J Neurotrauma. 2015 Jul 15;32(14):1101-8. doi: 10.1089/neu.2014.3611.
Interferon-Stimulated Gene 15 Upregulation Precedes the Development of Blood-Brain Barrier Disruption and Cerebral Edema after Traumatic Brain Injury in Young Mice.
Janet L. Rossi,1–3 Tracey Todd,4 Zachary Daniels,4 Nicolas G. Bazan,1 and Ludmila Belayev1,4
1Neuroscience Center of Excellence, 2Children’s Hospital of New Orleans, 3Department of Pediatrics, 4Department of Neurosurgery, Louisiana State, University Health Sciences Center, New Orleans, Louisiana.
Recent studies show that myosin light chain kinase (MLCK) plays a pivotal role in development of cerebral edema, a known complication following traumatic brain injury (TBI) in children and a contributing factor to worsened neurologic recovery. Interferon-stimulated gene 15 (ISG15) is upregulated after cerebral ischemia and is neuroprotective. The significant role of ISG15 after TBI has not been studied. Postnatal Day (PND) 21 and PND24 mice were subjected to lateral closed-skull injury with impact depth of 2.0 or 2.25 mm. Behavior was examined at 7 d using two-object novel recognition and Wire Hang tests. Mice were sacrificed at 6 h, 12 h, 24 h, 48 h, 72 h, and 7 d. ISG15 and MLCK were analyzed by Western blot and immunohistochemistry, blood–brain barrier (BBB) disruption with Evans Blue (EB), and cerebral edema with wet/dry weights. EB extravasation and edema peaked at 72 h in both ages. PND21 mice had more severe neurological deficits, compared with PND24 mice. PND24 mice showed peak ISG15 expression at 6 h, and PND21 mice at 72 h. MLCK peaked in both age groups at 12 h and co-localized with ISG15 on immunohistochemistry and co-immunoprecipitation. These studies provide evidence, ISG15 is elevated following TBI in mice, preceding MLCK elevation, development of BBB disruption, and cerebral edema.
KEYWORDS: ISG15; MLCK; behavior; cerebral edema; traumatic brain injury; young mice
It is well understood that children especially those less than 4 years of age have a worse outcome following TBI (traumatic brain injury). This has been established through clinical studies incorporating multiple tools of measurement, including physical, cognitive, emotional and socioeconomic factors1. One of the major reasons for this worse outcome in young children stems from the worsening secondary injury after the initial trauma. The secondary injury results in but is not limited to ischemia2, ion alterations such as calcium influx3, receptor activation4, mitochondria responses5, oxidative stress6, reactive oxygen specie generation7, edema and axonal swelling8. ISG15 (interferon Stimulating Gene 15) is an ubiquitin superfamily responsible for post-translationally modifying proteins. ISG15 can be protective as represented by the protective effects in a mouse stroke model9 and disruptive as represented in the invasion of breast cancer10. We hypothesize the reason for worse secondary injury stems from the reduced level of protective lipids and proteins. Here we hypothesize ISG15 is protective to the blood brain barrier following TBI in young mice. We compared the levels of ISG15 and ISGylation-proteins modified by ISG5, between PND24 mice, analogous to 6-8y children and PND21 mice, analogous to 2-3y children.
FIGURE 1. PND21 mice western blot 1a. showing decreased ISGylation through 6h 1b. PND21 a cyclic change of free ISG15 over timepoints <2m through 7d. PND24 mice western blot 1c. showing increased ISGylation from t<2m through24h 1d. PND24 mice showing increased free ISG15 over the 6h with cyclic changes in free ISG15 over 7d. * 2.25mm or 2mm impact depth compared to sham, # 2.25mm compared to 2mm. * or # p<0.01, ** or ## p<0.001, *** or ### p<0.0001. White bar-Sham, Red bar-2mm TBI, Blue bar-2.25mm TBI.
