A Decade of Research on the Environmental Impacts of Pulp and Paper Mill Effluents in Canada (1992-2002)
- Publishing Information
- 1.0 Executive Summary
- 2.0 General Information
- 3.1 Field Studies and Mechanistic Research - Summary
- 3.2 Canadian Research Leading Up to the 1992 Pulp and Paper Regulatory Package
- 3.3 Research Program to Identify the Causative Compounds, How to Eliminate Them, and Determine Their Short and Long-Term Environmental Effects
- 3.4 Evolution of the Research Questions
- 3.5 Evolution of the Research Questions: Monitoring Sites over the Long-term for Evidence of Recovery Following Process and Treatment Changes.
- 3.6 Evolution of the Research Questions: Need to Identify Process and Treatment Changes Responsible for Partial Recovery and Chemicals Involved
- 3.7 Evolution of the Research Questions : Cycle 2 EEM Results, What Were the Major Response Patterns and How Widespread Were They?
- 3.8 Conclusions
- 4.1 Development and Application of Bioassays - Summary
- 4.2 History
- 4.3 Mesocosms
- 4.4 Lifecycle Studies
- 4.5 Conclusions
- 5.1 Characterization of Bioactive Chemicals - Summary
- 5.2 Introduction
- 5.3 Causal Investigations of Bioactive Substances
- 5.4 Characteristics of bioactive substances revealed during field and laboratory studies
- 5.5 AOX: Regulation and relationship to effects
- 5.6 Effluent and Receiving Environment Chemistry
- 5.7 Conclusions
- 6.0 References
5.3 Causal Investigations of Bioactive Substances
Numerous studies have attempted to address the paucity of information regarding the identities of bioactive substances present in final effluents. Although a variety of effects in fish have been associated with exposure to effluents in the field (See Field Studies and Mechanistic Research) and under controlled conditions in the laboratory or on-site (See Development and Application of Bioassays), the specific chemicals associated with the array of reproductive responses have proven extremely difficult to identify. Although effluent constituents such as b -sitosterol (MacLatchy & Van Der Kraak, 1995; Tremblay & Van Der Kraak, 1999), abietic acid, pinosylvin and betulin (Mellanen et al., 1996) have the potential to affect fish reproduction when tested individually, definitive cause and effect relationships have not been established because of effluent complexity, differences in species response patterns (e.g. between laboratory species and wild fish), and a lack of information on the mechanisms of action (Van Der Kraak et al., 1998).
Research in the area of compound identification has progressed on a number of fronts. Work in the early 1990s was focused on the identification of substances associated with cytochrome P4501AI induction (measured as ethoxyresorufin-O-deethylase or EROD activity; Figure 5). Work focused on this area because little was known about the characteristics of reproductive dysfunction in wild fish exposed to effluents and sources of acute toxicity had already been attributed to resin acids and guaiacols (Kringstad & Lindstrom, 1984), the levels of which are substantially reduced after secondary treatment.
Subsequent work from the mid 1990s to the present day have attempted to address the more complex issue of reproductive effects in wild fish (Figure 5), and those responses have now been conclusively demonstrated on a national scale through the Environmental Effects Monitoring (EEM) program (Lowell et al., 2003; Munkittrick et al., 2002a). Causal investigations of reproductive effects ranged from pure compound exposures, investigations of individual waste streams within the manufacturing process to Toxicity Identification Evaluations (TIEs). The TIE approach has been adopted by many researchers to isolate and identify chemicals associated with EROD induction. There has also been development of tests using fish and cell lines to drive chemical fractionations to investigate the causative agents (See Development and Application of Bioassays). Although ecologically relevant, reproductive dysfunction in wild fish has represented a much more difficult problem to address because the mechanisms are not understood. The responses have shown effects on gonad size, depressed levels of circulating steroids (Munkittrick et al., 1998), perturbations in the sex steroid biosynthesis pathway (McMaster et al., 1995a) and effects on gonadotropin production and peripheral sex steroid metabolism (Van Der Kraak et al., 1992), indicating multiple mechanisms and chemicals are involved. In the late 1990s development of suitable bioassays, such as fish-specific sex steroid receptor assays (Van Der Kraak & Biddiscombe, 1999; Wells & Van Der Kraak, 2000), life cycle tests (Parrott et al., 2003a) and short- term in vivo tests for steroid effects (Dube & MacLatchy, 2001a; Hewitt et al., 2002) has provided the opportunity to couple mechanistically linked endpoints to chemical fractionations in the investigation of the identities of the causative substances. This has led to the ability to formulate questions regarding the characteristics of bioactive substances, their relationship to production type and whether compounds associated with sex steroid depressions are related to other reproductive impacts (Figure 5).
