Elevated cerebrospinal fluid galectin-3 and associated cytokines after severe traumatic brain injury in patients

Abstract

This study aimed to investigate the galectin-3 and associated cytokines levels in the cerebrospinal fluid (CSF) of severe traumatic brain injury (sTBI) patients. Temporal CSF expression of galectin-3 and associated cytokines levels in sTBI patients within 1-week post-injury were studied using the multiplex bead array. STBI patient group was stratified using the Modified Rankin Score (mRS) into 3 groups: mRS 6 (died), mRS 5 (severely disabled) and mRS 1–4 (mild-to-moderately disabled) group. Analysis for bead array data using Kruskal-Wallis test with post hoc Dunn's multiple comparisons test, and temporal changes and correlation analysis using Spearman's correlation were carried out. At day 1 post-injury, CSF galectin-3 and interleukin-6 (IL-6), interleukin-10 (IL-10), cysteine-cysteine motif chemokine ligand-2 (CCL-2), and cysteine-cysteine motif chemokine ligand-20 (CCL-20), but not interleukin-1β (IL-1β) and tumor necrosis factor (TNF-α) levels were significantly elevated in mRS 5 group compared to non-TBI controls. Temporal correlation analysis at 1–7 days showed decreased IL-10 level in the mRS 6 group, decreased IL-10 and CCL-2 levels in mRS 5 group, and decreased IL-6, CCL-2, and CCL-20 levels in the mRS 1–4 group. Receiver operating characteristic curve analyses revealed a significant area under the curve for comparison between mRS 6 and mRS 5 groups for galectin-3 and IL-6. No significant differences in sex, age, Glasgow Coma Scale score, C-reactive protein levels and types of TBI-induced hemorrhages were observed between the groups. CSF galectin-3 and associated cytokines, especially IL-6, CCL-2 and CCL-20 levels were different within sub-groups of sTBI patients, suggesting their potential use in sTBI prognostics.

1. Introduction

Traumatic brain injury (TBI) is a devastating disease affecting approximately 69 million people worldwide, with approximately 8% of these resulting in severe TBI (sTBI).^[1] Apart from sTBI being a major cause of mortality associated with a 30% to 70% death rate within 6 months, a large proportion of sTBI survivors suffer from debilitating disabilities, resulting in dependence on others.^[2] Therefore, identifying early prognostic factors would greatly benefit the stratification and management of sTBI, ultimately benefiting patients, families and healthcare providers. One strategy is to study the pathophysiology of sTBI in patients, of which neuroinflammation is a key process.

Recently, we reported that galectin-3, a 29–35 kDa β-galactoside binding lectin, is an important alarmin in the neuroinflammatory process that promoted neurodegeneration in a moderate TBI mouse model.^[3] Furthermore, galectin-3 has been suggested to function as a master regulator of other cytokines such as interleukin-1β (IL-1β),^[4] interleukin-6 (IL-6),^[5] interleukin-10 (IL-10),^[6] Tumor necrosis factor (TNF-α),^[7] cysteine-cysteine motif chemokine ligand-2 (CCL-2; formerly monocyte chemoattractant protein-1, MCP-1),^[8,9] and cysteine-cysteine motif chemokine ligand-20 (CCL-20; formerly macrophage inflammatory protein-3, MCP-3).^[10]

The aim of the study was to identify the cerebrospinal fluid (CSF) expression levels of galectin-3 and a selective set of associated cytokines, namely IL-1β, IL-6, IL-10, TNF-α, CCL-2, and CCL-20 in sTBI patients over a period of 7 days. CSF from non-TBI patients who did not have severe physical traumatic brain injury were used as controls.

2. Methods

2.1. Ethics

The present study was approved by the Institutional Review Board of Chang Gung Memorial Hospital in Taiwan (IRB No. 201700537B0C504). Written informed consent was obtained from subjects or legal representatives prior to study procedures. All methods were performed in accordance with the relevant guidelines and regulations.

