Early detection of cancer is important for survival, and multiple targeted contrast agents can be used for advanced cellular profiling. Photoacoustic (PA) imaging is effective at imaging optically-absorbing contrast agents. In this study, we present simulated and experimental results using spectral decomposition to separate the PA absorption spectrum into components from individual contrast agents. The original co-localized concentration of each agent is then simultaneously reconstructed. Monte Carlo simulations were performed to estimate wavelength-dependent fluences in the sample. In phantom experiments, the concentration of two contrast agents (gold nanorods with different absorption spectra) and a background absorbing dye (India Ink) were reconstructed. Finally, we implemented our algorithm in a mouse after injections of gold nanorods with absorption peaks of 750 and 810 nm. The concentration of each nanoparticle was identified from the endogenous background signal with a signal-to-noise ratio of greater than 25 dB.