Sound Science and Junk Science in the Proof of Causation: The Example of Benzene-Induced Leukemia

by Raphael Metzger, Esq.

Introduction

In the last decade there have been breathtaking advances in science that are rapidly becoming important for the proof of causation in toxic tort cases.  Whole new fields of science have sprung to life, with such exotic names as genomics, proteomics, and metabolomics, each of which has its own specialty journals, conferences, and jargon.  These new fields of science are largely the result of technological advances developed from the Human Genome Project.  These technologies provide virtually unlimited opportunities to generate huge amounts of data regarding the toxic effects of chemicals on the approximately 25,000 genes in the human genome.  Gene microarrays now enable scientists to determine the effects of chemicals on thousands of genes simultaneously.  Scientists working in these fields are presented with a substantial problem: how to evaluate and make sense of the huge amounts of data being generated.  Statisticians, using sophisticated methods of data analysis are being recruited to help the scientists manage all the data.  Data generated from studies using these new methods, coupled with data obtained from cytogenetic and immunophenotypic studies (the new sciences of the last decade), provide great promise for proving or disproving causation in toxic tort cases by objective means.  They also present great opportunities for mischief.

Most scientists using the new scientific technologies are proceeding cautiously, publishing their results without making broad claims regarding their import and effect until the limitations of the technologies and the data are better understood.  However, some scientists envision fool-proof methods for applying the data to prove or disprove causation and are seeking to capitalize on the legal market’s desire for certainty in a field that has thus far been fraught with uncertainty.  How are we in the legal profession to ascertain what is the good science and what is the junk science?

In this paper I will provide two examples of scientific thought based on the new technologies as they relate to benzene-induced leukemia.  One of the examples is based on years of experimentation and publication in the field of molecular epidemiology by a renowned scientist who has spent his career studying the effects of benzene on human chromosomes and genes.  The other example is based on the work of a physician who claims to have developed a fool-proof test for determining whether benzene exposure caused a particular worker’s leukemia.  I will explain why the methods applied by the renowned scientist constitute sound science, while the test marketed by the physician is either junk science based on a lack of understanding of science and  limitations of the new technologies, or is chicanery advocated by a charlatan hoping to make a quick buck by taking advantage of well-intentioned claims representatives and attorneys looking for sure answers to difficult questions, who are being misled by the supposed infallibility of the new technologies.

Benzene

Benzene is a known human carcinogen.  It is also a clastogen, which means that it has the ability to break chromosomes.  Many epidemiology studies have been conducted of benzene-exposed workers and it is universally understood and agreed that exposure to benzene can cause Acute Myelogenous Leukemia.  It is widely believed that benzene causes Acute Myelogenous Leukemia by the interaction of certain of its metabolites on chromosomes and on chromosome-repair enzymes.  However, benzene is not the only cause of Acute Myelogenous Leukemia.  Other causes have been ascertained including ionizing radiation, alkylating agents, topoisomerase inhibiting drugs, other chemicals such as 1,3-butadiene, and cigarette smoke.  It is also likely that other chemicals cause Acute Myelogenous Leukemia, but have not yet been proven to do so.  In any case where a worker exposed to benzene develops Acute Myelogenous Leukemia, the question arises: Did occupational exposure to benzene cause the worker’s leukemia, or was the cancer caused by something else?  Studies regarding the effects of benzene’s metabolites on chromosomes and specific genes have been and continue to be conducted to help answer this question.

