Introduction

Fraser syndrome is a rare autosomal recessive disorder characterized by developmental malformations evident before birth1. Individuals affected may exhibit various features such as cryptophthalmos (fused eyelids), syndactyly (fused fingers and toes), and unilateral or bilateral renal agenesis (failure of kidney development), alongside respiratory and ear abnormalities2. Despite its rarity, Fraser syndrome may contribute to early-term miscarriages, mainly due to renal or pulmonary complications at the embryonic stage1. Patients without life-threatening phenotypes can survive into adulthood and may undergo surgical interventions to resolve skin malformations such as syndactyly.

FRAS1, FREM1, and FREM2 are structurally similar proteins shown to serve as important components of the extracellular matrix (ECM) protein complex3. These proteins are predominantly localized in the skin within the sublamina densa, a component of the basement membrane zone between the epidermis and dermis of the skin, playing a critical role in preserving epithelial-mesenchymal integrity. GRIP1 is involved in trafficking the ECM proteins to their correct location, making it a crucial protein for mediating organ morphogenesis. Although it is not included in the ECM protein complex, it was suggested to play an important role in maintaining the structural integrity of tissues4. Pathogenic variants in the FRAS1, FREM1, FREM2, and GRIP1 genes cause epithelial detachment at the level of the sublamina densa3,4.

Previous studies reported that pathogenic variants in FRAS1, FREM2, and GRIP1 are causative of Fraser syndrome; however pathogenic variants in FREM1 alone do not induce the disorder5. Nonetheless, pathogenic variants in these genes result in similar phenotypes. The severity of these phenotypes in patients, caused by deleterious pathogenic variants or the absence of ECM proteins and GRIP1 highlights their requirement for assembling basement membranes across critical organs such as the skin, kidneys, testes, and trachea during embryogenesis3. Mouse models have proven an excellent tool for investigating the roles of key proteins in embryonic development and explaining their phenotypic correlations. Although FRAS1, FREM1, and FREM2 are predicted to be structurally similar, genetic studies in mouse models indicate that they are functionally non-redundant; loss of any component destabilizes the entire basement membrane complex, suggesting that these proteins are likely to form a complex and reciprocally stabilize one another at the epithelial basement membrane6.

Previously reported mouse models carrying mutations in this group of genes have displayed similar blistering phenotypes, often referred to collectively as “bleb” mutants3. Loss of function in Grip1 results in blistering over the eyes, mimicking the phenotype of the eye blebs (eb) mutant strain5. Mutations in Fras1 in mice produce paw blebs (bl) in the Fras1bl/bl mice, while Frem1 mutations result in head blebs (heb) in mice, often presenting as head blisters and malformed or absent eyes at birth7. The Myelencephalic bleb (my) mouse line, identified as early as the 1920s and currently available from The Jackson Laboratory, has been linked to the Frem2 gene, although the underlying genetic lesion remains undefined, as no coding region mutations were found in Frem26. Reduced FREM2 expression was observed in Frem2my/my embryos, but it remains unclear whether this results from cis-regulatory disruption or transcript instability, and the allele has not been confirmed as a functional null6. Another strain, carrying the Frem2-myUCL allele caused by a random transgene insertion event, was reported to localize to the Frem2 gene; however, the precise mutation has not been identified, the transgene was not detected, and no coding region mutations were found in Frem28. Fras1- and Grip1-knockout mouse models also show early-onset blistering, detectable by embryonic day 12–13, and both exhibit embryonic lethality9,10.

Mutations in Frem2 have been associated with a spectrum of Fraser syndrome-like phenotypes, including cryptophthalmos, epithelial blebbing, blood-filled blisters, renal agenesis, and bony syndactyly8,11. Among the previously reported Frem2 mouse models are the compound heterozygous Frem2R725X/R2156W mice, which carry patient-derived variants associated with cryptophthalmos, and the Frem2my-F11 mice, which harbor a nonsense mutation resulting in a premature stop codon in Exon 5 of 2411,12. Most, if not all, of these models carry variants predicted to produce truncated FREM2 protein, or single amino acid substitutions, both of which may retain partial function (Supplementary Table 1). These alleles are often associated with milder phenotypes, with the deficit limited to specific cell types or tissues, such as the cryptophthalmos phenotype of the Frem2R725X/R2156W mouse model11. Therefore, a constitutive Frem2 knockout (KO) mouse model is needed to fully elucidate FREM2’s role in development and to generate a more complete phenotypic picture of Fraser syndrome.

