Generation and characterization of a mouse model of conditional Chd4 knockout in the endometrial epithelium.

The endometrium is a dynamic, hormone-responsive tissue that forms the inner uterine lining and consists of luminal epithelium, glandular epithelium, and stroma. The primary physiological function of the endometrium is to prepare for and maintain pregnancy [ 1 ]. In response to ovarian steroid hormones, estrogen and progesterone, the endometrium undergoes cyclic proliferation and breakdown each menstrual cycle. Endometrial function is dependent upon the ability of endometrial tissue to respond to hormonal signals through coordinated transcriptional programs, a process mediated in part by chromatin remodeling [ 2 – 4 ]. Chromatin remodeling genes play an integral role in the maintenance of endometrial homeostasis [ 2 – 4 ], and their disruption through mutations or alterations is associated with endometrial pathologies, including endometriosis [ 5 ], endometrial hyperplasia [ 6 ], and 66% of endometrial carcinomas [ 7 ]. Chromatin remodeling proteins function by using energy from ATP hydrolysis to alter nucleosome structure, chromatin accessibility, and ultimately, gene transcription [ 8 , 9 ]. We previously showed that SWI/SNF chromatin remodeling proteins, ARID1A and BRG1, regulate endometrial epithelial identity and integrity, with the loss of SWI/SNF leading to an epithelial-to-mesenchymal transition (EMT) phenotype in mice [ 10 , 11 ]. Chromodomain-helicase-DNA-binding protein 4 (CHD4) is an ATP-dependent chromatin remodeling protein and integral subunit of the nucleosome remodeling and deacetylase (NuRD) complex, which regulates transcriptional repression, DNA damage response, and cell cycle progression [ 12 – 17 ]. CHD4 is the predominant chromatin remodeling gene mutated or altered in an aggressive form of endometrial carcinoma known as uterine serous carcinoma. Approximately 17% of cases harbor CHD4 mutations [ 18 , 19 ], with the majority of mutations resulting in loss of CHD4 function [ 20 , 21 ]. CHD4 has been shown to have tumor suppressive properties in human endometrial carcinoma cell lines, with CHD4 loss leading to increased invasion and stemness in vitro [ 20 ]. CHD4 has been associated with the acquisition of a metastatic phenotype in several cancer types, including ovarian, colorectal, papillary thyroid, and breast cancers [ 22 – 25 ]. However, little is known about the tissue-specific functions of CHD4 within the normal endometrium in vivo or the extent to which CHD4 loss may contribute to endometrial disease. To this end, we generated a conditional Chd4 floxed allele, which, when combined with BAC-Sprr2f-Cre, targets Chd4 knockout to the endometrial epithelium [ 26 ], the proposed cell of origin for endometrial carcinoma [ 27 ]. We created the first endometrial epithelial-specific conditional knockout of Chd4 and characterized the structural and functional consequences of CHD4 loss.

The knockout-first, conditional Chd4 floxed-neo allele ( Chd4 tm1a(EUCOMM)Wtsi /Chd4 fl-neo ) was obtained from the Mutant Mouse Resource and Research Center in the targeted mutation 1a (tm1a) configuration. The Chd4 fl-neo allele contains a lacZ reporter and neomycin selection cassette flanked by flippase recognition target (FRT) sites and loxP sites flanking critical exons 11 and 12 ( Fig 1a - b )(#037690-UCD) [ 28 ]. The presence of the Chd4 fl-neo allele was confirmed by X-gal staining. The LacZ reporter present in the Chd4 fl-neo allele hydrolyzes X-gal substrate, whose product produces a blue precipitate upon further oxidation [ 36 ]. This color change was apparent in the uterus, kidney, and liver from the mouse containing the Chd4 fl-neo allele ( Fig 1c ). a. Schematic of the knockout-first, conditional Chd4 floxed allele. The targeted mutation 1a (tm1a) allele configuration