Are other tick-borne infections overlooked in patients investigated for Lyme neuroborreliosis? A large retrospective study from South-eastern Sweden

l.


Introduction
Tick-borne infections are increasing in the Northern hemisphere, and new pathogens causing human disease are being identified (Socolovschi et al., 2009).In Europe, the hard tick Ixodes ricinus is the predominant tick species, and several bacterial species spread by I. ricinus are known to cause human diseases, including the Borrelia burgdorferi sensu lato (s.l.) group of spirochaetes, Borrelia miyamotoi, Anaplasma phagocytophilum, Neoehrlichia mikurensis, and several species of the spotted fever group (SFG) rickettsiae (Socolovschi et al., 2009;Wass et al., 2019;Welinder-Olsson et al., 2010).These microorganisms have all been reported with different prevalence in collected ticks in Sweden (Andersson et al., 2013;(Henningsson et al., 2015b); Lindblom et al., 2016;Wilhelmsson et al., 2013).The spread of B. burgdorferi s.l. to the central nervous system (CNS) causes Lyme neuroborreliosis (LNB).Meningitis due to Rickettsia helvetica, a SFG rickettsiae and the most frequently detected Rickettsia species in Swedish ticks, has been described in humans in Sweden (Lindblom et al., 2016;Nilsson et al., 2010;Nilsson et al., 2011).Although rare, infection by A. phagocytophilum may cause meningoencephalitis or affect the peripheral nervous system, e.g., brachial plexopathy or an isolated facial palsy (Dumler et al., 2007;Young and Klein, 2007).Since the CNS symptoms of these three infections are similar, they can be difficult to distinguish clinically.
The first case of human disease caused by N. mikurensis was reported in 2010 in Sweden (Welinder-Olsson et al., 2010).The organism mainly affects immuno-compromised patients, although immuno-competent individuals may also develop serious complications and experience symptoms of fever and myalgia as well as suffer from thrombo-embolic events (Grankvist et al., 2014;Welinder-Olsson et al., 2010;von Loewenich et al., 2010).The estimated prevalence of N. mikurensis in I. ricinus in Sweden is 6.0% (Andersson et al., 2013).For this pathogen, dissemination to the CNS has not yet been described in either humans or animals.
Borrelia miyamotoi, which belongs to the relapsing-fever group of Borrelia, is spread by different tick species belonging to the genus Ixodes, including I. ricinus, and it was first described to invade the CNS in humans in 2013 (Gugliotta et al., 2013;Hovius et al., 2013).Apparently, meningoencephalitis mostly affects immuno-compromised individuals, although CNS infection in an immuno-competent patient was recently described.The patients presented with LNB-associated symptoms, such as headache, neck pain and fever (Henningsson et al., 2019).The prevalence of B. miyamotoi in ticks in Southern Sweden is approximately 1% (Wilhelmsson et al., 2013).
In addition to bacteria, I. ricinus transmits tick-borne encephalitis virus (TBEV) in Europe.TBEV is a neurotrophic virus that may cause meningoencephalitis with serious and sometimes fatal neurological complications.The prevalence of TBEV in engorged ticks, which have bitten humans, in Southern Sweden, is estimated to be 0.2% (Lindblom et al., 2014).
Apart from B. burgdorferi s.l., B. miyamotoi and TBEV, the pathogenicity of other tick-borne microorganisms and their potentials to cause CNS infection in humans is poorly understood.Since knowledge of the clinical presentations of these infections is scarce amongst health-care providers in general, investigations of these pathogens are not commonly performed.Thus, these infections, and possible co-infections with multiple tick-borne pathogens, may be overlooked and symptoms may overlap, for which reason is why a broader diagnostic approach should be considered.Furthermore, as immuno-suppressive treatment is becoming more common in the population, co-infections, such as babesiosis, human granulocytic anaplasmosis (HGA) and infections caused by B. miyamotoi may pose a greater risk and the clinical picture may differ amongst patients (Sanchez et al., 2016).
The aim of the present study was to investigate the presence of tickborne pathogens other than B. burgdorferi s.l. in the cerebrospinal fluid (CSF), using both molecular and serological methods, in patients who have been investigated for LNB.

