Leveraging multi-omics data to empower quantitative systems pharmacology in immuno-oncology.

Understanding the intricate interactions of cancer cells with the tumor microenvironment (TME) is a pre-requisite for the optimization of immunotherapy. Mechanistic models such as quantitative systems pharmacology (QSP) provide insights into the TME dynamics and predict the efficacy of immunotherapy in virtual patient populations/digital twins but require vast amounts of multimodal data for parameterization. Large-scale datasets characterizing the TME are available due to recent advances in bioinformatics for multi-omics data. Here, we discuss the perspectives of leveraging omics-derived bioinformatics estimates to inform QSP models and circumvent the challenges of model calibration and validation in immuno-oncology.

Rate of Immune Complex Cycling in Follicular Dendritic Cells Determines the Extent of Protecting Antigen Integrity and Availability to Germinal Center B Cells.

Follicular dendritic cells (FDCs) retain immune complexes (ICs) for prolonged time periods and are important for germinal center (GC) reactions. ICs undergo periodic cycling in FDCs, a mechanism supporting an extended half-life of Ag. Based on experimental data, we estimated that the average residence time of PE-ICs on FDC surface and interior were 21 and 36 min, respectively. GC simulations show that Ag cycling might impact GC dynamics because of redistribution of Ag on the FDC surface and by protecting Ag from degradation. Ag protection and influence on GC dynamics varied with Ag cycling time and total Ag concentration. Simulations predict that blocking Ag cycling terminates the GC reaction and decreases plasma cell production. Considering that cycling of Ag could be a target for the modulation of GC reactions, our findings highlight the importance of understanding the mechanism and regulation of IC cycling in FDCs.

A mathematical modelling tool for unravelling the antibody-mediated effects on CTLA-4 interactions.

BACKGROUND: Monoclonal antibodies blocking the Cytotoxic T-lymphocyte antigen 4 (CTLA-4) receptor have revolutionized the field of anti-cancer therapy for the last few years. The human T-cell-based immune responses are modulated by two contradicting signals. CTLA-4 provides a T cell inhibitory signal through its interaction with B7 ligands (B7-1 and B7-2), while CD28 provides a stimulatory signal when interacting with the same ligands. A previous theoretical model has focused on understanding the processes of costimulatory and inhibitory complex formations at the synapse. Nevertheless, the effects of monoclonal antibody (mAb)-mediation on these complexes are relatively unexplored. In this work, we expand on the previous model to develop a new mathematical framework for studying the effects of anti-CTLA-4 mAbs on the co-stimulatory (CD28/B7 ligands) and the co-inhibitory (CTLA-4/B7 ligands) complex formation at the immunological synapse. In particular, we focus on two promising anti-CTLA-4 mAbs, tremelimumab (from AstraZeneca) and ipilimumab (from Bristol-Myers Squibb), which are currently in clinical trials and the market, respectively, for targeting multiple tumors. METHODS: The mathematical model in this work has been constructed based on ordinary differential equations and available experimental binding kinetics data for the anti-CTLA-4 antibodies from literature. RESULTS: The numerical simulations from the current model are in agreement with a number of experimental data. Especially, the dose-curves for blocking the B7 ligand binding to CTLA-4 by ipilimumab are comparable with the results from a previous competitive binding assay by flow cytometry and ELISA. Our simulations predict the dose response and the relative efficacies of the two mAbs in blocking the inhibitory CTLA-4/B7 complexes. CONCLUSIONS: The results show that different factors, such as multivalent interactions, mobility of molecules and competition effects, could impact the effects of antibody-mediation. The results, in particular, describe that the competitive effects could impact the dose-dependent inhibition by the mAbs very significantly. We present this model as a useful tool that can easily be translated to study the effects of any anti-CTLA-4 antibodies on immunological synaptic complex formation, provided reliable biophysical data for mAbs are available.

From virtual patients to digital twins in immuno-oncology: lessons learned from mechanistic quantitative systems pharmacology modeling.

