Introduction

Ocular anterior segment dysgenesis (ASD) describes a spectrum of clinically and genetically heterogeneous congenital disorders affecting anterior structures including the cornea, iris, lens, ciliary body and ocular drainage structures that often lead to impaired vision. Clinical manifestations of ASD are variable and include corneal opacities, cataracts, iris hypoplasia and iridocorneal adhesions. In addition, ASD is associated with extraocular defects [1]. ASD is a multigenic disorder [2], including mutations in several genes such as Cyp1b1, Foxc1, Foxc2, Foxe3, Lmx1b, Maf, Pax6, Pitx2, and Pitx3 [3]. Many of these genes encode transcription factors responsible for development and maturation of cornea and lens [4] . Although these genes have long been recognized in the etiology of ASD, the underlying mechanism remains elusive.

Ocular anterior segment development involves a series of highly orchestrated and complex inductive interactions between tissues derived from different embryonic lineages: the surface ectoderm, neural ectoderm, neural crest and cranial paraxial mesoderm  [5, 6]. It is well known that homeobox genes function as essential transcriptional regulators in a variety of developmental processes [7]. Msh homeobox genes are defined as a group of highly conserved homeodomain proteins that generally function as transcriptional repressors, are expressed at epithelial-mesenchymal transition sites during embryogenesis and are responsible for the development of skull, hair follicles, teeth, heart and brain. The murine Msx family is comprised of 3 members, Msx1, Msx2, and Msx3. Msx3 is absent in the human genome [8,9,10,11,12]. Msx2 acts as an upstream regulator in optic vesicle development [13,14,15]. In our previous study, Msx2 traditional knockout was linked to Peter’s anomaly (a subtype of ASD) and severe microphthalmia, which further substantiates its role in controlling eye development [14, 15]. Changes in FoxE3 and Prox1 expression support the importance of Msx2 in controlling transcription of target genes critical for early eye development [14]. During eye development, Msx2 transcripts first appear in the optic vesicle (OV) and adjacent ectoderm at E9.5. Along with lens genesis, Msx2 transcripts are located in the lens epithelial cells at E10.5 and E11.5 and then expressed in differentiated lens fiber cells after E12.5 [15]. The decrease of Msx2 expression hinders normal surface ectoderm development through several signaling pathways. Further investigation into these complex interactions among Msx2, various transcriptional regulators and signaling molecules may help clarify the pathogenesis of ASD and other congenital eye diseases.

The limitation of our previous study is that Msx2 was knocked out in both surface ectoderm and neuro ectoderm. Microphthalmia and retina malformation were found in Msx2 traditional knockout mice. Moreover, it is hard to rule out the effect of retina-lens sequential induction during eye development, which could lead to misinterpretation of the Msx2 gene function in ocular development. Conditional gene knockout is a technique used to eliminate a specific gene in a specific tissue. The conditional gene knockout technique may eliminate many of the undesired effects induced by traditional gene knockout. Thus, in our present study, we used head surface ectoderm-specific Msx2 gene knockout mice as a model and found that Msx2 deficiency leads to ASD without cornea-lentoid adhesions. Loss of Msx2 in the surface ectoderm down-regulated Gja8 and crystallin expression and up-regulated Tgm2, Capn1, and Camk2b expression in mouse lens. Msx2 therefore acts as an upstream gene of a calcium signaling pathway in ocular development.

Material and methods

Mice

All animal experiments were performed in accordance with approved guidelines and regulations established by the China Medical University (16005 M). We conducted the studies according to the guidelines provided by the Care and Use of Laboratory Animals of China Medical University according the Chinese version of 8th Guide (http://202.118.40.32/sydwb/info/1834/1177.htm) which followed the US Public Health Management Policy. The mice used in this study were housed in the controlled specific-pathogen-free (SPF) environment and cared for according to the approved protocol. Mice used in the experiments were crested on the BALB/c, C57/B6 and FVB background and kept under C57/B6 background [16]. A previous study demonstrated that after Le-Cre+/− was backcrossed to CBA/Ca for seven generations, some Le-Cre+/−; Pax6+/+, as well as Le-Cre+/−; Pax6fl/+ mice exhibited significant eye abnormalities; and after 2 generations of backcrossing Le-Cre+/− mice to the original FVB/N strain, eye abnormalities in Le-Cre+/−; Pax6+/+ mice were diminished [17]. However, in our study, no eye abnormalities were found in the 15th generations of Le Cre+/−; Msx2 fl/fl; and Le Cre+/-; Msx2 fl/+ mice, which is shown in Supplementary Fig. 1. To address whether Msx2 function in lens development is intrinsic to local Msx2 function, we generated Msx2 conditional knockout mice (Msx2 CKO) by crossing Msx2 floxed mice with Le-Cre mice, whose Cre-recombinase expression is regulated by Pax6-Le tissue-specific regulatory elementals which is only active in surface ectoderm, as reported previously [18]. The experimental mice were genotyped by tail polymerase chain reaction (PCR). The Le-Cre heterozygous mice were bred with mice carrying floxed Msx2 alleles (Msx2 flox/flox). The male mice carrying Le-Cre;Msx2flox/+ were subsequently crossed with females carrying Msx2flox/flox to generate Le-Cre; Msx2flox/flox (Msx2 CKO) mice. Littermates were used as the experimental group and littermates carrying Msx2flox/flox (Msx2 control) were served as controls. Msx2 heterozygous transgene mice were not used as the control group. PCR primers for Le-Cre genotyping are as follows: Le-Cre (forward), 5′-TAATCGCCATCTTCCAGCAG-3′, Le-Cre (reverse), 5′-CTCTGGTGTAGCTGATGATC-3′; forward, 5′-GTTGAGCCGAGTCTCCCACCT-3′ and reverse 5′-GATTCCTTGGGCGGCTTCTT-3′) for floxed alleles of Msx2.

