Intestinal transit-amplifying cells require METTL3 for growth factor signaling and cell survival

Abstract

Intestinal epithelial transit-amplifying cells are essential stem progenitors required for intestinal homeostasis, but their rapid proliferation renders them vulnerable to DNA damage from radiation and chemotherapy. Despite these cells' critical roles in intestinal homeostasis and disease, few studies have described genes that are essential to transit-amplifying cell function. We report that RNA methyltransferase-like 3 (METTL3) is required for survival of transit-amplifying cells in the murine small intestine. Transit-amplifying cell death after METTL3 deletion was associated with crypt and villus atrophy, loss of absorptive enterocytes, and uniform wasting and death in METTL3-depleted mice. Sequencing of polysome-bound and methylated RNAs in enteroids and in vivo demonstrated decreased translation of hundreds of methylated transcripts after METTL3 deletion, particularly transcripts involved in growth factor signal transduction such as Kras. Further investigation verified a relationship between METTL3 and Kras methylation and protein levels in vivo. Our study identifies METTL3 as an essential factor supporting the homeostasis of small intestinal tissue via direct maintenance of transit-amplifying cell survival. We highlight the crucial role of RNA modifications in regulating growth factor signaling in the intestine with important implications for both homeostatic tissue renewal and epithelial regeneration.

Keywords: Adult stem cells; Gastroenterology; Homeostasis; RNA processing.

Figure 1. Constitutive METTL3 deletion causes growth retardation and small intestinal epithelial distortion.

(A) Growth curves from postnatal day 15 to 29. (B) Gross appearance at postnatal day 29. (C) Composite appearance and behavior score at postnatal day 29. (D) Kaplan-Meier survival curves through postnatal day 29; P value corresponds to log-rank (Mantel-Cox) test. (E) Representative small intestine and colon H&E images. Regenerative and atrophic crypts are highlighted. “L” denotes lymphocytic infiltrate. (F and G) Composite histological score for small intestine (SI) and colon. (H) Representative images of Ki67 in distal small intestine and number of hypoproliferative crypts (< 10 Ki67⁺ cells) per 1 mm distal half small intestine. Hypo- and hyperproliferative crypts are highlighted. (I) Representative images and quantification of TUNEL staining. (J–L) Representative images and quantification of intestinal secretory markers MUC2, LYZ, and CHGA. (M) Representative images and quantification of percentage alkaline phosphatase–positive (ALPI) villus length. Each plotted point corresponds to 1 mouse and depicts the mean of 3 representative sections imaged per mouse with bar at median value. Unless otherwise noted, P value represents unpaired parametric Student’s t test. Immunofluorescence staining and quantification performed in distal half small intestine. All scale bars 100 μm. ECAD, epithelial cadherin; MUC2, mucin 2; LYZ, lysozyme; CHGA, chromogranin A.

Figure 2. Inducible METTL3 deletion causes mortality and small intestinal epithelial disruption in adult mice.

(A) Experimental schematic depicting sacrifice 9 days after final tamoxifen injection. (B) Weight curves through 9 days after tamoxifen injection, mean ± SD. (C) Kaplan-Meier survival curves through 22 days after final tamoxifen injection; P value corresponds to log-rank (Mantel-Cox) test. (D) Representative small intestine and colon H&E images. Regenerative and atrophic crypts are highlighted. “L” indicates lymphocytic infiltrate. (E and F) Composite histological score for small intestine (SI) and colon. (G) Representative images of Ki67 and quantification of hypoproliferative (< 10 Ki67⁺ cells) crypts per 1 mm intestine. Hypo- and hyperproliferative crypts are highlighted. (H) Representative images and quantification of TUNEL staining. (I–K) Representative images and quantification of intestinal secretory markers MUC2, LYZ, and CHGA. (L) Representative images and quantification of percentage alkaline phosphatase–positive villus length. Unless otherwise noted, each data point corresponds to 1 mouse and depicts the mean of 3 representative sections imaged per mouse with bar at median value. Unless otherwise noted, P value represents unpaired parametric Student’s t test. Immunofluorescence staining and quantification performed in distal half small intestine. All scale bars 100 μm.