To assess whether the absence of FREM2 recapitulates the reported phenotypes in Frem2 mutants, or perhaps, results in a more exacerbated phenotype, we developed and characterized a constitutive Frem2-KO mouse model. Upon anatomical and histological analysis, we found that the Frem2-KO mice exhibit neonatal mortality, which we associate with bilateral renal agenesis. In addition, the Frem2-KO fetuses develop blood-filled blisters on the eyes and paws that progress into hemorrhages and missing eyelids. A single Frem2-KO mouse survived to adulthood and displayed unilateral renal agenesis, while exhibiting syndactyly, cryptophthalmos, and microphthalmia. Our findings confirm FREM2’s role as an important protein for the formation of the skin, kidneys, and eyes. This Frem2-KO mouse model provides a valuable tool for further, more detailed prenatal investigation of its critical roles in organ development and underlying phenotypes resulting from its absence.

Results

Genetic analysis of Frem2-KO mouse model

Mouse ES cell clones were purchased from the European Conditional Mouse Mutagenesis Program and used to generate the Frem2tm1a(EUCOMM)Hmgu mouse line carrying the knockout-first allele with conditional potential (F2KCR allele) (Fig. 1A). First, the FRT-flanked Neo cassette was excised by crossing with a FLP deleter strain to generate the conditional Frem2fl/fl allele. Although the generated mouse line was designed to carry a floxed Frem2 allele to enable cell-type-specific FREM2 functional studies (Fig. 1A), after several breeding steps with pan-Cre and tissue-specific Cre lines (see Methods), we were unable to generate adult homozygous Frem2-floxed mice. Upon further investigation of the mouse genetics, and the sequencing results of the insertion site of the floxed-Frem2 mouse, we discovered mutations within the targeted insertion site. These included two in-frame insertions (27 bp and 30 bp) and one in-frame deletion (9 bp) (Fig. 1A). Subsequently, we analyzed the open reading frame in exon 1 of the floxed Frem2 allele. We identified a premature stop codon (Fig. 1B) within exon 1 of the Floxed-Frem2, which likely results in a truncated or missing FREM2 protein, converting the floxed-Frem2 allele effectively into a constitutive null allele. All animals evaluated in this study were obtained from Frem2+/- x Frem2+/- crosses.

Fig. 1
figure 1

Frem2-KO mouse design strategy. A ES cells carrying the Frem2-KO-first, conditional-ready allele was used to generate the Floxed-Frem2 mouse line. The floxed allele shows the FRT-flanked Neo cassette (FLP and FRT system) inserted into exon 1, dividing it into 1a and 1b regions. The conditional-KO mouse harbors the following in-frame mutations (indicated as a red line): two insertions (27 bp and 30 bp) and one deletion (9 bp), with sequences provided below. B A portion of the sequence (15,100–15,120 bp) from exon 1 of Floxed-Frem2 is shown. The premature TGA stop codon sequence is highlighted in blue near the middle of the reading frame.

Full size image

Frem2-KO mice have hemorrhagic blisters and skeletal malformations on their paws

Upon our examination of fetuses (E13–16), we immediately observed the skin defects on their paws, in agreement with previous studies reporting FREM2’s critical role in epidermis development. Hemorrhagic blisters were observed on the digits of Frem2-KO fetuses at E15 and E16 (Fig. 2A, B) which allowed us to phenotype them when compared to Frem2 wild-type and heterozygote fetuses that appeared normal. To observe the postnatal development of the limbs, newborn Frem2 pups were collected. When we assessed blood-filled blisters on the digits of newborn Frem2-KO pups, they were still present but often appeared dry (Fig. 2C, arrowheads).