Patients
Through laboratory databases, 600 patients were identified at the three accredited laboratories in Linköping, Jönköping and Kalmar, respectively, and they were consecutively chosen from analyses performed on patients under the investigation of LNB during the years 2009-2013.The selection criterion used was that for each patient at least 1 mL of the frozen sample volume of CSF and serum, respectively, was available.White blood cell counts in the CSF and serological analyses regarding B. burgdorferi s.l. were performed as part of the initial investigation at the time of sampling.All the patient samples had been stored at -20 For the 55 patients with CSF mononuclear pleocytosis (>5 × 10 mononuclear cells per litre of CSF), the sera were analysed for antibodies to B. miyamotoi, A. phagocytophilum, SFG rickettsiae, TBEV, and B. microti.Three patient sera were not analysed for antibodies to B. miyamotoi due to lack of sera.We did not have access to serological assays for other Babesia spp.and a serological assay detecting antibodies to N. mikurensis has not yet been developed.
Medical charts were reviewed for all the patients with CSF mononuclear pleocytosis and for the patients in whom N. mikurensis and Borrelia spp.DNA was detected in the serum by real-time PCR.

Total nucleic acid extraction and cDNA synthesis from CSF and serum samples
Total nucleic acids (NA) were extracted from 200 µL of CSF and  (Barbour, 1984).RNase-free water was used as a negative control.The NA were reverse-transcribed to cDNA using the illustra™ Ready-to-Go RT-PCR Beads kit (GE Healthcare Life Sciences, Little Chalfont, Buckinghamshire, UK).Fifteen microlitres of NA and 10 µL pd (N)6 random hexamer primers (0.25 µg/µL) were incubated for 5 min at 97 • C and then mixed with one RT-PCR bead dissolved in 25 µL RNase-free water.The mixture was incubated for 30 min at 42 • C, followed by 5 min at 97 • C, producing 50 µL of cDNA.All PCR analyses were performed at the Department of Clinical Microbiology, County Hospital, Jönköping, Sweden.

Detection of Borrelia spp. in CSF and serum samples
Detection of Borrelia spp.(including B. miyamotoi) was performed using a genus-specific TaqMan real-time PCR assay, as described in the supplementary.The primers Borrelia_F and Borrelia_R, and the probe Borrelia_P were designed to target the Borrelia spp.16S rRNA gene, to amplify a 116-bp long amplicon (Table 1).Positive samples were reanalysed using a B. miyamotoi species-specific TaqMan real-time PCR assay, as previously described (Hovius et al., 2013).The primers Bm_F and Bm_R, and the probe Bm_P are designed to target the B. miyamotoi flagellin B gene (flaB) to amplify a 156-bp long amplicon (Table 1).To determine other B. burgdorferi s.l.species of the samples positive in the real-time PCR assay, a nested, conventional PCR assay using primers targeting the intergenic spacer region between 5S and 23S rRNA genes, was applied as previously described (Postic et al., 1994;Wilhelmsson et al., 2010).Nucleotide sequencing of the PCR products amplified by the conventional PCR assay to determine species of Borrelia was performed by Eurofins GATC Biotech (Konstanz, Germany).All sequences were confirmed by sequencing both strands.The obtained chromatograms were initially edited and analysed using BioEdit Software v7.0 (Tom Hall, Ibis Therapeutics, Carlsbad, CA), and the sequences were examined using Basic Local Alignment Tool.

Detection of Anaplasma phagocytophilum in CSF and serum samples
Detection of A. phagocytophilum was performed using a TaqMan realtime PCR assay, as previously described ((Henningsson et al., 2015a).The primers ApF and ApR and the probe ApM are designed to target the A. phagocytophilum citrate synthase gene (gltA), to amplify a 64-bp long amplicon (Table 1).As a positive control, a synthetic plasmid containing the target sequence of the TaqMan real-time PCR assay was used.The plasmid contained the target sequence, spanning nucleotides 304-420 of the A. phagocytophilum gltA gene (GenBank: AF304137), synthesised and cloned into the pUC57 vector (GenScript Inc., Piscataway, NJ, USA).

Detection of Rickettsia spp. in CSF and serum samples
Detection of Rickettsia spp. was performed using a TaqMan real-time PCR assay, as previously described (Stenos et al., 2005).The primers CS-F and CS-R, and probe CS-P are designed to target the Rickettsia spp.citrate synthase gene (gltA), to amplify a 74-bp long amplicon (Table 1).As a PCR-positive control, a synthetic plasmid containing the target sequence of the TaqMan real-time PCR assay was used.The plasmid contained the target sequence, spanning nucleotides 1102-1231 of the R. rickettsii gltA gene (GenBank: 59,729), synthesised and cloned into the pUC57 vector (GenScript).

