Abstract
Our objective was to determine if relative concentrations of neurodegenerative and inflammatory biomarkers differed between traumatic brain injury (TBI) and age-matched controls. Among individuals with TBI, we sought to evaluate differences in biomarker expression by injury severity, investigate changes in biomarkers over the first 2 weeks post-injury, and assess associations with functional outcome. Plasma samples from 126 individuals with TBI were collected at day 1 (n = 105) and 2 weeks (n = 67) post-injury, and 34 healthy controls had a single blood draw. TBI severity was defined using the Glasgow Coma Scale (GCS). The Alamar NULISAseq™ CNS Panel was used to obtain expression levels of 122 different proteins associated with neurodegenerative and/or inflammatory processes. Protein levels were compared between groups using adjusted differential expression analysis. Principal component (PC) analysis was performed, and associations between PCs and functional outcome (Glasgow Outcome Scale–Extended, score 7–8 vs. 1–6) were evaluated using adjusted logistic regression. Participants were a median age of 33 years. Compared with controls, day-1 post-injury samples had 17 upregulated and 10 downregulated proteins, with glial fibrillary acidic protein (GFAP), serum amyloid A1, and interleukin-6 showing the greatest increased expression in TBI. At 2 weeks post-injury, 3 proteins were elevated compared with controls (neurofilament light, neurofilament heavy, GFAP) and no proteins were downregulated. When stratified by TBI severity, there were 15 upregulated day-1 post-injury proteins in individuals presenting with GCS 3–12 compared with GCS 13–15. In PC analysis, the first 4 PCs accounted for a total of 45.7% of the variance. Among individuals with TBI, PC2 was associated with higher odds of better 2-week functional outcome (odds ratio [OR] = 1.20, 95% confidence interval [CI] = 1.02–1.43) and PC3 was associated with lower odds of better 2-week (OR = 0.82, 95% CI = 0.67–0.98) and 6-month (OR = 0.79, 95% CI = 0.63–0.96) functional outcome. In conclusion, our results suggest that proteins of neurodegeneration, synaptogenesis, angiogenesis, and chemokines are associated with TBI outcomes, providing mechanistic insights.
Introduction
Traumatic brain injury (TBI) remains one of the leading causes of injury-related death and disability worldwide, with an estimated 50–60 million incident TBIs occurring globally each year. 1 TBI is increasingly recognized as a chronic condition 2 and has been shown to be associated with long-term sequelae, including increased long-term risk of epilepsy, 3 stroke,4,5 and other neurodegenerative diseases.6–9 However, given the complex and heterogeneous nature of TBI, traditional classification methods, such as injury severity defined by the Glasgow Coma Scale (GCS), remain limited in predicting individual patient prognosis. 10 To this end, the National Institutes of Health (NIH)–National Institute of Neurological Disorders and Stroke (NINDS) Classification and Nomenclature Initiative recently proposed a new, more patient-centered framework for the classification of acute TBI that is anchored in four pillars: clinical, blood-based biomarker, imaging, and modifiers (CBI-M).11,12
The blood-based biomarker pillar of the CBI-M framework is currently comprised of three primary biomarkers, glial fibrillary acidic protein (GFAP), ubiquitin C-terminal hydrolase L1 (UCHL1), and S100 calcium-binding protein B (S100B), that are cleared by the US Food and Drug Administration (GFAP and UCHL1) and/or are CE certified in Europe (GFAP, UCHL1, and S100B) for clinical use in the acute (within 24 h) injury setting to aid in the diagnosis of TBI. 11 No biomarkers are currently approved for clinical use in the subacute (weeks) or chronic (months-years) time periods after TBI. Plasma biomarkers of amyloid, tau, astrogliosis, and axonal degeneration are biomarkers of growing interest given their acute and long-term associations with TBI and potential prognostic value.13–17 However, the advent of new highly sensitive, high-dimensional biomarker assays has the potential to gain further insight into the heterogenous TBI pathophysiology and may inform prognosis and future therapeutic targets.
