Heterozygous variants in SIX3 and POU1F1 cause pituitary hormone deficiency in mouse and man

Abstract

Congenital hypopituitarism is a genetically heterogeneous condition that is part of a spectrum disorder that can include holoprosencephaly. Heterozygous mutations in SIX3 cause variable holoprosencephaly in humans and mice. We identified two children with neonatal hypopituitarism and thin pituitary stalk who were doubly heterozygous for rare, likely deleterious variants in the transcription factors SIX3 and POU1F1. We used genetically engineered mice to understand the disease pathophysiology. Pou1f1 loss-of-function heterozygotes are unaffected; Six3 heterozygotes have pituitary gland dysmorphology and incompletely ossified palate; and the Six3+/-; Pou1f1+/dw double heterozygote mice have a pronounced phenotype, including pituitary growth through the palate. The interaction of Pou1f1 and Six3 in mice supports the possibility of digenic pituitary disease in children. Disruption of Six3 expression in the oral ectoderm completely ablated anterior pituitary development, and deletion of Six3 in the neural ectoderm blocked the development of the pituitary stalk and both anterior and posterior pituitary lobes. Six3 is required in both oral and neural ectodermal tissues for the activation of signaling pathways and transcription factors necessary for pituitary cell fate. These studies clarify the mechanism of SIX3 action in pituitary development and provide support for a digenic basis for hypopituitarism.

Graphical Abstract

Figure 1

Children with CPHD have a SIX3 variant that reduces transactivation. (A) Pedigree of three children with CPHD. Heterozygotes are indicated for variants in SIX3 (p.P74R) and POU1F1 (p.S50A) inherited from their unaffected father and mother, respectively. (B) Sanger sequence validated the heterozygous variant in SIX3 (c.C221G) and segregation within the pedigree, including an unaffected sibling heterozygous for the SIX3 variant. (C) The proline at codon 74 of SIX3 is conserved in mammals. (D) SIX3 p.P74R is in the regulatory domain N-terminal to other variants associated with HPE. SIX3 p.F87E affects repressor activity and V92G affects transactivation. (E) Cells were transfected with SIX3 expression vectors bearing the indicated amino acid substitutions and a reporter gene, SBE2-Luc. (F) Same as E except the expression vector was Foxg1-Luc. Asterisks indicate a significant difference from wild type (WT) (P < 0.05).

Figure 2

Six3^+/−; Pou1f1^+/dw double heterozygous mice exhibit pituitary dysmorphology and impaired palate ossification. (A) Immunostaining reveals SIX3 colocalization with PROP1 at e13.5. (B) SIX3 and POU1F1 only co-localize in a few cells. (C) Six3^+/−; Pou1f1^+/dw mice weigh the same at 3 weeks compared with single mutants and controls. Each symbol represents an individual mouse. Symbols colored in red represent mice that did not survive past weaning. (D) Whole pituitary glands dissected at 6 weeks show abnormal shape of the Six3^+/−; Pou1f1^+/dw double heterozygotes. (E) Images of the pituitary gland within the head with rostral aspects at the top. (F) Ventral view of skeletal preps from P0 pups. Red outline indicates area of presphenoid bone that is ossified, and blue outline indicates area of palatal cartilage measured. (G) Bone area was calculated for each individual and genotype as a percentage of bone (red outline) compared with the overall area of bone plus cartilage (blue outline). Each symbol represents an individual neonate. Statistical significance was determined using a Student t test. *** = P < 0.001. C = control, P^+/dw = Pou1f1^+/dw, S^+/− = Six3^+/−, DH = Six3^+/−; Pou1f1^+/dw. Scale bar in panel A represents 50 μm and is applicable to panels A and B. Scale bar in panel D represents 100 μm. The scale bar in panels E and F represents 500 μm.

Figure 3

Six3 haploinsufficiency causes dysmorphology. (A) Representative hematoxylin and eosin staining is shown for sagittal sections from e14.5 embryos of each genotype. The posterior lobe and dorsal aspect of Rathke’s pouch are outlined. Cells between these tissues are present in Six3^+/− and Six3^+/−; Pou1f1^+/dw embryos. Arrow shows growth of Rathke’s pouch through the underlying cartilage plate. (B) Representative immunostaining for CCND2 is shown for each genotype, Six3^+/− (N = 3) and Six3^+/−; Pou1f1^+/dw (N = 5). (C) Staining for EdU is shown for embryos collected at e13.5, Six3^+/− (N = 6) and Six3^+/−; Pou1f1^+/dw (N = 5). (D) Immunostaining for p57 reveals cells exiting the cell cycle, arrows; Six3^+/− (N = 6) and Six3^+/−; Pou1f1^+/dw (N = 5). The scale bar in panels A and B represents 50 μm and is applicable to all panels.

