Analysis of knockout mice reveals critical female-specific roles for the Hippo pathway component PTPN14.

To understand the role of PTPN14 in development and tissue homeostasis, we generated Ptpn14 conditional knockout (cKO) mice using Cre-LoxP technology. In the design of the cKO allele, we chose to flank exon 3 with LoxP sites, as deletion of exon 3 would shift the remainder of the protein out of frame ( Fig. 1 A). After obtaining founders (see the Materials and Methods), we performed multiple diagnostic PCRs to identify mice carrying the Ptpn14 cKO allele before expanding the colony. First, we confirmed the presence of the 5′ and 3′ LoxP sites on either side of exon 3 by PCR ( Supplemental Fig. S1A ). We then verified integration into the correct location in the genome by designing 5′ and 3′ external PCRs in which one primer was located outside the region corresponding to the donor DNA used during the CRISPR targeting of the locus ( Supplemental Fig. S1A ). In addition to the cKO mice, we also generated constitutive Ptpn14 knockout mice by crossing Ptpn14 fl/fl mice with C57BL/6J Deleter-Cre mice. The Deleter-Cre transgene drives recombination in the germline, resulting in Ptpn14 -null alleles. After breeding out the Cre transgene, we propagated the colony and confirmed genetic recombination at the Ptpn14 locus in both Ptpn14 +/ − and Ptpn14 − / − mice ( Supplemental Fig. S1B ). Ptpn14 deficiency compromises viability. ( A ) Strategy for generating Ptpn14 conditional and complete knockout strains. ( i ) The Ptpn14 locus with a magnified view of exon 3. ( ii ) sgRNAs flanking exon 3 and the donor DNA template containing floxed exon 3 are indicated. ( iii ) Initial targeting results in a Ptpn14 conditional knockout allele are shown. ( iv ) Upon expression of the Cre recombinase, Cre recombines the LoxP sites surrounding exon 3, leading to exon 3 deletion from the genome and a complete Ptpn14 knockout allele. Sites for PCR primers (A–H) used to confirm the allele (see the text) are highlighted. ( B ) Expected and actual pup numbers obtained at weaning after intercrossing Ptpn14 +/ − mice. The χ 2 analysis indicates a significant disenrichment of Ptpn14 +/ − and Ptpn14 −/− mice (χ 2 = 9.45, the threshold for P -value < 0.05 is 5.99 for two degrees of freedom). ( C ) Body weight of Ptpn14 +/+ ( n = 14), Ptpn14 +/ − ( n = 23), and Ptpn14 − / − ( n = 16) mice at weaning (days 20–25). P -value = 0.0025. ( D ) Kaplan–Meier curve showing the survival of Ptpn14 +/+ ( n = 15) and Ptpn14 − / − ( n = 17) mice. Mice still living at the conclusion of the study are marked with black boxes ( P -value = 0.0001 by Gehan–Breslow–Wilcoxon test). ( E ) Phenotypes observed in the cohorts used in the Kaplan–Meier aging study in D . One wild-type mouse died from a seizure, and one developed a skin infection that necessitated sacrifice. To verify that Cre-mediated recombination of exon 3 resulted in loss of the PTPN14 protein, we intercrossed Ptpn14 +/fl mice and made mouse embryonic fibroblasts (MEFs) at E13.5. After delivering adenovirus-Cre (Ad-Cre) or control empty adenovirus (Ad-Empty) to Ptpn14 +/+ , Ptpn14 +/fl , and Ptpn14 fl/fl MEFs, we found that only Ad-Cre transduction into the Ptpn14 fl/fl MEFs caused loss of the PTPN14 protein ( Supplemental Fig. S1C ). These findings confirmed that recombination of the Ptpn14 cKO allele generates a Ptpn14 -null allele after Cre introduction. To evaluate whether PTPN14 deficiency affects organismal viability, we intercrossed Ptpn14 +/ − mice and quantified the genotypes of their progeny at weaning ( Fig. 1 B). Although a number of Ptpn14 − / − mice survived to adulthood, χ 2 analysis indicated there were significantly fewer Ptpn14 − / − and Ptpn14 +/ − mice than expected by Mendelian ratios ( Fig. 1 B). Ptpn14 − / − mice also displayed a lower body weight at weaning, a sign that Ptpn14 deficiency may compromise developmental processes and/or homeostasis ( Fig. 1 C). Upon tracking the success of the breeding pairs in our colony, we found that out of 16 Ptpn14 +/ − female breeders, two females reached morbidity due to dystocia (an inability to push the pups out of the birthing canal). We also bred two Ptpn14 − / − female mice, which did not successfully reproduce, with one carrying a single litter before succumbing to dystocia and the other never becoming pregnant at all. These observations of Ptpn14 +/ − and Ptpn14 − / − mice suggest a negative effect on reproduction with PTPN14 loss. Because many Ptpn14 − / − mice reached adulthood, we aged these mice and assessed whether Ptpn14 loss altered overall survival or caused any deleterious phenotypes. We generated cohorts of 15 wild-type mice (eight males and seven females) and 17 Ptpn14 − / − mice (seven males and 10 females) and measured their life span and any associated phenotypes upon morbidity. The Ptpn14 − / − mice displayed reduced survival relative to their wild-type counterparts ( Fig. 1 D). Of the 17 Ptpn14 − / − mice, three died prematurely and nine had severe eye defects necessitating sacrifice ( Fig. 1 E). These defects manifested as large white protrusions emerging from the center of the eye. Although no other external signs of poor health appeared in these knockout mice at morbidity, we observed other deleterious phenotypes, such as hydrometra and kidney cysts, after sacrifice (see below). These findings underscore the role of PTPN14 in normal tissue function. The dystocia observed in the Ptpn14 − / − female mice suggested a potential defect in the female reproductive system. Indeed, at morbidity, all Ptpn14 − / − female mice from the aging cohort and four mice sacrificed from the colony, all of which were virgins, exhibited distended uterine horns filled with fluid upon dissection, a condition known as hydrometra ( n = 12) ( Fig. 2 A). The material filling the uterine horns was usually red, though in some mice it was white and cloudy, and upon histological evaluation, the liquid appeared to be a mixture of proteinaceous material and blood ( Fig. 2 B,C). Although the wall of the uterine horns was severely distended to accommodate the liquid accumulating there, no other abnormalities were detected in the myometrium or the luminal epithelium ( Fig. 2 D), suggesting that the liquid might be trapped there by an obstruction further down in the uterus or the cervix. We also noted severely dilated blood vessels located within the uterine wall that could have resulted from blocked blood flow, further suggesting an obstruction occurring somewhere in the organ ( Fig. 2 E). Although hydrometra is typically caused by a blockage of the female reproductive tract, histological evaluation failed to detect the presence of an obstruction in the uterus, cervix, or vagina. However, these membranes can rupture immediately upon dissection, and therefore the cause