Biology of Infection with Borrelia burgdorferi

The spirochete Borrelia burgdorferi is a tick-borne obligate parasite whose normal reservoir is a variety of small mammals [1]. Whereas infection of these natural hosts does not lead to disease, infection of humans can result in Lyme disease, as a consequence of the human immunopathological response to B. burgdorferi [2, 3]. Consistent with the pathogenesis of Lyme disease, bacterial products that allow B. burgdorferi to replicate and survive, rather than true “virulence factors,” appear to be primarily what is required for the bacterium to cause disease in a susceptible host. In support of this idea, the genome sequence of B31, the type strain of B. burgdorferi sensu stricto [4, 5], revealed that the bacterium lacks factors common to many bacterial pathogens, such as lipopolysaccharide, toxins, and specialized secretion systems. In this chapter, we will describe the basic biology of B. burgdorferi, and some of the bacterial components required to infect and survive in the mammalian and tick hosts.

Natural history of the Lyme disease spirochete

The causative agent of Lyme disease is a member of the eubacterial phylum Spirochaetes. Members of this group of organisms share a distinctive morphology that includes a spiral or wavelike body and flagella (organs of motility) enclosed between the outer and inner membranes. The spirochetes include several human pathogens, including Treponema pallidum (agent of syphilis), Leptospira interrogans (leptospirosis), and several Borrelia spp. that cause relapsing fever. Although these other spirochetes had long been known to medicine, it was relatively recently that the bacterial agent of Lyme disease was identified [1].

Lyme disease was clinically described as an infectious illness by Dr. Alan Steere and colleagues in 1977 and is currently the leading vector-borne disease in the United States [6, 7]. Steere et al. suggested that the epidemiology of Lyme disease indicated transmission by an arthropod vector due to the geographic clustering of patients in rural areas and the seasonal occurrence of the symptoms [7]. Subsequently, Dr. Willy Burgdorfer and coworkers observed spirochetes in the midgut tissues from ticks collected in a Lyme disease endemic area [1]. These spirochetes produced a skin rash resembling erythema migrans when injected into rabbits, and sera from Lyme disease patients reacted with the bacteria in indirect immunofluorescence assays. In recognition of this discovery, the bacterium was named Borrelia burgdorferi [8].

In Europe and Asia, the three species of Borrelia responsible for most human cases of disease are B. burgdorferi sensu strictu (s.s.), B. garinii and B. afzelii, which are collectively referred to as B. burgdorferi sensu lato [9]. Lyme disease in the U.S. is caused by the single species B. burgdorferi s.s. The clinical manifestations of European and North American Lyme disease share some common features such as an erythema migrans rash and an influenza-like illness. Subsequently, other symptoms may develop that roughly correlate with the infecting species. Arthritis frequently accompanies B. burgdorferi s.s. infection, whereas neurological symptoms are correlated with B. garinii, and skin disorders with B. afzelii, although these clinical associations are not absolute [10].

Genome and molecular genetics of B. burgdorferi

Although the underlying genetic traits contributing to the differences in disease have not been identified, the release of three Borrelia genome sequences (one from each 21 species) should aid in this endeavor [5, 11–13]. These genomes have several common features, including a linear chromosome and a large number of smaller DNA molecules (plasmids), some of which are linear and others circular [14–17]. The linear structure of the chromosome and many of the plasmids is unusual in the bacterial world, although the evolutionary advantage of this form of DNA is unknown. However, it likely confers some benefit to the genus Borrelia, as all characterized members retain linear DNA molecules.

Despite the atypical DNA form, the majority of genes encoded on the B. burgdorferi chromosome are commonly found in other bacterial genomes [4]. The genes encoded on the plasmid component of the genome are less recognizable and the majority of these appear to be unique to the genus Borrelia [4, 5]. Additionally, several plasmid-encoded genes have been shown to be required for infectivity or persistence in the tick or the mammalian host (discussed below).

