First evidence of infectious hematopoietic necrosis virus (IHNV) in the Netherlands

In spring 2008, infectious hematopoietic necrosis virus (IHNV) was detected for the first time in the Netherlands. The virus was isolated from rainbow trout, Oncorhynchus mykiss (Walbaum), from a put-and-take fishery with angling ponds. IHNV is the causative agent of a serious fish disease, infectious hematopoietic necrosis (IHN). From 2008 to 2011, we diagnosed eight IHNV infections in rainbow trout originating from six put-and-take fisheries (symptomatic and asymptomatic fish), and four IHNV infections from three rainbow trout farms (of which two were co-infected by infectious pancreatic necrosis virus, IPNV), at water temperatures between 5 and 15 °C. At least one farm delivered trout to four of these eight IHNV-positive farms. Mortalities related to IHNV were mostly <40%, but increased to nearly 100% in case of IHNV and IPNV co-infection. Subsequent phylogenetic analysis revealed that these 12 isolates clustered into two different monophyletic groups within the European IHNV genogroup E. One of these two groups indicates a virus-introduction event by a German trout import, whereas the second group indicates that IHNV was already (several years) in the Netherlands before its discovery in 2008.

Development and validation of a real-time PCR assay for the detection of anguillid herpesvirus 1

Anguillid herpesvirus 1 (AngHV1) causes a haemorrhagic disease with increased mortality in wild and farmed European eel, Anguilla anguilla (L.) and Japanese eel Anguilla japonica, Temminck & Schlegel). Detection of AngHV1 is currently based on virus isolation in cell culture, antibody-based typing assays or conventional PCR. We developed, optimized and concisely validated a diagnostic TaqMan probe based real-time PCR assay for the detection of AngHV1. The primers and probe target AngHV1 open reading frame 57, encoding the capsid protease and scaffold protein. Compared to conventional PCR, the developed real-time PCR is faster, less labour-intensive and has a reduced risk of cross-contamination. The real-time PCR assay was shown to be analytically sensitive and specific and has a high repeatability, efficiency and r(2) -value. The diagnostic performance of the assay was determined by testing 10% w/v organ suspensions and virus cultures from wild and farmed European eels from the Netherlands by conventional and real-time PCR. The developed real-time PCR assay is a useful tool for the rapid and sensitive detection of AngHV1 in 10% w/v organ suspensions from wild and farmed European eels.

Ostreid herpesvirus OsHV-1 μVar in Pacific oysters Crassostrea gigas (Thunberg 1793) of the Wadden Sea, a UNESCO world heritage site

The Wadden Sea is an extensive wetland area, recognized as UNESCO world heritage site of international importance. Since the mid-1990s, the invasive Pacific oyster Crassostrea gigas (Thunberg 1793) population in the area has grown exponentially, having a distinct impact on the ecosystem. The recent spread of the emerging oyster pathogen Ostreid herpesvirus OsHV-1 μVar worldwide and specifically in the oyster culture areas in the south of the Netherlands raised the question whether the virus may also be present in the Wadden Sea. In the summer of 2012 juvenile Pacific oysters were collected from five locations in the Dutch Wadden Sea. The virus was shown to be present in three of the five locations by real-time PCR and sequencing. It was concluded that OsHV-1 μVar has settled itself in Pacific oyster reefs in the Wadden Sea. These results and the recent discoveries of OsHV-1 microvariants in Australia and Korea indicate that OsHV-1 μVar and related variants might be more widespread than can be deduced from current literature. In particular in regions with no commercial oyster culture, similar to the Wadden Sea, the virus may go undetected as wild beds with mixed age classes hamper the detection of mortality among juvenile oysters.