of hydrometra can often not be definitively determined. To further investigate this phenotype, we conducted RNAscope in the uteri of wild-type female mice to assess which uterine cell types express Ptpn14 mRNA ( n = 3) ( Fig. 2 F). Our results detected robust expression of Ptpn14 in uterine glandular and luminal epithelia, as well as in the myometrium, but showed little to no expression in the uterine stroma ( Fig. 2 F). In support of these results, high expression of Ptpn14 mRNA has been detected in bulk tissue analyses of mouse uterine cells ( Supplemental Fig. S2 ). Characterization of hydrometra in Ptpn14 − / − mouse uteri. ( A ) Whole-mount images of representative Ptpn14 +/+ ( left ) and Ptpn14 − / − ( right ) mouse uteri. A total of 12 female mice was dissected and evaluated for hydrometra. These mice comprised those in the aging cohort ( n = 8, excluding the mice that were found dead) and other mice in the colony ( n = 4). These ranged in age from 171 to 499 days. ( B ) Diagram of the female mouse reproductive system. The black box indicates an example of the regions shown in C – E . ( C ) H&E image of the uterine wall and the fluid filling the uterus of a Ptpn14 − / − mouse with hydrometra. The uterine wall is denoted by an asterisk, and the fluid is indicated with arrows. Scale bar, 100 μm. ( D ) H&E images of longitudinal sections of the uterine walls of Ptpn14 +/+ ( left ) and Ptpn14 − / − ( right ) mice. Brackets indicate the bounds of the uterine wall region, with Ptpn14 − / − mice exhibiting a thinner wall due to the dramatic distension of the uterus. Scale bar, 100 μm. ( E ) H&E images of normal blood vessels in wild-type mice ( left ) and the distended blood vessels found in Ptpn14 − / − mice ( right ). Arrows indicate blood vessels. Scale bar, 100 μm. ( F ) RNAscope images of Ptpn14 expression in glandular epithelial cells ( top left ), luminal epithelial cells ( top right ), and myometrial cells ( bottom left ) in the wild-type mouse uterus. ( Bottom right ) Lower-magnification cross-section of the mouse uterus indicating the luminal epithelia (red-outlined region), glandular epithelia (yellow-outlined region), and myometrium (black-outlined regions) in the RNAscope staining. Black arrows indicate examples of positive staining. Scale bar, 50 μm. n = 3. ( G ) Representative YAP immunofluorescence staining in wild-type ( n = 2) and Ptpn14 −/− ( n = 3) mouse uteri. The white dotted lines indicate the epithelial regions counted in H for the quantification of the YAP staining. Scale bar, 50 μm. White arrows indicate examples of nuclear YAP staining. ( H ) Quantification of the ratio of YAP nuclear to cytoplasmic localization in mouse uteri ( P -value = 0.0222). Because PTPN14 is known to regulate YAP/TAZ nuclear localization and we detected high expression of Ptpn14 in the uterine epithelia, we investigated YAP/TAZ localization in the epithelium of aged female mouse uteri ( Wilson et al. 2014 ). PTPN14 deficiency led to an increase in the nuclear:cytoplasmic ratio of YAP in the epithelium of Ptpn14 −/− mouse uteri compared with wild-type mice ( Fig. 2 G,H), indicative of enhanced YAP/TAZ signaling. During the estrous cycle, YAP/TAZ oscillate between nuclear and cytoplasmic localization ( Moon et al. 2022 ). Although these mice were aged ≥1 year and should be in estropause, the increased YAP signaling observed in the knockout mice resembles that of younger mice in their estrus phase ( Moon et al. 2022 ). To further understand the role of PTPN14 in vivo, we next analyzed Ptpn14 − / − mice for abnormalities in vital organs. We thoroughly surveyed the effects of Ptpn14 loss on these organs by assessing H&E staining of tissues from wild-type and Ptpn14 − / − mice in our colony ( n ≥ 3 per genotype for each organ), combined with observations from our aging cohort. Further investigation of our aging mouse cohort revealed that one 171 day old female Ptpn14 − / − mouse developed an abnormal-looking kidney with dramatic glomerular cysts, suggesting that loss of PTPN14 may disrupt kidney function over time ( Fig. 3 A,B). However, we observed no further kidney abnormalities by gross inspection or histological analysis in the remaining mice in the aging cohort or in the other Ptpn14 − / − mice in the colony ( Fig. 3 C). This led us to hypothesize that the knockout mouse with kidney cysts was exposed to a unique stress or injury that revealed the lack of Ptpn14 expression in the kidney. Because Hippo pathway inactivation can cause increases in organ size, we also assessed the weight of the kidneys of these mice relative to their total body weight ( Fig. 3 D). These data indicate that Ptpn14 deficiency does not affect kidney size, which is consistent with the relatively low expression of Ptpn14 in mouse kidneys at homeostasis ( Fig. 7 A, below; Supplemental Fig. S2 ). Ptpn14 deficiency in organs typically affected in Hippo pathway knockout mice. ( A ) Whole-mount images of kidneys from a wild-type female mouse from the colony ( top ) and a Ptpn14 − / − female mouse from the aging study ( bottom ). ( B ) H&E images of the kidneys from a Ptpn14 +/+ mouse ( top ) and the Ptpn14 − / − mouse ( bottom ) pictured in A , showing dramatic glomerular cysts. Scale bar, 250 μm. ( C ) Representative H&E images of the kidneys from age-matched Ptpn14 +/+ ( top ) and Ptpn14 − / − ( bottom ) mice (both male aged 100 and 98 days, respectively). Scale bar, 100 μm. ( D ) The ratio of the right and left kidney weight to the body weight of Ptpn14 +/+ ( n = 7 [five males and two females]) and Ptpn14 − / − ( n = 9 [six males and three females]) mice. ( E ) Representative H&E images of the hearts from Ptpn14 +/+ ( top ) and Ptpn14 − / − ( bottom ) mice. Scale bar, 50 μm. ( F ) The ratio of the heart weight to the body weight of Ptpn14 +/+ ( n = 5 [five males and two females]) and Ptpn14 − / − ( n = 10 [five males and five females]) mice ( P -value = 0.0099). ( G ) Representative H&E images of livers from Ptpn14 +/+ ( top ) and Ptpn14 − / − ( bottom ) mice. Scale bar, 100 μm. ( H ) The ratio of the liver weight to the body weight of Ptpn14 +/+ ( n = 7 [five males and two females]) and Ptpn14 − / − ( n = 10 [five males and five females]) mice. For organ weight figures, all organs measured were from mice aged >220 days. As previous studies reported abnormal heart growth and cardiac hyperplasia with either Hippo pathway knockout or YAP/TAZ overexpression, we next analyzed the hearts in Ptpn14 − / − and wild-type control cohorts ( Johnson et al. 1990 ; Monroe et al. 2019 ). We measured heart weight relative to the body weight of each mouse and generated H&E sections of the hearts for histological evaluation. Although histological analysis failed to show a difference in heart tissue morphology ( Fig. 3 E), these analyses revealed that Ptpn14 − / − mice