Abstract
Background and Objectives
Autologous hematopoietic stem cell transplantation (AHSCT) can induce long-lasting immune tolerance and disease quiescence in patients with severe multiple sclerosis (MS), but the underlying mechanisms are not well understood. We hypothesized that AHSCT would induce persistent decreases in inflammatory cytokines, chemokines, and axoglial damage biomarkers in serum and CSF.
Methods
Longitudinal serum and CSF samples were obtained from 24 participants with severe MS undergoing AHSCT in the HALT-MS trial (NCT00288626), with serum collected at 9 time points from pretreatment through month 60 and CSF collected at 3 time points, including pretreatment, month 24, and month 48. Approximately 50 soluble biomarkers were analyzed to determine treatment-induced changes and associations with disease activity, including clinical and MRI parameters.
Results
AHSCT resulted in decreased expression of many proinflammatory cytokines, chemokines, and neuroinflammatory markers for several years in both serum and CSF, including neurofilament light chain (NfL). CSF levels of glial fibrillary acid protein and IL-7, on the contrary, increased after treatment. Expansion of cytomegalovirus and/or Epstein-Barr virus early after AHSCT resulted in transient increases in inflammatory cytokines in serum, but these changes did not persist. Changes in serum were largely distinct from changes in CSF, and concentrations of cytokines and chemokines in serum rarely correlated with concentrations in CSF, except for NfL.
Discussion
These results confirm the hypothesis that AHSCT results in long-lasting reductions in inflammatory cytokines, chemokines, and axoglial damage markers, consistent with persistent disease quiescence. The lack of relationships between biomarkers in serum and CSF suggests that these samples largely reflect distinct immunologic processes and compartments, and that analysis of CSF may be warranted for deeper understanding of disease processes in the CNS despite the challenges associated with performing lumbar punctures. Expression of IL-7 in CSF may be an underappreciated biomarker of the inflammatory state in the CNS and warrants further study.
Introduction
Autologous hematopoietic stem cell transplantation (AHSCT) can induce immune tolerance in severe multiple sclerosis (MS),1 but the underlying processes are not well understood. Changes in inflammatory and immunoregulatory mediators in serum and CSF after AHSCT may reveal possible mechanisms.2 Inflammation in MS causes myelin, glial, and axonal injury through both direct and indirect mechanisms. Neurofilament light chain (NfL) and glial fibrillary acid protein (GFAP) are byproducts of axonal degradation and glial cell injury, respectively, which reflect 2 distinct processes involved in neuronal damage and associate with or predict disease activity in MS.3 Previous studies have demonstrated decreased NFL after AHSCT, whereas GFAP remained unchanged.4
Little is known about the concomitant effects of AHSCT on a large panel of inflammatory cytokines, chemokines, and neuro-axo-glial damage biomarkers in both serum and CSF, together with their relation to clinical outcomes. Furthermore, cytomegalovirus (CMV) and Epstein-Barr virus (EBV) viremia are common early after AHSCT, but the effects on serum and CSF biomarkers have not been previously examined. Against this background, we investigated these relationships.
In the HALT-MS study, 24 participants with severe relapsing-remitting MS (RRMS) underwent AHSCT, of whom 17 achieved durable remission and 7 met the primary end point for recurrence of disease activity.5 We previously demonstrated that treatment resulted in the removal of most of the preexisting T-cell repertoire in both CSF and circulation,6 transiently increased the frequency of activated Treg, and decreased the frequency of Th17.1 cells.7
We hypothesized that reduced MS activity would be associated with decreased proinflammatory cytokines and increased regulatory mediators. Similarly, we hypothesized that AHSCT would reduce NfL and GFAP, reflecting reduced neuroglial injury. To test these hypotheses, we examined approximately 50 cytokines, chemokines, and biomarkers of neuroinflammation and neurodegeneration, in both serum and CSF, before and after AHSCT. We analyzed serum for both acute effects (up to 6 months after transplant) and lasting changes (one to five years after transplant) and compared these with changes in CSF (two and four years after transplant). Analysis of longitudinal CSF samples is particularly valuable, given that CSF may be more representative of CNS immunopathology than serum. Longitudinal serum and CSF also allowed us to assess the relationship between the 2 samples over time.
Methods
Patient Cohort and Sample Collection
Demographics, treatment regimen, and clinical outcomes of the HALT-MS study (NCT00288626) were previously described.5 In brief, in this prospective, open-label, single-arm, multicenter phase 2 trial of AHSCT for severe RRMS, participants underwent stem cell mobilization with filgrastim, followed by leukapheresis. Conditioning included carmustine, etoposide, cytarabine, melphalan (BEAM regimen), and rabbit anti-thymocyte globulin. CD34-selected cells were infused on day 0, followed by filgrastim and prednisone.
Brain MRI was performed at screening, baseline, 2 months, 6 months, and annually through 5 years on scanners with a 1.5T field strength. Scans were analyzed centrally (NeuroRx, Montreal, Canada).
Serum was collected at baseline (prior to mobilization and conditioning, 1–2 months before transplant) and at 1, 2, 6, 12, 24, 36, 48, and 60 months after transplant.
Lumbar punctures were performed at screening (2–4 months prior to transplant) and 24 and 48 months after transplant.
CSF samples were also collected from 6 non-MS, noninflammatory controls undergoing lumbar puncture for reasons such as normal pressure hydrocephalus or idiopathic intracranial hypertension (eTable 1).
Definition of New Disease Activity
Of 24 participants, 7 had new disease activity, as defined in the HALT-MS protocol.5 Each met at least one of the following criteria during the 5-year period after transplant:
Change in Extended Disability Status Scale (EDSS) score of >0.5 compared with baseline; progression was confirmed 3 months later.
Two or more independent lesions indicative of MS on brain MRI performed 1 year or more after transplant; lesions were either gadolinium-enhancing lesions and/or new T2-weighted lesions.
Relapse, defined as the development of a new neurologic sign and corresponding symptom, or worsening of an existing neurologic sign and symptom, localized to CNS white matter, resulting in neurologic deficit/disability and lasting over 48 hours.