Ultrastructural comparison of Bonamia spp. (Haplosporidia) infecting ostreid oysters

The ultrastructure of Bonamia from Ostrea angasi from Australia, Crassostrea ariakensis from the USA, O. puelchana from Argentina and O. edulis from Spain was compared with described Bonamia spp. All appear conspecific with B. exitiosa. The Bonamia sp. from Chile had similarities to the type B. exitiosa from New Zealand (NZ), but less so than the other forms recognized as B. exitiosa. Two groups of ultrastructural features were identified; those associated with metabolism (mitochondrial profiles, lipid droplets and endoplasmic reticulum), and those associated with haplosporogenesis (Golgi, indentations in the nuclear surface, the putative trans-Golgi network, perinuclear granular material and haplosporosome-like bodies). Metabolic features were regarded as having little taxonomic value, and as the process of haplosporogenesis is not understood, only haplosporosome shape and size may be of taxonomic value. However, the uni-nucleate stages of spore-forming haplosporidians are poorly known and may be confused with Bonamia spp. uni-nucleate stages. The many forms of NZ B. exitiosa have not been observed in other hosts, which may indicate that it has a plastic life cycle. Although there are similarities between NZ B. exitiosa and Chilean Bonamia in the development of a larger uni-nucleate stage and the occurrence of cylindrical confronting cisternae, the clarification of the identity of Chilean Bonamia must await molecular studies.

Vibrio vulnificus outbreaks in Dutch eel farms since 1996: strain diversity and impact

Vibrio vulnificus is a potentially zoonotic bacterial pathogen of fish, which can infect humans (causing necrotic fasciitis). We analysed 24 V. vulnificus isolates (from 23 severe eel disease outbreaks in 8 Dutch eel farms during 1996 to 2009, and 1 clinical strain from an eel farmer) for genetic correlation and zoonotic potential. Strains were typed using biotyping and molecular typing by high-throughput multilocus sequence typing (hiMLST) and REP-PCR (Diversilab®). We identified 19 strains of biotype 1 and 5 of biotype 2 (4 from eels, 1 from the eel farmer), that were subdivided into 8 MLST types (ST) according to the international standard method. This is the first report of V. vulnificus biotype 1 outbreaks in Dutch eel farms. Seven of the 8 STs, of unknown zoonotic potential, were newly identified and were deposited in the MLST database. The REP-PCR and the MLST were highly concordant, indicating that the REP-PCR is a useful alternative for MLST. The strains isolated from the farmer and his eels were ST 112, a known potential zoonotic strain. Antimicrobial resistance to cefoxitin was found in most of the V. vulnificus strains, and an increasing resistance to quinolones, trimethoprim + sulphonamide and tetracycline was found over time in strain ST 140. Virulence testing of isolates from diseased eels is recommended, and medical practitioners should be informed about the potential risk of zoonotic infections by V. vulnificus from eels for the prevention of infection especially among high-risk individuals. Additional use of molecular typing methods such as hiMLST and Diversilab® is recommended for epidemiological purposes during V. vulnificus outbreaks.

Tularaemia in a brown hare (Lepus europaeus) in 2013: first case in the Netherlands in 60 years

Tularaemia has not been reported in Dutch wildlife since 1953. To enhance detection, as of July 2011, brown hares (Lepus europaeus) submitted for postmortem examination in the context of non-targeted wildlife disease surveillance, were routinely tested for tularaemia by polymerase chain reaction (PCR). Francisella tularensis subspecies holarctica infection was confirmed in a hare submitted in May 2013. The case occurred in Limburg, near the site of the 1953 case. Further surveillance should clarify the significance of this finding.

Bonamia exitiosa (Haplosporidia) observed infecting the European flat oyster Ostrea edulis cultured on the Spanish Mediterranean coast

Bonamia exitiosa and Bonamia ostreae are parasites that reproduce within the haemocytes of several oyster species. In Europe, the host species is the flat oyster Ostrea edulis. The parasite B. ostreae has been responsible for mortalities since the late 1970s throughout the European Atlantic coast. B. exitiosa was first detected, in 2007, on this continent in flat oysters cultured in Galicia (NW Spain). Since then, the parasite has also been detected in France, Italy and the United Kingdom. The bays of the Ebro Delta in the south of Catalonia represent the main bivalve culture area in the Mediterranean coast of Spain. Previous information from the area includes reports of several flat oyster pathogens, including the notifiable parasite Marteilia refringens. However, the status with regard to Bonamia parasites was uncertain. In the present study, a Bonamia parasite was observed in flat oysters cultured in the Alfacs Bay of the Ebro Delta by histology and real-time PCR. PCR-RFLP and sequencing suggested the presence of B. exitiosa. Finally, phylogenetic analyses of the studied Bonamia isolates corroborated B. exitiosa infection. M. refringens was also observed in the same oyster batch, and co-infection with both parasites was also detected. This is the first detection of B. exitiosa, in Catalonia and the Spanish Mediterranean coast. The impact of the parasite on the Mediterranean flat oyster activity needs to be urgently addressed.

