Monocyte, neutrophil, and whole blood transcriptome dynamics following ischemic stroke

BACKGROUND

After ischemic stroke (IS), peripheral leukocytes infiltrate the damaged region and modulate the response to injury. Peripheral blood cells display distinctive gene expression signatures post-IS and these transcriptional programs reflect changes in immune responses to IS. Dissecting the temporal dynamics of gene expression after IS improves our understanding of immune and clotting responses at the molecular and cellular level that are involved in acute brain injury and may assist with time-targeted, cell-specific therapy.

METHODS

The transcriptomic profiles from peripheral monocytes, neutrophils, and whole blood from 38 ischemic stroke patients and 18 controls were analyzed with RNA-seq as a function of time and etiology after stroke. Differential expression analyses were performed at 0-24 h, 24-48 h, and >48 h following stroke.

RESULTS

Unique patterns of temporal gene expression and pathways were distinguished for monocytes, neutrophils, and whole blood with enrichment of interleukin signaling pathways for different time points and stroke etiologies. Compared to control subjects, gene expression was generally upregulated in neutrophils and generally downregulated in monocytes over all times for cardioembolic, large vessel, and small vessel strokes. Self-organizing maps identified gene clusters with similar trajectories of gene expression over time for different stroke causes and sample types. Weighted Gene Co-expression Network Analyses identified modules of co-expressed genes that significantly varied with time after stroke and included hub genes of immunoglobulin genes in whole blood.

CONCLUSIONS

Altogether, the identified genes and pathways are critical for understanding how the immune and clotting systems change over time after stroke. This study identifies potential time- and cell-specific biomarkers and treatment targets.

Abstract WP229: Hub Genes Drive Specific Gene Expression Changes Seen In Intracerebral Hemorrhage And Ischemic Stroke

Gene expression changes in peripheral leukocytes display distinctive profiles after intracerebral hemorrhage (ICH) and ischemic stroke (IS), differentiating both conditions at the molecular level. The breadth of data produced by high-throughput transcriptomic analyses can identify groups of genes and main gene expression drivers that have meaningful functional associations with the disease. This can help prioritize the investigation of key genes for diagnosis and treatment. Thus, we performed whole transcriptome analyses on ICH and IS samples and constructed gene networks from a genome-wide perspective. RNA-seq was performed on peripheral blood (WB) and isolated monocytes (MON) and neutrophils (NEU) (n=6 ICH, n=33 IS and n=9 vascular risk factors control (VRFC) subjects). Gene expression results were used to construct separate co-expression networks for all datasets analyzed (ICH + VRFCs, and IS + VRFCs, for MON, NEU and WB) using Weighted Gene Co-expression Network Analysis. Modules of genes significantly associated with ICH in the ICH + VRFCs network, and with IS in the IS + VRFCs network, were identified. The most highly interconnected genes in each of these modules were identified, representing hub genes that are potential master regulators. Functional annotation of the modules and hubs were done using gene ontology. From the significantly associated modules for ICH and IS in all sample types analyzed, there was little overlap in genes between diagnoses (≤2% in MON, ≤21 % in NEU and ≤16% in WB), and no overlap of ICH or IS hub genes from MON and NEU. In WB, ≤2% of the hubs were common to ICH and IS. It is plausible that these potential master regulators drive diagnosis-specific gene expression profiles. ICH hubs were associated with RNA splicing and mRNA processing (MON), cell adhesion (NEU) and NF-κβ signaling (WB). IS hubs were associated with cell migration (MON), T cell chemotaxis (NEU) and transcription factor activity (WB). In addition, most hubs in IS MON were noncoding RNA. The gene networks and their respective hub genes provide novel cell-specific pathophysiological insights and could represent potential key pharmacological targets and biomarkers.

Gene expression changes implicate specific peripheral immune responses to Deep and Lobar Intracerebral Hemorrhages in humans

The peripheral immune system response to Intracerebral Hemorrhage (ICH) may differ with ICH in different brain locations. Thus, we investigated peripheral blood mRNA expression of Deep ICH, Lobar ICH, and vascular risk factor-matched control subjects (n = 59). Deep ICH subjects usually had hypertension. Some Lobar ICH subjects had cerebral amyloid angiopathy (CAA). Genes and gene networks in Deep ICH and Lobar ICH were compared to controls. We found 774 differentially expressed genes (DEGs) and 2 co-expressed gene modules associated with Deep ICH, and 441 DEGs and 5 modules associated with Lobar ICH. Pathway enrichment showed some common immune/inflammatory responses between locations including Autophagy, T Cell Receptor, Inflammasome, and Neuroinflammation Signaling. Th2, Interferon, GP6, and BEX2 Signaling were unique to Deep ICH. Necroptosis Signaling, Protein Ubiquitination, Amyloid Processing, and various RNA Processing terms were unique to Lobar ICH. Finding amyloid processing pathways in blood of Lobar ICH patients suggests peripheral immune cells may participate in processes leading to perivascular/vascular amyloid in CAA vessels and/or are involved in its removal. This study identifies distinct peripheral blood transcriptome architectures in Deep and Lobar ICH, emphasizes the need for considering location in ICH studies/clinical trials, and presents potential location-specific treatment targets.