Virtual patients and digital patients/twins are two similar concepts gaining increasing attention in health care with goals to accelerate drug development and improve patients' survival, but with their own limitations. Although methods have been proposed to generate virtual patient populations using mechanistic models, there are limited number of applications in immuno-oncology research. Furthermore, due to the stricter requirements of digital twins, they are often generated in a study-specific manner with models customized to particular clinical settings (e.g., treatment, cancer, and data types). Here, we discuss the challenges for virtual patient generation in immuno-oncology with our most recent experiences, initiatives to develop digital twins, and how research on these two concepts can inform each other.

Quantitative systems pharmacology modeling of macrophage-targeted therapy combined with PD-L1 inhibition in advanced NSCLC.

Immune checkpoint inhibitors remained the standard-of-care treatment for advanced non-small cell lung cancer (NSCLC) for the past decade. In unselected patients, anti-PD-(L)1 monotherapy achieved an overall response rate of about 20%. In this analysis, we developed a pharmacokinetic and pharmacodynamic module for our previously calibrated quantitative systems pharmacology model (QSP) to simulate the effectiveness of macrophage-targeted therapies in combination with PD-L1 inhibition in advanced NSCLC. By conducting in silico clinical trials, the model confirmed that anti-CD47 treatment is not an optimal option of second- and later-line treatment for advanced NSCLC resistant to PD-(L)1 blockade. Furthermore, the model predicted that inhibition of macrophage recruitment, such as using CCR2 inhibitors, can potentially improve tumor size reduction when combined with anti-PD-(L)1 therapy, especially in patients who are likely to respond to anti-PD-(L)1 monotherapy and those with a high level of tumor-associated macrophages. Here, we demonstrate the application of the QSP platform on predicting the effectiveness of novel drug combinations involving immune checkpoint inhibitors based on preclinical or early-stage clinical trial data.

Virtual patient analysis identifies strategies to improve the performance of predictive biomarkers for PD-1 blockade.

Patients with metastatic triple-negative breast cancer (TNBC) show variable responses to PD-1 inhibition. Efficient patient selection by predictive biomarkers would be desirable, but is hindered by the limited performance of existing biomarkers. Here, we leveraged in-silico patient cohorts generated using a quantitative systems pharmacology model of metastatic TNBC, informed by transcriptomic and clinical data, to explore potential ways to improve patient selection. We tested 90 biomarker candidates, including various cellular and molecular species, by a cutoff-based biomarker testing algorithm combined with machine learning-based feature selection. Combinations of pre-treatment biomarkers improved the specificity compared to single biomarkers at the cost of reduced sensitivity. On the other hand, early on-treatment biomarkers, such as the relative change in tumor diameter from baseline measured at two weeks after treatment initiation, achieved remarkably higher sensitivity and specificity. Further, blood-based biomarkers had a comparable ability to tumor- or lymph node-based biomarkers in identifying a subset of responders, potentially suggesting a less invasive way for patient selection.

Generating immunogenomic data-guided virtual patients using a QSP model to predict response of advanced NSCLC to PD-L1 inhibition.

Generating realistic virtual patients from a limited amount of patient data is one of the major challenges for quantitative systems pharmacology modeling in immuno-oncology. Quantitative systems pharmacology (QSP) is a mathematical modeling methodology that integrates mechanistic knowledge of biological systems to investigate dynamics in a whole system during disease progression and drug treatment. In the present analysis, we parameterized our previously published QSP model of the cancer-immunity cycle to non-small cell lung cancer (NSCLC) and generated a virtual patient cohort to predict clinical response to PD-L1 inhibition in NSCLC. The virtual patient generation was guided by immunogenomic data from iAtlas portal and population pharmacokinetic data of durvalumab, a PD-L1 inhibitor. With virtual patients generated following the immunogenomic data distribution, our model predicted a response rate of 18.6% (95% bootstrap confidence interval: 13.3-24.2%) and identified CD8/Treg ratio as a potential predictive biomarker in addition to PD-L1 expression and tumor mutational burden. We demonstrated that omics data served as a reliable resource for virtual patient generation techniques in immuno-oncology using QSP models.

