Elicit: B-cell Biomarkers and Ocrelizumab Response

B-cell Biomarkers and Ocrelizumab Response

Which biomarkers of B-cell depletion correlate with clinical response to ocrelizumab?

CD19+ B-cell counts, plasmablast levels, baseline regulatory cell populations (CD49d+ T-cells and B-1a cells), and CD8+ T-cell depletion correlate with clinical response to ocrelizumab, with the relative importance of each biomarker varying by MS subtype and baseline inflammatory status.

Abstract

Multiple biomarkers of B-cell depletion correlate with clinical response to ocrelizumab, though their predictive value varies by clinical context. Total CD19+ B-cell depletion showed consistent associations with outcomes in RRMS, with higher ocrelizumab exposure producing lower B-cell counts and reduced disability progression (HR 0.64, p=0.0135). Baseline CD19+ counts predicted repopulation kinetics, with counts ≥12-14% identifying patients at risk for fast repopulation and higher MRI activity (17.39% vs. 2.53%, p=0.008). Plasmablast levels emerged as particularly informative: responders at 24 months had 3-fold lower plasmablasts at 6 months compared to non-responders (7.69% vs. 22.66%, p=0.043), and plasmablast populations correlated positively with 12-month EDSS scores (R=0.564, p=0.004). However, in PPMS, baseline inflammatory status determined which biomarkers mattered—B-cell repopulation and serum neurofilament levels predicted outcomes in Gd+ patients, while serum IgA levels and T-cell remodeling were more relevant in Gd- patients. Elevated baseline regulatory populations (CD49d+ T-cells, B-1a cells) predicted achievement of NEDA-3 status, while CD8+ T-cell modulation correlated with disability outcomes (OR=1.01, p=0.02). Treatment-naive patients demonstrated stronger biomarker-outcome correlations than previously treated patients (HR=2.53 for radiological activity with prior treatment, p=0.039), suggesting prior therapies modify B-cell compartment dynamics and diminish predictive accuracy of standard depletion metrics.

Methods

We analyzed 10 sources from an initial pool of 200, using 8 screening criteria. Each paper was reviewed for 5 key aspects that mattered most to the research question.

Records from Elicit search
n = 200
Papers screened using: Ocrelizumab Treatment, B-cell Depletion Biomarkers, Clinical Response Outcomes, Quantitative Correlation Analysis, Follow-up Biomarker Measurements, Adequate Sample Size, Full-text Publication, Ocrelizumab-specific Data
n = 200 Papers screened out
n = 190 Papers included for extraction
n = 10

Paper search

We performed a semantic search across over 138 million academic papers from the Elicit search engine, which includes all of Semantic Scholar and OpenAlex.

We ran this query: "Which biomarkers of B-cell depletion correlate with clinical response to ocrelizumab?"
The search returned 200 total results from Elicit. We retrieved 200 papers most relevant to the query for screening.

Screening

We screened in sources based on their abstracts that met these criteria:

Data extraction

Results

Characteristics of Included Studies

All 10 studies evaluated biomarkers of B-cell depletion in patients treated with ocrelizumab, though they varied in MS subtype, sample size, and follow-up duration.

Study Full text retrieved? MS Subtype Sample Size Follow-up Duration Previous Treatment Baseline EDSS (median/mean)
S. Hauser et al., 2023 Yes RRMS and PPMS Not specified 96 weeks (RRMS), variable (PPMS) Randomized to ocrelizumab or comparator Not specified
Ece Akbayır et al., 2025 Yes RRMS 31 12 months Previously treated, resistant to first-line agents Not specified
M. Boziki et al., 2022 Yes Active PPMS 22 24 months Treatment-naive for immunosuppressants 4.91 (mean)
J. I. Fernández-Velasco et al., 2021 Yes PPMS 53 6 months Not specified 6 (median)
Alice Willison et al., 2025 Yes RRMS 34 OCR, 25 OFA, 20 treatment-naive 12 months Median 1 previous DMT (OCR and OFA groups); 25 treatment-naive 2.00 (median)
M. Cellerino et al., 2021 Yes RRMS (n=93), PPMS (n=43), SPMS (n=17) 153 1.9 years (median) 76.95% previously treated Lower baseline EDSS associated with better outcomes
Hany Mohammed Amin Aref et al., 2024 No RRMS 30 12 months Most had two previous DMTs; 13.3% treatment-naive 4.5 (median)
G. Abbadessa et al., 2021 Yes RRMS (n=127), SPMS (n=43), PPMS (n=48) 218 (155 at 12 months) At least 6 months; 155 with 12-month follow-up 76.95% previously treated; 23.04% treatment-naive Not specified
J. I. Fernández-Velasco et al., 2022 Yes PPMS 69 12 months Not specified 5.5 (median)
Nicola Capasso et al., 2022 Yes RRMS, SPMS, PPMS 78 36.5 months (mean) Previous DMTs collected at baseline Not specified

