One major challenge when developing new biomaterials is translating in vitro testing to in vivo models. We have recently shown that a single formulation of a bone tissue adhesive, phosphoserine modified cement (PMC), is safe and resorbable in vivo. Herein, we screened many new adhesive formulations, for cytocompatibility and bioactive ion release, with three cell lines: MDPC23 odontoblasts, MC3T3 preosteoblasts, and L929 fibroblasts. Most formulations were cytocompatible by indirect contact testing (ISO 10993-12). Formulations with larger amounts of phosphoserine (>50%) had delayed setting times, greater ion release, and cytotoxicity in vitro. The trends in ion release from the adhesive that were cured for 24 h (standard for in vitro) were similar to release from the adhesives cured only for 5–10 min (standard for in vivo), suggesting that we may be able to predict the material behavior in vivo, using in vitro methods. Adhesives containing calcium phosphate and silicate were both cytocompatible for seven days in direct contact with cell monolayers, and ion release increased the alkaline phosphatase (ALP) activity in odontoblasts, but not pre-osteoblasts. This is the first study evaluating how PMC formulation affects osteogenic cell differentiation (ALP), cytocompatibility, and ion release, using in situ curing conditions similar to conditions in vivo.