A possible role in chemical carcinogenesis for epigenetic, heritable changes in gene expression

DNA methylation changes from primary cultures through senescence-bypass in Syrian hamster fetal cells initially exposed to benzo[a]pyrene

Current chemical testing strategies are limited in their ability to detect non-genotoxic carcinogens (NGTxC). Epigenetic anomalies develop during carcinogenesis regardless of whether the molecular initiating event is associated with genotoxic (GTxC) or NGTxC events; therefore, epigenetic markers may be harnessed to develop new approach methodologies that improve the detection of both types of carcinogens. This study used Syrian hamster fetal cells to establish the chronology of carcinogen-induced DNA methylation changes from primary cells until senescence-bypass as an essential carcinogenic step. Cells exposed to solvent control for 7 days were compared to naïve primary cultures, to cells exposed for 7 days to benzo[a]pyrene, and to cells at the subsequent transformation stages: normal colonies, morphologically transformed colonies, senescence, senescence-bypass, and sustained proliferation in vitro. DNA methylation changes identified by reduced representation bisulphite sequencing were minimal at day-7. Profound DNA methylation changes arose during cellular senescence and some of these early differentially methylated regions (DMRs) were preserved through the final sustained proliferation stage. A set of these DMRs (e.g., Pou4f1, Aifm3, B3galnt2, Bhlhe22, Gja8, Klf17, and L1l) were validated by pyrosequencing and their reproducibility was confirmed across multiple clones obtained from a different laboratory. These DNA methylation changes could serve as biomarkers to enhance objectivity and mechanistic understanding of cell transformation and could be used to predict senescence-bypass and chemical carcinogenicity.

On the intrinsic inevitability of cancer: from foetal to fatal attraction

The cracks in the paradigm of oncogenic mutations and somatic evolution as driving force of tumorigenesis, lucidly exposed by the dynamic heterogeneity of "cancer stem cells" or the diffuse results of cancer genome sequencing projects, indicate the need for a more encompassing theory of cancer that reaches beyond the current proximate explanations based on individual genetic pathways. One such integrative concept, derived from first principles of the dynamics of gene regulatory networks, is that cancerous cell states are attractor states, just like normal cell types are. Here we extend the concept of cancer attractors to illuminate a more profound property of cancer initiation: its inherent inevitability in the light of metazoan evolution. Using Waddington's Epigenetic Landscape as a conceptual aid, for which we present a mathematical and evolutionary foundation, we propose that cancer is intrinsically linked to ontogenesis and phylogenesis. This explanatory rather than enumerating review uses a formal argumentation structure that is atypical in modern experimental biology but may hopefully offer a new coherent perspective to reconcile many conflicts between new findings and the old thinking in the categories of linear oncogenic pathways.

Deterministic and stochastic models of genetic regulatory networks

Traditionally molecular biology research has tended to reduce biological pathways to composite units studied as isolated parts of the cellular system. With the advent of high throughput methodologies that can capture thousands of data points, and powerful computational approaches, the reality of studying cellular processes at a systems level is upon us. As these approaches yield massive datasets, systems level analyses have drawn upon other fields such as engineering and mathematics, adapting computational and statistical approaches to decipher relationships between molecules. Guided by high quality datasets and analyses, one can begin the process of predictive modeling. The findings from such approaches are often surprising and beyond normal intuition. We discuss four classes of dynamical systems used to model genetic regulatory networks. The discussion is divided into continuous and discrete models, as well as deterministic and stochastic model classes. For each combination of these categories, a model is presented and discussed in the context of the yeast cell cycle, illustrating how different types of questions can be addressed by different model classes.

Elevated p53 and p21waf1 mRNA expression in blood lymphocytes from lung cancer patients with chemoresistance

BACKGROUND

Clinical and experimental studies suggest that alteration of the expression level of the p53 gene and other damage responsive genes may be associated with chemoresistance in cancer patients.