Histological preparation of mice embryos and eyeballs

Pregnant female mice were sacrificed at various time points after conception. The mice were anesthetized by sevoflurane and then sacrificed by cervical dislocation before the procurement of their embryos or eyeballs. The embryos and eyeballs were fixed in 4% paraformaldehyde overnight at 4 °C, then dehydrated through graded alcohol, cleared in xylene, and embedded in paraffin. Sections were cut to 4 µm and stained with hematoxylin and eosin. Sections were photographed using an Olympus microscope (BX51, Olympus, Japan) with a SPOT camera. Three mice/six eyes from each group were evaluated to compare eye structures between these two groups and the experiment was repeated three times.

Lens measurement

Electronic balance (Acculab ALC, Germany) was used to measure mice lens quality. After obtaining mice lenses, the residual surface water was removed and the lens were placed in the same EP tube carefully for testing. Six mice/12 lenses from each group were evaluated to compare the lens wet weight of these two groups and was repeated three times.

Whole-mount in situ hybridization

Embryos of various ages were fixed overnight in 4% paraformaldehyde in PBS. Whole-mount in situ hybridization was performed according to standard protocols [19]. The Msx2 cDNA plasmid was purchased from ATCC (Manassas, VA) [18]. All RNA probes were labeled with digoxigenin-UTP according to the manufacturer’s recommendations (Roche Applied Science). Three mice/6 lenses from each group were evaluated to compare Msx2 mRNA expression between these two groups and the experiment was repeated times.

BrdU labeling and TUNEL assay

Pregnant mice were sacrificed at various time points after conception. One hour before sacrificing, the mice were injected intraperitoneally with 100 µg BrdU (Sigma, St. Louis, USA) per gram of body weight. For post-natal mice, BrdU was injected intraperitoneally 2 h before sacrifice [20]. Then the embryos were put into ice-cold PBS.

Apoptotic cells were detected by using the fluorescein in situ Cell Death Detection Kit (Roche Applied Science, Indianapolis, IN). Briefly, 4% PFA-fixed tissue sections were treated with Proteinase K (20 μg/ml) for 20 min. Fragmented DNA was labeled with fluorescein-dUTP, then cell nuclei were counterstained with DAPI. Histological photos were taken by Olympus fluorescence microscope equipped with a Spot CCD camera. Three mice/six lenses from each group were evaluated to detect lens cell proliferation and apoptosis in each of the two groups and the experiment was repeated three times.

RNA-seq analysis and quantitative PCR

Lens RNA was extracted from Msx2 control and Msx2 CKO mice at P60 using RNAeasy™ micro kit (R0024, Beyotime, China). RNA purity was checked using the Nanophotometer spectrophotometer (IMPLEN, CA, USA). RNA concentration was measured using Qubit RNA Assay kit in Qubit 2.0 flurometer (Life Technologies, CA, USA), and RNA integrity was assessed using the RNA nano 6000 Assay kit of the Bioanalyzer 2100 system (Agilent Technologies, CA, USA). The RNA-seq was carried out by Illumina HiSeq sequencing (Novogene Bioinformatics Institute, Beijing, China). Differential expression analysis: prior to differential gene expression analysis, for each sequenced library, the read counts were adjusted by edgeR program package through one scaling normalized factor. Differential expression analysis of two conditions was performed using the edgeR package (3.12.1). The P values were adjusted using the Benjamini and Hochberg method. Corrected P-value of 0.05 and absolute fold change of 2 were set as the threshold for significantly differential expression. Gene Ontology (GO) enrichment analysis of differentially expressed genes was implemented by the clusterProfiler R package, in which gene length bias was corrected. GO terms with corrected P value less than 0.05 were considered significantly enriched by differential expressed genes. KEGG is a database resource for understanding high-level functions and utilities of the biological system, such as the cell, the organism and the ecosystem, from molecular-level information, especially large-scale molecular datasets generated by genome sequencing and other high-through put experimental technologies (http://www.genome.jp/kegg/). We used clusterProfiler R package to test the statistical enrichment of differential expression genes in KEGG pathways.

