The role of DCs (dendritic cells) as potent mediators of inflammation has not been sufficiently investigated in stroke. Therefore, in the present study, circulating mDCPs (myeloid DC precursors), pDCPs (plasmacytoid DCPs) and tDCPs (total DCPs) were analysed by flow cytometry in (i) healthy controls (n=29), (ii) patients with ACI-S (asymptomatic cerebral infarction stenosis; n=46), (iii) patients with TIA (transient ischaemic attack; n=39), (iv) patients with AIS (acute ischaemic stroke; n=73), and (v) patients with AHS (acute haemorrhagic stroke; n=31). The NIHSS (National Institutes of Health Stroke Scale) and infarction size on a CT (computer tomography) scan were evaluated after stroke. In a patient subgroup, post-mortem immunohistochemical brain analyses were performed to detect mDCs (CD209), pDCs (CD123), T-cells (CD3) and HLA-DR. In AIS and AHS, the numbers of circulating mDCPs (P<0.005), pDCPs (P<0.005) and tDCPs (P<0.001) were significantly reduced. A significant inverse correlation was found between the NIHSS and circulating DCPs (P<0.02), as well as between hsCRP (high-sensitivity C-reactive protein) and circulating DCPs (P<0.001). Patients with large stroke sizes on a CT scan had significantly lower numbers of mDCPs (P=0.007), pDCPs (P=0.05) and tDCPs (P=0.01) than those with smaller stroke sizes. Follow-up analysis showed a significant recovery of circulating DCPs in the first few days after stroke. In the infarcted brain, a dense infiltration of mDCs co-localized with T-cells, single pDCs and high HLA-DR expression were observed. In conclusion, acute stroke leads to a decrease in circulating DCPs. Potentially, circulating DCPs are recruited from the blood into the infarcted brain and probably trigger cerebral immune reactions there.
- dendritic cell
- ischaemic brain injury
- leucocyte infiltration
- T-cell activation
Inflammation is a major secondary pathophysiological factor after stroke [1,2]. Local cerebral inflammation is involved in focal brain oedema, haemorrhage and tissue necrosis. These events are associated with further impairment of neurological outcome. Moreover, studies have described the phenomenon of the so-called CIDS [CNS (central nervous system)-injury-induced immunodepression], which is a result of a systemic reduction in circulating lymphocytes and impaired T-cell and natural killer cell activity, and frequently leads to severe systemic infections with an increase in mortality [3,4]. Systemic infections are important in the context of stroke. On the one hand, it is well known that preceding infections are a significant risk factor for cerebral infarction , whereas, on the other hand, it has been shown that SAIs (stroke-associated infections) are a predictor of a poor functional outcome . However, the fact that SAIs are not independently associated with neurological outcome after stroke suggests that they do not contribute to cerebral damage, but are a simple marker of stroke severity . This fact is reflected in clinical trials showing the inefficacy of antibiotic prophylaxis in improving neurological outcome after stroke .
Under normal conditions, the immune system, particularly T-cells, are anergic to cerebral antigens. Following stroke, cerebral inflammation may lead to an autoimmune response directed against neuroepitopes, which can in turn result in further brain injury. Supporting this hypothesis, recent studies have shown that administration of LPS (lipopolysaccharide) in a rat model of stroke induced a Th1 immune response towards MBP (myelin basic protein), resulting in persisting severe post-ischaemic brain injury . On the other hand, the induction of tolerance by nasal vaccination to MBP or myelin oligodendrocyte glycoprotein led to a reduction in infarct size and neurological improvement [9,10]. Thus antigen-presenting cells may play a crucial role for the neurological damage after stroke.
DCs (dendritic cells) are effective antigen-presenting cells with the unique ability of stimulating memory and naïve T-cells. This provides them with the ability to induce an immune response against antigens which were so far not recognized by the immune system . Thus DCs might be the best candidates for the supposed primary immune response against neuroepitopes after stroke. Two lineages of DCs can be differentiated: mDCs (myeloid DCs), which respond to bacteria and fungi releasing IL-12 (interleukin-12), as well as pDCs (plasmacytoid DCs), which release IFN-α (interferon-α) upon viral infection . Both lineages can be detected as DCPs (DC precursors) in blood, patrolling through the circulation and invading the tissue in response to a local infection or other inflammatory situation. After migration into the tissue as immature DCs, they are enabled to take up antigens and, upon subsequent maturation, acquire the ability to stimulate T-cells against those antigens (mature DCs). In recent times, the role of mDCs or pDCs has been implicated in several pro-inflammatory diseases, for example atherosclerosis [13,14]. Furthermore, in multiple sclerosis, it has been shown that mDCs invade the human brain, subsequently triggering cerebral inflammation .
