Anti-inflammatory genes in PBMCs post-spontaneous intracerebral hemorrhage

2.1 Patient selection, sample collections, and outcome measures

Patients who were admitted to the Neurointensive Care Unit for management of acute spontaneous ICH confirmed on computed tomography (CT) scan were screened and consented. Demographic and clinical information were collected as shown in Table 1a and b.

Venous blood samples (5 mL) were collected in sodium citrate tube within 48 h of presentation. Venous blood samples and PBMCs were categorized as “early” if collected within 48 h of symptom onset and “delayed” if collected greater than 72 h from symptom onset. The median time of blood collection for early onset was 28.88 h (Table 1). PBMCs were prepared from whole blood per manufacturer’s protocol using the Lymphoprep Ficoll and Sepmate Tube (StemCell Technologies, Vancouver, CN). PBMC counts were determined using a TC20 cell counter (Bio-Rad, Hercules, CA) and prepared for freezing in RMPI media (ThermoFisher Scientific, Waltham, MA) with 10% heat-inactivated fetal bovine serum (FBS, Atlanta Biologicals) and 10% DMSO (Corning Inc., Corning, NY) using standard protocol for freezing live cells. Samples were stored in liquid nitrogen until use.

Data on demographics, clinical, radiological characteristics, and hospital length of stay were collected. Patients were contacted via a telephone call at 90 days from clinical presentation to determine a modified Rankin score (MRS) using a standard questionnaire. We defined favorable outcome as 90-day MRS ≤3 and unfavorable outcome as 90-day MRS 4–6.

1. Informed consent: Informed consent has been obtained from all individuals included in this study.

2. Ethical approval: The research related to human use has been complied with all the relevant national regulations, institutional policies, and in accordance with the tenets of the Helsinki Declaration and has been approved by the authors’ institutional review board or equivalent committee.

2.2 Total RNA extraction and real-time RT-PCR analysis

Frozen PBMCs were rapidly thawed in a 37°C water bath with gentle agitation for 1–2 min, transferred to a 15 mL centrifuge tube containing 5 mL of RMPI media +10% FBS, and centrifuged at 400 × g at room temperature for 10 min. The pellet was quickly prepared for total RNA extraction using the miRNeasy microRNA kit per manufacturer’s protocol (Qiagen, Valencia, CA). Briefly, the PBMC pellet was lysed in Qialysis buffer and phase-separated using chloroform, transferred to a column, washed, and resuspended in 14 µL RNase-free water. RNA concentration was determined using the NanoDrop spectrophotometer (Thermo Fisher Scientifics, Waltham, MA). Total RNA (100 ng) was used for cDNA synthesis using the iScript cDNA Synthesis Kit (Bio-Rad). The cDNA was diluted in sterile water and 10 ng was used for real-time PCR using Taqman probe (Thermo Fisher Scientifics). The target genes include PPARG, IL10, NFKB1 (P105/p55), REL (c-Rel), SOD1, and NFE2L2. PCR amplification was carried out in a Applied Biosystem 7,500 Fast Thermal Cycler (Thermo Fisher Scientifics) in triplicate using recommended parameters for 40 cycles. Cycle threshold (Ct) data were normalized to the GAPDH control gene and relative fold change was determined using the 2^−ΔΔCt method. Differential gene expression was determined based on a 1.5-fold cutoff.

2.3 Statistical methods

Demographic and clinical characteristics were compared between ICH volume groups (<30 vs ≥30 mL) by means of a Fisher exact test (corresponding OR were obtained from a logistic regression model with Firth correction) or a t-test with Satterthwaite correction. Similar analyses were carried out to compare these characteristics between favorable and unfavorable outcome groups (MRS 0–3 and 4–6). Potential covariation of ICH volume at 24 h and Ct values (normalized to GAPDH) for each gene (IL10, NFKB1, NFE2L2, PPARG, SOD1, and REL) at 0–2 days – and separately at 3–5 days – was estimated by means of raw Spearman correlations. Covariation of ICH volume at 24 h and fold change by the 2^−ΔΔCt method from 0–2 days to 3–5 days was also assessed by raw Spearman correlations. P-values are considered as indicative of potential associations. Gene expression was estimated using mean normalized Ct values for clinical outcome and for early (0–2 days) to delayed (3–5 days) and analyzed by the Holm-Sidak t-test with alpha = 0.05 in Prism 8.0 (Graphpad, San Diego, CA).

