The Need for, and the Role of the Toxicological Chemist in the Design of Safer Chemicals

Abstract

During the past several decades there has been an ever increasing emphasis for designers of new commercial (non-pharmaceutical) chemicals to include considerations of the potential impacts a planned chemical may have on human health and the environment as part of the design of the chemical, and to design chemicals such that they possess the desired use efficacy while minimizing threats to human health and the environment. Achievement of this goal would be facilitated by the availability of individuals specifically and formally trained to design such chemicals. Medicinal chemists are specifically trained to design and develop safe and clinically efficacious pharmaceutical substances. No such formally trained science hybrid exists for the design of safer commercial (non-pharmaceutical) chemicals. This article describes the need for and role of the ”toxicological chemist”, an individual who is formally trained in synthetic organic chemistry, biochemistry, physiology, toxicology, environmental science, and in the relationships between structure and commercial use efficacy, structure and toxicity, structure and environmental fate and effects, and global hazard, and trained to integrate this knowledge to design safer commercially efficacious chemicals. Using examples, this article illustrates the role of the toxicological chemist in designing commercially efficacious, safer chemical candidates.

Introduction

Recognition of the need for chemists to design chemicals that are not only useful but of minimal hazard can be traced back to at least 1928, when Alice Hamilton, the well known physician and pioneer in industrial medicine, made the following statements in her chapter, Protection Against Industrial Poisoning, in the book Chemistry in Medicine (Hamilton, 1928):

“Chemistry and medicine have thus made possible real progress in the protection of working men and women against industrial poisons........ Much remains to be done in this field, even in the light of our present knowledge, and greater progress will be made possible in the future through advances in chemistry. For instance, substitutes which are relatively non-toxic may be found to take the place of toxic compounds now in use…. Toxicology must join with chemistry in testing the new compounds which chemistry introduces into industry.... Synthetic chemistry must have as one of its great objectives the further safeguarding of health and of life in the industries into which chemistry itself has introduced new poisons.“

In the era these statements were made they were quite bold, if not radical, and likely to have been received with much indifference and perhaps opposition, especially since the statements were made by a woman. In 1928, only eight years had elapsed since the Nineteenth Amendment to the United States (U.S.) Constitution had, after intense debate, become law and allowed women to vote. The U.S. economy was doing well and jobs were plentiful. Although it undoubtedly existed, pollution was not viewed as a problem by the general population or the federal government. As such, very few federal laws or regulatory authorities existed that regulated the development and marketing of commercial industrial chemicals, pesticides or pharmaceutical substances to protect human health and the environment from risks posed by such substances.

Synthetic chemistry, involving both organic and inorganic chemistry, has been a prominent discipline among the physical sciences for over two centuries. The basic scientific literature on chemical synthesis has been a virtual watershed of knowledge that has fueled the development of large numbers of new industrial and commercial products and intermediates by the chemical industry. For many years following the early introduction of chemicals for use in commerce and industry, the emphasis was primarily on efficacy. Motivated by highly competitive markets, chemists focused principally on how well a chemical functioned in its intended use, and designed chemicals accordingly. Developers of new chemicals have continuously explored new structural designs and configurations in their attempts to develop better-performing, lower cost products and intermediates. For many years only limited consideration was given to the effects that parent molecular structures and their functional groups, or metabolites thereof, might have on human health and the environment.

As Alice Hamilton noted in 1928, it was apparent that many chemicals have the potential to pose serious risks to human health and the environment (Hamilton, 1928). In the United States, these concerns were addressed many years later through establishment of federal regulatory agencies such as the Food and Drug Administration (FDA) and the Environmental Protection Agency (EPA). These organizations are empowered by many laws to control the risks posed by new and existing drug substances (FDA), pesticides (EPA) and industrial chemicals (EPA). Similar organizations were established and laws enacted in many other countries.

The Need for the Toxicological Chemist

It was eventually realized that the above approach to chemical design led to the development and marketing of many chemicals that posed unreasonable risks to human health and the environment, and that efforts to control these risks through traditional regulation of exposure are highly resource-intensive approaches with no real advantages to industry, environmental authorities, or society as a whole. Hamilton’s views on the importance of chemical safety and the need for synthetic chemists to collaborate with toxicologists to develop chemicals that are “relatively non-toxic“ were both brilliant and very far ahead of her time.

During the past several decades there has been an ever increasing emphasis for designers of new chemicals to include considerations of the potential impacts a planned chemical may have on human health and the environment as part of the design of the chemical, and to design chemicals such that they possess the desired use efficacy while minimizing threats to human health and the environment (Iles and Mulvihill, 2012). Some major commercial chemical companies have teams composed of chemists, toxicologists, environmental fate specialists, policy analysts and others devoted to the design of safe and efficacious commercial chemical substances (Settivari, et al., 2017). But most new commercial chemicals are not designed from such multidisciplinary expertise.

A current impediment is that few academic graduate programs exist that offer curricula or formal training to design safer commercial chemicals (Iles and Mulvihill, 2012). Garrett (1996) has discussed the need for academic institutions to establish multidisciplinary curricula that provides individuals with firm grounding in organic chemistry, biochemistry, toxicology and other coursework and research directly related to the chemistry/biology relationships involved in designing safer chemicals, and described individuals that receive such training as “toxicological chemists”.

An analogous impediment once existed within the field of drug development. In the United States the “Federal Food Drug and Cosmetic Act“ (FFDCA) became effective in 1938, and has since been amended several times to address emerging societal concerns regarding the safety of drug substances. Among other provisions, this law, as amended, authorizes the FDA to require that pharmaceutical firms provide evidence of safety and efficacy of new drug substances before such substances can be marketed. Through the FFDCA, FDA requires pharmaceutical firms to conduct extensive testing to identify and characterize a candidate drug substance’s: clinical pharmacological efficacy; bioavailability; bodily distribution; metabolites; excretion; and any adverse or toxic effects the substance may cause in experimental animals and in humans during pre-market clinical trials. Pharmaceutical firms have to develop these data, ostensibly as proof that their new drug is safe and effective.

