Immunopathogenesis of Lyme disease

INTRODUCTION — Advances in the understanding of the interactions that occur between Borrelia burgdorferi and its mammalian and tick hosts have led to important insights into the pathogenic mechanisms that underlie the manifestations of human Lyme disease. With the sequencing of the B. burgdorferi genome came several surprising insights. Among these is the fact that B. burgdorferi has a very small genome compared with other bacteria. It does not encode for any toxins or lipopolysaccharides, but does encode a large number of lipoproteins relative to other bacteria. Also missing from its genome are genes that enable bacteria to synthesize such essential products as amino acids, fatty acids, enzyme cofactors, and nucleotides. As such, B. burgdorferi is dependent upon its environment to provide these nutrients and has evolved specialized mechanisms for adapting to its different environments.

The lifecycle of B. burgdorferi requires that it survive in two distinct environments — that found in a tick host and that found in a mammalian host. The challenges posed by these environments differ greatly. Because ticks do not thermoregulate, while in the tick host B. burgdorferi must survive at the extremes of ambient temperatures found in winter and summer in its areas of geographic distribution. In addition, ticks take only one blood meal every 6 to 12 months, and as a result, B. burgdorferi must be able to survive with minimal nutrition for long periods of time.

In contrast, life in a mammalian host provides a stable temperature and an abundance of nutrients. However, compared with ticks, which have rudimentary immune systems, mammals have highly sophisticated immune defenses. The ability of B. burgdorferi to cause long-term infection in mammalian hosts requires that it implement strategies to successfully evade and subvert host immune defenses. Optimal host immune control of B. burgdorferi infection requires both innate and adaptive immune systems. However, neither system, alone or in combination, is able to completely eradicate the organism in mice despite resolution of signs of infection and inflammation. In humans, it appears that the immune system can eventually eradicate the infection, since almost all patients eventually resolve their symptoms even without antibiotics [1]. In Europe, B. afzelii has been cultured from skin lesions of patients in Europe with acrodermatitis atrophicans after greater than 10 years [2], but this genotype is not present in the United States.

In this topic, we will describe some of the mechanisms that B. burgdorferi uses to establish infection as well as the host immune responses to invasion by the organism that both help control the infection and also cause some of the manifestations of Lyme disease. The microbiology, epidemiology, clinical manifestations, diagnosis, prevention, and treatment of Lyme disease are discussed separately. (See "Epidemiology of Lyme disease" and "Clinical manifestations of Lyme disease in adults" and "Lyme disease: Clinical manifestations in children" and "Diagnosis of Lyme disease" and "Prevention of Lyme disease" and "Evaluation of a tick bite for possible Lyme disease" and "Treatment of Lyme disease".)

ACUTE LOCALIZED DISEASE

Bacterial adaptations

Outer surface protein variations — B. burgdorferi exit the tick while it feeds on a mammalian host and establish infection in the skin. To transit from the tick midgut, where the B. burgdorferi are located prior to initiation of feeding, to the tick salivary gland, the organism rapidly downregulates a protein called outer surface protein A (OspA) and upregulates another protein called outer surface protein C (OspC) [3,4]. Downregulation of OspA accomplishes two tasks:

●Since OspA is involved in the binding to the tick midgut, downregulation facilitates detachment from the midgut and transit to the salivary glands

●Downregulation of OspA removes a potent target for the immune system

OspA was used as the antigen in the human Lyme disease vaccine, which is no longer available, and antibodies to OspA efficiently target and kill bacteria expressing this protein [5-8]. (See "Prevention of Lyme disease", section on 'Vaccination'.)

Upregulation of OspC facilitates invasion of bacteria into skin, and, in the absence of OspC, bacteria are unable to establish infection in the mammalian host efficiently [9,10], although it is possible that other proteins may substitute for the function of OspC [11]. One role for OspC is in the binding of a tick salivary protein, Salp15. Binding of Salp15 may protect the spirochete by inhibiting antibody-mediated killing. Salp15 also inhibits activation of T lymphocytes by binding to the CD4 receptor [12]. After the establishment of infection, OspC is no longer required and is turned off [13]. Another protein, VlsE, which is capable of antigenic variation, may be able to replace the functions of OspC during later infection [14].

