Reply to Tarchini et al.: Evidence for a GPR156 role in central stress responses in addition to its inner-ear function

In their comments on our study, Tarchini et al. point out that homozygous loss of function of Gpr156 in mice leads to hearing loss and vestibular dysfunction, and they raise the question of whether the phenotypes we observed are due to inner-ear dysfunction rather than abnormal stress responses (1). We agree that homozygous loss of function of Gpr156 causes inner-ear dysfunction (2–7). The human E533D variant carriers in our study are heterozygous, and they display an episodic illness, with baseline normal mood punctuated by depressive episodes. For our in vivo functional studies, we generated homozygous mutant mice as a probe of GPR156’s effects on stress-related behavior. While we did not test inner-ear function in these mice, their behavioral phenotypes could reflect inner-ear dysfunction, abnormal stress responses, or both.

Several of our observations nevertheless suggest that altered central stress responses contribute to the phenotypes (8). We found that Gpr156 is highly expressed in the medial habenula, a region of the central nervous system implicated in depression and stress responses (9, 10). Moreover, Gpr156 mutations cause medial habenula neuronal hyperactivity, establishing a central electrophysiological phenotype that could underlie altered stress responses. Behaviorally, homozygous mutants have intact locomotion and initial tail hanging in the land-based tests, consistent with prior reports (4, 11). This suggests that the phenotypes we found in those tests involve stress-related processes rather than strictly vestibulomotor impairment. In the forced swim test, chronic fluoxetine treatment partially rescued the mutant phenotype, supporting a stress-related component while not excluding the role of inner-ear GPR156 in swimming (7).

To test more directly whether GPR156 acts in the medial habenula to regulate stress responses, we are conducting studies to selectively knock out Gpr156 in this brain region. By preserving GPR156 function in the inner ear, this model will directly test whether loss of GPR156 in the medial habenula is sufficient to alter stress-related behavior.

We independently generated the constitutive Gpr156 knockout mouse line used in our study as described. We used a targeting strategy similar to that used for the KOMP line, but with a different 129-derived BAC. The two lines are expected to produce comparable null alleles, consistent with their concordant phenotypes.