PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Chem Biol Interact. Author manuscript; available in PMC Mar 19, 2011.
Published in final edited form as:
PMCID: PMC2846187
NIHMSID: NIHMS170945
Systems biology of human benzene exposure
Luoping Zhang,a* Cliona M. McHale,a Nathaniel Rothman,b Guilan Li,c Zhiying Ji,a Roel Vermeulen,d Alan E. Hubbard,a Xuefeng Ren,a Min Shen,b Stephen M. Rappaport,a Matthew North,e Christine F. Skibola,a Songnian Yin,c Christopher Vulpe,e Stephen J. Chanock,b Martyn T. Smith,a and Qing Lanb
a Division of Environmental Health Sciences, School of Public Health, University of California, Berkeley, CA 94720-7356, USA
b Division of Cancer Epidemiology and Genetics, National Cancer Institute, NIH, 6120 Executive Boulevard, MSC 7242, Bethesda, MD 20892, USA
c Chinese Center for Disease Control and Prevention, 27 Nan Wei Road, Beijing 100050, China
d Environmental Epidemiology, Institute of Risk Assessment Sciences, Jenalaan 18d, 3584 CK, Utrecht, Netherlands
e Department of Nutritional Science and Toxicology, University of California, Berkeley, 119 Morgan Hall #3104, Berkeley, CA 94720, USA
* Corresponding Author: Tel: (510) 643–5189; Fax: (510) 642-0427 ; luoping/at/berkeley.edu (Luoping Zhang)
Toxicogenomic studies, including genome-wide analyses of susceptibility genes (genomics), gene expression (transcriptomics), protein expression (proteomics), and epigenetic modifications (epigenomics), of human populations exposed to benzene are crucial to understanding gene-environment interactions, providing the ability to develop biomarkers of exposure, early effect and susceptibility. Comprehensive analysis of these toxicogenomic and epigenomic profiles by bioinformatics in the context of phenotypic endpoints, comprises systems biology, which has the potential to comprehensively define the mechanisms by which benzene causes leukemia. We have applied this approach to a molecular epidemiology study of workers exposed to benzene. Hematotoxicity, a significant decrease in almost all blood cell counts, was identified as a phenotypic effect of benzene that occurred even below 1ppm benzene exposure. We found a significant decrease in the formation of progenitor colonies arising from bone marrow stem cells with increasing benzene exposure, showing that progenitor cells are more sensitive to the effects of benzene than mature blood cells, likely leading to the observed hematotoxicity. Analysis of transcriptomics by microarray in the peripheral blood mononuclear cells of exposed workers, identified genes and pathways (apoptosis, immune response, and inflammatory response) altered at high (>10ppm) and low (<1ppm) benzene levels. Serum proteomics by SELDI-TOF-MS revealed proteins consistently down-regulated in exposed workers. Preliminary epigenomics data showed effects of benzene on the DNA methylation of specific genes. Genomic screens for candidate genes involved in susceptibility to benzene toxicity are being undertaken in yeast, with subsequent confirmation by RNAi in human cells, to expand upon the findings from candidate gene analyses. Data on these and future biomarkers will be used to populate a large toxicogenomics database, to which we will apply bioinformatic approaches to understand the interactions among benzene toxicity, susceptibility genes, mRNA, and DNA methylation through a systems biology approach.
Keywords: benzene, toxicogenomics, epigenetics, systems biology, hematotoxicity, human
Systems biology is a recent and evolving interdisciplinary field that focuses on the systematic study of complex interactions in biological systems [1, 2]. Systems biology employs a holistic approach to study all components and interactions in the network of DNA (genes), RNA, proteins and biochemical reactions within a cell or organism. This new field utilizes powerful tools that include toxicogenomics, epigenomics, bioinformatics, and phenomics, classical toxicological or phenotypic endpoints (Figure 1).
Figure 1
Figure 1
Overview of systems biology and its components
Toxicogenomics combines toxicology with molecular profiling technologies, including genomics (DNA), transcriptomics (mRNA), proteomics (proteins) and metabolomics (chemical metabolites) to elucidate molecular mechanisms involved in chemically-induced toxicity. Chemically-induced alterations in the transcriptome, proteome, and metabolome are analyzed in the context of the stable, inherited genome, which is assessed by genomics. Toxicogenomic studies of human populations are crucial to understanding gene-environment interactions, and can provide the ability to develop novel biomarkers of exposure (exposome), early effect (responsome), and susceptibility (genome) [35]. Epigenomics is the study of epigenetic elements, including DNA methylation (methylomics), non-coding microRNA (miRNAomics) along with small interfering RNA (siRNA) and short hairpin RNA (shRNA) for RNA interference (RNAi), and histone modification. Epigenetic modifications play an essential role in regulating gene expression and biological and molecular functions in living cells, without altering the genome.
Another tool central to systems biology is bioinformatics, the application of computational information technology to the field of molecular biology to understand how cells and cell systems work [6, 7]. Bioinformatics facilitates the analysis of complex biological data (toxicogenomic and epigenomic endpoints) and applies knowledge from annotated functions, pathways and networks to describe normal and perturbed biological states, also known as phenomics, the study of outcomes (phenotypic endpoints). Together, these omic technologies can each provide a “molecular signature” or “fingerprint” of chemical exposure, early effect or genetic susceptibility, which may enhance our understanding of gene-environment interactions. Thus, this holistic approach known as systems biology has the potential to comprehensively define the mechanisms contributing to disease. The purpose of this review paper is to describe how and why it is important to apply the “Systems Biology” approach to benzene mechanistic studies and future directions.
