Human Borrelia miyamotoi infection in California: Serodiagnosis is complicated by multiple endemic Borrelia species

Abstract

To determine whether human Borrelia miyamotoi infection occurs in the far-western United States, we tested archived sera from northwestern California residents for antibodies to this emerging relapsing fever spirochete. These residents frequently were exposed to I. pacificus ticks in a region where B. miyamotoi tick infection has been reported. We used a two-step B. miyamotoi rGlpQ assay and a B. miyamotoi whole-cell lysate (WCL) assay to detect B. miyamotoi antibody. We also employed Borrelia hermsii and Borrelia burgdorferi WCL assays to examine if these Borrelia induce cross reacting antibody to B. miyamotoi. Sera were collected from 101 residents in each of two consecutive years. The sera of 12 and 14 residents in years one and two, respectively, were B. miyamotoi rGlpQ seroreactive. Sufficient sera were available to test 15 of the 26 seropositive samples using B. miyamotoi and B. hermsii WCL assays. Two residents in year one and seven residents in year two were seroreactive to both Borrelia antigens. Although discernible differences in seroreactivity were evident between the B. miyamotoi and B. hermsii WCL assays, infection with one or the other could not be determined with certainty. Sera from two Borrelia burgdorferi /B. miyamotoi seropositive subjects reacted strongly against B. miyamotoi and B. hermsii WCL antigens. Ecological, epidemiological, and clinical data implicated B. miyamotoi as the probable cause of infection among those whose sera reacted against both antigens. Our findings suggest that human B. miyamotoi infection occurs in northern California and that B. hermsii and B. burgdorferi infections produce antibodies that cross-react with B. miyamotoi antigens. Health care professionals in the far-western United States should be aware that B. miyamotoi disease may occur throughout the geographic distribution of I. pacificus and that improved relapsing fever group spirochete antibody assays are urgently needed.

Introduction

Borrelia miyamotoi is a relapsing fever-group spirochete that was discovered in Ixodes persulcatus ticks in Japan more than 20 years ago and later determined to cause clinical illness in humans [1–9]. This spirochete can cause a febrile viral-like illness that relapses in up to 10% of patients [2, 5–6]. Immunocompromised patients may experience meningoencephalitis [3, 7–8]. B. miyamotoi is widespread in the United States in Ixodes scapularis ticks in the Northeast and upper Midwest and in Ixodes pacificus ticks in the Far West [10–16]. Human cases of relapsing fever due to B. miyamotoi (hard tick-borne relapsing fever) have been described in the Northeast and upper Midwest, as well as in Russia, the Netherlands, Germany, and Japan [2–9]. In the Northeast, B. miyamotoi seroprevalence is estimated to be approximately 1 to 3%, which is about one-tenth to one-third that of Lyme disease [4, 17]. No human cases of B. miyamotoi previously have been reported from the western United States even though I. pacificus ticks in northern California have a spirochete-infection prevalence similar to or exceeding that of I. scapularis ticks in the Northeast and upper Midwest [11–16].

To determine whether human B. miyamotoi infection occurs in the far-western United States, we used a two-step B. miyamotoi rGlpQ antigen-based antibody assay to test archived sera from residents of a small rural community (population ~150) in northern California. This particular community was located in ecologically diverse Mendocino County, a region where B. miyamotoi and Borrelia burgdorferi-infected I. pacificus ticks repeatedly have been found [Massachusetts General Hospital Tick-borne Diseases Conference; June 17–20, 2016, Boston, Massachusetts, USA] [11, 14–16, 18–19]. Seroprevalence determination with the B. miyamotoi GlpQ assay is not affected by Lyme disease infection because B. burgdorferi does not produce GlpQ antigen, however, several soft tick-borne relapsing fever Borrelia species that are endemic in the western United States do produce GlpQ and thus might elicit cross-reacting antibodies against rGlpQ or other B. miyamotoi antigens [20]. We therefore tested the same archived sera for antibodies against the relapsing fever spirochetes Borrelia hermsii and B. miyamotoi in whole cell lysate (WCL) assays.

Materials and methods

Human study population

Serum samples were obtained in 1988 and 1989 from 101 residents of a community at high risk for Lyme disease (CHR) located in the Ukiah area of southern Mendocino County [18–19]. These sera previously were tested for B. burgdorferi seroreactivity as part of an inter- and intra-laboratory comparative study [18]. Since then, the sera were maintained at -80°C, although they were frozen and thawed a few times prior to B. miyamotoi testing. In the initial serosurvey [19], subjects were asked to recall any previous Lyme disease diagnosis, history of tick bite in the previous two years, and signs or symptoms suggestive of Lyme disease. The study was carried out with the approval of the Committee for Protection of Human Subjects at the University of California, Berkeley.