5.3.1 TIE studies
TIEs, developed by the US EPA, are an approach by which the active substances of interest in a complex matrix are characterized in three phases (USEPA, 1991, 1993a,b, 1997). The TIE approach is based on bioassay-directed chemical fractionations to identify unknowns and was developed for municipal sewage investigations in concert with Toxicity Reduction Evaluations (TREs) to ameliorate effluent acute and chronic toxicity. Phase I of a TIE involves determining the broad characteristics of the active agents through manipulations of the effluent. Phase II involves specific methods to isolate the active chemicals and propose structures for their identification. Phase III involves techniques to confirm that the proposed substances are in fact responsible for the observed toxicity.
Bioassay-directed effluent fractionations and TIE studies have been employed to isolate and characterize compounds associated with EROD induction in fish exposed to effluents (Burnison et al., 1996, 1999; Hewitt et al., 1996; Martel et al., 1997). In the early 1990s there was much speculation that polychlorinated dibenzo-dioxins and dibenzofurans (PCDD/DFs) present in final effluents as well as in biota from receiving environments were solely responsible for EROD induction observed in fish exposed to effluents (Figure 5). Although the physiological consequences of EROD induction were intensely debated (and still are), work was driven by the fact that enzyme induction represented exposure to compounds with the potential for dioxin-like toxicity. However, in the last decade an increasing body of evidence has associated non-dioxin classes of chemicals interacting with the Aryl Hydrocarbon Receptor (AhR). This evidence is collectively derived from i) the demonstration of EROD induction in fish after removal of defoaming agents from the manufacturing process that contained dioxin precursors, ii) the persistence in EROD induction following alterations in bleaching processes that do not promote polychlorinated dioxin formation, iii) transient EROD induction in fish upon cessation of exposure which is not characteristic of the sustained nature of induction associated with PCDD/DFs, iv) decreases in PCDD/DF levels in fish tissues collected following mill process changes, v) discrepancies between EROD induction and dioxin equivalents in fish measured in cell lines suggested that additional compounds may be involved (van den Heuvel et al., 1996) and vi) several studies which eventually associated EROD activity with non-dioxin effluent constituents. It is important to note that in the causal investigations conducted in the last decade, the total EROD activity of the effluents has not been accounted for and additional AhR ligands presently remain unidentified.
In the course of TIE studies conducted over the last decade, characteristics of the causative agents and selected individual compounds causing MFO induction have been identified (Figure 6). Hewitt et al. (1996) fractionated effluents before and after treatment and after a maintenance shutdown at a bleached kraft mill as one of the first studies to address the role of secondary treatment in affecting EROD activity. Laboratory rainbow trout were exposed to treated and untreated effluent, whole and filtered (<1 m m) effluent, resuspended solids and two fractions of effluent that had been generated by nanofiltration. It was assumed that a comparison of relative EROD activity levels in the different effluents with chemical levels in the samples would provide insight into correlations of chemicals with the biological responses. For example, any chemicals found in whole effluent and in a fraction that was not associated with EROD activity could be eliminated as candidate inducers. Similarly, any chemical removed in a fraction or sample that retained EROD induction, could also be eliminated as a potential candidate. These analyses eliminated resin aids, fatty acids, bacterial fatty acids, terpenes, chlorophenolics, aliphatic alkanes, plant sterols and chlorinated dimethylsulphones as candidates. Although the comparisons found correlations of EROD activity with tetrachloroguaiacol, 4,5,6-trichlorotrimeth-oxybenzene and 2,4,6-trichloroanisole, only tetrachloroguaiacol was substantially above detection limits. Subsequent exposures confirmed that tetrachloroguaiacol did not cause induction, but Hewitt et al. (1996) concluded that the correlations might indicate the potential source of the compounds. Phenolics are among the major products of the oxidation of residual lignin during bleaching. When chlorine is used in bleaching, some of these phenolics may become chlorinated. The observed correlation with lignin-derived phenolics suggested that the inducers might also be lignin-derived.