The genome sequence of the Lyme disease spirochete revealed several interesting features [4, 5]. First, no classically-defined virulence factors were identifiable, probably because B. burgdorferi did not evolve to cause disease in mammals. However, a sizable number of putative lipoproteins were recognized and some have been shown to trigger components of the mammalian innate immune system (discussed below). Finally, the plasmids appear to be in a state of rapid evolution, as many genes appear to be mutationally inactivated, whereas others have been duplicated and may be functionally diverging from each other [5]. Perhaps reflecting this evolving state, the number of plasmids, their sizes, and the gene order varies substantially among strains and between species; in some cases, plasmids or portions of some plasmids are absent [18, 19]. The implications of these variations in plasmid structure and content to infectivity or disease remain unclear.

Many genes are present across all three Lyme disease spirochetes but display sequence variation both between and within species. Genes required for infectivity or persistence within vertebrate hosts (such as ospC and vlsE, see below) may vary significantly at both the nucleotide and amino acid level [13, 20–22]. The variability within a given gene product complicates the design of effective vaccines based on these proteins since the elicited immune response may not protect against all variants [23, 24].

Analysis of the B. burgdorferi genome sequence found that greater than 6% of the chromosomal genes are involved in motility and chemotaxis [4]. The flagella of spirochetes traverse the length of the cell body and are “hidden” beneath the outer membrane, in contrast to other organisms that have external flagella radiating outward. A potential advantage to the flagellar arrangement of spirochetes is the shielding of the highly conserved and immunogenic flagella from the host immune system. Also, the morphology and motility of spirochetes allows these organisms to swim in highly viscous media that immobilize other bacterial species [25–28]. The structural form of spirochetes may aid pathogenic species in penetrating host tissues and disseminating throughout the host.

In contrast to those of free-living bacteria, the genome of B. burgdorferi is relatively small, probably reflecting its lifestyle as an obligate parasite. B. burgdorferi lacks the conventionally recognizable machinery for synthesizing nucleotides, amino acids, fatty acids, and enzyme cofactors, apparently scavenging these necessities from the host [4]. The limited metabolic capacity of B. burgdorferi requires a complex and chemically undefined growth medium for in vitro cultivation.

Over the last 15 years, substantial advances have been made in the ability to genetically modify B. burgdorferi (for a recent review see Rosa et al. [29]). Currently, targeted genes can be selectively inactivated and the resulting mutant strain can be tested for infectivity in an experimental mouse-tick cycle. In this way, putative virulence factors can be assessed in the laboratory. Using the opposite approach, transposon mutagenesis randomly inactivates a large number of genes, with each mutation occurring in an individual clone [30, 31]. Potentially, batches of these random mutants can be tested together for infectivity in the mouse model and specific mutants that no longer infect or persist can be isolated. Use of this method on a limited scale identified several genes that appear to contribute to B. burgdorferi infectivity in the mouse [32].

B. burgdorferi life cycle

B. burgdorferi infects a wide range of vertebrate animals including small mammals, lizards, and birds [33–39]. Ticks of the genus Ixodes transmit B. burgdorferi between hosts and are the only natural agents through which humans have been shown to become infected [2, 40]. Worldwide geographic distribution of Lyme disease correlates to the overlapping ranges of both a competent reservoir host for B. burgdorferi and the tick vector. In the northeastern and midwestern United States, the primary tick species for human disease is Ixodes scapularis (the black-legged tick) and in the western states I. pacificus (the western black-legged tick) is the main agent of dissemination [41]. European and Asian Lyme disease agents are primarily transmitted by I. ricinus (the European sheep tick) and I. persulcatus (the taiga tick), respectively [42, 43].

