B-Cell Lymphoma 2 (Bcl-2) Gene Is Associated with Intracranial Hypertension after Severe Traumatic Brain Injury

Abstract

Severe traumatic brain injury (TBI) activates the apoptotic cascade in neurons and glia as part of secondary cellular injury. B-cell lymphoma 2 (Bcl-2) gene encodes a pro-survival protein to suppress programmed cell death, and variation in this gene has potential to affect intracranial pressure (ICP). Participants were recruited from a single clinical center using a prospective observational study design. Inclusion criteria were: age 16-80 years; Glasgow Coma Scale (GCS) score 4-8; and at least 24 h of ICP monitoring treated between 2000-2014. Outcomes were mean ICP, spikes >20 and >25 mm Hg, edema, and surgical intervention. Odds ratios (OR), mean increases/decreases (B), and 95% confidence intervals (CIs) were reported. In 264 patients, average age was 39.2 years old and 78% of patients were male. Mean ICPs were 11.4 ± 0.4 mm Hg for patients with homozygous wild-type (AA), 12.8 ± 0.6 mm Hg for heterozygous (AG), and 14.3 ± 1.2 mm Hg for homozygous variant (GG; p = 0.023). Rs17759659 GG genotype was associated with more ICP spikes >20 mm Hg (p = 0.017) and >25 mm Hg (p = 0.048). Multi-variate analysis showed that GG relative to AA genotype had higher ICP (B = 2.7 mm Hg, 95% CI [0.5,4.9], p = 0.015), edema (OR = 2.5 [1.0, 6.0], p = 0.049) and need for decompression (OR = 3.7 [1.5-9.3], p = 0.004). In this prospective severe TBI cohort, Bcl-2 rs17759659 was associated with increased risk of intracranial hypertension, cerebral edema, and need for surgical intervention. The variant allele may impact programmed cell death of injured neurons, resulting in elevated ICP and post-traumatic secondary insults. Further risk stratification and targeted genotype-based therapies could improve outcomes after severe TBI.

Introduction

Traumatic brain injury (TBI) is a leading cause of death and disability worldwide. In the United States, it results in 2.8 million annual emergency department visits, 282,000 hospitalizations, and 56,000 deaths.¹ The heterogeneity of TBI hinders accurate clinical assessment and optimal treatment, and individuals with similar injuries frequently have divergent outcomes.^2,3 Following injury, onset of secondary sequelae can often worsen the course of recovery.⁴ Neuro-intensive care focuses on advanced modalities of intracranial monitoring to reduce the risks of intracranial hypertension, edema, hypoxia, and metabolic dysregulation.^5–8 Collaborative efforts, such as the prospective multi-center Transforming Research and Clinical Knowledge in TBI (TRACK-TBI) to improve detection of TBI, identify patients at risk for persistent deficits, and develop targeted treatments are ongoing.^3,9-11

A growing body of literature points toward a role for genetic factors in individual variability after brain injury. Single-nucleotide polymorphisms (SNPs) that affect outcome after TBI occur in genes that regulate neural and glial apoptosis, neuroinflammatory pathways, and neurobehavioral capacity.^12–16 B-cell lymphoma 2 (Bcl-2) is a pro-survival protein that was first identified in follicular lymphoma. It is shown to be differentially expressed by neuronal tissue in the setting of ischemia and trauma.^17,18 Together with apoptosis promoters (i.e., Bax and Bad), the Bcl-2 family genes regulate programmed cell death in response to cellular injury via mitochondrial membrane permeability, and injured neurons that lack Bcl-2 exhibit apoptotic morphology.¹⁹

Bcl-2 expression increases after TBI,^20,21 and higher cerebrospinal fluid Bcl-2 concentrations in pediatric patients are associated with lower mortality and better Glasgow Outcome Scale (GOS) scores.¹⁸ The Bcl-2 gene is located on chromosome 18q21.3, and includes three exons. In a prior candidate gene study of 17 BCL-2 SNPs, we identified the SNP rs17759659, an intronic SNP near the 5′ end of intron II (Fig. 1), as a predictor of 3-month functional outcomes.¹² The variant allele was associated with poorer outcomes as measured by the GOS, Neurobehavioral Rating Scale-Revised (NRS-R), and mortality compared with the wildtype.¹² However, the mechanism through which this SNP impacts pathophysiology and outcome after head trauma remains unexplored and warrants further investigation.

