The mature mammalian gut epithelial lining requires continuous cell proliferation that enables the replacement of lost cells and almost weekly tissue renewal (1
). This renewal driven by the Wnt–β-catenin pathway must be tightly controlled to ensure the proper differentiation of intestinal epithelial cells (IECs). IEC differentiation is required to maintain normal intestinal functions and to prevent the development of colorectal cancers (2
). Development of these cancers often depends on the interaction of the host with commensal microbiota and elicited tissue inflammation (4
). Inflammatory cytokines regulate the innate immune responses shaped by the microbiota and affect the rate of intestinal epithelial cell proliferation (6–9
). Among these cytokines are type I interferons (IFN) that activate a cognate cell surface receptor (consisting of the IFNAR1 and IFNAR2 chains) and signal to induce the transcription of IFN-stimulated genes (ISGs), some of which are known for their antiproliferative properties (10
). Despite the known suppressive effects of IFN on cell proliferation and constitutive induction of IFN in the gut (11
), the role of these cytokines in regulating intestinal epithelium proliferation and function remains poorly understood.
Adequate expression of the IFNAR1 chain of the IFN receptor is required for all IFN effects (13–15
). Genetic studies using the Ifnar1
knockout in mice have yet to establish the role of IFN in regulating intestinal epithelial cell proliferation. Ablation of Ifnar1
, specifically in IECs (Ifnar1ΔIEC
), resulted in only a modest increase in bromodeoxyuridine (BrdU) labeling, and this increase was dependent on the changes in the commensal microbiota (16
). However, neither alteration in the microbiome profile (17
) nor increased IEC proliferation (17
) was observed in mice lacking Ifnar1
in all tissues compared to wild-type mice. These data suggest either that IFN do not play an important role in regulating intestinal epithelium proliferation or that genetic differences between Ifnar1+/+
animals are masked by additional factors.
Indeed, high levels of cell surface IFNAR1 are specifically required for the antiproliferative effects of IFN as opposed to their ability to elicit an antiviral state (14
). These levels of IFNAR1 are tightly regulated by its phosphorylation-dependent ubiquitination and subsequent endocytosis and lysosomal degradation (20–22
). The rate-limiting event in these processes is the phosphorylation of serine residues within the IFNAR1 degron that enables the recruitment of beta-transducin repeat containing protein ( βTrcp) E3 ubiquitin ligase and IFNAR1 ubiquitination (21
). Importantly, while this phosphorylation can be induced by IFN via the activation of protein kinase D2 (24
), there is also a ligand-independent pathway that removes IFNAR1 from the surface of cells that are yet to encounter IFN (23
). This pathway was shown to be activated by inflammatory cytokines (27
) and may contribute to the lack of phenotypic differences between wild-type and Ifnar1
knockout mice under inflammatory conditions (28
). We previously purified and characterized casein kinase 1α (CK1α) as a major ligand-independent kinase that is capable of phosphorylating IFNAR1 in vitro
Importantly, CK1α is also a critical mediator of β-catenin ubiquitination and degradation (2
). Priming phosphorylation of β-catenin by CK1α greatly increases the phosphorylation of the β-catenin degron by glycogen synthase kinase 3β (GSK3β) (30
) that is required for its recognition by the βTrcp E3 ubiquitin ligases and subsequent ubiquitination and proteasomal degradation (32–36
). Our recent studies demonstrated that gut-specific knockout of the Csnk1a1
gene that encodes CK1α leads to robust stabilization of β-catenin and activation of Wnt target genes (37
). Intriguingly, ablation of Csnk1a1
alone did not lead to either epithelial cell hyperproliferation or tumorigenesis. Instead, inactivation of CK1α induced the DNA damage response (DDR) and p53/p21-dependent senescence. These events appear to prevent tumorigenesis driven by hyperactive β-catenin because the concurrent ablation of CK1α with either p53 or p21 resulted in hyperproliferation and a rapid development of aggressive and invasive intestinal tumors (37
Here we determined the role of CK1α in the regulation of IFNAR1 ubiquitination and levels in vivo. We found that despite the accumulation of IFNAR1 protein (but not mRNA), the ubiquitination of IFNAR1 was decreased in CK1α-deficient intestinal tissues. In addition, the expression of IFN-stimulated genes was increased in the gut upon Csnk1a1 ablation. As the lack of CK1α stabilized both β-catenin and IFNAR1, the phenotype associated with the loss of Csnk1a1 highlighted the contribution of IFN signaling to control IEC proliferation and function. Intriguingly, IFN signaling was required for the effective activation of p53 and p21 and induction of senescence and apoptosis in the CK1α-deficient intestinal epithelium. Furthermore, the concurrent ablation of CK1α with Ifnar1 led to unrestricted IEC proliferation to an extent that caused profound aberrations of gut barrier function and rapid animal death. These results demonstrate that IFN play an important role in restricting intestinal epithelial cell proliferation elicited by the activated β-catenin pathway.
