| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
FJ
EXPRESS SUMMARY ARTICLE The Full-length version of this article is also available, published online September 4, 2003 as doi:10.1096/fj.02-1180fje. |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Laboratoire d'Organogénèse Expérimentale, Université Laval, CHA, Hôpital du Saint-Sacrement, Québec; and
Montreal Neurological Institute, McGill University, Montreal, Canada
2Correspondence: Laboratoire d'Organogénèse Expérimentale, Hôpital du Saint-Sacrement, 1050 chemin Sainte-Foy, Québec, QC, Canada G1S 4L8. E-mail: Francois.berthod{at}chg.ulaval.ca
SPECIFIC AIMS
Nerve regeneration in the peripheral nervous system is a critical step, especially to enhance the recovery of organ functionality after transplantation, when it is mostly or partly monitored by the nervous system, such as the sense of touch in skin. The goal of this study was to develop and characterize the first tissue-engineered 3-dimensional culture system that physiologically reproduces a peripheral nerve regeneration process in vitro by coculturing fibroblasts, endothelial cells, and dorsal root ganglion neurons, with or without keratinocytes, in a collagen sponge for up to 31 days.
PRINCIPAL FINDINGS
1. Development of the tissue-engineered models of peripheral nerve regeneration (MPNR)
Animal and cell culture models are developed to further understand processes of neurotoxicity, neuronal dysfunction, and nerve degeneration/regeneration. None of the in vitro and in vivo models proposed in the literature allows the long-term study of neurite growth in a well-defined 3-dimensional environment that truly mimics the living connective tissue in which peripheral nerves grow in vivo. The goal of this study was to develop and characterize the first 3-dimensional culture system that closely reproduces a peripheral nerve regeneration process in vitro.
We focused on preparing a connective tissue that could support neurite outgrowth. We took advantage of the dermal part of our exclusive model of endothelialized reconstructed skin as a basis to prepare an endothelialized connective tissue. Indeed, this model has the unique ability to encourage the spontaneous reconstruction by human endothelial cells of a network of capillary-like tubes that mimics the skin vasculature. This allowed us to study the close association between nerves and capillaries in vitro in a very physiological environment.
Peripheral nerves are found mostly in vivo in connective tissues made of fibroblasts, endothelial cells, and the extracellular matrix they produce. These elements should play a major role in the maintenance and regeneration of nerves by cell–cell and cell–matrix contacts and the release of growth factors and signaling molecules. Fibroblasts and endothelial cells produce NGF, BDNF, and VEGF, which are known to activate axonal growth. Moreover, the deposition in high amounts of extracellular matrix in our model of connective tissue should help to support neurite outgrowth.
To assess the effect of a tissue-engineered connective tissue on neurite outgrowth in vitro, we seeded sensory neurons extracted from mouse dorsal root ganglia at the top of human fibroblasts or fibroblasts and human endothelial cells cultured for 17 days in a collagen-chitosan sponge. This construct was then maturated at the air–liquid interface for 14 or 31 days to establish a gradient of nutrients and growth factors that might favor neurite outgrowth from the top to the bottom of the sponge.
2. Neurite growth in the tissue-engineered MPNR: effect of nerve growth factor and endothelial cells
The neurite outgrowth process in the MPNR was assessed in the presence or absence of NGF by staining neuronal cells with the 150 kDa neurofilament, an intermediate filament expressed only in neurons. In models containing endothelial cells, a staining of PECAM-1 was also performed. PECAM is a cell adhesion molecule expressed in endothelial cells when they form a vascular network. Neurons survived without addition of NGF in the cell layer on the top of the sponges in the model with fibroblasts only (Fig. 1
A) and in the model with fibroblasts and endothelial cells (Fig. 1B
) for 14 days, but no neurite extension was observed. However, peripheral neurons are known to require NGF for survival in vitro. This unexpected survival in our model should be due to the coculture with fibroblasts, which could secrete growth factors that support neuron survival in vitro, such as NGF and BDNF. When 10 ng/mL of NGF was added to the culture media, neurites were seen deep down the sponges in the model with fibroblasts (Fig. 1C
) and with fibroblasts and endothelial cells (Fig. 1D
). These neurites also expressed the calcitonin gene-related peptide, a neurotransmitter found in sensory neurons (Fig. 1F
).
|
To further characterize the nerve network in the MPNR, an observation of the 150 kDa neurofilament-positive neurites was made by confocal microscopy. This technique allowed us to scan a 30 µm-thick section of the model to observe the neurite network full length. The neurites did not only go down deep inside the sponge, but extended in all directions to innervate large areas of the reconstructed tissue (Fig. 1I
).
