Neurologische Klinik und Poliklinik
Refine
Has Fulltext
- yes (69)
Is part of the Bibliography
- yes (69)
Year of publication
- 2022 (69) (remove)
Document Type
- Journal article (51)
- Doctoral Thesis (18)
Keywords
- ischemic stroke (6)
- stroke (6)
- Parkinson’s disease (5)
- Fabry disease (4)
- GFAP (4)
- Polyneuropathie (4)
- Schlaganfall (4)
- neurodegeneration (4)
- neuroinflammation (4)
- Autoantikörper (3)
Institute
- Neurologische Klinik und Poliklinik (69)
- Institut für diagnostische und interventionelle Neuroradiologie (ehem. Abteilung für Neuroradiologie) (8)
- Medizinische Klinik und Poliklinik I (6)
- Rudolf-Virchow-Zentrum (5)
- Deutsches Zentrum für Herzinsuffizienz (DZHI) (4)
- Institut für Virologie und Immunbiologie (4)
- Klinik und Poliklinik für Kinder- und Jugendpsychiatrie, Psychosomatik und Psychotherapie (4)
- Klinik und Poliklinik für Psychiatrie, Psychosomatik und Psychotherapie (4)
- Pathologisches Institut (4)
- Institut für Klinische Epidemiologie und Biometrie (3)
Sonstige beteiligte Institutionen
Rag1\(^{−/−}\) mice, lacking functional B and T cells, have been extensively used as an adoptive transfer model to evaluate neuroinflammation in stroke research. However, it remains unknown whether natural killer (NK) cell development and functions are altered in Rag1\(^{−/−}\) mice as well. This connection has been rarely discussed in previous studies but might have important implications for data interpretation. In contrast, the NOD-Rag1\(^{null}\)IL2rg\(^{null}\) (NRG) mouse model is devoid of NK cells and might therefore eliminate this potential shortcoming. Here, we compare immune-cell frequencies as well as phenotype and effector functions of NK cells in Rag1\(^{−/−}\) and wildtype (WT) mice using flow cytometry and functional in vitro assays. Further, we investigate the effect of Rag1\(^{−/−}\) NK cells in the transient middle cerebral artery occlusion (tMCAO) model using antibody-mediated depletion of NK cells and adoptive transfer to NRG mice in vivo. NK cells in Rag1\(^{−/−}\) were comparable in number and function to those in WT mice. Rag1\(^{−/−}\) mice treated with an anti-NK1.1 antibody developed significantly smaller infarctions and improved behavioral scores. Correspondingly, NRG mice supplemented with NK cells were more susceptible to tMCAO, developing infarctions and neurological deficits similar to Rag1−/− controls. Our results indicate that NK cells from Rag1−/− mice are fully functional and should therefore be considered in the interpretation of immune-cell transfer models in experimental stroke. Fortunately, we identified the NRG mice, as a potentially better-suited transfer model to characterize individual cell subset-mediated neuroinflammation in stroke.
Idiopathic Parkinson’s disease (PD) is characterized by a progredient degeneration of the brain, starting at deep subcortical areas such as the dorsal motor nucleus of the glossopharyngeal and vagal nerves (DM) (stage 1), followed by the coeruleus–subcoeruleus complex; (stage 2), the substantia nigra (SN) (stage 3), the anteromedial temporal mesocortex (MC) (stage 4), high-order sensory association areas and prefrontal fields (HC) (stage 5) and finally first-order sensory association areas, premotor areas, as well as primary sensory and motor field (FC) (stage 6). Autoimmunity might play a role in PD pathogenesis. Here we analyzed whether anti-brain autoantibodies differentially recognize different human brain areas and identified autoantigens that correlate with the above-described dissemination of PD pathology in the brain. Brain tissue was obtained from deceased individuals with no history of neurological or psychiatric disease and no neuropathological abnormalities. Tissue homogenates from different brain regions (DM, SN, MC, HC, FC) were subjected to SDS-PAGE and Western blot. Blots were incubated with plasma samples from 30 PD patients and 30 control subjects and stained with anti-IgG antibodies to detect anti-brain autoantibodies. Signals were quantified. Prominent autoantigens were identified by 2D-gel-coupled mass spectrometry sequencing. Anti-brain autoantibodies are frequent and occur both in healthy controls and individuals with PD. Glial fibrillary acidic protein (GFAP) was identified as a prominent autoantigen recognized in all plasma samples. GFAP immunoreactivity was highest in DM areas and lowest in FC areas with no significant differences in anti-GFAP autoantibody titers between healthy controls and individuals with PD. The anti-GFAP autoimmunoreactivity of different brain areas correlates with the dissemination of histopathological neurodegeneration in PD. We hypothesize that GFAP autoantibodies are physiological but might be involved as a cofactor in PD pathogenesis secondary to a leakage of the blood–brain barrier.