Quantitative PCR (qPCR) was used to verify the results by using SYBR Premix Ex TaqTM II kit (Takara, China) and analyzed based on the equation RQ = 2−ΔΔCT. The sequences of real-time qPCR primers were listed in Table 1. Ten mice/20 lenses at P60 from each group were evaluated to compare gene expression and the experiment was repeated three times.

Table 1 List of primers used for real-time quantitative PCR
Full size table

Immunofluorescence and immunohistochemistry

After deparaffinization and rehydration, the sections were boiled for 10 min in diluted with ddH2O from ×100 to ×1 antigen repair solution (MXB Biotechnologies, China) and blocked with 5% BSA for 1 h. Anti-Capn1 monoclonal antibody (ab108400, ABCAM, USA, 1:100), Anti-Tgm2 polyclonal antibody (ab421, ABCAM, USA, 1:500), Anti-Camk2b polyclonal antibody (ab34703, ABCAM, USA, 1:200), Anti-Gja8 polyclonal antibody (ab222885, ABCAM, USA, 1:100); Anti-Cryab polyclonal antibody (ab5577, ABCAM, USA, 1:200), anti-Cryba1 polyclonal antibody (PA5-71690, ThermoFisher Scientific, USA, 1:500), anti-Crybb1 polyclonal antibody (bs-12582R, Bioss Antibodies, China, 1:500), and anti-Crybb3 polyclonal antibody (21009-1-AP, Proteintech Systems, USA, 1:100) were used as primary antibody. Anti-Rabbit antibody488 (A-21206, Invitrogen, USA), and anti-Rabbit antibody594 (A-21207, Invitrogen, USA) as secondary antibodies were incubated at room temperature for 2 h. Cell nuclei were counterstained with DAPI. DAB (TA-060-QHDX ThermoFisher Scientific, USA) was used for color development followed by hematoxylin counterstaining in immunohistochemistry assay. Three mice/six lenses from each group were evaluated to compare these proteins expression of these two groups and repeated 3 times.

Statistical analysis

Data were recorded as mean ± standard deviation (SD), and analyzed using SPSS for Windows, version 16.0 (SPSS Inc. IL, USA). Significant difference was evaluated by analysis of unpaired Student’s-test (two-tailed). Statistical significance was defined as indicated in the figure legends.

Results

Targeted disruption of Msx2 on eye surface ectoderm led to ASD

Our previous study using Msx2 germline knockout mice as an experimental model showed that loss of Msx2 can affect eye development [14]. In this study, Msx2 was conditionally deleted using Le-Cre to analyze its function in lens development. Morphological analysis revealed abnormal eye development characterized by eye socket depression, small eye, lack of eyelashes and narrow palpebral fissure in Msx2 CKO mice (Fig. 1A) but not in Msx2 control mice (Msx2 floxed only; Fig. 1B) and Le-Cre heterozygous transgene mice (Fig S1) at postnatal day 60 (P60). Moreover, corneal opacities, iris cornea synechia (virtual box) were found in Msx2 CKO mice by slit lamp under mydriatic condition but not in in Msx2 control mice (Fig. 1C vs. D) and Le-Cre heterogenous transgene mice (Fig S1). Furthermore, cornea stroma thickening; cornea-iris adhesion (virtual box) and disappearance of the anterior chamber (arrowhead) were found in Msx2 CKO mice but not in in Msx2 control mice (Fig. 1E vs. F). Small irregularly shaped lenses and lens opacity (arrowhead) were found in Msx2 CKO mice but not in in Msx2 control mice (Fig. 1G vs. H), which are the typical phenotypes observed in human ASD. To more precisely quantify the abnormality of lens development induced by Msx2 CKO, lens wet weight was compared between Msx2 CKO and control mice at P60. Lenses were significantly smaller in Msx2 CKO mice (1.09 ± 0.11) compared with Msx2 control (6.72 ± 0.05) (Fig. 1I), further confirming the abnormal development of lens after loss of Msx2 function.

