Industrial emissions and other urban air pollutants pose a major threat to human health. Ozone, nitrogen dioxide, particulate matter and sulfur dioxide (SO₂)⁴ have all been implicated in causing damage to pulmonary tissues and in initiation of asthmatic symptoms (Gong 1992). SO₂ Is a highly injurious agent which, in contact with moist alveolar membranes, will form sulfuric acid. Regulation of SO₂ production has reduced concentrations in urban air to approximately 1.4 µg/m³, but local variations and occupational exposures may result in asthmatics and other susceptible individuals suffering SO₂-related pulmonary injury (Gong 1992, Stjernberg et al. 1985, Tewari and Shukla 1991).

Within the lung SO₂ is detoxified through the sulfitolysis of oxidized glutathione (GSSG) (Kagedal et al. 1986, Mannervik et al. 1974). GSSG is generated from reduced glutathione (GSH) through the action of glutathione peroxidase (GPx) on free radical species (Lawrence and Burk 1976). Sulfitolysis of GSSG and the subsequent elimination of sulfitolysis products in the urine, prevents the regeneration of GSH through glutathione reductase (GRed) (Mannervik et al. 1974, Winell and Mannervik 1969) and results in the depletion of GSH from all tissues (Langley-Evans et al. 1996). This may leave individuals more susceptible to tissue injury caused by secondary agents, or to long term pulmonary disorders associated with chronic gas exposure. Chronic exposure to low SO₂ concentrations is implicated in the development of childhood asthma (Tseng and Li 1990).

Previous work with rat models in our laboratory has demonstrated that glutathione metabolism is determined in utero by aspects of the maternal diet (Langley et al. 1994, Langley-Evans et al. 1995). Activities of GPx and GRed are strongly associated with maternal protein intake in rats undergoing mild food restriction. In otherwise untreated weanling rats, maternal diet is a determinant of tissue GSH concentrations, pulmonary and hepatic GRed and -glutamylcysteine synthetase (GCS) activities (Langley-Evans et al 1995). Given the influence of maternal diet on GSH metabolism and the role of GSH in the detoxification of SO₂, we have proposed that maternal diet during pregnancy may determine the later susceptibility of the offspring to SO₂-induced tissue injury. In this article we report the effects of chronic SO₂ exposure on rats of different gestational protein nutrition.

MATERIALS AND METHODS

When body weight reached 200-225 g rats were fed purified diets containing 180 g casein/kg diet (control diet) or 120, 90 or 60 g casein/kg diet (experimental), as previously described (Langley et al. 1994). After habituation to the diets for 14 d, rats were mated and maintained on the habituated level of dietary protein until parturition. Within 12 h of parturition all dams were transferred to standard nonpurified diet (CRMX, Special Diet Services, UK: 189 g/kg protein, 23 g/kg corn oil), and the same food was used to wean the pups. These rats differed, therefore, only in terms of prenatal protein nutrition.

The gas treatment protocol of 5 h/d for 28 d was designed to resemble occupational exposures in humans. A concentration of 286 µg SO₂/m³ had been demonstrated in preliminary studies to induce measurable lung injury. To achieve the specified SO₂ concentration, gas delivered from cylinders containing 1144 µg SO₂/m³ (BOC, Guildford, Surrey, UK), was mixed with compressed air (BOC) using a clinical gas mixing unit. Concentration of SO₂ was monitored using an industrial monitor (Neotox, Harlow, Essex, UK).