contains flippase recognition targets (FRT) flanking a lacZ reporter and a neomycin resistance cassette. LoxP sites flank the neomycin cassette and critical exons 11 and 12. The tm1b allele configuration (gray box) is produced when Cre recombinase is introduced prior to FLP1 recombination (not generated in this study). The tm1c allele configuration is produced following FLP1 recombination. The tm1d allele configuration is only produced in cells expressing Cre recombinase. The use of BAC-Sprr2f-Cre drives conditional knockout of the Chd4 floxed allele in the endometrial epithelium. b. Tabular summary of the tm1a, tm1c, and tm1d allele configurations for the Chd4 fl-neo allele. c. X-gal staining was used to assess the functionality of the lacZ reporter gene in mice within the Chd4 fl-neo allele in Chd4 fl-neo / + and control ( Chd4 + / + ) mice. The lacZ reporter gene allows cells with the Chd4 fl-neo allele to produce β-galactosidase, which hydrolyses X-gal and produces a blue/green color. The control uterus and the kidney had some background staining, which could result from endogenous β-galactosidase activity known to occur in these two organs [ 37 ]. Scale bar = 1000 µm. d. A selective breeding scheme was used to refine the Chd4 fl-neo allele to the Chd4 fl allele to ultimately lead to endometrial epithelial Chd4 knockout. i. A female R26 Fki homozygote was crossed with a Chd4 fl-neo/ + sire. ii. FLP1 recombination occurred in offspring that inherit both the R26 Fki and Chd4 fl-neo alleles, creating the Chd4 fl allele. The R26 Fki allele was no longer needed and was bred out by crossing Chd4 fl/ + ; R26 Fki with a wild-type C57BL/6 mouse. iii. The resultant Chd4 fl/ + offspring were crossed to create female Chd4 fl-only homozygotes. iv. A BAC-Sprr2f-Cre   0/+ Chd4 + / + sire is crossed with a Chd4 fl/fll dam. v. The resultant BAC-Sprr2f-Cre 0/+ ; Chd4 fl/ + sires were then crossed with Chd4 fl/fl dams. vi. The resultant BAC-Sprr2f - Cre 0 / + ; Chd4 fl/fl female offspring had endometrial epithelial Chd4 knockout. e. - h. Representative PCR genotyping gels to confirm: the presence of the Chd4 fl-neo allele at 536 bp ( e ), the R26 Fki at 725 bp ( f ), the Chd4 fl allele at 413 bp and the wild-type allele at 219 bp ( g ), and the Cre allele for BAC-Sprr2f-Cre mice at 300 bp ( h ). A targeted breeding scheme was performed to achieve conditional knockout of Chd4 exclusively in the endometrial epithelium driven by BAC-Sprr2f-Cre (JAX: #037052) ( Fig 1d ) [ 26 ]. The lacZ reporter and neomycin selection cassette present in the Chd4 fl-neo allele are not needed in vivo and were removed to prevent off-target activity [ 38 ]. The first step toward FLP1 recombination was crossing a sire heterozygous for the Chd4 fl-neo allele with a dam homozygous for the R26-Flp knock-in allele ( R26 Fki ; JAX #016226) ( Fig 1d , i ) [ 29 ]. FLP1 recombination occurred in the resultant offspring that inherit both the R26 Fki and Chd4 fl-neo alleles, producing the Chd4 fl allele ( Fig 1d , ii). To prevent future FLP1 recombination events, Chd4 fl-only ; R26 Fki offspring were crossed with a wild-type C57BL/6 mouse to remove the R26 Fki allele ( Fig 1d , ii). The resultant offspring that were heterozygous for the Chd4 fl allele were then crossed to produce female Chd4 fl- homozygotes ( Fig 1d , iii). Next, the BAC-Sprr2f-Cre allele was introduced by crossing a Chd4 fl/fl dam with a sire hemizygous for the BAC-Sprr2f-Cre ( Fig 1d , iv). Lastly, the resultant BAC-Sprr2f-Cre 0/ + ; Chd4 fl/ + male offspring were crossed with Chd4 fl- homozygotes ( Fig 1d , v) to produce BAC-Sprr2f-Cre ; Chd4 fl/fl females ( Fig 1d , vi). The resultant BAC-Sprr2f 0/ + Chd4 fl/fl female offspring were conditional knockout (cKO) mice, lacking CHD4 expression in the Cre-expressing, SPRR2F-positive cells of the endometrial epithelium. The