The pathogenesis of and susceptibility to malabsorption syndrome in broilers is associated with heterophil influx into the intestinal mucosa and epithelial apoptosis

Abstract

Malabsorption syndrome (MAS) in broilers is characterized by enteritis and reduced body weight gain. The pathogenesis of the intestinal lesions and the reasons for susceptibility differences between broiler lines are not clear. We studied the development of enteric lesions, epithelial apoptosis, and cell proliferation in relation to susceptibility. One-day-old chickens from two broiler lines were orally inoculated with intestinal homogenate derived from MAS-affected chickens. Vacuolar degeneration and apoptosis of the villous epithelium and infiltration of heterophils into the lamina propria occurred from day 1 post-inoculation. Following heterophil accumulation, at day 4 to 6 post-inoculation, there was severe apoptosis of the crypt epithelium and villous atrophy. The susceptible broilers had a significantly greater influx of heterophils and, subsequently, severe epithelial apoptosis and cystic damage to the crypts. There appeared to be a causal relationship between heterophil influx and the onset of apoptosis. Coincident with the epithelial apoptosis, MAS-affected chickens had crypt hyperproliferation and faster epithelial turnover. Heterophil infiltration and epithelial apoptosis appear to be critical in the pathogenesis of MAS. Heterophil recruitment may be a major factor in differences in susceptibility to MAS.

Pathogénie du syndrome de malabsorption et sensibilité des poulets de chair à ce syndrome qui est associé à un influx d'hétérophiles dans la muqueuse intestinale et à une apoptose épithéliale

Le syndrome de malabsorption (MAS), chez le poulet de chair, est caractérisé par une entérite et une diminution du gain de poids. La pathogénie des lésions intestinales et les raisons des différences de sensibilité des lignées de poulets de chair ne sont pas claires. Nous avons étudié le développement des lésions entéritiques, l'apoptose épithéliale et la prolifération cellulaire en relation avec la sensibilité. Des poulets de chair, âgés d'un jour appartenant à deux lignées, ont été inoculés par voie orale avec un broyat d'intestins de poulets affectés par le MAS. Un jour après inoculation, il a été observé une dégénérescence vacuolaire et une apoptose de l'épithélium villeux ainsi qu'une infiltration d'hétérophiles au niveau de la lamina propria. Quatre à six jours après l'inoculation et la phase d'accumulation des hétérophiles, il y a eu une sévère apoptose au niveau des cryptes de l'épithélium et une atrophie villeuse. Les poulets de chair sensibles ont présenté un afflux d'hétérophiles significativement supérieur et par la suite, une apoptose épithéliale sévère et des lésions cystiques des cryptes. Il est apparu une relation causale entre l'afflux d'hétérophiles et le début de l'apoptose. De façon identique à l'apoptose épithéliale, le MAS a affecté les poulets qui avaient une hyperprolifération de cryptes et un turnover de l'épithélium plus rapide. L'infiltration des hétérophiles et l'apoptose épithéliale apparaissent être cruciales dans la pathogénie du MAS. Le recrutement des hétérophiles peut être un facteur majeur dans les différences de sensibilité au MAS.

Die Pathogenese des Malabsortionssyndroms und die Empfänglichkeit von Broilern ist mit dem Einstrom von Heterophilen in die intestinale Mukosa und mit epithelialer Apoptose verbunden