Detection of Neoehrlichia mikurensis in CSF and serum samples
Detection of N. mikurensis cDNA was performed using a SYBR Green real-time PCR assay, as previously described (Labbe Sandelin et al., 2015).The primers Neo_16S_F and Neo_16S_R are designed to target the N. mikurensis 16S rRNA gene, to amplify a 107-bp long amplicon (Table 1).As a positive control, cDNA samples positive for N. mikurensis, as confirmed by sequencing in an earlier study (Labbe Sandelin et al., 2015), were used in each run.Confirmatory PCR (using the GroEL target gene) of one positive serum sample was conducted at the Department of Clinical Microbiology, Sahlgrenska University Hospital, Gothenburg, Sweden, as previously described (Grankvist et al., 2015).

Detection of TBEV in CSF samples
Detection of TBEV was performed using a duplex TaqMan real-time PCR assay, as previously described (Lindblom et al., 2014).The primers and probes are designed to target all three subtypes of TBEV, to amplify a 68-bp and an 88-bp long amplicon, respectively (Table 1).

Detection of Babesia spp. in CSF and serum samples
Detection of Babesia spp. was performed using a SYBR Green realtime PCR assay, as previously described (Wilhelmsson et al., 2020).Primers BJ1 and BN2 are designed to target the Babesia spp.18S rRNA gene, to amplify a 411-452 bp long amplicon depending on the species of Babesia (Table 1).As a PCR-positive control, a synthetic plasmid containing the target sequence of the SYBR Green real-time PCR assay was used.The plasmid contained the target sequence, spanning nucleotides 467-955 of the B. divergens 18S rRNA gene (GenBank: J439713), synthesized and cloned into the pUC57 vector (GenScript).

Detection of antibodies to Borrelia burgdorferi s.l. in CSF and serum
All 600 patients were previously analysed for Borrelia-specific antibodies in paired CSF and serum using the IDEIA Lyme Neuroborreliosis Kit (Oxoid, Hampshire, UK).The serological tests were performed according to the manufacturer´s instructions at the Departments of Clinical Microbiology in Linköping, Jönköping, and Kalmar, respectively.

Detection of antibodies to Borrelia miyamotoi in serum
Serum samples from patients with CSF mononuclear pleocytosis (N=52) were analysed for the presence of B. miyamotoi IgM and IgG antibodies by previously described two-tier serology, using an enzymelinked immunosorbent assay (ELISA) followed by confirmatory Western blot assays (WB), based on the recombinant protein Glycerophosphodiester phosphodiesterase (GlpQ) (Boyer et al., 2020;Hoornstra et al., 2018;Wagemakers et al., 2016).GlpQ is able to distinguish relapsing fever Borrelia from B. burgdorferi s.l.Additional whole cell lysate (WCL) WB analysis was performed using the B. miyamotoi strains NL-IR-1 and Izh-4 isolated from a Dutch tick (Kuleshov et al., 2019) and a patient with PCR-confirmed B. miyamotoi disease (BMD) in European Russia (Kuleshov et al., 2020), respectively.The amount of antigens used was 500 ng recombinant protein and 1 μg WCL.As IgM and IgG positive controls a pool of sera from 16 PCR-confirmed Russian BMD patients with high α-GlpQ antibody levels were used.The negative controls consisted of 10 serum samples obtained from healthy blood donors from a region in Norway non-endemic for ticks (Hoornstra et al., 2018).The cut-off for an ELISA positive sample was arbitrarily set at four standard deviations above the median read-out of the negative controls (Hoornstra et al., 2018;Krause et al., 2013).The interpretation of the confirmatory WB was assessed as published (Hoornstra et al., 2018;Koetsveld et al., 2018).The analyses were performed at the centre for Experimental and Molecular Medicine, Amsterdam, the Netherlands.