2.2. Participants

Prospective data were collected from 40 patients admitted to the neurosurgery department of Chang Gung Memorial Hospital in Taiwan between April 2017 and December 2019. Inclusion criteria were aged between 18 to 90 years old, had evidence of intracranial damage on a computed tomography (CT) scan, and had brain injury requiring neurosurgical intervention. Individuals with head injuries but no evidence of intracranial bleeding (e.g., depressed skull fracture) were excluded. The reasons for excluding under 18 years of age are that it has been shown cytokine variation occurs between children and adults,^[11] and many clinical TBI studies only involve adult patients.^[12] STBI patients were managed in accordance with the standard clinical practice guidelines for sTBI in Taiwan, which recommended intracerebral pressure (ICP) monitoring and external ventricular drain placement for ICP control via CSF drainage.^[13] Additional surgical interventions were performed on some patients based on the clinical judgment of the attending neurosurgeon. C-reactive protein (CRP) level in peripheral blood was also measured on admission. The reference non-TBI control group in the study consisted of non-TBI patients with the criteria 1-aged between 18 and 90 years old, 2-had no acute evidence of intracranial damage on a CT scan, and 3- received ventriculoperitoneal shunt placement for normal pressure hydrocephalus.

2.3. Collection, processing and analysis of CSF samples

CSF from sTBI patients (n = 35 subjects) were collected directly via the extraventricular tube every other day from day 1 to 7 post-TBI. CSF sampling was chosen over blood sampling as CSF is minimally influenced by peripheral cytokine production and liver clearance.^[14] A single CSF sample was obtained from each non-TBI patient (n = 5 subjects) during ventriculoperitoneal shunt implantation for treatment of normal pressure hydrocephalus. The acquired CSF samples were centrifuged, aliquoted, and frozen at–80 °C until subsequent analysis.

In order to efficiently, specifically, and simultaneously study various CSF inflammatory protein markers, a customized human Procarta 7-plex^TM panel multiplex bead array (ThermoFisher Scientific, Cat No. PPX-07-MXYMKCP, Lot No. 231624-000) was utilized to analyze the following 7 protein targets: galectin-3, IL-1β, IL-6, IL-10, TNF-α, CCL-2, and CCL-20 to avoid batch variations between runs.^[15] A Luminex 200™ clinical diagnostic system with xPONENT™ 3.1 software was used to generate standard curves for all targets, and the lower and upper limits of quantification at 30% bias were derived. Readings from sTBI and non-TBI patients were calibrated against standard curves and within the lower and upper limits of quantification to obtain inflammatory marker concentration values.

2.4. Functional outcomes

Modified Rankin Scale (mRS), which is a global measurement of disability widely used in evaluating neurological recovery, including after TBI, was used.^[16] Although mRS analysis is often dichotomized to mRS 0 to 2 (good outcome) and 3 to 6 (poor outcome), this study focuses specifically on the severe cases of TBI, and there were no patients with mRS 0 (no symptoms) in this study. Therefore, we stratified sTBI patients into 3 groups consisting of mRS 6 (died), mRS 5 (severely disabled), and mRS 1–4 (mild-to-moderately disabled) at 6 months post-injury. The non-TBI patient controls had a range of mRS 1 to 4.

2.5. Statistical analyses

All graphical and statistical analyses were performed with GraphPad Prism v8 using non-parametric tests because the data sets did not follow normal distribution according to the Shapiro–Wilk test. Statistical analysis for clinicodemographical factors, galectin-3 and associated cytokines bead array data at day 1 were determined using the Kruskal-Wallis test with post hoc Dunn's multiple comparisons test. For all statistical comparisons, P value was set at <.05 threshold for significance.

Missing data were observed in mRS 6 (died, n = 11), with day 1, 5 and 7 contained 1, 2 and 5 missing data, respectively. In mRS 5 (severely disabled, n = 18), day 1, 5, and 7 contained 2, 1, and 4 missing data, respectively. In mRS 1–4 (mild-moderate disabled, n = 6), day 5 and 7 contained 1 and 3 missing data, respectively. Non-TBI control had only 1 sample due to ethical constrain as these patients do not require further treatment and would require invasive procedure to collect additional sample.