During ischemic stroke, infarct growth before recanalization diminishes functional outcome. Hence, adjunct treatment options to protect the ischemic penumbra before recanalization are eagerly awaited. In experimental stroke targeting two different pathways conferred protection from penumbral tissue loss: (1) enhancement of hypoxic tolerance of neurons by deletion of the calcium channel subunit Orai2 and (2) blocking of detrimental lymphocyte–platelet responses. However, until now, no preclinical stroke study has assessed the potential of combining neuroprotective with anti-thrombo-inflammatory interventions to augment therapeutic effects. We induced focal cerebral ischemia in Orai2-deficient (Orai2\(^{-/-}\)) mice by middle cerebral artery occlusion (MCAO). Animals were treated with anti-glycoprotein Ib alpha (GPIbα) Fab fragments (p0p/B Fab) blocking GPIbα–von Willebrand factor (vWF) interactions. Rat immunoglobulin G (IgG) Fab was used as the control treatment. The extent of infarct growth before recanalization was assessed at 4 h after MCAO. Moreover, infarct volumes were determined 6 h after recanalization (occlusion time: 4 h). Orai2 deficiency significantly halted cerebral infarct progression under occlusion. Inhibition of platelet GPIbα further reduced primary infarct growth in Orai2\(^{-/-}\) mice. During ischemia–reperfusion, upon recanalization, mice were likewise protected. All in all, we show that neuroprotection in Orai2\(^{-/-}\) mice can be augmented by targeting thrombo-inflammation. This supports the clinical development of combined neuroprotective/anti-platelet strategies in hyper-acute stroke.
Recent advances in proteomic technologies now allow unparalleled assessment of the molecular composition of a wide range of sample types. However, the application of such technologies and techniques should not be undertaken lightly. Here, we describe why the design of a proteomics experiment itself is only the first step in yielding high-quality, translatable results. Indeed, the effectiveness and/or impact of the majority of contemporary proteomics screens are hindered not by commonly considered technical limitations such as low proteome coverage but rather by insufficient analyses. Proteomic experimentation requires a careful methodological selection to account for variables from sample collection, through to database searches for peptide identification to standardised post-mass spectrometry options directed analysis workflow, which should be adjusted for each study, from determining when and how to filter proteomic data to choosing holistic versus trend-wise analyses for biologically relevant patterns. Finally, we highlight and discuss the difficulties inherent in the modelling and study of the majority of progressive neurodegenerative conditions. We provide evidence (in the context of neurodegenerative research) for the benefit of undertaking a comparative approach through the application of the above considerations in the alignment of publicly available pre-existing data sets to identify potential novel regulators of neuronal stability.
Cerebrospinal fluid (CSF) and serum biomarkers are critical for clinical decision making in neurological diseases. In cerebral small vessel disease (CSVD), white matter hyperintensities (WMH) are an important neuroimaging biomarker, but more blood-based biomarkers capturing different aspects of CSVD pathology are needed. In 42 sporadic CSVD patients, we prospectively analysed WMH on magnetic resonance imaging (MRI) and the biomarkers neurofilament light chain (NfL), glial fibrillary acidic protein (GFAP), chitinase3-like protein 1 (CHI3L1), Tau and Aβ1-42 in CSF and NfL and GFAP in serum. GFAP and CHI3L1 expression was studied in post-mortem brain tissue in additional cases. CSVD cases with higher serum NfL and GFAP levels had a higher modified Rankin Scale (mRS) and NIHSS score and lower CSF Aβ1-42 levels, whereas the CSF NfL and CHI3L1 levels were positively correlated with the WMH load. Moreover, the serum GFAP levels significantly correlated with the neurocognitive functions. Pathological analyses in CSVD revealed a high density of GFAP-immunoreactive fibrillary astrocytic processes in the periventricular white matter and clusters of CHI3L1-immunoreactive astrocytes in the basal ganglia and thalamus. Thus, besides NfL, serum GFAP is a highly promising fluid biomarker of sporadic CSVD, because it does not only correlate with the clinical severity but also correlates with the cognitive function in patients.