Generating immunogenomic data-guided virtual patients using a QSP model to predict response of advanced NSCLC to PD-L1 inhibition.

Generating realistic virtual patients from a limited amount of patient data is one of the major challenges for quantitative systems pharmacology modeling in immuno-oncology. Quantitative systems pharmacology (QSP) is a mathematical modeling methodology that integrates mechanistic knowledge of biological systems to investigate dynamics in a whole system during disease progression and drug treatment. In the present analysis, we parameterized our previously published QSP model of the cancer-immunity cycle to non-small cell lung cancer (NSCLC) and generated a virtual patient cohort to predict clinical response to PD-L1 inhibition in NSCLC. The virtual patient generation was guided by immunogenomic data from iAtlas portal and population pharmacokinetic data of durvalumab, a PD-L1 inhibitor. With virtual patients generated following the immunogenomic data distribution, our model predicted a response rate of 18.6% (95% bootstrap confidence interval: 13.3-24.2%) and identified CD8/Treg ratio as a potential predictive biomarker in addition to PD-L1 expression and tumor mutational burden. We demonstrated that omics data served as a reliable resource for virtual patient generation techniques in immuno-oncology using QSP models.

Germinal centers are permissive to subdominant antibody responses.

Introduction: A protective humoral response to pathogens requires the development of high affinity antibodies in germinal centers (GC). The combination of antigens available during immunization has a strong impact on the strength and breadth of the antibody response. Antigens can display various levels of immunogenicity, and a hierarchy of immunodominance arises when the GC response to an antigen dampens the response to other antigens. Immunodominance is a challenge for the development of vaccines to mutating viruses, and for the development of broadly neutralizing antibodies. The extent by which antigens with different levels of immunogenicity compete for the induction of high affinity antibodies and therefore contribute to immunodominance is not known. Methods: Here, we perform in silico simulations of the GC response, using a structural representation of antigens with complex surface amino acid composition and topology. We generate antigens with complex domains of different levels of immunogenicity and perform simulations with combinations of these domains. Results: We found that GC dynamics were driven by the most immunogenic domain and immunodominance arose as affinity maturation to less immunogenic domain was inhibited. However, this inhibition was moderate since the less immunogenic domain exhibited a weak GC response in the absence of the most immunogenic domain. Less immunogenic domains reduced the dominance of GC responses to more immunogenic domains, albeit at a later time point. Discussion: The simulations suggest that increased vaccine valency may decrease immunodominance of the GC response to strongly immunogenic domains and therefore, act as a potential strategy for the natural induction of broadly neutralizing antibodies in GC reactions.

A transcriptome-informed QSP model of metastatic triple-negative breast cancer identifies predictive biomarkers for PD-1 inhibition.

Triple-negative breast cancer (TNBC), a highly metastatic breast cancer subtype, has limited treatment options. While a small number of patients attain clinical benefit with single-agent checkpoint inhibitors, identifying these patients before the therapy remains challenging. Here, we developed a transcriptome-informed quantitative systems pharmacology model of metastatic TNBC by integrating heterogenous metastatic tumors. In silico clinical trial with an anti-PD-1 drug, pembrolizumab, predicted that several features, such as the density of antigen-presenting cells, the fraction of cytotoxic T cells in lymph nodes, and the richness of cancer clones in tumors, could serve individually as biomarkers but had a higher predictive power as combinations of two biomarkers. We showed that PD-1 inhibition neither consistently enhanced all antitumorigenic factors nor suppressed all protumorigenic factors but ultimately reduced the tumor carrying capacity. Collectively, our predictions suggest several candidate biomarkers that might effectively predict the response to pembrolizumab monotherapy and potential therapeutic targets to develop treatment strategies for metastatic TNBC.

BTG1 mutation yields supercompetitive B cells primed for malignant transformation.