Das Malabsorptionssymdrom (MAS) in Broilern ist durch Enteritis und reduzierte Gewichtszunahme gekennzeichnet. Die Pathogenese der intestinalen Läsionen und die Gründe für Unterschiede in der Empfänglichkeit verschiedener Broilerlinien sind nicht klar. Wir haben die Entwicklung der Darmläsionen, der Apoptose von Epithelzellen und der Zellproliferation in Relation zur Empfänglichkeit untersucht. Eintagsküken von zwei Broilerlinien wurden oral mit Homogenat von Darminhalten, die aus an MAS erkrankten Hühnern stammten, inokuliert. Vakuoläre Degeneration, Apoptose des Zottenepithels und Infiltration von Heterophilen in die Lamina propria trat ab dem ersten Tag post inoculationem auf. Im Anschluss an die Akkumulation von Heterophilen vier bis sechs Tage post inoculationem wurde eine hochgradige Apoptose des Kryptenepithels und eine Athrophie der Zotten beobachtet. Die empfänglichen Broiler wiesen einen signifikant stärkeren Einstrom von Heterophilen und nachfolgend hochgradige Apoptose des Epithels sowie Zerstörung der Zysten in den Krypten auf. Es schien ein kausaler Zusammenhang zwischen dem Influx der Heterophilen und dem Beginn der Apoptose zu bestehen. Gleichzeitig mit der epithelialen Apoptose zeigten die MAS-betroffenen Tiere Hyperproliferation der Krypten und einen schnelleren Epithelumsatz. Die Infiltration von Heterophilen und die Apoptose des Epithels scheinen kritische Punkte in der Pathogenese von MAS zu sein, wobei die Verfügbarkeit von Heterophilen der Hauptfaktor hinsichtlich der Unterschiede in der Empfänglichkeit für MAS sein könnte.

La patogenia y susceptibilidad al síndrome de malabsorción en pollos de engorde se asocia con la entrada de heterófilos a la mucosa intestinal y con la apoptosis de células epiteliales

El síndrome de malabsorción (MAS) en pollos de engorde se caracteriza por enteritis y reducción de la ganancia de peso. La patogénesis de las lesiones intestinales y las razones para las diferencias en susceptibilidad entre las diferentes líneas de pollos de engorde no son claras. Se estudió el desarrollo de las lesiones entéricas, apoptosis epitelial, y proliferación celular en relación a la susceptibilidad. Pollos de un día de edad de dos líneas distintas fueron inoculados vía oral con homogenizado intestinal de pollos afectados con MAS. Se observó degeneración vacuolar, apoptosis del epitelio de las vellosidades e infiltración con heterófilos en la lámina propia desde el día 1 post-inoculación. Tras la acumulación de heterófilos, de los 4 a los 6 días post-inoculación, se observó marcada apoptosis del epitelio de las criptas y atrofia de vellosidades. Los pollos de engorde susceptibles presentaron una entrada significativamente mayor de heterófilos y, consecuentemente, mayor apoptosis epitelial y daño quístico a las criptas. Parece que existe una relación causal entre la entrada de heterófilos y el inicio de la apoptosis. En coincidencia con la apoptosis epitelial, los pollos afectados de MAS presentaron hiperproliferación de criptas y un recambio epitelial más rápido. La infiltración con heterófilos y la apoptosis epitelial parecen ser críticas en la patogenia del MAS. El reclutamiento de heterófilos podría ser un factor importante en las diferencia en susceptibilidad al MAS.

Introduction

Malabsorption syndrome (MAS), also known as infectious stunting syndrome, is a gastrointestinal disease affecting young broilers, and causes growth reduction, stunting, and enteritis (Kowenhoven et al., 1978; Bracewell & Randall, 1984). The aetiology is associated with different enteric viruses, including reovirus, rotavirus, enterovirus, parvovirus, calicivirus, and others (Decaesstecker et al., 1986; Goodwin et al., 1993). These viruses are commonly isolated from MAS-affected chickens, but a single agent has not been established to be the cause. Experimentally, MAS can be induced by oral inoculation of 1-day-old chickens with crude intestinal homogenate derived from MAS-affected chickens and, by using the same crude homogenate from MAS-affected chickens, the MAS syndrome can be reproduced repeatedly (Songserm et al., 2002).

Affected chickens develop enteric lesions similar to those of field cases, including dilation of the crypts of Lieberkühn (crypt) and atrophy of the villi (Reece & Frazier, 1990). We previously showed that there is a genetically linked difference in susceptibility to MAS in broiler lines (Zekarias et al., 2002). However, the basis for the differences in susceptibility and the pathogenesis of the intestinal mucosal lesions is not clear.

