- Top of page
- Materials and Methods
Purpose: Endotoxins have been proved to be responsible for acute anterior uveitis (AAU) in animals in a well-established experimental model of endotoxin-induced uveitis (EIU). The purpose of our study was the detection of antibodies against endotoxins of selected enterobacteria in the serum of patients with idiopathic AAU and searching for correlations between the levels of these antibodies and the presence of HLA-B27 antigen as well as characteristic signs of EIU such as bilaterality and the absence of spontaneous recurrences of the disease.
Methods: Reactions of serum IgG antibodies with lipopolysaccharides (LPSs) of Escherichia coli O1, E. coli O10, E. coli O111, E. coli J5, and Klebsiella pneumoniae O3 were determined for 60 patients with idiopathic AAU and 40 healthy volunteers. The presence of HLA-B27 antigen in patients was determined. Documentation of the frequency of recurrences of AAU during a follow-up period of 8 years was collected.
Results: We have observed that the sera of patients with a first attack of AAU reacted stronger with the LPS of K. pneumoniae O3 than the sera of patients with relapse of the disease. Patients with bilateral AAU had markedly higher levels of antibodies against four of the five used LPSs than patients with one eye involved. A multiply comparison showed higher levels of IgG reacting with LPS of E. coli O111 in patients with bilateral eye inflammation admitted with the first attack of AAU comparing to controls. The incidence of recurrent form of AAU was significantly increased in HLA-B27-positive patients compared to HLA-B27-negative patients. However, we found in HLA-B27 carriers that those with the bilateral form of AAU had over three times smaller risk of recurrence and showed stronger immunization by endotoxins than patients with unilateral inflammation.
Conclusion: Our results suggest a potential role of endotoxins in the aetiology of the nonrecurrent bilateral form of AAU. We suggest that not only HLA-B27 status but also determination of number of involved eyes may be useful to assess the risk of recurrence of the idiopathic AAU.
- Top of page
- Materials and Methods
Many factors responsible for aetiopathogenesis of acute anterior uveitis (AAU) have been already identified but AAU of unknown aetiology still constitutes an important problem in the clinical practice. It is estimated that in as many as 60–80% of patients, the cause of the illness is not diagnosed and half of the cases are idiopathic AAU (Curi et al. 2005). Numerous investigations indicate a significant role of HLA-B27 antigen as a factor predisposing to AAU (Chang et al. 2005; Careless & Inman 1995). According to Feltkamp, the presence of HLA B27 increases eight times the relative risk of acquiring AAU (Feltkamp 1990). Some authors stress that having the antigen may be connected with a more severe course of the disease and a more serious prognosis, which in such cases is often characterized by unilateral recurrences, and it is even suggested to treat so-called HLA-B27-positive AAU as a separate clinical entity (Chang et al. 2005; Rothova et al. 1987; Power et al. 1998).
One of the common features of those bacteria is the presence of endotoxin lipopolysaccharide (LPS), which is the main surface antigen of Gram-negative bacteria. Although its role in AAU remains unclear, endotoxins have been proved to be responsible for AAU in animals (Dyster-Aas et al. 1969). In an experimental animal model endotoxin-induced uveitis (EIU), symptoms of AAU could be induced by LPS isolated from different bacterial species, e.g. Escherichia, Shigella, and Salmonella. Subcutaneous injection of endotoxin into a rat’s extremity resulted in the accumulation of proteins and inflammatory cells in the aqueous humour of the eye after 4–8 hr. (Rosenbaum et al. 1980; Okumura & Mochizuki 1988; Herbort et al. 1988; Chen & Gordon 2005) This effect reaches a maximum after 24–48 hr and gradually decreases. The cellular infiltration usually disappears in a week. The mechanism of this phenomenon is very complex – the injection of LPS induces an increase or reduction in the expression of over 1900 genes in the iris and ciliary body, including those for IL-1β, IL-6, the RANTES chemokines, and induced nitric oxide synthase and for the transcription factors Jun-b, c-Fos, and c-Jun. (Ohta et al. 2005) One of the characteristic features of EIU is simultaneous inflammation in both eyeballs. A meta-analysis of the results of different experiments (unpublished data) in which we estimated clinical signs of eye inflammation in rats (n = 56) 24 hr after induction of EIU (subcutaneous injection of 50 μg of LPS of H. alvei 981 into the pad of the rat’s hind leg) revealed that in 87.5% of cases clinical signs of inflammation occurred in both eyes, in 7.1% in one eye, and in 5.4% there were no signs of EIU in biomicroscopic evaluation. Even considering the absence of clinical symptoms of EIU, the number of cells in the aqueous humour of the rats treated with LPS was significantly greater than in the control group. It is interesting that recurrences are not observed in EIU, in contrast with experimental melanin-induced uveitis, in which spontaneous recurrences occur in some of the animals (Broekhuyse et al. 1991, 1992).
According to some authors, bacterial infections, including Gram-negative ones, could be one of the factors engaged in the initiation of AAU in predisposed persons (Huhtinen et al. 2002a,b). The expression of toll-like receptor 4 (TLR4) on the endothelial cells of the iris and TLR-4 and CD14 on the ciliary body cells of the human eye has been presented.(Brito et al. 2004) TLR-4 and CD14 were also detected on antigen-presenting cells that reside in the uvea.(Chang et al. 2004) These receptors are crucial in the initiation of the inflammatory reaction to endotoxins. This supports on the molecular level that the iris and ciliary body in humans, similarly to animal models, could respond directly to the presence of endotoxin, using the mechanism of innate immunological response. That is why research is directed towards elements which could combine the experimental model of AAU with clinical observations to prove the hypothesis about the possibility of the induction of uveitis in humans by endotoxins. So far, patients with AAU possibly related to the contact with endotoxin have not been characterized.
The purpose of our study was the detection of antibodies against endotoxins of selected bacterial species in the serum of patients with idiopathic AAU and searching for correlations between the levels of these antibodies and the presence of HLA-B27 antigen as well as characteristic signs of EIU such as bilaterality and the absence of spontaneous recurrences of the disease.
- Top of page
- Materials and Methods
The detailed characteristics of the study group of patients with idiopathic AAU (n = 60) and the control group (n = 40) are illustrated in Table 1. HLA-B27 antigen was present in the majority of patients (65%). Patients who were admitted with the first attack of the disease constituted 60% of all cases, and bilateral AAU was diagnosed in 30%.
Table 1. Characteristic features of examined patients and control group.
|Patients with idiopathic AAU|
|Number of patients||60|
| Female||29 (48.3%)|
| Male||31 (51.7%)|
| Mean ± SD||45 ± 18|
|Presence of HLA-B27|
| Patients HLA-B27-positive||39 (65.0%)|
| Patients HLA-B27-negative||21 (35.0%)|
|Number of attacks|
| Patients with 1 attack||36 (60.0%)|
| Patients with 2 attacks||15 (25.0%)|
| Patients with 3 attacks|| 7 (11.7%)|
| Patients with 4 or more attacks|| 2 (3.3%)|
|Form of disease|
| Patients with unilateral AAU||42 (70.0%)|
| patients with bilateral AAU||18 (30.0%)|
|Number of people||40|
| Female||14 (35%)|
| Male||26 (65%)|
| Mean ± SD||39 ± 11|
In Table 2, the percentages of recurrences (the second or subsequent attack of the disease) are presented depending on different factors. The percentage of patients with AAU recurrence among the HLA-B27-positive patients was 61.5%, while in all patients who were HLA-B27-negative AAU occurred for first time (p < 0.001). The frequency of noted recurrences in all patients with bilateral AAU was lower than in all patients presenting unilateral AAU, but this was not statistically significant (OR: 0.31, 95% CI: 0.09–1.08, n.s.). However, in the subgroup of HLA-B27-positive patients, this difference was significantly lower (OR: 0.07, 95% CI: 0.02–0.31, p < 0.001).
Table 2. Occurrence of the relapses of idiopathic AAU in patients depending on HLA-B27 carriage and AAU laterality during a follow-up period of 8 years.
|Clinical feature||Percentage of patients admitted with relapse of AAU (n/total)||Fischer’s exact test||Percentage of patients with relapse of AAU during the last 8 years (n/total)||Fischer’s exact test|
| HLA-B27 positive||61.5% (24/39)||p < 0.001||51.3% (20/39)||p < 0.002|
| HLA-B27 negative||0.0% (0/21)||10.0% (2/20)|
| Unilateral||47.6% (20/42)||p = 0.088||41.5% (17/41)||p = 0.389|
| Bilateral||22.2% (4/18)||27.8% (5/18)|
|AAU laterality in patients with HLA-B27|
| HLA-B27 positive/unilateral||83.3% (20/24)||p < 0.001||70.8% (17/24)||p < 0.004|
| HLA-B27 positive/bilateral||26.7% (4/15)||20.0% (3/15)|
|AAU laterality in patients without HLA-B27|
| HLA-B27 negative/unilateral||0.0% (0/18)||p = 1.000||0.0% (0/17)||p < 0.02|
| HLA-B27 negative/bilateral||0.0% (0/3)||66.7% (2/3)|
The analysis of AAU recurrences during following eight years proved the previously observed dependencies: the risk of recurrence of AAU among HLA-B27 patients was significantly higher in patients with unilateral disease then in patients with bilateral disease (RR: 0.28, 95%CI: 0.13–0.63, p < 0.004).