METHODS

The present study evaluated the differences of the basal levels of lymphocytic p53 and p21waf1 mRNA expression collected before receiving cisplatin-based chemotherapy between 48 chemo-ineffective lung cancer patients and 39 chemo-effective lung cancer patients using an optimized semi-quantitative multiplex reverse transcriptase-polymerase chain reaction (RT-PCR).

RESULTS

The data indicated that the mean mRNA level of p53 gene in chemo-ineffective patients (0.66) was 26.9% higher than that of the chemo-effective patients (0.52) with statistical significance (P=0.03), and a significantly higher level of p21waf1 mRNA expression in the chemo-ineffective patients (P=0.03) was also observed. In addition, by the multiplex long quantitative PCR analysis, we demonstrated that chemo-ineffective and chemo-effective patients have similar amounts of UV damage on their p53 gene of lymphocyte DNA through equal UV treatment.

CONCLUSION

Our results suggest that elevated levels of p53/p21waf1 mRNA in blood lymphocytes collected before chemotherapy may predict the chemoresponses of lung cancer patients.

Gamma irradiation of Type B spermatogonia leads to heritable genomic instability in four generations of mice

Mice conceived 6 weeks after paternal exposure to ionizing radiation were fathered by sperm that were Type B spermatogonia at the time of irradiation. Previous studies of these offspring showed that this paternal F0 germ cell irradiation led to decreased embryonic cell proliferation rates, altered enzyme activities, protein levels and whole-body weights. In the present study, we examined four generations of CD1 mice following paternal F0 irradiation of the Type B spermatogonia (1.0 Gy, (137)Cs gamma rays) to determine the stability of the heritable effects. Offspring were evaluated for changes in protein kinase C and mitogen-activated protein kinase enzyme activities and Trp53 and p21(waf1) protein levels. Two or more endpoints were significantly altered in all four generations of offspring from the irradiated F0 sire (P <or= 0.05). To test the hypothesis that these heritable biochemical effects are random stochastic responses rather than some predictable uniform response, each endpoint was also evaluated in terms of a variability index (VI). Results of VI analyses show that the observed heritable phenotype is unpredictable in magnitude and direction of change for an endpoint between generations and within generations. These results indicate that irradiated spermatogonia develop a capacity to transmit a type of heritable genomic instability to four generations of offspring.

Regulatory mechanism of glutathione S-transferase P-form during chemical hepatocarcinogenesis: old wine in a new bottle

The expression of glutathione S-transferase P-form (GST-P) is markedly up-regulated in the initial phase of chemical hepatocarcinogenesis. It is unlikely that a specific genetic change is associated with this common response to a variety of carcinogens. Here, we describe how GST-P gene expression is induced by carcinogenic treatment, focusing on the changes in the network of liver-enriched transcription factors, including CCAAT/enhancer-binding proteins. Although the balance of positive and negative transcription factors regulates the expression of the GST-P gene, additional factors such as the altered regulation of growth control may certainly be necessary for these cells to develop into preneoplastic foci. Furthermore, our genetic analyses on the tumor susceptibility of (F344 x DRH)F2 rats support the hypothesis that the formation of GST-P-positive lesions is required but is not directly associated with final malignant transformation.

Prediction and measurement of an autoregulatory genetic module

The deduction of phenotypic cellular responses from the structure and behavior of complex gene regulatory networks is one of the defining challenges of systems biology. This goal will require a quantitative understanding of the modular components that constitute such networks. We pursued an integrated approach, combining theory and experiment, to analyze and describe the dynamics of an isolated genetic module, an in vivo autoregulatory gene network. As predicted by the model, temperature-induced protein destabilization led to the existence of two expression states, thus elucidating the trademark bistability of the positive feedback-network architecture. After sweeping the temperature, observed population distributions and coefficients of variation were in quantitative agreement with those predicted by a stochastic version of the model. Because model fluctuations originated from small molecule-number effects, the experimental validation underscores the importance of internal noise in gene expression. This work demonstrates that isolated gene networks, coupled with proper quantitative descriptions, can elucidate key properties of functional genetic modules. Such an approach could lead to the modular dissection of naturally occurring gene regulatory networks, the deduction of cellular processes such as differentiation, and the development of engineered cellular control.