Reduction in the absorptive surface area and nutrient malassimilation can be caused by atrophy of the mucosa due to increased epithelial apoptosis or inhibition of cell proliferation. Normally, the gut mucosa is maintained by regular renewal of the surface epithelium by proliferation of stem cells located at the base of crypts (Mayhew et al., 1999). These progenitor cells proliferate and differentiate while migrating to the tip of the villus, where they undergo physiological apoptosis (Pritchard & Watson, 1996). While apoptosis is a normal process in intestine mucosal epithelium cell turnover, it also plays an important role in the pathogenesis of several gastrointestinal diseases, such as those caused by rotavirus infection and in ulcerative colitis (Iwamoto et al., 1996; Shirin & Moss, 1998; Boshuizen et al., 2003). A disturbance in stem cell proliferation, apoptosis or epithelial cell turnover is common to the pathogenesis of most gastrointestinal tract infections (Guy-Grand et al., 1998). A similar mechanism could play a role in the pathogenesis of the mucosal lesions in MAS. Here we studied the early pathogenesis of the mucosal lesions in experimentally MAS-affected chickens by comparing the leukocyte response in the intestinal mucosa, epithelial apoptosis and proliferation in susceptible and resistant broiler lines.

Materials and Methods

Chickens

Fertile eggs of two commercial broiler lines were obtained from a commercial breeder and hatched at our institute. Line S, a White Plymouth Rock line, is more susceptible to MAS and line R, a Cornish line, is “resistant” (less susceptible) to MAS (Zekarias et al., 2002).

Experimental procedures

Inoculation

On day 1 after hatching, 40 chicks from each line were orally inoculated with 0.5 ml intestinal homogenate from MAS-affected chickens (MAS homogenate) (Songserm et al., 2002). In the control group, chicks were inoculated with saline.

Housing and feed

All birds were kept on litter. MAS homogenate-inoculated and saline-inoculated groups were kept in separate houses and each line was kept in a separate pen. Standard broiler feed and water were provided ad libitum.

Sampling

Five chickens from each group were killed at 8 h post-inoculation (p.i.), and on days 1, 2, 3, 4, 5, 7, 9, 11, 15 and 21. The chickens were individually weighed, before cervical dislocation. Gender differences were not considered. Autopsies were performed and about 1-cm lengths of the duodenum, mid-jejunum and ileum were collected in buffered 4% formaldehyde. The intestinal sections were cut open for better fixation and were fixed for a maximum of 24 h before processing for paraffin embedding. Serial sections were cut for histopathology and imunohistochemistry.

Histopathology

Intestinal tissue sections were stained with haematoxylin and eosin, and were examined for inflammatory and morphological changes in the crypts and villi. The infiltration of heterophils and lymphocytes, crypt degeneration and cystic dilation of crypts, and atrophy of villi were specifically examined.

The lesions were graded semi-quantitatively using a scale of 0 to 5. For heterophil infiltration into the lamina propria of the crypts and villi, sections were examined under the 20× objective and the categories were: 0 = no heterophils, 1 = < 5 heterophils seen in some fields, 2 = < 5 heterophils seen in all fields examined, 3 = 5 to 10 heterophils in some fields, 4 = 10 to 20 heterophils in several fields, and 5 = > 20 heterophils in all parts of the section. For scoring of crypt degeneration, several fields per transverse sections were examined under the 10× objective and the categories were: 0 = normal, 1 = mild degeneration of the crypt epithelium involving a couple of crypts in some fields, 2 = mild degeneration of the crypt epithelium in a large number of crypts in several fields, 3 = moderate numbers of crypts showing epithelial degeneration in several fields, 4 = severe degeneration of the crypt epithelium in several fields, and 5 = severe degeneration of crypts in all fields. For cystic deformation of crypts, the sections were examined under 10×objective and the categories were: 0 = normal, 1 = mild cystic dilation of crypts in a few fields, 2 = a few cystic crypts in all fields, 3 = many cystic crypts in several fields, 4 = severe cystic dilation of crypts in several fields, and 5 = cystic crypts in all fields

Morphological measurement

Crypt depth and villus height were measured on sections of the mid-jejunum using a graded eyepiece. The fields measured were selected based on the alignment of the villi and crypts. Ten villi and crypts per section were measured.

