Tibial Bone Compression Promotes Recovery After Brain Injury Through Osteocyte PIEZO1

Long summary of Tibial Bone Compression Promotes Recovery After Brain Injury Through Osteocyte Piezo1 by Zhiqing Cai et al.; with some added context; Audio by Paper2Audio.

Additional context

This document... places neurorehabilitation within the emerging view that recovery after brain injury is shaped not only by neural circuits and local immune responses but also by signals exchanged among distant organs. It builds on the established concept of bone as an endocrine organ, extending research on osteokines and skeletal regulation of metabolism and brain function by showing how mechanical forces can alter bone-derived signals.

Its focus on osteocyte Piezo1 connects to broader work on mechanotransduction, in which mechanically gated ion channels translate physical stimuli into cellular and systemic responses. The findings also intersect with research on secondary injury after stroke and traumatic brain injury, where neuroinflammation, neuronal loss, and impaired plasticity are recognized therapeutic targets beyond the initial lesion.

More broadly, this work complements exercise and rehabilitation research by suggesting that controlled skeletal loading may provide a peripheral, potentially noninvasive route to influence brain repair, while raising questions about which circulating factors and dosing strategies can translate safely to patients.

Paper Summary: We investigated whether mechanically stimulating bone could improve recovery after traumatic brain injury and stroke. These injuries can trigger bone healing and other skeletal responses, but whether bone activity can in turn benefit the injured brain was unclear. Bone is not only structural: it also acts as an endocrine organ, releasing signaling proteins known as osteokines, and its activity responds to mechanical forces.

We focused on osteocytes, the long-lived cells embedded in bone that account for more than 95 percent of bone cells and sense mechanical loading. We proposed that these cells might translate bone compression into circulating signals that influence brain repair. The study therefore tested both the therapeutic effects of tibial compression and the cellular and molecular processes that might explain them.

Our central approach was dynamic compressive tibial axial loading, abbreviated D.C.T.A.L: repeated, controlled compression applied along the long axis of the tibia. We examined its effects in mouse models of traumatic brain injury and stroke, tested the underlying mechanisms in cells and genetically modified mice, and assessed therapeutic effects in pigs with traumatic brain injury.

In mice, our standard moderate loading regimen applied four newtons of force at two cycles per second for 300 cycles, five days per week. We also examined lower forces of 0.5 and one newton. For moderate traumatic brain injury, loading began 24 hours after injury and was generally given for five days; separate experiments tested loading before or after stroke induction.

We used controlled cortical impact to produce mild, moderate or severe traumatic brain injury, and induced stroke by temporarily blocking the middle cerebral artery for 45 minutes. Motor function was evaluated with a pole test, in which we measured the time needed to turn and descend. Spatial learning and memory were assessed with the Morris water maze, including a probe trial that counted crossings of the former platform location and time spent in its target quadrant.

For the large-animal experiments, we used four-month-old male Bama miniature pigs with controlled cortical impact injury. Their tibiae received compression at two times body weight, two cycles per second and 400 cycles, five days per week. We measured learning-related performance in a T-maze, exploration in an open field, brain structure by magnetic resonance imaging, and tissue and blood markers associated with brain and bone responses.

In mice with severe traumatic brain injury, loading begun 24 hours after injury produced 90 percent survival at 14 days, compared with 20 percent in untreated injured controls. The survival difference was statistically significant, and the protective effect extended beyond the immediate post-injury period.

In moderate traumatic brain injury, five days of loading improved pole-test performance at both one and eight weeks after injury. The benefits included shorter movement-initiation latency and total descent time. At one week, loaded mice crossed the former water-maze platform location 4.88 times as often as untreated injured mice; at eight weeks, the increase was 1.54-fold. Preference for the target quadrant improved by 62.2 percent at one week and 28.6 percent at eight weeks.

Loading also improved outcomes in mild traumatic brain injury, and lower-force regimens of 0.5 or one newton benefited mice with moderate injury on cognitive measures and total pole-test time. In stroke experiments, loading before the event reduced pole-test return time by 54.3 percent and total time by 52.3 percent relative to stroke controls. Loading after stroke also improved both motor measures, showing benefit in preconditioning and post-injury treatment paradigms.

We found less hippocampal neuron loss after loading in traumatic brain injury mice. Measures of neuronal structure and cell abundance both improved: hippocampal M.A.P.2 signal and the number of Nissl-stained cells were higher in loaded animals than in untreated injured animals. In stroke mice, loading before or after the event reduced the volume of the cerebral infarct.

The treatment also shifted the post-injury inflammatory environment. One week after traumatic brain injury, loaded mice had 78.6 percent fewer I.B.A-1-positive cells, a marker of activated microglia, and 75.6 percent fewer G.F.A.P-positive reactive astrocytes in damaged hippocampal areas. In brain tissue, inflammatory cytokines T.N.F, interleukin 1 beta and interleukin 6 decreased, while anti-inflammatory cytokines including interleukin 4 and transforming growth factor beta increased.