Benzene, a ubiquitous chemical, is an established cause of acute myeloid leukemia (AML), myelodysplastic syndromes (MDS), and probably lymphocytic leukemias and non-Hodgkin lymphoma (NHL) in humans [811]. Benzene-induced toxicity in blood-forming systems has been known for more than a century [12]. In 1982, the International Agency for Research on Cancer (IARC) stated, “There is sufficient evidence that benzene is carcinogenic to man” [13], and when a new IARC classification system was established in 1987, benzene was immediately placed in the Group 1 human carcinogen category [14].
Potential mechanisms of benzene toxicity have been investigated primarily in the following areas [15]: 1) benzene metabolism in the liver (CYP2E1, etc.) and transportation to the bone marrow for secondary metabolism (MPO, NQO1) [16, 17]; 2) oxidative stress from reactive oxygen species generated by redox cycling [18, 19]; 3) chromosome alterations including translocations, deletions, and aneuploidy [12]; 4) protein damage to tubulin, histone proteins, topoisomerase II, etc. [15]; and 5) immune system dysfunction (TNF-α, INF-γ, AhR, etc.) [2022]. Benzene induces chromosomal alterations similar to those found in therapy-related MDS and AML (t-MDS/AML), and in de novo leukemia [23]. Distinct chromosome effects arise following exposure to alkylating agents (5q-/-5 or 7q-/-7 and associated genetic abnormalities) and topoisomerase II inhibitors (recurrent balanced translocations or inversions) used in chemotherapy treatment. Exposure to benzene or its metabolites has been associated with loss and long (q) arm deletion of chromosomes 5 and 7 [24] and translocations involving t(21q) [25, 26], further suggesting that benzene induces leukemia through multiple different mechanisms.
Studies to date have provided evidence for multiple potential mechanisms using approaches that rely on limited research tools that analyze only one or a few, a priori selected genes, pathways or metabolites at a time. A systems biology approach is necessary to interrogate all potential mechanisms by which benzene exposure contributes to disease, through the application of unbiased omic-based technologies in an integrated manner. Since the last international benzene conference at Munich in 2004, we have adopted such an approach to understand the mechanisms underlying human benzene toxicity. This review summarizes our findings published over the last 5 years and preliminary data from recent pilot studies (Table 1). It also provides an overview of our current understanding of benzene-induced hematotoxicity and suggestions for further research.
Table 1
Table 1
Summary of benzene studies applying systems biology approach
We describe the studies in the context of systems biology as defined in Figure 1. First, we discuss hematotoxicity as a phenotypic outcome of benzene exposure, with implications for adverse future health effects. We then discuss findings from two toxicogenomic studies, transcriptomics and proteomics, followed by preliminary epigenomics data. As these toxicogemomic and epigenomic responses to exposure are likely influenced by susceptibility, next we describe how we investigated human susceptibility genes using a yeast genomic screening approach with validation of homologous human genes in human cells. We also describe genotyping studies of candidate genes in human exposed populations. Finally, we discuss how the current and future omic datasets could be integrated, using sophisticated bioinformatics approaches in progress, into one consolidated model of the perturbations effected by benzene. This could identify robust biomarkers and help to clarify the molecular and cellular networks impacted by benzene, yielding a more comprehensive understanding of the mechanistic effects of benzene.
Although benzene was known to have toxic effects on the hematopoietic system (hematotoxicity) at high, occupational doses for over a century [12], the degree of hematotoxicity at low levels of exposure was largely unknown. Recently, a study of 250 workers exposed to varying levels of benzene and 140 unexposed controls in Tianjin, China, during which benzene and other chemical exposure levels were monitored repeatedly for up to 12 months, was conducted. Air, urine and blood samples were collected and complete blood counts (CBC) analyzed [27]. In comparison with the non-exposed controls (n=140), a significant decrease was observed in almost all blood cell counts, such as white blood cells (WBC), granulocytes, lymphocytes, platelets etc, in workers exposed to benzene (n=250), even at exposures below 1 ppm (n=109), the current occupational standard in the U.S. Additionally, lymphocyte subset analysis showed significant, dose-dependent, decreases in CD4+-T cells, CD4+/CD8+ ratio, and B cells at < 1 ppm benzene exposures. These findings, based on the differentiated blood cell counts, provide evidence of bone marrow toxicity in workers exposed to benzene at or below 1 ppm [27].
Because all types of WBC counts were suppressed, it was suspected that the number or functionality of hematopoietic stem and/or progenitor cells generated in the bone marrow had been reduced by benzene. To test this hypothesis, we cultured the progenitor cells circulating in peripheral blood and examined the effects of benzene on different types of progenitor cell colony formation (CFU-GM, BFU-E, CFU-GEMM). The results showed highly significant dose-dependent decreases in colony formation from all three types of progenitor cells, especially when compared to the corresponding decreased levels of differentiated WBC and granulocytes [27]. This suggests that early myeloid progenitor cells are more sensitive than mature cells to the effects of benzene, and clarifies the role of benzene in reduced blood cell counts.