presence of the Chd4 fl-neo ( Fig 1e ), R26 Fki ( Fig 1f ), Chd4 fl ( Fig 1g ), wild-type ( Fig 1g ), and Cre ( Fig 1h ) alleles was confirmed by PCR. The R26 mT/mG fluorescent Cre-reporter allele (JAX #: 007676) was used to characterize the activity of BAC-Sprr2f-Cre across the female reproductive tract in 12-week-old BAC-Sprr2f- Cre 0 /+ ; R26 mT/mG and control ( BAC-Sprr2f- Cre 0 /+ ; R26 +/+ ) mice [ 30 ]. In mice with the R26 LSL-mT/mG allele, membrane-targeted fluorescent Tomato is ubiquitously expressed in all cells. In BAC-Sprr2f- Cre 0 /+ ; R26 mT/mG mice, Cre-recombination occurs in SPRR2F-expressing cells, resulting in a switch from mTomato to enhanced green fluorescent protein (EGFP) expression ( Fig 2a ). Expression of the BAC-Sprr2f-Cre allele begins at the onset of puberty in a stochastic manner, with initial reports finding a 50% recombination in 6-week-old mice [ 26 ]. In concordance with previously published results, BAC-Sprr2f-Cre was expressed exclusively in the endometrial epithelium, and expression was absent in the endometrial stroma, oviducts, and ovaries ( Fig 2b ) [ 26 ]. The variegated expression characteristic of the BAC-Sprr2f-Cre allele appears stronger in the luminal epithelium compared to the glandular epithelium by anti-GFP IHC in BAC-Sprr2f- Cre 0 /+ ; R26 mT/mG ; Arid1a fl/fl mice when compared to control mice ( Fig 2c ). a. Schematic of the R26 mT/mG allele. The R26 LSL-mT/mG allele consists of a membrane-targeted Tomato fluorescent reporter (mT) and a Lox-STOP-Lox (LSL) cassette, flanked by lox P sites. Before Cre-recombination, all cells ubiquitously express mTomato fluorescence (channel not shown). In SPRR2F- expressing cells, BAC-Sprr2f-Cre-mediated recombination excises the LSL cassette, driving a switch from mTomato to membrane-targeted enhanced green fluorescent protein (mEGFP/mG). b. Photomicrograph of uteri from 12-week-old control ( BAC-Sprr2f-Cre 0/+ ; R26 +/+ ) and BAC-Sprr2f-Cre 0/+ ; R26 mT/mG mice. Cre-positive cells within the endometrial epithelium fluoresce green in BAC-Sprr2f- Cre 0/+ ; R26 mT/mG mice. Cre expression is absent in the oviducts and ovaries. Scale bar = 100 µm. c. Immunofluorescence staining for epithelial marker keratin 8 (KRT8-AF647 pseudo-colored red), green fluorescent protein (GFP), and DAPI nuclear stain was performed on uterine cryosections (from the mice in b ). In the three-color merged image, co-expression is shown as yellow. The BAC-Sprr2f-Cre allele expression drives R26 mT/mG expression exclusively in the endometrial epithelium and is absent in the stroma. Maximum intensity projection images at 400x magnification; scale bar = 10 µm. We have created a novel mouse model of endometrial epithelial-specific CHD4 loss driven by BAC-Sprr2f-Cre . For the initial pilot cohort, three Chd4 conditional knockout (cKO, BAC-Sprr2f-Cre 0/+ ; Chd4 fl/fl ) mice and three Cre-negative control ( BAC-Sprr2f-Cre +/+ ; Chd4 fl/fl or Chd4 fl/+ ) mice were sacrificed at 12 weeks of age to confirm loss of CHD4 in endometrial epithelium and to assess endometrial histology. The 12-week-old mice from the pilot cohort were not estrous cycle staged. Knowing that the expression of the BAC-Sprr2f-Cre allele is estrogen-dependent and cumulative estrogen exposure increases with age, we wanted to examine the extent of CHD4 loss in a cohort of 26-week-old diestrus-staged Chd4 cKO (n = 3) and control (n = 3) mice [ 26 ]. In both 12- and 26-week-old mice, control mice showed strong nuclear expression of CHD4 in the luminal and glandular epithelia ( Fig 3a ). The Chd4 cKO mice at both ages exhibited variegated expression of CHD4 in the luminal and glandular epithelia, with notable variation observed between mice ( Fig 3a - b ). The extent of CHD4 loss was determined by manually counting DAB-positive and DAB-negative cells from anti-CHD4 IHC using