Reactive arthritis (ReA)

Certain bacterial infections have unequivocally been demonstrated to be causative of ReA in a percentage of individuals, who are exposed to these organisms. The percentage of individuals affected is termed the ‘attack rate’. Chlamydia trachomatis (Ct), Salmonella, Shigella , Campylobacter and Yersinia are all known triggers of ReA . C. trachomatis is thought to be the most common bacteria to cause ReA . There is mounting evidence to suggest that Chlamydia (Chlamydiophila) pneumoniae (Cpn) is another, albeit less frequent, causative agent . The pathophysiology of ReA represents the classic interplay between environment and genetics. The causative organisms are Gram-negative with lipopolysaccharide (LPS) as a key component to their cell wall. Elicitation of this disease unquestionably results largely from interaction at the genetic level between host and pathogen. While we are beginning to understand aspects of the bacterial products and their effects during infection, we have very little understanding of the host’s response to those products.

Polymerase chain reaction (PCR) technology has documented that all of definite bacterial triggers of ReA, or their bacterial products, are present in the synovial tissue or fluid of patients with ReA. This has been demonstrated in multiple studies involving many different laboratories, often years after the initial infection . This somewhat remarkable, and under-appreciated, fact, demonstrates that the entire bacteria or bacterial components traffic from the initial site of infection to the joints, and possibly other affected tissues of patients with ReA. While patients with ReA do not have systemic infection in the traditional sense, these microbes disseminate with particular tropism for certain organs. In spite of the fact that all of these triggering microbes disseminate from their initial site of infection, there are important differences. The most important distinction involves viable bacterial persistence. In patients with post-chlamydial ReA, the disseminated chlamydial organisms enter a state of persistence, whereas the post-enteric organisms have generally not demonstrated viability in synovial samples. Besides the definite bacterial triggers of ReA, there are many other bacteria that have been implicated as potential causes. These include Ureaplasma urealyticum, Clostridium difficile, Escherichia coli , intravesicular Bacillus Calmette-Guerin and various intestinal parasites . This article will focus on the definitive triggers, that is, chlamydiae , Salmonella, Shigella, Campylobacter and Yersinia .

Chlamydiae are Gram-negative, obligate intracellular organisms. Synovial tissue analyses from patients affected with post-chlamydial ReA have shown that the persistent form of chlamydiae exist in a morphologically aberrant, but metabolically active viable state . The pattern of gene expression is attenuated and significantly different than that seen during normal active infection. For example, during persistence of Ct expression of the major outer membrane protein ( omp1 ) gene, several genes required for the cell division process are severely down-regulated. This is coupled with differential regulation of the three paralogue genes specifying C. trachomatis heat shock proteins (HSP)-60 (Ct110, Ct604 and Ct755) . In vivo , synovial cells that are infected with persistent Ct display moderate up-regulation of Ct110, significant up-regulation of Ct604 and down-regulation of Ct755. The principal host cell in vivo for persistent synovial Ct is the macrophage.

It is important to remember that the above findings apply only to Ct and not Cpn. There are differences even within the Chlamydia genus. Differences in cytokine and chemokine mRNA profiles have been demonstrated in human synovial tissue chronically infected with Ct and Cpn . Further, a detailed gene expression profile of intracellular viable Ct and Cpn revealed different transcriptional response, which was no longer present when the organisms were ultraviolet (UV)-inactivated . These differences suggest that more than innate immunity is involved, and may explain the apparent higher risk of ReA with Ct, as opposed to Cpn. The different chlamydial species could also have an additive or synergistic effect in determining ReA attack rate or incidence. Given that Cpn is a common infection, previous exposure to Cpn could have an effect on a subsequent response to a Ct infection, or vice versa. Reports have demonstrated that prior Cpn infection primes a Th1 T-cell response to Ct antigens . In spite of these differences in the persistent state of synovial-based Ct compared with Cpn, the clinical sequelae that result are felt to be indistinct.

The classic form of ReA consists of the clinical triad of symptoms including the synovium, urethra and conjunctiva; however, the majority of patients do not present this with classic triad . Other organ systems are also frequently involved, especially the skin. Two characteristic rashes that are part of ReA include keratoderma blenorrhagicum and circinate balanitis. Recent data have shown that chlamydial organisms can also traffic to the skin in patients with suspected keratoderma blenorrhagicum . Therefore, it is logical to speculate that the persistent chlamydiae might traffic to all involved organ systems. Taken together, these data disprove the theory that post-chlamydial arthritis represents a sterile arthritis; these findings also challenge the notion that the role of the bacteria is simply one of triggering an auto-immune response and does not participate in perpetuating disease.

