Transcranial amelioration of inflammation and cell death after brain injury



Traumatic brain injury (TBI) is increasingly appreciated to be highly prevalent and deleterious to neurological function1,2. At present, no effective treatment options are available, and little is known about the complex cellular response to TBI during its acute phase. To gain insights into TBI pathogenesis, we developed a novel murine closed-skull brain injury model that mirrors some pathological features associated with mild TBI in humans and used long-term intravital microscopy to study the dynamics of the injury response from its inception. Here we demonstrate that acute brain injury induces vascular damage, meningeal cell death, and the generation of reactive oxygen species (ROS) that ultimately breach the glial limitans and promote spread of the injury into the parenchyma. In response, the brain elicits a neuroprotective, purinergic-receptor-dependent inflammatory response characterized by meningeal neutrophil swarming and microglial reconstitution of the damaged glial limitans. We also show that the skull bone is permeable to small-molecular-weight compounds, and use this delivery route to modulate inflammation and therapeutically ameliorate brain injury through transcranial administration of the ROS scavenger, glutathione. Our results shed light on the acute cellular response to TBI and provide a means to locally deliver therapeutic compounds to the site of injury.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Figure 1: Pathology associated with compression injury.
Figure 2: Innate immune response to a compression injury.
Figure 3: Metrics of transcranial diffusion through the skull bone.
Figure 4: Purinergic receptor signalling mediates the innate immune response to compression injury.
Figure 5: Transcranial administration of glutathione reduces inflammation and cell death after compression injury.


  1. 1

    DeKosky, S. T., Blennow, K., Ikonomovic, M. D. & Gandy, S. Acute and chronic traumatic encephalopathies: pathogenesis and biomarkers. Nature Rev. Neurol. 9, 192–200 (2013)

    CAS  Article  Google Scholar 

  2. 2

    Roozenbeek, B., Maas, A. I. & Menon, D. K. Changing patterns in the epidemiology of traumatic brain injury. Nature Rev. Neurol. 9, 231–236 (2013)

    Article  Google Scholar 

  3. 3

    Zetterberg, H., Smith, D. H. & Blennow, K. Biomarkers of mild traumatic brain injury in cerebrospinal fluid and blood. Nature Rev. Neurol. 9, 201–210 (2013)

    CAS  Article  Google Scholar 

  4. 4

    McConeghy, K. W., Hatton, J., Hughes, L. & Cook, A. M. A review of neuroprotection pharmacology and therapies in patients with acute traumatic brain injury. CNS Drugs 26, 613–636 (2012)

    CAS  Article  Google Scholar 

  5. 5

    Xiong, Y., Mahmood, A. & Chopp, M. Animal models of traumatic brain injury. Nature Rev. Neurosci. 14, 128–142 (2013)

    CAS  Article  Google Scholar 

  6. 6

    Finnie, J. W. Neuroinflammation: beneficial and detrimental effects after traumatic brain injury. Inflammopharmacology 21, 309–320 (2013)

    CAS  Article  Google Scholar 

  7. 7

    Hall, E. D., Andrus, P. K. & Yonkers, P. A. Brain hydroxyl radical generation in acute experimental head injury. J. Neurochem. 60, 588–594 (1993)

    CAS  Article  Google Scholar 

  8. 8

    Lewén, A., Matz, P. & Chan, P. H. Free radical pathways in CNS injury. J. Neurotrauma 17, 871–890 (2000)

    Article  Google Scholar 

  9. 9

    Young, W. Role of calcium in central nervous system injuries. J. Neurotrauma 9 (suppl. 1). S9–S25 (1992)

    PubMed  Google Scholar 

  10. 10

    Bullock, R. et al. Evidence for prolonged release of excitatory amino acids in severe human head trauma. Relationship to clinical events. Ann. NY Acad. Sci. 765, 290–297 (1995)

    ADS  CAS  Article  Google Scholar 

  11. 11

    Bullock, R. et al. Factors affecting excitatory amino acid release following severe human head injury. J. Neurosurg. 89, 507–518 (1998)