3. Results

3.1. Clinicodemographic factors

The clinical characteristics of a total of 35 sTBI patients and 5 control patients with non-TBI used in this study are summarized in Table 1. There was no significant difference in the age, GCS, various types of TBI-induced hemorrhages, and the systemic inflammatory marker CRP levels between the various sTBI patient groups and non-TBI control group (Table 1). As TBI is a male-dominant disease, there were more males in all groups of sTBI in comparison to a larger proportion of females (80%) in the non-TBI control group (Table 1).

Table 1.

Clinicodemographics of sTBI and non-TBI control patient groups subdivided according to the modified Rankin Scale for neurologic disability.

mRS score (functional outcome)	mRS 6 (died)	mRS 5 (severe disabled)	mRS 1–4 (mild-moderate disabled)	mRS 1–4 (non-TBI control)
Age (yr ± SD)	59.3 (±26.4)	61.0 (±15.9)	36.7 (±13.6)	68.2 (±6.9)
Sex ratio (F:M) (%)	3:8 (27.2%)	1:17 (6%)	2:4 (33.3%)	4:1 (80%)
GCS (±SD)	7.6 (±3.6)	7.2 (±1.6)	7.7 (±2.2)	13.2 (±4.0)
EDH	3/11 (27.3%)	2/18 (1.1%)	2/6 (33.3%)	0/5 (0%)
SDH	7/11 (63.6%)	14/18 (7.8%)	6/6 (100%)	0/5 (0%)
ICH	7/11 (63.6%)	11/18 (61.1%)	4/6 (66.7%)	0/5 (0%)
SAH	10/11 (90.9%)	15/18 (3.3%)	4/6 (66.7%)	0/5 (0%)
IVH	1/11 (9.1%)	3/18 (16.7%)	0/6 (0%)	0/5 (0%)
CRP (mg/mL ± SD)	91.4 (±15.1)	91.5 (±71.9)	29.4 (±15.1)	ND

CRP = C-reactive protein, EDH = epidural hematoma, F = female, GCS = Glasgow coma scale, ICH = intracerebral hemorrhage, IVH = intraventricular hemorrhage, M = male, mRS = Modified Rankin Score, ND = not determined as not part of standard care, SAH = subarachnoid hemorrhage, SD = standard deviation, SDH = subdural hematoma, sTBI = severe traumatic brain injury.

3.2. CSF galectin-3 and associated cytokines at day 1 post-injury

CSF galectin-3 (P = .005), IL-6 (P = .005), IL-10 (P = .004), CCL-2 (P = .004) and CCL-20 (P = .008) levels were significantly greater in the mRS 5 group at day 1 post-injury compared to non-TBI control group (Fig. 1A–E). Furthermore, CSF galectin-3 level was significantly higher in the mRS 5 group compared to the mRS 6 group (P = .049) (Fig. 1A). CSF IL-6 showed a significant increase in the mRS 1–4 group compared to the non-TBI control group (P = .002) (Fig. 1B). The CSF IL-10 levels were significantly higher in the mRS 5 group compared to the non-TBI control group (P = .004) (Fig. 1C). CSF TNF-α and IL-1β levels were not significantly higher between any of the sTBI groups or the non-TBI control group (Fig. 1F–G).

Figure 1.

Galectin-3 and associated cytokine levels in the CSF of sTBI patients and non-TBI patients acting as control. (A) Galectin-3, (B) IL-6, (C) IL-10, (D) CCL-2, (E) CCL-20, (F) TNF-α, and (G) IL-1β levels at day 1 post-injury. STBI group was sub-grouped into their mRS score of mRS 6 (died, blue open circle), mRS 5 (severely disabled, red closed circle) and mRS1–4 (mild-to-moderately disabled, green triangle). Non-TBI group with an mRS of 1–4 was only collected at time 0 during surgical intervention (black square). CCL-2 = cysteine-cysteine motif chemokine ligand-2, CCL-20 = cysteine-cysteine motif chemokine ligand-20, IL-1β = interleukin-1β, sTBI = severe traumatic brain injury, TNF-α = tumor necrosis factor.