Fig. 1
figure 1

Generation of surface ectoderm-specific conditional Msx2 knockout mouse. A–H Representative Msx2 CKO and Msx2 control mice eyes at postnatal-day-60. (A vs. B). Gross examination of Msx2 CKO (arrow) and control eyes (arrowhead). Eye socket depression, small eye, lack of eyelashes and narrow palpebral fissure were seen in Msx2 CKO mice A. C vs. D Msx2 CKO mice indicated corneal opacities, iris cornea synechia (virtual box) was found in Msx2 CKO mice by slit lamp under mydriatic conditions but not in in Msx2 control mice. E vs. F There was cornea stroma thickening; cornea-iris adhesion (virtual box); and the anterior chamber was absent (arrowhead) in Msx2 CKO mice but not in in Msx2 control mice. G vs. H A small irregular shaped lens and lens opacity (arrowhead) were found in Msx2 CKO mice but not in in Msx2 control mice (arrow). Three mice/six eyes from each group were evaluated to compare general phenotype of these two groups; N = 3; Co cornea, Le lens, Ir iris, AC anterior chamber (cavity between posterior cornea, iris, lens). I Lens wet weight of Msx2 CKO was much less than Msx2 control mice at P60. Six mice/12 lenses from each group were evaluated to compare the wet weights; N = 3; ***P < 0.0001

Full size image

Conditional Msx2 knockout specifically affected lens development

To confirm that Msx2 gene conditional disruption can specifically delete Msx2 gene expression in lens, we first isolated total RNA from P60 mice lens and analyzed Msx2 mRNA level by RT-qPCR. Our data show a significant difference in Msx2 mRNA expression level between Msx2 CKO and control mice (Fig. 2A). From E9 to E9.5, Pax6 protein was eliminated in the Le-mutant surface ectoderm (SE) [15], and from E10.5 it was undetectable [15, 17]. Our whole-mount in situ hybridization results showed identical results. Msx2 mRNA transcripts in Msx2 CKO at E9.5 optic vesicle (arrow). However, transcripts were seen in the developing lens vesicle from E10.5 to E12.5 in the Msx2 control group (arrow), but was not found in Msx2 CKOs (arrowhead) (Fig. 2B). These results confirmed that the tissue specificity of conditional Msx2 knockout.

Fig. 2
figure 2

Verification of Msx2 deficiency in Msx2 CKO mice. A RT-qPCR quantification of Msx2 mRNA of Msx2 CKO and control mice lens at postnatal-day-60. gapdh serves as a loading control. Msx2 CKO is normalized to Msx2 control which was arbitrarily set as 1. Msx2 mRNA expression iswas absence in Msx2 CKO compared to Msx2 control group. Ten mice/20 eyes from each group were evaluated to compare Msx2 expression; N = 3; ****P < 0.00001. B Whole mount in situ hybridization shows that Msx2 mRNA transcripts were not detectable in Msx2 CKO at E9.5 optic vesicle (arrow) and absence of Msx2 mRNA in the Msx2 CKO lens vesicle (arrowhead) compared to Msx2 control group (arrow) at E10.5, E11.5, and E12.5. Three mice/6 eyes from each group were evaluated to compare Msx2 mRNA; N = 3

Full size image

Next, we compared the phenotype induced by traditional Msx2 knockout mice (Msx2 KO) and Msx2 CKO at E14.5 and P60 . Consistent with our previous findings [14], Msx2 KO eyes possessed defective lenses (Fig. 3C arrow). The eye phenotype of Msx2 CKO was less severe than that of Msx2 KO mice at E14.5 (Fig. 3B vs. C). Msx2 KO mice at E14.5 demonstrated much smaller eyes with thickened retina. (Fig. 3D). In contrast, Msx2 CKO mice at E14.5 displayed slightly thickened retina, but much less severe than the Msx2 KO (Fig. 3E vs. F). Smaller lenses with thickened retina (arrowhead) and abnormally proliferating primary vitreous (arrow) were observed in Msx2 KO mice at E14.5 (Fig. 3F). Smaller lens, uneven thickness and detached retina were detected in Msx2 CKO and Msx2 KO mice at P60, compared with the Msx2 control group (Fig. 3H, I vs. G). Retinal folds and detachment were observed in Msx2 KO retina at P60 (Fig. 3I). Therefore, it appears that Msx2 CKO specifically disrupts lens development while minimally affects other eye structures; this is most apparent at E14.5.