Fig. 2
figure 2

Frem2-KO fetuses and newborn pups exhibit hemorrhagic blisters and syndactyly. A Frem2+/+(wild-type) E15 fetus (left) with normal paws (arrow) and a Frem2-/- (KO) littermate (right) with a blood-filled blister over the digits (arrowhead). B E16 heterozygous Frem2+/- fetus (left) with normal paws (arrow) and E16 Frem2-/- fetus littermate (right) with a large hemorrhagic bleb on the digits (arrowhead). C P0 Frem2+/+ (left) and Frem2+/- (middle) have normal paws, and a Frem2-/- (right) has dried blood-filled blisters on both hind paws. D Newborn Frem2-/- pups with a malformation of hind limbs, soft-tissue syndactyly, and blood-filled blisters on digits. Reference Frem2+/+ and Frem2+/- paws showing normal paw morphology. E Hematoxylin and eosin-stained histology section of paws from wild-type (Frem2+/+) and several knockout (Frem2-/-) mice with limb malformation and blood-filled blisters (arrowhead). Scale bars: A–C: 2 mm, D, E: 1 mm.

Full size image

In newborn Frem2-KO pups, most hind paws revealed soft tissue syndactyly, independent of whether any blisters were observed. Affected paws often also displayed dorsal flexure, an anatomical malformation where the paws curl upwards (Fig. 2D). We next assessed the skeletal formation of the paws and other morphological features using histological sections, which were prepared in the same orientation across all samples (Fig. 2E). We determined that newborn Frem2-KO mice with paws that had blood-filled blisters, syndactyly, or dorsal flexure also exhibited skeletal malformations in the digits, with the surrounding skeletal structures often compromised. Interestingly, these regions of skeletal malformations typically occurred bilaterally on the hind limbs.

FREM2 deficiency causes ocular abnormalities in Frem2-KO mice

Frem2-KO fetuses consistently exhibited pronounced bubble-like blood-filled blisters over their eyes as early as E15 (Fig. 3A). Coronal histological sections of E15 Frem2-KO fetus heads localized the blisters near or within the eyelids (Fig. 3B, C). In some cases, the eyelids in Frem2-KO fetuses were thinned and hemorrhaged (Fig. 3D, E). Such hemorrhages and blood-filled blisters appeared bilaterally or unilaterally on the eyelids of Frem2-KO fetuses (Fig. 3E, Supplementary Table 2).

Fig. 3
figure 3

Frem2-KO mice lack eyelids and develop blood-filled blisters covering their eyes. A Gross image of an E15 Frem2+/+ fetus with a normal eye. A Frem2-/- fetus (below) with a blood-filled blister (arrowhead) covering the eye. B Hematoxylin and eosin-stained coronal head section of a Frem2+/- E15 fetus and a higher magnification image below display normal eyelid morphology. C Histological section of a Frem2-/- E15 fetus with a blood-filled blister, and a higher magnification image displays the absence of the cornea and eyelid. D Low- and high-magnification histology images of a Frem2-/- E16 fetus show a normal right eyelid and a hemorrhage covering the left eyelid. E Gross images of Frem2+/+ and Frem2+/- P0 pups show normal eyelids. F Gross images of Frem2-/- P0 pups display missing eyelids (arrowhead). G Histology sections of P0 Frem2+/+ and Frem2+/- heads show normal eyelids, with corresponding higher magnification images of the right (R) and left (L) eyes presented below. H Histology section of a P0 Frem2-/- pup shows thinning of the eyelid (R) and a missing eyelid with a hemorrhage over the eye (L). I Images from a newborn Frem2-/- pup show a missing eyelid (R) and thinning of the eyelid (L). J Images from a Frem2-/- pup show a missing eyelid (R) and a normal contralateral eyelid (L). Scale bars: A: 2 mm, B–G: 1 mm, H–J: 1 mm.

Full size image

In newborn Frem2-KO pups, however, no blood-filled blisters were observed. Instead, Frem2-KO pups predominantly displayed missing eyelids, with some hemorrhages often observed on the skin around the eyes (Fig. 3F). The heads of newborn pups were sectioned coronally to assess eye morphology using hematoxylin and eosin staining. In wild-type and heterozygous Frem2 pups, the epidermal layers of the eyelids were properly formed (Fig. 3G). In Frem2-KO pups, however, the eyelids were often missing (Fig. 3H, I) or thinned (Fig. 3I, J). Periocular hemorrhages and the absence of eyelids were bilateral or unilateral in Frem2-KO pups. Blood-filled blisters on the eyelids that occur during embryonic stages, as well as the missing eyelids in newborn Frem2-deficient animals, suggest that proper FREM2 expression is critical for the epidermal development of the eyelids.

