Immunosuppression in adult female B6C3F1 mice by chronic exposure to ethanol in a liquid diet

Immunopharmacology, 26 (1993) 31-51 © 1993 Elsevier Science Publishers B.V. All rights reserved 0162-3109/93/$06.00

31

I M P H A R 00647

Immunosuppression in adult female B6C3F1 mice by chronic exposure to ethanol in a liquid diet Michael P. Holsapple a, Micah Eads a, Wayne D. Stevens a, Steve C. Wood a, Norbert E. Kaminski a, Dale L. Morris a, Alphonse Poklis b, Edward J. Kaminski c and Stephen D. Jordan a Department of PharmacoloKy and Toxicology, bDepartment of Pathology, Medical College of Virginia/Virginia Commonwealth University, Richmond, VA, USA and cDepartment of Pathology, Northwestern University Medical School, Chicago, IL, USA (Received 4 February 1992; accepted 16 October 1992)

Abstract: The overall objective of these studies was to characterize the effects of ethanol on the immunocompetence of adult female B6C3F1 mice. To obtain a significant suppression in the antibody response to SRBC, splenocytes from untreated mice had to be directly exposed to concentrations of ethanol from 0.3 % to 3.0%, or to acetaldehyde at concentrations greater than 0.03 To. We do not believe that these results are consistent with a role by a direct effect by either ethanol or its primary metabolite because these concentrations are higher than what could be obtained as reasonable blood levels. For in vivo exposure, we employed a pair-feeding regimen which was based on a liquid diet containing 5% ethanol (v/v) that provided 36% of the caloric intake as ethanol. Our results indicated that there was a definite temporal relationship to the consequent suppression of the antibody response to SRBC in that no effect was observed after 14 days exposure, and that the magnitude of the suppression increased from 18% after 21 days to 70% after 42 days. We also monitored the liver for histopathology and observed that the ethanol-induced liver damage was restricted to steatosis (fatty liver), which was also manifested with time and which was most pronounced after 42 days exposure. In contrast to our results with the in vivo antibody response, we saw no effect on mitogen-induced proliferation by splenocytes from ethanol-treated mice. These results prompted us to measure in vitro antibody responses by splenocytes from ethanol-treated mice. We saw no suppression of the in vitro antibody responses to SRBC, DNP-Ficoll or LPS after any length of exposure to ethanol, and speculated that the basis for the suppression of the in vivo antibody response was an indirect consequence of exposure. We subsequently determined that when normal splenocytes were cultured in 5% serum from ethanol-exposed mice (42-day group), there was a > 8 0 % suppression relative to the serum from the pair-fed controls. As important controls for these studies, we have demonstrated that there was no difference between the responses of normal lymphocytes cultured in 5% normal mouse serum and in 5% serum taken from the pair-fed restricted controls. A determination of the ethanol content in the serum from ethanol-exposed mice (42-day group) indicated that the amount of ethanol present in these cultures was < 0.003 %. These results suggest that the mechanism of ethanol-induced immunosuppression is at least in part due to an indirect consequence of chronic exposure, which is possibly mediated by a serum factor. Key words:

Chronic ethanol exposure; Pair-feeding paradigm; Immune function; Steatosis

Correspondence to: M.P. Holsapple, Department of Pharmacology and Toxicology, P.O. Box 613, Medical College of Virginia/Virginia Commonwealth University, Richmond, VA 23298, USA. Abbreviations: Ab, antibody; AFC, antibody forming cell; Con A, Concanavalin A; EBSS, Earle's balanced salt solution; FCS, fetal calf serum; HBSS, Hank's balanced salt solution; LPS, lipopolysaecharide; 2-ME, 2-mercaptoethanol; MS, mouse serum; NMS, normal mouse serum; PCL, polyclonal, as in AFC response to LPS; psi, pounds per square inch; RC, restricted control; SRBC, sheep erythrocytes; TD, T-dependent, as in AFC response to SRBC; TI, T-independent, as in AFC response to DNP-Ficoll.

32 Introduction

The possibility that a chemical or xenobiotic may target the immune system is suggested whenever exposure to that chemical/xenobiotic is associated with an increased susceptibility to infections and/or an increased incidence of cancer. As previously reviewed (Palmer, 1981; MacGregor, 1986), chronic alcoholics demonstrate an increased susceptibility to a wide spectrum of infections including bacterial pneumonias, listeria, tuberculosis and viral disorders, and an increased incidence of cancers. Alcoholics are also characterized by a number of abnormalities in immune function including anergy to skin tests (Berenyi et al., 1974), reduced responsiveness to mitogenic stimulation (Hsu and Leevy, 1971; Lundy et al., 1975), abnormal thymocyte function (Bernstein et al., 1974; Roselle and Mendenhall, 1984) and hypergamma-globulinaemia (i.e., increases in serum IgM, IgG and IgA) associated with polyclonal activation of B-cells (B ailey et al., 1976; Morgan et al., 1980; Drew et al., 1984). Studies by Watson and co-workers have indicated that short-term chronic exposure (i.e., 8 days) to ethanol in mice results in significant alterations in lymphocyte subsets, with decreases noted in total T-cells, T-suppressor cells and surface IgM-positive B-cells (Watson etal., 1988), and that long-term exposure to ethanol (i.e., several weeks to several months) in rats produces decreased thymus weights and the general loss of the T-cell population from the spleen (Mufti et al., 1988). Taken together, these results support the hypothesis that exposure to ethanol can have an effect on immune function. Because of the diversity of experimental conditions and of immunological parameters shown to be affected, the mechanism for the action is still not established. Similar to the situation with ethanol-induced hepatotoxicity, there has been an active debate as to the relative roles played by ethanol, and by other factors associated with alcohol abuse, such as malnutrition, in the immunosuppression associated with the chronic alcoholic. However, be-

cause a number of studies in man as well as a number of studies in animal models have demonstrated that exposure to ethanol can depress immunocompetence under conditions where the diet can be controlled (see the Discussion), it is apparent that at least a component of ethanolinduced immunotoxicity is independent of nutritional deficiencies. Nonetheless, the relative contributions by the direct and indirect consequences of exposure to ethanol are still debated. For example, Jerrells and co-workers have demonstrated that the immunosuppression associated with acute exposure to high doses of ethanol or following withdrawal of ethanol from dependent animals was primarily an indirect effect mediated through the adrenal neuraxis (Jerrells et al., 1986, 1989). The overall objective of the present investigation was to characterize the effects of chronic exposure to ethanol in the mouse, which has heretofore not been widely studied. Because mice have a natural aversion for ethanol, a pairfeeding exposure paradigm based on a liquid diet containing ethanol was employed in order to minimize, if not eliminate, nutritional differences between treated and control mice. Besides the central nervous system, the liver has represented the most extensively studied organ targeted by ethanol. Therefore, we attempted to correlate the effects on immune function with ethanol-induced liver damage in the pair-feeding paradigm. In the mouse, we anticipated that the extent of ethanolinduced liver damage would be limited to steatosis (fatty liver).

Materials and Methods Animals' Female (C57BL/6 x C3H) F1 (hereafter referred to as B6C3F1) mice were purchased from the Frederick Cancer Research Center (Frederick, MD). The mice arrived at 5-6 weeks of age (i.e., body weight of ~ 18 g) and were quarantined for at least one week. Upon arrival, mice were randomized, weighed and housed 4 per cage.

33 Until mice were placed on the liquid diet, all animals were maintained with Purina Lab Chow and tap water ad libitum. The animal quarters were maintained at 21-24 °C and 40-60% relative humidity. A 12-h light/dark cycle was used. Chemicals and reagents Ethanol was purchased from the MCV Hospital Pharmacy as a 95 % solution. Acetaldehyde was purchased from Aldrich Chemical Co. (Milwaukee, WI). For the direct addition studies, ethanol was added directly to cultured splenocytes from untreated mice at the concentrations indicated. The antigen, either SRBC or DNP-Ficoll, was added immediately. Initial studies indicated that acetaldehyde was extremely toxic when added under similar conditions. Therefore, for the direct addition studies with acetaldehyde, splenocytes from untreated mice were exposed for 4 h to the concentrations of acetaldehyde indicated. As a comparative control in these studies, parallel cultures of splenocytes were exposed to ethanol under similar preincubation conditions. After the 4 h incubation, the cells were washed, adjusted for use in the in vitro antibody response and the antigen, SRBC, was added. The liquid diet formula was based on previous work by Lieber and DeCarli (1982) and was purchased commercially from Dyets, Inc. (Bethlehem, PA). The preparation of the liquid diet is described below. Media, media supplements and balanced salt solutions were obtained from the following sources: R P M I 1640 medium without NaHCO3, EBSS without NaHCO3, H B S S without N a i l CO3, CaC12, MgC12"6 H20, MgSO2"7H20 and L-glutamine (Gibco, Grand Island, NY); penicillin-streptomycin-fungizone solution and M E M Spinners without L-Gln (Hazelton Research Products, Inc., Denver, PA), 2-mercaptoethanol (2-ME; Bio-Rad Laboratories, Richmond, CA), fetal calf serum (FCS; Hyclone Laboratories, Inc., Logan, UT), sheep erythrocytes (SRBC; Colorado Serum, Denver, CO), DNP-Ficoll (Biosearch, San Raphael, CA); lipopolysaccharide (LPS; S. typhosa 0901; D I F C O Laboratories, Detroit, MI); and con-