Epigenetic Theories of Cancer Initiation

I argue that carcinogenic insults injure many cells rather than mutate a few. This results from evidence that such insults convert too many cells to a precancerous state and that too many of the converted cells then revert to plausibly involve mutation and its repair; from evidence that the delays between such insults and chemically demonstrable mutations are long enough to easily allow nonmutational mechanisms to work; from evidence that even ionizing radiation first acts on the cytoplasm and mainly affects cells unhit by it; from the fact that such insults induce proto-oncogene expression far too quickly to do so by mutation; and from the fact that fusions of various cells and cell parts show that the tumorous or nontumorous nature of the product depends on its cytoplasmic rather than its nuclear component. I further argue that reduced DNA methylation, modifications of the histone code, and tissue disorganization are the three main mechanisms of epigenetic cancer initiation. Hypomethylation would result from DNA excision repair. Moreover, a methyl-deficient diet is carcinogenic and demethylation is also known to be carcinogenic via the histone code. Finally, I strongly argue for tissue disorganization as a mechanism of cancer initiation. This results from evidence that skin carcinogens disrupt the dermal/epidermal connection and from the fact that tumorigens swiftly disrupt gap junctions, as well as from evidence that such disruption is tumorigenic.

Loss of heterozygosity, microsatellite instability, and mismatch repair protein alterations in the radial growth phase of cutaneous malignant melanomas

Little is known about genomic alterations during development of the radial growth phase (RGP) of cutaneous malignant melanomas (CMMs). In this investigation polymerase chain reaction-based microsatellite assays were applied to analyze 13 RGP-CMMs with 18 microsatellite markers at six chromosomal regions: 1p, 3p, 4q, 6q, 9p, and 10q. Loss of heterozygosity (LOH) was found in eight cases (62%), at 9p22, 1p36, and 10q11, suggesting the presence of tumor-suppressor genes at these regions. LOH was encountered frequently at the interferon-alpha (31%) and D10S249 loci (15%). Low-level microsatellite instability (MSI) (11-16% of investigated loci unstable) was noted in three cases (23%). Two MSI banding patterns were seen: band shift and the presence of additional bands. To investigate the underlying mechanisms of the low-level MSI pattern, we analyzed the lesions for expression of mismatch repair (MMR) proteins with immunoperoxidase methods and mouse monoclonal antibodies. The average percentages of positively stained cells for human MutL homolog 1 (hMLH1), human MutS homolog 2 (hMSH2), and human MutS homolog 6 (hMSH6) in RGP-CMM (75.6 +/- 3.4%, 67.20 +/- 7.71%, and 76.6 +/- 2.1%, respectively) were reduced compared with benign nevi. No statistically significant differences in MMR protein expression were found between microsatellite-stable and low-level MSI lesions (P = 0.173, P = 0.458, and P = 0.385 for hMLH1, hMSH2, and hMSH6, respectively). There was a direct correlation between values for percentages of positively stained cells for hMSH2 and hMSH6 (r = +0.9, P = 0.03), suggesting that common mechanisms regulate their expression. In conclusion, LOH, MSI, and reduced MMR protein expression appear to be present in at least some RGP-CMMs and may play a role in their pathogenesis. Further studies are necessary to support these finding and to determine their diagnostic and prognostic significance.

