Prevention of guanylyl cyclase–B dephosphorylation rescues achondroplastic dwarfism

Activating mutations in the fibroblast growth factor receptor 3 (FGFR3) or inactivating mutations in guanylyl cyclase–B (GC-B), also known as NPR-B or Npr2, cause short-limbed dwarfism. FGFR3 activation causes dephosphorylation and inactivation of GC-B, but the contribution of GC-B dephosphorylation to achondroplasia (ACH) is unknown. GC-B7E/7E mice that express a glutamate-substituted version of GC-B that cannot be inactivated by dephosphorylation were bred with mice expressing FGFR3-G380R, the most common human ACH mutation, to determine if GC-B dephosphorylation is required for ACH. Crossing GC-B7E/7E mice with FGFR3G380R/G380R mice increased naso-anal and long (tibia and femur), but not cranial, bone length twice as much as crossing GC-B7E/7E mice with FGFR3WT/WT mice from 4 to 16 weeks of age. Consistent with increased GC-B activity rescuing ACH, long bones from the GC-B7E/7E/FGFR3G380R/G380R mice were not shorter than those from GC-BWT/WT/FGFR3WT/WT mice. At 2 weeks of age, male but not female FGFR3G380R/G380R mice had shorter long bones and smaller growth plate hypertrophic zones, whereas female but not male GC-B7E/7E mice had longer bones and larger hypertrophic zones. In 2-week-old males, crossing FGFR3G380R/G380R mice with GC-B7E/7E mice increased long bone length and hypertrophic zone area to levels observed in mice expressing WT versions of both receptors. We conclude that preventing GC-B dephosphorylation rescues reduced axial and appendicular skeleton growth in a mouse model of achondroplasia.


Introduction
Achondroplasia (ACH) is the most common form of human skeletal dwarfism, which affects about 1 in 20,000 live births worldwide (1)(2)(3). ACH is disproportionate dwarfism. Both the appendicular and axial skeleton are affected. Skeletal defects include rhizomelic bowing of long bones, macrocephaly, midface hypoplasia, maxillary hypoplasia, depressed nasal bridges, and spinal foramen magnum stenosis. Adults with ACH have an average height of just over 4 feet, and more than 80% of ACH cases spontaneously occur through mutations in sperm associated with increased paternal age (4). The vast majority of ACH cases result from a single amino acid substitution in fibroblast growth factor receptor 3 (FGFR3) that increases the tyrosine kinase activity of the receptor (2,3). The most common missense mutation in humans is p.Gly380Arg. Cell culture studies determined that mutations that cause more deleterious ACH phenotypes result in greater increases in phosphorylation and activation of FGFR3 (5). Murine transgenic studies demonstrated that expression of Fgfr3 gain-of-function mutations (p.Gly380Arg) in chondrocytes and osteoblasts results in ACH (6,7).