3.3. Temporal changes in CSF galectin-3 and associated cytokines from day 1 to 7

There was no significant correlation in the temporal change in the levels of galectin-3 (r² = 0.05–0.71, P = .157–0.775, Fig. 2A), TNF-α (r² = 0.01–0.02, P = .462–0.741, Fig. 2F) and IL-1β (0.04–0.10, P = .050–0.396, Fig. 2G) for all sTBI groups. In contrast, significant negative correlation in the temporal IL-6 level was observed in mRS 1–4 group (r² = −0.27, P = .020) but non-significantly in mRS 6 (r² = 0.01, P = .637) and mRS 5 (r² = 0.0002, P = .905) groups (Fig. 2B). Significant negative correlation in the temporal IL-10 levels was detected in mRS 6 (r² = −0.12, P = .044) and mRS 5 (r² = −0.07, P = .033), but not in mRS 1–4 (r² = 0.15, P = .088) groups (Fig. 2C). Significant negative correlation in the temporal CCL-2 levels was detected in mRS 5 (r² = −0.19, P = .0003) and mRS 1–4 (r² = −0.244, P = .027), but not in mRS 6 (r² = 0.09, P = .080) groups (Fig. 2D). Significant negative correlation in the temporal CCL-20 levels was detected in mRS 1–4 (r² = −0.32, P = .009), but not in mRS 6 (r² = 0.02, P = .459) and mRS 5 (r² = 0.04, P = .119) groups (Fig. 2E).

Figure 2.

Temporal correlation analysis of galectin-3 and associated cytokine levels in the CSF of sTBI patients during the initial 7 days post-injury. (A) Galectin-3, (B) IL-6, (C) IL-10, (D) CCL-2, (E) CCL-20, (F) TNF-α, and (G) IL-1β levels at 1, 3, 5, and 7 d post-injury. STBI group was sub-grouped into their mRS score of mRS 6 (died, blue open circle), mRS 5 (severely disabled, red closed circle) and mRS 1–4 (mild-to-moderately disabled, green triangle). CCL-2 = cysteine-cysteine motif chemokine ligand-2, CCL-20 = cysteine-cysteine motif chemokine ligand-20, IL-1β = interleukin-1β, sTBI = severe traumatic brain injury, TNF-α = tumor necrosis factor.

3.4. Correlation of galectin-3 with associated cytokines

Given that the multiplex bead array assay was able to analyze various protein targets simultaneously, the Spearman correlation between galectin-3 and associated cytokines was studied. IL-6 and CCL-20 levels were the only 2 of the 6 galectin-3 associated cytokines tested that demonstrated a significant positive correlation with galectin-3 levels throughout days 1 to 7 post-injury (Table 2). IL-1β and IL-10 levels also showed a significant positive correlation with galectin-3 levels, but only at days 3 and 5, and days 1 and 5, respectively (Table 2). A significant positive correlation was observed between TNF-α and galectin-3 levels at day 5 post-injury. In contrast, CCL-2 levels did not show any significant correlation with galectin-3 levels at any time point (Table 2).

Table 2.

Correlation of galectin-3 and associated cytokines in CSF from sTBI patients.

Cytokine	Day post-injury	Spearman r	95% CI (LL)	95% CI (UL)	P value	Significance
IL-6	1	0.533	0.222	0.746	.001	**
	3	0.535	0.225	0.747	.001	**
	5	0.586	0.262	0.791	.001	**
	7	0.582	0.144	0.830	.011	*
IL-10	1	0.443	0.119	0.682	.008	**
	3	0.055	−0.304	0.399	.763	ns
	5	0.384	0.002	0.668	.044	*
	7	0.449	−0.037	0.764	.061	ns
CCL-2	1	0.189	−0.164	0.499	.277	ns
	3	0.315	−0.043	0.601	.075	ns
	5	−0.034	−0.412	0.354	.864	ns
	7	0.169	−0.336	0.599	.502	ns
CCL-20	1	0.339	−0.004	0.610	.047	*
	3	0.403	0.059	0.662	.020	*
	5	0.601	0.283	0.800	.0007	***
	7	0.620	0.201	0.847	.006	**
TNF-α	1	−0.090	−0.420	0.260	.607	ns
	3	0.041	−0.316	0.388	.820	ns
	5	0.423	0.056	0.690	.022	*
	7	0.348	−0.175	0.717	.170	ns
IL-1β	1	−0.038	−0.376	0.308	.828	ns
	3	0.358	0.006	0.631	.041	*
	5	0.503	0.141	0.747	.008	**
	7	0.292	−0.217	0.676	.240	ns

Statistical analysis:

^*P < .05;

^**P < .01,

^***P < .001.