Abstract TMP114: Time-based Dynamic Analyses Of Gene Expression In Monocytes, Neutrophils And Whole Blood Identify Key Hub Genes And Functional Processes Following Acute Ischemic Stroke

Gene expression changes in peripheral blood reflect injury and repair processes occurring post ischemic stroke (IS). Our study explored the dynamic time-dependent expression of key genes involved in the immune response after IS to better understand the biology and to identify specific diagnostic biomarkers. Using RNA-sequencing, we analyzed gene expression profiles of 38 IS patients and 18 controls with at least one vascular risk factor (VRFC) including diabetes and/or hypertension and/or hypercholesterolemia in isolated monocytes, neutrophils and whole blood. We used two approaches: Weighted Gene Co-expression Network Analysis (WGCNA) with respect to time after stroke onset; and differential expression analyses with subject samples split into time points (TPs) from stroke onset (TP0=VRFC; TP1=0-24 h; TP2=24-48 h; and TP3≥48 h). In WGCNA, highly interconnected “hub” genes were identified for modules significant to time (p<0.05). Differentially expressed genes (DEGs) with diagnosisхTP p<0.02 and fold-change>

Abstract WP133: Sex Differences In The Human Intracerebral Hemorrhage Peripheral Blood Transcriptome Implicate Differential Immune And Inflammatory Responses

Sex differences in immune and inflammatory pathways have been shown in health and disease. However, little research has been done on sex differences following human Intracerebral Hemorrhage (ICH). We sought to unveil transcriptome differences in blood between Male (M) and Female (F) ICH responses. We evaluated 33 ICH patients and 33 vascular risk factor matched controls (VRFC), 9F and 24M each. Peripheral blood expression of 21,175 genes was analyzed at the co-expression network level with WGCNA - separate F (F-ICH and F-VRFC) and M (M-ICH and M-VRFC) networks generated; and per gene level (ANCOVA: Age, Time, and Sex*Dx (Diagnosis)). Five F (F-6, F-20, F-21, F-19, F-5) and 2 M (M-32, M-26) WGCNA modules were significant to Dx (p < 0.05) (Fig 1A). 105 genes were significant for the direct comparison (F-ICH vs M-ICH p < 0.05, |FC| > 1.2; Sex*Dx FDR < 0.2), which separated Sex*Dx groups (Fig 1B). 1,425 genes were differentially expressed in F-ICH (F-ICH vs F-VRFC p < 0.05, |FC| > 1.2; Sex*Dx FDR < 0.2) and 421 in M-ICH (M-ICH vs M-VRFC p < 0.05, |FC| > 1.2; Sex*Dx FDR < 0.2) (Fig 1C). F-ICH response was enriched in Monocyte and T Cell Specific genes and M-ICH in B Cell and Erythroblast specific genes; both were enriched in Neutrophil specific genes (Fig 1A). Overall, F modules and gene lists were significantly enriched in 236 GO terms (FDR < 0.1) and M modules in 55; 20 were common (Fig 1D). The ICH response unique to F-ICH included Inflammatory Response, T Cell Activation, Autophagy, Apoptotic Process, and RNA Splicing. M-ICH unique included B Cell Receptor Signaling, Immunoglobulin Receptor Binding, Antigen Binding, Complement Activation, and Receptor Mediated Endocytosis. Common responses included Innate Immune Response, Blood Coagulation, and Fc-ã Receptor Signaling Involved in Phagocytosis (Fig 1E). We found sex differences in ICH transcriptome responses in human peripheral blood. These implicate specific cell types in each sex that could represent novel sex-specific treatment targets.

Lipopolysaccharide, Identified Using an Antibody and by PAS Staining, Is Associated With Corpora amylacea and White Matter Injury in Alzheimer's Disease and Aging Brain