Flow cytometry of injured brain tissue showed fewer macrophages, inflammatory macrophages and neutrophils after loading. At the same time, measures associated with neurogenesis, the production of new neurons, increased: immature-neuron marker D.C.X-positive cells rose 4.5-fold, while Nestin-positive and Ki67-positive cell populations each rose 5.67-fold. Together, the tissue findings show that loading was associated with reduced secondary injury and increased indicators of neuronal regeneration.

Piezo1 is a mechanically activated ion channel that allows cells to convert physical force into intracellular signals. To test whether osteocytes require this channel for the benefits of loading, we conditionally deleted Piezo1 in osteocytes using tamoxifen-inducible mice expressing Cre under the sclerostin promoter. Protein analysis and tissue staining confirmed reduced Piezo1 in the targeted bone cells.

In control mice, loading improved motor and water-maze performance after traumatic brain injury and increased hippocampal neuronal markers. These benefits were lost or substantially diminished when Piezo1 was deleted in osteocytes. The genetic intervention therefore identified osteocyte Piezo1 as necessary for the measured effects of tibial compression in this model.

We also tested whether activating Piezo1 locally in bone could reproduce part of the loading effect. Intratibial administration of the Piezo1 agonist M.C.B-22 to 174 at two milligrams per kilogram improved pole-test and water-maze performance after moderate traumatic brain injury and increased hippocampal M.A.P.2 signal. This complementary intervention supports the role of bone Piezo1 activation in the observed protection.

To determine whether signals carried in blood could transmit the effects of loading, we injected serum from loaded or untreated traumatic brain injury mice into other injured mice. Serum from loaded donors improved pole-test performance and spatial-memory measures compared with control serum. Recipients also showed higher hippocampal M.A.P.2 signal and more Nissl-stained cells.

We tested the bone as a direct source of protective activity by collecting supernatant from cortical bone of loaded mice. This material improved neuronal M.A.P.2 measures after oxygen-glucose deprivation, a laboratory model of cellular stress caused by reduced oxygen and glucose. Supernatant from bones with osteocyte-specific Piezo1 deletion did not reproduce the same protection, linking the activity to Piezo1-dependent bone responses.

Transection of the femoral and sciatic nerves did not eliminate the improvement in hippocampal neuronal measures produced by loading. Taken together with the serum-transfer and bone-supernatant experiments, this result indicates that peripheral nerve perception was not required for the protective effect tested here, while circulating and bone-released factors could convey it.

After loading, serum concentrations of brain-derived neurotrophic factor, or B.D.N.F, platelet factor 4, or P.F.4, and dopamine increased relative to injured controls. B.D.N.F supports neuronal survival and growth, P.F.4 is a platelet-derived signaling protein, and dopamine participates in motor and cognitive processes. These changes suggest that loading altered several circulating signals in addition to factors released directly by bone.

Bone gene-expression analysis identified Hspa1a, which encodes heat shock protein 70, or H.S.P.70, as strongly responsive to loading. Hspa1a expression in tibial cortical bone increased 30-fold, and circulating H.S.P.70 rose over the course of treatment. Deleting Piezo1 in osteocytes prevented the loading-associated increase in bone and serum H.S.P.70.

Cell experiments further connected Piezo1 activation to H.S.P.70 production through A.K.T signaling, an intracellular pathway involved in cell regulation. Activating Piezo1 in osteocyte-like cells increased H.S.P.70 in cells and culture supernatant, while an A.K.T inhibitor prevented this increase. These results support a mechanistic route from mechanical sensing in bone to release of a circulating neuroprotective factor.

We assessed whether H.S.P.70 contributed to the benefits of loading by administering a neutralizing antibody to loaded traumatic brain injury mice. The antibody partially reversed improvements in pole-test performance and spatial memory and reduced the loading-associated preservation of hippocampal neuronal markers. This intervention supports a functional contribution of H.S.P.70 without implying that it accounts for every effect of loading.

Conversely, recombinant H.S.P.70 administered to injured mice improved motor and water-maze outcomes and reduced neuronal loss. In cultured neurons exposed to oxygen-glucose deprivation, H.S.P.70 also increased M.A.P.2 signal across the tested treatment conditions. These findings were consistent with the serum and bone experiments showing that mechanically stimulated bone could provide protective activity.

We found elevated H.S.P.70 in cortical bone and blood after loading in both mice and pigs. The mouse experiments further tied this response to osteocyte Piezo1, while the cell experiments implicated A.K.T signaling. Together, the perturbation and replacement experiments identify H.S.P.70 as one mediator of the broader systemic response, rather than a complete explanation for the treatment's multiple outcomes.

