Activatable Photoacoustic Probe for Imaging Infection: Gold Nanorod Dissociation In Vivo Reports Bacterial Protease Activity

We present a strategy for constructing activatable photoacoustic imaging (PAI) probes for in vivo enzyme activity measurements, based on a dissociation strategy previously applied to in vitro sensing. Unlike conventional nanoparticle aggregation strategies, dissociation minimizes false positives and functions effectively in complex biological environments. Overcoming the challenge of dissociating nanostructure aggregates, which arises from the strong van der Waals forces at short distances, we demonstrate the controlled assembly and dissociation of citrate-capped gold nanorods (AuNRs-citrate) using a diarginine peptide additive and a thiolated polyethylene glycol (HS-PEG-OMe), respectively. This assembly dissociation mechanism enables precise control of the optical and photoacoustic (PA) properties of AuNRs in both in vitro and in vivo settings. Building on these findings, we engineered an enzyme-sensitive PAI probe (AuNRs@RgpB) composed of AuNR assemblies and a PEG-peptide conjugate with a protease-specific cleavage sequence. The probe detects Arg-specific gingipain (RgpB), a protease expressed by Porphyromonas gingivalis associated with periodontal disease and Alzheimer's disease. Proteolytic cleavage of the peptide sequence triggers AuNR dissociation, resulting in enhanced PA signal output. The probe was designed to be injected intrathecally for preclinical trials to image gingipains and investigate the value of gingipain inhibitors developed for Alzheimer's disease. The probe's performance was characterized in vitro using UV-vis spectroscopy and PAI, achieving detection limits of 5 and 20 nM, respectively. In vivo studies involved intracranial injection of AuNRs@RgpB into RgpB-containing murine models, with PA monitoring over time. RgpB activity produced a four-fold PA signal increase within 2 h, while P. gingivalis-infected mice showed similar signal enhancement. Specificity was confirmed by negligible responses to Fusobacterium nucleatum, a non-RgpB-producing bacterium. Additionally, the system demonstrated utility in drug development by successfully monitoring the inhibition of RgpB activity using RgpB inhibitors (leupeptin and KYT-1) in vivo models. Beyond its immediate application to RgpB detection, this modular approach to plasmonic-based sensing holds significant potential for detecting other proteases, advancing both nanotechnology and protease-targeted diagnostics.