Burnison et al. (1996) attempted to directly isolate chemicals inducing EROD activity in fish by following a classical TIE approach on final effluent from two bleached kraft mills located in Ontario. Using centrifugation, tangential flow filtration and C18 solid phase extraction (SPE), effluents after secondary treatment were investigated using a 4 d rainbow trout in vivo bioassay. It was determined that methanol extracts of particulates/colloidal material and SPE fractions contained active substances. Work focused on the particulate material and showed that activity could be isolated using HPLC. HPLC isolations determined that the active substances were present in a relatively nonpolar region of the chromatographic separation, with a logarithmic octanol/water partition coefficient (Kow) of 4.6-5.1. As a result of follow-up studies using rainbow trout exposures and incubations with a rat hepatic carconoma cell line (H4IIE) which directed HPLC fractionations of the methanol extract of the high molecular weight material, a chlorinated pterostilbene structure was postulated for an unknown compound strongly associated with induction (Figure 6; Burnison et al., 1999). This was significant in that it showed a natural product, modified in the bleach plant, was eliciting the biological response.
In a parallel line of work, Martel et al. (1997) determined the source and identities of two subs tances associated with induction present in the primary-treated effluent of a newsprint thermomechanical pulp (TMP) mill. To determine the sources of activity within the mill, rainbow trout were exposed static for 96 h to TMP condensate, deinking, and paper machine effluents and TMP whitewater and tested various process effluents sampled throughout the mill. Exposure concentrations were based on the flow of these process streams in relation to final effluent. Contaminated TMP steam condensates were identified as the major process source of EROD-inducing substances. Using conventional extraction, silica gel fractionation and preparative thin-layer chromatography (TLC) procedures, an EROD-inducing fraction was isolated. The major constituents were identified by gas chromatography/mass spectrometry (GC-MS) as juvabione, dehydrojuvabione, and manool, all naturally occurring extractives in balsam fir (Abies balsamea). After extraction and isolation from balsam fir and TMP condensates, trout exposed to juvabione and dehydrojuvabione exhibited significant hepatic EROD induction (Figure 6). This study further determined that secondary treatment in an activated sludge system effectively eliminated the EROD-inducing potential of the combined mill effluent consistent with a corresponding 90% reduction of both juvabione and dehydrojuvabione.
While the aforementioned study on TMP effluents appeared to have identified the source and identities of chemicals elevating EROD activity in fish, other researchers in Canada have investigated the causative agents at mills employing different pulping processes and wood species (Figure 5). One of the main reasons for doing so was the observation that EROD induction and reproductive dysfunction in wild fish persisted after installation of secondary treatment at a well studied bleached kraft mill (Munkittrick et al., 1992b).
Further studies of final effluents at other sites have proven problematic. Difficulties encountered include: i) fractionation experiments conducted on "grab" samples of effluent do not reflect temporal fluctuations in active chemicals, ii) toxicological potencies of effluent samples were influenced by sample handling and storage conditions, iii) the large amount of high molecular weight lignin material proved a significant interference when investigating low molecular weight extractives, iv) the complexity of low molecular weight effluent extractives and v) uncertainties regarding the bioavailability of identified bioactive components.
In the late 1990s, a different approach was adopted to address some of the obstacles associated with investigations of final effluents (Figure 5). Parrott et al. (2000c) used caged fish to investigate the uptake of AhR ligands from effluent from a bleached kraft mill. Ligands were recovered in two solvents used and non-dioxin ligands were found in tissue extracts using EROD induction in H4IIE cells as the indicator. This led to the development of a bioaccumulation model to investigate bioactive substances in complex mixtures. In these investigations the approach has been to focus on what is bioavailable to the organism by using controlled exposures to investigate bioactive substances in tissue residues (Hewitt et al., 2000a, b, 2003).