had significantly larger hearts than the wild-type controls ( Fig. 3 F). Ptpn14 is highly expressed in multiple cell types of the mouse heart ( Supplemental Fig. S2 ; The Tabula Muris Consortium 2018 ). As hepatomegaly and hepatocellular carcinoma manifest in many Hippo pathway knockout mice, we next investigated the effects of Ptpn14 loss in the liver ( Zhou et al. 2009 ; Lee et al. 2010 ; Lu et al. 2010 ; Song et al. 2010 ; Zhang et al. 2010 ; Driskill and Pan 2021 ). Upon dissection, we observed no obvious difference between Ptpn14 − / − livers and their wild-type counterparts by visual inspection and no evidence of hepatocellular carcinoma or hyperplasia in the Ptpn14 − / − livers by histopathological evaluation ( Fig. 3 G). There was also minimal difference in the liver to body weight ratios of these mice ( Fig. 3 H). Consistent with these observations, liver cells show little to no expression of Ptpn14 mRNA ( Supplemental Fig. S2 ). Together, these findings suggest that PTPN14 does not significantly influence liver health under basal conditions. Similarly, we detected no clear differences between Ptpn14 − / − and Ptpn14 +/+ mice in other major organs, including the lungs, pancreata, and skin, by histological analysis ( Supplemental Fig. S3 ) despite being highly expressed in some of these tissues, such as the skin and lungs. This high expression level suggests that PTPN14 only plays a role in these tissues under specific conditions, which may explain the lack of a clear knockout phenotype in these tissues at homeostasis ( Supplemental Fig. S2 ). One of the most prominent phenotypes observed in the Ptpn14 − / − mice was the development of white eye lesions upon aging. The eyes also appeared more sunken into the skull, with swelling and hair loss in the surrounding eyelid region, suggesting irritation of the orbit. Of the 17 Ptpn14 − / − mice in our aging cohort, nine (53%) of them presented with severe eye lesions ( Fig. 4 A). Moreover, eight out of the nine mice with eye lesions were female, suggesting that sex-specific differences may influence this phenotype. The single male that we identified had a much less severe phenotype, with no protrusion from the eye and only a subtle clouding of the cornea with cells in the epithelium that appeared slightly more proliferative. To fully characterize these eye lesions, we used knockout mice from our aging study and the rest of the colony that developed the phenotype as well as wild-type control mice from the colony. Fluorescein staining of these eyes showed extensive damage to the epithelial layer of the cornea ( Fig. 4 B). Analysis of H&E-stained eyes revealed an irregular expansion of the basal epithelial cells in the cornea ( Fig. 4 C,D). We confirmed this observation by costaining for keratin 12 (K12), a marker of corneal epithelial cells, and Ki67, a proliferation marker, which indicated that the basal epithelial cells of the cornea were indeed more proliferative in Ptpn14 − / − mice than in wild-type mice ( Fig. 4 E,F). Ptpn14 mRNA is detected in the cornea and other cell types of the eye, supporting its role as a regulator of this tissue ( Supplemental Fig. S2 ). Ptpn14 −/− mice present with overproliferative corneal cells. ( A ) Whole-mount images of eyes from Ptpn14 +/+ ( top ) and Ptpn14 −/− ( bottom ) mice. Ptpn14 −/− mice presented with large white protrusions that extended from the center of the eye. These particular mice were not part of the aging study, though those in the aging study presented with similar lesions. ( B ) Fluorescein staining of eyes pictured in A . Dramatically increased staining in Ptpn14 −/− mice indicates severe damage to the corneal epithelial layer of these eyes. ( C ) Diagram of the various tissue layers of the mouse cornea. ( D ) H&E images of Ptpn14 +/+ ( top ) and Ptpn14 −/− ( bottom ) mouse corneas. Black brackets indicate the stromal layer, and red brackets denote the epithelial layer. Arrows indicate regions of overproliferation in the Ptpn14 −/− mouse corneas. Mice presenting with these lesions were aged between 171 and 499 days, with a median of 219 days when the lesion formed. Scale bar, 100 μm. ( E ) Quantification of Ki67-positive cells in the basal epithelial layer of the wild-type ( n = 3 [from the colony]) and Ptpn14 −/− ( n = 4 [two from the aging study and two from the colony]) mouse corneas ( P -value = 0.0428). ( F ) Representative K12, Ki67, and PCAD immunofluorescence staining in wild-type ( n = 3) and Ptpn14 −/− ( n = 4) mouse corneas quantified in E . The white dotted lines indicate the basal epithelial regions counted in E for the quantification of the Ki67 staining. Scale bar, 100 μm. Because of the stochastic nature of these eye lesions, we hypothesized that a random stress, such as a scratch, triggered hyperproliferation and an aberrant injury response that resulted in the observed lesions. Indeed, in both Ptpn14 +/+ and Ptpn14 − / − male and female neonatal mice, the cornea developed normally, suggesting that the eye phenotypes that we observed in adults were not due to developmental defects but rather an uncontrolled injury response ( Supplemental Fig. S4A ). In the adult mouse, the cornea is a highly regenerative organ, and its cells are in a state of constant renewal ( Kao 2020 ). At homeostasis, the limbal epithelium, a stem cell compartment of the cornea, releases cells that slowly differentiate into mature corneal epithelial cells, which migrate toward the outer edge of the corneal epithelium and are eventually exfoliated off of the ocular surface at the end of their life ( Supplemental Fig. S4B ; Kao 2020 ). Throughout the differentiation process, these cells lose expression of keratin 19 (K19) and P-cadherin (PCAD), severely reduce their expression of keratin 14 (K14), and start to express K12 when they complete the transition to mature epithelial cells. This process quickens during injury, when the corneal epithelial stem cells increase their proliferation 40-fold and increase their migration to heal the corneal epithelium ( Supplemental Fig. S4B ; Park et al. 2019 ). During this injury response, YAP/TAZ also displays increased nuclear localization to enhance proliferation and migration ( Li et al. 2021 ). As PTPN14 is a negative regulator of YAP/TAZ that also controls progenitor cell responses to injury in Drosophila , we hypothesized that PTPN14 loss and increased YAP/TAZ signaling may have disrupted the injury response in the corneal epithelium. To assess the injury response in the Ptpn14 − / − mouse corneas, we stained for markers of several known cell types in the cornea. We found that the corneal epithelial cells in the Ptpn14 − / − mice expressed PCAD ( Figs. 4 F, 5 A) and K19, markers of corneal progenitor cells normally lost in mature corneal epithelial cells ( Fig. 5 A). Notably, the cells