A 3-dimensional extracellular matrix as a delivery system for the transplantation of glioma-targeting neural stem/progenitor cells

Abstract

Neural stem/progenitor cells (NSPCs) display inherent pathotropic properties that can be exploited for targeted delivery of therapeutic genes to invasive malignancies in the central nervous system. Optimizing transplantation efficiency will be essential for developing relevant NSPC-based brain tumor therapies. To date, the real-world issue of handling and affixing NSPCs in the context of the neurosurgical resection cavity has not been addressed. Stem cell transplantation using biocompatible devices is a promising approach to counteract poor NSPC graft survival and integration in various types of neurological disorders. Here, we report the development of a 3-dimensional substrate that is based on extracellular matrix purified from tissue-engineered skin cultures (3DECM). 3DECM enables the expansion of embedded NSPCs in vitro while retaining their uncommitted differentiation status. When implanted in intracerebral glioma models, NSPCs were able to migrate out of the 3DECM to targeted glioma growing in the contralateral hemisphere, and this was more efficient than the delivery of NSPC by intracerebral injection of cell suspensions. Direct application of a 3DECM implant into a tumor resection cavity led to a marked NSPC infiltration of recurrent glioma. The semisolid consistency of the 3DECM implants allowed simple handling during the surgical procedure of intracerebral and intracavitary application and ensured continuous contact with the surrounding brain parenchyma. Here, we demonstrate proof-of-concept of a matrix-supported transplantation of tumor-targeting NSPC. The semisolid 3DECM as a delivery system for NSPC has the potential to increase transplantation efficiency by reducing metabolic stress and providing mechanical support, especially when administered to the surgical resection cavity after brain tumor removal.

The prognosis of patients with malignant glioma remains poor despite advances in neurosurgical techniques and the introduction of radiotherapy plus concomitant and adjuvant temozolomide.¹ The time of survival significantly depends on the extent of the surgical tumor resection.^2,3 The highly infiltrative nature of glioma cells often renders a complete surgical removal impossible, inevitably leading to tumor recurrence. Therefore, the eradication of invading glioma cells before they give rise to a recurrent tumor is a highly desirable therapeutic goal for the treatment of low- and high-grade gliomas.

Neural stem/progenitor cells (NSPCs) display an inherent and extensive tropism to areas of neuropathology such as brain malignancy, which render them ideal vectors for the targeted delivery of therapeutic gene products to areas of tumor cell invasion.^4,5 Previous studies have demonstrated that NSPCs migrate long distances within the host brain and specifically enrich in the tumor mass of gliomas or brain metastasis.^6–9 When genetically modified to express therapeutically relevant molecules transplanted NSPCs were able to induce a significant inhibition of tumor growth in preclinical brain tumor models.^6,10,11 A localized therapy using tumor-targeting NSPCs has the advantage of circumventing barriers of systemic drug delivery and the potential to increase the therapeutic efficiency while minimizing the risk of systemic toxicity.⁴

The implantation of biodegradable carmustine-releasing wafers directly into the resection cavity after surgical removal of a tumor mass is a clinically proven local therapy that is safe and offers a survival benefit to patients with newly diagnosed or recurrent malignant gliomas.^12,13 However, in contrast to chemotherapeutics diffusing out of intracavitary applied wafers the drug delivery by tumor-targeting NSPCs may offer additional therapeutic benefits, because they are able to migrate throughout the brain and specifically colocalize with invasive brain tumor cells distant from the resection cavity.

Ideally, a motile NSPC-based local glioma therapy would be initiated intraoperatively directly after surgical tumor resection to destroy the remaining infiltrated glioma cells thus preventing tumor recurrence. However, the surgical resection is a traumatic injury to the brain tissue negatively affecting local cell-based therapies.^14,15 Even without a major surgical wound, reduced survival and integration of grafted cells after direct intraparenchymal transplantation into the central nervous system (CNS) when applied as a cell suspension is a well-recognized problem in various types of neurological disorders.^16,17 In addition, the injection of cell suspensions into a surgical resection cavity or into the wall of a resection site poses a mechanical problem, which most likely results in a reduction of available cells. In clinical gene therapy trials for glioblastoma, it became apparent that local injections of viral or cellular suspensions into the surgical resection wall frequently result in loss of suspension due to reflux of fluid.^18,19 To minimize the fluid loss multiple and deep local intraparenchymal injections into the surgical resection wall were performed nonetheless potentially increasing the risk of neurological deficits, especially when the tumor is localized close to functionally significant brain regions. In order to overcome the disadvantages of transplanting cells as cell suspensions and to support survival and integration of grafted cells, various biomaterial-based devices have been developed to aid tissue restoration in neurological disorders such as spinal cord injury, stroke, and Parkinsons's disease.^16,20–22

