecular weight components from larger biomolecules (such as proteins). Interestingly, PAH-GNR did not aggregate upon mixing with filtrate solutions (Figure S4, Supplementary data). To test whether depletion of aggregation-inducing proteins can prevent the observed nanoparticle aggregation, we pre-incubated skin with blank water for 6 or 24 hours followed by discharging the water and replace it with PAH-GNR solution. Interestingly, the aggregation of PAH-GNR showed similar rate and extent upon exposure to either pre-incubated or “as is” skin discs (Figure 2E). These results indicate that the release of the aggregation-inducing biomolecules is a continuous process and cannot be prevented by simple pre-cleaning or pre-exposure.

Synthesis and characterization of PEG-Cys-GNR and their colloidal stability upon exposure to excised human skin In order to prepare cationic GNR with superior colloidal stability compared to the aggregating PAH-GNR, we evaluated the displacement of CTAB on GNR with two moieties: 1) thiolated cationic ligand (cystamine) to provide effective positive surface charge; 2) thiolated polyethylene glycol (PEG-SH) to maintain colloidal stability of GNR via steric repulsion (non-electrostatic mechanism). Sequential ligand exchange strategy starting by displacement of CTAB with PEGSH followed by the addition of cystamine was developed to prepare PEG-Cys-GNR (Figure 5A). UV-vis spectra and solution color and zeta potential analysis confirmed successful surface functionalization (Figure 5B & 5C). PEG-Cys-GNR showed high positive effective surface charge (zeta potential~ +55 mV; Figure 5C) with exceptional colloidal stability. PEG-Cys-GNR did not aggregate when exposed to skin up to 6 hours and only slightly aggregated after 24 hours as evident from UV-vis spectra and spectral broadening analysis (PPBI% ~170 for PEG-Cys-GNR compared to ~1290 for PAH-GNR; Figure 2D & 2E). Unlike PAH-GNR, PEG-Cys-GNR did not aggregate when mixed with conditioned solution (Figure 4). It is worth mentioning that the colloidal stability of cationic PEG-Cys-GNR

12

upon exposure to skin is comparable to the stability of the nonionic PEG-GNR (Figure 2C & 2E) indicating that the excellent colloidal stability of PEG-Cys-GNR originates from the outstanding stabilization effect of PEG (Gao et al., 2012). However, The surface charge of PEG-Cys-GNR was also decreased after 6 hours of skin exposure (from ~+55 to ~+30 mV) but to a less extent compared to PAH-GNR (from ~+60 to ~+5.5 mV; Figure 2F), which may suggest that PEG-CysGNR also interact with the skin protein but maintain excellent stability due to the presence of PEG shell that stabilized the particles by steric repulsion (electrostatic-independent mechanism) (Karakoti et al. 2011).

Stability of PAH-GNR and PEG-Cys-GNR upon exposure to isolated layers of human skin. Considering the complex anatomy of human skin, we assumed that the outermost skin layer, SC, is the responsible layer for the observed skin-induced aggregation of PAH-GNR. To test this hypothesis, we evaluated the colloidal stability of PAH-GNR exposed to isolated skin layers (whole skin, SC, epidermis and dermis). Surprisingly, PAH-GNR incubated with SC or epidermis was stable over 24 hours and there was no detected peak broadening (Figure 6A). However, PAH-GNR incubated with whole skin or dermis show peak broadening after 1 hour of exposure and they extensively aggregated and precipitated out from solution after 24 hours of skin exposure (Figure 6A & 6C). Interestingly, all PEG-Cys-GNR incubated with different skin layers did not aggregate over 24 hours of exposure to isolated skin layers (Figure 6B & 6C). However, those incubated with dermis show slight peak broadening after 24 hours (PPBI~162%; Figure 6B & 6C) (Figure S5 in the Supplementary data shows UV-vis spectra of PAH-GNR and PEG-Cys-GNR after 1 and 3 hours of skin exposure). Collectively, our results may suggest that the aggregation of cationic PAHGNR is not a result of a direct interaction between nanoparticles and “bound” components on the surface of the skin (SC layer). Indeed, we belief that deeper skin layers are most likely the responsible compartments for the observed aggregation. Moreover, this hypothesis is supported by the aggregation kinetics where 6 hours of exposure to skin was needed to induce PAH-GNR aggregation when exposed to skin.