Fig. 3
figure 3

Msx2 CKO mice showed defective lens only, while the entire eye was abnormal in KO mice. Lens morphology and histology of Msx2 control, Msx2 CKO and Msx2 KO mice at E14.5 and P60. A–C Phenotypes of Msx2 CKO mice, Msx2 KO mice and control mice. Whole mount embryo of Msx2 CKO mice show a slightly smaller eye (B). Msx2 KO mice with much smaller eyes and severe ocular abnormality (arrow) (C). D–F Histology analysis of ocular development at E14.5 of Msx2 CKO mice, Msx2 KO mice and control mice. Msx2 CKO mice at E14.5 stage displayed a little bit thickened retina but much less severe than Msx2 KO (E vs. F). Bottom panels showed the enlarged view of the boxed regions (D–E). Smaller lenses with thickened retina (arrowhead) and abnormally proliferated primary vitreous (arrow) wascould be observed in Msx2 KO mice (F). Small lenses were seen in Msx2 CKO mousee eyes (E). G–I Histology analysis of control mice, Msx2 CKO and Msx2 KO mice eyes at postnatal-day-60. Smaller lens, uneven thickness and detached retina detected in Msx2 CKO and Msx2 KO mice at P60 compared with Msx2 controls (H, I vs. G). Retinal folds and detachment were observed in Msx2 KO retina at P60 (I). Bottom panels showed an enlarged view of the boxed regions (G–I). Three mice/six eyes from each group were evaluated to compare eye structures; N = 3; Co cornea; Le lens; Ir iris, Re retnia; Scale bars: 100 um (D–F), 200 um (G–I)

Full size image

Abnormal lens development in Msx2 CKO mice at early embryonic stages

Next, we systemically investigated Msx2 CKO mice lens development by gross examination and histological analysis of Msx2 CKO mice lens at different embryonic development stages. Subtle changes in the eyes of the Msx2 CKO mice were first detected as early as E12.5, compared with littermate controls (Fig. 4A, B). Examination of histologic sections revealed that the lens of the Msx2 CKO mice appeared to be smaller but completely developed from E12.5 to E18.5 (Fig. 4A, B). After delivery, abnormal development of eyes in the mutants could be easily recognized as the anterior expansion of the iris pigmented epithelium. Smaller lens and forward movement of lens were found from P2 (Fig. 4A, B). In severe case, lens and cornea adhesion was found after delivery. The severity of lens and cornea defects in the eyes of Msx2 CKO mice varied among animals and even between eyes of the same animals.

Fig. 4
figure 4

Lens development in Msx2 CKO prenatally and postnatally. A Morphological and histological analysis of eyes in Msx2 CKO and control mice from E12.5 to P8. The boxed area is enlarged in B. C Schematic diagram showing how to measure anteroposterior axis (D), rectilinear axis (E) and corneal thickness (F). Measurement of anteroposterior axis and rectilinear axis of lens in Msx2 CKO and control mice from E12.5 to P8 (D–E). N = 3; *P < 0.05, **P < 0.001, ***P < 0.0001, ****P < 0.00001. F Measurement of corneal thickness of Msx2 CKO and control mice from P2 to P8. Three mice/six eyes from each group were evaluated to compare eye structures; N = 3; Scale bars: 100 um (A), 50 um (B); ***P < 0.0001

Full size image

In order to more clearly demonstrate lens defects caused by loss of Msx2 function, lens size was quantified by measuring the anteroposterior and horizontal diameters in both groups from E12.5 to P8. A schematic diagram (Fig. 4C) shows the measurements of anteroposterior axis (D), rectilinear axis (E) and corneal thickness (F). There was statistically significant difference in anteroposterior diameter from E14.5 to P4 between these two groups (Fig. 4D). Similarly, horizontal diameter difference between Msx2 control and CKO mice was also observed later, at E16.5 (Fig. 4E). We also examined and quantified the corneal thickness of mice from E12.5 to P8 and did not observe a statistical difference between the two groups before birth (data not shown). However, after birth, corneal thickness was significantly decreased in Msx2 CKO mice (Fig. 4F). In summary, Msx2 CKO-induced aberrant lens development was observable at early embryonic stages.

Abnormal proliferation and apoptosis in Msx2 CKO lens

In order to determine the phenotypes of the Msx2 conditional knockout, lens cell proliferation and apoptosis were analyzed by immunofluorescence. BrdU positive cells (arrow in Fig. 5A) were observed at anterior lens epithelium cell as early as E12.5 in two groups. From E12.5 to E16.5, the percentage of BrdU-positive lens epithelial cells in both two groups was significantly increased (Fig. 5C). At E16.5, the number of BrdU positive lens epithelial cells was significantly reduced in Msx2 CKO mice compared with Msx2 control mice (Fig. 5A, C). From E16.5 to P8, the percentage of BrdU-positive lens epithelial cells in both two groups was decreased (Fig. 5C). The ratio of BrdU-positive cells at P8 between the two groups was virtually identical (Fig. 5C). The percentage of BrdU positive lens epithelial cells in Msx2 CKO mice was significantly less than the control group, especially from E16.5 to P2 (Fig. 5A, C). Msx2 CKO mice have fewer BrdU positive cells than Msx2 control mice at each developmental stage (Fig. 5C). Lens cell apoptosis rates increased from E12.5 to E14.5 and then decreased with age until P8 in Msx2 CKO mice (Fig. 5D). The apoptosis rate was significantly decreased in Msx2 CKO mice at each stage compared with the control group. Few apoptotic lens cells were found in Msx2 control mice from E12.5 to P8 (Fig. 5B, D). In contrast, Msx2 CKO mice have more apoptotic lens cells at each stage (Fig. 5D). These results indicate that Msx2 deficiency impairs lens development by suppressing cell proliferation and promoting cell apoptosis.