The results of the reaction in ELISA (mean OD) of IgG antibodies in the serum of the patient and control groups with selected LPSs are presented in Figs 1 and 2. In both the patient and healthy control groups (Fig. 1), the greatest values of mean OD were observed for the LPS of E. coli O1 (0.851 ± 0.045 and 0.822 ± 0.063 respectively) and the lowest for the LPS of E. coli J5 (0.234 ± 0.033 and 0.254 ± 0.042). Comparison of the mean OD values for the whole group of patients with the control group revealed a significantly stronger reaction of IgG with the LPS of E. coli O111. When patients with a first attack of AAU were compared to patients with the recurrent form of the disease, a significant result was a weaker reaction of serum with the LPS of K. pneumoniae O3 (Fig. 1). A multiply comparison procedure (Fig. 2) revealed markedly higher levels of IgG reacting with LPS of E. coli O111 in patients with bilateral eye inflammation admitted with the first attack of AAU when compared with controls. The presence of HLA-B27 had no influence on differences in the reactions of patients’ serum with the LPS, while the serum of patients with bilateral AAU reacted significantly stronger even with the four LPSs derived from E. coli O1, E. coli O10, E. coli O111, and K. pneumoniae O3 (Fig. 1). In the subgroup of HLA-B27-positive persons, patients with bilateral AAU had significantly higher levels of antibodies against LPS E. coli O10, E. coli O111, E. coli J5, and K. pneumoniae O3 than the HLA-B27-positive patients with the unilateral form of the disease. In the subgroup of HLA-B27-negative patients, there were no statistically important differences in the levels of antibodies regarding the unilateral or bilateral form.
Figure 1. The comparison of the levels of antibodies against the lipopolysaccharides of E. coli O1, E. coli O10, E. coli O111, E. coli O111 and E. coli J5, and K. pneumoniae O3 in patients with idiopathic AAU and healthy controls. Average values of extinction (±SE) are presented in the ELISA test. *p < 0.05 (Mann–Whitney U-test).
Download figure to PowerPoint
Figure 2. The multiply comparison of the levels of antibodies against the lipopolysaccharides of E. coli O1, E. coli O10, E. coli O111, E. coli O111 and E. coli J5, and K. pneumoniae O3 in patients and healthy controls considering the laterality and number of attacks of idiopathic AAU. Average values of extinction (±SE) are presented in the ELISA test. *p < 0.05 (Kruskal–Wallis analysis of ranks).
Download figure to PowerPoint
The results obtained for serum in dilutions 1:32 and 1:128 were comparable to those described earlier, which is why they are not presented.
- Top of page
- Materials and Methods
Most of the results suggesting a bacterial aetiology of AU are based on investigations of the presence of specific antibodies against bacterial antigens because it is hardly possible to identify bacteria directly among patients with AU without reactive arthritis. In our study, we chose the LPSs isolated from five bacterial species to search for antibodies against the O-antigen, which is present on the surface of Gram-negative bacteria, in the sera of patients and healthy volunteers. The LPSs of E. coli O1, E. coli O10, and E. coli O111 represent three different types of core, i.e. R1, R2, and R3, respectively, which do not react in the cross-tests. (Ługowski et al. 1996) The LPS of E. coli J5 (a rough mutant of E. coli O111:B4 of the type Rc) is a strain which is not pathogenic to humans, and the antibodies against this LPS are characterized by wide cross-reactivity. (Łukasiewicz et al. 2002) We chose the LPS of K. pneumoniae O3 with regard to some authors’ suggestions that infection with Klebsiella plays a role in AAU.(Mäki-Ikola et al. 1995) Previously we showed that the LPSs of K. pneumoniae O3 and E. coli J5 can induce EIU in animal models (M. Misiuk-Hojło, R. Międzybrodzki & C. Ługowski, unpublished). Significantly higher levels of IgG, IgM, and IgA antibodies against antigenous extracts from E. coli in patients with AAU were observed by Szanto et al. (1991), who examined 114 patients with AAU (22% of whom were patients with idiopathic AAU). Anti-E. coli LPS antibodies in patients with iritis and spondyloarthropathy were evaluated by Mäki-Ikola et al. (1995), who observed higher levels of these antibodies of the IgA class if spondyloarthropathy was accompanied by iritis in comparison with a group without eye inflammation. In our study, we have described the strongest reaction of serum for the LPS of E. coli O1 (core type R1), in patients and healthy persons. E. coli bacteria with of this type LPS are most often isolated from patients with bacteremia (68%) and the faeces of healthy humans (63%). (Appelmelk et al. 1994; Gibbs et al. 2004) However, general comparison with the control group revealed significantly higher levels of IgG against the LPS of E. coli O111 in patients with AAU, which could suggest an infectious background of the disease or a direct role of endotoxins in AAU.