MATERIALS AND METHODS
All experiments with animals were carried out under protocol 803995 approved by the IACUC of the University of Pennsylvania. All mice were on the C57BL/6 background, had water ad libitum
, and were fed regular chow. Ifnar1−/−
mice (a kind gift of Dong-Er Zhang, UCSD) were crossed with Csnk1a1lox/lox
mice, which bear floxed Csnk1a1
), to generate Csnk1a1lox/lox
mice bearing Vil1
in genotype identification is omitted for simplicity). Genotyping was performed on the tail of 4-week-old pups according to standard protocols using previously described primers (37
Tamoxifen (Sigma) was dissolved in corn oil (Sigma), and mice were injected intraperitoneally (120 mg/kg of body weight) on two consecutive days. On day 5 after the last injection, mice were euthanized. The jejunum, the ileum, and the entire large intestine were flushed with ice-cold phosphate-buffered saline (PBS); cut open longitudinally; and subjected to fixation in 4% formaldehyde and paraffin embedding. Small pieces of the jejunum were embedded in Tissue-Tek OCT compound (Sakura) and frozen at −80°C. IECs were isolated from the middle part of the small intestine as described previously (39
) but with the following slight modifications: intestinal cells were separated into single cells in Hanks' balanced salt solution containing 5 mM EDTA at 4°C for 30 min.
For antibiotic treatment, Csnk1a1Δgut; Ifnar1−/− mice were gavaged with 100 mg streptomycin (Sigma), and the drinking water was immediately replaced with filter-sterilized water containing ampicillin (1 g/liter; American Bioanalytical), vancomycin (0.5 g/liter; MP Biomedicals), neomycin (1 g/liter; Sigma), metronidazole (1 g/liter; Sigma), and 1% sucrose (Fisher). Antibiotic-containing water was replaced at least once a week during the course of the experiment. For all experimental groups, either mice were cohoused or their feces were swapped daily between cages to minimize potential differences in the gut microbiota.
Histology and immunotechniques.
Sections (5 μm) were cut for hematoxylin and eosin (H&E) staining and immunohistochemistry analysis. For immunohistochemistry, sections were incubated with antibodies to detect CKIα (C-19 [1:1,000]; Santa Cruz Biotechnology), β-catenin (1:200; Cell Signaling), cyclin D1 (SP4 [1:100]; Thermo Scientific), Ki67 (2.5 mg/ml; BioLegend), cleaved caspase-3 (1:100; Cell Signaling Technology), and p53 (1:500; NovoCasta). Secondary antibodies were horseradish peroxidase (HRP)-conjugated anti-rabbit, anti-goat, and anti-rat antibodies (Millipore, Cell Signaling Technology). 3,3′-Diaminobenzidine (DAB) chromogen (Lab Vision) was used for detection. For whole-tissue immunofluorescence, pieces of small intestine harvested from mice were frozen in Tissue-Tek OCT compound, cryosectioned by using Leica CM3050 S cryostats, fixed in acetone, washed, and blocked with PBS containing 5% goat serum. The sections were incubated for 1 h with primary antibodies to detect IFNAR1 (2 μg/ml; Sino Biological), γH2AX (1:100; Millipore), E-cadherin (1:500; Millipore), or TJP1/ZO-1 (1:100; Thermo Fisher). The sections were then washed, incubated with the corresponding secondary antibodies labeled with Alexa Fluor 488 or 594 (Invitrogen) for 1 h, washed again, and mounted onto coverslips by using mounting solution with 4′,6-diamidino-2-phenylindole (DAPI) (Prolong Gold). For senescence-associated β-galactosidase (SA-βGal) staining (described in detail in reference 40
), 10-μm sections were cut from OCT-embedded frozen tissue and allowed to adhere to coated slides at 25°C for 1 min before fixation for 15 min. Staining was performed according to the instruction provided with the senescence β-galactosidase staining kit (catalog number 9860S; Cell Signaling). After staining, sections were counterstained with nuclear fast red, dehydrated, and mounted. Periodic acid-Schiff (PAS) staining for goblet cell determination was performed according to standard protocols. Numbers of γH2AX-, IFNAR1-, and Ki67-positive foci per crypt/villus axis were determined by counting foci in 40 or 20 low-power fields (magnification of ×200 or ×400). The number of positive cells per crypt (Ki67, cleaved caspase-3, and SA-βGal) was determined by counting foci in 20 low-power fields (magnification of ×200).
