A method for the isolation of neural stem cells and oligodendrocyte progenitor cells from the brains of live rats is presented here in experimental detail. It allows multiple collections of these cells from the same animals without compromising their well-being.
Tissue-specific neural stem cells (NSCs) remain active in the mammalian postnatal brain. They reside in specialized niches, where they generate new neurons and glia. One such niche is the subependymal zone (SEZ; also called the ventricular-subventricular zone), which is located across the lateral walls of the lateral ventricles, adjacent to the ependymal cell layer. Oligodendrocyte progenitor cells (OPCs) are abundantly distributed throughout the central nervous system, constituting a pool of proliferative progenitor cells that can generate oligodendrocytes.
Both NSCs and OPCs exhibit self-renewal potential and quiescence/activation cycles. Due to their location, the isolation and experimental investigation of these cells is performed postmortem. Here, we describe in detail "brain milking", a method for the isolation of NSCs and OPCs, amongst other cells, from live animals. This is a two-step protocol designed for use in rodents and tested in rats. First, cells are "released" from the tissue via stereotaxic intracerebroventricular (i.c.v.) injection of a "release cocktail". The main components are neuraminidase, which targets ependymal cells and induces ventricular wall denudation, an integrin-β1-blocking antibody, and fibroblast growth factor-2. At a second "collection" step, liquid biopsies of cerebrospinal fluid are performed from the cisterna magna, in anesthetized rats without the need of an incision.
Results presented here show that isolated cells retain their endogenous profile and that NSCs of the SEZ preserve their quiescence. The denudation of the ependymal layer is restricted to the anatomical level of injection and the protocol (release and collection) is tolerated well by the animals. This novel approach paves the way for performing longitudinal studies of endogenous neurogenesis and gliogenesis in experimental animals.
Tissue-specific stem cells are partially committed cells that can give rise to all cell populations that constitute the respective tissues. Apart from being multipotent, they are self-renewing cells and crucial for maintaining the homeostasis and the regenerative capacity of tissues1. Some tissue-specific stem cells remain in an active, strongly proliferative state, such as intestinal or hematopoietic stem cells. Others, such as brain stem cells, remain largely quiescent or dormant2. In the adult brain, neural stem cells (NSCs) can be found in specialized areas, often called niches. Two such well described areas exist in the subependymal zone (SEZ) of the lateral ventricles and in the dentate gyrus of the hippocampus. The SEZ niche generates the highest numbers of cells, primarily neuroblasts that migrate toward the olfactory bulbs and contribute to the local interneuron population; in contrast, generated oligodendroblasts migrate to the adjacent corpus callosum (CC)3. Oligodendrocyte progenitor cells (OPCs) are mitotically active cells, widely distributed throughout the central nervous system, that: i) are committed to the oligodendroglial lineage, ii) can migrate to sites of demyelination, and iii) can differentiate into myelinating oligodendrocytes. OPCs also exhibit self-renewal potential and quiescence4.
Until now, the isolation and study of NSCs and OPCs required postmortem dissociation of the dissected brain and spinal cord tissue. To circumvent this experimental limitation, we established a method that allows, for the first time, the isolation of brain NSCs and OPCs from live animals. We call this method "milking", because it enables multiple collections of cells as their pools are not depleted. The protocol was developed in rats, due to their large brain size, targeting mainly the SEZ, or the CC, and includes two major steps. First, NSCs or OPCs are "removed" from the tissue via i.c.v. injection of a "release cocktail" containing neuraminidase, a toxin that induces ventricular wall denudation, an integrin-β1-blocking antibody, and fibroblast growth factor 2 (FGF2). The cocktail is stereotaxically injected bilaterally within the lateral ventricles. If the intended use is the isolation of NSCs, rostral areas of the lateral ventricles are targeted. If the aim is to isolate OPCs more purely, the cocktail is injected caudally in the area of the hippocampal fimbria. At a second "collection" step, liquid biopsies of cerebrospinal fluid (CSF) are performed from the cisterna magna of anesthetized rats, without the need of an incision. The liquid biopsy is mixed with NSC culture medium and can be kept at 4 °C until plating.
Animal breeding, maintenance, and experimental procedures were conducted in accordance with the UK Animals (Scientific Procedures) Act 1986, authorized by the Home Office, and with the Presidential Decree 56/2013 of the Hellenic Republic, scrutinized by the Animal Welfare and Ethical Review Bodies of the Universities of Cambridge and Patras, as well as approved and scrutinized by the local Prefectural Animal Care and Use Committee (Protocol number: 5675/39/18-01-2021). Male and female Sprague-Dawley, Wistar, and Long-Evans rats, with ages varying between 2 and 4 months and with body weights between 150 g and 250 g, were used. The protocol is graphically summarized in Figure 1.