Terminal deoxynucleotide transferase-mediated nick end labelling assay

Apoptosis was studied using a terminal deoxynucleotide transferase-mediated nick end labelling (TUNEL) assay kit (Promega, Madison, Wisconsin, USA) to detect DNA fragmentation by incorporation of biotin-labelled dUTP (2′-deoxyuridine 5′-triphosphate). The incorporation of biotinylated dUTP was visualized using horseradish-peroxidase-conjugated streptavidin (Stadelmann & Lassmann, 2000). Briefly, paraffin-embedded sections from the mid-jejunum mounted on poly-l-lysine-coated slides were dewaxed in xylene, hydrated and treated with proteinase K (10 µg/ml) for 30 min at room temperature and the endogenous peroxidase activity quenched by immersing the slide in 3% H₂O₂ for 10 min. The slides were then incubated for 1 h at 37°C with the biotin-conjugated dUTP and terminal deoxynucleotidyl transferase. Slides were then incubated with streptavidin-peroxidase (DAKO, Glostrup, Denmark) for 30 min and the peroxidase activity was visualized by incubation in 0.05% diaminobenzidine in 0.1 M Tris–HCl (pH 7.5) containing 0.01% H₂O₂ (Sigma, St Louis, Missouri, USA). Slides were examined under the light microscope and apoptosis was detected as dark-brown nuclear staining.

The number of TUNEL-positive nuclei in crypts or villi were examined under the 10× objective and scored using a semi-quantitative scale. The scores were: 0 = none, 1 = < 3 positive cells in a few fields, 2 = a few (<3) positive cells in each of several fields, 3 = moderate numbers (<10) of positive cells in several fields, 4 = a large number (>10) of positive cells dispersed in several fields, and 5 = a large number (>10) of positive cells in all fields.

Proliferating cell nuclear antigen

Proliferating cell nuclear antigen (PCNA) is a cell proliferation marker expressed by cells throughout the S to M phases of the cell cycle (Yu et al., 1992). PCNA was detected by immunohistochemistry using a monoclonal antibody against recombinant rat PCNA (PC-10; DAKO). Briefly, paraffin-embedded tissue sections from the mid-jejunum were dewaxed in xylene, washed and incubated with the monoclonal antibody PC-10 at a 1:200 dilution overnight at 4°C. Endogenous peroxidase was inhibited by incubation in with 1% H₂O₂ in 0.1 M Tris–HCl. The sections were then incubated with the biotinylated secondary antibody (rabbit anti rat antibody) followed by peroxidase-conjugated streptavidin (DAKO) for 30 min each. Peroxidase activity was detected using 0.05% diaminobenzidine in 0.1 M Tris–HCl (pH 7.5) containing 0.01% H₂O₂ (Sigma). The slides were examined using light microscopy to detect the dark-brown nuclear staining of PCNA-positive cells.

The number of PCNA-positive cells in the crypts or villi were scored using a semi-quantitative scale. The scores were: 0 = none, 1= a few stained (<10) cells in a few fields, 2 = 10 to 50 stained cells in several fields, 3 = 10 to 50 stained cells uniformly distributed in several fields, 4 = a large number (>50) of stained cells with a variable distribution in the section present in several fields, and 5 = a large number of positive cells present in all microscopic fields.

Statistical analysis

The lesion score data were analysed with a threshold model for ordered categorical data (McCullagh, 1980). An underlying (unobserved) variable (y) was assumed with threshold values (t ₁, t ₂, …) such that score 1 was observed when y<t ₁, score 2 was observed when t ₁<y<t ₂, and so forth. The latent variable y was assumed to follow a logistic distribution. When the mean of this distribution was high there would be an increased probability of high scores, and when the mean of y was low there would be an increased probability of low scores. The mean was assumed to depend on linear and quadratic components for time and main effects for lines and treatments and interaction between lines and treatments. So, for example, when a treatment effect was large, the mean of the latent variable y would move to the right, the distribution of y will be located near to the higher threshold values and there will be an increased probability of higher scores compared with treatments with a lower effect. When there are only two possible outcomes, the model is equivalent to ordinary logistic regression. Significance tests were based on the likelihood ratio test employing an appropriate chi-squared distribution. The analysis was performed with the statistical package GenStat (2000). The body weight, villus height, and crypt depth were analysed with analysis of variance. Means were compared pairwise with Fisher's least significant difference method, employing the statistical package SPSS. In all tests, P<0.05 was considered statistically significant.