The included studies represented diverse MS populations, with sample sizes ranging from 22 to 218 patients. Most studies (8 of 10) had full-text availability. Follow-up durations ranged from 6 months to 36.5 months. The majority of patients across studies had prior treatment exposure, though several studies included treatment-naive cohorts. Baseline disability varied substantially, with median/mean EDSS scores ranging from 2.00 to 6.0, reflecting heterogeneity in disease severity.

B-Cell Depletion Biomarkers Measured

Studies employed diverse approaches to quantify B-cell depletion, though CD19+ B cells emerged as the most commonly measured biomarker. S. Hauser et al. measured CD19+ B cells in cells/μL and categorized patients by median B-cell levels (0, 1-5, and >5 cells/μL). Hany Mohammed Amin Aref et al. reported baseline CD19 B cell counts of 65.98 ± 79.04 cells/μl, which depleted to 3.59 ± 2.94 cells/μl at 1 month, 25.6 ± 43.98 cells/μl at 6 months, and 11.11 ± 19.53 cells/μl at 12 months. G. Abbadessa et al. measured CD19 as a percentage, reporting baseline values of 12.94% that decreased to 0.45% after 12 months.

Several studies examined specific B-cell subsets beyond total CD19+ counts. Ece Akbayır et al. quantified plasma cells (CD19+ CD38+ CD138+), which were significantly suppressed at 12 months, along with naive B-cells (CD19+ CD27-IgD+) and switched memory B-cells (CD19+ CD27+ IgD-). M. Boziki et al. used 9-color multiparametric flow cytometry to assess plasmablasts, transitional B-cells, marginal-zone-like B-cells, class-switched memory B-cells, non-switched CD27+ memory B-cells, and naive B-cells, with particular focus on plasmablast percentages of CD19+ cells. J. I. Fernández-Velasco et al. reported drastic reduction in naive and memory B cells, along with significant increases in the proportions of plasmablasts and transitional B-cells relative to total CD19+ B cells.

Alice Willison et al. documented significant reductions in CD20+ T and B cells, various B-cell subsets (naive B cells, regulatory B cells, marginal zone B cells, memory B cells, transitional B cells), and noted an expansion of CD5+CD19+CD20− B cells at months 1 and 12. Nicola Capasso et al. measured total lymphocyte counts and CD19+ and CD20+ subpopulations, reporting reductions that remained stable after the first infusion.

Two studies categorized patients based on repopulation kinetics. G. Abbadessa et al. identified fast repopulation (FR) and slow repopulation (SR) groups, with FR patients showing higher baseline CD19 cell counts (12% for 6-month cut-off, 14% for 12-month cut-off). S. Hauser et al. noted complete depletion with slow repopulation, with median time to B-cell repletion of 72 weeks.

Clinical Response Measures

Studies used heterogeneous outcome measures to assess clinical response to ocrelizumab. Disability progression was measured using confirmed disability progression (CDP) at 24 weeks, confirmed disability accumulation (CDA), confirmed disability worsening, and EDSS progression. EDSS scores were reported at various timepoints across studies.

Relapse-related outcomes included annualized relapse rate (ARR), time to first relapse, and number of attacks. ARR was significantly reduced from 20 to 0 after 12 months in the Aref et al. study and reached very low levels (0.13-0.18) across exposure quartiles in the Hauser et al. analysis.

