Megaoesophagus in Rassf1a-null mice

Abstract

Megaoesophagus, or oesophageal achalasia, is a neuromuscular disorder characterized by an absence of peristalsis and flaccid dilatation of the oesophagus, resulting in the retention of ingesta in the dilated segment. The aetiology and pathogenesis of idiopathic (or primary) megaoesophagus are still poorly understood and very little is known about the genetic causes of megaoesophagus in humans. Attempts to develop animal models of this condition have been largely unsuccessful and although the ICRC/HiCri strain of mice spontaneously develop megaoesophagus, the underlying genetic cause remains unknown. In this report, we show that aged Rassf1a-null mice have an enhanced susceptibility to megaoesophagus compared with wild-type littermates (∼20%vs. ∼2% incidence respectively; P = 0.01). Histological examination of the dilated oesophaguses shows a reduction in the numbers of nerve cells (both ganglia and nerve fibres) in the myenteric plexus of the dilated mid and lower oesophagus that was confirmed by S100 immunohistochemistry. There was also a chronic inflammatory infiltrate and subsequent fibrosis of the myenteric plexus and the muscle layers. These appearances closely mimic the gross and histopathological findings in human cases of megaoesophagus/achalasia, thus demonstrating that this is a representative mouse model of the disease. Thus, we have identified a genetic cause of the development of megaoesophagus/achalasia that could be screened for in patients, and may eventually facilitate the development of therapies that could prevent further progression of the disease once it is diagnosed at an early stage.

Achalasia is a disorder of the oesophagus characterized by decreased lower oesophageal sphincter (LOS) tone, absence of LOS relaxation with swallowing and aperistalsis of the body of the oesophagus. It is a rare disease with an annual incidence of approximately 1/100,000 and a prevalence of 1/10,000. The disease can occur at any age, with a similar incidence in men and women, but is usually diagnosed between 25 and 60 years (Farrokhi & Vaezi 2007). There are two types of achalasia: primary achalasia (PA) or idiopathic disease of unknown aetiology and secondary achalasia arising as a result of Chagas’ disease, diabetic autonomic neuropathy, or infiltration of the wall of the oesophagus by malignant tumours, amyloidosis or sarcoidosis, and diseases affecting the vagal motor nuclei. Primary achalasia accounts for the majority of cases worldwide, however, the aetiology of the disease remains unknown (Farrokhi & Vaezi 2007). It is now generally accepted that the clinical manifestations of this disease most likely result from dysfunction of Auerbach's plexus (with the majority of pathological studies revealing variable loss of ganglionic cells) (Cross 1952; Adams et al. 1976; Goldblum et al. 1996), however, the role of inflammatory infiltrations in the pathogenesis of the disease remains a controversial issue. Although it is a rare disease in humans, megaoesophagus has also been reported in rats, dogs, cats and mice; all these reports pertain to histopathological changes in the myenteric plexus (Randelia & Lalitha 1988). Megaoesophagus has been observed in the inbred ICRC/HiCri strain of mice (with a wide range of expression from a slightly enlarged to a dilated oesophagus), and although it was found to be an autosomal recessive condition, no investigation as to the genetic origin of the disease was carried out (Randelia & Lalitha 1988).

The RASSF1 gene (Ras association family member 1), located on chromosome 3p21.3, encodes several isoforms through the use of alternative promoters and differential splicing. The most predominant isoforms are RASSF1A and RASSF1C, and selective hypermethylation of the RASSF1A promoter is one of the most frequent epigenetic events in human tumours (reviewed in Donninger et al. 2007). RASSF1A contains a Ras association (RA) domain and is potentially an effector of the Ras oncoprotein. RASSF1A modulates multiple apoptotic and cell cycle checkpoint pathways (reviewed in van der Weyden & Adams 2007) and its role as a tumour suppressor gene was confirmed by Rassf1a-null mice showing enhanced susceptibility to tumourigenesis both spontaneously and after exposure to mutagens (Tommasi et al. 2005; van der Weyden et al. 2005). We have previously reported that Rassf1a-null mice show an increased incidence of tumours of the gastrointestinal tract (van der Weyden et al. 2005), and Rassf1a-null mice on a Min (Apc^+/Min) background show elevated levels of intestinal polyp formation and decreased survival compared with wild-type littermates (van der Weyden et al. 2008). Examination by necropsy of Rassf1a mice on tumour watch from three different cohorts (spontaneous/untreated, irradiated and irradiated on a Blooms-deficient background) led us to observe that many showed a dilated oesophagus, and that this anomaly occurred more frequently in Rassf1a-null mice than their wild-type littermates. In this report, we describe the histology of this phenotype in detail and show that the Rassf1a-null mouse represents a good model of this disease in humans.

