Nanopore sequencing identifies Borrelia miyamotoi as an unexpected cause of meningitis after B cell depletion

Key points.

• Metagenomic nanopore sequencing identified Borrelia miyamotoi as a rare cause of meningitis in an immunocompromised patient, undetected by standard PCR and serological tests.

• Nanopore sequencing provided a swift, unbiased diagnosis, proving critical in cases where routine methods are inconclusive.

• The case underscores the diagnostic challenge posed by Borrelia miyamotoi, with MRI features resembling tuberculous meningitis, emphasising the role of advanced sequencing in differential diagnoses.

Bacterial infections of the central nervous system (CNS) are often severe and can be potentially fatal, particularly in immunocompromised patients. Early detection is crucial for rapid and tailored antibiotic treatment, which often determines prognosis. Routine diagnostics of cerebrospinal fluid (CSF) identifies pleocytosis, which defines meningitis, and pathogen detection usually involves microbial culture or serology, yet the cause remains unidentified in approximately 50% of acute meningitis/meningoencephalitis cases [1, 2]. Molecular diagnostics using pathogen‐specific polymerase chain reaction (PCR) (e.g., meningitis/encephalitis multiplex PCRs) provides an alternative approach but is hypothesis‐driven, requiring prior suspicion of the pathogen. Metagenomic next‐generation sequencing represents an unbiased alternative and has been successfully established in CSF [3, 4, 5]. Bacterial infections with Borrelia miyamotoi, an emerging disease in the northern hemisphere (including the United States, Canada and Asia), have only been documented in six cases in Europe. CNS involvement is generally rare and has been reported in 1% of B. miyamotoi infections [6, 7]. Serological tests are currently under development, and most diagnoses are made using specific PCR tests [7] while this pathogen is assumed to rarely cause infection in Europe and is thus not routinely tested.

Here, we report the case of a 41‐year‐old European male with primary‐progressive multiple sclerosis (PPMS) who was under stable treatment with ocrelizumab (9 cycles, 3 years) at our neurology centre in Germany. Upon presentation to our department, the patient described a 3‐month history of a gradually worsening bifrontal throbbing headache, horizontal diplopia and loss of appetite. A detailed history revealed that he had visited Japan six times in the past 10 years, and his last visit was approximately 5 months before the onset of his symptoms. Clinical examination revealed mild spastic hemiparesis of the right side, clonus in the right foot and an extensor plantar response (positive Babinski sign). There was no meningism, and Lhermitte's sign was negative. Peripheral CD19+ B lymphocytes were fully depleted; blood tests showed no signs of systemic inflammation and were otherwise unremarkable (Data S1).

Cerebral magnetic resonance imaging (MRI) at admission revealed disseminated lesions in the subarachnoid space, as well as contrast enhancement of the supra‐ and infratentorial meninges and cranial nerves, including the bilateral oculomotor, trigeminal, facial and vestibulocochlear nerves (Figure 1A). Compared to previous cerebral MRIs from 6 weeks and 5 months before admission (Figure S1), no new intra‐parenchymal T2‐hyperintense or gadolinium (Gd)‐enhancing lesions were detected, and importantly, these earlier MRIs also did not show evidence of the current contrast‐enhancing lesions, despite the patient having had symptoms 3 months ago. Spinal MRI showed contrast enhancement of the caudal spinal nerves, indicating polyradiculitis (Figure 1A). Given that the MS lesions remained stable and the patient's symptoms had not improved following high‐dose corticosteroid pulse therapy administered 6 weeks earlier, a relapse of MS was considered unlikely. However, the detection of novel contrast enhancement of the meninges and cranial and caudal spinal nerves indicated an infectious aetiology. The pattern of contrast enhancement was reminiscent of tuberculous meningitis. First, lumbar puncture was performed 1 day after hospital admission and revealed elevated CSF opening pressure (40 cmH₂O; reference: <20 cmH₂O); increased cell count with mixed lymphocyte/granulocyte pattern (183 lymphocytes/μL, 117 granulocytes/μL; reference: <5 cells/μL); blood–brain barrier dysfunction with strongly elevated CSF protein (3700 mg/L; reference: <500 mg/L); increased CSF lactate (5.04 mmol/L; reference: 1.5–2.1 mmol/L); and decreased CSF glucose (45.7 mg/dL; reference: 49–75 mg/dL). Flow cytometry of CSF revealed an increased proportion of activated CD4+ T‐lymphocytes (HLA DR‐CD4+ 38.2%; reference: <23.17%), a highly increased CD4/CD8 ratio (33.8; reference: 2.21–7.19), an elevated proportion of natural killer cells (16.7%; reference: <2.91%) and the presence of plasma cells (Figure 1B). Similar changes have been described previously in neurosarcoidosis [8]. However, laboratory tests for sarcoidosis and vasculitis were negative (Data S1), but highly elevated CXCL13 levels (466 pg/mL; reference: <250 pg/mL) indicated an infectious aetiology, though it can also be elevated in multiple sclerosis [9].