This information is submitted to FDA as part of an application for new drug approval, and undergoes extensive review. If the FDA determines that the drug substance is clinically efficacious and has minimal adverse effects, it will approve its marketing and use. Even with the streamlined processes currently used, the development and marketing of a new drug product is time consuming and resource intensive. Typically, for every new drug that reaches the market more than 8,000 potential drug candidates were synthesized, tested to varying extents along the way, and judged to be unsuitable. The identification of a candidate drug substance, its testing, and FDA approval usually takes many years and, nowadays, costs upwards of hundreds of millions of dollars. Because of the costs and rigorous approval process outlined above, relatively few new drug substances are approved and registered by FDA on an annual basis.

Promulgation of the FFDCA in 1938, specifically the pre-market testing it mandates, led to the publication of many studies that reported the metabolism, pharmacological and toxicological properties of many classes of chemicals undergoing evaluation as potential pharmaceuticals. This wealth of information enabled characterization of relations between structure, pharmacological activity, potency, and toxicity of many classes of organic chemicals. Identification of these relationships would provide organic chemists with a rational basis from which molecular modifications expected to maximize the desired pharmacologic effect while minimizing toxicity could be inferred and, therewith, used to design new molecules in which therapeutic effectiveness was maximized and toxicity minimized.

The problem was that individuals that developed new pharmaceutical products received none of the academic training in the biological or physical sciences that was needed to enable them to analyze, interpret and integrate such information for the design of new and improved drug substances. There was a need for a new type of scientist, a “medicinal chemist“: a chemist hybrid who received extensive training not only in synthetic organic chemistry but also in biochemistry, pharmacology and toxicology, and the relationships between chemical structure with physical properties, pharmacological action and toxicologic effects. Such an individual would be well prepared to design new clinically efficacious drug substances of low toxicity.

The noted biochemist R. Tecwyn Williams, and noted organic chemist Alfred Burger recognized this need. In 1947, Williams published the first edition of his classic text on mechanisms of drug metabolism: Detoxication Mechanisms of Drugs and Allied Organic Compounds (Williams, 1947), which is an extensive compilation of the metabolic pathways that many of the drugs and industrial chemicals in use at the time undergo in experimental animals and humans. Burger, in 1951 and 1952, wrote and published a two volume book set that he titled Medicinal Chemistry: Chemistry, biochemistry, therapeutic and pharmacological action of natural and synthetic drugs (Burger, 1951,1952), to provide graduate students majoring in organic chemistry who plan to pursue careers in drug development, and organic chemists working for pharmaceutical firms, a framework from which safe and efficacious drug substances could be designed (Burger, 1988).

Burger’s book helped to establish the field of medicinal chemistry. Soon after publication of the two volumes in 1951 and 1952, respectively, many drug companies formed departments of medicinal chemistry, and many colleges, particularly colleges of pharmacy, did the same. This eventually led to the availability of professionals who were specifically and formally trained to design and develop therapeutically useful but safe drug susbstances. Within 10 years the Journal of Medicinal Chemistry was founded, also by Burger (1988), and the American Chemical Society established its section on Medicinal Chemistry.

Design and development of commercial chemicals is done quite differently, despite the widespread recognition of the need for safe, commercially efficacious chemicals to be designed, developed and marketed, in that few people receive any formal training to design such chemicals (Iles and Mulvihill, 2012). Within the field of commercial chemical development, there is no “toxicological chemist”, in sharp contrast to the field of pharmaceutical development, in which there exists the medicinal chemist. The need for toxicological considerations to be a routine component of commercial chemical design has been emphasized by many over many decades (Hamilton, 1928; Ariëns and Simonis, 1982; Garrett, 1996; DeVito, 1996, 2012, 2016; Iles and Mulvihill, 2012; Anastas, 2016; Coish, et. al., 2016).

The toxicological chemist is one who is formally trained in synthetic organic chemistry, biochemistry, physiology, toxicology, environmental science, and in the relationships between structure and commercial use efficacy, structure and toxicity, and structure and environmental fate and effects, and global hazard. The toxicological chemist will be well prepared to work closely and collaboratively with experts in other disciplines to design commercially useful chemicals of low overall hazard.

Using examples, this article illustrates the role of the toxicological chemist in designing safer, commercially efficacious chemicals of minimal toxicity. This article is intended primarily for synthetic chemists and toxicologists in industry, academia and government, as these individuals represent the principal architects in the design, redesign, or evaluation of chemicals that ultimately fulfill the needs of an ever-advancing technological society. Space limitations do not permit a comprehensive discussion herein on how safer chemicals can be designed or on the many factors that need to be considered when designing a chemical to be useful and safe. But many of the approaches used by medicinal chemists to design safer and efficacious pharmaceutical products can be used to design safer and efficacious commercial chemicals. Such discussions are available, and include discussions on how to design chemicals to have reduced ecological hazard, global hazard, environmental persistency, among other undesired properties (Ariëns and Simonis, 1982; DeVito, 1996, 2012, 2016; Boethling and Voutchkova, 2012).

What Makes a Chemical Toxic?

Understanding the biochemical basis of why an existing chemical is toxic is extremely helpful in the design of new and safer chemicals, as it often enables one to infer structural modifications that are expected to reduce toxicity. The mechanisms of toxicity of some commercial chemicals and classes of chemicals have been elucidated or postulated, and put forth in many books or other reference sources. While this information is invaluable, a dilema that confronts the chemical designer is that the mechanisms of toxicity caused by the majority of commercial chemicals and classes of chemicals are unknown, or at least not well understood.

Worse yet, all toxic properties of many of the chemicals used in commerce remain uncharacterized or have not been fully investigated, let alone the biochemical mechanisms by which they are caused. A dilema that chemists will have, even with chemicals for which toxic mechanisms have been characterized, is that it is unlikely that the typically trained chemist is able to understand the toxicological and biochemical literature that describe toxic mechanisms. Without formal training in toxicology, or at least access to a toxicologist to help explain and interpret such literature, many chemists who attempt to use such information on their own will struggle in discerning structural modifications that are expected to reduce toxicity.