RESULTS

Table 1. Recovery of leukocytes in bronchoalveolar lavage fluid from rats exposed in utero to maternal diets of different protein concentration, in response to sulfur dioxide treatment1 [View Table]

Consistent with maternal diet-related differences in leucocyte populations in the lung, circulating cell populations also varied among rats born of dams fed different amounts of protein during pregnancy (Table 2). The interaction of maternal diet with SO₂ treatment influenced total cell number (P < 0.01), and maternal diet, SO₂ and their interaction influenced lymphocyte number (P < 0.05). Although rats exposed in utero to the 60 g casein/kg maternal diet had numbers of lymphocytes and neutrophils in circulation similar to controls, rats exposed in utero to 120 g casein/kg maternal diet had higher total leucocyte counts, and rats exposed in utero to 90 g casein/kg maternal diet had significantly fewer lymphocytes. Treatment with 286 µg SO₂/m³ reduced total cell numbers in the 180 g casein/kg and 120 g casein/kg maternal diet groups, but not in rats prenatally exposed to lower levels of maternal dietary protein. This effect was attributable to large SO₂-induced declines in lymphocyte populations. No significant effects of SO₂ on circulating leucocytes were observed in rats exposed to lower maternal dietary protein concentrations.

Table 2. Leukocyte populations in circulation of rats exposed in utero to maternal diets of different protein concentrations, in response to sulfur dioxide treatment1 [View Table]

BALF protein concentration provides a measure of damage to capillary epithelial cells in the lungs because such damage allows leakage of vascular fluid into the alveoli. There was no significant variation in baseline BALF protein concentrations among the different maternal diet groups that were not treated with SO₂ (Figure 1). SO₂ Treatment did, however, modulate BALF protein concentrations (P < 0.005). Following SO₂ treatment, significantly higher BALF protein concentration was observed in rats exposed to 90 or 60 g casein/kg diets in utero, relative to the air-treated controls of the same maternal dietary groups. No significant differences were noted in the 120 g casein/kg or 180 g casein/kg maternal diet groups. In the 60 g casein/kg maternal diet group the concentration of protein in BALF following SO₂ treatment was significantly higher than in SO₂-treated control diet (180 g casein/kg maternal diet) rats (P < 0.05). This was not true of SO₂-treated rats of the 120 or 90 g casein/kg maternal diet groups. In the latter group of rats however, baseline BALF protein concentrations were low, and the magnitude of SO₂-induced elevation of BALF protein concentration (92%) was comparable to that observed (88%) in rats exposed to 60 g casein/kg maternal diet. Following SO₂ treatment, BALF protein concentration was inversely correlated with maternal protein intake (r = 0.38, P < 0.05), suggesting that SO₂-induced tissue injury was greater in rats exposed to low protein maternal diets.

Table 3. Tissue, blood and bronchoalveolar lavage fluid (BALF) glutathione concentrations in rats exposed in utero to maternal diets of different protein concentrations, in response to sulfur dioxide treatment1 [View Table]

Blood GSH concentrations were significantly modulated by maternal diet (P < 0.02), SO₂ treatment (P < 0.001) and their interaction (P < 0.03). In air-treated rats, blood GSH concentrations were similar among the 180, 120 and 90 g casein/kg maternal diet groups, but elevated in the 60 g casein/kg group relative to the 180 and 90 g casein/kg groups (Table 3). Following SO₂ treatment, rats exposed to 180 and 60 g casein/kg maternal diet had lower blood GSH concentrations than did untreated controls. No alterations due to SO₂ were noted in the 120 and 90 g casein/kg maternal diet groups.

Figure 2 shows pulmonary activities of GPx, GRed, GCS and GT. Maternal diet had no influence on GPx or GCS activities in this tissue. Gestational exposure to the 120 g casein/kg maternal diet however, led to a postnatal elevation of lung GRed and GT activities relative to control and 90 g casein/kg maternal diet groups. Treatment with SO₂ had a number of maternal diet-specific effects on pulmonary enzyme activities. Rats born of dams fed 120 g casein/kg diet had markedly lower lung GPx activity following SO₂ treatment. No other maternal diet groups responded in this manner. Lung GCS activity was elevated significantly by SO₂ exposure in 180 and 60 g casein/kg exposed rats (P < 0.05), but GRed and GT activities were unaltered. Lung GST (Fig. 3) activity was not responsive to maternal diet. Following exposure to SO₂, GST activity was between 29 and 40% lower in all maternal diet groups, except the 60 g casein/kg group in which the effect was not statistically significant (P = 0.18).

