TL;DR
This study reveals distinct patterns of how lentiviral and gamma-retroviral vectors integrate into the genome of CAR T-cells, impacting gene expression differently. Understanding these integration "hotspots" is crucial for improving CAR T-cell therapy safety and efficacy, as it can influence clinical outcomes.
What Problem Does This Paper Address?
Chimeric Antigen Receptor (CAR) T-cell therapy has emerged as a powerful tool against hematological cancers. A key step in creating these therapeutic cells involves genetically modifying a patient's T-cells using viral vectors, most commonly lentiviruses and gamma-retroviruses, to express the CAR. While these vectors efficiently integrate into the host cell's DNA, the precise locations of these integrations (integration sites or IS) and their potential impact on CAR T-cell function and patient outcomes have remained unclear. Previous studies have shown that some integrations can lead to unintended consequences, like uncontrolled cell growth. This paper aims to comprehensively map these integration sites across the entire genome in pre-infusion CAR T-cell products and investigate how these patterns differ between lentiviral and gamma-retroviral vectors, and whether they correlate with clinical responses.
Methods
The researchers employed a modified Vector Integration Sites Analysis (VISA) pipeline to identify viral integration sites. This pipeline involves three main phases:
- Amplification of hDNA-vDNA junctions: Genomic DNA from CAR T-cell products is fragmented, and then specific PCR primers are used to amplify the junctions where the viral DNA has integrated into the human DNA.
- Library preparation and sequencing: The amplified DNA fragments are prepared into sequencing libraries and then sequenced using a high-throughput platform (Illumina Miseq).
- Bioinformatic analysis: Raw sequencing reads are processed to identify and annotate the precise locations of viral integration sites across the genome.
In addition to integration site analysis, the study also utilized RNA-sequencing to assess gene expression changes and performed gene ontology (GO) analysis to understand the functional implications of these integration events.
Datasets
The study analyzed a total of 75 CAR T-cell and 6 TCR T-cell products. These included:
- 57 products manufactured with lentiviral vectors (targeting CD22, CD19/CD22, CD30, and FGFR4).
- 24 products manufactured with gamma-retroviral vectors (targeting BCMA, SLAMF7, CD19, and E7-TCR).
These products were sampled before infusion into patients. Paired non-transduced cells were used as controls for gene expression analysis.
Key Insights and Ideas
- Vector Integration is Not Entirely Random: While integration is often described as semi-random, the study identified specific "hotspot" genes and genomic regions where integration events are preferentially enriched, suggesting non-randomness influenced by viral vector and host factors.
- Lentiviral vs. Gamma-Retroviral Differences: A significant finding is the distinct integration preferences between the two vector types. Gamma-retroviruses favor promoters, untranslated regions (UTRs), and exons, while lentiviruses lean towards introns and intergenic regions.
- Impact on Gene Expression Varies: Gamma-retroviral integrations appear to have a stronger impact on the host transcriptome, potentially by affecting gene expression more directly through promoter/UTR insertions. Lentiviral integrations, often in introns, might influence gene expression more through post-transcriptional modifications like alternative splicing.
- Clinical Outcome Association: Integration patterns in specific genes were associated with different clinical outcomes (response vs. non-response, cytokine release syndrome, HLH). This suggests that integration site analysis could potentially serve as a predictive marker.
- Functional Impact Beyond mRNA: For genes with differential integration and clinical outcome association, the impact might be through disrupting amino acid sequences and generating abnormal proteins, rather than solely altering mRNA levels.
Key Results
- Chromosome 19 Enrichment: Both vector types showed the greatest enrichment of integration sites on Chromosome 19, a gene-dense region. Lentiviral vectors showed a higher enrichment on this chromosome compared to gamma-retroviral vectors.
- Genomic Feature Preferences: Gamma-retroviral vectors were more likely to integrate into promoter (14% vs. 4%), UTR (7% vs. 4%), and exon (21% vs. 13%) regions compared to lentiviral vectors. Lentiviral vectors were more likely to integrate into introns (33% vs. 23%) and intergenic regions (46% vs. 35%).
- Hotspot Genes Identified: Specific genes like NRBF2P1 (promoter) and TNNI1 (UTR) were identified as integration hotspots for gamma-retroviral and lentiviral vectors, respectively.
- Differential Gene Expression: Gamma-retroviral vector products showed a higher number of differentially expressed genes (DEGs) compared to lentiviral vector products.
- Clinical Outcome Correlations: The study identified sets of genes with differential integration enrichment associated with clinical response, high-grade CRS, and HLH. For example, genes involved in neutrophil activation were enriched in non-responders.
Limitations and Future Work
- Limited Real-Time Data: The study analyzed pre-infusion CAR T-cell products. Tracking the CAR T-cells post-infusion to assess clonal expansion and long-term integration impacts was not possible due to the inability to obtain patient samples after infusion.
- Indirect Effects: While some integration events directly affected gene expression, many DEGs without direct integration in their loci suggest indirect regulatory mechanisms are at play.
- Future Directions: The authors suggest that future research should focus on studying the long-term impacts of viral integration events and exploring the possibility of studying post-infusion samples. Further investigation into how integration events affect amino acid sequences and protein function, rather than just mRNA expression, is also warranted. The findings establish a foundation for further analyses into the clinical relevance of these integration patterns.
Summary updated