Etiology and Pathogenesis

Introduction

The etiopathogenesis of ankylosing spondylitis (AS) is not fully understood. However, there is strong evidence indicating that genetic factors play a pivotal role in disease susceptibility, and that interaction with epigenetic and environmental factors results in immune-mediated events to release proinflammatory cytokines that lead to disease. Twin studies have shown that nearly 90% of the risk of developing AS is determined genetically, indicating the highly heritable nature of this disease. Moreover, disease activity, functional impairment, radiographic severity and age of symptom onset all show significant heritability in AS. Recurrence rate of the disease among first-degree relatives is quite high, and there is a substantially higher transmission rate to children from mothers than fathers.

Genetic Factors

Genome wide association studies (GWAS) in AS-only analyses have, to date, identified 113 genetic susceptibility loci for AS, 39 of them reaching genome-wide significance. However, they only contribute to 28% of the total genetic risk, of which a little over 20% is contributed by the major histocompatibility complex (MHC), primarily by the HLA-B27 gene and the remaining 7.4% by non-MHC loci, such as ERAP1, ERAP2, IL23R and MEFV. The identified risk genes and their putative functions are listed in Table 4-1.

Enlarge 

Role of HLA-B27 and Other MHC Genes

HLA-B27 is a perfectly normal gene that is distributed worldwide but with variable prevalence. It resides in the MHC locus on the short arm of chromosome 6 (6p21 region) and is the primary disease susceptibility gene for AS, first reported 48 years ago. But it is not a prerequisite for AS, as the disease also affects individuals who do not possess this gene. Moreover, among those who possess this gene, only 1% to 5% develop AS and they most likely possess additional disease predisposing genes. But HLA-B27 homozygous individuals are up to three times more likely to get the disease than heterozygotes. A 35-year follow-up study of patients (probands) with axial spondyloarthritis (axSpA) diagnosed in 1985 and their first-degree relatives (FDR) observed a lifetime recurrence rate (RR) of 27.1% among HLA-B27(+) FDR, compared to 1.6% among HLA-B27(-) FDR, illustrating the relative importance of HLA-B27 among other genetic factors in the etiology of axSpA. The same study also revealed that the lifetime relative risk (RR) is much higher among children of probands with AS compared to children of probands with non-radiographic axial spondyloarthritis (nr-axSpA). Figure 4-1 shows a simplified diagram of the antigen-binding cleft of an HLA-Class I molecule, such as HLA-B27, with a 9 amino acid long bound peptide with its side chains that fit into the side pockets A to F.

HLA-B27 is a highly polymorphic gene with 332 known alleles based on nucleotide sequence differences, as of September 2020, although an overwhelming majority of them are extremely rare. Some of the genetic substitutions are silent, i.e., they do not cause amino acid changes in the translated protein structure, some occur in untranslated regions and are not coded, and a few other are null alleles that are not expressed. Therefore, this polymorphism can be encompassed by a numbering system—HLA-B*27:01 to HLA-B*27:235. Three old assignments, HLA-B*27:051, HLA-B*27:22 and B*27:215N have been withdrawn, when they were found to be identical to B*27:05:02, HLA-B*27:06 and B*27:05:02:01, respectively.

The currently known 224 transcribed proteins (HLA-B27 subtypes or allotypes) listed in Table 4-2. They differ by one or more amino acid substitutions, and show an extremely varied racial/ethnic prevalence throughout the world. HLA-B*27:05 (specifically the HLA-B*27:05:02 allele) is the most widely distributed and disease-associated subtype in the world (considered the “mother subtype”) and has been the subject of most studies. The other common disease-associated subtypes are HLA-B*27:02 (more common among Mediterranean populations) and HLA-B*27:04 (more common among the Chinese and some of the other southeast other Asian populations). HLA-B*27:01, HLA-B*27:03, HLA-B*27:07, HLA-B*27:08, HLA-B*27:10, HLA-B*27:11, HLA-B*27:13, HLA-B*27:14, HLA-B*27:15, HLA-B*27:18, HLA-B*27:19, HLA-B*27:23, HLA-B*27:24, HLA-B*27:25 and HLA-B*27:49 are also known to be disease associated, or at least one or more AS patients possessing these subtypes have been observed. Most of the remaining subtypes are either too rare or are too recently described to have been evaluated for disease presence or association.

