We have extended the TLA+ proof system TLAPS with a new backend to improve the automation of proof steps that involve higher-order reasoning. The current support for such steps is poor, requiring the user to break down proofs into unnecessarily small steps. We defined a translation from TLA+ to THF, the TPTP dialect for higher-order logic, and evaluated several higher-order solvers on proof obligations generated from the standard library of TLA+. Our results demonstrate that the solvers are able to handle much coarser proof steps than the other strategies provided by TLAPS, reducing the amount of necessary user interactions by a significant margin.