Skip to main content
NCBI home page
Journal of Cerebral Blood Flow & Metabolism logoLink to Journal of Cerebral Blood Flow & Metabolism
. 2011 Mar 30;31(9):1886–1896. doi: 10.1038/jcbfm.2011.31

CSF Bcl-2 and cytochrome C temporal profiles in outcome prediction for adults with severe TBI

Amy K Wagner 1,2,*, Krutika B Amin 1, Christian Niyonkuru 1, Brett A Postal 1, Emily H McCullough 1, Haishin Ozawa 1, C Edward Dixon 1,2,3, Hulya Bayir 2,4, Robert S Clark 2,4, Patrick M Kochanek 2,4, Anthony Fabio 1,5
PMCID: PMC3185877  PMID: 21448217

Abstract

The biochemical cascades associated with cell death after traumatic brain injury (TBI) involve both pro-survival and pro-apoptotic proteins. We hypothesized that elevated cerebrospinal fluid (CSF) Bcl-2 and cytochrome C (CytoC) levels over time would reflect cellular injury response and predict long-term outcomes after TBI. Cerebrospinal fluid Bcl-2 and CytoC levels were measured for 6 days after injury for adults with severe TBI (N=76 subjects; N=277 samples). Group-based trajectory analysis was used to generate distinct temporal biomarker profiles that were compared with Glasgow Outcome Scale (GOS) and Disability Rating Scale (DRS) scores at 6 and 12 months after TBI. Subjects with persistently elevated temporal Bcl-2 and CytoC profiles compared with healthy controls had the worst outcomes at 6 and 12 months (P⩽0.027). Those with CytoC profiles near controls had better long-term outcomes, and those with declining CytoC levels over time had intermediate outcomes. Subjects with Bcl-2 profiles that remained near controls had better outcomes than those with consistently elevated Bcl-2 profiles. However, subjects with Bcl-2 values that started near controls and steadily rose over time had 100% good outcomes by 12 months after TBI. These results show the prognostic value of Bcl-2 and CytoC profiles and suggest a dynamic apoptotic and pro-survival response to TBI.

Keywords: Bcl-2, biomarker, cerebral spinal fluid, cytochrome C, outcome, traumatic brain injury

Introduction

Secondary injury has a prominent role in pathophysiology associated with traumatic brain injury (TBI). Cellular injury results from secondary injury pathways that involve inflammation, excitotoxicity, and ischemia and result in oxidative stress (Love, 1999). These secondary injury pathways can initiate cell death pathways including apoptosis (Zou et al, 1997). Apoptosis after TBI has been confirmed by multiple experimental studies in vivo (Clark et al, 1997b; Colicos and Dash, 1996; Rink et al, 1995). Within the intrinsic (mitochondrial mediated) pathway associated with caspase-dependent apoptosis, stress on cellular organelles leads to initiation of apoptosis via mitochondrial permeability transition pore formation, and subsequent release of cytochrome C (CytoC). Pro-apoptotic molecules (e.g., from some Bcl-2 family proteins) translocate to mitochondria, resulting in mitochondrial membrane permeabilization. This, in turn, provides a route for release of intermembrane space proteins like CytoC into the cytosol. Cytosolic CytoC promotes ‘apoptosome' formation, a molecular platform for caspase-9 activation. In turn, active caspase-9 catalyzes the proteolytic activation of effector caspases including caspase-3 (Boehning et al, 2003). Previous work shows CytoC is released into the extracellular medium from ‘suicidal' cells and represents a specific apoptosis biomarker in humans (Renz et al, 2001).

The Bcl-2 family proteins are important apoptosis modulators, of which Bcl-2 is the founding member. They comprise both pro-apoptotic and pro-survival proteins that regulate mitochondrial permeability through transition pore formation (Clark et al, 2005; Liou et al, 2003; Zhang et al, 2005). Bcl-2 is increased in injured neurons and reduces programmed cell death in rats (Clark et al, 1997a) and mice (Raghupathi et al, 1998) by inhibiting release of mitochondrial proteins (Zhang et al, 2005). Bcl-2 overexpression in experimental models can be neuroprotective and reduce cortical neuron loss (Nakamura et al, 1999; Raghupathi et al, 1998). However, tissue sparing is not necessarily associated with improved behavioral outcomes (Tehranian et al, 2006).

Clinical studies have detected both CytoC and Bcl-2 in human TBI. A pediatric TBI study showed relatively higher cerebrospinal fluid (CSF) levels of CytoC in females and in those with inflicted injuries (Satchell et al, 2005). Another study showed time-dependent increases in CytoC and activated caspase-9 levels in nine adults with TBI. This case series showed that activated caspase-9, but not CytoC, was correlated to outcome based on the Glasgow Outcome Scale (GOS) (Darwish and Amiridze, 2010). Some clinical studies have reported increased CSF Bcl-2 levels post-TBI (Clark et al, 1999, 2000; Uzan et al, 2006). A study comparing CSF Bcl-2 in 14-TBI subjects showed no correlation with other markers of apoptosis (Fas and caspase-3) (Uzan et al, 2006). In pediatric patients with TBI, higher CSF Bcl-2 was associated with increased survival (Clark et al, 2000). Finally, genetic variation in the Bcl-2 gene has been linked to adult TBI outcomes (Hoh et al, 2010).

Apoptosis in TBI can occur in a delayed manner. Apoptotic bodies are observed between 24 hours and 1 week after TBI in experimental models (Conti et al, 1998; Sullivan et al, 2002), and peak levels of apoptosis occur 24 to 72 hours after injury (Zhang et al, 2005). This time-dependent progression of apoptosis may provide a therapeutic window from which to prevent further injury (Liou et al, 2003; Robertson et al, 2006). These studies also suggest that time-dependent profiles in apoptosis-related biomarkers may be informative about secondary injury and outcomes. Studies to date do not provide adequate insight as to whether Bcl-2 or CytoC have value as prognostic indicators of TBI outcome. Also, the capacity of temporal Bcl-2 and CytoC profiles to inform TBI pathophysiology has not been addressed. Reports from our center (Berger et al, 2010; Salonia et al, 2010; Ozawa et al, 2009) show that time-dependent modeling strategies can effectively link biomarker profiles and outcome.

Therefore, the purpose of this study was to evaluate temporal profiles for CSF Bcl-2 and CytoC in a population of adults after severe TBI in relation to demographics, injury severity, and outcome. CytoC can serve as a specific apoptosis biomarker in humans that reflects cell death (Renz et al, 2001), and CSF Bcl-2 levels may reflect the amount of ongoing cell death as well as some attempt to counter apoptosis through increased translation of this pro-survival protein. Our results show that there are patient subgroups with unique profiles for both CytoC and Bcl-2. These profiles can predict long-term outcome, and capture dynamic changes to apoptosis and pro-survival information after severe TBI.

Materials and methods

Subjects

The study was approved by the University of Pittsburgh Institutional Review Board. Patients with severe TBI (n=76) were admitted to our level 1 trauma center. A diagnosis of severe TBI required a Glasgow Coma Scale (GCS) score ⩽8 and a positive cranial CT (computerized tomography) scan. On admission, patients received aggressive medical treatment as outlined by The Guidelines for the Management of Severe Head Injury (Bullock et al, 1996). In addition to severe TBI, enrollment in the study depended on the following inclusion criteria: (1) patient age between 16 and 65 years, (2) placement of an external ventricular drain for intracranial pressure monitoring, (3) availability of at least two CSF samples for analysis, and (4) signed consent from next-of-kin.