Inter-relationships of haplosporidians deduced from ultrastructural studies

We reviewed papers reporting haplosporidian ultrastructure to compare inter-relationships based on ultrastructure with those based on molecular data, to identify features that may be important in haplosporidian taxonomy, and to consider parasite taxonomy in relation to host taxonomy. There were links between the following: (1) the plasmodia of an abalone parasite, Haplosporidium nelsoni and Urosporidium crescens in the release of haplosporosomes; (2) H. costale and H. armoricanum in haplosporosome shape and presence and shape of Golgi in spores; (3) basal asporous crustacean haplosporidians which form haplosporosomes from formative bodies (FBs) in vegetative stages--H. nelsoni, which forms haplosporosomes from FBs in plasmodial cytoplasm, and H. louisiana, Minchinia spp. and Bonamia perspora, which form haplosporosomes from FBs in spores; (4) crustacean haplosporidians, Bonamia spp. and M. occulta in the predominance of uni- and binucleate stages; and (5) lipid-like vesicles in sporoplasms of H. costale, H. armoricanum, H. lusitanicum, H. pickfordi, H. montforti, and B. perspora. In general, these relationships reflect phylogenies based on molecular studies. As well as spore form and ornamentation, haplosporogenesis in spores appears to be taxonomically important. Parasite and host taxonomy were linked in the infection of lower invertebrates by Urosporidium spp., the infection of oysters by Bonamia spp., and of molluscs by Minchinia spp. Haplosporidium spp. are patently an artificial, paraphyletic group probably comprising many taxa. Consequently, the taxonomy of haplosporidians needs a thorough revision.

Ultrastructure of sporulation in Haplosporidium armoricanum

An ultrastructural study was carried out on the tissues of an oyster (Ostrea edulis), heavily infected with Haplosporidium armoricanum, that had been fixed in Carson's fixative. The well-fixed tissues revealed details of sporulation and of the spores, which had not been previously reported from H. armoricanum. These include the initial presence of sparse haplosporosomes after thickening of the plasma membrane in early sporonts, division of sporont nuclei by multiple fission, cup-like indentations in the nuclear surface associated with putative nuclear material in both the sporonts and spores, and cytoplasmic multi-vesicular bodies in the cytoplasm of sporonts and spores. The spore wall and operculum were formed from a light matrix that occurred in short cisternae of smooth endoplasmic reticulum in the episporoplasm, and parallel bundles of microfibrils were present in some spores. Spores were rarely bi-nucleate with the nuclei occurring as a diplokaryon, with putative nuclear material at the junction of the 2 nuclei. Nuclear membrane-bound Golgi (NM-BG) cisternae were common in spores, and they appeared to synthesise a light granular material into lysosome-like granules. Dense bodies similar to those reported from H. lusitanicum, H. pickfordi and H. monforti occurred in, or outside, the peripheral endosporoplasm, which was closely apposed to the spore wall. Spore haplosporosomes were frequently axehead-shaped, more like those of H. costale than those previously reported from H. armoricanum, and in some haplosporosomes there was a small round lucent patch with a dark point near the centre of the lucent patch. Overall, H. armoricanum appears to be closely related to H. costale and Bonamia spp. Although the endosporoplasm of H. armoricanum has NM-BG and it resembles the uni-nucleate stage, it appears to be unlikely that they are the same, as the axehead-shaped haplosporosomes of the spore differ considerably from the spherical haplosporosomes of vegetative stages.

Toxicity of PCB-126 in European flounder ( Platichthys flesus ) with emphasis on histopathology and cytochrome P4501A induction in several organ systems

