CARD10 cleavage by MALT1 restricts lung carcinoma growth in vivo

CARD-CC complexes involving BCL10 and MALT1 are major cellular signaling hubs. They govern NF-κB activation through their scaffolding properties as well as MALT1 paracaspase function, which cleaves substrates involved in NF-κB regulation. In human lymphocytes, gain-of-function defects in this pathway lead to lymphoproliferative disorders. CARD10, the prototypical CARD-CC protein in non-hematopoietic cells, is overexpressed in several cancers and has been associated with poor prognosis. However, regulation of CARD10 remains poorly understood. Here, we identified CARD10 as the first MALT1 substrate in non-hematopoietic cells and showed that CARD10 cleavage by MALT1 at R587 dampens its capacity to activate NF-κB. Preventing CARD10 cleavage in the lung tumor A549 cell line increased basal levels of IL-6 and extracellular matrix components in vitro, and led to increased tumor growth in a mouse xenograft model, suggesting that CARD10 cleavage by MALT1 might be a built-in mechanism controlling tumorigenicity.


CRISPR
CRISPR knock-in was performed on the parental A549 cell line from ATCC at passage 4 after purchase.
Cells were treated for 5 hours with 5 μM of the DNA-PK inhibitor NU-7441 (Novartis) to reduce the frequency of NHEJ and increase the efficiency of homology-directed repair (HDR) 1 . Three millions cells were then nucleotransfected with the Neon system (Thermofisher) according to manufacturer instruction with the following mix in 100 μl total volume : (i) 2 µM of CARD10 gRNA Monitor clone growth and image plates every 2-3 days to identify single clones. For each confluent well, PCR was performed on CARD10 exon 11 (F: GGCCTCAAACCTGCCAAGG and R: CCATGCAAAAAGGGTCATCATCTCC) and digestion with AvaI (present in WT sequence) and NaeI (modified in edited sequence). NaeI digested/AvaI non-digested clones were then confirmed by sequencing of exon 11 of CARD10. normalized expression values using version 3.6.1 of the R scripting language. Differential expression analysis was done with the limma R package 2 (version 3.42.2) using the formula "~ 0 + genotype + clone + (1|replicate)" where genotype refers to WT or KI, clone to one of three WT clones or two KI clones, and replicate to the sample triplicates. Multiple testing correction was done using the Benjamini-Hochberg false discovery rate.

Microarray
For gene set enrichment analysis of the microarray data, the R package fgsea was used 3 . In cases where a gene was represented by more than one probe set, the probe set with the best t score was used. As gene sets, the MSigDB version 7.0.1 collections hallmark, C2-CP, C2-BIOCARTA, C2-KEGG, C2-PID, C2-REACTOME and the MetaBase version 19.4.69900 collection "maps" were used 4-6 (see http://software.broadinstitute.org/gsea/msigdb/collections.jsp for a description of the MSigDB collections, and https://www.gsea-msigdb.org/gsea/msigdb/collection_details.jsp for their references; MetaBase is a commercial product of Clarivate). The analysis was performed over 100,000 permutations with gene sets having at least 10 and not more than 1,000 genes. Differential protein expression analysis was also done with the limma R package 2 (version 3.42.2), using the formula "~ 0 + genotype_stimulation + clone" where genotype_stimulation refers to the combination of genotype (WT or KI) and stimulation (baseline or 24 hours of stimulation with P/I), and clone refers to one of two WT or KI clones. Multiple testing correction was done using the Benjamini-Hochberg false discovery rate with a cutoff of < 0.05.

SomaScan
Gene set enrichment analysis of the SomaScan data was done with the R package fgsea 3 (version 1.12.0) and using the SOMAmer R protein with the best t score when a gene was represented by more than one SOMAmer. The same MSigDB and MetaBase gene set collections described above were used, with the addition of MSigDB C5-BP. These were filtered for ECM-related gene sets by using the keywords "cell adhesion", "ECM", "extracellular matrix", "matrisome" and "integrin", resulting in 124 gene sets.
Furthermore, to adjust for the smaller universe of proteins covered by the SomaScan platform, the selected gene sets were filtered to contain only genes encoding for proteins covered in the SomaScan platform. The analysis was performed over 100,000 permutations with gene sets having at least 10 and not more than 1,000 genes.