Borrelia miyamotoi rGlpQ ELISA and Western blot assays

Serum samples were diluted 1:320 and tested for IgG antibodies against recombinant B. miyamotoi glycerophosphodiester phosphodiesterase antigen (rGlpQ) using an IgG ELISA [17]. As a negative control for each ELISA plate, we used sera from three healthy participants who had no history of tick bite or tick-borne disease but who lived in an area where Lyme disease is endemic. A signal ≥3 SD above the mean of three non-infected serum controls was considered positive for B. miyamotoi antibody. As positive controls, we used sera from patients who were confirmed by PCR or serology to have B. miyamotoi infection.

Serum samples that were equivocal or positive by ELISA were retested for antibodies against rGlpQ using a Western blot IgG antibody assay [17]. Nitrocellulose membrane strips were individually incubated with human serum at a 1:250 dilution. Samples with a 39-kDa band were counted as Western blot positive. Serum samples were considered B. miyamotoi seropositive if both ELISA IgG and Western blot IgG tests yielded positive results.

Borrelia miyamotoi and Borrelia hermsii WCL Western blot antibody assays

Serum samples diluted 1:100 were assayed for antibodies against B. miyamotoi and B. hermsii WCL antigens in Western blot assays [21]. For both the B. miyamotoi and B. hermsii set of blots, we included serum from a subject who resides in a non-Lyme disease endemic area and had no history of prior B. miyamotoi infection (negative control). For both sets of blots, we also included serum from a subject who had PCR-confirmed B. miyamotoi infection (positive control) and from another subject who had confirmed B. hermsii infection (positive control). Two of us (PJK and SN) independently counted the number of bands for the positive and negative control sera and the CHR sera that were tested using B. miyamotoi and B. hermsii whole lysate assays. A seronegative serum was defined as one with the same or fewer bands as the negative control sera. A seroreactive serum was defined as one that had one or more bands than the negative control. Seroreactive sera were further classified into “moderately reactive” if they had one or more bands than the negative control but fewer bands than the positive control or “strongly reactive” if they had at least as many bands as the positive control.

Borrelia burgdorferi WCL Western blot antibody assay

Serum samples from patients with physician diagnosed Lyme disease (erythema migrans rash and B. burgdorferi seropositive) and no history of B. miyamotoi infection, and the same B. miyamotoi seropositive and seronegative controls as described above were diluted 1:100 and assayed for antibodies against B. miyamotoi and B. burgdorferi. We used WCL antigens prepared from in vitro cultivated B. miyamotoi or B. burgdorferi in separate Western blot assays [22–23].

Results

Seroreactivity to B. miyamotoi GlpQ antigen

Clinical and laboratory data were evaluated among 101 of the original 119 CHR subjects whose archived serum samples were available for B. miyamotoi serologic testing. The mean age of the 101 subjects was 32 years (range 4 to 52 years) and participants included similar percentages of females (51%) and males (49%). Using the two-step ELISA-Western blot rGlpQ assay, we found that sera from 12 subjects in 1988 and 14 subjects in 1989 reacted against B. miyamotoi rGlpQ antigen (Table 1 and S1 Table).

Table 1. B. miyamotoi rGlpQ ELISA and Western blot test results among CHR subjects.

Study subject	ELISA optical density	ELISA optical density value > positive control cutoff	rGlpQ Western blot
CHR-23 88	0.147	0.037*	POSITIVE
CHR-24 88	0.141	0.031	POSITIVE
CHR-39 88	0.078	0.001	POSITIVE
CHR-40 88	0.082	0.005	POSITIVE
CHR-47 88	0.107	0.030	POSITIVE
CHR-50 88	0.082	0.005	POSITIVE
CHR-52 88	0.087	0.010	POSITIVE
CHR-58 88	0.087	0.010	POSITIVE
CHR-64 88	0.112	0.045	POSITIVE
CHR-68 88	0.093	0.026	POSITIVE
CHR-81 88	0.092	0.025	POSITIVE
CHR-84 88	0.175	0.041	POSITIVE
CHR-26 88	0.059	negative	negative
CHR-45 88	0.093	0.016	negative
CHR-60 88	0.057	negative	negative
CHR-78 88	0.059	negative	negative
CHR-23 89	0.175	0.065	POSITIVE
CHR-26 89	0.080	0.013	POSITIVE
CHR-39 89	0.133	0.056	POSITIVE
CHR-45 89	0.090	0.013	POSITIVE
CHR-47 89	0.143	0.066	POSITIVE
CHR-58 89	0.101	0.034	POSITIVE
CHR-59 89	0.082	0.015	POSITIVE
CHR-60 89	0.123	0.056	POSITIVE
CHR-65 89	0.130	0.063	POSITIVE
CHR-69 89	0.095	0.028	POSITIVE
CHR-78 89	0.080	0.013	POSITIVE
CHR-92 89	0.176	0.060	POSITIVE
CHR-112 89	0.125	0.019	POSITIVE
CHR-116 89	0.109	0.003	POSITIVE