Fig. 1.

The Bcl-2 gene on chromosome 18q21.3. Detailed map of the Bcl-2 gene, the three exons, and single-nucleotide polymorphisms (SNPs) including rs17759659, an intronic SNP near the 5′ end of intron II on chromosome 18q21.3, based on HapMap data and the National Center for Biotechnology Information SNP database.

In this study, we explored the association between Bcl-2 rs17759659 and intracranial pressure (ICP) in a cohort of severe TBI patients while adjusting for injury severity and patient demographics. Given the regulatory role of Bcl-2 in neuronal apoptosis, we hypothesized that Bcl-2 rs17759659 would be associated with elevated ICP, radiographic evidence of edema, and need for neurosurgical intervention.

Methods

Study design

All study patients were enrolled in the Brain Trauma Research Center (BTRC) database, a prospective observational, cohort study that enrolls patients from a single level I trauma center of a tertiary care institution. The BTRC is located in the University of Pittsburgh Medical Center (UPMC), with the goal of using blood and cerebrospinal fluid (CSF) biomarkers to determine biochemical changes and TBI recovery. This study utilized consecutively enrolled patients from May 2000 and April 2014 who met the inclusion criteria and the legal-authorized representative (LAR) were approached for proxy consent. Inclusion criteria for the study were the following: 1) age of 16 and 80 years; 2) severe TBI with Glasgow Coma Scale (GCS) score 4-8 with clinically indicated head computerized tomography (CT) scan; 3) placement of an external ventricular drain (EVD) for intracranial monitoring and CSF drainage and sampling per standard care; and 4) signed written consent from health care proxies. Exclusion criteria included pregnancy, imminent brain death and initial GCS of 3. Patients with GCS score 3 were enrolled but not included in the present study because of overall poor prognosis. For the purposes of this study, other exclusion criteria were missing genotype data for Bcl-2 rs17759659 and less than 24 h of ICP monitoring data available through the medical record. The study was approved by the University of Pittsburgh Institutional Review Board.

Biospecimen acquisition and Bcl-2 genotyping

Specimen acquisition and genetic analysis were performed as previously described.¹² In brief, blood samples were collected via peripheral venipuncture or existing arterial catheters within 24 h of presentation and stored at the biospecimens repository at the UPMC Department of Neurological Surgery. DNA analysis was conducted using white blood cells using a salting out technique. Genotyping of Bcl-2 rs17759659 was performed using ABIPrism 7000 Sequence Detection System with commercially available assays and SDS 2.0 software (Applied Biosystems Inc., Carlsbad, CA). Quality control included independent duplication of 10% of genotypes and independent double calls for each sample. Genotyping was performed by lab technicians blinded to demographic data and clinical outcomes.

Demographic and outcome measures

Demographic and clinical information (age, gender, admission GCS score, mechanism of injury), as well as hourly ICP measurements, were collected by trained research assistants blinded to Bcl-2 genotypes. Age in years was categorized in accordance with the Medical Subject Headings in the National Library of Medicine.²² Admission GCS scores were dichotomized into 4-6 and 7-8. Hourly ICP measurements in mm Hg from 1 to 120 h (standard duration per institutional protocol) after monitor placement were obtained from clamped EVD recordings. The pressure transducers of all EVDs were leveled to the anatomical tragus. Per institutional protocol for neurocritical monitoring, hourly ICP measurements at a minimum were documented by the nursing staff after 5 min of equilibration with the ICP transducer, confirming continuity of the drainage system, and visualizing an accurate waveform of the patient. Mean ICP, proportion of ICP recordings >20 mm Hg, and proportion of recordings >25 mm Hg were calculated for each patient.