    CAS  Article  Google Scholar 

  12. 12

    Mazzeo, A. T., Beat, A., Singh, A. & Bullock, M. R. The role of mitochondrial transition pore, and its modulation, in traumatic brain injury and delayed neurodegeneration after TBI. Exp. Neurol. 218, 363–370 (2009)

    CAS  Article  Google Scholar 

  13. 13

    Janowitz, T. & Menon, D. K. Exploring new routes for neuroprotective drug development in traumatic brain injury. Sci. Transl. Med. 2, 27rv1 (2010)

    CAS  Article  Google Scholar 

  14. 14

    Mamourian, A. C., Hoopes, P. J. & Lewis, L. D. Visualization of intravenously administered contrast material in the CSF on fluid-attenuated inversion-recovery MR images: an in vitro and animal-model investigation. AJNR Am. J. Neuroradiol. 21, 105–111 (2000)

    CAS  PubMed  Google Scholar 

  15. 15

    Xu, H. T., Pan, F., Yang, G. & Gan, W. B. Choice of cranial window type for in vivo imaging affects dendritic spine turnover in the cortex. Nature Neurosci. 10, 549–551 (2007)

    CAS  Article  Google Scholar 

  16. 16

    Eltzschig, H. K., Sitkovsky, M. V. & Robson, S. C. Purinergic signaling during inflammation. N. Engl. J. Med. 367, 2322–2333 (2012)

    CAS  Article  Google Scholar 

  17. 17

    Davalos, D. et al. ATP mediates rapid microglial response to local brain injury in vivo. Nature Neurosci. 8, 752–758 (2005)

    CAS  Article  Google Scholar 

  18. 18

    Chen, G. Y. & Nunez, G. Sterile inflammation: sensing and reacting to damage. Nature Rev. Immunol. 10, 826–837 (2010)

    CAS  Article  Google Scholar 

  19. 19

    Nimmerjahn, A., Kirchhoff, F. & Helmchen, F. Resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo. Science 308, 1314–1318 (2005)

    ADS  CAS  Article  Google Scholar 

  20. 20

    Haynes, S. E. et al. The P2Y12 receptor regulates microglial activation by extracellular nucleotides. Nature Neurosci. 9, 1512–1519 (2006)

    CAS  Article  Google Scholar 

  21. 21

    McDonald, B. et al. Intravascular danger signals guide neutrophils to sites of sterile inflammation. Science 330, 362–366 (2010)

    ADS  CAS  Article  Google Scholar 

  22. 22

    Jung, S. et al. Analysis of fractalkine receptor CX3CR1 function by targeted deletion and green fluorescent protein reporter gene insertion. Mol. Cell. Biol. 20, 4106–4114 (2000)

    CAS  Article  Google Scholar 

  23. 23

    Zhuo, L. et al. Live astrocytes visualized by green fluorescent protein in transgenic mice. Dev. Biol. 187, 36–42 (1997)

    CAS  Article  Google Scholar 

  24. 24

    Faust, N., Varas, F., Kelly, L. M., Heck, S. & Graf, T. Insertion of enhanced green fluorescent protein into the lysozyme gene creates mice with green fluorescent granulocytes and macrophages. Blood 96, 719–726 (2000)

    CAS  PubMed  Google Scholar 

  25. 25

    Yang, G., Pan, F., Parkhurst, C. N., Grutzendler, J. & Gan, W. B. Thinned-skull cranial window technique for long-term imaging of the cortex in live mice. Nature Protocols 5, 201–208 (2010)

    CAS  Article  Google Scholar 

  26. 26

    Chuang, K. H., Koretsky, A. P. & Sotak, C. H. Temporal changes in the T1 and T2 relaxation rates (ΔR1 and ΔR2) in the rat brain are consistent with the tissue-clearance rates of elemental manganese. Magn. Reson. Med. 61, 1528–1532 (2009)

    Article  Google Scholar 

Download references


The study was supported by the National Institutes of Health (NIH), the National Institute of Neurological Disorders and Stroke (NINDS), and the Center for Neuroscience and Regenerative Medicine (CNRM) at the Uniformed Services University of the Health Sciences—a collaborative effort between the NIH, the Department of Defense and the Walter Reed National Military Medical Center to develop innovative approaches for brain injury diagnosis and recovery.

