Preterm intraventricular hemorrhage in vitro: modeling the cytopathology of the ventricular zone

Background Severe intraventricular hemorrhage (IVH) is one of the most devastating neurological complications in preterm infants, with the majority suffering long-term neurological morbidity and up to 50% developing post-hemorrhagic hydrocephalus (PHH). Despite the importance of this disease, its cytopathological mechanisms are not well known. An in vitro model of IVH is required to investigate the effects of blood and its components on the developing ventricular zone (VZ) and its stem cell niche. To address this need, we developed a protocol from our accepted in vitro model to mimic the cytopathological conditions of IVH in the preterm infant. Methods Maturing neuroepithelial cells from the VZ were harvested from the entire lateral ventricles of wild type C57BL/6 mice at 1–4 days of age and expanded in proliferation media for 3–5 days. At confluence, cells were re-plated onto 24-well plates in differentiation media to generate ependymal cells (EC). At approximately 3–5 days, which corresponded to the onset of EC differentiation based on the appearance of multiciliated cells, phosphate-buffered saline for controls or syngeneic whole blood for IVH was added to the EC surface. The cells were examined for the expression of EC markers of differentiation and maturation to qualitatively and quantitatively assess the effect of blood exposure on VZ transition from neuroepithelial cells to EC. Discussion This protocol will allow investigators to test cytopathological mechanisms contributing to the pathology of IVH with high temporal resolution and query the impact of injury to the maturation of the VZ. This technique recapitulates features of normal maturation of the VZ in vitro, offering the capacity to investigate the developmental features of VZ biogenesis.


Introduction
Intraventricular hemorrhage (IVH) is a severe neurological disorder of preterm infants, affecting about 20 percent of the preterm infants born at or below 32 gestational weeks [1]. Up to one half of infants with IVH develop post-hemorrhagic hydrocephalus (PHH), which is the most common etiology of pediatric hydrocephalus in North America [2]. Despite advances in the medical and surgical care of these infants, the prognosis of preterm IVH with PHH shows unacceptable high rates of persistent neurocognitive deficits (up to 85%) and cerebral palsy (up to 70%) [3]. Current treatment methods focus on draining cerebrospinal fluid (CSF) to reduce the excess pressure regardless of hydrocephalus etiology. There have been no substantive advances in the treatment

Open Access
Fluids and Barriers of the CNS *Correspondence: lcastan@wustl.edu 1 Department of Neurological Surgery, Washington University School of Medicine and the St. Louis Children's Hospital, Campus Box 8057, 660 South Euclid Ave., St. Louis, MO 63110, USA Full list of author information is available at the end of the article of PHH in recent years, and few viable targeted therapeutic strategies have been proposed. To improve the care and outcomes for these patients, we must first define the fundamental mechanisms underlying the pathophysiology of IVH and its sequelae, and develop experimental models for testing therapeutic interventions.
An association between IVH, PHH, and neurodevelopmental impairment is well-established, but the mechanisms linking these disorders remain unclear. Recent evidence implicates impairment of cell junction complexes within the ventricular zone (VZ) [4][5][6][7][8][9][10] with associated ciliopathy in the etiology of congenital, nonhemorrhagic hydrocephalus [11][12][13][14][15][16][17][18][19] in both experimental models [4-6, 10, 11, 20-24] and humans [25][26][27][28]. VZ disruption is also identified as a key feature in IVH in humans [29]. Therefore, an understanding of the complex effects of the blood on the VZ is needed to develop preventive treatments for neurological sequelae of IVH/ PHH. This protocol describe an in vitro experimental strategy that provides the ability to control the environment of murine VZ development. The reductionist approach allows a rapid and direct method to address questions related to mechanisms and effects of bloodrelated VZ disruption with high temporal and physiological resolution.
The methods detailed herein are uniquely applied to IVH, PHH, or hydrocephalus more generally. While other groups have cultured ependymal cells (EC) previously [30,31], the notion of using VZ cultures to study the effects of IVH is a novel strategy reported from our group [32]. This protocol is a major step forward in IVH/PHH research, as it enables rigorously controlled mechanistic experiments, with the added flexibility of mimicking various human physiological parameters. For example, studies of the VZ from patients with PHH [29] and congenital hydrocephalus [27,28] indicate that alterations in adherens junction molecules play an important role in EC damage. Furthermore, the VZ culture system eliminates problems such as drug bioavailability, the blood brain barrier, and it allows high-throughput testing of the effect of drugs. Testing of pharmacological agents in vitro is an important first step to not only assess efficacy but also determine optimal ranges of dosing prior in vivo testing.
It is well known that normal multiciliated EC differentiate from monociliated stem cells; however, the complex developmental mechanisms governing this process are still not completely clear. For example, how polarity is established in multiciliated EC and how trafficking of adhesion molecules to the cell membrane occurs, requires further study. Our model will be useful for the study of normal developmental neurobiological processes, the myriad of secondary newborn brain disorders (e.g. hypoxia, inflammation, nutrition and metabolic disorders), and therapeutics explorations designed to mitigate the neurological effects of VZ injury.
The VZ is a layer of tissue mainly composed of neurostem cells that line the fetal ventricular system [33,34]. This layer is a proliferative area involved in the neurogenesis (intermediate progenitors, neurons) and gliogenesis (astrocytes, oligodendrocytes, ependymocytes) [35] of the brain; therefore representing an important platform for brain development. Our model enables investigation of the effect of blood on these precursor cells and specifically the impact on differentiation from VZ to multiciliated EC, the final step in the transition of the VZ. EC are involved in control of fluid movement between brain extracellular space and the CSF, in part to clear metabolites from the interstitial fluid [36]. EC also promote localized movement of fluid and its contents within the ventricles through the synchronized beating of cilia [37]. In fact, deficits in the formation of cilia cause serious disorders in the central nervous system, including hydrocephalus (reviewed in [38][39][40]). Despite the importance of the VZ in brain development, there are few reported methods to differentiate EC from brain neural stem cells (NSC) [30,[41][42][43]. We have recently reported on the effects of syngeneic blood on the VZ using an in vitro model [32], but the details of our procedures have not been published.

