Detection of Helicobacter-like DNA in the gastric mucosa of Thoroughbred horses


  • M. Contreras,

    1.  Laboratorio de Trombosis Experimental, Centro de Biofísica y Bioquímica, Instituto Venezolano de Investigaciones Científicas, Miranda, Venezuela
    Search for more papers by this author
  • A. Morales,

    1.  Departamento de Patología Veterinaria, Facultad de Ciencias Veterinarias, Universidad Central de Venezuela, Aragua, Venezuela
    Search for more papers by this author
  • M.A. García-Amado,

    1.  Laboratorio de Trombosis Experimental, Centro de Biofísica y Bioquímica, Instituto Venezolano de Investigaciones Científicas, Miranda, Venezuela
    Search for more papers by this author
  • M. De Vera,

    1.  Departamento de Patología, Instituto Nacional de Hipódromos “La Rinconada”, Intercomunal Coche, Caracas, and Hipódromo Nacional de Valencia, Carabobo, Venezuela
    Search for more papers by this author
  • V. Bermúdez,

    1.  Departamento de Patología Veterinaria, Facultad de Ciencias Veterinarias, Universidad Central de Venezuela, Aragua, Venezuela
    Search for more papers by this author
  • P. Gueneau

    1.  Laboratorio de Trombosis Experimental, Centro de Biofísica y Bioquímica, Instituto Venezolano de Investigaciones Científicas, Miranda, Venezuela
    Search for more papers by this author

P. Gueneau, Laboratorio de Trombosis Experimental, Centro de Biofísica y Bioquímica, Instituto Venezolano de Investigaciones Científicas, Km 11 Carretera Panamericana, Altos de Pipe, Miranda, Venezuela.


Aims:  To assess the presence of Helicobacter DNA in the gastric mucosa Thoroughbred horses.

Methods and Results:  Squamous and glandular mucosa samples were collected from 20 Thoroughbreds. None of these horses had shown any clinical symptoms of gastrointestinal disease. Necropsy tissues were analysed using histopathological techniques and a Helicobacter genus-specific PCR assay followed by sequencing of the amplicons. Seven horses were diagnosed with gastric ulceration, five with gastritis and six with both pathologies. Only two horses had a healthy gastric mucosa. Helicobacter-like DNA was detected in two out of seven horses with gastric ulcers, three out of five horses with gastritis, five out of six horses with both pathologies and one horse with normal gastric mucosa. The sequences of 1195 and 1237 bp fragments of the 16S rRNA gene shared 99% identity with the Helicobacter pylori 16S rRNA gene. However, all the samples were negative when tested with H. pylori-specific PCR assays targeting the cagA and glmM genes.

Conclusions:  The Helicobacter genus might colonize the gastric mucosa of horses.

Significance and Impact of the Study:  This is the first report of Helicobacter-like DNA in the gastric mucosa of horses and the pathogenic potential of these organisms requires further investigation.


Helicobacter species have been detected in both humans and a broad range of animal hosts suffering gastritis or gastric ulcers (Fox 2002). Recently, a new enterohepatic Helicobacter species, H. equorum, was isolated from faecal samples of two clinically healthy horses (Moyaert et al. 2007a). The prevalence of H. equorum DNA was significantly higher in hospitalized horses (7·9%) than in healthy, privately owned horses (0·8%, = 0·02) (Moyaert et al. 2007b).

The incidence of gastric ulcers in racehorses in active training has been shown to exceed 90% in some circumstances (Murray et al. 1996; Bezdùkováet al. 2005). The cause of this is still unknown. It has been proposed that various stress factors are involved, while diet, management and training practices are regarded as potential risk factors (Murray et al. 1996). Surprisingly, in horses, where gastritis and gastric ulcers are very common, there exist very few reports of Helicobacter infection in the gastric mucosa (Hepburn 2004; Andrews et al. 2005). The aim of this study was to detect Helicobacter spp. in the gastric mucosa of Thoroughbred horses from Venezuela.

Materials and methods


We studied the stomachs of 20 Thoroughbred racehorses (Equus caballus) with ages between 2 and 6 years; 13 (seven stallions and six mares) were from the La Rinconada Racecourse in Caracas and the other seven (five stallions and two mares) were from the Hipódromo Nacional de Valencia. All the horses had been sacrificed because they had open fractures of their legs (usually the proximal sesamoids). The euthanasia in these animals was performed according to previously described protocols by the American Association of Equine Practitioners (Euthanasia 1995, American Equine Practitioner). The stomach samples were obtained by necropsy, as described previously (King et al. 1981) and were immediately processed. None of these horses had previously shown any clinical symptoms of gastro-intestinal disease.

From each stomach, two sections (approx. 1 cm2) of gastric tissue were taken from both the squamous and glandular mucosa near to the margo plicatus, with particular care to avoid inter-sample contamination: sterile instruments were used in each sampling. One sample of each type was fixed in 100% ethanol and the others were fixed in 10% formalin.


The tissue samples fixed in formalin were processed by conventional histological techniques: dehydration, inclusion in paraffin, microtome slicing and haematoxylin–eosin and Warthin-Starry staining (Luna 1968). Morphometric study was performed by comparing regeneration vs tissue injury by using Image Tool and Stesys Test programs (Karen et al. 1998). Briefly the measures were obtained from the tip of the epithelium of the basal membrane of each gastric sample and from the ulcers along the margo plicatus and compared with healthy gastric areas from the fundic regions. Measures were obtained from the dorsal aspect of the muscular mucosa to the basal surface of the bottom of the ulcers cavity. All the measures were made in microns.

DNA extraction

DNA was extracted from the tissue samples fixed in ethanol by using the Invisorb spin Tissue Mini Kit (Invitek, Berlin, Germany) according to the manufacturer’s instructions.

Helicobacter genus-specific and H. pylori-specific PCR assays

Helicobacter DNA was detected using a genus-specific PCR assay targeting a 399 bp fragment of the 16S rRNA gene (Germani et al. 1997) (Table 1). The presence of H. pylori DNA was ruled out using specific assays targeting a 294 bp fragment of the glmM gene and a 128 bp fragment of the cagA gene (Kansau et al. 1996; Rugge et al. 1999) (Table 1). PCR was performed using the Ready To-Go PCR beads kit (Amersham Biosciences, Piscataway, NJ) and the thermal cycler model GeneAMP PCR System 9700 (Applied Biosystems, Foster City, CA). The positive control was H. pylori DNA and the negative control was a no-template PCR reaction. The amplicons were visualized by running the reaction mixture in a TBE agarose gel (at 2·0%), staining it with ethidium bromide, and observing them using an UV transilluminator.

Table 1.   Primers used in the PCR assays and DNA sequencing
PrimersSequence (5′–3′)Source or reference (nt positions)
515FGTGCCAGCMGCCGCGGTAAE.coli 16S rRNA, nt 515–533
519RGWATTACCGCGGCKGCTGE.coli 16S rRNA, nt 519–536
926FGGGCCCGCACAAGCGGTE.coli 16S rRNA, nt 926–942
926RACCGCTTGTGCGGGCCCE.coli 16S rRNA, nt 926–942
1525RAAGGAGGTGATCCAGCCE.coli 16S rRNA, nt 926–942

16S rRNA gene sequencing and sequence analysis

Amplicons (399 bp) obtained from the squamous mucosa DNA of five horses selected at random (1E, 2e, 3e, 5e and 5E) using the Helicobacter genus-specific PCR assay (Germani et al. 1997) were purified for sequencing using the CONCERTTM Rapid PCR Purification System kit (Gibco BRL-Invitrogen, Gaithersburg, MD), according to the manufacturer’s recommendations. The primers used for sequencing of both DNA strands were HeliF and HeliR (Table 1).

