Unsupervised discovery of dynamic cell phenotypic states from transmitted light movies.

Identification of cell phenotypic states within heterogeneous populations, along with elucidation of their switching dynamics, is a central challenge in modern biology. Conventional single-cell analysis methods typically provide only indirect, static phenotypic readouts. Transmitted light images, on the other hand, provide direct morphological readouts and can be acquired over time to provide a rich data source for dynamic cell phenotypic state identification. Here, we describe an end-to-end deep learning platform, UPSIDE (Unsupervised Phenotypic State IDEntification), for discovering cell states and their dynamics from transmitted light movies. UPSIDE uses the variational auto-encoder architecture to learn latent cell representations, which are then clustered for state identification, decoded for feature interpretation, and linked across movie frames for transition rate inference. Using UPSIDE, we identified distinct blood cell types in a heterogeneous dataset. We then analyzed movies of patient-derived acute myeloid leukemia cells, from which we identified stem-cell associated morphological states as well as the transition rates to and from these states. UPSIDE opens up the use of transmitted light movies for systematic exploration of cell state heterogeneity and dynamics in biology and medicine.

Mutation Patterns Predict Drug Sensitivity in Acute Myeloid Leukemia.

PURPOSE: The inherent genetic heterogeneity of acute myeloid leukemia (AML) has challenged the development of precise and effective therapies. The objective of this study was to elucidate the genomic basis of drug resistance or sensitivity, identify signatures for drug response prediction, and provide resources to the research community. EXPERIMENTAL DESIGN: We performed targeted sequencing, high-throughput drug screening, and single-cell genomic profiling on leukemia cell samples derived from patients with AML. Statistical approaches and machine learning models were applied to identify signatures for drug response prediction. We also integrated large public datasets to understand the co-occurring mutation patterns and further investigated the mutation profiles in the single cells. The features revealed in the co-occurring or mutual exclusivity pattern were further subjected to machine learning models. RESULTS: We detected genetic signatures associated with sensitivity or resistance to specific agents, and identified five co-occurring mutation groups. The application of single-cell genomic sequencing unveiled the co-occurrence of variants at the individual cell level, highlighting the presence of distinct subclones within patients with AML. Using the mutation pattern for drug response prediction demonstrates high accuracy in predicting sensitivity to some drug classes, such as MEK inhibitors for RAS-mutated leukemia. CONCLUSIONS: Our study highlights the importance of considering the gene mutation patterns for the prediction of drug response in AML. It provides a framework for categorizing patients with AML by mutations that enable drug sensitivity prediction.

Accurate detection of subclonal variants in paired diagnosis-relapse acute myeloid leukemia samples by next generation Duplex Sequencing.

Mutations characterize diverse human cancers; there is a positive correlation between elevated mutation frequency and tumor progression. One exception is acute myeloid leukemia (AML), which has few clonal single nucleotide mutations. We used highly sensitive and accurate Duplex Sequencing (DS) to show now that AML, in addition, has an extensive repertoire of variants with low allele frequencies, < 1%, which is below the accurate detection limit of most other sequencing methodologies. The subclonal variants are unique to each individual and change in composition, frequency, and sequence context from diagnosis to relapse. Their functional significance is apparent by the observation that many are known variants and cluster within functionally important protein domains. Subclones provide a reservoir of variants that could expand and contribute to the development of drug resistance and relapse. In accord, we accurately identified subclonal variants in AML driver genes NRAS and RUNX1 at allele frequencies between 0.1% and 0.3% at diagnosis, which expanded to comprise a major fraction (14-53%) of the blast population at relapse. Early and accurate detection of subclonal variants with low allele frequency thus offers the opportunity for early intervention, prior to detection of clinical relapse, to improve disease outcome and enhance patient survival.

Very late antigen-4 function of myeloblasts correlates with improved overall survival for patients with acute myeloid leukemia.