A series of experiments was set up to elucidate the effects of pollution on marine and estuarine fish health, since the European flounder (Platichthys flesus) has shown a relatively high prevalence of (pre)neoplastic liver lesions and lymphocystis virus disease in Dutch coastal and estuarine waters. The hypothesis of a causal relationship between pollution and the above-mentioned diseases was supported by results from semi-field experiments. Therefore several laboratory experiments were carried out to substantiate causality further and to identify the xenobiotics that may play a major role in the field. The present study focuses on polychlorinated biphenyls (PCBs). European flounders (Platichthys flesus) were orally exposed to a single dose of 0, 0.5, 5 or 50 mg PCB-126/kg body weight under controlled laboratory conditions. The effects on liver, gills, gastrointestinal tract, gonads, spleen and mesonephros were examined histologically after 16 days. Induction and localization of cytochrome P4501A (CYP1A) immunoreactivity, and effects on hepatocyte proliferation were visualized immunohistochemically. Effects on thymus size were examined by morphometric analysis of serial sections. Three out of five animals of the highest dose group showed haemorrhages in the fins and tail after 16 days. All animals showed reduced activity in the later stages of the experiment, and some animals of the highest dose group discontinued feeding 14 days after exposure. Strong and exposure-related induction of CYP1A immunoreactivity was noted in hepatocytes, endothelium in all organs examined, and epithelium of the digestive tract and mesonephros at PCB-126 levels of 0.5, 5 and 50 mg/kg. In addition, the strong induction of CYP1A immunoreactivity in a distinct population of haematopoietic cells in the mesonephros and in circulating blood is remarkable, and has not been described previously in other fish species. Furthermore, a morphometrically determined significant reduction in relative thymus size was noted in animals exposed to 50 mg PCB-126/kg. Although the functional implications for the immune system of this reduction need to be further investigated, an impact on the specific resistance against infectious diseases as observed in the field, e.g. viral lymphocystis disease, is not implausible. In addition, a significant increase in absolute liver weight, in hepatosomatic index, and in number of proliferating hepatocytes [measured as immunoreactivity against proliferating cell nuclear antigen (PCNA)] was noted in animals of the highest dose group. From these findings we suppose that PCB-126 (and related chemicals) may play a role in the promotion of tumour development in the liver of European flounders as observed in the field. The results of the present experiment show relatively stronger effects than effects previously reported from experiments with TCDD, suggesting that the TEF of 0.005 assigned to PCB-126 from early life stage mortality experiments in rainbow trout (Oncorhynchus mykiss), underestimates the toxic potential of PCB-126.

Regulation of interleukin 1 beta RNA expression in the common carp, Cyprinus carpio L

The intron-exon organisation of the carp IL-1beta gene consists of 2455bp and comprises seven exons. Three IL-1beta RNA transcripts have been found in carp: (1) a fully spliced product; (2) exon 1-7 with introns 5 and 6; and (3) exon 1-7 with intron 5 only. The intron-containing products probably represent partially spliced transcripts. IL-1beta mRNA expression in carp was semi-quantitatively analysed by RT-PCR in multiple organs, including brain and pituitary. Constitutive expression of the IL-1beta mRNA was found in these organs with a predominant expression in the immune organs head kidney and spleen. Furthermore, a scattered distribution of IL-1beta producing cells was shown by in situ hybridisations of head kidney tissue. Administration of phorbol-myristate-acetate (PMA), lipopolysaccharide (LPS) or retinoic acid (RA), to phagocytes isolated from the head kidney, resulted in expression of IL-1beta intron-containing transcripts. Of these, only PMA and LPS were stimulators that induced the fully spliced transcript. A role for the nuclear factor (NF)-kappaB pathway in carp IL-1beta expression was shown with suppression of the LPS-induced IL-1beta expression by NF-kappaB inhibitor pyrrolidine dithiocarbamate (PDTC). Cortisol was able to inhibit in vitro constitutive expression of IL-1beta transcripts. Addition of cortisol simultaneously with LPS could not substantially inhibit transcription.

Toxicity of TCDD in European flounder (Platichthys flesus) with emphasis on histopathology and cytochrome P450 1A induction in several organ systems

