YTHDF3 Induces the Translation of m6A-Enriched Gene Transcripts to Promote Breast Cancer Brain Metastasis

Abstract

Brain metastasis is a major cause of cancer mortality, but its molecular mechanisms are severely understudied. In addition, little is known regarding the role of m⁶A reader YTHDF3 in human diseases. Here, we show that YTHDF3 overexpression clinically correlates with brain metastases in breast cancer patients. YTHDF3 promotes cancer cell interactions with brain endothelial cells and astrocytes, blood-brain barrier extravasation, angiogenesis, and outgrow. Mechanistically, YTHDF3 enhances the translation of m⁶A-enriched transcripts for ST6GALNAC5, GJA1, and EGFR, all associated with brain metastasis. Furthermore, overexpression of YTHDF3 in brain metastases is attributed to increased gene copy number and the autoregulation of YTHDF3 cap-independent translation by binding to m⁶A residues within its own 5' UTR. Our work uncovers an essential role of YTHDF3 in controlling the interaction between cancer cells and brain microenvironment, thereby inducing brain metastatic competence.

Keywords: YTHDF3; brain metastasis; epigenetic regulation; gene amplification; m(6)A RNA methylation.

Figure 1.. YTHDF3 Expression Is Directly Correlated with Decreased Brain Metastasis-Free Survival in Breast Cancer Patients. YTHDF3 Depletion Impairs Brain Metastasis Formation

(A) Heat-map showing RNA differential expression of 47 common genes between clinical breast cancer brain metastases and primary tumors (left), and between MDA-MB-231 and MDA-MB-231BrM cells (right). (B) Kaplan-Meier analysis for brain metastasis-free survival of 454 breast cancer patients with primary tumors expressing high or low levels of YTHDF3 mRNA. p = 0.0447 by log-rank test. (C) Representative IHC staining (3 samples; Insets: high-magnification images) and histograms of YTHDF3 protein levels in primary breast tumors (n=50) and unmatched brain metastases (mets) (n=59). Lines show mean and SD, t-test, ***p < 0.001. (D) Histograms of YTHDF3 protein levels in primary breast tumors and matched brain metastases from 26 patients. Lines show mean and SD, t-test, ***p<0.001. (E) Western blotting for YTHDF3 in 4T1, MDA-MB-231, and MDA-IBC3 cells. P, parental; BR, brain metastasis. (F) Western blotting for YTHDF3 in 4T1Br, MDA-MB-231Br, and MDA-IBC3 cells stably expressing control or YTHDF3 shRNAs. (G) Brain metastases counts 30 days after intracardiac injection of the cells in (F). Large metastases are >300 mm on the longest axis. No micrometastasis detected in mice injected with MDA-IBC3 cells. In the boxplot, the middle line represents the median, the bottom and top lines correspond to the 25th and the 75th percentiles, and the whiskers represent the maximum and minimum numbers. N=10 mice, two-sided Mann-Whitney test, **p<0.01, ***p<0.001. (H) Bioluminescent imaging (BLI) at 8, 16 and 26 days after intracardiac injection of YTHDF3 knockdown and control MDA-MB-231Br cells. The dots represent the photon flux of the BLI signal within a region that corresponds to the brain for each mouse. The line represents mean, n=8 mice; Two-sided Mann-Whitney test, ***p<0.001. (I) Kaplan-Meier survival analyses for mice injected with shcontrol or shYTHDF3 4T1Br, 231Br, and MDA-IBC3 cells (mean ± SD, n=10; log-rank test, **p<0.01). See also Figure S1.

Figure 2.. YTHDF3 Promotes Brain Endothelial Adhesion, Extravasation, Cancer Cell-Astrocyte Interaction, and Angiogenesis of Breast Cancer Cells

(A) BLI at 0–72 hours after intracardiac injection of YTHDF3 KD and control MDA-MB-231Br cells. Data are mean ± SD, n=8; Two-sided Mann-Whitney test, **p < 0.01***p<0.001. (B) Tumor cell seeding in brain. CMFDA (green) labeled cells were intracardially injected into mice for 48 hrs. Data are mean ± SEM, n=4 brains and 10 random fields of each brain; t-test, *p < 0.05, ***p<0.001. (C) Quantification of percentage reduction of tumor cells adhered to human brain microvascular endothelial (HBMEC) cells after knockdown of YTHDF3. (D) Extravasation of tumor cells into brain parenchyma. Brains were collected 48 hours post intracardiac injection of the indicated cells labeled with CMFDA (green). Brain slices were IF stained with CD31 (red). The percentage of extravasated tumor cells over total tumor cells was calculated. Data are mean ± SEM, n=4 brains and 10 random fields of each brain; t- test, *p<0.05, **p<0.01. (E) Quantitation of fold decrease of transwell invasion after knockdown of YTHDF3 in MDA-MB-231Br cells. (F) In vitro BBB transmigration activity of the MDA-MB-231Br or 4T1Br control and YTHDF3 KD cells was analyzed using Trans-BBB assay as showing in Figure S2I. (G) Percentage reduction of tumor cells adhered to astrocyte monolayer after knockdown of YTHDF3. (H) Cancer cell-astrocyte interaction in brain metastatic lesion formed by 4T1Br and MDA-MB-231Br shcontrol and shYTHDF3 cells was analyzed by IF staining. Tumor cells, astrocyte cells and nuclei are stained with CK18 (red), GFAP (green) or DAPI (blue), respectively. Quantitation of tumor-associated GFAP positive astrocytes. Data are mean ± SEM, n=5 mice and 10 fields per mouse; t-test, **p<0.01, ***p<0.001. (I and J) Representative image (I) and quantitation of blood vessel densities in brain metastases formed by 4T1Br and MDA-MB-231Br shcontrol and shYTHDF3 cells (J). Blood vessel density were analyzed from 10 random 0.25-μm² tumor fields per mouse. Data are mean ± SEM, n=5 mice; t-test, **p<0.01. (K) In vitro angiogenesis assay using HBMEC cells treated with conditioned media from 4T1Br and MDA-MB-231Br shcontrol and shYTHDF3 cells. Data in C, E, F, G and K are mean ± SEM of 3 biological replicates; t-test, *p<0.05, **p<0.01, ***p<0.001. See also Figure S2.

