Isolation of Enteric Ganglia from the Myenteric Plexus of Adult Rats

Enteric neurons and glia cells were isolated from adult Sprague Dawley rats. A procedure is described using a combination of microdissection and mechanical dissociation after enzyme treatment which yields large numbers of cell clusters suitable for tissue culture and grafting into the injured spinal cord. Differentiated enteric ganglia remained viable for at least 5 days in vitro Cultured neurons expressed histochemical reactivity for acetylcholinesterase and nicotinamide adenine dinucleotide phosphate diaphorase. Nestin positive glia, which represented a population of non-myelinating enteric Schwann cells, could also be identified in cultures maintained 5 days or longer in vitro. The myenteric plexus of adult rats can provide a readily available source of neurons and Schwann cells for grafting to the central nervous system.

survive as neural gratis with the exception of enteric ganglia and other types of mature peripheral ganglia /40/. Gratis from such sources consist of fully differentiated neurons and glia cell types which may provide important alternatives to fetal tissue in their application to modify the microenvironment of injured CNS.
Tissue and explant cultures of the myenteric plexus have been used previously to study physiology and neurochemistry of the enteric nervous system. For most of this work enteric ganglia were isolated from immature mammals, such as newborn guinea-pigs /2,6,22,23,33/, neonatal hamsters/25/or newborn rats/30,31,39/. More recently, important physiological and pharmacological characteristics of emetic neurons and gila were established using preparations from mature guinea-pigs /1,12,41/. Isolating the enteric ganglia from the gut wall musculature of mature animals requires difficult and time consuming dissections which typically yield only small quantities of intact ganglia /1,41/. The need to harvest large quantities of enteric ganglia for transplantation to injured CNS /20,26,37/ has provided an incentive to simplify available protocols. A method is described here which allows relatively rapid isolation of sufficient quantities of enteric ganglia from the myenteric plexus of adult rats.

METHODS
Twenty adult rats (150-250 g) of the Sprague Dawley strain were used in separate experiments for the isolation of enteric ganglion cell clusters.
Isolation of gut neurons from adult animals Rats were deeply anesthetized with nembutal (50 mg/kg). The abdomen was shaved, disinfected, and draped. The intraperitoneal cavity was opened by a midline incision, the small intestine was excised and the rat was killed. The jejunum was divided into 2-3 cm long segments, rinsed (in and out) with sterile Hank's balanced salt solution (HBSS) and placed in ice-cold HBSS. For microdissection, each segment of gut was placed separately in a 60 mm Petri dish and thoroughly rinsed with HBSS. The longitudinal muscle layer was carefully lifted and separated from underlying circular muscle and mucosa along one end of the segment by teasing the layers apart with fine tweezers avoiding excessive stretching. In many of the segments the outer muscle layer was peeled off like a sleeve. This microdissected tissue contained longitudinal smooth muscle sheets with attached myenteric ganglia (Fig. 1). Tissue was pooled from the segments, placed in 0.1% collagenase in cell dissociation fluid (Sigma) or HBSS and incubated at 37C for periods from 30 minutes to four hours. Following incubation, the enzyme solution was aspirated and replaced with Dulbecco's modified Eagle's medium containing 5% horse serum, 5% fetal bovine serum and 50 units penicillin-streptomycin per milliliter. Tissue Four rats were used to estimate the recovery and survival of ganglion cell clusters over a period of five days. For this a quantitative evaluation of cell clusters was carried out as follows. Tissue was dissected from the jejunum (> 60 cm length) as described above. All dissociated ganglia were collected in 2.6 or 3.6 ml tissue culture medium.
Identical volumes (40 tl)of dissociated cell clusters were plated into 13 or 18 compartments of five multiwell (with 24 wells each) plates coated with collagen. Cell clusters were allowed to settle and rinsed once with fresh medium. After 24 hours of incubation at 37C, clusters were rinsed again with tissue culture medium to remove floating cells and debris. Subsequently, the culture medium was replaced with fresh medium once per day. One of the plates was fixed (see below) each day for five consecutive days. Subsequently, all the cell clusters were counted in each well and added to obtain an JOUIONAL  completely detached from "round cell" containing clusters unlike the preparation shown in Figure 3 below. Scale bar-180 tm.
Spindle shapes characterized single smooth muscle cells (Figs. 2,3). Enteric neurons had rounded shapes and they were larger in diameter (Fig. 3). On suitable substrates (such as adjacent non-neuronal cells) the rounded cell clusters grew processes after several days in culture ( dissociation yielded a population of cell clusters that varied in size containing between 5-60 cells per cluster. Optimal mechanical dissociation of enteric ganglia from microdissected muscularis externa was achieved following enzymatic digestion for 30 to 120 minutes (Fig. 2). Longer times of enzymatic treatment (4 hours) resulted in more complete cell dissociation but limited survival of enteric neurons.  enteric neurons in ganglion cell clusters. Following AChE staining, a dark brov reaction product labeled larger cell clusters and a few single cells (Fig. 3). A proportion of small cell clusters with rounded somata and some single cells failed to express AChE positive stain. After three to five days in vitro the majority of cell clusters had become flattened in shape but they continued to distribute in cell groups or "colonies". These colonies had variable sizes and they contained AChE positive neurons (Fig. 5) and AChE negative non-neurons or glia cells. A proportion of the myenteric neurons in enteric cell clusters stained with NADPH-diaphorase histochemistry (Fig. 4) Figure 2, were negative for ED 1. Flattened polygonal cells never stained with ED h any of the cultures. ED positive cells maintained their round shapes. A few of these were observed near the surface of rarely noted large colonies, which appeared to be derived from enteric ganglia that had been poorly or not at all dissociated.
Cluster cultures maintained in vitro for five days or longer contained cell groups that stained positively with Rat 401 monoclonal antiserum (Figs.