3.1 Patient characteristics

A total of 20 consecutive patients with spontaneous ICH between November 2017 and January 2019 who met the inclusion criteria were recruited for the study. Overall, 13 were male (65%) and 90% had a prehospital diagnosis of hypertension; 9 patients (45%) had a favorable outcome (MRS 0–3). Table 1a and b describe patients’ characteristics based on ICH volume at 24 h and the comparison between patients with favorable and unfavorable outcomes based on Modified Rankin Scale. A lower ICH score at admission (≤2) was more likely in patients with favorable outcome (75 vs 8.3%). Also, there was a tendency towards younger age in the favorable outcome group and in ICH volume <30 mL group. The median time from onset to blood collection was 28.88 and the interquartile range (IQR) 20.05 h. When subset based on favorable and unfavorable outcome, the median time was 28.35 vs 15.17 h and was significantly different for favorable vs unfavorable, respectively. When the early onset (first 48 h) was subset into <12 and >12 h for the two outcome groups (MRS 0–3 vs MRS 4–6), the difference between the two groups was not significant.

3.2 Differential expression of mRNA in PBMCs and outcome

Several genes were found statistically significant in PBMCs of patient with favorable outcome vs unfavorable outcome. The early onset (first 48 h) and delayed (day 3–5) gene expression patterns for favorable and unfavorable outcomes are shown (Figure 1). Expression of IL10 at the delayed time points was statistically significant (p-value < 0.05) and 3-fold higher for patients with favorable compared with unfavorable outcome. In patients with favorable outcomes, the expression of IL10 increased 1.5-fold at the delayed time points over the level of the early time points, whereas in patients with unfavorable outcome, the levels remained constant from the early to delayed time points. Expression of SOD1 at the delayed time points was statistically significant (p-value < 0.05) and 5-folds higher in favorable than unfavorable outcomes. SOD1 expression in unfavorable outcome decreased 3.2-fold from early to delayed time points, whereas in favorable outcome, a gradual increase was seen over time. Expression of IL10 and SOD1 were not different between groups at early time points.

Figure 1

Expression of IL10, SOD1, NFKB1, REL, and PPARG in PBMCs at Day 0–2 and Day 3–5 of ICH patients with unfavorable outcome (n = 7) relative to ICH patients with favorable outcome (n = 10). A smaller normalized Ct value indicates a higher gene expression level. (a) Expressions of IL10 and SOD1 were significantly higher at Day 3–5 in favorable outcome than unfavorable outcome (p-value < 0.05). (b) Expressions of NFKB1 (p = 0.046) and PPARG (p = 0.027) at Day 0–2 were significant. Expressions of NFKB1 and PPARG were higher at Day 0–2 in ICH patients with unfavorable outcome. REL expression was significantly higher at Day 3–5 in favorable outcome compared to unfavorable outcome (p = 0.032). Expression of NFE2LE was not significant at either time points. Data are expressed as mean ± standard error of mean (SEM).

At the early time points, expression of NFKB1 and PPARG were significant (p < 0.05) and exhibited a 2.3-fold and 5-fold increase, respectively, in unfavorable outcome than favorable outcome. No significant difference was seen at the delayed time points. For NFKB1 in unfavorable outcome, expression of NFKB1 significantly decreased 2.3-fold from the early time point over the delayed time points (p < 0.05). The opposite pattern of 2.3-fold increase from the delayed time points over the early time points for patients with favorable outcome is noteworthy. In unfavorable outcome, PPARG expression decreased slightly at 1.5-fold from the early to the delayed time points, in contrast to favorable outcome in which PPARG expression significantly increased greater than 9-fold (p < 0.05).