Adhesion of acute myeloid leukemia (AML) blasts in the bone marrow microenvironment confers protection from chemotherapy-induced apoptosis. One mechanism for retention of blasts within the bone marrow is adhesion via very late antigen-4 (VLA-4), the alpha(4)beta(1) integrin heterodimer that binds to its main ligands, fibronectin, and vascular cell adhesion molecule-1 (VCAM-1). To examine the relationship of functional expression of VLA-4 to prognosis in AML, we studied marrow samples from 175 adult AML patients who underwent induction chemotherapy with anthracycline and cytarabine on Southwest Oncology Group trials. The studies included flow cytometry and functional in vitro assays for ligand binding and maximal beta(1) activation. VLA-4 expression varied widely, with mean expression 60.6% for alpha(4), and was not significantly associated with response to chemotherapy, relapse-free, or overall survival (OS). However, increased binding of soluble VCAM-1 via VLA-4 was significantly associated with longer OS, corrected for age (P = .033). Estimated 5-year OS was 31% (95% confidence interval, 14%-48%) in 30 patients with soluble VCAM-1 binding greater than or equal to 40%, compared with 10% (confidence interval, 3%-17%) in 72 patients with lower binding. Adhesion and migratory properties of AML blasts thus appear to influence chemosensitivity and therefore may be therapeutic targets.

Restoring RUNX1 deficiency in RUNX1 familial platelet disorder by inhibiting its degradation.

RUNX1 familial platelet disorder (RUNX1-FPD) is an autosomal dominant disorder caused by a monoallelic mutation of RUNX1, initially resulting in approximately half-normal RUNX1 activity. Clinical features include thrombocytopenia, platelet functional defects, and a predisposition to leukemia. RUNX1 is rapidly degraded through the ubiquitin-proteasome pathway. Moreover, it may autoregulate its expression. A predicted kinetic property of autoregulatory circuits is that transient perturbations of steady-state levels result in continued maintenance of expression at adjusted levels, even after inhibitors of degradation or inducers of transcription are withdrawn, suggesting that transient inhibition of RUNX1 degradation may have prolonged effects. We hypothesized that pharmacological inhibition of RUNX1 protein degradation could normalize RUNX1 protein levels, restore the number of platelets and their function, and potentially delay or prevent malignant transformation. In this study, we evaluated cell lines, induced pluripotent stem cells derived from patients with RUNX1-FPD, RUNX1-FPD primary bone marrow cells, and acute myeloid leukemia blood cells from patients with RUNX1 mutations. The results showed that, in some circumstances, transient expression of exogenous RUNX1 or inhibition of steps leading to RUNX1 ubiquitylation and proteasomal degradation restored RUNX1 levels, thereby advancing megakaryocytic differentiation in vitro. Thus, drugs retarding RUNX1 proteolytic degradation may represent a therapeutic avenue for treating bleeding complications and preventing leukemia in RUNX1-FPD.

High-Throughput Drug Screening and Multi-Omic Analysis to Guide Individualized Treatment for Multiple Myeloma.

Multiple myeloma (MM) is a genetically heterogeneous malignancy characterized by variable treatment responses. Although numerous drugs have been approved in recent years, the ability to predict treatment response and tailor individual therapy is limited by the absence of robust predictive biomarkers. The goal of this clinical trial was to use ex vivo, high-throughput screening (HTS) of 170 compounds to predict response among patients with relapsed or refractory MM and inform the next treatment decisions. Additionally, we integrated HTS with multi-omic analysis to uncover novel associations between in vitro drug sensitivity and gene expression and mutation profiles. MATERIALS AND METHODS: Twenty-five patients with relapsed or refractory MM underwent a screening bone marrow or soft tissue biopsy. Sixteen patients were found to have sufficient plasma cells for HTS. Targeted next-generation sequencing was performed on plasma cell-free DNA from all patients who underwent HTS. RNA and whole-exome sequencing of bone marrow plasma cells were performed on eight and seven patients, respectively. RESULTS: Results of HTS testing were made available to treating physicians within a median of 5 days from the biopsy. An actionable treatment result was identified in all 16 patients examined. Among the 13 patients who received assay-guided therapy, 92% achieved stable disease or better. The expression of 105 genes and mutations in 12 genes correlated with in vitro cytotoxicity. CONCLUSION: In patients with relapsed or refractory MM, we demonstrate the feasibility of ex vivo drug sensitivity testing on isolated plasma cells from patient bone marrow biopsies or extramedullary plasmacytomas to inform the next line of therapy.