Our bone transcriptomic analysis also highlighted interleukin 1 receptor type 2, or I.L-1.R.2, a decoy receptor that binds the inflammatory signals interleukin 1 alpha and interleukin 1 beta. Bone was a major source of I.L-1.R.2 among the tissues examined, and loading increased its expression in cortical bone and its concentration in serum. Osteocyte-specific Piezo1 deletion prevented the loading-associated increase in bone expression.

The inflammatory findings in injured brains were consistent with a role for this signal: loading reduced inflammatory cytokines and the recruitment of inflammatory cells, including macrophages and neutrophils. The study therefore identifies I.L-1.R.2 as a mechanically responsive bone-derived candidate for limiting inflammatory signaling after brain injury.

We also observed increased expression of I.L-1.R.2 in cortical bone from loaded pigs. While the mouse experiments connect the signal to osteocyte Piezo1, the reported results do not establish that I.L-1.R.2 alone accounts for the reduction in brain inflammation. We interpret it as one component of a coordinated set of bone-derived and systemic factors.

Apolipoprotein L 11a, or A.P.O.L.11.a, emerged from bone gene-expression analysis as another mechanically responsive factor. Its expression was prominent in bone relative to the other major tissues examined, increased after loading, and depended on osteocyte Piezo1. The study investigated A.P.O.L.11.a because cholesterol can support brain tissue during injury and other apolipoproteins participate in lipid transport.

To test whether loading affected movement of cholesterol from bone toward the brain, we introduced fluorescently labeled N.B.D-cholesterol into bone marrow. After loading, the injured brain contained 8.89 times more N.B.D-cholesterol signal than the corresponding control condition. In cell experiments, supernatant from Piezo1-activated or A.P.O.L.11.a-overexpressing osteocyte-like cells increased cholesterol-associated signal in stressed neuronal cells; reducing A.P.O.L.11.a weakened the effect.

We then manipulated Apol11a in mouse tibiae using viral vectors. Tibial A.P.O.L.11.a overexpression improved motor and memory outcomes after traumatic brain injury, whereas A.P.O.L.11.a knockdown weakened the benefits of loading and reduced neuronal-cell preservation. These results support a role for A.P.O.L.11.a in the recovery response and are consistent with its participation in cholesterol delivery, although they do not by themselves define every step of that transport process.

We tested the approach in pigs because their brains more closely resemble human brains in size and structure than rodent brains. In the pig protocol, tibial loading was applied at two times body weight, two cycles per second and 400 cycles, five days per week. In severe injury, loading extended median survival by five days compared with untreated controls.

In pigs with moderate traumatic brain injury, loading improved T-maze task times at one and four weeks and helped restore open-field exploration, measured as moving and sniffing a toy. Brain imaging at four weeks showed better preservation of brain tissue and less hemispheric atrophy in loaded animals. Tissue analysis also found reduced neuronal loss in the injured brain.

After five days of loading, pig cortical bone showed increased H.S.P.70 and I.L-1.R expression, while blood levels of H.S.P.70, B.D.N.F and P.F.4 rose. Serum from loaded pigs, administered to immunodeficient mice with traumatic brain injury, improved motor and water-maze performance and increased hippocampal M.A.P.2 signal. This transfer experiment provided evidence that the pig response included circulating activity capable of improving outcomes in the mouse model.

We assessed potential effects on bone, joints and peripheral nerves. In pigs, five days of loading did not produce a significant change in tibial bone mineral density, and examination found no detectable adverse changes in joint or meniscus structure. In mice, loading caused a marginal increase in the bone-resorption marker C.T.X-1 but did not alter the bone-formation marker P.1.N.P.

In healthy mice, five days of loading did not impair pole-test or water-maze performance and did not produce detectable structural damage in the sciatic nerve. Circulating H.S.P.70, I.L-1.R.2 and dopamine increased, whereas B.D.N.F and P.F.4 did not significantly change in these healthy animals. These observations describe the safety measures assessed in the experiments and do not establish safety in human use.

We compared loading with tibial fracture because fractures can accompany traumatic brain injury and may worsen outcomes. Fracture produced a distinct, inflammation-associated bone gene-expression response, did not increase serum H.S.P.70, I.L-1.R.2, P.F.4 or dopamine, and did not induce the loading-associated bone expression of Hspa1a, Il-1r2 or Apol11a. Unlike D.C.T.A.L, fracture did not reduce neuronal loss in the injured brain in the reported comparison.

Our results support a model in which mechanical compression activates osteocyte Piezo1 and changes bone's endocrine output. Directly responsive bone factors included H.S.P.70, I.L-1.R.2 and A.P.O.L.11.a, while serum levels of B.D.N.F, P.F.4 and dopamine also increased after loading in injured mice. Their combined effects were associated with neuronal preservation, reduced inflammation and improved behavioral recovery.