It should be noted that not all B. burgdorferi in the population turn off OspA and turn on OspC as the tick takes its blood meal. Studies have shown that there is heterogeneity in protein expression, likely due to differences in either the microenvironments encountered by the bacteria, slow turnover of proteins, or subtle genetic differences between the bacteria themselves [4].

Factors in tick saliva — Tick saliva contains several factors that are important in altering the host inflammatory response, which is important for the tick to feed successfully, but may also be exploited by B. burgdorferi to establish infection. Induction of Th1 type, pro-inflammatory cytokines has been shown to improve early host control of borrelial infection, whereas induction of Th2 type cytokine responses are more permissive for the establishment of infection [15-18]. Tick saliva has been shown to promote the release of Th2 type cytokines (eg, interleukin 4 and 10) that suppress cellular immunity [19,20]. Tick saliva also contains components that can inactivate antimicrobial peptide release, chemotaxis of neutrophils, complement activation and the generation of reactive oxygen species [21-24].

Host response — The earliest host immune responses to the entry of B. burgdorferi into the skin are from the innate immune system.

Cellular immune response — B. burgdorferi is effectively killed by professional phagocytes such as neutrophils and macrophages [25-27]. In animal models, there is early recruitment of neutrophils and basophils to the site of inoculation, but this is rapidly replaced by eosinophils and macrophages [28]. Histologic examinations of human erythema migrans lesions show a marked lack of neutrophils [29]. Artificially increasing the number of neutrophils infiltrating an erythema migrans lesion by increasing amounts of neutrophil chemokines in the lesion may improve early control of infection [28], but it is unknown whether Borrelia specifically suppress neutrophil infiltration as a mechanism of immune evasion. Two tick salivary proteins have been shown to suppress neutrophil adherence and killing [21]. Recognition and processing of B. burgdorferi by macrophages and dendritic cells result in release of cytokines and recruitment of T cells and B cells to the site of infection. Induction of cytokines and chemokines occurs largely through recognition by pattern recognition receptors (PRRs) [30,31].

Toll-like receptors — Innate immune effector cells such as neutrophils and macrophages recognize pathogens through identification of pathogen-associated molecular products (PAMPs) by pattern recognition receptors PRRs. Among the best studied PRRs are the toll-like receptors (TLRs). (See "Toll-like receptors: Roles in disease and therapy".)

The human genome encodes for 10 known TLRs, each of which recognize specific microbial products (eg, lipopolysaccharide, microbial DNA, peptidoglycan). Borrelial lipoproteins are recognized by heterodimers of TLR1 and TLR2 [32,33]. Engagement of TLR1/TLR2 by borrelial lipoproteins results in activation of multiple downstream signaling pathways as well as translocation of nuclear factor kappa B (NFkB), leading to the release of inflammatory cytokines and chemokines [32]. Recognition of different B. burgdorferi products by other TLRs also appears to contribute to the inflammatory response [31,34]. Among the chemokines and cytokines found to be elevated in erythema migrans lesions are interleukin (IL)-6, interferon-gamma, IL-1b, tumor necrosis factor-alpha (TNF-alpha), IL-12 and transforming growth factor-beta (TGF-beta) [29,35].

Consistent with the histological findings in erythema migrans, levels of the macrophage and T cell chemokines are increased over neutrophil chemokines [29,36]. In animal models, TLRs have been shown to play an important role in controlling infection with B. burgdorferi, but TLR-deficient animals show high levels of inflammation and cytokine release, suggesting that there are redundant pathways for detecting Borrelia and activating inflammatory pathways [37-39].

TLR1 and TLR2 may also affect B. burgdorferi gene expression. A microarray analysis of B. burgdorferi gene expression in a mouse model showed that several genes were altered in TLR1/2-deficient mice compared with wild-type mice [40]. This suggests that B. burgdorferi may be altering protein expression in response to the inflammatory environment mediated by TLR signaling. It is currently unknown how the genes that have been identified as altered in TLR1/2-deficient mice are involved in the adaptation to either the wild-type or the TLR1/2-deficient milieu.