The present study is part of a series of experiments, set up to elucidate the impact of aquatic pollution on fish health in the marine and estuarine environment. In the Dutch coastal and estuarine waters, European flounder (Platichthys flesus) showed a relatively high prevalence of (pre)neoplastic liver lesions and lymphocystis virus disease. The hypothesis of a causal relationship between pollution and these diseases was supported by semi-field experiments. Therefore a series of laboratory experiments was performed to further substantiate causality and identify the xenobiotics that may play a major role in the field. Polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) are important environmental pollutants. They are highly persistent, highly lipophilic, and have shown to induce several toxic effects in mammalian and non-mammalian species at relatively low concentrations. This report describes a study in which European flounder were orally exposed to the most toxic PCDD congener, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) or to harbor sludge extract under controlled laboratory conditions. The effects on several organs (liver, gills, gastro-intestinal tract, thyroid gland, gonads, spleen and mesonephros) were examined microscopically. Induction and localization of cytochrome P4501A (CYP1A) immunoreactivity, and effects on hepatocyte-proliferation were visualized immunohistochemically. Effects on thymus size were examined by morphometric analysis. Oral exposure of flounder to 0.0125 or 0.3125 µg TCDD/kg bw, or to 0.3125 µg TEQ/kg bw of a harbor sludge extract, weekly for 8 weeks, induced a significant increase in CYP1A immunoreactivity in hepatocytes. Single administration of higher doses (20, 100 and 500 µg/kg bw) of TCDD also induced a significant increase CYP1A immunoreactivity in the endothelium in all organs examined, and in the epithelium of the digestive tract, liver, and mesonephros. Remarkably, strong immunoreactivity was noted in a distinct cell population of the hematopoietic tissue in the mesonephros and spleen, which has not been described in fish previously. Moreover, oral exposure to 20 µgTCDD/kg bw resulted in an increased mitotic activity, and an increased hepatosomatic index was found after exposure to 500 µgTCDD/kg bw. In the thymus only a trend in size reduction was noted, again in the highest dose group. Nevertheless, no marked pathology was detected even in fish exposed to a single dose of 500 µg TCDD/kg body weight. The present experiments show that, under the actual experimental conditions, European flounder is relatively insensitive to the toxic effects of TCDD. However, we assume that exposure to TCDD (and related substances) may promote the development of tumors in the field.

Indications for a distinct putative T cell population in mucosal tissue of carp (Cyprinus carpio L.)

A monoclonal antibody against carp intestinal T cells (WCL38; of IgM class) was produced by immunization of mice with isolated membrane molecules of carp intestinal intraepithelial lymphoid cells. Flow cytometric analysis showed that WCL38 reacted with 50-70% of the lymphoid cells isolated from intestine, gills or skin, with less than 6% of lymphoid cells isolated from thymus, head kidney or spleen and with a negligible number of PBL. WCL38+ cells were abundant in the intestinal epithelium and less numerous in the lamina propria. Immunogold labelling confirmed that WCL38 reacted with lymphoid cells; in gills and skin some of them have the morphology of large granular lymphoid cells. Immunochemical analysis showed that WCL38 reacted with dimeric membrane molecule on mucosal lymphoid cells with an Mr of 76 kDa, consisting of two 38 kDa subunits. WCL 38+ lymphoid cells are postulated to T cells, since WCL38 does not react with B cells, macrophages or non-specific cytotoxic cells. In conclusion, like higher vertebrates, carp seem to have a distinct (Putative) T cell population in their mucosal tissues.

Induction of oral tolerance in carp (Cyprinus carpio L.) after feeding protein antigens

Induction of oral tolerance against ferritin, recombinant surface glycoprotein of viral haemorrhagic septicemia virus (KLG18) and ovalbumin (OVA) was studied in carp. Feeding of ferritin or KLG18 resulted in lower Ab titres compared to unprimed controls when animals were intramuscularly (i.m.) injected with protein 10 weeks later and sampled 21 days after this injection. After administration of OVA by different routes (oral, anal, i.m.) and i.m. injection with OVA + Freund's incomplete adjuvant 2 months later, only a few fish responded to OVA as measured by serum Ab titres. Responsiveness to OVA appeared to be carp strain dependent. When an isogenic carp strain was selected for an optimal response to i.m. injection with OVA, this carp strain did not develop oral tolerance after feeding. In contrast, 6 x feeding high doses of OVA on subsequent days, resulted in immunological memory formation. Oral tolerance can be induced in carp, but differences in tolerance induction may depend on the protein used. A possible role of genetic factors in the induction of oral tolerance in fish is discussed.