Conclusion This study demonstrates that colloidal stability of different surface charge have different tendency to aggregate when interacting with skin. Anionic (PAA-GNR) and pegylated (PEG-GNR) nanorods

13

solution exhibited excellent stability over 24 hours upon exposure to human skin. Cationic nanorods (PAH-GNR) solution demonstrates high extent and very fast rate of aggregation upon exposure to human skin. However, cationic PEG-Cys-GNR solution showed good stability compared to cationic PAH-GNR solution due to the presence of the PEG shell that provide steric repulsion (non-electrostatic stabilization mechanism). Interestingly, the deeper (dermis) but not the outermost (stratum corneum) skin layer is responsible for the observed aggregation. Our results suggest that biomolecules such as proteins may leach out from skin, adsorb on the surface of the cationic-GNR, neutralize their surface charge and induce nanoparticle aggregation. However, direct contact of GNR with skin significantly increased the rate and extent of aggregation. Since nanoparticles size is well known to control the cellular uptake and skin penetration, our group currently is conducting a bio-study to investigate the skin penetration of PAH-and PEG-Cys-GNR, which will be reported in a separate contribution. Our preliminary results indicate a dramatic effect of nanoparticle aggregation on their skin penetration parameters. We recommend that the colloidal stability of nanomaterial upon exposure to skin should be carefully evaluated to prevent data misinterpretation and minimize conflicts in reported results. A systematic evaluation of nanoparticle stability upon exposure to skin will also enrich our understanding to the nano-skin interface toward engineering effective nanomedicine for skin.

Acknowledgements The authors thank the Scientific Research Deanship (SRD) at the University of Jordan for financial support (grants number: 11/2012-2013 and 152/2014-2015).

Supplementary data Table showing some recent studies evaluating the penetration and diffusion of GNPs into skin and the availability of evaluation to their colloidal stability; protein concentrations in conditioned solution and estimation of amount of proteins adsorbed on gold nanorods; stability of PAH-GNR mixed with filtrates of conditioned solution and UV-vis absorption spectra of PAH-GNR and PEGCys-GNR exposed to isolated skin layers are available as supplementary data.

14

References Albanese A., Walkey CD., Olsen JB., Guo H., Emili A., Chan WC., 2014. Secreted Biomolecules Alter the Biological Identity and Cellular Interactions of Nanoparticles. ACS Nano. 8, 5515-26. Alkilany AM., Abulateefeh SR., Mills KK., Yaseen AI., Hamaly MA., Alkhatib HS., Aiedeh KM., Stone JW., 2014. Colloidal Stability of Citrate and Mercaptoacetic Acid Capped Gold Nanoparticles upon Lyophilization: Effect of Capping Ligand Attachment and Type of Cryoprotectants. Langmuir. 30, 13799-808. Alkilany AM., Lohse SE., Murphy CJ., 2013. The Gold standard: Gold Nanoparticle Libraries to Understand the Nano-bio Interface. Acc. Chem. Res. 46, 650-61. Alkilany AM., Nagaria PK., Hexel CR., Shaw TJ., Murphy CJ., Wyatt MD., 2009. Cellular Uptake and Cytotoxicity of Gold Nanorods: Molecular Origin of Cytotoxicity and Surface Effects. Small. 5, 701-8. Boulos SP., Davis TA., Yang JA., Lohse SE., Alkilany AM., Holland LA., Murphy CJ.,2013. Nanoparticle-protein interactions: a thermodynamic and kinetic study of the adsorption of bovine serum albumin to gold nanoparticle surfaces. Langmuir. 29, 14984-96. Docter D., Distler U., Storck W., Kuharev J., Wünsch D., Hahlbrock A., Knauer SK., Tenzer S., Stauber RH., 2014. Quantitative profiling of the protein coronas that form around nanoparticles. 9, 2030-44. Dykman L. and Khlebtsov N., 2012. Gold Nanoparticles in Biomedical Applications: Recent Advances and Perspectives. Chem. Soc. Rev. 41, 2256-82. Fernandes R., Smyth NR., Muskens OL., Nitti S., Heuer-Jungemann A., Ardern-Jones MR., Kanaras AG., 2015. Interactions of Skin with Gold Nanoparticles of Different Surface Charge, Shape, and Functionality. Small. 11, 713-21. Filon FL, Crosera M, Adami G, Bovenzi M, Rossi F, Maina G., 2011. Human skin penetration of gold nanoparticles through intact and damaged skin. Nanotoxicology. 5, 493-501. Franz TJ., 1975. Percutaneous absorption on the relevance of in vitro data. J Invest Dermatol. 64, 190-5. Gao J., Huang X., Liu H., Zan F., Ren J., 2012. Colloidal Stability of Gold Nanoparticles Modified with Thiol Compounds: Bioconjugation and Application in Cancer Cell Imaging., 2012. Langmuir. 28, 4464-71. Gole A. and Murphy CJ., 2005. Polyelectrolyte-Coated Gold Nanorods:  Synthesis, Characterization and Immobilization. Chem. Mater. 17, 1325-30.