FIGURE 1.

Summary of clinical and molecular features. (A) Representative cerebral and spinal MRI images. Cerebral MRI with Gd application revealed extended FLAIR hyperintensities, indicated disseminated granulomatous lesions in the subarachnoid space (white arrowhead), leptomeningeal Gd enhancement (blue arrow) and further Gd enhancement in cranial nerves, including oculomotor nerves, facials nerves and vestibulocochlear nerves on both hemispheres (red arrowhead). The MS lesions were constant, and there was no MS lesion‐associated Gd enhancement. Additionally, spinal MRI with Gd application showed diffuse Gd enhancement of the ventral roots of caudal spinal nerves (white arrow), indicating polyradiculitis. (B) CSF flow cytometry revealed elevation of natural killer cells (16.3%, ref. <2.91%), positive plasma cells and elevation of activated CD4+ T‐lymphocytes (38.2%, ref. <23.17%) with strongly increased CD4/CD8 ratio (33.8, ref. 2.21–7.19), indicating inflammation and infection of the CNS. (C) Depth and breadth of Borrelia miyamotoi genome coverage. B. miyamotoi assigned reads and contigs from nanopore sequencing (top) and Illumina sequencing (bottom) were mapped against reference B. miyamotoi NL‐IR‐2 (GenBank #CP044625.1). (D) Cumulative sum of bases assigned to B. miyamotoi over time. CNS, central nervous system; CSF, cerebrospinal fluid; FLAIR, fluid‐attenuated inversion recovery; Gd, gadolinium; MRI, magnetic resonance imaging; MS, multiple sclerosis; ref., reference values.

Upon suspicion of infectious meningitis, empirical intravenous therapy with ceftriaxone, ampicillin and acyclovir was initiated. However, PCR‐based pathogen screening remained negative for viral (HSV1, HSV2, varicella zoster virus [VZV], EBV, CMV, tick‐borne encephalitis virus, HIV, enterovirus), bacterial (Escherichia coli K1, Haemophilus influenzae, Listeria monocytogenes, Neisseria meningitidis, Streptococcus agalactiae, Streptococcus pneumoniae, Mycobacterium tuberculosis complex, Borrelia burgdorferi sensu lato group and B. miyamotoi [Asian genotype], Brucella spp., Mycoplasma pneumoniae, Tropheryma whipplei and pan‐bacterial 16S rRNA gene) fungal (Cryptococcus neoformans, Cryptococcus gattii and pan‐fungal ITS1/2 genes) or parasitic (Toxoplasma gondii) infection. Both Gram and modified Kinyoun staining were performed on a CSF sample and were negative. Serological and CSF antibody screening (IgM, IgG) for infection with Treponema pallidum, B. burgdorferi, Brucella, T. gondii, Histoplasma capsulatum and Francisella tularensis as well as tick‐borne encephalitis by ELISA was negative. Culture analysis of CSF to detect bacteria, mycobacteria and fungi was performed according to the laboratory standard and remained negative. A specific culture for Borrelia species was not carried out.

We considered a meningeal biopsy, but since infectious meningitis remained the most likely differential diagnosis, metagenomic sequencing from the CSF obtained during the initial lumbar puncture (1 day after admission) was performed. Library preparation of gDNA isolated from CSF was performed with the Rapid Barcoding Kit V14 (SQK‐RBK114.24, Oxford Nanopore Technologies). Due to the low DNA concentration (1 ng/μL), tagmentation was carried out in four reactions using the same barcode. The final library was sequenced on a MinION Mk1C device with an R10.4.1 flow cell (FLO‐MIN114).