retained high-level expression of K14, a protein that is highly expressed in corneal progenitor cell but decreases its expression in mature corneal epithelial cells ( Fig. 5 A). We also noted that the intensity of the PCAD and K19 staining correlated positively with the severity of the eye lesions, with lower expression in the apparently unaffected Ptpn14 − / − eyes ( Fig. 5 A). Interestingly, when we conducted RNAscope in wild-type mouse eyes to determine which cell types would be most affected by Ptpn14 deletion, we found robust expression of Ptpn14 mRNA in the limbal stem cells of the cornea ( Fig. 5 B). This expression was sustained moving out of the limbal compartment and into the mature corneal epithelial cells, suggesting a role for PTPN14 in controlling the proliferation of both the progenitor and mature cell populations ( Fig. 5 B). Finally, the corneal epithelium in Ptpn14 − / − mice displayed increased YAP protein levels in both the cytoplasm and the nucleus compared with the corneal epithelium in wild-type mice ( Fig. 5 C). Taken together, these data indicate a disruption in the normal differentiation process of the corneal epithelial progenitor cells and implicate YAP signaling in their increased proliferative capacity. Ptpn14 loss alters corneal epithelial cell maturation and inhibits lacrimal gland development ( A ) PCAD, K19, and K14 immunofluorescence staining in wild-type ( n = 3; top row) and Ptpn14 −/− ( n = 3; middle and bottom rows) mouse corneas. Representative images of Ptpn14 −/− mouse corneas with no obvious eye lesion ( middle row) and representative images of a severe lesion ( bottom ) are shown. Black arrows indicate examples of regions with specific K19 staining. Scale bar, 40 μm. ( B ) RNAscope images of Ptpn14 mRNA expression (red) in the limbal stem cells ( top panel) and mature corneal epithelial cells ( bottom panel) of wild-type mice. DAPI staining is shown in blue. White arrows indicate examples of positive staining for Ptpn14 mRNA. Scale bar, 50 μm. ( C ) Representative YAP immunofluorescence staining in wild-type ( n = 3; top ) and Ptpn14 −/− ( n = 3; bottom ) mouse corneas. The dotted line demarcates the barrier between the epithelial and stromal layers of the cornea. Scale bar, 40 μm. ( D ) H&E images of the lacrimal glands in neonate Ptpn14 +/+ ( top ) and Ptpn14 −/− ( bottom ) mice ( n = 3 for both genotypes). Scale bar, 200 μm. Interestingly, we also noted that the lacrimal gland, which is responsible for producing tears, was markedly smaller or absent entirely in neonate Ptpn14 − / − mice compared with wild-type controls, indicating that PTPN14 is essential for normal lacrimal gland development ( Fig. 5 D). This is consistent with the pattern of Ptpn14 mRNA expression, which is detected in the lacrimal glands of mice ( Supplemental Fig. S2 ). Although this defect likely leads to severe dry eye in the Ptpn14 − / − mice, it is unlikely to be the primary stress inducing eye lesions, as the Ptpn14 − / − female mice typically developed the eye lesions in only one eye, but both lacrimal glands developed aberrantly. Moreover, the severe cornea phenotype is exclusively seen in females, but analysis of our neonate cohort indicated that both male and female lacrimal glands developed improperly. Instead, we hypothesize that PTPN14 loss reprograms the corneal epithelial cells to a progenitor-like state and that injury to the corneal epithelium triggers a proliferation of these primed progenitor cells that is unable to halt at the conclusion of corneal wound healing ( Supplemental Fig. S4B ). To investigate whether, in some organs, PTPN14 may be dispensable for development but play an integral role in the injury response, we used a liver injury model using 3,5-diethoxycarbonyl-l,4-dihydrocollidine (DDC), a hepatotoxin that causes widespread hepatic injury and induces proliferation of biliary epithelial cells (BECs) near portal vein tracts to regenerate the liver ( Preisegger et al. 1999 ; Yi et al. 2013 ). Under hepatotoxic conditions, these BECs assume a rounded morphology, which led the BEC injury response to be termed “the oval cell response” ( Shinozuka et al. 1978 ). We chose to conduct the injury model in the liver because the liver appears to develop normally without Ptpn14 under basal conditions and because a similar study showed that the Hippo component AMOT plays a critical role in liver injury repair but not during homeostasis ( Fig. 3 G,H; Yi et al. 2013 ). Both male and female age-matched mice were fed either their normal diet or a 0.2% DDC-supplemented diet for 3 weeks and then sacrificed for liver analysis ( Fig. 6 A). We stained liver sections for CD31 to mark the portal veins and quantified the K19-positive area, which marked the activated BECs in the oval cell response surrounding these portal veins ( Fig. 6 B). Although both Ptpn14 +/+ and Ptpn14 − / − mice initiated an oval cell response, Ptpn14 − / − mice had a small but significant increase in K19 positivity, indicative of an enhanced oval cell injury response relative to wild-type mice ( Fig. 6 C). We also quantified the number of YAP-positive nuclei relative to the area of each oval cell response and found that this ratio was significantly higher in Ptpn14 −/− mice, indicating an increase in YAP signaling in these mice following injury ( Fig. 6 D,E). These data support a role for PTPN14 in the injury response of the liver and suggest that other organs that appear unaffected by Ptpn14 inactivation during homeostasis may reveal knockout phenotypes under perturbation. PTPN14 regulates the injury response of biliary epithelial cells. ( A ) Schematic of the liver injury study and cohort (control diet: n = 6 Ptpn14 +/+ mice [three males and three females]; DDC diet: n = 8 Ptpn14 +/+ mice [five males and three females] and n = 6 Ptpn14 −/− mice [three males and three females]). ( B ) Representative immunofluorescence staining for CD31 and K19 in livers from the liver injury mouse cohort outlined in A . Scale bar, 50 μm. ( C ) Quantification of the total K19 + area surrounding the portal veins of mice in B . At least three portal veins were quantified per mouse ( P -value = 0.0043 using a t -test). ( D ) Representative images of YAP immunohistochemistry staining in oval cell response regions in DDC-fed mice in the liver injury study. Scale bar, 50 μm. ( E ) Quantification of the number of YAP-positive nuclei per square millimeter in the oval cell response regions of the mice fed DDC in the liver injury study. Three regions were quantified for each mouse in the DDC group of the study. One Ptpn14 knockout mouse was excluded due to poor YAP staining quality ( P -value = 0.0218). We next sought to understand whether the many roles that PTPN14 plays in maintaining the health of various mouse tissues are conserved in humans. Analysis