5. BIOLOGIC and toxic effects of PCBs

The discussion in these sections is primarily based on the reviews by Parkinson and Safe (1987), Hansen (1987), and McFarland and Clarke (1989). The aspects considered in these sections include (a) molecular structure versus toxicity relationships for individual congeners, (b) toxicity of PCB mixtures, and (c) environmental significance of individual PCBs.

Because of their low solubility's in water, sublethal and chronic toxic effects of PCB contamination in the environment are more likely than acute and lethal effects. Growth, moulting, and reproduction are primary functions that have been shown to be affected by exposure of aquatic organisms to PCBs. The ability of organisms to eliminate foreign organic compounds or endogenous waste products may also be affected. In both fish and higher vertebrates, the metabolic activities such as steroid biosynthesis and the degradation and bio-transformation of foreign organic compounds are strongly influenced by terminal oxidase activities of the microsomal cytochrome P-450 systems. Induction of hepatic microsomal enzymes is one of the earliest and most sensitive indicators of PCB response. Some, although not all, PCB congeners are mixed-function oxidase (MFO) inducers in fish, mammals, and birds, and to a lesser extent in aquatic invertebrates.

5.1 Mixed-function oxidase induction

Liver endoplasmic reticulum (microsomes) contains a family of 12 cytochrome P-450 isozymes (i.e., group of enzymes which are chemically distinct but functionally alike; also referred to as mixed-function oxidase or MFO systems). The function of these isozymes or hemoproteins is to catalyse the bio-transformation of lipophilic xenobiotics (e.g., by hydroxylating, epoxidating, dealkylating or oxygenating, and in some cases by dehalogenating and reducing) to metabolites that are more readily eliminated from the body. However, bio-transformation of xenobiotics by cytochrome P-450 is not always a beneficial process; there are many cases where the metabolites are more toxic or biologically active than the parent compound.

Several cytochrome P-450 isozymes are highly induced by various xenobiotics (e.g., phenobarbital (PB), 3-methylcholanthrene (3-MC), etc.). Frequently, the cytochrome P-450 systems are characterised (e.g., PB-type, 3-MC-type, etc.) by reference to model chemicals that stimulate (induce) or inhibit the production of these enzymes. The group of microsomal cytochrome P-450-dependent enzyme systems that catalyse oxidative bio-transformations of aromatic ring-containing compounds falls in the category of mixed function oxidases (MFOs). The MFOs that are induced by the industrial mixtures of PCBs in vertebrates are characterised as phenobarbital-type (PB-type), 3-methylcholanthrene-type (3-MC-type), or possessing catalysing properties of both (mixed-type). Both PB-type and 3-MC-inducible enzymes have the potential for producing toxicity through bio-activation. However, the potential for contributing to toxicity is greatest with the pure 3-MC-type and mixed-type inducers while PB-type inducers are considered potentially more toxic than weak inducers and non-inducers. The MFOs of fish and apparently of invertebrates, are qualitatively similar to the 3-MC-inducible MFOs of vertebrates. The PB-type induction has been reported in fish (e.g., mummichog, rainbow trout, and carp) only in a few investigations. The enzymes, aryl hydrocarbon hydroxylase (AHH) and ethoxyresorufin O-deethylase (EROD) which are examples of MFO systems, are characteristic in fish as well as in 3-MC-induced mammals (McFarland and Clarke, 1989).

5.2 Molecular structure and PCB toxicity

The toxic reactions of PCBs have been studied with regard to their receptor binding avidities and several receptor-mediated responses, including body weight loss, thymic atrophy, and the induction of the cytochrome P-450-dependent mono-oxygenases, AHH, and EROD. The specificity for MFO induction (and correlatively, for potential toxicity) of the individual PCB isomers and congeners differs greatly and can be related to how closely the PCB isomers approach the molecular spatial configuration and distribution of forces (i.e., are isosteres) of 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD). The dioxin 2,3,7,8-TCDD is one of the most potent synthetic environmental toxicants, and is regarded as a standard for comparison for other organic toxicants that are more or less iso-steric. The cytosolic receptor that binds 2,3,7,8-TCDD is a soluble protein produced by the Ah (aryl hydrocarbon) gene locus. Translocation of the inducer-receptor complex to the nuclear Ah locus is considered to initiate the synthesis of AHH, EROD, and related enzymes that may be involved in either bio-transformation, conjugation and removal, or bio-activation of certain lipophilic foreign compounds.

