Oral Presentation ANZBMS-MEPSA-ANZORS 2022

Gram negative bacterial infections drive neurogenic heterotopic ossification after spinal cord or brain injury in mice and humans (#43)

Kylie A Alexander 1 , Hsu-Wen Tseng 1 , Marjorie Salga 2 3 , Selwin G Samuel 1 , Aurélien Dinh 4 , Sébastien Kerever 5 , Paer-Sélim Abback 6 , Catherine Paugam 6 7 , François Genêt 2 3 , Jean-Pierre Levesque 1
  1. Mater Research, The University of Queensland, Brisbane, Queensland, Australia
  2. Department of Physical Medicine and Rehabilitation, Raymond Poincaré Hospital, APHP, CIC 1429, Garches, France
  3. University of Versailles Saint Quentin en Yvelines, END:ICAP U1179 INSERM, UFR Simone Veil-Santé, Montigny le Bretonneux, France
  4. Department of Infectious Disease, Raymond Poincaré Hospital, APHP, CIC 1429, Garches, France
  5. Department of Anesthesiology and Critical Care, Lariboisière University Hospital, AP-HP, Paris, France
  6. Department of Intensive Care Unit, Beaujon Hospital, Clichy, France
  7. University of Paris VII Denis-Diderot, Paris, France

Neurogenic heterotopic ossifications (NHO) are incapacitating complications of traumatic brain and spinal cord injuries (SCI) which manifest as abnormal heterotopic bones in periarticular muscles.  NHOs are debilitating, causing pain, joint ankylosis, as well as vascular and nerve compression. The mechanisms leading to NHO are largely unknown and the only effective treatment remains surgical resection. To elucidate NHO pathophysiology we developed the first model of NHO following SCI, in genetically unmodified mice. Using this model, we have established numerous inflammatory pathways that drive NHO pathogenesis. Several retrospective studies have shown that NHO prevalence is higher in patients who suffer concomitant infections. However, it is unclear whether these infections directly contribute to NHO development or reflect the immunodepression observed in SCI patients.

Using our NHO mouse model consisting of a SCI (T11-13) and a muscle injury via an intramuscular injection of cardiotoxin (CDTX), we demonstrate that infection with gram-negative bacteria, mimicked via the intramuscular administration of lipopolysaccharide (LPS, purified from the wall of gram-negative Escherishia coli), resulted in an exacerbation of NHO volumes measured via micro-computed tomography. Systemic LPS administration to mimic endotoxemia also increased NHO. The effect of LPS was TLR4-dependent as LPS did not increase NHO in TLR4-/- mice, and TLR4 transcripts and protein were expressed by muscle fibro-adipogenic progenitors. Using mice knocked-out for MyD88 or TRIF/Ticam1 adaptors, we also established that LPS enhances NHO via the endosomal TLR4/TRIF pathway, not via the plasma membrane TLR4/MyD88 pathway. Finally, we examined the presence of infection and NHO development in patients admitted to ICU after neurological injury. Interestingly, the presence of gram-negative bacterial infections (more specifically Pseudomonas aeruginosa) correlated with increased NHO development. Together these data demonstrate a correlation between infection and NHO development, and further identify specific bacterial agents and pathways associated with the formation of NHO in mouse and human.