Deciphering the genetic interactions between Pou4f3, Gfi1 and Rbm24 in maintaining cochlear hair cell survival

Guangqin Wang
Yunpeng Gu
Zhiyong Liu author has email address

Institute of Neuroscience, State Key Laboratory of Neuroscience, CAS Center for Excellence in Brain Science and Intelligence Technology, Chinese Academy of Sciences, Shanghai, 200031, China
University of Chinese Academy of Sciences, Beijing, 100049, China
Shanghai Center for Brain Science and Brain-Inspired Intelligence Technology, Shanghai, 201210, China

https://doi.org/10.7554/eLife.90025.1

Open access
Copyright information

Figures and data

Rbm24 expression is repressed in cochlear HCs in the absence of Pou4f3.
(A) A simple illustration of how Pou4f3 ^-/- strain was generated. (B) One exampled tail PCR gel image that helped us to distinguish WT (Pou4f3^+/+), heterozygous (Pou4f3^+/-) and homozygous (Pou4f3^-/-) mice. (C-F’’’) Triple straining of Rbm24, Pou4f3 and Bcl11b in WT (C and E-E’’’) and Pou4f3^-/- (D and F-F’’’) mice (n=3 for each) at E16.5, with only Rbm24 (green) channel was visualized in (C and D) at low magnification. (E-E’’’) and (F-F’’’) were the high magnification images of middle turn WT and Pou4f3^-/- mice, respectively. In the absence of Pou4f3, Rbm24 expression completely disappeared, despite nascent Bcl11b+ OHCs were present. OHC: outer hair cell; IHC: inner hair cell. Scale bars: 100 μm (D), 20 μm (F’’’).

Gfi1 is dispensable for Rbm24 expression.
(A) A simple cartoon representing how the Gfi1 was inactivated in the one-cell stage mouse zygotes derived from male Gfi1^HACre/+ by the CRISPR-stop approach. (B-C) With sgRNA-6 as an example, the Sanger sequencing chromatograms of control Gfi1^HACre/+ (B) and F0 Gfi1^-/- (C) were presented. The base ‘C’ (black arrow in B) was converted in ‘T’ (red arrow in C), resulting in pre-emergence of protein translation stop codon TAG. The red ‘T’ was a single peak, meaning that the pre-emergence of stop codon occurred in both alleles. (D-E’’) Dual staining of HA (Gfi1) and Rbm24 in cochleae from control Gfi1^HACre/+(D-D’’, n=3) and F0 Gfi1^-/- (E-E’’) mice (n=4) at E16.5. Rbm24 expression (E’) was not altered in the absence of the HA (Gfi1) expression (E). OHC: outer hair cell; IHC: inner hair cell. Scale bar: 20 μm.

Pou4f3 and Gfi1 expressions are normal in the absence of Rbm24.
(A) A simple cartoon to elaborate how we used CRISPR-stop approach to induce mosaic inactivation of Rbm24 in the one-cell stage mouse zygote harboring Gfi1^HACre/+. Notably, the injection was in the late one-cell stage of the zygotes, resulting in higher rates of final mosaic inactivation. (B-C) The Sanger sequencing chromatograms of control Gfi1^HACre/+; Rbm24^+/+ (B) and Gfi1^HACre/+; Rbm24^-/- (mosaic) (C). The base ‘C’ (black arrow in B) was converted into ‘T’ (red arrow in C) in a fraction of cochlear cells. Notably, the mosaic inactivation of Rbm24 was evidenced by the double peaks of ‘T’ and ‘C’ (red arrow in C). (D-E’’’) Triple staining of Rbm24, HA and Pou4f3 in cochlear samples (basal turns) of control Gfi1^HACre/+; Rbm24^+/+(D-D’’’, n=3) and Gfi1^HACre/+; Rbm24^-/- (mosaic) (E-E’’’, n=3) at E17. Consistent with the Sanger sequencing results (C), Rbm24 was inactivated in cochlear HCs in a mosaic pattern (E). The orange arrows (E-E’’’) pointed to one OHC that lost Rbm24, whereas the blue arrows (E-E’’’) marked one OHC in which the Rbm24 expression was maintained. Notably, regardless of whether Rbm24 was inactivated or not, HA (Gfi1) and Pou4f3 expression were not altered. OHC: outer hair cell; IHC: inner hair cell. Scale bar: 20 μm (E’’’).

