Genetic Variation in the TP53 Gene and Patient Outcomes Following Severe Traumatic Brain Injury

Abstract

Traumatic brain injury (TBI) is a leading cause of death and disability, with more than 5 million people in the United States living with long-term complications related to TBI. This study examined the relationship between TP53, the gene that codes for the protein p53, and outcome variability following severe TBI. The p53 protein impacts neuronal apoptosis following TBI, thus investigation into TP53 genetic variability as a prognosticator for TBI outcomes (mortality, Glasgow Outcome Scale [GOS], Neurobehavioral Rating Scale [NRS], and Disability Rating Scale [DRS]) is warranted. Participants (N = 429) with severe TBI (Glasgow Coma Scale score ≤8) were enrolled into a prospective study with outcomes assessed over 24 months following injury. The single-nucleotide polymorphism Arg72Pro (rs1042522), a functional missense polymorphism for which the CC homozygous genotype is most efficient at inducing apoptosis, was investigated. Individuals with the CC genotype (arginine homozygotes) were more likely to have poorer outcomes at 24 months following TBI compared to individuals with CG/GG genotypes (GOS: p = .048, DRS: p = .022). These findings add to preliminary evidence that p53 plays a role in recovery following TBI and, if further replicated, could support investigations into p53-based therapies for treating TBI.

Traumatic brain injury (TBI) is a leading cause of death and disability in the United States, with approximately 2.87 million emergency department visits, 288,000 hospitalizations, and 56,800 deaths related to TBI annually (Centers for Disease Control and Prevention, 2019). The Centers for Disease Control and Prevention reported a 54% increase in TBI-related emergency department visits and a 6% decrease in death rates from 2006 to 2014, suggesting an increase in the number of survivors of TBI. Survivors often face chronic complications that reduce their quality of life (Conley et al., 2014; Faul et al., 2010; Hasan et al., 2017; National Institute of Neurological Disorders and Stroke, 2015; Weaver & Portelli, 2014), and an estimated 5.3 million people in the United States are living with long-term complications or disabilities related to TBI (National Institute of Neurological Disorders and Stroke, 2015).

Gene activation that promotes apoptosis is a cause of cell death following TBI in the hours to days after the initial injury, though research has not completely identified the mechanisms of neuronal apoptosis (Maas et al., 2008; Martinez-Lucas et al., 2005). Apoptosis after TBI is a target of therapies under development (Chuang et al., 2012; Kim et al., 2010), including agents targeting TP53, which plays a critical role in neuronal cell apoptosis and autophagy (He et al., 2017) as well as transactivating genes that play a role in neuronal cell repair and regeneration (Napieralski et al., 1999; Wan et al., 2013; Wan et al., 2014). p53’s role in apoptosis includes recognition of DNA damage, DNA repair, and cell cycle regulation. Given that the p53 protein is crucial to neuronal apoptosis and that increased neuronal apoptosis following TBI is well documented, focusing on a role for p53 in patient outcomes following TBI is warranted. TP53, the gene that codes for the p53 protein, is located on chromosome 17q13.1. The Arg/Arg homozygous (CC) genotype of a common functional polymorphism in the gene, Arg72Pro (rs1042522), is at least 5 times more efficient at inducing apoptosis than the heterozygous or Pro/Pro homozygous genotypes (CG/GG; Dumont et al., 2003). The purpose of the current study was to explore the relationship between the Arg72Pro (rs1042522) functional polymorphism of TP53 and variability in patient outcomes up to 2 years following severe TBI.

Materials and Methods

Participants

We derived data from a prospective observational study of adult patients with severe TBI that was approved by the local institutional review board. Data collected in that study from patients admitted to a hospital affiliated with a university medical center included demographic, acute hospitalization, and neurological outcome data and fluid biospecimens. Inclusion criteria were as follows: admission for a closed head injury with a Glasgow Coma Scale (GCS) score ≤8 without effects of drugs, alcohol, paralytics, or sedatives; age 16–80 years old; and placement of an external ventricular drain as standard of care. Patients with brain-death concerns were excluded.