Detection of antibodies to Anaplasma phagocytophilum in serum
Serum samples from patients with CSF mononuclear pleocytosis (N=55) were analysed for the presence of A. phagocytophilum IgG antibodies using a commercially available indirect immunofluorescence assay (IFA) (Focus Diagnostics Inc., Cypress, CA, USA), and processed according to the manufacturer's instructions, except that dilutions of 1:80, 1:160, 1:320, etc. were used.The end-point titre was recorded as the reciprocal of the last serial dilution at which specific apple-green fluorescence of Anaplasma inclusion bodies focally located in the cytoplasm of the infected cells.IFA titre ≥1:160 was interpreted as positive.This cut-off value was based on the results obtained for local Swedish serum samples that were used when validating the method (Henningsson et al., 2015b).The analyses were performed at the Department of Clinical Microbiology, County Hospital, Jönköping, Sweden.

Detection of antibodies to SFG rickettsiae in serum
Serum samples from patients with CSF mononuclear pleocytosis (N=55) were analysed for the presence of IgM and IgG antibodies to R. rickettsii, a member of the SFG rickettsiae, using a commercially available IFA (Focus Diagnostics) and processed according to the man-ufacturer´s instructions.IFA titre ≥1:64 for IgM and ≥1:512 for IgG were interpreted as positive; a higher titre than that set by the manufacturer, in order to achieve higher specificity and reduce the risk of crossreactivity (Kantso et al., 2009).The analyses were performed at the Statens Serum Institut (SSI), Copenhagen, Denmark.

Detection of antibodies to TBEV in serum
Serum samples from the patients with CSF mononuclear pleocytosis (N=55) were analysed for the presence of IgM and IgG antibodies to TBEV using a commercially available enzyme immuno-assay (EIA) (Enzygnost anti-TBE IgG and IgM; Siemens Healthcare Diagnostics Products GmbH, Erlangen, Germany) according to the manufacturer´s instructions.The analyses were performed at the Department of Clinical Microbiology, County Hospital, Jönköping, Sweden.

Detection of antibodies to Babesia microti in serum
Serum samples from patients with CSF mononuclear pleocytosis (N=55), were analysed for the presence of IgM and IgG antibodies to B. microti using a commercially available IFA (Focus Diagnostics), according to the manufacturer´s instructions.IFA titre ≥1:64 was interpreted as positive.The analyses were performed at the SSI, Copenhagen, Denmark.

PCR findings
One female patient (53 years of age) had detectable N. mikurensis in the serum by two PCR methods targeting different genes (16S rRNA and groEL).This patient was splenectomised during childhood due to spherocytosis and consulted health-care due to a couple of hours of headache, vertigo and vomiting.Information regarding tick bites was missing.Computed tomography (CT) and magnetic resonance imaging (MRI) of the brain did not reveal any pathology.The CSF analysis was devoid of pleocytosis.The intrathecal Borrelia-specific antibody index (AI) was not elevated.The patient recovered spontaneously without a conclusive diagnosis.
One male patient (45 years of age) had detectable Borrelia spp. in serum by real-time PCR targeting the 16S rRNA gene (Cq-value >40).Based on the sequencing chromatograms of the intergenic spacer region between 5S and 23S rRNA genes, it was not possible to distinguish the Borrelia species, however the PCR targeting the B. miyamotoi flaB was negative.This patient had a peripheral facial palsy together with headache and neck pain for three days.The CSF was without CSF pleocytosis, the intrathecal Borrelia-specific AI was not elevated and the chemokine CXCL13 was below cut-off (<250 ng/L, ReaScan CXCL13, Reagena Ltd, Toivala, Finland).A CT of the brain revealed no pathology.He did not receive any antibiotic treatment and the facial palsy persisted to some extent for more than six months. None

Patients with CSF mononuclear pleocytosis and elevated intrathecal Borrelia-specific antibody index
Thirteen patients (seven females and six males; median age, 10 years) had both mononuclear pleocytosis and an elevated intrathecal Borrelia-specific AI.Eleven patients had symptoms suggestive of LNB and were treated accordingly (Table 2).One patient received the diagnosis of suspected cytomegalovirus (CMV) myelitis with fever, symptoms of cauda equine syndrome and both IgM and IgG antibodies to CMV and one patient was without any conclusive diagnosis.These two patients were not treated with antibiotics and recovered within a few months.