MRI outcomes varied across studies. S. Hauser et al. assessed new/enlarging T2 lesions, T1 gadolinium-enhancing lesions, and brain volume changes. M. Boziki et al. measured new/enlarging T2 lesions and brain volume changes, with significant effects on gray matter cerebellar volume (F=4.342, p=0.23) and normalized gray matter cerebellar volume (F=4.279, p=0.024). Several studies evaluated the presence or absence of Gd-enhancing lesions.

Composite measures were employed in multiple studies. NEDA-3 (no evidence of disease activity) was defined as absence of relapses, confirmed disability worsening, and MRI activity. Ece Akbayır et al. achieved NEDA-3 in 19 of 31 patients (61.3%). M. Boziki et al. defined optimal response as patients free of relapses, CDA, and new/enlarging T2 lesions, achieved by 63.6% at 12 months and 59.1% at 24 months. Hany Mohammed Amin Aref et al. used the Multiple Sclerosis Functional Composite score, which showed significant improvement across three timepoints.

Biomarker-Clinical Response Correlations

Total B-Cell and CD19+ Depletion

S. Hauser et al. demonstrated a dose-response relationship between ocrelizumab exposure, B-cell depletion, and clinical outcomes. Higher ocrelizumab exposure led to lower blood B-cell counts, and higher median B-cell levels were associated with greater rates of disability progression. The hazard ratio for disability progression was 0.64 (95% CI: 0.45-0.92; p=0.0135) favoring higher ocrelizumab exposure. Patients with more complete B-cell depletion showed greater reduction in risk of disability progression.

Hany Mohammed Amin Aref et al. found a statistically significant correlation between CD19 B cell counts and EDSS scores, with sustained CD19 B cell depletion associated with marked clinical and radiological improvement. However, no significant correlation was found with the Multiple Sclerosis Functional Composite score or MRI disease activity.

G. Abbadessa et al. identified baseline CD19 cell count as a predictor of repopulation kinetics and clinical outcomes. Higher baseline CD19 counts were associated with fast repopulation rate (FR). FR patients had significantly higher active MRI scans at 12 months compared to slow repopulation (SR) patients (17.39% vs. 2.53%; p=0.008). ROC curve analysis established cut-offs of 12% CD19 cells at baseline for predicting fast repopulation at 6 months (AUC=0.62) and 14% for 12 months (AUC=0.71).

B-Cell Subset Correlations

Plasmablasts emerged as a significant predictor of treatment response in multiple studies. M. Boziki et al. found that responders at 24 months had significantly reduced peripheral blood plasmablasts (7.69 ± 4.4% of CD19+ cells) compared to non-responders (22.66 ± 7.19%) at the 6-month timepoint (p=0.043). This was the first study to link reduced plasmablast depletion to suboptimal response in active PPMS.

Ece Akbayır et al. demonstrated that month 12 EDSS scores correlated positively with plasmablast populations (R=0.564, p=0.004). Additionally, reduced post-treatment populations of switched memory B-cells were negatively correlated with treatment efficacy, suggesting that failure to suppress these populations indicates reduced treatment response.

Memory B-cell dynamics showed complex patterns. While Ece Akbayır et al. found reduced switched memory B-cells associated with better outcomes, J. I. Fernández-Velasco et al. reported significant reductions in both naive and memory B cells that correlated with decreased serum neurofilament light chain (sNfL) levels (p<0.0001 for B-cell reduction; p=0.008 for sNfL reduction).

J. I. Fernández-Velasco et al. (2022) observed that low repopulation rate of inflammatory B cells and reduced sNfL levels were associated with NEDA status in patients with gadolinium-enhancing lesions (Gd+). In Gd- patients, low B-cell numbers and sNfL levels were observed, but increases in serum IgA levels and IgA/IgM ratio were more significant for NEDA status.

Regulatory Cell Populations

Elevated baseline populations of regulatory cells predicted better outcomes. Ece Akbayır et al. found that baseline regulatory CD49d+ T-cells and B-1a cells were positively correlated with achieving NEDA-3 status. Alice Willison et al. reported that CD5 expression on B cells was negatively correlated with EDSS at baseline (r=-0.496, p=0.001) and positively correlated with time since last relapse (r=0.456, p=0.004).

Naive T-cell populations also showed protective associations. Ece Akbayır et al. found increased post-treatment populations of naive T-cells were positively correlated with lower EDSS scores at month 6 (R=-0.515 and R=-0.602, p=0.003) and higher rates of NEDA-3 status.