The bioaccumulation model was first developed at a bleached kraft mill that has been associated with reproductive dysfunction in wild fish since the late 1980s. Both naïve white sucker (Catostomus commersoni) and sucker collected adjacent to the effluent outfall were held in a concentrated effluent stream (50% v/v) for 4 d. Hepatic tissue extracts from exposed fish were fractionated according to lipophilicity using HPLC. Fractions with different octanol/water (Kow) partition coefficients were tested for the presence of bioavailable chemicals that function as ligands for the AhR in H4IIE cells, rainbow trout hepatic estrogen receptors (ER), goldfish testicular androgen receptors (AR) and goldfish sex steroid binding protein (SSBP). It was shown that fish rapidly accumulate multiple non-dioxin ligands for the AhR and fish sex steroid receptors after a 4 d exposure (Hewitt et al., 2000a, 2003). PCDD/DF equivalents measured by EROD activity in H4IIE cells and by high resolution GC-MS showed that in all fish historically exposed to effluent, the contributions to total toxic equivalents (TEQs) from TCDD was >80% and that in naïve fish held in effluent accumulated 1,2,3,7,8-pentachlorodibenzo-furan which accounted for a major portion of TEQ (Hewitt et al., 2000b). This study also showed that when fish normally residing in the effluent plume leave for a brief period to spawn in an uncontaminated stream, hepatic burdens of ligands for the AhR and sex steroid receptors decrease to the levels found in fish at reference locations. For wild populations historically exposed to bleached kraft mill effluent this suggests that a sustained exposure is required to maintain tissue concentrations. This study provided a plausible mechanistic linkage between uptake of ligands for sex steroid receptors and effects in wild fish and provided evidence of multiple effluent compounds functioning as ligands for multiple biological receptors.
A follow-up study at a bleached sulfite/groundwood mill was conducted to determine if the accumulation profiles of bioavailable substances were related to process type (Figure 5; Hewitt et al. 2003). White sucker again accumulated ligands for each receptor after 4 d exposure to effluent, and the pattern of accumulated substances was very similar to that previously obtained at the bleached kraft mill. Accumulated ligands were evaluated after exposure of fish to effluent for two different durations and following a depuration period. Hepatic EROD activity and whole-liver hormonal activity, measured as binding to SSBP, returned to background following 6 d depuration and were reduced but still significant after 12 d exposure to effluent. Whole-liver extract affinities for the AR were maintained after extended exposure and depuration, indicating the potential for AR ligands to bioaccumulate. By conducting these experiments at a bleached sulfite mill, this study provided further evidence that ligands for sex steroids and the AhR are associated with effects in wild fish.
5.3.2 Studies with pure compounds
In addition to systematically investigating effluents and waste streams using TIE approaches, researchers have followed hypotheses regarding effluent constituents with the potential to affect EROD activity or reproductive function in fish. A series of investigations have focused on retene (alkylated phenanthrene), an anaerobic degradation product of abietic and dehydroabietic acid, and its ability to induce EROD (Figure 6). Both parent chemicals are resin acids, common coniferous wood extractives, which are a family of acidic compounds commonly found in effluents. Resin acids are reduced, but not eliminated during effluent treatment and have historically been associated with effluent acute toxicity. Retene forms in treatment ponds and in sediments within waterways receiving pulp mill effluents. Retene has been found to cause elevated EROD activity in fish (Fragaoso et al., 1998), which was not sustained following transfer to clean water and is consistent with field observations (Munkittrick et al., 1992b). Spiking studies and studies with sediments collected in the vicinity of a bleached kraft mill contaminated with retene showed the bioavailability of retene and additional AhR ligands (Oikari et al., 2002). Further studies have shown that sediment retene is bioavailable and metabolites are detectable in bile of exposed fish (Billiard et al., 2000). Prolonged exposure results in a variety of dioxin-like toxicities to larval fish, including blue sac disease.
In the past six years, several researchers have hypothesized wood components are involved in reproductive responses (Figure 5). The focus on wood products has been based on the observation that reproductive responses in wild fish occur downstream of mills with very disparate processes (Munkittrick et al. , 1998, 2002a), implying the effects are related to chemicals common to pulp mills, e.g., the wood furnish. This hypothesis is also supported by studies which have linked bioactivity in digest wastes, black liquor and bleachery process streams to lignin degradation products (see Process Streams and Conference Summaries below). When tested individually, b-sitosterol, the dominant plant sterol consistently measured in final effluents, exhibits estrogenic activity in vitro and can induce vitellogenin production and reduce plasma sex steroids in vivo (MacLatchy & Van Der Kraak, 1995; Tremblay & Van Der Kraak, 1999). Estrogenic potential has also been associated with abietic acid, pinosylvin and betulin (Melanen et al., 1996) and the flavonoid genistein (Pelissaro et al., 1991), which has recently been detected in bleached kraft effluent (Kiparissis et al., 2001). Studies with chemical recovery condensates have associated condensate components chemically resembling kraft lignin with testosterone depressions in fish (Hewitt et al., 2002). The confounding aspect of studies using pure compounds is that while effluent concentrations of many of these constituents are above threshold levels for effects, the same responses are not observed following effluent exposures. Suspected reasons for this involve response differences between laboratory and wild species, response pathways affected by other effluent constituents (Van Der Kraak et al., 1998) and interactions between effluent constituents.