Figure 4

Pituitary-specific deletion of Six3 results in Rathke’s pouch hypoplasia and impaired ventral diencephalon signaling. (A) Sagittal sections of embryos representing three genotypes were collected at e11.5 and stained with hematoxylin and eosin. The Six3^fl/fl; Prop1-cre embryos had a range of abnormalities from hypoplastic Rathke’s pouch with no evidence of infundibulum formation (7/16) to multiple invaginations at ectopic sites (9/16). (B) Sections from the same genotypes and age were immunostained for PITX1 and (C) LHX3. Some regions of Six3^fl/fl; Prop1-cre embryos expressed both PITX1 and LHX3 (arrows), and other regions expressed PITX1, but without detectable LHX3 immunostaining (arrowheads). (D) Sections from the same genotypes and age were used for in situ hybridization for Bmp4 with RNAscope probes, and traditional in situ probes for (E) Fgf10 and (F) Shh. Boundaries of Bmp4 and Fgf10 expression are expanded in Six3^fl/fl; Prop1-cre embryos relative to controls (arrowheads). Shh transcripts represent in both the ventral diencephalon and oral ectoderm in controls, but Six3^fl/fl; Prop1-cre mutants had reduced or absent Shh staining in the ventral diencephalon (asterisks) despite detectable staining in the oral ectoderm (arrowheads). Two examples from the Six3^fl/fl; Prop1-cre mutants are shown for each marker, with the left panel representing the more severe phenotype. The fraction of embryos reflecting the severe and milder phenotypes is indicated. The scale bar in panel A represents 100 μm. The scale bar in panel B–F represents 50 μm.

Figure 5

Hypothalamus-specific Six3 knockout embryos have severe pituitary hypoplasia and stalk disruption. (A) Hematoxylin and eosin staining of sagittal sections at e11.5 reveals variable pituitary gland hypoplasia in Six3^fl/fl; Nkx2.1-cre embryos. (B) LHX3 and PITX1 immunostaining are reduced in Six3^fl/fl; Nkx2.1-cre embryos at e11.5. (C) TLE4 immunostaining ranged from absent or reduced on in the distal side of the infundibulum (white arrow) in Six3^fl/fl; Nkx2.1-cre embryos. (D, E) Six3^fl/fl; Nkx2.1-cre mice have PSIS at E18.5. (D) Six3^fl/fl; Nkx2.1-cre mice had no pituitary or small anterior pituitary gland without a posterior lobe (black arrow). (E) All Six3^fl/fl; Nkx2.1-cre mutants showed AVP immunopositive cells that do not project to the pituitary gland (white arrows). Two examples from the Six3^fl/fl; Nkx2.1-cre mutants are shown for each marker, with the panel to the left representing the more severe phenotype. Incidence of observation is shown in the upper left corner of all mutant panels. Scale bars: 100 μm.

Figure 6

Altered patterning of the ventral diencephalon in Six3^fl/fl; Nkx2.1-cre mice underlies pituitary hypoplasia and agenesis. (A) Axin2 RNAscope detects anterior extension of Axin2 expressing area in Six3^fl/fl; Nkx2.1-cre mutants. (B) Hes1 in situ hybridization detects transcripts only in the rostral part of the ventral diencephalon in Six3^fl/fl; Nkx2.1-cre embryos. (C) Fgf10 in situ hybridization did not detect transcripts in 5/5 Six3^fl/fl; Nkx2.1-cre embryos. (D) Immunostaining for pSMAD 1/5/9, a marker of BMP signaling, was weakly positive (one out of four samples) or absent in the diencephalon of mutants. (E) Tbx3 in situ hybridization is absent or reduced in Six3^fl/fl; Nkx2.1-cre embryos. (F) Shh in situ hybridization in the diencephalon ranges from absent to comparable in Six3^fl/fl; Nkx2.1-cre embryos. Two examples from the Six3^fl/fl; Nkx2.1-cre mutants are shown for each marker, with the panel to the left representing the more severe phenotype. Incidence of observation is shown in the upper left corner of all mutant panels. Scale bars: 100 μm.

Figure 7

Genetic hierarchy of SIX3 action in the ventral diencephalon that affects infundibulum formation and Rathke’s pouch expansion. Deletion of Six3 in the ventral diencephalon has no effect on Rax expression, but it permits elevated canonical WNT signaling and reduces Notch signaling (Hes1). This causes reduced expression of Otx2, which reduces downstream signaling by BMP and FGF, and failure to activate signature transcription factor expression for infundibulum development, namely Tbx3 and Lhx2. It is also associated with reduced expression of the pituitary fate marker Lhx3, and apoptosis of Rathke’s pouch.

Figure 8

SIX3 expression is required in the oral ectoderm and neural ectoderm for the anterior and posterior lobes of the pituitary gland and the pituitary stalk. Six3 expression is indicated at e11.5 in shaded areas. The transcription factors Lhx2 and Tbx3 are critical in the neural ectoderm, and Lhx3 is essential for Rathke’s pouch development. WNT, BMP, FGF and SHH signaling stimulates growth of Rathke’s pouch and anterior pituitary development. Disruption of Six3 expression in Rathke’s pouch with Prop1-cre and in the ventral diencephalon using Nkx2.1-cre reveals that Six3 has critical role in both tissues. The mechanism of action in each tissue has similarities and differences. Six3 is required to activate lineage specific transcription factors in each tissue, i.e. Lhx3 in Rathke’s pouch and Lhx2 and Tbx3 in the ventral diencephalon. Both knockouts caused reduced BMP, FGF and SHH signaling, but failure to suppress WNT signaling (i.e. Axin2 expression) was only detected in the ventral diencephalon knockout.