ISG15 and ISGylation
Both PND21 (Fig. 1a) and PND24 (Fig. 1c) brains show oscillation in free ISG15 and ISGylation. However PND24 ISGylation is greater in the early timepoints and the PND21 ISGylation is greatest in the later timepoints. The free ISG15 is up initially following TBI at T<2m in the 2.25mm PND24 mice (4.00 vs 10.17) (Fig. 1d) and returns to baseline at 5m following TBI and does not rise above baseline until 30m in the 2.25mm depth (3.88 vs 15.50) and between depths (6.33 vs 15.50) and continues to rise to the peak at 6h elevated in both the 2mm (3.33 vs 11.71) and 2.25mm (3.33 vs 15.50) compared to sham and between depths (5.50 vs 15.50). ISG15 decreases from that point to baseline at 7days. PND21 free ISG15 (Fig. 1b) decreases in the 2.25mm at T<2m below the sham level (3.83 vs -14.83) and continues to decrease to a low at 30m sham vs 2.25 (3.5 vs 15.5), sham vs 2.00mm (3.5 vs 12.0) and 2.00 vs 2.25mm (5.50 vs 15.50) with return to baseline only at 15m. It then rises above baseline in the 2.25mm depth at 4h (3.667 vs 15.50) returning to baseline until again at 48h in both 2mm (2.33 vs -10.33) and 2.25mm ((2.33 vs – 12.67) decreasing, then back up to peak expression at 7days in the 2.25mm depth (2.50 vs 11.5).
ISG15 shows an oscillatory pattern in both PND21 and PND24 mice11,12. The oscillations in PND21 mice are greater and more frequent than the PND24 mice. Free ISG15 is responsible for the ISGylation of other proteins. The amount of free ISG15 is greater in PND24 mice than PND21 mice. This allows for ISGylation to begin sooner in PND24 mice compared to PND21 mice. The decrease in the free ISG15 from T<2m in PND21 mice suggests this is a developmental protein that becomes upregulated at baseline with age. It also raises the question of what does all this ISGylation do to the stability of the BBB and the neurological signaling at baseline and following injury.
In an effort to begin to answer some of these questions we compared PND21 male C57bl6 WT mice to the PND21 male ISG15-/- mice on a C57bl6 background and the disruption of the BBB. We evaluated this using three different molecular weight dyes and behavior studies. We chose these dyes based on size to evaluate the different pathways molecules traverse the BBB. Texas Red (TR) is a 4000da dextran requiring a glucose transporter to cross the BBB13-15. Fluorescein sodium salt (FITC) is less than 500da which can cross the BBB through several sodium transporters and channels16, and the Evans blue (EB) as bound to albumin 69,000da which requires the disruption of the junctional proteins to extravasate into the brain parenchyma17,18.
Method: In an effort to reduce the number of mice used and the pain of injection on the mice we injected all three dyes at one time. After under going a closed head traumatic brain injury, one hour prior to sacrifice, we injected the mice i.p Texas Red Dextran 4000 da (TR), FITC sodium salt 500 da (FITC) and Evans blue 69,000 da (EB). TR and FITC were reconstituted with Phosphate Buffer Solution (PBS) into 10mg/ml concentration, and EB dye was reconstituted to a 2% concentration in dH20. We combined all three dyes into one vial at equal volumes, vortexing aggressively. We injected 30 μl/g, i.p. to sham and 2.25mm TBI ISG15-/- and WT mice. After sacrifice, at time of processing, brain tissue was excised from below the impact site, or like area in shams, weighed, sonicated in formamide (Sigma Aldrich) and then incubated overnight at 37°C. Samples were then centrifuged and the supernatant aliquoted to a 96-well plate. Fluorescence of TR was measured using an excitation (ex) at 595 and emission (em) at 600, FITC at 460 ex and 515 em, and EB at 620 ex and 675 em with a Spectramax M5 microplate reader using a strict cutoff for each (Molecular Devices). Fluorescence data was compared to a standard curve of decreasing values of ng/ml for each dye flourescence and then analyzed using Softmax Pro 5.4 (Analytical Technologies.
FIGURE 2. Blue dot represents impact point. Black outline represents area of brain taken for analysis (cartoon). 2b.Triple Dye – Brain examples of PND21 WT mice at 72h timepoint, 2.25mm impact depth. 2c. Triple Dye –brain example of PND21 ISG15 -/- mice at 72 h timepoint 2.25mm impact depth. 2d. TR PND21 WT and ISG15-/- mice sham (White), Wt TBI (Black) and ISG15 -/- TBI (Gray). 2e. FITC PND21 WT and ISG15-/- mice sham (White), Wt TBI (Black) and ISG15 -/- TBI (Gray). 2e. FITC PND21 WT and ISG15-/- mice sham (White), Wt TBI (Black) and ISG15 -/- TBI (Gray). * p<0.01, ** p<0.001, *** p<0.0001, **** p<0.00001.