1. Release cocktail preparation
NOTE: Prepare fresh on the day of the procedure and keep on ice. The quantities are given per 2 µL to be i.c.v. injected in each lateral ventricle. Prepare an additional 1 µL per intended injection.
2. Injection of release cocktail
NOTE: The whole process can be performed within 20 min. Take care to perform the surgery in aseptic conditions. Clean all surfaces with antiseptic (e.g., 3% or 6% hydrogen peroxide). Use autoclaved or readily sterile tools, gloves, gowns, and drapes.
3. Cerebrospinal fluid (CSF) liquid biopsy
NOTE: The whole process can be performed within 10 min. The liquid biopsy described here is performed 3 days post injection of the release cocktail but can be performed in exactly the same way whenever required. Take care to perform the surgery in aseptic conditions. Clean all the surfaces with antiseptic (e.g., 3% or 6% hydrogen peroxide). Use autoclaved or readily sterile tools, gloves, gowns, and drapes.
4. Processing of tissue for immunofluorescence
5. Processing isolated cells for immunofluorescence
6. Microscopy and image analysis
Release and collection of NSCs
NSCs of the SEZ are separated from the CSF only by the monolayer of ependymal cells, albeit they remain in direct contact with the ventricular content via intercalating mono-ciliated processes8,9. Neuraminidase acts specifically on ependymal cells via cleavage of sialic acid residues and can induce denudation of the ventricular wall. This leads to neuroblast clustering on the surface of the ventricle10,11. Moreover, a flow of neuroblasts has been observed in the CSF after the i.c.v. injection of an integrin-β1-blocking antibody, probably due to loosening of the inter-ependymal cell junctions12. These observations have led to the development of a protocol that enables the isolation of the brain's stem and progenitor cells via controlled compromise of the integrity of the lateral ventricle's walls. In a first step, the release from the parenchyma and the subsequent flow of NSCs or OPCs inside the CSF are induced via i.c.v. injection of the release cocktail. The cocktail is stereotaxically injected at a rate of 1 µL/min, bilaterally (2 µL per injection) in the lateral ventricles (coordinates targeting SEZ NSCs: AP = 0.3 mm, L = ± 1.2 mm, D = 3.5 mm; coordinates targeting CC OPCs: AP = 1.5 mm, L = ± 2 mm, D = 3.5 mm). The second ("collection") step involves the performance of CSF liquid biopsies from the cisterna magna. The rats need to be anesthetized and the biopsy can be done with 1 mL syringes. The use of the stereotaxic device allows almost complete success in retrieving approximately 100 µL of CSF, without the need for incisions. The liquid biopsy is added to iced culture medium and is kept at 4 °C until plating for <3 h (NSC culture medium contains Dulbecco's modified Eagle medium [DMEM], B27 supplement [2%], N2 supplement [1%], FGF2 [20 ng/mL], and epidermal growth factor [EGF; 20 ng/mL]).
Histological assessment of the periventricular area after the injection of the release cocktail
A first cohort of experiments revealed that when injecting more than 3 µL of liquid, the ventricles could be damaged due to non-specific, mechanical injury (Figure 2B,C). The slow injection of 2 µL of release cocktail led to the emergence of clusters of Dcx-immunopositive neuroblasts at the ventricular surface. These clusters remained visible even at 8 months post injection (Figure 2D). As the method was intended for longitudinal studies, aiming at long-term follow-up of the animals, the tissue damage caused by the release cocktail was assessed. Immunostaining was performed on 14 µm thick cryostat brain sections for ependymal cell markers, such as S100β and β-catenin13, and the overall integrity of the ependymal layer was assessed. Sites of denudation of the ependymal layer were present only close to the rostrocaudal level of the injections, with a gradual decline in ependymal layer perturbations detected at more posterior and more anterior areas of the SEZ (Figure 2E,F) and becoming absent after a distance of ±2 mm from the site of the injection. The above-mentioned results show that the partial denudation of the ependymal layer caused by i.c.v. injection of the release cocktail is focal, restrained in the proximity of the injection site, and leaves the rest of the periventricular ependymal layer intact.