In the liver (Fig. 4), maternal diet influenced activities of GCS and GRed. Activity of GCS was elevated in the 120 and 60 g casein/kg maternal diet groups relative to 180 g casein/kg and 90 g casein/kg maternal diet groups. Liver GRed activity was elevated in rats born of dams fed 60 g casein/kg diet relative to dietary controls. Treatment with 286 µg SO₂/m³ had no effect on hepatic GST, but GCS activity was significantly greater due to SO₂ exposure in rats whose dams were fed 180 and 90 g casein/kg diet. Following SO₂ treatment liver GRed activity was higher in 120 and 90 g casein/kg maternal diet groups than in their corresponding air exposed (untreated) controls, and GPx activity was similarly higher in SO₂ treated rats in the 60 g casein/kg maternal diet group.

DISCUSSION

The injury induced by SO₂ is, however, somewhat difficult to define. The variability of the response is most likely attributable to the magnitude and duration of the exposure. Human studies indicate that acute exposures induce bronchoconstriction in the vulnerable population (Sandstrom et al. 1989a and 1989b, Stjernberg et al. 1985), while chronic exposure, such as the SO₂ treatment employed in this study, may be responsible for the development of asthmatic symptoms (Tseng and Li 1990). In rats, very high (2288 µg SO₂/m³) exposure also leads to bronchoconstriction and damage to the upper airways (Stratmann et al. 1991) characterized by mucous metaplasia and sloughing of tracheal epithelial cells (Basbaum et al. 1990). Previous studies that utilized the offspring of rats fed non-purified diets (Langley-Evans et al. 1996), indicated that over 7-14 d exposure to 14-286 µg SO₂/m³ for 5 h/d, the lung was invaded by neutrophils. After resolution of this response a smaller secondary influx of neutrophils and macrophages was observed. There was no apparent evidence of capillary epithelial damage, and GSH was depleted by 25-40% in all tissues.

In this study the pattern of injury just described was not observed, despite using the same SO₂ treatment regimen. A preliminary histological examination of the lungs of two rats from each of the eight groups revealed no evidence of inflammatory cell infiltration of the airways, which parallels the lack of differences in cell numbers recovered by bronchoalveolar lavage. Tissue GSH concentrations were largely maintained in most SO₂-treated rats, a phenomenon which appears to stem from SO₂-induced changes in enzyme activities discussed below. In addition to differences in maternal diets, an important contrast between this study and our earlier work (Langley-Evans et al. 1996) is the timecourse over which the injury was studied. In this study, SO₂ exposure was for 28 d, and by this time elements of the injury, for example the influx of inflammatory cells, may well have been resolved. A transient neutrophil influx into the lungs with later resolution was noted in guinea pig neonates exposed to a hyperoxic atmosphere over 28 d (Phillips 1995). Moreover, the variation in numbers of particular cell types recovered in BALF and in blood was high, possibly masking trends within the data. In BALF in particular, this variability has been previously noted (Kelly et al. 1991) and reflects the fact that in some cases, neutrophils may be completely absent.

