Preconditioned mesenchymal stem cells treat myasthenia gravis in a humanized preclinical model
Muriel Sudres, Marie Maurer, Marieke Robinet, Jacky Bismuth, Frédérique Truffault, Diane Girard, Nadine Dragin, Mohamed Attia, Elie Fadel, Nicola Santelmo, Camille Sicsic, Talma Brenner, Sonia Berrih-Aknin
Muriel Sudres, Marie Maurer, Marieke Robinet, Jacky Bismuth, Frédérique Truffault, Diane Girard, Nadine Dragin, Mohamed Attia, Elie Fadel, Nicola Santelmo, Camille Sicsic, Talma Brenner, Sonia Berrih-Aknin
View: Text | PDF
Research Article Immunology

Preconditioned mesenchymal stem cells treat myasthenia gravis in a humanized preclinical model

Abstract

Myasthenia gravis (MG) with anti–acetylcholine receptor (AChR) Abs is an autoimmune disease characterized by severe defects in immune regulation and thymic inflammation. Because mesenchymal stem cells (MSCs) display immunomodulatory features, we investigated whether and how in vitro–preconditioned human MSCs (cMSCs) could treat MG disease. We developed a new humanized preclinical model by subcutaneously grafting thymic MG fragments into immunodeficient NSG mice (NSG-MG model). Ninety percent of the animals displayed human anti-AChR Abs in the serum, and 50% of the animals displayed MG-like symptoms that correlated with the loss of AChR at the muscle endplates. Interestingly, each mouse experiment recapitulated the MG features of each patient. We next demonstrated that cMSCs markedly improved MG, reducing the level of anti-AChR Abs in the serum and restoring AChR expression at the muscle endplate. Resting MSCs had a smaller effect. Finally, we showed that the underlying mechanisms involved (a) the inhibition of cell proliferation, (b) the inhibition of B cell–related and costimulatory molecules, and (c) the activation of the complement regulator DAF/CD55. In conclusion, this study shows that a preconditioning step promotes the therapeutic effects of MSCs via combined mechanisms, making cMSCs a promising strategy for treating MG and potentially other autoimmune diseases.

Authors

Muriel Sudres, Marie Maurer, Marieke Robinet, Jacky Bismuth, Frédérique Truffault, Diane Girard, Nadine Dragin, Mohamed Attia, Elie Fadel, Nicola Santelmo, Camille Sicsic, Talma Brenner, Sonia Berrih-Aknin

×

Figure 1

Characterization of the novel NSG-MG model.

Options: View larger image (or click on image) Download as PowerPoint
Characterization of the novel NSG-MG model. (A) Human acetylcholine rece...
(A) Human acetylcholine receptor–specific (AChR-specific) Abs were measured in mouse serum in the 3 myasthenia gravis (MG) groups and the control (CTRL) group. Each symbol represents the mean value ± SEM of human anti-AChR Ab titers measured in mice for each experiment (CTRL group, 12 grafts, 51 mice; MG neg group, 4 grafts, 14 mice; MG low, 4 grafts 16 mice, MG high group, 9 grafts, 31 mice). (B) The mice displayed MG-like clinical symptoms. The Kaplan-Meyer curve shows the occurrence of the disease (score >1) in the high-titer anti-AChR seropositive MG group (n = 31), in the low-titer seropositive MG group (n = 18), in the seronegative MG group (n = 14), and in the CTRL group (n = 51). (C) The endplate AChR content of the diaphragmatic muscle (measured by 125I-α-bungarotoxin labeling) was diminished in the MG groups. The cpm data are normalized using the mean cpm values of the CTRL group (levels set at 100%). Each symbol represents the cpm value mean ± SEM of each graft experiment (CTRL group, 10 grafts, 37 mice; MG low group, 3 grafts, 10 mice; MG high group, 8 grafts, 25 mice). (D) Endplate AChR loss correlated with MG severity. Each symbol represents 1 mouse from the MG groups (seronegative and seropositive). (E) Patient Ab titers correlated with mouse Ab titers. Each symbol represents the AChR-specific Ab titer measured in 1 MG patient and the corresponding mean Ab titer value measured in mice for each experiment (P < 0.0003). (F) The patient score correlated with the mouse score. Each symbol represents the score of 1 MG patient (untreated by corticosteroids) and the corresponding mean score obtained in mice for each experiment (P < 0.03). (G and H) In humans (G) and mice (H), Ab titers did not correlate with clinical scores. P values were determined using the 1-way ANOVA test (A and C) or the log-rank (Mantel-Cox) test (B). (D–H) R and P values were determined using the linear regression test. *P < 0.05; ***P < 0.001; ****P < 0.0001. ns, not significant.