Overall, these hematologic effects could reflect events in the bone marrow that may be associated with adverse health effects in the future [28]. Having established that hematotoxicity, specifically effects myeloid progenitor cells, as a phenotypic anchor of benzene toxicity, we began to examine the molecular mechanisms underlying these effects, through the comprehensive systems biology approach proposed above.
Transcriptomic studies are useful in determining the impact of environmental or occupational exposure to chemicals on the transcriptome, the set of all mRNA transcripts expressed within a cell. To better understand the risks of benzene in humans, the peripheral blood mononuclear cell (PBMC) transcriptomes from occupationally exposed workers in China were examined by microarray (Affymetrix). Analysis of six exposed-control pairs revealed differential expression in 29 genes in the exposed individuals, compared to the controls. Four genes, CXCL16, ZNF331, JUN, and PF4, were shown to be potential biomarkers of early response to benzene exposure as they were confirmed by quantitative-polymerase chain reaction (q-PCR) [29].
A later study of 8 exposed-control pairs confirmed these results, using 2 different microarray platforms (Affymetrix & Illumina) to identify global gene expression changes. The differential expression of 2692 genes and 1828 genes was found by Affymetrix and Illumina, respectively, and the 4 genes, CXCL16, ZNF331, JUN, and PF4, were among the most significantly altered, validating the findings from the earlier Forrest et al. study. This study additionally identified biological pathways that were associated with high benzene exposure, including genes involved in apoptosis and lipid metabolism. This study used a two-platform approach that identified robust changes in the PBMC transcriptome of benzene-exposed individuals [30].
The effects of exposure to high levels of benzene are well documented compared to low-level exposure, the latter being more challenging due to confounders. More recently, we have shown, in an expanded study of 125 factory workers, that low-dose benzene exposure (<1 ppm, n=59) can also cause widespread subtle, yet highly significant, perturbation of gene expression in PBMC. This study was designed with sufficient power to detect robust expression changes, accounting for technical variability as well as age, gender and other confounders. Our microarray analysis revealed significant dysregulation of more than 2500 genes by low dose benzene exposure, over 70 of which had differential expression ratios exceeding 1.5. Several of the detected genes exhibited significantly altered expression only at low levels of benzene exposure, and are thus potential biomarkers of low-dose exposure. The findings show that even low levels of occupational benzene exposure cause a significant perturbation of expression of genes involved in immune and inflammatory responses [31].
Another important toxicogenomic tool, proteomics, can be used to measure alterations in the proteome (e.g. protein levels, posttranslational modifications) associated with exogenous chemical exposure. Effects on the blood proteome may reflect effects at distal body sites. As with transcriptomics, proteomics can be used to discover biomarkers of exposure and early effect, as well as increase our understanding of the mechanisms underlying disease.
We examined the impact of benzene on the human serum proteome in exposed factory workers and controls to obtain insight into the mechanism of action of benzene [32]. Serum samples were fractionated and proteins were bound to surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF-MS) chips. Three proteins (4.1, 7.7, and 9.3 kDa) were consistently down-regulated in the exposed (n=10) compared to the control (n=10) individuals in two separate sets of study subjects (40 subjects total). All proteins were highly inversely correlated with individual estimates of benzene exposure. The 7.7- and 9.3-kDa proteins were identified as platelet factor 4 (PF4), also down-regulated at the gene expression level described above, and connective tissue activating peptide CTAP-III, respectively, both platelet-derived CXC chemokines. Thus, reduced protein levels of PF4 or CTAP-III are potential biomarkers of the early biologic effects of benzene. Future proteomic studies could identify further biomarkers of benzene exposure, and elucidate the mechanisms underlying benzene toxicity and associated disease.
Gene expression and ultimately protein expression is regulated at the epigenetic level by processes including DNA methylation, histone modification and miRNA (microRNA) expression. The epigenome, while stable through cell division and even in some cases reproduction, can be reprogrammed by nutritional, chemical, and physical factors [33]. Thus, the study of toxic effects on the epigenome is crucial to understanding mechanisms of action. Further, epigenetic modifications represent more stable biomarkers and fingerprints of exposure than altered gene or protein expression [34]. While epigenetics refers to the study of individual or specific gene activity, epigenomics focuses on global analyses of epigenetic changes across the entire genome. A recent study reported that hyper-methylation in p15 and genome-wide hypo-methylation assessed by LINE-1 (Long Interspersed Nuclear Element-1) were associated with very low benzene exposures (~22 ppb), in healthy subjects including gas station attendants and traffic police officers, although the corresponding effects on methylation were very low [35]. To determine whether epigenetics plays a role in the hematotoxicity of benzene, we have recently performed several pilot epigenomic studies including DNA methylation and miRNA expression arrays in the blood of workers occupationally exposed to benzene. The results described below are very preliminary, and serve mainly as a proof of principle of this epigenomic approach.
