Response to Immune Interventions
The immune models permit simulation of therapeutic interventions.
1 B Cell Depletion (Rituximab)
Rituximab depletes CD20⁺ B cells, modeled as an exponential reduction in \(B_n\) and \(B_a\) with a rate proportional to drug concentration. The model reproduces the key clinical observation from the Fluge et al. trials (Fluge et al. 2011) (Fluge et al. 2015): clinical improvement is delayed by 3–6 months because pre-existing plasma cells (\(P\), which are CD20⁻ and not depleted by rituximab) continue producing autoantibodies until they reach the end of their natural lifespan. The model predicts that the subsequent RituxME phase III trial failure (Fluge et al. 2019) could result from insufficient depletion duration relative to the plasma cell half-life, patient heterogeneity in autoantibody-dependence, or—most parsimoniously—that B cell-mediated autoimmunity is not a primary driver in the unselected ME/CFS population.
2 Low-Dose Naltrexone
Low-dose naltrexone (LDN) is hypothesized to modulate immune function through transient opioid receptor blockade, leading to upregulation of endogenous opioid production and downstream immunomodulation (Polo et al. 2019). In the model, LDN is represented as a shift in the cytokine network parameters: reduced TNF-\(\alpha\) production (\(\sigma_{\text{TNF}}^{\text{mono}}\)) and increased IL-10 production. The predicted effect is a gradual shift from the high-inflammation attractor toward a lower-inflammation state, with a timescale of weeks consistent with clinical reports.
The B cell model treats autoantibodies as a single species. In reality, ME/CFS patients harbor autoantibodies against multiple targets—\(\beta_2\)-adrenergic receptors, muscarinic receptors, and others (Loebel et al. 2016) (Freitag et al. 2021). Do different autoantibody specificities contribute independently to distinct symptom domains? If so, the clinical response to B cell depletion would depend on which autoantibody populations are most reduced, requiring antigen-specific modeling.
3 Plasma Cell Depletion (Daratumumab)
Daratumumab, an anti-CD38 monoclonal antibody, targets plasma cells directly—unlike rituximab, which depletes only CD20⁺ B cells and spares CD20⁻ plasma cells. A pilot study reported a 60% response rate in ME/CFS patients with elevated GPCR autoantibodies (Fluge et al. 2025). The model represents daratumumab as simultaneous depletion of both \(B_a\) and \(P\) populations:
\[ \frac{d P}{d t}|_{\text{dara}} = k_{\text{plasma}} B_a - d_P P - k_{\text{dara}} \cdot C_{\text{dara}}(t) \cdot P \tag{1}\]
where \(k_{\text{dara}} \cdot C_{\text{dara}}(t)\) is the drug-mediated killing rate. The model yields a quantitative prediction impossible without the mathematics: clinical response to daratumumab should appear faster than rituximab by a factor of approximately \(d_P^{-1} \\/ \tau_{\text{dara}}\), where \(d_P^{-1}\) is the natural plasma cell half-life (months to years) and \(\tau_{\text{dara}}\) is the daratumumab-mediated depletion time constant (days to weeks). For typical parameter values, the model predicts a response latency of 2–6 weeks for daratumumab versus 3–6 months for rituximab. Furthermore, the model predicts that daratumumab will show a clearer dose–response relationship because it acts directly on the autoantibody-producing cells, whereas rituximab’s effect is indirect and delayed through the plasma cell half-life bottleneck.