Standard Protocol Approvals, Registrations, and Patient Consents
The HALT-MS study (NCT00288626) was approved by local institutional review boards, and all participants provided written informed consent.5
Cytokine Analysis by Meso Scale Diagnostics, LLC. (MSD)
V-PLEX® and S-PLEX® assays were used (Meso Scale Diagnostics, LLC., Rockville MD). Additional information on panels used is provided in eMethods. Summary statistics are provided in eTables 2 (CSF) and 3 (serum).
Cytokine Analysis by Muraro Lab
Soluble CD27 (sCD27) measurement used the DuoSet ELISA (R&D Systems, UK), following the manufacturer's instructions. MSD plates were customized for CCL19, CXCL10, and CXCL13, and the assays were performed according to the manufacturer's instructions. All samples were analyzed in duplicate.
CSF IgG and IgM Oligoclonal Band (OCB) Analysis
Serum and CSF albumin, IgG, and IgM were quantified by nephelometry on a BN ProSpec analyzer (Siemens Healthcare Diagnostics).
Oligoclonal IgG and IgM bands were measured in CSF and serum by isoelectric focusing and immunoblotting, as previously described.8,9
Intrathecal IgG and/or IgM was demonstrated by the presence of 2 or more oligoclonal bands in CSF that were not present in paired serum.
Detection of CMV and EBV Viremia
CMV viral loads were monitored by pp65 antigenemia or DNA assay from day 14 until 3 months after transplant if seronegative for CMV at baseline, or until 6 months after transplant if seropositive for CMV at baseline. EBV viral loads were monitored by PCR on a weekly basis from day 14 through day 100 after transplant and then every other week until month 6.
Statistics
Paired t tests were used for changes in CSF. Linear mixed-effects models were used for changes in serum and relationship with EBV or CMV, using the nlme package for R. Intraindividual associations between cytokines across multiple time points were assessed using the rmcorr package for R. Pearson correlation coefficients were calculated for relationships between cytokines. Association between clinical measures and cytokines was assessed using linear regression, with age and disease duration included as covariates. All statistical analyses were conducted in R 4.4.2.
Data Availability
All data will be publicly available from the Immune Tolerance Network TrialShare website.
Results
AHSCT Results in Decreased Expression of Disease-Associated Cytokines, Chemokines, and Neuroinflammatory Markers in Both CSF and Serum
The effects of AHSCT on the inflammatory milieu in the brain and spinal cord were assessed by comparing CSF samples obtained before transplant and at months 24 and 48 after transplant. At month 24, 11 mediators significantly decreased, while 2 significantly increased (Figure 1B). Among those downregulated were many inflammatory mediators associated with MS, including IFNγ, IL-12/23p40, IL-12p70, IgM, IgG, and IgG index (Figure 1, A and B). IL-10, which may play both immunostimulatory and immunoregulatory roles, was also downregulated at month 24. Other inflammatory markers that were downregulated included soluble ICAM1, CCL22, and CCL19. NfL was significantly downregulated at both 24 and 48 months after transplant, consistent with suppression of disease activity (Figure 1B). We hypothesized that AHSCT would increase anti-inflammatory or immunoregulatory mediators in CSF, but the only 2 factors increased at month 24 were IL-7, a T-cell growth and survival signal, and GFAP, a marker of astrocyte activation; both factors were no longer increased at month 48 (Figure 1B). NfL was the only biomarker that remained significantly different from baseline at month 48, although the sample size was small (n = 24 at baseline, n = 18 at 24 months, n = 7 at 48 months).
Figure 1. AHSCT Reduces Inflammatory Mediators in CSF of Patients With RRMS.
(A) Volcano plot of log-transformed fold changes in CSF biomarker concentrations from baseline to month 24 after transplant vs log-transformed p values. The horizontal dotted line represents p < 0.05 (not adjusted for multiplicity). Any data that occurred after a participant met the primary end point for new disease activity were removed because of confounding effects of additional treatment. N = 24 at baseline, n = 18 at month 24. (B) Changes in CSF levels for a selected subset of biomarkers—NfL, GFAP, IL-12/23p40, and IL-7—at baseline (n = 24), month 24 (n = 18), and month 48 (n = 8) after transplant. Red lines indicate participants who met the primary end point for new disease activity before the end of study at month 60 (n = 3), and black lines indicate participants without evidence of new disease activity (n = 15 at month 24, n = 8 at month 48). Hollow circles at baseline represent participants with baseline samples only. Hollow red circles indicate participants who met the primary end point for new disease activity before the end of study at month 60 (n = 4), and hollow black circles indicate participants without evidence of new disease activity. *p < 0.05 vs baseline by the paired t test. (C.) Selected CSF cytokine concentrations at baseline, month 24, and month 48 compared with CSF samples from noninflammatory neurologic disease (NIND) controls. Biomarker concentrations in study samples were normalized with NIND controls by the standard score method, followed by statistical analysis using the Wilcoxon rank-sum test. *p < 0.05. GFAP = glial fibrillary acid protein; NfL = neurofilament light chain; RRMS = relapsing-remitting MS.
We found a significant decrease in IgG concentration, IgM concentration, and IgG index in CSF at month 24 (Figure 1A and eFigure 1). Given the importance of IgM as a biomarker of MS severity, we also assessed IgM OCBs.10 Six of 24 participants were positive for IgM OCBs at baseline, and 3 of those 6 became negative for IgM OCBs by 48 months (data not shown). Decreased CSF IgG and IgM, as well as eradication of IgM OCB in 3 of 6 positive participants, suggest that AHSCT may have long-term impacts on the development or maintenance of plasma cells in the brain and/or spinal cord.
To examine the disease relatedness of changes in CSF cytokines, we compared cytokine concentrations in our cohort with CSF samples from patients with noninflammatory neurologic disorders (NINDs), including normal pressure hydrocephalus, idiopathic intracranial hypertension, and samples collected to rule out MS or infection (eTable 1, n = 6). The goal was to explore how the change in CSF biomarkers observed after therapy compared with “normal” levels, i.e., levels seen in CNS disorders that are not characterized by an overt inflammatory process. Consistent with reduced immune activation in the CNS after AHSCT, IFNγ, IL-12/23p40, and IL-10 were reduced to levels equivalent to NIND controls (Figure 1C). Of interest, increased IL-7 concentrations found in CSF after AHSCT were consistent with higher levels in NIND controls. Higher levels of IL-7 in the CSF in the absence of inflammation may be attributable to reduced consumption, reflecting reduced T-cell infiltration in the CNS. Although sample size for NIND controls was limited by the difficulty in acquiring CSF samples, the results are consistent with a “normalization” of inflammatory cytokines after AHSCT.
