Venous blood from healthy individuals was collected, and peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation using Lymphoprep (Axis-Shield, Oslo, Norway) and washed twice in cold phosphate buffered saline (PBS)/2 mM ethylenediaminetetraacetic acid. For cryopreservation, PBMCs were frozen in fetal bovine serum (Thermo Fisher Scientific, MA, USA)/10% dimethyl sulfoxide (DMSO; Sigma-Aldrich, MO, USA) at −150°C; some were stained with carboxyfluorescein succinimidyl ester (CFSE; Molecular Probes, MA, USA) prior to freezing. As described previously, cord blood was collected from newborns and PBMC separated as for venous blood (1). Cerebrospinal fluid (CSF) from treatment-naïve patients with relapsing remitting multiple sclerosis (RRMS) was collected in a polypropylene tube on ice and immediately centrifuged for 10 min at 400 g to separate cells from fluid. All patients were diagnosed with RRMS based on the 2017 revised McDonald criteria (33).

Cells applied to flow cytometry were incubated with FcR blocking reagent (Miltenyi Biotec, Bergisch Gladbach, Germany) to prevent nonspecific Ab binding, and thereafter stained in PBS/2% FBS/0.02% NaN₃ with a combination of fluorochrome-conjugated antibodies and live/dead stain (Thermo Fisher Scientific, MA, USA). Isotype-matched controls were used to correct for nonspecific Ab binding and spectral overlap, where appropriate. Data were acquired on a FACSymphony or FACSCanto II flow cytometer, both from BD Biosciences (CA, USA), and data analysis was performed using the software FlowJo (TreeStar, OR, USA). Antibodies used are listed in Supplementary Table S1 . Gating strategy to identify CD20⁺ T cells include gating of lymphocytes in a FSC/SSC dot plot, single cells in a FSC-A/FSC-H dot plot, live cells (corresponding to cells negative for the live/dead stain), and hereafter CD20⁺ T cells in a CD3/CD20 dot plot.

For trogocytosis, purified untouched B cells and untouched CD20⁻ T cells from freshly drawn blood were used. B cells were first isolated from PBMCs using a negative selection kit, followed by depletion of potential remaining T cells to ensure removal of all CD20⁺ T cells; both kits were from Stem Cell Technologies (Vancouver, Canada). The purity test showed <0.15% T cells, <0.25% NK cells, and <0.25% monocytes. For CD20⁻ T cell isolation, T cells were first negatively selected using a kit from Stem Cell Technologies followed by fluorescence-activated cell sorting based on antibodies recognizing CD20 (PE-Cy7, 2H7) to discriminate CD20⁻ from CD20⁺ T cells and CD19 (APC, HIP19) to ensure exclusion of potential remaining B cells, both antibodies from BioLegend (CA, USA). The purity of isolated CD20⁻ T cells was 100%. An example is shown in Figures 1A–D . To accommodate T:B cell interaction and trogocytosis, 300,000 B cells were pulsed for 20 h with 30 µg/ml glatiramer acetate and 300,000 CD20⁻ T cells added. The T:B cell co-culture was incubated for 1 h and cells analyzed by flow cytometry for identification of transferred proteins, as described above. Growth media used was RPMI 1640 (Thermo Fisher Scientific, MA, USA)/5% human AB serum (Invitrogen, MA, USA)/pen-strep (Thermo Fisher Scientific, MA, USA). The assay was performed two independent times with different donors.

T cells were negatively selected from three healthy donors using a kit from Stem Cell Technologies, and CD20⁺ and CD20⁻ T cells were isolated by fluorescence-activated cell sorting using antibodies targeting TCRαβ (APC; IP26) and CD20 (PE-Cy7, 2H7) both from BioLegend. The purified T cells were lysed in QIAzol (QIAGEN, Hilden, Germany) lysis buffer, chloroform added, and samples separated into an aqueous and an organic phase by centrifugation. RNA was precipitated from the aquas phase using an RNeasy kit from Qiagen and used for quantitative polymerase chain reaction (qPCR) or affymetrix analysis. For real-time qPCR analysis, synthesis of cDNA was performed on total RNA using the High-Capacity cDNA Reverse Transcription Kit (Life Technologies, Carlsbad, USA) as prescribed by the manufacturer, followed by amplification on a ViiA™ 7 real-time PCR System using TaqMan™ Fast Advanced Master Mix and commercially manufactured primer/probe kits (MS4A1: Hs00544819_m1, SDHA: Hs00417200_m1, RNA18S5, Hs03928989_g1) (Thermo Fisher Scientific, MA, USA). MS4A1 and SDHA were amplified on undiluted cDNA, whereas a 1:1000 dilution was used for RNA18S5. Index qPCR values were calculated using the 2^{-ΔΔ C T} method with SDHA and RNA18S5 as reference genes relative to mean MS4A1 from CD20⁻ T cells. Affymetrix analysis was performed at the Core facility for Genomic Medicine, the Kennedy Center, Rigshospitalet, Denmark, according to the manufacturer. The array used was the Human Gene 2.0 ST Array.