Three Rbm24 enhancers are sufficient to drive specific EGFP expression in cochlear HCs.
(A) Visualization of ATAC-seq (top row, orange color) of P1 cochlear HCs (from our previous study, GSE181311) and Pou4f3 Cut&Run assay (bottom row, blue color) of E17.5 cochlear HCs (from another study, GSE150391) by Integrative Genomics Viewer. The black arrow marked the Rbm24 promoter, and the three Rbm24 enhancers (Eh1, Eh2 and Eh3) were marked by the dotted square. (B-G’’) Dual staining of EGFP and Myo7a in Eh1-EGFP+ (B-C’’, n=7), Eh2-EGFP+ (D-E’’, n=6) and Eh3-EGFP+ (F-G’’, n=4) at P1. Only the EGFP channel was visualized in whole mount staining (B, D, F). In the cryosection cochlear samples, EGFP was overlapped with HC marker Myo7a. OHCs: outer hair cells; IHC: inner hair cell. Scale bars: 100 μm (F), 20 μm (G’’).

Forced expression of Rbm24 fails to rescue the OHC degeneration in Pou4f3^-/- mutants.
(A) Generation of Rosa26-Loxp-stop-Loxp-Rbm24*3ξHA/+ (Rosa26^Rbm²⁴^/+) mouse stain. The C-terminus of the ectopic Rbm24 was tagged with three HA tags. (B) One exampled tail PCR gel image that helped us to distinguish WT (+/+) and KI (Rosa26^Rbm²⁴^/+) mice. (C-F’’) Dual staining of HA (Rbm24) and hair cell marker Myo7a in mice with four different genotypes at P1: Pou4f3^+/+ (C-C’’, n=3), Pou4f3^-/- (D-D’’, n=3) and Pou4f3^+/+; Atoh1^Cre/+; Rosa26^Rbm²⁴^/+(E-E’’, n=3) and Pou4f3^-/-; Atoh1^Cre/+; Rosa26^Rbm²⁴^/+ (F-F’’, n=3). The ectopic HA (Rbm24) was not detected in Pou4f3^+/+ (C) and Pou4f3^-/- (D), but was detected in both Pou4f3^+/+; Atoh1^Cre/+; Rosa26^Rbm24/+(E) and Pou4f3^-/-; Atoh1^Cre/+; Rosa26^Rbm24/+(F). The white arrows in (E-E’’) pointed to one Myo7a+ IHC that expressed the ectopic HA (Rbm24). The HCs seemed normal by P1, despite the ectopic Rbm24 expression was induced. The orange arrows in (E-E’’) labelled one SC that expressed HA (Rbm24) but not Myo7a. The white arrows in (F-F’’) should be one remaining Myo7a+ HC that expressed ectopic Rbm24. Overall, many of the Pou4f3^-/- HCs degenerated, albeit ectopic Rbm24 was induced. OHC: outer hair cell; IHC: inner hair cell; HC: hair cell. Scale bar: 20 μm (F’’).

Rbm24 expression is repressed in cochlear HCs at E16.5 when Pou4f3 is inactivated.
(A) An illustration of how Pou4f3^-/- mice (Founder 0, F0) were rapidly generated by CRISPR-stop. (B-C) Sanger sequencing chromatograms of WT (B) and F0 Pou4f3^-/- (C) were presented. The base ‘C’ (black arrow in B) was completely converted to ‘T’ (red arrow in C), resulting in pre-emergence of protein translation stop codon. The Pou4f3^-/-mutants (F0 mice) were immediately ready for phenotypic analysis. (D-E’’’) Triple staining of pan-HC marker Pou4f3, Rbm24 and OHC marker Insm1 in whole mount prepared cochlear samples (basal turns) of WT (D-D’’’, n=3) and Pou4f3^-/- (E-E’’’, n=9) at E16.5. Relative to WT mice, both Pou4f3 and Rbm24 expression disappeared in Pou4f3^-/- HCs, whereas the Insm1 expression was normal. The orange arrows in (E-E’’’) pointed to one Insm1+ OHC that lost Pou4f3 and Rbm24. OHC: outer hair cell; IHC: inner hair cell. Scale bar: 20 μm (E’’’).

Severe degeneration of Pou4f3^-/- HCs happens by perinatal ages.
(A-C’’) Dual staining of Pou4f3 and Myo7a in middle turn regions of the cochleae of WT (A-A’’) at P1 and Pou4f3^-/- mice at P0 (B-B’’) and at P1 (C-C’’). Notably, here the Pou4f3^-/- mice were germ line stable strain produced by CRISPR/Cas9-mediated large DNA fragment deletion (Figure 1A). Arrows in (B-C’’) represented the surviving HCs in the absence of Pou4f3. (D) Quantification of the percentage of the remaining HCs at basal, middle, and apical turns in Pou4f3^-/- mice at P0 (n=4) or at P1 (n=3), relative to WT mice at P1 (n=3). Data were presented as Mean ± Sem. Student’s t test was used for statistical analysis. ** p<0.01; **** p<0.0001; n.s. no significant difference. HC degeneration at middle and basal turns was more severe at P1 than that at P0, but there was no significant difference in apical turns (D). (E-F’’) Dual staining of Rbm24 and Myo7a in WT (E-E’’, n=3) and Pou4f3^-/- (F-F’’, n=3) mice at P1. The white arrow in (F-F’’) marked one remaining Pou4f3^-/- HC that lost Rbm24 expression. OHC: outer hair cell; IHC: inner hair cell; HC: hair cell. Scale bar: 20 μm (C’’ and F’’).