canavalin A (Con A; Pharmacia, Piscataway, N J). Complete R P M I media was prepared as a 1 x solution by the addition of 2 mM L-Gln, 1% (v/v) penicillin-streptomycin-fungizone solution, 5 x 10 5 M 2-ME and 5% FCS. Chronic exposure to ethanol The exact compositions of the diets were previously described by Lieber and DeCarli (1989) and were provided by the manufacturer. One liter of the control diet was prepared with 221.78 g of the control diet mix that was q.s. to one liter with cold tap water and mixed in a blender. The control diet contains 1 kcal/ml of which 35 % are derived from fat, 18% are derived from protein and 47% are derived from carbohydrate. One liter of a 5% ethanol diet was prepared with 132.18 g of the ethanol diet mix and 50.47 ml 95% ethanol, which was q.s. to one liter with cold tap water and mixed in a blender. The 5 % ethanol diet contains 1 kcal/ml of which 35 % are derived from fat, 18 % are derived from protein, 11% are derived from carbohydrate and 36% are derived from ethanol. When mice were switched from the Lab Chow to the liquid diet, they were re-randomized into cages of two mice per cage. Each cage of ethanoltreated mice was assigned a cage of control mice as the pair-fed restricted control. All mice were given free access to the control diet for one week in order to acclimate them to the liquid diet. After one week, one half of the mice were switched to the ethanol diet and were given free access to the ethanol diet throughout the experiment. Both the control diet and the ethanol diet were replaced each day. The amount of ethanol diet consumed was recorded for each ethanol cage. The pair fed restricted control mice for that ethanol cage received that exact amount of control diet the following day. Mice were kept on the liquid diets, i.e., either the control diet or the ethanol diet, for 14, 21, 28, 35 or 42 days. In order to avoid any problems with withdrawal from the ethanol, mice were kept on their respective diets until they were sacrificed

34 by cervical dislocation, 24 h after the day of exposure as designated above. Unless otherwise indicated, there were 4 mice per treatment group (i.e., either control or ethanol) per timepoint.

Body and organ weights and histopathology In studies in which the mice were not immunized with SRBC during the exposure (i.e., the mice used in the in vitro antibody responses described below and the mice used in the mitogen responses described below), after sacrificing the following organs were removed and weighed: liver, spleen and thymus. In addition, the livers were fixed in 10~o neutral buffered formalin, embedded in paraffin, sectioned and stained with hemotoxylin and eosin. The slides were read by a pathologist (Dr. E.J. Kaminski) who was not aware of the treatments ahead of time. Mitogen-induced proliferation Splenocyte single cell suspensions were prepared as described above from restricted control and ethanol-treated mice. The suspension from each mouse was adjusted to 1 x 10 6 cells/ml in RPMI complete media. Aliquots of 200/~1 (i.e., 2 x 105 cells) were added to individual wells of 96-well microtiter plates (Costar, Cambridge, MA) so that there were quadruplicate wells for media controls and for each of the concentrations of both the T-cell mitogen, Con A, and the B-cell mitogen, LPS. The Con A concentrations ranged from 0.5 #g/ml to 2.0 #g/ml, while the LPS concentrations ranged from 1.0 #g/ml to 100.0 ~ug/ml. All cultures were pulsed with 1 /~Ci of [3H]thymidine (ICN, Irving, CA) on day 3. The cultures were harvested onto glass fiber filters using a P H D Cell Harvester (Cambridge, MA) 17 h later. The filters were air-dried, placed in 2 ml of scintillation fluid (Ready-Solv, Beckman) and counted in a liquid scintillation counter (Beckman). Results were expressed as absolute cpm. In vitro antibody responses For the direct addition studies, splenocytes were aseptically prepared into single-cell suspensions

from several untreated mice (i.e., the exact number of mice used in any given study depended on the size of the experiment; but was always > 4 mice) as previously described (Holsapple et al., 1983) and were pooled. For the chronic exposure studies, splenocytes from each mouse were aseptically prepared into single-cell suspensions after the indicated lengths of exposure. Cells were suspended in RPMI 1640 complete media and quadruplicate cultures (0.5 ml/well) for each animal, or for each treatment group in the direct addition studies, were set up in 48-well culture plates (Costar, Cambridge, MA). Either 7.5 x 106 splenocytes/well with an equal concentration of SRBC, 7.5 x 106 splenocytes/well with 100 ng/ml (50 ng/well) of DNP-Ficoll, or 5 x 106 splenocytes/well with 100 #g/ml (50 gg/well) of LPS were cultured for the T-dependent (TD), T-independent (TI) or polyclonal (PCL) antibody forming cell (AFC) response, respectively. All cultures were incubated with rocking at 37 °C in an atmosphere of 10~o CO2, 7~o 02 and 83~/o N2 at 6 psi. Splenocyte and antigen concentrations, sensitization intervals and conditions were shown to be optimal in preliminary experiments. AFC responses to SRBC, DNP-Ficoll and LPS were measured on days 5, 3 and 2, respectively using the Jerne plaque assay as described below. The SRBC AFC response was measured against the same antigen used to immunize the cultures. The DNP-Ficoll and LPS AFC responses were measured against SRBCs which were densely coupled with trinitrophenyl (TNP) following the method of Rittenberg and Pratt (1969). Aliquots of recovered spleen cell suspensions were used to determine cell counts using a Coulter Counter (Model ZBI) and viabilities were determined using the pronase method as previously described in this laboratory (Johnson et al., 1987). Results for all AFC responses were analyzed both as AFCs/106 recovered splenocytes and as AFCs/culture. Because there was never any difference between the results in the two forms, only the AFCs/106 recovered splenocytes are presented.

35 In vitro antibody responses using mouse serum

In some studies, as indicated below, we replaced FCS with mouse serum (MS) which was derived from either untreated mice (i.e., designated normal mouse serum or NMS), mice maintained on the control liquid diet or mice maintained on the ethanol liquid diet. Mice were deeply anesthesized with CO2 and blood was drawn by cardiac puncture into 10 x 75 mm glass tubes. Mice were then sacrificed by cervical dislocation. The blood was allowed to clot at 4 °C for one h. The clot was dislodged and pelleted by centrifugation for 25 min at 2000 x g. The remaining serum was heat-inactivated (i.e., to destroy complement) for 30 min at 56 °C, filter-sterilized and stored at - 7 0 ° C. The various batches of MS were used in the in vitro antibody responses as described above for FCS. Quantitation of ethanol in mouse serum In separate studies, mice were maintained on the liquid diets for 14-42 days, and were deeply anesthesized with CO2; blood was drawn and serum samples were prepared as described above. The ethanol content in the serum samples was determined on a Shimadzu 14-A gas chromatograph (Shimadzu Co., Kyoto, Japan) equipped with a 3 m x 2 mm i.d. G P 60/80 Carbopack B/5 To Carbowax ® glass column. The flame ionization detector response was recorded with an H P 3396A integrator (Hewlett Packard, Avondale, PA). A stock solution of GC internal standard was prepared by diluting 0.30 ~1 1-propanol with 100 ml distilled water. The stock solution was diluted one to ten with distilled water for a working internal standard of 2.34 g/1 1-propanol. All reagents were reagent grade. A G C calibrator was prepared by diluting absolute ethanol, methanol, isopropanol and acetone to 10.0 ml with serum to produce 1.57 g/1 ethanol, 0.79 g/l methanol, 0.78 g/1 isopropanol and 0.79 g/1 acetone. The linearity of the calibrator solution was checked against aqueous working standards containing 0.50, 1.00, 1.50, 2.00, 3.00 and 5.00 g/1 ethanol. The G C assay was linear to 6.00 g/l; linear regression analysis of the GC standard

curve yielded y = 0 . 9 5 1 x + 0 . 0 4 g/1 ethanol; r = 1.00. Serum ethanol control solutions of 0.70 g/1 and 1.92 g/l, with acceptable ranges of 0.62-0.78 g/l, and 1.70-2.14 g/1 obtained from Sigma Chemical Co. (St. Louis, MO), were used as controls for G C ethanol determinations. All calibrators, reagents and controls were stored at 4 oC. For the actual analyses, an aliquot of 0.10 ml of the serum samples was placed in a test tube containing 0.90 ml internal standard solution. The mixture was vortexed for 20 s and 1/~1 of it was injected into the GC. The helium capillary flow was set at 20 ml/min linear velocity. The G C temperature was set at 75 ° C for the oven, 200 ° C for the injector and 230 °C for the detector. The retention time for ethanol was 1.49 min and the retention time for the 1-propanol internal standard was 3.23 min. In vivo antibody response

