- Email: [email protected]
Effect of antibiotics on clinical, pathologic and immunologic responses in murine Potomac horse fever: Protective effects of doxycycline
Veterinary Microbiology, 19 (1989) 253-262 Elsevier Science Publishers B.V., Amsterdam - - Printed in The Netherlands
253
Effect of A n t i b i o t i c s on Clinical, P a t h o l o g i c and I m m u n o l o g i c R e s p o n s e s in M u r i n e P o t o m a c Horse Fever: P r o t e c t i v e Effects of D o x y c y c l i n e YASUKO RIKIHISA and BAOMING M. JIANG
Department of Veterinary Pathobiology, College of Veterinary Medicine, The Ohio State University, I925 Cof[ey Road, Columbus, OH 43210-1092 (U.S.A.) (Accepted for publication 27 October 1988)
ABSTRACT Rikihisa, Y. and Jiang, B.M., 1989. Effect of antibiotics on clinical, pathologic and immunologic responses in murine Potomac horse fever: protective effects of doxycycline. Vet. Microbiol., 19: 253-262. Effects of three antibiotics on clinical, pathologic and immunologic responses in murine Potomac horse fever caused by Ehrlichia risticii infection were examined. When antibiotics were given after the development of clinical signs, antibiotics ranked in the order of reducing clinical signs and in preventing body weight loss and an intestinal enlargement were doxycycline, demeclocycline and rifampin. Infected mice treated with doxycycline and demeclocycline developed greater splenomegaly than rifampin-treated or untreated infected mice. All antibiotics used prevented thymic atrophy due to E. risticii infection. Indirect fluorescent antibody titers were highest with doxycycline treatment. Mice treated with demeclocycline and rifampin produced higher antibody titer than those without treatment. Ehrlichia risticii was reisolated from the spleens of both untreated and rifampin-treated infected mice. The effects of administering single doses of doxycycline at different times after infection were examined. Body weight loss was prevented by the drug given at every treatment day examined, i.e. Days 3, 5 and 7 post-infection (PI). Thymic atrophy was minimum in mice treated at Day 5 PI, while splenomegaly was found on every treatment day. Splenocyte proliferative response to concanavalin A and lipopolysaccharide, and specific antibody development against E. risticii was best in mice treated at Day 5 PI followed by those treated at Day 3 and Day 7 PI.
INTRODUCTION
Infection with Ehrlichia risticii, the etiologic agent of Potomac horse fever (Rikihisa and Perry, 1985; Holland et al., 1985 ), results in high morbidity and mortality (Knowles et al., 1983). Signs in horses include loss of appetite, depression, fever, leukopenia, dehydration, explosive diarrhea, and, in some cases, laminitis (Knowles et al., 1983). The effect of specific antimicrobial therapy against E. risticii has had limited investigation. Oxytetracycline inter0378-1135/89/$03.50
© 1989 Elsevier Science Publishers B.V.
254
fered with the development of Potomac horse fever, but did not prevent ponies from developing the disease (Palmer et al., 1988). Demeclocycline, doxycycline and oxytetracycline were found to be the most effective with respect to their inhibitory activity on E. risticii in vitro; minocycline, tetracycline and rifampin showed a moderate effect; erythromycin and nalidixic acid had relatively poor inhibitory effects on this organism (Rikihisa and Jiang, 1988). Our previous study also showed that mice were susceptible to E. risticii and infected mice developed acute and profound immunodepression (Rikihisa et al., 1987). Thus, the present study examined the effect of two of the most effective antibiotics in vitro, demeclocycline and doxycycline, and one of moderately effective antibiotics, rifampin, on mice infected with E. risticii. The results were assessed in terms of clinical signs, gross pathologic changes and host immune depression. MATERIALS AND METHODS
Mice
Female Sprague-Dawley CF-1 mice 4-8 weeks old, were obtained from Harlan Sprague Dawley, Indianapolis, IN. They were given commercial laboratory chow and water ad libitum. Mice were conditioned for 1 week before being used in experiments. Antibiotics
Doxycycline hydrochloride (Sigma Chemical Co., St. Louis, MO) was dissolved in phosphate buffered saline (2 mM KH2P04, 6 mM NaeHPO4, 2 mM KC1 and 136 mM NaC1, PBS pH 5.5). Demeclocycline kindly provided by American Cyanamid Co. (Pearl River, NY) and rifampin (Sigma) were dissolved in dimethyl sulfoxide (DMSO) and diluted in PBS, pH 5.5. Antibiotic stock solutions (10 mg m1-1) were prepared and further diluted in Hanks' balanced salt solution (HBSS) (GIBCO Laboratories, Grand Island, NY) on the day of use. Ehrlichia culture Ehrlichia risticii was propagated in P388D1 murine macrophage cells (American Type Culture Collection, Rockville, MD ) in 150-cm 2 plastic tissueculture flasks (Corning Glass Works, Corning, NY) in RPMI 1640 media (GIBCO) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (GIBCO) and 2 mM L-glutamine (GIBCO) at 37°C in a humidified atmosphere of 5% COz and 95% air. The cells were harvested when they were more
255
than 90% infected as assessed by Diff-Quik or indirect fluorescent antibody staining using positive anti-E, risticii serum (Rikihisa and Perry, 1985 ). Ehrlichia infection and antibiotic treatment of mice A total of 14 8-week-old mice were each inoculated intraperitoneally (i.p.) with 0.2-ml H B S S containing 3 X 10 ~ E. risticii-infected P388D1 cells. Four control mice received the same number of uninfected P388D1 cells. Three groups of three mice each were given i.p. 0.2 ml of either demeclocycline, doxycycline, or rifampin in H B S S at Day 5 post-infection (PI). Five non-antibiotic-treated infected mice and all control uninfected mice were injected with 0.2 ml of H B S S at Day 5 PI. On 10 or 11 days PI, the mice were anesthetized with ether. After blood was collected from the axillary blood vessels, mice were killed by cervical dislocation. At necropsy, the whole animal, spleen, thymus and intestine were weighed.
Administration of doxycycline at different times P I A group of 17 4-week-old mice were each inoculated with 106 E. risticii-infected P388D1 cells, and five mice received the same number of uninfected P388D1 cells. Groups of four infected mice received doxycycline (10 mg kg -1 ) at 3, 5 or 7 days PI. Five infected and five uninfected mice were not treated with the antibiotic and served as positive and negative controls. All of the mice were sacrificed at Day 15 PI. Organ and body weights were measured, serum antibodies were titrated and spleen lymphocyte blast transformation assays were performed. The effects of doxycycline at 10 mg k g - 1 in four normal control mice on the morphologic changes of organs, and on immune responses were examined in a separate experiment.
