SETD8 inhibition targets cancer cells with increased rates of ribosome biogenesis

Abstract

SETD8 is a methyltransferase that is overexpressed in several cancers, which monomethylates H4K20 as well as other non-histone targets such as PCNA or p53. We here report novel SETD8 inhibitors, which were discovered while trying to identify chemicals that prevent 53BP1 foci formation, an event mediated by H4K20 methylation. Consistent with previous reports, SETD8 inhibitors induce p53 expression, although they are equally toxic for p53 proficient or deficient cells. Thermal stability proteomics revealed that the compounds had a particular impact on nucleoli, which was confirmed by fluorescent and electron microscopy. Similarly, Setd8 deletion generated nucleolar stress and impaired ribosome biogenesis, supporting that this was an on-target effect of SETD8 inhibitors. Furthermore, a genome-wide CRISPR screen identified an enrichment of nucleolar factors among those modulating the toxicity of SETD8 inhibitors. Accordingly, the toxicity of SETD8 inhibition correlated with MYC or mTOR activity, key regulators of ribosome biogenesis. Together, our study provides a new class of SETD8 inhibitors and a novel biomarker to identify tumors most likely to respond to this therapy.

Fig. 1. A virtual screen to identify compounds inhibiting 53BP1 foci.

A Pipeline of the screening strategy. A virtual screen of around 1 M compounds was performed to find chemicals mimicking the binding of a methylated H4K20 peptide bound to the TUDOR domain of 53BP1, using a published structure (PDB: 2LVM). 25 hits were next tested by HTM in U2OS cells exposed to 10 Gy and evaluated for 53BP1 foci formation 45 min post-IR. B Heatmap (red-to-black) illustrating the number of 53BP1 foci per nucleus in each well of a plate from the screen, where each compound was tested in triplicate. Note that there was only a clear hit (C23) that abrogated 53BP1 formation. C Representative image from the experiment defined in B, where 53BP1 foci (green) were analyzed by immunofluorescence. DAPI (blue) was used to stain DNA. C23 was used at 10 μM. Scale bar (white) indicates 5 μm.

Fig. 2. C23 impairs 53BP1 functions.

A Immunofluorescence of 53BP1 (green) and γH2AX (red) in U2OS cells, treated or not with C23 (10 μM), 45 min after exposure to 10 Gy of IR. Scale bar (white) indicates 5 μm. B WB of KAP, phosphorylated KAP (pKAP), CHK2 and phosphorylated CHK2 (pCHK2) in U2OS cells 45 min after exposure to 10 Gy of IR, in the presence or absence of the indicated concentrations of C23. Levels of β-ACTIN are shown as a loading control. C HTM-dependent quantification of the mean signal of γH2AX per nucleus, in U2OS cells at various times after being exposed to 10 Gy of IR, in the presence or absence of C23 (10 μM). D Flow cytometry in CH12 cells illustrating the effect of C23 (10 μM) in the efficacy of IL4-induced CSR (as measured by IgA expression). Numbers indicate the percentage of IgA+ cells.

Fig. 3. C23 phenocopies the impact of the SETD8 inhibitor UCN0379 in DNA replication.

A Immunofluorescence of an HA-tagged SETD8 (SETD8^HA, green) and PCNA (red) in U2OS cells, illustrating the localization of SETD8 to DNA replication foci. Scale bar (white) indicates 2,5 μm. B HTM-dependent quantification of the mean signal of EdU per nucleus, in U2OS cells exposed to increasing concentrations of C23 or UNC0379 for 45 min. C HTM-dependent quantification of number of 53BP1 foci per nucleus, in U2OS cells exposed to 10 Gy of IR and grown in the presence or absence of increasing concentrations of C23 or UNC0379 for 45 min. D Representative flow cytometry-based analysis of the DNA content in U2OS cells exposed for 48 h to 10 μM of UNC0379 or C23. Numbers indicate the fraction of cells in G2 and the polyploid fraction.

Fig. 4. C23 is a SETD8 inhibitor.

A WB illustrating the distribution of SETD8 in the cytoplasmic, nucleoplasmic and chromatin-bound fractions of U2OS cells treated with increasing concentrations of C23 or UNC0379. H2A and TUBULIN levels are shown as control of the fractionation. B WB of H4K20me1 and total H4 in U2OS cells treated with 5 and 15 μM of C23 or UNC0379 for 8 h. B HTM-dependent quantification of H4K20me1 levels per nucleus, in U2OS cells arrested with 1 mM HU for 24 h, and then released in the presence of increasing concentrations of C23 for 6 h. Red lines indicate median values. D Representative images from the experiment defined in C. Scale bar (white) indicates 10 μm.