As a comparison group, healthy adult control subjects (n=10) were also enrolled for biomarker analysis. Control subjects' CSF was obtained via lumbar puncture for research purposes and not as a part of another clinical work-up. Criteria for control subject enrollment included age between 18 and 70 years and no current or past history of brain injury, neurologic disease, or bleeding disorder. Patients were excluded if pregnant, taking oral contraceptives or hormone therapy, or had any history of reproductive or endocrine disorder.

Cerebrospinal Fluid Sample Collection and Measurements

Treatment for severe TBI subjects included initial placement of an external ventricular drain catheter for intracranial pressure monitoring and treatment of increased intracranial pressure by CSF drainage. The external ventricular drain catheter was inserted as part of standard of care, and CSF was drained passively. Cerebrospinal fluid was collected for up to 6 days after initial injury, and CSF was stored at 4°C until processing. Cerebrospinal fluid samples were then centrifuged at 3,000 r.p.m. for 10 minutes and aliquoted. Samples were stored at −80°C until batch analysis. Enzyme-linked immunosorbent assay (ELISA) was used to measure Bcl-2 (EMD Biosciences, Inc., Cat No. QIA23 © 2005) and CytoC levels (Assay Designs, Cat. No. 900-141 © 2004) in CSF (Clark et al, 2000; Satchell et al, 2005). Enzyme-linked immunosorbent assays were specific for human biospecimens and performed following the manufacturer's instructions. The inter-ELISA and intra-ELISA coefficient of variation was <10%. Cerebrospinal fluid samples were run in duplicate, with the average value for each sample used in analysis. Measurements from both healthy control and TBI CSF were within the range of detection using manufacturer standards, a linear calibration curve, and sample dilution as appropriate. Cerebrospinal fluid Bcl-2 and CytoC concentrations were measured daily for up to 6 days after injury (Bcl-2 n=255; CytoC n=271 measurements made from n=277 samples).

Demographic and Clinical Injury Variables

Independent variables included injury type, sex, age, initial GCS, injury severity score, hypothermia treatment, the number of acute care complications, mechanism of injury, hospital length of stay (LOS), and discharge disposition. Information on injury type was based on clinical head CT radiology reports. The GCS was taken within 8 hours of injury to limit the influence of alcohol, sedatives, or paralytics. Age was grouped by decile for analysis. In addition to the number of complications, acute complications (e.g., sepsis) were abstracted from the medical record and grouped into 11 categories for analysis. Categories included neurologic/endocrine, pulmonary, cardiac, nonspecific infectious disease, hematologic, musculoskeletal, renal, integument, gastrointestinal, multisystem organ failure, and other. Neurologic/endocrine complications included stroke, anoxic brain injury, seizure, infection of the central nervous system (CNS), syndrome of inappropriate antidiuretic hormone, and diabetes insipidous. A small number of subjects were enrolled in a hypothermia clinical trial (n=7). Of these, five subjects received hypothermia and two subjects received normothermia. Subjects not in this clinical trial were treated to maintain normothermia.

Outcome Variables

The GOS score and Disability Rating Scale (DRS) were the primary outcome measures. Glasgow Outcome Scale is a frequently used measurement developed by Jennett and Bond (1975). Glasgow Outcome Scale scores were assigned to TBI subjects' at 6 and 12 months after injury. Patients were assigned good recovery (5), moderate disability (4), severe disability (3), persistent vegetative state (2), or death (1). For the purposes of this study, GOS categories were collapsed into 1 versus 2/3 versus 4/5. Mortality during acute care hospitalization and mortality after acute care discharge was recorded after reviewing medical records and the Social Security Death Index (http://ssdi.rootsweb.ancestry.com/). The DRS score, developed by Rappaport et al (1982), was also assigned at 6 and 12 months after injury, and patients were assigned a DRS of disability (0) to dead (30). For this scale, subjects were analyzed in three groups based on scores: (1) 0 to 3 with partial to no disability, (2) 4 to 14 with moderate or severe disability and (3) 15 to 30 with extremely severe disability, vegetative state, or dead. In all, 67 and 60 subjects had complete data for GOS at 6 and 12 months, respectively, while 66 and 56 subjects had complete data for DRS at 6 and 12 months.

Statistical Analysis

Statistical analyses were performed using SAS version 9.2 (Cary, NC, USA) and SPSS version 17.0 (Chicago, IL, USA). Summary statistics, including mean values, s.e.m., and median values were generated for continuous variables. Frequencies and percentages were determined for categorical variables. Data were checked for errors and normality assessed for all continuous variables using the Kolmogorov–Smirnov test. Bcl-2 and CytoC levels were graphed by day postinjury and trajectory group membership. Group differences in daily biomarker values, ratios, and other continuous data were compared using the Wilcoxon Rank Sum and Kruskal–Wallis tests. Trajectory group differences with regard to categorical data were assessed using χ2 analysis with Fisher's exact test when appropriate.

Group based trajectory analysis (TRAJ) was used to explore biomarker levels over time in CSF using PROC TRAJ Macro (Jones et al, 2001) available in SAS software (version 9.2 of the SAS system for windows. Copyright © (2002 to 2008) SAS Institute Inc.). In general, the TRAJ algorithm identifies clusters of individuals following a similar progression of some measure over time. The TRAJ procedure determines trends in longitudinally collected data by identifying trajectory groups on a likelihood basis. This method uses a probability function to discern a set of trajectories that closely resemble one another. TRAJ also assumes the existence of latent subpopulations and leverages repeated subject sampling to relate temporal patterns. We have used this approach in previous studies assessing the ability of temporal biomarker patterns to predict TBI outcome (Berger et al, 2010; Salonia et al, 2010; Ozawa et al, 2009). Bcl-2 TRAJ groups were formulated using a cohort of 70 patients and CytoC TRAJ used a cohort of 73 patients.

Daily biomarker levels were not normally distributed due to the high variance of the sample set, a lack of an upper bound for maximum values, and clusters of values around the detection limits. As such, multiple modes of data transformation were explored to better approximate a normal distribution. After examining multiple transformation methods, biomarker data for each postinjury day were ranked before deriving TRAJ groups for the population. The probability distribution of the data was determined using a censored normal approach, which is typically applied when there is a minimal detectable limit for the data. The number of TRAJ groups for each biomarker was determined by evaluating the BIC (Bayesian information criterion), AIC (Akaike information criterion), and by using clinical knowledge about each biomarker. Model diagnostics, like average group posterior probability, odds of correct classification, and group membership comparisons, were used to assess model accuracy. The average posterior probability in our final models for each group and for each biomarker ranged from 0.88 to 0.96, well above 0.7 usually recommended for this type of analysis (Nagin, 2005).

TRAJ group assignments for each subject and biomarker were compared with demographic and injury variables. Bivariate analyses were performed using the Mann–Whitney U, Kruskal–Wallis tests, or χ2 analysis to determine TRAJ group associations with other variables like age, gender, injury type, injury severity, and outcome. TRAJ group assignments were linked with actual biomarker concentrations for each subject over time to compare the shape of the concentration data over time with the shape of the time trends formed using the TRAJ procedure. Differences in concentration for each group were assessed at each point in time, and TRAJ group average daily values were compared with values observed in healthy controls. After TRAJ model selection and bivariate analyses, Bcl-2 TRAJ group findings suggested that the two groups having consistently high Bcl-2 levels were similar. As such, these groups were combined for bivariate analysis (see Figures 2A and B). Preliminary analyses (not shown) comparing TRAJ Group membership to weekly means for Bcl-2 and CytoC showed TRAJ groups as better outcome predictors.