The aim of the present study was to investigate whether circulating DCPs are transiently reduced in patients with acute cerebral ischaemia due to their recruitment into the infarcted brain, and if there is evidence for a DC-triggered cerebral (auto)immune response.
MATERIALS AND METHODS
Patients and controls
Between February 2006 and July 2007, 39 patients with TIA (transient ischaemic attack), 73 patients with AIS (acute ischaemic stroke) and 31 patients with AHS (acute haemorrhagic stroke) were enrolled, who were admitted within 24 h after symptom onset to the Department of Neurology, University Hospital Erlangen, Erlangen, Germany. Standard diagnostic clinical analyses on admission included cranial CT (computer tomography), repeated after 24–48 h if the initial CT scan did not show pathological changes, to evaluate regions with cerebral ischaemia or haemorrhage. Moreover, Doppler sonography of extra- and intra-cranial carotid arteries, electrocardiography and echocardiography were routinely performed. Pathological findings (ischaemia and haemorrhage) were analysed on the CT scan by a radiologist, and their maximum diameter was assessed. According to the maximum diameter, acute cerebral ischaemia was classified radiologically as (i) small stroke (<10 mm), (ii) medium stroke (10–30 mm), or (iii) large stroke (>30 mm). Furthermore, the aetiology of stroke was classified according to TOAST (Trial of Org 10172 in Acute Stroke Treatment) criteria . The clinical severity of the stroke was evaluated at admission according to NIHSS (National Institute of Health Stroke Scale) . Patients with AIS admitted within 3 h after the onset of symptoms and with 4–22 points on NIHSS (n=18) underwent rt-PA (recombinant tissue plasminogen activator) thrombolysis according to the National Institute of Neurological Disorders and Stroke criteria and American Heart Association guidelines. The control group consisted of 29 gender- and age-matched cerebrovascular healthy subjects who presented themselves for a health check-up. Additionally, 46 patients were analysed with significant ACI-S (asymptomatic cerebral infarction stenosis) (>70%), but without acute cerebral ischaemia, who were awaiting routine carotid endarterectomy. Exclusion criteria were acute or chronic infections, malignancies, autoimmune diseases, hyperthyroidism, acute coronary syndromes and immunosuppressive medication.
Routine blood analyses were performed according to clinic standards. Standardized blood sampling for analysis by flow cytometry was carried out in the first few hours after admission and, in case of an acute cerebral ischaemia, at follow-up after 2 days (TIA) or 4 days (stroke). Immediately after blood withdrawal, samples were analysed by flow cytometry.
Informed consent was obtained from all patients. The study was approved by institutional ethics committee of the University Hospital Erlangen.
Serum analysis of pro-inflammatory markers
Serum concentration of hsCRP [high-sensitivity CRP (C-reactive protein)] was measured using an immunonephelometric assay on a BN II analyser, according to the manufacturer's instructions (Dade Behring).
Identification of DCPs by FACS
Fresh blood samples collected on EDTA were analysed using a blood dendritic cell enumeration kit (Miltenyi Biotec). Three-colour staining and FACS analysis were performed as described previously . FACS analysis was performed using the FACSCalibur flow cytometer with CellQuest software (Becton Dickinson). As circulating DCPs comprise only 0.1–1% of WBCs (white blood cells), a special gating strategy (Figure 1) was used to analyse mDCPs, pDCPs and tDCPs accurately, as described previously . The relative cell numbers of circulating DCPs were assessed as a percentage of WBCs. The absolute cell numbers (cells/μl) were calculated using relative cell numbers multiplied by the WBC count.
Histological characterization of the infarcted brain after stroke
Serial slides of brain tissue of patients who died from stroke and control slides were kindly provided by Professor I. Blümcke (Department of Neuropathology, University Hospital Erlangen, Erlangen, Germany). Brain morphology was analysed by an experienced neuropathologist using haematoxylin/eosin staining. No major autolytic changes were observed after a thorough microscopy inspection. Minor changes were present in all specimens and included an increased extracellular space. Subsequently, specimens were classified as acute ischaemic (n=17), acute haemorrhagic (n=12) or past (>4 weeks; n=12) stroke, and healthy tissue (n=11).