Immunofluorescent antibody titration of mouse sera Approximately 103 E. risticii infected P388D1 cells were fixed in acetone in each well of 12-well teflon-coated multi-well slides (Carlson Scientific, Peotone, IL) and incubated with 2-fold serial dilutions of mouse sera in PBS. The cells were then incubated with fluorescein isothiocyanate ( F I T C ) -labeled goat anti-mouse IgG (United State Biochemical Corp., Cleveland, O H ) diluted 1:200 in P B S (pH 7.4). The slides were washed in P B S containing 0.002% Tween 20, counterstained with Evans blue solution, and again washed with PBS. The covered glass slides were examined using a Nikon Microphot-FX fluorescent microscope at X 400. The highest dilution of serum which showed a positive reaction was defined as the titer of the antibody. The control assay was per-
256
formed by incubating infected P388D1 cells with serum of uninfected mice or by incubating uninfected P388D1 cells with serum of E. risticii-infected mice.
L ymphoblast transformation assay The spleen was asceptically removed and placed in a 50-cm 2 petri dish to which 20 ml of HBSS was added. The spleen was smashed with the blunt end of a 10-ml syringe through a Cellector (Thomas Scientific, Swedesboro, NJ) to release the cells. Each cell suspension was drawn into a 10-ml syringe through a 21-gauge needle and forced back into the petri dish. After this procedure had been repeated twice, the cell suspension was centrifuged at 430 × g for 5 min. The cell pellet was resuspended in 3 ml 0.85% NH4C1 by vortexing and left at room temperature for 2 min. Sterile FBS (1.5 ml) was slowly released to underlay the cell suspensions. After centrifugation at 430 × g for 5 min, the pellet was resuspended and vortexed in 5-ml RPMI 1640 containing 5% FBS, 2 mM L-glutamine, 50 mM Hepes (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid) buffer (GIBCO), 5 mM sodium pyruvate (GIBCO), 100 IU m1-1 penicillin (GIBCO), 100/~g ml- 1 streptomycin (GIBCO). The cell concentration was adjusted to 1.0 × 107 cells ml-1 and 0.1 ml cell suspension was plated in individual wells of 96-well microtiter plates (Flow Lab., McLean, VA). The cells were stimulated by the addition of a predetermined optimum final concentration of concanavalin A (Con A; 2.5 ~g ml-1; Miles Laboratories, Naperville, IL ) and lipopolysaccharide (LPS; 32/~g ml- 1; from E. coli 0111:B4 Sigma). Each assay was done in triplicate for each mouse. The plates were incubated for 66 h at 37°C in a humidified atmosphere of 5% C02 and 95% air. One-half zCi [3H] thymidine (New England Nuclear, Boston, MA) was added to each culture for 6 h. The cells were harvested onto glass microfiber 934-AH filters (Whatman, Clifton, NJ) by a Model M-24 Brandel Cell Harvester (Biomedical Research and Development Lab, Gaithersburg, MD). Samples were dissolved in pre-mixed scintillation solution (Econofluor, NEN Research Products, Boston, MA) and the amount of radioactivity incorporated into the cellular DNA was determined by standard liquid scintillation counting using a TriCarb Liquid Scintillation counter (Model 2000CA; Packard Instrument Co., Downers Grove, IL).
Reisolation of E. risticii from spleen Spleens were asceptically removed from the mice and the splenic cell suspensions were prepared as previously described. The cells were suspended in RPMI 1640 supplemented with 10% heat-inactivated FBS and 2 mM L-glutamine and added to 150-cm2 confluent P388D1 cell monolayers. The cultures were incubated at 35 ° C in a humidified atmosphere of 5% CO2 and 95% air for 7 days. At the end of the incubation period, presence of any E. risticii in the
257
culture was determined by Diff-Quik and indirect fluorescent antibody staining using positive anti-E, risticii serum.
Statistical analysis Statistical significance was determined by analysis of variance and Tukey's studentized range test for all experiments; P < 0.05 was considered significant. RESULTS
Comparative effects of three antibiotics on E. risticii-infected mice At Day 5 PI all mice inoculated with 3 X 106 E. risticii-infected P388D1 cells showed clinical signs of illness. In untreated mice, ruffled fur, crouching, inactivity, anorexia, followed by diarrhea which occurred near death. One mouse died at Day 10 PI. In comparison, infected mice treated with demeclocycline, doxycycline and rifampin exhibited less ruffling of fur and lethargy; no diarrhea was observed (Table 1). No mortality had occurred in antibiotic-treated infected mice by Day 11. Loss of body weight was evident in untreated and rifampin-treated infected mice (Table 2). Doxycycline, however, prevented body weight loss due to E. risticii infection whereas demeclocycline had an intermediate effect in preventing body weight loss (Table 2). Mice inoculated with uninfected P388D1 cells manifested no clinical signs of illness (Table 1 ). Untreated, infected mice had higher whole intestinal weight than uninfected mice (Table 2). Infected mice treated with doxycycline had significantly reTABLE 1 Signs of illness in E. risticii-infected mice after t r e a t m e n t with doxycycline, demeclocycline and rifampin Mouse t r e a t m e n t '
Uninfected control Infected control Demeclocycline Doxycycline Rifampin
No. of animals 4 5 3 3 3
Clinical signs~ Ruffled fur
Lethargy
. + + + + + + + + + +
. + + + + + + + + + +
.
Diarrhea
Death
+ :~ ----
+ :~ ----
.
'Mice were given i.p. 3 X 106 uninfected or E. risticii-infected P388D1 murine macrophage cells. After 5 days, three groups of infected mice were given a single injection i.p. of demeclocycline, doxycycline or rifampin at 10 mg k g - 1 body weight. 2The severity of signs was expressed, in a decreasing order: + + + + , + + + , + + , + , with - indicating the absence of signs of illness. :~One mouse developed diarrhea and died 10 days PI.