The objective of this study was to assess if relative concentrations of neurodegenerative and inflammatory biomarkers measured by the Alamar NUcleic Acid Linked Immuno-Sandwich Assay (NULISAseq™) CNS disease panel 120 18 differed between TBI and age-matched controls in the acute (day 1) and subacute (2 weeks) post-injury time periods. In addition, among individuals with TBI, we sought to evaluate differences in biomarker levels by injury severity, to investigate changes in biomarkers over the first 2 weeks post-injury, and to assess associations with subacute (2 weeks) and chronic (6 months) functional outcomes.
Methods
Study population
Individuals over 18 years presenting with an acute nonpenetrating TBI, as per the American Congress of Rehabilitation Medicine criteria (head impact with positive head computed tomography and/or one or more of the following: loss of consciousness, post-traumatic amnesia, altered mental status, focal neurological deficit), 19 to a single Level 1 trauma center in Philadelphia, Pennsylvania, USA, were eligible for enrollment within 72 h of injury. Exclusion criteria included pre-existing neurological disorders, prisoners or individuals in police custody, and pregnancy. Healthy controls from the local community, as well as friends and family of TBI participants, were recruited, given they did not have a TBI within the prior year. All participants or their legally authorized representative provided written informed consent, and the study was approved by the University of Pennsylvania Institutional Review Board.
A random subset of age-matched (on a group level) TBI and control participants with available plasma samples was selected for inclusion in the present analysis. Specifically, plasma from 126 total individuals with TBI was assayed from samples collected at day 1 (n = 105) and 2 weeks (n = 67) post-injury timepoints. In addition, plasma samples collected at study enrollment from 34 healthy controls were assayed.
Clinical variables
Demographic and injury-related characteristics (for participants with TBI) were ascertained at study enrollment. Age (continuous), sex (female; male), race (Black; White; Other), education (high school graduate or less; more than high school, unknown), and history of prior TBI (yes; no; unknown) were self- or proxy-reported. Injury-related characteristics were ascertained from electronic health records and included GCS score (score 13–15; score 3–12), radiological signs of intracranial injury on head CT (positive; negative), injury cause (fall; road traffic accident; other [nonintentional injury or violence/assault; unknown]), loss of consciousness (yes; suspected; no), and post-traumatic amnesia (yes; suspected; no). Among participants with TBI, 2-week and 6-month functional outcomes were assessed using the Glasgow Outcome Scale–Extended (GOSE, categorized as 1–6 [dead; vegetative state; lower/upper severe disability; lower/upper moderate disability] vs. 7–8 [lower/upper good recovery]). 20
Blood sample collection, processing, and storage
EDTA plasma samples were collected from participants with TBI by peripheral venipuncture at two timepoints post-injury (day 1 and 2 weeks post-injury). The median time of collection for the first blood draw (i.e., day 1) was 1 day following injury (interquartile range [IQR]: 1–1), and for the second blood draw (i.e., 2 weeks) it was 17 days post-injury (IQR: 15–26). Control participants had one blood draw at the time of enrollment. Samples were incubated at room temperature for 30 min after collection. Tubes were then centrifuged at 1500 × g for 15 min at room temperature. Plasma samples were then aliquoted into 500 µL cryovials and stored in −80°C freezers until shipment over dry ice to Alamar Biosciences for biomarker measurement (Fremont, California, USA).
Alamar NULISAseq CNS disease panel 120
Relative expression levels of 124 neurodegenerative and inflammatory proteins were measured using the Alamar NULISAseq CNS disease panel 120 (Supplementary Table S1). The Alamar NULISAseq assays utilize matched capture and detection antibody pairs, one conjugated to partly double-stranded DNA with a poly-A oligonucleotide tail and the other with a biotinylated oligonucleotide, to produce an immunocomplex when the targeted protein is present (both antibodies in a pair are conjugated with part of a sequence specific to each target). The immunocomplex undergoes a series of capture and release steps, which reduce noise and increase sensitivity, enabling a final amplification of the sequence specific for each target protein, which is quantified using next-generation sequencing. 18 Our samples were run in two batches. Batch one was run on two plates and had a median intraplate coefficient of variation (CV) of 7.8% and a median interplate CV of 9.1%. Batch two was run on one plate with a median intraplate CV of 6.2%.