of PTPN14 expression in the GTEX gene expression data set across human tissues confirmed robust expression of the RNA in the uterus, with weaker but detectable expression in the kidney and heart ( Fig. 7 A), the tissues where we observed phenotypes in the Ptpn14 knockout mice. Although the eye was not included in this data set, we found that PTPN14 mRNA is expressed in the human lacrimal gland and cornea according to the Human Eye Transcriptome Atlas ( Wolf et al. 2022 ). PTPN14 exhibits similar expression and phenotypes in humans and mice. ( A ) Expression levels of human PTPN14 across organs from the GTEX data set. Tissues where Ptpn14 knockout mouse phenotypes were identified are highlighted in light blue. ( B ) Table of genome-wide association studies (GWAS) from the GWAS catalog that identified phenotypes correlated with PTPN14 variation that parallel phenotypes in the Ptpn14 knockout mouse. These PTPN14 GWAS associations were taken from a list of 59 total PTPN14 associations identified across 49 different GWAS. ( C ) The top 15 human phenotypes associated with PTPN14 variation from the Common Metabolic Diseases Knowledge Portal data sets. Also included is the underpowered anecdotal observation that PTPN14 variation correlates with primary angle closure glaucoma, as this may relate to the eye phenotype in the Ptpn14 knockout mouse. Human genetic evidence (HUGE) scores were used to evaluate the amount of genetic evidence for each association, with values ≥30 corresponding to “very strong” evidence, values ≥10 indicating “strong” evidence, values ≥3 corresponding to “moderate” evidence, values ≥1 indicating “anecdotal” evidence, and values ≤1 representing no evidence for the given association ( Dornbos et al. 2022 ). ( D ) The top 10 PheCode traits associated with PTPN14 variation in humans from the ExPheWas data portal. We then used data from the 49 studies in the genome-wide association studies (GWAS) catalog with significant PTPN14 associations to examine potential phenotypes associated with PTPN14 variation in humans. Of the 59 associations with PTPN14 variation identified in these GWAS, nine were kidney-related (either glomerular filtration rate or serum creatine levels), reinforcing our hypothesis that PTPN14 plays an important role in kidney health ( Fig. 7 B). GWAS also associated PTPN14 variation with low birth weight and endometriosis ( Fig. 7 B). Our Ptpn14 knockout mice similarly weighed less at birth than their wild-type counterparts, and the endometriosis observed in humans could relate to the uterine abnormalities that developed in female Ptpn14 − / − mice. Analysis of data from the Common Metabolic Diseases Knowledge Portal reinforced these conclusions. We noted that PTPN14 variation was significantly associated with adult height and weight as well as birth weight, which we observed in the Ptpn14 knockout mice ( Fig. 7 C). The subjects in this data set also had a significant association between PTPN14 variation and several uterus phenotypes, as well as many kidney health biomarkers, paralleling our observations in the knockout mouse uterus and kidney ( Fig. 7 C). Furthermore, PTPN14 variation was anecdotally associated with primary angle closure glaucoma, which, when combined with our data, supports our hypothesis that PTPN14 also plays a role in eye health in humans ( Fig. 7 C). Finally, analysis of the ExPheWas data portal associated PTPN14 variation with traits relating to reproduction, renal health, and eye health, further reinforcing our observations in the knockout mice ( Fig. 7 D; Legault et al. 2022 ). Together, these data from many large human cohorts suggest that PTPN14 plays an evolutionarily conserved role between mice and humans in many tissues and that the phenotypes in the Ptpn14 knockout mice will provide insight into the basis of these human conditions.

Here, we investigated the physiological roles of PTPN14 by generating Ptpn14 knockout mice. Although numerous Ptpn14 − / − mice survive to adulthood, we observed significant selection against Ptpn14 − / − pups prior to weaning, indicating that PTPN14 is essential for organismal viability. In contrast to other Hippo pathway protein knockout mice, which commonly fail to survive postnatally, the survival of Ptpn14 − / − mice provides a unique opportunity to study the Hippo pathway across tissues in adult mice. These analyses revealed that Ptpn14 loss is associated with a host of phenotypes in varied tissues, including the eye, uterus, heart, and kidney. Notably, by studying both male and female mice, we discovered that Ptpn14 deficiency triggered female-specific phenotypes in the eye and uterus. Together, our analyses significantly expand our knowledge of PTPN14 and the role it plays in maintaining organ function at homeostasis. One of the most dramatic phenotypes that we observed in the Ptpn14 knockout mice was the development of large eye lesions caused by an overproliferation of corneal epithelial cells. These lesions typically arose in only one eye, suggesting that a stochastic event, such as a scratch, elicited a subsequent injury response that was not effectively controlled. YAP/TAZ has an established role in the injury response, during which the Hippo pathway becomes inactivated to allow YAP/TAZ to initiate the growth-promoting gene expression necessary for wound healing ( Zanconato et al. 2016 ; Yui et al. 2018 ; Moya and Halder 2019 ; Deng et al. 2022 ). Therefore, it follows that deletion of a known negative regulator of YAP/TAZ, like PTPN14, could cause an aberrant response to injury. Indeed, YAP/TAZ has specifically been studied in corneal wound healing, where it becomes activated to promote proliferation and migration and to disrupt cell–cell adhesion ( Li et al. 2021 ). Therefore, when PTPN14 is lost, the YAP/TAZ overgrowth phenotype initiated during wound healing is sustained after the injury heals, leading to a severe overproliferation of cells in the Ptpn14 knockout mouse eyes and the consequent lesions. Although many female Ptpn14 − / − mice developed eye lesions, 100% of Ptpn14 − / − female mice presented with hydrometra. Although this disease is poorly understood in women, current research implicates hormonal dysregulation or obstruction or inflammation of the cervix/vagina/fallopian tubes as potential causes for hydrometra ( Wu et al. 2023 ). Our findings suggest that inappropriate YAP/TAZ-induced proliferation may contribute to the development of hydrometra in the Ptpn14 −/− mice. Indeed, it is known that during the estrus phase of the mouse estrous cycle in response to hormone signaling, increased nuclear YAP/TAZ is associated with peak luminal epithelial cell proliferation ( Jin 2019 ; Moon et al. 2022 ; Zhang et al. 2023 ) , . YAP/TAZ also regulates injury repair in the uterus following pregnancy ( Zhang et al. 2023 ). As proliferation during both the estrous cycle and uterine injury repair is regulated by YAP/TAZ, the increased YAP/TAZ signaling that we observed upon deletion of Ptpn14 may reflect perturbations in the normal hormonal signaling in the uteri of Ptpn14 knockout mice. Defects in the female reproductive tract have also been reported in knockout mice for other Hippo pathway members. For example, Lats1 knockout mice have incorrectly developed ovaries, and conditional knockout of Lats1 and Lats2 in the mesenchymal cells that form the female reproductive tract results in malformation of the uterus, with thicker and shorter uterine horns combined with abnormal development of the tissue demarcating endometrial and myometrial cells ( St John et al. 1999 ; McPherson et al. 2004 ; Yabuta et al. 2007 ; St-Jean et al. 2019 ). As PTPN14 is only one upstream effector of LATS1/2, it is not surprising that deletion of Ptpn14 would result in a uterine phenotype less dramatic than complete ablation of LATS1/2. Mice with depletion of NF2 signaling, another upstream Hippo pathway effector, also display uterine phenotypes, such as uterine tumors, infertility, and disrupted formation of endometrial glands ( Messerli et al. 2002 ; Lopez et al. 2018 ). Interestingly, PTPN14 is highly mutated in female-specific cancers in humans, including uterine corpus endometrial carcinoma (9.28%), uterine carcinosarcoma (9.09%), and endometrial carcinoma (7.41%) ( Supplemental Fig. S5A,B ), further supporting the role of PTPN14 in female-specific tissues ( Cerami et al. 2012 ; Gao et al. 2013 ; Jones et al. 2014 ; Liu et al. 2018 ; Dou et al. 2020 ; de Bruijn et al. 2023 ). Because obstruction of the lower reproductive track is related to hydrometra development, it is notable that PTPN14 regulates cervical epithelial function. Specifically, PTPN14 is bound and degraded by the E7 protein of high-risk human papilloma viruses that cause cervical cancer ( White et al. 2016 ). These observations suggest that PTPN14 restrains proliferation and serves as a tumor suppressor in the female reproductive system ( White et al. 2016 ). Indeed, HPV E7-mediated PTPN14 degradation prevents proper differentiation of primary human keratinocytes and promotes oncogenic transformation ( Hatterschide et al. 2019 ). Although we failed to identify a blockage histologically, it remains possible that PTPN14 deficiency induced an aberrant proliferation of cervical epithelial cells that restricted the fluid flow out of the uterus, leading to hydrometra. Importantly, hydrometra is a predictor of cervical cancer, which, given the role of PTPN14 as a tumor suppressor in the cervix, suggests that the mice may have developed this cancer if we let them age longer ( Wu et al. 2023 ). Further investigation of the effects of PTPN14 loss during different stages of the mouse estrous cycle and at different reproductive ages will be necessary to fully establish PTPN14 as a regulator of mouse fertility, the estrous cycle, and the aging female reproductive system. An understanding of this aspect of uterine health is critical, as hydrometra is estimated to affect 14.1% of postmenopausal women ( Bar-Hava et al. 1998 ). Other phenotypes that we observed may also relate to defective injury responses, such as the glomerular cysts that developed in the kidney. The stochastic nature of this phenotype suggests that the mice may have experienced stress that revealed a lack of PTPN14 function. Indeed, a large number of GWAS associate human PTPN14 variation with kidney abnormalities and diseases. Moreover, in a study of the mouse kidney, deletion of Ptpn14 reduced inflammation in diabetic nephropathy and allowed podocytes to better heal from injury ( Lin et al. 2021 ). These data support a role for PTPN14 in controlling the injury response and suggest that inactivating PTPN14 may allow for more efficient tissue repair in the kidney and other tissues. While we identified many organs with phenotypes, the majority of the tissues in the Ptpn14 − / − mice appeared to develop and function normally. Although PTPN14 deficiency may not affect the function of these organs, it may also be that PTPN14 signaling is dispensable at homeostasis and yet essential under certain conditions, such as in response to an injury or other stress. The data from our liver injury study support this hypothesis, as livers in Ptpn14 −/− mice develop normally but display an enhanced hepatic biliary epithelial cell injury response when exposed to the hepatotoxin DDC. This observation also reinforces the idea that the corneal ulcers observed in female Ptpn14 − / − mice developed as an aberrant response to injury, whereas in most cases the other eye developed and functioned normally. This phenotypic pattern is not without precedent in Hippo pathway knockout mice, as mice with liver-specific knockout of Amot have no apparent defects in liver function at homeostasis but display reduced biliary epithelial cell proliferation following liver injury ( Yi et al. 2013 ). In another example, Mst1 knockout mice display proper kidney development but, when subjected to ischemia reperfusion injury, exhibit reduced apoptosis in the injured cells ( Li et al. 2019 ). Although these amplified injury responses may more efficiently heal tissues immediately following injury, sustained activation of these pathways, such as we observed with Ptpn14 deletion in the eye, could result in an overgrowth of regenerating cells when the injury response is not halted upon the completion of healing. Our study supports the model that increased YAP signaling from Ptpn14 deficiency results in the deleterious knockout phenotypes that we observed. Although genetic studies using conditional Yap knockout mice or pharmacological studies using TEAD inhibitors would definitively implicate YAP in mediating Ptpn14 knockout mouse phenotypes, many previous studies have already demonstrated that PTPN14 negatively regulates YAP ( Liu et al. 2013 ; Michaloglou et al. 2013 ; Wilson et al. 2014 ). Specifically, in pancreatic cancer cells, PTPN14 overexpression decreased YAP nuclear localization and cellular proliferation, whereas Ptpn14 knockdown in these cells enhanced colony growth, which was attenuated with the YAP inhibitor verteporfin ( Mello et al. 2017 ). Other studies showed that Ptpn14 knockdown increases nuclear YAP, YAP-dependent gene expression, and the ability of cells to seed at low density ( Liu et al. 2013 ; Michaloglou et al. 2013 ; Wilson et al. 2014 ). Conversely, overexpression of Ptpn14 decreased YAP nuclear localization, YAP-dependent gene expression, and cellular proliferation ( Liu et al. 2013 ; Michaloglou et al. 2013 ; Wilson et al. 2014 ; Mello et al. 2017 ). These observations are consistent with the findings in previous studies that PTPN14 binds YAP through its PPxY domains ( Liu et al. 2013 ; Michaloglou et al. 2013 ). Combined with the elevated nuclear YAP in the Ptpn14 knockout mouse tissues, these findings strongly support the idea that PTPN14 negatively regulates YAP signaling and that overactive YAP contributes to the phenotypes that we observed. This does not exclude the possibility that Ptpn14 inactivation could also affect other signaling pathways that contribute to the Ptpn14 − / − phenotypes. Indeed, other studies have mechanistically tied PTPN14 to the P13KA/AKT/mTOR pathway, PDGFRβ signaling, TGFβ signaling, VEGFR signaling, Roquin2 phosphorylation, and many other signaling pathways, all of which have broad biological roles across many organ systems and may be involved in the Ptpn14 knockout phenotypes that we observed ( Wyatt et al. 2007 ; Au et al. 2010 ; Benzinou et al. 2012 ; Belle et al. 2015 ; Choi et al. 2018 ; Bottini et al. 2019 ; Li et al. 2023 ; Ma et al. 2024 ). The sex bias in our phenotypes underscores the importance of conducting experiments in both male and female mice. For many years, female mice have been excluded from studies, and consequently, sexually dimorphic signaling pathways have been largely understudied ( Beery and Zucker 2011 ). Hydrometra was our most penetrant phenotype, affecting every female mouse. Ptpn14 − / − female mice also suffered from dystocia (difficulty giving birth), leading to compromised fertility. These phenotypes may relate to the various deleterious female reproductive phenotypes associated with PTPN14 variation in humans ( Fig. 6 B,C). Another predominant phenotype in female Ptpn14 − / − mice was development of severe hyperproliferative eye lesions that were absent in the male mice. We originally hypothesized that these lesions might occur because PTPN14 loss disrupts lacrimal gland development and causes dry eye, but the male mice also have this defect and do not develop the severe eye lesions that we observed in females. The female tendency toward this phenotype could be explained by the same factors that cause human females to experience greater frequencies of dry eye than human males, including lower androgen signaling and the increased duration of wound healing in female human corneas ( Matossian et al. 2019 ; Kamil and Mohan 2021 ). Because our knockout mice mirror phenotypes observed in women with PTPN14 variation and highlight issues of great relevance to women, such as dry eye and uterine/birthing complications, our model may be useful to investigate how PTPN14 and the Hippo pathway contribute to sexual dimorphism in humans ( Matossian et al. 2019 ). Overall, the phenotypes that we observed suggest that PTPN14 plays an essential but context-specific role in regulating YAP/TAZ signaling, with a particular importance in female biology and the injury response. The many PTPN14 -associated phenotypes detected in humans through GWAS, the Common Metabolic Diseases Knowledge Portal, and ExPheWas indicate that many of the phenotypes in our knockout mice are also observed in humans, and the high evolutionary conservation of PTPN14 suggests that our findings could provide a framework for therapeutic innovations in humans. Specifically, modulation of the Hippo pathway through either PTPN14 inhibition or blockage of YAP/TAZ signaling through TEAD inhibitors could lead to better health outcomes for women in conditions like dry eye, corneal injury, hydrometra, and other uterine complications. Future studies will further elucidate the mechanisms underlying PTPN14 function, ultimately helping to design critical therapies for improving health in women.

All mouse studies were approved and performed in compliance with the Stanford University Institutional Animal Care and Use Committee, known as the Administrative Panel on Laboratory Animal Care (APLAC, protocol no. 10382). Mice were housed in Stanford's Comparative Medicine Pavilion and Shriram Center animal research facilities with a 12 h light–dark cycle (7:00–19:00) at 22°C ambient temperature with 40% humidity in compliance with practices detailed in the National Institutes of Health and the Institutional Animal Care and Use Committee (IACUC). The Association for Assessment and Accreditation of Laboratory Animal Care provides additional accreditation to Stanford University. Mice were maintained on a pure C57BL/6 background and genotyped by PCR. Mice of both sexes were used for all experiments. To design Ptpn14 conditional knockout mice, two LoxP sequences were inserted at Ptpn14 intron 2 and intron 3, floxing exon 3, such that deletion of exon 3 following Cre expression resulted in a downstream reading frame shift with 12 nonsense amino acids and a stop codon at exon 4. These Ptpn14 +/fl mice were generated using CRISPR-mediated genome recombination. Four guide RNAs were designed, and single-strand guide RNAs (sgRNAs) were purchased from Synthego. After in vitro validation by DNA cutting efficiency, two sgRNAs were chosen for generating Ptpn14 +/fl mice: Ptpn14 -g1 (GGCCTACGATCTC) and Ptpn14 -g3 (CCAGGAGGCTTCCGGCGAAG). Single-strand donor DNA containing the exon 3 locus with LoxP sites was synthesized by IDT. The donor DNA contained 150 nt homologous arms on each side, exon 3, and two LoxPs inserted at the gRNA PAM site floxing exon 3. A mixture of 10 ng/µL sgRNA, 25 ng/µL CAS9 protein (IDT), and 10 ng/µL donor DNA was injected into C57Bl/6J pronuclear mouse zygotes. Injected embryos were then implanted into pseudopregnant foster mothers. A total of 215 embryos was injected, 17 pups were born, and three pups containing the LoxP sites were identified by PCR and DNA sequencing. The genotyping PCR to confirm the 5′ LoxP site used one primer upstream of the donor DNA (CATCAACGAACCCCAGCTCT) and one primer on the 3′ LoxP site (CCAGCAGCTCCATAACTTCGT). The PCR product was then sequenced to confirm the presence and sequence of the 5′ LoxP site. The same PCR strategy confirmed the correct insertion of the 3′ LoxP site, with one PCR primer downstream from the donor DNA (AAGTGCAGAGTCCAACGGAG) and one primer at the 5′ LoxP site (GGCCTACGATCTCAATAACTTCGTA). This 3′ LoxP PCR product was then confirmed by sequencing. The Ptpn14 +/fl founder mice were bred with C57Bl/6J for heterozygotes. The Ptpn14 fl/fl mice were achieved by breeding Ptpn14 +/fl heterozygotes. Based on these analyses, we selected multiple founder mice, which we crossed to C57BL/6J mice to generate colonies of pure C57BL/6J Ptpn14 conditional knockout mice. After the establishment of founder mice, subsequent mice were genotyped using a set of primers to identify the 5′ LoxP site (forward primer: 5′-CTTTTAGCTTGTCCCACGCTG-3′, reverse primer: 5′-AAGCAGAGGCTCGTAGCAA-3′), the 3′ Loxp site (forward primer: 5′-CCCCCTCGAGTTCCCCATAA-3′, reverse primer: 5′-AAGTCAGGAGTCCAACGGAG-3′), the 5′ LoxP site and a region outside the donor DNA (forward