DNA Methylation Array
A DNA methylation array (GoldenGate Methylation Cancer Panel I, Illumina) was applied to determine the methylation status of the CpG islands of >800 genes in DNA isolated from the buffy coats of 6 benzene-exposed workers (2 male, 4 female) and 4 unexposed controls (2 male, 2 female). As expected gender-specific methylation patterns were seen for numerous genes including ELK1, EFNB1, MYCL2, VBP1, DNASE1L1, DKC1 and CDM. This pilot study also found altered methylation induced by benzene at many CpG sites. Decreased methylation of RUNX3 (AML2), a gene whose altered expression has been associated with myeloproliferative disorders [36] occurred at three different CpG sites (Figure 2A). Increased methylation of MSH3, a critical gene in the maintenance of genome integrity, and Sema3C, a secreted guidance protein implicated in tumorigenesis [37], was also observed (Figure 2A). There appeared to be a gender-specific effect of benzene on the methylation of several genes, although the sample number was small.
Figure 2
Figure 2
Heatmap of methylation profiles of (A) workers exposed to benzene and controls, and (B) TK6 cells exposed to hydroquinone (HQ)
We examined the methylation status of the same panel of genes in DNA from TK6 cells treated in vitro with hydroquinone (HQ) at 10, 15, and 20 μM, for 48 h. The most significantly hyper-methylated gene was IL12 (Figure 2B), whose expression has previously been shown to be down-regulated by HQ in mouse macrophages [38]. RUNX1T1 (runt-related transcription factor 1, also known as ETO), was hypo-methylated by HQ in vitro (Figure 2B). The protein encoded by this gene is commonly fused to RUNX1 (AML1) in the t(8;21)(q22;q22) translocation, which is one of the most frequent karyotypic abnormalities in AML [39]. MAGEA1 (melanoma antigen family A, 1), also known as MAGE-1, was also hypo-methylated by HQ in the present study (data not shown). Interestingly, this gene was reported recently to be hypo-methylated weakly in subjects exposed to increasing airborne benzene levels [35].
miRNA Microarrays
Human miRNA microarrays (Agilent), containing probes for 470 human and 64 human viral miRNAs, were used to analyze the differential expression of miRNAs in the total PBMC RNA from 7 exposed-control matched pairs, in a pilot study. Preliminary analysis showed upregulation of 4 miRNAs (miR-154*; miR-487a; miR-493-3p; and, miR-668) by benzene exposure. Upregulation of miR-154* expression, possibly through a change in the methylation and acetylation status of the 14q32 region, has been reported in patients with acute promyelocytic leukemia bearing the t(15;17) translocation [40].
While both of these studies are relatively small, and the data are very preliminary, the findings suggest that further studies to examine the epigenetic effects of benzene on gene-specific promoter methylation and miRNA expression, in a larger study of exposed workers, are warranted.
Genome-wide association studies (GWAS) are a well-established means to examine the association of genetic susceptibility, i.e. single-nucleotide polymorphisms (SNPs), with disease. Such studies have been performed for two diseases related to benzene toxicity, NHL [41] and t-AML [42]. A very small number of pharmacogenomic GWAS have been reported [43], while to our knowledge, no GWAS on occupational or environmental exposure and associated toxic outcomes have been performed. Given the relatively small effects observed in disease GWAS, such studies are very expensive to undertake for less well-defined, pre-disease, toxicological outcomes. We adopted an alternative genomic approach to discover human susceptibility genes, the aims of which are: 1) to discover/screen susceptible genes in yeast by genomics; 2) to select human homolog genes using bioinformatics; 3) to test identified gene (e.g. WRN etc) functions by RNAi in human cells; 4) to identify human susceptible genes by SNP genotyping in population studies.
Genomic screening in yeast
In order to reduce the complexity and expense of analyzing the human genome, while retaining the ability to systematically screen a genome highly relevant to human biology, we chose a screening system in yeast (Saccharomyces cerevisiae). As we reported recently, this genomics approach has been employed to discover novel biomarkers of benzene toxicity in yeast cells exposed to the major active metabolites of benzene, hydroquinone (HQ), catechol (CAT), and 1,2,4-benzenetriol (BT) [44]. Using a collection of yeast strains representing a complete set of non-essential gene deletions, genetically tagged so that individual strains can be identified in competitive growth experiments, fitness assays were performed to identify mutant strains whose fitness is significantly altered following treatment with benzene metabolites. A comparison of the global deletome profiles of the metabolites revealed that deletion of certain genes rendered yeast cells sensitive to all three compounds. Several of the genes identified in the yeast studies have human orthologs with conserved biological function, supporting the notion that the mechanisms of toxicity identified in yeast are relevant to human disease.
Selection of homologous human genes by bioinformatics
To select and prioritize likely human candidate genes from complex yeast genomic data, we applied bioinformatic analyses using specific computational programs including “clustering” by HOPACH (Hierarchical Ordered Partitioning And Collapsing Hybrid) algorithm methods [45], pathway analysis using Cytoscape with the BiNGO Gene Ontology identification plugin [46, 47] and a comparative genomics approach. It is suggested that toxicants of similar mechanisms of action most likely have similar profiles of genes required for tolerance. We thus employed a variety of computational “clustering” methods to analyze the sensitivity and resistance data and to identify yeast strains most sensitive to each of the benzene metabolites tested as well as the genes specifically involved in sensitivity to each toxicant. We also set out to identify biologically significant patterns and features involved in toxicant response between metabolites. Similarly, a comparative genomics approach has been applied to identify functional orthologs and pathways between evolutionarily distant organisms. This approach assists in the identification of yeast and subsequently human candidate genes and pathways for further evaluation in the mammalian cell culture system.