Multivariate ordinal regression models were used to evaluate how biomarker TRAJ groups, and other clinical variables, affected GOS and DRS scores at 6 and 12 months. To simplify interpretation, subjects in the riser Bcl-2 TRAJ group 1 were combined with the low Bcl-2 group for multivariate analysis. Also the high CytoC TRAJ group 3 was combined with the decliner group in the multivariate regression models. Clinical and demographic variables having a P value ⩽0.2 in bivariate analyses when compared with outcome at 6 and 12 months were first tested in multivariate ordinal regression models. Bcl-2 and CytoC were then added in the logistic regression models, and we evaluated their added outcome prediction value using the likelihood ratio test for nested models (Gouriéroux et al, 1982). The odds of better outcome were modeled for each outcome.

Results

Description of Population

In our TBI cohort (n=76), 20.0% of the subjects were women, the average patient age was 35±1.54 years, and the median GCS score was 6. The population had a median of two complications, and 22% had at least one neurologic/endocrine complication. There was a 24% acute care mortality rate. The mean hospital LOS was 23±1.5 days. In all, 54% sustained their injuries through automobile or motorcycle collisions, and 18% sustained their TBI through jump or fall. In all, 51% of the population was discharged to an acute rehabilitation facility, and 20% discharged to a skilled nursing or long-term acute care facility. The median for 6 and 12 months GOS scores was 3. The median 6 and 12 month DRS scores were 8 and 7.5, respectively. In the control population (n=10), 40% were women, and the average patient age was 33±5 years.

Description of Daily Cerebrospinal Fluid Bcl-2 and CytoC Biomarker Levels

Figure 1A shows Bcl-2 levels were elevated in TBI patients compared with controls (control levels=5.82±0.34 U/mL). Bcl-2 concentrations were elevated for days 1, 2, and 4 (P=0.041 all comparisons and trended toward significance on day 3 (P=0.051). Figure 1B shows daily CytoC concentrations also were elevated in TBI subjects compared with controls (control levels=1,298.74±152.33 pg/mL). CytoC levels were higher than controls for days 1 to 4 (P⩽0.043 all comparisons) and trended toward significance on day 0 (P=0.074). CytoC levels peaked 24 hours after injury (day 1=4,561.84±1,011.16 pg/mL) and were higher than day 0 (P=0.042) and days 3 to 5 (P⩽0.019 all comparisons).

Figure 1.

Figure 1

Open in a new tab

Daily biomarker levels over the first 6 days after severe traumatic brain injury (TBI) compared with healthy controls. (A) Daily mean cerebrospinal fluid (CSF) Bcl-2 levels are shown. Bcl-2 levels are significantly higher than controls on days 1, 2, and 4 (P<0.05 all comparisons). (B) Daily mean CSF cytochrome C levels by day. Cytochrome C levels are maximal at day 1 post-TBI, and levels are significantly higher than healthy controls on days 1 to 4 (P<0.05 all comparisons).

Trajectory Profile Development and Characterization

Four statistically distinct temporal biomarker profiles for Bcl-2 were identified as the best model, and TRAJ group profiles generated from ranked data are described in Figure 2A. Comprising ∼11% of the population, group 1 (riser) subjects had rising Bcl-2 levels. Group 2 (low) subjects had persistently low levels over time. As both groups 3 and 4 had consistently high Bcl-2 levels that were elevated compared with controls, TRAJ groups 3 and 4 were combined and labeled group 3 (high) for further analysis. Actual Bcl-2 concentrations were then graphed for each of the Bcl-2 TRAJ groups and compared with healthy control levels (Figure 2B). The actual biomarker levels correspond well with the relative relationships of the ranked profiles for each TRAJ group presented in Figure 2A. The riser group started below control levels, and by day 4, increased to levels comparable to the high group. The low group (31% of the population) had Bcl-2 levels that remained similar to control levels across time. Bcl-2 levels for each TRAJ were different from each other for all 6 days (P<0.004 all comparisons).

Figure 2.

Figure 2

Open in a new tab

(A) Trajectory groups for profiles over time and percent membership for each trajectory group for cerebrospinal fluid (CSF) Bcl-2 using ranked data. The group based trajectory analysis (TRAJ) procedure identified four groups for Bcl-2. Groups 3 and 4 were combined for further analysis. (B) Mean daily CSF Bcl-2 concentrations over time for three Bcl-2 trajectory groups compared with healthy controls. Group 3 includes subjects from groups 3 and 4 in (A). Significant differences between Bcl-2 levels for each trajectory group on days 0 to 5 (P<0.004 all comparisons).

Three statistically distinct temporal biomarker profiles for CytoC were identified as the best model, and TRAJ group profiles generated from the ranked data are described in Figure 3A. Group 1 (low) had consistently low levels, group 2 (decliner) initially had high levels that decreased over time, and group 3 (high) had a consistently high level of CytoC over time. Actual CytoC concentrations were graphed for each TRAJ group and compared with concentrations observed in healthy controls (Figure 3B). Like Bcl-2 TRAJ groups, the actual biomarker levels corresponded well to the relative relationships of the ranked profiles for each TRAJ group presented in Figure 3A. The low CytoC group (41% of the population) had a consistently low CytoC concentration over time that was similar to controls. CytoC concentrations for the decliner group (42% of the population) initially were higher than control levels and peaked on day 1. Levels decreased thereafter, returning to control levels by the last day. In contrast, CytoC concentrations for the high group (20% of the population) remained elevated compared with controls throughout the study period. CytoC levels for each TRAJ group were different from each other on all 6 days (P<0.047 all comparisons).

Figure 3.

Figure 3

Open in a new tab

(A) Trajectory groups for profiles over time and percent membership for each trajectory group for cerebrospinal fluid (CSF) cytochrome C using ranked data. The group based trajectory analysis (TRAJ) procedure identified three groups for cytochrome C. (B) Mean daily CSF cytochrome C concentrations over time for three cytochrome C trajectory groups compared with healthy controls. There were significant differences between cytochrome C levels for each trajectory group on days 0 to 5 (P<0.05 all comparisons).

Trajectory Group and Outcome Associations with Demographic and Clinical Variables After traumatic brain injury

TRAJ group comparisons were made with demographic and injury/clinical variables (see Table 1). Although there were some occasional trends in daily Bcl-2 levels based on sex, GCS, and mortality (data not shown), there were no differences by Bcl-2 TRAJ group membership with regard to these variables or injury severity score, hospital LOS, number of total complications, presence of neurologic complications, presence of sepsis, hypothermia treatment, and type of injury. Similarly, there were some occasional trends between daily CytoC levels and sex, GCS, age, and mortality (data not shown). However, CytoC TRAJ membership only showed trends between specific CytoC TRAJ groups and acute mortality, age, GCS, or SAH (Table 1).

Table 1. Bivariate trajectory group associations with demographic and clinical variables after TBI.