258 TABLE 2 Organ weights of E. risticii-infected mice after t r e a t m e n t with doxycycline, demeclocycline and rifampin Mouse t r e a t m e n t ~
Uninfected control Infected control Demeclocycline Doxycycline Rifampin
No. of animals
4 5 3 3 3
Body weight (g)2
25.2 ± 0.7 a 22.1 _+2.2 h 24.4 ± 1.6 a'~' 26.4 ± 0.5 a 22.3_+ 1.0"
Relative organ weight (g per 100 g body weight) 2 Intestine
Spleen
Thymus
8.1 _+0.5 ~ 15.3 ± 1.1 a 13.5 ± 1.9 a'~, 11.6_+0.5 ~''c 14.6 ± 1.7 a'b
0.37 _+0.02" 1.42 _+0.25 ~' 2.38 ± 0.33 a 2.20 ±0.19 a 1.50 ±0.13 b
0.28 _+0.03 a 0.058 + 0.035 b 0.27 ± 0.04 a 0.30 +0.08 ~ 0.21 ±0.01 a
1Mice were given i.p. 3 × 10 ~ uninfected or E. risticii-infected P388D1 murine macrophage cells. After 5 days, three groups of infected mice were given a single injection i.p. of demeclocycline, doxycycline or rifampin at 10 mg kg 1 body weight. ~Means _+SD. For each relative organ or body weight, means with the different superscript letter are significantly different by Tukey's studentized range test.
duced intestinal weights in comparison to untreated infected mice (Table 2). Demeclocycline and rifampin did not prevent the increase of intestinal weight (Table 2). When compared with mice injected with uninfected P388D1 cells, all infected mice had significant splenomegaly (Table 2). Splenomegaly was greatest in mice treated with demeclocycline and doxycycline (Table 2). Rifampin-treated infected mice had the same degree of splenomegaly as untreated infected mice (Table 2). Thymic atrophy was significant in untreated, infected mice (Table 2 ). All three antibiotics prevented the occurrence ofthymic atrophy due to E. risticii infection (Table 2 ). All the infected mice developed IgG antibody against E. risticii. Doxycyclinetreated mice had a slightly higher antibody titer (one mouse 1:160, two mice 1:320) than the other antibiotic-treated (three mice 1:160) or untreated (two mice 1:80, two mice 1:160), infected mice. E. risticii was reisolated from untreated and rifampin-treated infected mouse spleens but not from doxycycline or demeclocycline-treated infected mouse spleens. Effect of doxycycline administered at different days P I on infected mice The untreated mice began to manifest rough coats and squinty eyes at Day 9 PI with l0 s E. risticii-infected P388D1 cells. One untreated infected mouse died at Day 14 PI, and another untreated infected mouse developed diarrhea at Day 15 PI. In contrast, no clinical signs were observed when doxycycline was given at Day 3 or 5 PI but slightly rough fur coat was found among mice
259 TABLE 3 Effect of time of single dose a d m i n i s t r a t i o n of doxycycline on organ weights ofE. risticii-infected mice Mouse t r e a t m e n t 1
Uninfected control Infected control Doxycycline (Day 3 ) Doxycycline (Day 5 ) Doxycycline (Day 7 )
No. of animals
5 4 4 4 4
Body weight (g)2
17.2 _+0.6 ~ 12.3 +_ 1.8 b 17.0 _+ 1.4" 17.3 +_0.5 a 16.1 _+_1.2 a
Relative organ weight (g per 100 g body weight) 2 Spleen
Thymus
0.36 _+0.04" 1.39 _+0.26 b 1.53 _+0.34 b 1.71 + 0.20 b 1.62 + 0.14 b
0.57 _+0.09" 0.29 _+0.09 b 0.32 _+0.02 b 0.44 _+0.05 a'b 0.39 _+0.03 b
1Mice were given i.p. 106 uninfected or E. risticii-infected P388D 1 murine macrophage cells. T h r e e groups of infected mice were given a single injection i.p. of doxycycline at 10 mg kg-1 body weight on Days 3, 5 or 7 PI, respectively. 2Means_+ SD. For each organ or body weight, means with the same superscript letter are not significantly different by Tukey's studentized range test. TABLE4 Splenocyte proliferative response to mitogens in doxycycline-treated E. risticii-infected mice ~ Mouse t r e a t m e n t 2
Mitogen
3H-thymidine uptake (cpm)
Uninfected control
None Con A LPS
23631 _+2525 89248 +_3870 31518 +_6756
3.8 1.3
None Con A LPS
256 _+48 166 + 23 67_+ 15
0.7 0.3
Doxycycline (given on Day 3 PI)
None Con A LPS
540 _+ 188 7968 _+863 4199 _+332
14.8 7.8
Doxycycline (given on Day 5 PI)
None Con A LPS
545 _+88 25076 _+1949 4654 _+290
46.0 8.5
Doxycycline (given on Day 7 PI )
None Con A LPS
386 _+34 2864 _+650 143 +_47
7.4 0.5
Infected control
SI '~
~Mice were killed at Day 15 PI. Spleen cells were cultured for 66 h in the presence of 2.5 #g ml 1 Con A or 32 l~g m l - ~ LPS, t h e n pulsed with 3H-thymidine for 6 h. Results shown are mean cpm per 106 cells +_S.E. for three mice. ~Uninfected control mice received 106 uninfected P388D1 cells; infected a n d doxycycline treated mice received 106 E. risticii-infected P388D 1 cells. :*Stimulation index calculated by dividing the response in the presence of mitogen by t h a t found with cells alone.