Professor Martyn Smith’s Research

Martyn Smith is Professor of Toxicology at the University of California at Berkeley where he directs the Molecular Epidemiology and Toxicology Laboratory.  His laboratory has studied the effects of benzene and its metabolites on chromosomes for more than a decade, resulting in many publications in the scientific peer-reviewed literature.  Professor Smith has identified particular enzymes that metabolize benzene in the liver and has studied the mechanisms by which these metabolites induce chromosomal abnormalities and cause Acute Myelogenous Leukemia.  Dr. Smith’s research has shown that the metabolism of benzene is complex, that benzene’s metabolites induce various types of damage in multiple different chromosomes, and that benzene’s metabolites alter gene expression, but that gene expression appears also to be related to stage of disease and particular subtypes of Acute Myelogenous Leukemia.  Professor Smith has concluded that benzene (through its reactive metabolites) can cause multiple types of genetic damage, but that due to the complex and multiple mechanisms of injury involved, there is no single “fingerprint” of benzene-induced damage to chromosomes or to specific genes in the human genome.  In evaluating whether a worker's leukemia is caused by benzene exposure, Professor Smith carefully evaluates whether the worker’s chromosomes have damage that has been shown to be caused by benzene’s metabolites as well as other relevant information (including dose) in assessing whether a worker’s Acute Myelogenous Leukemia was likely caused by occupational exposure to benzene.  Professor Smith eschews the notion that there is or can be a definitive, infallible answer to the causation question.

In a recent case, Professor Smith has provided testimony supportive of causation of benzene-induced leukemia where the worker developed a particular chromosome translocation.  A chromosome translocation occurs when a chromosome breaks into two or more pieces and, in an unsuccessful effort to repair the damage, pieces of the damaged chromosome reassemble improperly.  Consequentially, one or more genes are fractured so that they may not perform their usual functions.  When these genes regulate cell growth, division or proliferation, or when they regulate apoptosis (programmed cell death), leukemia may result.

Dr. Smith explained that since benzene is clastogenic, it has the ability to break chromosomes and therefore induce chromosome translocations.  Dr. Smith’s research has also shown that certain reactive metabolites of benzene inhibit a particular DNA repair enzyme known as Topoisomerase II, and that chemicals which inhibit the functionality of this enzyme induce particular chromosome translocations, including t(15;17), the characteristic translocation of Acute Promyelocytic Leukemia, a subtype of Acute Myelogenous Leukemia.

Topoisomerase II (topo II) is an enzyme that is essential for the maintenance of proper chromosome structure and segregation; it removes knots and tangles from genetic material by passing an intact double helix through a transient double-stranded break that it creates in a separate segment of DNA.  To maintain genomic integrity during its catalytic cycle, topo II forms covalent bonds between active-site tyrosyl residues and the 5'-DNA termini created by cleavage of the double helix.  Normally, these covalent topo II-cleaved DNA complexes (known as cleavable complexes) are fleeting intermediates and are tolerated by the cell.  However, when the concentration or longevity of cleavage complexes increases significantly, DNA strand breaks and other genotoxic events occur.  Lindsey, R. H., et al., “1,4-Benzoquinoine Is a Topoisomerase II Poison,” Biochemistry 43(2):7563-7574 (2004) at p. 7564.

A variety of widely prescribed anticancer drugs, such as etoposide and mitoxantrone, kill cells by inhibiting topo II.  These drugs are referred to as topo II inhibitors.

