The role of plasma metagenomic sequencing in identification of Balamuthia mandrillaris encephalitis

Abstract

Balamuthia mandrillaris is a rare, free-living amoeba (FLA) that causes granulomatous amoebic encephalitis, a disease with close to 90% mortality. The geographical ranges of many FLA are expanding, potentially increasing human exposure to B. mandrillaris. Here, we report a case of a 58-year-old woman with progressive neurological symptoms, ultimately diagnosed postmortem with B. mandrillaris encephalitis through plasma metagenomic next-generation sequencing (mNGS) despite negative results on both cerebrospinal fluid (CSF) mNGS and CSF PCR testing. Histologic analysis and real-time PCR (qPCR) studies on postmortem brain tissue confirmed B. mandrillaris infection with significant vascular clustering of trophozoites. Retrospective analysis of CSF mNGS data demonstrated subthreshold reads for B. mandrillaris, emphasizing the challenges of interpreting low-level pathogen signals. A systematic review of 159 published B. mandrillaris cases revealed only two reports of B. mandrillaris diagnosed using plasma mNGS, both of which also had diagnostic CSF studies. This case demonstrates the diagnostic challenges of B. mandrillaris infections, highlights its vascular tropism, and suggests that plasma mNGS may warrant evaluation as a diagnostic tool for B. mandrillaris.

Introduction

Balamuthia mandrillaris is a free-living amoeba (FLA) that causes granulomatous amoebic encephalitis (GAE), a rare and highly fatal condition characterized pathologically by variable degrees of inflammation, granuloma formation, and tissue necrosis. Found in soil and water, B. mandrillaris infections typically arise from direct inoculation through skin wounds or inhalation [11]. GAE often presents with nonspecific neurological symptoms, including headache, fever, and focal deficits, which complicates diagnosis and can delay treatment [11]. Accurate diagnosis requires differentiation from Acanthamoeba spp., another cause of GAE.

The increasing prevalence of amoebic infections over recent decades has been associated with climate change [27], suggesting that the need for diagnostic tools for B. mandrillaris may increase. Traditional diagnostic methods such as cerebrospinal fluid (CSF) analysis and neuroimaging frequently fail to identify B. mandrillaris [92]. Recent advances in metagenomic next-generation sequencing (mNGS) have augmented diagnosis of many central nervous system infections [25, 29, 36, 104]. mNGS sequences all nucleic acids present in a sample at very high depth, allowing unbiased pathogen detection from diverse tissue samples [25, 29, 36, 104]. While mNGS has been applied successfully to CSF and brain tissue in many GAE cases, the diagnostic utility of plasma mNGS in GAE has not been comprehensively evaluated.

Here, we report a fatal case of B. mandrillaris GAE diagnosed via plasma mNGS, despite negative CSF mNGS results. Retrospective analysis of CSF mNGS revealed subthreshold reads for B. mandrillaris, emphasizing the challenges of interpreting low-level pathogen signals. This case and our review of the literature highlight the organism’s pronounced vascular tropism, which we hypothesize could underlie a potential utility for plasma mNGS as a diagnostic tool for GAE.

Case presentation

A 58-year-old woman, with no significant past medical history, presented with 10 days of progressive neurological symptoms, including right facial numbness, left-sided hemiparesis, imbalance, and tinnitus. She presented with no visible skin lesions; however, she reported having sustained superficial scratches on her arms from gardening several weeks prior, which had resolved. Upon examination, no skin lesions were observed. Brain MRI revealed a 2.1 cm rim-enhancing lesion in the right pons with surrounding edema and mass effect (Fig. 1a). CSF analysis showed lymphocytic pleocytosis (132–203 cells/µL), increased opening pressure (27 cm H₂O), elevated protein (64 mg/dL), and normal glucose levels. Comprehensive testing for bacterial, fungal, mycobacterial, viral, and autoimmune etiologies yielded negative results (see Table 1). Flow cytometry indicated reactive T-cells without aberrancy.

Fig. 1

Initial MRI and surgical biopsy. a Initial axial post-contrast T1-weighted MRI, performed the day after admission, showed a right pontine rim-enhancing lesion (arrow). Right middle cerebellar peduncle biopsy demonstrated b macrophage-rich chronic inflammation (H&E, 20x) with c CD68 stain highlighting numerous macrophages (CD68, 20x). d Rare possible granulomas were identified (H&E, 40x). Scale bars: b, c = 50 µm, d = 20 µm

Table 1 Laboratory test results

A biopsy of the right middle cerebellar peduncle showed chronic inflammation with macrophage infiltration (CD68-positive) and necrosis, without identifiable organisms and with rare possible granulomas (Fig. 1b–d). Stains for acid-fast bacilli (AFB), fungal organisms (GMS), Gram-positive/negative bacteria, and other pathogens were negative. Cultures of blood, CSF, and biopsy tissue—including anaerobic, fungal, and mycobacterial screens—were negative. Despite broad-spectrum antimicrobials and methylprednisolone for suspected demyelination, the patient’s condition deteriorated.