For the treatment of fast-growing tumors like glioblastomas, it will be crucial to provide sufficient numbers of tumor-targeting NSPCs in the tumor invasion area as soon as possible after cytoreductive surgery. This is the first study to demonstrate the feasibility of a matrix-supported intracerebral and intracavitary transplantation of glioma-targeting NSPC that has the potential to optimize transplantation efficiency. Our aim was to develop a 3-dimensional extracellular matrix (3DECM) by adapting purified human collagen products, which are derived from neonatal human fibroblasts using tissue-engineering technology, and that are already approved for clinical use in reconstructive and cosmetic surgery.^23–26 The collagen-based 3DECM allowed the transplantation of high numbers of viable NSPCs, while itself not inducing differentiation of 3DECM-embedded NSPCs. The 3DECM was permissive for external chemotactic signals from gliomas thus enabling a tumor-targeted NSPC migration out of the matrix. As bioengineered human collagen devices have been successfully manufactured under good manufacturing practice and on commercial scales (kg/year) and have proven to be both safe and biocompatible, our approach would provide for rapid translation into clinical studies in humans.

Materials and Methods

Cell Culture

The human glioblastoma cell line NCE-G55²⁷ and the murine glioma cell line GL261 (DCTDC Tumor Repository) were cultured in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen) supplemented with 10% fetal calf serum (FCS, Invitrogen). Cells were maintained in T-75 tissue culture flasks in 5% CO₂/95% air at 37°C in a humidified incubator and were routinely passaged at confluency. For the intracranial implantation experiments, glioma cells were dispersed with a 0.05% solution of trypsin/EDTA (Invitrogen), washed with phosphate-buffered saline (PBS) and adjusted to the final concentration in PBS.

Culture, Labeling, and Differentiation of NSPCs

NSPCs were harvested from the frontoparietal brain of 3-week-old C57BL/6 mice as described previously.²⁸ The cells were grown as neurospheres in complete NSPC growth medium containing neurobasal medium (Invitrogen) with B27 supplement (20 µL/mL; Invitrogen), Glutamax (10 µL/mL, Invitrogen), fibroblast growth factor-2 (20 ng/mL, FGF-2, Peprotech), epidermal growth factor (EGF, 20 ng/mL, Peprotech), and heparin (32 IE/mL, Ratiopharm). Growth factors and heparin were refreshed twice weekly. NSPCs were routinely split by mechanical dissociation when they reached 200–500 µm. EGFP-expressing NSPCs were established by retroviral transduction with pMSCV-EGFP using a commercially available kit (MSCV Retroviral Expression System, BD Biosciences). NSPC labeling using the lipophilic tracer DiI (Molecular Probes) was performed for 30 minutes according to the manufacturer's protocol. Freshly dissociated NSPCs were differentiated in 8-chamber Lab Tec-slides (Nalge Nunc International) at 5000 cells per well. Differentiation medium consisted of neurobasal medium with B27 supplement (20 µL/mL), Glutamax (10 µL/mL), 10% FCS (all from Invitrogen), 1 mM retinoic acid, and 10 mg/mL cAMP (both from Sigma-Aldrich).

Preparation of 3DECM

Human extracellular matrix purified from tissue-engineered skin cultures was used to produce 3DECM implants.^29,30 Implants were fabricated starting with atelopeptide collagens 0.012 N HCl pH ∼2.0 at a final concentration of 3.0 mg/mL (Orion Biosolutions). After neutralization to physiological pH (∼7.5) by the addition of sodium phosphate buffer, collagen fibrils were allowed to form after incubation at 37°C for 16–24 hours. The resulting collagen fibers were centrifuged at ∼10 000×g for 10 minutes, then the supernatant was removed, and the pellet was washed once in 1 mL of DMEM/F12 (Invitrogen) with N2 supplement (Invitrogen) and 0.1% (w/v) bovine serum albumin (Sigma-Aldrich) by centrifugation again at 10 000×g for 10 minutes. The resulting pellet was then allowed to partially dry overnight by incubation in a 37°C oven. Depending on the grade of hydration, the 3DECM could be prepared in various degrees of consistency ranging from semifluid to hard. In order to enhance cellular adhesion, the 3DECM was then coated with 200 µg/mL purified laminin from EHS sarcoma (Sigma-Aldrich), which consists primarily of laminin-1 and nidogen-1 complexed, at 37°C overnight.