Expression values from the Alamar panel were normalized using both an internal control and an interplate control. Normalized expression values were log base 2 transformed, resulting in NULISA Protein Quantification units. Median normalization was performed using 4 TBI and 4 control samples that were run in both batches to control for batch effects (Supplementary Fig. S2). Proteins where more than 50% of the samples had expression values below the limit of detection were excluded from analysis (n = 1: pleiotrophin). We additionally excluded APOE4 from analysis, since it is only measured dichotomously. This left 122 proteins included in the present analyses.
Additional biomarker data
A subset of participants’ samples had been previously quantified using Meso Scale Discovery (MSD, Rockville, Maryland, USA) and/or Quanterix (Billerica, Massachusetts, USA) assays. A total of 87 plasma samples from 68 individuals with TBI (day 1 post-injury) and 19 controls had data on the following MSD assay panels: S-plex neurology panel 1, S-plex proinflammatory panel 1, V-plex angiogenesis panel 1, V-plex vascular injury panel 2, V-plex cytokine panel 1, and V-plex chemokine panel 1. There were 34 overlapping proteins between the MSD and Alamar assays that were matched using UniProt identification codes. A total of 93 serum samples from 72 individuals with TBI (day 1 post-injury) and 21 controls had Quanterix Neurology 4-plex B assay measurements performed on the Simoa HD-X platform, 71 of which also had data from the Quanterix pTau-181 Advantage V2.1 assay. These five proteins overlapped between the Quanterix and Alamar assays (matched using UniProt identification codes). Biomarker data obtained from both the MSD and Quanterix assays were expressed in pg/mL and were log base 2 transformed for analyses.
Statistical analysis
Participant characteristics were summarized using medians and IQRs for continuous variables and numbers and percentages for categorical variables. Spearman’s correlations were used to compare the Alamar biomarker data to biomarker values from the MSD and Quanterix assays.
Differential expression analyses were performed to compare protein expression among individuals with TBI (day 1 and 2 weeks post-injury) versus controls. Among individuals with TBI who had biomarker measurements at both day-1 and 2-week timepoints (n = 46), we additionally investigated the acute to subacute post-injury change of the proteins using differential expression analyses. Secondary differential expression analyses were conducted to examine differences between TBI severity, as defined as GCS 3–12 versus 13–15, and to evaluate for differences between CT-positive and CT-negative participants among individuals with TBI and GCS 13–15. In these differential expression analyses, all linear regression models were adjusted by age and sex. In addition, in the model comparing TBI severity groups, we also included radiological signs of intracranial injury on head CT (positive vs. negative) as a covariate. False discovery rate correction was used within each contrast to account for multiple comparisons, and an adjusted two-sided p value <0.05 was considered statistically significant.
Linear principal component (PC) analysis was performed on the 122 Alamar biomarkers measured from day-1 post-injury samples from participants with TBI and samples from controls for data dimensionality reduction. A scree plot was generated to visualize the percent of variance accounted for by each PC and to identify the number of PCs to be used in subsequent analyses. Individual scores for each participant on each PC were calculated based on the eigenvectors for the PC. PC scatter plots were used to evaluate for clustering by TBI versus control status comparing the first, second, third, and fourth PC (PC1, PC2, PC3, and PC4) scores. Loadings, interpretable as the correlation between the protein and PC, were calculated by multiplying the eigenvector by the square root of the eigenvalue and were visualized using Syndromics plots.21,22 Logistic regression adjusted for age and sex was used to estimate the associations of PC1, PC2, PC3, and PC4 scores with 2-week and 6-month GOSE outcomes. Sixty-five participants with TBI had available 2-week GOSE data, and 43 participants had 6-month GOSE outcomes. A two-sided p value <0.05 was considered statistically significant.