260
treated at Day 7 PI. The administration of doxycycline at Day 3, 5 or 7 PI all significantly protected the mice from body weight loss (Table 3 ). Splenomegaly was identified among all infected mice (Table 3). Relative thymic weight was reduced in both untreated and doxycycline-treatedinfected mice. Less thymic atrophy however, was observed when doxycycline administered at Day 5 PI than when it was administered on either Day 3 or 7 PI (Table 3). Table 4 illustrates an impaired proliferative response to Con A and LPS in spleen cells of the untreated infected mice. 3H-thymidine uptake per spleen cell was lower than controls in both untreated and treated infected mice without mitogen. With doxycycline treatment, however, significantly higher proliferative responses to both mitogens were observed than in untreated infected or uninfected mice. The proliferative response was largest with doxycycline given on Day 5 PI, followed by Day 3 and Day 7 PI. At Day 15 PI pooled sera from mice treated at Day 5 PI showed higher antibody titer (1:320) to E. risticii than those treated at Day 3 PI (1:160) or 7 PI (1:80), or untreated (1:40). The mice inoculated with 106 uninfected P388D1 cells and control normal mice did not have detectable antibodies against E. risticii at 1:20 dilution by IFA tests. Neither significant morphologic change nor suppressed proliferation of spleen cells due to doxycycline treatment alone was observed in mice (data not shown). DISCUSSION
As opposed to in vitro results (Rikihisa and Jiang, 1988), doxycycline was superior to demeclocycline in treating experimentally induced Potomac horse fever in mice. This superior effectiveness may be owing to increased lipid solubility, longer half-life, and intestinal rather than renal excretion of doxycycline (Shaw and Rubin, 1986). In agreement with in vitro data, reisolation of E. risticii from infected and rifampin-treated mice indicates that a single dose of rifampin at Day 5 PI was not effective for treating Potomac horse fever in mice. All infected mice given antibiotics after the development of clinical signs had greater splenomegaly than untreated infected controls. This may have resulted from proliferation of lymphoid tissue. Catanzaro et al. (1976) observed similar splenomegaly due to enlargement of the white pulp in BALB/c mice infected with Karp strain after immunization with Gilliam strain of Rickettsia tsutsugamushi. A decreased Con A and phytohemagglutinin-induced splenocyte blastogenesis occurred in the mice inoculated with E. risticii-infected U937 cells (Rikihisa et al., 1987). The present study confirmed a marked reduction in blastogenesis of spleen cells in response to Con A and LPS in E. risticii infection. These results indicated that the responses of both T- and B-lymphocytes were
261
suppressed. Since hypoimmune responses were observed in mice injected with both E. risticii-infected murine P388D1 cells or E. risticii-infected human U937 cells, it seems that the immune response depended primarily upon the presence of Ehrlichiae, and that the host cell types were not a major factor. Doxycycline treatment counteracted this immunosuppression as demonstrated by higher antibody titer and stimulation index of splenocyte in response to mitogens. In experimental murine scrub typhus, the outcome of infections depended on the time chloramphenicol therapy was initiated. Initiation of chemotherapy early in the infection delayed development of immune responsiveness, whereas delay of chloramphenicol treatment, allowing initial proliferation of rickettsiae, resulted in rapid development of immunity and the accelerated ability to transfer protection with splenic lymphocytes (Shirai et al., 1977 ). The present experiment demonstrated that the development of the immune response during doxycycline treatment in Ehrlichia infection of mice was closely related to the time of initiation of treatment relative to infection. Mice treated at 3 or 7 days after infection developed lower splenocyte proliferative responses and serum antibody titers at Day 15 PI compared with the mice treated at Day 5 PI, in which significantly higher cellular and humoral immune responses were apparent. These results indicate that a delicate balance between the intensity of host's immunization and the proliferation of E. risticii appears to be important for effective antibiotic treatment. Earlier treatment apparently leads to inhibited replication of the Ehrlichiae and reduced immune stimulation. The treatment of infected mice at Day 5 PI resulted in an effective immune response. This may be due to the accumulation of a critical antigen mass. The depressed immune response of the mice receiving the late treatment may attribute to the host's inability to overcome the proliferation of E. risticii, which in turn causes lymphoid depletion (Rikihisa et al., 1987). Magnitude and onset of clinical, pathologic and immunologic changes and responses to antibiotic treatment varied depending on dosages of the organism given and ages of mice used. For example, higher dosage of the organism gives earlier and more severe clinical signs, pathologic changes and immunodepression (Rikihisa et al., 1987 ). Relative thymic weight is larger and thymus structure is more affected in E. risticii infection in younger mice. Differences observed in thymic weights in our two parts of the present study reflected the dosage and age differences between the two parts of the experiments. Although the representative data shown in this paper consist of relatively few numbers of mice in each group, repeated similar experiments gave similar results as shown in this paper. Present results demonstrated that pathologic and immunologic changes in mice infected with E. risticii serve as useful parameters in evaluating antibiotics and mice can serve as a useful laboratory model for screening antibiotics for treating horses with Potomac horse fever. Since there is a good correlation
262 between antibiotic effectiveness and positive immune responses in infected mice, measurement of immune responses in infected horses after treatment with antibiotics may also aid in evaluating antibiotics for Potomac horse fever. ACKNOWLEDGEMENTS
This work was supported by USDA Special Grant 85-CRSR-2-2735. We thank A. Walton for her editorial help with the preparation of this manuscript and Michele Martino for her technical assistance.
REFERENCES Catanzaro, P.J., Akira, S., Hildebrandt, P.K. and Osterman, J.V., 1976. Host defenses in experimental scrub typhus: histopathological correlates. Infect. Immun., 13: 861-875. Holland, C.J., Weiss, E., Burgdorfer, W., Cole, A.I. and Kakoma, I., 1985. Ehrlichia risticii sp nov: etiologic agent of equine monocytic ehrlichiosis (syn. Potomac horse fever). Int. J. Syst. Bacteriol., 35: 524-526. Knowles, R.C., Anderson, C.W., Shipley, W.D., Whitlock, R.H., Perry, B.D. and Davidson, J.P., 1983. Acute equine diarrhea syndrome (AEDS): A preliminary report. Proc. Annu. Meet. Am. Assoc. Equine Pract., pp. 353-357. Palmer, J.E., Whitlock, R.H. and Benson, C.E., 1988. Equine ehrlichial colitis: Effect of oxytetracycline treatment during the incubation period of Ehrlichia risticii infection in ponies. J. Am. Vet. Med. Assoc., 192: 343-345. Rikihisa, Y. and Perry, B.D., 1985. Causative ehrlichial organisms in Potomac horse fever. Infect. Immun., 49: 513-517. Rikihisa, Y. and Jiang, B.M., 1988. In vitro susceptibility of Ehrlichia risticii to eight antibiotics. Antimicrob. Agents Chemother., 32: 986-991. Rikihisa, Y., Johnson, G.C. and Burger, C.J., 1987. Reduced immune responsiveness and lymphoid depletion in mice infected with Ehrlichia risticii. Infect. Immun., 55: 2215-2222. Shaw, D.H. and Rubin, S.I., 1986. Pharmacologic activity of doxycycline. J. Am. Vet. Med. Assoc., 189: 808-810. Shirai, A., Catanzaro, P.J., Eisenberg Jr., G.H.G. and Osterman, J.V., 1977. Host defenses in experimental scrub typhus: effect of chloramphenicol. Infect. Immun., 18: 324-329.