Corpora amylacea (CA) increase in number and size with aging. Their origins and functions remain unknown. Previously, we found that Alzheimer's disease (AD) brains have more CA in the periventricular white matter (PVWM) compared to aging controls. In addition, CA is associated with neurodegeneration as indicated by colocalization of degraded myelin basic protein (dMBP) with periodic acid-Schiff (PAS), a CA marker. We also found that bacterial lipopolysaccharide is present in aging brains, with more LPS in AD compared with controls. Periodic acid-Schiff staining is used to identify CA by virtue of their high polysaccharide content. Despite the growing knowledge of CA as a contributor to AD pathology, the molecules that contribute to the polysaccharides in CA are not known. Notably, lipopolysaccharides (LPS) are important cell-surface polysaccharides found in all Gram-negative bacteria. However, it is unknown whether PAS could detect LPS, whether the LPS found in aging brains contribute to the polysaccharide found in CA, and whether LPS associate with myelin injury. In this study, we found that aging brains had a myelin deficit zone (MDZ) adjacent to the ventricles in PVWM. The MDZ contained vesicles, most of which were CA. LPS and dMBP levels were higher in AD than in control brains. LPS was colocalized with dMBP in the vesicles/CA, linking white matter injury with a bacterial pro-inflammatory molecule. The vesicles also contained oxidized fibers, C-reactive protein, NG2, and GALC, markers of oligodendrocyte precursor cells (OPCs) and oligodendrocyte cells (OLs), respectively. The vesicles/CA were surrounded by dense astrocyte processes in control and AD brains. LPS was co-localized with CA by double staining of PAS with LPS in aging brains. The relationship of LPS with PAS staining was confirmed by PAS staining of purified LPS on nitrocellulose membranes. These findings reveal that LPS is one of the polysaccharides found in CA which can be stained with PAS. In addition, vesicles/CA are associated with oxidized and damaged myelin. The LPS in these vesicles/CA may have contributed to this oxidative myelin damage and may have contributed to oxidative stress to OPCs and OLs which could impair the ability to repair damaged myelin in AD and control brains.

Bacterial lipopolysaccharide is associated with stroke

We aimed to determine if plasma levels of bacterial lipopolysaccharide (LPS) and lipoteichoic acid (LTA) are associated with different causes of stroke and correlate with C-reactive protein (CRP), LPS-binding protein (LBP), and the NIH stroke scale (NIHSS). Ischemic stroke (cardioembolic (CE), large artery atherosclerosis (LAA), small vessel occlusion (SVO)), intracerebral hemorrhage (ICH), transient ischemic attack (TIA) and control subjects were compared (n = 205). Plasma LPS, LTA, CRP, and LBP levels were quantified by ELISA. LPS and CRP levels were elevated in ischemic strokes (CE, LAA, SVO) and ICH compared to controls. LBP levels were elevated in ischemic strokes (CE, LAA) and ICH. LTA levels were increased in SVO stroke compared to TIA but not controls. LPS levels correlated with CRP and LBP levels in stroke and TIA. LPS, LBP and CRP levels positively correlated with the NIHSS and WBC count but negatively correlated with total cholesterol. Plasma LPS and LBP associate with major causes of ischemic stroke and with ICH, whereas LPS/LBP do not associate with TIAs. LTA only associated with SVO stroke. LPS positively correlated with CRP, LBP, and WBC but negatively correlated with cholesterol. Higher LPS levels were associated with worse stroke outcomes.

Abstract P787: Sexually Dimorphic Gene Expression Molecular Correlates of Improvement in Human Ischemic Stroke

Objective: Ischemic stroke (IS) is sexually dimorphic for risk factors, age, heritability, causes, treatment, and outcome. We identified transcriptional correlates with 90-day outcome that differed between male and female IS subjects.

Methods: RNA from 72 samples from 2 peripheral blood draws (at ≤3 and 24h post IS onset) was analyzed on Affymetrix U133 Plus 2 microarrays. These represented samples from 36 CLEAR trial IS patients treated with tPA with or without eptifibatide after the first blood sample within 3 hours of stroke onset. Changes in gene expression levels (deltaGE) between 3h and 24h were calculated and the association with percent NIH Stroke Scale (NIHSS) improvement from 3h to 90 days (% Improvement) examined. We used mixed-effects linear regression, including Treatment, Age, Sex, Vascular Risk Factors, 3h NIHSS, % Improvement, and a Sex * % Improvement interaction. Sex differences in association of gene expression with % Improvement were determined by examining the Sex * % Improvement interaction term, p<0.005 was considered statistically significant.

Results: 577 genes correlated differently with % Improvement in IS males and females. These included matrix metalloproteinases (MMPs), which play a major role in BBB dysfunction and outcomes post IS. MMP11 , MMP14 and MM17 correlated with % Improvement in opposite direction in males and females. Inflammatory genes like IL-27 , implicated in infarct volume and stroke outcome, and ABC transporters ( ABCC9 ) also had opposite correlation with % Improvement in males and females. Calmodulin 1 ( CAML1 ) was also sexually dimorphic, and a SNP in CALM1 has been implicated in IS risk and blood coagulation in female IS patients. EIF2 signaling, a major protein synthesis pathway was activated in males (adj. p = 1e-8), while suppressed in females (adj. p value = 1e-9). Protein synthesis and associated unfolded protein response cascade have previously been implicated in stroke outcome.