All analyses were performed using R version 4.4.1. Differential expression analyses were performed using the limma package. 23
Results
Participant characteristics
The median age of the 126 individuals with TBI was 33.0 years (IQR: 24.0, 52.8); 70.6% were male (Table 1). The majority of participants with TBI presented with GCS 13–15 (84.1%) and 57.9% had CT findings positive for radiological evidence of intracranial injury. The median age in the control group was 33.5 years (IQR: 26.8, 47.3), and 50.0% were male. Comparing participants by TBI severity (GCS 13–15 vs. 3–12) and by the presence versus absence of trauma-related intracranial CT findings (among individuals with GCS 13–15), differences in age between the two groups were larger (Supplementary Table S2).
Participant Characteristics Stratified by Traumatic Brain Injury Status
GCS, Glasgow Coma Scale; TBI, traumatic brain injury.
Spearman’s Correlations of Measured Proteins Between Different Assays
CXCL10, C-X-C motif chemokine 10; FGF2, fibroblast growth factor 2; GFAP, glial fibrillary acidic protein; IFNG, interferon-gamma; IL, interleukin; MAPT, microtubule-associated protein tau; pTau, phosphorylated Tau; SAA1, serum amyloid A-1; VCAM1, vascular cell adhesion protein 1; VEGFA, vascular endothelial growth factor A; VEGFD, vascular endothelial growth factor D.
Correlations between assays
Twenty-five of the 34 proteins measured using both Alamar and MSD assays had moderate to very strong (Spearman’s correlation [ρ] > 0.50) positive monotonic relationships (Table 2 and Supplementary Fig. S1). The highest Spearman’s correlations were observed for GFAP (ρ = 0.97), interleukin-6 (IL-6, ρ = 0.95), fibroblast growth factor 2 (FGF2, ρ = 0.94), neurofilament light chain (NEFL [NfL], ρ = 0.93), thymus and activation-regulated chemokine (CCL17, ρ = 0.92), and interferon-gamma (IFNG, ρ = 0.90). In contrast, there was weak to no correlation with IL-2, IL-4, vascular cell adhesion molecule 1 (VCAM1), and vascular endothelial growth factor D (VEGFD). Comparing proteins measured using both Quanterix and Alamar assays, all five biomarkers had Spearman’s correlation coefficients greater than 0.50, with GFAP (ρ = 0.96) and NEFL (ρ = 0.95) having the strongest correlations.
Differential expression analyses
In differential expression analyses, there were 17 upregulated and 10 downregulated proteins when comparing biomarkers from day-1 post-TBI samples to controls (Fig. 1A). GFAP, serum amyloid A-1 (SAA1), and IL-6 were higher in the day-1 TBI samples compared with controls, having the greatest log-fold differentials of 4.18 (95% confidence interval [CI]: 3.21–5.16), 3.27 (95% CI: 2.50–4.04), and 3.14 (95% CI: 2.36–3.92), respectively (Fig. 2). Neuronal pentraxin receptor (NPTXR), acetylcholinesterase (ACHE), and corticoliberin (CRH) were the most significantly downregulated proteins in day-1 post-TBI samples compared with controls. At 2 weeks post-injury, only three proteins remained significantly differentiated between TBI and controls, including NEFL (log-fold change: 2.67, 95% CI: 2.03–3.30), neurofilament heavy chain (NEFH, log-fold change: 3.42, 95% CI: 1.96–4.88), and GFAP (log-fold change: 0.86, 95% CI: 0.46–1.25), which were each elevated in the TBI samples compared with the controls (Fig. 1B and Fig. 3). Although GFAP was still upregulated at 2 weeks post-injury, it decreased in comparison to day-1 expression among individuals with TBI who had biomarker measurements at both day-1 and 2-week timepoints (n = 46) (Fig. 1C). In contrast, NEFL and NEFH significantly increased at 2 weeks post-injury among individuals with TBI when compared with their day-1 measures.
Volcano plots showing significantly upregulated proteins in red and significantly downregulated proteins in blue.
Differential expression analysis comparing day-1 post-TBI samples to controls.
Differential expression analysis comparing TBI 2-week samples to controls.
In secondary differential expression analyses comparing TBI severity, there were 15 upregulated proteins at day 1 post-injury among participants with GCS 3–12 when compared with GCS 13–15 (Fig. 1D). GFAP, IL-6, and NEFL had the largest log fold changes, and several tau proteins were also upregulated, including microtubule-associated protein tau (MAPT), phosphorylated tau 217 (pTau-217), pTau-181, and pTau-231. Among participants with GCS 13–15 TBI, when comparing CT positive to CT negative at day 1, there were 4 upregulated and 24 downregulated proteins (Fig. 1E).