253
Effect of A n t i b i o t i c s on Clinical, P a t h o l o g i c and I m m u n o l o g i c R e s p o n s e s in M u r i n e P o t o m a c Horse Fever: P r o t e c t i v e Effects of D o x y c y c l i n e YASUKO RIKIHISA and BAOMING M. JIANG
Department of Veterinary Pathobiology, College of Veterinary Medicine, The Ohio State University, I925 Cof[ey Road, Columbus, OH 43210-1092 (U.S.A.) (Accepted for publication 27 October 1988)
ABSTRACT Rikihisa, Y. and Jiang, B.M., 1989. Effect of antibiotics on clinical, pathologic and immunologic responses in murine Potomac horse fever: protective effects of doxycycline. Vet. Microbiol., 19: 253-262. Effects of three antibiotics on clinical, pathologic and immunologic responses in murine Potomac horse fever caused by Ehrlichia risticii infection were examined. When antibiotics were given after the development of clinical signs, antibiotics ranked in the order of reducing clinical signs and in preventing body weight loss and an intestinal enlargement were doxycycline, demeclocycline and rifampin. Infected mice treated with doxycycline and demeclocycline developed greater splenomegaly than rifampin-treated or untreated infected mice. All antibiotics used prevented thymic atrophy due to E. risticii infection. Indirect fluorescent antibody titers were highest with doxycycline treatment. Mice treated with demeclocycline and rifampin produced higher antibody titer than those without treatment. Ehrlichia risticii was reisolated from the spleens of both untreated and rifampin-treated infected mice. The effects of administering single doses of doxycycline at different times after infection were examined. Body weight loss was prevented by the drug given at every treatment day examined, i.e. Days 3, 5 and 7 post-infection (PI). Thymic atrophy was minimum in mice treated at Day 5 PI, while splenomegaly was found on every treatment day. Splenocyte proliferative response to concanavalin A and lipopolysaccharide, and specific antibody development against E. risticii was best in mice treated at Day 5 PI followed by those treated at Day 3 and Day 7 PI.
INTRODUCTION
Infection with Ehrlichia risticii, the etiologic agent of Potomac horse fever (Rikihisa and Perry, 1985; Holland et al., 1985 ), results in high morbidity and mortality (Knowles et al., 1983). Signs in horses include loss of appetite, depression, fever, leukopenia, dehydration, explosive diarrhea, and, in some cases, laminitis (Knowles et al., 1983). The effect of specific antimicrobial therapy against E. risticii has had limited investigation. Oxytetracycline inter0378-1135/89/$03.50
© 1989 Elsevier Science Publishers B.V.
254
fered with the development of Potomac horse fever, but did not prevent ponies from developing the disease (Palmer et al., 1988). Demeclocycline, doxycycline and oxytetracycline were found to be the most effective with respect to their inhibitory activity on E. risticii in vitro; minocycline, tetracycline and rifampin showed a moderate effect; erythromycin and nalidixic acid had relatively poor inhibitory effects on this organism (Rikihisa and Jiang, 1988). Our previous study also showed that mice were susceptible to E. risticii and infected mice developed acute and profound immunodepression (Rikihisa et al., 1987). Thus, the present study examined the effect of two of the most effective antibiotics in vitro, demeclocycline and doxycycline, and one of moderately effective antibiotics, rifampin, on mice infected with E. risticii. The results were assessed in terms of clinical signs, gross pathologic changes and host immune depression. MATERIALS AND METHODS
Mice
Female Sprague-Dawley CF-1 mice 4-8 weeks old, were obtained from Harlan Sprague Dawley, Indianapolis, IN. They were given commercial laboratory chow and water ad libitum. Mice were conditioned for 1 week before being used in experiments. Antibiotics
Doxycycline hydrochloride (Sigma Chemical Co., St. Louis, MO) was dissolved in phosphate buffered saline (2 mM KH2P04, 6 mM NaeHPO4, 2 mM KC1 and 136 mM NaC1, PBS pH 5.5). Demeclocycline kindly provided by American Cyanamid Co. (Pearl River, NY) and rifampin (Sigma) were dissolved in dimethyl sulfoxide (DMSO) and diluted in PBS, pH 5.5. Antibiotic stock solutions (10 mg m1-1) were prepared and further diluted in Hanks' balanced salt solution (HBSS) (GIBCO Laboratories, Grand Island, NY) on the day of use. Ehrlichia culture Ehrlichia risticii was propagated in P388D1 murine macrophage cells (American Type Culture Collection, Rockville, MD ) in 150-cm 2 plastic tissueculture flasks (Corning Glass Works, Corning, NY) in RPMI 1640 media (GIBCO) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (GIBCO) and 2 mM L-glutamine (GIBCO) at 37°C in a humidified atmosphere of 5% COz and 95% air. The cells were harvested when they were more
255
than 90% infected as assessed by Diff-Quik or indirect fluorescent antibody staining using positive anti-E, risticii serum (Rikihisa and Perry, 1985 ). Ehrlichia infection and antibiotic treatment of mice A total of 14 8-week-old mice were each inoculated intraperitoneally (i.p.) with 0.2-ml H B S S containing 3 X 10 ~ E. risticii-infected P388D1 cells. Four control mice received the same number of uninfected P388D1 cells. Three groups of three mice each were given i.p. 0.2 ml of either demeclocycline, doxycycline, or rifampin in H B S S at Day 5 post-infection (PI). Five non-antibiotic-treated infected mice and all control uninfected mice were injected with 0.2 ml of H B S S at Day 5 PI. On 10 or 11 days PI, the mice were anesthetized with ether. After blood was collected from the axillary blood vessels, mice were killed by cervical dislocation. At necropsy, the whole animal, spleen, thymus and intestine were weighed.
Administration of doxycycline at different times P I A group of 17 4-week-old mice were each inoculated with 106 E. risticii-infected P388D1 cells, and five mice received the same number of uninfected P388D1 cells. Groups of four infected mice received doxycycline (10 mg kg -1 ) at 3, 5 or 7 days PI. Five infected and five uninfected mice were not treated with the antibiotic and served as positive and negative controls. All of the mice were sacrificed at Day 15 PI. Organ and body weights were measured, serum antibodies were titrated and spleen lymphocyte blast transformation assays were performed. The effects of doxycycline at 10 mg k g - 1 in four normal control mice on the morphologic changes of organs, and on immune responses were examined in a separate experiment.
Immunofluorescent antibody titration of mouse sera Approximately 103 E. risticii infected P388D1 cells were fixed in acetone in each well of 12-well teflon-coated multi-well slides (Carlson Scientific, Peotone, IL) and incubated with 2-fold serial dilutions of mouse sera in PBS. The cells were then incubated with fluorescein isothiocyanate ( F I T C ) -labeled goat anti-mouse IgG (United State Biochemical Corp., Cleveland, O H ) diluted 1:200 in P B S (pH 7.4). The slides were washed in P B S containing 0.002% Tween 20, counterstained with Evans blue solution, and again washed with PBS. The covered glass slides were examined using a Nikon Microphot-FX fluorescent microscope at X 400. The highest dilution of serum which showed a positive reaction was defined as the titer of the antibody. The control assay was per-
256
formed by incubating infected P388D1 cells with serum of uninfected mice or by incubating uninfected P388D1 cells with serum of E. risticii-infected mice.
L ymphoblast transformation assay The spleen was asceptically removed and placed in a 50-cm 2 petri dish to which 20 ml of HBSS was added. The spleen was smashed with the blunt end of a 10-ml syringe through a Cellector (Thomas Scientific, Swedesboro, NJ) to release the cells. Each cell suspension was drawn into a 10-ml syringe through a 21-gauge needle and forced back into the petri dish. After this procedure had been repeated twice, the cell suspension was centrifuged at 430 × g for 5 min. The cell pellet was resuspended in 3 ml 0.85% NH4C1 by vortexing and left at room temperature for 2 min. Sterile FBS (1.5 ml) was slowly released to underlay the cell suspensions. After centrifugation at 430 × g for 5 min, the pellet was resuspended and vortexed in 5-ml RPMI 1640 containing 5% FBS, 2 mM L-glutamine, 50 mM Hepes (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid) buffer (GIBCO), 5 mM sodium pyruvate (GIBCO), 100 IU m1-1 penicillin (GIBCO), 100/~g ml- 1 streptomycin (GIBCO). The cell concentration was adjusted to 1.0 × 107 cells ml-1 and 0.1 ml cell suspension was plated in individual wells of 96-well microtiter plates (Flow Lab., McLean, VA). The cells were stimulated by the addition of a predetermined optimum final concentration of concanavalin A (Con A; 2.5 ~g ml-1; Miles Laboratories, Naperville, IL ) and lipopolysaccharide (LPS; 32/~g ml- 1; from E. coli 0111:B4 Sigma). Each assay was done in triplicate for each mouse. The plates were incubated for 66 h at 37°C in a humidified atmosphere of 5% C02 and 95% air. One-half zCi [3H] thymidine (New England Nuclear, Boston, MA) was added to each culture for 6 h. The cells were harvested onto glass microfiber 934-AH filters (Whatman, Clifton, NJ) by a Model M-24 Brandel Cell Harvester (Biomedical Research and Development Lab, Gaithersburg, MD). Samples were dissolved in pre-mixed scintillation solution (Econofluor, NEN Research Products, Boston, MA) and the amount of radioactivity incorporated into the cellular DNA was determined by standard liquid scintillation counting using a TriCarb Liquid Scintillation counter (Model 2000CA; Packard Instrument Co., Downers Grove, IL).
Reisolation of E. risticii from spleen Spleens were asceptically removed from the mice and the splenic cell suspensions were prepared as previously described. The cells were suspended in RPMI 1640 supplemented with 10% heat-inactivated FBS and 2 mM L-glutamine and added to 150-cm2 confluent P388D1 cell monolayers. The cultures were incubated at 35 ° C in a humidified atmosphere of 5% CO2 and 95% air for 7 days. At the end of the incubation period, presence of any E. risticii in the
257
culture was determined by Diff-Quik and indirect fluorescent antibody staining using positive anti-E, risticii serum.
Statistical analysis Statistical significance was determined by analysis of variance and Tukey's studentized range test for all experiments; P < 0.05 was considered significant. RESULTS
Comparative effects of three antibiotics on E. risticii-infected mice At Day 5 PI all mice inoculated with 3 X 106 E. risticii-infected P388D1 cells showed clinical signs of illness. In untreated mice, ruffled fur, crouching, inactivity, anorexia, followed by diarrhea which occurred near death. One mouse died at Day 10 PI. In comparison, infected mice treated with demeclocycline, doxycycline and rifampin exhibited less ruffling of fur and lethargy; no diarrhea was observed (Table 1). No mortality had occurred in antibiotic-treated infected mice by Day 11. Loss of body weight was evident in untreated and rifampin-treated infected mice (Table 2). Doxycycline, however, prevented body weight loss due to E. risticii infection whereas demeclocycline had an intermediate effect in preventing body weight loss (Table 2). Mice inoculated with uninfected P388D1 cells manifested no clinical signs of illness (Table 1 ). Untreated, infected mice had higher whole intestinal weight than uninfected mice (Table 2). Infected mice treated with doxycycline had significantly reTABLE 1 Signs of illness in E. risticii-infected mice after t r e a t m e n t with doxycycline, demeclocycline and rifampin Mouse t r e a t m e n t '
Uninfected control Infected control Demeclocycline Doxycycline Rifampin
No. of animals 4 5 3 3 3
Clinical signs~ Ruffled fur
Lethargy
. + + + + + + + + + +
. + + + + + + + + + +
.
Diarrhea
Death
+ :~ ----
+ :~ ----
.
'Mice were given i.p. 3 X 106 uninfected or E. risticii-infected P388D1 murine macrophage cells. After 5 days, three groups of infected mice were given a single injection i.p. of demeclocycline, doxycycline or rifampin at 10 mg k g - 1 body weight. 2The severity of signs was expressed, in a decreasing order: + + + + , + + + , + + , + , with - indicating the absence of signs of illness. :~One mouse developed diarrhea and died 10 days PI.
258 TABLE 2 Organ weights of E. risticii-infected mice after t r e a t m e n t with doxycycline, demeclocycline and rifampin Mouse t r e a t m e n t ~
Uninfected control Infected control Demeclocycline Doxycycline Rifampin
No. of animals
4 5 3 3 3
Body weight (g)2
25.2 ± 0.7 a 22.1 _+2.2 h 24.4 ± 1.6 a'~' 26.4 ± 0.5 a 22.3_+ 1.0"
Relative organ weight (g per 100 g body weight) 2 Intestine
Spleen
Thymus
8.1 _+0.5 ~ 15.3 ± 1.1 a 13.5 ± 1.9 a'~, 11.6_+0.5 ~''c 14.6 ± 1.7 a'b
0.37 _+0.02" 1.42 _+0.25 ~' 2.38 ± 0.33 a 2.20 ±0.19 a 1.50 ±0.13 b
0.28 _+0.03 a 0.058 + 0.035 b 0.27 ± 0.04 a 0.30 +0.08 ~ 0.21 ±0.01 a
1Mice were given i.p. 3 × 10 ~ uninfected or E. risticii-infected P388D1 murine macrophage cells. After 5 days, three groups of infected mice were given a single injection i.p. of demeclocycline, doxycycline or rifampin at 10 mg kg 1 body weight. ~Means _+SD. For each relative organ or body weight, means with the different superscript letter are significantly different by Tukey's studentized range test.
duced intestinal weights in comparison to untreated infected mice (Table 2). Demeclocycline and rifampin did not prevent the increase of intestinal weight (Table 2). When compared with mice injected with uninfected P388D1 cells, all infected mice had significant splenomegaly (Table 2). Splenomegaly was greatest in mice treated with demeclocycline and doxycycline (Table 2). Rifampin-treated infected mice had the same degree of splenomegaly as untreated infected mice (Table 2). Thymic atrophy was significant in untreated, infected mice (Table 2 ). All three antibiotics prevented the occurrence ofthymic atrophy due to E. risticii infection (Table 2 ). All the infected mice developed IgG antibody against E. risticii. Doxycyclinetreated mice had a slightly higher antibody titer (one mouse 1:160, two mice 1:320) than the other antibiotic-treated (three mice 1:160) or untreated (two mice 1:80, two mice 1:160), infected mice. E. risticii was reisolated from untreated and rifampin-treated infected mouse spleens but not from doxycycline or demeclocycline-treated infected mouse spleens. Effect of doxycycline administered at different days P I on infected mice The untreated mice began to manifest rough coats and squinty eyes at Day 9 PI with l0 s E. risticii-infected P388D1 cells. One untreated infected mouse died at Day 14 PI, and another untreated infected mouse developed diarrhea at Day 15 PI. In contrast, no clinical signs were observed when doxycycline was given at Day 3 or 5 PI but slightly rough fur coat was found among mice
259 TABLE 3 Effect of time of single dose a d m i n i s t r a t i o n of doxycycline on organ weights ofE. risticii-infected mice Mouse t r e a t m e n t 1
Uninfected control Infected control Doxycycline (Day 3 ) Doxycycline (Day 5 ) Doxycycline (Day 7 )
No. of animals
5 4 4 4 4
Body weight (g)2
17.2 _+0.6 ~ 12.3 +_ 1.8 b 17.0 _+ 1.4" 17.3 +_0.5 a 16.1 _+_1.2 a
Relative organ weight (g per 100 g body weight) 2 Spleen
Thymus
0.36 _+0.04" 1.39 _+0.26 b 1.53 _+0.34 b 1.71 + 0.20 b 1.62 + 0.14 b
0.57 _+0.09" 0.29 _+0.09 b 0.32 _+0.02 b 0.44 _+0.05 a'b 0.39 _+0.03 b
1Mice were given i.p. 106 uninfected or E. risticii-infected P388D 1 murine macrophage cells. T h r e e groups of infected mice were given a single injection i.p. of doxycycline at 10 mg kg-1 body weight on Days 3, 5 or 7 PI, respectively. 2Means_+ SD. For each organ or body weight, means with the same superscript letter are not significantly different by Tukey's studentized range test. TABLE4 Splenocyte proliferative response to mitogens in doxycycline-treated E. risticii-infected mice ~ Mouse t r e a t m e n t 2
Mitogen
3H-thymidine uptake (cpm)
Uninfected control
None Con A LPS
23631 _+2525 89248 +_3870 31518 +_6756
3.8 1.3
None Con A LPS
256 _+48 166 + 23 67_+ 15
0.7 0.3
Doxycycline (given on Day 3 PI)
None Con A LPS
540 _+ 188 7968 _+863 4199 _+332
14.8 7.8
Doxycycline (given on Day 5 PI)
None Con A LPS
545 _+88 25076 _+1949 4654 _+290
46.0 8.5
Doxycycline (given on Day 7 PI )
None Con A LPS
386 _+34 2864 _+650 143 +_47
7.4 0.5
Infected control
SI '~
~Mice were killed at Day 15 PI. Spleen cells were cultured for 66 h in the presence of 2.5 #g ml 1 Con A or 32 l~g m l - ~ LPS, t h e n pulsed with 3H-thymidine for 6 h. Results shown are mean cpm per 106 cells +_S.E. for three mice. ~Uninfected control mice received 106 uninfected P388D1 cells; infected a n d doxycycline treated mice received 106 E. risticii-infected P388D 1 cells. :*Stimulation index calculated by dividing the response in the presence of mitogen by t h a t found with cells alone.
260
treated at Day 7 PI. The administration of doxycycline at Day 3, 5 or 7 PI all significantly protected the mice from body weight loss (Table 3 ). Splenomegaly was identified among all infected mice (Table 3). Relative thymic weight was reduced in both untreated and doxycycline-treatedinfected mice. Less thymic atrophy however, was observed when doxycycline administered at Day 5 PI than when it was administered on either Day 3 or 7 PI (Table 3). Table 4 illustrates an impaired proliferative response to Con A and LPS in spleen cells of the untreated infected mice. 3H-thymidine uptake per spleen cell was lower than controls in both untreated and treated infected mice without mitogen. With doxycycline treatment, however, significantly higher proliferative responses to both mitogens were observed than in untreated infected or uninfected mice. The proliferative response was largest with doxycycline given on Day 5 PI, followed by Day 3 and Day 7 PI. At Day 15 PI pooled sera from mice treated at Day 5 PI showed higher antibody titer (1:320) to E. risticii than those treated at Day 3 PI (1:160) or 7 PI (1:80), or untreated (1:40). The mice inoculated with 106 uninfected P388D1 cells and control normal mice did not have detectable antibodies against E. risticii at 1:20 dilution by IFA tests. Neither significant morphologic change nor suppressed proliferation of spleen cells due to doxycycline treatment alone was observed in mice (data not shown). DISCUSSION
As opposed to in vitro results (Rikihisa and Jiang, 1988), doxycycline was superior to demeclocycline in treating experimentally induced Potomac horse fever in mice. This superior effectiveness may be owing to increased lipid solubility, longer half-life, and intestinal rather than renal excretion of doxycycline (Shaw and Rubin, 1986). In agreement with in vitro data, reisolation of E. risticii from infected and rifampin-treated mice indicates that a single dose of rifampin at Day 5 PI was not effective for treating Potomac horse fever in mice. All infected mice given antibiotics after the development of clinical signs had greater splenomegaly than untreated infected controls. This may have resulted from proliferation of lymphoid tissue. Catanzaro et al. (1976) observed similar splenomegaly due to enlargement of the white pulp in BALB/c mice infected with Karp strain after immunization with Gilliam strain of Rickettsia tsutsugamushi. A decreased Con A and phytohemagglutinin-induced splenocyte blastogenesis occurred in the mice inoculated with E. risticii-infected U937 cells (Rikihisa et al., 1987). The present study confirmed a marked reduction in blastogenesis of spleen cells in response to Con A and LPS in E. risticii infection. These results indicated that the responses of both T- and B-lymphocytes were
261
suppressed. Since hypoimmune responses were observed in mice injected with both E. risticii-infected murine P388D1 cells or E. risticii-infected human U937 cells, it seems that the immune response depended primarily upon the presence of Ehrlichiae, and that the host cell types were not a major factor. Doxycycline treatment counteracted this immunosuppression as demonstrated by higher antibody titer and stimulation index of splenocyte in response to mitogens. In experimental murine scrub typhus, the outcome of infections depended on the time chloramphenicol therapy was initiated. Initiation of chemotherapy early in the infection delayed development of immune responsiveness, whereas delay of chloramphenicol treatment, allowing initial proliferation of rickettsiae, resulted in rapid development of immunity and the accelerated ability to transfer protection with splenic lymphocytes (Shirai et al., 1977 ). The present experiment demonstrated that the development of the immune response during doxycycline treatment in Ehrlichia infection of mice was closely related to the time of initiation of treatment relative to infection. Mice treated at 3 or 7 days after infection developed lower splenocyte proliferative responses and serum antibody titers at Day 15 PI compared with the mice treated at Day 5 PI, in which significantly higher cellular and humoral immune responses were apparent. These results indicate that a delicate balance between the intensity of host's immunization and the proliferation of E. risticii appears to be important for effective antibiotic treatment. Earlier treatment apparently leads to inhibited replication of the Ehrlichiae and reduced immune stimulation. The treatment of infected mice at Day 5 PI resulted in an effective immune response. This may be due to the accumulation of a critical antigen mass. The depressed immune response of the mice receiving the late treatment may attribute to the host's inability to overcome the proliferation of E. risticii, which in turn causes lymphoid depletion (Rikihisa et al., 1987). Magnitude and onset of clinical, pathologic and immunologic changes and responses to antibiotic treatment varied depending on dosages of the organism given and ages of mice used. For example, higher dosage of the organism gives earlier and more severe clinical signs, pathologic changes and immunodepression (Rikihisa et al., 1987 ). Relative thymic weight is larger and thymus structure is more affected in E. risticii infection in younger mice. Differences observed in thymic weights in our two parts of the present study reflected the dosage and age differences between the two parts of the experiments. Although the representative data shown in this paper consist of relatively few numbers of mice in each group, repeated similar experiments gave similar results as shown in this paper. Present results demonstrated that pathologic and immunologic changes in mice infected with E. risticii serve as useful parameters in evaluating antibiotics and mice can serve as a useful laboratory model for screening antibiotics for treating horses with Potomac horse fever. Since there is a good correlation
262 between antibiotic effectiveness and positive immune responses in infected mice, measurement of immune responses in infected horses after treatment with antibiotics may also aid in evaluating antibiotics for Potomac horse fever. ACKNOWLEDGEMENTS
This work was supported by USDA Special Grant 85-CRSR-2-2735. We thank A. Walton for her editorial help with the preparation of this manuscript and Michele Martino for her technical assistance.
REFERENCES Catanzaro, P.J., Akira, S., Hildebrandt, P.K. and Osterman, J.V., 1976. Host defenses in experimental scrub typhus: histopathological correlates. Infect. Immun., 13: 861-875. Holland, C.J., Weiss, E., Burgdorfer, W., Cole, A.I. and Kakoma, I., 1985. Ehrlichia risticii sp nov: etiologic agent of equine monocytic ehrlichiosis (syn. Potomac horse fever). Int. J. Syst. Bacteriol., 35: 524-526. Knowles, R.C., Anderson, C.W., Shipley, W.D., Whitlock, R.H., Perry, B.D. and Davidson, J.P., 1983. Acute equine diarrhea syndrome (AEDS): A preliminary report. Proc. Annu. Meet. Am. Assoc. Equine Pract., pp. 353-357. Palmer, J.E., Whitlock, R.H. and Benson, C.E., 1988. Equine ehrlichial colitis: Effect of oxytetracycline treatment during the incubation period of Ehrlichia risticii infection in ponies. J. Am. Vet. Med. Assoc., 192: 343-345. Rikihisa, Y. and Perry, B.D., 1985. Causative ehrlichial organisms in Potomac horse fever. Infect. Immun., 49: 513-517. Rikihisa, Y. and Jiang, B.M., 1988. In vitro susceptibility of Ehrlichia risticii to eight antibiotics. Antimicrob. Agents Chemother., 32: 986-991. Rikihisa, Y., Johnson, G.C. and Burger, C.J., 1987. Reduced immune responsiveness and lymphoid depletion in mice infected with Ehrlichia risticii. Infect. Immun., 55: 2215-2222. Shaw, D.H. and Rubin, S.I., 1986. Pharmacologic activity of doxycycline. J. Am. Vet. Med. Assoc., 189: 808-810. Shirai, A., Catanzaro, P.J., Eisenberg Jr., G.H.G. and Osterman, J.V., 1977. Host defenses in experimental scrub typhus: effect of chloramphenicol. Infect. Immun., 18: 324-329.