CCL-2 = Cysteine-cysteine motif chemokine ligand-2, CCL-20 = cysteine-cysteine motif chemokine ligand-20, CI = confidence intervals, CSF = cerebrospinal fluid, IL-10 = interleukin-10, IL-1β = interleukin-1β, IL-6 = interleukin-6, LL = lower limits, n.s = non-significant, sTBI = severe traumatic brain injury, TNF-α = tumor necrosis factor, UP = upper limits.

3.5. Receiver operating characteristic (ROC) curve analysis

The receiver operating characteristics for the galectin-3 and associated cytokines used to predict between the poor outcomes of mRS 6 and mRS 5 after sTBI are summarized in Table 3. There were significant increased galectin-3 and IL-6 levels observed in mRS 5 compared to mRS 6 at days 1 to 3 post-injury (Table 3). Additionally, there were continued increases in galectin-3 levels in mRS 5 compared to mRS 6 at day 5 post-injury. IL-6 and CCL-20 levels were significantly increased in mRS 5 compared to mRS 1–4 group at day 5 post-injury, and IL-6 level was also significantly detected at day 3 post-injury (Table 4).

Table 3.

Diagnostic power of galectin-3 and associated cytokines within the CSF of sTBI patients that has a comparison between mRS 6 (died) and mRS 5 (severely disabled).

Cytokine	D	AUC	SE	P value	Cutoff (pg/mL)	Sensitivity (%)	Specificity (%)
Galectin-3	1	0.798	0.087	.008**	19454	77.8	81.8
	3	0.764	0.091	.023*	11884	55.6	90.0
	5	0.764	0.091	.023*	21268	55.6	90.0
	7	0.662	0.159	.301	16088	61.5	80.0
IL-6	1	0.733	0.097	.043*	1427	56.3	81.8
	3	0.733	0.095	.044*	1156	61.1	90.0
	5	0.706	0.101	.089	972	58.8	88.9
	7	0.667	0.142	.248	1253	64.3	83.3

Statistical analysis:

^*P < .05;

^**P < .01.

AUC = Area under the curve, D = day post injury, CSF = cerebrospinal fluid, IL-6 = interleukin-6, mRS = Modified Rankin Score, SE = standard error, sTBI = severe traumatic brain injury.

Table 4.

Diagnostic power of galectin-3 and associated cytokines within the CSF of sTBI patients that has a comparison between mRS 5 (severely disabled) and mRS 1–4 (mild-to-moderately disabled).

Cytokine	Day	AUC	SE	P value	Cutoff (pg/mL)	Sensitivity (%)	Specificity (%)
IL-6	1	0.646	0.137	0.302	2443	50.0	81.3
	3	0.796	0.092	0.033*	511	77.8	83.3
	5	0.859	0.084	0.017*	721	64.7	100
	7	0.833	0.112	0.078	1287	100	64.3
CCL-20	1	0.542	0.116	0.764	72.1	50.0	50.0
	3	0.736	0.118	0.089	40.5	66.7	55.6
	5	0.806	0.125	0.042*	24.6	80.0	58.8
	7	0.736	0.164	0.220	58.3	100	33.3

Statistical analysis:

^*P < .05.

AUC = area under the curve, D = day post injury, CCL-20 = cysteine-cysteine motif chemokine ligand-20, CSF = cerebrospinal fluid, IL-6 = interleukin-6, mRS = Modified Rankin Score, SE = standard error, sTBI = severe traumatic brain injury.

