The fractured genome of HeLa cells
© BioMed Central Ltd 2013
Published: 17 April 2013
Skip to main content
© BioMed Central Ltd 2013
Published: 17 April 2013
Whole-genome sequencing of the widely used HeLa cell line provides a nucleotide-resolution view of a greatly mutated and in some places shattered genome.
Human cell culture has proven an invaluable tool for revealing key aspects of human biology. The trouble with growing human cells in culture, or rather most primary cells, is that they eventually senesce or cease to divide. The trick to exploiting human cells in culture, for long-term studies, is to immortalize them so that they continue to divide indefinitely. Long before Shay and Wright demonstrated that ectopic expression of human telomerase reverse transcriptase (hTERT) enabled cells to proliferate endlessly, with seemingly normal phenotypes , scientists relied on naturally immortalized cells, often derived from the tumors of cancer patients. The first such immortalized cell, the HeLa cell line, was established more than half a century ago from the tumor of a cervical cancer patient called Henrietta Lacks. Although Henrietta Lacks eventually died from her cancer in 1951, the HeLa line has continued to proliferate in culture, becoming one of the most commonly used human cell lines in biomedical research. Roughly 60,000 scientific publications have cited the use of HeLa cells and major discoveries have been made using the cell line, including the development of the polio vaccine in 1952, the link between human papillomavirus (HPV) and cervical cancer, and the role of telomerase in chromosome maintenance . Now, for the first time, Lars Steinmetz and colleagues report a comprehensive genomic analysis and expression profile for the popular Kyoto version of the HeLa cell line .
Despite the tremendous value and widespread use of HeLa cells, it has been known for some time that these cells, like most cancer cells, are genetically abnormal, and perhaps even more so than typical cancer-derived cell lines. A previous study  combined spectral karyotyping, fluorescence in situ hybridization and conventional cytogenetic techniques to reveal extensive chromosomal aberrations, including hyper-triploid chromosome number and genetic abnormalities (often used as HeLa markers) on 20 chromosomes. Landry et al.  have now used deep DNA and RNA sequencing to define the HeLa genome and transcriptome, revealing the true extent of genetic abnormalities at nucleotide resolution. This study establishes a reference sequence for the HeLa genome, along with genetic variations identified in the cell line: valuable resources for the continued use of HeLa cells in biomedical research.
Finally, the authors  find extensive copy-number heterogeneity, with most loci found in three or more copies, which is consistent with previous studies reporting a 3n+ chromosome state . Surprisingly, although less than 1% of the HeLa genome is present at a copy number of one, there are large stretches of homozygosity non-uniformly distributed in the HeLa genome. The authors partitioned the genome into 100 kb bins and found that 23% of these bins were composed of mostly homozygous SNVs (purple in Figure 1). In contrast, they did not find any 100 kb bins of homozygosity in HapMap samples, which are more representative of normal human variation. Comparisons of copy numbers with transcriptome data, also generated in the study , indicated a correlation between copy number and expression level, suggesting that dosage compensation does not occur at a global scale in HeLa cells. Among the more highly expressed genes were those enriched for functions such as proliferation, transcription, and DNA repair - arguably valuable assets for life in culture or in a tumor.
A subset of cervical cancers is caused by HPV infection. The authors identified a known insertion of HPV18 on chromosome 8, consistent with previous studies, but also documented nine additional putative viral integration sites . Remarkably, they also found evidence that four of the HeLa chromosomes had been shattered into pieces, with many of the fragments reassembled randomly into highly rearranged chromosomes. This recently described phenomenon, known as chromothripsis, has been found to be associated with 2 to 3% of all cancers . In HeLa cells, evidence of chromothripsis was most pronounced in chromosome 11, which is a hotspot for loss of heterozygosity associated with cervical cancer, the cancer that killed Henrietta Lacks. Whole-genome sequencing was instrumental in revealing the abnormalities of the shattered chromosome 11. In a previous study that used cytogenetic techniques, it was reported that the rearrangement of chromosome 11 segments could not be fully resolved .
Human cell lines are important models for studying biological function and disease, but to maximize insight from model cell lines, it is critical to understand the genomic architecture of these cells. HeLa cells are particularly abnormal compared with non-cancer human cell lines, as well as compared with the human reference sequence. At the same time, most genomic studies in HeLa cells have used the human reference sequence. The authors highlight the critical importance of this in a case study in which they re-evaluate small interfering RNAs (siRNAs) designed for a large-scale screen in HeLa cells. Some of the siRNAs designed to target the human reference sequences failed to elicit effects in HeLa cells because they did not match the sequence of the HeLa cell genome. This shows the importance of validating targets of siRNAs and other reagents designed against a specific genomic sequence. It is equally important to understand the transcriptional (and ultimately the proteomic) profile of model systems to confirm expected properties before initiating a new study. High-throughput sequencing offers a fast and cost-effective way to characterize cell lines, with the price of exome and transcriptome sequencing already below $1,000 and dropping. For some popular cell lines, these data are already available in NCBI's Sequence Read Archive and other public repositories. The use of high-throughput sequencing data to characterize cell lines is timely, not just because of cost, but also because methods have matured for detecting SNVs, indels and more complex variants in the genome [5, 6]. The authors of the HeLa study  are working toward making available not only the HeLa reference sequence, but also all the read data from whole-genome and transcriptome sequencing.
In addition to characterizing cell lines, it is just as important to confirm the identity of the lines. Misidentification of cell lines, sometimes because of contamination from other cells, is a continuing concern . Historically, cell lines have been identified using short tandem repeat (STR) markers that can be assayed with commercial kits, or from validated marker lists provided by sample repositories such as the American Type Culture Collection (ATCC). In the HeLa study , for example, the authors reported that they first confirmed the identity of their HeLa cells using 16 STRs, of which nine were promoted as standards by the ATCC and the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ). They reported that more than 80% of the markers matched, the gold standard set by the ATCC for STR-based cell line identification. In the era of high-throughput sequencing, even exome sequencing provides sufficient data from which a multitude of additional markers can be established. Improved methods for accurately identifying STR genotypes from high-throughput sequencing data  make STR marker identification eminently feasible and open up the possibility of constructing databases with thousands of STRs that uniquely distinguish one cell line from another.
Sequencing the genome and transcriptome of the HeLa cell line is an important milestone in biomedical research. For more than 60 years, scientists have studied many biological processes in HeLa cells, publishing some 60,000 papers along the way. The detailed studies by Steinmetz and his colleagues  will undoubtedly foster even more productive research using HeLa cells. By documenting just how aberrant HeLa genomes are, however, they also heighten our awareness of exactly what it means to select a cell line for a particular study, and they raise the bar for making such decisions.
We thank Jonathan Landry, Lars Steinmetz, Paul Theodor Pyl, Raeka Aiyar, Jan Korbel and Wolfgang Huber for helpful discussions. This work was supported by an R21 grant (NS079926) from the NIH to DM and an R01 grant from the NIH (GM38219) to JHW.