Immunohistochemical detection of DCs and T-cells in infarcted human brain
For immunohistochemical staining, the following monoclonal antibodies were used: anti-CD209 for mDCs (Becton Dickinson), anti-CD123 for pDCs (BioLegend), anti-CD3 for T-cells (Dako) and anti-HLA-DR (Dako). For immunohistochemical staining of CD209, CD123 and HLA-DR, a Catalysed Signal Amplification kit was used according to manufacturer's instructions (CSA System™; Dako). For immunohistochemical analysis of CD3 the EnVision™ G/2 Detection System (Dako) was used. Double immunohistochemical stainings were performed with the EnVision™ Doublestain System (Dako). Negative controls were treated with irrelevant isotype-matched antibodies.
Stained cells were identified with a charge-coupled-device camera (Nikon DXM 1200) at a magnification of ×150 in each of six representative sections (each 0.25 μm2) in the perivascular or interstitial stroke area. For controls, corresponding areas were analysed. For each patient, the mean cell number of mDCs, pDCs, T-cells and HLA-DR+ cells was calculated. Perivascular and interstitial cells are shown separately.
All values are reported as medians. P<0.05 was considered statistically significant. Clinical data (see Table 1) were compared statistically between patients and controls by one-way ANOVA. The correlation of mDCPs, pDCPs and tDCPs with different parameters was analysed using the non-parametric Spearman rank order test (see Table 2). Unless stated otherwise, all other comparisons and statistical analyses were done using the non-parametric Mann–Whitney rank sum test.
The baseline characteristics are shown in Table 1. No significant differences were observed for age, gender, medical history, atherogenic risk factors and current medication. Most of the serum parameters did not differ between the study groups. Patients with acute cerebral ischaemia had a trend for higher serum glucose levels. Regarding inflammation, significantly higher levels of leucocytes and hsCRP were detected in patients with AIS and AHS compared with controls (Table 1). According to TOAST criteria, the cause of stroke was large-artery atherosclerosis in 20% of the patients, cardioembolism in 25% of the patients, small-artery occlusion in 23% of the patients, other determined aetiology in 2% of the patients and undetermined aetiology in 30% of the patients.
Decrease in circulating DCPs in patients with AIS and AHS
Compared with healthy controls, patients with TIA, AIS and AHS had a significant decrease in relative and absolute levels of circulating mDCPs and tDCPs (Figure 2). Regarding pDCPs, a statistically significant reduction in their relative and absolute levels was observed in patients with AIS and AHS, whereas, in patients with TIA, the values were slightly lower compared with controls. As both the relative and, more importantly, the absolute values of circulating DCPs were reduced, we were able to exclude that their decrease might be caused by a secondary dilution phenomenon due to an increase in another WBC population. In a follow-up analysis of patients with acute cerebral ischaemia, a significant reconstitution of circulating DCPs was observed after 2 days (TIA) or 4 days (AIS and AHS), suggesting that their decrease is a transient phenomenon in stroke (results not shown). In contrast, patients with ACI-S did not have any significant alterations in circulating DCPs compared with the healthy controls (Figure 2).
Subgroup/correlation analysis of circulating DCPs and various clinical factors
Regarding TOAST criteria, the decrease in circulating DCPs in patients with acute cerebral ischaemia was independent of their aetiology (results not shown). In healthy controls, the values of circulating DCPs did not differ significantly between the sexes. In contrast, females with acute cerebral ischaemia had significantly lower mDCPs (9.4 compared with 10.5 cells/μl; P=0.04), pDCPs (4.9 compared with 6.2 cells/μl; P=0.01) and tDCPs (15.0 compared with 18.4 cells/μl; P=0.02) when compared with the corresponding male patients. In diabetic patients, significantly lower circulating tDCPs were observed (Table 2). Patients taking clopidogrel had significantly higher values of circulating mDCPs and tDCPs. Moreover, circulating pDCPs were significantly elevated in patients undergoing treatment with AT1 (angiotensin II type 1 receptor) antagonists.
To analyse whether the decrease in circulating DCPs was directly associated with other clinical factors, a correlation analysis was performed (Table 2). The age of the patients inversely correlated with circulating DCP levels. Regarding the clinical severity of stroke, a significant inverse correlation was observed between all subtypes of circulating DCPs and NIHSS (Table 2 and Figure 3). A significant inverse correlation between circulating DCPs and leucocytes or hsCRP was observed (Table 2 and Figure 3). Blood glucose or HbA1c (glycated haemoglobin) levels were significantly inversely correlated with circulating DCPs (Table 2).