Sequencing produced 2,937,807 raw reads, and high‐accuracy basecalling with dorado (v0.5.3) yielded 5.8 Gb. Raw .fastq files were analysed using the ID‐Seq pipeline [10]. After quality and host filtering, 539,150 (18.35%) non‐human reads remained. Of these, 3153 loose reads and 7 de‐novo‐assembled contiguous sequences (contigs), totalling 2,290,110 bases with an average length of 6781 bases, were assigned to B. miyamotoi. Assembly‐based alignment against the latest NCBI index (6 February 2024) showed that most aligned reads/contigs corresponded to the European NL‐IR‐2 strain of B. miyamotoi (GenBank #CP044625.1, Figure 1C). Contigs were further analysed using the BLASTn suite from NCBI against the core_nt database. Phylogenetic analysis was conducted based on the contig with the highest number of matches (n = 47) that demonstrated at least 95% sequence identity and a minimum of 90% query coverage. The phylogenetic tree was constructed using the neighbour‐joining method and shows clustering with records of European origin (Figure S2). These results suggest that the infection likely originated in Europe. Nanopore sequencing enables real‐time analysis of basecalled data, with a substantial number of bases matching B. miyamotoi being sequenced within the first few hours (Figure 1D). Remarkably, 113 B. miyamotoi reads (77 kb) were obtained within the first hour, with 16 reads (4.3 kb) captured in the first 10 min of sequencing. These results were further confirmed by direct PCR (of target genes 16S rRNA, flagellin B and p41, species identification using multi‐locus sequence typing [MLST] including clpA, clpX, nifS, pepX, pyrG, uvrA, rplB and recG) (Data S1) and paired‐end next‐generation Illumina sequencing, which generated 41,338,324 reads. Among these, 16,409 reads and 257 contigs were identified as belonging to B. miyamotoi, showing the greatest similarity to the European NL‐IR‐2 strain (Figure 1C). Upon completion of intravenous therapy with ceftriaxone for 21 days, the patient's clinical symptoms (headache, neck pain and visual disturbance) completely subsided. Follow‐up CSF analysis (7 and 19 days after admission) showed a decrease in cell counts (86 and 47/μL) and protein (2700 and 1420 mg/L), normalisation of lactate (2.46 and 2.09 mmol/l) and no plasma cells present. The patient was discharged on Day 23. Right‐sided hemiparesis remained but was considered to be a consequence of the pre‐existing PPMS.

In this case, nanopore sequencing of CSF facilitated the rapid detection of B. miyamotoi in a B cell‐depleted MS patient. B. miyamotoi, discovered in Japan in 1995 [11], is a genetically and ecologically distinct species belonging to the Borrelia relapsing fever group that is transmitted by the same Ixodes tick species as B. burgdorferi, though it is less prevalent. Human infections were first reported in 2011 in Russia [12], often presenting with relapsing fever and mainly affecting immunocompetent patients [11, 12, 13]. However, neurological manifestations have been described in immunocompromised patients, particularly those receiving B cell‐depleting treatments such as rituximab [13, 14, 15, 16, 17]. Clinical symptoms range from acute meningeal signs such as headache, vomiting, dizziness and neck stiffness [16, 17] to prolonged neurologic decline over months, characterised by cognitive impairment, confusion, gait instability and hearing difficulties [13, 14]. Some patients also experience intermittent febrile episodes, fatigue and progressive sensory impairments, complicating the clinical picture [15, 17]. PCR tests for B. miyamotoi are not routinely performed, and thus, the prevalence might be underestimated.