PC analyses
In PC analysis using day-1 post-TBI samples and control samples, PC1 accounted for 19.8% of the variation, PC2, 11.5%, PC3, 9.7%, and PC4, 4.6% for a total of 45.7% across the first four PCs (Fig. 4). Using scatter plots to visually assess the separation between the controls and TBI cases, PC1 and PC4 did not provide much discrimination. However, PC2 and PC3 showed some separation, with controls having slightly higher PC2 scores and lower PC3 scores. The five proteins with the highest loadings to PC1 included phosphoglycerate kinase 1 (PGK1), superoxidase dismutase (SOD1), IL-18, malate dehydrogenase (MDH1), and peroxiredoxin 6 (PRDX6), which were all negatively correlated with PC1 (Fig. 5). The five proteins with the highest loadings to PC2 included agrin (AGRN), C-X-C motif chemokine 10 (CXCL10), NPTXR, eotaxin (CCL11), and C-C motif chemokine 22 (CCL22), which were all positively correlated with PC2. NEFL, MAPT, pTau-181, fatty-acid-binding protein (FABP3), and pTau-231 had the highest loadings and were positively correlated with PC3. In PC4, VEGFD, periostin (POSTN), gamma-enolase (ENO2), and APOE had loadings greater than 0.5.
Principal component (PC) plots.
Principal component analysis.
Among individuals with TBI, in age- and sex-adjusted logistic regression models, PC2 was associated with higher odds of better 2-week functional outcomes (GOSE score 7–8 vs. 1–6, odds ratio [OR]: 1.20, 95% CI: 1.02–1.43), and PC3 was associated with lower odds of better 2-week functional outcomes (OR: 0.82, 95% CI: 0.67–0.98) (Table 3). PC3 was also associated with lower odds of better functional outcomes at 6 months (OR: 0.79, 95% CI: 0.63–0.96). PC1 and PC4 were not associated with 2-week or 6-month functional outcomes.
Results of Logistic Regression Showing Odds of a GOSE Score of 7 or 8 at 2 Weeks and 6 Months Post-injury
CI, confidence interval; GOSE, Glasgow Outcome Scale–Extended; OR, odds ratio.
Discussion
In this cohort of individuals with TBI presenting to a Level 1 trauma center, we assessed neurodegenerative and inflammatory proteomic profiles in the acute (1 day) and subacute (2 weeks) post-injury time periods. Utilizing the Alamar NULISAseq CNS disease panel, we identified key plasma biomarkers that were differentiated between controls and individuals who experienced a TBI, including GFAP, SAA1, and IL-6 on day 1 post-injury and NEFL, NEFH, and GFAP at 2 weeks post-injury. In addition, at day 1 post-injury, we found differences of protein expression between TBI severity groups defined by GCS, including differential expression of GFAP, NEFL, IL-6, and tau-related proteins. Our study also found that proteins of synaptogenesis, angiogenesis, and chemokines were associated with better 2-week outcomes, while tau and other primarily neurodegenerative proteins were associated with worse 2-week and 6-month outcomes. Taken together, the results of this study provide insight into the pathological processes occurring in the acute and subacute post-TBI time periods and the mechanisms underlying the sequelae of TBI.
Given the limited use of the Alamar panel in TBI research to date, we first assessed the correlation between the Alamar expression values and measurements obtained from the MSD and Quanterix panels, which have been used extensively in prior TBI studies.24–27 The reporting of assay results in relative units by Alamar and in absolute units by MSD and Quanterix meant that we were only able to assess correlation between assays. When comparing Alamar to MSD, the majority of the proteins had moderate to very strong positive monotonic relationships. Of note, there were several proteins that had very weak correlations, including IL-2, IL-4, VCAM1, and VEGFD. This finding is consistent with another study in the BIO-AX-TBI cohort that compared Alamar to OLINK and reported weak correlations of IL-2 and IL-4. 14 The five proteins assayed on both Quanterix and Alamar platforms all had moderate to very strong correlation coefficients. Taken together, these results support the reliability of these measurements across these different assays, especially for key biomarkers of interest in TBI (e.g., GFAP, NEFL, and IL-6).