Proteins were extracted from intestinal epithelial cell pellets in protein lysis buffer containing protease and phosphatase inhibitors according to whole-cell extract protocols. IFNAR1 was immunoprecipitated from whole-cell lysates by using MAR1-5A3 (Leinco Technologies, Inc.), as previously described (28
). Membranes were incubated with antibodies to detect IFNAR1 (2 μg/ml; Sino Biological), ubiquitin (Ub) (P4D1 [1:1,000]; Santa Cruz Biotechnology), CKIα (C-19 [1:1,000]; Santa Cruz Biotechnology), p21CIP1/WAF1
(ab7960 [1:1,000]; Abcam), interferon regulatory factor 7 (IRF7) (ab62505 [1:1,000]; Abcam), ZO-1 (1:100; Thermo Fisher), and β-actin (AC-74 [1:5,000]; Sigma). Secondary antibodies conjugated to horseradish peroxidase were purchased from Millipore Bioscience Research Reagents. Blots were processed as previously described (23
) and developed by using ECL (GE Healthcare).
Total RNA was extracted from cell pellets by using TRIzol reagent and phenol-chloroform methods. RNA (1 μg) was subjected to reverse transcription using a first-strand cDNA synthesis kit (Thermo Scientific), and mRNA expression levels were measured by quantitative real-time PCR using an Applied Biosystems 7500 Fast real-time PCR system. Relative quantities of gene transcripts were normalized to β-actin transcript levels. Sequences of PCR primers are as follows: 5′-TAGGCGGAATGAAGATGGAC (forward primer for Axin2), 5′-CTGGTCACCCAACAAGGAGT (reverse primer for Axin2), 5′-CAGTATCTCCCGGACTGAGG (forward primer for Cd44), 5′-GCCAACTTCATTTGGTCCAT (reverse primer for Cd44), 5′-GGTGCGGAAGATCGGATCT (forward primer for Csnk1a1), 5′-TTCACTGCCACTTCCTCGC (reverse primer for Csnk1a1), 5′-TTGACTGCCGAGAAGTTGTG (forward primer for cyclin D1), 5′-CCACTTGAGCTTGTTCACCA (reverse primer for cyclin D1), 5′-CTGCTGCCTGGGCTTCATAG (forward primer for Ifitm3), 5′-GGATGCTGAGGACCAAGGTG (reverse primer for Ifitm3), 5′-TCCACAGCGATATCCAGACA (forward primer for Cdkn1a), 5′-AGACAACGGCACACTTTGCT (reverse primer for Cdkn1a), 5′-GGAGCTCAGCAAGACTCTGG (forward primer for Sox9), 5′-TGTAATCGGGGTGGTCTTTCT (reverse primer for Sox9), 5′-GCCTACTCGTCGGAGGAA (forward primer for Ascl2), 5′-CCAACTGGAAAAGTCAAGCA (reverse primer for Ascl2), 5′-CCCTGTGAAGGAAGTGGCTA (forward primer for Oas2), 5′-CTGTTGGAAGCAGTCCATGA (reverse primer for Oas2), 5′-GTCAGAGTGGAAATCCTAAG (forward primer for Ifnβ), 5′-ACAGCATCTGCTGGTTGAAG (reverse primer for Ifnβ); 5′-CGACCAAGTGTGAATTCTCTTTAC (forward primer for Ifnar1), 5′-ATCAACCTCATTCCACGAAGAT (reverse primer for Ifnar1), 5′-ACCCGAAACTGATGCTGTGGATAG (forward primer for Tjp1), 5′-AAATGGCCGGGCAGAGACTTGTGTA (reverse primer for Tjp1), 5′-TGAAACGCCGACCTATCCTTA (forward primer for Trp53), 5′-GGCACAAACACGAACCTCAAA (reverse primer for Trp53), 5′-ATATTAACCGGCGCTACGAC (forward primer for Bak1), 5′-AGGCGATCTTGGTGAAGAGT (reverse primer for Bak1), 5′-ATGCTGTGGATCTGGGCTGTCCT (forward primer for Fas), 5′-GCATAATGGTTCTTGTCCATG (reverse primer for Fas), 5′-CCTCAAGTTTTGCCCTTTA (forward primer for Casp1), 5′-CCTTCTTAATGCCATCATCTT (reverse primer for Casp1), 5′-AGAGGGAAATCGTGCGTGAC (forward primer for β-actin), and 5′-CAATAGTGATGACCTGGCCGT (reverse primer for β-actin).