  Bcl-2 trajectory groups
 Cytochrome C trajectory groups
  Riser group 1 (N=6) Low group 2 (N=23) High group 3 (N=41) Bcl-2 statistics Low group 1 (N=30) Decliner group 2 (N=31) High group 3 (N=12) CytoC statistics
Radiological injury type
 SAH, % 50.0 73.9 75.6 χ2=1.841, P=0.463 63.3 83.9 58.3 χ2=4.453, P=0.105
 SDH, % 50.0 56.5 61.0 χ2=0.451, P=0.813 46.7 64.5 50.0 χ2=2.112, P=0.348
 EDH, % 16.7 8.7 9.8 χ2=0.856, P=0.692 6.7 12.9 8.3 χ2=0.774, P=0.866
 Contusion, % 50.0 56.5 51.2 χ2=0.290, P=0.934 40.0 58.1 58.3 χ2=2.339, P=0.310
 IVH, % 50.0 21.7 39.0 χ2=2.760, P=0.294 33.3 45.2 25.0 χ2=1.709, P=0.421
 ICH, % 33.3 30.4 29.3 χ2=0.224, P>0.999 40.0 19.4 25.0 χ2=3.164, P=0.191
 DAI, % 16.7 39.1 39 χ2=1.033, P=0.651 40.0 41.9 16.7 χ2=2.505, P=0.321
 Other, % 16.7 4.3 14.6 χ2=1.957, P=0.409 13.3 12.9 8.3 χ2=0.222, P>0.999
                 
Injury severity
 GCS, median (IQR) 6 (3.8) 6 (5.7) 6 (4.7) χ2=2.18, P=0.34 6 (6.7) 6 (4.7) 5 (3.6) χ2=5.58, P=0.062
                 
 ISS, x̄±s.e. 33.0±3.7 35.8±2.0 34.1±1.5 F=0.366, P=0.716 34.7±1.3 34.0±1.8 34.3±3.1 F=0.041, P=0.960
                 
Demographic and clinical variables
 Total complications, median (IQR) 2 (1.4) 2 (1.3) 2 (1.3) χ2=0.494, P=0.781 2 (0.3) 2 (1.3) 2 (1.3) χ2=2.75, P=0.253
 Neurological complications, % present 16.7 21.7 22.0 χ2=0.093, P>0.99 16.7 22.6 33.3 χ2=1.35, P=0.491
 Sepsis, % present 16.7 8.7 21.9 χ2=1.99, P=0.386 13.3 12.9 33.3 χ2=2.56, P=0.278
 Age, years, x̄±s.e. 31.3±5.1 33.5±2.7 35.0±2.4 F=0.199, P=0.820 30.4±1.9 35.2±2.9 40.1±5.1 F=2.084, P=0.132
 Sex % female 0.0 26.1 14.6 χ2=2.144, P=0.336 20.0 22.6 8.3 χ2=0.988, P=0.671
 Acute mortality, % dead 0.0 17.4 26.8 χ2=1.995, P=0.326 10.0 25.8 41.7 χ2=5.499, P=0.062
 Hypothermia treatment, % yes 16.7 0.0 9.8 χ2=4.403, P=0.178 6.7 9.7 0.0 χ2=2.052, P=0.838
 Length of stay in hospital, days, x̄±s.e. 31.4±5.0 22.9±2.9 22.7±1.9 χ2=2.34, P=0.311 22.3±1.8 25.9±2.7 23.2±3.8 χ2=0.704, P=0.703
Open in a new tab

DAI, diffuse axonal injury; EDH, epidural hematoma; GCS, Glasgow Coma Scale; ICH, intracerebral hemorrhage; IQR, interquartile range; ISS, injury severity score; IVH, intraventricular hemorrhage; SDH, subdural hematoma; TBI, traumatic brain injury.

Bivariate analyses showed that higher GCS scores at the time of injury were associated with better outcomes (P⩽0.043 all comparisons). There was no effect of sex on outcome. Having diffuse axonal injury was associated with better GOS-6 and GOS-12 scores (P=0.032 and P=0.009, respectively) and DRS-6 (P=0.037). Acute care LOS was associated with outcome at 6 and 12 months with GOS (P<0.006 all comparisons). In each case, the outcome groups that included mortality had shorter LOS than those who survived their injuries (GOS-6: GOS group 1=14.98±2.77; GOS groups 2 to 3=27.66±2.33; GOS groups 4 to 5=23.94±2.15 and GOS-12: GOS group 1=16.33±2.96; GOS groups 2 to 3=31.67±3.95; GOS groups 4 to 5=23.17±2.01).

Table 2 shows that Bcl-2 and CytoC TRAJ groups were associated with 6 and 12 months outcome with GOS and DRS (P<0.015 all comparisons) with bivariate analysis. The riser Bcl-2 group was associated with the best outcomes at 6 and 12 months for the GOS and DRS. Five of six subjects in this riser Bcl-2 group had a good outcome defined as a GOS score of 4 to 5 and a DRS score of 0 to 3 at 6 months. By 12 months after injury, all (100%) of subjects in the riser Bcl-2 TRAJ group had a GOS 4 to 5 score and a DRS 0 to 3 score. In Table 2, subjects in the low CytoC group had the best GOS and DRS outcomes at 6 and 12 months. Those in the high CytoC TRAJ group had a 50% mortality rate at 6 months and a 67% mortality rate at 12 months based on GOS scores. Additionally, 100% of those in the high CytoC group were in the worst DRS group (15 to 30) by 12 months after injury.

Table 2. Bivariate outcome associations with Bcl-2 and CytoC trajectory groups.

  Bcl-2 trajectory groups
 Cytochrome C trajectory groups
  Riser group 1 (N=6) Low group 2 (N=23) High group 3 (N=41) Bcl-2 statistics Low group 1 (N=30) Decliner group 2 (N=31) High group 3 (N=12) CytoC statistics
GOS-6, %
 1 0.0 20.0 37.1 χ2=13.4, P=0.009 12.0 32.1 50.0 χ2=10.8, P=0.015
 2/3 16.7 30.0 45.7   40.0 35.7 50.0  
 4/5 83.3 50.0 17.1   48.0 32.1 0.0  
                 
GOS-12, %
 1 0.0 19.1 46.7 χ2=12.82, P=0.012 12.0 41.7 66.7 χ2=14.70, P=0.002
 2/3 0.0 14.3 26.7   20.0 16.7 33.3  
 4/5 100 66.7 26.7   68.0 41.7 0.0  
                 
DRS-6, %
 0–3 83.3 45.0 17.7 χ2=13.27, P=0.009 58.3 25.0 0.0 χ2=20.4, P<0.001
 4–14 16.7 30.0 29.4   29.2 32.1 16.7  
 15–30 0.0 25.0 52.9   12.5 42.9 83.3  
                 
DRS-12, %
 0–3 100 52.4 15.3 χ2=15.60, P=0.003 57.1 33.3 0.0 χ2=19.55, P<0.001
 4–14 0.0 23.8 23.1   28.6 25.0 0.0  
 15–30 0.0 23.8 54.6   14.3 41.7 100  
Open in a new tab

DRS, Disability Rating Scale; GOS, Glasgow Outcome Scale.