T-Cell and Exhaustion Markers

M. Cellerino et al. identified CD8+ cells as a potential biomarker, with the decrease in CD8+ cells being less pronounced in patients with inflammatory activity (p=0.022). CD8+ cell count was independently associated with early inflammatory activity (OR 1.005, 95% CI 1.001-1.009, p=0.019).

Nicola Capasso et al. found that higher probability of EDSS progression was associated with reduced modulation (i.e., higher levels) of CD8 lymphocytes (OR=1.01, 95% CI=1.00-1.05, p=0.02), suggesting that progressive reduction of CD8 cytotoxic T-lymphocytes correlates with clinical stability.

Alice Willison et al. documented extensive correlations with exhaustion markers. TIGIT expression on CD8+ T cells was negatively correlated with EDSS at baseline (r=-0.35, p=0.029) and positively correlated with time since disease manifestation (r=0.422, p=0.007). TIM-3 expression on CD4+ T cells showed negative correlation with baseline EDSS (r=-0.393, p=0.013). PD-1 expression on B cells was positively correlated with EDSS at month 1 (r=0.37, p=0.0341), while CTLA-4 expression on NKT cells was negatively correlated with EDSS at month 1 in OCR-treated patients (r=-0.516, p=0.0487).

Double-negative T cells (KLRG1-HLADR+ DN T cells) showed a positive correlation with the number of relapses prior to baseline (r=0.322, p=0.01) and negative correlation with EDSS at month 12 (r=-0.573, p=0.0161).

Patient-Level Predictors

Lower baseline EDSS was independently associated with reduced risk of disability worsening (HR=1.45, 95% CI 1.05-2.00, p=0.024). Previous treatment exposure was independently associated with increased probability of radiological activity (HR=2.53, 95% CI 1.05-6.10, p=0.039), suggesting treatment-naive patients may respond better to ocrelizumab.

Synthesis

The studies reveal complex, context-dependent relationships between B-cell depletion biomarkers and clinical response to ocrelizumab. Rather than a single biomarker universally predicting outcomes, different markers appear relevant in different clinical contexts.

MS Subtype as a Context Modifier
The relevance of specific biomarkers varies by MS subtype. In RRMS populations, total B-cell depletion metrics (CD19+ counts, repopulation kinetics) showed strong associations with clinical outcomes. The Hauser et al. post-hoc analysis of the OPERA trials demonstrated clear dose-response relationships in RRMS, with hazard ratios for disability progression ranging from 0.70 in the lowest exposure quartile to 0.34 in the highest quartile versus interferon β-1a.

In PPMS populations, the pattern differed substantially. While Hauser et al. observed similar trends (hazard ratios ranging from 0.88 to 0.55 across quartiles), the PPMS-specific studies revealed that baseline inflammatory status fundamentally altered which biomarkers mattered. Fernández-Velasco et al. (2022) demonstrated that in Gd+ PPMS patients, B-cell repopulation rate and sNfL levels predicted NEDA status, whereas in Gd- patients, the same B-cell metrics showed no predictive value, with serum IgA levels and T-cell/innate immune remodeling instead correlating with outcomes. This suggests that in non-inflammatory PPMS, B-cell depletion enables immune modulation through indirect mechanisms rather than directly suppressing inflammatory B cells.

The Plasmablast Paradox
Multiple studies identified plasmablasts as paradoxically important despite representing a small fraction of circulating B cells. Boziki et al. found that responders had 3-fold lower plasmablasts (7.69% vs. 22.66% of CD19+ cells, p=0.043) at 6 months, while Akbayır et al. showed plasmablast populations correlated positively with 12-month EDSS scores (R=0.564, p=0.004). Fernández-Velasco et al. (2021) noted that plasmablast proportions increased significantly after ocrelizumab, seemingly contradicting their role as a negative prognostic marker.

This apparent contradiction resolves when considering that ocrelizumab depletes CD20+ B cells but spares CD20-negative plasmablasts and plasma cells. The increase in plasmablast proportion (relative to depleted total B cells) does not indicate absolute expansion. Studies measuring absolute plasmablast depletion efficiency—rather than proportional changes—consistently found that incomplete plasmablast suppression predicts worse outcomes. This suggests residual plasmablasts may sustain pathogenic antibody production or represent reservoirs for rapid B-cell reconstitution.