Several studies have shown that topo II inhibitors induce the t(15;17) translocation which is the characteristic feature of Acute Promyelocytic Leukemia.  See Andersen, M.K., et al., “Chromosomal Abnormalities in Secondary MDS and AML: Relationship to Drugs and Radiation with Specific Emphasis on the Balanced Rearrangements,” Haematologica 83:438-488 (1998); Andersen, M. K., et al., “Balanced Chromosome Abnormalities inv(16) and t(15;17) in Therapy-Related Myelodysplastic Syndromes and Acute Leukemia: Report from an International Workshop,” Genes, Chromosomes & Cancer 33:395-400 (2002); Anonymous, “Chemotherapy-related Acute Promyelocytic Leukemia: Role of Topoisomerase II,” Nat. Clin. Pract. Oncol. 2(6):282 (2005); Beaumont, M., et al., “Therapy-related Acute Promyelocytic Leukemia,” J. Clin. Oncol. 21(11):2123-2137 (2003); Bhavnani, M., et al., “Therapy-Related Acute Promyelocytic Leukaemia,” Brit. J. Haematol. 86(1):231-232 (1994); Carli, P.M., et al., “Increase Therapy-related Leukemia Secondary to Breast Cancer,” Leukemia 14(6):1014-1017 (2000); Di Renzo, A., et al., “Acute Promyelocytic Leukemia After Treatment for non-Hodgkin’s Lymphoma with Drugs Targeting Topoisomerase II,” Amer. J. Hematol. 60(4):300-304 (1999); Detourmignies, L., et al., “Therapy-related Acute Promyelocytic Leukemia: A Report on 16 Cases,” J. Clin. Oncol. 10(9):1430-1435 (1992); Felix, C.A., “Secondary Leukemias Induced by Topoisomerase-Targeted Drugs,” Biochim. Biophys. Acta 1400(1-3);233-255 (1998); Felix, C.A., “Leukemias Related to Treatment with DNA Topoisomerase II Inhibitors,” Medical and Pediatric Oncology 36:525-535 (2001); Fenaux, P., et al., “Therapy-Related Acute Promyelocytic Leukaemia,” Brit. J. Haematol. 87(2):445 (1994); Gillis, S., et al., “Acute Promyelocytic Leukaemia with t(15;17) following Treatment of Hodgkin’s Disease: A Report of 4 Cases,” Ann. Oncol. 6(8):777-779 (1995); Hoffman, L., et al., “Therapy-Related Acute Promyelocytic Leukemia with t(15;17)(q22;q12) Following Chemotherapy with Drugs Targeting DNA Topoisomerase II: A Report of Two Cases and a Review of the Literature,” Ann. Oncol. 6(8):781-788 (1995); Kroger, N., et al., “Secondary Acute Leukemia Following Mitoxantrone-Based High-Dose Chemotherapy for Primary Breast Cancer Patients,” Bone Marrow Transplantation 32:1153-1157 (2003); Kudo, K., et al., “Etoposide-related Acute Promyelocytic Leukemia,” Leukemia 12(8);1171-1175 (1998); Mistry, A.R., et al., “DNA Topoisomerase II in Therapy-Related Acute Promyelocytic Leukemia,” New Engl. J. Med. 352(15):1529-1538 (2005); Ogami, A., et al., “Secondary Acute Promyelocytic Leukemia Following Chemotherapy for non-Hodgkin’s Lymphoma in a Child,” J. Pediatr. Hematol. Oncol. 26(7):427-430 (2004); Pedersen-Bjergaard, J., et al., “The Balanced and the Unbalanced Chromosome Aberrations of Acute Myeloid Leukemia May Develop in Different Ways and May Contribute Differently to Malignant Transformation,” Blood 83:2780-2786 (1994); Pedersen, Bjergaard, J., “Acute Promyelocytic Leukemia with t(15;17) Following Inhibition of DNA Topoisomerase II,” Ann. Oncol. 6(8):751-753 (1995); Pedersen-Bjergaard, J., et al., “Genetic Pathways in Therapy-Related Myelodysplasia and Acute Myeloid Leukemia,” Blood 99(6):1909-1912 (2002).

These studies establish that etoposide, mitoxantrone, doxorubicin, and other topo II inhibiting drugs cause Acute Promyelocytic Leukemia.  The mechanism by which these drugs induce Acute Promyelocytic Leukemia has recently been clarified in a seminal article in the New England Journal of Medicine.  Mistry and co-workers showed that translocation breakpoints in Acute Promyelocytic Leukemia that developed after exposure to mitoxantrone, a topo II inhibitor, were rightly clustered in a specific region of the PML gene.  In functional assays, this “hot spot” and the corresponding RARA breakpoints were common sites of mitoxantrone-induced cleavage by topo II.  Etoposide and doxorubicin also induced cleavage by topoisomerase II at the translocation breakpoints in Acute Promyelocytic Leukemia arising after exposure to these agents.  Thus, drug-induced cleavage of DNA by topo II mediates the formation of specific non-random chromosomal translocation breakpoints in mitoxantrone-related Acute Promyelocytic Leukemia and in Acute Promyelocytic Leukemia that occurs after therapy with other topo II poisons.  Mistry, A.R., et al., “DNA topoisomerase II in therapy-related acute promyelocytic leukemia,” New Engl. J. Med. 352:1529-1538 (2005).