Frem2-KO mice die hours after birth

During routine genotyping of Frem2 litters across several years of the study, no Frem2-KO mice were identified. Interestingly, the dam was observed giving birth to pups exhibiting phenotypes of Frem2-KO embryos. However, a few hours after birth, the pups were discovered dead in the cage. We thus concluded that Frem2-KO pups do not survive after birth, and next sought to investigate the causes of this mortality. Multiple histological sections and post-mortem necropsies were performed in newborn Frem2-KO pups from 8 litters. In almost all newborn Frem2-KO pups we observed bilateral renal agenesis (Fig. 4 and Supplementary Table 2). Likely a result of the lack of urine production from renal agenesis, Frem2-KO pups had empty urinary bladders. We also commonly observed the absence of one or both adrenal glands (Fig. 4A, B). Parasagittal sections further confirmed the renal agenesis and the empty urinary bladder in newborn Frem2-KO pups (Fig. 4C–F).

Fig. 4
figure 4

Renal agenesis in Frem2-KO pups. Macroscopic images of P0 pups show kidneys (K), adrenal glands (AD), and urinary bladder (UB) present in a Frem2+/+ pup (A), and kidneys and an inconspicuous urinary bladder in a Frem2-/- pup (B). Hematoxylin and eosin-stained parasagittal sections of a Frem2+/+ (left column) and a Frem2-/- (right column) P0 pup show organs present to the left (C, D) and right (E, F) of the spinal cord. Locations of the deflated urinary bladders are labeled with asterisks. Sections are from similar regions to compare the presence and morphology of organs. Scale bars: A, B; 1 mm, C–F; 3 mm. Labels: AD adrenal glands, K kidney, H heart, LV liver, LUN lung, I intestines, S stomach, UB urinary bladder.

Full size image

We carried out serial transverse histological sections to increase confidence in our macroscopic results (Fig. 5 and Supplementary Movies). All Frem2-KO pups, except one, had bilateral renal agenesis, confirmed by the necropsies or with histological evaluation. A single newborn Frem2-KO pup displayed one small but developed kidney, confirming unilateral renal agenesis (Fig. 5D). The functional capacity of that kidney and chances of that animal’s survival remain unknown, although a single surviving Frem2-KO pup had been observed in previous years. All examined Frem2-KO pups also had empty urinary bladders. No major defects in other vital organs, such as the lungs and heart, were seen in Frem2-KO pups upon examination with necropsy or histological analysis. Lungs were submerged into phosphate-buffered saline to assess whether they were filled with air and were confirmed to float, suggesting that the animals were breathing before death. We conclude that the likely cause of postnatal mortality of Frem2-KO pups is renal agenesis.

Fig. 5
figure 5

Serial sectioning of newborn (P0) Frem2 pups reveals renal agenesis in knockout mice. Sections down the columns progress below the diaphragm to the pelvic region from the same pup. Sections across each row correspond to the same region of each pup’s body. Hematoxylin and eosin-stained histology sections from Frem2+/+ (A) and Frem2+/- (B) pups exhibit the presence of both kidneys (KID) and a comparable arrangement of organs. Frem2-/- pups display bilateral (C) and unilateral renal agenesis (D). Locations of missing kidneys are labeled with asterisks. Scale bar: A–D 1 mm. Labels: AD adrenal glands, KID kidney, LV liver, LUN lung, I intestines, SP spleen, UB urinary bladder.