Patients with CSF mononuclear pleocytosis and serum antibodies to Borrelia miyamotoi
Six of the 52 patients with CSF pleocytosis showed IgM reactivity in their serum samples using the GlpQ ELISA.WB could not confirm these results.However, one patient showed IgM reactivity in the WCL WB.Five patients showed IgG reactivity in the GlpQ ELISA, of which none could be confirmed by WB.However, in the additional WCL WB four out of five patients showed reactivity.One of these four patients showed reactivity to both IgM and IgG.The four patients with positive GlpQ ELISA and WCL WB are included in Table 3.The clinical picture of two patients consisted of vertigo and myalgia respectively and they both showed a low-grade mononuclear pleocytosis in the CSF (7 × 10 6 /L and 12 × 10 6 /L).Concomitantly they tested positive for SFG rickettsiae IgM antibodies (titre 512) and TBEV IgG antibodies, respectively.None of these patients were treated with antibiotics and they were diagnosed with multiple sclerosis (MS) and demyelinating polyneuropathy (DPNP), respectively.The third patient underwent a headache investigation without a diagnosis and recovered within a month.The fourth patient, a child of eight years old, excibited a three-day lasting peripheral facial palsy and had an elevated intrathecal Borrelia-specific AI.None of the above four patients reported either fever, a low white blood cell count, thrombocytopenia, or elevated liver enzymes (measured in three of four patients).

Patients with CSF mononuclear pleocytosis and serum antibodies to Anaplasma phagocytophilum
Four patients had a positive IgG titre (1:160) to A. phagocytophilum (Table 3).Of these, three patients had antibodies to several tick-borne pathogens (Table 4).One patient had low-grade mononuclear pleocytosis (8 × 10 6 /L) and reported five weeks of nausea and vertigo.However, no explanation for the CSF mononuclear pleocytosis was given in the medical charts and the diagnosis remained unclear.Two patients received a diagnosis of LNB and TBE, respectively, and the third patient underwent a headache investigation but did not receive any conclusive diagnosis.No follow-up serological assays regarding A. phagocytophilum were performed.

Patients with CSF mononuclear pleocytosis and serum antibodies to Babesia microti
None of the patients with CSF pleocytosis had detectable serum antibodies to B. microti.

Patients with serum antibodies to more than one pathogen
Nine patients had serum antibodies to more than one pathogen (Table 4).Cases 1, 3, 4 and 5 had symptoms compatible with LNB, had CSF mononuclear pleocytosis and an elevated AI to Borrelia (IgM and/or IgG), thereby fulfilling the diagnostic criteria for LNB.They received treatment for LNB and had concomitant IgG antibodies to B. miyamotoi, A. phagocytophilum, SFG rickettsiae and TBEV, respectively.They all had a short duration of disease prior to the lumbar puncture (LP) (range 0.4-6 weeks) and they recovered within three months following antibiotic (doxycycline) treatment.Case 2 had long-term headache, lowgrade CSF mononuclear pleocytosis (7 × 10 6 /L), an elevated Borreliaspecific AI (IgG) and antibodies to A. phagocytophilum.This patient recovered spontaneously within four months.Two patients (cases 6 and 7) had symptoms of vertigo and myalgia/arthralgia, respectively, and had low-grade CSF mononuclear pleocytosis (7 × 10 6 /L and 12 × 10 6 / L).They both had antibodies to B. miyamotoi and concomitant IgM antibodies to SFG rickettsiae (titre 1:512) and IgG antibodies to TBEV, respectively.These patients later received the diagnosis of MS and DPNP.Case 8 who had IgM and IgG antibodies to TBEV, as well as IgG antibodies to A. phagocytophilum had symptoms compatible with TBE.One patient who was previously vaccinated against TBE (Case 9) presented with head and neck pain and had IgG antibodies to TBEV, as well as IgG antibodies to SFG rickettsiae (titre 1:256).This patient was a forestry worker who later received the diagnosis of SDH.