Functional testing of candidate genes by RNAi in human cells
The roles of the human homologs of selected genes in benzene toxicity have been examined through mechanistic studies in human cell lines. SNPs in WRN, an important protein that plays a role in the maintenance of genomic stability, have been associated with an increased risk for some cancers and benzene hematotoxicity. We knocked down WRN protein using siRNA in HeLa cells and examined sensitivity to toxicity following exposure to the benzene metabolite, HQ [48]. Depletion of WRN led to decreased cell proliferation and increased HQ cytotoxicity, evident by increased necrosis. Additionally, these cells displayed increased DNA double-strand breaks (DSB), a potential biomarker of benzene hematotoxicity. Together, the results showed that WRN plays an important role in resistance to benzene toxicity in HeLa, and perhaps other cells.
More recently, we used shRNA to silence WRN in the human HL60 acute promyelocytic cell line [49]. Upon exposure to HQ, HL60 cell growth rates were accelerated, and DNA breaks and sensitivity to HQ-induced cytotoxicity and genotoxicity were increased, similar to the findings in HeLa cells. Loss of WRN also resulted in higher levels of early apoptosis. An accumulation of such genetic lesions can lead to the development of AML. The data from this study provides mechanistic support for the link between WRN and benzene-induced hematotoxicity, and possibly even benzene-induced leukemia. Studies are underway to confirm the role of other susceptibility genes in benzene toxicity.
Genotyping results in human population studies
In addition to the yeast genomic studies, large-scale human population genotyping studies have also been conducted. In collaboration between the National Cancer Institute and China CDC, such a study analyzed 1,395 SNPs in 411 potential carcinogenesis-related genes using an Illumina GoldenGate assay in 250 benzene-exposed workers and 140 unexposed controls in China [50]. One or more SNPs in five genes (WRN, BLM, TP53, RAD51, and WDR79) which play critical roles in DNA repair and genomic maintenance, were associated with highly significant 10–20% reductions (p values ranged from 0.0011 to 0.0002) in the WBC count among benzene-exposed workers but not controls, with evidence for gene-environment interactions for SNPs in BLM, WRN and RAD51. Earlier candidate gene studies identified a small number of SNPs in genes involved in benzene metabolism, cytokine and cellular adhesion molecules [51], and DNA DBS repair [52], which appear to confer susceptibility to benzene hematotoxicity. These studies were the first to provide evidence that genetic polymorphisms in certain genomic stability maintenance genes, like WRN [53], impact benzene-induced phenotypic outcomes such as hematotoxicity.
The human population studies confirm a critical role for DNA repair and genomic maintenance in susceptibility, and further support these effects as benzene-induced phenotypic outcomes. In addition, genetic variants in metabolizing enzymes responsible for activating and detoxifying benzene, in particular MPO and NQO1, have also been linked to increased susceptibility to benzene hematotoxicity [27]. Together, these genomic and genotyping studies provide important information regarding benzene toxicity and disease pathways.
From our omic studies to date, benzene appears to cause hematotoxicity through multiple mechanisms that may involve alterations in the expression of multiple genes and proteins, DNA methylation patterns and miRNA profiles even at low-doses. Transcriptomics has identified many genes, functions and pathways altered by benzene, offering insight into mechanisms and providing potential signatures of benzene exposure, and/or early effect. These data could be integrated with information on susceptibility genes to further understand gene-environment interactions and perhaps to identify the most susceptible individuals. We will expand our proteomic studies by conducting further analyses on different protein fractions and affinity chips to identify more altered proteins. As discussed earlier, expanded DNA methylation and miRNA profiling studies are necessary in larger populations of exposed individuals. All studies will be performed at a range of benzene exposures to examine dose-response effects. Several of these individual omic datasets are large, measuring e.g. the expression level of ~24,000 genes or the methylation status of >14,000 genes at multiple CpG islands and bioinformatic methods to analyze them continue to be refined.
Each individual omic dataset is anticipated to provide information on the effects of benzene, and potentially identify biomarkers of exposure and early effect. Through our systems biology approach, we will use sophisticated bioinformatics to integrate individual datasets into one consolidated model of the perturbations effected by benzene. From this model we will make inferences, specifically, we will aim to understand the interactions between benzene toxicity, SNPs, mRNA, miRNA, protein, and DNA methylation. This could identify robust biomarkers and help to clarify the molecular and cellular networks impacted by benzene, yielding a more comprehensive understanding of the mechanistic effects of benzene. While all the toxicogenomic endpoints have the potential to yield biomarkers, some endpoints such as DNA methylation and gene expression may reflect more upstream mechanistic effects while others such as proteomics may reflect more downstream, phenotypic effects, and might be more informative of actual molecular and cellular processes affected. Multiple regulatory mechanisms probably determine the phenotypic outcome (e.g. a gene could be up-regulated by DNA methylation and down-regulated by miRNA). The systems biology approach will require a high level of computing power and will capitalize on the ever-expanding knowledge of biological pathways and networks. In order to realize this approach, future studies need to be designed with sufficient power to robustly detect effects of benzene and to allow for analysis of the interrelationship among the different endpoints. With respect to sample processing, the ideal scenario is to analyze DNA, RNA and protein from the same cell population but this remains challenging with existing protocols and sample availability.