Alterations of mismatch repair protein expression in benign melanocytic nevi, melanocytic dysplastic nevi, and cutaneous malignant melanomas

Immunoperoxidase-staining methods were used to examine the expression of hMLH1, hMSH2, and hMSH6 mismatch repair (MMR) proteins in 50 melanocytic lesions. Microsatellite instability (MSI), screened previously in these lesions by polymerase chain reaction-based microsatellite assay, showed low-level microsatellite instability (MSI-L) in 11 of 22 melanocytic dysplastic nevi (MDN) and two of nine primary cutaneous malignant melanomas (CMMs) but not in the benign melanocytic nevi (BN). Mismatch repair proteins were widely expressed in the epidermis and adnexal structures. All lesions showed positive immunoreactivity with a gradual decrease in the MMR staining values during the progression from BN to MDN to CMMs. The average percentage of positively (PP) stained cells for hMLH1, hMSH2, and hMSH6 in BN was 85.50 +/- 1.95, 77.90 +/- 4.50, and 87.11 +/- 1.85, respectively. The PP cell values in CMMs were significantly reduced as compared with BN (75.22 +/- 3.57, p= 0.01; 56.11 +/- 8.73, p= 0.02; 65.22 +/- 6.47, p = 0.0002 for hMLH1, hMSH2, and hMSH6, respectively). No comparable significant difference was found between microsatellite stable and MSI-L lesions (p = 0.173, p = 0.458, and p = 0.385), suggesting a lack of correlation between MMR expression and MMR function. There was a direct correlation between PP cell values of hMSH2 and hMSH6 (R = 0.39, p = 0.008), implying that their expression could be regulated by a common mechanism. Thus, an important finding of these studies was the reduction of MMR protein levels in CMMs; whether this reflects underlying genetic or epigenetic mechanisms is still to be determined.

Reduced expression of hMSH2 and hMLH1 and risk of prostate cancer: a case-control study

BACKGROUND

Although prostate cancer is the most common incident cancer in men, not much is known about its etiology. We tested the hypothesis that expression levels of hMSH2 and hMLH1 in unaffected (normal) tissue play a role in the etiology of prostate cancer.

METHODS

Total RNA was extracted from peripheral blood lymphocytes of subjects ascertained by a case-control study (70 patients and 97 age- and ethnicity-matched controls). A multiplex reverse transcription-polymerase chain reaction assay was used to simultaneously evaluate the relative expression of hMSH2 and hMLH1, using beta-actin as the internal control.

RESULTS

The relative gene expression levels of hMSH2 and hMLH1 were significantly lower in cases than in controls (P < 0.05 for both genes). When compared with the highest tertile of the controls, low expression levels (the middle and lowest tertiles) of hMLH1 were associated with significantly increased risk of prostate cancer in a dose-response relationship (ORs = 2.68, and 4.31; 95% confidence interval = 1.00-7.23 and 1.64-11.30, respectively) after adjustment for age, ethnicity, smoking status, and family history of prostate cancer.

CONCLUSIONS

These results suggest that reduced expression of hMLH1 in peripheral lymphocytes may be a risk factor for prostate cancer. However, it cannot be ruled out that the reduced expression we observed may be caused by the disease status. Our findings and the factors that may affect the expression of hMLH1 need further confirmation in larger prospective studies.

Involvement of transcription factor HNF3gamma in the effect of o-aminoazotoluene on glucocorticoid induction of tyrosine aminotransferase in mice sensitive to its hepatocarcinogenic action

In the rodent liver, hepatocarcinogens inhibit the glucocorticoid induction of several liver-specific genes, including tyrosine aminotransferase (TAT). A distinct positive correlation exists in mice between the extent of inhibition of TAT induction after acute administration of o-aminoazotoluene (OAT) and the frequency of liver tumors after chronic exposure to the carcinogen. To elucidate the mechanism of the carcinogenic action, the effects of OAT on the DNA-binding activity of several transcription factors participating in the glucocorticoid regulation of TAT gene expression were studied. The experimental inbred male mice were sensitive (A/He and SWR/J, tumor induction frequency of 75-100%, TAT induction inhibition of 35-50%) and resistant (CC57BR/Mv and AKR/J, 0-6% and 10-15%, respectively) to OAT. Gel retardation experiments showed that hepatocyte nuclear factor 3 (HNF3)gamma DNA-binding activity was strongly reduced in nuclear extracts from the livers of OAT-treated A/He and SWR/J mice but only slightly reduced in CC57Br/Mv and AKR/J mice. The DNA-binding activities of Ets, AP1 family members, and GME binding proteins were unaffected. HNF3gamma DNA-binding activity was reduced by 1 h after OAT administration and remained low for 1 mo, as did inhibition of TAT induction in the liver. These results suggested that the inhibitory effect of OAT on the glucocorticoid induction of TAT is mediated by reduced HNF3gamma DNA-binding activity.

