Published online by Cambridge University Press: 19 October 2004
To date Paneth cells have not previously been reported to kill Giardia trophozoites and other protozoa in vivo. Here we report the first evidence for in vivo killing of Giardia trophozoites by intestinal Paneth cells. Transmission electron microscopic (TEM) examination of duodenal specimens taken from naturally infected mice revealed that only Giardia trophozoites harbouring peripheral bacterial endosymbionts were destroyed and lysed in the vicinity of the activated Paneth cells. Additionally, intestinal epithelium was more affected by Giardia harbouring bacterial endosymbionts than Giardia with no endosymbionts. Our findings imply that the bacterial endosymbionts within Giardia trophozoites have a role in both host protective and pathological mechanisms, probably through altering the trophozoite antigencity. These observations might shed light on the diversity in infectivity and host specificity of Giardia species.
Paneth cells are epithelial granulocytes located at the base of the crypts of Lieberkühn in the small intestine of many mammalian species (Ouellette & Selsted, 1996). These cells contribute actively to mucosal immunity through secretion of antimicrobial polypeptides including α-defensins (cryptdin) that are released from the intestinal Paneth cells in response to bacteria and bacterial antigens, but not to protozoa. These antimicrobial polypeptides kill a wide range of organisms including bacteria, viruses, fungi and tumour cells (Martin, Ganz & Lehrer, 1995; Ayabe et al. 2000).
Some in vitro studies have reported that Giardia trophozoites were destroyed and lysed by defensins and other antimicrobial polypeptides derived from the intestinal Paneth cells (Aley et al. 1994). However, no evidence has been reported so far for the in vivo killing of Giardia trophozoites when these cells are stimulated.
Giardia trophozoites, in common with some other protozoan parasites, demonstrate bacterial endosymbiosis (de-Souza & Motta, 1999). The significance of the bacterial endosymbionts within Giardia is not well defined. It is possible that they could alter trophozoite pathogencity, metabolism, range of infectivity and host specificity, as they do in other protozoa (Nemanic et al. 1979).
In this report the ultrastructural interaction between Giardia trophozoites harbouring bacterial endosymbionts and mouse intestinal Paneth cells is described.
Light and transmission electron microscopic (TEM) examinations were conducted for examination of the duodenal specimens taken from Giardia infected mice. Twenty-five laboratory reared albino mice (Tuck ordinary strain), naturally infected with Giardia muris as detected by screening of their faecal suspension (Roberts-Thomson, Stevens & Mahmmoud, 1976), were chosen. It was found that 10 of these 25 mice were heavily infected with G. muris trophozoites. These heavily infected mice were sacrificed.
Mice were dissected and duodenal segments were removed. Segments were fixed in alcoholic Bouin with 10% buffered neutral formalin for subsequent light microscopic examination. Similar segments were fixed in 2·5% glutaraldehyde for TEM examination.
Specimens prepared for TEM were cut into 2–3 mm2 sections and immediately fixed in 2·5% glutaraldehyde with 0·1 M sodium cacodylate buffer (pH 7·2) at 4 °C for 2–3 h. Specimens were post-fixed in 1% osmium tetroxide, dehydrated in ascending grades of ethyl alcohol, and embedded in Spurr's resin. Sections (0·5 μm) were stained with toluidine blue. Ultrathin sections were stained with uranyl acetate and lead citrate and examined in a Jeol 1200 EX TEM at 80 KV (Glauert, 1980).
TEM examination of the specimens taken from naturally infected mice demonstrated many Giardia trophozoites on the brush border, at the base of the crypts as well as among the intestinal microvilli. The internal fine structure of these trophozoites was clearly demonstrated (Fig. 1A, B). Severe blunting and atrophy of microvilli with occasional loss of basic epithelial morphology and distorted Goblet cells were observed in some of the examined duodenal sections (Fig. 1C, D).