Construction of Gfi1-3◊HA-P2A-Cre/+ (Gfi1^HA-Cre/+) mouse model by which HA can faithfully represent Gfi1 expression.
(A-C) A simple cartoon to illustrate how the Gfi1^HA-Cre/+ strain was constructed. Basically, the WT Gfi1 allele (A) was inserted a fragment containing 3◊HA-P2A-Cre (B) immediately before the TGA stop codon (red in A), resulting in a post-targeted Gfi1^HA-Cre/+ allele (C). (D) One exampled tail PCR gel image that helped us to distinguish Gfi1^+/+, Gfi1^HA-Cre/+ and Gfi1^{HA-Cre/HA-Cre} mice. (E-F’’) Dual staining of HA (Gfi1) and HC marker Myo7a of control Gfi1^+/+ (E-E’’, n=3) and homozygous Gfi1^{HA-Cre/HA-Cre}mice (F-F’’, n=3) at P1. Arrows in (F-F’’) marked one OHC that expressed HA (Gfi1) and Myo7a. (G-J’’) Fate mapping analysis of cells experiencing Gfi1 expression by Gfi1^HA-Cre/+; Rosa26-loxp-stop-loxp-tdTomato (Ai9)/+ (Gfi1^HA-Cre/+; Ai9/+), with confocal low magnification images being presented in (G-H’’’) and high magnification images in (I-J’’). In Ai9/+ control mice at P2 (n=5), no Tdtomato+ cells were detected (G-G’’’ and I-I’’). In contrast, in the Gfi1^HA-Cre/+; Ai9/+ mice (n=5), Tdtomato was detected in both non-sensory cells (white arrows in H) and the Myo7a+ HCs that expressed heterogeneous levels of HA (Gfi1), as represented in different arrows in (J-J’’): white arrows (#1) marked one OHC with the highest HA (Gfi1), the orange arrows (#3) with the lowest (or undetectable) HA (Gfi1), and the blue arrows (#2) with the intermediate level of HA (Gfi1). Scale bars: 100μm in (H’’’); 20 μm (F’’ and J’’).

Gfi1 expression is lost in the Pou4f3^-/- HCs.
(A) One-step generation of Pou4f3 mutants by CRISPR-stop, with the zygotes derived from the Gfi1^HA-Cre/+ mice. (B-C) Sanger sequencing chromatograms of tail DNA from the control Gfi1^HA-Cre/+ (B) and Gfi1^HA-Cre/+; Pou4f3^-/- (mosaic) mice (C). The base ‘C’ in (blue arrow in B) was converted to ‘T’ (red arrow in C). However, the conversion was not complete because double peak existed, supporting that the inactivation of Pou4f3 was mosaic. (D-E’’’) Triple staining of Pou4f3, Rbm24 and HA (Gfi1) in the basal region of cochleae of Gfi1^HA-Cre/+ (D-D’’’, n=3) and Gfi1^HA-Cre/+; Pou4f3^-/-(mosaic, E-E’’’, n=3) at E16.5. Blue arrows in (E-E’’’) marked one OHC that expressed Pou4f3, Rbm24 and HA (Gfi1), whereas the orange arrows labelled the other OHC that lost Pou4f3, Rbm24 and HA (Gfi1). Thus, it confirmed that not only HA faithfully represented the Gfi1 expression, but also HA expression was dependent on Pou4f3. Scale bar: 20 μm (E’’’).

Rbm24 expression is normal in the absence of Gfi1.
(A) The DNA fragment between the sgRNA-4 and sgRNA-9 within the Gfi1 locus was deleted by CRISPR/Cas9 method. (B) One exampled tail PCR gel image that helped us to distinguish Gfi1^+/+, Gfi1^+/- and Gfi1^-/- mice. (C-D’’) Dual staining of Pou4f3 and Rbm24 in middle turn cochleae of WT (C-C’’, n=3) and Gfi1^-/-(D-D’’, n=3) at P1. Severe HC, especially OHC, degeneration happened in Gfi1^-/- cochleae. The orange arrows in (D-D’’) marked one surviving Gfi1^-/- OHC that expressed Pou4f3 and Rbm24. (E-F’’’) Triple staining of Pou4f3, Rbm24 and OHC marker Bcl11b of cochleae of WT (E-E’’’, n=3) and Gfi1^-/- (F-F’’’, n=3) at E16.5. In terms of the expression patterns of Pou4f3, Rbm24 and Bcl11b, no significant difference was captured between WT and Gfi1^-/- cochleae at E16.5, suggesting that Gfi1^-/- HC degeneration did not start yet. Scale bars: 20 μm (D’’ and F’’’).

Sign up for email alerts