Conclusions: The identified sexually dimorphic gene expression associated with 90-day improvement might relate to sex differences in blood immune and clotting pathways. The findings expand our understanding of the genomic underpinnings associated with stroke outcome and may serve as potential sex-specific treatment targets.

Abstract P576: Plasma Bacterial Lipopolysaccharide Associates With Carotid Atherosclerosis, a Cause of Large Vessel Stroke

Introduction: Inflammation and infection are associated with cerebrovascular diseases including stroke due to carotid atherosclerotic plaques. C-reactive protein (CRP), an acute-phase protein, is upregulated in the plasma of patients with carotid atherosclerotic plaques. However, little is known about whether bacterial molecules trigger inflammation or play a role in patients with carotid atherosclerotic plaques. Recently, it has been recognized that inflammation associated with atherosclerosis and morbidity and mortality in cardiovascular diseases may be due to lipopolysaccharide (LPS) that is found in the outer wall of all Gram-negative bacteria. These findings prompted this study to explore whether plasma levels of LPS and LPS-binding protein (LBP) are elevated and correlated with CRP levels in patients with asymptomatic carotid plaques (ACP). We also compared LBP levels in patients with ACP to large vessel (LV) strokes due to carotid plaques and to matched controls.

Methods: Patients (n = 30) with ACP, LV stroke due to carotid atherosclerosis and age-, sex- matched healthy controls gave consent and had their blood drawn. Plasma was processed for LPS, LBP and CRP detection using separate ELISA for each.

Results: Plasma LBP level in ACP (22.7 ± 2.92 μg/ml) was similar to LV stroke (21.6 ± 1.56 μg/ml, p = 0.74, ACP vs LV) but greater than controls (13.6 ± 1.43 μg/ml, p = 0.011, ACP vs controls). In ACP patients, plasma LPS level (159.5 ± 30.5 μg/ml) was greater than controls (42.6 ± 11.7 μg/ml, p = 0.001); plasma CRP levels (20.2 ± 6.2 μg/ml) was higher than controls (5.3 ± 2.1 μg/ml, p = 0.011). There was a positive correlation between LPS levels and LBP levels (r = 0.86, p < 0.00001), LPS levels and CRP levels (r = 0.82, p = 0.00001), and LBP levels and CRP levels (r = 0.89, p < 0.00001) in ACP cases.

Conclusions: Plasma LPS, LBP and CRP associate with asymptomatic carotid plaques suggesting a pro-inflammatory state exists in patients with asymptomatic carotid plaques, a cause of large vessel stroke. LPS is postulated to directly upregulate both CRP and LBP. Elevated LBP in large vessel stroke patients suggests a Gram-negative bacteria associated post-stroke inflammatory state.

Abstract P744: Gene Transcript Clusters Distinguish Time-Dependent Expression Patterns in Monocytes, Neutrophils and Whole Blood After Ischemic Stroke Injury

Introduction: After ischemic stroke (IS), peripheral leukocytes infiltrate the damaged region and modulate the response to injury. We previously showed that peripheral blood cells display different gene expression profiles after IS and these transcriptional programs reflect the changes in immune processes in response to IS. Dissecting the temporal dynamics of gene expression after IS improves our understanding of the changes of molecular and cellular pathways involved in acute brain injury.

Methods: We analyzed the transcriptomic profiles of 33 IS patients in isolated monocytes, neutrophils and whole blood. RNA-sequencing was performed on all the stroke samples as well as 12 controls with vascular risk factors (diabetes and/or hypertension and/or hypercholesterolemia). To identify differentially expressed genes, subjects were split into time points (TPs) from stroke onset (TP1= 0-24 h; TP2= 24-48 h; and TP3= > 48 h), and controls were assigned TP0. A linear regression model including time and the interaction of diagnosis x TP with cutoff of p<0.02 and fold-change>|1.2| was used. Time dependent changes were analyzed using artificial neural networks to identify clusters of genes that behave in a similar way across TPs.

Results: Unique patterns of temporal expression were distinguished for the three sample types. These include genes not expressed in TP0 that peak only within the first 24 h, others that peak or decrease in TP2 and TP3, and more complex patterns. Genes that peak at TP1 in monocytes and neutrophils are related to cell adhesion and leukocyte differentiation/migration, respectively. Early peaks in whole blood occur in genes related to transcriptional regulation. In monocytes, interleukin pathways are enriched across all TPs, whereas there is a trend of suppression after 24 h in neutrophils. The inflammasome pathway is enriched in the earlier TPs in neutrophils, while not enriched in monocytes until over 48 hours.

Conclusion: Our analyses on gene expression dynamics and cluster patterns allow identification of key genes and pathways at different time points following ischemic injury that are valuable as IS biomarkers and may be possible treatment targets.