K-ras mutations in mouse lung tumors of extreme age: independent of paternal preconceptional exposure to chromium(III) but significantly more frequent in carcinomas than adenomas

Preconceptional exposure of male NIH Swiss mice to chromium(III) chloride resulted in increased incidence of neoplastic and non-neoplastic changes in their progeny, including lung tumors in females [Toxicol. Appl. Pharmacol. 158 (1999) 161-176]. Since mutations in the K-ras protooncogene are frequent, early changes in mouse lung tumors, we investigated possible mutational activation of this gene as a mechanism for preconceptional carcinogenesis by chromium(III). These offspring had lived until natural death at advanced ages (average 816+/-175 days for controls, 904+/-164 for progeny of chromium-treated fathers). Mutations of K-ras, analyzed by single-strand conformation polymorphism and sequencing, were, in codon 12, wild type GGT (glycine), to GAT (aspartic acid); to GTT (valine); and to CGT (arginine); and in codon 61, wild-type CAA (glutamine), to CGA (arginine). K-ras mutation frequencies in lung tumors were very similar in control progeny (4/14) and in progeny of chromium-treated fathers (5/15). Thus, germline mutation or tendency to spontaneous mutation in K-ras does not seem to be part of the mechanism of preconceptional carcinogenesis here. However, an additional interesting observation was that K-ras mutations were much more frequent in lung carcinomas (8/16) than in adenomas (1/13) (P=0.02), for all progeny combined. This was not related to age of the tumor-bearing mice or the size of the tumors. K-ras mutations may contribute to malignant tumor progression during aging, of possible relevance to the putative association of such mutations with poor prognosis of human lung adenocarcinomas.

Radiation-induced genomic instability: a paradigm-breaking phenomenon and its relevance to environmentally induced cancer

The existing paradigm governing radiobiology which is fundamental to the estimation of environmental radiation risk, cannot explain the phenomena of radiation induced genomic instability and the bystander effect. Both effects can, however, be understood in terms of the dynamical genome concept, qualitatively described herein. The dynamical genome concept may find further application in better understanding other aspects of biology, most notably the cancer process in general and the consequences of genetic modification.

Genetic and epigenetic alterations in carcinogenesis

Precise and deliberate observations on tumors stand true for decades, and then meet mechanistic explanations. The presence of genetic alterations in tumors is now widely accepted, and explains the irreversible nature of tumors. However, observations on tissue differentiation indicated that it shares something in common with carcinogenesis, that is, "epigenetic" changes. Now, DNA methylation in CpG sites is known to be precisely regulated in tissue differentiation, and is supposed to be playing key roles. Many tumor suppressor genes are known to be inactivated by the hypermethylation of their promoter regions. DNA methylation is connected to histone deacetylation and chromatin structure, and regulatory enzymes of DNA methylation are being cloned. Dedifferentiation, dis(dys)differentiation and convergence of cancer cells were studied phenotypically and biochemically, and are now explained from molecular aspects of disturbances in tissue-specific transcription factors. Spontaneous regression of malignant tumors enchanted researchers, and it is now noticed that genes inactivated by hypermethylation are frequently involved in tumors that relatively often undergo spontaneous regression. Carcinogenic mechanisms of some carcinogens seem to involve modifications of epigenetic switch, and some dietary factors also have the possibility to modify the switches. Based on the growing understanding of the roles of DNA methylation, several new methodologies were developed to make a genome-wide search for changes in DNA methylation. Now, a wave of new findings is in sight.