Marker profile and in vitro behavior of collected cells
Subsequently, the in vitro behavior and the marker profile of the cells isolated via the milking protocol were assessed. The average cell yield of each liquid biopsy is approximately 300 ± 45 cells (per biopsy of 100 µL)7. Milking biopsies resulted in NSC cultures with an average of 3.17 ± 0.45 passage potential. Cells isolated from saline-injected rats could be passaged on average 1.92 ± 0.76 times; in contrast, those isolated via milking even reached nine passages (p = 0.038, t test) (Figure 3A)7. The average passaging capacity of standard postmortem, SEZ-derived neurosphere cultures is higher than 12 passages in our hands. Because the in vivo expansion potential of SEZ NSCs, as revealed by in vivo cell-fate mapping experiments14, has been shown to be limited, milking produces cells with significantly different in vitro behavior than that of cells in standard cultures, albeit much closer to the behavior of endogenous NSCs. Collected cells were plated on poly-D-lysine-coated wells, where they grew both as adherent monolayers and more rarely as neurospheres (Figure 3B). Freshly isolated cells (collected 3 days post injection and fixed 24 h after plating) that were immunopositive for the astroglial marker GFAP were also immunopositive for ID3, a marker of quiescence, and had the characteristic for NSC bipolar morphology (Figure 3C). Moreover, as reported previously7, a more detailed immunocytochemical comparison of biopsy-derived cells and of postmortem-derived cells from the same animals, at 3 days post injection (looking for GFAP+ astrocytes, Dcx+ neuroblasts, PDGFRa+ oligodendrocyte progenitors, and SOX2+ cells of the neural lineage) showed that the profile of collected cells was similar to that of endogenous SEZ cells (Figure 3D). Notably, when biopsy- and tissue-derived cells were compared in different conditions (e.g., injection of a release cocktail with and without FGF2), it was found that the growth factor resulted in a concomitant and significant increase in the presence of SOX2+ cells, as well as in a significant decrease in the presence of Dcx+ neuroblasts in both samples (Figure 3D). These data confirmed that any changes appearing in the profile of endogenous populations of NSCs are mirrored in milking-generated cell samples.
Figure 1: Graphical summary of milking. A coronal section of one brain hemisphere with the major anatomical landmarks (the lateral ventricle, the overlying corpus callosum, and the anterior commissure below [white matter tracts in grey], and the subependymal zone at the lateral walls [in blue]). The release cocktail is injected in the lateral ventricle, leading to compromise of the integrity of the tissue and the release of postnatal brain neural stem cells in the cerebrospinal fluid, from which they can be collected via liquid biopsies. Abbreviations: SEZ = subependymal zone; LV = lateral ventricle; CSF = cerebrospinal fluid; pbNSCs = postnatal brain neural stem cells; β1-int = beta1 integrin. Please click here to view a larger version of this figure.
Figure 2: Histological assessment of the effects of i.c.v. injections. (A–D) Low-magnification images of the dorsal half of the lateral ventricle after immunostaining for Dcx (in red, to mark neuroblasts). (A) The simple i.c.v. insertion of a Hamilton syringe does not disturb the cyto-architecture of the SEZ, while the i.c.v. injection of 10 mL (infusion rate of 1 mL/min) leads to severe damage of the ventricular wall irrespective of its content. (B) Saline and (C) release cocktail. (D) The injection of 2 µL of the release cocktail leads to a controlled compromise of the ventricular wall, observed even at 8 months after surgery. Higher-magnification detail of the wall of the SEZ at 7 and 14 days post injection are shown in (E) and (F), respectively. There is a disorganized structure, with Dcx+ neuroblasts (in green) resting at the surface of the wall and the other cells of the niche (Sox2+, in white) resting deeper. The periventricular tissue is damaged at the rostrocaudal level of the injection at the 2 month time point (in G), while the tissue is intact at a more caudal level (in H) in the same animal. The ventricular wall is assessed by immunostaining for ependymal markers S100β and β-catenin. Detail of the typical architecture of the wall is shown in (I). Nuclear staining is performed using DAPI (shown in blue). Scale bars = 300 µm (A–C,G,H, insets), 150 µm (D), 30 µm (E,F), and 50 µm (I). This figure is modified from McClenahan et al.13. Abbreviations: SEZ = subependymal zone; i.c.v = intracerebroventricular; Dcx = doublecortin; DAPI = 4',6-diamidino-2-phenylindole. Please click here to view a larger version of this figure.