Isolation and culture of intestinal crypts.
Crypt culture was performed as previously described (41
). After intestinal crypt isolation, a total of 500 crypts were mixed with 50 μl of Matrigel (BD Biosciences) and plated into 24-well plates. After polymerization of Matrigel, 500 μl of crypt culture medium (Advanced Dulbecco's modified Eagle's medium [DMEM]–F-12 medium containing 50 ng/ml epidermal growth factor [EGF] [Invitrogen], 1 μg/ml R-spondin [Peprotech], and 100 ng/ml Noggin [Peprotech]) was added. Organoids were treated with 10 μM the CK1 inhibitor D4476 (Sigma), 10 μg/ml IFN-β neutralizing antibody (Leinco Technologies), and the corresponding vehicle or isotype controls. After 4 days of culture at 37°C, the numbers of live organoids were quantified.
Microarray analyses were performed with an Illumina whole-genome array. Total RNA was isolated from intestinal epithelial cells of Csnk1a1Δgut; Ifnar1+/+ (single knockout [SKO]) and Csnk1a1Δgut;Ifnar1−/− (double knockout [DKO]) mice at day 5 after the last tamoxifen treatment by using an miRNeasy minikit (Qiagen). Biotin-labeled cRNA samples were prepared by using a TargetAmp-Nano labeling kit (Epicentre) as recommended by the manufacturer. Thereafter, 0.75 μg cRNA was hybridized to Illumina Sentrix Mouse-6 v.1 BeadChips, which were scanned with an Illumina BeadStation 500 instrument (both from Applied Biosystems-Life Technologies, Inc.). Data were collected with Illumina BeadStudio 188.8.131.52 software, and statistical analyses were conducted with IlluminaGUI R-package15,70.
Intestinal permeability assay.
Barrier function was evaluated by measuring in vivo paracellular permeability to fluorescence-labeled dextran. Mice were fasted for 4.5 h and then gavage fed fluorescein isothiocyanate (FITC)-labeled 4.4-kDa dextran (FD4; Sigma). Plasma was obtained 5 h after gavage administration by terminal cardiac puncture after CO2 anesthesia. The plasma FD4 concentration was calculated by comparing samples with serial dilutions of known standards by using a Varioskan Flush fluorimeter (Thermo Scientific) with excitation at 485 nm and emission at 530 nm.
Data are presented as averages ± standard errors of the means (SEM). Statistical analysis was performed by using Microsoft Excel. Statistical significance was calculated by using a two-tailed Student t test. A P value of <0.05 was considered significant.
Microarray data accession number.
Raw data were deposited in the GEO database under accession number GSE76512
Our data presented here demonstrate that induction of IFN signaling appears to contribute to the activation of the DNA damage responses and apoptotic pathways as well as the suppression of intestinal epithelium proliferation that occurs upon the inactivation of CK1α. The latter event stabilizes both β-catenin and IFNAR1, thereby highlighting the conditions that determine the role of IFN signaling in restricting IEC proliferation. In addition, IFN contributes to the vitally important function of maintenance of intestinal barrier function.
CK1α is capable of phosphorylating numerous proteins and affecting a multitude of signaling pathways and transcriptional activities toward specific genes (reviewed in reference 48
). Whereas CK1α was capable of phosphorylating the IFNAR1 degron in vitro
), the role of other kinases in stimulating the recruitment of βTrcp to IFNAR1 and promoting its ubiquitination, endocytosis, and degradation was also demonstrated (24
). Our current data clearly characterize CK1α as a major regulator of IFNAR1 ubiquitination and stability in vivo
). Furthermore, these results implicate this kinase in the negative regulation of the IFN pathway in intestinal tissues and underscore the importance of CK1α function for the proliferation of the intestinal epithelium and its permeability.