A close association between the capillary-like network and neurites was observed within the sponge (Fig. 1H
) that resembled the one observed in the normal human skin (Fig. 1G
). Thus, endothelial cells seem to modulate the spatial organization of axons. Moreover, endothelial cells significantly increased of 37% the maximal depth of neurite elongation through the sponge after 14 days of maturation (505 µm±38 vs. 367 µm±7.6 in MPNR without endothelial cells). This effect was not maintained after 31 days of culture. It could be hypothesized that capillaries mostly played a role of enhancer, which should only be noticeable in the beginning of the neural network establishment. Indeed, neurites growth could be facilitated by migration on the laminin-rich basement membrane of the capillary-like tubes and/or by the secretion of NGF and BDNF by endothelial cells. However, axons might also influence the organization and the level of differentiation of the capillary-like network.
3. Neurite growth in the MPNR: effect of keratinocytes
Since keratinocytes are known to produce NGF, we hypothesized that the epidermis should play a role in axonal guidance. We prepared a model of innervated reconstructed skin with and without endothelial cells. Keratinocytes were added on the opposite side on models with and without HUVEC in order to reconstruct a skin model of nerve regeneration. Our model was flipped upside-down and the keratinocytes were seeded on the other side (with respect to neurons) (Fig. 3)
. We observed that the presence of a well-differentiated epidermis did not seem to influence neurite outgrowth in terms of neurite length, orientation, and number in these culture conditions after 14 days of culture (Fig. 1E
) compared with a MPNR without keratinocytes, but cultured in exactly the same conditions (named MPNR without keratinocytes in Fig. 2
). The presence of keratinocytes avoided the significant enhancement of neurite elongation induced by endothelial cells after 14 days of culture, probably because of a greater influence of these cells on axons compared with endothelial cells. However, a twofold decrease in the number of neurites was observed between 14 and 31 days of culture in models with and without endothelial cells (14.6±0.6 neurites/mm2 vs. 5.9±0.6; Fig. 2
). This result suggested that the MPNR failed to maintain for up to 31 days the neurites that grew during the first 2 wk, since half of them disappeared. This major decrease was avoided in the presence of keratinocytes (14.1±1 neurites/mm2 after 31 days with keratinocytes; Fig. 2
). Therefore, the epidermis could favor the stabilization of the neurite network for long-term cultures. This effect could be due to paracrine interactions with keratinocytes, natural target cells for sensory nerves in the epidermis, along with Merkel cells.
|
|
Hence, we demonstrated that a vigorous neurite outgrowth was obtained in our model of peripheral nerve regeneration. This axonal growth was mostly due to the presence of fibroblasts and to the addition of 10 ng/mL of NGF. However, neurite outgrowth did not necessitate a coculture with glial cells or addition of the B27 supplement, which is usually essential for neuronal survival in culture. Moreover, a long-term culture of neurons can be performed in this model, up to 4 wk and more, compared with less than the 1 week generally observed with other models of neuronal culture. This is a major advantage to study the effect of different molecules and cells on neurite outgrowth on a long-term basis, as we have shown by demonstrating an influence by keratinocytes on the neurite network detected only after 31 days of culture. The absence of a Schwann cells coculture requirement is also an advantage when studying a specific effect of drugs on axons, independent of glial cells. However, the addition of Schwann cells can also be performed in the model, as we plan to do.
CONCLUSION
In conclusion, we developed a unique model that mimicked the peripheral nerve regeneration process in vitro. This tissue-engineered model allowed a vigorous neurite outgrowth from dorsal root ganglia neurons from the top to the bottom of the tissue at a distance up to 770 µm. Neurites were attracted by NGF added in the culture medium underneath the tissue. Neurites migrated through a dense extracellular matrix made by human fibroblasts nearly similar to the normal connective tissue matrix. This fibroblast reconstructed matrix should reproduce a 3-dimensional environment much closer to the in vivo situation than an agarose gel, a collagen gel or a Matrigel©. This model is highly flexible since various types of cells can be cocultured in it, such as Schwann cells, dendritic cells, etc., while motor or sympathetic neurons can be used instead of sensory neurons. Various types of molecules can be added in the model such as inhibitory, repulsive, or attractive molecules. These molecules can be mixed with the culture medium or incorporated in the collagen sponge, taking advantage of its ionic bonds.
FOOTNOTES
1 To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.02-1180fje; doi: 10.1096/fj.02-1180fje 
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