We have previously shown that targeting endoneurial macrophages with the orally applied CSF-1 receptor specific kinase (c-FMS) inhibitor PLX5622 from the age of 3 months onwards led to a substantial alleviation of the neuropathy in mouse models of Charcot-Marie-Tooth (CMT) 1X and 1B disease, which are genetically-mediated nerve disorders not treatable in humans. The same approach failed in a model of CMT1A (PMP22-overexpressing mice, line C61), representing the most frequent form of CMT. This was unexpected since previous studies identified macrophages contributing to disease severity in the same CMT1A model. Here we re-approached the possibility of alleviating the neuropathy in a model of CMT1A by targeting macrophages at earlier time points. As a proof-of-principle experiment, we genetically inactivated colony-stimulating factor-1 (CSF-1) in CMT1A mice, which resulted in lower endoneurial macrophage numbers and alleviated the neuropathy. Based on these observations, we pharmacologically ablated macrophages in newborn CMT1A mice by feeding their lactating mothers with chow containing PLX5622, followed by treatment of the respective progenies after weaning until the age of 6 months. We found that peripheral neuropathy was substantially alleviated after early postnatal treatment, leading to preserved motor function in CMT1A mice. Moreover, macrophage depletion affected the altered Schwann cell differentiation phenotype. These findings underscore the targetable role of macrophage-mediated inflammation in peripheral nerves of inherited neuropathies, but also emphasize the need for an early treatment start confined to a narrow therapeutic time window in CMT1A models and potentially in respective patients.
Introduction/Aims
Schwann cell clusters have been described at the murine dermis-epidermis border. We quantified dermal Schwann cells in the skin of patients with small-fiber neuropathy (SFN) compared with healthy controls to correlate with the clinical phenotype.
Methods
Skin punch biopsies from the lower legs of 28 patients with SFN (11 men, 17 women; median age, 54 [range, 19-73] years) and 9 healthy controls (five men, four women, median age, 34 [range, 25-69] years) were immunoreacted for S100 calcium-binding protein B as a Schwann cell marker, protein-gene product 9.5 as a pan-neuronal marker, and CD207 as a Langerhans cell marker. Intraepidermal nerve fiber density (IENFD) and subepidermal Schwann cell counts were determined.
Results
Skin samples of patients with SFN showed lower IENFD (P < .05), fewer Schwann cells per millimeter (P < .01), and fewer Schwann cell clusters per millimeter (P < .05) than controls. When comparing SFN patients with reduced (n = 13; median age, 53 [range, 19-73] years) and normal distal (n = 15, median age, 54 [range, 43-68] years) IENFD, the number of solitary Schwann cells per millimeter (p < .01) and subepidermal nerve fibers associated with Schwann cell branches (P < .05) were lower in patients with reduced IENFD. All three parameters correlated positively with distal IENFD (P < .05 to P < .01), whereas no correlation was found between Schwann cell counts and clinical pain characteristics.
Discussion
Our data raise questions about the mechanisms underlying the interdependence of dermal Schwann cells and skin innervation in SFN. The temporal course and functional impact of Schwann cell presence and kinetics need further investigation.
Background and purpose
The aim was to characterize a combined vestibular, ocular motor and postural syndrome induced by deep brain stimulation (DBS) of the subthalamic nucleus in a patient with Parkinson's disease.
Methods
In a systematic DBS programming session, eye, head and trunk position in roll and pitch plane were documented as a function of stimulation amplitude and field direction. Repeat ocular coherence tomography was used to estimate ocular torsion. The interstitial nucleus of Cajal (INC), zona incerta (ZI) and ascending vestibular fibre tracts were segmented on magnetic resonance imaging using both individual and normative structural connectomic data. Thresholded symptom-associated volumes of tissue activated (VTA) were calculated based on documented stimulation parameters.