In these duodenal specimens bacterial endosymbionts within Giardia trophozoites were demonstrated. Bacteria were found in close contact with the outer surface of some Giardia trophozoites. In some others, they were inundating the trophozoite cell membrane. Electron-lucent spaces were detected surrounding the bacterial endosymbionts and bounded with aggregation of cytoplasmic granules (Fig. 2A, B). No evidence was found for digestion of these bacterial endosymbionts by Giardia trophozoites. There were, however, many electron-lucent vacuoles forming a track directed towards the periphery of the trophozoite. Endosymbionts were concentrated peripherally in direct contact with the intestinal microvilli, and were uncommon in the nuclear area (Fig. 2C).
No gross morphological changes were detected in the Giardia trophozoites as the result of these bacterial endosymbionts. However, Giardia trophozoites were found completely lysed at the base of the duodenal crypts and close to the Paneth cells (Fig. 3A–D).
Paneth cells were found to be highly active with many secretory granules inside, especially those in contact with Giardia trophozoites harbouring peripheral endosymbionts. On the other hand Paneth and Goblet cells associated with Giardia trophozoites and having no endosymbionts were found inactive and the trophozoites were demonstrated to be completely intact and discrete. Mucosal lining in close contact with such trophozoites was observed to be almost intact compared to the severely damaged mucosa observed with trophozoites containing endosymbionts (Figs 1D and 3E).
The precise nature of the interaction between Giardia and its host remains conjectural because of the paucity of published studies. The present ultrastructural report describes an interaction between Giardia trophozoites and intestinal Paneth cells in the presence of bacterial endosymbionts.
Although intestinal Paneth cells secrete antimicrobial peptides that affect many intestinal pathogens, Giardia as many other protozoan parasites have not previously been reported to be killed by Paneth cells or by their antimicrobial secretions (Jones & Bevins, 1992; Selsted et al. 1992; Aley et al. 1994).
In the present report we observed lysis of the cell membrane of Giardia trophozoites when bacterial endosymbionts came in close contact with Paneth cells. Intestinal Paneth cells associated with such trophozoites were found highly activated and with many internal secretory granules.
We could not detect lysis of any Giardia trophozoites free from bacterial endosymbionts and we therefore speculate that the bacterial endosymbionts stimulate Paneth cells. Giardia trophozoites did not stimulate Paneth cells when these cells were exposed to trophozoites in vitro (Ayabe et al. 2000). However, the authors did not exclude the possibility that these trophozoites and other eukaryotes may induce Paneth cell secretion by an unidentified mechanism in vivo.
The reported activation and lytic effect of Paneth cells were seen in all duodenal specimens examined with trophozoites harbouring bacterial endosymbionts. This lead to our proposal that bacterial endosymbionts activate Paneth cells, resulting in release of their lytic peptides mainly defensins. It is likely that defensins are responsible for Giardia lysis since they account for 70% of the released bactericidal activity of Paneth cells (Ayabe et al. 2000). In an earlier study, α-defensins demonstrated a remarkable lysis and killing of Giardia trophozoites in vitro (Aley et al. 1994).
The pathological changes observed in this study as the result of Giardia trophozoites indicated that intestinal epithelium was more affected by Giardia with bacterial endosymbionts than those without endosymbionts suggesting that bacterial endosymbionts might alter the trophozoite pathogenicity.
The reported variation in the pathogenicity and the stimulation of the Paneth cells to react actively with the trophozoites could be due to variation in the trophozoite surface antigens and consequently the immunopathological process (Meng, Hetsko & Gillin, 1993). Such variation in the Giardia surface antigens as the result of bacterial endosymbionts has been proposed earlier (Meyer, 1990).
Bacterial flora and other organisms that live in the intestine appear to provide the several necessary services for healthy development and the suggested functions of the bacterial endosymbionts described here could explain the significance of endosymbiosis in Giardia and other similar protozoa.
The authors would like to thank Dr Samir A. Zaki, Associate Professor of histology, Faculty of Medicine, Zagazig University, Egypt, for his kind assistance in this work.