Our historical peak response for the in vivo antibody response to SRBC occurs 4 days after sensitization. As described above, in order to avoid any problems with withdrawal from the ethanol, mice were kept on their respective diets until they were sacrificed. Therefore, for sensitization, mice were injected (i.v. into the tail vein) with 1 × 108 SRBC 4 days prior to sacrifice, i.e., days 11, 18, 25, 32 and 39 for the 14-, 21-, 28-, 35- and 42-day exposures, respectively. Single spleen cell suspensions were prepared from each spleen in 3 mls EBSS, washed and resuspended in 3 ml EB S S and counted on a Coulter Counter. Enumeration of AFC was performed using a modification of the Jerne plaque assay as previously described (Holsapple et al., 1984). Briefly, a 0.5 ~o melted agar (DIFCO, Detroit, MI) solution in EBSS was prepared containing 0.05~o DEAE-Dextran (Pharmacia, Piscataway, NJ) and was maintained at 47 °C. The melted agar solution was dispensed in 350 pl aliquots into 12 x 75 mm heated culture tubes. Each spleen cell suspension was diluted 30-fold, by resuspending a 100 #l aliquot in 2.9 ml of EBSS and held on ice. Each agar tube received 100/~l of the diluted spleen cell suspension, 25 ktl of guinea pig

36 complement (GIBCO, Grand Island, NY) and 25/~l of indicator SRBC. The tube was vortexed immediately, poured into a 100 x 15 mm Petri dish and covered with a 45 x 50 mm cover slip. After the agar solidified, the Petri dishes were incubated at 37 °C for 3 h. After the incubation, the A F C s were enumerated at 6.5 x magnification using a Bellco Plaque viewer. Results for all A F C responses were analyzed both as AFCs/106 recovered splenocytes and as AFCs/spleen. Because there was never any difference between the results in the two forms, only the AFCs/106 recovered splenocytes are presented. Statistics All results for the antibody responses and mitogen responses were evaluated using a one-way analysis of variance for vehicle and ethanol treatment groups. When significant differences were observed, a Dunnett's t-test was used to determine which groups were significantly different from controls.

Results Direct effects of ethanol and acetaldehyde on in vitro antibody responses Spleen cell suspensions from untreated mice were exposed to concentrations of ethanol from 0.1~o to 3.0~o and were stimulated with either SRBC or DNP-Ficoll. The T-dependent (TD) and T-independent (TI) A F C responses were measured 5 days and 3 days later, respectively. The results are shown in Fig. 1 and indicate that both A F C responses were suppressed in a doserelated fashion from 0.3 ~o to 3.0 ~o. Although the TD A F C response was only significantly suppressed at the highest concentration, while significant suppression of the TI AFC response was noted at the two highest doses, the magnitude of suppression was comparable for the two AFC response models; i.e., compare the 31 ~o suppression of the TD response with the 28~o suppression of the TI response at 1.0~o ethanol; and the 72~o suppression of the TD response with the

T-Dependent Antibody Response

T-Independent Antibody Response

1500"

300

1250"

~>" 1000"

250

ID

}~:i? !i~ i :. 750" o

<~O

O 0~ n"

150

<~O O

100

-K

500" C~ IJ.

U250"

N NA

0.1

0.3

1.0 Ethanol

(%)

3.0

NA

0.1

0.3

1,0 Ethanol (%)

i 3.0

Fig. 1. Direct effects of ethanol on in vitro antibody responses. Ethanol was added directly to splenocytes from B6C3FI mice at concentrations of from 0.1% to 3.0% and the cultures were immediately immunized with either SRBC (T-dependent response; T D ) or DNP-Ficoll (T-independent response; TI). The antibody response was measured after five days (TD) or three days (TI) of incubation. Results represent the mean _+ SE of quadruplicate wells for each treatment group and are representative of three separate experiments. ** = p < 0.01 and * = p < 0.05 as determined by Dunnett's t-test using the naive (untreated; NA) group as the control.

37 62~o suppression of the TI response at 3.0Yo ethanol. In all studies to date, concentrations of ethanol less than 1.0~o had no effect on viability, while the suppression by concentrations of 3.0~o was always associated with decreased viability ( > 40 ~o reduction in viability when compared to controls; data not shown). Spleen cell suspensions from untreated mice were exposed to concentrations of acetaldehyde from 0.03~o to 3.0~o or to concentrations of ethanol from 0.1~o to 10.0~o for 4 h. The cells were then washed, the cell count was adjusted, and the cells were added to individual wells of 48-well culture plates and stimulated with SRBC. The TD AFC response was measured five days later. The results are shown in Fig. 2 and indicate that ethanol was only suppressive at a concentration of 10 ~o, while acetaldehyde was markedly suppressive at concentrations as low as 0.1~o. The greater activity by acetaldehyde, relative to ethanol, is in agreement with two of the three previous studies in which the two chemicals have been compared. These activities include the sup-

1200

pression of a number of phagocytic functions (Schopf et al., 1985) and the suppression of both mitogen- and antigen-induced proliferation of human lymphocytes (Roselle and Mendenhall, 1982). In a third paper, ethanol, but not acetaldehyde, was shown to be capable of suppressing antibody-dependent cell-mediated cytotoxicity (Stacey, 1984). Although we have clearly shown that direct exposure to either ethanol, or its primary metabolite, can produce marked immunosuppression, it is important to emphasize that the concentrations required to do so were unreasonably high and would be difficult, if not impossible to obtain as blood levels. For example, in Virginia an individual is legally intoxicated with a blood alcohol content of 0.1 ~o. Effects of chronic exposure to ethanol via a liquid diet on body and organ weights All mice maintained on either liquid diet appeared normal throughout the study in terms of their gross physical appearance and behavior. The body weight and some select organ weights

1200

Ethanol

1000"

Acetaldehyde

1000 >,

_e

800"

800"

i~i!ii'i!)i:~!i:)i~;;i!

=> 8

600

ii:

o ¢D n-

re

=<

400-

o

O

200

='~'~'~'~

200 t¢

0 N\'~ NA

0.1

0.3

1.0

Concentration (%)

3.0

10.0

oN NA

0.03

0.1

0.3

Concentration

1.0

3.0

(%)

Fig. 2. Direct effects of ethanol or acetaldehyde on in vitro T-dependent aantibody response after 4 h of preincubation. Splenocytes from B6C3F1 mice were preincubated with concentrations of ethanol from 0.1% to 10.0% or of acetaldehyde from 0.03% to 3.0% for 4 h. The splenocytes were then washed and immunized with SRBC. The T-dependent antibody response was measured after five days of incubation. Results represent the mean + SE of quadruplicate wells for each treatment group and are representative of two separate experiments. ** = p < 0.01 as determined by Dunneu's btest using the naive (untreated; NA) group as the control.

38 for mice maintained on the restricted control liquid diet or the ethanol liquid diet for 14 days through 42 days are shown in Table I. As indicated in the third column, there was a measurable weight gain in both groups over the study duration, which would be less than that expected of an animal maintained .on an ab libitum diet (i.e., Purina lab chow). Similar observations regarding a slow weight gain with a restricted diet have been previously made by others, including Tabakoff et al. (1978). As indicated in Table I, there was no difference in any of the parameters between the restricted controls and the ethanolexposed mice when analyzed on an absolute basis. We also analyzed the parameters on a relative basis (i.e., percent of body weight) and received an identical profile (data not shown). Based on the body weights shown in Table I and on the amount of ethanol liquid diet consumed, as discussed above in the Materials and Methods, we have determined that the grams of ethanol consumed per kg mouse body weight per day was 18.92, 18.84, 19.44, 19.86 and 20.24 for the 14-day through 42-day exposure groups, respectively.