Activation of cytotoxic T lymphocytes in HLA-A, -B and -C-identical responder-stimulator pairs. II. New subtypes of HLA-Bw35

We compared five cytotoxic T lymphocytes raised by primary mixed lymphocyte cultures of HLA-A, -B and -C serologically identical Bw35-positive responder-stimulator combinations. When tested on a panel of third-party target cells, the reactivity pattern of these cytotoxic T lymphocytes allowed the distinction of three subtypes of HLA-Bw35. Cold-target inhibition experiments and analysis of CTL activity at the clonal level showed the existence of subsets of CTLs directed against distinct antigenic determinants associated with HLA-Bw35.

Activation of cytotoxic T lymphocytes in HLA-A, -B and -C-identical responder-stimulator pairs. I. Variations in generation of anti-class-II CTL in primary mixed lymphocyte cultures

Cytotoxic T lymphocytes were activated in primary one-way mixed lymphocyte cultures of cells matched for serologically defined HLA-A, -B and -C antigens. In 16 out of the 29 combinations mismatched for the HLA-D/DR antigens, cell-mediated lympholysis of the stimulator cells occurred. The specificity of 5 selected cytotoxic T lymphocytes was studied in detail. Three of these cytotoxic T lymphocytes recognize antigenic determinants associated with HLA-Bw35 (Breuning et al. 1984, II). The 2 other cytotoxic T lymphocytes failed to lyse T-target cells enriched by rosetting with sheep red blood cells, whereas target cells from the 'non-T' fraction were strongly lysed, indicating that antigenic determinants associated with Class-II HLA molecules were the targets recognized by these cytotoxic T lymphocytes. This notion was supported by a study of a panel of HLA-typed third-party target cells. One cytotoxic T-lymphocyte population preferentially lysed HLA-DR2-positive target cells. Family studies, including a family with a recombination between HLA-B and -D, showed that the target antigen recognized by the latter cytotoxic T lymphocyte segregated with DR2. The second cytotoxic T-lymphocyte population recognized a determinant associated with DRw8. However, in 13 of the 29 HLA-A-, -B- and -C-identical, D/DR-different combinations, cell-mediated lympholysis of stimulator target cells could not be detected, not even on enriched 'non-T' target cells. Thus, after primary mixed lymphocyte culture of HLA-A-, -B- and C-identical, HLA-D/DR-non-identical cells, cytotoxic T lymphocytes directed against sensitizing Class-II molecules can be detected in some combinations, but not in others.

Specificity of anti-HLA-B27 cytotoxic T lymphocytes

Sub-types of HLA-B27 were detected by cytotoxic T lymphocytes (CTL) generated between HLA-A, -B- and -C-identical B27-positive individuals. We now report the specificity of six independent CTL's generated by mixed lymphocyte culture (MLC) of HLA-A, -B and -C serologically identical B27-positive responder and stimulator cells. Three CTL's recognize one sub-type, and three the other. The combined reactivity of all CTL's allows unequivocal "typing" of B27-positive cells for the two different sub-types B27K and B27W. The specificity of two CTL's was analysed by cold-target inhibition. The results indicate that (1) no further sub-types of HLA-B27 can be detected by the CTL's raised in these combinations; (2) the majority of the CTL's is directed against the B27 antigens; and (3) "extra reactions" on B27-negative cells are caused by a subset(s) of CTL's recognizing unknown antigens shared between stimulator and target cells. CTL's raised by stimulation of HLA-B27-negative responder cells with B27-positive cells of either sub-type lysed all B27-positive target cells indiscriminately. In cold-target inhibition, however, B27-positive cells, carrying the sub-type of B27 different from that of the stimulator, could not inhibit the lysis of cells bearing the stimulator sub-type of B27. This indicates the activation, in B27-negative responders, of at least two different groups of CTL clones, one directed against shared determinants of HLA-B27, and one against the HLA-B27 sub-type. Heterogeneity of the HLA-B27 antigen may have implications for studies on the well-known association between this antigen and various diseases.

Subtypes of HLA-B27 detected by cytotoxic T lymphocytes and their role in self-recognition

In the present study cytotoxic T lymphocytes were generated in MLC of lymphocytes from two unrelated HLA-A, B, C-identical, B27-positive, but D/DR-different, individuals. These CTL were shown to detect subtypes of HLA-B27. CTL specific for influenza virus lysed infected target cells matched for HLA-B27 only when they shared the same subtype. This indicates that the two subtypes of HLA-B27 detected by CTL function also as distinct elements in a self-restricted CTL response. Both subtypes were found among patients with ankylosing spondylitis.