Measures

Evaluation of functional and neurobehavioral outcomes

Trained neuropsychology technicians supervised by a senior neuropsychologist completed outcome evaluations at 3, 6, 12, and 24 months (Conley et al., 2014). The outcomes of interest included mortality and scores on the Glasgow Outcome Scale (GOS), Neurobehavioral Rating Scale (NRS), and Disability Rating Scale (DRS). The GOS is a valid and reliable measure (McMillan et al., 2016) that categorizes TBI outcomes based on a survivor’s independence in daily functioning as follows: 1 = death, 2 = persistent vegetative state, 3 = severe disability, 4 = moderate disability, and 5 = good recovery (Christensen, 2014). The NRS rates 27 items on a scale from 0 to 6 assessing behavioral manifestations of TBI (0 = deficit absent, 1 = very mild, 2 = mild, 3 = moderate, 4 = moderately severe, 5 = severe, 6 = extremely severe). Some areas of interest covered by the scale include alertness, attention, fatigability, orientation, memory, motor behavior, expressive/reception language, mood disturbances, disinhibitory behavior or agitation, and capacity for self-insight (Vanier et al., 2000). The ratings for each item are tallied to create a summative score. Possible scores range from 0, or no deficit, to 162, or extremely severe deficits. The NRS has an average interrater reliability of 74.3% and an average κ statistic of .40 (Conley et al., 2014). The NRS requires individuals to be alive and able to participate in the assessment; therefore, participants who were dead (i.e., GOS of 1) or in a persistent vegetative state (i.e., GOS of 2) could not be evaluated using this scale. The DRS is an assessment used in both the acute hospital setting and the community to evaluate functional outcomes and ability following TBI. The assessment measures the three categories of impairment, disability, and handicap by rating subcategories of eye-opening, ability to communicate, motor responsiveness, cognitive skill necessary for self-care, overall dependence, physical and cognitive abilities, and employment. Scores range from 0, or no disability, to 30, or death. The DRS is both reliable and valid, with an interrater reliability ranging from 93% to 98% in the inpatient setting (Deepika et al., 2017).

Genotype data collection

DNA samples for this study were extracted from one of two sources: blood (preferred) or cerebral spinal fluid (CSF) (alternate). Researchers obtained 10 mL whole venous blood via venipuncture, centrifuged it to isolate the white blood cells, and extracted it using a salting out protocol (Miller et al., 1988). CSF drainage was collected in a ventriculostomy bag as a part of routine clinical care, and the DNA was extracted using the manufacturer’s instructions for the QIAamp Midi kit (Qiagen, Valencia, CA, USA). All DNA samples were stored in 1× TE buffer at 4 °C. Restriction fragment length polymorphism analysis was used to genotype participants. The primer set for rs1042522 was 5′-CTGGTAAGGACAAGGGTTGG-3′ as forward and 5′-ACTGACCGTGCAAGTCACAG-3′ as reverse. The 397-base-pair fragment was amplified by polymerase chain reaction and then treated with 5 units of BstUI restriction enzyme, creating two short fragments (166 and 213 base pairs) if the G allele was present. A 2% agarose gel was run to separate the DNA fragments. Arg homozygotes (CC) showed a single band with no cut at 379 base pairs; Pro homozygotes (GG) showed two bands at 213 and 166 base pairs; and heterozygotes (CG) showed all three bands at 379, 213, and 166 base pairs.

Statistical Analysis

The independent variable in this study is the single nucleotide polymorphism (SNP) genotype. We evaluated Hardy–Weinberg equilibrium for the SNP. The dependent variables include mortality and GOS, DRS, and NRS scores at 3, 6, 12, and 24 months. Potential covariates include age, sex, and severity of injury (GCS). We dichotomized GOS into poor outcomes (GOS 1, 2, 3) versus good outcomes (GOS 4, 5). For GOS, we used χ² analyses at each time point separately. To analyze DRS and NRS by genotype, we used one-way analysis of variance. To further break down severity of TBI, we dichotomized GCS scores (GCS 3–4, GCS 5–8).

To investigate if outcome differences existed between homozygous Arg (CC) and presence of Pro variants (CG and GG), we ran independent-samples t tests and χ² tests. We considered a p value ≤.05 significant. We ran multivariate regression analyses on all time points where outcome measures trended toward significance to explore the effects of potential covariates of age, race, sex, and initial GCS score. We calculated odds ratios (ORs) and 95% confidence intervals.

