Machine Learning Identifies Stemness Features Associated with Oncogenic Dedifferentiation.

Cancer progression involves the gradual loss of a differentiated phenotype and acquisition of progenitor and stem-cell-like features. Here, we provide novel stemness indices for assessing the degree of oncogenic dedifferentiation. We used an innovative one-class logistic regression (OCLR) machine-learning algorithm to extract transcriptomic and epigenetic feature sets derived from non-transformed pluripotent stem cells and their differentiated progeny. Using OCLR, we were able to identify previously undiscovered biological mechanisms associated with the dedifferentiated oncogenic state. Analyses of the tumor microenvironment revealed unanticipated correlation of cancer stemness with immune checkpoint expression and infiltrating immune cells. We found that the dedifferentiated oncogenic phenotype was generally most prominent in metastatic tumors. Application of our stemness indices to single-cell data revealed patterns of intra-tumor molecular heterogeneity. Finally, the indices allowed for the identification of novel targets and possible targeted therapies aimed at tumor differentiation.

Ordered and deterministic cancer genome evolution after p53 loss.

Although p53 inactivation promotes genomic instability1 and presents a route to malignancy for more than half of all human cancers2,3, the patterns through which heterogenous TP53 (encoding human p53) mutant genomes emerge and influence tumorigenesis remain poorly understood. Here, in a mouse model of pancreatic ductal adenocarcinoma that reports sporadic p53 loss of heterozygosity before cancer onset, we find that malignant properties enabled by p53 inactivation are acquired through a predictable pattern of genome evolution. Single-cell sequencing and in situ genotyping of cells from the point of p53 inactivation through progression to frank cancer reveal that this deterministic behaviour involves four sequential phases-Trp53 (encoding mouse p53) loss of heterozygosity, accumulation of deletions, genome doubling, and the emergence of gains and amplifications-each associated with specific histological stages across the premalignant and malignant spectrum. Despite rampant heterogeneity, the deletion events that follow p53 inactivation target functionally relevant pathways that can shape genomic evolution and remain fixed as homogenous events in diverse malignant populations. Thus, loss of p53-the 'guardian of the genome'-is not merely a gateway to genetic chaos but, rather, can enable deterministic patterns of genome evolution that may point to new strategies for the treatment of TP53-mutant tumours.

Copolymerization of single-cell nucleic acids into balls of acrylamide gel.

We show the use of 5'-Acrydite oligonucleotides to copolymerize single-cell DNA or RNA into balls of acrylamide gel (BAGs). Combining this step with split-and-pool techniques for creating barcodes yields a method with advantages in cost and scalability, depth of coverage, ease of operation, minimal cross-contamination, and efficient use of samples. We perform DNA copy number profiling on mixtures of cell lines, nuclei from frozen prostate tumors, and biopsy washes. As applied to RNA, the method has high capture efficiency of transcripts and sufficient consistency to clearly distinguish the expression patterns of cell lines and individual nuclei from neurons dissected from the mouse brain. By using varietal tags (UMIs) to achieve sequence error correction, we show extremely low levels of cross-contamination by tracking source-specific SNVs. The method is readily modifiable, and we will discuss its adaptability and diverse applications.

Multiplex accurate sensitive quantitation (MASQ) with application to minimal residual disease in acute myeloid leukemia.

Measuring minimal residual disease in cancer has applications for prognosis, monitoring treatment and detection of recurrence. Simple sequence-based methods to detect nucleotide substitution variants have error rates (about 10-3) that limit sensitive detection. We developed and characterized the performance of MASQ (multiplex accurate sensitive quantitation), a method with an error rate below 10-6. MASQ counts variant templates accurately in the presence of millions of host genomes by using tags to identify each template and demanding consensus over multiple reads. Since the MASQ protocol multiplexes 50 target loci, we can both integrate signal from multiple variants and capture subclonal response to treatment. Compared to existing methods for variant detection, MASQ achieves an excellent combination of sensitivity, specificity and yield. We tested MASQ in a pilot study in acute myeloid leukemia (AML) patients who entered complete remission. We detect leukemic variants in the blood and bone marrow samples of all five patients, after induction therapy, at levels ranging from 10-2 to nearly 10-6. We observe evidence of sub-clonal structure and find higher target variant frequencies in patients who go on to relapse, demonstrating the potential for MASQ to quantify residual disease in AML.

Tumour evolution inferred by single-cell sequencing.

Genomic analysis provides insights into the role of copy number variation in disease, but most methods are not designed to resolve mixed populations of cells. In tumours, where genetic heterogeneity is common, very important information may be lost that would be useful for reconstructing evolutionary history. Here we show that with flow-sorted nuclei, whole genome amplification and next generation sequencing we can accurately quantify genomic copy number within an individual nucleus. We apply single-nucleus sequencing to investigate tumour population structure and evolution in two human breast cancer cases. Analysis of 100 single cells from a polygenomic tumour revealed three distinct clonal subpopulations that probably represent sequential clonal expansions. Additional analysis of 100 single cells from a monogenomic primary tumour and its liver metastasis indicated that a single clonal expansion formed the primary tumour and seeded the metastasis. In both primary tumours, we also identified an unexpectedly abundant subpopulation of genetically diverse 'pseudodiploid' cells that do not travel to the metastatic site. In contrast to gradual models of tumour progression, our data indicate that tumours grow by punctuated clonal expansions with few persistent intermediates.