Follow up MRI, performed 28 days after the initial scan, demonstrated expansion of the initial lesion with new brainstem, cerebellar, cerebral, and leptomeningeal lesions (Fig. 2a–c), and multifocal vessel wall enhancement of the bilateral middle cerebral, internal carotid, and vertebral arteries (Fig. 4a–d). Cyclophosphamide was initiated for suspected primary CNS angiitis. Biopsies of the dura and cerebellum showed only chronic inflammation. mNGS of CSF was not reported to be positive for any organisms. The patient’s disease progressed to status epilepticus, obstructive hydrocephalus, brainstem herniation, and death. Plasma mNGS, ordered prior to the patient’s death, identified B. mandrillaris postmortem.

Fig. 2

Follow-up MRI findings and gross brain at postmortem examination. Axial post-contrast T1-weighted MRI performed on day 29 of hospitalization showed a expansion of the initial pontine lesion (arrow) with b new brainstem (arrow) and cerebellar lesions (arrows) as well as c cerebral (white arrow) and leptomeningeal (red arrows) lesions. d Brain at autopsy with cerebral edema, cerebellar herniation, hemorrhage, and brainstem necrosis (arrow)

Autopsy revealed severe cerebral edema, cerebellar herniation, hemorrhage, and necrosis of the occipital lobes, brainstem, and cerebellum (Fig. 2d). Histopathological evaluation demonstrated trophozoites and cysts in necrotic brain tissue and perivascular spaces (Fig. 3a–c). There was marked perivascular inflammation with fibrinoid necrosis of vasculature (Fig. 4e). The trophozoites, many of which were necrotic, were round to ovoid with foamy cytoplasm; most had a single, round nucleus (Fig. 3b). The cyst forms appeared to show varying stages of development, with mature cysts characterized by thick, multilayered capsules enclosing vacuolated cytoplasm, and a nucleus (Fig. 3c). Prominent perivascular clustering of trophozoites and vascular invasion were observed, primarily by cyst forms (Figs. 4f and 5). Giemsa (Figs. 3a, 5d) and PAS (Fig. 5a and b) stains highlighted amoebae. Targeted qPCR on formalin-fixed paraffin-embedded (FFPE) brain tissue confirmed B. mandrillaris DNA in the brainstem and cerebellum (Fig. 3d). Retrospective analysis of CSF mNGS revealed subthreshold reads for B. mandrillaris. PCR for B. mandrillaris on this CSF sample was negative.

Fig. 3

B. mandarallis trophozoites and cysts, and free-living amoeba PCR. a Trophozoites adjacent to cerebellar granule cells (Giemsa, 40x). High-power views of b trophozoites (H&E, 100x) and c cysts (H&E, 100x). d PCR amplification curves of B. mandrillaris specific nucleic acid in: (A) B. mandrillaris positive control (Cp = 28.3) and (B) patient’s cerebellar FFPE tissue (Cp = 31). (C) negative controls (Naegleria fowleri and Vero cell line controls). Scale bars: a = 200 µm; b, c = 10 µm

Fig. 4

Vessel wall enhancement on antemortem MRI and H&E stain of postmortem brain. Axial post-contrast T1-weighted MRI performed on day 29 of hospitalization showed vessel abnormalities, including a bilateral middle cerebral artery (MCA) and b bilateral internal carotid artery (ICA) wall enhancement (arrows). c and d provide closer views of vessel wall enhancement (arrows; brightness/contrast of image adjusted for visualization). e Fibrinoid necrosis of the vasculature (arrow, H&E, 20x). f Perivascular trophozoites in the cerebellum (arrow, H&E, 40x). Scale bars: e = 50 µm, f = 20 µm

Fig. 5

Vascular tropism of B. mandrillaris observed at postmortem brain examination. a–b Perivascular trophozoites in the cerebellum (arrows) (PAS; a: 20x, b: 40x). c–e Invasion of the arterial wall, predominantly by cysts (c: H&E, 10x; d: Giemsa, 40x; e: H&E, 100x). Scale bars: a = 50 µm, b, d = 20 µm, c = 100 µm, e 10 µm

Discussion and conclusions

GAE is a rare but often fatal infection caused by free-living amoebae (FLA) such as Acanthamoeba spp. and B. mandrillaris (previously known as “leptomyxid ameba”) [8]. Our autopsy findings included striking perivascular clustering of amoebae and vascular invasion (Figs. 4, 5). The organism’s ability to cross the blood–brain barrier involves interactions with human brain microvascular endothelial cells (HBMECs) via carbohydrate moieties, facilitating CNS invasion [53]. This vascular affinity likely contributed to the multifocal vessel wall enhancement observed on imaging (Fig. 4a–d). We hypothesized that these characteristics facilitated the detectability of circulating trophozoites or DNA fragments in plasma, leading to the positive plasma mNGS result in this case.