Ethanol-induced liver damage In addition to recording the liver weights, we also collected the livers in buffered formalin and determined the extent of ethanol-induced damage to the liver based on histopathology. The results for the 21-, 28-, 35- and 42-day exposure groups are shown in Table II and indicate that, as expected, the extent of ethanol-induced liver damage in mice was restricted to steatosis. Our results suggest that there was a temporal relationship between the magnitude of the liver damage and the duration of exposure to ethanol. In the 21-day group, only one mouse exhibited micro steatosis, while all ethanol-treated mice exhibited micro steatosis at both 28 and 35 days. At 42 days, all of the mice exhibited micro steatosis and 2/5 mice exhibited macro steatosis. Effects of chronic exposure to ethanol via a liquid diet on the in vivo antibody response In contrast with the lack of effect on thymus and spleen weight and spleen cell number (Table I), we determined that chronic exposure to ethanol in a liquid diet caused a decrease in the antibody response to SRBC (Fig. 3). More importantly,

TABLE I Efects of chronic exposure to ethanol on b o d y a n d organ weights (wt.) Treatment group

Body wt. (g)

Spleen wt. (mg)

Spleen cell no. ( x 107)

T h y m u s wt. (mg)

Liver weight (rag)

14 day cont. 14 day E t O H

20.2 + 1.0 21.6 _+ 1.0

76.5 + 2.9 65.5 i 3.0

12.4 _+0.9 12.0 + 0.5

48.0 + 6.5 48.8 _+ 6.7

934.5 _+77.0 1020.8 + 53.6

21 day cont. 21 day E t O H

22.1 + 0.5 22.6 + 0.2

62.5 _+4.3 61.8 + 3.1

12.2 _+0.5 12.6 + 0.4

56.8 _+6.2 54.0 i 2.0

1001.0 _+ 16.2 1097.5 ± 39.8

28 day cont. 28 day E t O H

22.7 + 1.0 24.3 + 0.6

74.5 + 4.1 75.2 _+ 2.3

11.8 + 0.9 12.9 + 0.2

59.0 _+3.4 64.8 + 5.1

1061.0 + 31.5 1094.0 + 19.5

35 day cont. 35 day E t O H

24.6 + 1.6 25.2 _+0.8

70.5 _+ 11.2 73.5 + 6.7

12.6 + 1.6 13.6 _+ 0.5

62.0 _+ 3.0 56.5 + 2.2

1040.2 + 84.6 1160.0 _+60.0

42 day cont. 42 d a y E t O H

22.3 _+0.6 24.6 _+0.9

63.0 +_2.7 61.0 + 3.3

13.2 + 0.6 11.5 + 0.7

64.2 + 4.3 66.8 _+ 3.6

948.0 +_ 38.5 I 121.0 + 31.2

39 T A B L E II Histological examination of liver specimens Treatment group

Micro steatosis

Macro steatosis

Comments

21 day control 21 day E t O H

6/6 = 0 1/6 = +

6/6 = 0 6/6 = 0

None None

28 day control 28 day E t O H

6/6 = 0 6/6 = +

6/6 = 0 6/6 = 0

None None

35 day control 35 day E t O H

6/6 = 0 4/4 = +

6/6 = 0 6/6 = 0

a b

42 day control 42 day E t O H

6/6 = 0 5/5 = +

6/6 = 0 2/5 = +

c None

a, 1 mouse had evidence of slight degenerative changes; b, 1 m o u s e had evidence of inflammatory cell proliferation; c, 1 m o u s e had Councilman's body.

we observed that the magnitude of suppression increased along with the length of exposure. In the Summary in the lower right-hand corner of Fig. 3, the values were calculated by dividing the response of each of the ethanol-treated mice by the respective mean of the control mice for that time point. We believe that the variability of the effects and responses were good, especially in light of the fact that these values are based on 4 mice per treatment group. We have repeated this experiment and obtained results which were almost identical to those obtained in the first study (data not shown). Interestingly, our results with an exposure regimen of 35 days in the current study (i.e., a 47~o suppression) are also in good agreement with preliminary results (data not shown) in which we observed a 54 ~o suppression after 38 days using the liquid diet in which the controls had unlimited access to the control diet (i.e., this study was conducted before we adopted the pair-feeding restricted control exposure paradigm). Effects of chronic exposure to ethanol via a liquid diet on mitogen-induced proliferation Spleen cell suspensions were prepared from mice maintained on either the control liquid diet or the

ethanol liquid diet for the indicated periods of time and were stimulated with either Con A or LPS. The results for the optimal responses to both mitogens are shown in Table III. In contrast with the results with the in vivo AFC response to a TD antigen (Fig. 3), there were no effects by ethanol on either mitogen at any timepoint. Similar results were observed at all mitogen concentrations (data not shown). Effects of chronic exposure to ethanol via a liquid diet on in vitro antibody responses At the time that the preceding results were generated, we believed that there were at least two possible explanations for the differential effects on the in vivo AFC response and the mitogen responses. First, these results could be indicative of a selective action of ethanol on the differentiation of the AFC, which appeared to occur in the absence of any defect in the proliferative capabilities of either B- or T-cells. Second, these results could be indicative of an action on immune function which was an indirect consequence of exposure in the animal as opposed to a direct action on lymphocyte function. In order to test these two possibilities we decided to evaluate the AFC response by splenocytes from ethanoltreated mice which were sensitized to the antigen in vitro, as opposed to an in vivo immunization. The results for the polyclonal (PCL), TI and TD AFC responses are shown in Fig. 4. The results are plotted as percent of control in order to show all three AFC response models on the same graphs. The Summary in the lower right-hand comer of Fig. 4 is for the TD AFC response. It is important to note that there was considerable variability among the various AFC responses measured at the various timepoints. The control responses for the polyclonal AFC response were 364 + 2, 200 + 27, 487 + 8, 244 + 51 and 232 + 9; the control responses for the TI AFC response were 550 + 21, 188 + 23, 661+ 5, 308 +42, and 298 + 22; and the control responses for the TD AFC response were 854 _+94, 283 + 38, 702 + 50, 987 + 211 and 1494 + 212 for the 14-, 21-, 28-, 35and 42-day exposure groups, respectively. It is

40 21-day Exposure

14-day Exposure O

1600

800

13. 03 "O qJ :=, O

03

1200

600

g

400

o

6oo

~D < 0

200'

< Q

400

0

0

Control

EtOH

Control

35-day Exposure

28-day Exposure o

,-1 O. 03

1000

C~

r,~o < o

u.

1600

.=1 13. 03

8OO

"o > 0 u

600

i/

400 2000

"1o

12 O0

0(J

800

ne ~D

400 u.

0 Control

EtOH

Control

SUMMARY

42-day Exposure .J

1200

EtOH Treatment

Treatment

c)

EtOH Treatment

Treatment

"-F-

120

03 m "1o

o 800

-W

O

o

o

-W

400

/

o

u.

"--7--"

90 6O

u

•

el

30 0

0

EtOH

Control Treatment

14

21

28

35

42

Exposure (days)

Fig. 3. Effects of chronic exposure to ethanol in a liquid diet on the in vivo antibody r e s p o n s e to sheep erythrocytes. Female B6C3F1 mice were administered either a 5% ethanol liquid diet or a resticted pair control liquid diet for either 14, 21, 28, 35 or 42 days. Mice were sensitized (by i.v. injection) with 1 × l0 s S R B C in 0.2 ml of EBSS. F o u r days "after sensitization and 24 h after the exposure regimens defined above, the mice were sacrificed a n d the T - d e p e n d e n t antibody r e s p o n s e was m e a s u r e d . T h e bars represent the m e a n + SE from 4 mice per t r e a t m e n t group. Similar results were obtained in a second experiment. ** = p < 0.0 l a n d

41 TABLE III Effects of chronic exposure to ethanol on mitogen-induced proliferation Treatment group

Concanavalin A ~t

Lipopolysaccharidea

14 day control 14 day EtOH

22.3 _+ 1.4 19.1 _+ 1.1

9.8 _+0.6 10.5 + 1.1

21 day control 21 day EtOH

29.7 -+ 1.2 27.5 _+2.6

18.4 -+ 1.3 18.6 _+ 1.4

28 day control 28 day EtOH

41.5 + 0.4 42.1_+0.6

12.0 _+0.7 12.1_+0.9

35 day control 35 day EtOH

21.8 _+ 1.5 24.3_+2.1

11.1 _+ 1.9 12.3_+0.4

42 day control 42 day EtOH

37.5 + 2.3 41.1+1.3

13.2 + 2.2 11.6_+0.6

Responses are reported as absolute cpms x 10 4. a Results are from the optimal concentrations of the mitogens, 2/~g/ml for Con A and 100/~g/ml for LPS.

also important to note that with one possible exception (i.e., the TD AFC response in the 21-day group), these control response, although admittedly variable, are within our historical ranges for these AFC response models. The most important observation derived from the results in Fig. 4 is that the effects of ethanol on the in vitro A F C response are markedly different from the effects on the in vivo A F C response. We saw no consistent or marked suppression of any of the A F C response models. In fact, it appeared that the in vitro TD A F C response was increased in the ethanol-treated mice (i.e., note in particular the significant increase in the TD AFC response at 28 days). Based on our original premise, these results suggest that the effects of ethanol are indirectly mediated in the whole animal as a consequence of chronic exposure.