Effects of macromolecular transport and stochastic fluctuations on dynamics of genetic regulatory systems

To predict the dynamics of genetic regulation, it may be necessary to consider macromolecular transport and stochastic fluctuations in macromolecule numbers. Transport can be diffusive or active, and in some cases a time delay might suffice to model active transport. We characterize major differences in the dynamics of model genetic systems when diffusive transport of mRNA and protein was compared with transport modeled as a time delay. Delays allow for history-dependent, non-Markovian responses to stimuli (i.e., "molecular memory"). Diffusion suppresses oscillations, whereas delays tend to create oscillations. When simulating essential elements of circadian oscillators, we found the delay between transcription and translation necessary for oscillations. Stochastic fluctuations tend to destabilize and thereby mask steady states with few molecules. This computational approach, combined with experiments, should provide a fruitful conceptual framework for investigating the function and dynamic properties of genetic regulatory systems.

Effect of oxidative DNA damage in promoter elements on transcription factor binding

Reactive oxygen species produced by endogenous metabolic activity and exposure to a multitude of exogenous agents impact cells in a variety of ways. The DNA base damage 8-oxodeoxyguanosine (8-oxodG) is a prominent indicator of oxidative stress and has been well-characterized as a premutagenic lesion in mammalian cells and putative initiator of the carcinogenic process. Commensurate with the recent interest in epigenetic pathways of cancer causation we investigated how 8-oxodG alters the interaction between cis elements located on gene promoters and sequence-specific DNA binding proteins associated with these promoters. Consensus binding sequences for the transcription factors AP-1, NF-kappaB and Sp1 were modified site-specifically at guanine residues and electrophoretic mobility shift assays were performed to assess DNA-protein interactions. Our results indicate that whereas a single 8-oxodG was sufficient to inhibit transcription factor binding to AP-1 and Sp1 sequences it had no effect on binding to NF-kappaB, regardless of its position. We conclude from these data that minor alterations in base composition at a crucial position within some, but not all, promoter elements have the ability to disrupt transcription factor binding. The lack of inhibition by damaged NF-kappaB sequences suggests that DNA-protein contact sites may not be as determinative for stable p50 binding to this promoter as other, as yet undefined, structural parameters.

Sequence dependence and characteristics of bends induced by site-specific polynuclear aromatic carcinogen-deoxyguanosine lesions in oligonucleotides

The tumorigenic metabolite of benzo[a]pyrene, the (+)-7R,8S,9S,10R enantiomer, and the nontumorigenic mirror-image isomer, (-)-7S,8R,9R, 10S, of r7,t8-dihydroxy-t9,10-epoxy-7,8,9, 10-tetrahydrobenzo[a]pyrene (anti-BPDE) bind covalently to the exocyclic amino group of deoxyguanosine (N2-dG) in native DNA. These adducts can cause structural perturbations such as DNA bends, which in turn may influence the cellular processing of these lesions. The characteristics of bends in site-specifically modified oligodeoxyribonucleotide duplexes induced by single (+)- and (-)-anti-[BP]-N2-dG lesions were examined by self-ligation and gel electrophoresis techniques. The modified residues (dG*) were centrally positioned in the 11-mer oligonucleotide d(CACAXG*XACAC) complexed with the natural complementary strands, with X = T or C, or in oligonucleotides 16 or 22 base pairs long with the same centrally positioned 11-mer. Among the four stereochemically distinct lesions, the 10S(+)-trans-anti-[BP]-N2-dG adducts were significantly more bent than any of the other three stereoisomeric adducts and were selected for detailed studies. In the TG*T sequence context (X = T), the retardation factor RL (apparent length of multimer/sequence length) is approximately independent of the phasing (distance, in base pairs, between the lesions) of the adducts with respect to the helical repeat (10.5 base pairs/helix turn). In contrast, in the CG*C sequence context (X = C), RL is markedly lower in the case of ligated 16-mers than in the case of ligated 11-mer duplexes. The dependence of RL on the phasing of the bends as a function of the helical repeat, indicate that the bends associated with (+)-trans-anti-[BP]-N2-dG lesions are relatively rigid in the d(...CG*C...).d(...GCG...) sequences, and flexible in the d(...TG*T...).d(...ACA...) sequence context. These differences are attributed to the orientations of the pyrenyl residues on the 5'-side of the modified deoxyguanosine residues in the minor groove and to the intrinsic roll and tilt characteristics of DNA dinucleotide steps CG, GC, TG, and GT. The influence of flanking bases on the extent and character of DNA bending suggest that base sequence effects may be important in the cellular processing of (+)-trans-anti-[BP]-N2-dG lesions.