Procedure
• Expose P0-P4 C57BL/6 mice to wet ice for 2 min or until the pup is not responsive or use methods as guided by the local IACUC for terminal studies. Once circulatory arrest by cooling is achieved, euthanize the pup by removing the head and gently strip the scalp. • Remove the brain through a lateral incision from the foramen magnum toward the external auditory canal (~ 3 mm) and a superficial sagittal incision from the foramen magnum all the way through the midline using Metzenbaum surgical scissors. After that remove the frontal and parietal bones with curved surgical forceps. • Deposit the brain in a Petri dish cooled in ice with HBSS, 1% PSA. • Working under a stereomicroscope, dissect both cortical hemispheres, as well as the brainstem and the olfactory bulbs by cutting in the midline using a single-edge razor blade (Fig 1). • Position the medial hemispheric side superiorly and using ultrafine forceps, introduce one tip into the lateral ventricle in an anteroposterior direction all the way through the ventricle until the para-hippocampal gyrus is reached. Keep the opposite tip of the forceps in a ventral position regarding the anatomical orientation of the brain hemisphere, and close the forceps completely, producing an incision that divides the hippocampus and thalamus from the lateral wall. With a single, gentle twist of the wrist, remove the hippocampus and thalamus from the lateral wall ( Fig. 1). Peel the meninges off the surface of the cerebral cortex, grasping them at the edge of the hemisphere and pulling away of the hemisphere. The wall dissection should not take longer than an hour and the tissue must not be stored or frozen before carrying out the VZ cell culture.  Dissection of the lateral ventricular wall to obtain VZ cells. Photographs of mouse brain gross morphology to illustrate isolation of the lateral ventricle wall using the following steps: (1) The brain is divided in the midline (dashed line) using a dorsal approach. (2) With the midsagittal surface down, the cerebral hemisphere is separated from the cerebellum (dashed line). (3) Beginning at the midsagittal surface, one sharp tip of ultrafine forceps is inserted into the dorsal lateral ventricle through the septum(s) between the anterior corpus callosum (CC) and the fornix (Fo), and angled so that the tip exits the occipital pole of the cerebral hemisphere; the other arm of the forceps should be external to the ventral surface of the brain. When the para-hippocampal gyrus is reached, the forceps are closed to detach the thalamus (T) and hippocampus medially from the lateral cerebral cortex. (4) By opening the lateral cerebral cortex, lateral wall of the ventricle (LW) is visualized and dissected to provide tissue for cell culture  From this point onward, the cells will progressively differentiate from stem cells to multiciliated EC. On days 3, 5, 7 and thereafter, the differentiation medium is added. About 4%, 23% and 52% of the cells, at days 3, 5, 7, respectively, should be multiciliated [32]. The cultures will remain stable for 10-12 days. • After cervical dislocation, extract blood from 3-month-old mice through an orbital sinus puncture (reviewed by [44]; this age corresponds to the most feasible time to obtain the quantity of blood required for 24 well-plates. Other ages may be used depending on the design and clinical relevance of the experiment. To acquire syngeneic blood, the donors must be from the same C57BL/6 strain. The blood is held in a non-heparinized 50 mL Falcon tube on ice less than 5 min before placing 30 µL of whole blood or serum onto the media that is already in the well, over the VZ cell cultures. This step is carried out under the laminar flow cabinet in sterile conditions to mimic a bleeding process. Room temperature PBS 7.4 was used as control. To select a specific developmental time, consider that the cells are mostly NSC by 3 to 7 days, and mostly EC from 7 days-onward [32].  [32] and the list of antibodies used are available in Table 1.

