Gut: first published as 10.1136/gut.27.1.73 on 1 January 1986. Downloaded from

Gut, 1986, 27, 73-77

Direct and dependent cell mediated cytotoxicity against Giardia lamblia by splenic and- intestinal lymphoid cells in mice

S S KANWAR, N K GANGULY, B N S WALIA, AND R C MAHAJAN From the Departments ofParasitology and Paediatrics, Postgraduate Institute ofMedical Education and Research, Chandigarh, India

SUMMARY Direct cytotoxicity and antibody dependent cell mediated cytotoxicity against Giardia lamblia trophozoites exhibited by splenic, intraepithelial and lamina propria populations isolated from G lamblia infected mice were studied. Different patterns of cytotoxicity were found. Intraepithelial showed a direct cytotoxic activity of 20*6±5-6% before infection. It was significantly higher on the 20th (p<0.01) and 30th (p<005) day postinfection. Lamina propria lymphocytes showed a significantly augmented level of both direct cytotoxicity and antibody dependent cell mediated cytotoxicity on the 20th and 30th postinfection days. Direct cytotoxicity by splenic lymphocytes remained unchanged during infection but antibody dependent cell mediated cytotoxicity was significantly increased.

The host response to G lamblia involves the immune weighing 10-12 g were used in this study. G lamblia system. Previous exposure to this infection is known cysts were obtained from the stool of a patient and a http://gut.bmj.com/ to increase resistance to a second challenge in both fixed inoculum of 10 000 cysts/0-2 ml was prepared man and animals.1 2 Smith et al3 reported that on a sucrose gradient9 and fed to the animals.1" Five human peripheral blood monocytes/macrophages animals were killed on each of the days 0, 10, 20, are spontaneously cytotoxic for G lamblia tropho- and 30 postinfection. zoites and this cytotoxicity is reduced in chronic giardiasis.4 In another study they reported defective QUANTITATION OF CYST AND TROPHOZOITE spontaneous cytotoxicity but normal antibody de- (a) Cyst count on September 28, 2021 by guest. Protected copyright. pendent cytotoxicity by resident macrophages iso- Two hour stool samples from each animal were lated from C3H/Hej mice.5 collected, homogenised in normal saline and layered Giardia lamblia has been shown to infect mice. on sucrose gradient.9 The results were expressed as Using a mouse model, we have examined the role of number of cysts/g of stool. cytotoxic effector cells in resistance of mice to this infection. Direct cytotoxicity of lymphocytes and (b) Trophozoite count antibody dependent cell mediated cytotoxicity The small intestine was flushed with a fixed amount against G lamblia are likely to be most relevant in of normal saline, and the trophozoites in the mucosae. We have therefore investigated the direct intestinal perfusate counted in a haemocytometer.6 and antibody mediated cytotoxicity of mucosal lymphocytes (intraepithelial lymphocytes and PREPARATION OF EFFECTOR CELLS lamina propria lymphocytes) against G lamblia. (a) Splenic lymphocytes These were removed and harvested according to the Methods method described earlier.6 They were purified by Ficoll-isopaque density gradient. ANIMALS AND MODE OF INFECTION Parasite free, 2-3 weeks old Swiss albino mice, (b) Intestinal lymphocytes Address for correspondence: Professor R C Mahajan, Head, Department of were isolated a modified Parasitology, Postgraduate Institute of Medical Education and Research, Gut lymphocytes by Chandigarh-160012, India. method of Davies and Parrott" and Taglibue et al.12 Received for publication 9 April 1985 Briefly, the small intestine was washed with calcium 73 Gut: first published as 10.1136/gut.27.1.73 on 1 January 1986. Downloaded from