Full size image

A single case of a Frem2-KO mouse surviving into adulthood

Over years of breeding, only one female Frem2-KO mouse was identified and survived to adulthood during this study. Although no apparent health conditions were identified and the animal produced two litters, the female was euthanized at 51 weeks of age due to ulcerative dermatitis, a condition common in older mice. Interestingly, the mouse displayed physical phenotypes, such as unilateral cryptophthalmos (Fig. 6A). The left eye of the mouse was closed, with no visible eyelid crease. The contralateral eye, however, appeared normal. A post-mortem incision through the skin of the closed eyelid revealed a significantly smaller eyeball, which was sent for histological analysis along with the contralateral eye. The histological analysis revealed a normally developed right eye and a malformed left eye (Fig. 6B). The left eye was dramatically smaller, and the retina appeared to have folds, a feature often observed in microphthalmia (Fig. 6B’).

Fig. 6
figure 6

Characterization of the Frem2-KO adult mouse. A Images show a Frem2-/- adult mouse with a normal right eye and a left eye affected by cryptophthalmos. B Hematoxylin and eosin-stained histological section through the middle of the normal right eye and the defective left eye. The defective left eye reveals the lens with reduced thickness and retinal dysplasia. B’ Higher magnification image of the defective left eye showing microphthalmia. C Images of normal front paws, affected right (D) and left (E) hind paws by syndactyly. F, G Gross anatomy and histology images of the adult Frem2-/- mouse show one left kidney (K), ovaries (OV), uterine horns (UH), uterine epithelium (UE), and urinary bladder (UB) present. H Histological section of the normal left kidney and adrenal gland. I–J’ High-magnification images of the adrenal gland and different regions of the kidney showing tubules, collecting ducts, and nephrons. Scale bars: B 2 mm, B’ 250 um, G 4 mm, H 3 mm, I–J‘ 500 um.

Full size image

Consistent with our observations in newborn Frem2-KO mice, the front paws of the adult mouse appeared normal (Fig. 6C), while both hind paws were affected by syndactyly (Fig. 6D, E). The hind paws were also slightly curled, similar to the paws observed in newborns. We next examined the renal system of this adult mouse and found that it had one functional kidney. There was no gross evidence of renal tissue on the right side (Fig. 6F). The left kidney was fully attached to the ureter along with the rest of the urinary system. Hematoxylin and eosin-stained histological sections of the left kidney confirmed normal morphology of the adrenal gland and kidney (Fig. 6H). Higher magnification images of the histological sections reveal normal adrenal gland and kidney morphology, with preserved tubules and nephrons (Fig. 6I, J’). Histological evaluation also revealed the absence of the right ureter (Fig. 6G). Although missing a kidney, the reproductive functions were not affected in the Frem2-KO female, allowing her to breed and produce litters. The renal pelvis was also dilated, according to gross examination. As a result of the functional kidney, the urinary bladder was full (Fig. 6F, G). Although this female survived with one functional kidney, the Frem2-KO adult mouse displayed other prominent Fraser syndrome phenotypes, such as cryptophthalmos and syndactyly (Fig. 6A, C and D).

Discussion

In this study, we developed a constitutive knockout (KO) Frem2 mouse model to evaluate the associated phenotypes. Our findings reveal that the absence of the FREM2 protein results in significant Fraser syndrome-like phenotypes, including cryptophthalmos, syndactyly, and blood-filled blisters, observable as early as during the embryonic stage. Frem2-KO pups displayed distinct phenotypes that were evident within the litter, even before genotyping was performed. Although Frem2-KO mice can survive until birth, they die shortly thereafter, likely due to bilateral renal agenesis. To date, only one Frem2-KO mouse with unilateral renal agenesis from our animal colony has survived into adulthood. These results highlight the critical role of FREM2 in the development of the epidermis, eyes, skeletal structure, and kidneys.

The extracellular matrix (ECM) is well-known for its role in providing structural support. However, it also plays a crucial role in maintaining tissue integrity by regulating cell proliferation, differentiation, and survival. FRAS1, FREM1, and FREM2 serve as important components of the ECM protein complex and are structurally similar proteins. Previous studies show that absence of any one of these proteins causes severe phenotypes, suggesting that they are unable to substitute for each other in any major capacity. As previously reported in Frem2 mutant mice, not only FREM2 but FRAS1 and QBRICK/FREM1 were depleted from the basement membrane zone, suggesting that loss of one complex member leads to depletion of the others, indicating interdependence6.