FGFR3 is a tyrosine kinase receptor that signals through multiple downstream pathways including: ERK1/2-MAPK, p38-MAPK, PI3K-AKT, STATs, and PLCγ (1). In the growth plate, FGFR3 activates both STATs and the ERK1/2 kinases (8). Although disruption of the Stat1 gene rescued the chondrocyte proliferation deficiency in Fgfr3 G374R/+ mice, these mice still developed an ACH-like phenotype (9). In contrast, activated MEK1 expression in chondrocytes of mice lacking active Stat1 yielded an ACH-like phenotype with a decreased hypertrophic chondrocyte zone but no decrease in chondrocyte proliferation (9). Numerous ACH mouse models have been generated with bone and growth plate cartilage anomalies associated with small proliferative and hypertrophic chondrocyte area and less matrix secretion (6,7). Recently, the mouse Activating mutations in the fibroblast growth factor receptor 3 (FGFR3) or inactivating mutations in guanylyl cyclase-B (GC-B), also known as NPR-B or Npr2, cause short-limbed dwarfism. FGFR3 activation causes dephosphorylation and inactivation of GC-B, but the contribution of GC-B dephosphorylation to achondroplasia (ACH) is unknown. GC-B 7E/7E mice that express a glutamatesubstituted version of GC-B that cannot be inactivated by dephosphorylation were bred with mice expressing FGFR3-G380R, the most common human ACH mutation, to determine if GC-B dephosphorylation is required for ACH. Crossing GC-B 7E/7E mice with FGFR3 G380R/G380R mice increased naso-anal and long (tibia and femur), but not cranial, bone length twice as much as crossing GC-B 7E/7E mice with FGFR3 WT/WT mice from 4 to 16 weeks of age. Consistent with increased GC-B activity rescuing ACH, long bones from the GC-B 7E/7E /FGFR3 G380R/G380R mice were not shorter than those from GC-B WT/WT /FGFR3 WT/WT mice. At 2 weeks of age, male but not female FGFR3 G380R/G380R mice had shorter long bones and smaller growth plate hypertrophic zones, whereas female but not male GC-B 7E/7E mice had longer bones and larger hypertrophic zones. In 2-week-old males, crossing FGFR3 G380R/G380R mice with GC-B 7E/7E mice increased long bone length and hypertrophic zone area to levels observed in mice expressing WT versions of both receptors. We conclude that preventing GC-B dephosphorylation rescues reduced axial and appendicular skeleton growth in a mouse model of achondroplasia.
Both FGFR3 and GC-B are highly expressed in proliferating chondrocytes, and reciprocal antagonism between the FGF-FGFR3 and CNP-GC-B pathways has been reported by multiple investigators (33,45,49,50). Specifically, CNP-GC-B signaling was shown to inhibit FGFR3-dependent MAPK activation via protein kinase G regulation of Raf-1 (49,51). Conversely, FGF inhibits CNP-dependent cGMP elevations by increasing GC-B dephosphorylation (24,52). Importantly, FGF failed to inactivate GC-B-7E in cell culture and in live tibial tissue, consistent with FGF-dependent inactivation of GC-B requiring decreased receptor phosphorylation (24,52). Since inactivation of GC-B causes a similar phenotype to that observed with FGFR3 gain-of-function mutations, and FGFR3 activation causes GC-B dephosphorylation and inactivation, we hypothesized that GC-B dephosphorylation is required for the FGFR3-dependent ACH. If this hypothesis is correct, then preventing GC-B dephosphorylation by substituting glutamates for known phosphorylation sites should prevent or reduce the severity of FGFR3-dependent ACH.
To test this hypothesis, we crossed mice where 1 allele of the FGFR3 gene was replaced with the human FGFR3-G380R gain-of-function gene (10) with mice expressing one GC-B-7E allele (22) to generate mouse lines of all possible genetic combinations. We observed that both naso-anal and long bone length of the FGFR3 G380R/G380R GC-B 7E/7E mice was greater than that for FGFR3 G380R/G380R GC-B WT/WT mice. Importantly, we also observed that the hypertrophic zones of 2-week-old growth plates of the FGFR3 G380R/G380R GC-B 7E/7E male mice were the same size as those in WT male mice. Thus, prevention of GC-B dephosphorylation at an early age improved defective chondrocyte differentiation sufficiently to allow normal long bone growth.

Results
Generation of the mouse lines. To evaluate the contribution of GC-B dephosphorylation to ACH, we crossed the GC-B 7E/+ mice as originally described as Npr 7E/+ by Shuhaibar et al. (22) or as GC-B 7E/+ by Robinson et al. (53) with the FGFR3 G380R/+ mice originally described as FGFR3 ACH/+ by Lee et al. (10) to generate all the murine lines shown in Figure 1 and Figure 2. The F1 cross-generated F2 littermates of every possible combination of the FGFR3 and GC-B genes in a mixed C57BL/6 and 129 Sv background. The F2 pups were weaned and genotyped 21 days after birth. Naso-anal measurements were performed once weekly from 4 to 16 weeks of age. Initial visual characterization of 16-week-old male mice indicated that expression of 2 GC-B-7E alleles rescued the majority of the naso-anal length defect but failed to rescue the midface hypoplasia defect in mice expressing 2 FGFR3 G380R alleles that was obvious to the naked eye ( Figure 1). JCI Insight 2021;6(9):e147832 https://doi.org/10.1172/jci.insight.147832

FGFR3-G380R and GC-B-7E mice exhibit sex-specific gene-dosage effects.