Fiji/ImageJ. The percentage of cells expressing CHD4 was calculated as the number of CHD4 (DAB)-positive cells divided by the total number of cells (CHD4-positive + CHD4-negative) multiplied by 100% and was quantified for the luminal, glandular, and total (luminal + glandular) epithelia ( Fig 3c - h ). Therefore, the percentage of cells with CHD4 loss was calculated as 100% minus the percentage of cells expressing CHD4. a. CHD4 loss was assessed by anti-CHD4 immunohistochemistry (IHC). Representative images of the endometria from 12- and 26-week-old Chd4 cKO ( BAC-Sprr2f-Cre 0/+ ; Chd4 fl/fl ) and ( BAC-Sprr2f-Cre +/+ ; Chd4 fl/fl or Chd4 fl/+ ) control mice. Images were acquired at 20x magnification; Scale bar = 200 µm. Dashed boxes indicate regions shown at 40x magnification in the inset below each image (scale bar = 50 µm). The black and red arrowheads label the luminal and glandular epithelia, respectively. b. Representative images of anti-CHD4 IHC showing the variation in CHD4 expression in the 12-week-old (cKO1–3) and 26-week-old (cKO4–6) mice. Images were taken at 40x magnification; scale bar = 50 µm. c. - e. The percentage of CHD4 expression between 12- and 26-week-old Chd4 cKO mice was assessed in the luminal ( c ), glandular ( d ), and total ( e ) epithelia. f. - h. Comparison of the percentage of CHD4 expressing cells between Chd4 cKO (n = 6) and control (n = 6) mice in the luminal ( f ), glandular ( g ), and total ( h ) epithelia. c. - h. The percentage of CHD4-expressing cells was calculated as: the number of CHD4-positive cells divided by the total number of cells multiplied by 100% and was calculated for the luminal, glandular, and total epithelial cells. In all scatterplots, the horizontal line represents the mean; error bars indicate ± 1 standard deviation. The statistic used was a two-tailed, unpaired t test. Welch’s Correction was only used for unequal variance. Significance thresholds: ns = not significant; p < 0.05 = *, p < 0.01 = **, p < 0.001 = ***, p < 0.0001 = ****. To assess whether age affects BAC-Sprr2f-Cre expression, we assessed the percentage of cells with and without CHD4 expression in 12- and 24-week-old Chd4 cKO mice. At both 12 and 26 weeks old, Chd4 cKO mice had a lower percentage of cells expressing CHD4, and therefore, a higher percentage of cells with CHD4 loss, in the luminal epithelium ( Fig 3c ) compared to the glandular epithelium ( Fig 3d ). There was no significant difference in the percentage of cells expressing CHD4 in the luminal ( Fig 3c ), glandular ( Fig 3d ), or total epithelia ( Fig 3e ) between 12- and 26-week-old Chd4 cKO mice. With age-related differences in the extent of CHD4 loss ruled out, the 12- and 26-week-old mice were combined into six Chd4 cKO and six control mice to assess the percentage of cells with CHD4 expression loss. The percentage of cells with CHD4 loss in the luminal ( Fig 3f ), glandular ( Fig 3g ), and total epithelia ( Fig 3h ) was 56.1.0%, 31.2%, and 47.5%, respectively, when compared to control mice ( Table 2 ). Hematoxylin and eosin (H&E) staining showed no apparent differences in the endometrial architecture between Chd4 cKO and control mice at 12 or 26 weeks of age ( Fig 4a ). Chd4 cKO mice had normal endometrial histology by H&E staining, similar to that of control mice at both 12 and 26 weeks of age ( Fig 4b ). The expression of epithelial marker E-cadherin appeared similar in the glandular and luminal epithelia between genotypes in the 12- and 26-week-old mice ( Fig 4c ). IHC showed minimal expression of cleaved caspase 3 (CC3) in both 12- and 26-week-old mice, independent of genotype ( Fig 4d ). ( Fig 4c - d ). Variable expression of Ki-67 occurred across the non-estrous cycle staged 12-week-old mice from the pilot cohort, consistent with known changes in proliferation throughout the murine estrous cycle( Fig 4e ) [ 39 ]. The expression of Ki-67 was assessed in the diestrus-staged 26-week-old Chd4 cKO and control mice ( Fig 4e ). There was no significant difference in the percentage of Ki-67-positive cells in the luminal ( Fig 4f ), glandular ( Fig 4g ), or total ( Fig 4h ) epithelia between 26-week-old Chd4 cKO and control mice. a. - b. Representative images of hematoxylin and eosin (H&E) stained endometria from 12 and 26-week-old Chd4 cKO and control mice. a. Low-power (4x magnification; scale bar = 600 µm) images showing similar endometrial architecture between genotypes at both 12- and 26-week-old mice. b. Higher magnification (20x magnification; scale bar = 200 µm) image of the slide shown in ( a ). D a shed boxes indicate regions shown at 40x magnification in the inset below each image. The black and red arrowheads label the luminal and glandular epithelia, respectively. c. - e. Representative images of IHC staining of endometria from 12- and 26-week-old Chd4 cKO and control mice. Images were acquired at 20x magnification; Scale bar = 200 µm. IHC was performed to assess the expression of epithelial (E-cadherin) ( c ), apoptotic (Cleaved Caspase 3/CC3) ( d ), and proliferative (Ki-67) ( e ) markers between genotypes at 12 and 26 weeks of age. f. - h. The percentage of Ki-67 expression was assessed in the luminal ( f ), glandular ( g ), and total ( h ) epithelia of the diestrus-staged, 26-week-old mice (n = 3 Chd4 cKO; n = 3 control). The percentage of cells expressing Ki-67 was calculated as the number of Ki-67-positive cells divided by the total number of cells multiplied by 100% and was calculated separately for the luminal, glandular, and total epithelia. In all scatterplots, the horizontal line represents the mean; error bars indicate ± 1 standard deviation. The statistic used was a two-tailed, unpaired t test. Significance thresholds: ns = not significant; p < 0.05 = *, p < 0.01 = **, p < 0.001 = ***, p < 0.0001 = ****. The functional effects of CHD4 loss on endometrial function were assessed by a 6-month breeding trial in which Chd4 cKO (n = 8) and control (n = 8) dams were serially bred to a C57BL/6 sire ( Fig 5a ). Over the course of the breeding trial, all Chd4 cKO and control dams successfully mated and gave birth to live pups, which were weighed 24 or 72 hours post-parturition ( Fig 5b ). Dam age ranged from 8 to 14 weeks of age at the start of the trial, and there was no significant difference in the dam age between genotypes ( Fig 5c ). There was no difference in the average litter size ( Fig 5d ), the average litter size per dam ( Fig 5e ), the number of litters per dam ( Fig 5f ), or mean pup weight between genotypes when weighed 24 or 72 hours post-parturition or when combined ( Fig 5g )( Table 3 ). a. Visual summary of experimental design for the Chd4 breeding trial. b. Representative timeline of mouse pregnancy containing both timepoints to assess pup weight: 24 and 72 hours post-parturition. c. There was no difference in dam age between genotypes upon starting the breeding trial. d. There was no difference in the average litter size between genotypes. Each datapoint represents one litter. e. There was no difference in the average litter size per dam between genotypes. Each datapoint represents one dam. f. There was no genotype-specific difference in the number of litters per dam. g. There was no difference between genotypes in pup weight at 24 or 72 hours post-parturition or when the timepoints were combined. Blue and black datapoints denote pups weighed 24 and 72 hours post-parturition, respectively. c. - g. In all scatterplots, the horizontal line represents the mean; error bars indicate ± 1 standard deviation. The statistic used was a two-tailed, unpaired t-test. Significance thresholds: ns = not significant; p < 0.05 = *, p < 0.01 = **, p < 0.001 = ***, p < 0.0001 = ****.