Another member of the NF-κB family of transcription factors, REL expression at the early time points was not different for patients with favorable and unfavorable outcome. However, at the delayed time point, REL expression was statistically different (p < 0.05). Expression of REL increased 2.9-fold from the early to the delayed timepoints in patients with favorable outcome (p < 0.05), while in patients with unfavorable outcome, REL expression did not change. Though NFE2L2 was 2-fold higher in unfavorable outcome compared to favorable outcome, NFE2L2 expression at the delayed time point was similar between outcome groups. No change in NFE2L2 was seen for unfavorable outcome; however, in patients with favorable outcome, NFE2L2 expression increased 3.2-fold at delayed time point.

In general, at the early time point the expression of NFKB1, PPARG, and NFE2L2 in unfavorable outcome was higher than that in favorable outcome, while at the delayed time point expression of IL10, SOD1, and REL was higher in patients with favorable outcome. Interestingly, the observed higher early expression of PPARG in unfavorable outcome was reversed at the delayed time point, with a higher expression in the favorable outcome group. In summary, gene expression increased from the early time points to the delayed time points in patients with favorable outcome.

3.3 Correlation of PBMCs gene expression, ICH volume at 24 h, and clinical/demographic characteristics

It is known that ICH volume is an independent predictor of outcome. We analyzed the effect of each gene expression with known covariates of ICH outcome such as age, gender, ICH location, and diabetes mellitus (Table 2, Suppl Table S1). The table shows unadjusted and partialled Spearman correlation of genes and ICH volume at 24 h. Unadjusted analysis showed a potential direct association between ICH volume and each of IL10 and SOD1 (rho = 0.41, p = 0.073).The association for ICH volume and IL10 expression increases slightly from 0.4091 (p = 0.0732) to 0.4801 (p = 0.0375) and becomes significant when adjusted for age, which suggests the strength of association becomes more likely for IL10 expression and ICH volume (Table 2). The strength of the association is stronger with ICH location (rho = 0.4666, p = 0.0440) and surgical intervention (rho = 0.5587, p = 0.0129; rho = 0.5339, p = 0.0185). When ICH volume was correlated to the temporal fold change from the early day 0–2 time points to delayed 3–5 time points for the individual genes, only IL10 expression showed a significant and negative correlation (Table 3). The inverse correlation indicates that larger values of ICH volume at 24 h tend to associate with smaller fold change of IL10 from early 0–2 days to delayed 3–5 days. Although IL10 was not found statistically significant (p = 0.068), its expression and direction of change suggest increasing ICH volume has a negative impact on IL10 expression (Figure 2). Interestingly, the association between ICH volume and SOD1 expression increases slightly when adjusting for diabetes mellitus, but remains unchanged when adjusted for every other covariates. Lastly, REL expression showed a significant (p = 0.042) and moderate association when adjusted for hemicraniectomy. Due to the small sample size, a multivariate analysis was not done to identify the effect of the various covariates on gene expressions.

Figure 2

Plot of ICH volume at 24 h vs IL10 fold change. Expression of IL10 showed a negative correlation to high hematoma volume (>30 mL).

4.1 Study strengths and limitations

It is recognized that ICH volume is an independent predictor of clinical outcome after spontaneous ICH. Our study suggests that expression levels of IL10, an anti-inflammatory, M2 monocytes-associated cytokine, correlate to ICH volume and outcome. The result from the study also showed that measuring pattern of gene expression (PPARG, REL, IL10, NFE2L2) in PBMCs at early onset and 3–5 days post-admission may provide indication for clinical intervention and to predict outcome. Our small sample size limits the ability to account for variations in gene expression influenced by age, sex, race, and premorbid conditions. Although we looked at temporal gene expression changes at early onset and delayed, our sampling method of combining and grouping PBMC samples as day 0–2 and day 3–5 may also account for gene expression variations and may not provide a clear, temporal changes from day 0 to day 5. Even though we did not see a statistically significant difference between the two groups, time as a continuous variable should be considered when evaluating the fluctuation of gene expression of PBMCs during the first 48 h. Also, our study did not take into account the gene-gene interaction effects.