In Vivo Studies

Orthotopic glioblastoma xenografts were established in 4- to 6-week-old NMRI-nu/nu or in C57BL/6 mice for the syngeneic glioma model (both Charles River). Mice were anesthetized (100 mg/kg ketamine and 5 mg/kg xylazine) and received a stereotactically guided injection of 4.5 × 10⁴ human NCE-G55 glioblastoma cells or 1.4 × 10⁵ murine GL261 glioma cells into the left forebrain (2 mm lateral and 1 mm anterior to the bregma, at a 2.5 mm depth from the skull surface). Eight days after tumor cell injection, mice were anesthetized and a ∼2 × 2 mm cortical incision into the contralateral right forebrain was performed. One 3DECM implant loaded with NSPCs (from cultures less than passage 7) was placed into the cortical incision. Control groups received NSPC-loaded 3DECM implants in the absence of a glioblastoma xenograft. For experiments assessing transplantation efficiency NSPC-loading of 3DECM implants was checked by histological analysis directly prior to implantation and only those implants containing approximately 1 × 10⁵ GFP-positive NSPCs per 1 mm³ were used for transplantation in vivo. Intracerebral delivery of NSPCs as a single cell suspension was performed by stereotactically guided injection of 1 × 10⁵ freshly dissociated GFP-positive NSPCs in 3 µL of PBS into the cortex of the right forebrain.

In order to mimic the clinical scenario of a glioma resection, we also used the recently developed glioma surgical resection model as described previously.³¹ In brief, 12 days after NCE-G55 cell injection, established human glioblastoma xenografts were surgically removed using a microsurgical technique. At that time, some degree of tumor cell invasion had occurred resulting in the formation of small tumor extensions and satellites distant from the main tumor mass.²⁷ Microsurgical removal was pursued until clear resection margins were visible. After achieving hemostasis, NSPC-loaded 3DECM was placed into the resection cavity. Skin incisions were closed by suturing. All animals were anesthetized and perfused with 4% paraformaldehyde 16 days after tumor cell injection. Brains were removed, embedded in tissue tek O.C.T. (OCT), and stored at −80°C until further processed for histological analysis. All animal studies were performed in accordance with institutional guidelines.

Histological Analyses

Frozen brains and 3DECM embedded in OCT were cut into serial 10-µm sections and counterstained with hematoxylin and eosin (H&E) or 4′-6-diamidino-2-phenylindole (DAPI) for histological evaluation. Cells or frozen sections were fixed with 4% paraformaldehyde and permeabilized with 3% Triton X-100 in PBS (except for GalC staining) and blocked with 5% horse serum. Primary antibodies were mouse anti-nestin (1:500; BD Biosciences), mouse anti-MAP2 (1:50; Chemicon), mouse anti-NF (1:50; Dako), rabbit anti-GFAP (1:40; Dako), mouse anti-GalC (1:100; Chemicon), and mouse anti-MIB5 (1:50; Dako). After incubation for 90 minutes, slides were washed with 5% horse serum. Secondary antibodies, donkey anti-mouse IgG rhodamine (1:50; Chemikon) and donkey anti-rabbit IgG fluorescein (1:50; Chemicon) were added for 30 minutes. For MIB5 staining, we used the DAKO EnVision™+ System HRP kit. Slides were mounted using Vectashield Hard Set mounting medium with DAPI (Vector Laboratories). For double staining, mouse antibodies were added first for 90 minutes, followed by the addition of the rabbit antibody for 60 minutes and simultaneous detection with secondary antibodies. Negative control slides were obtained by omitting the primary antibody. The proliferation index was quantified by counting the number of positively MIB-5 stained cells of 100 nuclei in 5 randomly chosen high-power fields. Quantification of migrated NSPCs localized within the tumor mass was performed by counting GFP-positive cells in 5 randomly chosen high-power fields (HPF = 0.1584 mm²) of different tumor sections for each glioma xenograft. Tumor volumes were calculated with the formula: volume = (square root of maximal tumor cross-sectional area).²⁷

Results

The NSPCs used in this study were isolated from the frontoparietal brain of 3-week-old C57BL/6 mice. Cells grew primarily as neurospheres (Fig. 1A) and cultures became expandable for up to 50 passages. The NSPCs expressed the stem cell marker nestin (Fig. 1B) and musashi-1 (data not shown), while lacking the expression of mature markers of glial and neuronal lineage (data not shown). After 1 week under culture conditions favoring differentiation, numerous cells were immunoreactive for the astroglial marker GFAP, the neuronal marker MAP2 (Fig. 1C), or the oligodendroglial marker GalC (Fig. 1D), indicating multipotency.