The findings from the systems biology study of benzene could also contribute more generally to the field of risk assessment. Comparison of the toxicogenomic, epigenomic and genomic profiles associated with different exposures, e.g. suspected leukemogens or carcinogens, and diseases (NHL, AML), may help to clarify the connection between chemicals, genes/proteins, pathways/networks, and disease. Initiatives such as the Comparative Toxicogenomics Database [54] and Chemical Effects in Biological Systems [55] have been developed towards this goal. It has been also discussed how omic data/measurements obtained through a systems biology approach can be applied to identify all potential mechanisms of action and serve as an information base for subsequent evaluation of these mechanisms when conducting risk assessment [56].
In conclusion, the systems biology approach described here should help inform the mechanisms underlying benzene hematoxicity and associated disease, and identify robust biomarkers of exposure, early effect, susceptibility and disease development.
Acknowledgments
We thank the participants for taking part in this study. We are extremely grateful to Heather Lustig and Kshama Mehta at Agilent Technologies, for facilitating the miRNA pilot study. We thank Ms. Suzanne May for her technical assistance with the DNA methylation analysis. We are grateful to Ms. Anh Duong for her tireless contribution and editorial assistance to this manuscript, and to Mr. Patrick Chang for his technical talent to illustrate the Systems Biology figure. This work was supported by NIH grants RO1ES06721 and P42ES04705 (MTS), P42ES05948 and P30ES10126 (SMR). MTS has received consulting and expert testimony fees from law firms representing both plaintiffs and defendants in cases involving exposure to benzene. GL has been a consultant for the American Petroleum Institute on benzene-related health research.
Abbreviations
AMLAcute Myeloid Leukemia
BFU-EBurst Forming Unit – Erythroid
BiNGOBiological Network Gene Ontology
BT1,2,4-Benzenetriol
CATCatechol
CBCComplete Blood Counts
CFU-GEMMColony Forming Unit – Granulocyte, Erythrocyte, Monocyte, Megakaryocyte
CFU-GMColony Forming Unit – Granulocyte, Monocyte
CYP2E1Cytochrome P450 2E1
GWASGenome-Wide Association Studies
HQHydroquinone
IARCInternational Agency for Research on Cancer
MDSMyelodysplastic Syndromes
miRNAmicroRNA
MPOMyeloperoxidase
NHLNon-Hodgkin Lymphoma
NQO1NAD(P): HQuinone Oxidoreductase 1
PBMCPeripheral Blood Mononuclear Cells
PF4Platelet Factor 4
q-PCRQuantitative-Polymerase Chain Reaction
RNAiRNA Interference
SELDI-TOF-MSSurface-Enhanced Laser Desorption/Ionization Time-Of-Flight Mass Spectrometry
shRNAshort hairpin RNA
siRNAsmall interfering RNA
SNPsSingle-Nucleotide Polymorphism
t-MDS/AMLtherapy-related MDS and AML
WBCWhite Blood Cells

Footnotes
For submission to: Chemico-Biological Interactions
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1. Kirschner MW. The meaning of systems biology. Cell. 2005;121:503–504. [PubMed]
2. Kitano H. Systems biology: a brief overview. Science (New York, NY) 2002;295:1662–1664. [PubMed]
3. McHale C, Zhang L, Hubbard AE, Smith M. Toxicogenomic Studies in Human Populations. In: Boverhof DR, Gollapudi BB, editors. Applications of Toxicogenomics in Safety Evaluation and Risk Assessment. John Wiley & Sons, Inc; Hoboken, New Jersey: 2009.
4. Smith M, Rappaport S. Building Exposure Biology Centers to Put the E into “G × E” Interaction Studies. Environmental health perspectives. 2009;117:A334–A335. [PMC free article] [PubMed]
5. Vlaanderen J, MLE, Smith MT, et al. Application of omics technologies in occuptional and environmental health research: Current status and projections. Occupational and Environmental Medicine. 2009
6. Bayat A. Science, medicine, and the future: Bioinformatics. BMJ. 2002;324:1018–1022. [PMC free article] [PubMed]
7. NRC, National Research Council. Committee on Toxicity Testing and Assessment of Environmental Agents. National Academies Press; Washington DC: 2007. Toxicity Testing in the 21st Century: A Vision and a Strategy.