Materials and methods

Mice

Mice were housed in accordance with Home Office regulations (UK). The mice were housed in individually ventilated cages within an air-conditioned animal house, and were free from infection. They were fed a diet of mouse pellets and water ad libitum. Rassf1a wild-type and -null mice (on a mixed 129/Sv-C57BL/6J background) were generated by interbreeding of Rassf1a heterozygous mice and were placed on tumour watch (spontaneous) or irradiated with 3.5 Gy at 4–5 weeks of age before being placed on tumour watch (irradiated) (van der Weyden et al. 2005). Rassf1a heterozygous mice were also interbred with Blooms-deficient mice (Blm homozygous-null mice), which show an increased rate of loss of heterozygosity resulting from mitotic recombination (Luo et al. 2000) to generate Rassf1a wild-type and -null mice on a Blm-deficient background, which were sub-lethally irradiated with 3.5 Gy at 4 weeks of age before being placed on tumour watch (Blm^−/− & irradiated). For each of the three cohorts, the mice were examined twice daily for signs of morbidity at which time they were killed and a full necropsy performed. Genotyping for the Rassf1a (van der Weyden et al. 2005) and Bloom (Luo et al. 2000) alleles was performed as described previously.

Histology

At the time of necropsy, all tissues were collected and placed in 10% neutral-buffered formalin. All specimens were then routinely processed for conventional paraffin embedding, sectioned and stained with haematoxylin and eosin. For the oesophagus, longitudinal and transverse sections were examined.

Immunohistochemistry

Rabbit polyclonal anti-S100 antibody (which stains nerve fibres and ganglion cells; Dako, Ely, UK) was used for immunoperoxidase detection on the Techmate system using Dako ‘Real’ reagents (Dako) with 4 mm formalin-fixed, paraffin-embedded tissue sections. Sections were deparaffinized in xylene (3× 5 min) and rehydrated through graded alcohols to distilled water. Antigen retrieval was performed using 0.1% chymotrypsin in 0.1% calcium chloride, pH 7.8, for a range of digestion times from 20 to 50 min, with 30 min as the optimum time. Endogenous peroxidase activity was blocked with 1% H₂O₂ solution for 15 min. The anti-S100 primary antibody (1/500 dilution in Dako ‘Real’ antibody diluent) was incubated with sections for 45 min at room temperature. Following washing of sections with ‘Real’ wash reagent, the sections were further incubated for 45 min in swine anti-rabbit biotinylated antibody (1/250; Dako) for the second layer. The Avidin–Biotin staining protocol was used according to the manufacturers’ guidelines using the ‘Real’ reagents (Dako). The third layer was a strepavidin conjugate incubated for 30 min. Slides were further developed using 3, 3′-diaminobenzidine (DAB), which served as the chromogen generating a brown end product. Sections were counterstained with Mayer's haematoxylin. No signal was detected in sections when the primary antibody was omitted.

Results

Prevalence of megaoesophagus observed by gross examination

At birth, the oesophagus of Rassf1a-null and wild-type mice was normal in appearance, and dilated oesophaguses were only observed in aged mice that had been placed on tumour watch. As shown in Table 1, megaoesophagus was observed in Rassf1a mice from three different cohorts (spontaneous, irradiated and irradiated on a Blooms-null background). The anomaly showed a variable penetrance with not all Rassf1a-null mice developing the condition, although Rassf1a-null mice developed the condition statistically more significantly than their wild-type littermate controls (∼20%vs. ∼2% respectively; P = 0.01 using a two-tailed Fisher's Exact test).

Table 1.