Fig. 1.

Morphological and immunocytochemical characterization of NSPCs derived from the brains of 3-week-old C57BL/6 mice. (A) Phase contrast image of a typical neurosphere culture in the presence of 20 ng/mL FGF-2 and EGF. (B) The majority of cells displayed immunoreactivity for nestin (red). (C) Multipotency was demonstrated by differentiation into GFAP-positive astrocytes (green), MAP2-positive neuronal cells (red), and (D) GalC-positive oligodendrocytes (green). Cells were counterstained with DAPI. Bar: (A) 100 µm, (B and D) 50 µm, (C) 20 µm.

In Vitro Culture and Characteriziation of 3DECM-Embedded NSPCs

We have developed a 3DECM preparation soft enough to allow a simple cell loading procedure, whereas still having a consistency hard enough to ensure practical handling during a surgical procedure. Loading of 3DECM for a final implant volume of 5–8 mm³ was performed by adding 3.5 × 10⁵ NSPCs, followed by centrifugation at 1500 × g for 5 minutes to encase the NSPCs with the 3DECM, while simultaneously forming an implant that could be handled with forceps. Immediately after the loading procedure, the majority of NSPCs were found as single cells adherent to fibers within the 3DECM (Fig. 2A). 3DECM implants containing NSPCs were cultured separately in 24-well plates with complete neural stem cell growth medium under the same conditions used for normal neurosphere cultures. Within 7 days, NSPCs embedded in a 3DECM implant formed multiple neurosphere-like clusters (Fig. 2B) and were spreading between the collagen fibers as cell planes (Fig. 2C), indicating that the 3DECM had no negative impact on the cells. The whole 3DECM implant was interspersed with NSPCs (Fig. 2D), whereas maintaining a consistency that allows the surgical handling necessary for transplantation procedures. The collagen fibers within the matrix were the origin of NSPC growth and the formation of neurosphere-like clusters (Fig. 2E and F). The proliferation rate of NSPCs embedded in 3DECM implants was 39% ± 4.9% (mean ± standard deviation, n = 4), indicating that diffusion of nutrients was possible throughout the 3DECM implant, allowing the expansion of cells without any negative impact (Fig. 2G). This proliferation rate was not significantly different from NSPCs that were cultured under normal NSPC conditions growing as neurospheres (41% ± 11%). As the 3DECM is intended as a delivery system of undifferentiated and motile NSPCs, we next asked whether the 3DECM is changing the differentiation status of the embedded NSPCs. Similar to NSPCs growing under normal cell culture conditions, the majority of NSPCs embedded in 3DECM displayed immunoreactivity for nestin (Fig. 2H), while being negative for mature neuronal (NF and MAP2) and oligodendroglial markers (GalC). Only rarely were GFAP-positive cells (<1%) found (Fig. 2I).

Fig. 2.

In vitro culture of 3DECM-embedded neural stem cells. (A) Phase contrast images at the rim of an NSPC-loaded 3DECM implant demonstrated the distribution of single NSPCs (arrow) around collagen fibers within the 3DECM immediately after loading. One week after in vitro culture under neural stem cell conditions (B) numerous neurosphere-like clusters (arrow) and (C) cell planes (arrow) between fibers were apparently distributed (D) throughout the 3DECM implant. H&E stained cryosections of 3DECM implants 7 days after NSPC loading demonstrated a growth pattern of NSPCs around collagen fibers (arrows) as (E) neurosphere-like clusters or (F) multiple cell layers. (G) Proliferating NSPCs within the 3DECM were identified by the expression of MIB-5 (brown cell nuclei). (H) One week after in vitro culture, nestin immunoreactivity (red) confirmed the uncommitted state of the 3DECM-embedded NSPCs. (I) Astroglial differentiation as demonstrated by GFAP immunoreactivity (green) was a rare event (<1%). Cells in (H) and (I) were counterstained with DAPI. Bar: (A–C, E, F) 100 µm, (D) 1 mm, (G) 50 µm.