*The negative control values used in the test run for CHR-23 88 were 0.085, 0.065, 0.060. The ≥3 SD cutoff value above the mean of the negative control values was 0.110 (positive cutoff). The optical density reading for CHR-23 88 was 0.147. Thus, the CHR-23 88 value was 0.037 above the positive cutoff value.

Seroreactivity to B. miyamotoi and B. hermsii WCL antigen

A sufficient volume of sera was available from 15 of 26 B. miyamotoi GlpQ seroreactive subjects for B. miyamotoi and B. hermsii WCL antibody testing. Although discernible differences in seroreactivity against B. miyamotoi and B. hermsii WCL were noted between the CHR study subjects, infection with one or the other of them could not be distinguished with certainty (Fig 1). Two coauthor analysts (PJK and SN) independently assessed the number of bands on the Western blots to determine bands that represented specific reactivity compared to the negative control sera. They scored the specimens similarly (Table 2). Two of four B. miyamotoi GlpQ antibody positive subjects and seven of 11 B. miyamotoi GlpQ antibody positive subjects were seroreactive against B. miyamotoi and B. hermsii WCL antigens in 1988 and 1989, respectively.

Fig 1. B. miyamotoi and B. hermsii WCL antigen Western blot reactivity in serum samples from California subjects whose sera reacted against B. miyamotoi rGlpQ antigen.

Numbers at the top of the rows are subject numbers. For both the B. miyamotoi and B. hermsii sets of blots, lane 1 displays the test result for a subject who resided in a non-Lyme disease endemic area and had no history of prior B. miyamotoi infection (negative control). The blots in lane 2 display the test results for a subject who had confirmed B. hermsii infection (positive control). The blot in lane 3 displays the test results for a subject who had PCR-confirmed B. miyamotoi infection (positive control). The positive and negative control sera are the same as those described in Table 2. The other blots display the test results for 15 of 26 B. miyamotoi GlpQ seropositive CHR subjects who had a sufficient volume of serum remaining to perform B. miyamotoi and B. hermsii WCL antibody testing.

Table 2. B. miyamotoi and B. hermsii WCL seroreactivity in CHR subjects who were B. miyamotoi rGlpQ seroreactive.

Subject	B. miyamotoi WCL	B. hermsii WCL
1988
CHR-52	Seroreactive (moderate)a	Seroreactive (moderate)
CHR-84	Seroreactive (moderate)	Seroreactive (moderate)
CHR-23	Seronegative	Seronegative
CHR-50	Seronegative	Seronegative
1989
CHR-26	Seroreactive (moderate)	Seroreactive (moderate)
CHR-39	Seroreactive (moderate)	Seroreactive (moderate)
CHR-45	Seroreactive (strong) b	Seroreactive (moderate) b
CHR-65	Seroreactive (strong) b	Seroreactive (strong) b
CHR-69	Seroreactive (moderate)	Seroreactive (moderate)
CHR-78	Seroreactive (moderate)	Seroreactive (moderate)
CHR-116	Seroreactive (moderate)	Seroreactive (moderate)
CHR-58	Seronegative	Seronegative
CHR-59	Seronegative	Seronegative
CHR-60	Seronegative	Seronegative
CHR-112	Seronegative	Seronegative

a Determination of seroreactivity based on the number of bands on the Western blot by two independent analysts. See Results above for definition of strong, moderate and negative reactivity.