Inflammatory mediators were analyzed in serum at pretransplant and at 8 time points from month 1 (n = 24) through month 60 (n = 10) after transplant. At 1 month after transplant, there were significant changes in 30 of the 45 detectable factors (Figure 2A). While many of these changes waned over time, a few remained consistently altered at months 48 and/or 60 (Figure 2A). Many changes in the first 6 months were likely associated with immune ablation and may not necessarily represent the persistent “immune reset” associated with AHSCT. For example, NfL significantly increased at 1 and 2 months after transplant but significantly decreased vs baseline from month 24 onward (Figure 2B). GFAP similarly increased at months 1 and 2 but then returned to baseline levels (Figure 2B). To address changes associated with lasting immune re-balancing, we focused on serum biomarkers that were significantly different from baseline at 2 or more time points from month 12 or later. Biomarkers persistently decreased included 2 cytokines (TNFα and IL-4), one chemokine (CCL4), and 4 neuroinflammatory markers (NfL, BDNF, NPTX1, and aMMP9) (Figure 2C). Biomarkers persistently increased include 4 cytokines (IL-7, IL-2, GM-CSF, and TNFβ), one chemokine (CCL17), and one neuroinflammatory marker (NfH). Of interest, only 2 changes were consistent between serum and CSF—decreased NfL and increased IL-7 (Figure 2C).
Figure 2. AHSCT Results in Lasting Changes in Inflammatory Mediators in Serum of Patients With RRMS.
(A) Volcano plot showing changes in serum cytokines, chemokines, inflammatory markers, and neuroinflammatory mediators from baseline (n = 24) to month 6 (n = 24), month 12 (n = 22), month 24 (n = 22), and month 48 (n = 17) after transplant. (B) Changes in serum NfL and GFAP after transplant. Red lines indicate participants who met the primary end point for new disease activity before the end of study at month 60 (n = 7), and black lines indicate those without evidence of new disease activity (n = 17). *p < 0.05 vs baseline using the linear mixed-effects model. BL = baseline. (C) Venn diagrams depicting the overlap between biomarkers that are persistently changed in serum (definition of persistently changed = significantly changed at more than one time point from month 12 onward) and significantly changed in CSF at month 24 and/or month 48.
Strong Correlations Between Serum and CSF Cytokine Concentrations Are Rare
The degree to which serum biomarkers are representative of the CNS is an important question, given the limitations in performing lumbar punctures. The connection between serum and CSF may depend on permeability of the blood-brain barrier, which may be affected by the inflammatory state.11 We assessed correlation between serum and CSF, including at baseline, month 24, and month 48. There was a strong correlation between serum NfL (sNfL) and CSF NfL (cNfL) (r = 0.86, p = 2 × 10−8, Figure 3), confirming that serum is a reliable surrogate for CSF throughout different phases of disease activity. Of interest, there was a significant correlation between serum GFAP (sGFAP) and CSF GFAP (cGFAP) at baseline (r = 0.70, p = 1.2 x 10^-4), but not when all 3 time points were included (r = 0.18, p = 0.386, Figure 3), suggesting that sGFAP may be less reliable for assessment of glial cell activity in the CNS during disease quiescence. Of the 12 cytokines that were detectable in both serum and CSF, only 2 had a statistically significant, albeit weak, correlation: IL-12/23p40 (r = 0.48, p = 0.01) and TNFα (r = 0.41, p = 0.035). In addition, there was a significant correlation between serum and CSF for IgM (r = 0.64, p = 4.8 × 10−4), but not IgG (r = 0.26, p = 0.22). For comparison, there was a strong correlation between serum and CSF for serum amyloid A (SAA, r = 0.84, p = 7 × 10−8) and C-reactive protein (CRP, r = 0.56, p = 0.003), demonstrating that a strong correlation between the 2 compartments is feasible. The lack of correlation for most cytokines suggests that serum and CSF may reflect distinct immunologic niches, reinforcing the observation that AHSCT-induced changes in serum and CSF were largely nonoverlapping.
Figure 3. Persistent Strong Correlations Between Biomarker Concentrations in Serum and CSF Are Rare.
Correlations between paired serum and CSF concentrations are shown for selected biomarkers, with serum concentration on the x-axis and CSF concentration on the y-axis (log scales). Time points are represented by color. Repeated measure correlation analysis between serum and CSF are shown in the lower right of each plot, with correlation coefficient (R) and p value. Of 31 biomarkers that were measured and detectable in both serum and CSF, only 7 were significantly correlated (p < 0.05).
Many Inflammatory Mediators in CSF Are Coordinately Regulated and Independent of NfL
To further characterize relationships between cytokines and MS pathophysiology, we assessed the correlation between cytokines and established biomarkers of neurodegeneration, NfL and GFAP. At baseline, there was a strong positive correlation between cNfL, cNfH, IgM index, and IgG index (Figure 4A). Surprisingly, however, there were no strong positive correlations between cNfL and any cytokines/chemokines in CSF. CSF IL-7 (cIL-7), on the contrary, was negatively associated with cNfL. cGFAP levels were largely independent of all other biomarkers. Strikingly, there were strong positive correlations between CSF concentrations of many cytokines, including IFNγ, IL-12/23p40, TNFα, GM-CSF, IL-2, IL-10, ICAM1, CCL17, CCL19, and CCL22, suggestive of a co-regulated inflammatory network. IL-10 correlated strongly with IgG index, suggesting that, in this setting, IL-10 may play an inflammatory role via positive effects on intrathecal B cells. cIL-7, on the contrary, was negatively correlated with many inflammatory mediators, including IgG index, IL-12/23p40, IL-10, ICAM1, and CCL22.
Figure 4. Subsets of Inflammatory Mediators Are Correlated in CSF but Not Serum.