Figure 3: Assessment of the cells isolated via milking. (A) Graph showing the maximum number of passages obtained per liquid biopsy sample from saline-injected animals (gray bars, total of 12 samples), or after milking (black bars, total of 29 samples, release cocktail with neuraminidase and integrin-β1-blocking antibody). (B) Representative confocal microscopy image of primary cells obtained via milking, at 3 days post injection, immunostained against GFAP and ID3. (C,D) Brightfield images of cells isolated 3 days post injection, plated on poly-D-lysine-coated wells and allowed to grow in NSC proliferation medium for 7 days. (E) Graph showing the cell-type profile of cells isolated via milking of the SEZ and of the endogenous population of SEZ NSCs from the same experimental animals. The SOX2+ and the Dcx+ fractions were significantly increased and decreased, respectively, after the co-injection of FGF2. (One-way ANOVA analysis per marker, followed by post hoc analysis; n = 4-6 animals per experimental group.) Scale bars = 100 µm (C,D) and 10 µm (B). This figure is modified from McClenahan et al.13. Abbreviations: SEZ = subependymal zone; DPI = days post injection; NSC = neural stem cells; Dcx = doublecortin; GFAP = glial fibrillary acidic protein. Please click here to view a larger version of this figure.
Stem and progenitor cells are relatively sparse in mammalian brain tissue. In addition, NSCs are located in areas inaccessible for easy and safe biopsies (ventricular walls, hippocampus). Therefore, the only way to work experimentally with such cells, so far, has been their postmortem isolation. A method allowing the single or repeated collection of NSCs and OPCs from live rats, named milking, is described here step by step. The method is based on two key features: i) NSCs or OPCs are separated by the ependymal cell monolayer from the CSF, flowing within the ventricular system of the brain; ii) ependymal cells and neural progenitors retain contact between them and with other neighboring cell types via integrins. Therefore, the release cocktail injected in specific areas of the ventricles to enable the disengagement of NSCs, or OPCs, from the brain parenchyma contains neuraminidase, a toxin that specifically targets and kills ependymal cells10,11, and an integrin-β1-blocking antibody. It also contains FGF2, as this growth factor is essential for the survival and maintenance of NSCs in the niche and in culture15,16. It has been previously shown7 that this protocol is well tolerated by the animals; in this report, it is further confirmed that the damage to ependymal cells, induced by the release cocktail as a prerequisite for the release of NSCs/OPCs in the CSF, is limited around the rostrocaudal level of the injection. This has been of key importance, since the protocol should preserve the homeostatic function of the brain, including the stem cell niche itself, and should not compromise the long-term well-being of the animals. The stereotaxic injection of the release cocktail is of mild severity and has never resulted in death. Furthermore, the performance of liquid biopsies from the cisterna magna is a quick and minimally invasive procedure that can be repeated several consecutive times.
Importantly, the NSC samples isolated via milking closely reflect the endogenous pool of NSCs. The profile of SEZ progenitors can be influenced by the i.c.v. injection of FGF2. When this happens, the profile of milking-derived cells is affected in the same way. Moreover, previous experimental work has indicated that NSCs isolated via milking of the SEZ have limited self-renewal potential, a behavior similar to that of endogenous NSCs as revealed with in vivo, transgenic, cell-fate strategies14,17. Postnatal brain NSCs are retained in quiescence, rarely transiting toward mitotic activation to generate neurogenic and gliogenic progeny18,19,20. Here, evidence is provided that the fraction of milking-derived cells that express GFAP and are expected to include bona fide NSCs co-express ID3, a marker of quiescent NSCs21. The enriched presence of quiescent NSCs, which are then plated in the NSC medium typically used for the culture of NSCs isolated postmortem, seems to limit the expansion potential of collected NSCs (e.g., a process crucial if cells were to be used for autologous transplantations). This is because these media are designed specifically for the survival and mitotic expansion of activated NSCs; thus, the generation of protocols enabling the efficient and rapid mitotic activation of quiescent NSCs will increase the value and scope of use of milking.
Overall, these data indicate that milking enables the sampling of postnatal brain NSCs and OPCs (depending on the rostrocaudal level of injecting the release cocktail) from live rats. This work indicates the feasibility of conducting successive liquid biopsies in the same animal, even at significant time lengths after injection of the release cocktail (thus, to plan and perform longitudinal experimental studies), since neuroblasts remain clustered on the ventricular walls even at 8 months post injection. Such methods are of critical importance because they allow the investigation of events within individual animals resulting in reduced biological noise generated by physiological variation. This, in turn, leads to enhanced accuracy and to the implementation of the principles of the "3Rs" (replacement, reduction, and refinement). Future key steps for the improvement of the method will be the determination of the maximum number and the timeframe that liquid biopsies can be performed after one release procedure, albeit the performance of additional release procedures should be feasible, if necessary.
The authors have nothing to disclose.
This work was supported by an Action Medical Research (UK) grant (GN2291) to R.J.M.F. and I.K. The research work was also partly supported (animal costs and support to D.D) by the Hellenic Foundation for Research and Innovation (H.F.R.I.) under the "First Call for H.F.R.I. Research Projects to support Faculty members and Researchers and the procurement of high-cost research equipment grant" (Project Number: 3395).