The mechanism by which benzene induces leukemias has not been fully elucidated, but it is known that benzene is first metabolized to benzene oxide in the liver by cytochrome P450 2E1.  While some of the oxide is cleared by conjugation to glutathione, a large proportion is converted to phenol by a nonenzymatic rearrangement.  Phenol is further metabolized by cytochrome P450 2E1 to 1,4-hydroquinone, which is carried throughout the body in the bloodstream.  When transported to the bone marrow, 1,4-hydroquinone ultimately is converted to 1,4-benzoquinone by the high concentration of endogenous myeloperoxidase in the marrow.  Thus, 1,4-benzoquinone is thought to be a critical leukemogenic metabolite of benzene.  Smith, M.T., “The Mechanism of Benzene-Induced Leukemia: A Hypothesis and Speculations on the Causes of Leukemia,” Environ. Health Perspect. 104 (Suppl. 6):1219-1225 (1996); Whysner, J., et al., “Genotoxicity of Benzene and its Metabolites,” Mutation Res. 566:99-130 (2004).

Several studies have shown that reactive benzene metabolites inhibit topo II.  See, e.g., Chen, H., et al., “Topoisomerase Inhibition by Phenolic Metabolites: A Potential Mechanism for Benzene’s Clastogenic Effects,” Carcinogenesis 16(10):2301-2307 (1995); Hutt, A. M., et al., “Inhibition of Human DNA Topoisomerase II by Hydroquinone and p-Benzoquinone, Reactive Metabolites of Benzene,” Environ. Health Perspect. 104(Suppl. 6):1265-1269 (1996); Frantz, C. E., et al., “Inhibition of Human Topoisomerase II in vitro by Bioactive Benzene Metabolites,” Environ. Health Perspect. 104(Suppl. 6):1319-1323 (1996); Eastmond, D. A., et al., “Characterization and Mechanisms of Chromosomal Alterations Induced by Benzene in Mice and Humans,” Health Effects Institute Research Report 103 (2001); Baker, R. K., et al., “Benzene Metabolites Antagonize Etoposide-Stabilized Cleavable Complexes of DNA Topoisomerase II α,” Blood 93(3):830-833 (2001); Hoffmann, M. J., et al., “The Potential Role of Topoisomerase II Inhibition in Hydroquinone-Induced Alterations in the Maturation of Mouse Myeloblasts,” Adv. Exp. Med. Biol. 500:315-318 (2001); Fung, J., et al., “Inhibition of Topoisomerase II in 32D.3(G) Cells by Hydroquinone is Associated with Cell Death,” J. Appl. Toxicol. 24:183-188 (2004); Hasinoff, B. B., et al., “Structure-Activity Study of the Interaction of Bioreductive Benzoquinone Alkylating Agents with DNA Topoisomerase II,” Cancer Chemother. Pharmacol. 57(2):221-233 (2006).

Studies have shown that the benzene metabolites 1,4-hydroquinone and 1,4-benzoquinone are also poisons of topo II, enhancing DNA cleavage mediated by human topoisomerase II.  See, e.g., Lindsey, R. H., et al., “1,4-Benzoquinone Is a Topoisomerase II Poison,” Biochemistry 43(2):7563-7574 (2004); Lindsey, R. H., et al., “Stimulation of Topoisomerase II-Mediated DNA Cleavage by Benzene Metabolites,” Chem. Biol. Interact. 153-154: 197-205 (2005); Lindsey, R. H., et al., “Effects of Benzene Metabolites on DNA Cleavage Mediated by Human Topoisomerase II:  1,4-Hydroquinone Is a Topoisomerase II Poison,” Chem. Res. Toxicol. 18:761-770 (2005).  It has therefore been proposed that topoisomerase II inhibition by myeloperoxidase-activated hydroquinone is the mechanism underlying the genotoxic and carcinogenic effects of benzene.  Eastmond, D.A., et al., “Topoisomerase II inhibition by meyloperoxidase-activated hydroquinone: a potential mechanism underlying the genotoxic and carcinogenic effects of benzene,” Chem. Biol. Interact. 153-154:207-216 (2005).