(A) Relationships between individual inflammatory mediators in CSF at baseline. Blue denotes strong positive relationships, while red denotes strong negative relationships by Pearson correlation. (B.) Relationships between individual inflammatory mediators in serum at baseline, as in A.
The correlations between baseline cytokines, chemokines, and neurodegenerative markers in CSF were far more striking than in serum. Serum NfL was weakly positively correlated with GFAP and TNFα but negatively correlated with ICAM1 (Figure 4B). Serum GFAP positively correlated with IL-2, while serum IL-7 negatively correlated with NPTX1, TNFα, CCL4, NfH, and IL-10. These results suggest that the relationship between inflammatory mediators is stronger or more easily detected in CSF than in serum.
CMV and EBV Viremia Early After Transplant Alter the Serum Cytokine Milieu
The conditioning regimen for AHSCT results in transient lymphopenia, and infection or reactivation of both CMV and EBV is common.12-14 CMV and EBV viral loads were assessed frequently for the first 3 to 6 months after transplant for safety. Of 24 participants, 10 developed detectable CMV viremia and 12 developed detectable EBV viremia (5 developed both). Our previous studies demonstrated that CMV viremia after transplant can have lasting impacts on T-cell reconstitution, with decreased T-cell receptor diversity and increased ratio of memory-to-naïve T cells.6 Thus, we investigated whether CMV or EBV viremia affected cytokine concentrations. We found that CMV viremia was associated with increased inflammatory mediators in serum 1 month after transplant, including IFNγ, IL-12p70, CXCL10, IL-2, and TNFα (Figure 5A and eFigure 2A). EBV viremia also affected serum biomarkers at 1 month after transplant, with significant increases in GM-CSF, IL-10, and TNFα and significant decreases in CCL17, CCL22, and BDNF (Figure 5B and eFigure 2B). Virus-associated cytokine changes were most robust at 1 month after transplant, with few significant differences at later time points. There were no significant associations between CMV or EBV viremia and biomarker concentrations in CSF (data not shown).
Figure 5. CMV and EBV Viremia After AHSCT Are Associated With Transient Changes in Serum Cytokines and Chemokines.
(A) Association between CMV viremia early after transplant and selected serum biomarker concentrations over time. *p < 0.05 comparing viremia-positive and viremia-negative populations using the linear mixed-effects model. Y-axis scales were log transformed for biomarkers with wide ranges in values. (B) Association between EBV viremia early after transplant and selected serum biomarkers, as in A. Additional plots in Figure S2. AHSCT = autologous hematopoietic stem cell transplantation; CMV = cytomegalovirus; EBV = Epstein-Barr virus.
NfL and GFAP Are the Strongest Biomarkers Associated With Clinical Parameters at Baseline
To better understand the link between soluble mediators in CSF and disease activity in this cohort at baseline, we examined associations with MRI T1 and T2 lesion volume, the number of gadolinium-enhancing (Gd+) lesions, the number of relapses in the 18 months prior to baseline, EDSS scores, and Multiple Sclerosis Functional Composite (MSFC) scores, using linear regression adjusted for age and duration of MS. cNfL was positively associated with T1 and T2 lesion volume, as well as EDSS scores (Figure 6, A and B). Similarly, cGFAP was positively associated with T1 and T2 volumes while negatively associated with MSFC scores (lower MSFC score = worse disease). Both IgG index and CSF IL-10 (cIL-10) were positively associated with the number of relapses in the prior 18 months, while CSF IL-12/23p40, CCL17, and CCL22 were positively associated with the number of Gd+ lesions (Figure 6, A and B). Finally, cICAM1 was inversely associated with MSFC scores, similar to cGFAP. Thus, among participants with active RRMS, increased concentrations of inflammatory biomarkers in CSF were associated with increased disease activity.
Figure 6. CSF and Serum Biomarkers Associated With Disease Activity at Baseline.
(A) Relationships between CSF biomarkers and clinical variables at baseline. Blue lines represent linear regression. The shaded area represents the pointwise 95% confidence interval. Each plot lists the linear regression estimate (t value) and p value (not corrected), adjusted for age and duration of MS. (B) Linear regression estimates (t value) for relationship between baseline CSF biomarkers and baseline clinical parameters, including T1 lesion volume, T2 lesion volume, number of Gd+ lesions, EDSS score, number of relapses in the prior 18 months, and MSFC score, as shown in A. Heatmap only includes statistically significant relationships (p < 0.05); red = positively associated, blue = negatively associated. Lower MSFC scores indicate more severe disease; thus, all significant relationships depicted are positive associations between biomarker concentration and disease activity. (C) Relationships between serum biomarkers and clinical parameters at baseline, as in A. (D) Linear regression estimates showing relationship between serum biomarkers and clinical parameters at baseline for those with p < 0.05, as in B. EDSS = Expanded Disability Status Scale; MSFC = Multiple Sclerosis Functional Composite.
In serum, NfL was positively associated with baseline MRI T1 volume, T2 volume, EDSS score, and lower MSFC score, while GFAP was positively associated with T2 volume and EDSS score (Figure 6, C and D). Among cytokines and chemokines, serum TNFβ was positively associated with the number of Gd+ lesions, while serum CCL4 was positively correlated with MRI T1 and T2 lesion volume (Figure 6, C and D). Only NfL, NfH, and GFAP were positively associated with disease activity in both serum and CSF.
Relationship Between Changes in Biomarker Concentration and Changes in Clinical Variables
AHSCT resulted in a significant decrease in EDSS scores at years 1, 2, 3, and 4, but the increased T1 volume and decreased T2 volume did not reach statistical significance until year 55. Thus, in our cohort, changes in biomarkers at month 24 may precede changes in clinical parameters. Using linear regression adjusted for age and MS duration, we assessed relationships between changes in CSF biomarkers and changes in clinical parameters from baseline to month 24, including T1 and T2 lesion volume, number of GD-enhancing lesions, EDSS score, and MSFC score. Although AHSCT resulted in a significant decrease in cNfL, we were unable to detect any significant relationships between changes in cNfL and changes in clinical parameters. Similarly, we were unable to detect significant relationships between changes in cGFAP and changes in clinical parameters. Decreased cNfH, on the contrary, was associated with disease improvement as measured by the decreased number of GD+ lesions and increased MSFC scores (eFigure A and B). Decreased IgM index was associated with increased T1 lesion volume. Of 13 inflammatory mediators that changed in CSF, only 3 were significantly associated with changes in clinical parameters: decreased IL-12/23p40, ICAM1, and CCL17 were all associated with a decreased number of Gd+ lesions, although this was largely driven by a small subset of participants (eFigure 3, A and B).
