Intrathecally Expanding B Cell Clones in Herpes Simplex Encephalitis: A Case Report

The 28-year-old female patient presented with headache, fevers, behavioral changes, decreased consciousness, and left hemiparesis for three days. The cranial MRI showed a T2/FLAIR hyperintensity, particularly in the right temporal lobe with leptomeningeal gadolinium enhancement (Fig. 1). The initial CSF analysis showed pleocytosis (292 cells/μl); HSV-1 was detected in the CSF by PCR (411,000 genome equivalents/ml). No CSF cells were collected from this initial lumbar puncture. The patient was initially treated with acyclovir 750 mg i.v. every 8 h. After mild clinical improvement, the patient developed a vertical gaze palsy 5 days after the beginning of treatment. An epileptic seizure occurred and was treated with levetiracetam. Electroencephalography showed isolated sharp waves in the right hemisphere, and a follow-up MRI showed an expansion of FLAIR hyperintensity in the right temporal and parietal lobe. The second CSF analysis, 5 days after the first lumbar puncture, showed 338 cells/μl and an increased viral load (1.4 Mio. genome equivalents/ml). Assuming acyclovir resistance, treatment was changed to foscarnet 40 mg/kg bodyweight i.v. every 8 h, which led to partial recovery with a residual headache and vertical gaze palsy within 7 days. A third CSF analysis 5 days after the second showed reduced pleocytosis (76 cells/μl), and no HSV genes were detected by PCR. The patient was discharged 13 days after the last lumbar puncture for rehabilitation in her home country. CSF cell pellets and PBMCs extracted from peripheral blood samples at the second lumbar puncture (first time point, t1) and at the third lumbar puncture (second time point, t2) were stored in liquid nitrogen for further analysis.

CSF cells at t1 and t2 were analyzed using a single-cell immune profiling technology. PBMCs at t1 and t2 were analyzed using a conventional immune profiling approach featuring bulk sequencing of immunoglobulin (Ig) heavy chains. We determined the expanding CSF B cell clones and the overlap with the peripheral blood B cell repertoire.

After thawing and washing the cryopreserved t1 and t2 CSF samples, a single-cell suspension was prepared using the 10× Genomics Sample Preparation protocol (CG00039, Rev C). Cell count and viability were measured using an automated cell counter (Cellometer Auto 2000, Nexcelom Bioscience), and the volume was adjusted to obtain 1 × 10⁶ cells/ml. Cells were loaded onto a 10× Genomics Chromium chip per the manufacturer’s recommendations, with a targeted cell recovery of 1 × 10⁴ cells. Reverse transcription and library preparation was performed using the 10× Genomics Chromium Single Cell Immune Profiling (v1) kit, following the 10× Genomics protocol (CG000086, Rev J). After quantitative PCR-based library quantification using primers targeting the p5 and p7 (p5: AATGATACGGCGACCACCGAGAT, p7: CAAGCAGAAGACGGCATACGA) regions of Illumina adapters and KAPA Library Quantification Standards (KK4903, Kappy Biosystems), single-cell BCR V(D)J libraries were prepared using the MiSeq Reagent Kit v2 (300-Cycles, Illumina) and sequenced on a Miseq sequencing platform (Miseq Control Software, version 2.6.2., Illumina) with paired-end reads (151 × 151 bp).

PBMC samples from t1 and t2 were thawed and RNA was extracted using the RNeasy® Mini Kit (Qiagen). cDNA libraries were prepared using the Takara SMARTer® Human BCR IgG IgM H/K/L Profiling Kit (Takara) according to the manufacturer’s protocol. Following quantification using quantitative PCR, the library was prepared using the MiSeq Reagent Kit v3 (600-Cycles, Illumina) sequenced on a Miseq platform (Illumina) with paired-end reads (301 × 301 bp).

The CSF B cell libraries (single-cell data) were processed using Cell Ranger (v3.1.0, 10× Genomics) and aligned with human V(D)J reference (vdj_GRCh38_alts_ensembl-3.1.0-3.1.0). The libraries of PBMC samples (bulk sequencing data) were processed with Immune Profiler, modifying the external library to use the latest version of the tool mixcr-3.0.12. Only productive full-length sequences were considered for analysis. After quality filtering (excluding wrong cell types and clonotypes with no V-gene match), we focused only on paired chains. Overlapping B cell clonotypes (based on the same complementary determining region (CDR) 3 nucleotide sequence) between t1 and t2 were selected. We created new clonal groups using a hamming distance of ≤ 1 between CDR3 amino acid sequences and V and J gene identity. We selected the most expanded clones (based on the ratio t2/t1) using the percentage instead of the cell count to correct for sample size bias. Full-length sequences of the chains were extracted from the 10× data and re-analyzed using the IMGT HighVQuest online tool [15].

We looked for B cell clones based on the CDR3 nucleotide sequence (Fig. 2). In the CSF sample from t1 (1964 B cells) we found 1354 clones, and in the CSF sample from t2 (333 cells) we found 260 clones. Among the clones in these two groups, 76 (4.9% of all clones found in CSF) overlapped between t1 and t2. In the PBMCs from t1 (2476 B cells) we found 329 clones, and at t2 (1867 B cells) there were 340 clones. Among the clones in these two groups, 127 (23.4% of all clones in PBMCs) overlapped. This suggests a greater change over time (5 days) in the CSF than in PBMCs. However, it has to be taken into consideration that we applied different sequencing techniques for CSF and PBMC samples, which may represent a limitation. Moreover, upon comparing CSF and PBMCs, we found 71 overlapping clones at t1 (4.4% of all clones; 3.6% overlap of CSF B cells with peripheral B cells) and 14 overlapping clones at t2 (2.4% of clones; 4.2% overlap of CSF B cells with peripheral B cells).

We created clonal groups using a hamming distance of ≤ 1 between CDR3 amino acid sequences. We focused on the clonal groups that showed the greatest expansion based on the ratio between t2 and t1 (Fig. 3). Since the cell counts at t1 (n = 1964 cells) and t2 (n = 333 cells) were different, we calculated the ratio as the percentage of cells in one clonal group at t2 in relation to the percentage of cells in the same clonal group at t1. We found 87 clonal groups with a ratio above 1 (i.e., expanding clonal groups) and 4 clonal groups with a ratio below 1. Expanding clonal groups of CSF B cells suggest an ongoing intrathecal B cell response between t1 and t2. The mean expansion ratio was 4.7 and the median of the expansion ratio was 5.9. Among 87 clonal groups with expansion ratios ≥ 1, we found 16 that overlapped with the BCR repertoire of PBMCs. The CDR3 amino acid sequences of a selection of interesting clonal groups are shown in Table 1, including information about isotype and V(D)J gene usage. This selection includes the six clones with the highest expansion ratio (11.8), two clones (nos. 6 and 8) with a high expansion ratio that were present in a high amount of B cells, and two clones (nos. 141 and 53) that were also a dominant presence in B cells but had lower expansion ratios. This table also shows two clones (nos. 14 and 25) with high expansion ratios that overlapped with PBMCs, and clone no. 2, which was present in a high percentage of the cells at both t1 and t2, but presented a limited expansion ratio. Supplementary Table 1 provides the full nucleotide sequences of CDR3 for all the clones/clonal groups with expansion ratios ≥ 1.