Characterization of the effects of mouse serum on the in vitro antibody response

In an attempt to begin to characterize the possible basis for the indirect effects of ethanol in chronically exposed mice, we decided to set out to determine the role by a serum factor. Historically, mouse serum has been difficult to work with in regards to supporting a variety of in vitro immune functional responses (Veit and Michael, 1973). We initially decided to eliminate any problem that endogenous complement may play by destroying these serum-borne factors by heatinactivation. In our initial preliminary results (data not shown), we were able to get surprisingly good results, in terms of the magnitude of the A F C response (i.e., within the historical range of A F C responses which are done in the presence of FBS), but we were disappointed by the lack of consistency. We noted in those cultures which had very low responses, a number of large cells with long pseudopods, which appeared to be firmly attached to the bottom of the wells. We speculated that these cells were some type of macrophage-like cell, which were not consistently present in the splenocyte cultures; but which could become highly activated by some component of mouse serum and cause suppression of the A F C response when they were present. In order to reduce the inherent variability associated with these cells, we decided to deplete them from our cultures using a 4-h adherence step in plastic petri dishes prior to adjusting the cells for use in the AFC response with mouse serum. In the initial studies for this investigation, we decided to determine the effects of adding normal mouse serum to splenocytes from untreated mice cultured with 5~o FCS. The results are shown in Fig. 5 and indicate that the response by 5.0~o FCS is unaffected by the addition of N M S at concentrations up to 3.0~o. We have extensively

* = p < 0 . 0 5 as determined by Dunnett's t-test using the restricted control group from each time point as the control. In the summary, the values were calculated by dividing the response of each of the ethanol-treated mice by the respective mean of the control mice for that time point.

in

t-

0

rll

0

ca

"

4 J

D

0

I~ C~

0

0

0

0

0

.~ 0

0"} C,

¢o 0

0 0

,...& I',0 0

._L

PERCENT CONTROL

0

PERCENT CONTROL

~ 0

._&

0

e-

Z

0

13

5

"t

<

0

0"} 0

,.~

0

j

-

~

o

0

~

O

0

~

0

0

- -

~

0

_

~

O

~

I

O

PERCENT CONTROL

0

PERCENT CONTROL

O

0

i

~

•

I"

o

Q

&

0

J J

.~

O

O

O

O

O

O

PERCENT CONTROL

.=L

PERCENT CONTROL

.I

O

,--~

!

I

&

O O~ C

X

43 characterized this type of titration experiment and have observed that there is little to no effect when NMS, at concentrations up to 5.0~o, is added to cultures supplemented with 5.0~o FCS (data not shown). We do consistently observe a decreased response when the concentration of NMS approaches 10.0~o (data not shown). Because we observed that as much as 5.0~o NMS could be added to FCS-supplemented cultures without any reduction in the response, we decided to determine the feasibility of culturing cells in the presence of NMS without any FCS supplement. These results are shown in Fig. 6 and indicate that with some very simple manipulations (i.e., a 4-h adherence step, as described above), we can obtain antibody responses with NMS only, which are as good as any responses that we routinely get with the more traditionally used FCS. It should be noted that with normal mouse splenocytes cultured in the presence of 3.0~o NMS only (i.e., no FCS, or any other type of supplement, serum or otherwise) we observed a response of over 2500 A F C / 1 0 6 recovered splenocytes. Having demonstrated the feasibility of using mouse serum in the AFC response model, we decided to compare the responses in the presence of NMS, serum taken from mice maintained on the restricted control diet and serum from mice maintained on the ethanol diet. Two of our initial trials using serum from ethanol-treated mice and the restricted controls are presented in Fig. 7. The serum samples for the restricted controls and the ethanol-treated mice were from the 42-day exposure period from our two studies to-date; i.e., these samples were from mice exposed in two

3000 O z w

Q W

2500

2000 1500

O

1000

o

500

5% F C S only

5% FCS + 0.5% NMS

5% FCS + 1.0% NMS

5% F C S + 3.0% NMS

TREATMENT

Fig. 5. Titration of normal mouse serum (NMS) into cultures supported with 5% fetal calf serum (FCS). Splenocytes from untreated B6C3F1 mice were pooled and cultured in the presence of either 5% fetal calf serum (FCS) alone, or 5% FCS plus concentrations of normal mouse serum (NMS) from 0.5% to 3.0%. The cultures were immunized with SRBC and the T-dependent antibody response was measured after five days of incubation. Results represent the mean _+ SE of quadruplicate wells for each treatment group and are representative of three separate experiments.

different studies. Moreover, in addition to using serum from two different treatment groups, it is important to emphasize that these studies were conducted over two months apart. The results are notable for several reasons. (1) We were able to obtain an unexpectedly large antibody response with cells cultured in NMS only, i.e., in Trial # 2, we were approaching a response with a magnitude in the low end of our historical range of responses with FCS-supplemented cultures (~ 1000 AFC/106 recovered splenocytes). (2)

Fig. 4. Effects of chronic exposure to ethanol in a liquid diet on in vitro antibody responses. Female B6C3F1 mice were administered either a 5% ethanol liquid diet or a resticted pair control liquid diet for either 14, 21, 28, 35 or 42 days. Twenty-four h after the last day of exposure, the mice were sacrificed and splenocyte suspensions were prepared for individual mice. The splenocytes were immunized with either lipopolysaccharide (LPS), DNP-Ficoll or SRBC for the polyclonal (PCL), T-independent (TI) and T-dependent (TD) antibody responses, respectively. The PCL, TI and TD antibody responses for each mouse were determined after 2 days, 3 days or 5 days of culture, respectively. The bars represent the mean _+ SE from 4 mice per treatment group. The values were calculated by dividing the response of each of the ethanol-treated mice by the respective mean of the control mice for that time point. * = p < 0.05 as determined by Dunnett's t-test using the restricted control group from each time point as the control. In the summary, the values for the T-dependent response from the ethanol-treated mice at all time points is plotted again for comparative purposes to the summary in Fig. 3.

44

3000 l

m

2500

®

L) O

,,=, K. U3 r~ LU LLI

O
RPMI only

5% FCS

2% NMS

3% NMS

4% NMS

5% NMS

TREATMENT

Fig. 6. Comparison of the response of normal mouse splenocytes when cultured with fetal calf serum (FCS) vs normal mouse serum (NMS). Splenocytes from untreated B6C3F1 mice were pooled and cultured in the presence of either media only (RPMI), 5% fetal calf serum (FCS) alone, or concentrations of normal mouse serum (NMS) alone from 2.0% to 5.0%. The cultures were immunized with SRBC and the T-dependent antibody response was measured after five days of incubation. Results represent the mean + SE of quadruplicate wells for each treatment group and are representative of three separate experiments.

There was no appreciable difference in the response of cells cultured in NMS when compared to those cultured with serum from the restricted control (i.e., RC) mice, i.e., we observed a 21~, increase with the RC serum in Trial # 1 and a 28 ~o decrease with the RC serum in Trial # 2. (3) In spite of the fact that the ethanol-treated serum was from two different experiments, there was excellent agreement, in that both lots produced almost exactly the same magnitude of suppression (i.e., 83~o in Trial 1 and 84~o in Trial 2) when compared to the appropriate restricted control serum. Ethanol content in serum f r o m mice maintained on the liquid diet

Having demonstrated that serum from ethanoltreated mice (i.e., 42-day group) could suppress the response of cells from untreated mice, we

determined the ethanol content in serum samples from mice maintained on the liquid diets for 14-42 days. As anticipated, there was no ethanol detected in the serum samples from any of the mice maintained on the restricted control diet. The amount of ethanol present in the serum samples from mice maintained on the ethanol liquid diet was highly variable from mouse to mouse (note: in particular the variability in the mice from the 35-day ethanol treatment). Four mice maintained on the ethanol liquid diet for 14 days had serum ethanol contents of 610, 290, 470 and 120 mg/1. Six mice maintained on the ethanol liquid diet for 21 days had ethanol contents of 10, 10, 10, 0, 10 and 30 mg/1. Four mice maintained on the ethanol liquid diet for 28 days had ethanol contents of 10, 20, 600 and 40 mg/1. Six mice maintained on the ethanol liquid diet had ethanol contents of 20, 20, 1270, 100, 910 and 40 mg/1. Six mice were maintained on the ethanol diet for 42 days. One serum sample was inadvertently discarded. The ethanol content in the five remaining samples were 220, 640, 480, 0 and 910 mg/l. Although there was great mouse to mouse variability in the ethanol content, this type of profile was not unexpected for an animal with a small body mass and high rate of metabolic activity. We have no explanation for the extremely low ethanol content in the 21-day treatment group, or for the fact that two mice maintained on the ethanol liquid diet had no ethanol in their serum samples (i.e., one each from the 21-day treatment and the 42-day treatment groups). One could speculate that these mice did not consume any of the ethanol diet during the hour immediately prior to the time that they were sacrificed and were therefore able to clear the ethanol. Because the serum samples used in the in vitro antibody responses, as described above, were pools from several mice maintained on a given treatment, we also calculated the mean serum ethanol content for the 5 treatment groups: 14-day, 372 mg/1; 21-day, 12 mg/1; 28-day, 167 mg/1; 35-day, 393 mg/1 and 42-day, 450 mg/1. Although there was obvious variability in some of the groups (i.e., 21-day and 28-day), there was a

45 TRIAL

TRIAL

#1

U)

#2

1000

>0 0

o

Z =,

Q. if) Q

(n a LU

0

o

,o Iz o

o

83%~

84%~, <

<

2o0,

×\\\\\\\\\\\~

5% NMS

5% Res. Cont.