Here is where having been trained as a toxicological chemist will have enormous value. From their training, toxicological chemists will be able to contemplate a chemical’s structure and infer how its molecules will behave or react in the presence of other chemicals. This skill is critical in the design of safer, commercially useful chemicals because the basis of toxicity of any chemical ultimately distills down to an interaction of a chemical’s molecules or those of its metabolite(s) with molecules of another chemical, typically a cellular macromolecule.

Fundamentally, the individual molecules of an organic chemical can be viewed as individual clouds of electrons; each cloud held together in an identical shape unique to that molecule. The arrangements of atoms within a molecule’s structure determine the electron densities in different areas of the molecule (i.e., the shape of the cloud) and, therewith, the physical properties of that molecule. These structural arrangements and physical properties will determine whether and to what extent a molecule will react or interact with a molecule of another chemical species. Chemicals that are electrophiles, for example, will react with chemicals that are nucleophiles, whether in a test tube, reaction vessel, lake, soil, or a bodily organ such as the liver or kidney. When such reactions occur in humans or other living organisms, toxicity often ensues. This a priori knowledge ultimately enables the toxicological chemist to envision how a planned chemical will fulfill its intended use purpose, behave in a living or environmental system, and interact with other chemicals.

Toxicophores: Recognizing a Toxic Chemical When You See One

There are a number of structural substituents or moieties that are known to bestow toxicity, either directly or indirectly through metabolic conversion. Such structural moieties are known as toxicophores. For some toxicophores, the mechanisms by which they bestow toxicity are known, whereas with other toxicophores the mechanisms are unknown. In either case, the toxicological chemist will recognize known or potential toxicophoric substituents in a chemical, or obtain information pertaining to the mechanism by which a given toxicophoric substituent bestows toxicity and is in a much better position to design safer structural analogs of the chemical. From such knowledge one can often infer structural modifications needed to design a new substance to be much less toxic than the unmodified toxic substance. Structural changes intended to reduce toxicity should, of course, not interfere with the commercial usefulness of the substance, nor bestow another unwanted property (e.g., environmental persistence). Many of the more commonly encountered toxicophores found in existing commercial chemical substances, the mechanisms by which they bestow toxicity, and how an awareness of such toxicophores in molecules and knowledge of their toxic mechanisms can be used to design safer alternative substances, have been discussed (DeVito, 1996, 2012).

Structural moieties that are electrophilic or that are metabolized to electrophilic species (i.e., are proelectrophilic) are often toxicophoric because they are capable of reacting covalently with nucleophilic substituents of cellular macromolecules, such as DNA, RNA, enzymes, proteins, and others (DeVito, 1996, 2012). These cellular nucleophilic substituents include, for example: thiol groups of cysteinyl residues in protein; sulfur atoms of methionyl residues in protein; primary amino groups of arginine and lysine residues, or secondary amino groups (e.g., histidine) in protein; amino groups of purine bases in RNA and DNA; oxygen atoms of purines and pyrimidines; and, phosphate oxygens (P=O) of RNA and DNA (Vermeulen and te Koppele, 1993). Electrophilic or proelectrophilic toxicophores are of particular concern because they interact covalently and, therewith, irreversibly with important cellular macromolecules to cause disruptions in the functions of macromolecules. Depending upon the type of electrophile (or proelectrophile) and the cellular macromolecule, such covalent interactions can lead to a variety of serious toxic effects including cancer, hepatotoxicity, hematotoxicity, nephrotoxicity, reproductive toxicity, and developmental toxicity.

Fortunately, the mammalian body has several defense systems that minimize absorption of chemicals into the systemic circulation (e.g., the skin) or, if absorption has occurred, offer “sacrificial” nucleophiles that can react with foreign electrophiles. These natural defense systems are located in the liver and other organs, and include, among others, the glutathione transferase system and the epoxide hydratase system (Ariëns and Simonis, 1982). Reactions of electrophiles with the nucleophilic sites of these systems form readily-excretable water soluble products, and allow the safe elimination of the electrophiles before they can react with nucleophiles in more biologically-critical cellular macromolecules. The nucleophiles of these defense systems may become depleted or overburdened, however, upon continuous or high exposure to electrophilic substances. Under these conditions toxicity ensues.

Depending upon the structural characteristics of an electrophile, its reaction with a nucleophile will proceed either via an S_N1 mechanism or an S_N2 mechanism. Electrophiles reacting via an S_N1 mechanism, such as benzyl, allyl, or tertiary alkyl halides for example, do so through the formation of a carbonium ion intermediate and a displaced anion. The rate by which these electrophiles react with nucleophiles is not dependant on the amount of nucleophilic sites, but rather on how quickly the carbonium ion intermediate can separate from the displaced anion. Consequently, chloro-electrophilic moieties of commercial chemicals that react via an S_N1 mechanism will do so more readily with biological nucleophiles located within a cell than with extracellular nucleophiles such as those that may be found in the blood, since chloride ion concentration is normally lower in cells than it is in the blood and, hence, the carbonium ion intermediate will form more readily within a cell (Ross, 1962).

The rates by which electrophiles react with nucleophiles via an S_N2 mechanism, such as α,β-unsaturated carbonyl moieties, is determined primarily by the concentration of nucleophilic sites. When in the body, electrophilic substituents of commercial substances that react by an S_N2 mechanism tend to react more so in areas where there are higher concentrations of nucleophilic sites, such as near nucleic acid chains of DNA or RNA, or proteins rich in thiol moieties (cysteine residues) (Ross, 1962).

Examples and discussions of specific electrophilic chemical substances and the toxic effects they cause are available (DeVito, 1996, 2012 and references cited therein; Blagg, 2010). Although the presence of a toxicophoric substituent raises the possibility that the substance may be toxic, it should not be inferred that the presence of such a substituent always means that the substance is toxic. Whether the substance is toxic is also dependent on factors such as its overall bioavailability, its metabolism and the presence of other substituents that may enhance or attenuate the reactivity of the electrophilic substituent.

Designing Safer Electrophilic Substances

Ideally, electrophilic substituents should never be incorporated into a substance. An electrophilic group is often necessary, however, for the intended commercial use of the substance. This presents a challenge for one who wishes to design an electrophile to react with a nucleophile necessary for intended commercial use but not with biological nucleophiles in individuals exposed to the substance.