Fig. 5
figure 5

Defective proliferation and enhanced apoptosis of lens in Msx2 CKO mice. A BrdU staining of Msx2 CKO and Msx2 control mice eyes at E16.5 and P2. Arrows indicate the BrdU positive lens epithelial cells. Bottom panels show the enlarged view of the boxed regions in P2 panels. The percentage of BrdU positive lens epithelial cells is calculated in C. B Lens cell apoptosis was determined by TUNEL assay at E16.5 and P4. Arrows indicate the apoptotic lens epithelial cells. Bottom panels show the enlarged view of the boxed regions in P4 panels. The percentage of apoptosis cells is calculated in D. Three mice/six eyes from each group were evaluated to compare BrdU+ staining and lens cell apoptosis; N = 3; Scale bars: 100 um; *P < 0.05, **P < 0.001, ***P < 0.0001

Full size image

Variation of gene expression in Msx2 conditional knockout lens

In order to gain a deeper understanding of how lens phenotypes result from conditional Msx2 deletion, we conducted RNA-seq analysis of lenses isolated from p60 Msx2 CKO and Msx2 control mice. The RNA-seq data and protocols have been submitted to NCBI’s Gene Expression Omnibus (GEO) database with a GEO accession number of GSE114854. The volcano plot detected 1911 differentially expressed genes (DEGs) in Msx2 CKO mice compared to the control mice. The selection of these 1911 genes was based on greater than 2-fold change in expression of genes between Msx2 CKO and Msx2 control mice. Among them, 1586 genes were upregulated (red) and the rest were downregulated (green) (Fig. 6A). Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway and Reactome Pathway were applied to analyze differentially expressed mRNAs. GO enrichment analysis comprised 3 structured networks including biological processes (BP), cellular components (CC) and molecular function (MF). Among them, eye development in BP, ion channel complex in CC and ion channel activity, calcium ion binding and structural constituent of lens in MF closely regulate lens development (Fig. 6B) [21, 22]. Pathway analysis showed that 20 pathways were highly enriched among the abnormally expressed mRNAs (Fig. 6C). Among them, calcium signaling pathway was directly related to the lens opacification [23], ranking the fifth and associated with 36 DEGs. Reactome pathway analysis also confirmed that gene Msx2 might regulated eye development through visual photo transduction pathway (Fig. 6D). Reactome enrichment analysis classified 18 genes (Rcvrn, Gnat1, Rho, Gngt1, Pde6g, Sag, Pde6a, Guca1a, Guca1b, Slc24a1, Gucy2e, Grk1, Pde6b, Rgs9bp, Gucy2f, Rgs9, Gnb5, Gnb1) into visual phototransduction category. There is no enough reference about the expression and function of these genes in lens. Future work is needed to clarify their expression and function in lens. Therefore, Msx2 deficiency dramatically disturbed the gene expression profile of developing lens.

Fig. 6
figure 6

RNA-seq analysis of Msx2 CKO and Msx2 control mice lens. A Volcano plot of RNA-Seq analysis of lens isolated from p60 Msx2 CKO and Msx2 control mice. The volcano plot detected 1911 differentially expressed genes (DEGs) in Msx2 CKO mice compared to the control mice. Among them, 1586 genes were upregulated (red) and the rest were downregulated (green). B–D Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) Pathway and Reactome Pathway were applied to analyze differentially expressed mRNAs. The red boxes highlighted the related pathways

Full size image

Dysregulated calcium signaling pathways after Msx2 conditional deletion

Previous studies have demonstrated that a calcium signaling pathway is essential for lens development [23]. To verify the RNA-Seq analysis results, we performed qPCR to determine the mRNA expression of Calpain 1 (Capn1), Transglutaminase 2 (Tgm2), calcium/calmodulin-dependent protein kinase 2b (Camk2b), and Gja8 at P60 in lenses of Msx2 CKO mice and Msx2 control mice (Table 1). We found that Capn1, Tgm2, and Camk2b were significantly upregulated in Msx2 CKO mice compared to control mice, and Gja8 expression was significantly downregulated (P < 0.01) in Msx2 CKO mice (Fig. 7A). These results were also verified by immunofluorescence assays of lenses from Msx2 CKO and Msx2 control lenses which were performed at E14.5. Capn1 expression slightly increased in differentiated lens cells from Msx2 CKO mice. Tgm2 was expressed in all lens cells, and its expression increased in Msx2 CKO mice at E14.5. Camk2b expression increased mainly in differentiated lens cells compared with Msx2 control mice. However, Gja8 expressed at both lens epithelium cell and lens fiber cells, and its expression decreased in Msx2 CKO mice (Fig. 7B). The above data indicate that Msx2 is vital for a calcium signaling pathway in lens development.