A machine learning approach to integrate big data for precision medicine in acute myeloid leukemia.

Cancers that appear pathologically similar often respond differently to the same drug regimens. Methods to better match patients to drugs are in high demand. We demonstrate a promising approach to identify robust molecular markers for targeted treatment of acute myeloid leukemia (AML) by introducing: data from 30 AML patients including genome-wide gene expression profiles and in vitro sensitivity to 160 chemotherapy drugs, a computational method to identify reliable gene expression markers for drug sensitivity by incorporating multi-omic prior information relevant to each gene's potential to drive cancer. We show that our method outperforms several state-of-the-art approaches in identifying molecular markers replicated in validation data and predicting drug sensitivity accurately. Finally, we identify SMARCA4 as a marker and driver of sensitivity to topoisomerase II inhibitors, mitoxantrone, and etoposide, in AML by showing that cell lines transduced to have high SMARCA4 expression reveal dramatically increased sensitivity to these agents.

Cyclophosphamide promotes engraftment of gene-modified cells in a mouse model of Fanconi anemia without causing cytogenetic abnormalities.

A major hurdle for hematopoietic stem cell (HSC) gene therapy for inherited bone marrow disorders, including Fanconi anemia (FA), is adequate engraftment of gene-modified cells. A phenotypic defect in DNA repair renders FA patients sensitive to alkylating agents such as cyclophosphamide (Cy); however, at lower doses, Cy is well tolerated in the FA transplant setting. We tested whether non-alkylating agents could replace Cy for pretransplant conditioning to enhance engraftment of FANCA gene-modified hematopoietic cells. We compared Cy preconditioning with fludarabine (Flu) or cytarabine (AraC) or no conditioning as a control in fanca ( -/- ) mutant mice receiving gene-modified bone marrow cells. Only mice conditioned with Cy exhibited appreciable engraftment of gene-modified cells by PCR and resistance to mitomycin C (MMC). Cy administration following transplantation increased gene marking levels in all animals treated, but highest gene marking and corresponding MMC resistance were observed in mice receiving Cy pre- and posttransplantation. Importantly, no cytogenetic abnormalities were observed in Cy-treated mice. We conclude that Cy is an effective and superior preparative regimen with respect to engraftment of lentivirus-transduced cells and functional correction in fanca ( -/- ) mice. Thus, appropriately dosed Cy may provide a suitable conditioning regimen for FA patients undergoing HSC gene therapy.

Differences in F36VMpl-based in vivo selection among large animal models.

Animal models are indispensable tools for understanding physiological and pathological processes, as well as for developing new therapies. Ultimately, the results of animal experimentation must provide information that can guide the development of therapeutic approaches in humans. Significant differences have been reported comparing a gene therapy approach between different animal models. However, little information exists describing differences among the available large animal models. Here we evaluated, in the hemopoietic cells of baboons, a system of selection that has previously demonstrated activity in mice, in dogs, and in human cells ex vivo. This system employs a derivative of the murine thrombopoietin receptor (F36Vmpl), which is conditionally activated in the presence of a small-molecule drug called a chemical inducer of dimerization (CID). Whereas cultured mouse, human, and, to a lesser extent, dog hemopoietic cells all proliferate in response to the F36Vmpl signal, we observed only a minor and variable response to the F36Vmpl signal in the cultured cells of baboons. Similarly, we have noted significant rises in the frequency of transduced hemopoietic cells in mice and in dogs upon CID administration in vivo; however, here we show that responses to CID administration in three baboons were modest and variable. These findings have general implications for the evaluation and development of new strategies for gene therapy.