Multicellular life requires altruistic cooperation between cells. The adaptive immune system is a notable exception, wherein germinal center B cells compete vigorously for limiting positive selection signals. Studying primary human lymphomas and developing new mouse models, we found that mutations affecting BTG1 disrupt a critical immune gatekeeper mechanism that strictly limits B cell fitness during antibody affinity maturation. This mechanism converted germinal center B cells into supercompetitors that rapidly outstrip their normal counterparts. This effect was conferred by a small shift in MYC protein induction kinetics but resulted in aggressive invasive lymphomas, which in humans are linked to dire clinical outcomes. Our findings reveal a delicate evolutionary trade-off between natural selection of B cells to provide immunity and potentially dangerous features that recall the more competitive nature of unicellular organisms.

Investigating the Mechanism of Germinal Center Shutdown.

Germinal centers (GCs) are transient structures where affinity maturation of B cells gives rise to high affinity plasma and memory cells. The mechanism of GC shutdown is unclear, despite being an important phenomenon maintaining immune homeostasis. In this study, we used a mathematical model to identify mechanisms that can independently promote contraction of GCs leading to shutdown. We show that GC shutdown can be promoted by antigen consumption by B cells, antigen masking by soluble antibodies, alterations in follicular dendritic cell (FDC) network area, modulation of immune complex cycling rate constants, alterations in T follicular helper signaling, increased terminal differentiation and reduced B cell division capacity. Proposed mechanisms promoted GC contraction by ultimately decreasing the number of B cell divisions and recycling cells. Based on the in-silico predictions, we suggest a combination of experiments that can be potentially employed by future studies to unravel the mechanistic basis of GC shutdown such as measurements of the density of pMHC presentation of B cells, FDC network size per B cell, fraction of cells expressing differentiation markers. We also show that the identified mechanisms differentially affect the efficiency of GC reaction estimated based on the quantity and quality of resulting antibodies.

Antibody Mediated Intercommunication of Germinal Centers.

Antibody diversification and selection of B cells occur in dynamic structures called germinal centers (GCs). Passively administered soluble antibodies regulate the GC response by masking the antigen displayed on follicular dendritic cells (FDCs). This suggests that GCs might intercommunicate via naturally produced soluble antibodies, but the role of such GC-GC interactions is unknown. In this study, we performed in silico simulations of interacting GCs and predicted that intense interactions by soluble antibodies limit the magnitude and lifetime of GC responses. With asynchronous GC onset, we observed a higher inhibition of late formed GCs compared to early ones. We also predicted that GC-GC interactions can lead to a bias in the epitope recognition even in the presence of equally dominant epitopes due to differences in founder cell composition or initiation timing of GCs. We show that there exists an optimal range for GC-GC interaction strength that facilitates the affinity maturation towards an incoming antigenic variant during an ongoing GC reaction. These findings suggest that GC-GC interactions might be a contributing factor to the unexplained variability seen among individual GCs and a critical factor in the modulation of GC response to antigenic variants during viral infections.

In Silico Analysis of the Longevity and Timeline of Individual Germinal Center Reactions in a Primary Immune Response.

Germinal centers (GCs) are transient structures in the secondary lymphoid organs, where B cells undergo affinity maturation to produce high affinity memory and plasma cells. The lifetime of GC responses is a critical factor limiting the extent of affinity maturation and efficiency of antibody responses. While the average lifetime of overall GC reactions in a lymphoid organ is determined experimentally, the lifetime of individual GCs has not been monitored due to technical difficulties in longitudinal analysis. In silico analysis of the contraction phase of GC responses towards primary immunization with sheep red blood cells suggested that if individual GCs had similar lifetimes, the data would be consistent only when new GCs were formed until a very late phase after immunization. Alternatively, there could be a large variation in the lifetime of individual GCs suggesting that both long and short-lived GCs might exist in the same lymphoid organ. Simulations predicted that such differences in the lifetime of GCs could arise due to variations in antigen availability and founder cell composition. These findings identify the potential factors limiting GC lifetime and contribute to an understanding of overall GC responses from the perspective of individual GCs in a primary immune response.

Ymir: A 3D structural affinity model for multi-epitope vaccine simulations.