74 Kanwar, Ganguly, Walia, and Mahajan and magnesium free Hank's balanced salt solution. hours at 37°C in a shaking water bath. 0*1 ml of The mesentery, adherent connective tissue and fat labelled trophozoites (5 x 10),5 suspended in incuba- were removed from the intestine and Peyers patches tion medium, were placed in sterile small test tubes. were dissected. The intestine was opened longitu- 0*1 ml of effector cells were also added (effector cell dinally and cut into 1-2 cm long pieces and washed density was adjusted to achieve 25:1, 50:1 and 100:1 thoroughly in calcium and magnesium free Hank's E:T ratios). To antibody dependent cell balanced salt solution. To isolate intraepithelial mediated cytoxicity, 0.1 ml of immune serum lymphocytes, gut pieces were incubated at 37°C for (1:3000 diluted) was added. The effector cells, 15 minutes in calcium and magnesium free Hank's labelled trophozoites and antiserum were mixed and balanced salt solution containing 5 mmol EDTA tubes were centrifuged at 100 g for five minutes. with constant stirring. This process was repeated These were then incubated at 37°C in a 5% CO2 three times. The supernatants were pooled, centri- humidified atmosphere for two hours. After incuba- fuged, resuspended in RPMI-1640, filtered through tion, the tubes were centrifuged and the amount of glass wool column and washed. Most of the lympho- radio-activity in the supernatant was measured in a cytes obtained by this method were from the gamma counter. The maximum release was ob- epithelial layer. " tained by treating the labelled trohozoites with For isolation of lymphocytes from the lamina 10% Triton-xlOO followed by processing as outlined propria, gut pieces were washed with calcium and above. The spontaneous release varied from 10- magnesium free Hank's balanced salt solution and 15% of the maximal release. The percentage specific incubated for 20 minutes in 25 ml RPMI-1640 cytotoxicity was calculated by the following formula: containing 5% heat inactivated fetal calf serum % specific cytotoxicity=experimental release (RPMI-FCS) and 2 mmol L-glutamine with stirring. (cpm)-spontaneous release (cpm) x 100/maximal re- These pieces were then incubated in 15 ml RPMI- lease (cpm)-spontaneous release(cpm). FCS containing 25 U/ml collagenase (Boehringer, W. Germany) for 15 minutes. The whole procedure STATISTICAL ANALYSIS was repeated three times. The supernatants were Student's t test was used to analyse all the data. pooled, filtered through a glass wool column and washed in RPMI-FCS. Results

The lymphocyte rich preparations of intra- http://gut.bmj.com/ pethelial lymphocytes and lamina propria lympho- Figure 1 shows the kinetics of cyst excretion and cytes were loaded on a percoll density gradient trophozoite counts in the intestine during the 30 day (Pharmacia, Sweden) and the lymphocytes reco- period of study. The peak cyst excretion and peak vered from the 1*050/1 085 g/ml density interface. trophozoite count occurred on the 10th postinfec- Their viability was checked by the 0.2% trypan blue tion day. By the 30th day, cyst excretion ceased in dye exclusion test. The yield of intraepithelial all the animals, although a few trophozoites could

lymphocytes obtained was 0-6-2Slx 10.7 still be detected in 30-40% of infected animals in the on September 28, 2021 by guest. Protected copyright. intestinal washings. ANTISERUM AGAINST G lamblia Different effector:target cell (E:T) ratios were Antiserum to G lamblia trophozoites was raised in used (25:1, 50:1 and 100:1) to test for direct rabbits13 and titrated in a cytotoxicity assay. The cytotoxicity and antibody dependent cell mediated serum was used at a dilution of 1:3000 in all cytoxicity. There was a slight but consistent increase experiments. At this concentration, the antiserum in direct cytotoxicity at E:T ratios of 50:1 and 100:1. itself had no giardiacidal activity. Normal rabbit A very low level of direct cytotoxicity (6.34±2.15%) serum was used at the same dilution. Both normal was observed at E:T ratios of 25:1 with splenic and immune sera were heated at 56°C for 30 minutes lymphocytes, while with intraepithelial lymphocytes before use. and lamina propria lymphocytes, the respective values were 19 03±2 08% and 18-30±3-92%. With CYTOTOXICITY ASSAY lamina propria lymphocytes, the antibody depen- Giardi lamblia trohozoites were purified by the dent cell mediated cytoxicity was much higher Feely and Erlandsen'4 method of sticking them to (44.98%), in comparison to the values obtained with polystyrene petri dishes, and washed with RPMI- intraepithelial lymphocytes and splenic lyphocytes 1640 containing 25 mM HEPES, 10% heat inacti- at E:T ratios of 100:1 (Fig. 2). vated fetal calf serum and 50 ,g/ml gentamicin. The Intraepithelial lymphocytes showed a cytotoxic cytotoxicity assay was a modified technique of Smith activity of 20-6±5*63% before infection. On the et al.S Trophozoites were labelled with sodium5' 20th postinfection day, the values were increased to chromate using 70 gCi/5 x 106 parasite/ml for 12 39*82±5*87% (p

Mediated cytotoxicity against Giardia lamblia in mice 75

50. p

0C -0 / / 0 / U U ._ 30 N 0, 0. 0. O- CLi in 20

10 Days after inocculation Fig. 1 Giardia lamblia cyst and tropozoite counts (log,o) on differentpostinfection days. Each point represents the mean ±SD offive animals. * @ 25:1 50:1 100:1 Effector: dent cell mediated cytoxicity for intraepithelial target ratio http://gut.bmj.com/ lymphocytes showed no significant change from the Fig. 2 Direct (continuous line) and antibody dependent preinfection value. In contrast, lamina propria (dotted line) cell mediated cytotoxic activity exhibited by lymphocytes showed a significantly higher direct splenic (L), IEL (@) and LPL(O) at different E:T ratios cytotoxicity and antibody dependent cell mediated against G lamblia trophozoites. Values represent the mean cytoxicity on day 20 and 30 (Table 1). ofthree experiments before infection. The level of direct cytotoxicity exhibited by splenic lymphocytes at E:T ratios of 50:1 before infection was 12.9±5-4%; this remained unchanged Discussion on September 28, 2021 by guest. Protected copyright. after infection (Table 2). In the presence of immune serum, cytotoxic activity was augmented to a signifi- In our mouse model, G lamblia infection persisted cant level (p<001 and p<002 respectively on 20th for 30 days. The selfresolution of the infection and 30th day). appears to be immune mediated. We have already