Results

Bodyweight

All the chickens inoculated with the intestinal homogenate derived from MAS-affected chickens had reduced weight gain compared with the controls. The MAS-affected susceptible chickens had a significantly lower mean body weight compared with the body weight of the MAS-affected resistant chickens, relative to uninfected controls (line R) (Table 1). In line S, the greatest difference between control and MAS-affected chickens (54% difference in body weight relative to the control chickens) occurred at day 11 p.i. In the resistant line, the body weight was 24% below the weights of the controls at day 7 p.i (Table 1).

Table 1.  Mean body weights of two broiler lines

	1 day p.i.		3 day p.i.		7 day p.i.		11 day p.i.		15 day p.i.
	Control chickens	MAS-affected chickens	Control chickens	MAS-affected chickens	Control chickens	MAS-affected chickens	Control chickens	MAS-affected chickens	Control chickens	MAS-affected chickens
Line S	53.5±5.9	44.8^C±3.2	69.8^B±3.2	58.2^C±3.8	128.6^B/C±9.3	66.4^C±6.0	219.8^B/C±18.4	102.0^C±10.7	343.6^B/C±17.1	212.8^C±30.3
Line R	53.6±3.7	56.8±1.9	75.4±7.1	68.2±7.6	151.1^B±7.7	114.2±25,7	288.0^B±28.8	242.8±40.8	415.0±30.2	360.8±30.1
Line S, MAS-susceptible broiler line; line R, MAS-“resistant” broiler line (n=5).
Control chickens were inoculated with saline; MAS affected chickens were inoculated with intestinal homogenate derived from MAS-affected chickens.
^BSignificant difference on the same day between control and MAS-homogenate-inoculated chickens in either line.
^CSignificant difference on the same day between line S versus line R (within same column).

Histopathology

Enteritis, cystic distension of crypts, and villous atrophy were observed in both lines of chickens inoculated with the intestinal homogenate from MAS-affected chickens. The lesions were more severe in line S than in line R. Lesions were more severe in the jejunum than in the proximal or distal intestine. There were no lesions at any time point in the intestinal mucosae of the control chickens. In chickens inoculated with the intestinal homogenate derived from MAS-affected chickens there was hyperaemia in the lamina propria and in the villus corum at 8 h p.i. At day 1 p.i. there was vacuolar degeneration of the epithelial cells of the villi and erosions of the tips of the villi. An influx of heterophils into the lamina propria of the crypts and villi was first seen at 8 h p.i. in line S (Figure 1). Heterophil infiltration was most pronounced at day 3 p.i. The number of heterophils in the mucosa was significantly higher in line S. Heterophil infiltration diminished from day 5 p.i. Degeneration of the crypt epithelium and dilation of the crypt lumens was detected from day 3 p.i. onwards. Analysis across the collected time points showed that line S developed more severe crypt degeneration than line R (P < 0.05). There was remarkable cystic distension of crypts at day 5 and 7 p.i., as well as severe villous atrophy. No significant differences were found between the MAS-affected lines in crypt depth or villus height. Analysis across the collected time points showed that the susceptible chickens developed more severe crypt degeneration and cystic crypts than the resistant chickens (P<0. 05) (Figure 1). The incidence of cystic crypts decreased from day 7 p.i. in both lines. From day 15 p.i. there was crypt regeneration. The mucosal recovery was more protracted in line S than line R.

Figure 1.  Medians and ranges of lesions scores in birds affected with MAS. The difference in severity of crypt degeneration and of cystic crypts between the two lines differed significantly over the collective time points (P < 0.05).