Vaccine development is challenged by the hierarchy of immunodominance between target antigen epitopes and the emergence of antigenic variants by pathogen mutation. The strength and breadth of antibody responses relies on selection and mutation in the germinal center and on the structural similarity between antigens. Computational methods for assessing the breadth of germinal center responses to multivalent antigens are critical to speed up vaccine development. Yet, such methods have poorly reflected the 3D antigen structure and antibody breadth. Here, we present Ymir, a new 3D-lattice-based framework that calculates in silico antibody-antigen affinities. Key physiological properties naturally emerge from Ymir such as affinity jumps, cross-reactivity, and differential epitope accessibility. We validated Ymir by replicating known features of germinal center dynamics. We show that combining antigens with mutated but structurally related epitopes enhances vaccine breadth. Ymir opens a new avenue for understanding vaccine potency based on the structural relationship between vaccine antigens.

Germinal Centre Shutdown.

Germinal Centres (GCs) are transient structures in secondary lymphoid organs, where affinity maturation of B cells takes place following an infection. While GCs are responsible for protective antibody responses, dysregulated GC reactions are associated with autoimmune disease and B cell lymphoma. Typically, 'normal' GCs persist for a limited period of time and eventually undergo shutdown. In this review, we focus on an important but unanswered question - what causes the natural termination of the GC reaction? In murine experiments, lack of antigen, absence or constitutive T cell help leads to premature termination of the GC reaction. Consequently, our present understanding is limited to the idea that GCs are terminated due to a decrease in antigen access or changes in the nature of T cell help. However, there is no direct evidence on which biological signals are primarily responsible for natural termination of GCs and a mechanistic understanding is clearly lacking. We discuss the present understanding of the GC shutdown, from factors impacting GC dynamics to changes in cellular interactions/dynamics during the GC lifetime. We also address potential missing links and remaining questions in GC biology, to facilitate further studies to promote a better understanding of GC shutdown in infection and immune dysregulation.

Synchronous Germinal Center Onset Impacts the Efficiency of Antibody Responses.

The germinal center reaction is an important target for modulating antibody responses. Antibody production from germinal centers is regulated by a negative feedback mechanism termed antibody feedback. By imposing antibody feedback, germinal centers can interact and regulate the output of other germinal centers. Using an agent-based model of the germinal center reaction, we studied the impact of antibody feedback on kinetics and efficiency of a germinal center. Our simulations predict that high feedback of antibodies from germinal centers reduces the production of plasma cells and subsequently the efficiency of the germinal center reaction by promoting earlier termination. Affinity maturation is only weakly improved by increased antibody feedback and ultimately interrupted because of premature termination of the reaction. The model predicts that the asynchronous onset and changes in number of germinal centers could alter the efficiency of antibody response due to changes in feedback by soluble antibodies. Consequently, late initialized germinal centers have a compromised output due to higher antibody feedback from the germinal centers formed earlier. The results demonstrate potential effects of germinal center intercommunication and highlight the importance of understanding germinal center interactions for optimizing the antibody response, in particular, in the elderly and in the context of vaccination.

Mathematical modeling identifies Lck as a potential mediator for PD-1 induced inhibition of early TCR signaling.

Programmed cell death-1 (PD-1) is an inhibitory immune checkpoint receptor that negatively regulates the functioning of T cell. Although the direct targets of PD-1 were not identified, its inhibitory action on the TCR signaling pathway was known much earlier. Recent experiments suggest that the PD-1 inhibits the TCR and CD28 signaling pathways at a very early stage ─ at the level of phosphorylation of the cytoplasmic domain of TCR and CD28 receptors. Here, we develop a mathematical model to investigate the influence of inhibitory effect of PD-1 on the activation of early TCR and CD28 signaling molecules. Proposed model recaptures several quantitative experimental observations of PD-1 mediated inhibition. Model simulations show that PD-1 imposes a net inhibitory effect on the Lck kinase. Further, the inhibitory effect of PD-1 on the activation of TCR signaling molecules such as Zap70 and SLP76 is significantly enhanced by the PD-1 mediated inhibition of Lck. These results suggest a critical role for Lck as a mediator for PD-1 induced inhibition of TCR signaling network. Multi parametric sensitivity analysis explores the effect of parameter uncertainty on model simulations.