b These sera were also B. burgdorferi seropositive

Seroreactivity to B. miyamotoi and B. burgdorferi WCL antigens

Two CHR residents (CHR 45 and CHR 65) who were seropositive for B. miyamotoi GlpQ and WCL also were seropositive for B. burgdorferi. They had coinfection or sequential infection with these two Borrelia. Their sera reacted more strongly to B. miyamotoi WCL antigen than did any of the other CHR B. miyamotoi seropositive sera (Fig 1). To determine if the strong reaction against B. miyamotoi WCL antigen among these two B. miyamotoi and B. burgdorferi-infected subjects was due in part to B. burgdorferi cross-reacting antibody, we tested sera from patients who had experienced B. miyamotoi infection alone (positive B. miyamotoi control subject), Lyme disease alone (erythema migrans rash and B. burgdorferi seropositive using the standard ELISA-Western blot assay), and neither infection (negative control) using a B. burgdorferi WCL assay and a B. miyamotoi WCL assay using a different B. miyamotoi WCL assay than that described above. The sera from patients with Lyme disease were reactive against B. burgdorferi WCL and against several B. miyamotoi bands in the B. miyamotoi WCL assay (Fig 2B) but did not react to B. miyamotoi GlpQ antigen in a Western blot assay (Fig 2A). The sera from patients with B. miyamotoi were seroreactive against several B. burgdorferi bands in the B. burgdorferi WCL assay (Fig 2B).

Fig 2. B. burgdorferi and B. miyamotoi WCL Western blot reactivity in serum samples from New England residents whose sera were reactive against B. burgdorferi and B. miyamotoi antigen.

2A. B. miyamotoi rGlpQ Western blot results of sera from patients who previously had B. miyamotoi sensu lato infection alone (Bmsl positive control subject, PCR confirmed B. miyamotoi infection) or Lyme disease alone (Bb positive control, erythema migrans rash and B. burgdorferi seropositive using the standard two-step ELISA and Western blot assay). The arrow indicates the rGlpQ specific band. 2B. B. miyamotoi and B. burgdorferi WCL Western blot results of sera from the same patients as in Fig 2A.

Ecological and epidemiological evidence of B. miyamotoi infection

Ecological and epidemiological evidence indicates that CHR subjects whose sera reacted against B. miyamotoi GlpQ and WCL antigens and B. hermsii WCL antigen were much more likely to have been infected with B. miyamotoi than either B. hermsii or Borrelia parkeri, two other relapsing fever spirochetes present in northern California. B. miyamotoi has been detected in ticks from Mendocino County but not B. hermsii or B. parkeri. Indeed, human cases caused by the latter two spirochetes are acquired distant from the CHR community and the incidence of human infection with them is either low (B. hermsii) or rare (B. parkeri) statewide. A high percentage (78%) of CHR residents reported tick bite, which is more consistent with hard-bodied than soft-bodied ticks. The latter feed for less than an hour at night and are very seldom noticed in people who suffer soft tick-borne relapsing fever. One or more relapsing fever episodes typically occur in patients infected with all soft tick-borne relapsing fever infections but with only 10 percent or fewer people experiencing B. miyamotoi infection. No relapsing fever episodes were reported by CHR residents in 1988 and 1989.

Discussion

Our principal aim was to determine whether residents of the far-western United States are at risk of human B. miyamotoi infection. The CHR residents were an ideal group to screen because B. miyamotoi-infected I. pacificus ticks have been identified from nearby areas in Mendocino County. Moreover, the members of this rural community experienced extraordinarily high tick exposure rates that far exceed published exposure rates from the Northeastern United States [19, 24, 25]. It was therefore not surprising that 2 of 101 CHR residents in 1988 and 7 of 101 CHR residents in 1989 were seroreactive to B. miyamotoi GlpQ and WCL antigens. Although these residents were also seroreactive to B. hermsii antigens, this was likely a result of B. miyamotoi cross reacting antibody, rather than previous B. hermsii infection. Human B. hermsii infection is infrequently reported in mountainous areas far to the east of Mendocino County and none of the residents reported having experienced relapsing fever episodes that are characteristic of B. hermsii during the two-year study period. Although we found that B. burgdorferi, like B. hermsii, can elicit cross-reacting antibody against B. miyamotoi antigens, none of our putative B. miyamotoi seropositive CHR residents could have had Lyme disease alone because they reacted against B. miyamotoi GlpQ antigen and B. burgdorferi does not produce GlpQ [20]. In sum, our data suggests that human B. miyamotoi infection occurs in the Far West and that B. miyamotoi, B. hermsii and B. burgdorferi spirochetes share many proteins that elicit cross reacting antibodies, which complicate the serodiagnosis of B. miyamotoi and other relapsing fever infections.