Details of the Rassf1a mice on tumour watch that developed megaoesophagus

	Spontaneous cohort*		Irradiated cohort**		Blm−/− & irradiated cohort***
	Rassf1a+/+	Rassf1a−/−	Rassf1a+/+	Rassf1a−/−	Rassf1a+/+	Rassf1a−/−
No. of necropsied mice showing megaoesophagus	1/50 (2%)	8/35 (23%)	0/44 (0%)	7/35 (19%)	1/49 (2%)	10/50 (20%)

^*Rassf1a^+/+vs. Rassf1a^−/−, P = 0.01.

^**Rassf1a^+/+vs. Rassf1a^−/−, P = 0.005.

^***Rassf1a^+/+vs. Rassf1a^−/−, P = 0.01.

Food particles were sometimes found in the lowermost part of the oesophagus, below the level of the diaphragm. The anomaly showed a wide range of expression from being slightly dilated to being severely grossly dilated and distended (Figure 1). The severity of the dilatation appeared to reflect the age of the mouse at the point of examination (many of the mice developed tumours and it was this that was responsible for their signs of morbidity, which led to their killing), and the older the age of the mouse the more dilated the oesophagus. The age ranges (median) of the mice were 14–76 (47), 28–82 (69) and 63–78 (75) weeks for the three cohorts respectively (spontaneous, irradiated and irradiated on a Blooms-null background).

Figure 1.

Macroscopic analysis of dilated oesophagus from Rassf1a-null mice. Representative macroscopic views of Rassf1a-null mice showing a healthy/normal oesophagus (a) and a dilated oesophagus (b) in situ. Direct comparison of the excised oesophagus and stomach from three individual Rassf1a-null mice at time of necropsy (c) with normal on the left, and examples of megaoesophagus in the centre and on the right. The ages of the mice were 81, 76 and 63 weeks, from left to right.

Analysis of oesophageal changes by histological examination

In humans, the muscle coat in the upper third of the oesophagus is composed of striated muscle, the lower third composed of smooth muscle and the middle third has both types, with a transition from striated to smooth (Weisbrodt 1976). The smooth muscle is innervated by the autonomic nervous system, which includes the submucous (Meissner's) and myenteric (Auerbach's) plexuses. In the mouse (as with the rat and dog), the muscle coat of the entire oesophagus consists of striated muscle, which is innervated by motor axons of both vagus nerves (Gruber 1978). In healthy Rassf1a-null and wild-type mice (Figure 3a,c), four layers were seen throughout the length of the oesophagus, namely mucosa, submucosa, muscularis and serosa. The mucosa was lined by stratified squamous epithelium and the muscular coat consisted of both circular and longitudinal striated muscle fibres. Auerbach's myenteric plexus, whose ganglia were seen as groups of pale cells, were located between the two muscle coats.

Figure 3.

Histological analysis of dilated oesophaguses from Rassf1a-null mice on a Bloom-null background. Representative examples of Rassf1a-null, Bloom-null mice showing a healthy/normal oesophagus (a) and a dilated oesophagus with luminal surface bacteria and a chronic inflammatory infiltrate around the wall (b); the ages of the mice were 79 and 76 weeks respectively. In a healthy oesophagus (c), there were no inflammatory cells surrounding the muscle or nerve cells, whereas the dilated oesophagus (d) showed chronic inflammatory cells underlying the muscular wall of the oesophagus; the ages of the mice were 84 and 63 weeks respectively. In a mild case (early stage of the disease), a few inflammatory cells were seen approaching ganglion cells (e; highlighted by dashed white ellipse), whereas later stages of the disease showed many inflammatory cells surrounding a ganglion cell (f; highlighted by dashed white ellipse); both mice were aged 76 weeks. Chronic inflammatory cells were commonly seen infiltrating skeletal muscle in the intra-abdominal oesophagus at early/mid stages of the disease (g) and chronic inflammatory cells and fibrosis infiltrating and replacing the muscle wall of oesophagus were commonly observed in later stages of the disease (h); the ages of the mice were 63 and 78 weeks respectively. Panels a, b are 200× magnification and panels c–h are 400× magnification.