There are also differences among some of the subtypes when it comes to disease association. HLA-B*27:05 and HLA-B*27:02 confer almost equal susceptibility to AS in Caucasian populations, but among Chinese populations, HLA-B*27:04 shows a stronger association than HLA-B*27:05. It is of extreme interest that HLA-B*27:06 (a common subtype in southeast Asia) and HLA-B*27:09 (a rare subtype and found primarily on the Italian island of Sardinia) seem to lack association with classical AS, although they differ from their closely related but disease-associated subtypes by only one or two amino acids (Figure 4-2 and Figure 4-3).

The precise explanation for the association of HLA-B27 with ankylosing spondylitis/axial spondyloarthritis (AS/axSpA) has defied researchers but several hypotheses have been proposed. An “arthritogenic peptide” hypothesis, based on the natural immunologic function of presenting antigenic peptides to restricted CD8+ cytotoxic T cells, suggests that this ability to bind such a peptide (or a set of peptides), either microbial or self-derived, results in disease from the cytotoxic T-cell response to such peptide(s) found only in affected sites. A “molecular mimicry” hypothesis proposes that HLA-B27 itself becomes autoantigenic due to its sequence homology with certain bacterial proteins.

As discussed later, accumulation of misfolded HLA-B27 in the endoplasmic reticulum (ER) can cause intracellular stress, unfolded protein response (UPR) and autophagy that lead to proinflammatory events. Moreover, modification of gut microbiome and intracellular microbial-handling of putative arthritogenic organisms by HLA-B27 that can result in an aberrant or impaired immune response in conjunction with an upregulated production of cytokines, such as IL-23, that leads to AS or related spondyloarthritis (SpA).

Our current knowledge about the effect of dysbiosis (homeostasis imbalance) in the gut microbiome has been reviewed by So and Tam. There may be alteration of intracellular handling of microbes due to HLA-B27 that may lead to a less effective elimination of microbes, such as salmonella, in conjunction with an upregulated production of cytokines. As shown in a recent model, salmonella replication is not enhanced by the expression of HLA-B27 alone, but by its misfolding, which influences the ER stress environment and provides a favorable environment for replication of S enterica. The presence or absence of HLA-B27 may also have an influence on the composition of the gut microbiome. It is possible that HLA-B27 may predispose to AS via more than one mechanism, or by different mechanisms in different patients, based on additional genetic, epigenetic and environmental factors.

The arthritogenic peptide hypothesis has been strongly revived by new data. There is now strong evidence for the presence of clonally expanded CD8+ T lymphocytes carrying TRBV9–CDR3–TRBJ2.3 motif encoding the variable domain of the TCRβ chain in HLA-B*27(+) patients with AS compared with healthy HLA-B*27(+) controls. However, the corresponding TCRα chains were not described. The strongest evidence to date for this hypothesis has now been published by Yang and colleagues. Using single-cell RNA sequencing, the authors isolated orphan TCRs expressing the disease-associated BV9–CDR3β-TRBJ2.3 motif from blood and synovial fluid samples obtained from HLA-B*27(+) patients with AS and from the blood and eye (aqueous humor) of those with acute anterior uveitis (AAU). They showed that these TCRβ chains consistently paired with TRAV21-positive TCRα chains. Moreover, bulk TCR sequencing of these samples showed at least a 10-fold enrichment of the disease-associated β-chain motif in synovial fluid or in ocular fluid compared with blood, respectively. This suggests that T cells expressing the disease associated TCRs have undergone clonal expansion in the inflamed tissues. They then used single-cell sequencing of the TCRs of these clonally expanded CD8+ T-cells and expressed them as recombinant TCRs to screen for potential HLA-B*27:05-restricted peptides that engage these disease-associated TCRs and were able to narrow down the millions of possibilities to a very short list of human and microbial proteins. Structural analysis revealed that a shared binding motif present in both self-antigens and microbial antigens engages the BV9–CDR3β TCRs.