Radiographic edema was determined by the presence of at least one of the five following findings on head CT: parenchymal herniation, sulcal effacement, ventricular compression, basilar cistern compression or effacement, and loss of gray–white differentiation. Final radiology reports were reviewed and confirmed by two trained research technicians, and any discrepancies were adjudicated by an independent blinded physician. All participants received medical treatment in accordance with the Brain Trauma Foundation guidelines for the management of severe TBI.²³ Neurosurgical decompression was performed in instances of intracranial hypertension refractory to tiered medical therapy, clinical decline with corresponding new radiographic findings, or focal mass lesion meeting operative criteria.

Statistical analysis

Descriptive statistics are presented as means, standard deviations (SDs) and standard errors (SEs) for continuous variables and proportions for categorical variables. Group differences in patient demographics and injury characteristics across Bcl-2 rs17759659 genotypes were assessed by Pearson's chi-squared test (X²) for categorical variables and analysis of variance (ANOVA) for continuous variables. Fisher's exact test was used to assess for differences in categorical variables when expected counts ≤5. We conducted multi-variable linear regression analyses testing the association between each of three ICP measures and rs17759659 genotypes using homozygous wildtype genotype (AA) as the reference group while controlling for gender, age, GCS score, and mechanism of injury. Binary logistic regressions were to assess the relationship between genotype and the development of edema and the need for neurosurgical decompression. The adjusted mean differences (B) and their associated 95% confidence intervals (CIs) are reported for predictors in each linear regression analysis. In binary logistic regression, the odds ratio (OR) and associated 95% CIs are reported. Significance was assessed at p = 0.05. All analyses were performed using the Statistical Package for Social Sciences (SPSS; version 25; IBM Corporation, Chicago, IL).

Results

Demographics and injury characteristics

In total, 264 patients were included in the current analysis as illustrated in Figure 2. A summary of the patient demographics and injury characteristics are listed in Table 1. The average age (SD) was 39.2 years, 78.0% of patients were male, and 90.5% of patients were LAR-identified as Caucasian. By age groups, adults 25-44 years old were the most common. Median GCS score was 6, with 48.9% of patients having admission GCS 4-6 and 51.1% having GCS 7-8. Mechanisms of injury included motor vehicle accident (MVA, 48.9%), fall (22.7%), motorcycle/all-terrain vehicle (ATV, 17.0%), hit by blunt object or assault (4.2%), and other (8.0%). No statistically significant differences were observed for any demographic or injury characteristics across SNP rs17759659 genotypes. Figure 1 provides a detailed map of the Bcl-2 gene with SNPs on chromosome 18q21.3 based on HapMap data and the National Center for Biotechnology Information SNP database. There were 105 (39.8%) participants with homozygous wild-type (AA), 127 (48.1%) with heterozygous (AG), and 32 (12.1%) with homozygous variant (GG) genotypes, conforming to Hardy-Weinberg equilibrium (X² = 0.46, p > 0.05).

Fig. 2.

Study profile of patients enrolled between 2000 and 2014. Flow chart depicting the patients with severe traumatic brain injury in the Brain Trauma Research Center who were included in the current analysis.

Table 1.

Demographic and Clinical Characteristics of the Patients in the Study by ICP Findings

