Nanoparticles, due to their smaller sizes and associated unique properties, provide unprecedented opportunities to interrogate cellular and molecular processes with realistic clinical applications.^1–10 Specific types of nanoparticles are being utilized as drug delivery vehicles, cellular biomarkers, and cancer imaging and therapy agents.^1–10 The multifaceted applications of nanoparticles are the direct result of their ability to deliver high pay loads of drugs or biomarkers to the desired sites within the body.^1–10 Design and development of tumor specific nanoparticles could significantly amplify the delivering capacity to a specific target of interest, without affecting healthy cells.^1–10 The target specificity in nanoparticles could be imparted by tagging with certain biovectors, which navigate them to desired organ or site under in vivo conditions. The most commonly used target vectors are monoclonal antibodies and receptor-specific peptides.^11–13 Although, both biomolecules have shown high targeting abilities, the (in vivo) transport properties of monoclonal antibodies and peptides differ drastically. Monoclonal antibodies, due to their larger sizes show poor in vivo mobility resulting in time delayed, and reduced uptake over the desired target. Moreover, monoclonal antibodies are highly immunogenic, which lead to harmful side effects. In sharp contrast, peptides being smaller in size bring various advantages; namely, rapid blood clearance, ease in penetration of tumor vascular endothelium, increased diffusion rates in tissue, and low immunogenicity. Receptors for peptides are highly expressed on a variety of neoplastic and non-neoplastic cells.^14–15 Furthermore, receptor targeting peptides have shown high level of internalization within the tumor cells via receptor mediated endocytosis.¹⁶ The ability to internalize probes within tumor cells is important for delivering maximum pay loads to tumor cells.^14–16 These attractive physical properties coupled with their smaller size make peptides ideal candidates for developing new target specific nanoparticles. Therefore, we have designed and developed peptide conjugated nanoparticles that may circumvent some of the currently encountered problems.

Synthetic analogues of bombesin have been proven to show a similar binding affinity as the nascent bombesin peptide with GRP receptors. ^28–31 Based on our extensive previous studies, we have selected the peptide sequence shown in Scheme 1 for conjugation with gold nanorods. In the present study, GNRs of aspect ratio 3.1, capped with surfactant CTAB, have been employed. These GNRs show a strong absorption band at 725 nm due to longitudinal oscillations of electrons. The conventional method of surface conjugation of biomolecules to GNRs involved replacement of capping CTAB ligands with monodentate thiolated biomolecules to form Au-S bonds. Although the conjugation is very effective, it suffers a major drawback as the Au-S bond undergoes an irreversible oxidation and consequent elimination under biological conditions. Elimination of these surface bound ligands from nanoparticles leads to instant destabilization of these conjugates to form macro aggregates. In order to circumvent this problem, we have utilized a cyclic dithio-moiety, thioctic acid (TA), as a linker between GNR and bombesin peptide (Figure 1 and Scheme 1). This dithio-moiety chelates with gold atoms on the surface of GNRs leading to a more stable cyclic structure. Indeed, TA conjugated gold nanospheres (AuNPs) have shown extraordinary stability when compared with that of monothiol bound gold nanoconjugates.^35–44 For example, Abad et al. have used 3-nm TA-AuNPs as highly stable scaffolds for conjugation with Cobalt(II) complex for subsequent interaction with histidine-tagged protein;³⁵ and TA-AuNPs have also been reported as biological probe by attaching with electroluminescent luminol.³⁶ Mirkin and Graham have independently established the high stability attained by AuNPs upon conjugation with TA-oligonucleotides.^37,38 Recently, Li and coworkers have demonstrated the high in vitro and in vivo stabilities of AuNPs coated with TA-PEG than that of AuNPs conjugated with monothiol-PEG.⁴⁴ These literature examples suggest that TA can serve as ideal ligands for covalent binding of AuNPs with antibodies or peptides for in vivo imaging and therapy applications.⁴⁴ Another reason for the choice of disulfide as a starting material is based on the fact that, this moiety is stable under peptidic reaction conditions. In sharp contrast, free thiols on peptidic reaction conditions form disulfides to yield mixtures. This complication poses additional problems during the synthesis of thiols containing peptides. In this study, we have synthesized dithioctic acid conjugate of bombesin peptide using standard Fmoc methodology using solid phase support. We chose commercially available thioctic acid as a ligand to covalently link gold and bombesin peptide. Thioctic acid contains two functional groups, a disulfide and a carboxylic group. The carboxyl group of thioctic acid was attached with N-terminal of bombesin peptide via conventional solid-phase peptide synthesis to yield thioctic-bombesin conjugate (SS-BBN, Figure 1). The dithio-bombesin (SS-BBN) was purified by standard HPLC methods. The SS-BBN peptide is remarkably stable at 4 °C, and shows no detectable decomposition even after one year. The disulfide group in SS-BBN was utilized to conjugate with GNRs.

