Periodic Reporting for period 2 - IMPACT (High-Dimensional single cell mapping of inflammatory disease signatures in monozygotic twins)
Reporting period: 2022-07-01 to 2023-12-31
Multiple Sclerosis (MS) is a chronic inflammatory disease, where immune cell invasion into the central nervous system causes immunopathology and neurological deficit. Although disease-modifying therapies dramatically reduce disease activity, they hold the potential for severe adverse effects while long-term disability prospects remain poor. Moreover, there is to date no biomarker for monitoring the disease activity and to guide therapy decisions. I propose that the key to identifying such biomarkers is to combine single-cell mapping of leukocytes across well-curated patient cohorts with unbiased machine-learning based data interrogation. Using such an approach, we have already delineated a disease signature in a helper T cell population specific for MS. However, the immune compartment of cross-sectional cohorts is influenced by the individual genetic make-up, which masks disease-specific signals and hinders a more precise characterisation of involved immune cell populations.
To eliminate genetic influences, we pursue the following specific goals:
- Aim 1 to interrogate the immune compartment of a unique cohort of monozygotic twin pairs -discordant for MS- and deeply analyse peripheral blood lymphocytes by single-cell mass cytometry, combined TcR and single cell sequencing, and epigenetic profiling.
- Aim 2 to develop representation-learning methods to account for the paired genetics of twins or longitudinal samples and to include clinical covariates into the high-dimensional data set.
- Aim 3 to use well-defined patient samples of MS-like disorders (MS-Mimics) and longitudinal samples of patients undergoing disease-modifying therapy (e.g. B cell depletion, autologous stem cell transplant) using single-cell mass cytometry.
Ultimately, the goal is to reduce the dimensionality of disease signature(s) towards a clinically translatable low-dimensional biomarker that could be identified and quantified by routine methods available in the clinics.
To eliminate genetic influences, we pursue the following specific goals:
- Aim 1 to interrogate the immune compartment of a unique cohort of monozygotic twin pairs -discordant for MS- and deeply analyse peripheral blood lymphocytes by single-cell mass cytometry, combined TcR and single cell sequencing, and epigenetic profiling.
- Aim 2 to develop representation-learning methods to account for the paired genetics of twins or longitudinal samples and to include clinical covariates into the high-dimensional data set.
- Aim 3 to use well-defined patient samples of MS-like disorders (MS-Mimics) and longitudinal samples of patients undergoing disease-modifying therapy (e.g. B cell depletion, autologous stem cell transplant) using single-cell mass cytometry.
Ultimately, the goal is to reduce the dimensionality of disease signature(s) towards a clinically translatable low-dimensional biomarker that could be identified and quantified by routine methods available in the clinics.
We made great progress towards the interrogation of the immune compartment in monozygotic twin pairs -discordant for MS-. The work for aim 1 is finished and published (see Ingelfinger, Gerdes et al. Nature, 2022).
The development of representation-learning methods to account for paired genetics or longitudinal samples is an ever-evolving endeavour. Several breakthroughs were accomplished by collaborations with our teams computational biologists and biologists together with collaborating labs. Some of these accomplishments can be read in (Diebold et al., 2022; Ingelfinger et al., 2022a; Ingelfinger et al., 2022b; Ingelfinger et al., 2022c; Kreutmair et al., 2022a; Kreutmair et al., 2022b; Nuñez et al. 2023).
Aim 3, using patient samples of MS-like disorders (MS-Mimics) and longitudinal samples of patients undergoing disease-modifying therapy is ongoing. We use single cell spectral flow cytometry over CyTOF because of the superior dynamic range, sensitivity, and acquisition speed. We developed the expertise and tools in our lab to push that technology to the limit by creating 40+ parameters panels. We already acquired 2 cohorts of patient samples with MOG-AD, a rare MS mimic with pathogenic autoantibodies against MOG as disease drivers. Data analysis is ongoing. Equally, the identification of predictive biomarkers for disease response has progressed fast and is already published (Diebold et al., 2022). Lastly, the identification of the mechanistic underpinnings of B cell depleting therapies in MS has pointed towards upregulated CD27 expression by T cells to be most potently and reproducibly affected by the loss of peripheral B cells in response to BCDTs (Ulutekin, Galli et al. MedRxiv). We could validate this finding in an independent cohort which we analysed recently.