Evaluation of the blood brain barrier
Fig. 2b is a PND24 WT mouse brain 2.25mm depth 72h timepoint Fig 2c is a PND24 ISG15-/- mouse brain 2.25mm 72h timepoint Fig. 2d shows TR in PND21 mice brains decreases at 72h in WT 2.25mm compared to sham brains (3.50 vs 16.83). In the ISG15-/- brain the TR showed no difference between TBI and Sham but overall the same TR as the sham WT brains. Fig. 2e shows FITC in PND21 mice brains increases at 72h in WT 2.25mm depth compared to sham (3.50 vs 12.67). In the ISG15-/- mice FITC increases slightly but is not significant. The total amount of FITC in sham mice was significantly higher in the ISG15-/- mice compared to WT mice (3.50 vs 16.17).
Fig. 2f shows EB in PND21 increases at 72h in WT 2.25mm depth compared to sham (3.50 vs 21.50). In the ISG15 -/- brains the EB shows no significance between sham and 2.25mm TBI. There is no difference between sham EB brains.
We have showed ISG15 formed a protein-protein interaction with GLUT111. This could explain the reason for the decrease in TR transport into the brain of the TBI PND21 WT mice compared to the sham. However, it does not explain the constant level of TR in the ISG15-/- mice. We recently also showed that phosphorylated GLUT1 correlated to ISG15 and to a more constant expression of glucose in the brains in PND 24 mice following TBI, compared to PND21, which showed decreased levels of ISG15 and ISGylation and decreased p-GLUT1 with increased glucose11. This could potentially explain the constant level of TR ISG15 in the sham and TBI brains of the ISG15-/- mice. FITC and EB both showed and increase in the WT 2.25mm depth over the sham mice but showed no difference between the sham and ISG15 -/- 2.25mm depth. The shams of the ISG15-/- mice show and increase over the WT shams in both FITC and EB, which implies the BBB has increased permeability at baseline in ISG15 -/- mice. The FITC is a small molecule that because it is a sodium salt has many pathways across the BBB16,19 and would suggest the BBB in ISG15-/- mice are leaker than WT. EB bound to albumin requires disruption of the cellular junctions to penetrate the BBB17,20. The increased level of EB in the sham ISG15-/- mice would suggest the junctions are less tight than the shams of the WT mice. The difference in the sham vs 2.25mm ISG15-/- mice may actually be a result of the skull thickness. The ISG15-/- mice have thicker and potentially more ossified bones, which would prevent the force of the impact from propagating as diffusely as it does in the WT mice of the same age. It does not however explain the behavior results.
We examined WT and ISG15-/- mice in both the wire hang and 2-Object Novel Recognition to evaluate motor control and strength and recognition memory following TBI.
FIGURE 3. 2-Object Novel Recognition: Top Fig. WT: sham (White), TBI (Black). Bottom Fig. ISG15-/-: sham (White) TBI (Gray). Wire Hang: WT and ISG15-/-: Sham (White), WT TBI (Black) and ISG15-/- TBI (Gray). * p<0.01 ** p<0.001 ***
2-Object Novel Recognition
The PND21 sham WT mice show preference for the new object over the old object (34.40 vs 72.90). The PND21 2.25mm WT mice show no preference between the old and new objects. The PND21 ISG15-/- shams show no preference between the old and new objects and a significant decrease in the preference for the new object compared to the old object (28.32 vs 17.13).
WT mice show a decrease in the latency to fall between sham and 2.25mm (20.33 vs 3.50). The ISG15-/- mice show no difference between the sham and 2.25mm depth in latency to fall. The average hang time from sham WT and sham ISG15 -/- shows no significant difference but is trending toward significance (WT 20.33 vs 12.17).
ISG15-/- mice have a shorter latency to fall in both the sham and 2.25mm TBI mice compared to WT sham mice, but no difference between the sham and TBI ISG15-/- mice. Also the 2-Object novel recognition test the ISG15 -/- mice appear to be unable to differentiate old from new objects in the sham group and are more interested in the old object than the new object in the 2.25mm TBI. This inability for the ISG15-/- mice to differentiate between old and new objects means there may be role for ISG15 in learning and memory as well as muscle strength and coordination.