The ECM includes basement membranes that form sheets underlying epithelial and endothelial cell layers. FREM2 is localized within the epithelial basement membrane. Previous studies using immunogold labeling have demonstrated the clustered localization of FREM2 within the sublamina densa of embryonic skin, highlighting its importance in tissue integrity3. Mouse models with a deficiency of ECM proteins, such as FREM2, reveal how disruptions in ECM components can lead to blistering phenotypes, highlighting the intricate relationship between genetic mutations and the assembly of essential basement membranes13. Although we do not expect a complete loss of Frem2 transcript and were unable to confirm a complete loss of protein due to the lack of reliable commercially available anti-FREM2 antibodies, the resulting consistent and severe phenotype observed in Frem2-KO mice suggests that this allele is likely to represent a functional null. We acknowledge this limitation to the study and note that custom antibody generation may be necessary for definitive protein-level confirmation, as done in prior studies6,14. Please refer to Materials and Methods for a list of antibodies used in this study.

Our observations support previously reported evidence from other studies that use Frem2 mouse models, showing that mutations in Frem2 can cause blood-filled blisters and hemorrhages on the skin. For instance, in a study involving a Frem2 mouse model of the myUcl strain, homozygote mutant mice exhibited epithelial blebbing beginning at E11.58. By E14, these blebs had progressed into hemorrhagic lesions. Another study focused on embryonic development suggests that the onset of angiogenesis, the process of new blood vessel formation involving the growth and differentiation of endothelial cells, may explain the timing of these hemorrhages12. Inadequate adhesion of endothelial cells to adjacent structures, particularly in cells deficient in FREM2 protein, could explain deficits in angiogenesis12. Although FREM2 is not directly localized within vascular structures and is instead located in the membrane lining of the blood vessels, it appears to play a critical role in stabilizing them during development12. The absence of FREM2 in surrounding tissues likely contributes to the hemorrhages observed on the skin, aligning with the parallel occurrence of syndactyly and blood-filled blisters observed on the paws and eyes of Frem2-KO mice.

The skeletal malformations, syndactyly, and dorsal flexure observed in Frem2-KO mice are unlikely to result from primary defects in bone development, as Frem2 has not been directly implicated in osteogenesis. This further supports the conclusion that FREM2 is essential not only for vascular stability but also for tissue remodeling and skeletal development during gestation. Supporting this interpretation, studies involving mouse models lacking ubiquitous basement membrane proteins, such as Nidogen 1 and Nidogen 2, have shown that disruption of the ectodermal basement membrane alone is sufficient to cause limb malformations15. This reinforces the idea that basement membrane integrity is essential for normal limb development. However, the potential relationship between these anomalies remains speculative and warrants further investigation through targeted histological and morphological analyses of skin, bone, or cartilage to elucidate the underlying developmental mechanisms.

Patients with Fraser syndrome who exhibit cryptophthalmos are born with skin covering their eyes or with fused eyelids. Instead, in Frem2-deficient mice, this phenotype more often presents as the absence rather than the fusion of eyelids. We observed that Frem2-KO fetuses developed blood-filled blisters or hemorrhages that covered their eyes, ultimately resulting in the absence of one or both eyelids at birth. A study on Frem2 mice carrying the myF11 mutation, which is predicted to result in a truncated protein, indicated that embryonic hemorrhages might lead to localized tissue necrosis, which could explain the absence of eyelids12. Given the proposed role of FREM2 in vascular stability, it is likely that the loss of FREM2 during key morphogenic events, like eyelid development, could lead to such hemorrhages. The subsequent blood-filled blisters and hemorrhages likely inhibit normal skin development, causing the observed phenotypes.

Moreover, a study utilizing a compound heterozygous mutation derived from a Fraser syndrome patient to generate mice that mimic the human cryptophthalmos phenotype found significant abnormalities in eyelid development during the critical stages at E13-1411. Specifically, the lower eyelid fold was poorly defined in Frem2 mutant fetuses, in contrast to wild-type mice, where the grooves of the ectoderm that eventually form the upper and lower eyelids are visible. These mutants exhibited dysplasia and microphthalmia, with reductions in the eye’s axial length and lens size. Eyes affected by microphthalmia typically have a thicker cornea, an absent or severely underdeveloped lens, and a retina prone to folding or filling the vitreous body. These features likely exert pressure on the lens, potentially exacerbated by the presence of blisters or hemorrhages16. In our Frem2-KO model, we observed microphthalmia where one eye was smaller than the other, as well as retinal folding. In Fraser syndrome patients, these ocular abnormalities often lead to impaired vision. This shows that FREM2 is directly involved in eyelid development and indirectly in the development of the lens, retina, and other parts of the eye. While the scRNAseq data suggests that FREM2 is also expressed at early stages of human retinal development17, in the current study we could not detect early changes in the developing retina.