Characteristics	n (%)	Mean ICP mm Hg (SE)	p	% ICP >20 mm Hg (SE)*†	p	% ICP >25 mm Hg (SE)*	p
Gender			0.237		0.401		0.429
Male	206 (78.0)	12.7 (0.4)		12.9 (1.4)		6.1 (1.0)
Female	58 (22.0)	11.7 (0.6)		10.5 (2.1)		4.4 (1.3)
Age, years			<0.001		0.036		0.679
16-25	73 (27.7)	14.1 (0.6)		15.9 (2.2)		6.6 (1.5)
25-44	98 (37.1)	12.9 (0.6)		13.9 (2.1)		6.3 (1.4)
45-64	65 (24.6)	11.2 (0.7)		9.0 (2.2)		4.8 (1.9)
65-80	28 (10.6)	9.5 (0.9)		5.6 (2.5)		3.4 (1.8)
Admission GCS			0.714		0.408		0.983
4-6	129 (48.9)	12.3 (0.5)		11.4 (1.7)		5.7 (1.3)
7-8	135 (51.1)	12.6 (0.5)		13.3 (1.7)		5.7 (1.1)
Mechanism of injury			0.894		0.305		0.076
MVA	127 (48.1)	12.2 (0.4)		9.9 (1.4)		3.7 (0.1)
Fall	60 (22.7)	12.9 (0.9)		15.7 (3.2)		9.6 (2.6)
Motorcycle/ATV	45 (17.0)	12.8 (0.9)		15.1 (3.2)		6.5 (1.7)
Object/assault	11 (4.2)	11.6 (2.0)		12.4 (7.5)		8.0 (5.8)
Other	21 (8.0)	12.7 (1.0)		12.0 (3.3)		4.0 (1.9)
rs17759659			0.023		0.017		0.048
AA	105 (39.8)	11.4 (0.4)		8.8 (1.3)		3.2 (0.1)
AG	127 (48.1)	12.8 (0.6)		13.5 (1.9)		7.0 (1.4)
GG	32 (12.1)	14.3 (1.2)		19.4 (4.5)		8.6 (3.3)

Bold represents significance assessed at p = 0.05.

^*Percentage of all ICP measurements >20 mm Hg or >25 mm Hg.

†Study enrollment prior to the Brain Trauma Foundation guidelines to treat ICP >22 mm Hg published in 2016.

ICP, intracranial pressure; SE, standard error; GCS, Glasgow Coma Scale; MVA, motor vehicle accident; ATV, all-terrain vehicle.

ICP monitoring

The mean ICP (SD) of all patients was 12.55.6 mm Hg during the course of 120 h (5 d) from ICP monitor placement. Over this period, the proportions of ICPs >20 mm Hg and >25 mm Hg were 12.4% and 5.7%, respectively. As listed in Table 1, male and female patients did not statistically vary by mean ICPs (p = 0.237), the proportion of ICPs >20 mm Hg (p = 0.401), or ICPs >25 mm Hg (p = 0.429). Mean ICP (SE) was 14.1 mm Hg for patients age 16-25 years, 12.9 mm Hg age 25-44 years, 11.2 mm Hg age 45-64 years, and 9.5 mm Hg age 65-80 years (p < 0.001). Older age had decreased frequency of ICP recordings >20 mm Hg (p < 0.001) and >25 mm Hg (p = 0.036). Multiple linear regression analyses showed that relative to 16-25 years, patients age 45-64 years (B = -3.7 mm Hg, 95% CI [-5.6, -1.8]; p < 0.001) and 65-80 years (B = -6.6 mm Hg [-9.4, -3.9]; p < 0.001) had lower mean ICPs. Similarly, patients age 45-64 years (B = -10.8% [-17.5, -4.1]; p = 0.002) and 65-80 years (B = -19.0% [-28.3, -9.6]; p < 0.001) had decreased ICP elevations >20 mm Hg. Patients age 65-80 years also had decreased ICP elevations >25 mm Hg compared with those 16-25 years (B = -9.5%, [-16.2, -2.8]; p = 0.006). Admission GCS score and the mechanism of injury were not associated with differences in ICP.