This study is limited by the small number of animals, n=6 for each group, and the single time point following TBI of 72h. However this study sheds light on many areas that need further investigation. The answers to these questions may help understand not only the signaling of the brain following TBI and other neurological diseases, but also may shed light on the learning and memory, motor strength and coordination, and possibly psychiatric disorders.
1 Rosema, S. et al. Agreement on and predictors of long-term psychosocial development 16 years post-childhood traumatic brain injury. J Neurotrauma 31, 899-905 (2014).
2 Strbian, D. et al. The Blood– Brain Barrier is continuously open for several weeks following transient focal cerebral ischemia. Neurosci 153, 175-181 (2008).
3 Weber, J. Altered calcium signaling following traumatic brain injury. Front Pharmacol doi: 10.3389/fphar.2012.00060, 1-15 (2012).
4 Baeten, K. & Akassoglou, K. Extracellular Matrix and Matrix Receptors in Blood-Brain Barrier Formation and Stroke. Dev Neurobiol 7, 1018-1039 (2011).
5 Fiskum, G. Mitochondrial Participation in Ischémie and Traumatic Neural Cell Death. J Neurotrauma 17, 843-855 (2000).
6 Bayir, H. et al. Assessment of Antioxidant Reserves and Oxidative Stress in Cerebrospinal Fluid after Severe Traumatic Brain Injury in Infants and Children. Pediatr Res 51, 571-578 (2002).
7 Adibhatla, R. & Hatcher, J. Phospholipse A2, reactive oxygen species, ad lipid peroxidation in cerebral ischemia. Free Rad Biol Med 40, 367-387 (2006).
8 Inglese, M. et al. Diffuse axonal injury in mild traumatic brain injury: a diffusion tensor imaging study. J Neurosurg 103, 298–303 (2005).
9 Nakka, V. et al. Increased cerebral protein ISGylation after focal ischemia is neuroprotective. J Cerebr Blood F Met 31, 2375–2384 (2011).
10 Desai, S. et al. ISG15 disrupts cytoskeletal architecture and promotes motility in human breast cancer cells. Ex Biol Med 237, 38-49 (2012).
11 Rossi, J., Todd, T., Belayev, L. & Bazan, N. ISGylation Downregulation and Increased Blood-Brain Barrier Permeability at the Onset of Traumatic Brain Injury (TBI) in a Mouse Model of Pediatric TBI. submitted for review.
12 Rossi, J., Todd, T., Daniels, Z., Bazan, N. & Belayev, L. Interferon Stimulated Gene 15 upregulation precedes the development of blood brain barrier disruption and cerebral edema after traumatic brain injury in young mice. J Neurotrauma 32, 1101–1108 (2015).
13 Devraj, K. et al. GLUT-1 glucose transporters in the blood-brain barrier: Differential phosphorylation. J Neurosci Res 89, 1-24 (2011).
14 Morgello, S., Uson, R., Schwartz, E. & Haber, R. The human blood-brain barrier glucose transporter (GLUT1) is a glucose transporter of gray matter astrocytes. Glia 14, 43-54 (1995).
15 Salvador, E., Burek, M. & Förster, C. Stretch and/or oxygen glucose deprivation (OGD) in an invitro traumatic brain injury (TBI) model induces calcium alteration andi nflammatory cascade. Front Cell Neurosci 9, 1-14 (2015).
16 Betz, A. Transport of ions across the blood-brain barrier. Fed Proc 45, 2050-2054 (1986).
17 Dudek, S. & Garcia, J. Cytoskeletal regulation of pulmonary vascular permeability. J Appl Physiol 91, 1487-1500 (2001).
18 Rossi, J., Todd, T., Bazan, N. & Belayev, L. Inhibition of Myosin Light-Chain Kinase Attenuates Cerebral Edema After Traumatic Brain Injury in Postnatal Mice. J Neurotrauma 30, 1-8 (2013).
19 Goldstein, G., Betz, A., Bowman, P. & Dorovini-Zis, K. In vitro studies of the blood-brain barrier using isolated brain capillaries and cultured endothelial cells. Ann A Y Acad Sci 481, 202-213 (1986).
20 Rossi, J. et al. Albumin causes increased myosin light chain kinase expression in astrocytes via p38 mitogen-activated protein kinase. J Neurosci Res 89, 852-861 (2011).