For serum (eFigure 3, C and D), decreased concentrations of NfL, TNFα, and CCL4 were associated with changes in MRI parameters, including decreased number of Gd+ lesions, decreased T2 lesion volume, and/or increased T1 lesion volume. Increased serum IL-7, IL-10, TNFβ, and BDNF, on the contrary, were associated with improvement in disease symptoms, as measured by decreased EDSS or increased MSFC scores.
In the HALT-MS study, 7 of 24 transplanted participants met the primary end point for new disease activity, defined as (1) EDSS increase of >0.5 vs baseline, (2) 2 or more independent gadolinium-enhancing and/or new T2-weighted lesions, or (3) worsening neurologic status associated with clinical relapse.5 Of the 7 participants, 2 had an increase in EDSS score, 2 had new MRI activity, and 3 had clinical relapse. Timing of these events was variable, ranging from 5 to 48 months after transplant. Although the small sample size precludes strong conclusions, we assessed whether any inflammatory mediators predicted or were associated with disease activity. Simple comparison of baseline levels of inflammatory mediators in the 7 participants who met the primary end point vs the 17 participants who did not suggested that new disease activity was associated with higher pretreatment levels of sIL-6, sNPTX1, cTNFα, and cIL-2, as well as lower pretreatment levels of sIL-7 (eFigure 4, A and B). There was no association between NfL or GFAP and treatment outcome (not shown). We also explored shifts in inflammatory mediators at time points leading up to new disease activity, focusing on serum because it was sampled more frequently. While no statistical associations could be found, spikes in sNfL were observed in 2 participants who met the primary end point (eFigure 4C). Trends in sIL-7 were heterogeneous, with 2 participants who met the primary end point showing early spikes in sIL-7, while several others who met the primary end point showing lower sIL-7 levels (eFigure 4C).
Discussion
The HALT-MS study demonstrated that AHSCT can induce long-lasting immune tolerance in patients with active RRMS. Consistent with “immune reset,” we previously found removal of most of the preexisting TCR clones in both serum and CSF,6 as well as changes in circulating T-cell phenotypes.7 We hypothesized that this change should be accompanied by changes in the inflammatory milieu. Immune cell depletion and reconstitution is expected to alter expression of cytokines produced by circulating lymphocytes. In turn, suppression of CNS inflammation should result in changes in markers of axonal and glial pathology. Thus, we examined changes in cytokines, chemokines, and axoglial damage markers to better understand mechanisms mediating durable immune regulation after AHSCT.
As expected, AHSCT resulted in decreased expression of many proinflammatory cytokines, chemokines, and neuroinflammatory markers for several years in both serum and CSF. NfL decreased in CSF at both 2 and 4 years after transplant. In serum, NfL was temporarily elevated (months 1 and 2), consistent with previous reports of brief neuroinflammation associated with the conditioning regimen and therapy, followed by a significant decline through year 5.4,15,16 The persistent reduction in NfL is consistent with lasting clinical and radiologic disease remission in HALT-MS.5 The inflammatory cytokines IFNγ and IL-12/23p40 were markedly downregulated in CSF, pointing to a role for the Th1 axis in pathophysiology, as previously suggested.17,18 Previous reports have demonstrated a role for the Th17 axis in MS,19,20 and we have seen reductions in Th17 cells after AHSCT.7 However, the Th17 signature cytokine, IL-17A, was largely undetectable in CSF despite the use of a high-sensitivity assay and was not consistently altered in serum. TNF family cytokines have complex roles in MS,21 but our data support a potential pathogenic role for TNFα. At baseline, serum TNFα correlated with NfL and Gd+ lesions. Moreover, serum TNFα decreased after AHSCT, which was associated with decreased Gd+ lesions. One previous study reported decreased CSF concentrations of CCL3, IL-12B, and CXCL10 after AHSCT.22 IL-12 was the only inflammatory molecule decreased in CSF in both studies, but methodological aspects or the small sample size of our study may explain the differences.
Intrathecal B-cell activation is associated with MS, and the role of B cells in MS has been reinforced by the clinical efficacy of B-cell–targeted therapies.23 We and others previously demonstrated that AHSCT resulted in a durable increase in the ratio of naïve-to-memory B cells in the periphery,7,24 but effects on intrathecal B cells await further investigation. In this study, we confirm that AHSCT resulted in decreased CSF IgG concentration and IgG index.5 We expanded these observations to include decreased CSF IgM concentration and eradication of IgM OCBs in 3 of 6 participants. These results suggest that AHSCT can remove some, but not all, of the clonally expanded antibody-producing cells in the CNS. It is unclear to what extent this is a direct effect of immune ablation vs a downstream effect of changes in the inflammatory milieu. We found decreased CSF IL-10, which can act as a B-cell growth factor, as well as CCL19, which can act as a chemoattractant for both T and B cells, suggesting decreased recruitment and/or decreased growth or survival of intrathecal B cells.25 An indirect mechanism involving the regeneration of plasma blasts and plasma cells from newly generated B cells, and the reduction of B-cell–activating factors, may explain why intrathecal immunoglobulin production declines slowly over years after AHSCT.26
Two increased CSF biomarkers were GFAP and IL-7. Increased GFAP after AHSCT is consistent with one previous report, which attributed it to reactive astrogliosis.15 A moderate increase over 48 months could also be explained by time because it has been demonstrated that serum GFAP increases with age.27 IL-7 is a lymphocyte growth factor that can be expressed by a variety of stromal cells. Increased IL-7 after AHSCT could be a homeostatic response to lymphodepletion, or the result of decreased consumption due to reduced T-cell infiltration. Of interest, other studies have found increased IL-7 in CSF of patients with MS after treatment with natalizumab, IFNβ, and methylprednisolone.28,29 Similar to our findings, change in IL-7 after natalizumab and methylprednisolone was inversely correlated with change in NFL.29 In our cohort at baseline, IL-7 was negatively associated with inflammatory mediators and disease activity. Similarly, other studies29 found a negative correlation between IL-7 and NfL in patients with MS after treatment. These findings support the hypothesis that increased CSF IL-7 is attributable to decreased consumption secondary to decreased lymphocyte infiltration. However, genetic evidence suggests that IL-7 may play a complex role in MS. A polymorphism in the IL-7 receptor alpha chain increases the risk of MS, which is associated with increased expression of soluble IL-7R protein.30,31 The effects of increased soluble IL-7R expression are unclear because some suggest that it decreases IL-7 activity,32 while others suggest that it increases IL-7 activity.33 The role of IL-7 in CNS inflammation requires further research.