Although the effects of FGFR3-G380R and GC-B-7E expression on naso-anal length of mixed male and female populations have been described (10,24), the effects of sex have not been reported for either murine receptor. Using repeated measures analysis from 4 to 16 weeks of age, we observed that expression of 1 allele of FGFR3-G380R decreased skeletal length in male mice by 4.7% but only decreased skeletal length by 1.3% in female mice compared with WT controls ( Figure 2, A and B). Expression of 2 FGFR3-G380R alleles further decreased skeletal length in both sexes, but bone length was decreased to a greater extent (11.4%) in males than in females (8.6%) ( When comparing naso-anal length differences at a single age of 16 weeks, single and double FGFR3-G380R alleles decreased naso-anal length of male mice 5.7% and 11.0%, but females were reduced by 1.7% and 10.3%, respectively ( Figure 2E). In contrast, single and double alleles of GC-B-7E increased naso-anal length 1.7% and 5.4% in males and 3.3% and 5.0% in females, respectively ( Figure 2F). Thus, a single allele of FGFR3-G380R produces a significant decrease in long bone length in adult male but not female mice. Conversely, a single allele of GC-B-7E significantly increased long bone growth in females but not males.
GC-B 7E/7E expression increases naso-anal length more in ACH mice than in WT mice. To evaluate the contribution of GC-B dephosphorylation to ACH, we generated the mouse lines shown in Figure 3, A and B. These data indicate that expressing 2 alleles of GC-B-7E in mice with 2 alleles of FGFR3-G380R increases naso-anal length to a greater extent than expressing 2 alleles of GC-B-7E in mice with 2 WT FGFR3 alleles. Specifically, GC-B 7E/7E expression in an ACH background increased naso-anal length 8.4% and 6.7% in male and female mice, respectively, compared with 4.1% and 5.3% increases, respectively, in FGFR3 WT/WT mice. When examining naso-anal growth at a single time period of 16 weeks, expression of 2 GC-B-7E alleles increased growth by 5.3% in males and 4.7% in females in FGFR3 WT/WT mice but by 8.4% and 8.0%, respectively, in FGFR3 G380R/G380R mice (Figure 3, C and D). The fact that GC-B-7E expression increases naso-anal length  of GC-B WT/WT mice, and the cranial length of the FGFR3 G380R/G380R mice with 2 GC-B-7E alleles (24.15 mm) was slightly greater than that for mice with 2 GC-B-WT alleles (23.00 mm) (P = 0.15) ( Figure 4F), post hoc power calculations indicated that our study was significantly underpowered for this analysis (power, 36%). Based on the differences observed in our initial study, we would need 50 mice per group to reach 80% power, which is prohibitive. Consistent with the lack of differences in cranial length data, GC-B 7E/7E expression failed to rescue the midface hypoplasia and macrocephaly exhibited by the FGFR3 G380R/G380R mice in Figure 1.  Figure 5). Interestingly, we observed that tibias and femurs from 2-week-old mice expressing 2 FGFR3-G380R alleles were only shorter in male mice. Conversely, tibias and femurs were only longer in female 2-week-old mice expressing 2 GC-B-7E alleles. As with the older mice, GC-B-7E expression rescued tibia and femur reductions observed in the FGFR3 G380R/G380R 2-week-old male mice. Hence, the single gene effects of FGFR3-G380R and GC-B-7E on tibia and femur length occur as early as 2 weeks of age but only in a sex-specific manner. Importantly, GC-B-7E expression rescued the long bone reductions observed for the 2-week-old FGFR3 G380R/G380R male mice, as it did for 16-week-old male and female mice.