Additionally, 16S rRNA genes from two horses samples (5E and 5g), were amplified by nested PCR; the first with two external universal primers (8F and 1525R), as stated above, and the second PCR using two internal primers, one Helicobacter genus-specific (HeliF) and the other targeting the Epsilon branch of Proteobacteria (EpsilR) (Germani et al. 1997) (Table 1), yielding 1288 bp amplicons. For a typical one-step reaction, 6 μl of purified DNA was used for 16S rRNA gene detection. Amplification was carried out in 50 μl using the Ready To-Go PCR beads Kit according to conditions mentioned above. Next, 5 μl aliquots of the first amplification product were added to a new batch of 50 μl PCR. Amplification was performed according to the author’s recommendations (Germani et al. 1997). The amplicons were then purified in gel using the CONCERTTM Gel Extraction System Kit (GibcoBRL-Invitrogen) and directly sequenced. The primers used for sequencing were those mentioned above (HeliF and EpsilR) plus the internal universal primers 515F, 519R, 926F and 926R (Table 1).

Both strands of all seven of the purified amplicons were sequenced at the CeSAAN facility (IVIC, Altos de Pipe, Venezuela) with on ABI PRISMTM 3130xl Sequencer (Applied Biosystems, Foster City, CA). Sequences were compared with the compilation of 16S rRNA gene sequences available in the GenBank nucleotide library by BLAST searching. The 16S rRNA gene sequence of amplicons 5E and 5g were deposited in GenBank under the accession numbers DQ829805 and DQ829806 respectively.

Statistical analyses

Non-parametric Kruskal-Wallis tests were performed to compare the measures between injured gastric mucosa (ulcers gastric) and the healthy gastric mucosa (fundic region) in each animal obtained from 13 horses affected by ulcers used in the morphometric study, and a P-value < 0·05 was defined as significant.


Histopathological observation

Of the 20 Thoroughbred horses studied, 2 (10%) presented normal mucosae and 18 (90%) had lesions in their stomachs. Seven (35%) of the 20 horses had gastric ulcerations around the margo plicatus (Fig. 1a), five out of 20 (25%) had gastritis (Fig. 1b) and six out of 20 (30%) were presented with both types of lesions (Table 2). Moreover, the histologic slices revealed a loss of continuity of the gastric mucosa, with submucosa exposure and oedema with parakeratotic hyperkeratosis together with a mixed lymphoplasmocytic mononuclear infiltrate. With regard to ulcer distribution, both regions of the stomachs nearly to the margo plicatus showed similar patterns of lesions.

Figure 1.

 Gastric histopathology demostrating ulcer and gastritis in two horses (a and b). (a) Squamous region from a horse with gastric ulcer adjacent to margo plicatus. (b) Glandular region from other horse with superficial gastritis, oedema and slight lymphoplasmocytic infiltrate; haematoxylin-and-eosin (H&E) staining in both figures.

Table 2.   The histopathological results in relation to the Warthin-Starry (WS) stain and the presence of Helicobacter (PCR) in the horses’ stomachs
LesionsNumber of horsesWS stainHelicobacter PCR
Ulcers (U)722
Gastritis (G)543
G and U655

The morphometric studies revealed a significant loss of gastric mucosa in 13 horses affected by ulcers (data not shown from 20 samples stomachs), as an average damage of 351·4 ± 41·4 μm in depth could be observed. Using the Warthin-Starry special stain, spiral-shaped bacteria were found in two out of seven horses with ulcers, four out of five animals with gastritis and five out of six horses with both pathologies (Table 2).

PCR assays

Helicobacter-like DNA was detected by PCR in the two mucosal regions (squamous and/or glandular mucosae) of two out of seven horses with ulcers, three out of five animals with gastritis, five out of six horses with both pathologies and one out of two with normal mucosa. Every sample that was positive for the Warthin-Starry stain was also PCR-positive for the Helicobacter PCR assay, except in one case (2E) (Table 2). The cagA and glmM genes were not detected in any of the 40 gastric samples (data not shown). In addition, comparing the presence of Helicobacter DNA with the morphometric studies, we noted that the Helicobacter-positive samples had less profound ulcerous wounds than the gastric ulcers without Helicobacter (mean depth 335·2 ± 42·5 and 360·3 ± 39·4 μm respectively). This difference was not statistically significant (= 0·239, by the Kruskal-Wallis test), but a larger set of samples would be required to ascertain this observation.