Cryopreserved PBMCs from healthy controls were thawed and T cells negatively selected using a kit from Stem Cell Technologies, according to the manufacturer. Purified T cells were then cultured at a concentration of 1 × 10⁶ cells/ml for 48 h in RPMI 1640/5% human AB serum/pen-strep with 0–20 µM FoxO1-inhibitor (AS1842856; Merck, Darmstadt, Germany). Hereafter, CD20 expression on T cells was analyzed by flow cytometry as described above. The experiment was performed three independent times with a total of six individuals.

Cryopreserved carboxyfluorescein succinimidyl ester (CFSE)–labeled PBMC from three healthy donors were thawed, with T cells negatively selected, followed by depletion of potential remaining B cells using kits from Stem Cell Technologies, according to the manufacturer. Complete removal of B cells was ensured by flow cytometry prior to T-cell cultivation in flat-bottomed plates coated over night at 4°C with 0–4 µg/ml anti-CD3 antibodies (OKT3; BioLegend) and 1 µg/ml anti-CD28 antibodies (CD28.2; BD Biosciences). The purified T cells were stimulated at a concentration of 1 × 10⁶ cells/ml RPMI 1640/5% human AB serum/pen-strep for 24 h, and TCRαβ and CD20 expression as well as cell death were analyzed by flow cytometry. Annexin-V (BioLegend) and live/dead staining (Thermo Fisher Scientific, MA, USA) were used to measure early and late apoptosis, respectively. The purified T cells were also stimulated for 4 days to induce cell proliferation and CD20 expression measured on proliferated (CFSE-low and CFSE-intermediate) and non-proliferated (CFSE-hi) T cells by flow cytometry as described above.

Cryopreserved CFSE-labeled PBMC from healthy donors were thawed and B cells depleted using a kit from Stem Cell Technologies, according to manufacturer. Complete removal of B cells was ensured by flow cytometry prior to cultivation of 500,000 PBMC in 20 µg/ml myelin oligodendrocyte glycoprotein (MOG; AnaSpec Inc, CA, USA), 2.5 × 10⁶ candida albicans cells/ml (CA; InvivoGen, CA, USA), 1.5 µg/ml varicella-zoster virus antigen from infected cell extracts (VZV; Medix Biochemica Group, MO, USA), or RPMI 1640/5% human AB serum/pen-strep with no additives for 6 days. Hereafter, PBMC were stained for flow cytometry and CD20 expression measured on proliferated (CFSE-low) and non-proliferated (CFSE-hi) T cells. The experiment was performed three independent times with a total of five individuals.

To compare MS4A1-mRNA levels between CD20⁻ and CD20⁺ T cells, a student’s t-test was applied. For statistical analysis of CD20⁺ T cell frequency and CD20 MFI following FoxO1-inhibition, a Friedman test and post hoc Dunn’s test for multiple comparisons was performed. In the Dunn’s test the mean rank of unstimulated cells was compared to the mean rank of each of the concentrations applied. To compare CD20⁺ T-cell frequencies in the proliferation studies, paired t-tests were performed.

All participants gave informed, written consent to participation. The study was approved by the regional scientific ethics committee (protocol number H-17005703 and H-16047666).

To confirm the ability of human CD20⁻ T cells to receive CD20 from B cells and hence develop into CD20⁺ T cells through trogocytosis, we cultured isolated CD20⁻ T cells with autologous antigen-primed B cells ( Figures 1A–D ). Within 1 h, we observed that CD20 ( Figure 1E ) together with a selected set of B-cell surface molecules, including CD21, CD22, and CD180 ( Figure 1F ), were transferred to the surface of the T cells. CD21, CD22, and CD180 were only detected on T cells that acquired CD20 ( Figure 1F ); not on the remaining CD20⁻ T-cell population ( Figure 1G ).

If all CD20⁺ T cells in the blood of humans are a product of trogocytosis, a co-transfer of other B-cell molecules would be expected (34, 35). Analyzing the co-expression of CD20 with other B-cell markers on T cells from healthy donors showed that only CD20 was expressed on human blood T cells, not the B-cell markers CD21, CD22, and CD180 observed in the trogocytosis assay ( Figures 1H–J ). As the scenario may differ in tissue during inflammation, we additionally measured the expression of CD20 and other B-cell markers on T cells in the CSF of newly diagnosed untreated patients with relapsing-remitting multiple sclerosis (RRMS). As observed in the blood, neither CD21, CD22 nor CD180 was co-expressed with CD20 on CD20⁺ T cells in CSF ( Figures 1K–M ). The flow cytometry experiments were performed three independent times; representative examples are shown.