Somatic mutations in CHD4 are found in the endometria of healthy women and those with endometrial cancer [ 7 , 40 ]. Knowing that endometrial carcinomas and other pathologies of the endometrium, including endometrial hyperplasia [ 6 ] and endometriosis [ 41 ], are thought to arise from the endometrial epithelium, we created an endometrial epithelial-specific mouse model of conditional CHD4 loss. Through targeted breeding, FLP1 recombination excised the FRT site-flanked lacZ reporter and neomycin cassette, creating the Chd4 fl allele. At the onset of puberty, BAC-Sprr2f-Cre induced Cre recombination of the Chd4 fl allele, leading to conditional loss of CHD4 in the Cre-expressing cells of the endometrial epithelium. Mice with conditional Chd4 knockout exhibited a variegated expression pattern characteristic of BAC-Sprr2f-Cre activity reported in the literature [ 26 , 42 ]. Although variegated, CHD4 loss was more extensive in the luminal epithelium than in the glandular epithelium and did not change with age. Chd4 cKO mice did not show any gross anatomical or histological differences when compared to control mice. Endometrial function was assessed by a 6-month breeding trial, which showed no difference in fertility between genotypes. Collectively, these results demonstrate that variegated CHD4 loss does not alter endometrial structure or function in 6-month-old mice. Is this lack of phenotype due to the incomplete knockout caused by variegation, or is Chd4 dispensable for endometrial epithelial function? In mice, germline knockout of Chd4 ( Chd4 -/- ) is embryonic lethal due to implantation failure [ 43 ]. In contrast, Chd4 +/- (heterozygous) mice survive into adulthood but show altered growth, as well as altered neurological, cardiovascular, and reproductive development [ 44 ], similar to what is observed in Sifrim-Hitz-Weiss syndrome in humans [ 45 , 46 ]. Importantly, mice heterozygous for CHD4 loss can still reproduce, which may explain why the Chd4 cKO mice with 52.5% remaining CHD4 expression in endometrial epithelia exhibited normal fertility. While CHD4 expression is haploinsufficient during development, the consequences of post-developmental CHD4 loss appear to be tissue specific. In malignancy, CHD4 has context-dependent functions as a tumor suppressor and an oncogene. CHD4 overexpression is associated with oncogenic activity, poor prognosis, and increased risk of colorectal cancer and ovarian cancer metastases [ 23 , 25 ]. In endometrial carcinoma, CHD4 has been reported to have both tumor suppressive [ 20 ] and oncogenic activity [ 47 ], supporting the notion that CHD4 gene dosage may be key to understanding the context-dependent CHD4 activity. BAC-Sprr2f-Cre was selected due to its reported specificity to the endometrial epithelium and absence in the endometrial stroma, myometrium, ovaries, oviducts, and kidneys [ 26 ]. To elucidate the extent to which variegated CHD4 loss contributed to the observed absence of a phenotype, an alternative endometrial epithelial-specific Cre driver, such as Lactoferrin i-Cre ( Ltf-iCre ), could be used to further evaluate the extent to which endometrial epithelial CHD4 is required for endometrial epithelial structure and function [ 48 ]. However, lactoferrin is expressed in other tissues, including mammary glands and neutrophils, the off-target effects of which may alter endometrial functions such as fertility [ 48 , 49 ]. The majority of CHD4 mutations in endometrial carcinoma are missense mutations and are thought to lead to reduced or loss of CHD4 function [ 20 , 21 , 50 ]. By targeting Chd4 knockout to the endometrial epithelium, we took a functional approach to both model the purported effects of these mutations and understand the tissue-specific consequences of CHD4 loss in this cell type. Phenotypic differences between somatic mutation and deletion of cancer-associated genes have been reported, including differences in tumorigenicity between TP53 R172H mutation and TP53 fl in a mouse model of KRAS G12V -mutant rhabdomyosarcoma [ 51 ]. Accordingly, it is conceivable that the consequences of a CHD4 mutation, even a loss-of-function mutation, may not be entirely congruent with the effects of a conditional knockout, as modeled here. Endometrial carcinoma most commonly arises in peri- and post-menopausal women [ 50 ]. The mice used in this study were still cycling and, therefore, may not have represented the atrophic endometrial epithelium common in post-menopausal women. Further, CHD4 is always co-mutated with other cancer driver genes and is frequently co-mutated with PTEN , PIK3CA , PIK3R1 , ARID1A , TP53 , and KRAS [ 50 ]. Many chromatin remodeling gene mutations are necessary but not sufficient to drive endometrial tumorigenesis, suggesting that CHD4 may require an additional genetic alteration or activating mutation to promote tumor formation [ 11 , 52 , 53 ].