The histological features of megaoesophagus cases in Rassf1a-null mice were consistent across all the three cohorts, with Figure 2 showing cases from Rassf1a-null spontaneous and irradiated mice, and Figure 3 shows cases from Rassf1a-null mice on a Bloom-null background. H&E-stained sections from megaoeosphagus mice showed a dilated oesophagus with the surface lining of the oesophagus covered with bacteria (consistent with food stasis; Figures 2a and 3b). The dilated oesophaguses also showed mixed chronic inflammation, which in the early stages of the disease was characterized by infiltration into the muscle of the oesophageal wall (Figures 2d and 3g) and in the later stages of the disease was characterized by fibrosis infiltrating and replacing the muscle wall of oesophagus (Figure 3h). Inflammation was also found to be approaching the ganglion cells and nerve fibres (in the early stages of the disease; Figures 2b,c and 3d,e) and completely surrounding them in the later stages of the disease (Figures 2b,c and 3f). There was also evidence of food stuff (plant-based material) present in the bronchial lumena of the lungs of some mice with megaoesophagus, consistent with aspiration pneumonia, which may be seen in human cases of megaoesophagus (as when the oesophagus loses all tone and dilates, the reflex protecting the airways is disrupted and aspiration pneumonia may follow); see Figure 2e,f.

Figure 2.

Histological analysis of dilated oesophaguses from Rassf1a-null mice. The key features of megaoesophagus seen in Rassf1a-null mice included (a) a dilated oesophagus with bacteria lining the squamous epithelial surface, (b) chronic inflammation surrounding nerve fibres and (c) approaching a ganglion cell (indicated by arrow), (d) chronic inflammatory cells invading the oesophagus wall muscle layer, and (e, f) aspiration pneumonia with large infiltrates of neutrophils with food particles in the bronchial lumena of the lung (indicated by arrows). Panel a is 100× magnification, panels b, c are 400× magnification and panels d–f are 200× magnification. Panels a–d were taken from a mouse aged 69 weeks and panels e, f from a mouse aged 62 weeks.

In humans, endoscopic examination is important to rule out malignancy as the cause of achalasia. In our cohorts, most of the mice with megaoesophagus also had cancer (leukaemia/lymphoma, carcinoma and/or adenocarcinoma; n = 7/9 (78%), 5/7 (71%) and 8/11 (72%) cases in the spontaneous, irradiated and irradiated on a Blooms-null background cohorts respectively). In all but two of these cases, there was no evidence of tumourigenesis of the oesophagus and the cause of the oesophageal dilatation was not because of blocked passage of food due to the presence of a tumour (either primary or secondary), which is consistent with the findings observed in the ICRC/HiCri mice (which develop megaoesophagus irrespective of the concurrence of mammary tumours or leukaemias) (Randelia & Lalitha 1988). For the two cases, which showed the presence of megaoesophagus and mild to moderate amount of lymphoma around the wall of oesophagus and gastro-oesophageal junction, it is most likely to be Rassf1a-null associated megaoesophagus with incidental lymphomatous infiltrate in the wall (as the lymphoma was widespread amongst the tissues); however, we cannot exclude lymphoma as a contributory factor to the development of megaoesophagus. There were 7/27 (22%) cases of megaoesophagus (n = 2/9 (22)%, 2/7 (28%) and 2/11 (18%) cases from the spontaneous, irradiated and irradiated on a Blooms-null background cohorts respectively), in which there was a complete absence of tumours in all the tissues examined. Interestingly, these cases did not show greater severity in the degree of oesophageal dilatation/inflammation observed and showed a similar age distribution to cases in which tumours were present.

Immunohistochemical analysis

S100 immunostaining of oesophageal sections confirmed that megaoesophageal cases showed a marked reduction in the numbers of S100 positive nerve cells (both ganglia and nerve fibres) in the myenteric plexus compared with control (healthy) oesophageal cases (Figure 4). In some megaoesophageal cases, there were almost no S100 positive neural cells in the muscle layer of the oesophageal wall (Figure 4b,d). This is what was most likely responsible for the failure to correctly transport the bolus of food by peristalsis and failure to correctly open the LOS with subsequent dilatation of the oesophagus, resulting in megaoesophagus/achalasia. The reduction in nerve cells was also accompanied by a chronic inflammatory infiltrate and subsequent fibrosis of the myenteric plexus and the muscle layers.