Apoptosis

In the control chickens, only a few TUNEL-positive cells were detected at the tip of the villi at the extrusion zone (Figure 2a). In birds inoculated with the intestinal homogenate derived from MAS-affected birds many apoptotic cells were detected, all along the villous surfaces starting at day 2 p.i. (Figure 2b). From day 3 p.i. onwards, several TUNEL-positive cells were detected in the crypts and along the villus–crypt axis. The most remarkable increase in TUNEL reactivity in the crypt epithelium was between day 4 and 6 p.i. (Figure 2c). Analysis across the collected time points showed that line S had significantly more TUNEL-positive cells in the villi (P<0.05) and in the crypts (P<0.05) than line R (Figure 3). TUNEL-positive cells were also abundant in the lamina propria of the crypts and villi from day 3 to 5 p.i., apparently corresponding to apoptotic heterophils (Figure 2d). There was a reduction in the number of TUNEL-positive cells when the crypts were cystic, and only a few TUNEL-positive cells were detected in the crypts after day 9 p.i.

Figure 2.  Apoptosis detection by TUNEL assay. Sections of the mid-jejunum of broiler chickens (line S) orally inoculated at 1 day of age with an intestinal homogenate derived from MAS-affected chickens (MAS) or with saline (controls). Dark-brown staining indicates TUNEL-positive cells (counter-stained with haematoxylin). (2a) Day 4 p.i., control chicken. A few TUNEL-positive cells at the tip of the villi (arrow). In control chickens at all ages, apoptotic cells (TUNEL-positive) were limited to the tip of the villi (magnification 10× objective). (2b) Day 2 p.i., MAS chicken. Several TUNEL-positive cells along the apical surfaces of the villi (magnification 10× objective). (2c) Day 4 p.i., MAS chicken. Abundant TUNEL-positive cells in the crypt epithelium (magnification 20× objective). (2d) Day 4 p.i., MAS chicken. TUNEL-positive cells in the lamina propria, probably corresponding to heterophils (magnification 20× objective).

Figure 3.  Medians and ranges of TUNEL-positive and PCNA-positive scores in birds affected with MAS. The differences in incidence of apoptosis over the collective time points in the crypts and villi (TUNEL-positive cells) significantly differed between lines R and S (P < 0.05).

Cell proliferation

In control chickens, PCNA immunoreactivity was detected in the crypts and at the base of the villi (Figure 4a). In contrast, in chickens inoculated with the MAS homogenate the epithelial cells of the villi also expressed PCNA (Figure 4b). This was seen from day 1 p.i., but at days 5, 7 and 9 p.i. there was a decrease in PCNA-positive cells in the crypts, associated with depletion of proliferating cells in the crypts and a low cell density. Strong PCNA reactivity was detected from day 11 p.i., corresponding to crypt regeneration (Figure 4c). This biphasic PCNA reactivity in the crypt during the initial period after inoculation, and later during crypt regeneration, seemed to correspond with diminution and later a rise in the number of proliferating cells. The intensity of PCNA staining in the crypts of control chickens increased with age and was related to the increase in crypt numbers and diameter. More crypt PCNA staining was seen in the crypts of chickens from line S affected with MAS than in those from line R affected with MAS. There was only a slight difference in PCNA staining of the villi between the two broiler lines when affected with MAS (Figure 3).

Figure 4.  PCNA staining of mid-jejunal sections from broiler chickens (line S) that were orally inoculated at 1 day of age with an intestinal homogenate derived from MAS-affected chickens (MAS) or saline (controls). Dark-brown staining indicates cell proliferation (counter-stained with haematoxylin). (4a) Day 3 p.i., control chicken. PCNA staining in the crypts indicating cell proliferation and maturation of cells along the villus as they migrate to the tip (magnification 10× objective). (4b) Day 3 p.i., MAS chicken. PCNA staining of villous epithelial cells indicating faster cell migration along the villus wall (accelerated transit). No counter staining (magnification 10× objective). (4c) Day 11 p.i., MAS chicken. PCNA staining related to crypt regeneration and development of new crypts (magnification 20× objective).