The in vitro stabilities of GNR-BBN conjugates were evaluated by treating aqueous solutions of nanoconjugates with various biologically relevant molecules for 12 hours, followed by monitoring the plasmon resonance band (Figure 3). The molecules that are commonly encountered under in vivo conditions include NaCl, HSA, cysteine, histidine, and biological buffer additives such as dithiothreitol (DTT). These molecules possess the ability to dislodge thiolated ligands from gold-thiol nanoconjugates. The removal of thiolated ligands would lead to aggregation of gold nanoparticles, resulting in broadening of the plasmon resonance band. Therefore, this study may provide greater understanding on the in vivo stability of the conjugates. The aqueous solutions of 2c were mixed with 0.9% NaCl, 0.5% cysteine, 0.2 M histidine, 0.5% HSA or 0.5% BSA solutions and stirred for 12 hours and the UV-Visible spectra of the resulting solutions were recorded (Figure 3(a)). It is important to recognize that the plasmon absorption bands of 2c did not show any change in peak width or shape (Figure 3).

The cellular localizations of GNR-BBN nanoconjugates in PC3 and T47D cells, which overexpress GRP receptors, were evaluated using dark field optical microscopy and transmission electron microscope image analysis. It is widely accepted that the internalization of nanoparticles strongly depends on their physical characteristics including size, shape, and charge. Recently, DeSimone and coworkers have demonstrated that nanoparticles with either larger size or negative zeta potential exhibit no cellular internalizations.⁴⁵ In that study, authors concluded that investigating the internalization of nanoparticles bearing negative zeta potential and conjugated with receptor-stimulating ligand would be interesting. In order to deliver negatively charged DNAs inside the cells, researchers have utilized ammonium ions as vectors; the ammonium cation interacts effectively with a negatively charged cell membrane, triggering charge mediated endocytosis. It is important to note here that GNR-BBN conjugate 2c has a negative ζ-potential of −32.0 mV. As a negative zeta potential candidate, 2c is expected to have very minimal or no interaction with negatively charged cell surface. In addition, 2c does not have any positively charged ions on the surface, thus possesses no ability to trigger charge mediated endocytosis. These results indirectly suggest that internalization of 2c would be possible only by means of specific internalization events. This behavior is also expected because the bombesin sequence chosen in the present study has agonist properties, which means they have specific receptor (in this case, GRP receptor) triggering characteristics to internalize within the cells. Our investigations focused on two important aspects and they are as follows: (i) location of 2c within the cellular matrices, and (ii) understanding the mechanism of internalization of conjugate 2c in GRP receptors expressing cancer cells.

The mechanism of either phagocytosis or receptor-mediated endocytosis is considered as the “holy grail” of internalization of nanoparticles within cells. The widely accepted internalization mechanisms for nanoparticles are via phagocytosis or receptor-mediated endocytosis. In order to evaluate the internalization mechanism, we have carried out a detailed examination of TEM images of cancer cells treated with 2c. TEM images, shown in Figure 6, strongly suggest that the internalization occurs via receptor-mediated endocytosis, and phagocytosis is not observed. It is important to note that the cellular membrane did not show the formation of phagocytic cups, which are usually present in the cell wall if the phagocytosis has occured. In addition, the conjugates of 2c were not found in endosomes, but were present in cytosol (see Figure 6). It is commonly observed that phagocytosed gold nanoparticles form a dense mass of aggregated nanoparticles, secluded from the rest of the cells in endosomal vesciles. Recently, Brust and coworkers reported their results on the interaction of citrate-stabilized nanoparticles and CALNN-coated nanoparticles with HeLa cells; they observed the formation of densely packed aggregated nanoparticles in endosomes.⁴⁶ Further, these studies have unequivocally demonstrated the delivery of nanoparticles conjugated with cell penetrating peptides selectively to cytosol without any uptake within the endosomes.⁴⁶ In sharp contrast, the PEG-ylated nanoparticles, which are devoid of any targeting vector, did not show any sign of internalization.⁴⁶ Therefore, we can infer that nanoparticles coated with receptor specific peptides can be endocytosed predominantly via non-phagocytosis pathway. The generalized view is that nanoparticles that are susceptible for interaction with serum proteins, such as citrate-stabilized nanoparticles tend to be prone for phagocytosis. On the other hand, if the nanoparticles are stable towards serum proteins (for example, PEG-ylated nanoparticles) which lack cell targeting capabilities, will not be internalized into the cells. In the light of these findings, it is expected that nanoconjugate 2c, which resists interaction with serum protein, should not undergo phagocytosis. However, the targeting vector bombesin in 2c is anticipated to stimulate GRP receptors in the cellular surface; thus, providing opportunity to penetrate into cells. In order to corroborate that the internalization of GNR-BBN conjugates in PC-3 and T-47D occurs via receptor mediated endocytosis pathway, our studies focused on understanding the interaction of GNR-BBN conjugates with cells which lack GRP receptors. For this experiment, we used NIH-3T3 cells, which do not express GRP receptors.⁴⁷ GNR-BBN conjugates 2c were incubated with NIH-3T3 cells and monitored through TEM images. As shown in Figure 7, significantly fewer GNRs internalize in NIH-3T3 cells. Most importantly, the internalized GNRs were present in endosomes, and not in cytosol. Our experimental results, as discussed above, have clearly shown the uptake of bombesin based nanoparticles in cytosol when incubated with T47D and PC-3 cells, both of which are known to overexpress GRP receptors. These observations corroborate that the mode of internalization of GNRs in NIH-3T3 cells is predominantly via phagocytosis, whereas, internalization within PC3 and T47D proceeded via receptor mediated endocytosis pathway.