Discussion
In this large retrospective study, both molecular and serological techniques were used to investigate the presence of B. burgdorferi s.l. and other tick-borne pathogens in 600 patients with clinically suspected LNB in a highly Borrelia-and TBEV-endemic region of South-eastern Sweden.
Thirteen patients fulfilled the diagnostic criteria for LNB, in accordance with the EFNS guidelines (Mygland et al., 2010).However, B. burgdorferi s.l. was not detected in the CSF samples in these patients by PCR.Although PCR is a sensitive detection method on itself, its sensitivity to detect Borrelia spp. in clinical CSF samples is low (Mygland et al., 2010).This can possibly be attributed to the spirochetal affinity to myelin and the likely low number of Borrelia spirochaetes circulating in the CSF (Cerar et al., 2008;Rupprecht et al., 2008).Additionally the relatively longer period of time between disease onset and CSF sampling may have been a contributing factor to the reduced PCR sensitivity in this study (Barstad et al., 2018).The patients with LNB had a median disease duration of two weeks prior to LP, and only two patients reported a symptom duration of less than one week.
Neoehrlichia mikurensis was detected by real-time PCR in the serum of one patient, who was previously splenectomised.This is in accordance with previous reports, describing an association of this pathogen to immune-suppressive conditions (Grankvist et al., 2014;Höper et al., 2020;Quarsten et al., 2020).In this particular case, however, no immuno-suppressive medication, history of autoimmune disease or haematological malignancy was reported.The finding is interesting, as it potentially implicates an under-reported pathogen.So far, around 65 cases of neoehrlichiosis have been diagnosed in Sweden (A.Grankvist, personal communication, March 2021), of whom 40 patients have been described in a recent study (Höper et al., 2020).Neoehrlichia mikurensis has been detected in human blood, plasma and serum samples but so far not in the CSF (Portillo et al., 2018).Our finding highlights the necessity to consider N. mikurensis in tick-exposed immuno-compromised individuals in Sweden, since the symptoms may be variable (Grankvist et al., 2014).
In this study, Borrelia spp.were detected in one patient by real-time PCR in the serum, but none in the CSF samples.This patient had no CSF pleocytosis, but had symptoms compatible with LNB with peripheral facial palsy, headache and neck pain.Information regarding any erythema migrans (EM) were missing.The CSF was devoid of pleocytosis and negative regarding Borrelia AI, which could be in line with an infection caused by B. afzelii.This Borrelia spp.has been described to cause a CNS infection without CSF pleocytosis (Strle et al., 2006).Unfortunately, we were not able to sequence the PCR product in order to determine the Borrelia species in this particular patient, possible due to a low amount of Borrelia bacteria.
We were not able to detect A. phagocytophilum by PCR in either CSF or serum samples.In previous studies, Anaplasma-specific PCR has generally been performed on whole blood samples (Krause et al., 2002;Lee et al., 2018), and studies performed on serum have shown considerably lower sensitivity (Heo et al., 2002).The duration of symptoms may also be important to take into consideration, since Schotthoefer and co-workers reported a significantly higher sensitivity with PCR from patients with symptom duration of less than a few days (Schotthoefer et al., 2013).Nevertheless, the present study is, to our knowledge, the first to investigate the presence of A. phagocytophilum in CSF using molecular detection methods.
Rickettsia spp.were not detected by PCR in either CSF or serum samples.In a recent Danish study, in which CSF samples from patients with suspected LNB were investigated with molecular methods for Rickettsia spp.and serological methods for SFG rickettsiae, eight patients were found to be positive for Rickettsia spp.by PCR in the CSF, and most of them had symptoms and laboratory findings characteristic of LNB (Ocias et al., 2018).Rickettsia spp.DNA was also found in samples from patients without CSF mononuclear pleocytosis but in whom other non-infectious diagnoses were suspected.None of the PCR-positive patients in the Danish study had serum antibodies to SFG rickettsiae.All patients recovered spontaneously without any treatment.Similar findings have also been found in a Dutch study (Koetsveld et al., 2016).
In none of the 600 patients included in our study TBEV could be detected in the CSF by PCR.Only two patients were regarded as having a current TBEV infection according to their symptoms, CSF analysis and serology.We did not perform any PCR analysis of the sera, since the samples were drawn from patients with suspected neurological symptoms and TBEV is usually not detectable in the serum during the meningoencephalitic phase of the disease (Saksida et al., 2005).Interestingly, a few studies including a small number of patients iwith meningoencephalitis have detected TBEV with PCR in tissues other than the CSF, such as sera and urine (Nagy et al., 2018;Saksida et al., 2005;Veje et al., 2018).