Figure 3.. YTHDF3 Binds to mRNA of Key Brain Metastasis Genes and Enhances Their Translational Efficiency

(A) Supervised hierarchical clustering analysis of RPPA results. The results indicated 34 proteins with p <0.05 were differentially expressed between control and YTHDF3 KD 231Br and 4T1Br cells. (B) Scatter plot of log2 fold enrichment (IP/total) of YTHDF3 RIP-seq versus transcript abundance. Red points indicate transcripts with YTHDF3 enriched binding (log₂FC>1 and p<0.05). There are 3054 transcripts from 36239 transcripts having YTHDF3 enriched binding. Arrow indicates some of YTHDF3 targets identified by both RIP-seq and RPPA. (C) Plots showing YTHDF3 binding to individual mRNAs of brain metastasis genes in MDA-MB-231Br cells, as measured by RIP-seq. The normalized reads distribution for input (blue) and YTHDF3 (red) along mRNAs. (D) Schematic workflow of YTHDF3 downstream targets analysis. (E) CLIP-qPCR showing the indicated mRNA transcripts bound with YTHDF3 in 4T1Br cells. (F) Western blot analysis of protein levels of the indicated YTHDF3 target genes in 4T1Br and MDA-MB-231Br control and shYTHDF3 cells. (G) CLIP-qPCR showing the association of eIF3a with the indicated transcripts in 4T1Br cells with YTHDF3 depletion or overexpression. (H) Polysome profiling of 4T1Br control and shYTHDF3 cells. Fractions of the cell lysates are shown. (I) Relative mRNA distribution of the target genes in each ribosome fractions from (H) was analyzed by qPCR and normalized to RCN2 mRNA. Data are from three independent polysome-profiling experiments. Data in E and G are mean ± SEM of 3 biological replicates; t-test, *p<0.05, **p<0.01, ***p<0.001. See also Figure S3.

Figure 4.. ST6GALNAC5, GJA1 and EGFR Are Functionally Essential Targets of YTHDF3 in Breast Brain Metastasis

(A) Western blotting for YTHDF3 in the indicated cell lines. (B) Time course of invasion for MDA-MB-231 control and YTHDF3 OE cells. Representative images are shown. (C) In vitro angiogenesis assay using HBMEC cells treated with conditioned media from 4T1 and MDA-MB-231 control or YTHDF3 OE cells. (D) BLI at 8, 16 and 26 days after intracardiac injection of YTHDF3 OE and control MDA-MB-231 cells. The dots represent the photon flux in brain for each mouse at day 26. The line represents mean, n=8 mice; Two-sided Mann-Whitney test, ***p<0.001. (E) BLI at 8 weeks or when mice were moribund after mammary fat pad injection of YTHDF3 OE and control 4T1 cells. The line represents mean, n=8 mice; Two-sided Mann-Whitney test, ***p<0.001. (F) Brain metastasis counts in histopathology evaluation of the brains from the experiments in (E). In the boxplot, the middle line represents the median, the bottom and top lines correspond to the 25th and the 75th percentiles, and the whiskers represent the maximum and minimum number; N=8 mice; Two-sided Mann-Whitney test, *p<0.05, **p<0.01. (G) MDA-MB-231 YTHFD3 OE cells were transfected with two independent shRNAs for GJA1, ST6GALNAC5, or EGFR. Then the knocking down efficiency for these shRNAs was analyzed by western blotting. (H) In vitro trans-BBB transmigration activity of MDA-MB-231 control and YTHDF3 OE cells with knockdown of ST6GALNAC5 or GJA1 alone, or both together. Data are mean ± SEM, 3 biological replicates, t-test, ***p<0,001 as compared with YTHDF3 OE + shcontrol cells. (I) Brain extravasation of MDA-MB-231Br cells upon knocking down ST6GALNAC5 or GJA1 at 48 hours post injection. The percentages of extravasated tumor cells over total tumor cells were calculated. (J) Brain extravasation of MDA-MB-231 control and YTHDF3 OE cells upon knocking down ST6GALNAC5 or GJA1 at 48 hours post injection. (K) Brain extravasation of 4T1 control and YTHDF3 OE cells with knockdown of ST6GALNAC5 or GJA1 at 48 hours post injection. Data in I, J and K are mean ± SEM, n=4 mice and 10 fields per mouse; t-test, *p<0.05, **p<0.01. See also Figure S4.