Fig. 5. C23 has a preferential impact on nucleolar factors.

A Impact of increasing the temperature on SETD8 levels as assessed by TPP in U2OS cells treated with C110 (20 μM) for 25 min. Data from 2 biological replicates is shown. B Enrichr analysis from the list of factors that presented an altered thermostability in the presence of C23. The most significant classes from “REACTOME” pathways and “GO cellular component” are shown, indicating a distinct effect of C23 on nucleoli and rRNA metabolism. C Immunofluorescence images from UBF, NPM and FBL (green) in U2OS cells exposed for 45 min to C23, C110 or UNC0379 at 10 μM. DAPI (blue) was used to stain DNA. Scale bar (white) indicates 5 μm. D HTM-dependent quantification of UBF1 levels per nucleus, from the experiment defined in C. Black lines indicate median values. E Representative TEM images of U2OS cells treated with C23 or UNC0379 (25 μM) for 45 min. Control cells exhibit typical reticulated nucleoli with several small fibrillar centers (FC), surrounded by a shell of dense fibrillar component (DFC), and small irregular masses of granular component (GC). Treatment with the drug (C23 or UNC0379) induced the alteration of nucleolar integrity, including the loss of compartmentalization (nucleolar segregation), together with the formation of an enlarged fibrillar centers (FC) and accumulation of large masses of the granular component (GC). The formation of intranucleolar vacuoles (V) was also a recurrent observation in the presence of either drug. Scale bar (black) indicates 2 μM. Additional examples are provided in Supplementary Fig. S4C. ***P < 0.001; t-test.

Fig. 6. SETD8 deletion impairs nucleolar structure and function.

A Representative images of the colocalization of SETD8^HA (green) and UBF (red) in U2OS cells, which was lost upon treatment with C23. Scale bar (white) indicates 2.5 μM. DAPI (blue) was used to stain DNA. B WB illustrating the loss of SETD8 expression in Setd8^lox/− mESC 24 h after exposure to 4-OHT (1 μM). Targeted deletion was mediated by the expression of a tamoxifen inducible Cre recombinase (CreER). β-ACTIN levels are shown as a loading control. C Immunofluorescence images from UBF (green) in Setd8^+/− and Setd8^lox/− mESC 24 h after exposure to 4-OHT (1 μM). DAPI (blue) was used to stain DNA. Scale bar (white) indicates 2.5 μM. D HTM-dependent quantification of UBF nuclear levels, from the experiment defined in C. Dashed lines indicate median values. E Immunofluorescence images of EU levels (green) in Setd8^+/− and Setd8^lox/− mESC 24 h after exposure to 4-OHT (1 μM). EU was added for the last 30 min before fixation. DAPI (blue) was used to stain DNA. Scale bar (white) indicates 2.5 μM. F HTM-dependent quantification of EU levels per nucleus, from the experiment defined in E. Dashed lines indicate median values. G Representative TEM images from Setd8^+/− and Setd8^lox/− mESC 24 h after exposure to 4-OHT (1 μM). Whereas cells that preserved SETD8 expression lacked nucleolar alterations with several small fibrillar centers (FC), depletion of SETD8 induced severe nucleolar abnormalities, including prominent nucleolar segregation, appearance of both giant fibrillar centers (GFC) and intranucleolar vacuoles (V) and nucleolar fragmentation, as seen with SETD8 inhibitors. Scale bar (white) indicates 2 μm. Additional examples are provided in Supplementary Fig. S4D. n.s. non-significant; ***P < 0.001; t-test.

Fig. 7. The toxicity of SETD8 inhibitors correlates with ribosome biogenesis.

A sgRNA enrichment scores from a CRISPR resistance screen for C23 (10 μM) in KBM7 cells, 10 days after starting the treatment. Bubble plots illustrate the median enrichment of all sgRNAs targeting a given gene over DMSO, bubble sizes indicate significance. Significantly enriched sgRNAs (blue bubbles) were used for the enrichment analysis shown in B. B Gene Ontology “Biological process” pathways that are significantly enriched among the list of significantly enriched sgRNAs selected from A. Note that most pathways are related to RNA metabolism and ribosome biogenesis. C Effect of increasing doses of C23 on the viability of Ba/F3^MYCER cells, in the presence or absence of 4-OHT to promote the nuclear translocation of MYC^ER. D Effect of increasing doses of C23 on the viability of U2OS cells, previously transfected with siRNAs targeting MYC or a control siRNA. E Effect of increasing doses of C23 on the viability of Tsc2^+/+ and Tsc2^−/− MEF. Note that both cell lines were also p53 deficient, which is necessary to enable the growth of Tsc2^−/− MEF. ***P < 0.001; ANOVA.