Quantitative structure-activity relationship (QSAR) studies have shown that PCB congeners with chlorine substitution at both para (4 and 4') and two or more meta (3,3',5, and 5') positions (see Figure 1) are very potent mimics of 2,3,7,8-TCDD both in cytochrome P-450 induction and toxic effects, which are mediated through initial binding to the Ah (aryl hydrocarbon) receptor. These laterally-substituted congeners (a total of 6 congeners; i.e., PCB # 15, 37, 77, 81, 126, and 169- see Table 1) are true coplanar PCBs as they lack the bulky ortho-chloro substituent that restricts free rotation about the phenyl-phenyl bond. Three of the so-called coplanar PCBs, namely 3,3',4,4'-tetra (PCB 77), 3,3',4,4',5-penta (PCB 126), and 3,3',4,4',5,5'-hexachlorobiphenyl (PCB 169) are the most toxic PCBs known. With the exception of PCB # 81, the non-ortho-coplanar congeners are potent inducers of AHH and EROD activities in in vitro rat hepatoma cell preparations.

The addition of a single ortho -chloro substituent to the coplanar PCBs decreases the co-planarity between the two phenyl rings and hence the toxicity of the chlorinated derivatives of coplanar PCBs. This group of mono-ortho coplanar PCBs includes congeners 105, 114, 118, 123, 156, 157, 167, and 189. The loss of toxic potency is accompanied by a qualitative shift in cytochrome P-450 induction; all of the mono-ortho -substituted PCBs are mixed-type inducers (PB-type + 3-MC-type) in contrast to the parent coplanar PCBs, which are predominantly 3-MC-type inducers (except PCB # 81 which is a mixed-type inducer). Among the mono-ortho -chloro substituted PCBs, the group having ortho-chloro adjacent to a meta-chloro substituent (as in 2,3,4- and 2,3,4,5-substituted PCBs) is more toxic than the group with ortho-chloro adjacent to a meta-hydrogen (as in the 2,4- and 2,4,5-substituted PCBs).

The trend toward phenobarbital (PB)-type characteristics and away from 3-methylcholanthrene (MC)-type characteristics continues as two or more ortho -chloro substituents are added to the coplanar PCBs. The di-ortho coplanar PCB congeners (i. e., PCB 128, 137, 138, 153, 158, 166, 168, 170, 180, 190, 191, 194, and 205) are less potent than their non-ortho coplanar and mono-ortho coplanar congeners.

Although there have been few systematic studies, toxicities other than those associated with the Ah locus do not appear to follow structure/activity patterns. Toxic responses unrelated to Ah locus effects and less intensive than those for Aroclor 1254 have been reported for PCB congeners 4, 28, 31, 49, 52, 84, 95, 110, 136, and 153 (Hansen, 1987).

The presence of the most toxic PCB congeners #77, #126, and #169 has been reported in various commercial PCB preparations and tissue samples from a wide range of terrestrial and aquatic (freshwater and marine) organisms, including humans (Tanabe et al.,1987; Niimi and Oliver, 1989; Kannan et al., 1988). However, the concentration of these congeners in water and animals was low, ranging from the non-detectable level (<0.01 ng/L) to (a) 1.0 ng/L in water samples from the Hudson River, N.Y. (Bush et al., 1985), and (b) 0.007 µg/g (wet weight) in the muscle tissue of lake trout (Salvelinus namaycush ; Niimi and Oliver, 1989) or 0.055 µg/g in the fat tissue (wet weight) of killer whale (Tanabe et al.,1989).