In an extension of the work on sex steroid ligands accumulated by fish, in the late 1990s effluent samples were collected from 10 mills across Canada of various pulping and effluent treatment technologies to determine any relationship between pulping process, treatment systems and wood furnish (Figure 5; McMaster et al., 2002b). Ligands for fish sex steroid receptors were detected with higher activities observed for sex steroid binding protein and the androgen receptor, which reflects earlier observations in bioavailability studies (Hewitt et al., 2000a, b; 2002). Additional studies have shown the potential for effluents from several mills to exhibit antiestrogenic activity when extracts were co-incubated with estradiol in primary cultures of rainbow trout hepatocytes (Marlatt, 2003). Ligands for fish retinoic acid receptors have also been detected in effluent extracts from half of the mills surveyed, with high activity associated with effluents from thermomechanical facilities (Alsop et al., 2003). Active extracts were further partitioned into acidic and base-neutral fractions and it was found that most of the retinoic acid activity was associated with acidic components. At the present time it is not clear if ligands for retinoic acid receptors are also associated with depleted retinoid stores in fish collected from receiving environments but a mechanistic linkage has been established that can be addressed in future studies. In these series of effluent studies using in vitro tests there were no correlations with production type or treatment and it is unclear what effects, if any, conventional effluent treatment is having on levels of hormone ligands in effluent.
More recent studies on wood furnish have also attempted to address the question of whether bioactive substances can be isolated from wood chips prior to pulping (Figure 5). This question has arisen from previous work isolating natural products associated with EROD activity (Martel et al. 1997; Burnison et al. 1999) and early work which examined the effects of wood furnish on toxicity. When O'Connor et al. (1992) varied the wood furnish, simulated mechanical pulping effluents varied in acute and chronic toxicity. Chronic toxicity was measured using Ceriodaphnia reproduction and balsam fir effluents had the highest potency. Recent investigations have since determined that chemicals functioning as sex steroid ligands can be extracted using two different solvents from wood chips used in pulp manufacture (McMaster et al., 2002b). Ligands for fish retinoid receptors are also present in extracts of wood furnish (Alsop et al., 2003). Crude extracts were tested in these initial studies and the active compounds presently remain unidentified.
5.3.3 Process Stream Investigations
Beginning in the mid 1990s, several researchers have investigated individual waste streams within the papermaking process to determine the source(s) of EROD induction and compounds affecting steroid levels in fish (Figure 5). The principal challenge in these investigations has been to avoid acute toxicity associated with individual wastes, particularly those from the bleach plant. Researchers have primarily used in vivo fish tests in these investigations (See Development and Application of Bioassays) and the results will be summarized here in the context of what knowledge was gained on the sources of bioactive substances.
Black liquor was the subject of investigations involving EROD activity and hormonal endpoints. The pulping process digests lignin, the complex phenolic polymer that binds cellulose fibres together. The spent cooking liquor, known as black liquor, contains the degradation products of lignin and cellulose as well as wood extractives such as resin and fatty acids. Zacharewski et al. (1995) found that the methanol extract of black liquor particles and colloids > 0.1 m m from a bleached kraft mill contained AhR ligands which also displayed antiestrogenic effects via the AhR in vitro . Hodson et al. (1997) investigated the potential of black liquor from hardwood and softwood pulping at a bleached kraft mill to induce EROD activity in rainbow trout and found significant activity. A higher potency liquor was associated with alcohol digestion of wood chips as well as solvent extracts of wood.