Immunohistochemistry (IHC)
• Rinse the cells three times with PBS, fix the cells cultured on coverslips with 4% paraformaldehyde in PBS PH 7.4 for 7 min. Wash three times with PBS to remove the remaining fixative. Note, that the fixative used should be optimized for the antibodies employed for IHC.
• Permeabilize the cells with 5% Bovine Serum Albumin in PBS, 1% Triton X-100 for 1 h. • Established IHC routines can be followed to perform IHC, see [30,32]. The list of antibodies used are available in Table 1.

Image analysis
The immunofluorescent images were taken using Pascal confocal microscopy (Oberkochen, Germany). The roundness plug-in of ImageJ (NIH public software) was used to quantify the morphology of the cells labeled with anti-GFAP (see Additional file 1: Figure S1). The roundness is a magnitude that ranges from 1 when the shape is a perfect sphere to 0 when it is a line. For WB images, the molecular imager Chemidoc XRS and image lab 3.0 software were used to quantify band density (Bio-Rad, Hercules, CA).

Statistical analysis
Student's t test or Mann-Whitney U-test were used for parametric or non-parametric data respectively.

EC differentiation
VZ cells in vitro progressively differentiate from monociliated NSC to multiciliated EC [30,32]. To define the development of the EC and monitor the validity this IVH in vitro model, we used well-known VZ/EC markers  (Table 1). Since EC represent a glial lineage, characteristic astrocytic cell markers such as GFAP and S100β were present in the VZ (Fig. 2a-c′). S100β labels mature ependymal cells [45] and GFAP labels NSC and immature EC [9,10,29,46]. Cilia markers are also appropriate to discriminate between NSC (monociliated) and EC (multiciliated) [47]. Double labeling with GFAP and βIV tubulin is appropriate in this model because allows distinction between NSC (if monociliated), immature EC (if multiciliated and positive for GFAP) and mature EC (if multiciliated and negative for GFAP) (Fig. 2a-b′). In this model the cells also express other markers such as epithelial membrane antigen (EMA) and Aquaporin 4 (AQP4) that show selective expression in mature ependymal cells (Fig. 2d, e). VZ/EC also elaborate cell junction proteins such as N-cadherin and β-catenin, indicating attachment and viability of the cell [48]; these proteins are expressed in both NSC and EC (Fig. 2c-d′). An increase in multiciliated ependymal cells over time is detected in control conditions (Fig. 2a-b′ and f ). To confirm this, we quantified the percentage of cells expressing EMA on days 5 (37.51 ± 11.38), 6 (42.93 ± 10.77), and 7 (58.15 ± 10.66) (Fig. 4e, f ); these percentages were similar to those previously reported by our group [32].

Representative results of cell junction pathology
Impairment of cell junctions is a cardinal feature of IVHrelated VZ disruption in human infants [29] and our in vitro model faithfully reproduces this phenomenon, with a decrease in the expression of N-cadherin after 48 h of blood exposure. Figure 3a-b′ illustrates the typical expression of adherens junctions in control (PBS) compared to blood-exposed cultures. There is diminution and dislocation of N-cadherin in blood-treated cultures, while control cultures show a characteristic polygonal shape as previously reported [42]. This model also allows the use of quantitative techniques such as WB to quantify targets of interest to understand this disease; Fig. 3c shows representative results of n-cadherin expression normalized with the expression of GADPH after 48 h in control and IVH conditions. In this case significant differences are found in the ratio of N-Cad/GADPH in control (0.66 ± 0.07) vs blood (0.30 ± 0.04) p < 0.05.

Ventricular zone disruption and glial activation
Periventricular gliosis is another important feature of IVH/PHH in humans and in vivo models, and the VZ culture model also consistently demonstrates such glial activation [10,29]. VZ disruption itself represents an alteration in the NSC/EC with a loss or detachment of these cells, followed by astroglial proliferation. Figure 4 shows a representative IHC in which a decrease in multiciliated EC, labeled with βIV tubulin, is present after 48 h of blood induction.  (Fig. 4e, f ). There was also a change in the morphology of the cells expressing GFAP, with branch-like projections resembling reactive astrocytes (Fig. 4a-b′). To identify this astrocytosis, the morphology (roundness, measured using an ImageJ software application) of the cells seems to be a reliable variable in which the values close to 1 define a perfect sphere compared to values close to 0 which indicate a completely flat profile. Figure 3c shows a graphical representation of the roundness, being significantly lower after blood induction compared to control (respectively 0.44 ± 0.01 vs 0.53 ± 0.01 p < 0.001) due to the increase of elongated projections in blood conditions. The levels of GFAP can also be quantified through WB. Figure 3d shows results from a single WB of control compared to blood, in which an increase in the ratio of GFAP/β actin was found after 48 h of blood (control; 0.87 ± 0.16 vs blood; 1.60 ± 0.44), confirming that blood increases the amount of GFAP in the VZ. Other Inflammatory markers may be used to more completely understand these cytopathological mechanisms.