5"/o EtOH

TREATMENT

5% NMS

5% Res. Cont.

5% EtOH

TREATMENT

Fig. 7. Mouse serum comparison: normal mouse serum (NMS), serum from restricted control mice and serum from ethanol-treated mice. Serum was collected from unimmunized mice maintained on either the ethanol liquid diet or the restricted control diet for 42 days. Splenocytes from untreated B6C3F1 mice were pooled and cultured in the presence of 5 % normal mouse serum (NMS), 5 % serum from mice maintained on the restricted control diet (Res. Cont.) or 5 % serum from mice maintained on the ethanol diet (EtOH). The cultures were immunized with SRBC and the T-dependent antibody response was measured after five days of incubation. Results represent the mean + SE of quadruplicate wells for each treatment. Serum collected from exposed mice in two separate studies are shown.

consistency in the ethanol content for other groups (i.e., 14-day, 35-day and 42-day). Most importantly, unlike the results shown in Fig. 3 and Table II, there was little or no change in the ethanol content associated with the length of exposure on the ethanol liquid diet. Compare in particular the ethanol content in serum from mice exposed to ethanol for 14 days versus 42 days.

Discussion

As highlighted in the Introduction, our studies to characterize the effects of ethanol on the immunocompetence of female B6C3F1 mice were initiated in light of results, which suggested that the basis for the observed effects of ethanol on immunity in other animal models and in man maybe a combination of direct and indirect effects. Two lines of evidence supported the interpretation that the immunosuppression associated

with exposure to ethanol is simply not an indirect consequence associated with poor nutrition. First and most importantly, ethanol-induced immunotoxicity can be demonstrated in animal studies where the diet can be controlled. Rats exposed to ethanol had a decreased hypersensitivity response to dinitrofluorobenzene and a delayed antibody response to both typhoid H. and B. abortus antigens (Tennenbaum et al., 1969), a decreased antibody response to sheep erythrocytes (Loose et al., 1975) and a decreased lymphoproliferative response to both B- and T-cell mitogens (Jerrells et al., 1986). In agreement with these studies using the rat, humans drinking chronically in a controlled environment also failed to develop antibody to a new antigen (Gluckman et al., 1977). The second line of evidence is that ethanol-induced immunotoxicity can be demonstrated by direct exposure to immunocompetent cells in culture and a number of labs have reported on the direct effects of ethanol

46 on immune function. Ethanol can decrease mitogen-induced proliferation of lymphocytes from humans (Lundy et al., 1975; Roselle and Mendenhall, 1982; Glassman et al., 1985) and guinea pigs (Roselle and Mendenhall, 1984), antibody-dependent cell-mediated cytotoxicity (Stacey, 1984; Walia and Lamon, 1989), and the activity of natural killer (NK) cells (Ristow et al., 1982). More recently, Aldo-Benson (1989) demonstrated that direct exposure to ethanol inhibited the antibody response by antigen-stimulated B-cell lines. However, it is important to emphasize that in our hands and others (Jerrells et al., 1989), the concentrations of ethanol and acetaldehyde required to produce direct suppression were quite high. As such, it is difficult to ascribe a role for direct effects by either ethanol or its primary metabolite on immune function because the concentrations required to produce such effects are unreasonably high to be obtained as blood alcohol or blood acetaldehyde levels. We believe that the key observations of our paper are fourfold: first, that the immunosuppression associated with chronic exposure to a liquid diet of ethanol is characterized by a definite temporal relationship (see Fig. 3); second, that there was at least the suggestion that the immunosuppression and the liver pathology associated with chronic exposure to ethanol were correlated with time (compare Fig. 3 and Table II); third, that the suppression of the antibody response was manifested when the antigen was administered to ethanol-treated mice (see Fig. 3), and not when added directly to cultured splenocytes prepared from mice exposed to ethanol under identical conditions (see Fig. 4); and four, that cells from untreated mice cultured in the presence of serum from ethanol-treated mice were suppressed relative to cells cultured in serum from the pair-fed restricted controls (see Fig. 7). We have formulated a model in which the basis for the ethanol-induced changes in immune function is an indirect consequence of exposure in the animal. As discussed above, because we employed a pair-feeding paradigm, we do not believe that the indirect effects are related to differences

in nutrition. A type of indirect model, which is unrelated to nutritional concerns, was also recently described by Jerrells et al. (1989). Their studies strongly implicated a role for an elevation in endogenous corticosteroids in the suppression associated with acute exposure to ethanol and upon withdrawal from ethanol in dependent animals. While their evidence is quite convincing, we do not believe that a similar mechanism can account for our results in a chronic exposure model. As shown in Table I, there were no differences in body and organ weights between the ethanol-treated mice and their restricted pair fed controls. The lack of an effect by ethanol on spleen weight, spleen cell number and thymus weight are in marked contrast to previous work described by Jerrells et al. (1986) using a multiple gavage regimen for ethanol exposure in the rat. Although Tabakoff et al. (1978) had previously demonstrated that the levels of glucocorticoids were increased in mice maintained on a liquid diet, our results are not consistent with that interpretation. A marked decrease in thymus weight is a hallmark indicator of an increase in adrenal glucocorticoids in the mouse, and as shown in Table I, there was no such effect by ethanol in our study. The lack of an effect on thymus weight in mice chronically exposed to ethanol in the liquid diet is consistent with previous results by Guaza etal. (1983) who showed that acute, but not chronic, exposure to ethanol caused an elevation in serum corticosterone.

The model that we have formulated is that the ethanol-induced immunotoxicity is mediated by an indirect effect via the release of serum borne immunomodulatory factors, possibly as a consequence of the ethanol-induced damage to the liver. There are several observations in the literature, which are consistent with such a mechanism. Guinea pigs which were fed alcohol for 4 weeks at doses that were low enough to maintain normal liver histology were successfully sensitized to keyhole limpet hemocyanin (Caiazza and Ovary, 1976). A similar, i.e., negative, study was reported in mice in which the long-term patho-

47 logical effects of alcohol were avoided and the immunocompetence was not affected (Caren et al., 1983). Perhaps even more pertinent to this proposal are the reports in the literature which describe individuals having either liver damage or liver disease and simultaneously being markedly immunocompromised. Immunosuppression accompaning hepatic injury has often been reported to be associated, in part, with a factor present in the serum of patients. Studies utilizing peripheral blood lymphocytes from human patients with alcohol induced liver cirrhosis demonstrated that these lymphocytes exhibited subnormal P H A stimulation, but only when these lymphocytes were cultured in the presence of serum from cirrhotic donors (Tobias et al., 1967; Winters et al., 1966). Lymphocytes from cirrhotic donors cultured in the sera from normal donors were capable of normal stimulation. These results were later confirmed by Hsu and Leevy (1971). Similar results were observed in patients with biliary cirrhosis (Fox et al., 1969). Studies evaluating the proliferative responses of lymphocytes from patients with acute hepatitis and aggressive chronic hepatitis also exhibited decreased responses to P H A stimulation (Rossler et al., 1969; and Willems et al., 1969). The possible involvement of a serum factor has also been implicated in the immunosuppression by ethanol in alcoholics in a number of investigations (Hsu and Leevy, 1971; Newberry et al., 1973; Young et al., 1978 and 1979; Roselle et al., 1988). A recent paper from Roselle and co-workers has demonstrated a similar profile of activity when old rats (i.e., 18 months) were exposed to an ethanol liquid diet for three months - cells from the ethanol-treated rats were most immunocompromised when they were cultured in serum from these animals (Roselle etal., 1989). More importantly, this study demonstrated that cells from normal rats were also markedly suppressed when they were cultured in the presence of serum from the ethanol-treated animals. While this earlier study used rats as the animal model and mitogeninduced proliferation as the immune parameter, the profile of activity is in excellent agreement