Ordinal Logistic Regression Models

Table 3a shows the multivariate ordinal logistic regression for predictors of GOS score at 6 and 12 months. Age was inversely related with outcome for both GOS-6 (P=0.025) and GOS-12 (P=0.013) scores. For each decile, the odds of better outcome decreased by ∼37% for GOS-6 and ∼47% for GOS-12. Higher GCS was also significantly associated with better odds of good outcome (P=0.008 and P=0.029 for GOS-6 and GOS-12, respectively). Moreover, hospital LOS was associated with GOS-6 (P=0.044) and trended toward significance for GOS-12 (P=0.097). Patients in the low CytoC group had better GOS-12 scores when compared with the high and decliner CytoC groups combined (P=0.023). For example, patients in the low CytoC TRAJ group were ∼5 times more likely to have better outcome at 12 months than patients in the combined high and decliner CytoC groups. The combined low and riser Bcl-2 groups were significantly associated with better GOS-6 and GOS-12 scores (P=0.009 and P=0.002, respectively) when compared with the high Bcl-2 group. In fact, patients in the low and riser Bcl-2 groups were ∼9 times more likely to have better outcome 12 months after injury. Injury type, based on CT data, did not influence GOS.

Table 3. Multivariate logistic regression result.

(a) Independent variables GOS
  6 Months (N=58)
 12 Months (N=52)
  Odds ratio (95% CI) P value Odds ratio (95% CI) P value
Age 0.63 (0.42, 0.94) 0.025 0.53 (0.32, 0.87) 0.013
GCS 1.67 (1.14, 2.44) 0.008 1.62 (1.05, 2.48) 0.029
Bcl-2 group LOW+RISER versus HIGH 4.8 (1.49, 15.19) 0.009 8.9 (2.29, 34.41) 0.002
CytoC group LOW versus DECLINER+HIGH 2.2 (0.69, 6.95) 0.182 4.8 (1.20, 19.14) 0.026
Length of hospital stay
 1.1 (1.00, 1.10)
 0.044
 1.04 (0.99, 1.10)
 0.097
(b) Independent variables DRS
  6 Months (N=58)
 12 Months (N=49)
 
 Odds ratio (95% CI)
 P value
 Odds ratio (95% CI)
 P value
Age 0.65 (0.43, 0.96) 0.032 0.64 (0.41, 1.00) 0.049
GCS 1.48 (1.03, 2.12) 0.033 XXXa XXX
Bcl-2 group LOW+RISER versus HIGH 3.83 (1.25, 11.76) 0.019 8.69 (2.45, 30.85) 0.001
CytoC group LOW versus DECLINER+HIGH 3.64 (1.17, 11.35) 0.026 4.13 (1.20, 14.19) 0.025
Open in a new tab

CI, confidence interval; DRS, Disability Rating Scale; GCS, Glasgow Coma Scale; GOS, Glasgow Outcome Scale.

a

XXX denote that GCS is not included in the final model for DRS at 12 months.

Table 3b represents the multivariate ordinal logistic regression for predictors of DRS at 6 and 12 months. Similar to that observed with the GOS, older age was associated with worse outcome for both DRS-6 and DRS-12 scores (P=0.032 and P=0.05, respectively). Higher GCS scores were associated with better DRS-6 scores (P=0.033), but they were not associated with DRS-12. Compared with the low CytoC group, the combined high and decliner groups had worse DRS-6 and DRS-12 scores (P=0.026 and P=0.025, respectively). Patients in the combined high and decliner CytoC groups were ∼4 times more likely to have bad outcome at 6 and 12 months as measured by DRS. Injury type did not influence DRS.

Using the likelihood ratio test, the 2*log-likelihood value increase for models with one and two biomarkers included for each outcome at each time point was highly significant (P<0.001 all comparisons). These results indicate that the successive inclusion of each biomarker into the basic model significantly added to the predictive power of each model.

Discussion

CytoC release initiates apoptosis via the instrinsic caspase-dependent pathway, while the Bcl-2 protein modulates mitochondrial permeability transition pore formation through its ability to inhibit the pore forming function of the BH-3 domains of other pro-apoptotic Bcl-2 family proteins and at subsequent CytoC release. Other markers of apoptosis have been measured in human CSF and resected tissue after TBI (Cardali and Maugeri, 2006; Clark et al, 1999; Darwish and Amiridze, 2010; Fink et al, 2008; Lenzlinger et al, 2002; Uzan et al, 2006; Zhang et al, 2003), and previous studies have established that CytoC and Bcl-2 are elevated in CSF after pediatric TBI (Clark et al, 2000; Satchell et al, 2005). However, to our knowledge, this is the largest characterization of CSF CytoC and Bcl-2 levels in TBI, and it is the first study to assess temporal profiles for these markers. We show that Bcl-2 and CytoC profiles are effective predictors of GOS and DRS after severe TBI, with subjects who have sustained high levels of each biomarker having worse outcomes. Further, no clinical studies have attempted to characterize dynamic changes in the profiles for these biomarkers and their relationship to outcome. The results show that while sustained high levels of CytoC and Bcl-2 over time are associated with worse outcomes, there is a distinct riser group for Bcl-2 and a distinct decliner group for CytoC that each has unique associations with outcome. Bcl-2 modulates CytoC release in experimental studies, in part, by countering the pore forming effects of pro-apoptotic factors like Bcl-2-associated X protein (Antonsson et al, 1997). So while it is likely that the overall concentrations of these CSF markers largely reflects cell death occurring after TBI, it is tempting to speculate that these profiles over time may also reflect the degree to which CNS cells can dynamically influence apoptosis via the initiation of intrinsic, caspase-dependent pathways. Based on our data, the riser Bcl-2 group and the decliner CytoC group are most suggestive of this case.

The results show that, as a whole, CSF Bcl-2 levels are fairly stable over time and are elevated above controls for days 1 to 4, while CSF CytoC levels peak on day 1 after injury and decline over time. This peak at day 1 after injury is consistent with experimental work estimating peak apoptosis occurs 24 to 72 hours after injury (Zhang et al, 2005). Cerebrospinal fluid biomarker measurements reflect protein and other small molecule content CNS cells at the time of cell death. Thus, increased CSF levels for both Bcl-2 and CytoC after TBI are expected compared with healthy controls. From the overall temporal profile, however, unique TRAJ groups were developed for each biomarker. Also, the relative relationships of the TRAJ groups derived from ranked data followed very closely with the actual concentrations observed over time for each group. The low TRAJ groups for both Bcl-2 and CytoC had levels that were similar to healthy controls, while the high TRAJ groups for both biomarkers were consistently elevated compared with controls. The riser Bcl-2 group had levels that increased overtime, while decliner CytoC group had levels that decreased over time, suggesting some attempt by injured cells to counter the activation of cell death pathways (Clark et al, 1997a, 2000).

Changes in Bcl-2 described both within and across TRAJ groups could be attributable to transcriptional changes in response to cellular injury, as observed in other human studies evaluating resected tissue (Clark et al, 1999, 2000) and rodent models of TBI (Clark et al, 1997a). The concept that CSF Bcl-2 levels represent both degree of cell death and regulation of pro-survival influences may be one reason why the riser Bcl-2 TRAJ group, whose levels at early time points are near controls and levels at later time points are similar to the high Bcl-2 group, has the best outcomes at both 6 and 12 months after TBI. Similarly, relative changes in Bcl-2 regulation over time may be an important factor contributing to the dynamic decrease in CytoC levels for the decliner group.