Other recent data relating to Chlamydia -induced ReA force us to reconsider our traditional paradigms. It is known that about 5% of subjects exposed to Ct will develop ReA . Since these are genital infections, it was logical to assume that the genital strains of C. trachomatis were responsible for triggering ReA. Nearly all efforts that have been spent in trying to determine why only a small fraction of individuals exposed to this organism who developed ReA have focussed on the host. However, it is important to remember that there are several serovars of C. trachomatis , specifically serovars A through K. Serovars A, B and C are ocular (trachoma) serovars and the remainders (serovars D through K) are genital. Remarkably, a recent study analysing the chlamydial serovars of 36 subjects with known C. trachomatis -induced ReA demonstrated that all 36 synovial tissue samples were positive for the ocular serovars (2 A, 1 B, and 33 C serovars), not the genital serovars . It is known that genital infection with the ocular strains do occur, but are rare . The infrequent rate of genital infections with the ocular strain might explain the low attack rate of ReA in patients with acute chlamydial infections. Whether, and, if so, by what means, the A, B and C serovars are, in fact, uniquely arthritogenic, as opposed to genital serovars, remains to be established.

The findings relating to the apparent arthritogenic propensity of the ocular strain of C. trachomatis suggests that a great deal of emphasis has been placed on host response in terms of ReA disease genesis. However, genetic susceptibility clearly plays, at least, a partial role in ReA pathophysiology. An extensive review of all available data (including post-enteric ReA) suggests that the prevalence of HLA-B27 in affected individuals ranges from ∼30% to 50% . Yet, some studies with post-enteric ReA suggest that HLA-B27 has no role in disease predilection, particularly with Campylobacter ; this further accentuates the role of the triggering bacteria. The possibility has been raised that HLA-B27 might be more important in determining disease- severity, rather than disease- predilection . Therefore, this HLA antigen might serve as a diagnostic bias, rather than a true susceptibility locus. It does appear, however, that this HLA antigen plays a role in those individuals who develop chronic disease .

Salmonella is a rod-shaped, motile bacterium widespread in animals and environmental sources. It is the most frequently studied enteric bacteria associated with ReA. The attack rate of Salmonella -induced ReA has ranged between 6% and 30% in different studies . As with the other causative organisms, efforts have been made to detect Salmonella in synovial tissue or fluid. Salmonella bacterial degradation products have been detected in the synovial fluid from patients with Salmonella -induced ReA, but, unlike chlamydiae , no viable organisms have been detected . In other studies, synovial fibroblasts have been infected with Salmonella and analysed by electron microscopy and fluorescence in situ hybridisation . Mirroring the in vivo results from synovial samples, intracellular bacterial replication was followed by degradation, leading to ‘ghosts’ possessing LPS, but not DNA. However, some bacteria survived for more than 2 weeks. This suggests synovial-based Salmonella organisms have the ability to exist for a prolonged period of time. Since the viable bacteria do not appear to persist in synovial tissues, as is the case with chlamydiae , efforts have been made in attempting to locate certain portions of the Salmonella bacteria that might be particularly arthritogenic. Preliminary studies indicate that certain proteins in the outer membrane protein (Omp) might cause ReA .

After salmonellosis, individuals of Caucasian descent may be more likely than those of Asian descent to develop ReA ; this indicates that genetics play a role in disease predilection. However, children appear to be less susceptible to developing ReA after a Salmonella infection compared with adults ; this suggests that repeat exposures might increase the likelihood of disease development. A large study in Denmark comparing the different enteric pathogens known to cause ReA suggested that Salmonella was the second most common triggering infection (after Campylobacter ) and the second most arthritogenic, after Yersinia .

Investigations into host genetics and host response have also been performed with Salmonella -induced ReA. There have been large outbreaks of Salmonella typhimurium and Salmonella enteritidis with rheumatological follow-up of affected individuals. The HLA-B27 prevalence has ranged from 17% to 50% in these subjects . Since all of the triggering bacteria of ReA, including Salmonella , are Gram-negative organisms with a LPS component to their cell wall, it would be logical to assume that innate immunity occurs through Toll-like receptor 4 (TLR-4). Surprisingly, a recent study of patients with ReA after salmonellosis revealed that generic variants of TLR-2, but not TLR-4, were associated with acute ReA and extra-articular features . Interestingly, TLR-2 recognises many infectious agents and substances, one of which is peptidoglycan. Some of the Salmonella cell wall Omp proteins thought to be immunogenic, cited above, contain peptidoglycan .