These peptide-binding and presentation characteristics may be considered a hallmark of the HLA-B*27:05 subtype (associated with the disease) but are absent in the HLA-B*27:09 subtype (not associated with the disease); the two subtypes differ by only a single amino acid. This discovery is a major advance in solving the 50-year-old mystery of how HLA-B*27 contributes to the development of AS. These findings are expected to greatly accelerate identification of self- and microbial-derived peptides that instigate autoimmunity, which should aid the development of better targeted treatments. Additionally, the discovery of improved biomarkers, together with the emerging technology of electric field molecular fingerprinting, is expected to significantly improve early diagnosis. A ground-breaking paper published in 2023 demonstrated a remarkable effectiveness of targeting and eliminating disease-causing T cells in a HLA-B*27 patients with AS. This represents a true paradigm shift and bolsters the potential for preventing AS in individuals with high-risk genetic variants.

Studies beginning as far back as 45 years ago have indicated genotypic influence of HLA-B27 on clinical phenotype of AS, exemplified by the difference between HLA-B27(+) and HLA-B27(-) patients. Although there are many similarities, HLA-B27(+) patients with AS exhibit the following differentiating features:

• Younger age of onset
• Earlier age at diagnosis
• More acute anterior uveitis
• Greater familial occurrence
• A better clinical response to TNF inhibitors
• Less often associated psoriasis, UC and CD
• Genetic associations are similar, but association with ERAP1 is present only in HLA-B27(+) patients.

The above-mentioned 35-year follow-up study of axSpA patients and their first-degree relative (FDR) also observed an increased prevalence of discomfort of possibly inflammatory origin among “healthy” (never diagnosed) HLA-B27(+) relatives, with up to 25% experiencing at least one symptom, compared to <9% among “healthy” HLA-B27(-) relatives. This finding suggests an association of HLA-B27 with cryptic (possibly undiagnosed) axSpA. Another analysis of the same cohort confirmed a strong association of HLA-B27 with radiographic sacroiliitis, particularly in the male probands. A review of 60 worldwide studies revealed significant inter-regional heterogeneity in the prevalence of individual clinical manifestations of axSpA; for example, peripheral arthritis was relatively more common in North and South America and South-East Asia, while inflammatory bowel disease was comparatively more common in North Africa and the Middle East. One possible explanation of these divergent patterns of clinical phenotypes is inter-regional heterogeneity in the prevalence of HLA-B27 and other genetic markers, although more research and higher quality individual studies will be required to more fully elucidate the cause.

A study of data from a 35-year Swiss AS follow-up study and UK Biobank data for patients of European ancestry revealed that patients with AS in the long-term follow-up cohort who were positive for HLA-B27 had a shortened life expectancy compared to the general Swiss population. Interestingly, patients with nr-axSpA from the same cohort who were positive for HLA-B27 did not show increased mortality, nor did an increased mortality signal emerge among patients of European ancestry in the UK Biobank.

Other genes, including, HLA-B40, HLA-B47, HLA-B51, HLA-A2 and DRB1*0103 also show some increased risk for AS, whereas HLA-B7 and HLA-B57 may have some protective effect.

Enlarge 

Enlarge 

Enlarge 

Enlarge 

Role of Non-MHC Genes

As mentioned earlier, 7.4% genetic risk for AS is contributed by non-MHC loci, such as ERAP1, ERAP2, IL23R and MEFV. (Table 4-1). However, it should be emphasized that a small contribution of a risk locus may have substantial functional importance. For example, the TNFRSF1A polymorphism explains 0.06% of the genetic risk for AS, and yet tumor necrosis factor (TNF) inhibition therapy is highly effective in the treatment of this disease. The identified AS susceptibility loci are involved in several immunomodulatory pathways that are implicated in antigen presentation, IL-23R signaling, TNF-α/NF-κB signaling, lymphocyte development and activation, IL-1 cluster genes, G-protein coupled receptor and IL-17/IL-22 mediated immune response. Not surprisingly, some of them also confer susceptibility to Crohn’s disease (CD) and ulcerative colitis (UC) and to a much less extent, psoriasis.