Discussion
Despite the high prevalence and poor outcomes of PHH, the pathogenesis of this disorder remains poorly understood [49,50]. One major factor limiting scientific progress in this field is a lack of appropriate models to probe the basic biology of the disease and to test novel therapeutic strategies. The current effort describes the protocol of the first in vitro model of preterm IVH/PHH and represents a major step forward for investigators. The in vitro IVH VZ model is a reliable tool to define the biological mechanisms involved in this devastating disorder. To develop this model, an accurate isolation of the lateral wall of the lateral ventricle is a fundamental requirement. We have developed a straightforward and rapid technique to dissect the lateral ventricle by using two ultrafine forceps with just a single step to remove the thalamus, the medial wall of the lateral ventricle, and the hippocampus, versus the 5 steps previously described ( [30,31]).
The skull dissection must be done on cooled ice (Fig. 1) to provide enough stiffness to the fresh brain tissue; otherwise, the tissue may lose its shape and attach to the frontal and parietal bones complicating the bone removal.
To obtain a cell culture that mimics the cellular composition of the VZ with fidelity it is critical to totally remove the hippocampus and the meninges to prevent proliferation of neuronal-committed cells and fibroblasts, respectively. The cells are expanded for 3-5 days or until confluence. At this time point the differentiated cells are mechanically discarded. At the time the differentiation media is added, the cultures are composes of undifferentiated NSC that correspond to newborn lateral wall VZ development reported by Delgehyr et al. [30].
The most innovative process in this model is the addition of blood onto the cultured VZ to mimic the effects IVH on maturing neuroepithelial/ependymal cells. It is key to apply freshly obtained, whole blood onto the VZ cells as soon as possible (i.e., within 5 min). Within those 5 min, it is highly advised to keep the blood in ice cooled 25 mL falcon tubes to prevent coagulation. The presence of clots will impair pipetting the designated amount of blood onto the cells due to a pipet tip obstruction.
Likewise, this model is useful for studying the bloodrelated VZ cytopathology but does not take into account other physiological factors active in hydrocephalus such as increased intracranial pressure, which has been recently modeled using a 3-D neural cell cultures and a newly developed pressure controlled cell culture incubator [51].
While other authors have shown in vitro congenital hydrocephalus-dependent cytopathology [24,42], this technique is a highly specific method to study the cytological mechanisms involved in the VZ after IVH. In this protocol paper, our aim is to describe our model [32] and   blood (b). a shows a qualitatively higher expression of βIV (arrows) and a lower expression of GFAP when compared to blood (b). a′, b′ High magnification (×63) of a, b in which a′ shows several multiciliated EC (asterisk) vs b′ that shows a lack of multiciliated EC and changes in the morphology of the VZ cells with the increase of GFAP projections (arrowhead), as well as dislocation (asterisk) and diminution of the expression of N-Cad. c Graphical representation of the roundness of GFAP-labeled cells after 48 h of treatment (control vs. blood), In which the cells are significantly less round in blood conditions (p < 0.001). Data are mean of the roundness of 2 different experiments with at least n = 10 d GFAP protein levels quantified by WB normalized with β actin after 48 h of treatment, in which blood exposure increases the expression of GFAP. Data are mean from a single experiment with n = 8. e Comparisons of the expression of EMA at day 5 after 3 h of treatment at day 6 after 24 h of treatment and at day 7 after 48 h of treatment. The immunocytochemistry shows a decrease in the expression of EMA at 24 and 48 h after blood induction. f Quantification of EMA-positive cells. After 24 and 48 h treatment the mean percentage of cells expressing EMA was significantly decreased. Data are mean of the percentage of cells expressing EMA, from 2 experiments with at least n = 5 wells at 3, 24, and 48 hours, respectively (***p < 0.001; Student t-test). Scale bars: a, b = 50 μm; a′, b′ = 10 μm; e = 20 μm show representative results. Despite only showing representative results from 3 to 48 h after blood induction, this technique provides a high temporal-resolution tool that opens the possibility to study the cellular response within seconds or minutes, which is problematic in vivo. This protocol also allows investigators to employ drugs to modulate and understand normal VZ development compared to IVH conditions without in vivo impediments, such us crossing the blood brain barrier in drug delivery or tissue dissections.