Prevention of GC-B dephosphorylation rescues reduced hypertrophic zone area in 2-week-old FGFR3 G380R/G380R male mice. Growth plates from 2-week-old femurs were sectioned, stained, and analyzed as previously described (10). The growth plates from the FGFR3 WT/WT GC-B WT/WT mice were compared with growth plates from FGFR3 WT/WT GC-B 7E/7E , FGFR3 G380R/G380R GC-B WT/WT , and FGFR3 G380R/G380R GC-B 7E/7E mice to determine how increased GC-B activity affects hypertrophic zone area in mice expressing mutant and WT FGFR3. Specifically, we examined if the expression of GC-B-7E increased hypertrophic zone area to a greater extent in FGFR3 G380R/G380R growth plates compared with FGFR3 WT/WT growth plates. Previous murine studies indicated that CNP primarily increases chondrocyte hypertrophy (differentiation), as opposed to proliferation (54).
Analysis of growth plates from male mice of each genotype indicated that GC-B 7E/7E expression increased the hypertrophic zone area in animals expressing either FGFR3 WT/WT or FGFR3 G380R/G380 ( Figure 6). However, the GC-B 7E/7E -dependent increase in hypertrophic zone area was greater in the FGFR3 G380R/G380R mice (70.2% increase, P = 0.003) compared with the FGFR3 WT/WT mice (29.6% increase, P = 0.11). In fact, the hypertrophic zones were not significantly different between the FGFR3 G380R/G380R GC-B 7E/7E mice and the FGFR3 WT/WT GC-B WT/WT mice (P = 0.80), which is consistent with GC-B 7E/7E expression rescuing the hypertrophic zone defect in the FGFR3 G380R/G380R mice.
In female mice, hypertrophic zone decreases were not observed in FGFR3 G380R/G380R compared with FGFR3 WT/WT mice at 2 weeks (P = 0.87). However, GC-B-7E expression increased hypertrophic zone areas in both FGFR3 WT/WT (27.7% increase, P = 0.007) and FGFR3 G380R/G380R (18.2% increase, P = 0.032) female mice (Figure 7, A and B). In contrast to the males, hypertrophic zone area in FGFR3 G380R/G380R GC-B 7E/7E female mice was significantly larger than that observed in FGFR3 WT/WT GC-B WT/WT female mice (P = 0.0099), possibly because the FGFR3-G380R mutations failed to decrease long bone growth in the female mice at this early age. Together, these data indicate that the inhibitory effects of the FGFR3-G380R mutations occur earlier in male than female mice, whereas the positive effects of GC-B-7E expression on growth plate area occur earlier in female than male mice.

Discussion
In this report, skeletal phenotypes of mice expressing GC-B-7E alleles and FGFR3-G380R alleles were analyzed and described as a function of age. Our results demonstrate that blocking dephosphorylation- dependent inactivation of GC-B rescues the disproportionate dwarfism of the FGFR3 G380R/G380R mice by preventing the FGFR3-G380R-dependent decreases in growth plate hypertrophic zone area as graphically summarized in Figure 8. The rescue was observed at multiple biological levels, including naso-anal length and individual femur and tibia length, as well as growth plate hypertrophic zone area. Specific interactions between GC-B and FGFR3 on long bone growth were also identified. For example, crossing the GC-B 7E/7E mice with the FGFR3 G380R/G380R mice had a greater effect on naso-anal and long bone length than crossing the GC-B 7E/7E mice with FGFR3 WT/WT mice.