The above studies show that reactive metabolites of benzene inhibit the functions of topoisomerase II.  Since topoisomerase II inhibiting drugs not only cause Acute Myelogenous Leukemia and Acute Promyelocytic Leukemia specifically, through the induction of specific breaks in the genes involved in Acute Promyelocytic Leukemia, and since metabolites of benzene have been shown to be topoisomerase inhibitors, there is strong biological and mechanistic evidence that exposure to benzene causes Acute Promyelocytic Leukemia through its quinone metabolites via their ability to inhibit topo II and cause chromosome translocations involving the retinoic acid receptor-alpha (RARαa) gene on Chromosome 17 associated with Acute Promyelocytic Leukemia.

To summarize, Dr. Smith’s research has shown that benzene breaks chromosomes and induces chromosome translocations, that the t(15;17) translocation (the characteristic genetic feature of Acute Promyelocytic Leukemia) is induced by topoisomerase inhibitors, and that benzene’s reactive metabolites are topoisomerase inhibitors capable of inducing this chromosomal damage.  

Professor Smith therefore holds the opinion that the presence of the t(15;17) translocation in a worker with a substantial occupational exposure to benzene, who has not been prescribed radiotherapy or topoisomerase inhibiting drugs, and who does not have an appreciable smoking history, is supportive of benzene induction of his leukemia.  His opinion is based on experimental research findings and the synthesis of data regarding the metabolism of benzene and the mechanisms by which it induces chromosomal abnormalities, in particular the t(15;17) translocation through the inhibition of the topoisomerase II enzyme by benzene’s reactive metabolites.  Professor Smith’s opinion is based on sound science and is well supported by various lines of scientific evidence.  His opinion is not overreaching.  He does not opine that the leukemias of all benzene-exposed workers who have the t(15;17) translocation are caused by occupational exposure to benzene, but rather that this particular chromosomal translocation is supportive of induction of leukemia by benzene.

Dr. Bruce Gillis

Bruce Gillis is a Los Angeles physician who has a large workers’ compensation practice in which he is usually engaged by insurance companies to evaluate workers claiming that they have developed cancer and other diseases from occupational exposure to toxic chemicals.  Dr. Gillis claims to have developed a test using gene microarrays which is infallible in determining whether a worker’s leukemia was caused by exposure to benzene.  The test has thus far been used in workers’ compensation cases to deny benzene-exposed workers compensation benefits, although uninformed applicants’ attorneys have embraced the test as a means of proving whether their clients’ leukemias were caused by exposure to benzene.  Dr. Gillis claims that his test has been approved by the ABA and the ABA has recently sponsored a legal conference featuring Dr. Gillis and his benzene test.

Dr. Gillis’ test is based on a study that he and some researchers from his alma mater, the University of Illinois, recently published: Gillis, B., et al., “Identification of Human Cell Response to Benzene and Benzene Metabolites,” Genomics 90(3):324-333 (Sept. 2007).  In this article, his first publication in the field of genomics, Dr. Gillis makes some outlandish claims.  For example, he concludes that “the information presented in this study not only offers clarity for determining injurious exposures, but it also provides a methodology for eliminating incorrect diagnoses and conclusions following such an exposure [to benzene].”  Unfortunately, the article provides no data to support a claim of such breadth.  Indeed, such sweeping claims are inherently unscientific and can only be established when experimental data actually supporting them are presented and the experimental results are replicated by other scientific researchers.

In his study, Dr. Gillis examined cytokine production by, and gene expression changes in, human peripheral blood mononuclear cells upon their exposure to the benzene metabolites catechol, hydroquinone, 1,2,4-benzenetriol, and p-benzoquinone.  Based solely on these findings in human blood cells in petri dishes following short-term exposure to benzene metabolites, Dr. Gillis claims to have developed a test which is foolproof for determining whether a worker's cancer or other disease was caused by exposure to benzene!  This is simply ridiculous.

While Dr. Gillis may have identified certain cellular changes following short-term exposure to high levels of benzene metabolites, the study provides no data to prove that these effects have any correlation with benzene-induced leukemia or other disease.    Dr. Gillis has not evaluated gene expression changes in benzene-exposed workers who have developed Acute myelogenous leukemia or any other disease.  Gene expression following exposure to a toxic chemical is not static.  A disease like benzene-induced leukemia develops over a period of years, and the gene expression profile changes once the exposure to benzene ceases and the disease begins to develop.  Since Dr. Gillis has not examined gene expression patterns in workers with benzene-induced hematologic disease, his extrapolation that the immediate effects of benzene exposure in blood cells in petri dishes must be the same as in benzene-induced leukemia or other disease is unfounded speculation. 