Bcl-2 and CytoC were evaluated together because of the relationship with each other that has been described. Other markers like Bcl-2-associated X protein, a pro-apoptotic member of the Bcl-2 family proteins involved in mitochondrial pore formation and CytoC release, and Bcl-2-associated death promoter, another pro-apoptotic member of the Bcl-2 family that inactivates Bcl-2 protein, would be interesting candidate markers to study. Genetic variation within the BCL-2 gene was recently associated with outcomes in severe TBI (Hoh et al, 2010). However, the role that genetic variability has in modulating Bcl-2 levels and in mediating the apparent effects of Bcl-2 profiles on outcome is not yet known. The riser Bcl-2 group did not have any distinguishing clinical or demographic characteristics, except that they were all male. However, this group may have some genetic variant, perhaps in the untranslated promoter region, which facilitates Bcl-2 upregulation after TBI and leads to better outcomes.

Previous work evaluating CSF Bcl-2 levels in children with severe TBI has shown that higher Bcl-2 levels are associated with favorable outcomes (Clark et al, 2000). In light of the results from this study, it is possible that CSF Bcl-2 levels in children with severe TBI primarily reflect the degree to which pro-survival proteins are activated rather than reflecting the amount of dying cells whose contents eventually emerge in CSF. However, the influence of Bcl-2 levels on pediatric outcomes other than survival remains to be studied. The pediatric TBI population may have important age-related differences in the degree of apoptosis, which has been shown to be greater in developing versus adult animals. Despite the fact that there are sex differences in Bcl-2 expression that enhance cell survival in experimental brain injury (Alkayed et al, 2001), Bcl-2 TRAJ group membership also was not significantly different based on sex.

While our work provides new insights into the possible role of Bcl-2 and CytoC in clinical TBI, the work also shows their utility as biomarkers for TBI prognosis. Both Bcl-2 and CytoC TRAJ group membership were important predictors of GOS and DRS scores. This work also adds to other recent reports suggesting the utility of group-based TRAJ analyses in modeling outcome (Berger et al, 2010; Salonia et al, 2010; Ozawa et al, 2009). We compared average daily values for subjects across TRAJ groups and in relation to controls and found distinct differences in biomarker levels over time for each group. It is notable that the TRAJ algorithm was able to discriminate unique groups of patients whose biomarker levels change significantly over time, whose levels remain near control levels across the time course, and still others whose levels were persistently elevated. When taking both the bivariate and multivariate models into consideration, it was evident that each longitudinal profile carries relevant information about the evolution of the injury and relationship to outcome. Also, each biomarker added unique information to the predictive capacity of the multivariate models beyond what was provided with standard demographic and clinical/injury variables.

We also assessed single point estimates, like weekly mean and maximum, in relationship to outcome (data not shown) and TRAJ was superior in predicting outcome. Unlike TRAJ, this approach did not allow for the differentiation of relevant subgroups or the estimation of missing data. In order for clinical biomarker assays to be useful at the bedside, it is critical to evaluate these markers at early time points, in addition to time trends, to maximize management and treatments that could impact outcome. As such, future work will assess the ability of early biomarker levels to accurately predict TRAJ group membership (dynamic pattern recognition) for markers like Bcl-2 and CytoC that are known to discriminate outcome. Further work also will be required to determine the validity of these TRAJ-based biomarker concentration differences across independent severe TBI populations and their generalizeability to other TBI populations.

Limitations

The findings here show the potential for Bcl-2 and CytoC to serve as prognostic biomarkers. However, some limitations should be considered. Although CSF Bcl-2 and CytoC levels appear promising as predictive biomarkers, the availability of CSF for biomarker analysis is not practical for many with severe injuries, and external ventricular drain placement is not routine in patients with mild to moderate injury. Previous reports show, though, that both serum CytoC and Bcl-2 have some utility as a biomarker in cancer care (Renz et al, 2001; Camlica et al, 2008; Kavathia et al, 2009). Thus, future work may also include the viability of serum CytoC and Bcl-2 in outcome prediction across the range of injury severity for TBI. Other factors (e.g., sepsis and age) have been shown to influence CytoC levels (d'Avila et al, 2008; Comim et al, 2008).

From a technical perspective, it is not possible to know if/how the sensitivity of detection for each biomarker differentially changes over time from injury, which could impact the accuracy of any biomarker tested. The analysis assumes that CytoC and Bcl-2 elevations occur in CSF after TBI primarily due to cell death. However, we did not specifically assess this hypothesis by correlating these markers with objective measures of cell death in CNS tissue. Moreover, biomarker studies are observational and not derived from traditionally mechanistic in vitro studies or ex vivo assays using transgenic models or pharmacological manipulations. Yet, the information from the longitudinal riser and decliner groups does provide some insight, from a translational research perspective, about how Bcl-2 might be a biomarker that represents both cell death and the dynamic pro-survival modulatory influence present in CNS tissue during the first week postinjury. The small numbers of subjects in the riser Bcl-2 and the decliner CytoC groups precluded extensive analyses about specific characteristics of these subpopulations or about their unique relationship to outcome. As such, this work would benefit from replication in larger sample sets. CytoC also may have some capacity to modulate apoptosis. For example, cytosolic CytoC can act as a peroxidase and form α-synuclein–cytoC complexes that protect against apoptosis (Bayir et al, 2009).

As both Bcl-2 and CytoC are intracellular molecules, their source in CSF after TBI is not clear. Apoptotic cells are often silently phagocytosed, so the contribution of apoptosis to overall biomarker levels is unclear. It is possible that CSF levels emerge primarily from intracellular contents derived from necrosed cells of a variety of cell types (e.g., glia and leukocytes) other than neurons. Importantly, human studies examining CSF levels of specific biomarkers of neuronal death such as neuron-specific enolase, and its relationship to Bcl-2 and CytoC should be conducted to substantiate the hypothesis that Bcl-2 and CytoC profiles reflect degree and mechanism of neuronal death.

Although Bcl-2 and CytoC are mitochondrial proteins, CSF levels for these markers per se are probably not informative with regard to degree of mitochondrial dysfunction in CNS tissue. However, CytoC and/or Bcl-2 may have some potential as informative markers of mitochondrial dysfunction if linked to markers of mitochondrial respiration and bioenergetics. Although beyond the technical feasibility of this initial study, linkage of Bcl-2 and CytoC to physiological parameters like intracranial pressure also may further define what these biomarkers mean in terms of TBI pathophysiology. Future work should also assess the prognostic potential of these markers in predicting functional and cognitive outcomes for TBI survivors.

Conclusions

Our work indicates that CSF Bcl-2 and CytoC levels are elevated in adults after severe TBI. The work also shows that sustained high levels of both Bcl-2 and CytoC are indicative of poor outcomes. Importantly, temporal profiles for Bcl-2 and CytoC, using group-based TRAJ analysis, are effective predictors of outcome and disability after severe TBI. TRAJ groups with declining levels of CytoC and rising levels of Bcl-2 suggest that CSF levels of these biomarkers reflect dynamic regulation of mitochondrial mediated apoptotic pathways. Further work assessing other Bcl-2 family proteins and also how genetic variation influences these biomarker levels over time is warranted. Further work will be required to validate these findings in other populations and to determine their comparative utility to other putative prognostic biomarkers currently being studied in the field.

The authors declare no conflict of interest.