In the process of trogocytosis, T cells interact with B cells and become activated by signals induced through the TCR. Therefore, T cells that have gone through trogocytosis are antigen-experienced. To further investigate the possibility that CD20⁺ T cells arise from trogocytosis, we measured the activation and differentiation status of CD20⁺ T cells in the blood of healthy individuals. For this, PBMCs were stained with fluorescence labeled antibodies identifying CD20⁺ and CD20⁻ T cells along with the differential markers CD45RA and CCR7 and analyzed the samples by flow cytometry. To ensure proper identification of naïve T cells according to CD45RA and CCR7, we included staining of cord blood cells, which represents a naïve phenotype. This gating strategy identified a population of blood CD45RA⁺⁺CCR7⁺⁺ CD20⁺ T cells, corresponding to the naïve T cell population found in cord blood ( Figures 2A–E ). To confirm that these blood T cells were truly naïve, they were concomitantly stained with a range of antibodies against surface markers frequently used to identify antigen-naïve T cells (CD31, CD62L, CD27/CD38) and antigen-primed T cells (CD45RO, PD-1, CD49d, CD57). This showed that naïve T cells identified both within the population of CD20⁺ and CD20⁻ T cells stained positive for CD31, CD62L, and CD27^hiCD38^hi and negative for CD45RO, PD-1, CD49d, and CD57, strengthening the classification as a true naïve CD20⁺ T-cell population ( Figure 2F ). As a staining control of activation and differentiation markers, effector memory T cells (CD45RA⁻CCR7⁻) are shown ( Figure 2G ). The data shown are from three independent experiments with n = 3 healthy donors.

As our data indicate that CD20 on T cells may not only originate from trogocytosis in humans, we next measured endogenous MS4A1-mRNA, which translates into CD20 protein. For this, CD20⁺ and CD20⁻ T cells were isolated by fluorescence-activated cell sorting and qPCR applied. In contrast to CD20⁻ T cells, MS4A1-mRNA was detected in CD20⁺ T cells ( Figure 3A ). Also, an Affymetrix analysis of sorted cells showed that MS4A1-mRNA was measurable in both CD4⁺ and CD8⁺ CD20⁺ T cells in contrast to CD19, CD21, CD22, and CD180 ( Figures 3B, C ), strengthening our qPCR data.

CD20 expression is regulated both by epigenetic and transcription factors (28). The half-time of CD20 protein in human primary resting B cells is 71 h; hence, transcriptional activity of MS4A1 is to be expected within this timeframe (36). FoxO1 is a negative MS4A1 transcription factor, and the FoxO1-inhibitor AS1842856 has been shown to induce CD20 expression in B cells (37). To investigate if CD20⁺ T cells have endogenous CD20 production, we therefore incubated T cells with AS1842856 for 48 h and measured the frequency of CD20⁺ T cells and the CD20 surface expression level (MFI). This showed that FoxO1 inhibition induced both the frequency of CD20⁺ T cells and the MFI of CD20 ( Figures 3D–F ), further indicating endogenous CD20 production. The experiment was performed three independent times with a total of six healthy individuals.

To further analyze endogenous production of CD20 in CD20⁺ T cells, we investigated whether CD20 on CD20⁺ T cells is diluted following antigen-induced cell proliferation. For this analysis to be conclusive, a stable surface expression of CD20 is a prerequisite, that is, CD20 must not be shed or internalized following TCR priming. Likewise, CD20⁺ T cells must not suffer from excessive activation-induced cell death. To ensure this, we stimulated purified T cells from three healthy donors for 24 h with increasing concentrations of plate-bound anti-CD3 and anti-CD28 antibodies and observed that CD20 was stably expressed on the surface of T cells in contrast to TCRαβ, which was strongly downregulated ( Figures 4A–C ). Staining of the T cells with a life/dead stain and annexin-V as a marker of early apoptosis showed that CD20⁺ T cells were more prone to activation-induced cell death than their CD20⁻ counterparts ( Figures 4D, E ). A previous study has shown a transient increase in CD20 expression on isolated T cells after 4 days of in-vitro stimulation in non-human primates (9). We therefore stimulated purified and CFSE-labeled T cells for 4 days with anti-CD3 and anti-CD28 antibodies and observed an increased CD20-expression in the early proliferation state of T cells where the cells increase in size (blasts) ( Figures 4F–H ; CFSE-intermediate population). Following cell retraction, we observed a normalized CD20-expression pattern of the T cells ( Figures 4F–H ; CFSE-low population).

To validate that CD20 is not diluted on the surface of proliferated T cells in a more physiologically appropriate setting, we next purified and CFSE-labeled PBMCs from five healthy donors and removed all B cells from the samples ( Figure 4I ). Candida albicans (CA), varicella-zoster virus antigen (VZV), or myelin oligodendrocyte glycoprotein (MOG) was added, and the cells were cultured for 6 days ( Figure 4J ). Due to low antigen-reactivity for some donors, we analyzed only non-proliferated and proliferated T cells (i.e., CFSE-intermediate T cells were included in the proliferated population). This experiment confirmed that CD20 is not diluted on daughter cells ( Figure 4K ).