Author information




T.L.R., D.N. and D.B.M. designed all murine experiments and interpreted the data. L.L.L. provided and interpreted the human TBI data. T.A. and A.P.K. contributed the transcranial manganese data. T.L.R. and D.B.M. wrote the paper.

Corresponding author

Correspondence to Dorian B. McGavern.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Extended data figures and tables

Extended Data Figure 1 Compression injury model.

a, An intact mouse skull bone consists of three layers: cortical, cancellous, cortical. b, For standard skull thinning, a microdrill (not shown) and a microsurgical blade are used to manually thin all three layers of bone until only 30 µm of lower cortical bone remains. The diameter of the thinned area is 1,000 µm. This preparation permits intravital imaging of the unperturbed meninges and brain parenchyma. c, Compression injury is induced by thinning the bone to 20–30 µm and then using the flat blunt edge of a microsurgical blade to apply minimal downward pressure, which promotes concavity in the thinned skull region. d, This entire procedure is performed in 1–2 min, after which the cellular response to the compression injury can be continuously imaged by TPM as soon as 5 min after injury.

Extended Data Figure 2 Parenchymal cell death following compression injury.

Immunohistochemical analyses of coronal brain sections 12 h after compression injury (bottom) shows that cell death (red) extends several layers into the neocortex. Cell death at this time point was indiscriminate, as the lesion site was largely devoid of neuronal (NeuN+), astrocytic (GFAP+), oligodendrocyte (APC+), and microglia (Iba1+) staining (all in green). Dead cells were labelled by transcranially applying propidium iodide (red) 30 min before fixation. A normal uncompressed thinned skull preparation was performed on the contralateral hemisphere as a control (top). Note that parenchymal cell death is not observed in any of the control images. Cell nuclei are shown in blue. Images are representative of three mice per group.

Extended Data Figure 3 Microglia response to compression injury.

a, b, Representative 5 μm maximum projections in the xz plane (a, b) or 25 μm projections in the xy plane (b) were captured from CX3CR1gfp/+ mice at 60 min (a) or 6 h (b) after a normal thinned skull preparation (uncompressed) or compression injury. a, Microglia (green) extend processes into the meningeal space (white arrow) only after compression injury. The glial limitans is denoted with white dotted lines. b, After compression injury, SR101-labelled astrocytes (red) at the glial limitans are surrounded by a network of honeycomb microglia (top, xy projection). An example is indicated with a white arrowhead. In contrast, holes in the astrocytic coverage at the glial limitans are filled by jellyfish microglia (white arrow). A side view (xz projection) reveals microglia (green; white arrows) to be an integral part of the glial limitans after compression injury. Data are representative of three mice per group and at least three independent experiments.

Extended Data Figure 4 Pharmacology of transcranial fluorescent compound administration.

ac, Steady-state concentrations (Conc) of fluorescent compounds in aCSF were calculated by two-photon microscopy above the skull (blue) as well as in the meninges (orange) and brain parenchyma (green) for 1 h after transcranial administration through a surgically thinned (uncompressed) skull. Background fluorescence was established by imaging for 30 min before adding the fluorescent compounds. Low molecular weight compounds established higher steady-state concentrations in the meninges than larger compounds. The hydrophobic, low molecular weight compound Nile Red was also able to pass through the glial limitans into the brain parenchyma. See Methods for additional details about the applied fluorescent compounds. Oligonuc., oligonucleotide. d, A 100 nM SR101 solution was applied to an intact mouse skull for 30 min, then immediately thinned and imaged while the skull was submerged in aCSF to determine the decay in the meningeal SR101 concentration over time. e, The bar graph shows a comparison of the meningeal SR101 concentration (mean ± s.d.) after transcranial application to a thinned versus an intact skull. Data are representative of three mice per group and at least three independent experiments.