In summary, the intestinal mucosal lesions in MAS-affected chickens started with vacuolar degeneration and apoptosis of the villous epithelial cells, accompanied by crypt hyperplasia and heterophil infiltration. Acute heterophilic inflammation was a major feature of the early phases of MAS prior to apoptosis of the epithelium. Subsequently, there was cystic dilation of crypts and villous atrophy. The difference in susceptibility to MAS between susceptible (line S) and resistant (line R) chickens was correlated with differences in the amount of heterophil recruitment, epithelial apoptosis and the severity of the intestinal lesions.

Discussion

This study showed that heterophil infiltration into the intestinal mucosa and epithelial apoptosis played an early role in the pathogenesis of cystic enteritis and villous atrophy in MAS. Inoculation with intestinal homogenate from MAS-affected chickens caused acute vacuolar degeneration of villous epithelial cells and heterophil infiltration. The intestinal heterophil infiltration could be triggered by pro-inflammatory cytokines. These cytokines could be produced by the affected villous epithelium and leukocytes in the lamina propia (Madara et al., 1991; Jung et al., 1995). Interestingly, the heterophil infiltration was significantly more pronounced in the susceptible broilers than the resistant broilers. We have reported previously that at 1 day of age the chickens of line S have a higher proportion of heterophils in peripheral blood (Zekarias et al., 2002). This may be related to the greater heterophil infiltration into the intestinal mucosa in the MAS-affected chickens of line S compared with chickens of line R.

Heterophil infiltration can be beneficial in host defence (Kogut et al., 1994). However, tissue damage often results from the accumulation of heterophils in tissues (Madara et al., 1991; Harmon, 1998). In MAS-affected chickens the heterophil infiltration slightly preceded apoptosis of epithelial cells, followed by cystic distension of the crypts and atrophy of the villi. Cystic deformation of the crypts of Leiberkhun is commonly found in MAS (Reece, 2001). The apoptosis in the crypts and villi and the formation of cystic crypts were more severe in the susceptible chickens, which also had greater heterophil infiltration into the mucosa.

Infiltration of the intestinal mucosa by polymorphonuclear leukocytes (PMNL), which is associated with epithelial apoptosis, occurs commonly in gastrointestinal infections (Iwamoto et al., 1996; Shirin & Moss, 1998). The exact role of PMNL/heterophils in the pathogenesis of epithelial apoptosis is not clear. An in vitro study indicated that PMNL transmigration induced apoptosis in colonic epithelial cells (Le'Negrate et al., 2000). Heterophils (PMNL) might produce tumour necrosis factor-α and nitric oxide, followed by induction of apoptosis (Buttke & Sandstorm, 1994; Le'Negrate et al., 2000; Choi & Chae, 2002). Release of soluble Fas-ligand by PMNL can trigger apoptosis in lung epithelial cells (Serrao et al., 2001). Another explanation for crypt apoptosis is that specific components in the MAS inoculum might directly trigger epithelial apoptosis. Concurrent with the epithelial apoptosis, there is hyperproliferation of crypt epithelium in MAS-affected chickens. The crypt hyperproliferation could be a reaction to the epithelial apoptosis or could be stimulated by inflammatory cytokines (Stappenbeck et al., 1998).

In conclusion, heterophil infiltration and subsequent epithelial apoptosis is probably important in the pathogenesis of MAS. Heterophil recruitment into the mucosa is likely to be critical in the onset of crypt apoptosis. Susceptible chickens have greater heterophil infiltration into the mucosa and more severe apoptosis and cystic crypt lesions compared with resistant chickens. The results suggest that heterophil recruitment during inflammation is likely to be a major factor in differences in susceptibility to MAS. Heterophil recruitment mechanisms and functional comparisons between the broiler lines and their role in epithelial apoptosis need to be further elucidated.

Translations of the abstract in French, Germany and Spanish are available on the Avian Pathology website.

The authors thank Nutreco (Boxmeer, The Netherlands) for the supply of the chicken lines, Theo Niewold and Mark Mueller for critical reading of the manuscript, and Bas Engel for his help in the statistical analysis.