Genome-wide transcriptional analysis of the human cell cycle identifies genes differentially regulated in normal and cancer cells

Ziv Bar‐Joseph, Zahava Siegfried, Michael Brandeis, Benedikt Brors, Yong Lu, Roland Eils, Brian David Dynlacht, Itamar SimonProceedings of the National Academy of Sciences 2008 · DOI

TL;DR

This paper tackles the challenge of accurately identifying genes that cycle with the cell cycle in normal human cells, a crucial step for understanding cancer. They developed a novel computational method to "deconvolve" noisy microarray data from partially synchronized cells, revealing 480 periodically expressed genes, many of which are unique to normal cells and dysregulated in cancer.

What Problem Does This Paper Address?

Understanding the cell cycle is fundamental to biology, and its dysregulation is a hallmark of cancer. Genome-wide studies have aimed to identify genes that oscillate with the cell cycle by measuring gene expression in synchronized cell populations. However, synchronizing normal mammalian cells is notoriously difficult. Unlike yeast, normal mammalian cells often lose synchrony quickly after release from arrest, and a significant portion may not re-enter the cell cycle at all. This leads to "noisy" microarray data where the average expression across the population masks the true, individual cell cycle profiles, making it hard to distinguish genuine cycling genes from random fluctuations. Previous studies often relied on transformed cell lines (like HeLa), which are easier to synchronize but may not reflect the normal cell cycle. This paper addresses the critical need for a reliable method to identify cell cycle-regulated genes in primary, non-transformed human cells.

Methods

The core innovation here is a combined experimental and computational approach to overcome the limitations of partial synchrony.

  1. Experimental Synchronization: They synchronized primary human foreskin fibroblasts using standard methods (serum starvation and thymidine block) but acknowledged the inherent partial synchrony.
  2. Flow Cytometry (FACS): Crucially, they used FACS analysis at multiple time points to quantify the degree of synchrony and the percentage of cells re-entering the cell cycle. This empirical data is key.
  3. Synchronization Loss Model: They developed a computational model that uses the FACS data to infer the distribution of cells around the "ideal" time point. This model accounts for cells that don't re-enter the cycle and variations in individual cell cycle progression rates.
  4. Deconvolution Algorithm: This model is then applied to the raw microarray expression data. The algorithm "deconvolves" the mixed signals from the population to reconstruct an "average single-cell" expression profile for each gene. This is the novel computational step.
  5. Cyclicity Scoring: A scoring method (based on Fourier transform) was applied to these deconvolved profiles to identify genes with statistically significant periodic expression.

Datasets

  • Primary Human Foreskin Fibroblasts (HFFs): This is the primary experimental dataset generated by the authors. RNA was collected at 2-hour intervals over 32 hours after release from synchronization.
  • Published Datasets:
    • Whitfield et al. (2002) dataset of synchronized HeLa cells (transformed cell line).
  • Expression data from normal fibroblasts (IMR-90 cells) in proliferating and senescent states.
  • Expression data from various normal tissues and cancer cell lines.
  • Gene expression profiles of normal epithelial cells and fibrosarcoma cell lines.
  • Data from HUVECs (primary cells) and fibrosarcoma cell lines.

Key Insights and Ideas

  • The "Noisy" Reality of Primary Cell Synchronization: The paper strongly emphasizes that partial synchrony in mammalian cells is not a minor inconvenience but a fundamental challenge that significantly distorts gene expression data.
  • Deconvolution as a Solution: The deconvolution approach is presented as a powerful way to extract meaningful biological signals from inherently imperfect experimental data, effectively creating "virtual" single-cell profiles.
  • A Distinct Set of "Normal" Cycling Genes: A major finding is the identification of a substantial number of genes that cycle specifically in primary cells and not in transformed cell lines. This suggests that cancer cells may have fundamentally altered cell cycle regulation.
  • Implications for Cancer Research: The identification of these unique normal cycling genes provides a new avenue for understanding how their dysregulation contributes to cancer development and progression.

Key Results

  • Identification of 480 Periodically Expressed Genes: The authors identified 480 genes that show periodic expression across the cell cycle in primary human foreskin fibroblasts.
  • Comparison with Transformed Cells: Analysis revealed a significant overlap but also a distinct set of genes cycling exclusively in primary cells (118 genes, termed "Primary FF") compared to those cycling in HeLa cells.
  • Functional Enrichment: The "common" and "Primary FF" gene sets were highly enriched for cell cycle-related Gene Ontology (GO) categories and transcription factor binding motifs (e.g., E2F, NFY), reinforcing their biological relevance.
  • Differential Expression in Cancer: Genes in the "Primary FF" group showed similar expression levels in normal proliferating and arrested cells, but this pattern diverged significantly in cancer cells, suggesting dysregulation. Conversely, "common" genes were generally higher in cancer.
  • Validation in Other Cell Types: Experiments with HUVECs and fibrosarcoma cells supported the finding that some genes cycle primarily in normal cells.

Limitations and Future Work

  • Limitations of Deconvolution: While powerful, the deconvolution method relies on assumptions about cell cycle progression (e.g., normal distribution of rates). The paper acknowledges that the accuracy of the model was validated against time-lapse cinematography, but further validation across diverse cell types and conditions would be beneficial.
  • "HeLa" Genes May Not Be Cycling: The authors suggest that some genes previously identified as cycling in HeLa cells might not be truly cycling but rather exhibiting altered expression due to other cellular perturbations in transformed cells. This implies a potential refinement of existing cell cycle gene lists.
  • Future Work: The authors propose that further investigation into the "Primary FF" genes, particularly those that are dysregulated in transformed cells, could pinpoint genetic elements crucial for normal cell growth and cellular transformation. This opens doors for targeted studies on specific genes and their roles in cancer.

Summary updated

Abstract

Characterization of the transcriptional regulatory network of the normal cell cycle is essential for understanding the perturbations that lead to cancer. However, the complete set of cycling genes in primary cells has not yet been identified. Here, we report the results of genome-wide expression profiling experiments on synchronized primary human foreskin fibroblasts across the cell cycle. Using a combined experimental and computational approach to deconvolve measured expression values into "single-cell" expression profiles, we were able to overcome the limitations inherent in synchronizing nontransformed mammalian cells. This allowed us to identify 480 periodically expressed genes in primary human foreskin fibroblasts. Analysis of the reconstructed primary cell profiles and comparison with published expression datasets from synchronized transformed cells reveals a large number of genes that cycle exclusively in primary cells. This conclusion was supported by both bioinformatic analysis and experiments performed on other cell types. We suggest that this approach will help pinpoint genetic elements contributing to normal cell growth and cellular transformation.
BibTeX
@inproceedings{barjoseph2008genome,
  author = {Ziv Bar‐Joseph and Zahava Siegfried and Michael Brandeis and Benedikt Brors and Yong Lu and Roland Eils and Brian David Dynlacht and Itamar Simon},
  title = {{Genome-wide transcriptional analysis of the human cell cycle identifies genes differentially regulated in normal and cancer cells}},
  booktitle = {Proceedings of the National Academy of Sciences},
  year = {2008},
  doi = {10.1073/pnas.0704723105},
}
Figure