As impossible as this may seem, there are approaches that chemists can use to design safer, commercially-useful electrophilic substances. Acrylates, for example, contain an α,β-unsaturated carbonyl system and as such undergo 1,4-Michael addition reactions. This is believed to be the basis of the carcinogenic properties of acrylates (US EPA, 2002). Incorporation of a methyl (-CH₃) group onto the α-carbon (to provide a methacrylate) decreases the electrophilicity (i.e., reactivity) of the β-carbon (Osman, et al., 1988) and, hence, methacrylates do not undergo 1,4-Michael addition reactions as readily. Methacrylates often have commercial efficacy similar to acrylates in many applications, but are less likely to cause cancer because they are less reactive. This point can be demonstrated by comparing methyl methacrylate, 1, which does not cause cancer in experimental animals (NTP, 1986a), to ethyl acrylate, 2, which causes cancer in experimental animals in assays similar to those used to test 1 (NTP, 1986b).

It seems logical that placement of a methyl group onto the α-carbon of similar α,β-unsaturated systems may also decrease toxicity without significantly compromising commercial utility. This certainly appears to be true with methacrylonitrile, 3, a noncarcinogen in animal assays (NTP, 2001a), when compared to acrylonitrile, 4, which was found to be carcinogenic in identical assays (NTP, 2001b).

Because 3 does not cause cancer, but undergoes many of the same nucleophilic addition reactions as 4 at the β-carbon (3 is a less reactive monomer than 4 in polymerization reactions, Webster, O.W. (1990)), it is sometimes used as a safer commercial replacement for 4, such as in the manufacture of an acrylonitrile/butadiene/styrene-like polymer that provides improved barrier properties to gases such as carbon dioxide in carbonated beverage containers.

Exploiting Mechanisms of Toxicity for the Design of Safer Chemicals

It should be noted that when attempting to design a safer analog of an existing chemical of known toxicity, the toxicological chemist will often be confronted with situations where the mechanism by which the chemical elicits its toxicity is not intuitively discernable from structure; no obvious or known toxicophoric substitutents may be present, or it may otherwise not be known. In such cases the toxicological chemist who wishes to design a safer analog of the chemical should make an attempt to learn what the mechanism of toxicity is, or at least develop plausible possibilities as to what the toxic mechanism may be.

The toxicology and related literature goes deep on certain compounds, and shallow or doesn’t exist for a great many others that may be of interest to developers. This is especially true for proprietary substances under patent for which the structure may not be made known. For chemicals for which studies are available, the bits of information contained in these studies are much like pieces to a puzzle, that when carefully analyzed and integrated they form a clearer and more complete picture that may reveal the mechanism of toxicity. The toxicological chemist who takes the time to identify and obtain such reference materials, and carefully reviews and analyzes the information contained therein in totality, may very well be rewarded by the elucidation of the mechanism of toxicity and be better poised to design safer chemicals.

The most difficult circumstance for the toxicological chemist is in the attempt to identify a novel substance or novel class of substances that has less overall hazard than an existing, structurally unrelated substance or class of substances. Here the toxicological chemist is often challenged by having little or no empirical information to go on in regards to the toxicity of the novel substance, the toxicity characteristics of the chemical class to which it belongs and, if toxic, the mechanism(s) of any toxicities. While structurally manipulating or avoiding incorporation of toxicophores, and introducing structural modifications to reduce absorption or enhance elimination from the body will serve to minimize toxicity (DeVito, 1996), without adequate test data there is no way to know with a comfortable degree of certainty whether the novel chemical or chemicals in its class are in fact safer than the chemicals of which they are intended to replace. There are available, however, a number of computer software programs that can predict a priori the physicochemical, pharmacokinetic (or toxicokinetic), and toxic properties of planned or untested existing chemicals. These tools have emerged over the past several decades, as advancements in computational chemistry and molecular modeling have improved our understanding of the molecular basis of interactions between xenobiotics and biological macromolecules, and the corresponding sequalae of such interactions with respect to human health and the environment. A discussion of these computer programs is beyond the scope of this paper, but excellent reviews are available (Clark and Grootenhuis, 2002; Dearden, 2007; Greene, 2002; Helguera, et al., 2008; Jackson, 1995; Milne, et al., 1996; Mohan, et al., 2007; Richard, 1999; Ridings, et al., 1996; Valerio Jr., 2009; Voutchkova, et al., 2010; Wilson, et al., 2003).

While these programs have their limitations and are far from being perfect, they continue to be improved upon and have been shown to be useful in reducing uncertainty when estimating toxicities of planned and/or untested chemicals. The tools or methods that have been developed for estimation of toxicity are mostly directly or indirectly statistics based. The in silico approaches to toxicity predictions have been classified by some as heuristic or quantitative structure-activity relationship (QSAR)-based and so called “expert systems”, with the latter technically defined as a program that mimics the judgment of experts by following sets of knowledge rules. These “knowledge rules” are derived based on studies of toxicity mechanisms in animals and humans. Others have classified them as knowledge-based systems versus automated rule induction systems, based on whether the system is fed human knowledge or whether it derives its own prediction rules based on data-derived patterns.

These computational tools, as well predictive techniques involving the use of in vitro data and information are routinely used in the design and development of pharmaceuticals and pesticide chemicals (Settivari, et al., 2017), yet their applications in designing safer commercial chemicals appear to be substantially less. There is no reason to believe, however, that they would not serve as useful aids in the development of safer commercial chemical substances. Although, one should be careful not to allow any computer-assisted tool to supplant his or her own wisdom and intuition in chemical design and discovery, nor serve as a definitive substitute for experimental data.

Reducing Absorption from the Gastrointestinal Tract

If oral exposure is expected to be significant, the chemical should be modified to reduce absorption from the gastrointestinal tract. Simple modifications, such as increasing particle size or keeping the substance in an unionized form (i.e., free base, free acid), should in many cases reduce oral absorption. Designing the substance such that its octanol-water partitition coefficient (log P) is greater than 5 should ensure that the substance is not sufficiently water soluble for oral absorption. However, chemicals that are highly hydrophobic (log P >5) are more likely to persist in the environment unless their structures lend themselves to biotic or abiotic degradation in the environment. When designing chemicals to be more hydrophobic, one should avoid incorporating substituents (e.g, halogen atoms) that increase hydrophobicity which also impede environmental degradation. Designing the substance to be greater than 500 daltons, to have a melting point above 150 °C (for non-ionic substances) or to be a solid rather than a liquid should also reduce the likelihood of absorption from the oral route (DeVito, 1996).