The development of representation-learning methods to account for paired genetics or longitudinal samples is an ever-evolving endeavour. Several breakthroughs were accomplished by collaborations with our teams computational biologists and biologists together with collaborating labs. Some of these accomplishments can be read in (Diebold et al., 2022; Ingelfinger et al., 2022a; Ingelfinger et al., 2022b; Ingelfinger et al., 2022c; Kreutmair et al., 2022a; Kreutmair et al., 2022b; Nuñez et al. 2023).
Aim 3, using patient samples of MS-like disorders (MS-Mimics) and longitudinal samples of patients undergoing disease-modifying therapy is ongoing. We use single cell spectral flow cytometry over CyTOF because of the superior dynamic range, sensitivity, and acquisition speed. We developed the expertise and tools in our lab to push that technology to the limit by creating 40+ parameters panels. We already acquired 2 cohorts of patient samples with MOG-AD, a rare MS mimic with pathogenic autoantibodies against MOG as disease drivers. Data analysis is ongoing. Equally, the identification of predictive biomarkers for disease response has progressed fast and is already published (Diebold et al., 2022). Lastly, the identification of the mechanistic underpinnings of B cell depleting therapies in MS has pointed towards upregulated CD27 expression by T cells to be most potently and reproducibly affected by the loss of peripheral B cells in response to BCDTs (Ulutekin, Galli et al. MedRxiv). We could validate this finding in an independent cohort which we analysed recently.
We gladly report that goal 1, the interrogation of the immune compartment in monozygotic twin pairs -discordant for MS- was completed much faster than anticipated. The elimination of the genetic influence observed in cross-sectional cohorts allowed for the distillation of clearly environmentally driven immune signatures for MS. Beyond the strong genetic predisposition endowed by a variant of the IL-2Ra, we found that this pathway (incl. the ligand IL-2) is also highly related to the environmental risk of developing MS. Additional dysregulations were observed in the myeloid compartment with elevated GM-CSF sensing across monocytic cells of MS individuals.
Several emerging technologies did coalesce to allow for a better and unbiased interrogation of the immune compartment of paired samples (twins and longitudinal). We adapted the diffcyt analysis to be able to account for the paired design. Also, the inclusion of clinical covariates is possible with good reproducibility. We will continue to improve existing or develop new methods for data-driven analysis. Current therapies, such as Dimethylfumarate or the depletion of B cells have proven to be very efficacious, but the mechanisms of their action remain essentially elusive. We have made great progress in better understanding how these therapies work and identified biomarkers for disease response. The analysis of the MS mimic MOG-AD has proven to be one of the most exciting aspects of this project in that we found dramatic systemic changes in the immune compartment of patients afflicted with MOG-AD compared with healthy individuals. The verification of an overall trafficking defect is ongoing as are preclinical experiments to better understand the involvement of chemokine receptors in this disease.
Several emerging technologies did coalesce to allow for a better and unbiased interrogation of the immune compartment of paired samples (twins and longitudinal). We adapted the diffcyt analysis to be able to account for the paired design. Also, the inclusion of clinical covariates is possible with good reproducibility. We will continue to improve existing or develop new methods for data-driven analysis. Current therapies, such as Dimethylfumarate or the depletion of B cells have proven to be very efficacious, but the mechanisms of their action remain essentially elusive. We have made great progress in better understanding how these therapies work and identified biomarkers for disease response. The analysis of the MS mimic MOG-AD has proven to be one of the most exciting aspects of this project in that we found dramatic systemic changes in the immune compartment of patients afflicted with MOG-AD compared with healthy individuals. The verification of an overall trafficking defect is ongoing as are preclinical experiments to better understand the involvement of chemokine receptors in this disease.