Big Questions in 2020: Single-cell Sequencing in Myeloid Malignancies Brings Us Closer to Answers

The past decade has seen remarkable progress in both research and clinical care for patients with myeloid diseases. This includes advances in understanding of and treatment for clonal hematopoiesis (CH), myelodysplastic syndromes (MDS), myeloproliferative neoplasms (MPNs), and acute myeloid leukemias (AMLs). The year 2020 has been no different. Despite the global impact of the COVID-19 pandemic, valiant and game-changing research, both basic and clinical, has continued in the myeloid space. The innovations discussed below note key areas of progress, but importantly, they also highlight critical gaps in our understanding of the biology of the myeloid disorders.

Our myeloid arsenal has dramatically expanded in the past few years to include several targeted therapies in AML.^1-4  Now, 2020 has brought a new combination standard-of-care for older patients with AML,⁵  and the approval of the first new drug for MDS in 14 years, luspatercept.⁶  Additionally, the combination of decitabine and cezaduridine⁷  was approved as the first oral hypomethylating agent treatment for higher-risk MDS, while oral azacitidine was approved as maintenance therapy for patients post-induction for AML.⁸  All of these agents represent increasingly available treatments for patients. However, it is noteworthy that in the era of personalized medicine, most of these therapies are not treating the specific biology of myeloid diseases based on clonal evolution in an individual. Furthermore, the trial designs may not have given us guidance toward sequencing or combining these drugs rationally in human beings for future use. This critical gap may soon be addressed as we ask relevant biologic questions to advance bedside care.

Can we avoid groupthink and focus on a single cell? In a recent article in Nature,⁹  Dr. Linde A. Miles and colleagues provide insight into the clonal architecture and evolution of myeloid disorders to expand our understanding of the genomic landscape of myeloid malignancies. This article represents a marked innovation, not just in 2020 but also in the field. The authors make significant strides toward deciphering the question of somatic clonal evolution through single-cell sequencing. Compared to prior efforts with bulk sequencing, this novel methodology can distinguish both clonal complexity and order of mutations. These granular data are critical for thoughtful and rational drug design, to both choose biologically plausible targets and elucidate rational therapeutic regimens from the approved drugs we have. Whereas bulk sequencing provides global biological and prognostic information, single-cell sequencing has the potential to allow us to specifically target myeloid disease profiles and sequence therapies in individual patients in nearly real time.

Does a single mutation a myeloid malignancy make? Dr. Miles and colleagues performed single-cell DNA sequencing on 146 samples from 123 patients with clonal CH, AML, and MPN, focusing on 31 genes known to be mutated in these diseases. This allowed the authors to delineate mutations at single clone resolution and to build maps of clonal complexity as well as architecture in all three disorders. The results demonstrate that CH is a disease of oligoclonality with different mutations in different clones. In contrast, AML and MPNs are diseases of complex clonal evolution with progressive mutational acquisition, clonal branching, and emergence of dominant clones and significant subclones. Most patients with AML have one or two dominant clones, and these clones frequently harbor co-occurring mutations in epigenetic regulators. Interestingly, FLT3/RAS mutations in this dataset were rarely in the dominant clones and were commonly subclinical — key as we think about where to sequence our FLT3 inhibitors. A Markov model was used to predict the optimal clonal trajectory in each individual sample, allowing the authors to show which mutational combinations are most likely to drive clonal expansion. At the bedside, this could be paramount as we develop therapeutic algorithms for individual patients and perhaps target the mutations before they combine to cause disease progression and even prevention at the CH level.

Which myeloid mutation came first, and why does it matter? Dr. Miles’ group then suggested predictions for the initiating mutation. Specific mutational combinations such as co-occurring DNMT3A/IDH or NPM1/FLT3 mutations promote clonal expansion. Next, the group used combined single-cell DNA sequencing as well as cell surface profiling to characterize cell surface protein expression based on clone genotype. Specific subclones, especially those with signaling mutations, were associated with specific immunophenotypes. One piece of data with potential implications for bedside care is the observation that the dominant clones had a distinct immunophenotype from the minor subclones, suggesting that cell-surface targeting in leukemia may need to be clone directed and not just directed at bulk disease. This may have ramifications for targeted immune effector cell therapies.

The publication by Dr. Miles and colleagues represents a striking advance in our knowledge of mutational cooperativity and our resultant understanding of myeloid disorder biology. While exploratory, these data are an example of the potential power of this technology. We can expect further mechanistic and functional studies from this group and others that will truly bring the bench to the bedside and influence patient care far beyond 2020. As such, with the style of questions above, I will evaluate many of the myeloid basic and clinical abstracts at the 2020 ASH Annual Meeting for consideration of future application of single-cell techniques at the bedside.

While many will be eager to forget 2020 for the pandemic and other complexities, one happy memory of this year may be the remarkable growth in our basic understanding of myeloid disorders, which I believe will guide use of current therapies and future drug and trial design.