Talk:Paper - Further experiments on the development of peripheral nerves (1906)

From Embryology

Fuether Experiments on The Development Of Peripheral Nerves


Ross Granville Harrison.

From the Anatomical Laboratory of the Johns Hopkins University.

With Five Figures.

Two main questions have arisen in connection with the study of the development of the peripheral nerves. The one concerns the constitution of the nerve fiber, i. e., whether it is a process of a single cell or derived from a chain of cells. The other has to do with the manner in which the connection between center and periphery is established, whether there is a continuity ab initio (protoplasmic bridges) or whether the connection is secondarily brought about by outgrowth from the center towards the periphery.

Prior to the year 1904 all attempts to solve these problems were based on observations made upon successive stages of normal embryos. When one compares the careful analyses of their observations, as given by various authors, one cannot but be convinced of the futility of trying by this method to satisfy everyone that any particular view is correct. The only hope of settling these problems definitely lies, therefore, in experimentation.

The question of the constitution of the nerve fiber, whether a cell process or a cell chain, may here be considered first.

If one examines a developing nerve, one sees that there are numerous spindle shaped cells (cells of Schwann, sheath cells) throughout its course, and that these are very closely attached to the young nerve fiber ; on the other hand, it is also found that the nerve is connected with ganglion cells. The disputed point with which we here have to deal concerns primarily the respective roles played by these two kinds of cells in the genesis of the fiber. Some time ago I described a series of experiments ""

^ Read before the Association of American Anatomists at the meeting held at Ann Arbor, Mich., December 29, 1905.

' Harrison, Neue Versuche und Beobachtungen iiber die Entwiclvlung der peripheren Nerven aer Wirbeltiere. Sitzungsber. d. niederrheinischen Ges. f. Natur u. Heilkunde. Bonn, 1904.

American Journal of Anatomy. — Vol. V. 9

122 Experiments on tlie Uevelopnient of Peripheral Nerves

in which the spindle shaped sheath cells were eliminated by the removal of their source, at least their principal source, in an early embryonic stage, before nerves of any kind are visible. The experiment consisted in removing the ganglion crest. This was done by cutting off a thin strip from the dorsal side of the body of embryos {Rami esculenta) from 2.7 to 3 mm. in length (Fig. 1). Since this operation removes the source of the spinal ganglia also, the embryos develop without sensory nerves and ganglia, but the motor nerves do develop, and instead of being cellular in structure, as is the case in normal specimens (Figs. 2 and 3), they consist of naked fibers, which can be traced in a number of cases as far as the extreme ventral part of the musculature, i. e., as far as the nerves extend in the adult organism (Fig. 4).^

The first experiments were made upon Rana esculenta; they have since been confirmed upon the embryos of two American species, R. syJvatica and R. palustris. These experiments concerned only the spinal nerves.

Fig. 1. Profile view of frog embryo (Rana esculenta, 2.7 mm. long) at the stage of operation; the line (ab) indicates the incision.

Last season an attempt was made to corroborate the results in cranial nerves. For this purpose the cranial ganglia, the skin covering the side of the head^ and the dorsal part of the brain were excised from one side of the embryo before closure of the medullary folds. With one exception

^ The experiment was based upon the assumption that the ganglion crest is the source of the sheath cells. The result of the experiment proves this to be true, as far as the early stages of development are concerned. In certain lower vertebrates, particularly in Elasmobranchs, it has been shown that large numbers of cells are given off from the ventral part of the medullary cord, wandering out along the motor roots of both the cranial and spinal nerves, and giving rise at least in part to the nerve sheaths. The literature bearing upon this subject has recently been considered by Neal (Mark Anniversary .Volume, New York, 1903). In the frog such cells are not given off until the yolk is nearly gone but after this period cells do wander out singly along the motor roots, and in these features the frog embryo resembles closely the salmon (Harrison, Archiv f. mikrosk. Anat., Bd. 57, 1901). The cells do not, however, begin to come off until the motor nerves are well developed and have reached the extreme end of their course. Thus it happens that in the experiment the nerves are developed without sheath cells.