Sex-specific, single allele differences for both genes on long bone growth were also described for the first time, to our knowledge. For single FGFR3-G380R alleles in adult mice, greater skeletal length effects were observed in males, whereas for GC-B-7E, greater effects were observed in females. Increased CNP-GC-Bdependent bone growth in females is consistent with studies showing that CNP expression is more sensitive to exogenous CNP infusion in female compared with male rats (55). Unexpectedly, analysis of 2-weekold femurs and tibias indicated that only males had FGFR3-G380R-dependent reductions in long bone growth, whereas only females exhibited GC-B-7E-dependent increases in bone growth at 2-weeks of age. Histological examination of 2-week-old growth plates confirmed that FGFR3-G380R expression decreased hypertrophic zone area of male but not female growth plates. In contrast, GC-B-7E expression increased the hypertrophic zone area in female but not male growth plates. Importantly, GC-B-7E expression rescued the hypertrophic zone defects in the male FGFR3 G380R/G380R mice. In fact, there was no difference in the size of the growth plates from the FGFR3 G380R/G380R /GC-B 7E/7E mice compared with mice expressing WT versions of both receptors, which is consistent with a rescue of the reduced hypertrophic zone by GC-B-7E expression. Sex-specific effects of ACH on limb lengthening and bone age delay in children have been described, but whether these effects are related to our results in mice is currently unknown (56,57).
The differences in naso-anal growth between the various genetic lines changed very little over the 4to 16-week measurement period, which indicates that the primary effects of GC-B on long bone growth occurs before 4 weeks of age. Large differences in both 2-week-old bones and growth plate hypertrophic zones between mice expressing GC-B-7E and FGFR3-G380R alleles support the idea that the major skeletal effects resulting from the activation of these receptors occurs long before closure of the epiphyseal growth plate. Previously reported increases in murine CNP mRNA at 3 weeks compared with 6 weeks supports the notion that early GC-B activity facilitates endochondral bone growth (38). Identifying when GC-B activation exerts its greatest effect on the growth plate may be important for determining optimal timing of BMN-111 treatment of children with ACH.
While the FGFR3 G380R/G380R mice with 2 alleles of GC-B-7E are not dwarfed, they retain the midface hypoplasia that is a hallmark of ACH. This finding suggests that, although GC-B dephosphorylation is a primary mechanism by which the appendicular and axial skeleton is regulated, the majority of the cranial abnormality is not rescued by increased GC-B activity. However, treatment with BMN-111 and the phosphatase inhibitor LB-100 rescued skull base abnormalities in an ex vivo embryonic skull culture assay from Fgfr3 Y367C/+ mice (58). Thus, a reasonable interpretation of our cranial length data is that, since only a small portion of the skull is stimulated by GC-B, differences in cranial length between GC-B WT/WT and GC-B 7E/7E mice are difficult to demonstrate when measuring the total length of the skull. Hence, the majority of the skull growth occurs by membranous ossification, but a small portion at the base of the skull grows by endochondral ossification, which is stimulated by GC-B.
We suggest that GC-B dephosphorylation is both necessary and sufficient for ACH-dependent reductions in long bone growth. Regarding the "sufficiency" claim, a plethora of reports have shown that inactivating mutations in only GC-B result in dwarfism in mice, rats, and humans (17,(26)(27)(28)(59)(60)(61). Hence, it is clear that inactivation of GC-B, as occurs with dephosphorylation (15,16), is sufficient to cause ACH-like dwarfism. The data supporting the "necessary" claim are that mice expressing alanine substitutions for the phosphorylated serines and threonines in GC-B are dwarfed (17), and the GC-B 7E/7E mice that express a version of GC-B that cannot be inactivated by dephosphorylation are immune to the long bone and growth plate reductions seen in the FGFR3 G380R/G380R mice. Thus, if GC-B is never phosphorylated, the mice are dwarfed. Conversely, if GC-B cannot be inactivated by dephosphorylation, as is the case in the GC-B 7E/7E mice, then they are resistant to ACH.
Finally, multiple investigators have reported that CNP or BMN-111 (vosoritide) injections increase long bone growth in animal models of ACH (33,45,46,(62)(63)(64)(65)(66). Recent clinical phase III studies also indicate that vosoritide increases growth curves in achondroplastic children (67)(68)(69). The fact that BMN-111 is effective in these various ACH models means that a significant fraction of GC-B is phosphorylated, since dephosphorylated GC-B is not activated by CNP or, presumably, CNP analogs (15,16). According to our hypothesis, GC-B dephosphorylation resulting from constitutive activation of FGFR3 is a significant part of the molecular explanation for ACH, although we cannot rule out that other FGFR3-activated signaling pathways also participate in this process. Importantly, if preventing GC-B dephosphorylation rescues ACH as we suggest, then therapies that preserve or increase the phosphorylation state of GC-B, should increase long bone growth in ACH models. Consistent with this hypothesis, the phosphatase inhibitor LB-100 was recently shown to synergistically work with BMN-111 to improve long bone growth in mice (58). Future experiments are required to determine if this combination therapy will be safe and effective in other mammalian species, including humans.