4. Discussion

In this study, we demonstrated that CSF galectin-3, and associated cytokines IL-6, IL-10, CCL-2, and CCL-20 levels were significantly increased in sTBI patients with the functional outcome of mRS 5 (severely disabled) rather than mRS 6 (died) group when compared to non-TBI control at day 1 post-injury. Furthermore, galectin-3 level was significantly greater in the mRS 5 group than mRS 6 group, and IL-6 level was significantly higher in mRS 1–4 group (mild-moderate disabled) compared to non-TBI control. Temporal correlation analysis showed that within the mRS 5 group, there were significant negative correlation with time (i.e., 1–7 days) in the IL-10 and CCL-2 levels. In addition, significant negative correlation with time for IL-6, IL-10, CCL-2 and CCL-20 levels were also observed in the mRS1–4 group. However, IL-1β and TNF-α levels were no significantly altered after sTBI.

The increase in CSF galectin-3 level at 1 day supported the data we observed in mice with TBI at 1 day post-injury.^[3] Although changes in CSF galectin-3 levels in sTBI patients have not been previously documented, it has been reported that plasma galectin-3 levels are approximately 3-fold higher in sTBI patients than in controls.^[17] Furthermore, it was reported that a 4.3-fold increase in plasma galectin-3 level was also observed in patients with mild TBI at 1 to 8 hours post-injury compared to non-TBI controls.^[18] The lack of significant increase in galectin-3 level in sTBI patients that have died suggest the brain may be too severely injured to respond. We have previously demonstrated that some sTBI patients that have died within 1 month post-injury had significant loss of microglia in their brain tissue, thus limiting a source of galectin-3 from the brain.^[19] In contrast, in sTBI patients that had a mild-moderate disability, the non-significant increase in galectin-3 level and a non-significant temporal reduction in galectin-3 level suggests the galectin-3 associated neuroinflammation level was limited to induce a detrimental effect on the functional outcome.

In this study, galectin-3-associated cytokines IL-6, IL-10, CCL-2, and CCL-20, but not TNF-α or IL-1β, showed significant increases in sTBI patients with mRS 5 group compared to the non-TBI control group. CSF IL-6, IL-8, IL-10, IL-1β, and TNF-α have been previously reported to be significantly elevated at 6 to 12 hours and subsequently returned to baseline levels by 4 days post-injury.^[20] Similarly, another study reported increased CSF and serum IL-6 levels within the first few days before it plateaued after 1 week.^[21,22] Intriguingly, higher levels of acute and sub-acute CSF IL-6 were associated with higher risks of unfavorable outcomes at 6 and 12 months post-injury.^[15] However, in our study, only the mRS 5 and mRS 1–4 groups had significantly elevated CSF IL-6 levels compared to non-TBI controls at day 1, and only mRS 1–4 group had a significant temporal reduction in IL-6 level. Kumar and colleagues defined unfavorable outcomes using the Glasgow Outcome Scale (GOS), which spans death and severe disability.^[15] However, our data suggests that only the severely disabled group contributed to the elevations in CSF IL-6, instead of the mRS 6 group that died. A potential explanation for this is that acute IL-6 elevations have neuroprotective properties thus facilitate functional recovery, but chronic IL-6 elevations are associated with neurodegeneration and breakdown of the blood-brain barrier.^[23] Therefore, the prolonged elevated IL-6 levels for up to 7 days post-injury observed in our study for mRS 5 and 6 groups may suggest its contribution to the severely disabled outcome.

We observed in this study a significant increase in IL-10 level in the mRS 6 and mRS 5 groups compared to the non-TBI control group. Similarly, it was reported in another human sTBI study that the temporal profile of CSF IL-10 level was highest at day 1 and gradually decreased throughout the 2 week duration of the study.^[21] Interestingly, the temporal CSF and serum IL-10 levels were previously shown to be inversely correlated to TNF-α.^[21] This was also observed in this study involving the mRS 6 and mRS 1–4 groups, but not with the mRS 5 group (data not shown). IL-10 is an anti-inflammatory cytokine that dampens T cell and macrophage responses.^[24] Therefore, this suggests that when substantial injury is observed in sTBI patients, an acutely elevated IL-10 level may not be beneficial and potentially leads to unfavorable neurological functional recovery.