Novel genomic alterations and clonal evolution in chronic lymphocytic leukemia revealed by representational oligonucleotide microarray analysis (ROMA).

We examined copy number changes in the genomes of B cells from 58 patients with chronic lymphocytic leukemia (CLL) by using representational oligonucleotide microarray analysis (ROMA), a form of comparative genomic hybridization (CGH), at a resolution exceeding previously published studies. We observed at least 1 genomic lesion in each CLL sample and considerable variation in the number of abnormalities from case to case. Virtually all abnormalities previously reported also were observed here, most of which were indeed highly recurrent. We observed the boundaries of known events with greater clarity and identified previously undescribed lesions, some of which were recurrent. We profiled the genomes of CLL cells separated by the surface marker CD38 and found evidence of distinct subclones of CLL within the same patient. We discuss the potential applications of high-resolution CGH analysis in a clinical setting.

Novel insights into breast cancer copy number genetic heterogeneity revealed by single-cell genome sequencing.

Copy number alterations (CNAs) play an important role in molding the genomes of breast cancers and have been shown to be clinically useful for prognostic and therapeutic purposes. However, our knowledge of intra-tumoral genetic heterogeneity of this important class of somatic alterations is limited. Here, using single-cell sequencing, we comprehensively map out the facets of copy number alteration heterogeneity in a cohort of breast cancer tumors. Ou/var/www/html/elife/12-05-2020/backup/r analyses reveal: genetic heterogeneity of non-tumor cells (i.e. stroma) within the tumor mass; the extent to which copy number heterogeneity impacts breast cancer genomes and the importance of both the genomic location and dosage of sub-clonal events; the pervasive nature of genetic heterogeneity of chromosomal amplifications; and the association of copy number heterogeneity with clinical and biological parameters such as polyploidy and estrogen receptor negative status. Our data highlight the power of single-cell genomics in dissecting, in its many forms, intra-tumoral genetic heterogeneity of CNAs, the magnitude with which CNA heterogeneity affects the genomes of breast cancers, and the potential importance of CNA heterogeneity in phenomena such as therapeutic resistance and disease relapse.

Cells in the body remain healthy by tightly preventing and repairing random changes, or mutations, in their genetic material. In cancer cells, however, these mechanisms can break down. When these cells grow and multiply, they can then go on to accumulate many mutations. As a result, cancer cells in the same tumor can each contain a unique combination of genetic changes. This genetic heterogeneity has the potential to affect how cancer responds to treatment, and is increasingly becoming appreciated clinically. For example, if a drug only works against cancer cells carrying a specific mutation, any cells lacking this genetic change will keep growing and cause a relapse. However, it is still difficult to quantify and understand genetic heterogeneity in cancer. Copy number alterations (or CNAs) are a class of mutation where large and small sections of genetic material are gained or lost. This can result in cells that have an abnormal number of copies of the genes in these sections. Here, Baslan et al. set out to explore how CNAs might vary between individual cancer cells within the same tumor. To do so, thousands of individual cancer cells were isolated from human breast tumors, and a technique called single-cell genome sequencing used to screen the genetic information of each of them. These experiments confirmed that CNAs did differ ­ sometimes dramatically ­ between patients and among cells taken from the same tumor. For example, many of the cells carried extra copies of well-known cancer genes important for treatment, but the exact number of copies varied between cells. This heterogeneity existed for individual genes as well as larger stretches of DNA: this was the case, for instance, for an entire section of chromosome 8, a region often affected in breast and other tumors. The work by Baslan et al. captures the sheer extent of genetic heterogeneity in cancer and in doing so, highlights the power of single-cell genome sequencing. In the future, a finer understanding of the genetic changes present at the level of an individual cancer cell may help clinicians to manage the disease more effectively.

Two Distinct Categories of Focal Deletions in Cancer Genomes.

