*3.1. Synthesis and Functionalization of Nanostar-Shaped AuNPs Optimized for In Vivo Use*

To be able to use nanostars as a theranostic agent in vivo, they need to be stable under physiological conditions, preferably absorb the light in the NIR-range and have a size around 100 nm for optimal retention in tumors. To meet these requirements, we optimized a two-step method by changing the flow rate of the HAuCl4 to 12.5 μL/min during synthesis and by functionalizing them with a self-assembled monolayer to stabilize their specific shape (Figure 1) [13]. The resulting nanostars showed an improved plasmon absorption band around 670 nm and 679 nm (Supplementary Figure S1). This red shift indicated a successful chemical functionalization with the disulfide, resulting from a local refractive index change due to the chemisorption of the disulfide onto the nanostars. These results were confirmed by DLS, showing an average diameter of 66.3 ± 7.8 nm for the synthesized nanostars and 75.0 nm ± 5.6 nm for the further functionalized nanostars. After functionalization with disulfide-PEG-maleimide, the zeta potential was −41.3 ± 1.2 mV, indicating negatively charged nanostars (Supplementary Figure S2). The branched shape of the nanostars was confirmed using TEM (Figure 1). Incubating the nanostars in biological medium for one week at 37 ◦C did not show any indication of instability, confirmed by the absence of peak broadening of the LSPR band or increase in diameter of the nanostars over time (Supplementary Figure S2). Still, an increase of the diameter to 100.6 ± 1.8 nm was seen immediately after incubation in cell culture medium due to the formation of a protein corona. As a consequence, these optimized nanostars are suitable for in vitro and in vivo theranostic applications.

**Figure 1.** UV-Vis absorption spectroscopy of nanostars using different flow rates for the gold salt. Note the shift to the near infrared region (NIR) region with lower flow rates. TEM images of the nanostars that were generated with flow rates of 50 μL/min, 25 μL/min and 12.5 μL/min (from top to bottom). These images suggest that the amount and length of spikes change with different flow rates. The diameter of the nanostars did not significantly change as confirmed by dynamic light scattering (DLS) where no significant increase was noticed with lower flow rates (50 μL/min: 66.0 ± 1.5 nm; 25 μL/min: 67.5 ± 2.7 nm: 12.5 μL/min: 66.3 ± 7.8 nm; data not shown).

After functionalization, the capability of the nanostars to generate image contrast was studied in water using a concentration of 1.55 mg Au/mL. A PAI signal of 0.33 ± 0.04 a.u. was measured for the nanostars while water shows a signal of 0.13 ± 0.01 a.u. (Figure 2). For CT, a signal of 175.67 ± 2.74 and 57.33 ± 0.17 for the nanostars and water was measured, respectively. As a consequence, both imaging modalities showed almost an identical CNR of 32.85 for PAI and 32.53 for CT, respectively.

**Figure 2.** (**A**) Photoacoustic images of nanostars and water in tubes, which were quantified by plotting the signal amplitudes. (**B**) Computed tomography (CT) images of microcentrifuge tubes either filled with water or nanostars suspension. The corresponding signal amplitudes were used for quantification. (Error bars represent SD of triplicate samples; \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001).