Sequence analysis

The 16S rRNA partial gene sequences from Helicobacter spp. detected in seven horses were amplified by PCR using Helicobacter genus-specific primers and primers targeting the Epsilon branch of Proteobacteria. First, we sequenced 399 bp amplicons from five squamous mucosa (with gastric ulcers or gastritis) DNA samples that were positive for the presence of Helicobacter. The sequence of these amplicons was 98–99% identical to Helicobacter genus 16S rRNA gene sequence. A larger fragment of the Helicobacter 16S rRNA gene was amplified and sequenced from two samples (one squamous mucosa (5E) and one glandular mucosa (5 g)). The sequences were 1195 and 1237 bp, corresponding to approx. 79% of the total 16S rRNA gene sequence. Both sequences (5E and 5g) shared 99% identity to H. pylori 16S rRNA gene sequence and 93% identity to Helicobacter equorum 16S ribosomal RNA gene, complete sequence.


A high prevalence (90%) of ulcers and gastritis was found in Thoroughbred racehorses during our study even though none of them had a previous record of gastrointestinal disorders at the time of their euthanasia. This agrees with other studies reporting the occurrence of gastric ulcers in 80–90% of Thoroughbred racehorses (Murray et al. 1996; Jones 2002; Fink et al. 2006).

The proximal half of a horse’s stomach is entirely lined by a stratified squamous mucosa. The more distal glandular portion is similar to that of other mammalian species. The junction between squamous and glandular mucosae is the margo plicatus. The greatest proportion of such lesions appears on the squamous mucosa region close to the margo plicatus, with fewer lesions on the glandular mucosa portion (Murray 1994; Pagan 1999). However, our results show that the gastric ulcers as well as the gastritis were present on both mucosal regions nearly to the margo plicatus.

Although diverse species of Helicobacter have been linked to ulcers and to non-ulcerative gastritis in other animals (Fox 1997, 2002) none has been cultured from the stomach of horses. To date, only two studies have reported the presence of Helicobacter-specific DNA in the squamous and glandular mucosae of a total of 19 horses (Hepburn 2004; Andrews et al. 2005) and one new species of Helicobacter named H. equorum was isolated from the faeces of two asymptomatic horses (Moyaert et al. 2007a).

Helicobacter genus-specific PCR amplicons were only obtained in 11 of the 20 horses samples. Five of these 399 bp amplicons were sequenced and shared 98–99% identity with Helicobacter genus 16S rRNA genes. A larger fragment of this gene was obtained from two of the stomach samples and shared 99% identity with H. pylori. However, the H. pylori specific genes glmM and cagA were never detected in any of the samples. The 16S rRNA gene sequence of the species H. cetorum, isolated from marine mammals, also shares 99% sequence identity with the H. pylori gene sequence (Harper et al. 2002). It is well known that different bacterial species might share almost identical 16S rRNA gene sequences, which is why a new species cannot be described solely on the basis of that gene sequence (Dewhirst et al. 2005). Interestingly, our sequences were less similar (93%) to sequences reported from the faeces of horses (H. equorum, DQ307735 and DQ307736) (Moyaert et al. 2007a). Unfortunately, we have not been able to isolate the bacteria that we see in the histological sections or detect using PCR.

Ninety per cent (10/11) of the horses infected by Helicobacter examined in this study showed gastric lesions expressed as gastric ulcers, gastritis or both pathologies, except in one case in which we detected Helicobacter in the normal mucosa. This suggests that Helicobacter species are present in stomach of Thoroughbred horses. However, 39% of our horses with gastric pathologies did not show Helicobacter or other bacteria, indicating that lesions also may be due to other causes besides the presence of micro-organisms. Another explanation of these lesions in horses in training is due to increased exposure of the squamous mucosa to the highly acidic contents of the distal stomach (Lorenzo-Figueras and Merritt 2002).

In conclusion, this is the first report of Helicobacter-like DNA in the squamous and glandular mucosae of Thoroughbred horses. In other mammals, Helicobacter spp. were shown to be the etiological agent of gastric ulcers and gastritis. However, in Thoroughbred horses the pathogenic potential of Helicobacter genus requires further investigation.


This study was funded by the Instituto Venezolano de Investigaciones Científicas and in part by FONACIT S1-1999000102 to P.G. The authors gratefully acknowledge the CeSAAN Service at IVIC. We thank H. Rojas for technical assistance with the histological photomicrographs.