The critical indicator of ongoing pulmonary injury was the elevated BALF protein concentration noted in the low protein maternal diet groups (60 and 90 g casein/kg). Oxidative injury to capillary epithelial cells results in increased capillary permeability. The resulting exudation of vascular fluid into the alveoli is detectable as an elevated protein concentration, 50% of which is serum albumin (Kelly et al. 1991). Rats exposed to 180 g casein/kg maternal diet in utero did not exhibit increases in BALF protein concentration following 28 d of SO₂ exposure, consistent with previous observations (Langley-Evans et al. 1996). All SO₂-treated, low protein maternal diet groups had elevated BALF protein concentrations (significant in 90 and 60 g casein/kg diet groups), relative to air-treated controls. Although the 90 g casein/kg maternal diet group had similar BALF protein concentrations following SO₂ treatment to the rats from the 180 and 120 g casein/kg groups, the untreated 90 g casein/kg group tended to have a lower baseline BALF protein level. A nonsignificant (23%) SO₂-induced increase in BALF protein concentration in the 120 g/kg maternal diet group was also observed. Given the protocol followed, it is unclear whether the observed rise in vascular permeability within the lungs occurred earlier in the dietary control group and was subsequently repaired, or whether it is a late-stage phenomenon which occurs first in the more susceptible, low protein-exposed groups. The latter seems likely, because in rats exposed to SO₂ concentrations of between 14 and 286 µg SO₂/m³, no evidence of increased capillary permeability was apparent after 7, 14, 21 or 28 d exposure (Langley-Evans et al. 1996). By comparison, guinea pigs chronically exposed to hyperoxia develop such an injury early in their exposure and subsequently repair the epithelial damage (Phillips 1995). Although statistically significant, the relationship between maternal protein intake and post-SO₂ BALF protein concentration was weak. Maternal protein intake accounted for only 14% of the variation in BALF protein concentration. This relationship, however, considered only BALF protein concentration following SO₂ exposure. The relationship between the difference in protein concentration between SO₂ exposed and untreated rats and maternal dietary protein was much stronger (r = 0.967, P = 0.03) and explains 93% of the variation in BALF protein concentration. The remaining variability may be explained by different rates of repair and different timecourses of tissue damage. As stated, this study considers only the late phase of the injury elicited by SO₂ and may well overlook a complex development of injury and repair processes over a 4 wk time period.

It is possible that differences in SO₂-induced injury observed between this and earlier work (Langley-Evans et al. 1996) relate to maternal dietary influences on GSH metabolism. Our pilot studies utilized the offspring of rats fed non-purified diet, and such animals have lower hepatic and pulmonary GCS activities than do the offspring of rats fed casein-based purified diets (Langley-Evans et al. 1995). Since the maternal diet may broadly alter overall GSH metabolism, the subsequent utilization of GSH for the detoxification of SO₂ may also be modified. In this study, there was little evidence of the potent GSH depletion previously observed in SO₂-treated rats (Langley-Evans et al. 1996). This difference almost certainly relates to different responses to SO₂ exposure of rats born to dams fed different levels of dietary protein during pregnancy. -Glutamyl cycle enzymes in tissues of rats born to dams fed non-purified diet show no adaptive response other than to decrease liver GRed activity, which is likely to exacerbate GSH depletion (Langley-Evans et al. 1996). Rats exposed to purified diets in utero, in addition to differing in baseline hepatic GRed and GCS activities (Langley-Evans et al. 1996), appear to increase GSH conserving activities.

These data are, however, broadly consistent with previous reports of intrauterine programming of -glutamyl cycle enzyme activities in rats (Langley et al. 1994, Langley-Evans et al. 1995). Glutathione reductase appears to be most sensitive to maternal protein intake, most notably in the liver where higher activity was noted in rats exposed to the low protein diet (60 g casein/kg) in utero. Tissue GSH concentrations fluctuate less consistently, but heart and liver appear to be susceptible to maternal influences (Langley et al. 1994, Langley-Evans et al. 1995). Tissue GSH concentrations represent a balance of the rates of synthesis, transport and utilization. Differences in GSH concentrations among the groups with different maternal protein intake may be attributable to previously observed (Langley et al. 1994) programmed changes in these processes. Earlier work indicated that in rats prenatally exposed to a 120 or 60 g casein/kg maternal diet, capacity to transport GSH is impaired.