Patients with rs17759659 GG and AG genotypes had higher average ICPs relative to AA patients (14.3 mm Hg, 12.8 mm Hg and 11.4 mm Hg, respectively; p = 0.023) as shown in Figure 3A. As was also the case, GG and AG genotypes had higher proportion of ICP >20 mm Hg (19.4%, 13.5% and 8.8%, respectively; p = 0.017) compared with homozygous wildtype, in addition to ICP >25 mm Hg (8.6%, 7.0% and 3.2%, respectively; p = 0.048; Fig. 3B). Multi-variate regression showed that controlling for demographic and injury parameters, rs17759659 GG genotype had increased mean ICP (B = 2.7 mm Hg, [0.5, 4.9]; p = 0.015), greater ICP >20 mm Hg (B = 10.0%, [2.5, 17.4]; p = 0.009) compared with AA genotype. Figure 4 demonstrates the ICP recordings by SNP rs1775965 genotype over the course of 120 h after monitor placement post-trauma. A prior study characterized six distinct ICP groups in a risk-adjusted model to further assist with prognostication.²⁴

Fig. 3.

Intracranial pressure (ICP) measurements and outcomes by the patient Bcl-2 rs17759659 genotypes. (A) Mean ICP of patients with homozygous wild-type (AA) relative to heterozygous (AG) and homozygous variant (GG). (B) The proportion of measurement with spikes >20 mm Hg, >25 mm Hg, and >30 mm Hg. (C) The proportion of patients who developed radiographic edema and the need for surgical decompression.

Fig. 4.

Intracranial pressure (ICP) recordings over the course of 5 days post-trauma in severe traumatic brain injury (TBI) patients. The hourly ICP measurements of patients with severe TBI injury by their single-nucleotide polymorphism rs1775965 genotype following the placement of intracranial monitor.

Edema development and need for craniectomy

In total, 104 (43.7%) of patients showed signs of edema on CT. We found that the proportion of those with edema was not different by gender (p = 0.717), age group (p = 0.713), and admission GCS score (p = 0.093; Table 2). By the mechanism of injury, eight (72.7%) patients injured by object/assaults developed edema, followed by 30 (54.5%) injured by falls, 42 (38.2%) by MVAs, 13 (31.0%) by motorcycle/ATV, and 11 (55.0%) other (p = 0.020). Fifteen (55.6%) of patients with rs17759659 GG genotype had acute edema, followed by 56 (49.1%) of those with AG genotypes and 33 (34.0%) of AA genotype (p = 0.037). Multi-variate regression adjusting for patient characteristics showed that those struck by object/assaults were associated with increased odds of developing acute edema compared with MVA injuries (OR 4.7 [1.1, 19.5]; p = 0.034). Patients of Bcl-2 rs17759659 GG genotype experienced higher odds of developing edema than those of AA genotype (OR 2.5 [1.0, 6.0]; p = 0.049; Fig. 3C).

Table 2.

Characteristics of Severe TBI Patients Developing Acute Cerebral Edema and Those Requiring Surgical Decompression

Characteristics	Acute edema n (%)	p	Surgical decompression n (%)	p
Gender		0.717		0.021
Male	82 (44.3)		76 (36.9)
Female	22 (41.5)		12 (20.7)
Age, years		0.713		0.012
16-25	26 (41.3)		16 (21.9)
25-44	34 (40.5)		30 (30.6)
45-64	31 (49.2)		28 (43.1)
65-80	13 (46.4)		14 (50.0)
Admission GCS		0.093		0.037
4-6	58 (49.2)		51 (39.5)
7 – 8	46 (38.3)		37 (27.4)
Mechanism of injury		0.020		<0.001
MVA	42 (38.2)		28 (22.0)
Fall	30 (54.5)		31 (51.7)
Motorcycle/ATV	13 (31.0)		12 (26.7)
Object/Assault	8 (72.7)		6 (54.5)
Other	11 (55.0)		11 (52.4)
rs17759659		0.037		0.020
AA	33 (34.0)		28 (26.7)
AG	56 (49.1)		43 (33.9)
GG	15 (55.6)		17 (53.1)

Bold represents significance assessed at p = 0.05.

TBI, traumatic brain injury; SE, standard error; GCS, Glasgow Coma Scale; MVA, motor vehicle accident; ATV, all-terrain vehicle.