6,7). Immunocytochemical reactivity to Rat 401
was not detected in cell clusters acutely isolated from enteric ganglia and in paraformaldehyde fixed whole mounts of myenteric plexus attached to longitudinal oriented smooth muscle of the muscularis externa.

DISCUSSION
These studies demonstrate the feasibility of isolating cell clusters frown the myenteric plexus of adult rats. Partially dissociated ganglia were obtained from the intestinal wall after microdissection and subsequent enzymatic treatment of the tissue /1,12,20,30,31,41/. During microdissection the nyenteric plexus remains attached to the outer longitudinal layer of the bowel musculature frown which it can be extracted by treatment with various enzymes. The present study employed a simple separation of enteric ganglion cell clusters by differential adhesion to collagen coated tissue culture plastic. During this step most of the dissociated smooth muscle cells were removed. The cell clusters that adhered to the dish within one day of culture were separated from floating cells by several  Histogram of cell cluster counts from three experiments, series to 3, respectively. Day represents counts taken 24 hours after initial plating of the isolated cell clusters. medium, cells with originally romded shapes assumed flattened polygonal forms. Enteric neurons reactive for AChE and NADPH-diaphorase extended their processes on enteric glia cells which served as substrates.
Previously, several methods were developed in other laboratories for isolating the immature myenteric plexus /7,22,23,25,30/. The myenteric plexus is free of connective tissue, extracellular collagen and blood vessels/15,17/. Thus, treatment of intact segments of the gut wall with highly purified collagenase allows the isolation of interconnected myenteric ganglia following gentle separation of the muscularis extema of the bowel /22,23,33,35/. The layers of smooth muscle are not dissociated by this method and ganglia of the plexus remain connected causing minimal disruption of myenteric neurons and little, if any, contamination by adherent smooth muscle cells. Other procedures have also relied on enzymatic digestion of connective tissue and muscle following various mechanical operations. For example, mincing the gut wall prior to collagenase treatment was used to isolate myenteric neurons from newborn hamsters /25/. An efficient microdissection method was introduced by Nishi and Willard/30,31/by which the outer longitudinal muscle layer of the muscularis externa was separated from the intestine of newborn rats. Stretching the gut segment over a piece of glass tubing aided in the dissection. Subsequently, the tissue was treated with Dispase and mechanically dissociated. This method yields partially dissociated ganglia which contain clusters of enteric neurons and enteric glia cells. In newborn rats the gut wall. is thin and transparent and contains relatively few smooth muscle cells which rarely adhere to isolated neuron clusters /30,31/. However, application of this method to mature animals with increased muscle mass in their gut wall poses a greater challenge regarding the elimination of smooth muscle cells. These cells will adhere to the isolated ganglia and grow unless they are damaged by the enzyme treatment or removed from the initial cell isolate. For example, myenteric ganglia from adult rats dissociated by Nishi and Willard's procedure contained large quantities of smooth muscle and connective tissue after transplantation to the spinal cord/20/.
The method described in the present study combined microdissection and enzymatic treatment and it included several modifications of previous protocols. Most importantly, an "adhesion" separation of ganglion cell clusters from dissociated muscle was tested. The initial steps of the isolation procedure were similar to Nishi and Willard's /30/ method. However, separation of the two muscle layers ofthe muscularis extema in adult rats seemed to be facilitated in low calcium buffer and partly contracted gut segments kept at 4C. Furthermore, tissue dissociation with purified collagenase and optimal treatment times yielded preparations in which smooth muscle was dissociated into single cells whereas cells ofthe myenteric plexus remained in clusters. Initially, smooth muscle cells failed to attach to the culture dish. This allowed removal of dissociated smooth muscle cells, by several rinses with tissue culture medium, from ganglion cell clusters, which adhered to the substrate. Significant reduction of smooth muscle cells and enrichment of ganglion cell clusters was achieved. Nevertheless, it remains to be shown whether a small proportion of smooth muscle cells aggregated with the ganglion cell clusters. However, it was observed that contamination of ganglion cell clusters by macrophages was unlikely. Macrophages normally occur in adult intestine/28/. Partial dissociation of enteric ganglia during their isolation appeared to reduce ED 1 reactive macrophages.