The rats exposed to low protein maternal diet (60 g casein/kg) exhibited greater SO₂-induced lung injury, as assessed by recovery of protein in bronchoalveolar lavage, than did 180 g casein/kg maternal dietexposed rats. One of two explanations for this intrauterine modulation of susceptibility relates to GSH metabolism. GSH undergoes sulfitolysis with SO₂ following oxidation to glutathione disulphide (GSSG) by GPx (Kagedal et al. 1986, Mannervik et al. 1974, Winell and Mannervik, 1969). The product of sulfitolysis, S-sulphoglutathione (GSSO₃) cannot be recycled through the GRed reaction and is excreted as thiosulfate. GSSO₃ inhibits glutathione S-transferases (Pool-Zobel et al. 1990) and accordingly, pulmonary GST activities were markedly lower in the SO₂-exposed rats. Clearly all rats exposed to the four purified diets in utero were able to mount an adaptive response of the -glutamyl system to SO₂ exposure, a feature absent in rats prenatally exposed to non-purified diets (Langley-Evans et al. 1996). As a consequence of this response, all rats exposed to SO₂ were able to maintain tissue GSH levels, with the exception of the 60 g casein/kg group in which liver GSH fell by 27%. The wide variety of changes in enzyme activity observed in SO₂-treated rats shifted metabolism towards GSH conservation, and this response was apparently modified by the maternal diet. Rats exposed to the control maternal diet had greater GCS activity in both lung and liver after treatment with SO₂. Rats of the 120 g casein/kg maternal diet group did not demonstrate SO₂-induced changes in GCS activity, but had high activity in liver prior to the gas exposure. The adaptive response of this group appeared dependent upon an elevation in GRed activity in liver that recycled GSH, and a lower GPx activity in lung that diminished loss of GSH through sulfitolysis of GSSG. Rats in the SO₂-treated 90 g casein/kg group had greater hepatic GRed and GCS activities than untreated rats. Thus, three of the diet groups maintained liver GSH concentrations in response to SO₂ through alterations in the pattern of -glutamyl enzyme acitivies. In the 60 g casein/kg maternal diet group in which hepatic GSH concentrations were lower following SO₂ treatment, pulmonary GSH did not respond; in the lung, only GCS activity was greater than that of untreated rats. Importantly, activity of hepatic GPx was elevated, which may have contributed to the observed depletion of GSH from liver. The role of the GSH system in determining susceptibility to injury is again difficult to interpret in view of the likely complexity of the timecourse of injury. It does appear, however, that in the case of the 60 g casein/kg maternal diet group which demonstrated the greatest level of pulmonary injury, the ability to conserve GSH was less than in other groups.

The mechanism underlying differential susceptibility to SO₂-induced injury may alternatively relate to the acute phase response mounted following initial tissue damage. An inflammatory response is elicited by SO₂, demonstrated by recruitment of neutrophils into the airways. This response is presumably cytokine mediated. Cytokine production (Tappia et al. 1994) and the acute phase response to an endotoxin injection (Langley et al. 1994) are related in a non-linear manner to maternal dietary protein intake. Specific elements of the response to endotoxin are enhanced or blunted by intrauterine exposure to low protein diets matenal (Langley et al. 1994). The cell counts reported in this work may also reflect impairment of immune functions in these animals. Rats exposed in utero to 90 g casein/kg maternal diet had fewer circulating lymphocytes than did control animals. Additionally, as a proportion of the total cell number, rats from all the low protein groups tended to have a greater number of circulating neutrophils (control: 11%, 120 g casein/kg: 18%, 90 g casein/kg: 15%, 6 g casein/kg: 18% neutrophils). Moreover, such rats have also been observed in our lab to have low spleen weights. Others have demonstrated that the immune system is extremely sensitive to manipulations of the maternal diet during fetal development (Beach et al. 1982 and 1983).

The susceptibility of the young adult rat to sulfur dioxide induced lung injury, as evidenced by BALF protein concentrations, is programmed by aspects of the maternal diet. Fetal exposure to a low protein maternal diet increased the magnitude of injury, concomitant with modulation of glutathione metabolizing enzyme activities. As we have reported previously, rats of different prenatal nutrition respond differently to an inflammatory stimulus. This may reflect the programmed state of the immune system and aspects of antioxidant defences.