Eighty-eight (33.3%) of patients required craniectomies (36.9% male and 20.7% female patients; p = 0.021; Table 3). Craniectomy was most prevalent in patients age 65-80 years (50.0%), followed by age 45-64 years (43.1%), 25-44 years (30.6%) and 16-25 years (21.9%; p = 0.012). Patients with a GCS score 4-6 had higher rate of craniectomy (39.5%) compared with those with a GCS score 7-8 (27.5%; p = 0.037). Craniectomy by injury mechanism was 54.5% for struck by object/assault, 51.7% for falls, 26.7% for motorcycle/ATV, 22.0% MVA, and 52.4% for other (p < 0.001). Craniectomy increased from 26.7% in rs17759659 AA to 33.9% in AG and 53.1% in GG genotypes (p = 0.020; Fig. 3C). Multi-variate analysis showed that female patients were less likely to undergo surgery relative to male counterparts (OR 0.4 [0.2, 0.9]; p = 0.018), as were those with GCS score 7-8 relative to GCS score 4-6 (OR 0.6 [0.3, 1.0]; p = 0.046). Age 45-64 years had increased likelihood of craniectomy compared with 16-25 years (OR 2.5 [1.1, 5.9]; p = 0.033), and falls had higher odds of craniectomy (OR 2.8 [1.2, 6.2]; p = 0.013). Controlling for patient characteristics, patients displaying GG genotype had higher odds of craniectomy compared with AA (OR 3.7 [1.5, 9.3]; p = 0.004).

Table 3.

Multivariable Analysis of Patient Outcomes by ICP Monitoring, Findings of Acute Edema, and the Need for Craniectomy

	Mean ICP mm Hg		% ICP >20 mm Hg		% ICP >25 mm Hg		Acute edema		Surgical decompression
	B (95% CI)	p	B (95% CI)	p	B (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p
Gender
Male	Reference		—		—		—		—
Female	-0.3 (-1.9, 1.3)	0.703	0.1 (-5.6, 5.7)	0.984	-0.5 (-4.5, 3.6)	0.825	0.9 (0.5, 1.7)	0.678	0.4 (0.2, 0.9)	0.018
Age, years								0.737		0.188
16-25	Reference		—		—		—		—
25-44	-1.5 (-3.3, 0.2)	0.076	-3.8 (-9.7, 2.1)	0.208	-1.6 (-5.8, 2.6)	0.460	0.9 (0.4, 1.8)	0.721	1.5 (0.7, 3.3)	0.274
45-64	-3.7 (-5.6, -1.8)	<0.001	-10.8 (-17.5, -4.1)	0.002	-4.5 (-9.3, 0.3)	0.064	1.2 (0.5, 2.5)	0.716	2.5 (1.1, 5.9)	0.033
65-80	-6.6 (-9.4, -3.9)	<0.001	-19.0 (-28.3, -9.6)	<0.001	-9.5 (-16.2, -2.8)	0.006	0.7 (0.2, 2.1)	0.513	2.3 (0.3, 1.0)	0.151
Admission GCS
4-6	Reference		—		—		—		—
7-8	0.2 (-1.1, 1.5)	0.755	1.9 (-2.6, 6.5)	0.404	0.3 (-3.0, 3.6)	0.861	0.7 (0.4, 1.1)	0.139	0.6 (0.3, 1.0)	0.046
Mechanism of injury								0.061		0.011
MVA	Reference		—		—		—		—
Fall	3.1 (1.1, 5.1)	0.002	13.0 (6.2, 19.8)	<0.001	9.1 (4.2, 14.0)	<0.001	1.9 (0.9, 4.3)	0.117	2.8 (1.2, 6.2)	0.013
Motorcycle/ATV	1.0 (-0.9, 3.0)	0.288	6.5 (-0.2, 13.1)	0.057	3.1 (-1.6, 7.9)	0.195	0.7 (0.3, 1.6)	0.450	1.1 (0.5, 2.6)	0.783
Object/assault	1.2 (-2.2, 4.6)	0.498	8.0 (-3.7, 19.8)	0.180	6.6 (-1.8, 15.1)	0.122	4.7 (1.1, 19.5)	0.034	3.8 (1.0, 14.5)	0.051
Other	0.8 (-1.7, 3.4)	0.516	3.1 (-5.8, 11.9)	0.493	0.5 (-5.8, 6.8)	0.878	1.9 (0.7, 5.2)	0.217	4.1 (1.5, 11.6)	0.007
rs17759659								0.069		0.015
AA	Reference		—		—		—		—
AG	1.2 (-0.3, 2.6)	0.111	3.5 (-1.4, 8.5)	0.162	3.0 (-0.5, 6.6)	0.095	1.8 (1.0, 3.2)	0.064	1.2 (0.6, 2.2)	0.575
GG	2.7 (0.5, 4.9)	0.015	10.0 (2.5, 17.4)	0.009	5.0 (-0.3, 10.4)	0.066	2.5 (1.0, 6.0)	0.049	3.7 (1.5, 9.3)	0.004