Results
Ipsilateral ocular tilt reaction and body lateropulsion as well as contralateral torsional nystagmus were elicited by the right electrode in a current-dependent manner and subsided after DBS deactivation. With increasing currents, binocular tonic upgaze and body retropulsion were observed. Symptoms were consistent with an irritative effect on the INC. Symptom-associated VTA was found to overlap with the dorsal ZI and the ipsilateral vestibulothalamic tract, while lying rather distant to the INC proper. A ZI-to-INC ‘incerto-interstitial’ tract with contact to the medial-uppermost portion of the VTA could be traced.
Conclusion
Unilateral stimulation of INC-related circuitry induces an ipsilateral vestibular, ocular motor and postural roll-plane syndrome, which converts into a pitch-plane syndrome when functional activation expands bilaterally. In this case, tractography points to an incerto-interstitial pathway, a tract previously only characterized in non-human primates. Directional current steering proved useful in managing this rare side effect.
Repetitive transcranial magnetic stimulation (rTMS) is a non-invasive brain stimulation technique that has the potential to treat a variety of neurologic and psychiatric disorders. The extent of rTMS-induced neuroplasticity may be dependent on a subject’s brain state at the time of stimulation. Chronic low intensity rTMS (LI-rTMS) has previously been shown to induce beneficial structural and functional reorganisation within the abnormal visual circuits of ephrin-A2A5\(^{-/-}\) mice in ambient lighting. Here, we administered chronic LI-rTMS in adult ephrin-A2A5\(^{-/-}\) mice either in a dark environment or concurrently with voluntary locomotion. One day after the last stimulation session, optokinetic responses were assessed and fluorescent tracers were injected to map corticotectal and geniculocortical projections. We found that LI-rTMS in either treatment condition refined the geniculocortical map. Corticotectal projections were improved in locomotion+LI-rTMS subjects, but not in dark + LI-rTMS and sham groups. Visuomotor behaviour was not improved in any condition. Our results suggest that the beneficial reorganisation of abnormal visual circuits by rTMS can be significantly influenced by simultaneous, ambient visual input and is enhanced by concomitant physical exercise. Furthermore, the observed pathway-specific effects suggest that regional molecular changes and/or the relative proximity of terminals to the induced electric fields influence the outcomes of LI-rTMS on abnormal circuitry.
Ischemic disorders are the leading cause of death worldwide. The extracellular signal-regulated kinases 1 and 2 (ERK1/2) are thought to affect the outcome of ischemic stroke. However, it is under debate whether activation or inhibition of ERK1/2 is beneficial. In this study, we report that the ubiquitous overexpression of wild-type ERK2 in mice (ERK2\(^{wt}\)) is detrimental after transient occlusion of the middle cerebral artery (tMCAO), as it led to a massive increase in infarct volume and neurological deficits by increasing blood–brain barrier (BBB) leakiness, inflammation, and the number of apoptotic neurons. To compare ERK1/2 activation and inhibition side-by-side, we also used mice with ubiquitous overexpression of the Raf-kinase inhibitor protein (RKIP\(^{wt}\)) and its phosphorylation-deficient mutant RKIP\(^{S153A}\), known inhibitors of the ERK1/2 signaling cascade. RKIP\(^{wt}\) and RKIP\(^{S153A}\) attenuated ischemia-induced damages, in particular via anti-inflammatory signaling. Taken together, our data suggest that stimulation of the Raf/MEK/ERK1/2-cascade is severely detrimental and its inhibition is rather protective. Thus, a tight control of the ERK1/2 signaling is essential for the outcome in response to ischemic stroke.
Both nerve injury and complex regional pain syndrome (CRPS) can result in chronic pain. In traumatic neuropathy, the blood nerve barrier (BNB) shielding the nerve is impaired—partly due to dysregulated microRNAs (miRNAs). Upregulation of microRNA-21-5p (miR-21) has previously been documented in neuropathic pain, predominantly due to its proinflammatory features. However, little is known about other functions. Here, we characterized miR-21 in neuropathic pain and its impact on the BNB in a human-murine back translational approach. MiR-21 expression was elevated in plasma of patients with CRPS as well as in nerves of mice after transient and persistent nerve injury. Mice presented with BNB leakage, as well as loss of claudin-1 in both injured and spared nerves. Moreover, the putative miR-21 target RECK was decreased and downstream Mmp9 upregulated, as was Tgfb. In vitro experiments in human epithelial cells confirmed a downregulation of CLDN1 by miR-21 mimics via inhibition of the RECK/MMP9 pathway but not TGFB. Perineurial miR-21 mimic application in mice elicited mechanical hypersensitivity, while local inhibition of miR-21 after nerve injury reversed it. In summary, the data support a novel role for miR-21, independent of prior inflammation, in elicitation of pain and impairment of the BNB via RECK/MMP9.