Figure 3. Intestinal distortion and mortality after METTL3 deletion are microbiota independent.

(A) Log₂ fold-change in serum cytokines in mice 9 days after final tamoxifen injection; ns indicates P > 0.05 for the KO versus control comparison for each individual cytokine. Statistical outliers were removed in GraphPad Prism v9.3 using ROUT method (Q = 1%). (B) Spleen weight 9 days after final tamoxifen injection. (C and D) Small intestine and colon length 9 days after final tamoxifen injection. (E) Experimental schematic depicting timing of antibiotic (abx) treatment, tamoxifen injection, and sacrifice. (F) Log₂ fold-change in 16S rRNA amplified from fecal bacterial DNA on final day of tamoxifen injection in antibiotic-treated or water vehicle mice. (G) Kaplan-Meier survival curves through 15 days after final tamoxifen injection in antibiotic-treated mice; P value corresponds to log-rank (Mantel-Cox) test. (H) Weight change after final tamoxifen injection in antibiotic-treated mice, presented as mean ± SD. P value represents unpaired parametric Student’s t test for values at 8 days after final tamoxifen injection. (I) Representative H&E images from matched sections of distal small intestine in n = 3 antibiotic-treated mice per genotype. Unless otherwise noted, each data point represents a single mouse with bar at median value, and P denotes value of unpaired parametric Student’s t test. Scale bar 100 μm.

Figure 4. METTL3 deletion rapidly induces death of transit-amplifying cells.

(A) Experimental schematic depicting early sacrifice 2 days after tamoxifen injection. (B) Representative images and quantification of crypt height. (C) Representative images and quantification of Ki67 staining. (D) Log₂ fold-change in qPCR quantification of indicated genes in Mettl3^{VilCreERΔ/Δ} distal small intestinal crypts relative to the mean of Mettl3^fl/fl controls and normalized to Actb. Data presented as mean ± SD. (E) Representative images and quantification of TUNEL staining. (F) Representative image and quantification of distribution of TUNEL staining in villus, transit-amplifying (TA) zone, and crypt base in Mettl3^{VilCreERΔ/Δ} mice. Data presented as mean ± SD. Each plotted data point corresponds to 1 mouse. For IF, each data point depicts the mean of 3 representative sections imaged per mouse with bar at median value. P value represents unpaired parametric Student’s t test. All data from distal small intestine of mice 2 days after final tamoxifen injection. All scale bars 100 μm.

Figure 5. METTL3 deletion triggers growth arrest and death in intestinal epithelial enteroids and colonoids.

(A) Intestinal epithelial enteroids or colonoids were treated with 1 μm 4-OHT 48 and 24 hours before beginning of time course and analyzed at days 1 through 5. (B) Percentage live enteroids or colonoids from indicated tissue regions at 5 days after 4-OHT treatment. Each point represents n = 9 technical replicates across n = 3 passage separated biological replicates per genotype. Data presented as median ± SD. P value represents unpaired parametric Student’s t test. (C) Western blot for METTL3 in surviving distal colonoids 6 days after 4-OHT treatment. (D) Representative images at 5 days after 4-OHT treatment corresponding to quantification in B. Scale bar 500 μm. (E and F) ImageJ (NIH) quantification of average enteroid 2D area and percentage live enteroids in each of the 5 days after 4-OHT treatment of Villin-CreERT2 (VilCreERT2) and Mettl3^{VilCreERΔ/Δ} ileal enteroids. Each point represents n = 9 technical replicates across n = 3 passage-separated biological replicates per genotype. Data presented as mean ± SD. P value represents unpaired parametric Student’s t test at day 3 and day 5. (G) Representative images of ileal enteroids in the 5 days after 4-OHT treatment. Scale bar 200 μm.

Figure 6. METTL3 deletion leads to a global decrease in mRNA TE with impacts on growth factor signaling.

(A) Volcano plot of all transcripts with log₂ fold-change in TE > 0.5 or < –0.5 and –log₁₀ P > 1. Red marks all transcripts with increased TE, and blue marks all transcripts with decreased TE. (B) Volcano plot of all transcripts displayed in A, now filtered for transcripts containing at least 1 m⁶A peak. (C) Pathway enrichment analysis comparing transcripts with downregulated TE (log₂FC < –1) and at least 1 m⁶A peak against Gene Ontology Biological Process (GOBP) gene sets. Circle color and size both scale with number of genes overlapping between the tested gene set and the GOBP gene set. (D) Heatmap depicting z scores for TE. Genes presented are all 42 unique genes from the 4 most significantly enriched pathways in C. Genes are presented in order of greatest decrease in mean TE to smallest decrease. All data from RNA-Seq and polysome-Seq in n = 3 Mettl3^fl/fl (CTRL) and n = 3 Mettl3^{VilCreERΔ/Δ} (KO) ileal enteroids 72 hours after initiation of 4-OHT treatment.

Figure 7. METTL3 deletion downregulates KRAS and induces cellular senescence.

(A) Integrated Genomics Viewer depiction of read density for m⁶A-RIP (red) and input RNA (blue) for the Kras transcript as determined by m⁶A-seq in distal small intestinal crypts of n = 3 wild-type mice. CDS, coding sequence; RIP, RNA immunoprecipitation. (B–E) m⁶A enrichment determined by m⁶A-RIP-qPCR with primers targeting Kras 5′ UTR, CDS exons 1–2, CDS exons 3–4A, and 3′ UTR in crypt-enriched lysates from Mettl3^fl/fl and Mettl3^{VilCreERΔ/Δ} mice 3 days posttamoxifen. Data presented as mean ± SD. Dotted line at m⁶A enrichment = 1. (F) qPCR for Kras transcript in crypt-enriched lysates from Mettl3^fl/fl and Mettl3^{VilCreERΔ/Δ} mice 3 days posttamoxifen. Data normalized to Actb and the mean of Mettl3^fl/fl controls. Data presented as mean ± SD. (G) Western blot for top targets with downregulated TE in crypts of Mettl3^fl/fl and Mettl3^{VilCreERΔ/Δ} mice 2 days after final tamoxifen injection (n = 2 mice per genotype). (H) Representative images and quantification of p-ERK staining in atrophic small intestinal crypts in Mettl3^{VilCreERΔ/Δ} mice and region-matched Mettl3^fl/fl controls 9 days after final tamoxifen injection. “p-ERK low” crypts contain < 5 p-ERK⁺ cells. (I–K) Representative images and quantification of p21, γH2AX, and β-galactosidase staining in distal half small intestine of Mettl3^fl/fl and Mettl3^{VilCreERΔ/Δ} mice 2 days after final tamoxifen injection. (L) β-Galactosidase staining in control VilCreERT2 and Mettl3^{VilCreERΔ/Δ} enteroids 3 days after 4-OHT. For all plots, each data point represents a single mouse, and P denotes value of unpaired parametric Student’s t test. Unless otherwise noted, immunofluorescence data are from areas of most severe histological distortion in distal small intestine of mice 2 days after final tamoxifen injection. For immunofluorescence, each data point is the mean of 3 representative sections imaged per mouse with bars at median value. Scale bar 100 μm.

Figure 8. METTL3 maintains growth factor signaling and survival in intestinal transit-amplifying cells.

Proposed model. In wild-type mice, METTL3 methylates Kras and other transcripts involved in transducing growth factor signaling. Methylation promotes translation of these transcripts, enhancing proliferation and survival in transit-amplifying (TA) cells downstream of external growth factors. In the absence of METTL3, a decreased response to extracellular growth factors in METTL3-knockout TA cells leads to cellular senescence and death. Loss of transit amplification results in reduced crypt and villus size and diminished production of absorptive cells. Green protein represents KRAS and other proteins regulating the intracellular response to growth factors. ECM, extracellular matrix; RTK, receptor tyrosine kinase.