In the course of the diagnostic work‐up of our patient, we assumed possible interference of diagnostic results due to the B cell‐depleted state and thus considered a wide spectrum of potential causes of meningitis. Our case with contrast enhancement of the supratentorial and infratentorial meninges and cranial nerves (including the oculomotor, trigeminal, facial and vestibulocochlear nerves) aligns with previous reports of cranial nerve involvement [15, 17]. In contrast, other cases have shown either no acute MRI findings or normal cranial MRIs [13, 14]. Notably, our study has a temporal aspect, as cerebral MRI performed 5 weeks before admission was unremarkable despite the patient having experienced symptoms for the past 3 months. The imaging characteristics in our patient, with contrast enhancement resembling tuberculous meningitis, underscore the variability of neuroimaging presentations in B. miyamotoi infection. Immune cell composition obtained by CSF flow cytometry indicated an infectious aetiology of meningitis, and elevation of CXCL13 (466 pg/mL) in CSF strongly hinted towards neuroborreliosis; therefore, we extended diagnostic work‐up primarily targeting other/atypical pathogens.

The use of metagenomics sequencing using Illumina next‐generation [17] or nanopore third‐generation sequencing [18] in routine diagnosis offers the possibility to detect any pathogen in the CSF. This approach is especially valuable when bacterial culture, serology and PCR results are negative, despite a clinical suspicion of meningitis/meningoencephalitis. In our case, the patient's B cell‐depleted state presented a twofold diagnostic challenge: Serological tests were negative due to the lack of functional B cells, and immunosuppression increased susceptibility to unusual infections. Negative results in generic pan‐16S PCR testing for B. miyamotoi may be due to low bacterial load or primer mismatch, which can prevent adequate amplification of the organism's specific genetic sequences. Additionally, the GeneProof Borrelia PCR Kit was negative, likely because it specifically detects the Asian genotype of B. miyamotoi, which differs from the genotype present in our patient. Despite the patient's travel history to Japan, sequencing data showed the closest similarity to the European strain NL‐IR‐2, and PCR tests identified sequence type 35, which has thus far only been identified in Europe. While most metagenomic studies have utilised Illumina sequencing platforms, which have a sample‐to‐answer turnaround time of 48–72 h and require large case numbers for batch processing, nanopore sequencing offers a faster alternative. It can identify pathogens within minutes of initiating the sequencing process and takes less than 2 h for library preparation following DNA extraction. Additionally, nanopore sequencing is highly adaptable for diagnostic use, with the ability to reuse flow cells, adjust sequencing duration and employ barcoding for parallel sequencing of multiple samples. Given the low cost of the nanopore sequencing device and minimal additional requirements, this workflow is easily implementable in diagnostic laboratories.

In summary, this case highlights the efficacy of metagenomic nanopore sequencing in diagnosing rare CNS infections such as B. miyamotoi and the benefit of novel diagnostic approaches for patients in immunocompromised states.

AUTHOR CONTRIBUTIONS

Christine Anna Dambietz, Sonja Suntrup‐Krueger and Gerd Meyer zu Hörste were in charge of clinical care, diagnostic work‐up and treatment of the patient. Christine Anna Dambietz performed clinical examinations including lumbar punctures, data acquisition and analysis and drafted the manuscript. Tim Kintzinger and Franziska Schuler performed microbial diagnostics and were involved in the interpretation of results. Volker Fingerle performed extended microbial diagnostics, including real‐time PCR of Borrelia miyamotoi‐specific genes. Anne Albers and Christian Thomas performed neuropathological evaluation as well as nanopore sequencing and next‐generation Illumina sequencing. Christian Thomas and Gerd Meyer zu Hörste jointly supervised this case study and wrote this manuscript.

CONFLICT OF INTEREST STATEMENT

All authors declare that they have no conflicts of interest.

ETHICS APPROVAL

This study was performed in line with the Declaration of Helsinki and with the consent of the patient.

Supporting information

Data S1. CSF flow cytometry and extensive laboratory results.

Figure S1. Previous MRI findings.

Figure S2. Phylogenetic analysis of Nanopore sequencing data.

Dambietz CA, Kintzinger T, Schuler F, et al. Nanopore sequencing identifies Borrelia miyamotoi as an unexpected cause of meningitis after B cell depletion. Neuropathol Appl Neurobiol. 2024;50(6):e13017. doi: 10.1111/nan.13017

DATA AVAILABILITY STATEMENT

The sequencing data that support the findings of this study are openly available in NCBI at https://www.ncbi.nlm.nih.gov/sra/PRJNA1169805.