We observed little overlap between changes in serum and changes in CSF, as well as few correlations between biomarker concentrations in serum vs CSF, suggesting that the inflammatory milieu between these 2 spaces may be independent, although the small sample size prevents strong conclusions. One explanation could be distinct effects of treatment on cells in the CNS. Little is known about the effects of AHSCT on myeloid cells, which can play an important role in MS pathophysiology.34 CNS-resident microglia, macrophages, and astrocytes can be a major source of cytokines and chemokines, including several that were altered by AHSCT.35,36 Further studies are needed to understand differences between immunoregulatory mechanisms that occur in the CNS vs the circulation.
EBV is known to play a role in the development of MS,37,38 but little is known about the interplay between EBV and the efficacy of AHSCT. Both EBV and CMV viremia are known complications after AHSCT, but we and others have not observed a relationship between viral reactivation and MS activity after AHSCT.12-15,39 We previously demonstrated that CMV viremia after transplant alters T-cell reconstitution, and we can now add that this is accompanied by increased Th1-type inflammatory cytokines. EBV viremia, on the contrary, may have more subtle effects on immune reconstitution and warrants further study. In this study, we observed an association between EBV viremia and increased inflammatory cytokines. Thus, both EBV and CMV reactivation may add to the inflammatory milieu during early immune reconstitution. Further studies and larger cohorts are necessary to examine the relationship between viremia and immune reconstitution after AHSCT.
The main limitation of our study is the small sample size. In the HALT-MS study, 7 of 24 transplanted participants met the primary end point for new disease activity: 3 by clinical relapse, 2 by increase in EDSS score, and 2 by new MRI activity. There was no single biomarker that clearly distinguished the 7 participants with new disease activity from those who maintained disease quiescence, which may reflect heterogeneity in underlying mechanisms and timing of the events. Despite these challenges, our study has multiple strengths, including longitudinal analysis through 5 years after treatment, CSF samples at multiple time points, rich clinical and MRI data, and analysis of diverse biomarkers in both serum and CSF.
A main question remains whether a universal mechanism may exist coordinating or explaining the observed multifactorial changes, which include several anti-inflammatory as well as regulation-enhancing changes. While our study cannot address the question, we suggest that our results fit well in a model postulating that radical unselective ablation of mature adaptive immunity eliminates the proinflammatory cell subpopulations and recruits homeostatic mechanisms to quickly repopulate the immune system. The reconstitution and redifferentiation of adaptive immunity is accompanied by enhancement of regulatory pathways, resulting in a net anti-inflammatory and tolerogenic change after AHSCT.1
In conclusion, our findings support the hypothesis that AHSCT results in long-lasting changes in inflammatory cytokines, chemokines, and axoglial damage factors, with distinct alterations in the CNS vs the circulation.
Acknowledgment
This study was funded in part by the Immune Tolerance Network, which is sponsored by the National Institute of Allergy and Infectious Diseases (NIAID), of the National Institutes of Health (NIH), under award UM1AI109565. Research reported in this publication was also supported by the National Institute of Mental Health (NIMH), of the NIH, under award U24AI118663 to MSD. YMD was supported by grants from NIH NIAID Autoimmune Center of Excellence 2UM1AI144292-06, NIH NCATS 9R44 TR005293-02, and NIH NIDA R01-DA060226. This manuscript is the result of funding in whole or in part by the NIH. It is subject to the NIH Public Access Policy. Through acceptance of this federal funding, NIH has been given a right to make this manuscript publicly available in PubMed Central upon the Official Date of Publication, as defined by NIH. The content and opinions expressed are solely the responsibility of the authors and do not necessarily represent the official policy or position of the NIAID, NIMH, NIDA, NIH, DHHS, or any other agency of the US government.
Glossary
- AHSCT
autologous hematopoietic stem cell transplantation
- CMV
cytomegalovirus
- EBV
Epstein-Barr virus
- EDSS
Extended Disability Status Scale
- GFAP
glial fibrillary acid protein
- MS
multiple sclerosis
- MSFC
Multiple Sclerosis Functional Composite
- NfL
neurofilament light chain
- NIND
noninflammatory neurologic disorder
- OCB
oligoclonal band
- RRMS
relapsing-remitting MS
- sGFAP
serum GFAP
- sNfL
serum NfL
Author Contributions
L.A. Cooney: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. N. Lim: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. K.M. Harris: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design. D.E. Smilek: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. P. Bathala: major role in the acquisition of data. M. Stengelin: major role in the acquisition of data; analysis or interpretation of data. G. Sigal: major role in the acquisition of data. J.N. Wohlstadter: major role in the acquisition of data. M.T. Cencioni: major role in the acquisition of data. J.I. Fernández Velasco: major role in the acquisition of data. Y. Mao-Draayer: major role in the acquisition of data. D.A. Fox: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. L.M. Griffith: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. R. Nash: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. L.M. Villar: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. P.A. Muraro: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.
Study Funding
Funding sources include grants from the following divisions of the NIH: NIAID (award numbers UM1AI109565 and 2UM1AI144292-06), NIMH (award number U24AI118663) NCATS (award number 9R44 TR005293-02) and NIDA (award number R01-DA060226).