Frequency selectivity, multistability, and oscillations emerge from models of genetic regulatory systems

To examine the capability of genetic regulatory systems for complex dynamic activity, we developed simple kinetic models that incorporate known features of these systems. These include autoregulation and stimulus-dependent phosphorylation of transcription factors (TFs), dimerization of TFs, crosstalk, and feedback. The simplest model manifested multiple stable steady states, and brief perturbations could switch the model between these states. Such transitions might explain, for example, how a brief pulse of hormone or neurotransmitter could elicit a long-lasting cellular response. In slightly more complex models, oscillatory regimes were identified. The addition of competition between activating and repressing TFs provided a plausible explanation for optimal stimulus frequencies that give maximal transcription. Such optimal frequencies are suggested by recent experiments comparing training paradigms for long-term memory formation and examining changes in mRNA levels in repetitively stimulated cultured cells. In general, the computational approach illustrated here, combined with appropriate experiments, provides a conceptual framework for investigating the function of genetic regulatory systems.

High-affinity binding of the cell cycle-regulated transcription factors E2F1 and E2F4 to benzo[a]pyrene diol epoxide-DNA adducts

Previous studies indicated that DNA adducts formed by a carcinogenic diol epoxide, 7r,8t-dihydroxy-9t, 10t-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene (BPDE), can increase the affinity of the transcription factor Sp1 for DNA sequences that are not normally specific binding sites. It was suggested that adduct-induced bends in the DNA were responsible for this behavior. The cell cycle-regulated transcription factor E2F is also known to bend DNA upon binding. When partially purified E2F was tested in a gel mobility-shift assay, binding to a target DNA containing two consensus E2F-binding sites was enhanced by prior modification of the DNA with BPDE. Recombinant human E2F1, E2F4, and DP1 fusion proteins were affinity purified from bacteria expressing these genes. A combination of either E2F1 or E2F4 with their dimerization partner, DP1, gave preparations that exhibited binding to the E2F site-containing DNA fragment. In both cases, the proteins exhibited much higher apparent affinity for BPDE-modified DNA than for unmodified DNA. In addition, BPDE-modified DNA was a better competitor for the binding than unmodified DNA. Heterologous DNA that contained no consensus E2F binding motifs also competed well for E2F binding when modified with BPDE. In contrast, transcription factor that does not bend DNA appreciably (GAL4) did not show enhanced affinity for BPDE-modified DNA. These findings suggest that numerous transcription factors that bend DNA may bind with anomalously high affinity to sequences that contain carcinogen-DNA adducts.