Footnotes

This work was supported by DODW81XWH-071-0701 (AKW, AF), DOD PT073914 W81XWH-08-1-0237 (AKW, AF), R49 CCR323155-03 (AKW, AF, PMK, HB), NIH 5P01NS030318-16 (CED, PMK, RSB).

References

  1. Alkayed NJ, Goto S, Sugo N, Joh HD, Klaus J, Crain BJ, Bernard O, Traystman RJ, Hurn PD. Estrogen and Bcl-2: gene induction and effect of transgene in experimental stroke. J Neurosci. 2001;21:7543–7550. doi: 10.1523/JNEUROSCI.21-19-07543.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Antonsson B, Conti F, Ciavatta A, Montessuit S, Lewis S, Martinou I, Bernasconi L, Bernard A, Mermod JJ, Mazzei G, Maundrell K, Gambale F, Sadoul R, Martinou JC. Inhibition of Bax channel-forming activity by Bcl-2. Science. 1997;277:370–372. doi: 10.1126/science.277.5324.370. [DOI] [PubMed] [Google Scholar]
  3. Bayir H, Kapralov AA, Jiang J, Huang Z, Tyurina YY, Tyurin VA, Zhao Q, Belikova NA, Vlasova II, Maeda A, Zhu J, Na HM, Mastroberardino PG, Sparvero LJ, Amoscato AA, Chu CT, Greenamyre JT, Kagan VE. Peroxidase mechanism of lipid-dependent cross-linking of synuclein with cytochrome C: protection against apoptosis versus delayed oxidative stress in Parkinson disease. J Biol Chem. 2009;284:15951–15969. doi: 10.1074/jbc.M900418200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Berger RP, Bazaco MC, Wagner AK, Kochanek PM, Fabio A. Trajectory analysis of serum biomarker concentrations facilitates outcome prediction after pediatric traumatic and hypoxemic brain injury. Dev Neurosci. 2010;32:396–405. doi: 10.1159/000316803. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Boehning D, Patterson RL, Sedaghat L, Glebova NO, Kurosaki T, Snyder SH. Cytochrome c binds to inositol (1,4,5) trisphosphate receptors, amplifying calcium-dependent apoptosis. Nat Cell Biol. 2003;5:1051–1061. doi: 10.1038/ncb1063. [DOI] [PubMed] [Google Scholar]
  6. Bullock R, Chesnut RM, Clifton G, Ghajar J, Marion DW, Narayan RK, Newell DW, Pitts LH, Rosner MJ, Wilberger JW. Guidelines for the management of severe head injury. Brain Trauma Foundation. Eur J Emerg Med. 1996;3:109–127. doi: 10.1097/00063110-199606000-00010. [DOI] [PubMed] [Google Scholar]
  7. Camlica H, Duranyildiz D, Tas F, Yasasever V. Statistical interpretation of CA125 and Bcl-2 in serum of patients with late stage ovarian cancer. Am J Clin Oncol. 2008;31:585–588. doi: 10.1097/COC.0b013e318174dbd2. [DOI] [PubMed] [Google Scholar]
  8. Cardali S, Maugeri R. Detection of alphaII-spectrin and breakdown products in humans after severe traumatic brain injury. J Neurosurg Sci. 2006;50:25–31. [PubMed] [Google Scholar]
  9. Clark RS, Bayir H, Jenkins LW. Posttranslational protein modifications. Crit Care Med. 2005;33:S407–S409. doi: 10.1097/01.ccm.0000191712.96336.51. [DOI] [PubMed] [Google Scholar]
  10. Clark RS, Chen J, Watkins SC, Kochanek PM, Chen M, Stetler RA, Loeffert JE, Graham SH. Apoptosis-suppressor gene bcl-2 expression after traumatic brain injury in rats. J Neurosci. 1997a;17:9172–9182. doi: 10.1523/JNEUROSCI.17-23-09172.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Clark RS, Kochanek PM, Adelson PD, Bell MJ, Carcillo JA, Chen M, Wisniewski SR, Janesko K, Whalen MJ, Graham SH. Increases in bcl-2 protein in cerebrospinal fluid and evidence for programmed cell death in infants and children after severe traumatic brain injury. J Pediatr. 2000;137:197–204. doi: 10.1067/mpd.2000.106903. [DOI] [PubMed] [Google Scholar]
  12. Clark RS, Kochanek PM, Chen M, Watkins SC, Marion DW, Chen J, Hamilton RL, Loeffert JE, Graham SH. Increases in Bcl-2 and cleavage of caspase-1 and caspase-3 in human brain after head injury. FASEB J. 1999;13:813–821. doi: 10.1096/fasebj.13.8.813. [DOI] [PubMed] [Google Scholar]
  13. Clark RS, Kochanek PM, Dixon CE, Chen M, Marion DW, Heineman S, DeKosky ST, Graham SH. Early neuropathologic effects of mild or moderate hypoxemia after controlled cortical impact injury in rats. J Neurotrauma. 1997b;14:179–189. doi: 10.1089/neu.1997.14.179. [DOI] [PubMed] [Google Scholar]
  14. Colicos MA, Dash PK. Apoptotic morphology of dentate gyrus granule cells following experimental cortical impact injury in rats: possible role in spatial memory deficits. Brain Res. 1996;739:120–131. doi: 10.1016/s0006-8993(96)00824-4. [DOI] [PubMed] [Google Scholar]
  15. Comim CM, Rezin GT, Scaini G, Di-Pietro PB, Cardoso MR, Petronilho FC, Ritter C, Streck EL, Quevedo J, Dal-Pizzol F. Mitochondrial respiratory chain and creatine kinase activities in rat brain after sepsis induced by cecal ligation and perforation. Mitochondrion. 2008;8:313–318. doi: 10.1016/j.mito.2008.07.002. [DOI] [PubMed] [Google Scholar]
  16. Conti AC, Raghupathi R, Trojanowski JQ, McIntosh TK. Experimental brain injury induces regionally distinct apoptosis during the acute and delayed post-traumatic period. J Neurosci. 1998;18:5663–5672. doi: 10.1523/JNEUROSCI.18-15-05663.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. d'Avila JC, Santiago AP, Amâncio RT, Galina A, Oliveira MF, Bozza FA. Sepsis induces brain mitochondrial dysfunction. Crit Care Med. 2008;36:1925–1932. doi: 10.1097/CCM.0b013e3181760c4b. [DOI] [PubMed] [Google Scholar]
  18. Darwish RS, Amiridze NS. Detectable levels of cytochrome C and activated caspase-9 in cerebrospinal fluid after human traumatic brain injury. Neurocrit Care. 2010;12:337–341. doi: 10.1007/s12028-009-9328-3. [DOI] [PubMed] [Google Scholar]
  19. Fink EL, Lai Y, Zhang X, Janesko-Feldman K, Adelson PD, Szabo C, Berger RP, Sarnaik AA, Kochanek PM, Clark RS. Quantification of poly(ADP-ribose)-modified proteins in cerebrospinal fluid from infants and children after traumatic brain injury. J Cereb Blood Flow Metab. 2008;28:1523–1529. doi: 10.1038/jcbfm.2008.52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gouriéroux C, Holly A, Monfort A. Likelihood ratio test, Wald test, and Kuhn-Tucker test in linear models with inequality constraints on the regression parameters. Econometrica. 1982;50:63–80. [Google Scholar]