Disclosure
L.A. Cooney, N. Lim, K.M. Harris, D.E. Smilek, M.T. Cencioni, J.I. Fernández Velasc, D.A. Fox, L.M. Griffith, R. Nash, and L.M. Villar report no disclosures. P. Bathala, M. Stengelin, and G. Sigal are current or former employees of Meso Scale Diagnostics, LLC. Jacob N. Wohlstadter is the president and chief executive officer and ultimate beneficial owner of Meso Scale Diagnostics, LLC, which was awarded the National Institute of Mental Health (NIMH) U24AI118663 grant that supported some of the work reported in this manuscript. Y. Mao-Draayer has served as a consultant and/or received grant support from Biogen Idec, Celgene/Bristol Myers Squibb, EMD Serono, Sanofi-Genzyme, Genentech-Roche, Novartis, Questor, Janssen, Horizon, and Teva Neuroscience. P. A. Muraro reports research funding from NIHR-EME (UK) and the Benaroya Research Institute (Seattle, USA), and consulting fees from Magenta Therapeutics, Jasper Therapeutics, Teva, Glenmark, and Cellerys AG. Go to Neurology.org/NN for full disclosures.
References
- 1.Muraro PA. Resetting tolerance in autoimmune disease. Science. 2023;380(6644):470-471. doi: 10.1126/science.adg7489 [DOI] [PubMed] [Google Scholar]
- 2.Hendrawan K, Khoo MLM, Visweswaran M, et al. Long-term suppression of circulating proinflammatory cytokines in multiple sclerosis patients following autologous haematopoietic stem cell transplantation. Front Immunol. 2021;12:782935. doi: 10.3389/fimmu.2021.782935 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Meier S, Willemse EAJ, Schaedelin S, et al. Serum glial fibrillary acidic protein compared with neurofilament light chain as a biomarker for disease progression in multiple sclerosis. JAMA Neurol. 2023;80(3):287-297. doi: 10.1001/jamaneurol.2022.5250 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Zjukovskaja C, Larsson A, Cherif H, Kultima K, Burman J. Biomarkers of demyelination and axonal damage are decreased after autologous hematopoietic stem cell transplantation for multiple sclerosis. Mult Scler Relat Disord. 2022;68:104210. doi: 10.1016/j.msard.2022.104210 [DOI] [PubMed] [Google Scholar]
- 5.Nash RA, Hutton GJ, Racke MK, et al. High-dose immunosuppressive therapy and autologous HCT for relapsing-remitting MS. Neurology. 2017;88(9):842-852. doi: 10.1212/WNL.0000000000003660 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Harris KM, Lim N, Lindau P, et al. Extensive intrathecal T cell renewal following hematopoietic transplantation for multiple sclerosis. JCI Insight. 2020;5(2):e127655. doi: 10.1172/jci.insight.127655 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Karnell FG, Lin D, Motley S, et al. Reconstitution of immune cell populations in multiple sclerosis patients after autologous stem cell transplantation. Clin Exp Immunol. 2017;189(3):268-278. doi: 10.1111/cei.12985 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sádaba MC, González Porqué P, Masjuan J, Alvarez-Cermeño JC, Bootello A, Villar LM. An ultrasensitive method for the detection of oligoclonal IgG bands. J Immunol Methods. 2004;284(1-2):141-145. doi: 10.1016/j.jim.2003.09.018 [DOI] [PubMed] [Google Scholar]
- 9.Villar LM, González-Porqué P, Masjuán J, Alvarez-Cermeño JC, Bootello A, Keir G. A sensitive and reproducible method for the detection of oligoclonal IgM bands. J Immunol Methods. 2001;258(1-2):151-155. doi: 10.1016/s0022-1759(01)00492-6 [DOI] [PubMed] [Google Scholar]
- 10.Villar LM, Masjuan J, González–Porqué P, et al. Intrathecal IgM synthesis in neurologic diseases: relationship with disability in MS. Neurology 2002;58(5):824-826. doi: 10.1212/wnl.58.5.824 [DOI] [PubMed] [Google Scholar]
- 11.Cramer SP, Simonsen H, Frederiksen JL, Rostrup E, Larsson HBW. Abnormal blood-brain barrier permeability in normal appearing white matter in multiple sclerosis investigated by MRI. Neuroimage Clin. 2014;4:182-189. doi: 10.1016/j.nicl.2013.12.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Alexander T, Roldan E, Del Papa N, et al. Autologous haematopoietic stem cell transplantation for rheumatic diseases: best practice recommendations from the EBMT Practice Harmonization and Guidelines Committee. Bone Marrow Transpl. 2025;60(11):1451-1464. doi: 10.1038/s41409-025-02695-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Nash RA, Dansey R, Storek J, et al. Epstein-Barr virus-associated posttransplantation lymphoproliferative disorder after high-dose immunosuppressive therapy and autologous CD34-selected hematopoietic stem cell transplantation for severe autoimmune diseases. Biol Blood Marrow Transplant. 2003;9:583-591. doi: 10.1016/s1083-8791(03)00228-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mehra V, Rhone E, Widya S, et al. Epstein-Barr virus and monoclonal gammopathy of clinical significance in autologous stem cell transplantation for multiple sclerosis. Clin Infect Dis. 2019;69(10):1757-1763. doi: 10.1093/cid/ciz047 [DOI] [PubMed] [Google Scholar]
- 15.Ruder J, Dinner G, Maceski A, et al. Dynamics of inflammatory and neurodegenerative biomarkers after autologous hematopoietic stem cell transplantation in multiple sclerosis. Int J Mol Sci. 2022;23(18):10946. doi: 10.3390/ijms231810946 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Thebault S, R Tessier D, Lee H, et al. High serum neurofilament light chain normalizes after hematopoietic stem cell transplantation for MS. Neurol Neuroimmunol Neuroinflamm. 2019;6(5):e598. doi: 10.1212/nxi.0000000000000598 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Magliozzi R, Howell OW, Nicholas R, et al. Inflammatory intrathecal profiles and cortical damage in multiple sclerosis. Ann Neurol. 2018;83(4):739-755. doi: 10.1002/ana.25197 [DOI] [PubMed] [Google Scholar]