Several receptor-mediated responses, as noted above, have been used to study the structure/activity relationships. Table 6 shows the ED₅₀ and EC₅₀ values, respectively, for in vivo inhibition of body weight gain in the immature male Wistar Rat and in vitro induction of AHH and EROD enzyme activities in the rat hepatoma cell cultures (Sawyer and Safe, 1982; Safe, 1987). Both the in vivo mammalian toxicity response and the in vitro induction of the enzyme systems were positively correlated. The results in Table 6 show that 3,3',4,4',5 pentachlorobiphenyl (PCB 126) is the most toxic congener with a potential for toxicity (expressed as TEF or toxic equivalent factor) approaching that of the dioxin, 2,3,7,8-TCDD. Toxic equivalent factors have been used to assess the toxic potential of (a) coplanar PCBs in the environment relative to dioxins and furans (Niimi and Oliver, 1989; Olafsson and Bryan, 1987; Olafsson et al., 1988) and (b) PCB formulations (Kannan et al., 1988). At present, however, the AHH and/or EROD induction data cannot be used for setting water quality guidelines for PCBs. The reasons being:

(i) the data are specific to Wistar rats and are scant on other organisms (e.g., aquatic life),
(ii) the concentrations (in µg/L) and doses (in µg/g body weight) used in these experiments are much higher than those to which organisms are subjected to in the terrestrial and aquatic environments,
(iii) the difficulty in translating AHH or EROD response to a no adverse effect concentration, which is usually obtained from chronic effects or bioaccumulation information. It has been suggested that the increase in the hepatic microsomal AHH activity of fish may be used as an indicator of environmental contamination; but, should a concentration of PCBs in water, however small, be considered toxic (or undesirable) if it induces the enzymatic response in the organism? The linkage between chronic effects and AHH induction is currently unknown,
(iv) lack of relevant information on PCB congeners other than those shown in Table 6. Should other congeners which are structurally unlike coplanar PCBs or 2,3,7,8-TCDD

* EC₅₀ (or ED₅₀) represents half the concentration (or dose) of the toxic congener required to produce the maximum effect; TEF-AHH is the toxic equivalent factor based on AHH induction and is the ratio of the EC₅₀ value for the 2,3,7,8-TCDD divided by that of the PCB congener. TEF-EROD is calculated in a similar manner.
+ The range in AHH-based TEFs indicate that several sources were used to derive these factors.

that do not induce AHH or EROD activities, be considered non-toxic? The discussion in section 7 on Aquatic Life suggests they should not be considered non-toxic, and
(v) lack of relevant information on the toxicity of isomers and individual congeners to organisms using accepted chronic and acute tests. This type of information along with toxic equivalent factors could be useful in the development of criteria for those congeners which behave alike (i.e., showing structure/activity relationships).

5.3 Mixtures of PCB isomers and congeners

Industrial PCB mixtures contain several mono- and di-ortho derivatives of coplanar PCBs. The abilities of these derivatives to induce several forms of rat liver microsomal cytochrome P-450 and to produce a toxic syndrome similar to that caused by TCDD intoxication suggests that the mono- and di-ortho derivatives of the coplanar PCBs contribute substantially to the biologic and toxic properties of industrial PCB mixtures. The presence of relatively more toxic coplanar PCBs such as # 77, 126, and 169 also has been reported in various commercial PCB preparations and tissue samples from a wide range of organisms including marine mammals and humans (Tanabe et al.,1987; Niimi and Oliver, 1989; Kannan et al., 1988).

