Pinching antenna (PA) systems have recently emerged as a promising flexible-antenna technology, which can reconstruct the wireless propagation environment by dynamically adjusting the positions of pinching elements along dielectric waveguides, thereby providing new spatial degrees of freedom (DoFs) for enhancing wireless system performance. This paper investigates a multi-waveguide PA-based multi-cell communication system, focusing on the joint optimization of precoding matrices, waveguide power allocation, and antenna placement to maximize the weighted sum rate (WSR). In multi-cell scenarios, inter-cell interference typically leads to a highly coupled and nonconvex WSR maximization problem. To address this challenge, an efficient alternating optimization framework is adopted to optimize each variable in an iterative way. Specifically, fractional programming is first employed to reformulate the original problem by introducing auxiliary variables that decouple the signal and interference terms. Based on this reformulation, block coordinate descent is then applied to optimize the precoding matrices and power allocation, leading to closed-form or semi-closed-form updates. For the high-dimensional and nonconvex PA placement problem, particle swarm optimization (PSO) is utilized to perform an efficient search and improve scalability. Numerical results demonstrate that, under various system configurations, the proposed scheme significantly outperforms baseline methods, including average power allocation, fixed antenna placement, conventional multiple-input multiple-output (MIMO), and massive MIMO. These results highlight the strong potential of PA systems for large-scale multi-cell wireless communications.