Intracerebral Transplantation of 3DECM-Embedded NSPCs

In order to assess whether 3DECM-embedded NSPCs are able to migrate out of the 3DECM implant and to target a growing glioma, we transplanted an NSPC-containing 3DECM into the right murine forebrain in the presence of an established glioma in the left contralateral hemisphere (Fig. 3A and B). 3.5 × 10⁵ DiI-labeled or GFP-expressing NSPCs were loaded to an implant volume of ∼5 mm³ and cultured in complete NSPC growth medium overnight. The following day NSPC-containing 3DECM implants were transplanted to the hemisphere contralateral to a growing glioma using the syngeneic murine glioma cell line GL261 in C57BL/6 mice (n = 3) or the human glioblastoma NCE-G55 cell line in nude mice (n = 3). One week after intracerebral 3DECM application, the NSPCs had migrated out of the implant (Fig. 3C) towards the GL261 or NCE-G55 glioma (Fig. 3D and E). DiI-labeled or GFP-expressing cells were dispersed throughout the entire tumor mass. Other than the 3DECM transplantation site and the tumor mass, no NSPCs were observed elsewhere, which suggests the 3DECM-embedded NSPCs were able to detect and respond to chemotactic signals as has been described elsewhere for injected NSPCs by ourselves⁸ and others.^6,7,10 In control animals without tumor (n = 4), DiI-labeled NSPCs remained at the transplantation site and were enriched in the contact zone of the brain parenchyma and 3DECM (Fig. 3F). No NSPCs were detected in the contralateral hemisphere. The consistency of the 3DECM implants enabled close contact with the adjacent brain parenchyma; and in the immunocompetent C57BL/6 mice, no obvious inflammatory reaction was observed based on H&E stained cryosections 1 week after 3DECM implantation (Fig. 3G). Already 36 days after transplantation of 3DECM-embedded NSPCs, no remnants of the 3DECM implant were detectable, whereas a few GFP-positive NSPCs were still found only at the transplantation site (Fig. 3H). The animals did not show any signs of distress and histological analysis did not reveal any significant increase of cellular infiltration as a sign of inflammatory response.

Fig. 3.

Glioma-targeted migration of 3DECM-embedded NSPCs after intracerebral transplantation. (A and B) Tumor tropism of NSPC out of 3DECM implants was assessed by the administration of implants into the contralateral cortex of growing GL261 or NCE-G55 glioma. (C) One week after surgery DiI-labeled NSPCs (red) migrated out of the 3DECM implant (*), enriched in the adjacent surrounding brain parenchyma, and migrated towards glioma growing in the contralateral hemisphere. DiI-labeled (red) and GFP-expressing (green) NSPCs were observed throughout the tumor mass of GL261- (D) (arrows = tumor border) and NCE-G55 glioma (E). (F) In control animals with no tumor, DiI-labeled NSPCs remained at the adjacent brain parenchyma surrounding the 3DECM implant (*) and did not display any distant migration. (G) The plasticity of 3DECM implants ensured close contact to the surrounding brain parenchyma as demonstrated by an H&E counterstained cryosection at the transplantation site. (H) Thirty-six days after 3DECM implantation in control animals without tumor, no 3DECM remnants or signs of a significant inflammatory response were detected at the transplantation site. A few surviving GFP-positive NSPCs (green) were found only at the transplantation site (*cortical defect; arrows = outlining former 3DECM implantation site). Sections (B–F and H) were counterstained with DAPI. Bars: 100 µm.

Comparison of Intracerebral NSPC Transplantation Efficiency

We next asked whether the intracerebral transplantation of NSPCs embedded in a 3DECM is more efficient to deliver glioma-targeting NSPCs than the intracerebral injection of NSPCs as a single cell suspension. Eight days after intracerebral GL261 glioma cell injection into the left forebrain of C57BL/6 mice, 1 × 10⁵ GFP-expressing NSPCs were delivered to the contralateral hemisphere either by single injection as a cell suspension (n = 4) or by transplantation of an NSPC-containing 3DECM implant (n = 4). One week after intracerebral transplantation, we compared the efficiency of both delivery methods by the quantification of intratumoral GFP-positive NSPCs. The application of NSPCs embedded within 3DECM implants led to statistically significantly higher numbers of intratumoral GFP-positive NSPCs than the application as cell injections (mean ± SD: 27.3 ± 5.8 vs 18.9 ± 8.4 NSPCs per HPF; Fig. 4). Tumor volumes of glioma xenografts were not significantly different in both groups excluding different tumor-sized induced chemotactic effects (NSPCs in 3DECM: 0.79 mm³ vs NSPC injection: 0.85 mm³). This experiment demonstrated that the transplantation of 3DECM-embedded NSPCs for the intracerebral delivery of tumor-targeting NSPCs is superior to the conventional intraparenchymal injection of cell suspensions.