TLR signaling may also play an important role in phagocytosis of Borrelia, antigen presentation, and the subsequent development of an adaptive immune response to the organism. Other receptors that may be important in the early recognition of Borrelia and induction of inflammatory mediators include integrin-type receptors such as complement receptor 3 (CR3) and integrin alpha(3)beta(1) [41,42].

ACUTE DISSEMINATED DISEASE

Bacterial adaptations — After inoculation into the skin, Borrelia begins to multiply rapidly. Dissemination to distant sites begins to occur within two to three days [43-45]. It appears that some strains of Borrelia are more prone to dissemination than others, although this remains controversial and the specific mechanisms involved remain unknown [46]. Other studies have shown that the distribution of subtypes in distal tissues is similar to the distribution of subtypes in skin, suggesting that the differences in the presence of subtypes in blood may be due to the ability to detect these strains in blood and not to differences in the capacity for dissemination [47].

Adherence to host cells and ECM proteins — During the early phase of infection after inoculation, B. burgdorferi use several different strategies to establish infection and disseminate. Adherence to host cells and extracellular matrix proteins is an important first step in the establishment of infection. B. burgdorferi expresses proteins that bind to host integrins, proteoglycans and glycoproteins. Among the identified proteins are P66, BBK32, and decorin binding protein. P66 is a borrelial protein that binds to integrin receptors such as integrin alpha(IIB)beta(3), which is expressed on platelets and which may facilitate dissemination of the bacteria on entry into the bloodstream [48,49]. BBK32 and decorin binding protein adhere to fibronectin and decorin, respectively; these proteins are found in the extracellular matrix of tissues where Borrelia have a predilection to disseminate [50,51]. Bacteria that have deletions of BBK32 or decorin binding protein have defects in their ability to establish infection and to colonize distant tissues [52-54]. Binding to decorin may protect the organism from clearance by both the early innate immune response and the adaptive immune response. Whether variation in the sequence of these binding proteins between species of Borrelia accounts for differences in the frequency of specific clinical manifestations among strains of B. burgdorferi sensu lato has not yet been closely examined [55].

Utilization of host proteins — In order to disseminate from the site of inoculation, Borrelia need to be able to move through the dense extracellular matrix of the skin and other tissues. Most other bacteria that are inoculated into skin and disseminate produce proteases capable of digesting extracellular matrix proteins. B. burgdorferi does not produce any known exported proteases capable of digesting extracellular matrix proteins. Instead, it appears to utilize host proteins for this function. B. burgdorferi bind plasmin, a serine protease capable of digesting many extracellular matrix components. In animal models, binding of plasmin has been shown to aid in the movement of B. burgdorferi into the bloodstream [56]. In addition, plasmin is an activator of some matrix metalloproteinases. Matrix metalloproteinases are host proteases that are induced as part of the inflammatory response to Borrelia infection [57]. As a family, matrix metalloproteinases can degrade every component of the extracellular matrix and are the major proteases capable of digesting collagen.

Evasion of antibody-mediated killing — During the early phase after inoculation, the organism must begin to evade additional components of the immune system. IgM antibodies to prominent borrelial surface proteins, including outer surface protein C (OspC) and flagellin, begin to develop very early in infection. Antibodies against OspC, but not flagellin, are capable of killing the spirochete. While OspC is critical for the early establishment of infection, it is no longer necessary after the initial inoculation and is rapidly downregulated to evade antibody-mediated killing [10,58]. Bacteria that do not downregulate OspC are likely rapidly selected against. Another antigenic target, a surface lipoprotein called VlsE, is able to undergo antigenic variation during infection by recombination of 15 silent gene cassettes into the central expressed cassette to escape antibody-mediated killing [59].

Evasion of complement-mediated killing — Complement-mediated killing of Borrelia can occur either with or without the presence of specific antibody to B. burgdorferi, via the classical pathway or alternative pathway, respectively. To evade complement-mediated killing, Borrelia contain multiple proteins including outer surface protein E (OspE)-related proteins (Erps) and complement regulator-acquiring surface proteins (CRASPs) that bind complement factor H and complement factor H-like protein 1 [60,61]. Factor H and factor H-like protein 1 bind and inactivate C3b and protect the organism from complement-mediated killing. The different strains of Borrelia express different variants of CRASPs and Erps that bind to Factor H from different species of mammals with varying efficiency [62]. It has been shown that strains of CspA, a B. burgdorferi CRASP protein, differentially bind to factor H from different species and that binding correlated with transmissibility of strains to a species [63].

Host response — As described above, the early host response involves recognition by pattern recognition receptors, phagocytosis by macrophages and dendritic cells, and complement-mediated lysis. Each of these contributes to the control of infection, but ultimately, is not able to contain infection. As the organisms multiply and disseminate, it is presumed that the increased activation of immune effector cells and the release of inflammatory cytokines and pyrogens result in many of the non-specific signs of Lyme disease during this period including fevers, myalgias, and arthralgias. In animal models, localized findings, such as carditis, that occur during the acute disseminated phase are directly correlated with local levels of cytokines and often, but not always (depending upon the animal model), with the number of bacteria present in the tissue.

Cellular immune response — The role of specific types of cells in the development of local inflammation in response to Borrelia has been examined extensively in mouse models. While not all strains of mice develop significant pathology following infection with B. burgdorferi (eg, inbred C57BL/6 mice or the wild host Peromyscus mice), some strains such as inbred C3H mice develop carditis and arthritis after infection with the organism. Although arthritis is considered a manifestation of late disseminated disease, many of the immune processes are similar between arthritis and carditis, so the impact of the host immune system on these manifestations will be discussed together. The genes that control the severity of the response in different mouse strains have not been identified to date, although several loci on chromosomes have been proposed [64]. There are two major identifiable effects of the immune response during this stage — the generation of inflammation and the control of infection. Although the two are linked (ie, there is no inflammation in the absence of infection and conversely failure to control infection can overcome genetic resistance to the development of arthritis), different cells and different processes can contribute to each.

Bacteria that have been phagocytized by macrophages or dendritic cells are processed and presented to CD4+ T helper cells. Depending upon the genetics of the mouse, these activated T cells may produce Th1 type responses where interferon-gamma is generated or Th2 type responses that generate anti-inflammatory cytokines such as IL-10. Mice of different genetic backgrounds are predisposed to generate either Th1 or Th2 type responses to infections with different pathogens. In general, Th1 type responses result in increased inflammation, but better control of infection whereas Th2 type responses suppress inflammation at the cost of control of the infection [15,65]. Human T cell responses do not differentiate along the lines of Th1 and Th2 type responses as clearly as mice. CD4 cells in human peripheral blood and synovial fluid from patients with Lyme arthritis contain mixed populations of Th1, Th2, Th17, and T regulatory cells [66]. In general, in mice both Th1 and Th2 type cytokines may be induced by infection with Borrelia. This is true even in mice with more differentiated Th1 and Th2 type responses and it is the relative balance between pro- and anti-inflammatory cytokines that determines the phenotype of the infection.

Using the mouse model, the importance of specific cell types to control of inflammation and to control of infection has been delineated. Natural killer T cells are activated by diacylglycerol, a borrelial glycolipid, presented by CD1d-expressing cells such as macrophages [67]. Mice with a beige mutation, which results in a lack of natural killer cells and impaired macrophage function, develop more severe arthritis during the early period, but resolve their symptoms in a time course similar to wild-type congenic mice, suggesting that there is still development of a normal adaptive immune response that clears the organism.

Both adaptive T and B cell responses contribute to the control of infection, as severe combined immunodeficiency (SCID) mice or Rag-/- mice, which lack both T and B cells have high tissue levels of bacteria following infection. On an arthritis-susceptible genetic background (eg, C3H), the absence of T and B cells leads to severe arthritis. However, on a resistant background (eg, C57BL/6), the loss of T and B cells results in only minimal to moderate arthritis despite higher tissue numbers of bacteria [68,69]. Mice with a deletion of B cells but not T cells develop more severe arthritis and T and B cell depleted mice reconstituted with T cells alone also develop more severe arthritis, confirming a role for T cell in mediating inflammation that is separate from the control of infection [69]. B cells play a major role in clearing of infection but do not affect inflammation.

