Acute myeloid leukemia: from NGS, through scRNA-seq, to CAR-T. dissect cancer heterogeneity and tailor the treatment

Abstract

Acute myeloid leukemia (AML) is a malignant blood cancer with marked cellular heterogeneity due to altered maturation and differentiation of myeloid blasts, the possible causes of which are transcriptional or epigenetic alterations, impaired apoptosis, and excessive cell proliferation. This neoplasm has a high rate of resistance to anticancer therapies and thus a high risk of relapse and mortality because of both the biological diversity of the patient and intratumoral heterogeneity due to the acquisition of new somatic changes. For more than 40 years, the old gold standard "one size fits all" treatment approach included intensive chemotherapy treatment with anthracyclines and cytarabine.The manuscript first traces the evolution of the understanding of the pathology from the 1970s to the present. The enormous strides made in its categorization prove to be crucial for risk stratification, enabling an increasingly personalized diagnosis and treatment approach.Subsequently, we highlight how, over the past 15 years, technological advances enabling single cell RNA sequencing and T-cell modification based on the genomic tools are affecting the classification and treatment of AML. At the dawn of the new millennium, the advent of high-throughput next-generation sequencing technologies has enabled the profiling of patients evidencing different facets of the same disease, stratifying risk, and identifying new possible therapeutic targets that have subsequently been validated. Currently, the possibility of investigating tumor heterogeneity at the single cell level, profiling the tumor at the time of diagnosis or after treatments exist. This would allow the identification of underrepresented cellular subclones or clones resistant to therapeutic approaches and thus responsible for post-treatment relapse that would otherwise be difficult to detect with bulk investigations on the tumor biopsy. Single-cell investigation will then allow even greater personalization of therapy to the genetic and transcriptional profile of the tumor, saving valuable time and dangerous side effects. The era of personalized medicine will take a huge step forward through the disclosure of each individual piece of the complex puzzle that is cancer pathology, to implement a "tailored" therapeutic approach based also on engineered CAR-T cells.

Keywords: Acute myeloid leukaemia (AML); Intratumoral heterogeneity; Next generation sequencing (NGS); Personalized medicine; Single cell sequencing (scRNA-seq); Therapeutic targets.

Fig. 1

Leukemogenesis is a multistep process in which normal hematopoiesis is impaired by the accumulation of preliminary mutational events, founder mutations, which characterize the Pre-SLC state. The SLC condition develops following mutational events defined as driver mutations which can give rise to clones resistant to anticancer treatments and therefore responsible for the relapse of the disease (figure edited with biorender.com)

Fig. 2

Description of the experimental process for RNA sequencing in single cells. A suspension of cells (A) is used to separate each cell into individual reactors. The figure describes the approach based on the formation of microbubbles (B), which contain beads to allow the capture of RNA released by cell lysis within the reactor (C). RNA capture by the bead in the reactor is permitted by covering the beads with oligonucleotides that have a T-tail in their 3’ terminal portion (D). Each bead contains millions of these oligonucleotides that have common sequences (oligod(T) and R1), a unique sequence for each oligonucleotide (UMI) and a different sequence for each bead (cell barcode) (E). The UMIs allow for the digital counting of gene expression because different UMIs can be associated with the same RNA sequence, based on the abundance of the RNA. The cell barcode allows understanding whether the RNAs being evaluated are derived from the same bead, and therefore cell, or from different beads, and therefore different cells. Oligod(T) allows the capture of polyadenylated RNAs and their retrotranscription (F). Once retrotranscriptase (RT) reaches the 3’ terminal of RNA, due to its 3’ terminal transferase activity, it adds 3 cytosines to the cDNA (G). The overhang of C allows binding of the template switch primer (TSP) and continuation of retrotranscription of this primer (H) as well. At the end of retrotranscription, the cDNA will be characterized by having known ends given by R1 and the template switch primer (I and J). This whole process takes place within each reactor and therefore separately for each cell. Since at the end of the retrotranscription the cDNAs will be labeled with the cell barcode, it is possible to destroy the bubbles and continue the rest of the protocol in bulk. To amplify the low amount of RNA released from each individual cell, a PCR is done by exploiting primers complementary to R1 and TSP (K). This will yield a dsDNA with two known ends (L) that can be used for the sequencing of the full length by binding sequencing adapters to the ends of cDNA by using a PCR or can be fragmented and sequenced the portion corresponding to the 3’ end of RNA.

Fig. 3

Description of gene editing. (A) Cas9 is a protein that can induce double-stranded brakes in a target DNA (blue in the figure). Cas9 is guided at the DNA site through an RNA guide (gRNA) that has a proper 3D conformation and a complementary to a DNA sequence near a PAM motive that is different for different Cas types. (B) Architecture of TALEs. The N-terminal region of a TALE contains a type III secretion signal (T3SS) and four non-canonical repeats (NCR), the C-terminal part a transcription factor binding site (TFB), two nuclear localisation signals (NLS) and an acidic activation domain (AAD). Each TALE also contains a repeat region with a variable number of highly conserved 33-35-aa repeats arranged in tandem. The amino acid sequence of a consensus 34-aa TALE repeat is shown and the amino acids responsible for the DNA specificity of a TALE, the variable repeat di-residues (RVDs), are highlighted. The five most commonly used RVDs and the nucleotides they specify are also shown in the table. In both cases a double strand break in the target DNA sequence is induced. The cell activate repair mechanisms that cause the introduction of an insertion or a deletion at the site where the DNA was broken. If the site is in a coding sequence, the protein will not be produced because the introduction of wrong codon stops