The environment is contaminated with mixtures of PCB congeners. The variation in the biologic and toxic properties of PCB congeners necessitates a careful assessment of the impact of PCBs in mixtures on the environment. The composition of PCB mixtures changes dramatically in the environment, particularly in the food chain. For instance, it has been shown that the mixture of PCBs secreted in human milk differs from the PCBs in the industrial mixtures that contaminate the environment. The mixture of PCBs in breast milk is considerably enriched in many of the mono- and di-ortho substituted analogues of coplanar PCBs. The preferential metabolism and subsequent excretion of lower chlorinated PCBs and the poor absorption of higher chlorinated (those having seven to ten chlorines) PCBs likely accounts for this relative enrichment. As a result of this enrichment with mono- and di-ortho substituted PCBs, it has been shown that a mixture of PCBs secreted in human breast milk is more biologically active than the industrial PCB mixtures. Obviously, a knowledge of isomeric and con-generic composition in various compartments of the environment would render a more meaningful assessment of the potential risk posed by mixtures of PCBs than total PCB analyses in a single compartment.

There is another important question that must be considered in addressing the toxicity of the mixtures of PCBs; that is, how does one PCB isomer or congener influence the biologic or toxic reaction of another? Two potential mechanisms of synergistic interaction between individual PCBs have been identified. The first involves a sparing effect, whereby a non-toxic PCB congener occupies binding sites (non-specific binding sites, receptors, proteins and/or enzymes), thereby decreasing the removal of a toxic PCB congener. A second possibility involves an interaction at the level of the cytosolic Ah receptor. Both synergistic and antagonistic effects of PCBs in mixtures have been reported in the literature; more work is, however, needed in this direction. For instance, more than additive response (measured as induction of hepatic microsomal benzo(a)pyrene hydroxylase and EROD activities) was noted in immature male Wistar rats pre-treated with 2,4,5,2',4',5'-hexachlorobiphenyl and followed by an exposure to 3,4,4,3',4'-penta or 3,4,5,3',4',5'-hexachlorobiphenyl (Parkinson and Safe, 1987). On the other hand, less than additive response (based on observed AHH and EROD activities for individual congeners) was noted in Wistar rat hepatoma cells exposed to several environmentally significant reconstituted mixtures of PCBs (Sawyer and Safe, 1985).

Recently, assessments of the effects of PCBs relative to furans and dioxins on aquatic environments have been carried out by several investigators based on AHH or EROD enzyme-inducing capabilities of various toxic congeners (Niimi and Oliver, 1989; Bryan et al., 1987; Kannan et al., 1988). This procedure requires that the AHH-inducing capability of all congeners be known. However, not all PCB congeners are AHH or EROD inducers, and non-inducer congeners seem to be very toxic as well (Bridgham, 1988). Also, some results of the recent investigations have indicated that the PCB mixture Aroclor 1254 is a dioxin antagonist in rat hepatoma H-4-II E cells, in vitro, and in C57BL/6J mice, in vivo, if administered at non-effective (below toxic threshold) doses (McFarland and Clarke, 1989).

5.4 Carcinogenic and mutagenic effects

Evidence relating to the carcinogenicity of PCBs can be arbitrarily considered in three main categories: (i) influence on mutagenicity and initiation of carcinogenesis, (ii) influence on promotion and progression of carcinogenesis, and (iii) epidemiological evidence for carcinogenesis in naturally exposed populations. The majority of evidence comes from experimental studies, and its value depends upon the quality of various studies and the conceptual framework within which the studies have been designed and interpreted.

A number of rats exposed to dietary levels of 25 to 100 µg/g PCBs (e.g., Aroclor 1254 and 1260) were found to develop various types of cancer (Kimbrough et al., 1975; NCI, 1977; Weltman and Norback, 1981; Kimbrough, 1985). The US Food and Drug Administration (FDA) has used this information to determine the risk to humans from consuming fish contaminated with PCBs. However, the use of such information for risk analysis has been criticised for several reasons which include: (a) the FDA methodology employed high doses (25 to 100 µg/g in diet) in experiments on animals and extrapolated the observed rates of certain types of cancer at these elevated doses to the low doses found in human diets, and (b) the FDA assumed that humans and test animals are equally sensitive to PCB ingestion when measured on a µg/g in diet basis. Extrapolation on an equivalent consumption per unit of body weight is preferred and results in much lower health risks (Maxim and Harrington, 1984).