Fibromyalgia syndrome (FMS) is a heterogeneous chronic pain syndrome characterized by musculoskeletal pain and other key co-morbidities including fatigue and a depressed mood. FMS involves altered functioning of the central and peripheral nervous system (CNS, PNS) and immune system, but the specific molecular pathophysiology remains unclear. Anti-cholinergic treatment is effective in FMS patient subgroups, and cholinergic signaling is a strong modulator of CNS and PNS immune processes. Therefore, we used whole blood small RNA-sequencing of female FMS patients and healthy controls to profile microRNA regulators of cholinergic transcripts (CholinomiRs). We compared microRNA profiles with those from Parkinson's disease (PD) patients with pain as disease controls. We validated the sequencing results with quantitative real-time PCR (qRT-PCR) and identified cholinergic targets. Further, we measured serum cholinesterase activity in FMS patients and healthy controls. Small RNA-sequencing revealed FMS-specific changes in 19 CholinomiRs compared to healthy controls and PD patients. qRT-PCR validated miR-182-5p upregulation, distinguishing FMS patients from healthy controls. mRNA targets of CholinomiRs bone morphogenic protein receptor 2 and interleukin 6 signal transducer were downregulated. Serum acetylcholinesterase levels and cholinesterase activity in FMS patients were unchanged. Our findings identified an FMS-specific CholinomiR signature in whole blood, modulating immune-related gene expression.
Background
Long-term support of stroke patients living at home is often delivered by family caregivers (FC). We identified characteristics of stroke patients being associated with receiving care by a FC 3-months (3 M) after stroke, assessed positive and negative experiences and individual burden of FC caring for stroke patients and determined factors associated with caregiving experiences and burden of FC 3 M after stroke.
Methods
Data were collected within TRANSIT-Stroke, a regional telemedical stroke-network comprising 12 hospitals in Germany. Patients with stroke/TIA providing informed consent were followed up 3 M after the index event. The postal patient-questionnaire was accompanied by an anonymous questionnaire for FC comprising information on positive and negative experiences of FC as well as on burden of caregiving operationalized by the Caregiver Reaction Assessment and a self-rated burden-scale, respectively. Multivariable logistic and linear regression analyses were performed.
Results
Between 01/2016 and 06/2019, 3532 patients provided baseline and 3 M-follow-up- data and 1044 FC responded to questionnaires regarding positive and negative caregiving experiences and caregiving burden. 74.4% of FC were older than 55 years, 70.1% were women and 67.5% were spouses. Older age, diabetes and lower Barthel-Index in patients were significantly associated with a higher probability of receiving care by a FC at 3 M. Positive experiences of FC comprised the importance (81.5%) and the privilege (70.0%) of caring for their relative; negative experiences of FC included financial difficulties associated with caregiving (20.4%). Median overall self-rated burden was 30 (IQR: 0–50; range 0–100). Older age of stroke patients was associated with a lower caregiver burden, whereas younger age of FC led to higher burden. More than half of the stroke patients in whom a FC questionnaire was completed did self-report that they are not being cared by a FC. This stroke patient group tended to be younger, more often male with less severe stroke and less comorbidities who lived more often with a partner.
Conclusions
The majority of caregivers wanted to care for their relatives but experienced burden at the same time. Elderly patients, patients with a lower Barthel Index at discharge and diabetes are at higher risk of needing care by a family caregiver.
Trial registration
The study was registered at “German Clinical Trial Register”: DRKS00011696. https://www.drks.de/drks_web/navigate.do?navigationId=trial.HTML&TRIAL_ID=DRKS00011696
Background
Regulatory CD4\(^+\)CD25\(^+\)FoxP3\(^+\) T cells (Treg) are a subgroup of T lymphocytes involved in maintaining immune balance. Disturbance of Treg number and impaired suppressive function of Treg correlate with Parkinson’s disease severity. Superagonistic anti-CD28 monoclonal antibodies (CD28SA) activate Treg and cause their expansion to create an anti-inflammatory environment.