To further investigate diagnostic testing regimens and vascular association of B. mandrillaris, we performed a PubMed review with key words “Balamuthia mandrillaris” or “leptomyxid ameba” and identified 159 unique published cases of B. mandrillaris infection that had clear CNS or skin involvement and also reported patient demographics, outcome, and diagnostic methods (Table 2). This revealed 120 cases of B. mandrillaris diagnosed in brain tissue, 18 in skin, and 30 in CSF, with some cases diagnosed at more than one site. There was a high mortality rate of approximately 89%. The average patient age was 36 years, with cases spanning all age groups: children (< 18 years, 59 cases), adults (18–64 years, 72 cases), and older adults (65 + years, 28 cases). Gender distribution included 100 males and 58 females. One case lacked gender information. Within the United States, cases were geographically widespread, with the highest concentrations reported in Texas and California. Notably, the vascular tropism of B. mandrillaris observed in our case aligned with many prior studies. We identified both in vitro evidence for a predilection for perivascular invasion and vascular inflammation [53] as well as at least 10 cases with explicit mention of perivascular or vascular involvement on neuropathologic tissue analysis [24, 32, 33, 50, 56, 79, 81, 120,121,122].

Table 2 A systematic review of 159 reported cases of B. mandrillaris infection: demographics, initial presentation, outcome, and diagnostic methods

First identified in 1989 in a pregnant mandrill baboon at the San Diego Zoo Safari Park though use of immunofluorescence assay (IFA) [87], B. mandrillaris was retrospectively identified in human cases dating back to 1974 [87]. Since then, diagnostic modalities have evolved and include immunohistochemistry, immunofluorescence, and molecular techniques such as PCR, qPCR, RT-PCR, Sanger sequencing, and enzyme-linked immunosorbent assay. These tools have been applied to various specimens, including FFPE tissue, amoebic cultures, blood, and CSF, with varying success. While PCR is a cornerstone of infectious disease diagnostics, its sensitivity for B. mandrillaris remains suboptimal, with reported detection rates as low as 6%, underscoring the limitations of traditional diagnostic methods [11].

In recent years, mNGS has emerged as a potential diagnostic tool for B. mandrillaris and has been applied to specimens including brain tissue, CSF, and, recently, blood. We identified 20 cases reporting positive B. mandrillaris mNGS testing results, with 6 positive tests performed on brain tissue, 14 on CSF, 2 on plasma, 1 on serum, and 1 that did not identify the specimen (see Table 2), with some testing more than one tissue type. One report identified B. mandrillaris using mNGS on serum, also with positive mNGS on CSF and brain [59]. One case reported negative mNGS on CSF in the setting of positive mNGS on brain biopsy in a patient who survived [105]. Our analysis found only two other reported instance of B. mandrillaris detected via plasma mNGS. In both of those cases, CSF was also positive via mNGS and/or PCR [36, 120]. In contrast, in our case, plasma mNGS successfully identified the pathogen, while CSF B. mandrillaris PCR and CSF mNGS were both reported negative.

Upon retrospective review of the CSF mNGS raw data for this case, two reads mapping to B. mandrillaris were identified—far below the assay’s reporting threshold of 50 reads. PCR on this sample was also negative. Our observation of subthreshold reads highlights the challenges of interpreting mNGS data and may reflect several factors, including pathogen distribution between compartments and timing of sample collections. We note that CSF was collected 7 days prior to plasma and thus mNGS of CSF sampled concurrently with plasma may have yielded a positive result. While mNGS of CSF shows utility for many organisms, its sensitivity for B. mandrillaris is unknown. A comprehensive study of nearly 5,000 CSF mNGS results from patients with suspected CNS infections detected subthreshold B. mandrillaris reads in CSF in 2 of 3 identified cases [104]. Detection of B. mandrillaris DNA in plasma but not in CSF raises plasma mNGS as a possible diagnostic tool if CSF testing is inconclusive. However, the sensitivity and specificity of plasma mNGS for the diagnosis of B. mandrillaris GAE are not established. In the absence of these data, results should be interpreted with caution. Diagnostic laboratories offering mNGS testing on CSF could potentially consider reporting rare and difficult-to-diagnose pathogens like B. mandrillaris even if below the usual threshold for reporting.