It is not known whether Babesia spp.can infect the human CNS, and this study was not able to contribute to clarifying this issue, since Babesia spp.DNA was not detected in the CNS samples of any of the patients.Babesia spp.primarily infects red blood cells, causing mainly a systemic illness, which may be severe or lethal in immuno-suppressed and splenectomised patients (Krause, 2019).
In contrast to direct pathogen detection using molecular methods, the interpretation of serological results may be more challenging due to difficulties in differentiating between an on-going and a previous infection, especially in a tick-endemic region and without the availability of convalescent serum samples for evaluation of serodynamics.
In this study, four of the 52 patients with CSF pleocytosis had IgG, and one of these patients also had IgM antibodies to B. miyamotoi in serum, shown through both GlpQ ELISA and WCL WB reactivity.Although the conventional GlpQ WB was not reactive in any of these patients, the WCL WB was suggestive of previous exposure to B. miyamotoi.It is unclear why the GlpQ ELISA results could not be confirmed by WB, but it should be mentioned that all B. miyamotoi antibody assays are still experimental, including the GlpQ serology.It has been proposed that a GlpQ ELISA followed by GlpQ WB can be used for serodiagnosis of B. miyamotoi infection (Krause et al., 2013).A recent paper expressing concerns about the specificity of GlpQ (Reiter et al., 2020) is in stark contrast to many previously published papers (Boyer et al., 2020;Jahfari et al., 2017;Koetsveld et al., 2018;Schwan et al., 1996).Sample selection and technical issues could perhaps explain part of the observed and alledged discrepancy.For the current study we have followed another approach.Since the GlpQ ELISA results could not be confirmed by GlpQ WB, we sought for another second-tier test, i.e. a WCL WB, which actually is more rational since antibodies to other targets than GlpQ are analysed.In our manuscript, we only consider patients who are positive in the GlpQ ELISA as well as reactive in the WCL WB to have specific antibodies to B. miyamotoi.However, it should be noted that the use of WCL WB in addition to the GlpQ ELISA needs further evaluation and validation and thus the seroreactivity should be interpreted with caution.Only one of these patients received a diagnosis of tick-borne disease (LNB).Two patients either recovered spontaneously or had a quick recovery after treatment with antibiotics (doxycycline).The other two patients were found to have chronic neurological diseases, MS and DPNP, respectively.The detected IgG antibodies might reflect a previous exposure to B. miyamotoi, although an on-going infection at the time of sampling cannot be excluded, especially in the patient showing both IgG and IgM reactivity.This patient, however, did not have any typical symptoms or laboratory results supporting BMD (Hoornstra et al., 2018).It has previously been shown that B. miyamotoi can cause disease, including CNS infection, in both immunocompromised and immunocompetent individuals (Henningsson et al., 2019;Hovius et al., 2013;Platonov et al., 2011), and BMD has, as yet only rarely been described to resolve spontaneously without antibiotic treatment (Hoornstra et al., 2018;Sudhindra et al., 2016).In the Netherlands, it has been shown that 2% of healthy blood donors and 10% of forestry workers have antibodies to B. miyamotoi (i.e.GlpQ), suggesting that subclinical cases occur.Supporting the data in the Dutch study, it has been reported that 2% of ticks that have bitten humans in Sweden contain B. miyamotoi (Wilhelmsson et al., 2013), but so far, no studies of seroprevalence in the Swedish population have been performed.
Four of the patients with CSF pleocytosis had a positive, albeit a low IgG titre to A. phagocytophilum, probably reflecting a past exposure.Two patients with both CSF pleocytosis and an elevated Borrelia spp.AI were diagnosed with LNB and had serum antibodies to A. phagocytophilum.Neither of these patients had a severe disease course.Studies on mice that were simultaneously infected with B. burgdorferi s.l. and A. phagocytophilum have though showed the opposite (Thomas et al., 2001).One patient was serologically and from a clinical point of view regarded as having TBE and had no laboratory findings suggestive of HGA.
Five patients with CSF pleocytosis had either IgM or IgG antibodies to SFG rickettsiae in serum.While the measured antibody levels probably reflect a previous infection, one patient was considered to have a possible current SFG rickettsiae infection with thrombocytopenia and a high titre of IgG antibodies to SFG rickettsiae (1:1024).Knowledge regarding the kinetics of SFG rickettsiae antibodies is, however, scarce and follow-up serological analysis may be needed to determine whether it is a current or past infection.Interestingly, one patient with IgM antibodies to SFG rickettsiae (titre 1:512) was diagnosed with Miller-Fisher syndrome, a variant of Guillain-Barre syndrome that includes typical neurological symptoms, such as acute ophthalmoplegia, ataxia, and generalized areflexia, and has in fact been described in patients with acute HGA (Bakken et al., 1998).In this case, there was no medical information regarding the typical skin manifestations indicative of SFG rickettsiae infection.