Reduced survival of enteric neurons after exposure to anoxia, which occurred during extended storage of the intestine, was an expected finding. Surprisingly, non-neuronal cells that expressed immunoreactivity to Rat 401 antiserum after in vitro culture, were isolated from such ganglion cell clusters. Rat 401 antiserum has been shown to label products of the gene named "nestin" /27/. Characteristically neuronal precursor cells were labeled with anti Rat 401. Schwann cells, which form myelin sheaths ofperipheral nerves, also react immunocytochemically with Rat 401 /13/. This is of interest because enteric glia do not form myelin in the myenteric plexus /15,17/. However, non-myelinating Schwann cells of the enteric nervous system have the capacity to form myelin after transplantation to chick/9/or rat spinal cord /18,20/. Expression ofRat 401 in Schwann cells and cultured enteric glia suggests that these cells share a common ancestor. Absence of staining with Rat 401 in the myenteric plexus in situ could indicate low levels of the antigen and conditions which suppress transcription of the nestin gene. Enteric ganglia provide a potential "cell reserve" in the body that could be utilized as a source of fully differentiated neurons and enteric glia cells, with defined properties /14,31,39/, to repair injuries in the central nervous system/20,26,37/. For example, the use of mature neurons has the advantage of supplying already differentiated phenotypes of known functional potential/3,14/. Enteric glia cells may supply myelin sheaths for demyelinated CNS fiber tracts. Although these glia cells do not form myelin in the myenteric plexus /15,17/, they may do so after transplantation to an altered microenvironment Separation of nonadherent cells provided an important purification step in the isolation of enteric ganglion cells from dissociated smooth muscle cells. Enzymatic digestion and mechanical dissociation resulted in cell suspensions in which the single muscle cells outnumbered the cell clusters. After plating on an adhesive substrate the nonadherent smooth muscle cells and cell debris were aspirated.
However, poor adhesion to the substrate may also have caused some loss of potentially 'iable" ganglion cell clusters. An average of 55% of the clusters plated remained attached to the culture dish for at least 5 days in vitro. Cell cluster survival will be influenced by culture conditions and by the enzymatic digestion which may potentially injure some cells. Damaged cells would tend to be less sticky, since their ability to replace extracellular matrix components could be impaired. Floating and poorly adherent cell clusters were removed during rinsing and medium replacement procedures. In addition, requirements of adult enteric neurons in culture may differ form those of more immature cells/2,7,32,34/.
Fiorica-Howells and coworkers /12/ used another method of separating intact myenteric ganglia from the intestine of adult guinea-pigs. The tissue suspension was filtered through 'ucleopore" filters of just 8 Bm pore size. All isolated smooth muscle cells were removed h the filtrate and intact ganglia remained on the filter surface. During the course of the present study a similar method was tried unsuccessfully and subsequently abandoned in favor of the adhesion procedure. A nitex membrane of 70 Bm mesh size was used as a sieve. This membrane retained isolated cell clusters of partial ganglia, which differentiated into enteric neuron and gila cell containing colonies. Single muscle cells were observed in the filtrate. However, cells retained by the nitex membrane, as tested by direct microscopical observation, included a significant number of smooth muscle cells. In their relaxed state smooth muscle cells isolated from the rat jejunum had diameters of less than 10 gm but measured 180-250 Bm in length. These different observations may relate to differences in filter properties and variable characteristics of smooth muscle cells in guinea-pig and rat.
The myenteric plexus is subjected to variable stretching and relaxation periods during normal peristaltic movements of the gut musculature. This activity may cause a certain amount of ongoing 'natural injury" on enteric ganglia. Thus, the essential ability to reconstitute cell processes and to migrate under suitable circumstances may persist in adult enteric neurons. Moreover, the tissue environment of enteric ganglia, which become embedded in smooth muscle, may promote release of growth factors and discourage the synthesis of growth inhibiting molecules in enteric gila cells. Examples of plastic changes in enteric ganglia have been described following lesions of the myenteric plexus /11/and reanastemosis of the intestine/38/. These studies showed that individual enteric neurons from intact ganglia migrate to regions of denervated smooth muscle and reestablish functional connections.
Several factors may be generated and stored by enteric ganglia which could regulate expression of plasticity and proliferation. A recent investigation has shown that neurotrophin-3 induces the differentiation of neural crest-derived cells in vitro while NGF, BDNF, and neurotrophin-4/5 do not /7/. Other studies have suggested that purines and fibroblast growth factor (FGF) may contribute signals for regulation of cell survival and differentiation of enteric neurons/36/. FGF is known to affect survival and proliferation of CNS neuron precursors /29/ in addition to stimulating process elongation of neurons and PC12 cells. The source of FGF in the emetic nervous system has not been established but interstitial cells and enteric gila may be involved in its synthesis. Basement lamina proteins that ensheathe the plexus in its entirety may also aid in the storage ofFGF. The newly developed procedures for isolating enteric gila and enteric neuron enriched cultures from ganglia of immature intestine/2,7,33,35/and the methods described here will complement each other to allow further studies of the mechanisms that regulate plasticity in the