The incorporation of substituents (e.g., -SO₃⁻) that remain strongly ionized at a pH of 2 or below should make the substance so polar that it cannot easily penetrate the lipid membranes of the intestinal lining and other membranes, and should significantly reduce absorption. Even if some absorption does take place, substances containing sulfonate or equally ionizable moieties have great difficulty in penetrating biological membranes of other tissues, and also should be excreted rapidly in the urine because of their extreme water solubility. This principle has been successfully applied to reducing the carcinogenicity of aromatic amines used to make azo dyes (Ariëns and Simonis, 1982).

The use of carcinogenic aromatic amines (e.g., 1- or 2-naphthylamine) to make azo dyes or pigments does not necessarily negate their carcinogenicity as they exist in the dyes or pigments, nor does it negate the carcinogenic potential of the dyes or pigments themselves. Many of the older azo dyes and pigments made from carcinogenic aromatic amines were later determined to also cause cancer or be classified as highly suspected of causing cancer. The azo bonds of dyes are often easily broken by chemicals or enzymes via reduction to form the free aromatic amine(s). Many strains of bacteria catalyze this process, including bacteria normally found in the human gastrointestinal tract (e.g., Escherichia coli), and other bacteria found in the environment (Brown and DeVito, 1993; Bae, et al., 2006). Azo reduction may also take place in the liver of humans and other mammals by reductase enzymes to liberate the aromatic amine constituents, but probably to less of an extent than in the intestines (Brown and DeVito, 1993).

The aromatic amine compounds can then be absorbed from the intestines and cause cancer (Brown and DeVito, 1993). Azo dyes or pigments released into the environment during dyeing of textiles are also a source of human exposure to carcinogenic amines from reductive degradation of the azo compounds in the environment, and subsequent contamination of drinking water with the free amines (Bae, et al., 2006).

Inclusion of sulfonate moieties on the aromatic amine feedstocks mitigates toxicity, as illustrated with the azo dye Brilliant Black BN (C.I. Food Black 1), 5, in Figure 1. The sulfonate moieties of 5 and each of its three reduction products are highly ionized in the gastrointestinal tract and at environmental pHs (i.e., pHs of 5–9). As such, 5 or any of its reduction products cannot penetrate membranes of the cells lining the intestines following oral exposure (Ariëns and Simonis, 1982). Consequently, these chemicals are poorly absorbed and any portion that is absorbed is rapidly excreted in urine.

Figure 1.

Brilliant Black BN (5) and its environmental or intestinal azo-reduction products. The sulfonate moieties of 5 and each of its three reduction products are highly ionized in the environment and in the gastrointestinal tract, and as such 5 or any of its reduction products cannot penetrate the membranes of the cells lining the intestines. Consequently the chemicals are poorly absorbed. Any portion that is absorbed is rapidly excreted in the urine.

Design of Safer Analogs of Tris(2,3-Dibromopropyl)Phosphate (Tris-BP)

There is no better example of a poster-child chemical that demonstrates the need for chemical manufacturers to determine the safety of a chemical before it is marketed than the halogenated alkyl phosphate tris(2,3-dibromopropyl)phosphate (Tris-BP), 6, Figure 2. Tris-BP was widely used in the U.S. during the 1970s as a flame retardant in children’s sleepwear as well as in other fabrics and plastics. Questions concerning its safety following human dermal exposure emerged in the early 1970s when it was observed that the chemical, known to be absorbed through the skin, was a potent mutagen in in vitro assays, and at least one of its metabolite was structurally similar to 1,2-dibromoethane, a known carcinogen (Blum and Ames, 1977).

Figure 2.

Metabolism of Tris-BP (6) to the potent mutagen 2-bromoacrolein (7).

The suspicion that Tris-BP causes cancer was confirmed in 1977, when it was shown that rats or mice who were fed the chemical in their diets developed tumors in their kidneys, in addition to other forms of nephrotoxicity (NCI, 1978; Reznik, et al., 1979). Consequently, on April 7, 1977, after virtually every child in the U.S. had been dermally exposed to the chemical for years every night as they slept, the U.S. Consumer Product Safety Commission banned the use of garments that had been treated with Tris-BP (U.S. Consumer Product Safety Commission, 1977a). Less than a month later, all uses of the chemical were banned (U.S. Consumer Product Safety Commission, 1977b).

In looking at the chemical structure of Tris-BP, one may understandably be tempted to ascribe its carcinogenic properties to covalent nucleophilic displacement of at least one of the three terminal alkyl (C-3) bromine atoms with some cellular macromolecular nucleophile. It was found, however, that Tris-BP undergoes cytochrome P450-mediated oxidative metabolism to the α,β-unsaturated carbonyl metabolite, 2-bromoacrolein (7), a potent direct acting mutagen believed to be responsible for causing the cancers associated with exposure to Tris-BP (Nelson, et al., 1984). 2-Bromoacrolein is formed by initial oxidation (i.e., hydroxylation and hydrogen radical abstraction) at one of the terminal –CH₂Br moieties (C-3 carbon atoms) of Tris-BP, followed by spontaneous dehydrohalogenation and dehydrophosphorylation (Nelson, et al., 1984; Omichinski, et al., 1987), as shown in Figure 2.

From this information one can infer three general structural modifications to incorporate into 6 at either the C-1, C-2 or C-3 positions that, at least in theory, should prevent its metabolism to 7 and, presumably, result in potentially safer substitutes for 6. These are, on each ester moiety: 1) replacing the two C-3 hydrogen atoms with moieties (e.g., methyl groups) that can not be abstracted by cytochrome P450 enzymes; 2) replacing the C-2 hydrogen atom with another moiety that will not eliminate from the C-2 carbon (e.g., a methyl group); or 3) insertion of at least one additional methylene (-CH₂-) group between the C-1 carbon and the phosphate oxygen. These structural modifications are not expected to interfere with the flame retardant properties of this class of compounds, which is believed to be due to formation of bromine radicals (·Br) that react with and quench the hydrogen radicals (·H) that are formed during combustion and necessary for flame propagation (Blum and Ames, 1977).