Fig. 7
figure 7

Expression of Capn1, Tgm2, Camk2b, and Gja8 in Msx2 CKO and Msx2 control group lens. A qPCR detection of the mRNA expression of Capn1, Tgm2, and Camk2b are upregulated and Gja8 is downregulated within Msx2 CKO lens compare to the control group at P60. gapdh serves as loading control. Msx2 CKO is normalized to Msx2 control which is arbitrarily set as 1. Ten mice/20 eyes at P60 from each group were evaluated to compare gene expression; N = 3; ***P < 0.0001. B Immunohistochemistry analysis protein expression of Capn1, Tgm2, Camk2b are upregulated and Gja8 is downregulated in Msx2 CKO lens compare to Msx2 control mice at E14.5. Right panels show the enlarged view of the boxed regions in the left panels. Three mice/six eyes at E14.5 from each group were evaluated to compare protein expression; N = 3; E embryonic, P postnatal, Scale bars: 400 um, 200 um

Full size image

Abnormal lens crystallins expression in Msx2 conditional knockout mice

Among the dysregulated genes found by RNA-Seq in Msx2 CKO mice, crystallin family members are notable because they are highly expressed during lens development. To verify the RNA-Seq analysis result mRNA level of crystallin ab (Cryab), crystallin ba1 (Cryba1), crystallin bb1 (Crybb1), and crystallin bb3 (Crybb3) were determined by RT-qPCR. As expected, all of them were downregulated in Msx2 CKO mice (Fig. 8A). Consistently, downregulation of Cryab, Cryba1, Crybb1, and Crybb3 proteins were found throughout the whole lens of Msx2 CKO mice (Fig. 8B). Since crystallin function as lens structure proteins, loss of crystallin affected lens structure and morphology.

Fig. 8
figure 8

Abnormal crystallin expression and defective lens architecture in Msx2 CKO lens. A Relative expression of Cryab, Cryba1, Crybb1 and Crybb3 are downregulated in Msx2 CKO compared to Msx2 control lens at P60 as assessed by RT-qPCR. Ten mice/20 eyes at P60 from each group were evaluated to compare gene expression; N = 3; **P < 0.001, ***P < 0.0001. B Cryab, Cryba1, Crybb1, and Crybb3 proteins in Msx2 CKO and Msx2 control lens at P60 are detected by immunohistochemistry. They are all downregulated in Msx2 CKO mice lens fiber cells compare to the control group. Three mice/ six eyes at E14.5 from each group were evaluated to compare protein expression; N = 3; Scale bars: 400 um, 100 um

Full size image

Discussion

Msx2 plays an important role in multiple organ development [8, 24,25,26,27,28,29]. To investigate possible functions of Msx2 in early ocular development, a precious study using transgenic mice overexpressing Msx2 found that forced expression of the Msx2 gene resulted in optic nerve aplasia and microphthalmia in all transgenic mice [13]. Marker analysis showed suppression of Bmp4 and induction of Bmp7 expression in the optic vesicle. In our previous study, germline knockout Msx2 in mice led to microphthalmia or anophthalmia, corneal and lens dysgenesis resembling Peters anomaly and microphthalmia can be seen in humans [14]. Lens vesicle growth and development were affected by Msx2 traditional knockout mice. Moreover, loss of Msx2 caused FoxE3 and Prox1 expression changes that further provided evidence of the important role of Msx2 in regulating ocular development [14].

After the optic vesicle forms at E9.5, the neuroectoderm thickening in the lateral wall of the optic vesicle is destined to become the neural retina [13]. The corresponding surface ectodermal thickening becomes the lens placode. According to a previous study, Msx2 gene expression was detected at optic vesicle and adjacent ectoderm at E9.5 [14]. Therefore, traditional knockout of Msx2 gene perturbed lens and retina development simultaneously. As previous studies have shown, induction between lens vesicle and optic vesicle exists through the ocular development process [13, 14]. We could not distinguish whether dysregulated Msx2 gene expression or abnormal lens-retina induction leads to lens dysgenesis in Msx2 traditional knockout mice. Use of Le-Cre mice, which is under the control of the lens ectoderm promoter of Pax6, helped us eliminate the effect of abnormal retina development caused by lens-retina induction, and demonstrate lens development changes when a gene was conditionally deleted at the surface ectoderm.