8. Glass DC, Gray CN, Jolley DJ, et al. Leukemia risk associated with low-level benzene exposure. Epidemiology (Cambridge, Mass) 2003;14:569–577. [PubMed]
9. Hayes RB, Songnian Y, Dosemeci M, Linet M. Benzene and lymphohematopoietic malignancies in humans. American journal of industrial medicine. 2001;40:117–126. [PubMed]
10. Savitz DA, Andrews KW. Review of epidemiologic evidence on benzene and lymphatic and hematopoietic cancers. American journal of industrial medicine. 1997;31:287–295. [PubMed]
11. Smith MT, Jones RM, Smith AH. Benzene exposure and risk of non-Hodgkin lymphoma. Cancer Epidemiol Biomarkers Prev. 2007;16:385–391. [PubMed]
12. Zhang L, Eastmond DA, Smith MT. The nature of chromosomal aberrations detected in humans exposed to benzene. Critical reviews in toxicology. 2002;32:1–42. [PubMed]
13. IARC, International Agency for Research on Cancer. Some Industrial Chemicals and Dyestuffs. Monogr Eval Carcinog Risks Hum. 1982
14. IARC, International Agency for Research on Cancer. Monographs on the Evaluation of Carcinogenic Risks to Humans Overall Evaluations of Carcinogenicity. An Updating of IARC Monographs. 1987 [PubMed]
15. Smith MT. The mechanism of benzene-induced leukemia: a hypothesis and speculations on the causes of leukemia. Environmental health perspectives. 1996;104(Suppl 6):1219–1225. [PMC free article] [PubMed]
16. Ross D, Siegel D, Schattenberg DG, Sun XM, Moran JL. Cell-specific activation and detoxification of benzene metabolites in mouse and human bone marrow: identification of target cells and a potential role for modulation of apoptosis in benzene toxicity. Environmental health perspectives. 1996;104(Suppl 6):1177–1182. [PMC free article] [PubMed]
17. Snyder R, Hedli CC. An overview of benzene metabolism. Environmental health perspectives. 1996;104(Suppl 6):1165–1171. [PMC free article] [PubMed]
18. Subrahmanyam VV, Ross D, Eastmond DA, Smith MT. Potential role of free radicals in benzene-induced myelotoxicity and leukemia. Free radical biology & medicine. 1991;11:495–515. [PubMed]
19. Zhang L, Smith M, Bandy B, Tamaki SJ, Davison AJ. Role of Quinones, active oxygen species and metals in the genotoxicity of 1,2,4-benzenetriol, a metabolite of benzene. In: Nhol H, Esterbauer H, Rice-Evans C, editors. Free Radicals in the environment, medicine and toxicology. Richelieu; London: 1994. pp. 521–562.
20. Cho JY. Suppressive effect of hydroquinone, a benzene metabolite, on in vitro inflammatory responses mediated by macrophages, monocytes, and lymphocytes. Mediators Inflamm. 2008;2008:298010. [PMC free article] [PubMed]
21. Ouyang Y, Virasch N, Hao P, et al. Suppression of human IL-1beta, IL-2, IFN-gamma, and TNF-alpha production by cigarette smoke extracts. J Allergy Clin Immunol. 2000;106:280–287. [PubMed]
22. Yoon BI, Hirabayashi Y, Kawasaki Y, et al. Aryl hydrocarbon receptor mediates benzene-induced hematotoxicity. Toxicol Sci. 2002;70:150–156. [PubMed]
23. Pedersen-Bjergaard J, Andersen MK, Andersen MT, Christiansen DH. Genetics of therapy-related myelodysplasia and acute myeloid leukemia. Leukemia. 2008;22:240–248. [PubMed]
24. Zhang L, Rothman N, Wang Y, et al. Increased aneusomy and long arm deletion of chromosomes 5 and 7 in the lymphocytes of Chinese workers exposed to benzene. Carcinogenesis. 1998;19:1955–1961. [PubMed]
25. Ji Z, Zhang L, Guo W, McHale CM, Smith MT. The benzene metabolite, hydroquinone and etoposide both induce endoreduplication in human lymphoblastoid TK6 cells. Mutagenesis. 2009;24:367–372. [PMC free article] [PubMed]
26. Smith MT, Zhang L, Wang Y, et al. Increased translocations and aneusomy in chromosomes 8 and 21 among workers exposed to benzene. Cancer research. 1998;58:2176–2181. [PubMed]
27. Lan Q, Zhang L, Li G, et al. Hematotoxicity in workers exposed to low levels of benzene. Science (New York, NY) 2004;306:1774–1776. [PMC free article] [PubMed]
28. Lan Q, Vermeulen R, Zhang L, et al. Benzene Exposure and Hematotoxicity. Science (New York, NY) 2006;312:998–999. [PubMed]
29. Forrest MS, Lan Q, Hubbard AE, et al. Discovery of novel biomarkers by microarray analysis of peripheral blood mononuclear cell gene expression in benzene-exposed workers. Environmental health perspectives. 2005;113:801–807. [PMC free article] [PubMed]
30. McHale CM, Zhang L, Lan Q, et al. Changes in the peripheral blood transcriptome associated with occupational benzene exposure identified by cross-comparison on two microarray platforms. Genomics. 2009;93:343–349. [PMC free article] [PubMed]
31. McHale C, Zhang L, Lan Q, et al. Low-dose, occupational exposure to the leukemogen benzene induces robust changes in the blood transcriptome associated with altered immune system biology. American Society of Hematology; San Francisco, CA: 2008.