Eoss Granville Harrison


these experiments gave inconclusive results, as small ganglia were always found later, showing either that their normal rudiment had not been entirely removed, or that they had regenerated from some other source. In one experiment, however, in which the embryo was preserved four days after the operation, an examination of the serial sections revealed no ganglia except several sporadic cells on the n. facialis and n. vagus. These nerves consist of naked fibers, except that several sheath cells are present near their origin. A nerve in front of the facial, probably the oculo-motor, but perhaps the motor part of the trigeminus * is entirely without sheath cells and the naked fibers may be traced from the brain to a mass of mesoderm cells in the region of the eye. The results

Hind Leg

-Abdominal Muscfe

Segmental Werve

Fig. 2. Profile view of frog larva (Rana palustris, 12 mm. long) after complete resorption of yolk. The relation of the segmental nerves and the primary abdominal muscles are shown.

of this experiment, therefore, confirm the first series, showing that the cranial nerves may develop without the aid of the sheath cells.

But while the sheath cells are thus denionstrated not to be a necessary factor in the formation of the nerve, it may still be urged that they, as well as the ganglion cells, might normally form some of the fibers. During the past year an effort was made to solve this question by studying the behavior of the sheath cells in the absence of processes from the nerve centers. The source of the motor nuclei (ventral half of the medullary cord) was removed from embryos of the same nge as in the previous experiments, leaving the dorsal part of the cord together with the ganglion crest intact. The object was to ascertain whether the sheath

Owing to the absence of most of the important landmarks on the injured side the exact determination of this nerve is doubtful.

124 Expoi'iiiu'uts on the Dcvolo])ment of Peripheral Nerves

cells from the ganglion crest would be able in the absence of the motor ganglion cells to form the purely motor rami of the spinal nerves. Thereare difficulties in the way of making this experiment because it is first necessary to cut off the dorsal half of the cord, leaving it attached at oneend, then by a second cut to remove the ventral half entirely, and

Fig. 3. Semidiagrammatic view of the nerves of the abdominal walls of the frog larva (normal specimen). Abel. M., abdominal muscle; HL., rudiment of hind leg; Mot. N., motor branch of segmental nerve running in inscriptiO' tendinea of the primary abdominal muscle; Mot. Nuc, motor nucleus (ventral horn cells) in spinal cord; Seg. N., segmental (spinal) nerve; Sen. N., sensory branch of spinal nerve running to integument outside of muscle; 8p. C, spinal cord; 8p. G., spinal ganglion.

finally to heal the first strip, which is very thin, l)aek in place. Even if this is done successfully, a scar is loft, which lies in the path of the

Boss Granville Harrison 125

spinal nerves, and which no doubt serves as a hindrance to their development. Again it seems to be practically impossible to remove entirely the motor elements from all regions of the cord. After several days the larva?, although almost completely paralyzed, regain some power of movement, showing usually a slight tremor in some part of their axial musculature, when stimulated mechanically. Sections show that in some

Fig. 4. Semidiagrammatic view of the nerves of the abdominal walls of a frog larva from which the ganglion crest had been removed as shown in Fig. 1. Only motor nerves are present and these consist of axis cylinders without sheath cells.

segments very fine motor roots are present, and the ventral part of the remaining medullary cord contains in these regions a few large motor •cells. These motor fibers supply, as the movements indicate, merely that part of the musculature lying close to the spinal cord. With the two exceptions below noted, no motor fibers whatever were found in the

12(5 Experiint'iits on tlu^ Dcvolopinciil dI' I'tM'iphci'iil Xcrves

abdominal walls, which were used especially for study, because it is only there that one can distinguish clearly between motor and sensory ninu ( Fig. 3 ) . The results of ten ' experiments were as follows : In seven cases sensory nerves were found in the abdominal walls, but no motor (Fig. 5), althougli the sheath cells, as shown particularly in one case, were in very

Fig. 5. Semidiagrammatic view of the nerves of the abdominal walls of a frog larva from which the ventral half of the spinal cord had been removed at the stage represented in Fig. 1. Absence of the purely motor rami, which normally run in the inscriptiones tendineas.

close proximity to the point where the terminal motor rami normally arise; often, however, the sensory nerves were not so well developed as in

^ Each side of each individual specimen is counted as a case, because, as far as the factors in the experiment are concerned, the two sides of the body are mutually independent.