One of the key questions about genomic alterations in cancer is whether they are functional in the sense of contributing to the selective advantage of tumor cells. The frequency with which an alteration occurs might reflect its ability to increase cancer cell growth, or alternatively, enhanced instability of a locus may increase the frequency with which it is found to be aberrant in tumors, regardless of oncogenic impact. Here we've addressed this on a genome-wide scale for cancer-associated focal deletions, which are known to pinpoint both tumor suppressor genes (tumor suppressors) and unstable loci. Based on DNA copy number analysis of over one-thousand human cancers representing ten different tumor types, we observed five loci with focal deletion frequencies above 5%, including the A2BP1 gene at 16p13.3 and the MACROD2 gene at 20p12.1. However, neither RNA expression nor functional studies support a tumor suppressor role for either gene. Further analyses suggest instead that these are sites of increased genomic instability and that they resemble common fragile sites (CFS). Genome-wide analysis revealed properties of CFS-like recurrent deletions that distinguish them from deletions affecting tumor suppressor genes, including their isolation at specific loci away from other genomic deletion sites, a considerably smaller deletion size, and dispersal throughout the affected locus rather than assembly at a common site of overlap. Additionally, CFS-like deletions have less impact on gene expression and are enriched in cell lines compared to primary tumors. We show that loci affected by CFS-like deletions are often distinct from known common fragile sites. Indeed, we find that each tumor tissue type has its own spectrum of CFS-like deletions, and that colon cancers have many more CFS-like deletions than other tumor types. We present simple rules that can pinpoint focal deletions that are not CFS-like and more likely to affect functional tumor suppressors.

Oncogenic cooperation and coamplification of developmental transcription factor genes in lung cancer.

We used high-resolution array analysis to discover a recurrent lung cancer amplicon located at 14q13.3. Low-level gain of this region was detected in 15% of lung cancer samples, and high-level amplification was detected in an additional 4% of samples. High-level focal amplification appears to be specific to lung cancers, because it was not detected in >500 samples of other tumor types. Mapping of the commonly amplified region revealed there are three genes in the core region, all of which encode transcription factors with either established lung developmental function (TTF1/NKX2-1, NKX2-8) or potential lung developmental function (PAX9). All three genes were overexpressed to varying degrees in amplified samples, although TTF1/NKX2-1 was not expressed in the squamous cancer subtype, consistent with previous reports. Remarkably, overexpression of any pairwise combination of these genes showed pronounced synergy in promoting the proliferation of immortalized human lung epithelial cells. Analysis of human lung cancer cell lines by both RNAi and ectopic overexpression further substantiates an oncogenic role for these transcription factors. These results, taken together with previous reports of oncogenic alterations of transcription factors involved in lung development (p63, CEBPA), suggest genetic alterations that directly interfere with transcriptional networks normally regulating lung development may be a more common feature of lung cancer than previously realized.

Copy-number variants in patients with a strong family history of pancreatic cancer.

Copy-number variants such as germ-line deletions and amplifications are associated with inherited genetic disorders including familial cancer. The gene or genes responsible for the majority of familial clustering of pancreatic cancer have not been identified. We used representational oligonucleotide microarray analysis (ROMA) to characterize germ-line copy number variants in 60 cancer patients from 57 familial pancreatic cancer kindreds. Fifty-seven of the 60 patients had pancreatic cancer and three had nonpancreatic cancers (breast, ovary, ovary). A familial pancreatic cancer kindred was defined as a kindred in which at least two first-degree relatives have been diagnosed with pancreatic cancer. Copy-number variants identified in 607 individuals without pancreatic cancer were excluded from further analysis. A total of 56 unique genomic regions with copy-number variants not present in controls were identified, including 31 amplifications and 25 deletions. Two deleted regions were observed in two different patients, and one in three patients. The germ-line amplifications had a mean size of 662 Kb, a median size of 379 Kb (range 8.2 Kb to 2.5 Mb) and included 425 known genes. Examples of genes included in the germ-line amplifications include the MAFK, JunD and BIRC6 genes. The germ-line deletions had a mean size of 375Kb, a median size 151 Kb (range 0.4 Kb to 2.3 Mb) and included 81 known genes. In multivariate analysis controlling for region size, deletions were 90% less likely to involve a gene than were duplications (p < 0.01). Examples of genes included in the germ-line deletions include the FHIT, PDZRN3 and ANKRD3 genes. Selected deletions and amplifications were confirmed using real-time PCR, including a germ-line amplification on chromosome 19. These genetic copy-number variants define potential candidate loci for the familial pancreatic cancer gene.

Strong association of de novo copy number mutations with autism.

We tested the hypothesis that de novo copy number variation (CNV) is associated with autism spectrum disorders (ASDs). We performed comparative genomic hybridization (CGH) on the genomic DNA of patients and unaffected subjects to detect copy number variants not present in their respective parents. Candidate genomic regions were validated by higher-resolution CGH, paternity testing, cytogenetics, fluorescence in situ hybridization, and microsatellite genotyping. Confirmed de novo CNVs were significantly associated with autism (P = 0.0005). Such CNVs were identified in 12 out of 118 (10%) of patients with sporadic autism, in 2 out of 77 (3%) of patients with an affected first-degree relative, and in 2 out of 196 (1%) of controls. Most de novo CNVs were smaller than microscopic resolution. Affected genomic regions were highly heterogeneous and included mutations of single genes. These findings establish de novo germline mutation as a more significant risk factor for ASD than previously recognized.