15

Graf C., Meinke M., Gao Q., Hadam S., Raabe J., Sterry W., Blume-Peytavi U., Lademann J., Rühl E., Vogt A., 2009. Qualitative Detection of Single Submicron and Nanoparticles in Human Skin by Scanning Transmission X-ray Microscopy. J. Biomed. Opt. 14, 021015. Huang Y., Yu F., Park YS., Wang J., Shin MC., Chung HS., Yang VC., 2010. Co-Administration of Protein Drugs with Gold Nanoparticles to Enable Percutaneous Delivery. Biomaterials. 31, 9086-91. Jain PK., Huang X, El-Sayed IH, El-Sayed MA., 2008. Noble Metals on the Nanoscale: Optical and Photothermal Properties and Some Applications in Imaging, Sensing, Biology, and Medicine. Acc. Chem. Res. 41, 1578-86. Kanavi, O., 2006. Human Skin Is As An Extra-Pulmonary Site Of Secretion Of Surfactant Proteins. Erciyes Medical Journal. 28, 82-91 Karakoti AS., Das S., Thevuthasan S., Seal S., 2011. PEGylated Inorganic Nanoparticles. Angew. Chem. Int. Ed. Engl. 50, 1980-94. Kassis V., Søndergaard J., 1982. Heat-Separation of Normal Human Skin for Epidermal and Dermal Prostaglandin Analysis. Arch. Dermatol. Res. 273, 301-6. Labouta HI., Kraus T., El-Khordagui LK., Schneider M, 2011. Mechanism and Determinants of Nanoparticle Penetration Through Human Skin. Nanoscale. 3, 4989-99. Labouta HI., Schneider M., 2013. Interaction of inorganic nanoparticles with the skin barrier: current status and critical review. Nanomedicine. 9, 39-54. Lee O., Jeong SH., Shin WU., Lee G., Oh C., Son SW., 2013. Influence of Surface Charge of Gold Nanorods on Skin Penetration. Skin Res. Technol. 19, 390-6. Liu CD., Raphael AP., Sundh D., Grice JE, Soyer HP., Roberts MS., Prow TW., 2012. The Human Stratum Corneum Prevents Small Gold Nanoparticle Penetration and Their Potential Toxic Metabolic Consequences. Journal of Nanomaterials. Article ID 721706 (DOI: 10.1155/2012/721706). Mahmoudi M., Lynch I., Ejtehadi M.R., Monopoli M.P., Bombelli F.B., Laurent S., 2011. Protein−Nanoparticle Interactions: Opportunities and Challenges. Chem. Rev. 111, 5610-5637. Mahmoudi M., Serpooshan V., 2011. Large Protein Absorptions from Small Changes on the Surface of Nanoparticles. J. Phys. Chem. 115, 37. Maiolo D., Del Pino P., Metrangolo P., Parak WJ., Baldelli Bombelli F., 2015. Does Protein Corona Route to the Target or off Road?. Nanomedicine Delivery. 10, 3231-47.