Extended Data Figure 5 Role of purinergic receptors and connexin hemichannels in the compression injury response.

a, Representative 25 μm maximum projections in the xy plane were captured in CX3CR1gfp/+ mice 3 h after compression injury. Two compression injuries were generated per mouse (over the left and right hemispheres). Before injury the left skull bone was treated with the indicated purinergic receptor antagonist and the right with vehicle (aCSF). When compared to the vehicle control, transcranial application of P2RY12 or P2RX4 receptor antagonists impeded the ability of microglia (green) to assume both honeycomb and jellyfish morphologies, whereas P2RY6 antagonism blocked the formation of jellyfish processes only. b, xy projections were captured in LysMgfp/+ mice 6 h after compression injury. Relative to the control group, transcranial administration of a P2RX7 antagonist markedly reduced the number of neutrophils at the injury site. c, Twenty-five micrometre xy projections were captured in CX3CR1gfp/+ mice 3 h after compression injury. Transcranial administration of the connexin hemichannel inhibitor CBX, but not the pannexin hemichannel inhibitor probenecid (image not shown), prevented the formation of jellyfish and honeycomb microglia (green) relative to vehicle-treated mice after compression injury. d, Five micrometre xz projections were captured in B6 mice 3 h after compression injury. SR101 (red) applied transcranially 3 h after injury diffused more heavily into the brain parenchyma of CBX-treated mice when compared to the vehicle control group. Data are representative of three mice per group and at least three independent experiments. See also Fig. 4.

Supplementary information

Dynamics of pathology following compression injury

Part 1: Representative B6 mice injected i.v. with Q-dots (red) were imaged 5 minutes following a normal thinned skull preparation (control; left panel) or a compression injury (compressed; right panel). Q-dot leakage into the perivascular space, visualized as a bright red accumulation around vessels (white arrowhead), is only observed after compression injury. Vascular breakdown, visualized as a diffuse red cloud erupting from a vessel (white arrow), also is only present following injury. A side view (xz) of the preceding compression injury video shows examples of perivascular Q-dot accumulation in blood vessels (white arrow) that descend from the meninges into the brain parenchyma. Skull bone is shown in blue. Part 2: Representative time lapses were captured in GFAP-GFP mice beginning 5 minutes after a normal thinned skull preparation (left panel) or a compression injury (right panel). Astrocytes (green) normally form a continuous barrier between the meninges and parenchyma referred to as the glial limitans. Following compression injury, holes (white asterisks) appear in the glial limitans due to astrocyte death. An enlarged view from the preceding time lapse (denoted with a white box) shows an astrocyte in the glial limitans dying (white arrowhead) 45 minutes following compression injury. Part 3: Representative side views (xz) of two-photon time lapses were captured in CX3CR1gfp/+ mice 5 minutes following a thinned skull preparation (control; top panel) or a compression injury (compression; bottom panel). Meningeal macrophages (green; white arrowhead) immediately below the skull (blue) are only visible in the control animal. Note how the sub-arachnoid space is no longer present following compression injury. The time lapses that follow show meningeal macrophages (green; long rod-like cells) undergoing rapid necrosis following compression injury (right panel). Note the green cloud generated as GFP leaks out of damaged meningeal macrophages, while microglia (green; ramified cells) beneath these macrophages remain intact. Blood vessels (red) were labeled by i.v. injection of Q-dots. Part 4: Representative side views (xz) of two-photon time lapses were captured in CX3CR1 mice 5 minutes following a thinned skull preparation (control; top panel) or a compression injury (compression; bottom panel). In control mice, microglia (green) never extend their processes above the glial limitans into the meningeal space. Following compression injury, however, microglial processes often extend through the glial limitans into the injured meningeal space (white arrows) as evidenced by the presence of microglial processes adjacent to the skull bone (blue). (MOV 17411 kb)

Compression injury induces a stable network of “honeycomb” microglia at the glial limitans