Figure 5.. Regulation of the Expression of Key Brain Metastatic Genes, Thereby Brain Metastasis by YTHDF3 Is Dependent on the Binding of YTHDF3 to m⁶A-Modified mRNA

(A) Top consensus motif identified by HOMER with m⁶A-seq peaks in MDA-MB-231Br cells. (B) m⁶A-seq of MDA-MB-231Br cells shows m⁶A peaks at the individual mRNAs of YTHDF3 target genes. The y-axis shows normalized reads distribution for input (blue) and m⁶A IP (red) along the transcripts. (C) MDA-MB-231 and 4T1 cells transfected with Flag-YTHDF3-WT or Flag-YTHDF3-Mut plasmids were analyzed for the expression of Flag-YTHDF3-WT or Flag-YTHDF3-Mut protein by western blotting. (D) Flag-immunoprecipitated ribonucleoprotein complexes from MDA-MB-231 YTHDF3 OE-WT or OE-Mut cells were analyzed for the m⁶A contents of total RNAs. (E) Binding of GJA1, ST6GALNAC5, or EGFR mRNA to YTHDF3 wild type or mutant in MDA-MB-231 YTHDF3 OE-WT or OE-Mut cells was analyzed by CLIP assays. Negative control PCDH7 that does not contain m⁶A exhibited no binding. (F) Redistribution of the above YTHDF3 targets in non-ribosome and polysome portions of mRNPs in MDA-MB-231 YTHDF3 OE-WT or OE-Mut cells. (G) Protein expressions of the above YTHDF3 targets in MDA-MB-231 and 4T1 YTHDF3 OE-WT or OE-Mut cells was analyzed by western blotting. (H) Quantitation of invasion of the cells in (G). (I) Brain metastasis counts 30 days after intracardiac injection of the cells in (G). In the boxplot, the middle line represents the median, the bottom and top lines correspond to the 25th and the 75th percentiles, and the whiskers represent the maximum and minimum numbers.. N=10 mice; Two-sided Mann-Whitney test, *p < 0.05, **p<0.01 as compared to YTHDF3 OE-WT cells. Data in D, E, F and H are mean ± SEM of 3 biological replicates; t-test, *p<0.05, **p<0.01. See also Figure S5.

Figure 6.. Brain Metastases of Breast Cancer Have Higher Rate of Copy Number Gain in YTHDF3 Gene and YTHDF3 Promotes Translation Efficiency of Its mRNA

(A) YTHDF3 gene copy numbers in the cancer tissues were determined using TaqMan Gene Copy Number Assay. Chi-square test, ***p<0.001. (B) Plot showing YTHDF3 binding to its own mRNA in MDA-MD-231Br cells, as measured by RIP-seq. The normalized reads distribution for input (blue) and YTHDF3 (red) along mRNA. (C) CLIP-qPCR showing the association of YTHDF3 protein with endogenous YTHDF3 transcripts and in MDA-MB-231 cells. (D) Western blot analysis showing increase of YTHDF3 expression in MDA-MB-231 cells with ectopic expression of Flag-YTHDF3. (E) qPCR analysis of total YTHDF3 mRNA using CDs-primers or endogenous YTHDF3 mRNA using 3’UTR-primers in MDA-MB-231 cells transfected with control or Flag-YTHDF3. (F) Western blot analysis showing increase of endogenous YTHDF3 expression in MDA-MB-231 lls with ectopic expression of Dsred-YTHDF3. (G) Redistribution of YTHDF3 in non-ribosome and polysome portions of mRNPs in MDA-MB-231 lls expressing control or Flag-YTHDF3. (H) m⁶A-seq of MDA-MB-231Br cells shows m⁶A peaks on YTHDF3 mRNA transcripts. The y-axis shows normalized reads distribution for input (blue) and m⁶A IP (red) along the transcripts. (I) CLIP-qPCR analysis was used to identify YTHDF3 mRNA segments associated with Flag-YTHDF3. (J) CLIP-qPCR showing the association of YTHDF3 and eIF3a with YTHDF3 mRNA transcripts in MDA-MB-231 cells expressing Flag-YTHDF3. (K) Relative fluorescence intensity of the YTHDF3 5’ UTR EGFP reporter in MDA-MB-231 cells transfected with wild-type and mutant YTHDF3 plasmids. (L) YTHDF3 expression in MDA-MB-231 cells transfected with wild-type and mutant YTHDF3 plasmids. Data in C, E, G, I, J and K are mean ± SEM of 3 biological replicates; t-test, **p<0.01, ***p<0.001. See also Figure S6.