Koss Granville Harrison 137

normal specimens, in two cases motor as well as sensory nerves were fonnd; once in one, and once in two seo^ments, tlioiiol-, in other segments only sensory nerves were present. Here the motor nnclei had been less completely removed than in the other cases. In one case neither sensoiT nor motor branches were found. The last named case is to be explained as due to imperfect union of the parts, as is also the fact that in other cases the sensory nerves were often scantily developed. The results of these experiments show, therefore, that durin'g the period in which the specimens were kept imde.r observation, the sheath cells are unable by themselves to form nerve fibers. The negative character of the result renders it necessary, however, to secure further cases befQre this conclusion can be regarded as established beyond all question. Should it be nrged that time enongh was not given the sheath cells to form the nerves, it may be pointed ont that in normal specimens the motor fibers develop at a much earlier stage and that if the sheath cells nonnally contribute to their formation they should unquestionably act in the period allotted. The purpose of the experiment was to determine tl^- normal behavior of the cells and not any possible regulative action on thir part, which might take place later.

The above experiments deal only with the motor nerves, and it has not been found practicable to experiment systematically with the sensory nerves because in the latter the ganglion and the sheath cells have a common place of origin. In studying the normal development of the sensory nerves in the amphibian embryos, we find important evidence bearing upon the question. For instance, the nerves derived from the dorsal (giant) cells of Eohon-Beard are formed without sheath cells. These fibers consist, in fact, of naked axis cylinders, wdiich branch and form a delicate plexus of nerves under the skin of the frog larva, and are entirely devoid of cells (or nuclei) . Again in the Triton larva, even some of the nerves derived from the spinal ganglia of the tail are for a short time devoid of sheath cells; these, together with the nerves from the dorsal cells form a non-cellular plexus in the fin folds." In the frog larva the nerves derived from the spinal ganglia have sheath cells from the beginning. Comparison of these instances show that these cells are a variable element in the voung nerve fiber: it may therefore, be concluded that they play no necessary part in

« Since this fact was disputed by O. Schultze (Archiv f. mikrosk. Anat., Bd. 66, 1905, p. 68), I have again examined the specimens in question and have nothing'to correct in my former statement. It may be added, however, that I did not intend to include the n. lateralis, which is independent of the cutaneous plexus. This nerve of course has sheath cells at this stage.

128 Experiments on tlie Development of Peripheral Nerves

its formation. In addition to these noi'inal eases there are several experiments at hand, whieli show that even in the froo; embryo the spinal ganglion cells are by themselves capable of forming long peripheral nerve fibers. The cases in question are those in which relatively small fragments of ganglia had been dislocated or transplanted. In one such case four ganglion cells, transplanted to the abdominal wall, were found giving rise to a long nerve, which ran free through the peritoneal cavity of the larva. This nerve consisted solely of bundles of fibrillae without cells and could be traced for a distance of nearly two millimeters.

These results differ from those recently reported by 0. Schultze (Op. cit.) based also upon the study of the amphibian larva. It is not possible to discuss this work in detail here, but it may be pointed out that by confining his studies to relatively late stages, Schultze has missed the early and fundamental phases of development and thus is led to consider the purely secondary connections of the sheath cells with the nerve fibers as a primary genetic relation.

We may now take up the second great question, viz., the origin of the connection between ganglion cell and end organ. According to the one view a protoplasmic process grows out from the ganglion cell, makes its way through tissues and ultimately reaches its end organ, gradually differentiating into a nerve fiber. According to the second view (Hensen's ' hypothesis) protoplasmic connections remain between cells after division; those that are used, i. e., that function as conducting paths, persist and differentiate into nerves, the remainder disappearing.