FREM2 protein was reported to be expressed in the epithelia in the renal cortex, or the outer layer, of mouse kidneys18. FREM2 expression begins in the ureteric epithelia of the metanephros at E11.5, with peak expression at the tips of the ureteric buds8. In adult mice, FREM2 is strongly expressed in the collecting ducts, proximal convoluted tubules, and arterioles within the kidneys8. This expression pattern, along with the development of renal cysts in Frem2 mutant animals, suggests that FREM2 is essential for maintaining renal integrity18. The absence of kidneys in Frem2-KO mice that fail to survive postnatally provides additional evidence of FREM2’s critical involvement in renal development. Interestingly, over multiple years of breeding, the only Frem2-KO mouse that survived into adulthood had one functional kidney, suggesting that perhaps while FREM2 is critical for renal development, it is not absolutely indispensable. This partial necessity may be due to protein–protein interactions within the ECM protein complex, where other proteins such as FREM1 or FRAS1 may, at least partially, compensate for the loss of FREM2 in the formation of the basement membrane. Although Fraser syndrome is associated with a risk of miscarriage, bilateral renal agenesis is a significant phenotype contributing to this, while patients with unilateral renal agenesis can survive with a single kidney 1. Since kidneys contribute to the amniotic fluid in humans, by the second trimester of pregnancy, low amniotic fluid levels serve as an important indicator of renal agenesis in the fetus19. Insufficient amniotic fluid, also known as oligohydramnios, in addition to the absence of kidneys, poses a significant risk for fetal death. In contrast, the impact of bilateral renal agenesis is markedly different in mice, resulting in death within 2 days after birth20. While the mouse phenotype we observed in our Frem2-KO mice is clearly more severe when compared to other previously reported Frem2 mouse lines, given the uncertain rate of miscarriages related to FREM2 dysfunction, it is difficult to directly compare the severity of the phenotype we observed to that of patients carrying pathogenic FREM2 variants.

In summary, to address the gap in Fraser syndrome research, we have developed and characterized a constitutive Frem2-KO mouse model that lacks the FREM2 protein. While previous studies have utilized various Frem2 mouse models, some of which have retained limited FREM2 function, here we report a mouse model predicted to lack FREM2 entirely. Our findings reveal that Frem2-KO pups die shortly after birth, limiting the opportunity for extensive study. Nevertheless, we have identified the most prominent phenotypes and the primary cause of neonatal mortality. We acknowledge the possibility of non-morphological phenotypes, particularly in the respiratory or cardiovascular systems, that have yet to be uncovered. Our study provides a unique, though severe, model of Fraser syndrome that can be utilized in future embryonic studies to advance the understanding of Fraser syndrome’s pathophysiology. Critically, our work can help tease apart the crucial role that FREM2 has in the development of important organs and also in uncovering the underlying mechanisms that drive Fraser syndrome.

Materials and methods

Animals

Mouse ES cell clones (Frem2tm1a(EUCOMM)Hmgu; clones HEPD0988-1-E06 and HEPD0988-1-H08) were purchased from the European Conditional Mouse Mutagenesis Program (https://www.mousephenotype.org/about-impc/about-ikmc/eucomm/, Clone# HEPD0988_1_E06, HEPD0988_1_H08). The clones were then used to generate the Frem2tm1a(EUCOMM)Hmgu mouse line carrying the knockout-first allele with the conditional potential (F2KCR allele) (Fig. 1A) at the Harvard Transgenic Animal Core. For clone HEPD0988-1-E06, 36 blastocysts were injected, 8 pups were born (5 died), and no chimera pups obtained from this injection. For clone HEPD0988-1-H08, 48 blastocysts were injected, 15 pups were born (5 died), and two of the 10 surviving pups were chimeric mice (1 male was 15% and 1 female was 40% chimeric). The two chimera mice were bred to get germline transmission. The founder mice were then imported to the Mass Eye and Ear animal facility to set up an animal colony. To generate the conditional Frem2fl/fl allele, the FRT-flanked Neo cassette was excised by crossing with a FLP deleter strain (The Jackson Laboratory strain #003946).