Two patients had typical symptoms and IgM and IgG antibodies consistent with a current TBEV infection.Another four patients had IgG antibodies to TBEV but lacked symptoms compatible with TBE.At least three of them were previously vaccinated against TBE.
We found no serological evidence of current or past infections caused by B. microti amongst the study subjects with CSF mononuclear pleocytosis.In a recent study from Southern Sweden, serological responses to B. microti or B. divergens were seen in 10% of Borrelia-seropositive patients, although this study used a different serological assay and may therefore not be fully comparable to our study (Svensson et al., 2019).In our study we could only analyse IgG antibodies to B. microti, an analysis reported to exhibit low cross-reactivity to B. divergens and B. venatorum (Lempereur et al., 2015).
In the present study, nine patients had antibodies in serum or CSF to more than one of the analysed tick-borne pathogens.While four patients were diagnosed with LNB and one with TBE, none of them had symptoms or laboratory evidence of an on-going co-infection.This is consistent with other studies in Europe, in which patients with LNB or TBE have varying seroprevalence of other tick-borne pathogens but seldom exhibit symptoms compatible with a current co-infection (Koetsveld et al., 2016;Moniuszko et al., 2014;Ocias et al., 2018).Moniuszko et al (2014) investigated with both serological and molecular methods the presence of co-infections with Borrelia spp, A. phagocytophilum and Babesia spp. in patients with TBE and found one patient who was PCR-positive for B. microti in blood.PCR analysis was not performed in the CSF.None of the patients had symptoms compatible with a co-infection.Koetsveld et al. (2016) conducted PCR analysis of R. helvetica and R. monacensis DNA in CSF samples from patients investigated for suspected LNB and identified four patients positive for R. helvetica and one for R. monacensis, respectively.None had CSF mononuclear pleocytosis or symptoms compatible with an on-going co-infection, and only one of the five patients had antibodies to SFG rickettsiae.So far, co-infection with B. miyamotoi has only been described in one case with Lyme borreliosis with concomitant symptoms of high fever, myalgia, arthralgia, the presence of EM and elevated liver enzymes (Oda et al., 2017).Of the three patients with a possible co-infection with B. miyamotoi, one had typical symptoms of LNB and the other two were part of a neurological investigation and did not receive any diagnosis of tick-borne infection(s).None of the patients had any signs of an ongoing BMD and the B. miyamotoi serology might also reflect a past exposure.
In this study the strengths are the high number of patients included in three different Borrelia-and TBE-endemic regions and the use of corresponding paired CSF and serum samples.The retrospective design and lack of convalescent samples are limitations of the study, as is the use of serum rather than whole blood for the molecular detection of the obligatory intracellular pathogens such as A. phagocytophilum, Rickettsia spp.and Babesia spp.The assessment of whether the infections were previous or current may have been hampered by the retrospective study design, the limited clinical information in some cases, as well as the incomplete laboratory test results.

Conclusion
Despite the presence of various pathogens in ticks in South-eastern Sweden, the risk of contracting a tick-borne CNS infection, including co-infections with multiple tick-borne pathogens, appears to be low.Amongst the patients being investigated for LNB, there were few indications of current tick-borne co-infections.However, the results need to be confirmed in a prospective study conducted in a tick-endemic area with a higher number of patients with CSF pleocytosis.
of serum using the MagNA Pure LC 2.0 Instrument (Life Science, Roche), and the MagNA Pure LC Total Nucleic Acid Isolation Kit (Life Science, Roche) using a 100-µL elution volume, according to the manufacturer's instructions.Each extraction batch of 32 samples consisted of 30 samples, one positive control, and one negative control.As a positive control, 20 µL of PBS solution (pH 7.2) that contained a known number of B. afzelii (strain Lu81) cells (ranging from 1 to 10 5 cells per 20 μl) were spiked into 180 µL CSF (obtained from routine diagnostic sec- P.Gyllemark et al.µL

Table 1
Oligonucleotide primers and probes used for the molecular analysis of tick-borne pathogens.

Table 4
Patients with CSF mononuclear pleocytosis and possible co-infections.