Methods
Using the AAV1/2-A53T-α-synuclein Parkinson’s disease mouse model that overexpresses the pathogenic human A53T-α-synuclein (hαSyn) variant in dopaminergic neurons of the substantia nigra, we assessed the neuroprotective and disease-modifying efficacy of a single intraperitoneal dose of CD28SA given at an early disease stage.
Results
CD28SA led to Treg expansion 3 days after delivery in hαSyn Parkinson’s disease mice. At this timepoint, an early pro-inflammation was observed in vehicle-treated hαSyn Parkinson’s disease mice with elevated percentages of CD8\(^+\)CD69\(^+\) T cells in brain and increased levels of interleukin-2 (IL-2) in the cervical lymph nodes and spleen. These immune responses were suppressed in CD28SA-treated hαSyn Parkinson’s disease mice. Early treatment with CD28SA attenuated dopaminergic neurodegeneration in the SN of hαSyn Parkinson’s disease mice accompanied with reduced brain numbers of activated CD4\(^+\), CD8\(^+\) T cells and CD11b\(^+\) microglia observed at the late disease-stage 10 weeks after AAV injection. In contrast, a later treatment 4 weeks after AAV delivery failed to reduce dopaminergic neurodegeneration.
Conclusions
Our data indicate that immune modulation by Treg expansion at a timepoint of overt inflammation is effective for treatment of hαSyn Parkinson’s disease mice and suggest that the concept of early immune therapy could pose a disease-modifying option for Parkinson’s disease patients.
Objectives
The pathogenesis of fibromyalgia syndrome (FMS) is unclear. Transcranial ultrasonography revealed anechoic alteration of midbrain raphe in depression and anxiety disorders, suggesting affection of the central serotonergic system. Here, we assessed midbrain raphe echogenicity in FMS.
Methods
Sixty-six patients underwent transcranial sonography, of whom 53 were patients with FMS (27 women, 26 men), 13 patients with major depression and physical pain (all women), and 14 healthy controls (11 women, 3 men). Raphe echogenicity was graded visually as normal or hypoechogenic, and quantified by digitized image analysis, each by investigators blinded to the clinical diagnosis.
Results
Quantitative midbrain raphe echogenicity was lower in patients with FMS compared to healthy controls (p<0.05), but not different from that of patients with depression and accompanying physical pain. Pain and FMS symptom burden did not correlate with midbrain raphe echogenicity as well as the presence and severity of depressive symptoms.
Conclusion
We found reduced echogenicity of the midbrain raphe area in patients with FMS and in patients with depression and physical pain, independent of the presence or severity of pain, FMS, and depressive symptoms. Further exploration of this sonographic finding is necessary before this objective technique may enter diagnostic algorithms in FMS and depression.
Differences in stem cell marker and osteopontin expression in primary and recurrent glioblastoma
(2022)
Background
Despite of a multimodal approach, recurrences can hardly be prevented in glioblastoma. This may be in part due to so called glioma stem cells. However, there is no established marker to identify these stem cells.
Methods
Paired samples from glioma patients were analyzed by immunohistochemistry for expression of the following stem cell markers: CD133, Musashi, Nanog, Nestin, octamer-binding transcription factor 4 (Oct4), and sex determining region Y-box 2 (Sox2). In addition, the expression of osteopontin (OPN) was investigated. The relative number of positively stained cells was determined. By means of Kaplan–Meier analysis, a possible association with overall survival by marker expression was investigated.
Results
Sixty tissue samples from 30 patients (17 male, 13 female) were available for analysis. For Nestin, Musashi and OPN a significant increase was seen. There was also an increase (not significant) for CD133 and Oct4. Patients with mutated Isocitrate Dehydrogenase-1/2 (IDH-1/2) status had a reduced expression for CD133 and Nestin in their recurrent tumors. Significant correlations were seen for CD133 and Nanog between OPN in the primary and recurrent tumor and between CD133 and Nestin in recurrent tumors. By confocal imaging we could demonstrate a co-expression of CD133 and Nestin within recurrent glioma cells. Patients with high CD133 expression had a worse prognosis (22.6 vs 41.1 months, p = 0.013). A similar trend was seen for elevated Nestin levels (24.9 vs 41.1 months, p = 0.08).
Conclusions
Most of the evaluated markers showed an increased expression in their recurrent tumor. CD133 and Nestin were associated with survival and are candidate markers for further clinical investigation.