Ticks most frequently acquire spirochetes from infected rodents during their larval feeding [36, 44]. After molting to the nymphal stage, infected ticks feed on a broad range of animals, including rodents, which become a new reservoir perpetuating the cycle [40]. After the nymphs molt to the adult stage, they exclusively feed on larger mammals, which are often not competent hosts for B. burgdorferi [40]. The spirochetes are rarely, if ever, transmitted trans-ovarially [45, 46], so the larval and nymphal feedings are crucial to maintaining the spirochete. Both nymphs and adults occasionally feed on humans, but the small size of the nymphs makes them difficult to detect and, hence, more likely to feed long enough to transmit the spirochete and cause Lyme disease.

The tick and mammalian hosts provide contrasting environments for bacterial growth. Notably, mammals regulate their body temperatures at about 37–39°C, whereas ticks vary with the ambient temperature, except when feeding on a mammal. Also, the pH of mammalian tissue and blood is neutral, whereas the tick midgut is a more basic environment [47, 48]. In order to cycle between two very different hosts, B. burgdorferi varies its gene expression, leading to different protein components and enabling physiological adaptation to these environments [49–53]. A number of studies have begun to delineate those changes.

Mammalian components affecting B. burgdorferi infection

Several different experimental animals have been used for laboratory studies of B. burgdorferi infection and disease (reviewed by Philipp and Johnson [54]). Hamsters become infected in multiple tissues but do not show signs of disease [55], although arthritis can be induced in vaccinated or irradiated hamsters [56, 57]. Rabbits develop skin manifestations similar to the rash erythema migrans found in human disease, but clear the infection relatively rapidly [58, 59]. Dogs develop arthritis and become persistently infected [60]. Infection of rhesus monkeys more closely resembles human infection, with some monkeys exhibiting skin rash, arthritis, and neuroborreliosis [61–64]. Because mice are small and have extensively characterized genetics, many scientists use a laboratory model of the natural infectious cycle involving mice and Ixodes ticks to assess host and bacterial components that allow perpetuation of the cycle and, in some cases, disease.

In the murine model, mice can be infected by needle inoculation or tick feeding. Most mouse strains, as well as the natural reservoir hosts, show no sign of disease, but do develop a serological response to B. burgdorferi proteins and become persistently infected [65, 66]. Some mouse strains are considered to be disease-susceptible, in that they develop ankle swelling and inflammation that resembles arthritis in response to B. burgdorferi infection [66]. Other strains are disease resistant and considerable effort has been made to determine the factors and genes responsible for differences in susceptibility among mice [67–75]. In some cases, the differential severity of joint pathology exists despite similar numbers of spirochetes in the affected tissues [68]. These data suggest that disease is a consequence of host immunopathology, rather than a strategy of the bacterium to facilitate its persistence or transmission.

When B. burgdorferi enter a mouse, many components of the host innate immune response could potentially recognize the bacteria and help control their numbers. Among these are antigen presenting cells (such as macrophages and dendritic cells) in the peripheral tissues (e.g., at the site of the tick bite). The antigen presenting cells may subsequently migrate to lymph nodes, and stimulate T cell and B cell responses. Killing of B. burgdorferi by the phagocytic cells resident in the periphery and perhaps neutrophils attracted to the feeding lesion or inoculation site may also contribute to control of the initial inoculum. Complement may also help control B. burgdorferi numbers by opsonizing the bacteria (facilitating phagocytosis) or by direct killing via the alternative pathway.

To determine what components of the innate and acquired immune response recognize B. burgdorferi, and limit bacterial numbers and their ability to cause disease, infection has been characterized in mouse strains with mutations affecting a number of those components. These and other studies have identified a particular pattern receptor (TLR2) as key to mammalian recognition of B. burgdorferi lipoprotein antigens [76–78]. Mice deficient in TLR2 have increased spirochete loads and ankle swelling. However, mice lacking MyD88, an adapter required for signaling through TLR2 and several other pattern receptors, have even higher spirochete loads, suggesting multiple pathways respond to B. burgdorferi antigens [79–82]. Most likely, the phenotypes of these mutations result from poor recognition of B. burgdorferi components by phagocytic and antigen-presenting cells. Complement was first implicated in controlling spirochete numbers and perhaps in determining preferred reservoir hosts in a study that correlated sensitivity to various host sera with particular spirochete genospecies [83]. Complement was further implicated by infection of mice defective in component C3 production, which results in somewhat higher numbers of spirochetes in tissues, especially at early times after infection [84, 85]. In summary, several innate immune system components recognize spirochetes and control their numbers but, in most cases, are inadequate to completely clear an infection.

After the initial stage of B. burgdorferi infection, mice develop antibodies to numerous bacterial proteins [86, 87]. When serum from infected mice was transferred to naïve mice, the recipient animals were protected from infection with the same strain of B. burgdorferi [88, 89]. Similar transfer of T cells did not confer protection [90], suggesting that T cells play a lesser role than antibody in protection. Despite the presence of neutralizing antibodies, the host acquired immune response limits spirochete numbers but does not eradicate B. burgdorferi infection and most mice become persistently infected after needle inoculation or the bite of an infected tick. Antibody does serve to limit disease and pathogenesis, since SCID mice, which lack both the cellular and antibody components of the acquired immune response, contain much higher spirochete loads in tissues and exhibit more severely arthitic joints than normal mice [91]. The ability of the bacteria to survive in the face of an antibody response suggests that either the bacteria “hide out” in sites protected from antibodies or that the bacteria evade antibody reactivity by varying antigens or otherwise masking reactive proteins. The low numbers of spirochetes typically found in infected mammalian blood limits direct detection of bacteria in human clinical samples, complicating diagnosis and analysis of the progress of an infection. Sensitive molecular techniques now allow amplification of bacterial targets, but the bacterial numbers are at the lower limit detectable by such methods, so they are difficult to apply in the clinical laboratory.

Studies of B. burgdorferi factors required for mammalian and tick infection began with identifying changes in protein composition as the spirochete alternated between these disparate environments [49, 51, 92]. As microarray analysis was developed, this technique was applied to measuring differences in gene expression between bacteria grown in conditions that mimic some characteristics of each host environment [93–95]. A variant on this has been to compare the gene expression of cultured spirochetes with that of bacteria grown in an immune-privileged chamber within a rat, allowing host-adaptation [96]. Other studies have used the polymerase chain reaction (PCR) to follow the in vivo expression of genes highlighted in the microarray analysis, using samples of infected mammalian and tick tissues [97–101]. These experiments demonstrated that the bacterial gene expression profile changes, not only between hosts, but also at different times within a single host. Finally, one group has corroborated some of the gene expression patterns identified in these studies by directly examining the proteins made by bacteria growing in SCID mouse and rabbit skin, taking advantage of the larger number of bacteria found in these models [102, 103]. Using these methods, researchers have identified genes whose products are, or are likely to be, produced in mice or ticks, some of which have been tested for their contribution to bacterial survival in the corresponding host (Fig. 1).

Fig. 1.

B. burgdorferi outer surface proteins are differentially regulated in response to host conditions. Spirochetes remodel their outer surface in different host environments, represented above by different colors. In the unfed tick, B. burgdorferi (represented in blue) produce a variety of proteins, such as OspA, to persist within the tick midgut for extended periods of time. Once a tick has attached to a vertebrate host, B. burgdorferi (now represented in red) expresses other proteins (e.g. OspC) in preparation for transmission to the new host. During infection of the mammalian host, B. burgdorferi (colored in yellow) expresses a variety of other proteins (including VlsE), presumably to survive attack by the host immune system, disseminate to distant sites within the host, and acquire specific nutrients. This figure shows representative proteins and is not meant to be a comprehensive list.

B. burgdorferi products required for mammalian infection

Efforts to identify B. burgdorferi products that play roles in mammalian infection began by serially passaging strains in culture and correlating loss of infectivity with loss of specific plasmids [55, 104–108]. As genetic techniques have become available for B. burgdorferi analysis, other approaches have been applied to identifying genes that contribute to mammalian infectivity and the roles of some such genes have been rigorously defined by inactivation and restoration (complementation). The products of these genes can be divided into ones that play physiological roles and those that contribute to survival in the face of other aspects of the host environment, including normal defenses. Since B. burgdorferi has a limited biosynthetic capability, the bacteria rely on their host (or culture medium) for many nutrients that other bacteria can synthesize. Some enzymes shown to be crucial for bacterial survival in a mammalian host (although dispensable for growth in rich culture medium) include PncA, a nicotinamidase involved in production of NAD [109], and two products involved in purine nucleotide synthesis [32, 110]. Another unusual feature of B. burgdorferi is that it does not contain intracellular iron and, hence, does not use iron as a co-factor for enzymes [111], thereby facilitating survival in the iron-poor mammalian environment.

Factors important for bacterial survival in a mammalian host can be subdivided into two groups. Some are involved in early infection, such as OspC [112–116], which are presumably required for host colonization or resistance to innate immunity. Others are involved in resistance to acquired immunity, such as VlsE [13, 109, 117, 118].

The OspC product has been shown to be required for an early stage of mammalian infection [113–116] and conflicting data have been presented regarding its importance in tick transmission [113, 114, 119, 120]. OspC production begins in feeding ticks (or immediately after needle inoculation, Fig. 1) and lasts for the first couple of weeks of mammalian infection [49, 51, 103, 121–123]. Once the bacteria are established in a host, OspC production is not required for persistence [114]. The molecular function of the OspC protein is undefined, although the crystal structure of the protein predicts that it binds a small ligand [124, 125]. The gene product contains a highly variable region that has allowed characterization of OspC types [126–128]. Studies of the significance of OspC type have reached contradictory conclusions. Some studies have found a correlation between OspC type and wild rodent host specificity [129], but others found no such link [130]. The OspC types of human clinical isolates have also been correlated with invasive and non-invasive phenotypes [128, 131]. However, similar studies demonstrated greater diversity among invasive types than previously recognized, calling into question any causative effect of OspC sequence on invasion [130, 132]. Carefully controlling for variables other than OspC has not been possible in such studies, since additional genome components of these isolates also differ. Recently, the invasive OspC protein types were shown to bind plasminogen [131], a trait shared with other B. burgdorferi proteins [133, 134] that was suggested to facilitate B. burgdorferi infection of mammals and ticks [135].

Considerable attention has been paid to several B. burgdorferi proteins that co-opt the complement regulators factor H and factor H-like proteins, which normally protect mammalian cells from attack by their own complement by blocking activation of the alternative pathway [136–141]. Surface coating with factor H could protect the bacteria from killing or opsonization by host complement, facilitating infection. Surprisingly, however, B. burgdorferi infection of mice deficient in factor H production did not differ from infection of wild type mice, suggesting that any protection conferred by factor H binding was unimportant or redundant [85]. Because of these findings, the significance of the factor H-binding proteins for B. burgdorferi remains unknown.

VlsE is a B. burgdorferi protein that is required for persistent mammalian infection [13, 109, 117, 118] and whose synthesis begins around the time that OspC production ceases (Fig. 1)[103]. Although its function is unknown, this lipoprotein has an elaborate system for variation [22]. Variation at the VlsE locus appears to be required for persistence [118, 142], probably because its (unknown) essential function requires its presence on the bacterial surface, where it will be targeted by the adaptive immune response of the mammalian host.

Since B. burgdorferi can survive in the face of a neutralizing antibody response, the idea that the bacteria shelter in tissues with little exposure to antibodies by interacting with the host extracellular matrix (ECM) has been investigated (reviewed by Cabello et al. [143] and Coburn et al.[144]). Among the B. burgdorferi proteins that bind ECM components are DbpA and DbpB, which bind decorin [145], BBK32, which binds fibronectin [146], Bgp, which binds proteoglycans [147], and P66, which binds integrins [148]. These proteins may help B. burgdorferi migrate through mammalian tissue and persist in joints and skin, where the spirochetes may be inaccessible to circulating antibodies. Genetic studies, however, have shown that spirochetes lacking DbpA have little if any defect in mammalian infectivity [149]. Two studies of mutants lacking BBK32 also found a small [150] or no [151] decrease in infectivity. These studies call into question whether interaction with host ECM protects the bacteria. Alternatively, the multiplicity of ECM-binding proteins may ensure such interaction by providing redundant binding proteins so that the loss of any single protein may fail to yield a noticeable phenotype.

An accumulating body of evidence demonstrates that a cascade of transcriptional regulators, encoded by the rpoS and rpoN genes, controls the production of a number of lipoproteins in response to changing environmental factors [152, 153]. A sensor-regulator pair of proteins also contributes to this regulation [154]. When B. burgdorferi mutants were tested in the infection model, rpoS and rpoN expression were found to be required for infecting mice [153, 155], demonstrating the importance of this pathway for survival in a mammal.

B. burgdorferi infection of ticks

While infected nymphal ticks feed, spirochetes in the midgut respond in several ways to the incoming blood and increased temperature. The population of spirochetes expands [156–158] and their protein synthesis alters [49, 99, 122, 159, 160]. Then, spirochetes migrate from the midgut to the salivary glands, allowing transmission into a new host. The B. burgdorferi outer surface protein OspA was shown to be abundant on the surface of bacteria resident in ticks (Fig. 1), but down-regulated during tick feeding and transmission to a mammal [49, 160]. Subsequent studies suggest that OspA is an adhesin, important for retaining spirochetes in the tick midgut until feeding [161–163]. OspB, another potential midgut adhesin [164], BptA, a lipoprotein of unknown function, and the product of the BB0690 gene, which is probably involved in resistance to oxidative stress [165], also appear to contribute to bacterial survival in ticks [166–168]. The RpoN-RpoS regulatory cascade appears to be required for migration of spirochetes to the salivary glands during transmission, but not for survival within the tick environment [153]. Since the tick and mammalian environments differ significantly from each other, there are likely to be other B. burgdorferi proteins that carry out important roles during bacterial growth and survival in ticks.

Complementary studies have begun to elucidate tick proteins that contribute to B. burgdorferi infection and transmission. Ribeiro and colleagues identified transcripts that were differentially regulated between B. burgdorferi-infected and uninfected I. scapularis salivary glands [169]. A recent genetic study showed that the midgut protein TROSPA is a receptor for OspA binding, whose presence enhances colonization by B. burgdorferi [170]. The salivary gland protein Salp15 is immunosuppressive and may facilitate infection by the low numbers of spirochetes that are transmitted during tick feeding [171, 172]. Recently, a tick antioxidant was shown to facilitate tick acquisition of spirochetes from infected animals [173]. The Ixodes scapularis genome project, which is in progress, should yield additional gene candidates whose influences, positive or negative, on B. burgdorferi infection remain to be elucidated.

Conclusions and Perspectives

B. burgdorferi produces a number of products that allow it to colonize and persist in its natural mammalian and tick hosts. Although the functions of only a few B. burgdorferi products have been clearly defined, some (such as OspC) are required for the bacteria to survive the initial attack of the mammalian innate immune system, while others (like VlsE) contribute to resisting the subsequent acquired immune response. Bacterial factors such as RpoS and RpoN are components of signaling cascades regulating gene expression for survival in different environmental conditions. With the powerful genetic techniques now available for manipulating the spirochete, the mouse, and even the tick, the interactions among these three that lead to infection and disease are beginning to emerge.