Kimbrough (1985) reviewed laboratory and human studies on PCBs and related compounds. It was noted that three attempts were made to determine whether exposure to PCBs (in capacitor and refinery plants) caused an increased incidence of cancer. In all cases, the results were inconclusive for reasons of (a) small sample size, (b) interference from other chemicals, and (c) short-term exposures. It was concluded that, in humans, adequate studies were not conducted to judge if the long-term exposure to PCBs was associated with cancer and reproductive impairment (in exposed females).

In reviewing the literature on carcinogenicity of PCBs, Hayes (1987) reached the following conclusions:

PCBs appear to be, at the worst, very weak genotoxicants or initiators of carcinogenesis in various systems. Their well established activity at moderately high levels as promoters of hepatocarcinogenesis (or liver cancer) in rodents should not necessarily be accepted as clear predictive evidence for a similar effect in humans. Many xenobiotics with no known hepatocarcinogenic effects in human are, like PCBs, strong promoters of liver tumour growth in rodents that appear inherently highly susceptible to this response. Evidence that PCBs enhance genotoxicity and mutagenicity of many other xenobiotics in various in vitro test systems is in direct contrast to their protective role against carcinogenicity of many genotoxic carcinogens in vivo.. Furthermore, PCBs should not be universally regarded as liver tumour promoters because they strongly prevent the promoting activity of other environmental carcinogens such as dimethylnitrosamine, diethylnitrosamine, azo dyes, 2-acetylaminofluorene, and aflatoxin B1.

Collectively, the evidence suggests that PCBs may potentially be carcinogenic under some specific conditions. However, under natural exposure circumstances, PCBs are perhaps more likely to prevent carcinogenesis than enhance it. Accordingly, the risk estimates based on the worst-case analysis of the potentially carcinogenic effects are likely to be substantial overestimates of the real risk.

While PCBs are clearly toxic and hazardous to humans exposed accidentally to high levels, there is a lack of clear epidemiological evidence for carcinogenicity of environmental PCBs in humans and animals. However, the number of thorough epidemiological studies conducted is small. More discriminating studies of PCBs and other potential carcinogenic risk factors are necessary before it would be reasonable to conclude that PCBs are not human carcinogens.

5.5 Environmental significance of PCB congeners

McFarland and Clarke (1989) selected 36 PCB congeners which were considered to be of environmental significance based on three factors: (a) potential for toxicity (inferred by MFO induction; PB-type inducers were considered less toxic than pure 3-MC-type and mixed-type inducers, but more toxic than weak and non-inducers), (b) frequency of occurrence in environmental samples, and (c) relative abundance in animal tissue (determined from a PCB congener database developed from information reported in the scientific literature) (Table 7). Group 1 contains the highest priority congeners which are most likely to contribute to adverse biological effects. The subgroup 1A contains the three most potent congeners (pure 3-MC-type inducers: PCB #77, #126, and #169); although these congeners have been reported rarely in the environment, they have been identified as components of technical formulations. Subgroup 1B congeners (PCB #105, #118, #128, #138, #156, and #170) are mixed-type inducers that (except PCB #105) have been reported frequently in the environment. The seven congeners in Group 2 are PB-type inducers with high relative abundance in avian and mammalian tissue, but they are found to a lesser extent (as little as 1.5%) in fish and invertebrates. Except for congeners #77 (a tetrachlorobiphenyl) and #194 (an octachlorobiphenyl), all of the congeners in Group 1 and 2 are member of the penta-, hexa-, and heptachlorobiphenyl isomer groups.