The knockout-first mouse line (Frem2-KO-first allele with conditional potential, F2KCP) derived from these ES cells contains a cassette in the middle of exon 1 of the wild-type FREM2 gene. Upon further investigation, we determined that the sequence “GTCCCAGGTCCCGAAAACCAAAGAAGAAGAACGCA,” referred to as “exon 2” in the vendor’s documentation, does not align with any region of the published wild-type Frem2 gene (Ensembl ENSMUSG00000037016.12). The F2KCR line was bred with the FLP mouse (JAX strain # 003946) to delete the “FRT-Exon2-LacZ-LoxP-NeoR” part and obtain a floxed FREM2 line containing “Exon1-FRT-LoxP-Exon3 (or Exon1b)-LoxP” structure.

All procedures and protocols were approved by the Institutional Animal Care and Use Committee of Mass Eye and Ear and were carried out in compliance with all relevant ethical regulations for animal use. All methods are reported in accordance with ARRIVE guidelines (https://arriveguidelines.org). All mice were kept on a 12:12 h light–dark cycle with unlimited access to food and water. The fetuses were collected by timed-breedings or by keeping track of the pregnancy by weighing the dam. For the collection of newborn pups, the breeding cages with pregnant females were closely monitored for litter birth by an infrared webcam setup below the cage with a remote access functionality.

Phenotypic analysis and fixation of fetuses

Mouse fetuses were collected by euthanizing pregnant adult females at the needed gestation stage. Fetuses were collected and fixed in Bouin’s fixative (Electron Microscopy Sciences). E12.6–16.5 were fixed for 4 h. E17.5–18.5 were fixed for 72 h. The fetuses were then rinsed in three changes of 70% alcohol and stored at 70% alcohol until processing.

Collection and fixation of newborn pups

Newborn mouse pups were collected at P0 once the litter was detected using an infrared webcam setup (ELP 1080p USB Camera). Newborn pups were anesthetized on ice, weighed, imaged, and the tip of the tail was collected for genotyping. Pups were then euthanized by decapitation and fixed in Bouin’s solution (Electron Microscopy Sciences, Cat# 15990–01) for 72 h. After fixation, samples were rinsed in three changes of 70% alcohol and stored at 70% alcohol until processing. Pups used for gross anatomy evaluation were euthanized, and a median laparotomy was performed from the pubic crest to the trachea.

Histological analysis

Fetal and newborn pup bodies were trimmed into three sections with cuts at the diaphragm and pelvis to ensure adequate paraffin embedding and tissue processing. Heads were trimmed for coronal embedding to the area of interest. Samples were processed using a Microm STP-120 processor on a routine program for manual paraffin embedding with a Microm EC-350 embedding center. After paraffin embedding, 5 µm sections were cut from the head and body for Hematoxylin and eosin staining. Sections from the bodies were collected distally every ~ 400 µm through the entire cavity. Slides were imaged with the Aperio AT2 automatic slide scanner (Leica Biosystems) equipped with Consolo v102.0.7.5 and Controller v102.0.8.704. The slides were digitized with a 20x/0.75NA Plan Apo objective lens and processed with the Aperio ImageScope software (version 12.4.6), available for download from https://www.leicabiosystems.com/us/digital-pathology/manage/aperio-imagescope/ free of charge. Anti-FREM2 immunolabeling was carried out using a previously reported labeling protocol21. The following antibodies against FREM2 were tested on the skin cryosectioned tissue samples of embryos: mouse monoclonal Cat. #SC-376555 (Santa Cruz Biotechnology); rabbit polyclonal Cat. #5831–1003 (ProScience) and rabbit polyclonal Cat. #5713, One World Lab (discontinued).