Baseline Biomarkers Versus Dynamic Markers
Studies diverged on whether baseline or post-treatment biomarker levels better predicted outcomes, likely reflecting different mechanistic questions. Baseline CD19+ counts predicted repopulation kinetics: Abbadessa et al. established that baseline counts ≥12-14% identified patients who would experience fast repopulation and higher MRI activity. This suggests baseline B-cell burden determines the “depth” of achievable depletion with standard dosing.

Conversely, post-treatment dynamics—particularly at 6 months—predicted clinical trajectories in multiple studies. Boziki et al. found 6-month plasmablast levels distinguished responders from non-responders at 24 months, while Akbayır et al. showed that 6-month switched memory B-cell levels and naive T-cell populations correlated with NEDA-3 status. This 6-month timepoint may represent a critical window when the balance between depletion efficacy and early reconstitution becomes apparent, before full disease reactivation occurs.

The T-Cell Connection: Indirect Effects of B-Cell Depletion
Several studies identified T-cell populations as biomarkers despite ocrelizumab targeting CD20. Cellerino et al. found CD8+ cell counts independently associated with inflammatory activity (OR 1.005, p=0.019), while Capasso et al. showed reduced CD8+ modulation predicted EDSS progression (OR=1.01, p=0.02). These findings reflect that ocrelizumab depletes CD20+ T cells alongside B cells, with Cellerino et al. explicitly linking CD8+ depletion to reduced CD20+ T cells.

Willison et al. provided the most comprehensive T-cell profiling, demonstrating that exhaustion markers (TIGIT, TIM-3, PD-1, CTLA-4) on T, B, NK, and NKT cells increased with ocrelizumab and correlated with clinical parameters. TIGIT expression showed particularly consistent protective correlations across cell types (negative correlation with EDSS, positive correlation with time since manifestation). These findings suggest ocrelizumab induces a broader immune remodeling toward an exhausted/regulatory phenotype, potentially explaining efficacy beyond simple B-cell depletion.

Regulatory T-cell expansion provided mechanistic insight: Akbayır et al. found elevated baseline regulatory CD49d+ T-cells and B-1a cells predicted NEDA-3 status, while Willison et al. documented CD5+CD19+CD20− B cell expansion. This suggests successful treatment may depend on concurrent expansion of regulatory populations that suppress residual effector cells.

Treatment History as a Confounding Factor
Previous treatment exposure consistently modified biomarker-response relationships. Cellerino et al. found prior DMT exposure independently increased radiological activity risk (HR=2.53, p=0.039), while treatment-naive patients had higher NEDA-3 probabilities. Abbadessa et al.’s cohort was 76.95% previously treated, potentially explaining why baseline CD19 counts varied so widely (sufficient to create meaningful quartiles for analysis).

This likely reflects that prior treatments alter the B-cell compartment’s composition and repopulation capacity. Patients who failed previous therapies may have more aggressive disease biology, enrichment for treatment-resistant B-cell clones, or altered B-cell homeostatic mechanisms. The stronger biomarker-outcome correlations in treatment-naive cohorts suggest that without prior perturbations, B-cell depletion metrics more directly reflect disease activity.

Dose-Response and Threshold Effects
The Hauser et al. exposure-response analysis revealed a critical asymmetry: while ARR and MRI lesions reached near-zero levels across all exposure quartiles (suggesting a ceiling effect), disability progression showed continuous exposure-dependent improvement. This suggests different disease processes have different B-cell dependency thresholds. Inflammatory relapses and focal lesions may require only moderate B-cell depletion (achievable at lower exposures), while disability progression—potentially driven by compartmentalized inflammation or chronic B-cell activity in CNS tissues—benefits from maximal depletion.

Abbadessa et al.’s ROC analysis established practical cut-offs (12% and 14% baseline CD19), but the modest AUCs (0.62 and 0.71) indicate substantial overlap between responders and non-responders. This suggests baseline counts provide risk stratification rather than deterministic prediction, and that other factors (genetic, compartmentalized immunity) substantially influence outcomes.