Part 1: A representative time lapse from a CX3CR1 mouse shows 3.5 hrs of continuous imaging through a normal uncompressed skull window. Afterward, a compression injury was induced and the same area was imaged for an additional 9 hrs. Prior to injury, note the presence of meningeal macrophages (green; long rod-like cells) and ramified microglia (also green) probing their surroundings. Meningeal macrophages die following compression injury (see Video 1), and microglia immediately retract their ramified processes and extend them toward the glial limitans, forming a vast network resembling a “honeycomb”. This network of microglial processes is entirely formed within one hour of injury. A magnified time lapse of the area highlighted by the white box shows individual microglia converting from a ramified to honeycomb morphology between 5 min and 2 hrs post-injury. Microglia processes in the video were labeled with the Filament tool in Imaris to facilitate visualization. Note at 2 hrs how microglial processes are no longer ramified, but only extend upward towards the glial limitans, forming a honeycomb network. Static top (xy) and side (xz) views of labeled microglial processes from the preceding section are shown at 5 min and 2 hrs following injury. Individual cells are shown in distinct colors. Part 2: A representative 3D projection was captured in a CX3CR1 mouse 3 hrs following compression injury. Astrocytes (red) were labeled by transcranial administration of SR101. Note that honeycomb microglia (green) surround individual astrocytes in the glial limitans. (MOV 23412 kb)

Generation and dynamics of phagocytic “jellyfish” microglia

A representative 11 hr time lapse from a CX3CR1 mouse was captured beginning 5 min post-compression injury. At 5 min, a microglial cell (green) in the center of the frame (white arrowhead) has already begun retracting its ramified processes. Afterward, the cell expands a single “jellyfish” process outward in all directions. Following complete retraction of processes at 2 hrs, the soma itself retracts into the jellyfish structure, and the amoeboid microglia becomes motile as other jellyfish microglia can be seen moving into the area. Note at 10 hrs post-compression that microglia in the field become significantly less motile and form a continuous barrier. Blood vessels are labeled red by i.v. injection of Q-dots. (MOV 27950 kb)

Three distinct microglial morphologies generated in response to compression injury

A representative time lapse showing the three distinct microglial morphologies was captured in a CX3CR1 mouse beginning 5 min following compression injury. The three distinct microglial (green) morphologies are highlighted with white boxes and magnified into individual time lapses. Box 1 shows an area with an immediate motile jellyfish response. Box 2 shows an area with microglial processes that extend and converge on a particular area, which is followed by the formation of a continuous barrier at the glial limitans similar to the jellyfish microglia in Box 2. Box 3 shows an area forming a network of honeycomb microglial. (MOV 26789 kb)

Morphological conversion of “honeycomb” into phagocytotic “jellyfish” microglia

The representative time lapse was captured in a CX3CR1 mouse beginning 5 hrs following compression injury. The time lapse contains examples (white arrowheads) of microglia (green) comprising honeycomb networks that collapse and immediately fill the holes in their network likely as a result of astrocyte cell death in the glial limitans. Note that large endosomes become visible in the resultant jellyfish microglia, a process taking approximately 5 minutes. (MOV 6868 kb)

Microglial reconstitution of the glial limitans

The representative 3D projection was captured in a CX3CR1 mouse 3 hrs following compression injury. Note that most microglia (green) in the center of this field have retracted their ramified processes and extended fine, thread-like processes toward the glial limitans that expand into large, circular structures with a jellyfish morphology. The jellyfish processes are juxtaposed with one another and form a continuous barrier along the glial limitans. Skull bone is shown in blue. (MOV 9317 kb)

Microglia form a continuous phagocytotic barrier along the glial limitans by assuming a jellyfish morphology

The representative 15 hr time lapse was captured in a CX3CR1 mouse beginning 5 min post-compression injury. Most microglia (green) in the field immediately retract their ramified processes, funneling all available cell mass into single, phagocytotic processes at the glial limitans. Microglial cell death was observed as early as 5 hrs post-injury (white arrowhead) and began to accelerate by 9 hrs (white arrowhead). Evidence of edema (a wave of swelling in the tissue) was observed at 11 hrs, which was followed shortly thereafter by deterioration in the overall health of the microglial barrier structure at Athe glial limitans. (MOV 16726 kb)

