Efficient spatial navigation in complex environments relies on the construction of an egocentric cognitive map. Recent studies have reported that subpopulations of excitatory neurons in the granular retrosplenial cortex (RSC) are selectively activated at specific distances and orientations relative to environmental boundaries and vertices, tuned by the heading directions of the animal, to perform egocentric vector coding of the environmental geometry. However, the neural circuit mechanisms explaining how egocentric representation of vertices and boundaries emerges are yet unknown. One possible neural circuit mechanism is GABAergic inhibition, which is well known for shape and even gate the emergence of many spatially tuned neurons. To directly test this, we performed longitudinal in vivo Ca2+ imaging of GCaMP-expressing RSC excitatory neurons and optogenetically silenced PV interneurons while PV-Cre mice freely explored a square open chamber. Analysis of spike-trajectory plots and firing rate maps revealed that RSC excitatory neurons represented the environmental boundaries and vertices using an egocentric reference frame, as previously reported. However, silencing of PV interneurons not only weakened the strength of the egocentric vector coding of environmental geometry but also abolished the egocentric vector coding in subpopulations of RSC excitatory neurons. Overall, the population of egocentric vector coding neurons significantly decreased when PV interneurons were silenced, indicating that PV interneurons may carry egocentric information to excitatory neurons. Indeed, Ca2+ imaging of GCaMP-expressing PV interneurons revealed that a large population of PV interneurons are egocentrically tuned. Overall, these findings suggest that RSC PV interneurons egocentrically represent space and gate the egocentric vector coding of RSC excitatory neurons, indicating their critical roles in in the construction of egocentric cognitive map.