Consistent with current literature, our findings using the Alamar NULISAseq CNS panel showed changes in blood-based biomarkers of neurodegeneration and inflammation in the acute and subacute post-TBI setting.12–17,24,28–33 Specifically, glial and neuronal filament proteins (GFAP and NEFL) were upregulated in TBI at day 1 post-injury and remained elevated at 2 weeks post-injury. However, NEFL increased in those with TBI from day 1 to 2 weeks, whereas GFAP decreased although it remained elevated in comparison to controls. Our findings are consistent with the literature where GFAP is elevated following TBI and peaks at approximately 20 h following injury. 12 Likewise, our findings with NEFL align with previous studies where NEFL is upregulated within 24 h after injury but continues to rise and is estimated to peak between 10 days and 6 weeks following injury. 31 In contrast to studies showing increased UCHL1 and S100B in the acute post-injury setting,34,35 our study did not find increased UCHL1 or S100B comparing day-1 post-TBI to controls. It is possible that this difference may be due to the earlier peak (within 8 h post-injury) of both of these biomarkers34,35 in combination with blood draws potentially occurring later in the acute time period in our study. However, the exact time between injury and first blood draw was not always available from the medical records, so we were unable to fully evaluate this possibility. Further supporting this potential explanation for our study not finding increased UCHL1 or S100B among day-1 post-TBI samples compared with controls, our eligibility criteria allowed for enrollment for up to 72 h post-injury. At 2 weeks post-injury, in addition to GFAP and NEFL, the other differentiated biomarker was NEFH, a protein responsible for axonal support. NEFH has shown to increase following TBI, and prior studies have shown associations of elevated phosphorylated NEFH with poor 6-month outcomes.32,33
Regarding inflammatory proteins, SAA1, IL-6, c-reactive protein (CRP), and IL-10 were upregulated in individuals with TBI at day 1 but not 2 weeks post-injury compared with controls. These inflammatory proteins have each been previously shown to be increased in the acute TBI setting.14,24,28–30 SAA1 and CRP are acute-phase proteins produced by the liver in response to trauma and injury. 36 In contrast to our study, other studies have shown potential discriminating abilities of these biomarkers for TBI severity categorized by GCS score as well as for the presence of trauma-related intracranial CT findings.24,30 Consistent with prior work, we found that IL-6, a pro-inflammatory cytokine, and IL-10, an anti-inflammatory cytokine, were differentially expressed by TBI severity (defined by GCS), with higher levels observed with more severe injuries. 14 These findings likely reflect the complex interplay between secondary injury and tissue repair from inflammatory cascades after TBI.
Further exploring by TBI severity, we found that individuals with TBI GCS 3–12 had upregulated expression of NEFL, GFAP, and IL-6 compared with those with GCS 13–15. In addition, MAPT, pTau-217, pTau-181, pTau-231 were also notably elevated in GCS 3–12 compared with GCS 13–15. Prior literature has been mixed regarding associations of various tau isoforms with TBI and TBI severity in the acute and subacute post-injury time periods, with some studies reporting associations, 14 and others not. 37
Using PC analysis, the first four PCs accounted for 45.7% of the variance. The variance captured by PC1 and PC4 did not help to distinguish cases and controls, nor did it show any association with 2-week and 6-month GOSE outcomes. In contrast, PC2 was associated with higher odds of a favorable GOSE score (7–8) at 2 weeks, while PC3 was associated with lower odds of better functional recovery at 2 weeks and at 6 months. In relationship to TBI, PC2 could represent a coordinated response between regenerative and inflammatory processes as synaptogenic and angiogenic proteins (AGRN, NPTXR, TEK, and platelet-derived growth factor receptor beta [PDGFRB]) as well as chemokines that recruit other immune cells (CXCL10, CCL11, CCL22) are highly correlated with this PC. In pre-clinical models, AGRN has been shown to be increased in injury-affected regions of the brain. 