Results

We included a total of 429 participants with severe TBI in the present analysis. Table 1 outlines participants’ demographic characteristics. The average age was 37.4 years old (range 16–77), 78.3% were male, and 85.5% were Caucasian, reflecting the population presenting to the clinical site. Frequency of participants who had a GCS of 3–4 was 24.4%. The SNP (rs1042522) was in Hardy–Weinberg equilibrium. The sample sizes for completed GOS were n = 391 at 3 months, n = 383 at 6 months, n = 362 at 12 months, and n = 310 at 24 months. The sample sizes for NRS were n = 150 at 3 months, n = 175 at 6 months, n = 170 at 12 months, and n = 101 at 24 months. For DRS, the sample sizes were n = 256 at 3 months, n = 251 at 6 months, n = 224 at 12 months, and n = 159 at 24 months. Minor allele frequency in our population was 30.2%, which is consistent with population frequencies reported in the SNP database (dbSNP; www.ncbi.nlm.nih.gov/snp).

Table 1.

Demographic Characteristics of Participating Patients With TBI by TP53 rs1042522 Genotype.

Characteristic	Genotype			p Value
	CC	CG	GG
	n = 51 (11.9%)	n = 157 (36.6%)	n = 221 (51.5%)
Age (years), mean ± SE	37.16 ± 2.17	36.43 ± 1.24	38.10 ± 1.19	.637
Sex (male), n (%)	38 (74.5)	114 (72.6)	184 (83.3)	.036*
Race, n (%)				.014*
White/Caucasian	39 (76.5)	129 (82.2)	199 (90.0)
Non-White	10 (19.6)	12 (7.6)	11 (5.1)
GCS score, n (%)				.203
3–4	9 (17.6)	45 (28.7)	50 (22.6)
5–8	42 (82.4)	112 (71.3)	171 (77.4)

Note. N = 429. GCS = Glasgow Coma Scale; TBI = traumatic brain injury; SE = standard error.

* Significant at p ≤ .05.

Distribution and analysis of mortality by genotype (Table 2) shows participants with the CG genotype to have the lowest mortality rates across all time points, with a statistically significant difference at the 24-month time point (p = .013). Analysis of the effects of homozygous Arg and homozygous Pro on mortality showed no significant difference between the two groups (data not shown).

Table 2.

Mortality (GOS Score of 1) Over Time by TP53 rs1042522 Genotype.

Time Point	Genotype			p Value
	CC	CG	GG
	n (%)	n (%)	n (%)
3 Months, n = 391	13 (27.7)	35 (24.6)	74 (36.6)	.053
6 Months, n = 383	14 (31.1)	38 (27.3)	77 (38.7)	.087
12 Months, n = 362	16 (36.4)	40 (29.4)	77 (42.3)	.062
24 Months, n = 310	16 (48.5)	41 (34.5)	83 (52.2)	.013*

Note. Mortality listed is cumulative over time (e.g., participants who died at 3 months are included in the subsequent time points as well). GOS = Glasgow Outcome Scale.

* Significant at p ≤ .05.

Table 3 outlines results from the χ² test comparing the dichotomized GOS groups by genotype. We found no significant differences in GOS group by genotype. In Table 4, we report results for NRS and DRS scores by genotype. We found a significant association between genotype and DRS score at the 24-month time point (p = .048).

Table 3.

Frequency of a Poor Outcome Over Time (GOS Score of 1–3) by TP53 rs1042522 Genotype.

Time Point	Genotype			p Value
	CC	CG	GG
	n (%)	n (%)	n (%)
3 Months, n = 391	37 (78.7)	107 (75.3)	159 (78.7)	.746
6 Months, n = 383	31 (68.9)	94 (67.6)	135 (67.8)	.987
12 Months, n = 362	28 (63.6)	79 (58.1)	112 (61.5)	.743
24 Months, n = 310	25 (75.8)	70 (59.3)	104 (65.4)	.117

Note. GOS = Glasgow Outcome Scale.

Table 4.

Mean ± SE Scores on NRS and DRS by TP53 rs1042522 Genotype Over Time.

Time Point	NRS (Mean ± SE)				DRS (Mean ± SE)
	CC	CG	GG	p Value	CC	CG	GG	p Value
3 Months	40.86 ± 2.05	44.16 ± 2.18	39.50 ± 0.92	.081	9.56 ± 1.46	9.74 ± 0.78	8.06 ± 0.68	.242
6 Months	40.95 ± 1.80	40.50 ± 0.94	39.44 ± 0.90	.621	7.94 ± 1.45	6.94 ± 0.71	5.98 ± 0.60	.315
12 Months	36.80 ± 2.43	41.40 ± 1.23	40.05 ± 1.22	.233	5.96 ± 1.50	5.55 ± 0.70	4.69 ± 0.66	.563
24 Months	38.89 ± 1.97	40.89 ± 1.71	41.87 ± 1.99	.783	7.69 ± 2.04	4.87 ± 0.86	3.50 ± 0.54	.048*

Note: NRS = Neurobehavioral Rating Scale; DRS = Disability Rating Scale; SE = standard error.

* Significant at p ≤ .05.

Table 5 summarizes the multivariate analyses for all outcome measures (GOS, DRS, and NRS scores) by Arg homozygotes (CC) versus Pro heterozygotes and homozygotes (CG and GG), as well as for the potential covariates of age, race, sex, and initial extent of injury (GCS). We found significant differences in the 24-month GOS score between participants with CG versus CC genotype (OR = 0.27, p = .014) and between those with initial GCS scores of 3–4 versus those with initial scores of 5–8 (OR = 5.16, p < .001) when controlling for potential covariates. Age was significant for GOS at 24 months and NRS at 3 months. Sex also showed trends toward significance (OR = 1.99, p = .056). Multivariate analysis of DRS score at 24 months shows significant difference in both CG versus CC (B = −3.82, p = .021) and GG versus CC genotypes (B = −3.31, p = .048) and in initial GCS score of 3–4 versus 5–8 (B = 3.08, p = .021).

Table 5.

Multivariate Analysis for Differences in Outcomes by Genotype, Potential Demographic Covariates, and Initial GCS Score.

Variable	GOS at 24 Months		DRS at 24 Months		NRS at 3 Months
	OR (95% CI)	p Value	Coefficient	p Value	Coefficient	p Value
CG versus CC	0.27 (0.09–0.76)	.014*	−3.82	.021*	3.62	.326
GG versus CC	0.45 (0.16–1.25)	.126	−3.31	.048*	−1.41	.695
Age	1.06 (1.04–1.08)	<.001*	0.05	.194	0.16	.025*
White versus non-White	0.32 (0.08–1.36)	.122	−1.63	.529	−5.85	.214
Sex	1.99 (0.98–4.02)	.056	1.54	.176	−2.08	.436
Initial GCS 3–4 versus 5–8	5.16 (2.51–10.61)	<.001*	3.08	.021*	2.63	.404

Note. DRS = Disability Rating Scale; GCS = Glasgow Coma Scale; GOS = Glasgow Outcome Scale; NRS = Neurobehavioral Rating Scale; OR = odds ratio.

* Significant at p ≤ .05.

Table 6 summarizes the multivariate analysis of Arg homozygotes (CC) and Pro heterozygotes and homozygotes (CG/GG) for GOS and DRS scores at 24 months. Arg homozygotes had significantly worse outcomes than the Pro variants for both measures (GOS: p = .048, DRS: p = .022). Initial GCS score was also significantly associated with both long-term outcome measures (GOS: p < .001, DRS: p = .023).

Table 6.

Multivariate Analyses for Differences in Long-Term Outcomes of GOS and DRS Scores at 24 Months Between Arginine Homozygotes and Proline Heterozygotes/Homozygotes and by Additional Covariates.

Variable	GOS (Poor)		DRS
	OR (95% CI)	p Value	Coefficient	p Value
Proline versus Arginine	0.36 (0.13–0.97)	.48*	−3.58	.022*
Age	1.06 (1.04–1.08)	<.001*	0.05	.191
White versus non-White	0.34 (0.08–1.47)	.149	−1.54	.550
Sex	1.86 (0.94–3.74)	.082	1.43	.199
Initial GCS 3−4 versus 5−8	4.89 (2.40–9.97)	<.001*	2.96	.023*

Note. DRS = Disability Rating Scale; GCS = Glasgow Coma Scale; GOS = Glasgow Outcome Scale; OR = odds ratio.

* Significant at p ≤ .05.

Discussion

TP53 rs1042522 Arg homozygotes (CC genotype) were more likely to have poorer outcomes at 24 months following severe TBI compared to individuals with CG or GG genotypes. Interestingly, this genotype is also more efficient at inducing apoptosis (Dumont et al., 2003). p53 plays a role in neuronal apoptosis following TBI (Maas et al., 2008; Martinez-Lucas et al., 2005), and variability in the TP53 gene was shown in a previous study to be associated with TBI recovery (Maas et al., 2008; Martinez-Lucas et al., 2005); however, previous studies had smaller sample sizes and more limited outcome evaluations than the currently study. Taken together, the findings of the current and prior studies support a role for p53 in recovery after severe TBI, implying a potential for genotype-based precision interventions.

Genotype and allele-based analyses in the current study indicated associations with DRS score at 24 months. The GG (homozygous proline) TP53 genotype was associated with better outcomes, as measured by the DRS, at 2 years following TBI, suggesting that individuals with GG genotype may have a better prognosis following injury. Conversely, our multivariate analysis showed that individuals with the CC genotype were more likely than individuals with either the CG or GG genotype to have a poor DRS score 24 months after injury (p = .021, p = .048). Comparison of alleles at 24 months showed the C allele was associated with a worse outcome compared to the G allele (p = .022).

Unadjusted analyses indicated there were no significant differences in GOS outcome among the three genotypes. However, when we conducted multivariate regression analyses, we found CG to be a protective genotype compared to CC (OR = 0.27, p = .014). Further, our data indicated that individuals with the CG genotype had the lowest mortality rates across all time points. These findings were somewhat unexpected, given Martinez-Lucas et al.’s (2005) findings in which GG was associated with the best outcome. The variance between our results and previously reported work might be explained by our larger sample size, which enabled a more thorough evaluation of heterozygotes.

Homozygotes for Arg (CC) were 2.7 times more likely to have a poor outcome based on 24-month GOS (OR = 0.36, p = .048) compared to individuals with genotypes with the Pro allele (CG/GG) in multivariate analysis. We noted significant differences in 24-month GOS score by age and initial GCS score, with initial GCS score having the strongest correlation (OR = 4.89, p < .001) to GOS outcome. This finding partially replicates the findings of Martinez-Lucas et al. (2005) in which there was a significant difference in GOS outcomes at 6 months between the Arg homozygotes and Pro homozygotes but not between homozygotes and heterozygotes (Martinez-Lucas et al., 2005).

The genetic associations we found in the current study, coupled with previous work and the biological plausibility of p53 playing a role in recovery after TBI, support continued investigation into p53 as a potential therapeutic target. TP53 is a known tumor suppressor gene and has been highly studied in human longevity, cancer risk, and survival (Genetics Home Reference, 2020; Orsted et al., 2007). Despite TP53 being well characterized in cancer populations, few researchers have examined potential effects of variations in the gene in the TBI population. p53 contributes to neuronal cell apoptosis and autophagy as well as transactivating genes that play a role in neuronal cell repair and regeneration (Napieralski et al., 1999; Wan et al., 2013; Wan et al., 2014). Exploratory animal studies have identified changes in expression of p53 following closed head injury (Lu et al., 2000), ischemic brain injury (Chopp et al., 1999), and lateral fluid-percussion injury (Kaya et al., 1999) as well as in p53-deficient mice (Martinez-Lucas et al., 2005; Morrison et al., 1996). Specifically, these animal models have found an upregulation of p53 in the nuclei of injured cells following TBI (Kaya et al., 1999). A study by Yang et al. (2016) supports the administration of p53 inactivators following TBI. In that study, researchers delivered a p53 inhibitor, pifithrin-α oxygen analogue, to lab control TBI-induced rats 5 hr postinjury and found it resulted in reduction of neuronal apoptosis. Results also showed an improvement of motor and cognitive functions and supported the use of p53 inhibitors as a targeted therapeutic strategy in TBI.

Limitations of the current study included attrition related to participant dropout or death. In addition, the results may not adequately represent non-White patients or females, though our sample demographics resemble the TBI patient population, which is mostly male. Finally, the relationship between p53 and neuronal apoptosis is complex, and by looking only at the rs1042522 SNP, we left out of consideration potential environmental impacts and the influences of other genes.

Conclusion

In the current study, we found that the rs1042522 SNP of TP53 was associated with patient outcomes following severe TBI, with CC (Arg) homozygotes having increased risk for poor outcome. Our genetic findings are consistent with what we know from the literature about the cellular role of p53. The CC genotype of rs1042522 is known to be more efficient in inducing apoptosis and was associated with poorer outcomes in the current and prior studies, which is consistent with the literature that finds that, as neural apoptosis declines, outcomes after neuronal injury improve. Our findings support additional p53-related investigations in TBI patients, which could lead to p53-based precision treatment of patients following TBI.