In this study, using conditional Msx2 knockout mice, we demonstrated that Msx2 was a significant contributor to lens development. Without the Msx2 gene in the surface ectoderm, mice were born with phenotypes consistent with ASD and congenital cataract. Compared with traditional Msx2 knockout mice, Msx2 CKO mice showed moderately abnormal phenotype during ocular development. The mouse eyes appeared moderately small, although all the ocular tissue was developed [14]. In addition, the corneal thickness decreased, and lens opacity and dysgenesis were observed. Moreover, we observed lens cell proliferation and apoptosis from E12.5. Compared with the control group, lens cell proliferation in Msx2 CKO mice significantly decreased through the lens development process. A significantly higher apoptosis rate was seen in in Msx2 CKO lens, consistent with the results found in conventional Msx2 KO mice [14]. A previous study showed that germline knockout of Msx2 may result in the persistent presence of lens stalk, due to increased proliferation rate in anterior lens epithelial cells after E14.5 [14]. However, lower cellular proliferation and ahigher apoptosis rate in the lens epithelial cells were found in Msx2 CKO mice, which might explain why Msx2 CKO mice did not develop lenses as small as those in the KO mice, and why the lens separated from the cornea at the appropriate developmental stage.

Further investigation into the complex interactions among Msx2 and various transcriptional regulators and signaling molecules in ocular development may help clarify the pathogenesis of ASD. We conducted the RNA-seq analysis on P60 eyes in order to determine the detailed mechanism of Msx2 conditional deletion in regulation of lens development. When we set the threshold at a 2-fold or greater change in expression of genes between Msx2 CKO and Msx2 control mice, 1911 DEGs were detected. Among them, 1586 genes were upregulated and the rest were downregulated. According to the RNA-seq results, calcium signaling pathway was one of the most dysregulated pathways in Msx2 CKO mice. Previous studies showed that calcium controls lens cell homeostasis and lens development [30,31,32]. Intracellular calcium homeostasis requires normal cellular gap junctions [33,34,35]. Gja8 (Cx50) and two alpha connexin family members, are expressed in ocular lens [36, 37]. Gja8 is an important gap junction factor for lens development and is responsible for calcium coupling and regulating Ca2+ in lens fiber cell membranes. Substitution of aspartate-47 (D47) of Gja8 has been linked to an autosomal dominant congenital cataract in several human pedigrees [38]. Smaller lenses and lens opacity were found in Gja8 mutant mice [39, 40]. Gja8 expression was down-regulated when Msx2 was conditionally deleted in mouse lens.

We also found that Capn1, Tgm2, Camk2b were up-regulated in Msx2 CKO lens, which maybe the results of increased calcium level in lens cells. Fodrin, filensin and Vimentin were known as substrates of calpain in lens [41, 42]. Calpain activation caused by intracellular calcium overloading was associated with several pathological conditions, including cataracts in animals [43]. Calpain-mediated proteolysis of crystallin may led to increased light scatter [44]. Moreover, alpha-crystallin, beta-crystallin, and vimentin can be cross-linked by Tgm2 when it is up-regulated [45, 46]. Tgm2 catalyzed dimerization of alpha-crystallin, which might be a key step of initiating protein aggregation.

Previous studies have shown that calmodulin (CAM) directly interacts with aquaporin 0 (AQP0) C-terminus in a calcium dependent manner to regulate water permeability of AQP0 [47, 48]. Another study identified a missense mutation (p.R233K) in the putative CAM binding domain of AQP0 C-terminus in a congenital cataract family [49]. Our results demonstrate that Camk2b was upregulated in Msx2 CKO mice, and the Camk2b upregulation might lead to CAM hydrolysis and affect AQP0 indirectly.

The shortcoming of this study is that P60 lenses were selected for RNA-sequencing, but defects in lens development were observed as early as E10.5 in our previous study [15] and more prominently at birth. At different stages of lens development, gene regulation and expression varied and the results at P60 were dissimilar to those at E14.5 . While RNA-seq at P60 identifies interesting potential transcriptional changes in Msx2 CKO lenses RNA-seq performed at an earlier stage could closer to the onset of the phenotype which will reveal more direct downstream targets of Msx2. The purpose of RNA-sequencing in this study was to provide clues to figure out the possible gene downstream genes regulated by Msx2 at early stage. Future research should include obtaining additional lens samples at earlier developmental stages.

In summary, the current work demonstrates that conditional deletion of Msx2 gene at the surface ectoderm influenced lens cell proliferation and apoptosis, leading to ASD and lens opacity. We provide the first direct genetic evidence that Msx2 gene influences lens development through a calcium signaling pathway. The results further support the importance of Msx2 in eye development. This study provides important additional understanding of the Msx2 gene function. In the future, we may recommend genetic testing of Msx2 mutations for patients with a clinical diagnosis of ASD. Further studies in investigating interactions between calcium signaling pathway and Msx2 may provide further insight into lens development.