This pathway differs from the skeletal response to fracture. The fracture experiments showed inflammatory gene activity without the protective factor pattern associated with compression, and fracture did not reproduce the observed neuronal preservation. The comparison indicates that bone injury and controlled mechanical stimulation cannot be treated as interchangeable interventions in these models.

The work also links skeletal mechanical sensing to several stages of brain repair rather than to a single outcome. Loading was associated with changes in survival, motor and cognitive behavior, lesion size, inflammation, astrocyte reactivity and markers of neurogenesis. Because the study examined multiple factors and processes, the findings support a coordinated bone-to-brain response rather than a single-molecule account.

We observed functional benefits in both traumatic brain injury and stroke mouse models. In stroke, loading before injury and shortly after injury reduced motor deficits, and both schedules reduced infarct volume. The similarity of these outcomes suggests that the intervention can influence processes shared across the two injury settings.

The mechanistic work was conducted primarily in traumatic brain injury models. Although the stroke results establish a beneficial effect in the tested mice, they do not identify whether the same bone-derived signals act through the same pathways after vascular occlusion. The biological causes of stroke and traumatic brain injury differ, and the experiments do not resolve the mechanisms specific to stroke.

Across the traumatic brain injury studies, behavioral benefits appeared alongside tissue-level changes, including less neuronal loss and inflammation and greater expression of regeneration-associated markers. In pigs, improved task performance coincided with imaging and histological evidence of tissue preservation. These converging observations strengthen the evidence for recovery effects in the tested models while leaving the precise contribution of each mediator to be determined.

The pig experiments demonstrate that tibial compression can be delivered after traumatic brain injury in a large animal and can be associated with improved functional and structural outcomes. The selected pig regimen used a force proportional to body weight, and loading was delivered while animals were anesthetized. This makes the approach experimentally distinct from voluntary exercise and shows that it can be administered without requiring active participation in these models.

The study proposes controlled tibial loading as a possible nonpharmacological strategy for neurorehabilitation, based on its effects in rodents and pigs and the measured absence of several short-term musculoskeletal harms. However, these are preclinical experiments, not clinical trials. They do not establish an effective or safe dose, schedule, or patient population for human treatment.

Several features also limit how far we can generalize the findings. The animal studies used male mice and male pigs, and the reporting states that investigators were not blinded to experimental conditions.

The central mechanisms were investigated mainly after traumatic brain injury, while the stroke mechanisms remain unresolved. These constraints define important considerations for testing the approach in future studies.

The evidence for osteocyte Piezo1 comes from complementary interventions: deleting Piezo1 in osteocytes weakened loading benefits, and activating Piezo1 locally in bone improved outcomes. Serum-transfer and bone-supernatant experiments further showed that circulating or bone-released material could reproduce protective effects in injured animals or stressed neuronal cultures. Together, these approaches connect mechanical sensing in bone with signals that act beyond the skeleton.

The factor-specific experiments provide different levels of causal evidence. H.S.P.70 neutralization partially weakened the benefits of loading, recombinant H.S.P.70 improved outcomes, and A.P.O.L.11.a overexpression and knockdown respectively improved and attenuated selected effects. I.L-1.R.2 was identified as a loading-responsive decoy receptor and linked to the inflammatory findings, but the reported work does not isolate its independent contribution to recovery to the same extent as the H.S.P.70 and A.P.O.L.11.a interventions.

Important questions remain about how the signals reach and affect the injured brain, how much each factor contributes, and whether the same interactions operate after stroke. The study also does not establish how loading intensity, timing and duration should be optimized beyond the tested animal protocols. These questions are necessary to interpret the mechanism fully and to assess whether the preclinical effects can be translated.

We show that repeated tibial compression improved survival and functional recovery in mouse traumatic brain injury models, improved motor outcomes in mouse stroke models, and promoted behavioral and tissue recovery after traumatic brain injury in pigs. The experimental regimen in mice was commonly four newtons at two cycles per second for 300 cycles over five days; the pig regimen was two times body weight at two cycles per second for 400 cycles over five days.

Our mechanistic findings identify osteocyte Piezo1 as a key sensor for the loading response and point to H.S.P.70, I.L-1.R.2 and A.P.O.L.11.a as bone-derived factors, alongside increases in circulating B.D.N.F, P.F.4 and dopamine. The resulting changes were associated with reduced neuronal loss, inflammation and astrogliosis, smaller lesions, and increased markers of neurogenesis. Serum transfer, genetic manipulation and factor-directed experiments support the contribution of this bone-to-brain signaling program.

We therefore establish mechanically responsive bone as a source of signals that can influence brain repair in the models tested. The pig results and the measured safety observations provide a preclinical basis for further evaluation, but human efficacy and safety remain untested. Advancing this approach requires resolving the mechanisms in stroke, refining the intervention, and determining whether the animal findings translate to people.