Table 1 Level of cytotoxicity shown by intraepithelial lymphocytes and lamina propria lymphocytes at E: T ratios of25:1 using radio labelled giardia trophozoites as target cells (Mean±SD) Postinfection days Effector cells 0 10 20 30 Direct 20-6 ±5-63 25-84±6-42 39 82±5 87 32-01±5-35 IEL (<0.01) (<0.05) ADCC 15-11±5-11 19-49±4-69 20-56±4 09 21-9 ±9-0 Direct 15-75±4-85 22-3 ±5i49 31-92±7-23 29-02±7-02 LPL (<0.05) (<0.05) ADCC 14-00±4-83 20-10±6-06 38-82±4-21 33-97±4-42 (<0.01) (<0.01) * p value vs 0 day value. Gut: first published as 10.1136/gut.27.1.73 on 1 January 1986. Downloaded from 76 Kanwar, Ganguly, Walia, and Mahajan

Table 2 Level of cytotoxicity shown by splenic lympho- antibacterial antibody dependent cell mediated cytes at E:T ratios of 50:1 using radio labelled G lamblia cytoxicity when bound to lymphocytes from gut trophozoites as target cells lymphoid tissue. It is possible that specific S-IgA Postinfection days may be more effective in augmenting antibody dependent cell mediated cytoxicity by lamina prop- 0 10 20 30 ria lymphocytes, or even by intraepithelial lympho- cytes, in our system. Direct 12-92±5-4 18-20±6-60 24-98±7-2 23-93±10-42 No increase in direct cytotoxicity occurred with ADCC 21-29±5-07 28 06±7 05 43-72±7 04 37-88± 8-47 the splenic lymphocytes but the antibody dependent (<0.01) (<0.02) cell mediated cytoxicitf was significantly aug- mented. Taglibue et al, in a preliminary study, p values vs 0 day value. reported an increase in splenic NK activity when mice were infected with Nippostrongylus brasiliensis reported the appearance of in these or Giardia muris. These workers did not give details animals during the course of primary and secondary of their observations thus making comparisons infection.15 2 In this study, we have shown that difficult. intraepithelial lymphocytes populations exhibit only We conclude that intraepithelial lymphocytes direct cytotoxicity, where as lamina propria lympho- exhibit direct cytotoxicity, whereas lamina propria cytes exhibit both direct and antibody dependent lymphocytes exhibit both direct and antibody de- cell mediated cytoxicity. The augmented level of pendent cell mediated cytoxicity against G lamblia. cytotoxicity of these two lymphocyte populations In contrast, splenic lymphocytes exhibit only anti- coincides with the decline of parasite numbers in the body dependent cell mediated cytoxicity. These intestine. Giardia lamblia infection in mice has been findings may help in explaining the mechanism of shown to increase the intraepithelial lymphocyte mucosal defence against G lamblia infection. count.16 17 The intraepithelial l1mphocytes popula- tion is dominated by T cells,' and in G lamblia infected mice they respond by blast transformation with G lamblia antigen.17 The presence of cytotoxic References T cells in the intraepithelial lymphocytes and lamina http://gut.bmj.com/ propria lymphocytes population has also been 1 Moore GT, Cross WM, McGuire D, et al. Epidemic shown.19 Owen et a120 showed the attachment of giardiasis at a ski resort. N Engl J Med 1969; 281: lymphocytes to G muris trophozoites in the lumen, 402-7. and thus the higher level of direct cytotoxicity 2 Kanwar SS, Ganguly NK, Mahajan RC, Walia BNS. Acquired resistance to Giardia lamblia infection in exhibited by intraepithelial lymphocytes may be mice. Trop Geog Med 1985; 37: 32-6. because of their continuous presence in close 3 Smith PD, Elson CO, Keister DB, Nash TE. Human vicinity to the parasite. Different patterns of cyto- host response to Giardia lamblia infection spontaneous on September 28, 2021 by guest. Protected copyright. toxicity between the two intestinal lymphocyte killing by mononuclear leucocytes in vitro. J Immunol populations may be explained either by population 1981; 128: 1372-6. heterogeneity or by rigorous selection of homoge- 4 Smith PD, Gillin FD, Spira WM, Nash TE. Chronic nous population at the basement layer of the giardiasis: studies on drug sensitivity, production epithelium (by the failure of antibody dependent and host immune response. Gastroenterology 1982; 83: cell mediated cytoxicity to penetrate the epithe- 797-803. a micro- 5 Smith PD, Keister DB, Wahl SM, Meltzer MS. lium). Alternatively specialised Defective spontaneous but normal antibody dependent environment might produce phenotypic changes. cytotoxicity for an extracellular protozoan parasite, In the human host, Smith et a121 have reported Giardia lamblia by C3H/Hej mouse macrophages. Cell that the peripheral blood granulocytes, but not Immunol 1984; 85: 244-51. lymphocytes, were cytotoxic for G lamblia. The 6 Ganguly NK, Vasudeva V, Anand BS, Chandnani RE, predominant isotype of the antibodies in antibody Mahajan RC. Study of lymphocyte sub-populations in dependent cell mediated cytoxicity was IgG, but Giardia infected mice. Ann Trop Med Parasitol 1981; when IgG depleted serum was used, there was still 75: 347-51. some cytotoxic acitivity. This raises the possibility 7 Vasudeva V, Ganguly NK, Anand BS, Radhakrishna V, Dilawari JB, Mahajan RC. A study of Giardia that other immunoglobulin isotypes may also medi- infection in irradiated and thymectomised mice. J Trop ate antibody dependent cell mediated cytoxicity. In Med Hyg 1982; 82: 119-24. mucosal secretions, secretory IgA is the predomi- 8 Hill DR, Guerrant RL, Pearson RD, Hewlett EL. nant isotype. It has been recently reported by Giardia lamblia infection of sucking mice. J Infect Dis Taglibue et a122 that S-IgA in mice can mediate 1983; 147: 217-21. Gut: first published as 10.1136/gut.27.1.73 on 1 January 1986. Downloaded from Mediated cytotoxicity against Giardia lamblia in mice 77