Bold represents significance assessed at p = 0.05.

ICP, intracranial pressure; B, mean increase or decrease; CI, confidence interval; OR, odds ratio; GCS, Glasgow Coma Scale; MVA, motor vehicle accident; ATV, all-terrain vehicle.

Discussion

Clinical management of severe TBI patients remains an evolving evidence-based practice with persisting knowledge gaps.^23,25–28 There is growing interest in the genetic contributions to outcomes and recovery with the goal of improving prognosis and therapeutic precision.^15,29–33 ICP monitoring is an integral monitoring modality after severe neurotrauma, and our present study shows that the presence of the Bcl-2 rs17759659 variant allele is associated with increased ICP in human subjects. The present study indicates that the Bcl-2 rs17759659 variant allele may predispose patients to an increased risk of intracranial hypertension and cerebral edema. These findings may provide new insights into the role for genetic factors to account for variability of secondary sequelae after severe TBI.

Bcl-2 rs17759659 and implications in elevated ICP

TBI injury burden is exacerbated by secondary insults after the initial trauma. Experimental models in rats noted that Bcl-2 expression is induced 15-fold up to 7 days post-trauma, in cortical neurons that are injured but could possibly survive.^30,34 The Bcl-2 family genes regulate programmed cell death,^35,36 with injured neurons lacking Bcl-2 exhibiting apoptotic morphology, whereas those expressing Bcl-2 do not show biochemical or morphologic features of apoptosis. As with findings in vivo and the recent characterization of its genetic variability in humans,¹² the current investigation analyzed Bcl-2 rs17759659 genotypes with hourly ICP recordings as a potential clinical marker for trauma-induced apoptosis and swelling.

Our present study of severe TBI patients indicates that presence of one or both Bcl-2 variant alleles was associated with increased ICP over 5 days after injury in human participants. Intracranial hypertension, observed as ICP spikes >20 mm Hg and >25 mm Hg, increased in frequency with the Bcl-2 rs17759659 variant allele (G). Controlling for patient and injury parameters, mean ICP was 2.7 mm Hg higher with GG than AA genotype, along with 10.0% additional ICP spikes >20 mm Hg. Patients with variant Bcl-2 alleles have higher baseline ICPs that remain below the current recommended threshold to treat (i.e., ICP greater than 22 mm Hg).²⁵ However, compensatory mechanisms in these patients may be limited in accordance with the Monroe-Kellie doctrine,²⁹ resulting in more frequent ICP spikes and intracranial hypertension that require additional medical or surgical treatment. Possibly due to increased apoptosis, patients with both variant alleles are also at increased risk of developing cerebral edema. Bcl-2 rs17759659 polymorphism may affect Bcl-2 expression in the injured gray matter, which could be directly quantified from the cerebrospinal fluid drainage and will be a topic for future investigation. Bcl-2 levels in the CSF begin to increase on post-injury Day 1, peak on Days 3 and 4, then gradually decline after Day 5.³⁷ Thus, changes in the CSF milieu can closely reflect the intracranial response to TBI. Thus, both intracranial hypertension and radiographic edema were more prevalent in these patients, raising concern for the known correlation between high ICPs and poor outcome.

TBI patients of any severity are more likely to be male,¹ as was observed in our study population. While it is plausible that demographics and TBI mechanism may be associated with differences in neuromonitoring measurements and risk of intracranial hypertension,³⁸ a larger study population with higher statistical power is warranted. In agreement with previous studies,^39,40 older TBI patients in our study have lower overall ICPs post-trauma relative to their younger counterparts.

Cerebral edema development and need for craniectomy

The development of cerebral edema is associated with unfavorable outcomes after TBI, occurring in the initial 24 h and could peak 3-5 days post-trauma.⁴¹ Edema can be focal or diffuse and of vasogenic or cytotoxic etiology, both of which contribute to secondary insults and elevated ICPs, which facilitates secondary insults and intracranial hypertension. Findings from experimental models have shown that Bcl-2 messenger RNA is upregulated 6 h after initial injury in pericontusional cortex and maintained up to 7 days, particularly in injured but surviving neurons.^42,43 Bcl-2 is located intracellularly in the mitochondria, nuclear membranes and endoplasmic reticula,^44-46 protecting cells from traumatic and ischemic stimuli that can activate apoptosis and result in the development of cytotoxic edema. Current recommendations for intracranial hypertension mainly consist of a nonspecific, tiered approach. More targeted therapies, such as sulfonylurea receptor-1 antagonists, may be on the horizon.⁴⁷ In our study, patients homozygous for the variant rs17759659 allele were more likely to develop edema and require craniectomy. Given the role that Bcl-2 plays in survival of injured neurons and reducing programmed cell death by inhibiting the release of mitochondrial proteins,^18,29,48 it is possible that rs17759659 genetic variability alters Bcl-2 expression in injured neurons, resulting in increased risk for edema formation and secondary insults more resistant to medical treatment.

Limitations

While findings from the current study adds to the body of knowledge implicating a role for Bcl-2 after severe TBI, the study is not without limitations. Given that this was a single-center study, race and gender parameters of the sample limits generalizability of the data to the overall patient population, despite the large sample size. With these findings, the role for Bcl-2 in post-TBI outcomes will be further investigated in TRACK-TBI. While this study explored the relationship of Bcl-2 on post-trauma ICP, it did not account for disparate medical management that may temporarily increase or decrease ICP recordings. ICP changes are part of a dynamic process that can be reflective of secondary sequelae after TBI. For this reason, a small sample of ICPs alone may not be sufficient and only patients with at least 24 h of data were included in the current study in order to understand overall trends in ICP during the 5 days post-trauma. There is equipoise in the literature regarding continuous CSF drainage for ICP management, but it is our institutional standard of care. This may in fact reduce the likelihood of false positive findings. Our data did not distinguish the indication for each craniectomy or the craniectomy size. While provider variations exist, the overall association between Bcl-2 polymorphism and intracranial hypertension is nonetheless significant. Also, Bcl-2 rs17759659 polymorphism may affect Bcl-2 expression in the injured gray matter, which could be directly quantified from the cerebrospinal fluid drainage and will be a topic for future investigation.

Conclusions

The Bcl-2 rs17759659 polymorphism was associated with elevated ICP in a cohort of severe TBI patients in the 5 days post-trauma. Bcl-2 modulates programmed cell death cascades in injured neurons, and we found that the presence of both rs17759659 alleles carry greater risks of intracranial hypertension, edema, and the need for surgical intervention. Further clarification of Bcl-2 genomics may improve risk stratification and lead to targeted Bcl-2 therapies to mitigate secondary injury after TBI.

Funding Information

This study was supported by the federally funded grants R00-NR013176, R01NR008424, R01NR013342, and 5P50NS30318.

Author Disclosure Statement

No competing financial interests exist.