ERAP1 and ERAP2

ERAP1, ERAP2 and NPEPPS are also associated with AS; and in fact, ERAP1 was the second strongest gene (after HLA-B27) to show a direct role in enhancing risk for AS. ERAP1 association is only observed in HLA-B27(+) and HLA-B40(+) patients with AS, but ERAP2 is associated with AS in HLA-B27(+) and HLA-B27(-) patients. There appears to be a genetic as well as a functional interaction or epistasis between ERAP1 with HLA-B27. The aminopeptidase ERAP1 in the ER trims peptides to the optimal length for HLA-B27 to present them to the cell surface via the protein loading complex (PLC) function (Figure 4-4 and Figure 4-5).

ERAP1 gene polymorphism results in variation of its enzymatic function in the ER that leads to production of oligomeric peptides with low affinity to bind to disease associated HLA-B27 subtypes (and HLA-B40) in the ER-derived cytoplasmic vesicles. Accumulation of these misfolded free heavy chains (FHC) along with β2m and ER chaperones causes intracellular stress, UPR and autophagy, resulting in HLA-B27 pathogenicity via this noncanonical mechanism. In addition, the abnormal HLA-B27 peptide complexes get presented on cell surface via the PLC function and stimulate CD8+ T cells. Moreover, the free heavy chains (FHC) thus presented on cell surface can form homodimers that are amenable to recognition by CD4+ T cells or NK cells via the killer immunoglobulin-like receptors (KIR) that are also present on the membranes of T-helper 17 cells (Th17) and myeloid cells. This promotes release of proinflammatory cytokines, such as TNFα, IL-1, IL-6 and IL-23. Activation of the unfolded protein response (UPR) in macrophages has been reported to correlate with the inflammatory disease in a rat model of SpA. Thus, in response to putative arthritogenic trigger(s), these events lead to pathogenic immune responses via immunomodulation of both innate and adaptive immunity.

Enlarge 

Enlarge 

IL23 Receptor and IL-23/Th17 Pathway

IL23R gene, that encodes the receptor for IL-23, is associated with AS. IL-23 regulates the differentiation and expansion of Th17 cells, which are the major players of inflammation and autoimmunity. An allotypic variant of IL-23R, and some other disease predisposing genes, are shared between AS, UC and CD (Figure 4-6). This suggests the involvement of IL-23/Th17 pathway in the development of AS and the other SpA related diseases. Th17 cells are derived from naïve CD4+ lymphocytes. They reside at barrier surfaces such as gut mucosa and skin to protect the host against invasion by extracellular microorganisms. However, these Th17 cells can cause inflammation and autoimmunity when they acquire a pathogenic phenotype mediated by IL-23R signaling. There are different stages of Th17 cell differentiation, and their modulation by several cytokines are known in murine and human cells (Figure 4-7).

In humans, TGF­β + IL­21 are the key cytokines for the initiation of Th17 cell differentiation. TGF­β + IL-6 + IL-21 or TGF­β + IL-6 + IL-23 are the alternative cytokine combinations that can induce the expression of human Th17 lineage specific transcription factor RORc. IL-6 and IL-1β are involved in the enhancement and IL-23 in expansion and stabilization of the Th17 cells. Differentiated Th17 cells produce IL-17A, IL-17F and IL-22, as well as chemokine receptor CCR6. IL-17A and IL-17F stimulate the production of proinflammatory cytokines, chemokines and metalloproteinases from various cell types and mediate tissue inflammation. IL-21 participates in the differentiation/amplification of Th17 cells. CCR6-expressing Th17 cells are preferentially recruited to the inflamed sites via its ligand CCL20. In mice TGF­β + IL­6 are the key cytokines for the initiation of Th17 cell differentiation and expression of RORγt. IL-21 is involved in the enhancement and IL-23 in the expansion and stabilization of the Th17 cells.

TGF­β is indispensable, but not sufficient for Th17 development. In the absence of IL-21 or IL-6, TGF­β cannot promote Th17 polarization; not even in the presence of IL-23 because naïve CD4+ T cells do not express IL-23 receptor. TGF­β plays a crucial role also in the differentiation of FoxP3+ regulatory T cells (Treg cells), which have a reciprocal relationship with Th17 cells (Figure 4-8). Presence of proinflammatory cytokines IL-6 or IL-21 inhibits FoxP3 expression and promotes Treg plasticity toward Th17 cells. In the absence of inflammation, TGF­β promotes Treg differentiation promoting tolerance; this is due to FoxP3-mediated inhibition of RORc.