EROD activity of final effluent from a bleached kraft mill with 60% ClO2 substitution was found to be reduced by > 90% using activated sludge treatment and an investigation within the mill showed bleach plant effluent to be the major contributor (Schnell et al., 2000). Follow-up work on bleach plants at two kraft mills in central Canada involved testing alternating chlorination (100% ClO2 ) and alkaline extraction filtrates to determine which stages were associated with production of AhR ligands (Coakley et al., 2001). Filtrates were found to increase in potency through the stages of bleaching and the final bleaching stage was the most potent. Softwood filtrates also possessed more potency than hardwood filtrates, lending credence to the theory that the source of AhR ligands is ultimately related to wood furnish. The authors speculated that recycled wash water from the paper mill was a source of compounds in filtrates (Coakley et al., 2001).
As mentioned previously (5.3.1 TIE studies), process waste evaluations were used to identify candidate streams for TIE investigations at a thermomechanical mill (Martel et al., 1997). In that study, waste streams were evaluated flow-proportionally and steam condensates were ultimately identified as the major source of compounds causing elevated levels of EROD in rainbow trout. Specific chemicals, juvabione and dehydrojuvabione, were confirmed as AhR ligands (Martel et al., 1997).
In the late 1990s a systematic waste stream approach was applied to investigate sources of chemicals with the ability to affect sex steroid levels and/or production in fish at 3 other pulp mills. Extensive investigations on waste streams within two mills were conducted to determine their potential to elicit effects on circulating steroids in fish (Parrott et al., 2000a). Effluents before and after treatment were evaluated at a bleached kraft mill (18 streams) and a bleached sulfite mill (14 streams). In both cases, individual process wastes within the mill did not affect steroid levels or steroid production in goldfish (Carassius auratus) but final effluent from both mills after secondary treatment did cause significant steroid depressions. It is also interesting to note that final effluent from the bleached sulfite mill did not affect steroid levels after process changes that included i) increased ClO2 substitution from 60-65%, ii) a reduction in solids losses from the bleach plant, iii) reduced liquor losses through spill management and iv) increased aeration within secondary treatment. It was unclear which of the process changes might be associated with the lack of steroid effects (Parrott et al., 2000a).
An extensive investigation was conducted at a bleached kraft mill in Saint John NB , which is one of a handful of pulp mills in Canada that does not employ secondary treatment. At the Saint John mill the study involved systematic exposures of mummichog (Fundulus heteroclitus) to 5 in-mill process wastes to determine the waste stream source(s) contributing to depressed sex steroids associated with exposure to final effluent. These exposures were first conducted in a field-based, mobile, artificial stream system (Dubé & MacLatchy, 2000a, b) and later confirmed with laboratory studies (Dubé & MacLatchy, 2001a). This work successfully resulted in the identification of chemical recovery condensates as a primary source of substances that depress circulating sex steroids in fish. Reverse osmosis treatment of condensates was also conclusively proven to remove the active substances prior to their re-use in brownstock washing and dilution before bleaching (Dube & MacLatchy, 2001b). RO treatment resulted in a non-acutely lethal final effluent and the sublethal toxicity of the final effluent was reduced in three different marine species (Dube & MacLatchy, 2000b).
Following the identification of condensates as a source of endocrine disruptors, TIE studies were initiated at the Saint John Mill to investigate the identities of the causative chemicals (Figure 5). Minimal high molecular weight material was found in the condensates, facilitating bioassay-directed fractionation studies (MacLatchy et al., 2001). Using steroid depressions in mummichogs, a two-step solid phase extraction (SPE) method was developed which completely recovers the active chemicals from the condensates in two fractions (Hewitt et al., 2002). Differences in the chemical composition and bioactivity of condensates generated during hardwood and softwood production were observed, which suggests a linkage to chemicals derived from wood furnish. Results also indicate that the responsible chemicals are polar, water soluble and bioavailable, which is supported by the steroid depressions that can be induced in vivo after a 7-d waterborne exposure. GC-MS profiles of both fractions revealed relatively simple mixtures of < 20 chemicals and the mass spectra of several unknowns appeared to be consistent with lignin degradation products (Hewitt et al., 2002). These findings are consistent with earlier studies which showed the onset of steroid perturbations in wild fish to be rapid and associated with multiple lesions in the steroid biosynthetic pathway (McMaster et al., 1995b).
- Date Modified: