Sage Journals: Discover world-class research

Abstract

Chronic consequences of mild traumatic brain injury (mTBI) are heterogeneous, but may be treatable with targeted medical and rehabilitation interventions. A biological signature for the likelihood of response to therapy (i.e., “predictive” biomarkers) would empower personalized medicine post-mTBI. The purpose of this study was to correlate pre-intervention blood biomarker levels and the likelihood of response to targeted interventions for patients with chronic issues attributable to mTBI. Patients with chronic symptoms and/or disorders secondary to mTBI >3 months previous (104 days to 15 years; n = 74) were enrolled. Participants completed pre-intervention assessments of symptom burden, comprehensive clinical evaluation, and blood-based biomarker measurements. Multi-domain targeted interventions for specific symptoms and impairments across a 6-month treatment period were prescribed. Participants completed a follow-up testing after the treatment period. An all-possible model's backward logistic regression was built to identify predictors of improvement in relation to blood biomarker levels before intervention. The minimum clinically important difference (MCID) of the change score (post-intervention subtracted from pre-intervention) for the Post-Concussion Symptom Scale (PCSS) to identify treatment responders from non-responders was the primary outcome. The MCID for total PCSS score was 10. The model to predict change in PCSS score over the 6-month intervention was significant (R² = 0.09; p = 0.01) and identified ubiquitin C-terminal hydrolase L1 (odds ratio [OR] = 2.53; 95% confidence interval [CI], 1.18–5.46; p = 0.02) and hyperphosphorylated tau (p-tau; OR = 0.70; 95% CI, 0.51–0.96; p = 0.03) as significant predictors of symptom improvement beyond the PCSS MCID. In this cohort of chronic TBI subjects, blood biomarkers before rehabilitation intervention predicted the likelihood of response to targeted therapy for chronic disorders post-TBI.

Introduction

Approximately 3.2 million persons in the United States have disability secondary to mild (mTBI) traumatic brain injury (TBI).¹ Chronic effects from mTBI are estimated to cost >$60 billion annually in the United States alone.² Clinical manifestations of mTBI are heterogeneous, and modern management is predicated on the targeted treatment of specific conditions. Validation of objective assessments for the diagnosis and prognostication of mTBI has achieved recent success.^3,4 Focus is expanding to investigate additional contexts of use for blood biomarkers or clinical characteristics to predict the therapeutic response to a targeted intervention for mTBI. The Centers for Disease Control and Prevention refers to this class of biomarkers as “predictive biomarkers.”⁵ The purpose of this exploratory study was to assess the relationship between pre-intervention blood biomarker concentrations and clinical characteristics with the magnitude of symptom improvement after a 6-month targeted intervention for chronic symptoms related to mTBI.

Methods

Design and participants

The TEAM-TBI (Targeted Evaluation Action and Monitoring of TBI) study was a prospective multiple interventional trial of patients with sequelae secondary to mTBI >3 months previous (NCT02657135). Inclusion criteria were 18–60 years of age and fluent in the English language with a documented history of TBI or blast exposure since 2001. Interventions were targeted to the patient's impairments over 6 months. Clinical results of this trial have been published.⁶ Results from earlier work focused on symptoms and clinical grading scales whereas the present analysis focused on measured biomarkers.

Pre- and post-intervention assessments were conducted, including demographics/medical history (e.g., sex, injury mechanism, number of mTBIs, time since index injury, and military status), the Post-Concussion Symptom Scale (PCSS), and blinded analysis of blood biomarkers (e.g., glial fibrillary acidic protein [GFAP], tau, hyperphosphorylated tau Thr231 [p-tau], von Willebrand factor [vWF], brain lipid binding protein [BLBP], ubiquitin C-terminal hydrolase-L1 [UCH-L1], vascular endothelial growth factor-a [VEGFa], and claudin-5 [CLDN5]). These biomarkers were chosen before study initiation for analysis, based on their empirical association with mTBI and/or common mTBI sequelae. This study was approved by the University of Pittsburgh Institutional Review Board for human subjects' research.

Blood biomarker analysis

Plasma samples were analyzed by using the reverse phase protein microarray (RPPM) system, a high-sensitivity, high-throughput proteomics platform according to established procedures.^7–16 Denatured samples were serially diluted in a 1:2 manner (five-step) and printed onto nitrocellulose film slides by using a Quanterix 2470 Arrayer (Quanterix, Billerica, MA). Slides were dried and blocked with Azure Protein-Free Blocking Buffer (Azure Biosystems, Inc, Dublin, CA) and incubated with primary antibodies overnight (8–12 h) at 4°C. A list of primary antibodies can be viewed in Supplementary Table S1. After washing, slides were incubated with biotinylated secondary antibodies (1:100,000 dilution). Slides were scanned in an Innopsys InnoScan 710-IR scanner for XDR (extended dynamic range) signal acquisition at 785 nm (Innopsys Inc, Chicago, IL). The reverse phase protein array (RPPA) or RPPM, a high-sensitivity, high-throughput antibody-based analytical platform, was adopted from MD Anderson Cancer Center's protocol.¹⁷ We have successfully used RPPA in numerous TBI-related projects during the last decade. After incubation with the primary antibodies (see Supplementary Table S1 for details), slides were washed, blocked, and incubated with fluorescent-tagged secondary anti-mouse or -rabbit antibodies (from Invitrogen) diluted at 1:6000 in antibody incubation buffer for 1 h at room temperature.

Fluorescence data were imported into a Microsoft Excel-based bioinformatics program (Microsoft Corporation, Redmond, CA), and net intensity versus dilution was plotted on a log2-log2 scale. Total amount of antigen was determined by the y-axis intercept or Y-cept (i.e., by extrapolating the regression line to zero). Here, we express the Y-cept values as log2-transformed Y-cept values, which correspond to the total net intensity of the undiluted plasma sample.

Outcome measures

Post-Concussion Symptom Scale

The PCSS is a 22-item survey completed on a 7-point Likert scale, from 0 (not at all) to 6 (severe). PCSS score was the primary outcome measure for the present study.

Statistical analysis

Pre- and post-intervention PCSS scores were converted into change scores (post-intervention score – pre-intervention score). The MCID of PCSS change score was determined by multiplying 0.5 (signifying a medium Cohen's d effect size) with the standard deviation of the total sample. Univariable logistic regressions were conducted to assess the association of individual biomarkers with improving PCSS score beyond the MCID. An all-possible backward logistic regression model was built to assess the association of pre-intervention blood biomarker concentrations and demographics/medical history variables on improving outcome scores (cutoff, p < 0.05). Odds ratios (ORs) and 95% confidence intervals (CIs) were derived from the models. For interpretation, ORs >1 are indicative of higher odds of improving greater than the MCID, whereas ORs <1 are indicative of lower odds of improving greater than the MCID. Post hoc diagnostics were conducted to evaluate model performance. Internal validity testing was conducted using the calibrationbelt command in Stata software (SE 17.0; StataCorp LLC, College Station, TX). A p value <0.05 indicates that the model was poorly calibrated. Predicted probabilities of the model were generated to calculate discriminative validity using receiver operating characteristic area under the curve (AUC) analysis and the Brier score to calculate the accuracy of probabilistic predictions. A Brier score of 0 indicates perfect accuracy, and 1 indicates perfect inaccuracy.

Results

Demographics

Participants (n = 74; age: 35.8 ± 8.6 years; 19% female) were 5.6 ± 3.6 years from their index mTBI and reported a median of five mTBI exposures at enrollment (Table 1). The sample was 67.4% former military (n = 64). Reported mechanisms of injury were blunt (n = 35; 36.8%), blast (n = 24; 25.3%) or both (n = 36; 37.8%).

Table 1.

Demographic and Clinical Outcome Variables for the Sample

	Mean ± SD No. (%)	PCSS⁺ n = 50	PCSS^– n = 34	p value
Age at enrollment	35.8 ± 8.6	34.3 ± 8.0	38.7 ± 9.0	0.02^*
Years since last TBI	5.6 ± 3.6	5.72 ± 35.00	5.02 ± 3.70	0.38
Mechanism of index TBI Blast Blunt Both	24 (24.7) 35 (36.1) 36 (37.1)	15 (30.0) 18 (36.0) 17 (34.0)	6 (17.6) 13 (38.2) 15 (44.1)	0.66 0.84 0.35
Female sex	18 (18.6)	10 (20.0)	6 (17.6)	0.79
Military status	64 (67.4)	36 (72.0)	22 (64.7)	0.48
Exposures	5.2 ± 3.5	5.1 ± 3.2	5.7 ± 3.9	0.53
Pre-intervention PCSS total	49.3 ± 22.9	55.4 ± 21.6	40.1 ± 22.1	0.002^*
Post-intervention PCSS total	35.4 ± 22.6	28.5 ± 21.0	45.5 ± 21.3	<0.001^*
Change in PCSS total	–13.8 ± 20.3	–26.9 ± 13.7	5.4 ± 10.7	<0.001^*

Indicates statistically significant at p < 0.05.

TBI, traumatic brain injury; PCSS, Post-Concussion Symptom Scale; SD, standard deviation.

The MCID for PCSS change score was −10. Participants whose symptoms improved by more than −10 (PCSS⁺; n = 50) were younger at enrollment (4.3 years), with higher pre-intervention PCSS total scores (+15.3) and lower post-intervention total scores (−17.0) than participants whose symptoms improved by −10 or less (PCSS^–; n = 34).

Univariable regression models

Logistic regression models to differentiate PCSS⁺ from PCSS^– with individual biomarkers can be viewed in Table 2. No biomarkers had a statistically significant relationship with symptom improvement beyond the MCID, but UCH-L1 had the highest OR (OR = 1.76; 95% CI, 0.91–3.40; p = 0.09).

Table 2.

Univariable Logistic Regression Results to Predict Concussion Symptom Improvement After Targeted Intervention Using Individual Blood Biomarkers

	Odds ratio	p value	95% CI
UCHL1	1.76	0.09	0.91–3.40
GFAP	1.15	0.67	0.61–2.14
vWF	1.03	0.90	0.68–1.56
VEGFa	1.19	0.51	0.71–2.01
p-tau	0.82	0.15	0.63–1.07
BLBP	1.61	0.16	0.83–3.13
CLDN5	1.57	0.17	0.82–3.03
Tau	1.18	0.48	0.74–1.89

UCHL1, ubiquitin C-terminal hydrolase L1; GFAP, glial fibrillary acidic protein; vWF, von Willebrand factor; VEGFa, vascular endothelial growth factor-a; p-tau, hyperphosphorylated tau; BLBP, brain lipid binding protein; CLDN5, claudin-5; CI, confidence interval.

Multi-variable regression model

The model to differentiate PCSS⁺ from PCSS^– over the 6-month intervention period was statistically significant (R² = 0.09; p = 0.01) and retained two predictors (Table 3). UCH-L1 (OR = 2.54; 95% CI, 1.18–5.46; p = 0.017) was associated with higher odds of symptom improvement beyond the MCID, whereas p-tau (OR = 0.70; 95% CI, 0.51–0.96; p = 0.026) was associated with lower odds of symptom improvement beyond the MCID. Post hoc internal validity testing indicated that the model was well calibrated (p = 0.35). The final model had acceptable discriminative validity (AUC = 0.69; 95% CI, 0.56–0.82; p = 0.006), with a sensitivity of 77.3% and specificity of 50%. The Brier score was 0.21, indicating reasonable forecasting ability. No clinical characteristics were significant predictors.

Table 3.

Multi-Variable Logistic Regression to Predict Concussion Symptom Improvement After Targeted Intervention

	Odds ratio	p value	95% CI
UCH-L1	2.54	0.017	1.18–5.46
p-tau	0.70	0.026	0.51–0.96

UCH-L1, ubiquitin C-terminal hydrolase L1; p-tau, hyperphosphorylated tau; CI, confidence interval.

Discussion

This study identified predictive biomarkers for both a positive and negative response to targeted intervention in chronic mTBI. Higher pre-intervention UCH-L1 was associated with 154% higher odds of improving global concussion symptoms (i.e., PCSS scores) beyond the MCID. Conversely, higher pre-intervention p-tau was negatively associated with 30% lower odds of symptom improvement beyond the MCID after targeted intervention. These preliminary results for predictive biomarkers of targeted intervention for chronic TBI require further study and validation in larger mTBI populations.

UCH-L1 is an enzyme found in nerve cells throughout the brain involved in the clearance of unneeded proteins.¹⁸ Blood level of UCH-L1 is elevated after subconcussive blast exposure during military training and can remain elevated for years after moderate and severe TBI. UCH-L1 is a potential marker of treatment efficacy in patients with chronic neurological injury. Perhaps, higher levels of UCH-L1 have the capacity to improve symptoms and function after mTBI or subconcussive exposure by removing cortical proteins associated with these exposures and poorer clinical outcomes (e.g., amyloid beta, tau, and p-tau). This hypothesis may explain why UCH-L1 and p-tau were not significantly associated with symptom improvement as univariate predictors, but were both significantly associated with symptom improvement when considered together. Indeed, impairment of the ubiquitin-proteasome pathway (which can occur after TBI)¹⁹ has been shown to enhance the toxic accumulation of proteins (such as p-tau) and is associated with neurodegenerative disease, such as Parkinson's and Alzheimer's diseases.^20,21

p-tau is the hyperphosphorylated tau protein typically associated with neurofibrillary tangles in Alzheimer's disease.²² p-tau detected in the blood can differentiate military veterans with an mTBI history from those with no history, and p-tau correlates significantly with post-concussive symptom burdens.²³ p-tau formation is associated with synaptic dysfunction and neuronal loss, both mechanisms of memory impairments.²² Based upon this evidence and the results of this preliminary study, higher levels of p-tau in the blood of chronic mTBI patients may reflect a neuronal environment not conducive to improvement from non-pharmacological treatment.²⁴

Although this seems a plausible hypothesis for these preliminary results, it should be noted that PCSS scores were 15 points higher before intervention and ∼17 points lower post-intervention in the group of responders compared to non-responders. This could represent a regression to the mean for the responder group who had much higher symptom burdens before treatment. Non-responders were roughly 4.5 years older than responders (Table 1), which may have impacted the degree of improvement, but age was not retained in the final regression model as a significant predictor. In the absence of other group differences which may have impacted the outcome (i.e., years since last TBI, proportion of females and veterans, and exposures), it is unclear why responders had higher symptom burdens before intervention.

Limitations

The final model only accounted for 9% of variance in PCSS improvement from targeted intervention, indicating that other predictors are necessary to understand who will improve from targeted treatment in this unique population of patients with chronic issues related to mTBI. The sample was approximately two thirds former military, and >50% of the index injuries were related to a blast mechanism. This limits the generalizability of these results to the general population or athlete cohorts who are typically not exposed to blast mechanisms or military training. Many of the participants had missing data for follow-up (n = 21; 22%), resulting in a potential bias of PCSS change scores, so only participants with complete data were retained for analysis in the present study. Adherence to the prescribed therapies was not reliably available for this analysis and could help explain differences between groups. The wide range of time since last mTBI for the included participants may have introduced a temporal variance, which we were unable to account for because of the smaller sample size.

Conclusion

In this secondary analysis of a prospective interventional trial in mTBI, higher pre-intervention levels of UCH-L1 served as a predictive biomarker of greater capacity to improve from targeted intervention for chronic issues related to mTBI. Higher pre-intervention levels of p-tau may reflect a biomolecular environment resistant to symptom improvement from chronic mTBI-related symptoms. Future research is needed to confirm and extend this finding in larger cohorts with generalizable populations, given that UCH-L1 and p-tau accounted for ∼10% of the variance in symptom improvement from a targeted intervention in a small cohort of participants with chronic issues from mTBI.

Footnotes

Authors' Contributions

Shawn R. Eagle: conceptualization (lead); writing original draft (lead); review and editing (equal). Ava M. Puccio: methodology (co-lead); review and editing (equal). Denes V. Agoston: conceptualization (supporting); methodology (co-lead); formal analysis (supporting); review and editing (equal). Michael Mancinelli: methodology (supporting); review and editing (equal). Rachel Nwafo: methodology (supporting); review and editing (equal). Peyton McIntyre: methodology (supporting); review and editing (equal). Allison Agnone: methodology (supporting); review and editing (equal). Savannah Tollefson: methodology (supporting); review and editing (equal). Michael Collins: methodology (supporting); review and editing (equal). Anthony P. Kontos: methodology (supporting); review and editing (equal). Walter Schneider: methodology (supporting); review and editing (equal). David O. Okonkwo: conceptualization (supporting); methodology (supporting); review and editing (equal). Ryan J. Soose: conceptualization (supporting); methodology (supporting); review and editing (equal).

Funding Information

This work was funded by the Department of Defense (W81XWH-14-2-0002).

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

Abbreviations Used

References

Zaloshnja

, Miller

, Langlois

, et al. Prevalence of long-term disability from traumatic brain injury in the civilian population of the United States, 2005. J Head Trauma Rehabil, 2008; 23(6):394–400.

, Chan

, Carruthers

. Incidence, prevalence, costs, and impact on disability of common conditions requiring rehabilitation in the United States: stroke, spinal cord injury, traumatic brain injury, multiple sclerosis, osteoarthritis, rheumatoid arthritis, limb loss, and back pain. Arch Phys Med Rehabil, 2014; 95(5):986–995.e1.

Korley

, Jain

, Sun

, et al. Prognostic value of day-of-injury plasma GFAP and UCH-L1 concentrations for predicting functional recovery after traumatic brain injury in patients from the US TRACK-TBI cohort: an observational cohort study. Lancet Neurol, 2022; 21(9):803–813; doi: 10.1016/s1474-4422(22)00256-3

Maas

AIR

, Fitzgerald

, Gao

, et al. Traumatic brain injury over the past 20 years: research and clinical progress. Lancet Neurol, 2022; 21(9):768–770; doi: 10.1016/s1474-4422(22)00307-6

FDA-NIH Biomarker Working Group. BEST (Biomarkers, EndpointS, and other Tools) Resource. U.S. Food and Drug Administration (Silver Spring, MD), National Institutes of Health (Bethesda, MD); 2016.

Eagle

, Kontos

, Collins

, et al. Targeted intervention improves symptoms and impairments in patients with mild traumatic brain injury with chronic symptom: a prospective, multiple interventional research trial. J Spec Oper Med, 2021; 21(2):61–66.

Gyorgy

, Walker

, Wingo

, et al. Reverse phase protein microarray technology in traumatic brain injury. J Neurosci Methods, 2010; 192(1):96–101; doi: 10.1016/j.jneumeth.2010.07.029

Gyorgy

, Ling

, Wingo

, et al. Time-dependent changes in serum biomarker levels after blast traumatic brain injury. J Neurotrauma, 2011; 28(6):1121–1126; doi: 10.1089/neu.2010.1561

Kamnaksh

, Kwon

, Kovesdi

, et al. Neurobehavioral, cellular, and molecular consequences of single and multiple mild blast exposure. Electrophoresis, 2012; 33(24):3680–3692; doi: 10.1002/elps.201200319

10.

Kovesdi

, Kamnaksh

, Wingo

, et al. Acute minocycline treatment mitigates the symptoms of mild blast-induced traumatic brain injury. Front Neurol, 2012; 3:111; doi: 10.3389/fneur.2012.00111

11.

Kwon

, Kovesdi

, Gyorgy

, et al. Stress and traumatic brain injury: a behavioral, proteomics, and histological study. Front Neurol, 2011; 2:12; doi: 10.3389/fneur.2011.00012

12.

Rostami

, Davidsson

, Ng

, et al. A model for mild traumatic brain injury that induces limited transient memory impairment and increased levels of axon related serum biomarkers. Front Neurol, 2012; 3:115; doi: 10.3389/fneur.2012.00115

13.

Ahmed

, Kamnaksh

, Kovesdi

, et al. Long-term consequences of single and multiple mild blast exposure on select physiological parameters and blood-based biomarkers. Electrophoresis, 2013; 34(15):2229–2233; doi: 10.1002/elps.201300077

14.

Wright

, Trezise

, Kamnaksh

, et al. Behavioral, blood, and magnetic resonance imaging biomarkers of experimental mild traumatic brain injury. Sci Rep, 2016; 6:28713; doi: 10.1038/srep28713

15.

Wright

, Brady

, Kamnaksh

, et al. Repeated mild traumatic brain injuries induce persistent changes in plasma protein and magnetic resonance imaging biomarkers in the rat. Sci Rep, 2019; 9(1):14626; doi: 10.1038/s41598-019-51267-w

16.

Major

, McDonald

, O'Brien

, et al. Serum protein biomarker findings reflective of oxidative stress and vascular abnormalities in male, but not female, collision sport athletes. Front Neurol, 2020; 11:549624; doi: 10.3389/fneur.2020.549624

17.

Akbani

, Becker

, Carragher

, et al. Realizing the promise of reverse phase protein arrays for clinical, translational, and basic research: a workshop report: the RPPA (Reverse Phase Protein Array) society. Mol Cell Proteomics, 2014; 13(7):1625–1643; doi: 10.1074/mcp.O113.034918

18.

Day

, Thompson

. UCHL1 (PGP 9.5): neuronal biomarker and ubiquitin system protein. Prog Neurobiol, 2010; 90(3):327–362.

19.

DeKosky

, Asken

. Injury cascades in TBI-related neurodegeneration. Brain Inj, 2017; 31(9):1177–1182.

20.

Tramutola

, Di Domenico

, Barone

, et al. It is all about (U) biquitin: role of altered ubiquitin-proteasome system and UCHL1 in Alzheimer disease. Oxid Med Cell Longev, 2016; 2016:2756068; doi: 10.1155/2016/2756068

21.

Maraganore

, Lesnick

, Elbaz

, et al. UCHL1 is a Parkinson's disease susceptibility gene. Ann Neurol, 2004; 55(4):512–521.

22.

Polanco

, Li

, Bodea

, et al. Amyloid-β and tau complexity—towards improved biomarkers and targeted therapies. Nat Rev Neurol, 2018; 14(1):22–39.

23.

Kenney

, Qu

, Lai

, et al. Higher exosomal phosphorylated tau and total tau among veterans with combat-related repetitive chronic mild traumatic brain injury. Brain Inj, 2018; 32(10):1276–1284.

24.

Shultz

, Wright

, Zheng

, et al. Sodium selenate reduces hyperphosphorylated tau and improves outcomes after traumatic brain injury. Brain, 2015; 138(5):1297–1313.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.01 MB

Evaluating Targeted Therapeutic Response With Predictive Blood-Based Biomarkers in Patients With Chronic Mild Traumatic Brain Injury

Abstract

Introduction

Methods

Design and participants

Blood biomarker analysis

Outcome measures

Post-Concussion Symptom Scale

Statistical analysis

Results

Demographics

Univariable regression models

Multi-variable regression model

Discussion

Limitations

Conclusion

Footnotes

Authors' Contributions

Funding Information

Author Disclosure Statement

Supplementary Material

Abbreviations Used

References

Supplementary Material