Figure 4.

S100 immunohistochemical analysis of normal oesophagus and megaoesophagus from Rassf1a-null mice. In contrast with the normal oesophagus, which shows clusters of S100 positive ganglion/neural cells in the myenteric plexus of the muscle layer (a & c), two megaoesophagus cases show either no or only an occasional S100 positive ganglion/neural cell within the muscle layer of the oesophageal wall (b & d), together with an S100-positive nerve fibre [lower right in both images] external to the oesophageal wall that is included as an internal positive S100 staining control. The ages of the mice were 79 weeks for a, 78 weeks for b, 84 weeks for c and 63 weeks for d. All panels are 400× magnification.

Discussion

The occurrence in humans of ‘achalasia’, a neuromuscular disorder of the oesophagus, has prompted a number of investigators to develop an animal model (Harris et al. 1960; Gruber 1978), but techniques such as vagotomy or destruction of ganglion cells in dogs, cats and monkeys were not successful. Indeed, even spontaneously occurring cases of megaoesophagus in dogs and rats are not suitable as evidenced from manometric, pharmacological and histopathological findings (Diamant et al. 1973; Harkness & Ferguson 1979). The inbred ICRC/HiCri mouse strain has been reported to spontaneously develop megaoesophagus, however, the genetic cause of this anomaly has never been identified (Randelia & Lalitha 1988; Randelia et al., 1990).

In this report, we show that Rassf1a-null mice, which carry a targeted deletion of the Rassf1a isoform of the Rassf1 gene, display an enhanced susceptibility to megaoesophagus compared with wild-type littermates. Histological analysis reveals a reduction in the numbers of nerve cells (both ganglia and nerve fibres) in the myenteric plexus leading to failure to correctly transport the bolus of food by peristalsis and failure to correctly open the LOS with subsequent dilatation of the oesophagus, resulting in megaoesophagus/achalasia. This reduction in nerve cells is accompanied by a chronic inflammatory infiltrate and subsequent fibrosis of the myenteric plexus and the muscle layers. These appearances closely mimic the gross and histopathological findings in human cases of megaoesophagus/achalasia, indicating that this is a representative mouse model of the disease.

Our findings of the abnormalities of the myenteric plexus with a reduction in ganglion cells are similar to that reported for the megaoesophagus phenotype observed in the ICRC/HiCri mouse strain (Randelia & Lalitha 1988; Randelia et al., 1990). Although studies on this mouse strain describe mostly fibrosis in the very old mice (Randelia & Lalitha 1988; Randelia et al., 1990), we saw more evidence of chronic inflammation than fibrosis, which most likely reflects the fact that we were looking at an earlier stage in evolution of the disease in the Rassf1a-null mice. The inflammatory infiltrate may play an active part in the destruction of nerve cells and muscle cells with replacement by fibrosis at a later stage, which would support the concept for the inflammatory, possibly autoimmune, aetiology of autonomic nervous system injury, in PA in humans (Raymond et al. 1999). However, although the possibility cannot be excluded that the inflammation may be a secondary response to loss of nerve cells by a genetic mechanism that leads to necrotic cell death of the nerves, it is tempting to speculate that RASSF1A may influence the pattern or duration of the inflammatory response resulting in chronic inflammation as other members of the RASSF family have recently been shown to regulate the immune system. For example, RASSF2A has been shown to suppress the expression of inflammatory cytokines, which may in turn suppress angiogenesis and invasion (Maruyama et al. 2008). Similarly, the RASSF6 locus is implicated in susceptibility to bronchiolitis induced by respiratory syncytial virus (Hull et al. 2004), and RASSF6 has been shown to be highly effective at suppressing the NF-κB pathway (Allen et al. 2007).

Thus, although further analysis is required to determine the exact nature of the mechanistic link between the absence of Rassf1a and nerve cell reduction, loss or inflammatory destruction, it appears that Rassf1a plays a critical role in maintaining oesophageal myenteric plexus nerve cell viability over the long term. Thus, in conclusion, we have identified a genetic cause of development of megaoesophagus/achalasia that could be screened for in patients, and further study of this may lead to mechanistic insights that eventually provide therapies that could prevent further progression of the disease once it is diagnosed at an early stage.