Reconciling Heterogeneity: Toward an Integrated Model
The apparent heterogeneity in biomarker findings reflects that B-cell depletion operates through multiple mechanisms that vary by clinical context:

  1. In inflammatory RRMS/active PPMS: Direct suppression of pathogenic B cells (measured by CD19+ counts, plasmablast depletion) correlates with reduced relapses and MRI activity.
  2. In non-inflammatory PPMS: B-cell depletion enables immune remodeling (measured by regulatory cell expansion, IgA changes, T-cell modulation) that indirectly affects progression.
  3. Across all subtypes: Completeness of depletion (measured by repopulation kinetics, residual plasmablasts) predicts durability of response, while baseline regulatory populations (CD49d+ T-cells, B-1a cells) predict capacity for beneficial immune remodeling.
  4. In previously treated patients: Altered B-cell compartments and potentially resistant clones diminish predictive value of standard depletion metrics.

Rather than seeking a single “best” biomarker, these findings suggest clinical response depends on achieving adequate depletion and favorable immune reconstitution. Baseline metrics (CD19 counts, regulatory cells) predict depletion depth and reconstitution quality, while early post-treatment measures (6-month plasmablasts, memory B-cells, T-cell profiles) predict the realized treatment response before clinical outcomes fully manifest.

References

Ece Akbayır, Tugce Kizilay, Ruziye Erol, Duygu OZKAN‐YASARGUN, E. Tuzun, and 2 more (2025). B‐Cell and T‐Cell Populations in Peripheral Blood Linked to Ocrelizumab Treatment Efficacy in Multiple Sclerosis. In Vivo

M. Boziki, C. Bakirtzis, Styliani-Aggeliki Sintila, E. Kesidou, Evdoxia Gounari, and 11 more (2022). Ocrelizumab in Patients with Active Primary Progressive Multiple Sclerosis: Clinical Outcomes and Immune Markers of Treatment Response. Cells

J. I. Fernández-Velasco, J. Kuhle, E. Monreal, V. Meca-Lallana, J. Meca-Lallana, and 20 more (2021). Effect of Ocrelizumab in Blood Leukocytes of Patients With Primary Progressive MS. Neurology: Neuroimmunology & Neuroinflammation

Alice Willison, Ramona Hagler, M. Weise, S. Elben, Niklas Huntemann, and 16 more (2025). Effects of Anti-CD20 Antibody Therapy on Immune Cell Dynamics in Relapsing-Remitting Multiple Sclerosis. Cells

M. Cellerino, G. Boffa, C. Lapucci, F. Tazza, E. Sbragia, and 11 more (2021). Predictors of Ocrelizumab Effectiveness in Patients with Multiple Sclerosis. Neurotherapeutics

Hany Mohammed Amin Aref, Azza Abd El-Nasser Abdelaziz, Abeer Al-Sayed Ali Shehab, D. Zamzam, Alaa Mohamed Abou Steit, and 1 more (2024). Follow Up of Total B Cells (CD 19) Count after B Cell Depleting Therapy in Relapse Remitting Multiple Sclerosis Disorder in Egyptian Patients. The Quarterly journal of medicine

G. Abbadessa, G. Miele, P. Cavalla, P. Valentino, G. Marfia, and 10 more (2021). CD19 Cell Count at Baseline Predicts B Cell Repopulation at 6 and 12 Months in Multiple Sclerosis Patients Treated with Ocrelizumab. International Journal of Environmental Research and Public Health

J. I. Fernández-Velasco, E. Monreal, J. Kuhle, V. Meca-Lallana, J. Meca-Lallana, and 26 more (2022). Baseline Inflammatory Status Reveals Dichotomic Immune Mechanisms Involved In Primary-Progressive Multiple Sclerosis Pathology. Frontiers in Immunology

Nicola Capasso, R. Palladino, Vincenza Cerbone, A. Spiezia, B. Covelli, and 8 more (2022). Ocrelizumab effect on humoral and cellular immunity in multiple sclerosis and its clinical correlates: a 3-year observational study. Journal of Neurology

S. Hauser, A. Bar-Or, Martin S. Weber, H. Kletzl, A. Günther, and 8 more (2023). Association of Higher Ocrelizumab Exposure With Reduced Disability Progression in Multiple Sclerosis. Neurology: Neuroimmunology & Neuroinflammation