- 18.Magliozzi R, Scalfari A, Pisani AI, et al. The CSF profile linked to cortical damage predicts multiple sclerosis activity. Ann Neurol. 2020;88(3):562-573. doi: 10.1002/ana.25786 [DOI] [PubMed] [Google Scholar]
- 19.Darlington PJ, Touil T, Doucet J-S, et al. Diminished Th17 (not Th1) responses underlie multiple sclerosis disease abrogation after hematopoietic stem cell transplantation. Ann Neurol. 2013;73(3):341-354. doi: 10.1002/ana.23784 [DOI] [PubMed] [Google Scholar]
- 20.Darlington PJ, Stopnicki B, Touil T, et al. Natural killer cells regulate Th17 cells after autologous hematopoietic stem cell transplantation for relapsing remitting multiple sclerosis. Front Immunol. 2018;9:834. doi: 10.3389/fimmu.2018.00834 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Fresegna D, Bullitta S, Musella A, et al. Re-examining the role of TNF in MS pathogenesis and therapy. Cells. 2020;9(10):2290. doi: 10.3390/cells9102290 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Burman J, Zjukovskaja C, Svenningsson A, Freyhult E, Wiberg A, Kultima K. Cerebrospinal fluid cytokines after autologous haematopoietic stem cell transplantation and intrathecal rituximab treatment for multiple sclerosis. Brain Commun. 2023;5(1):fcad011. doi: 10.1093/braincomms/fcad011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Cencioni MT, Mattoscio M, Magliozzi R, Bar-Or A, Muraro PA. B cells in multiple sclerosis - from targeted depletion to immune reconstitution therapies. Nat Rev Neurol. 2021;17(7):399-414. doi: 10.1038/s41582-021-00498-5 [DOI] [PubMed] [Google Scholar]
- 24.von Niederhäusern V, Ruder J, Ghraichy M, et al. B-cell reconstitution after autologous hematopoietic stem cell transplantation in multiple sclerosis. Neurol Neuroimmunol Neuroinflamm. 2022;9(6):e200027. doi: 10.1212/NXI.0000000000200027 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Krumbholz M, Theil D, Steinmeyer F, et al. CCL19 is constitutively expressed in the CNS, up-regulated in neuroinflammation, active and also inactive multiple sclerosis lesions. J Neuroimmunol. 2007;190(1-2):72-79. doi: 10.1016/j.jneuroim.2007.07.024 [DOI] [PubMed] [Google Scholar]
- 26.Larsson D, Åkerfeldt T, Carlson K, Burman J. Intrathecal immunoglobulins and neurofilament light after autologous haematopoietic stem cell transplantation for multiple sclerosis. Mult Scler. 2020;26(11):1351-1359. doi: 10.1177/1352458519863983 [DOI] [PubMed] [Google Scholar]
- 27.Benkert P, Meier S, Schaedelin S, et al. Serum neurofilament light chain for individual prognostication of disease activity in people with multiple sclerosis: a retrospective modelling and validation study. Lancet Neurol. 2022;21(3):246-257. doi: 10.1016/S1474-4422(22)00009-6 [DOI] [PubMed] [Google Scholar]
- 28.Lundström W, Hermanrud C, Sjöstrand M, et al. Interferon beta treatment of multiple sclerosis increases serum interleukin-7. Mult Scler. 2014;20(13):1727-1736. doi: 10.1177/1352458514532700 [DOI] [PubMed] [Google Scholar]
- 29.Romme Christensen J, Komori M, von Essen MR, et al. CSF inflammatory biomarkers responsive to treatment in progressive multiple sclerosis capture residual inflammation associated with axonal damage. Mult Scler. 2019;25(7):937-946. doi: 10.1177/1352458518774880 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lundmark F, Duvefelt K, Iacobaeus E, et al. Variation in interleukin 7 receptor alpha chain (IL7R) influences risk of multiple sclerosis. Nat Genet. 2007;39(9):1108-1113. doi: 10.1038/ng2106 [DOI] [PubMed] [Google Scholar]
- 31.Gregory SG, Schmidt S, Seth P, et al. Interleukin 7 receptor alpha chain (IL7R) shows allelic and functional association with multiple sclerosis. Nat Genet. 2007;39(9):1083-1091. doi: 10.1038/ng2103 [DOI] [PubMed] [Google Scholar]
- 32.Crawley AM, Faucher S, Angel JB. Soluble IL-7R alpha (sCD127) inhibits IL-7 activity and is increased in HIV infection. J Immunol. 2010;184(9):4679-4687. doi: 10.4049/jimmunol.0903758 [DOI] [PubMed] [Google Scholar]
- 33.Lundström W, Highfill S, Walsh STR, et al. Soluble IL7Rα potentiates IL-7 bioactivity and promotes autoimmunity. Proc Natl Acad Sci U S A. 2013;110(19):E1761-E1770. doi: 10.1073/pnas.1222303110 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Bar-Or A, Li R. Cellular immunology of relapsing multiple sclerosis: interactions, checks, and balances. Lancet Neurol. 2021;20(6):470-483. doi: 10.1016/S1474-4422(21)00063-6 [DOI] [PubMed] [Google Scholar]
- 35.Mado H, Adamczyk-Sowa M, Sowa P. Role of microglial cells in the pathophysiology of MS: synergistic or antagonistic? Int J Mol Sci. 2023;24(3):1861. doi: 10.3390/ijms24031861 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Ponath G, Park C, Pitt D. The role of astrocytes in multiple sclerosis. Front Immunol. 2018;9:217. doi: 10.3389/fimmu.2018.00217 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Bjornevik K, Cortese M, Healy BC, et al. Longitudinal analysis reveals high prevalence of Epstein-Barr virus associated with multiple sclerosis. Science. 2022;375(6578):296-301. doi: 10.1126/science.abj8222 [DOI] [PubMed] [Google Scholar]
- 38.Lanz TV, Brewer RC, Ho PP, et al. Clonally expanded B cells in multiple sclerosis bind EBV EBNA1 and GlialCAM. Nature. 2022;603(7900):321-327. doi: 10.1038/s41586-022-04432-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Massey J, Artuz C, Dyer Z, et al. Diversification and expansion of the EBV-reactive cytotoxic T lymphocyte repertoire following autologous haematopoietic stem cell transplant for multiple sclerosis. Clin Immunol. 2023;254:109709. doi: 10.1016/j.clim.2023.109709 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
All data will be publicly available from the Immune Tolerance Network TrialShare website.