16

Monopoli MP., Walczyk D., Campbell A., Elia G., Lynch I., Bombelli FB., Dawson KA., 2011. Physical-Chemical Aspects of Protein Corona: Relevance to in Vitro and in Vivo Biological Impacts of Nanoparticles. J. Am. Chem. Soc. 2, 2525-34. Nel AE., Mädler L., Velegol D., Xia T., Hoek EM., Somasundaran P., Klaessig F., Castranova V., Thompson M., 2009. Understanding Biophysicochemical Interactions at the Nano-bio Interface. Nat. Mater. 8, 543-57. Padamwar MN., Patole MS., Pokharkar VB., 2011. Chitosan-Reduced Gold Nanoparticles: a Novel Carrier for the Preparation of Spray-Dried Liposomes for Topical Delivery. J. Liposome Res. 21, 324-32. Rancan F., Gao Q., Graf C., Troppens S., Hadam S., Hackbarth S., Kembuan C., Blume-Peytavi U., Rühl E., Lademann J., Vogt A., 2012. Skin Penetration and Cellular Uptake of Amorphous Silica Nanoparticles with Variable Size, Surface Functionalization, and Colloidal Stability. ACS Nano. 6, 6829-42. Rosenthal SR., Samet C., Shkolnik S., Winzler RJ., 1957. Substances Released From the Skin Following Thermal Injury. I. Histamine and Proteins. J Clin Invest. 36, 38-43. Saptarshi SR., Duschl A., Lopata AL., 2013. Interaction of Nanoparticles with Proteins: Relation to Bio-Reactivity of the Nanoparticle. J. Nanobiotechnology. 11, 26. Sau TK., Murphy CJ., 2004. Seeded High Yield Synthesis of Short Au Nanorods in Aqueous Solution. Langmuir. 20, 6414-20. Sonavane G., Tomoda K., Sano A., Ohshima H., Terada H., Makino K., 2008. In Vitro Permeation of Gold Nanoparticles Through Rat Skin and Rat Intestine: Effect of Particle Size. Colloids Surf B Biointerfaces. 65, 1-10. Urban DA., Rodriguez-Lorenzo L., Balog S., Kinnear C., Rothen-Rutishauser B., Petri-Fink A., 2016. Plasmonic Nanoparticles and Their Characterization in Physiological Fluids. Colloids Surf B. 137, 39-49. Waldera-Lupa DM., Kalfalah F., Florea AM., Sass S., Kruse F., Rieder V., Tigges J., Fritsche E., Krutmann J., Busch H., Boerries M., Meyer HE., Boege F., Theis F., Reifenberger G., Stühler K., 2014. Proteome-wide analysis reveals an age-associated cellular phenotype of in situ aged human fibroblasts. Aging (Albany NY). 6, 856-78. Zanganeh S., Spitler R., Erfanzadeh M., Mahmoudi M., 2016. Protein Corona: Opportunities and Challenges. Int J Biochem Cell Biol. S1357-2725.

17

Figure 1. Characterization of CTAB-GNR, PAA-GNR, PAH-GNR and PEG-GNR solutions. A) UV-vis absorption spectra and optical graph of corresponding solutions as labeled. B) Effective surface charge for purified GNR solutions as labeled. C) TEM images (scale bar=50 nm) and dimensions (values represent mean±SD; n=150) of CTAB-GNR. D) Cartoon demonstrates experimental setup using FDC to evaluate GNR aggregation in solution and protein release from skin upon incubation of GNR solution with skin discs as indicated.

18

Figure 2. Characterization of GNR solutions after exposure to human skin. A) UV-vis absorption spectra of PAA-GNR solution after 3,6 and 24 hours of skin exposure. B) UV-vis absorption spectra of PAH-GNR solution after 3,6 and 24 hours of skin exposure. C) UV-vis absorption spectra of PEG-GNR solution after 3,6 and 24 hours of skin exposure. D) UV-vis absorption spectra of PEG-Cys-GNR solution after 3,6 and 24 hours of skin exposure. E) PPBI% of PAAGNR, PAH-GNR, PEG-GNR, PEG-Cys-GNR and PAH-GNR (pre-incubated skin) at different skin exposure times F) Effective surface charge of PAA-GNR, PAH-GNR, PEG-GNR and PEG-CysGNR solutions at different skin exposure times.

19

Figure 3. TEM images of PAH-GNR before (A) (scale bar=500) and after (B) exposure to human skin (scale bar=400nm). C. SDS-PAGE analysis of striped protein from the surface of PAH-GNR versus the condition solution in which nanorods were incubated.

20

Figure 4. Characterization of PAH-GNR and PEG-Cys-GNR exposed to conditioned solution. A) UV-vis absorption spectra of PAH-GNR and PEG-Cys-GNR after mixing with conditioned solution (conditioning time = 6,12 and 24 hours as labeled). PPBI% (B) and Effective surface charge (C) for PAH-GNR and PEG-Cys-GNR mixed with conditioned solution as function of conditioning time. Data in B&C are given as mean ± standard deviation; n=3 for each group; unpaired t-test was to evaluate the differences at 24 hours; ** represents p