All mouse experiments were performed under protocol #: 202500068, approved by the Michigan State University (MSU) Institutional Animal Care and Use Committee and in compliance with the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals. Mice were housed at the MSU Grand Rapids Research Center Vivarium on a standard 12-hour light-dark cycle with ad libitum food and water. Mice were humanely euthanized by CO 2 inhalation delivered by an automated Euthanex System. All mice were maintained on a C57BL/6 background. The C57BL/6N- Chd4 tm1a(EUCOMM)Wtsi /BayMmucd (knockout-first, conditional Chd4 floxed-neo allele, Chd4 fl-neo ) allele was obtained from Mutant Mouse Resource and Research Centers at the University of California Davis (MMRRC #037690-UCD) [ 28 ]. The BAC-Sprr2f-Cre (strain #: 037052) [ 26 ], R26-Flp knock-in ( R26 Fki ; strain #: 016226) [ 29 ], and R26-mT/mG ( R26 mT/mG ; strain #: 007676) [ 30 ] alleles were purchased from The Jackson Laboratory (Bar Harbor, ME). Inheritance of the BAC-Sprr2f-Cre , the R26 Fki , and the R26 mT/mG alleles was confirmed by PCR using published methods [ 30 – 32 ]. ; Ep400 fl / fl or Ep400 fl/+  floxed (fl/fl or fl/+) mice were used as control mice unless otherwise specified. The Chd4 fl-neo allele configuration was generated through a targeted breeding scheme that induced sequential recombination by FLP1 and Cre recombinase. The presence and configuration of the Chd4 fl-neo allele were confirmed by PCR in accordance with MMRC-UCD genotyping protocol guidelines [ 33 ]. Table 1 summarizes the primer sequences, primer combinations, and PCR product sizes for each allele configuration of the Chd4 fl-neo allele [ 33 ]. Reactions highlighted in grey were not performed or included. PCR reaction conditions consisted of initial denaturation at 94°C for 2 minutes, followed by 34 cycles at 94°C for 15 seconds, 58°C for 30 seconds, and 72°C for 1 minute, and ended with extension at 72°C for 5 minutes. Agarose gels were imaged using a ChemiDoc XRS+ System (Bio-Rad Laboratories). Vaginal lavage cytology was used to determine the murine estrous cycle stage as described [ 34 ]. Twelve-week-old pilot mice were not estrous cycle staged. Mice in the 26-week-old cohort were sacrificed in the diestrus phase of the estrous cycle. X-gal staining solution and wash buffer were prepared as described, except that the staining solution contained 20 mM Tris-HCl (pH 7.5) [ 35 ]. All wash steps were repeated three times using wash buffer. Isolated organs were fixed in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) overnight at 4°C, washed, stained with X-gal staining solution overnight, washed, and placed in 4% PFA overnight for post-staining fixation. Samples were then washed and cleared through a graded glycerol series (50%, 70%, 90%, 100%) with consecutive overnight incubations. Whole-mount tissue samples, including X-gal-stained tissues and R26 mT/mG mice, were imaged using a Nikon SMZ18 Stereo Microscope with either a 0.5x or 1x SHR Plan Apo objective, illuminated by a KL 1600 LED light source (SHOTT) for brightfield imaging or an X-Cite Series 120PC Q (EXFO) for fluorescence imaging. Images were captured using a Nikon DS-U3 camera and the NIS-Elements Software (Nikon Instruments Inc., Tokyo, Japan). Upon necropsy, one uterine horn was fixed in 10% neutral buffered formalin in PBS for 72 hours. Samples were then washed twice with PBS and transferred to 70% ethanol. Samples were submitted to the Van Andel Research Institute (VAI) Pathology & Biorepository Core for paraffin embedding, sectioning, and hematoxylin and eosin (H&E) staining. The other uterine horn