Fig. 4.

Quantification of intratumoral NSPCs (Box and Whisker chart). The intracerebral delivery of 1 × 10⁵ 3DECM-embedded NSPCs led to higher numbers of intratumoral NSPCs when compared with the delivery of 1 × 10⁵ NSPCs as a single intraparenchymal cell injection (P < 0.01; unpaired Student t-test).

Intracavitary Transplantation of 3DECM-Embedded NSPCs in the Glioma Resection Model

In order to mimic the clinical scenario of a stem cell application into the resection cavity directly after glioma surgery, we tested the 3DECM as a delivery system for NSPCs in an orthotopic glioma resection model (Fig. 5A; n = 3). Twelve days after intracerebral NCE-G55 glioma cell injection, the established tumor mass was microsurgically removed. 3DECM implants of ∼5–8 mm³ size loaded with 3.5 × 10⁵ GFP-expressing NSPCs were prepared the day before surgery. After achieving hemostasis, 1 NSPC-containing 3DECM implant was then administered directly into a tumor resection cavity. The 3DECM implants could be cut to a size matching the proportions of the surgical resection cavity. As expected, histological analysis a week later revealed that the invading glioma cells that remained in the infiltration zone after surgery had formed new tumor masses. Remnants of NSPC-containing 3DECM were found at the former resection site, which had been filled out by recurrent tumor (Fig. 5B). GFP-expressing NSPCs were observed at the rim of the recurrent tumor (Fig. 5C) and also dispersed throughout the glioma mass (Fig. 5D). No NSPCs were found elsewhere in the brain.

Fig. 5.

Transplantation of 3DECM-embedded NSPCs into the surgical resection cavity. (A) Microsurgical resection of human NCE-G55 glioblastoma xenografts was immediately followed by intracavitary administration of NSPCs within a 3DECM implant. (B) Residues of 3DECM containing GFP-expressing NSPCs (green) were found in the former tumor resection cavity. (C) GFP-expressing NSPCs (green) have migrated out of the 3DECM being enriched in the border zone of the recurrent tumor (*) and adjacent brain parenchyma. (D) Furthermore, GFP-expressing NSPCs were observed throughout the recurrent tumor mass regrown into the resection cavity. Sections were counterstained with DAPI. Bars: 100 µm.

Discussion

Optimized transplantation efficiency remains a challenge in stem cell-based therapeutic strategies for a variety of neurological disorders.¹⁶ Previous attempts at using biomaterial-based stem cell transplantation approaches had mainly aimed to provide a scaffold for tissue regeneration and restoration supporting appropriate differentiation and integration of grafted and host cells locally at the transplantation site.^16,20–22 In contrast, our aim was to develop a biomaterial-based device for the delivery of a high number of motile NSPCs to the dynamic invasion zone of infiltrating gliomas after tumor resection. In this study, we report the development of a 3DECM based on biodegradable material that enables the intracerebral and intracavitary transplantation of high numbers of viable, undifferentiated, and motile tumor-targeting NSPCs.

Numerous factors are known to compromise the survival rate of CNS-grafted cells including growth factor withdrawal, hypoxia, metabolic stress, loss of cell–cell and extracellular matrix contact, and mechanical stress from the transplantation procedure itself.^17,32–34 Sufficient access to the host microvasculature is necessary for the survival of grafted cells, especially if high numbers of cells are injected as a cell suspension.³⁵ Furthermore, injuries leaving a cavity behind as in brain or spinal cord injuries, ischemic lesions, or after surgical tumor resection lack any structural assistance for grafted cells. In this regard, it has been previously shown that the injection of an NSPC suspension resulted in significantly less survival of CNS-grafted cells when compared with a matrix-supported NSPC application.³⁶