Methods
Animals. Male and female WT and GC-B 7E/7E mice (originally described as Npr2 7E/7E ) were genotyped and maintained as previously described on a C57BL/6J background (22). The FGFR3 G380R/G380R mice were genotyped and maintained on a mixed C57BL/6 and 129 Sv line (10). Crosses between the GC-B 7E/+ C57BL/6J mice and the FGFR ACH/+ C57BL/6 and 129 Sv mice produced all the genotypes shown in Figure 1 and Figure  2. Mice were fed standard chow under specific pathogen-free conditions. Naso-anal, cranial, tibial, and femoral measurements. Naso-anal measurements were conducted on conscious male and female mice once weekly from 4 to 16 weeks of age for the duration of the study using a digital caliper. Mice were euthanized at 16 weeks, and their tibias, femurs, and craniums were isolated and measured using a digital caliper. In a separate study, tibia and femur length from 2-week-old mice were determined as described above.
Growth plate histology. Two-week-old male and female mice were euthanized; femurs were harvested and fixed in 4% paraformaldehyde for 24 hours and then stored in fresh 4% paraformaldehyde until they were embedded in paraffin and sectioned. The bone sections were then stained with H&E and imaged with a 5× or 10× objective lenses. The upper and lower bounds of the cross-sectional area of the hypertrophic zones were identified by a morphology of columnar, stacked cells that were measured in millimeters in an unbiased manner using ImageJ software (NIH). The freehand selection tool in ImageJ determined the number of square millimeters of the identified hypertrophic region from each section of bone as indicated by the blue outline in the images shown in Figure 6 and Figure 7. In a blinded manner, the square millimeter value from each femoral section was determined and plotted as a single value in Figure 6B and Figure 7B.
Statistics. Statistics and graphs were generated with Prism 8 software (GraphPad Software). For the naso-anal growth curves, a mixed-model 2-way ANOVA with repeated measures was applied using a Geisser-Greenhouse correction for sphericity and Tukey's correction for multiple comparisons. Therefore, the naso-anal length values described under results were derived from repetitive measurements on the same animal at multiple ages. The remaining analyses including the growth plate measurements were performed using a standard, 1-way ANOVA assuming Gaussian distribution and equal SD with Tukey's correction for multiple comparisons. Except for the naso-anal growth curves, each symbol in a figure represents a measurement from an individual mouse. The longer horizontal bar within each group of symbols represents the mean, and the shorter bars above and below the mean represent the SEM. For the naso-anal growth curves, each point represents the mean length measurement at the indicated age, and bars above and below the mean indicate SEM. Differences were considered significantly different when the P value was less than 0.05. JCI Insight 2021;6(9):e147832 https://doi.org/10.1172/jci.insight.147832 Study approval. Animal care was compliant with the University of Minnesota IACUC-approved protocol no. 1806A-35995.

Author contributions
Study design was contributed by JWR, LLM, and LRP. Naso-anal length measurements were contributed by JWR and BMW. Tibia and femur measurements were contributed by JWR and BMW. Histological data were contributed by NK. Hypertrophic zone measurement was contributed by JWR, BMW, and LLM. Analysis of data was contributed by BMW, JWR, LLM, and LRP. Drafting the manuscript was contributed by JWR, BMW, and LRP. Revising manuscript content was contributed by BMW, JWR, LRP, and LLM. YWL and YCL provided the ACH mouse line, and YCL edited the paper. All authors approved the final version of the manuscript.