A truly remarkable study was published by Sherlock and colleagues in 2012 that demonstrated that excess IL-23 is sufficient in generating specific prototypic SpA manifestations because mice injected with IL-23 genetic mini-circles (to overexpress IL-23) develop enthesitis and subsequently arthritis (including sacroiliitis), osteoproliferation, psoriasis and inflammation of the aortic root. Expression of TNF-α, IL-6, chemokines and matrix metalloproteinases was observed in the inflamed paws, but TNF blockers did not inhibit development of this IL-23–mediated disease. Inflammation occurred independently of the classic CD4+ Th17 cells. Rather, IL23R+RORyt+ CD4-CD8- innate lymphoid-like T cells were found to be residing in both the entheses and the aortic root.

Remarkably, treatment of these mice with anti-IL-17 or anti-IL-22 ameliorated enthesitis and arthritis, but it was most effective when given in combination. However, the enthesitis and osteoproliferation phenotype was only reproduced by IL-22 expressing genetic mini-circles. In this model, it is proposed that IL-23 mediates inflammatory process through IL-17 and TNF, while IL-22 predisposes to new bone formation. These results are well summarized in a figure by Lories and McInnes (Figure 4-9).

A recent study provided new evidence in support of a link between the joints and the gut. This study revealed that Type 3 innate lymphoid cells producing IL-17 and IL-22 are expanded in the gut and migrate in extra-intestinal sites where, through the production of IL-17 and IL-22, they may be responsible for the induction of inflammation. Notably, more than 60% of AS patients have subclinical gut inflammation which correlates with spine inflammation, and is characterized by overexpression of IL-23. This cytokine is known to regulate IL-22 production by lamina propria NKp44(+) NK cells, where they appear to play a tissue-protective role.

The results of clinical trials have confirmed the importance of IL-23/Th17 pathway in the treatment of AS/axSpA, as well as in other SpA related diseases. IL-17 inhibitors, secukinumab and ixekizumab, are approved both by the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) for the treatment of psoriasis, psoriatic arthritis (PsA), AS and nr-axSpA, and IL-23 targeting therapies are approved for the treatment of psoriasis but not for AS.

Enlarge 

Enlarge 

Enlarge 

Enlarge 

Other Non-MHC Risk Genes for AS

Several of the other non-major histocompatibility complex (non-MHC) susceptibility loci are also IL-23R–related; these include IL-12B (it encodes IL-12p40, a component of IL-12 and IL-23), TYK2, IL6R, IL27, PTGER4 and CARD9. TYK2 is a member of Janus kinases and it is involved in signal transduction from several inflammatory and anti-inflammatory cytokines, including IL-23. IL-6R mediates IL-6 signaling pathway which stimulates Th17 cells in the presence of TGF-β. IL-27 inhibits the differentiation of CD4+ T cells into Th17 cells and limits immunopathology in Th17 associated diseases. The association of PTGER4 (it encodes PGE2EP4 receptor, one of the four known PGE2 receptors named EP1 to 4) with AS and Crohn’s disease (CD) is particularly fascinating (Figure 4-6). PGE2 is produced in response to dectin-1 recognition of β-glucan and both PGE2 and PGE2EP4 have been shown to stimulate dendritic cell production of IL-23, in turn activating Th17 lymphocytes and potentially other IL-23R–expressing cell types.

It is worth emphasizing that nonsteroidal anti-inflammatory drugs (NSAIDs) have greater anti-inflammatory effect in AS than in rheumatoid arthritis (RA). CARD9 encodes a component of the dectin-1 innate immunity signaling pathway. Dectin-1 recognizes β-glucan, a component of fungal and some bacterial cell wall, and signals to the nucleus through CARD9, leading to activation of Th1 and Th17 lymphocyte subsets, thereby linking the innate and adaptive immune systems. Interestingly, exposure of SKG mice to β-glucan induces SpA, episodic unilateral uveitis and inflammatory bowel disease (IBD) resembling CD. In a colitis associated mouse model, CARD9‐signaling has been shown to increase the production of IL‐1β within the damaged intestine and subsequent generation of IL‐22 by group3 innate lymphoid cells.