In conjunction with the rarity of the unique cells suitable and available for clinical trials of cell transplantation in neurological disorders, a multitude of biosynthetic materials have been developed to improve the results of tissue regeneration.^16,20,21 A recent study reported the use of a synthetic biodegradable polymer polyglycolic acid (PGA) for the implantation of murine NSPCs in a mouse model of hypoxic-ischemic brain injury.²² NSPC administration on a PGA-based scaffold into a intracerebral infarction cavity significantly supported the reconstitution of cortical tissue by facilitating graft–host cell interactions. Although this material is biodegradable, over time intracavitary implantation filling a cavity with a scaffold in principle may cause problems in humans. Raised intracranial pressure is of major concern especially when implantation is performed into a tumor resection cavity that is predisposed towards postoperative edema formation. Therefore, we have developed a biodegradable and semisolid 3DECM preparation of almost gel-like plasticity, which allows the erratic surface of a surgical resection cavity to be covered rather than filling the complete cavity. Remarkably, the 3DECM implants still provide enough consistency to enable surgical handling and to tailor the implants to the extension of the cavities.

The 3DECM preparation tested in our study mainly consisted of human extracellular matrix purified from tissue-engineered neonatal dermal fibroblast cultures.^29,30 Equivalent extracellular matrix preparations from human sources have been the basis for biologically safe and well-tolerated biodegradable human dermal replacements, which are already clinically approved as dermal substitutes in wound healing^25,26 and as injectable dermal fillers for cosmetic procedures.^23,24 We have demonstrated that the 3DECM was non-toxic to NSPCs and supported their rapid in vitro expansion. As collagen type-I is a major component of 3DECM, our data are in line with previous in vitro findings supporting the role of collagen type-I for NSPC proliferation and maintaining their differentiation capacity.^37,38 Further in vivo studies have demonstrated that collagen implants promote the intrinsic regeneration capacity of injured spinal cord tissue,³⁹ supporting our observation that the 3DECM did not elicit any notable inflammatory response in immuncompetent C57BL/6 mice.

The 3DECM-embedded NSPCs grew as cell clusters resembling neurospheres, therefore making the transplantation of high numbers of NSPCs in a small volume possible. It seems likely that the transplantation of neurosphere-like clusters embedded in a 3DECM is advantageous because of the preserved cell–cell and cell–extracellular matrix contact possibly protecting from metabolic and mechanical stress associated with injections of cell suspensions.

Our in vivo data using 2 different orthotopic glioma models have demonstrated that the transplantation of 3DECM-embedded NSPCs allows targeted migration of NSPCs towards a distant growing tumor mass, whereas in the absence of a tumor the 3DECM transplanted NSPCs remained at the application site with no distant migration. Thus, it can be concluded that the 3DECM is (i) permissive for incoming chemotactic signals and (ii) preserving the motile status of embedded NSPCs. When implanted directly into the surgical resection cavity in our glioma resection model, GFP-expressing NSPCs displayed a widespread distribution throughout the recurring glioma with no migration into unaffected parenchyma. Although not the goal of this study, it can be envisioned that genetically modified 3DECM-embedded NSPCs could target the remaining invasive tumor cells delivering a therapeutic “payload” and could have potentially prevented tumor recurrence. Therapeutic targeting of the remaining tumor cells in the invasion zone after cytoreductive glioma surgery is a fundamental aim holding the reasonable hope of further improving the clinical outcome of this fatal disease. Optimized transplantation efficiency is necessary for clinically relevant stem cell-based glioma therapies. These efforts have to take into consideration the clinical realties that have been neglected in most preclinical studies so far. One instance that we addressed in the present study is the challenge of affixing NSPCs to the wall of the surgical resection cavity without immobilizing them.

In summary, we have demonstrated that the intracerebral and intracavitary transplantation of NSPCs within a biodegradable 3DECM for the delivery of high numbers of vital and motile tumor-targeting NSPCs is feasible and more efficient than the delivery of NSPCs by conventional intracerebral injection of cell suspensions. Equivalent 3DECM preparations have already gained regulatory approval for other clinical uses, which distinguishes this approach from other, more experimental technologies.¹⁶ We have demonstrated the 3DECM's potential to provide a unique platform for the clinical administration of cell grafts to the CNS. In order to move stem cell-based technology in brain tumor therapy and in a variety of other neurological disorders from bench to bedside optimization of transplantation efficiency in relation to complex disease processes will be essential.

Funding

This work was supported by a young investigator award of the Forschungsfoerderungsfonds Medizin of the University of Hamburg to NOS.