38 In our models, we found NPTXR to be downregulated in day-1 TBI samples when compared with controls, but it was upregulated in 2-week TBI samples when compared with day 1, suggesting that its role in synaptic organization may occur in the subacute period after injury. Higher levels of TEK (Tie2) have been previously associated with less severe injury, 39 and PDGFRB expressed on pericytes had a biphasic response in a murine TBI model, 40 suggesting the role of both of these proteins in vascular stability and integrity following injury. Whereas CXCL10, CCL11, and CCL22 have been previously shown to be upregulated following TBI41–43 indicating an integrated response between inflammation and recovery. PC3 is mainly characterized by proteins of tau pathology and neurodegeneration; proteins that loaded higher in PC3 included NEFL, MAPT, pTau-181, pTau-231, IL-6, and GFAP. These proteins, particularly NEFL and GFAP, have been consistently reported to be elevated following TBI, and higher levels are typically associated with more severe TBI and worse functional outcomes.12,14,17,24,34
In interpreting the results of this study, several limitations should be considered. Given the largely exploratory nature of our study, samples were selected to be assayed post hoc based on matching the median age between TBI cases and controls. However, considering that many of these proteins are age-dependent, 44 we deemed it an important tradeoff to the potential spurious results that could arise due to a large age gap between TBI cases and controls. Moving forward, appropriately designed age-matched studies accounting for comorbidities should be explored for investigating differences in plasma proteins after TBI. In addition, multicenter studies with larger sample size are needed to corroborate these results and further investigate the relationship between these proteomic findings and outcome data. In addition, our study did not include an orthopedic injury control group, nor do we currently have information regarding extracranial injury in our TBI group. Therefore, it is possible that some of our findings could be attributed to other orthopedic injuries sustained during trauma and not just TBI, as has been suggested in a prior study. 14 Lastly, our study included data from one site; more research is needed to validate our results and to compare these different biomarker assays and develop potential harmonization methods to combine biomarker data across different platforms and studies, increasing sample size and generalizability.
In conclusion, in this single-center cohort of individuals with TBI, our results not only support the broader literature regarding inflammatory and neurodegenerative protein elevations following TBI, but also suggest that proteins of synaptogenesis, angiogenesis, and chemokines are associated with outcomes following TBI. Through use of a sensitive, multiplex assay, we can begin to discern the mechanistic pathways involved in TBI and how proteins interact to impact outcomes.
Transparency, Rigor, and Reproducibility Summary
This study and analysis plan were not formally preregistered. Clinical protocols and outcome procedures are available from the corresponding author upon reasonable request. All data used for this investigation are available from the corresponding author upon reasonable request. All statistical codes and scripts used for this investigation are available from the corresponding author upon reasonable request.
Authors’ Contributions
S.A.: Conceptualization, methodology, formal analysis, visualization, and writing—original draft. R.P.: Writing—review and editing. A.E.W.: Writing—review and editing. C.E.L.: Writing—review and editing. D.K.S.: Writing—review and editing. D.H.S.: Writing—review and editing and funding acquisition. A.L.C.S.: Conceptualization, methodology, writing—review and editing, and supervision. R.D.-A.: Conceptualization, methodology, writing—review and editing, funding acquisition, and supervision.
Supplemental Material
sj-docx-1-ntr-10.1177_2689288X261446294 — Supplemental material for Proteomic Changes of Neurodegenerative and Inflammatory Plasma Biomarkers Following Traumatic Brain Injury
Supplemental material, sj-docx-1-ntr-10.1177_2689288X261446294 for Proteomic Changes of Neurodegenerative and Inflammatory Plasma Biomarkers Following Traumatic Brain Injury by Sabrina Abbruzzese, Ronit Patel, Alexa E. Walter, Cillian E. Lynch, Danielle K. Sandsmark, Douglas H. Smith, Andrea L.C. Schneider, and Ramon Diaz-Arrastia
Footnotes
Author Disclosure Statement
No competing financial interests exist.
Funding Information
This study was supported by NIH/NINDS grants
Abbreviations
References
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.