GPR65, which encodes G protein-coupled acid sensing receptor, is one of the risk genes that has been found to be associated with AS by GWAS. The GPR65 gene codes for a proton sensor present on Th17 cells and induces a pro-inflammatory phenotype. Increased numbers of GM-CSF-producing CD4 and CD8 lymphocytes, as well as increased numbers of IL-17A+GM-CSF+ double-producing CD4, CD8, γδ and NK cells have been found in the blood and joints of patients with SpA, suggesting a possible role also for GM-CSF in AS pathogenesis. In the same study, GPR65 has been shown to be upregulated in the GM-CSF+ and IL-17A+GM-CSF+ cells. Worth mentioning, GPR65 in T cells are activated when cultured in an acid environment and leads to increased GM-CSF production, whereas blocking GPR65 leads to a decrease in production of GM-CSF, but not in IL-17. GPR65 knockout mice are protected from the development of experimental autoimmune encephalomyelitis. This indicates that GPR65 is needed for Th17-cell pathogenicity. Furthermore, in the absence of GPR65 in T cells, IL-17A is produced in significantly reduced levels when activated by IL-23.

Thus, it may be speculated that activated GPR65 by local metabolic changes, such as low pH in the joints, may modulate local Th17 immune responses and the generation of GM-CSF. It has long been known that serum and glucocorticoid-induced kinase-1 (SGK1), a salt-sensitive serine/threonine kinase, plays critical role in the induction of pathogenic Th17 cells by regulating IL-23R expression. Based on all these data, a novel pathogenetic model has been proposed which postulated that changes in the local metabolic environment (pH, salt) may play a key role in the development of AS by induction of a pro-inflammatory phenotype. The differentiation of Th17 and Treg cells are linked on a reciprocal basis and there is considerable amount of plasticity between the two compartments. The authors of this model speculate on the possibility of the suppression (loss) of FoxP3 expression by activation of SGK1 and GPR65 which leads to trans-differentiation of Treg cells into Th17 cells with resultant induction of a pro-inflammatory state. The whole model is detailed in Figure 4-10, with potential therapeutic targets pointed out, including GPR65 and GM-CSF.

There is compelling evidence indicating that IL-1 gene complex is a susceptibility locus for AS. A recent genome side association studies ( GWAS) of Turkish–Iranian AS patients has demonstrated that the M694V variant of the Mediterranean Fever (MEFV) gene is associated with a substantially increased risk of AS in these populations, which is consistent with previous candidate gene studies and several case reports describing co-existing cases of SpA and Familial Mediterranean Fever (FMF). Penetrant mutations in the pyrin-encoding MEFV gene, such as M694V, lead to excessive IL-1β production through dysregulation of caspase-1 activation and the inflammasome. Notably, in the GWAS study serum IL-1β, IL-17 and IL-23 levels were higher in AS patients than controls and serum IL-1β and IL-23 levels were increased among AS patients who were MEFV M694V carriers compared with non-carriers.

Pyrin, encoded by MEFV has also been shown to be involved in autophagy, which has been proposed to be implicated in the pathogenesis of AS. These findings suggest that FMF and AS have overlapping etiopathogenic mechanisms. IL-1 inhibitors are known to be effective and indicated in the treatment of colchicine resistant FMF. However, their effectiveness in AS is limited at best, as suggested by the somewhat contradicting results of two small open studies, which nevertheless do not completely exclude a therapeutic role of IL-1 inhibition in selected cases of AS.

Genetic and clinical studies also support existence of HLA-B27–independent common link of AS/axSpA and inflammatory lesions in the gut and skin. It is unusual, especially among individuals of northern European extraction, to observe familial occurrence of primary AS when there is no familial occurrence of HLA-B27, UC, CD and psoriasis. Recent genetic data further highlight the clinical and genetic heterogeneity between AS and nr-axSpA and suggest that differences in genetic associations may explain the variation observed in responses to therapies within the same or related disease spectrum.