  21. Hoh NZ, Wagner AK, Alexander SA, Clark RB, Beers SR, Okonkwo DO, Ren D, Conley YP. BCL2 genotypes: functional and neurobehavioral outcomes after severe traumatic brain injury. J Neurotrauma. 2010;27:1413–1427. doi: 10.1089/neu.2009.1256. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Jennett B, Bond M. Assessment of outcome after severe brain damage. Lancet. 1975;1:480–484. doi: 10.1016/s0140-6736(75)92830-5. [DOI] [PubMed] [Google Scholar]
  23. Jones KE, Wessberg J, Vallbo AB. Directional tuning of human forearm muscle afferents during voluntary wrist movements. J Physiol. 2001;536:635–647. doi: 10.1111/j.1469-7793.2001.0635c.xd. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kavathia N, Jain A, Walston J, Beamer BA, Fedarko NS. Serum markers of apoptosis decrease with age and cancer stage. Aging (Albany, NY) 2009;1:652–663. doi: 10.18632/aging.100069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lenzlinger PM, Marx A, Trentz O, Kossmann T, Morganti-Kossmann MC. Prolonged intrathecal release of soluble Fas following severe traumatic brain injury in humans. J Neuroimmunol. 2002;122:167–174. doi: 10.1016/s0165-5728(01)00466-0. [DOI] [PubMed] [Google Scholar]
  26. Liou AK, Clark RS, Henshall DC, Yin XM, Chen J. To die or not to die for neurons in ischemia, traumatic brain injury and epilepsy: a review on the stress-activated signaling pathways and apoptotic pathways. Prog Neurobiol. 2003;69:103–142. doi: 10.1016/s0301-0082(03)00005-4. [DOI] [PubMed] [Google Scholar]
  27. Love S. Oxidative stress in brain ischemia. Brain Pathol. 1999;9:119–131. doi: 10.1111/j.1750-3639.1999.tb00214.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Nagin DS. Cambridge: Harvard University Press; 2005. Group-based modeling of development. [Google Scholar]
  29. Nakamura M, Raghupathi R, Merry DE, Scherbel U, Saatman KE, McIntosh TK. Overexpression of Bcl-2 is neuroprotective after experimental brain injury in transgenic mice. J Comp Neurol. 1999;412:681–692. doi: 10.1002/(sici)1096-9861(19991004)412:4<681::aid-cne9>3.0.co;2-f. [DOI] [PubMed] [Google Scholar]
  30. Ozawa H, Fabio A, Rogers EH, Niyonkuru C, Loucks T, Dixon CE, Wagner AK. Modeling trajectory patterns of serum hormone production in comparison to clinical variables in severe TBI. J Neurotrauma. 2009;26:A-71. [Google Scholar]
  31. Raghupathi R, Fernandez SC, Murai H, Trusko SP, Scott RW, Nishioka WK, McIntosh TK. BCL-2 overexpression attenuates cortical cell loss after traumatic brain injury in transgenic mice. J Cereb Blood Flow Metab. 1998;18:1259–1269. doi: 10.1097/00004647-199811000-00013. [DOI] [PubMed] [Google Scholar]
  32. Rappaport M, Hall KM, Hopkins K, Belleza T, Cope DN. Disability rating scale for severe head trauma: coma to community. Arch Phys Med Rehabil. 1982;63:118–123. [PubMed] [Google Scholar]
  33. Renz A, Burek C, Mier W, Mozoluk M, Schulze-Osthoff K, Los M. Cytochrome c is rapidly extruded from apoptotic cells and detectable in serum of anticancer-drug treated tumor patients. Adv Exp Med Biol. 2001;495:331–334. doi: 10.1007/978-1-4615-0685-0_46. [DOI] [PubMed] [Google Scholar]
  34. Rink A, Fung KM, Trojanowski JQ, Lee VM, Neugebauer E, McIntosh TK. Evidence of apoptotic cell death after experimental traumatic brain injury in the rat. Am J Pathol. 1995;147:1575–1583. [PMC free article] [PubMed] [Google Scholar]
  35. Robertson CL, Soane L, Siegel ZT, Fiskum G. The potential role of mitochondria in pediatric traumatic brain injury. Dev Neurosci. 2006;28:432–446. doi: 10.1159/000094169. [DOI] [PubMed] [Google Scholar]
  36. Salonia R, Empey PE, Poloyac SM, Wisniewski SR, Klamerus M, Ozawa H, Wagner AK, Ruppel R, Bell MJ, Feldman K, Adelson PD, Clark RS, Kochanek PM. Endothelin-1 is increased in cerebrospinal fluid and associated with unfavorable outcomes in children after severe traumatic brain injury. J Neurotrauma. 2010;27:1819–1825. doi: 10.1089/neu.2010.1402. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Satchell MA, Lai Y, Kochanek PM, Wisniewski SR, Fink EL, Siedberg NA, Berger RP, DeKosky ST, Adelson PD, Clark RS. Cytochrome c, a biomarker of apoptosis, is increased in cerebrospinal fluid from infants with inflicted brain injury from child abuse. J Cereb Blood Flow Metab. 2005;25:919–927. doi: 10.1038/sj.jcbfm.9600088. [DOI] [PubMed] [Google Scholar]
  38. Sullivan PG, Keller JN, Bussen WL, Scheff SW. Cytochrome c release and caspase activation after traumatic brain injury. Brain Res. 2002;949:88–96. doi: 10.1016/s0006-8993(02)02968-2. [DOI] [PubMed] [Google Scholar]
  39. Tehranian R, Rose ME, Vagni V, Griffith RP, Wu S, Maits S, Zhang X, Clark RS, Dixon CE, Kochanek PM, Bernard O, Graham SH. Transgenic mice that overexpress the anti-apoptotic Bcl-2 protein have improved histological outcome but unchanged behavioral outcome after traumatic brain injury. Brain Res. 2006;1101:126–135. doi: 10.1016/j.brainres.2006.05.049. [DOI] [PubMed] [Google Scholar]
  40. Uzan M, Erman H, Tanriverdi T, Sanus GZ, Kafadar A, Uzun H.2006Evaluation of apoptosis in cerebrospinal fluid of patients with severe head injury Acta Neurochir (Wien) 1481157–1164.discussion [DOI] [PubMed] [Google Scholar]
  41. Zhang X, Chen Y, Jenkins LW, Kochanek PM, Clark RS. Bench-to-bedside review: apoptosis/programmed cell death triggered by traumatic brain injury. Crit Care. 2005;9:66–75. doi: 10.1186/cc2950. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Zhang X, Graham SH, Kochanek PM, Marion DW, Nathaniel PD, Watkins SC, Clark RS. Caspase-8 expression and proteolysis in human brain after severe head injury. FASEB J. 2003;17:1367–1369. doi: 10.1096/fj.02-1067fje. [DOI] [PubMed] [Google Scholar]
  43. Zou H, Henzel WJ, Liu X, Lutschg A, Wang X. Apaf-1, a human protein homologous to C. elegans CED-4, participates in cytochrome c-dependent activation of caspase-3. Cell. 1997;90:405–413. doi: 10.1016/s0092-8674(00)80501-2. [DOI] [PubMed] [Google Scholar]

Articles from Journal of Cerebral Blood Flow & Metabolism are provided here courtesy of SAGE Publications

RESOURCES