9 Olveda RK, Andrews JS, Hewlett EL. Murine giar- cell kinetics and protozoal infection in mice. Gas- diasis localization of trophozoites and small bowel troenterology 1978;74: 496-500. histopathology during the course of infection. Am J 17 Kanwar SS, Ganguly NK, Walia BNS, Mahajan RC. Trop Med Hyg 1982; 31: 60-6. Enumeration of small intestinal lymphocyte population 10 Roberts Thomson IC, Stevens DP, Mahmoud AAF, in Giardia lamblia infected mice. J Diar Dis Res 1984; Warren KS. Giardiasis in the mouse: an animal model. 2: 243-8. Gastroenterology 1976; 71: 57-61. 18 Guy-Grand D, Gricellic C, Vassalli P. The mouse gut 11 Davies MDJ, Parrott DMV. Preparation and purifica- T lymphocyte, a novel type of . Nature, origin and tion of lymphocytes from the epithelium and lamina traffic in mice in normal and graft-versus host condi- propria of murine small intestine. Gut 1981; 22: 481-8. tions. J Exp Med 1978; 148: 1661-76. 12 Taglibue A, Befus AD, Clark DA, Beinenstock J. 19 Davies MDJ, Parrott DMV. The early appearance of Characteristics of natural killer cells in the murine cytotoxic T cells in murine gut mucosa. Clin Exp intestinal epithelium and lamina propria. J Exp Med Immunol 1980; 42: 273-79. 1982; 155: 1785-96. 20 Owen RL, Nemanic PC, Stevens DP. Ultrastructural 13 Craft JC, Nelson JD. Diagnosis of giardiasis by counter observation on giardiasis in a murine model 1. Intestin- immunoelectrophoresis of feces. J Infect Dis 1982; 145: al distribution, attachment and relationship to the 499-504. of Giardia muris. Gastroenterology 14 Feely DE, Erlandsen S. Isolation and purification of 1979; 76: 757-69. giardia trophozoites from rat intestine. J Parasitol 1981; 21 Smith PD, Keister DB, Elson CO. Human host 67: 59-64. response to Giardia lamblia II. Antibody-dependent 15 Ganguly NK, Mahajan RC, Radhakrishna V, Dilawari killing in-vitro. Cell Immunol 1983; 82: 308-15. JB, Walia BNS. Haemagglutinating antibodies in 22 Taglibue A, Nencioni L, Villa L, Keren DF, Lowell experimental giardiasis in normal and thymectomised GH, Boraschi D. Antibody dependant cell mediated mice. Ind J Med Res 1981; 74: 527-33. antibacterial activity of intestinal lymphocytes with 16 McDonald TT, Ferguson A. Small intestinal epithelial secretory IgA. Nature 1983; 306: 184-6. http://gut.bmj.com/ on September 28, 2021 by guest. Protected copyright.