Additionally, we investigated whether the decrease in circulating DCPs was associated with the radiologically evaluated stroke size (Figure 4). For this reason, patients with ischaemic and haemorrhagic stroke were subdivided according to the CT scan into three subgroups with small, medium and large infarction diameters. In concordance with the results for NIHSS, we observed significantly lower values of circulating mDCPs, pDCPs and tDCPs in patients with larger strokes. Separate analysis of patients with ischaemic and haemorrhagic stroke provided similar results, but without statistical significance due to the limited number of patients in the divided groups (results not shown). Taken together, these results suggest that the reduction in DCPs in stroke can be caused by their rapid recruitment from blood into the infarcted brain.
Emergence of mDCs co-localized with T-cells in the infarcted brain
To investigate whether circulating DCPs are recruited into the infarcted brain, we analysed human autopsy samples for the presence of DCs and T-cells, and their immunostimulatory capacity. Compared with control tissue, in the area affected by AIS or AHS numerous mDCs and T-cells, and to a smaller extent pDCs, were observed (Figure 5A). Regarding the cell distribution, mDCs and T-cells were often located around intracerebral vessels. In contrast, single pDCs were located at a distance from the vasculature. Surprisingly, in patients with past stroke persisting high numbers of mDCs and T-cells were observed. In the stroke area, a strong HLA-DR expression was detected. The number of HLA-DR-expressing cells by far exceeded the number of mDCs, pDCs and T-cells, suggesting that it is very likely that another, probably residential, cell type might be contributing to HLA-DR expression. Indeed, many of the HLA-DR-expressing cells had a very similar morphology to astrocytes. In past stroke, a persisting high expression of HLA-DR was detected in the infarcted brain, suggesting a long-lasting cerebral immune reaction.
In double immunohistochemical staining, co-localization of mDsC and T-cells and a high expression of HLA-DR close to mDCs was observed, indicating that mDCs are mature and able to activate T-cells in the infarcted brain (Figure 5B).
Recent studies have revealed that immunological mechanisms play an important role in the neurological and overall outcome after stroke. The breakdown of the blood–brain barrier contributes to an invasion of immune cells into the infarcted brain after stroke. These cells upon encountering novel brain antigens initiate an autoimmune response which leads to further destruction of neurological tissue . On the other hand, a systemic immunodepression syndrome emerging after stroke is the reason for frequent systemic infections [3,4]. The aim of our present study was to investigate whether local and systemic alterations in DCs contribute to the local immune response or to the systemic immunodepression after stroke.
In our present study, we have shown that circulating levels of mDCPs, pDCPs and tDCPs are significantly reduced in patients with TIA, AIS or AHS. The extent of their decrease significantly correlated with the clinical stage and the radiological size of stroke. Regarding TOAST criteria, no significant differences in circulating DCPs were observed betweens groups with different stroke aetiology, suggesting that it is not a specific pre-existing situation, but rather the common consequences of stroke, that leads to the decrease in circulating DCPs. In short-term follow-up analysis, a rapid recovery of circulating DCPs was observed, suggesting that their transient decrease might be a result of an enhanced migration from blood into the infarcted brain. Supporting this hypothesis, it is known from other inflammatory diseases that a decrease in circulating DCPs is mostly caused by their recruitment into areas of inflammation .
Unexpectedly, no significant changes in the levels of circulating DCPs were observed in patients with ACI-S compared with controls, although it is known that circulating DCPs migrate into atherosclerotic lesions [13,21]. In contrast, in patients with stable coronary artery disease a significant decrease in circulating DCPs was described that is dependent on the extent of coronary atherosclerotic burden . However, a reason that we were unable to detect any alterations may be due to the limited number of patients in the ACI-S group.
Regarding inflammation, a strong inverse correlation was found between circulating DCPs and hsCRP. This is not surprising, as it is known that CRP is able to directly induce an up-regulation of adhesion molecules on endothelial cells , as well as to induce the release of the chemokines CCL2, CCL3 and CCL4 from monocytes . Thus elevated CRP levels are probably a major contributing factor in the migration of circulating DCPs to the sites of inflammation.