Acute anterior uveitis concomitant with spondyloarthropathies is characterized by a high frequency of HLA-B27 in patients with recurrence of attacks and unilaterality.(Chang et al. 2005) Until now, infection was considered to be a causative factor in this group of patients with AAU. According to one hypothesis, the immunological cells of HLA-B27-positive patients react in an atypical manner in the inflammatory process. (Careless & Inman 1995) This seems to be supported by the observation that monocytes after transfection with HLA-B27 are activated more strongly by LPS than nontransfected cells. (Penttinen et al. 2002) The hypothesis of molecular mimicry, primarily described for spondyloarthropathy, assumes the presence of common determinants between HLA-B27 and peptide sequences on the surface of some bacteria. (Careless & Inman 1995) Such phenomenon was noted, for example, in K. pneumoniae. (Feltkamp 1990; Geczy & Yap 1979) It may be a cause of the cross-reaction (involving T lymphocytes) which, in response to bacterial antigens, may stimulate the production of antibodies that cross-react with HLA-B27 antigen of the host. Theoretically, this process could also promote the long-term survival of pathogens in host tissues because of cross-tolerance between the bacterial antigens and the host’s HLA system. (Feltkamp 1990) According to another hypothesis, HLA-B27 acts like an alternative receptor which specifically reacts with some bacterial antigens and in this way modulates the response to infection. (Careless & Inman 1995) Analysing our data, we found a tendency towards slightly greater levels of anti-LPS antibodies in HLA-B27-positive patients than in HLA-B27-negative ones, but the observed differences were not statistically significant. Similarly to other researchers, we noted the presence of HLA-B27 antigen in most of the patients. (Linssen et al. 1991; Brewerton et al. 1973; Monnet et al. 2004) Moreover, we have proved that the sera of patients with the first attack of AAU reacted more strongly with the LPS than the sera of patients with relapse of the disease. It also appeared that patients with bilateral involvement, another characteristic feature of EIU, had markedly higher levels of antibodies against even four of the five used LPSs, i.e. those isolated from E. coli O1, E. coli O10, E. coli O111, and K. pneumoniae O3, than patients with inflammation in one eye. Such strong reaction of sera with a whole ‘panel’ of antigens could be related to the presence of antibodies in serum against other LPSs than those evaluated and their cross-reactions. The results of multiply comparison suggested a potential involvement of LPS in the aetiology of the nonrecurrent bilateral form of AAU.
The analysis of the frequency of recurrences did not reveal that uni- or bilaterality of the disease itself is significantly associated with the risk of the relapse of idiopathic AAU. However, a detailed analysis revealed that HLA-B27-positive persons with the bilateral form of the disease had a three times smaller risk of recurrence than the HLA-B27-positive patients with unilateral inflammation. In the HLA-B27-negative patients, the relation between recurrences and the form of the disease was similar to the observations by Natkunarajah et al. (2007) for all patients with AAU, irrespective of the aetiology of the disease, where a bilateral process promoted relapses. These clinical observations could allow better identification of patients with a higher risk of recurrence of idiopathic AAU. Interestingly, in the group of HLA-B27-positive patients with bilateral AAU in whom relapses occur more rarely, we found stronger immunization by endotoxins than in HLA-B27-carriers with unilateral AAU. As already mentioned, rare incidents of recurrences of AAU and a high percentage of simultaneous bilateral inflammation are characteristic features of anterior uveitis induced in animals by endotoxin injection. This suggests that it would be reasonable to search for an infectious background of the disease in HLA-B27 carriers with bilateral idiopathic AAU. Confirmation of this hypothesis in future and introducing diagnostic tests for measuring the anti-endotoxin antibody level may have a prognostic value in AAU.