Release cocktail | |||
β1-integrin-blocking antibody | BD Biosciences | #555002 | purified NA/LE Hamster Anti-Rat CD29 Clone Ha2/5, 1 mg/mL. Any abntibody with blocking activity should be appropriate. |
Neuraminidase from Clostridium perfringens (Clostridium welchii) | Sigma-Aldrich | #N2876 | Neuraminidases fromother sources (e.g., from Vibrio cholerae) have not been tested. |
Recombinant Human FGF-basic (154 a.a.) | Peprotech | #100-18B | kept as a 1 μg/μL stock, diluted in sterile water at -20 °C |
Surgical procedures | |||
10 µL Syringe | Hamilton | #80330 | Model 701 RN, Small Removable Needle, 26s gauge, 2 in., point style 2 |
BD Micro-fine 1 mL insulin syringes | BD biosciences | 04085-00 | 29 G x 12.7 mm |
BETADINE CUT.SOL 10% FLx30ML | LAVIPHARM-CASTALIA | SKU: 5201048131168 | |
Bupaq | RICHTERPHARMA | 1021854AF | 10 mL (buprenophine 0.3 mg/mL) |
Digital New Standard Stereotaxic, Rat and Mouse | Stoelting | 51500D | |
Homeothermic Monitoring System | Harvard Apparatus | 55-7020 | |
ISOFLURIN 1,000 mg/g inhalation vapour, liquid | Vetpharma Animal Health | 32509/4031 | |
Ketamidor | RICHTER PHARMA | SKU: 9004114002531 | Ketamine 100 mg/mL |
Nylon suture, Ethilon | Ethicon | D9635 | Clear , size 5-0 |
Rechargeable Cordless Surgical Trimmers | Stoelting | Item:51472 | |
Scalpel blades, sterile | Swann Morton | AW050 | |
Scopettes Jr. 8-inch Swabs | Birchwood Laboratories | 34-7021-12P | |
Stereotaxic High Speed Drill | Foredom | 1474w/o1464 | |
Stoelting’s Stereotaxic Instrument Kit | Stoelting | Item: 52189 | |
Xylan 2% | Chanelle Pharmaceuticals | 13764/03/19-5-2004 | Xylazine, 25 mL |
Tissue and cells handling and immunostainings | |||
96-well plates appropriate for microscopy | Greiner | #655866 | Screen star microplate |
B27 supplement | ThermoFisher Scientific | A1486701 | |
Bovine Serum Albumin (BSA) | Merck | P06-1391100 | Fraction V, heat shock |
Citrate | Merck | 71497 | Sodium citrate monobasic |
Cryostat | Leica | CM1510S | |
DAPI | Merck, Calbiochem | 28718-90-3 | Nuclear staining, Dilution: 1/1,000 |
DMEM | ThermoFisher Scientific | 11995065 | High glucose, pyruvate |
donkey anti-goat | Biotium | 20016 or 20106 or 20048 | Dilution: 1/1,000 |
donkey anti-mouse | Biotium | 20014 or 20105 or 20046 | Dilution: 1/1,000 |
donkey anti-rabbit | Biotium | 20015 or 20098 or 20047 | Dilution: 1/1,000 |
EGF | Peprotech | 315-09 | |
FGF-2 (or bFGF) | Peprotech | 100-18B | |
goat anti-GFAP | Abcam | ab53554 | Dilution: 1/500 |
goat anti-SOX2 | Santa Cruz Biotecnology | sc-17320 | Dilution: 1/200 |
mouse anti-ID3 | Santa Cruz Biotecnology | sc-56712 | Dilution: 1/200 |
mouse anti-S100β | Sigma | S2532 | Dilution: 1/200 |
Mowiol | Merck, Calbiochem | 475904 | Mounting medium |
N2 supplement | ThermoFisher Scientific | 17502048 | |
Parafolmadehyde | Merck | 158127 | |
Poly-D-Lysine | Merck, Millipore | A-003-E | Solution, 1.0 mg/mL |
rabbit anti-Doublecortin (DCX) | Abcam | ab18723 | Dilution: 1/500 |
rabbit anti-PDGFRα | Abcam | ab51875 | Dilution: 1/200 |
rabbit anti-β- catenin | Abcam | ab16051 | Dilution: 1/500 |
Triton X-100 | Merck | X100 | |
Microscopy and image analysis | |||
Confocal microscope | Leica | SP6 and SP8 | |
Image analysis | NIH, USA | ImageJ | |
Image analysis | Leica | LasX |