Corrigendum to “Generation of a set of isogenic, gene-edited iPSC lines homozygous for all main APOE variants and an APOE knock-out line” [Stem Cell Res. 34/1873–5061 (2019) 101349–55] (Stem Cell Research (2019) 34, (S1873506118302794), (10.1016/j.scr.2018.11.010))

Abstract

© 2020 The Author(s) The APOE-gene edited lines previously published passed all quality controls required by journals and biorepositories at that time (Schmid et al., 2019). As part of the deeper characterization performed in our laboratories and by collaborators, these lines were subject to an extra round of quality checks including not only Sanger sequencing, but also SNP genotyping, RNAseq and qRT-PCR. Despite having correct sequence at the DNA and RNA levels, we observed mismatches with the standard SNP taqman assays used for APOE genotyping. The on-target effects of CRISPR-Cas9 gene editing have been reported elsewhere, and only recently has been described to be found in up to 40% of iPSC-gene edited lines (Weisheit et al., 2020). Our findings confirm these unwanted effects in some lines generated within ADAPTED, flagging the necessity of revising what is considered as 'standard QC' for gene edited lines across the iPSC research community by including screening for these on target effects going forward. From ADAPTED and EBiSC, we recognise the impact that these gene editing artefacts may have in the work performed in laboratories accessing the lines worldwide. It must be highlighted that this set of iPSC lines has also been used in our laboratories at our own risk, while the complex genetic abnormalities were characterized. Despite the technical genotyping issues described and these lines having only one functioning ApoE allele, these lines do express the correct APOE isoform at the mRNA and protein level and suggest that this may still be a valid tool to assess gain/loss of function of the most known variants of APOE. Nevertheless, we acknowledge that they do not match the standards of integrity required in the iPSC field and new clones for these lines are being generated and will be available from EBiSC. Based on the experiments we have performed so far we can conclude that: a) the APOE KO line (BIONi010-C-3) has the two alleles disrupted with an early stop codon that leads to absence of APOE expression; b) the APOE2/E2 (BIONi010-C-6) and APOE3/E3 (BIONi010-C-2) lines are hemizygous due to a 3.4 kb insertion from the selection plasmid in the coding part of one of the two APOE alleles; and c) the APOE4/E4 line (BIONi010-C-4) carries a 5.4 kb insertion from the selection plasmid in the coding part of one of the two APOE alleles. The remaining (non-disrupted) allele is correctly transcribed in these three lines and does drive expression of the correct APOE isoforms. The precise locations of these insertions/disruptions have been determined by a novel technology recently described (Xdrop™1).[Figure presented] We designed a pair of primers amplifying a 2,6kb region surrounding APOE exon 4 (chr19:45,411,114-45,413,774; hg19), where we suspected structural rearrangements may have occurred. PCR products were observed by electrophoresis in a 1% agarose gel (Fig. 1).[Figure presented] We found the presence of an oversized PCR product (>5kb) in BIONi010-C-4-derived cells after running PCR products in a 1% agarose gel. The band was confirmed to be stable in other clones derived from BIONi010-C-4 (eg C4K2, Fig. 2).[Figure presented] In order to understand the functional consequences of the gene editing issues reported above, we set out to evaluate the impact at the RNA level. To this end, we performed RNA sequencing in microglia differentiated from BIONi010-C-lines. We then called for APOE genotype in this dataset. As shown in the table 2, allelic disruption described above does not impact the genotype at the level of RNA: [Figure presented] After confirming the correct genotype in the RNA sequence, we next compared the expression levels of each isoform across lines. Fig. 3 shows the lack of significant differences in the levels of APOE mRNA across the BIONi010-C lines studied, except for the KO line in which is below detection:[Figure presented] From the experiments shared above, we can conclude that: • BIONi010-C-6 (APOE2/E2), BIONi010-C-2 (APOE3/E3) and BIONi010-C-4 (APOE4/E4) lines have only a single functional APOE allele due to the presence of on-target unintended insertions caused by CRISPR gene editing in one copy of the alleles.• A 5.4 kb insertion was detected in one of the alleles of the APOE4 clone.• All gene edited lines do express the correct APOE isoform (RNA = protein), and RNA expression levels are comparable across them.The authors would like to apologize for any inconvenience caused. It should be noticed that BIONi010-C replacement lines are being generated and will be made available through EBiSC and other APOE isogenic cohorts derived from different donors are already available at www.EBiSC.org.