with our results using the mouse and the T-dependent antibody response. In spite of the apparent similarities observed in a number of studies, which have used both human and animal models, there are no reports that anyone has attempted to further characterize the immunosuppressive serum factor described. We have measured the ethanol content in serum from mice maintained on the liquid diet for 14-42 days. We believe that these results are consistent with our interpretation that the consumption of ethanol produces minimal to no direct effects on immunocompetence. First, although there was tremendous mouse to mouse variability across the treatment groups, there was little to no suggestion of a temporal relationship in the serum ethanol concentrations which could account for the increasing suppression of the in vivo antibody response as the length of time on the ethanol liquid diet was increased. Compare the suppression observed in the 14- and 42-day treatments, no effect and 70 ~o, respectively, with the mean serum ethanol content in the same groups, 372 and 450 mg/1, respectively. Second, and perhaps most convincingly, the concentrations of ethanol in the serum from the 42-day treatment were far too low to produce any effects on the antibody response by cells from untreated mice when tested at 5~o. Assuming that the concentration of ethanol in the serum of mice exposed to ethanol for 42 days was 500 mg/1, if 5 ~o serum was added to the cultures of splenocytes from untreated mice, then the concentration of ethanol in the wells would be < 0.003 ~o. As shown in Fig. 1, this concentration was two orders of magnitude less than the concentration of ethanol, 0.3 ~o, which was devoid of activity. Even if all of the mice in a given group had serum levels of 1000 rag/1 (i.e., it is important to emphasize that only one mouse had serum which contained this level of ethanol), the legally impaired level in Virginia, the concentration in the cultures would be only 0.005 ~o, which is still off the dose-response curve depicted in Fig. 1. In fact, based on the results in Fig. 1, a serum sample added at 5 ~o would have to contain 20 ~o

48 or 200,000 rag/1 ethanol to reach the effective concentration of 1 ~o ethanol in the cultures. Therefore, we believe that we have ruled out a role by the ethanol content in the serum for the suppression associated with serum from ethanoltreated mice. We believe that a detailed analysis of the serum factor(s) was beyond the scope of the present investigation; but that it represents a primary focus for future studies by our laboratory. For reasons discussed above, we do not believe that our results are consistent with the possible involvement of corticosteroids as the serum factors. However, exposure to ethanol can produce other changes in neuroendocrine status, so we cannot rule out the involvement of other hormones. One of the initial steps in our characterization of the ethanol serum will be to determine if the suppressive factor is primarily associated with the protein or lipid component of the serum. The association of the serum-derived immunosuppression with a protein would be consistent with a possible role by a hormone whose levels are changed as a result of the effects of ethanol on the neuroendocrine system. Such an indirect mechanism of immune modulation was one of the primary focuses of a recent symposium (Sanders et al., 1991). A second focus of this symposium was that indirect immune modulation could also be mediated through immunomodulatory protein factors released into the serum as a result of damage to the liver. As demonstrated in our study, the development of the suppression of the in vivo antibody response and tt'.-e manifestation of liver damage (fatty liver) were correlated with the time of exposure to ethanol in the liquid diet. Therefore the association of the serum-derived immunosuppression with a protein would also be consistent with a possible role by an immunomodulatory factor of hepatic origin. There are a number of nonhormone proteins released following various types of liver damage, which produce potent effects on immune function, including serum amyloid A (SAA), alpha-fetoprotein (AFP) and liver-derived inhibitory protein (LIP). Previous results from this lab (Kaminski and Holsapple,

1987) demonstrated a marked immunosuppression in mice treated (s.c.) with casein, which is a potent trigger for the release of SAA from hepatocytes. As recently reviewed by Deutsch (1991), one of the most potent actions of A F P is its immunoregulatory activity, which can be demonstrated with serum or amniotic fluid containing measurable levels of AFP. Initial results from Schumacher and co-workers characterized LIP as a protein of 67 k D a to 80 k D a which could be isolated from the liver extracts of patients with chronic hepatitis (Grol and Schumacher, 1983). Subsequent results from the same laboratory identified LIP as L-arginase (Brusdeilins et al., 1983). Therefore, there are a number of immunomodulatory factors released into the serum as a result of damage to the liver. If the suppression by serum from ethanol-treated mice can be associated with the protein fraction, then it will be important to determine whether ethanol-induced liver damage (i.e., which in the mouse is restricted to the development of fatty liver) is a suitable trigger for the release of a liver-derived immunosuppressive factor. However, it is important to emphasize that the factor(s) need not be a protein. A number of recent papers have demonstrated that exposure to ethanol produces pronounced changes in serum lipid and lipoprotein content. In particular, Lin et al. (1989) demonstrated in the rat that chronic exposure to ethanol caused significant increases in the serum levels of triglycerides and cholesterol after a one-week exposure and that the levels remained elevated for up to seven weeks of ethanol feeding. Branchey and Buysden-Branchey (1990) demonstrated that plasma total cholesterol, free (unesterified) cholesterol and HDL-cholesterol were higher in rats maintained on an ethanol liquid diet for 4 weeks when compared to the pair-fed controls. In light of the immunomodulation by lipoproteins (Macy et al., 1983), the possibility that the lipid fraction of serum from ethanol-treated mice is responsible for the observed immunosuppression must be considered. To summarize, our results indicate that direct

49 addition of ethanol to cells from untreated mice is immunosuppressive; but only at concentrations which could not be obtained in the blood following consumption of ethanol; that the suppression of the in vivo antibody response associated with chronic exposure to ethanol increases with time and is not correlated with changes in the serum ethanol levels; that there, is no suppression of the in vitro antibody response by cells from ethanoltreated mice suggesting that these cells "recover" when cultured in a supportive environment; and that serum from ethanol-treated mice is immunosuppressive and that this suppression cannot be attributed to the ethanol content of the serum. While our results cannot completely rule out that ethanol exposure has a direct effect on immunocompetence, we believe that this profile of activity is most consistent with the possibility that the immunosuppression by chronic exposure to ethanol is an indirect consequence of the exposure which is unrelated to nutritional differences; but which is mediated by a serum-derived factor. Much more work is needed to further characterize the serum from ethanol-treated mice and it is anticipated that the studies to isolate and identify the immunosuppressive factor(s) will be quite labor-intensive. However, we are encouraged by the success that we have had to date incorporating mouse serum into in vitro assays with mouse splenocyte cultures.

Acknowledgements This work was supported in part by the Faculty Grants Program of the Commonwealth of Virginia Center on Drug Abuse and by Training Grant No. NIH-ES-07087.

References Aldo-Benson, M. (1989) Mechanisms of alcohol-induced suppression of B-cell response. Alcohol. Clin. Exp. Res. 13: 469. Bailey, R.J., Krasner, N., Eddleston, A.L.W.F., Williams, R., Tee, D.E.H., Donlach, D., Kennedy, L.A., Batchelor, J.R.

(1976) Histocompatibility antigens, autoantibodies, and immuno-globulins in alcoholic liver disease. Br. Med. J. 2: 727. Berenyi, M.R., Straus, B. and Cruz, D. (1974) In vitro and in vivo studies of cellular immunity in alcoholic cirrhosis. Am. J. Dig. Dis. 119: 199. Bernstein, I.M., Webster, K.H., Williams, R.C., Jr. and Strickland, R.G. (1974) Reduction in circulating T-lymphocytes in alcoholic liver disease. Lancet 2: 488. Branchey, M.H. and Buysden-Branchey, L. (1990) Are the effects of chronic ethanol administration on erythrocyte membrane mediated by changes in plasma lipids? Drug Alcohol Depend. 25: 67. Brusdeilins, M., Haple, C., Huberts, H.H. and Schumacher, K. (1983) Identification of the apparently lymphocytespecific human liver-derived protein (LIP) as cytoplasmic liver L-arginase. J. Immunol. 131: 2427. Caiazza, S.S. and Ovary, Z. (1976) Effects of ethanol intake on the immune system of guinea pigs. J. Stud. Alcohol. 37: 959. Caren, L.D., Leveque, J.A. and Mardel, A.D. (1983) Effect of ethanol on the immune system in mice. Toxicol. Lett. 19: 147. Deutsch, H. (1991) Chemistry and biology of alphafetoprotein. Adv. Cancer Res. 56: 253. Drew, P.A., Clifton, P.M., LaBroay, J.T. and Shearman, D.J.C. (1984) Polyclonal B cell activation in alcoholic patients with no evidence of liver dysfunction. Clin. Exp. Immunol. 57: 479. Fox, R.A., James, D.G., Scheuer, P.J., Sharma, O. and Sherlock, S. (1969) Impaired delayed hypersensitivity in primary biliary cirrhosis. Lancet 1: 959. Glassman, A.B., Bennett, C.E. and Randall, C.L. (1985) Effects of ethyl alcohol on human peripheral blood lymphocytes. Arch. Pathol. Lab. Med. 109: 540. Gluckman, S.J., Dwrak, V.C. and MacGregor, R.R. (1977) Host defenses during prolonged alcohol consumption in a controlled environment. Arch. Intern. Med. 137: 1539. Grol, M. and Schumacher, K. (1983) Purification and biochemical characterization of human-derived inhibitory protein (LIP). J. Immunol. 130: 323. Guaza, C., Torrellas, A. and Borrell, S. (1983) Adrenocortical response to acute and chronic ethanol administration in rats. Pschopharmacology 79: 173. Holsapple, M.P., Munson, A.E., Munson, J.A. and Bick, P.H. (1983) Suppression of cell-mediated immunocompetence following subchronic exposure to diethylstilbestrol in female B6C3F1 mice. J. Pharmacol. Exp. Ther. 227: 130. Holsapple, M.P., McNerney, P.J., Barens, D.W. and White, K.L. Jr. (1984) Suppression of humoral antibody production by exposure to 1,2,3,6,7,8-hexachlorodibenzo-pdioxin. J. Pharmacol. Exp. Ther. 231: 518. Hsu, C.C.S. and Leevy, C.M. (1971) Inhibition of PHA stimulated lymphocyte transformation by plasma from