* Congeners included in Canadian Standard CLB-1, developed under the Marine Analytical Standards Program by the Atlantic Research Laboratory, NRC, Halifax, Nova Scotia. The remaining congeners making up CLB-1 are nos. 15, 31, 40, 54, 60, 86, 103, 121, 129, 137, 141, 143, 154, 159, 171, 173, 182, 185, 191, 195, 196, 200, 202, 203, 205, 206, 207, 208, and 209.
+ Congeners suggested for inclusion in human foodstuff and tissue analyses by Jones (1988). Other congeners listed are 8, 28, 60, 66, 82, 166, 179, and 187.

Group 3 congeners are weak or non-inducers but occur frequently, particularly in fish and invertebrate tissues. This group spans the isomer groups from tri- through octachlorobiphenyl. Group 4 congeners are mixed-type inducers. Although scarce in the environment, they are considered to be of possible concern because of their potential toxicity. McFarland and Clarke suggested that toxicologically relevant evaluations of PCB-contaminated environmental materials can be better accomplished by analysing samples for specific congeners in the four groups as compared to total PCB or as Aroclor equivalent. The congeners in Groups 1 and 2 were considered to be the most environmentally threatening.

In studying PCB congener levels in the Lake Ontario ecosystem, Oliver and Niimi (1988) observed that twelve PCB congeners (#153, #101, #84, #138, #110, #180 #, #87, #97, #149, #187, #182, and #105) constituted over half the PCBs in fish. These congeners were suggested as a focus for research on health effects. The authors also concluded that congener-specific PCB analysis is not required for all applications, but that it is vital for pathway research.

Working with a data set which was largely different from that of McFarland and Clarke (1989), Jones (1988) derived a list of 32 congeners which were considered to be environmentally important. With the exception of PCB 194, the list included all of the congeners assigned to Groups 1 and 2 in Table 7. Jones' list also included 10 of the 20 congeners assigned to Groups 3 and 4 in Table 7.

5.6 Congener-specific analysis of PCBs

The discussion in this section is based on material presented in McFarland and Clarke (1989) and Safe et al. (1987).

In general, the analysis of PCBs in environmental and human samples is dependent on gas chromatograph (GC) peak- or pattern-matching of commercial mixtures with the specific PCB extract under investigation. However, the composition of PCB extracts from diverse environmental matrices can vary widely and such extracts do not resemble any commercial mixture. Therefore, the results obtained from the pattern-matching technique are at best a semi-quantitative estimate of the PCBs present in a sample, and yield very little indication of the relative concentrations of the individual congeners.

Isomeric- or congener-specific PCB analyses have been performed using the high-resolution capillary gas chromatography technique. The individual peaks are identified by using synthetic standards and/or by retention index addition methods (Ballschmiter and Zell, 1980). The latter technique relied on the relative retention times (RRTs) that have been determined for the limited number of available synthetic PCB standards. However, the accurate quantitation of the individual PCB components in a mixture can only be accomplished by comparing the observed relative retention time (RRT) and peak height (or area) data for a PCB-containing extract and the RRT and molar (or weight) response factor of all the PCB standards.

Recently, the unambiguous synthesis and chromatographic properties of the 209 PCBs have been reported by Mullin et al.(1985). These workers successfully separated 187 of the 209 PCB congeners by capillary gas chromatography. Eleven pairs (PCB 94/61, 70/76, 95/80, 60/56, 145/81, 144/135, 140/139, 133/122, 163/160, 202/171, and 203/196) exhibited similar retention times using a SE-54 coated glass capillary column. Some of these pairs, however, can be resolved chromatographically despite their comparable RRT values. Also, some of these compounds were not present in commercial PCBs.

An analytical instrument calibration standard for PCB congener-specific analysis by capillary column gas chromatography has been developed under the Marine Analytical Standards Program by the Atlantic Research Laboratory, National Research Council of Canada, Halifax, Nova Scotia (McFarland and Clarke, 1989). The Canadian Standard Mixture, CLB-1, contains 51 congeners which include 13 of 16 congeners in Groups 1 and 2, and 9 of 20 congeners in Groups 3 and 4 (Table 7). Not included are PCBs 126 and 169, two of the three most toxic congeners in Group 1A.