The functional matrix hypothesis revisited. 4. The epigenetic antithesis and the resolving synthesis

In two interrelated articles, the current revision of the functional matrix hypothesis extends to a reconsideration of the relative roles of genomic and of epigenetic processes and mechanisms in the regulation (control, causation) of craniofacial growth and development. The dialectical method was chosen to analyze this matter, because it explicitly provides for the fuller presentation of a genomic thesis, an epigenetic antithesis, and a resolving synthesis. The later two are presented here, where the synthesis suggests that both genomic and epigenetic factors are necessary causes, that neither alone is also a sufficient cause, and that only the two, interacting together, furnish both the necessary and sufficient cause(s) of ontogenesis. This article also provides a comprehensive bibliography that introduces the several new, and still evolving, disciplines that may provide alternative viewpoints capable of resolving this continuing controversy; repetition of the present theoretical bases for the arguments on both sides of these questions seems nonproductive. In their place, it is suggested that the group of disciplines, broadly termed Complexity, would most likely amply repay deeper consideration and application in the study of ontogenesis.

DNA methylation, heterochromatin and epigenetic carcinogens

This paper will explore emerging concepts related to alternative carcinogenic mechanisms of 'non-mutagenic,' and hence epigenetic, carcinogens that may heritably alter DNA methylation without changing the underlying DNA sequence. In this review, we will touch on the basic concepts of DNA methylation, and will elaborate in greater detail on related topics including chromatin condensation, and heterochromatin spreading that is well known to induce gene silencing by position effect variegation in Drosophila and other species. Data from our model transgenic G12 cell system will be presented to support our hypothesis that certain carcinogens, such as nickel, may be carcinogenic not primarily because of their overt mutability, but rather as the result of their ability to promote DNA hypermethylation of important cancer-related genes. We will conclude with a discussion of the broader relevance of our findings and its application to other so-called 'epigenetic' carcinogens.

DNA damage can alter the stability of nucleosomes: effects are dependent on damage type

We have investigated the effects of DNA damage by (+/-)-anti-benzo[a]pyrene diol epoxide (BPDE) and UV light on the formation of a positioned nucleosome in the Xenopus borealis 5S rRNA gene. Gel-shift analysis of the reconstituted products indicates that BPDE damage facilitates the formation of a nucleosome onto this sequence. Competitive reconstitution experiments show that average levels of 0.5, 0.9, and 2.1 BPDE adducts/146 bp of 5S DNA (i.e., the size of DNA associated with a nucleosome core particle) yield changes of -220, -290, and -540 cal/mol, respectively, in the free energy (delta G) of nucleosome formation. These values yield increases of core histone binding to 5S DNA (K(a)) of 1.4-, 1.6-, and 2.5-fold, compared with undamaged DNA. Conversely, irradiation with UV light decreases nucleosome formation. Irradiation at either 500 or 2500 J/m2 of UV light [0.6 and 0.8 cyclobutane pyrimidine dimer/146 bp (on average), respectively] results in respective changes of +130 and +250 cal/mol. This translates to decreases in core histone binding to irradiated 5S DNA (K(a)) of 1.2- and 1.5-fold compared with undamaged DNA. These results indicate that nucleosome stability can be markedly affected by the formation of certain DNA lesions. Such changes could have major effects on the kinetics of DNA processing events.

The non-random distribution of UV-induced photoproducts in the nuclear matrix and non-matrix DNA fractions

The formation of UV-induced photoproducts in the chromatin fractions of human lymphocytes was studied by 32P-post-labeling. A higher level of DNA lesions was found in the matrix-attached DNA fraction as compared to non-matrix DNA of irradiated cells (about 150 and 110 adducts per 10(6) nucleotides, respectively, at a 500 J/m2 254 nm-UV dose). Formation of photoproducts in a MAR (matrix attached region) sequence from the mouse kappa immunoglobulin gene irradiated in vitro was examined as well. The MAR sequence showed a two-fold higher level of adducts as compared to non-MAR DNA. The effect of photoproducts on complex-formation between MAR DNA and proteins of the nuclear matrix was studied in vitro. The amount of UV-induced adducts was 1.5-fold higher in matrix-bound fraction as compared to non-fractionated DNA (and five-fold higher as compared to unbound fraction), which possibly resulted from preferential binding of lesion-containing DNA fragments to the nuclear matrix proteins.