As it is possible that the transient decrease in circulating DCPs in acute cerebral ischaemia is caused by their rapid recruitment into the infarcted brain, human cerebral specimens were immunohistochemically analysed. We were able to show that in patients with AIS or AHS numerous mDCs and T-cells were located in the infarcted area. Those results are in concordance with the study of Kostulas et al. , who described mDCs in the rodent infarcted brain. We observed a distribution of mDCs, co-localized with T-cells, around the cerebral vessels in the stroke area, which is a strategic location between the immune system and the CNS. Thus it appears likely that mDCs are initiating an antigen-specific immune response through T-cell activation there. HLA-DR-expressing cells by far exceeded the number of mDCs in the stroke area. HLA-DR expression appeared to be associated with residential cells such as astrocytes. However, it is questionable whether those CNS-resident cells are able to sufficiently induce an immune response comparable with professional antigen-presenting cells. It is possible that the observed expression of HLA-DR on residential cells is only an epiphenomenon induced by invaded immune cells, as described for EAE (experimental autoimmune encephalomyelitis) . An important finding in the present study was that patients with older strokes had persisting mDCs and T-cells, and high HLA-DR expression in the infarcted area, suggesting a long-lasting immune response after stroke. In concordance, long-term persistence of macrophages has been shown previously in the infarcted area after stroke by PET (positron emision tomography) and [11C]-PK11195 .
Recent findings suggest that the immune pathomechanisms of ischaemic and haemorrhagic stroke are very similar . Our present study confirms this opinion, showing that circulating DCPs are reduced in both ischaemic and haemorrhagic stroke, and are recruited into the affected brain under both conditions. According to the general clinical observation of higher severity of intracranial haemorrhage, our results show that the reduction in circulating DCPs is more pronounced in haemorrhagic than ischaemic stroke.
Several limitations of our present study have to be acknowledged. Systemic infections often occur after acute stroke. However, in our present study we had to exclude patients with obvious systemic infections, as it is well known that circulating DCPs are recruited to the site of infection. Thus the exclusion of patients with systemic infections was, in this case, necessary to be able to investigate whether circulating DCPs are reduced in acute stroke due to their recruitment into the infarcted brain.
As described above, the most likely reason for the decrease in circulating DCPs in patients after stroke appears to be their recruitment into the infarcted brain. However, other theoretically possible reasons have to be taken into account: (i) a reduced production of DCPs in the bone marrow, or (ii) an increase in apoptosis of circulating DCPs. These other reasons could not be excluded definitely in our present study. To the best of our knowledge, no published results exist reporting a selective decrease in the production of circulating DCPs in the bone marrow without any impairment in the production of other blood leucocytes. In the case of lymphocytes, an increase in apoptosis has been described as a reason for their transient reduction after stroke ; however, apoptotic cells have a decreased size in flow cytometry. In our analysis, circulating mDCPs and pDCPs had a very constant size, so apoptosis appears to be an unlikely reason for their decrease in the circulation. On the other hand, it is possible that not all mDCs which were found in the infarcted brain were recruited from the circulation, but might have developed from local cerebral cells .
In conclusion, we show for the first time a significant transient decrease in circulating DCPs in patients with stroke. The reason for this decrease is probably an enhanced recruitment of DCPs from blood into the infarcted brain. Fewer circulating DCPs might lead to immunodepression, resulting in frequent systemic infections and worse outcome in stroke patients. On the other hand, mDCs invading into the brain after stroke may activate T-cells against newly recognized neuroepitopes, inducing a long-lasting immune response, which leads to further neurological damage. Thus novel strategies based on modulation of DC migration or function might be a promising therapeutic approach to prevent systemic infections or further neurological damage in stroke patients.
This work was supported by the ELAN Fonds; and the Interdisciplinary Center for Clinical Research of the University Hospital Erlangen.
Abbreviations: ACI-S, asymptomatic cerebral infarction stenosis; AHS, acute haemorrhagic stroke; AIS, acute ischaemic stroke; AT1, angiotensin II type 1 receptor; CNS, central nervous system; CIDS, CNS-injury-induced immunodepression; CRP, C-reactive protein; CT, computer tomography; DC, dendritic cell; DCP, DC precursor; HbA1c, glycated haemoglobin; hsCRP, high-sensitivity CRP; MBP, myelin basic protein; mDC, myeloid DC; NIHSS, National Institute of Health Stroke Scale; pDC, plasmacytoid DC; SAI, stroke-associated infection; TIA, transient ischaemic attack; TOAST, Trial of Org 10172 in Acute Stroke Treatment; WBC, white blood cell
- © The Authors Journal compilation © 2010 Biochemical Society