LATE DISSEMINATED DISEASE

Bacterial adaptations — During untreated late stage infection in mammals, symptoms persist at specific sites (joints, central nervous system), and it is suspected that B. burgdorferi may persist at low levels in these tissues. In untreated patients, B. burgdorferi has been cultured (rarely) from acrodermatitis lesions greater than 10 years after infection, and B. burgdorferi DNA can be amplified by PCR from synovial fluid for years after the onset of infection [70,71]. The sites in which B. burgdorferi are found during late stage disease may be influenced by the species of B. burgdorferi (eg, B. burgdorferi sensu stricto in the US may be found in the joints, but is not found in the skin, whereas B. afzelii in Europe may be found in the skin). In mouse models, during late stage disease, the bacteria have typically downregulated most of the outer surface lipoproteins that serve as strong antigenic targets. Intermittent flares of arthritis occur in both untreated humans and mice, suggesting that the spirochetes become recognized by the host immune system. It is not clear what the cause of these flares are, but the potential hypotheses would include escape from immune control (eg, through antigenic variation by VlsE), allowing increased replication of the bacteria and reactivation of inflammation through stimulation of innate immune mechanisms. Another possibility is that B. burgdorferi express a new antigenic target in the presence of a specific stimuli that results in recognition by the immune system. In the wild mouse host, Borrelia must be able to sense and successfully colonize new feeding ticks in order to complete its lifecycle. It may sense and utilize tick salivary factors [21,72] or inflammation in response to tick bites [73] to help facilitate re-entry into a tick, and it is possible that some of these triggering mechanisms may be responsible for activating expression of repressed proteins, such as OspA, that result in increased inflammation.

Host response — During late stage disease, the adaptive immune system is typically successful at greatly reducing the bacterial numbers in tissues. With these reduced numbers of bacteria, activation of innate and adaptive immune mechanisms that release inflammatory cytokines systemically is reduced, although local levels may remain quite high. As a result, systemic symptoms are typically minimal during late infection. Low level destruction of tissue may continue to occur at sites where the bacteria are localized. The slow destruction of tissue compared with other bacterial causes of arthritis or encephalopathy likely is due to the low bacterial numbers and the fact that Borrelia, unlike other bacteria that spread through skin and cause end organ tissue damage, do not produce proteases capable of destroying host tissues and extracellular matrix. The damage that is seen is likely due to host inflammatory factors, including induced proteases such as matrix metalloproteinases [74].

Antibody response — Despite the fact that Borrelia typically downregulate outer surface proteins during mammalian infection, patients with late stage arthritis typically develop a very broad antibody response to many antigens, including to outer surface protein A [75]. As a result, patients with Lyme arthritis are relatively resistant to reinfection with B. burgdorferi after antibiotic therapy, whereas those treated during earlier stages typically are not [76]. Although antibodies that develop early to proteins expressed during early infection such as outer surface protein C (OspC) are effective against the identical strain, sequence variation in OspC between strains make antibodies to OspC relatively non-protective against reinfection [77-79]. In contrast, the sequence of OspA is relatively stable among North American strains of B. burgdorferi and antibodies to OspA are cross-protective against most strains found in the United States. Antibodies to OspA protect by entering into the tick midgut with the blood meal while OspA is still being expressed, and killing bacteria within the tick before it has a chance to enter the mammalian host.

SUMMARY

●Borrelia burgdorferi has evolved sophisticated mechanisms for adapting to its tick and mammalian hosts including:

•Tight control of protein expression to minimize antigenic targets for immune recognition and

•Subversion of host mechanisms to establish infection and evade host immunity (see 'Introduction' above)

●Optimal host immune control of B. burgdorferi infection requires both innate and adaptive immune systems. (See 'Host response' above and 'Host response' above and 'Host response' above.)

●Because B. burgdorferi do not produce toxins, secreted proteases, or other destructive molecules, the majority of the symptoms seen with human Lyme disease are due to the effects of the host immune response. (See 'Host response' above and 'Host response' above and 'Host response' above.)