Enlarge 

Microbiome and Other Non-genetic Factors

The putative environmental trigger(s) for AS is(are) still unknown, but the close relationships between AS and clinical and asymptomatic forms of CD and UC suggest the potential involvement of an immune reaction in the gut that may be influenced by changes in the gut microbiome, as discussed previously. An infectious trigger for the disease remains an intriguing hypothesis. Studies with transgenic rat model for SpA show that when these animals are raised in a germ-free environment, they do not develop disease; however, in a regular environment, they develop a disease that is like AS/SpA.

To date, however, no infectious trigger has been unequivocally established for AS, although it is not the case with reactive arthritis that can be triggered by certain bacterial infections, and it does show an association with HLA-B27, although it is much weaker than that with AS. It has been suggested that AS may be triggered by gut infection with Klebsiella bacteria, but the evidence is circumstantial, based on the observation by some, but not all, investigators of elevated levels of antibodies against Klebsiella pneumoniae in the blood of patients with active disease. More convincing proof has been lacking. Recent studies examining the gut flora have demonstrated a discrete microbial signature in the bowels of patients with AS or SpA compared to controls, with evidence of dysbiosis. Of particular interest, restoration of the perturbed microbiome was found to be associated with anti-TNF therapy.

A potential interaction between microbiota and Th17 immune response in humans is suggested by the induction of Th17 cells in the murine intestine by specific bacterial strains and restoration of dysbiosis by IL-23p19 inhibition in SKG mice. Diet also affects the composition of the gut microbiota, but no specific dietary components have as yet been identified that can impact the risk of development of AS.

A report suggesting that non-breast-fed babies are more likely to develop AS than their breast-fed siblings or healthy controls needs further verification. If the local metabolic environment affects the immune response as suggested by the model mentioned earlier, changes in diet may affect the axis between the gut and the immune system. For example, some recent studies reported that high salt diet induces IL-17-dependent gut inflammation and aggravate colitis in experimental mouse models by activation of SGK1 kinase. Moreover, high-salt intake in mouse models and in humans has been shown to affect the microbiome and induce Th17 immune response. Genes playing a role in the innate immune response, and aberrant bacterial sensing have been implicated in UC and CD (Figure 4-6).

Nevertheless, a specific environmental trigger(s) for AS remain(s) unknown. It has also been proposed that non-self–derived peptide may trigger a cross-reactive cytotoxic T-cell response activating the clonotypes that have not been eliminated during ontogenesis in the thymus. This failure to eliminate the self-peptide–activated T cells during development may contribute to the autoimmune and inflammatory process in later years in AS patients.

Biomechanical stressing and microdamage at entheses or synovio-entheseal complex, independent of adaptive immune response, can lead to autoinflammatory response, defined as a self-directed inflammation where local factors at sites predisposed to disease initiate activation of innate immune cells, including macrophages and neutrophils. A mouse model for SpA has demonstrated that mechanical strain may indeed trigger entheseal inflammation and lead to new bone formation. Figure 4-11 illustrates that regions such as entheses that have high biomechanical stress can exhibit reactive bone remodeling and osteoproliferation that can lead to ankylosis.

Enlarge 

Cells and Cytokines

Various cell types have been shown to be involved in AS, including CD4+, CD8+ and regulatory T cells, Th17 cell, dendritic cells, macrophages, NK cells, γδ cells, innate lymphoid cells, B cells, mast cells and intestinal Paneth cells. These immune cells produce various cytokines that play a key role in the pathogenesis of AS (Figure 4-12).

Enlarge 

Concluding Remarks

The etiopathogenesis of AS/axSpA is, yet, not fully understood, although there is strong evidence indicating that HLA-B27 and other genetic factors play a pivotal role. It is possible that HLA-B27 may predispose to AS via more than one mechanism, or by different mechanisms in different patients, based on additional genetic, epigenetic and environmental factors. We have focused on AS (“radiographic axSpA”) since it is both clinically and genetically much more homogeneous than nr-axSpA. A better understanding of pathogenesis with incorporation of genomics data and polygenic risk scores (PRS) will lead to better classification of axSpA, help explain differences in response to treatment and also benefit clinical care with development of new therapies.