Table 1.
Integration sites within the malignant CAR T cells and their effect on expression of surrounding genes in patient 2
| Insert mapping | Genomic context and RNA sequencing | |||||
|---|---|---|---|---|---|---|
| Chromosome | Insert site* | Strand† | Transcriptional shadow,‡ kbp | Gene same strand within shadow§ | Observed Log2 FC,‖ malignancy CPM vs ProductCPM | In silico evidence of functional product# |
| chr1 | 26c145c520 | + | 18 | FAM110D | 7.01 | FAM110D |
| chr1 | 53c455c616 | + | 25 | DMRTB1 | 12.20 | |
| chr1 | 154c760c298 | + | 50 | |||
| chr1 | 184c030c856 | + | 60 | TSEN15 | 0.66 | |
| chr1 | 208c989c499 | − | 450 | |||
| chr1 | 214c453c094 | + | 40 | |||
| chr2 | 57c195c660 | + | 50 | |||
| chr2 | 62c309c188 | − | 40 | |||
| chr3 | 99c637c483 | + | 100 | COL8A1 | 9.72 | COL8A1 |
| chr3 | 111c596c609 | + | 0 | |||
| chr5 | 49c702c91 | + | 120 | LINC01020 | 8.00 | |
| chr6 | 12c023c195 | + | 120 | HIVEP1 | 2.98 | HIVEP1 |
| chr6 | 35c678c292 | + | 50 | ARMC12 | 2.99 | |
| chr6 | 90c251c539 | − | 50 | BACH2 | −0.17 | |
| chr6 | 111c857c020 | − | 100 | FYN | 3.14 | FYN |
| chr6 | 141c455c746 | + | 40 | |||
| chr7 | 7c259c055 | + | 110 | LOC101927354 | 5.93 | |
| chr7 | 83c404c188 | + | 50 | |||
| chr7 | 143c289c731 | + | 0 | |||
| chr8 | 70c401c154 | − | 100 | NCOA2 | 1.05 | |
| chr9 | 36c557c687 | + | 100 | MELK | 1.30 | |
| chr10 | 108c310c155 | − | 1000 | LINC01435*** | 11.07 | |
| chr12 | 57c836c644 | + | 100 | |||
| chr21 | 19c588c904 | + | 100 | |||
Insertion sites within the genome, size of transcriptional shadow, and impact on exonic expression of surrounding genes are shown. CAR19 insertions in the malignancy were detected by μ insertion analysis and confirmed on whole-genome sequencing by searching alignment files for paired-end split reads between the CAR19 “contig” and regions of the autosomes, excluding calls that did not indicate a full-length insertion of CAR19. Transcriptional shadow was directly visualized using the Integrative Genomics Viewer - Broad Institute. Expression of genes within the transcriptional shadow and on the same strand was evaluated based on total exon counts, and functional impact of intronic insertion sites was estimated by determining whether missing exons interrupted the ORF based on gene annotation. If open reading frame interruption was seen, in silico translation of the spliced product was performed (expasy.org) and compared with the wild-type (WT) protein sequence (www.ebi.ac.uk/Tools/msa/clustalo/). If the resultant sequence retained the WT sequence (in-frame deletion), functional domains (uniprot.org) were evaluated to estimate the likelihood of the altered product retaining activity.
chr, chromosome; FC, fold change; MalignancyCPM, counts per million in malignant T cells; ProductCPM, counts per million in CD4 CAR T cells from the product.
Insert site indicates the nucleotide position of the insert on the chromosome.
Strand (+ or −) indicates the orientation of the transgene.
Transcriptional shadow indicates the presence and length of transcriptional readthrough seen in kilobase pairs as directly visualized in IGV.
Gene same strand within shadow indicates the name of genes potentially impacted by the transcriptional readthrough seen.
Observed Log2FC is the total increase in transcription within the gene affected by the transcriptional shadow including exonic, intronic, and out-of-frame sequences compared with the CAR T-cell product expressed as log2 fold change.
In silico evidence of functional product indicates genes with increased functional protein compared with the CAR T-cell product based on splicing analysis.
LINC01435* is long noncoding RNA not associated with malignancy.