The available literature reflects a different gene expression profile in benzene-induced disease and in benzene-exposed workers than that which Dr. Gillis has identified for cells exposed to benzene in petri dishes.  See Zhang, L., et al., “Microarray Analysis of Mononuclear Cell Gene Expression in Workers Exposed to Benzene,” Abstract No. 160, 9th International Conference on Environmental Mutagens (2005); Smith, M.T., et al., “Use of 'Omic' technologies to study humans exposed to benzene,” Chem. Biol. Interact. 153-154:123-127 (2005); Forrest, M.S., et al., “Discovery of novel biomarkers by microarray analysis of peripheral blood mononuclear cell gene expression in benzene-exposed workers,” Environ. Health Perspect. 113(6):801-807 (2005); Wang, H., et al., [cDNA microarray and cluster analysis to identify the significance of immune genes associated with benzene poisoning] Zhonghua Lao Dong Wei Sheng Zhi Ye Bing Za Zhi 23(4):260-262 (2005); Xia, Y., et al., [Analysis on tumor related gene expression profiles in benzene poisoning using cDNA microarray] Zhonghua Lao Dong Wei Sheng Zhi Ye Bing Za Zhi 23(4):256-259 (2005); Wang, H., et al., [Differentially expressed genes of cell signal transduction associated with benzene poisoning by cDNA microarray] Zhonghua Lao Dong Wei Sheng Zhi Ye Bing Za Zhi 23(4):252-255 (2005).

Dr. Gillis seems oblivious to the scientific flaws of his claims and has heavily marketed his new test to the insurance industry as a simple and definitive test for disproving workers’ compensation claims that benzene and other toxic industrial chemicals caused harm. Apparently, the test has been uncritically accepted by the workers’ compensation bar and has thus far been used to deny workers compensation benefits without scientific challenge.

To date, the test has apparently not been used in civil litigation, where, if challenged, it would not pass muster under either the Frye test applied in many state courts or the Daubert standard applicable in federal courts.  See Frye v. U.S. (D.C. Cir. 1923) 293 F. 1013; Daubert v. Merrell-Dow Pharmaceuticals, Inc. (1993) 509 U.S. 579.   When subjected to scientific challenge, Dr. Gillis’ test will be shown to be chicanery and, like so many scientific tests purporting to impart an infallible truth, will be cast upon the Frye rubbish heap.

Conclusion

Two proponents of recent scientific research regarding the causation of leukemia by occupational exposure to benzene are presented.  Each of the researchers bases his opinion on recent research regarding genetic damage induced by benzene.  However, one of the researcher’s conclusions are based on sound science supported by a substantial body of medical and scientific research, while the other researcher’s conclusions are based on a fundamental misunderstanding of gene expression profiles, limitations of the new genetic technologies, and unfounded speculation skewed by a substantial profit motive.  The new technologies developed by the Human Genome Project hold great promise for elucidating the causes of benzene-induced leukemias and other cancers caused by exposure to carcinogens.  However, the new technologies also offer substantial opportunities for fraud and chicanery due to their ostensible infallibility and lack of transparency. Our task is to distinguish which opinions based on the new technologies are sound and which are just more junk science.  To do this, we must first learn the basics of the new scientific technologies.

Raphael Metzger is the principal of the Metzger Law Group, a boutique firm specializing in toxic tort litigation located in Long Beach, California.  Mr. Metzger is the Co-Chair of the Benzene Litigation Group of the American Association for Justice and has chaired benzene litigation conferences sponsored by Mealey’s Publications and HarrisMartin publications.

This article was presented under the title “Gene Expression Technology as a Medical Causation Tool: Critique of Dr. Gillis’ MSDS1 Test,” at HarrisMartin’s Benzene Litigation Conference in New Orleans, Louisiana, June 2-3, 2008.