Myelomonocytic cell recruitment following compression injury

Part 1: The representative 12 hr time lapse from a LysM mouse was captured beginning 5 min post-compression injury. Neutrophils (bright yellowish green) begin to swarm the area of injury at 1 hr following compression, both from adjacent areas of the meninges and by vascular extravasation (white arrowhead). Monocytes (light green) arrive later in the response (~7 hrs). Vessels (red) labeled by i.v. injection of Q-dots. Part 2: Representative xy and xz time lapses were captured in a LysM mouse beginning 1 hr after compression injury. Thirty minutes after injury, dead cells were labeled by transcranial application of PI (red). Neutrophils localize almost exclusively to the meninges, which contains an abundance of dead cells. Skull bone is shown in blue. An enlarged view following a single neutrophil from the preceding video reveals how neutrophils interact with multiple dead cells over the course of many hours. (MOV 26066 kb)

Purinergic receptor and connexin hemichannel signaling mediate the innate immune response to compression injury

Part 1: Representative time lapses were captured in a CX3CR1 mice pretreated transcranially with P2Y12, P2Y6, P2X4, or P2X7 receptor antagonists and imaged beginning 5 min post-compression injury. Note that P2Y12, P2Y6, and P2X4 inhibition prevents microglia (green) from transforming into jellyfish at the glial limitans. P2Y12 and P2X4 antagonism also reduces the formation of honeycomb networks, whereas inhibition of P2X7 has no impact on either microglial response to compression injury. In general, microglial retained more ramified processes following injury when inhibited with P2Y12, P2Y6, or P2X4 antagonists compared to P2X7. Time lapses in the next section were captured in LysM mice pretreated with a P2X7 receptor antagonist or a vehicle control. P2X7 antagonism almost entirely eliminates the neutrophil (green) response to compression injury, even 12 hrs following injury. Part 2: Representative time lapses were captured beginning 5 min post-compression injury in CX3CR1 mice pretreated transcranially with a connexin hemichannel inhibitor (CBX) or a vehicle control. Inhibition of astrocytic ATP-dependent ATP release by CBX prevents microglia (green) from generating jellyfish and honeycomb responses. Microglia instead maintain ramified processes along with a single, small, ill-defined circular process at the glial limitans (white arrowhead). The structure resembles an aborted attempt to transform into the jellyfish morphology. A typical jellyfish reaction is shown in the vehicle treated control time lapse for comparative purposes. (MOV 20459 kb)

Glutathione reduces inflammation and glial limitans damage following compression injury

Part 1: Representative time lapses were captured in CX3CR1 mice pretreated transcranially with the ROS scavenger, glutathione, or a vehicle control. The standard compression injury was followed by physically cracking the skull (blue) with the microsurgical blade to promote a more severe inflammatory reaction. Note that meningeal macrophages (green; thin rod-like cells) and microglia near the fractured skull (green; ramified) only survive in glutathione treated animal. In addition, microglia in the glutathione-treated time lapse retain ramified processes and do not appear to sense the massive injury above them. This contrasts with the vehicle control group in which microglia immediately retract all processes and form single motile jellyfish structures. Magnified time lapses in vehicle versus glutathione treated mice following compression injury show the stark difference between microglial. Note that microglia in the vehicle control time lapse rapidly generate jellyfish structures, whereas those in the glutathione time lapse remain ramified and in a “naïve” state. Part 2: Representative xz time lapses were captured in CX3CR1 mice pretreated transcranially with vehicle or glutathione before compression injury. Note that within 30 minutes of injury, the subarachnoid space refills, and the compressed skull (blue) bows upward. This was never observed in the vehicle treated control group. Microglial / macrophages are shown in green. Part 3: Representative time lapses were captured in LysMgfp/+ mice pretreated transcranially with vehicle or glutathione before compression injury. Glutathione pretreatment significantly reduces the number of neutrophils (green) that migrate into the injured area relative to the vehicle control. Neutrophils occasionally become sequestered in blood vessels of glutathione treated mice, but do not extravasate and swarm the meninges. (MOV 18609 kb)

PowerPoint slides

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Roth, T., Nayak, D., Atanasijevic, T. et al. Transcranial amelioration of inflammation and cell death after brain injury. Nature 505, 223–228 (2014).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Sign up for the Nature Briefing newsletter for a daily update on COVID-19 science.
Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing