IRCCS San Raffaele Scientific Institute Showcase

Cardiovascular dysregulation is a frequent non-motor manifestation of multiple sclerosis (MS), potentially linked to structural damage in autonomic control regions such as the insular cortex. This study examined the role of insular integrity in modulating cardiovascular fitness (CF) and ambulation, and its influence on the efficacy of aerobic training (AT) in people with MS. A total of 61 MS patients (20 with relapsing–remitting and 41 with progressive MS) were retrospectively analyzed and randomized into two intervention arms: an AT group (MS-AT, n = 31) performing moderate-intensity aerobic exercise, and a motor training control group (MS-C, n = 30) performing non-aerobic exercises. Each patient underwent clinical, cardiopulmonary, and MRI evaluations at baseline and after 24 training sessions over 2–3 months. Two separate healthy control groups were included: one for cardiopulmonary assessment (HC-clinic, n = 20) and another for imaging comparison (HC-MRI, n = 60). At baseline, MS patients demonstrated significantly reduced VO₂max, HR reserve (HRR), and 6-minute walk test (6MWT) performance compared to healthy controls, alongside widespread gray matter atrophy including the bilateral insula. Notably, the presence of left insular T2-hyperintense lesions correlated with impaired HRR. Following AT, the MS-AT group—particularly patients without insular lesions—showed significant improvement in 6MWT distance and preservation of insular volume. In contrast, the MS-C group exhibited progressive left insular atrophy. Moreover, greater gains in walking capacity were significantly associated with more limited loss of left anterior insular volume. These results highlight the insula’s role in modulating both baseline cardiovascular impairment and training-related functional improvements. Assessing insular integrity via MRI may offer predictive value for rehabilitation outcomes and guide individualized aerobic training strategies in MS care.

Spinal cord atrophy is a key contributor to disability in multiple sclerosis (MS), with early and progressive loss of cervical cord volume correlating with worsening clinical outcomes. This study explored whether aerobic capacity moderates the association between spinal cord atrophy and clinical disability in people with MS. A cross-sectional analysis was conducted on 51 MS patients with mild to moderate disability (Expanded Disability Status Scale [EDSS] ≤ 6.0) and 33 age- and sex-matched healthy controls (HCs). All participants underwent 3T brain MRI to assess normalized mean upper cervical cord area (nMUCCA), a surrogate for spinal cord atrophy. Aerobic capacity was assessed via peak oxygen consumption (VO₂peak) measured during cardiopulmonary exercise testing (CPET), with low aerobic capacity defined as a z-score < –1.64. MS patients demonstrated significantly reduced nMUCCA and VO₂peak compared to HCs. Within the MS cohort, 61% were classified as having low aerobic capacity. Although no major differences in brain volumes or clinical characteristics (aside from body mass index) were observed between low and high aerobic capacity subgroups, a significant negative correlation between nMUCCA and EDSS was found only in those with low aerobic capacity. Moderation analysis confirmed a significant interaction between spinal cord atrophy and aerobic capacity in predicting disability (β = –0.099, p = 0.012). Specifically, reduced nMUCCA was associated with higher disability only in MS patients with low aerobic capacity, but not in those with high aerobic capacity. These findings suggest that greater aerobic capacity may attenuate the detrimental impact of spinal cord atrophy on clinical disability in MS, acting as a physical reserve factor. This highlights the potential of aerobic fitness enhancement as a non-pharmacological strategy to buffer neurodegenerative effects and support functional outcomes in MS, especially in the early stages of the disease.

Spinal cord (SC) involvement, particularly cervical SC (cSC) atrophy and lesions, is increasingly recognized as a key determinant of disability in multiple sclerosis (MS), yet its role in specific sensorimotor impairments remains underexplored. This cross-sectional multiparametric MRI study aimed to assess the relative contribution of brain and cSC lesion burden and atrophy to global disability and specific sensorimotor functions in MS. A total of 151 MS patients and 69 age- and sex-matched healthy controls underwent 3T brain and cSC MRI and clinical evaluations including Expanded Disability Status Scale (EDSS), pyramidal and sensory functional system (P-FS, S-FS) scores, motor performance tests (9-hole peg test [9-HPT], finger tapping test [FTT], timed 25-foot walk test [T25FWT]), and vibration detection threshold (VDT). Random forest models identified MRI features most predictive of disability milestones (EDSS ≥3.0, ≥4.0, and ≥6.0), functional system impairments, and motor/sensory test performance. Across all outcomes, cSC gray matter cross-sectional area (GM CSA) emerged as the most informative variable, outperforming brain volumetric and lesion metrics. GM CSA was the strongest predictor of progressive MS phenotype and higher EDSS scores (AUC up to 0.900). It also best explained impairments in lower limb motor function (T25FWT), upper limb repetitive motor performance (FTT), and vibratory sensation (VDT). In contrast, 9-HPT impairment, a more complex dexterity task, was associated with a broader combination of cSC and brain abnormalities, including lesion burden and cortical and deep gray matter atrophy. VDT impairment was linked to volume loss in both cSC GM and white matter (WM), supporting the role of dorsal columns and posterior horns in somatosensory processing. Brain metrics, particularly thalamic volume, also contributed to VDT variability. The findings highlight the predominant role of cSC GM atrophy as a structural correlate of sensorimotor impairment and disability in MS, especially at higher levels of functional decline. These results support integrating spinal cord MRI, particularly quantification of GM damage, into routine clinical assessments to enhance prognostic accuracy and guide personalized interventions in MS management.

Fatigue is a prevalent and disabling symptom in people with multiple sclerosis (pwMS), affecting both cognitive and physical domains. This study explored the relationship between aerobic capacity, thalamic nuclei volumes, and fatigue, and assessed whether thalamic integrity mediates the effect of aerobic fitness on fatigue perception. The study retrospectively enrolled 83 pwMS (36 with relapsing–remitting MS and 47 with secondary progressive MS) and 63 age- and sex-matched healthy controls (HC). All participants underwent 3T brain MRI to quantify thalamic and global brain volumes. Fatigue was evaluated using the Modified Fatigue Impact Scale (MFIS), and aerobic capacity was assessed via peak oxygen uptake (VO₂peak) from cardiopulmonary exercise testing (CPET), available for all pwMS and 22 HC. Compared to HC, pwMS exhibited significantly lower VO₂peak and thalamic volumes and higher global, physical, and cognitive fatigue scores. In pwMS, higher VO₂peak was associated with lower fatigue (MFIS and pMFIS) and with greater volume in the laterodorsal thalamic nuclei cluster (Dor). Furthermore, lower Dor volume was linked to greater fatigue in all MFIS domains. Mediation analyses revealed that the Dor cluster partially mediated the beneficial effects of VO₂peak on global fatigue (21% indirect effect) and cognitive fatigue (32% indirect effect), but not physical fatigue. These findings suggest that aerobic capacity influences fatigue in pwMS, partly via structural preservation of specific thalamic subregions, especially the laterodorsal nuclei. While physical fatigue may be more directly influenced by systemic or extra-thalamic factors, cognitive fatigue appears partially dependent on thalamic integrity. This underscores the potential neuroprotective role of cardiorespiratory fitness and highlights the Dor thalamic cluster as a critical neural substrate in MS-related fatigue. Targeting aerobic capacity through rehabilitation may not only alleviate fatigue but also support thalamic structural integrity in pwMS.

This study investigated the impact of neuroinflammation and cardiorespiratory fitness (CRF) on hippocampal volume in patients with multiple sclerosis (MS), with a focus on differences between clinical phenotypes. A total of 81 MS patients were included, of whom 27 had relapsing–remitting MS (RRMS) and 54 had progressive MS (PMS). An additional 45 age- and sex-matched healthy controls were enrolled for structural brain imaging comparisons. All participants underwent high-resolution 3T MRI to assess normalized brain, gray matter, hippocampal, and thalamic volumes. In MS patients, aerobic capacity was assessed via maximum oxygen consumption (VO₂max) using cardiopulmonary exercise testing, while T2-hyperintense lesion volume (T2-LV) and choroid plexus volume (CPV) were quantified as MRI markers of neuroinflammation. Compared to healthy controls, both RRMS and PMS patients exhibited significantly lower brain and regional volumes, including the hippocampus, and higher T2-LV and CPV, reflecting increased neuroinflammation. Regression analyses revealed that T2-LV and CPV were the most influential predictors of global and regional brain atrophy in both MS phenotypes. Notably, VO₂max was identified as a significant predictor of hippocampal volume only in RRMS patients, explaining 16.9% of the variance and retained as a relevant predictor alongside T2-LV in multivariate models. In contrast, VO₂max had no significant explanatory value for hippocampal or other volumetric outcomes in PMS patients. These findings suggest a selective and clinically relevant association between aerobic fitness and hippocampal integrity in the early, relapsing phase of MS, possibly mediated by mechanisms such as neurogenesis or neuroprotection. In more advanced stages of the disease, such as PMS, widespread neurodegeneration may overshadow the potential benefits of aerobic capacity. The results underscore the importance of promoting CRF early in the disease course as a strategy to preserve hippocampal structure and potentially mitigate cognitive decline.

An estimated 30% of Major Depressive Disorder (MDD) patients exhibit resistance to conventional antidepressant treatments. Identifying reliable biomarkers of treatment-resistant depression (TRD) represents a major goal of precision psychiatry, which is hampered by the clinical and biological heterogeneity underlying MDD. To parse heterogeneity and uncover biologically-driven subtypes of MDD, we applied an unsupervised data-driven framework to stratify 102 MDD patients on their neuroimaging signature, including extracted measures of cortical thickness, grey matter volumes, and white matter fractional anisotropy. Our novel analytical pipeline integrated different machine learning algorithms to harmonize neuroimaging data, perform data dimensionality reduction, and provide a stability-based relative clustering validation. The obtained clusters were then characterized for immune-inflammatory peripheral biomarkers, TRD, history of childhood trauma and different profiles of depressive symptoms. Our results indicated two different clusters of patients, differentiable with 67% of accuracy: 1) one cluster (n=59) was associated with a higher proportion of TRD compared to the other, and higher scores of energy-related depressive symptoms, history of childhood abuse and emotional neglect; this cluster showed a widespread reduction in cortical thickness (d=0.43-1.80) and volumes (d=0.45-1.05), along with fractional anisotropy in the right superior fronto-occipital fasciculus, stria terminalis, and corpus callosum (d=0.46-0.52); 2) the second cluster (n=43) was associated with cognitive and affective depressive symptoms and thicker cortices and wider volumes compared to the other. Multivariate analyses revealed distinct brain-inflammation relationships between the two clusters, with increase in pro-inflammatory cytokines and chemokines being associated with decreased cortical thickness and grey matter volumes. Our stratification of MDD patients based on structural neuroimaging identified clinically-relevant subgroups of MDD with specific symptomatic, childhood trauma, and immune-inflammatory profiles, which can contribute to the development of tailored personalized interventions for MDD. For any clarification, please contact the corresponding author (Dr. Federica Colombo) at colombo.federica@hsr.it

Institutional barriers and lack of engagement in research-performing organisations (RPOs) may limit the development and impact of Citizen Science (CS) initiatives. In the present case study, we detail the transformative and multidisciplinary approach of Vita-Salute San Raffaele University (UniSR) through the European project TIME4CS, showcasing how tailored roadmaps and mutual learning with other RPOs with established support structures and expertise for CS were able to overcome these challenges. The approach involved several key steps: 1) creation of a de novo research organization area dedicated to Research Development; 2) formation of a multidisciplinary core team to implement TIME4CS activities; 3) mapping the initial and final levels of awareness of CS among UniSR researchers through surveys; 4) developing and implementing a detailed communication plan, including seminars, newsletters, articles, and a repository of CS resources; 5) involvement of UniSR students, professors, researchers but also research support officers in the initiatives; 6) establishment of a contact point for stakeholders interested in CS and active participation in ECSA groups; 7) support to the development of pilot initiatives and projects of CS. To collect information regarding the awareness and the interest of the research community in Citizen Science, after one year from the beginning of the TIME4CS project we shared a survey among researchers. To measure the impact of TIME4CS on the UniSR research community, 2 months after the end of the 36-month-long project (M36+2) we repeated as a follow-up the survey. The data here deposited refer to these surveys. The academic impact includes increased awareness and engagement in CS initiatives among UniSR researchers. The actions triggered by the TIME4CS project have led to the emergence of several new CS research projects, enhancing UniSR's research excellence and contributing to its strategic goals of internationalization and competitiveness. This case study provides a model for overcoming institutional barriers in the promotion of CS and enhancing research excellence. The wider impact of the initiatives includes fostering a more collaborative and inclusive research environment at UniSR. By involving researchers, students, professors, research support officers, and external stakeholders, the project promoted a culture of Open Science and Responsible Research and Innovation (RRI). The activities also contribute to the broader scientific community by participating in European Citizen Science Association (ECSA) groups and sharing resources and best practices, potentially influencing other institutions to adopt similar approaches.

Despite recent progress in cancer treatment, liver metastases persist as an unmet clinical need. Here, we show that arming liver and tumor-associated macrophages in vivo to co-express tumor antigens (TAs), IFNα, and IL-12 unleashes robust anti-tumor immune responses, leading to the regression of liver metastases. Mechanistically, in vivo armed macrophages expand tumor reactive CD8+ T cells, which acquire features of progenitor exhausted T cells and kill cancer cells independently of CD4+ T cell help. IFNα and IL-12 produced by armed macrophages reprogram antigen presenting cells and rewire cellular interactions, rescuing tumor reactive T cell functions. In vivo armed macrophages trigger anti-tumor immunity in distinct liver metastasis mouse models of colorectal cancer and melanoma, expressing either surrogate tumor antigens, naturally occurring neoantigens or tumor-associated antigens. Altogether, our findings support the translational potential of in vivo armed liver macrophages to expand and rejuvenate tumor reactive T cells for the treatment of liver metastases. Data availability: The MERFISH, single-cell RNA sequencing and bulk RNA sequencing data have been deposited in the GEO repository under the accession number GSE273615. Additionally, the WES data have been uploaded to the ENA portal with the accession number PRJEB78386. Code availability: Code is available at the following link: http://www.bioinfotiget.it/gitlab/custom/notaro_mouse_lm_2025.

Upon metastatic seeding in the liver, liver macrophages, including Kupffer cells, acquire a transcriptional profile typical of tumor-associated macrophages (TAMs), which support tumor progression. MicroRNAs (miRNAs) fine-tune TAM pro-tumoral functions, making their modulation a promising strategy for macrophage reprogramming into an anti-tumoral phenotype. Here, we analyze the transcriptomic profiles of liver and splenic macrophages, identifying miR-342-3p as a key regulator of liver macrophage function. miR-342-3p is highly active in healthy liver macrophages but significantly downregulated in colorectal cancer liver metastases (CRLMs). Lentiviral vector-engineered liver macrophages enforcing miR-342-3p expression acquire a pro-inflammatory phenotype and reduce CRLM growth. We identify Slc7a11, a cysteine-glutamate antiporter linked to pro-tumoral activity, as a direct miR-342-3p target, which may be at least partially responsible for TAM phenotypic reprogramming. Our findings highlight the potential of in vivo miRNA modulation as a therapeutic strategy for TAM reprogramming, offering an approach to enhance cancer immunotherapy. Data and code availability: Next-generation sequencing data are deposited at the Gene Expression Omnibus (GEO) with the following accession numbers: GEO: GSE274043 (scRNA-seq), GSE274044 (RNA-seq on iKCs), GSE274045 (small RNA-seq on splenic and hepatic cell populations), and GSE274046 (bulk RNA-seq on splenic and hepatic cell populations). The code is available at GitLab: http://www.bioinfotiget.it/gitlab/custom/Bresesti_Cell_Reports_2025.

The dataset includes the raw data of graphs shown in figures 1, 2, 3, 5, 6, and 7 and in supplementary figures 1, 2, 3, 4, 5, 7, and 8 of the paper: Human iPSC-derived neural stem cells displaying radial glia signature exhibit long-term safety in mice. Luciani M, Garsia C, Beretta S, Cifola I, Peano C, Merelli I, Petiti L, Miccio A, Meneghini V, Gritti A. Nat Commun. 2024 Nov 1;15(1):9433. doi: 10.1038/s41467-024-53613-7. PMID: 39487141 The bulk RNA-seq, SREBF1-deficient RNA-seq, and ChIP-seq data generated in this study have been deposited at GEO under accession number GSE239446. The single-cell RNA-seq data generated in this study have been deposited at GEO under accession number GSE238206. The processed RNA-seq, ChIP-seq, and scRNA-seq data (list of DEGs and GO terms) are available in Supplementary Data 1-4 files. Human induced pluripotent stem cell-derived neural stem/progenitor cells (hiPSC-NSCs) hold promise for treating neurodegenerative and demyelinating disorders. However, comprehensive studies on their identity and safety remain limited. In this study, we demonstrate that hiPSC-NSCs adopt a radial glia-associated signature, sharing key epigenetic and transcriptional characteristics with human fetal neural stem cells (hfNSCs) while exhibiting divergent profiles from glioblastoma stem cells. Long-term transplantation studies in mice showed robust and stable engraftment of hiPSC-NSCs, with predominant differentiation into glial cells and no evidence of tumor formation. Additionally, we identified the Sterol Regulatory Element Binding Transcription Factor 1 (SREBF1) as a regulator of astroglial differentiation in hiPSC-NSCs. These findings provide valuable transcriptional and epigenetic reference datasets to prospectively define the maturation stage of NSCs derived from different hiPSC sources and demonstrate the long-term safety of hiPSC-NSCs, reinforcing their potential as a viable alternative to hfNSCs for clinical applications.