The current study also recognized significant increases in CCL-2 in the mRS 5 and mRS 1–4 groups compared to the non-TBI control group. Recently, it has been shown that a sustained elevated CCL-2 level was detected in the CSF of sTBI patients for 10 days post-injury, and that CCL-2 knockout mice exhibited a delayed reduction in neuroinflammation and neurodegeneration, ultimately resulting in improved functional recovery compared to wild-type mice.^[25] Furthermore, we observed significant increases in CCL-20 level in the mRS 5 and mRS 1–4 groups compared to the non-TBI control group. CCR6 (receptor for CCL-20) knockout mice suffered from less neuroinflammation and neurodegeneration than wild-type mice in repetitive TBI.^[26] Therefore, these data suggest that the elevated CSF CCL-2 level contributed to the survival of the sTBI patients and that CCL-20 may exacerbate traumatic neuronal injury and give rise to an unfavorable functional outcome in sTBI.

In this study, the positive correlation between higher CSF galectin-3 and IL-6 levels in the mRS 5 group compared to the mRS 6 group is similar to previously reported by Winter and colleagues, which described that parenchymal IL-6 levels were elevated in sTBI patients that survived than those that died.^[27] The temporal correlation of CSF galectin-3 and IL-6 levels after sTBI may be linked by their close association. For example, it has been shown that IL-6 is upregulated via a galectin-3-dependent pathway in human neuroblastoma.^[5] Overall, the associations between galectin-3 and various cytokines including IL-6, CCL-2, and CCL-20 may be explained by increases in microglial activation. Animal TBI models have illustrated significant elevations in activated microglia at day 1 post-injury, which contributes to the production of various cytokines.^[3]

5. Limitations

This clinical study has a few limitations. Firstly, there was no follow-up of patients’ functional recovery beyond 6 months. Secondly, there was a low number of severe TBI patients (n = 6–18) within the sub-groups as others have published using patient groups ranging from 17 to as high as 90 sTBI patients.^[15,28] Thirdly, there were missing values due to conflicts with management, stability, or death of the patient. However, it is important to include data of patients with short survival duration to avoid creating a biased set of data focused on survived sTBI patients that may have suffered a milder injury, so does not reflect the true nature and severity of sTBI disease.

6. Conclusion

Acute CSF galectin-3 and associated cytokines, except IL-1β and TNF-α levels were elevated in sTBI patients. Sustained elevated levels of pro-inflammatory cytokines, especially galectin-3, IL-6, CCL-2, and CCL-20 at sub-acute stages of injury is detrimental to functional recovery. Additionally, elevated anti-inflammatory cytokine IL-10 at the acute stage is also deleterious to TBI patients. Overall, the levels and temporal profiles of CSF galectin-3 and associated cytokines can sub-divide severe TBI patients into sub-groups to support the prognostication of sTBI.

Acknowledgments

We would like to thank the patients and family for the contribution of cerebrospinal fluid for this study.

Author contributions

Conceptualization: Ping Yip, Zhuo-Hao Liu

Data curation: Ping Yip, Wing Sze Leung, Melisa Aria Cetin, Ting-Wei Chang, Mun-Chun Yeap, Chun-Ting Chen, Yu-Chi Wang, Ching-Chang Chen, Zhuo-Hao Liu.

Formal analysis: Ping Yip, Wing Sze Leung, Melisa Aria Cetin, Ting-Wei Chang, Mun-Chun Yeap, Chun-Ting Chen, Yu-Chi Wang, Ching-Chang Chen, Zhuo-Hao Liu.

Funding acquisition: Zhuo-Hao Liu.

Investigation: Zhuo-Hao Liu.

Project administration: Zhuo-Hao Liu.

Resources: Ting-Wei Chang, Mun-Chun Yeap, Chun-Ting Chen, Yu-Chi Wang, Ching-Chang Chen, Zhuo-Hao Liu.

Supervision: Ping Yip, Zhuo-Hao Liu.

Writing – original draft: Ping Yip, Zhuo-Hao Liu.

Writing – review & editing: Ping Yip, Wing Sze Leung, Melisa Aria Cetin, Zhuo-Hao Liu.