primer: 5′-CGGGTGGTGTCAGATCAGAC-3′, reverse primer: 5′-CCAGCAGCTCCATAACTTCGT-3′), and the 3′ LoxP site and a region outside the donor DNA (forward primer: 5′-GGCCTACGATCTCAATAACTTCGTA-3′, reverse primer: 5′-ACATCTTTTGCCACCATGCTC-3′). These Ptpn14 +/fl mice may now be ordered from The Jackson Laboratory (stock no. 040676). Mouse embryonic fibroblasts (MEFs) were generated from E13.5 embryos produced from a Ptpn14 +/ − intercross, as described previously ( Brady et al. 2011 ). MEFs were cultured in DMEM with high glucose (Gibco) supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin. To test deletion of the PTPN14 protein using the floxed allele, cells were plated in 6 well plates. Twenty-four hours after plating, cells were transduced with either Ad5-CMV-Cre (University of Iowa Viral Vector Core) or Ad5-CMV-empty (University of Iowa Viral Vector Core) at a multiplicity of infection of 100. Seventy-two hours later, protein was collected for Western blotting using a 2% SDS protein extraction buffer (2% SDS, 5 mM Tris at pH 6.8 in dH 2 O). The liver injury study was conducted as described previously ( Yi et al. 2013 ). Briefly, mice between 1 and 4 months of age were fed either their normal diet or a 0.2% DDC-supplemented diet for 3 weeks and then sacrificed for liver analysis. At sacrifice, mouse livers were perfused with PBS before fixation in formalin for 24 h. Liver tissue was then embedded and sectioned for staining using standard protocols. The number of YAP-positive nuclei per square millimeter was quantified by measuring the oval cell response area using FIJI image software and then dividing the total number of YAP-positive nuclei by this area. Empty areas within portal veins were subtracted from the total area measurement. In the uterus, RNAscope in situ hybridization (ISH) was performed on formalin-fixed, paraffin-embedded (FFPE) uterus tissue sections using the RNAscope 2.5 HD reagent kit-brown (Advanced Cell Diagnostics 322300) following the manufacturer's instructions. Briefly, tissue sections were baked for 1 h at 60°C, deparaffinized in xylene, and rehydrated through a graded ethanol series. Endogenous peroxidase activity was blocked by incubation in H 2 O 2 for 10 min at room temperature, followed by antigen retrieval in RNAscope target retrieval solution for 15 min at 99°C–103°C using a steamer and protease digestion with RNAscope Protease Plus for 30 min at 40°C in a HybEZ oven (Advanced Cell Diagnostics). Probe hybridization was performed for 2 h at 40°C, followed by sequential signal amplification and chromogenic detection according to the manufacturer's protocol. Sections were counterstained with hematoxylin, dehydrated, and mounted for imaging. The following probes were used: RNAscope Probe-Mm-Ptpn14 (Advanced Cell Diagnostics 493181), RNAscope Probe-Mm-Ppib (positive control; Advanced Cell Diagnostics 313911), and RNAscope Probe-DapB (negative control; Advanced Cell Diagnostics 310043). RNAscope for eye tissues was performed according to the manufacturer's protocol on fresh-frozen tissues. Following tissue sectioning, the slides were baked for 30 min at 60°C and postfixed in 4% PFA overnight at 4°C. The sections were pretreated with RNAscope hydrogen peroxide (using RNAscope H 2 O 2 and protease reagents from ACDBio 322381) for 10 min at room temperature followed by 15 min antigen retrieval in boiling RNAscope target retrieval reagents (ACDBio 322000). The tissues were incubated for 10 min at 40°C with RNAscope Protease III, and the Ptpn14 RNAscope probes were applied on the slides for 2 h at 40°C. Signal detection was performed using the RNAscope Plus smRNA-RNA kit (ACDBio 322780). The slides were stained with DAPI, and coverslips were applied using the ProLong Gold antifade reagent with DAPI (Invitrogen P36931 ) mounting media. Western blotting was performed according to standard protocols. Briefly, protein quantification was performed using the Pierce BCA protein assay kit (Thermo Fisher). For each sample, 40 µg of protein was loaded onto a 4%–12% SDS-PAGE gel (Bio-Rad) and transferred to a PVDF membrane (Immobilon, Millipore). The membrane was blocked with a solution of 5% milk and 0.1% Tween 20 in Tris-buffered saline and then probed for either PTPN14 (1:100; Santa Cruz Biotechnology sc-373766) or α-tubulin (1:5000; Sigma T6074). For secondary antibodies, blots were probed with antimouse or antirabbit HRP-conjugated secondary antibodies (1:5000; Vector Laboratories), developed with Clarity Western ECL substrate (Bio-Rad), and imaged using a ChemiDoc XRS + (Bio-Rad). Gels were analyzed using Image Lab v.3.0 (Bio-Rad). Mouse tissue was dissected and collected postmortem and fixed in formalin overnight. Fixed tissues were stored in 70% ethanol prior to embedding. Hematoxylin and eosin staining and immunofluorescence were performed using standard protocols. For tissue immunofluorescence experiments, the following primary antibodies were used: keratin 14 (1:400; Thermo Scientific ms-115), keratin 19 (1:100; Developmental Studies Hybridoma Bank TROMA-III), keratin 12 (1:500; Abcam ab18562), Ki67 (1:100; BD Pharmingen 550609), Yap (1:50;, Santa Cruz Biotechnology sc-101199), and P-cadherin (1:1000; R&D Systems AF76). The ratio of YAP nuclear to cytoplasmic staining was quantified using FIJI image software to measure the mean gray value of the nuclear area, which was divided by the mean gray value of the cytoplasmic area for regions of epithelial cells in the mouse uteri ( Schindelin et al. 2012 ). For evaluation of eye tissues, the eyes were carefully enucleated using a curved tweezer; immediately fixed in 4% PFA, 20% isopropanol, 2% trichloroacetic acid, and 2% zinc chloride (diluted in distilled water) for 1 h at room temperature; and then stored in 70% ethanol at 4°C prior to processing. The tissues were washed three times with 1× PBS for 5 min on the rotor, and eyes were placed in 15% sucrose until the tissues were settled at the bottom of the tubes and transferred to 30% sucrose overnight at 4°C. The eyes were placed inside a biopsy blocking well containing OCT at the correct orientation and then kept at −80°C until the OCT completely solidified. The eyes were sectioned on a Leica CM1850 UV-3-1 cryostat microtome (Leica Biosystems) at 10 µm thickness. Antigen retrieval was performed on the sections using a heat steamer and 10 mM sodium citrate buffer (pH 6.0), and the sections were blocked for 1 h in 10% horse serum diluted in PBS. The appropriate primary antibody solution was added to the sections and then incubated overnight at 4°C. The tissues were washed three times in PBS for 5 min each, and the appropriate secondary antibody was applied to the sections. The slides were mounted with n-propyl gallate antifading reagent (Sigma-Aldrich P3130) and imaged using a Leica DM5000-B fluorescence microscope.