All four of the species of Shigella ( Shigella flexneri, Shigella dysenteriae, Shigella sonnei and Shigella boydii ) can cause ReA. Shigella is a very motile organism with the ability to invade human enterocytes, lyse intracellular vacuoles to enter the cytoplasm and move from cell to cell. In 1944, Shigella was the first bacteria to be directly implicated as a cause of ReA . Interestingly, Shigella is phylogenetically indistinguishable from E. coli sharing all but 175 of 3235 open reading frames , and recent reports have suggested that E. coli might be an infrequent cause of ReA . The overall attack rate of Shigella -induced ReA was reported to be 7–9% in one study . This is also the only large outbreak study assessing the HLA-B27 prevalence in affected individuals. Thirty-six percent of the subjects who developed ReA were HLA-B27 positive.

Similar synovial sample analyses have been performed in patients with Shigella -induced ReA revealing bacterial DNA from Shigella . In contrast to studies with Chlamydiae , but similar to findings with Salmonella , there have been no studies to detect viable organisms, only bacterial fragments .

Campylobacter is another Gram-negative motile bacterium; it is a very common cause of enteric infections. Campylobacter jejuni is the main cause of bacterial food-borne disease in many developed countries . In spite of its frequency, it appears to be somewhat less arthritogenic with an attack rate of 1–5%. It also frequently causes a polyarthritis, in addition to an oligoathritis, and the symptoms are generally mild . Data also suggest patients with Campylobacter -induced ReA less often develop inflammatory back pain than with other types of ReA . Even less is known about the host genetics in patients with Campylobacter -induced ReA as HLA-B27 does not appear to influence disease predilection . Recent data suggests, however, that single nucleotide polymorphisms in the IFNG gene was an independent risk factor for developing ReA after Campylobacter (and Salmonella ) infections . Similar to Salmonella -induced ReA, children appear to be at lower risk of developing ReA after a Campylobacter infection, again suggesting the possibility that repeat exposures might increase the likelihood of disease development. Studies have been performed looking for specific isolates of Campylobacter that might be particularly arthritogenic. Although specific sequence types have been associated with other post-infectious sequelae, namely Guillian–Barre syndrome, no specific sequences were associated with ReA . By contrast, another recent study that genotyped Campylobacter strains found that there was an association with Class A lipo-oligosaccharide gene locus and ReA .

Although Yersinia infections are not as common as some of the other enteric pathogens, some data suggest that Yersinia is particularly arthritogenic. A study in Denmark suggests that it was the most-likely organism to cause ReA with an attack rate of 23% . Other studies have suggested an attack rate of 12% . As with the majority of the other known triggering microbes, two studies have attempted to localise Yersinia in the synovial tissue or fluid of affected individuals. While both studies demonstrated that Yersinia does indeed traffic to the joints , one suggested that these Yersinae are metabolically active and the other study only demonstrated bacterial degradation products . The possibility also exists of cross-reactivity between Yersinia outer membrane proteins and other organisms known to cause chronic arthritis, specifically Borellia . This, too, suggests repeat exposures can lend to disease development. In another study, Yersinia HSP60 CD4+ T-cell lines, derived from patients with Yersinia -induced ReA, were highly antigen-specific revealing single immunodominant peptide epitopes . In spite of this specificity, some cell lines displayed cross-reactivity with HSP60 from other organisms, in this case human HSP60, again suggesting that different aetiological starting points might lead to similar clinical manifestations.

Regarding the concept of bacteria or bacterial fragments being demonstrated in the synovial tissue of patients with ReA, many of the same organisms, and others, have been demonstrated in the synovial tissue of patients with other diseases, namely osteoarthritis, or even asymptomatic controls . These data have called into question the pathologic importance of these findings. However, the prevalence of these bacteria in the synovial tissue in other conditions is significantly less compared to synovial tissue of patients with ReA . Although this highlights the importance of host genetic variability and host tolerance, it should also underscore the recent findings regarding important potential arthritogenic differences of the various bacterial serovars, sequence complexes, epitopes, etc. These subtle differences, even within the same genus or species, have not been investigated in similar studies.

A review of the role that bacteria play in the pathophysiology of ReA demonstrates several salient features. First, ReA is unquestionably triggered by a set of known bacteria; therefore, it can be viewed as the prototype for the possible role that bacteria play in all of the spondyloarthritides. Second, these causative bacteria disseminate from their initial site of infection, demonstrating tropism for certain organs, namely synovial tissue; in the case of chlamydiae , these bacteria exist in a persistently viable state. Third, it appears that too much emphasis has been placed on host genetics and host response. While host mechanics is important, the bacteria themselves offer enough heterogeneity, even within the same genus or species, to help explain a large portion of disease predilection. Finally, by demonstrating that several different bacteria with different pathophysiologic features can lead to the same phenotypic disease, it suggests a need to broaden our concept of what role bacteria might play in disease aetiology and, perhaps more importantly, in disease perpetuation.