Furthermore, our data suggest that the role of CK1α in regulating intestinal homeostasis is at least in part mediated by its effects on the stability and levels of IFNAR1 and the ensuing alterations in IFN signaling. Although microbiota-supported constitutive tonic IFN signaling has been described in the gut (11
), the role of this signaling in regulating gut renewal and function was a challenge to evaluate due to the potential phenotypic similarity of mice lacking the Ifnar1
gene and wild-type mice exhibiting rapid IFNAR1 degradation under inflammatory conditions (28
). The concurrent stabilization of IFNAR1 and β-catenin upon CK1α inactivation enabled us to determine that IFN plays an important role in restricting the proliferation and viability of the intestinal epithelium (Fig. 4
) and contributes to the maintenance of the equilibrium of the host-microbiota interaction and barrier function (Fig. 7
Our data further indicate that the IFN pathway contributes to the expression of p53-driven genes (including proapoptotic genes and Cdkn1a
) stimulated in the absence of CK1α (Fig. 3
). These data are consistent with our recently reported data demonstrating that IFN can amplify DNA damage responses (43
) as well as with data from previous reports that exogenous IFN can trigger an increase in p53 activities (44
). Importantly, these effects of IFN may provide an additional mechanism for the activation of the p53 pathway in the CK1α-deficient gut in addition to the previously reported downregulation of MDMX and the ensuing stabilization of the p53 protein (37
Intriguingly, activation of the p53 pathway was not observed in the Apc
-deficient gut (37
). Given that the ablation of either CK1α or APC results in the stabilization of β-catenin and stimulation of the Wnt pathway, the difference between these phenotypes and underlying gene profile signatures can be explained by at least two diverse reasons. First, APC possesses important functions that do not depend on the timely degradation of β-catenin. For example, APC can bind RNA and regulate the microtubule scaffold (50
). Furthermore, the loss of Apc
leads to the upregulation of Musashi proteins, which are pleiotropic translational regulators that affect numerous important signaling cascades, including the mTorc1 pathway (51
). Second, characteristic for the deletion of Cnsk1a1
(but not Apc
) induction of the DNA damage response, activation of the p53-driven genes, and cell senescence are likely augmented by IFN signaling activated by CK1α deletion. Indeed, massive hyperproliferation of the intestinal epithelium leading to a defective barrier function and rapid animal death that was seen in Cnsk1a1/Ifnar1
-deficient mice is reminiscent of the phenotype observed upon intestinal deletion of APC (46
) and is characterized by a similar gene expression signature (Fig. 6
The restriction of proliferation of Csnk1a1
-deficient IECs could be lifted by the concurrent inactivation of either p53, p21CIP1/WAF1
), or, to a lesser extent, IFNAR1 (this work). Importantly, ablation of either p53 or p21CIP1/WAF1
in CK1α-deficient animals led to the development of malignant tumors (37
). However, we did not observe these tumors in any groups of CK1α/IFNAR1-deficient mice that developed a lethal disruption of intestinal barrier function (Fig. 7
), suggesting a contribution of IFN-independent pathways to p53 function as a tumor suppressor in CK1α-null intestines.
A few overlapping and non-mutually exclusive mechanisms by which IFN contributes to the maintenance of the barrier function of the gut can be proposed. First, IFN-mediated restriction of the rate of cell proliferation should enable better differentiation, maturation, and establishment of the cohesive sheet of enterocytes and colonocytes. Second, IFN can contribute to immune defenses against conditionally pathogenic microbiota and intestinal inflammation. Katakura and coauthors previously reported a greater sensitivity of Ifnar1
-null mice to intestinal inflammation caused by dextran sodium sulfate (53
). Those authors also mentioned a positive effect of IFN on the protection of barrier properties of the epithelial cell sheet in vitro
. Importantly, our results implicate IFN in regulating the expression of the ZO-1 protein involved in the formation of tight junctions that separate the basolateral epithelial space from the microbiota. Potential immune-related effects of IFN on the ability to withstand the pathogenic effects of microbiota are illustrated by the rescue of Cnsk1a1/Ifnar1
-deficient mice upon administration of antibiotics (Fig. 7
Genetic alterations in the IFNAR1 gene in humans were linked with the susceptibility locus for inflammatory bowel disease (54
), and IFN-based pharmaceutical formulations have been used to treat patients with these disorders albeit with variable results (reviewed in reference 55
). As evident from the literature and our current results, the effects of IFN on intestinal homeostasis are pleiotropic. Roles of IFN in antigen recognition and immune function, cell differentiation and proliferation, and gut barrier function are likely to contribute to the complexity of patient responses to IFN. Although some of the long-term IFN effects, such as suppression of tissue-regenerative abilities in the gastrointestinal tract, could be genuinely detrimental (17
), it is also plausible that rapid degradation of IFNAR1 may constitute an additional challenge for the efficacy of IFN-based therapies (reviewed in reference 13
). Thus, future studies may determine the potential application of poorly bioavailable gut-restricted CK1α inhibitors to stabilize IFNAR1 and improve intestinal barrier function.