32. Vermeulen R, Lan Q, Zhang L, et al. Decreased levels of CXC-chemokines in serum of benzene-exposed workers identified by array-based proteomics. Proceedings of the National Academy of Sciences of the United States of America. 2005;102:17041–17046. [PubMed]
33. Dolinoy DC, Jirtle RL. Environmental epigenomics in human health and disease. Environmental and molecular mutagenesis. 2008;49:4–8. [PubMed]
34. Jirtle RL, Skinner MK. Environmental epigenomics and disease susceptibility. Nature reviews. 2007;8:253–262. [PubMed]
35. Bollati V, Baccarelli A, Hou L, et al. Changes in DNA methylation patterns in subjects exposed to low-dose benzene. Cancer research. 2007;67:876–880. [PubMed]
36. Muller CI, Luong QT, Shih LY, et al. Identification of marker genes including RUNX3 (AML2) that discriminate between different myeloproliferative neoplasms and normal individuals. Leukemia. 2008;22:1773–1778. [PubMed]
37. Neufeld G, Shraga-Heled N, Lange T, et al. Semaphorins in cancer. Front Biosci. 2005;10:751–760. [PubMed]
38. Kim E, Kang BY, Kim TS. Inhibition of interleukin-12 production in mouse macrophages by hydroquinone, a reactive metabolite of benzene, via suppression of nuclear factor-kappaB binding activity. Immunol Lett. 2005;99:24–29. [PubMed]
39. Paschka P. Core binding factor acute myeloid leukemia. Semin Oncol. 2008;35:410–417. [PubMed]
40. Dixon-McIver A, East P, Mein CA, et al. Distinctive patterns of microRNA expression associated with karyotype in acute myeloid leukaemia. PloS one. 2008;3:e2141. [PMC free article] [PubMed]
41. Skibola CF, Bracci PM, Halperin E, et al. Genetic variants at 6p21.33 are associated with susceptibility to follicular lymphoma. Nature genetics. 2009;41:873–875. [PMC free article] [PubMed]
42. Knight JA, Skol AD, Shinde A, et al. Genome-wide association study to identify novel loci associated with therapy-related myeloid leukemia susceptibility. Blood. 2009;113:5575–5582. [PubMed]
43. Crowley JJ, Sullivan PF, McLeod HL. Pharmacogenomic genome-wide association studies: lessons learned thus far. Pharmacogenomics. 2009;10:161–163. [PMC free article] [PubMed]
44. North M, Shuga J, Loguinov A, et al. Utilizing functional genomics in yeast to discover novel biomarkers of benzene toxicity in humans. International Symposium on Benzene Health Effects and Mechanisms of Bone Marrow Toxicity; Munich, Germany. 2009.
45. Van der Laan MJ, Pollard KS. A new algorithm for hybrid hierarchial clustering with visualization and the bootstrap. Journal of Statistical Planning and Inference. 2003:275–303.
46. Cline MS, Smoot M, Cerami E, et al. Integration of biological networks and gene expression data using Cytoscape. Nature protocols. 2007;2:2366–2382. [PMC free article] [PubMed]
47. Maere S, Heymans K, Kuiper M. BiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks. Bioinformatics (Oxford, England) 2005;21:3448–3449. [PubMed]
48. Galvan N, Lim S, Zmugg S, Smith MT, Zhang L. Depletion of WRN enhances DNA damage in HeLa cells exposed to the benzene metabolite, hydroquinone. Mutation research. 2008;649:54–61. [PMC free article] [PubMed]
49. Ren X, Lim S, Smith MT, Zhang L. Werner syndrome protein, WRN, protects cells from DNA damage induced by the benzene metabolite hydroquinone. Toxicol Sci. 2009;107:367–375. [PMC free article] [PubMed]
50. Lan Q, Zhang L, Shen M, et al. Large-scale evaluation of candidate genes identifies associations between DNA repair and genomic maintenance and development of benzene hematotoxicity. Carcinogenesis. 2009;30:50–58. [PMC free article] [PubMed]
51. Lan Q, Zhang L, Shen M, et al. Polymorphisms in cytokine and cellular adhesion molecule genes and susceptibility to hematotoxicity among workers exposed to benzene. Cancer research. 2005;65:9574–9581. [PubMed]
52. Shen M, Lan Q, Zhang L, et al. Polymorphisms in genes involved in DNA double-strand break repair pathway and susceptibility to benzene-induced hematotoxicity. Carcinogenesis. 2006;27:2083–2089. [PubMed]
53. Singh DK, Ahn B, Bohr VA. Roles of RECQ helicases in recombination based DNA repair, genomic stability and aging. Biogerontology. 2009;10:235–252. [PMC free article] [PubMed]
54. Mattingly CJ, Rosenstein MC, Colby GT, Forrest JN, Jr, Boyer JL. The Comparative Toxicogenomics Database (CTD): a resource for comparative toxicological studies. J Exp Zoolog A Comp Exp Biol. 2006;305:689–692. [PMC free article] [PubMed]
55. Waters M, Stasiewicz S, Merrick BA, et al. CEBS--Chemical Effects in Biological Systems: a public data repository integrating study design and toxicity data with microarray and proteomics data. Nucleic Acids Res. 2008;36:D892–900. [PMC free article] [PubMed]
56. Edwards SW, Preston RJ. Systems biology and mode of action based risk assessment. Toxicol Sci. 2008;106:312–318. [PubMed]