An example of each of these structural modifications is represented by compounds 8, 9 and 10, respectively, shown in Figure 3. It is noteworthy that, unlike Tris-BP, none of these compounds are mutagenic (Omichinski, et al., 1987), and may be noncarcinogenic as well.¹ In addition to being non-mutagenic, compound 8 was reported to be less nephrotoxic than Tris-BP and to have superior flame retardant properties (Day and Suprunchuk, 1988). Compounds 9 and 10 have flame retardant properties about equal to Tris-BP (Day and Suprunchuk, 1988). Neither 8, 9, or 10 are used commercially as flame retardants, perhaps because of their resemblance to Tris-BP.

Figure 3.

Non-mutagenic analogs of Tris-BP. Compound 8 is also much less nephrotoxic than Tris-BP, and has superior flame retardant properties [39, 40].

Interestingly, some substances structurally related to Tris-BP are still used in the United States and elsewhere as flame retardants in children’s foam-padded slepping products. Notable examples of such substances include tris(1,3-dichloro-2-propyl)phosphate and tris(2-chloroethyl)phosphate. The California Department of Toxic Substances has recently proposed a rule that, if finalized, will designate tris(1,3-dichloro-2-propyl)phosphate and tris(2-chloroethyl)phosphate as “chemicals of concern” under California’s Safer Consumer Products regulations (Erickson, 2016). California state regulators claim that exposure to children’s foam-padded sleeping products containing either of these chemicals is associated with cancer, reproductive toxicity, developmental effects and neurotoxicity, particularly in children, day care workers and school employees. If the proposed rule is finalized as currently written, companies that sell children’s foam-padded sleeping products that contain either tris(1,3-dichloro-2-propyl)phosphate or tris(2-chloroethyl)phosphate will have to identify safer flame retardants or use flame retardant-free foam.

Designing Safer Benzidine-Based Azo Dyes: Can Such Chemicals Exist?

As discussed previously in this paper, the use of carcinogenic aromatic amines to make azo dyes does not necessarily negate the carcinogenicity of an aromatic amine as it exists in the dyes made from the amine, or of the dyes or pigments themselves. The azo linkages in the dyes can be reduced (broken) by bacterial enzymes naturally found in the human gastrointestinal tract or environmental media, to liberate the carcinogenic amines (Brown and DeVito, 1993; Bae, et al., 2006). Azo reduction may also take place in the liver of humans and other mammals by reductase enzymes to liberate the aromatic amine constituents, but probably occurs to less of an extent than in the intestines (Brown and DeVito, 1993).

One way to mitigate the toxicity of the azo compound and any reduction products thereof is to incorporate substituents (e.g., -SO₃⁻) that remain strongly ionized at a pH of 2 or below as exemplified in Figure 1 with the dye Brilliant Black BN (5). Such substituents make the substance so polar that it cannot easily penetrate the lipid membranes of the intestinal lining and other membranes, and should significantly reduce absorption. Even if some absorption does take place, substances containing sulfonate or equally ionizable moieties have great difficulty in penetrating biological membranes of other tissues, and also should be excreted rapidly in urine because of their extreme water solubility. While often quite useful in mitigating the toxicity of azo dye reduction products, this modification is not always universally applicable because it can affect the intended color or other properties of a given dye.

The elucidation of the biochemical basis of carcinogenicity of aromatic amines, and the identification of structural modifications that reduce or mitigate the carcinogenic and other toxicities of aromatic amines have enabled a more specific biochemical-based approach to develop safer aromatic amines and derivatives thereof (Lai, et al., 1996). This is especially true in the colorant industry, and considerable progress has been made within the past 10 years on the development of safer azo-based dyes (Bae, et al., 2002 and 2006; Hinks, et al., 2000; Bello, et al., 2000; Sokolowska, et al., 2001; Szymczyk, et al., 2007; Sun, et al., 2007; Hanna, et al., 2007; Wang, et al., 2007a and 2007b). Much of this progress has been reported by the research group of Harold S. Freeman of North Carolina State University, and is focused on the identification of safer benzidine derivatives, or safer substitutes for benzidines.

Chemically, benzidine (11) and its congeners substituted with relatively small identical substituents at both the 3 and 3’ positions [e.g., -CH₃ (12); -Cl (13); -OCH₃ (14)], as shown in Figure 4, make excellent components of azo dyes and pigments because they impart many desirable properties, such as esthetically pleasing color and fastness, among others. In addition, they are inexpensive and readily form azo linkages with other substituted aromatic compounds to produce the desired azo dye or pigment in good yield. The problem with benzidine is that it is a known mutagen and known to cause bladder cancer in humans (NTP, 2005a). The 3,3’ disubstituted congeners 12–14 are known to be mutagenic and have been shown to cause cancer in experimental animals, and, hence, are likely to cause cancer in humans (NTP, 2005b–d). Not surprisingly, benzidine-3,3’-disulfonic acid is not carcinogenic (Ashby, 1978).

Figure 4.

Structures of benzidine (11) and some of its 3,3’ congeners (12–14). Benzidine is known to cause cancer in humans, and the 3,3’ congeners shown are known to cause cancer in experimental animals and, hence, are reasonably anticipated to be human carcinogens.

Many dyes made from 11 (e.g., Direct Black 38), 12 (e.g., C.I. Acid Red 114), 13 (e.g., C.I. Direct Red 46) and 14 (e.g., C.I. Direct Blue 15), as illustrated in Figure 5, have been shown to undergo azo reduction in the environment or gastrointestinal tract to release the respective benzidine congener (Brown and DeVito, 1993, NTP, 2005a–d). The U.S. National Toxicology Program classifies those azo dyes that contain or are metabolized to 11 as “known to be a human carcinogen” (NTP, 2005a), and those that contain or are metabolized to 12–14 as “reasonably anticipated to be human carcinogens” (NTP, 2005b–d).