Several ecological and epidemiological observations support the possibility of human B. miyamotoi infection in northern California. We found that between 2 and 7 of the 101 CHR residents were seroreactive against B. miyamotoi GlpQ and WCL antigen in 1988 and 1989, respectively. This seropostitivity range approximates the 0.5% to 15% B. miyamotoi infection range that has been found in I. pacificus ticks in Northern California [11, 14, 15]. The CHR residents built their homes on a former cattle ranch during the 1970s where I. pacificus and its manifold vertebrate and reservoir hosts abound [19]. Over three quarters (78%) of them reported tick-bites 1–2 years prior to enrollment and in a follow-up study, 79% had elevated anti-I. pacificus saliva antibodies [24]. By comparison, these tick-exposure rates are nearly triple the tick-bite incidence reported for residents of Block Island, Rhode Island (29%) where Lyme disease and babesiosis are highly endemic [25, 26]. Other ecological factors likely contributing to the force of B. miyamotoi transmission to humans at the CHR include the prolonged activity periods of I. pacificus nymphs and adults (compared to the lesser year-round activity periods of I. scapularis in the Northeast) that collectively span most of the year, as well as the occurrence of transovarial transmission of B. miyamotoi in its primary Ixodes spp. tick vectors [27, 28]. It is not likely that CHR residents were infected with either B. hermsii or B. parkeri. Neither of these two soft tick-borne relapsing fever species have been detected in ticks from Mendocino County [11, 14, 15]. B. hermsii infection incidence averages only 5 to 7 reported cases per year in California, and it occurs mainly in high mountains located about 500 to 1000 km to the east of the CHR [29, 30]. B. hermsii typically causes 1–10 febrile episodes accompanied by a viral-like illness that tends to occur in clusters among family members or friends who frequent the same mountain cabins. None of the CHR residents reported a previous diagnosis of soft tick-borne relapsing fever or episodes of a relapsing fever-like illness 1–2 years pre-entry or 1 year post-entry. B. parkeri, was incriminated in only a single human case in the distant Central Valley during the1940’s [31]. Human cases have not been reported due to a third soft tick-borne relapsing spirochete, B. coriaceae, first isolated from the soft tick Ornithodoros coriaceus in southern Mendocino County decades ago [32].

Although we found that B. burgdorferi also can elicit cross-reacting antibody against B. miyamotoi antigens, none of our putative B. miyamotoi seropositive CHR residents could have had Lyme disease alone because they reacted against B. miyamotoi GlpQ antigen and B. burgdorferi does not produce GlpQ [20]. Interestingly, the strongest reactions against B. miyamotoi WCL antigen of all the B. miyamotoi/ B. hermsii seropositive CHR sera were in the two residents who also reported previous Lyme disease. The intense reactivity to B. miyamotoi and B. hermsii WCL antigens is likely due in part to B. burgdorferi cross-reacting antibody [33]. Recent studies have shown that sera from B. miyamotoi-infected patients cross react to B. burgdorferi C6 peptide [16, 34]. Cross reacting antibody against B. miyamotoi/B. herrmsii WCL antigens also may have resulted from previous Haemophilus influenza or certain other gram-negative bacterial infections that produce GlpQ. The genetic difference between the GlpQ of these pathogens and B. miyamotoi is so wide, however, that significant cross-reactivity is unlikely [17, 20].

Our study was subject to several limitations. We enrolled a relatively small number of subjects and controls. No PCR-amplifiable B. miyamotoi DNA was detected in sera from CHR subjects, although these samples were obtained a few weeks to months after acute illness was likely to have occurred. The positive control sera used for the B. miyamotoi set of blots in Fig 1 was intensely positive and a number of bands likely bled into one another. We counted this central confluence as a single band but this had no effect on the number of seroreactive subjects as they were defined based on comparison with the number of bands in the negative control sera. B. burgdorferi antibody testing of the CHR residents was carried out before the development of the current two-tiered assay but the IFA assay that was used was determined to be highly sensitive and specific [18, 35]. Finally, archived sera had been frozen and thawed a few times before B. miyamotoi and B. hermsii antibody testing, but this is unlikely to have reduced antibody concentration.

Conclusions

Taken together, our serological, ecological and epidemiological data provide presumptive evidence that B. miyamotoi occasionally infects residents of Mendocino County in northwestern California. The presence of two soft-tick borne borreliae in California that either occasionally (B. hermsii) or rarely (B. parkeri) infect people in other counties, plus the occurrence of a third species (B. coriaceae) of uncertain human significance, precludes absolute confirmation that the seroreactive CHR residents were infected with B. miyamotoi. Our findings also highlight the pressing need to develop a serodiagnostic assay capable of differentiating all members of the relapsing fever group.

Supporting information

S1 Table. B. miyamotoi rGlpQ ELISA- Western blot results in CHR subjects, 1988–1989.

(DOCX)

Data Availability

All relevant data are within the paper and its Supporting Information files.

Funding Statement

This research was supported in part by generous gifts from The Gordon and Llura Gund Foundation to PJK, Tom Eames and Geneva Anderson to RSL and by NIH grant 1R56AI114859-01 to PJK, SN, and EF. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.