was fixed in 4% paraformaldehyde (PFA) for at least 72 hours and stored at 4°C. Each uterine horn was washed twice with PBS, subjected to a sucrose gradient (24 hours in 15% sucrose in PBS followed by 24 hours in 30% sucrose in PBS), embedded in O.C.T. Compound (Fisher Health Care), and cryosectioned as described [ 11 ]. IF was performed as described [ 11 ]. Slides were blocked using donkey blocking solution: 5% normal donkey serum (Jackson Immunoresearch Laboratories # 017-000-121), 1% IgG-free bovine serum albumin (Jackson Immunoresearch Laboratories # 001-000-161), and 0.05% Tween 20 in PBS. Cryosections were co-stained with chicken anti-GFP (Abcam #ab13970, 1:500) and rat anti-KRT8 (DSHB #TROMA1, 1:100). The following secondary antibodies were purchased from Jackson Immunoresearch Laboratories and used at a 1:250 dilution: donkey anti-chicken AF 488 (JAX #703-545-155) and donkey anti-rat AF 647 (JAX #712-605-153). Autofluorescence reduction was performed using the TrueVIEW Auto-fluorescence Quenching Kit (Vector Laboratories #SP-8500). Slides were mounted and stained with DAPI using ProLong Gold Antifade Reagent with DAPI (Invitrogen). Slides were imaged using the Nikon C2 Plus Confocal microscope configured on a Nikon Eclipse Ti inverted microscope using a 40x Plan Fluor oil objective (NA 1.30). DAPI, GFP, and KRT8 were excited using 405, 488, and 637 nm laser lines, respectively. Emission was collected by PMT detectors at 500–550 nm for GFP and 660–1000 nm for AF647. DAPI emission was collected using the closest available detector window. Indirect immunohistochemistry (IHC) was performed as described [ 11 , 34 ]. Antigen retrieval was performed using 10mM sodium citrate buffer Primary antibodies were incubated overnight at the following dilutions: 1:200 anti-CHD4 (Cell Signaling Technology, CST #12011), 1:250 anti-Cleaved caspase 3 (CC3; CST # 9579), 1:400 anti-Ki-67 (CST #12202), 1:100 anti-Keratin 8 (KRT8; Developmental Studies Hybridoma Bank #TROMA-I), 1:400 anti-E-Cadherin (CST# 3195), and 1:400 anti-GFP (CST#2956). Stained slides were digitally scanned at 20x magnification and imported into Aperio eSlide Manager by the VAI Pathology & Biorepository Core. Histology images were acquired using Aperio ImageScope 12.4.6 software (Leica Biosystems). For each mouse, a representative anti-CHD4 and anti-Ki-67 IHC image was taken at 20x magnification. Aperio eSlide Manager was selected for quantification of endometrial epithelial CHD4 and Ki-67 expression. The multipoint tool in Fiji/Image J was used to manually count DAB-positive and DAB-negative endometrial epithelial cells for anti-CHD4 or anti-Ki-67 IHC. Counts for the luminal and glandular epithelia were enumerated separately and were combined to calculate the counts for total epithelia. The percentage of cells expressing CHD4 or Ki-67 was calculated by dividing the number of DAB-positive cells by the total number of cells (DAB-positive + DAB-negative cells) and multiplying by 100% for the luminal, glandular, and total epithelia. All statistical analyses were performed using GraphPad Prism 10. Scatter plots show the mean and standard deviation for each dataset. An unpaired, two-tailed t-test was used to assess differences between genotypes unless otherwise specified. Welch’s Correction was used for samples with unequal variance. Significance thresholds: ns = not significant; p < 0.05 = *, p < 0.01 = **, p < 0.001 = ***, p < 0.0001 = ****.

Primer sequences, product sizes, annealing temperatures, as well as genotyping protocol source, are also shown. (PDF) Raw data used to generate graphs shown in Figs 3 , 4 , and 5 . (XLSX)