Figure 5.

Examples of dyes made from 11 (e.g. Direct Black 38), 12 (e.g. C.I. Acid Red 114), 13 (e.g. C.I. Direct Red 46) and 14 (e.g. C.I. Direct Blue 15) that have been found to undergo azo reduction in the environment, or gastrointestinal tract or liver of mammals, to release the respective benzidine. Benzidine (11) is known to be carcinogenic in humans, as is Direct Black 38. The 3,3’ benzidine congeners 12–14 are reasonably anticipated to be human carcinogens, as are C.I. Acid Red 114, Direct Red 46, and Direct Blue 15.

Azo pigments, due to their very low water solubility, are generally much less susceptible to bacterial-mediated reduction than azo dyes, particularly within the gastrointestinal tracts of mammals. As a class, azo pigments made from benzidine or its 3,3’ congeners undergo very little, if any, metabolism in mammals to liberate benzidine or its congeners (Sagelsdorff, et al., 1996; Golka, et al., 2004).

Because of its toxicity, benzidine (11) is no longer manufactured or imported for commercial sale in the U.S. nor is it used any longer in the manufacture of dyes or pigments (NTP, 2005a). While benzidine congeners 12–14 are no longer used in the U.S. in the manufacture of dyes or pigments, they are still manufactured, imported and sold commercially in the U.S. for other purposes, albeit to much less of an extent than when they were used in the U.S. to manufacture dyes and pigments (NTP, 2005b–d).

The mechanism by which benzidines and other aromatic amines cause mutations and cancer have been elucidated, and are discussed in detail elsewhere (Lai, et al., 1996; Ioannides, et al., 1989; Kim and Guengerich, 2005). To summarize, benzidines require metabolic activation to be mutagenic and carcinogenic. The major bioactivating step is the initial enzyme-mediated hydroxylation of one of the amino groups. Cytochrome P450 enzymes are primarily involved with this hydroxylation, although other enzymes can be involved as well. The ability of the two phenyl rings to adopt a coplanar configuration appears to be important for N-hydroxylation. Substituents at either the 2 or 2’ positions of the diphenyl nucleus of benzidine interfere with the ability of the phenyl rings to be coplanar, and as such benzidine congeners tend to have less of an ability to undergo N-hydroxylation, and tend to be weaker mutagens and carcinogens (Lai, et. al, 1996).

The resulting N-hydroxylamine metabolite is further metabolized to highly reactive ester derivatives, such as N’-acetyl, N-acetoxy-benzidine via N-acetyltransferase. These ester derivatives can bind covalently to DNA, but they can also undergo heterolytic clevage to yield the even more reactive nitrenium ion species, which reacts more readily with DNA to form covalent adducts. In either scenario, the covalent binding to DNA initiates carcinogenesis. The overall bioactivation mechanism of benzidine is illustrated in Figure 6.

Figure 6.

Mechanism by which benzidine (11) causes mutations and cancer.

In the search for structual modificiations of the benzidine nucleus that reduce toxicity, it was observed that incorporation of alkyl or alkoxy substituents of three or four carbon atoms in length ortho to the amino groups result in non-mutagenic benzidines (Hinks, et al., 2000; Bae and Freeman, 2002), some of which are shown in Figure 7 as compounds 15–20. The lack of mutagenic properties of such analogs may be due to steric-induced interference with the pathways involved with the bioactivation of the amino groups (Lai, et al., 1996), or in the inability of the bioactivated amino groups to react covalently with DNA.

Figure 7.

Examples of some non-mutagenic benzidine analogs. The presence of alkyl or alkoxy substituents of three or four carbon atoms in length ortho to the amino groups mitigates the mutagenic properties of benzidine (11).

Additional testing needs to be performed to determine whether 15–20 and other non-mutagenic benzidine analogs of benzidine cause cancer, but the fact that compounds such as 15–20 are not mutagenic indicates that they may not be carcinogenic. An additional encouraging observation is that incorporation of alkyl or alkoxy substituents of three to four carbon atoms in length ortho to the amino groups of benzidine to mitigate mutagenicity and possibly carcinogenicity does not appear to detract from the desirable properties of azo dyes or pigments made from such substituted benzidines.

For example, azo dyes made from 16 or 19 of the type represented by compounds 21 and 22, respectively (Figure 8) show potential commercial promise (Bae and Freeman, 2002). Congeners such as 16 and 19 couple readily and efficiently with naphthalene-based compounds that are frequently used to prepare direct dyes, and the resultant dyes (e.g. 21 and 22) have color and fastness properties that are comparable to certain commercial direct dyes. The azo dye represented by 23 (Figure 8), also made from 16, is non-mutagenic and has found commercial use as a black dye in ink jet applications (Hinks, et al., 2000; Bae and Freeman, 2002).

Figure 8.

Examples of dyes made from 16 and 19 (non-mutagenic benzidine congeners) that show commercial promise. Dyes 21 and 22 have color and fastness properties that are comparable to certain commercial dyes. Dye 23 is non-mutagenic and is used as a black dye in ink jet applications.

Isosteric modifications to the phenyl rings of the benzidine nucleus have provided compounds that are much less mutagenic than benzidine (Calogero, et al., 1987). For example, replacement of each phenyl ring with a thiazyl ring provides the non-mutagenic compound 24. Replacement of both the 2 and 2’ carbon atoms with nitrogen atoms within the benzidine nucleus, as represented by 25, greatly reduces mutagenicity (Calogero, et al., 1987).

In fact, it has been found that 25 undergoes tetrazotization reactions similarly to benzidine, and generates nongenotoxic dyes that have essentially identical hues to the same dyes made from benzidine (Calogero, et al., 1987). It remains to be established whether the dyes made from 25 have other satisfactory dyeing properties, but it is clear that the dyes themselves, as well as their cleavage products from azo reduction, are also appreciably less genotoxic than the corresponding benzidine-based dyes. An example is illustrated here with the mutagenic (and carcinogenic) benzidine-based dye Direct Violet 43 (26) and its corresponding analog, 27, in which the benzidine moiety is replaced with 25. Other examples are available (Calogero, et al., 1987).