50 patients with advanced alcoholic cirrhosis. Clin. Exp. Immunol. 8: 749. Jerrells, T.R., Marietta, C.A., Eckardt, M.J., Majchrowiez, E. and Weight, F.F. (1986) Effects of ethanol administration on parameters of immunocompetency in rats. J. Leuk. Biol. 39: 499. Jerrells, T.R., Peritt, D., Marietta, C. and Eckardt, M.J. (1989) Mechanisms of suppression of cellular immunity induced by ethanol. Alcohol. Clin. Exp. Res. 13: 490. Johnson, K.W., Munson, A.E. Holsapple, M.P. (1987) Primary cellular target responsible for dimethylnitrosamineinduced immuno-suppression in the mouse. Immunopharmacology 13: 47. Kaminski, N.E. and Holsapple, M.P. (1987) Inhibition of macrophage accessory cell function in casein-treated B6C3FI mice. J. Immunol. 139: 1804. Lieber, C.S. and DeCarli, L.M. (1982)The fedding of alcohol in liquid diets: Two decades of applications and 1982 update. Alcohol. Clin. Exp. Res. 6: 523. Lieber, C.S. and DeCarli, L.M. (1986) The feeding of ethanol in liquid diets. Alcohol. Clin. Exp. Res. 10: 550. Lieber, C.S. and DeCarli, L.M. (1989) Review: Liquid diet technique of ethanol administration: 1989 update. Alcohol Alcoholism 24: 197. Lin, R.C., Lumeng, L. and Phelps, V.L. (1989) Serum high-density lipoprotein particles of alcohol-fed rats are deficient in apolipoprotein E. Hepatology 9: 307. Loose, L.D. Stege, T. and Diluzio, N.R. (1975) The influence of acute and chronic ethanol or bourbon administration on phagocytic and immune responses in rats. Exp. Mol. Pathol. 23: 159. Lundy, J., Raaf, J., Deakins, S. Wamebo, H., Jacobs, D., Lee, T. and Jacobwitz, D. (1975) The acute and chronic effects of alcohol on the human immune system. Surg. Gynec. Obstet. 141: 211. MacGregor, R.R. (1986) Alcohol and immune defense. J.Am. Med. Assoc. 256: 1474. Macy, M., Okano, Y., Cardin, A.D., Avila, E.M. and Harmony, J .A.K. (1983) Suppression of lymphocyte activation by plasma lipoproteins. Cancer Research 43: 2496. Morgan, M.Y., Ross, M.G.R., Ng, C.M., Adams, D.M., Thomas, H.C., and Sherlock, S. (1980) HLA-BS, immunoglobulins, and antibody responses in alcohol-related liver disease. J. Clin. Pathol. 33: 488. Mufti, S.I., Prabhala, R., Moriguchi, S., Sipes, I.G. and Watson, R.R. (1988) Functional and numerical alterations induced by ethanol in the cellular immune system. Immunopharmacology 15: 85. Newberry, W.M., Shorey, J.W., Sanford, J.P. and Combes, B. (1973) Depression of lymphocyte reactivity to phytohemagglutinin by serum from patients with liver disease. Cell. Immunol. 6: 87. Palmer, D.L. (1981) Alcoholism, infection and immunity. In: Infection and the Compromised Host. Ed. J.C. Allen, Raven, New York, 121.

Ristow, S.S., Starkey, J.R. and Hass, G.M. (1982) Inhibition of natural killer cell activity in vitro by alcohols. Biochem. Biophys. Res. Commun. 105: 1315. Rittenberg, M.B. and Pratt, K.L. (1969) Antinitrophenyl (TNP) plaque assay. Primary response of Balb/c mice to soluble and particulate immunogen. Proc. Soc. Exp. Biol. Med. 132: 575. Roselle, G.A., Mendenhall, C.L. and Grossman, C.J. (1989) Age dependent alterations of host immune response in the ethanol-fed rat. J. Clin. Lab. Immunol. 29: 99. Roselle, G.A., Mendenhall, C.L., Grossman, C.J. and Weesner, R.E. (1988) Lymphocyte subset alterations in patients with alcoholic hepatitis. J. Clin. Lab. Immunol. 26: 169. Roselle, G. and Mendenhall, C. (1984) Ethanol induced alterations in lymphocyte function in the guinea pig. Alcoholism Clin. Exp. Res. 8: 62. Roselle, G. and Mendenhall, C. (1982) Alteration of in vitro lymphocyte function by ethanol, acetaldehyde and acetate. J. Clin. Lab. Immun. 9: 33. Rossler, R., Havemann, K. and Dolle, W. (1969) Unterschiedliche Reaktion von Lymphocyten auf Phytohaemagglutinin (PHA) bei Lebererkrankungen. Klin. Wschr. 47: 803. Sanders, V.M., Fuchs, B.A., Pruett, S.B., Kerkvliet, N.I. and Kaminski, N.E. (1991) Symposium Overview: Symposium on indirect mechanisms of immune modulation. Fund. Appl. Tox. 17: 641. Schopf, R.E., Trompeter, M., Bork, K. and Morsches, B. (1985) Effects of ethanol and acetaldehyde on phagocytic functions. Arch. Dermatol. Res. 277: 131. Stacey, N.H. (1984) Inhibition of antibody-dependent cellmediated cytotoxicity by ethanol. Immunopharmacology 8: 155. Tabakoff, G., Jaffe, R.C. and Ritzmann, R.F. (1978) Corticosterone concentrations in mice during ethanol drinking and withdrawal. J. Pharm. Pharmacol. 30: 371. Tennenbaum, J.I., Ruppert, R.D., St. Pierre, R.L. and Greenberger, N.J. (1969) The effect of chronic alcohol on the immune response of rats. J. Allergy Clin. lmmunol. 44: 272. Tobias, H., Safran, A.P. and Schaffer, F. (1967) Lymphocyte stimulation and chronic liver disease. Lancet h 193. Veit, B. and Michael, J.G., (1973) Characterization of an immunosuppressive factor present in mouse serum. J. Immunol. 111: 341. Walia, A.S. and Lamon, E.W. (1989) Ethanol inhibition of cell-mediated lysis of antibody-sensitized target cells at a calcium-dependent step. Proc. Soc. Exp. Biol. Med. 192: 177. Watson, R.R., Prabhala, R.H., Abril, E. and Smith, T.L. (1988) Changes in lymphocyte subsets and macrophage functions from high, short-term dietary ethanol in C57BL/6 mice. Life Sci. 43: 865. Willems, F.T.C., Meinick, J.L. and Rawls, W.E. (1969) Viral

51 inhibition of the Phytohemagglutinin response of human lymphocytes and application to viral hepatitis. Proc. Soc. Exp. Biol. 130: 652. Winters, G.C.B., McCarthy, C.F., Reading, A.E. and Yoffey, J.M. (1966) Development of macrophages in Phytohememagglutinin cultures of blood from patients with idiopathic steatorrhoea and with cirrhosis. Br. J. Exp. Pathol. 49: 66.

Young, G.P., Dudley, F.J. and Van Der Weyden, M.B. (1979) Suppressive effects of alcoholic liver disease sera on lymphocyte transformation. Gut 20: 833. Young, G.P., Van Der Weyden, M.B. Rose, I.S. and Dudley, F.J. (1978) Lymphopenia and lymphocyte transformation in alcoholics. Experientia 35: 268.