These promising examples and the wealth of experimental evidence summarized in the preceding paragraphs notwithstanding, there appears to be hesitancy on the part of federal regulatory authorities in the U.S. and other countries in embracing the notion that benzidine analogs such as 15–20, 24 or 25 may quite possibly be commercially efficacious and safer substitutes for benzidine and its carcinogenic congeners (e.g., 12–14). While understandable, given the well-documented toxicity of benzidine, hopefully one day soon substances such as 15–20, 24 and 25, and dyes or pigments made from them will be subjected to more extensive toxicological evaluation. Doing so would better characterize any hazards they may pose.

Conclusion

From the above discussions it is apparent that the design of a new chemical to be commercially efficacious and pose minimal hazard to human health and the environment requires one to be able to integrate toxicity-related information with chemistry considerations. As discussed at the beginning of this paper, most designers of chemicals lack this skill. Change is warranted in the curricula of academic institutions to include and integrate courses on organic chemistry, toxicology and health and environmental sciences.

There is no better time than the present for academic institutions to implement such change. Anastas has been promoting the concept of green toxicology, a new subspeciality within the field of toxicology, that bridges the gap that currently exists between toxicologists and chemists in the design of safer chemicals, especially with the development and application of new types of data such as high throughput screening (HTS) data (Anastas, 2016). HTS data are generated from in vitro assays and provide insight into the interactions that a chemical may have with cellular macromolecules, influences it may have on cell homeostasis, and adverse effects that may ultimately result from pertubations the chemical causes in normal biochemical processes. There is growing interest in the generation and use of HTS data on many diverse chemicals for rapid identification of potential toxicities and the biochemical mechanisms responsible for these toxicities (Kavlock, et al., 2012; Tice, et al., 2013; Hartung and McBride, 2011; Richard, et al., 2016; Shah, et al., 2016).

HTS and other in vitro data are routinely used for early hazard screening purposes during the development of pharmaceutical substances. HTS data have a huge potential as an information source for informing molecular design of commercial chemical substances, and in fact HTS and other in vitro data are used by some chemical companies during the design of commercial chemical substances (Settivari, et al., 2017). Known or potential uses of HTS data include: profiling of chemicals that affect endocrine systems (Reif, et al., 2010), identification of chemicals that have estrogenic properties (Browne, et al., 2015), predicting carcinogenicity (Benigni, 2014), early hazard screening purposes during the development of pharmaceutical substances (Shah and Green, 2014), assessment of cytotoxicities of many diverse chemicals (Grimm, et al., 2015; Rusyn, et al., 2012; Sirenko, et al, 2013; Sirenko, et al., 2014a–b; Shen, et al., 2016), and prioritization (Auerbach, et al., 2016; Hu, et al., 2015), to name a few.

High throughput screening techniques have also been reported for the development of quantitative structure-use models (Phillips, et al., 2017). These models may have great utility in predicting commercial use efficacy of planned chemicals and, if used in combination with HTS toxicity information, in selecting chemicals that appear to be useful and safe (Vanderveen, et al., 2015). High throughput screening techniques are also being used to predict exposure (Egeghy, et al., 2016).

Which existing chemicals should we start with for further prioritization and potential replacement with safer alternatives? DeVito (2016) has recommended chemicals regulated by the U.S. Environmental Protection Agency’s (EPA’s) Toxics Release Inventory (TRI) Program. The TRI database is a rich source of information on many diverse toxic chemicals that are routinely manufactured, processed or otherwise used and released into the environment or otherwise managed as waste by facilities across the U.S. Many of these same chemicals are used commercially and released into the environment or otherwise managed as waste in countries throughout the world.

The TRI chemical list offers an opportunity to apply HTS data being generated by the U.S. federal government’s Tox21 program (Tice, et al., 2013) and the U.S. EPA’s ToxCast Program (Kavlock, et al., 2012). Of the 675 chemicals included on the TRI list of toxic chemicals, 506 have been screened by the Tox21 Program. Of these 506 TRI chemicals, 298 have also been evaluated through EPA’s ToxCast Program (DeVito, 2016). Much insight could be gained from an analysis and integration of available in vivo toxicity data, pharmacokinetic data and HTS data, with the intent of elucidating structure-toxicity relationships and inferring rules that can be used to design new chemicals of reduced hazard that may serve as viable substitutes to the chemicals included on the TRI chemical list.

The design of chemicals to be commercially useful and of low hazard to humans and the environment is not only feasible but also achievable. This is evident by the different approaches and examples described in this article and the references cited herein. Many more examples exist, but many more should exist. Much more could be done, and needs to be done. If Alice Hamilton were alive today she would be pleased to see that since 1928 commercial chemicals which are relatively non-toxic have been identified through advances in chemistry, as well as toxicology, biochemistry, and environmental chemistry. But Dr. Hamilton would probably be disappointed, perhaps shocked, to observe that given these advances much more has not been achieved, and that individuals still receive little to no formal training on how to design commercially useful chemicals that have a high degree of overall safety.

As for the future, a concerted effort must be made by academic institutions, industry and governmental authorities to address the largely unrecognized widespread problem that little formal training is offered in academic institutions on how to design commercially useful chemicals of reduced hazard. Progress in the design of safer commercial chemicals as a subspeciality within modern-day science has been quite slow in comparison to other specialties.

The toxicological chemist is one who will be formally trained in synthetic organic chemistry, biochemistry, physiology, toxicology, environmental science, and in the relationships between structure and commercial use efficacy, structure and toxicity, and structure and environmental fate and effects, and global hazard. The toxicological chemist will be well prepared to work closely and collaboratively with experts in other disciplines to design commercially useful chemicals of low overall hazard.

As we move further into the 21^st century, it is clear that we have the tools and the resources to further unlock the secrets of molecular toxicology and to integrate this knowledge with our understanding of the relationships between chemical structure and properties with industrial application. What is needed now is the collective resolve of individuals and organizations in both the public and private sectors to build the proper infrastructure, use the enormous amount of pharmacokinetic, toxicological, and HTS data with the computational tools now in-hand, and make the necessary changes to effectively implement the concept of designing safer chemicals widely and routinely throughout industry.