According to Hensen's hypothesis the nerve paths are thus developed much earlier than they seem to be, and they are present for some time before they become visible. If we consider the embryo of a stage just before the nerves do become visible, then the two theories might be distinguished as follows : according to the one, the center (ganglion cells) is the all important factor in forming the nerve; according to the other the nerve is formed in situ in the peripheral path. This difference affords the basis for experimentation, though unfortunately the distinction is not so clearly cut as could be desired, for the first view does not deny the importance of the periphery in forming paths along which the developing nerve grows, nor does the second altogether disclaim the influence of the ganglion cell upon the differentiation of the primitive protoplasmic connections into nerve fibers.

The first set of experiments consisted in the extirpation of the center.

' Virchow's Archiv, Bd. XXXI, 1864. Die Entwickelungsmechanik der Nervenbahnen im Embryo der Saugetiere. Kiel und Leipzig, 1903.

Eoss Granville HcTrrison 129

This was clone by removing the medullary cord of the trunk shortly after its closure. The result is always the total absence of peripheral nerves, ■except the cranial. In the second set of experiments the peripheral path was altered. The simplest way to accomplish this is to remove the spinal €ord before any nerves are visible. After this the wound heals readily and during the next week at least no regeneration takes place. Above the notochord in the trunk of the embryo there is thus left a small space which becomes filled with mesenchyme. Into this the longitudinal bundle fibers arising in the brain grow, and after a few days they may be followed as far as six or eight segments from the cut end of the medullary tube. In other words, fibers which normally develop in the walls of the latter, develop here within the mesenchyme, which is a tissue as unlike that forming the normal path as it could possibly be.

The third mode of experimentation, which is not formally different from the preceding, consisted in the transplantation of parts of the central organ. In one series of experiments the spinal cord of the embryo was extirpated, and in each case a small piece of the cord was transplanted under the skin of the abdominal walls. The normal nerves of the body ■of course do not develop in such cases, but small nerve trunks do arise from the transplanted pieces and run for some distance in various directions, usually remaining in the abdominal walls. Sometimes portions of the ganglion crest were transplanted with the cord, resulting in the formation of small ganglia. In one of these instances, already referred to •above, the nerve fibers, which were sheathless, ran free throu,gh the peritoneal cavity. While the great length of this nerve is due largely no doubt to the shifting of its peripheral attachment, it is nevertheless quite im|)ossible that preformed bridges could have been present in its course.

The foregoing results can be interpreted in but one way. The nerve •center (ganglion cells) is shown to be the one necessary factor in the formation of the peripheral nerve. When the former is removed from the body of the embryo the latter fails to develop. When it is transplanted to abnormal positions in the body of the embryo it then gives rise to nerves which may follow paths, where normally no nerves rim, and likewise when the tissues surrounding the center are changed entirely, nerves proceeding from tliat center may develop as normally. The nerve fil)cr is therefore a product of the ganglion cell. The histological findings indicate that it is an outflow of the substance of the ganglion cell and not 3. more activation by contact of indifferent extra ganglionic substance.

While Lewis's ' experiments upon the olfactory and optic nerves afford

130 Experiments on tiic I )t'\i'l(i|iiii('iit ol Peripheral Nerves

inq^ortant atldilional evidence I'or tliis view, the conclusion of Brans," who was the lii'st to exjieriment upon this phase of the question of iiei've clevelo])ment, are diametrically opposed to it. Ilraus interprets his results in accordance with Hensen's hypothesis. AVhile one cannot but admire his ingenuity of experimentation and argument, his results are not, in my opinion, in any way inconsistent with the outgrowth theory. The growth of strange (facial or pelvic) nerves into a transplanted fore limb can be accounted for on the assumption, for wliich there is good evidence, that the configuration of the various organs and tissues plays an important part in determining the course taken by growing nerve fibers. The failure of the nerves of the host to grow into " aneurogenic " buds, while they do grow into " euneurogenic " transplantations, might be due to the absence of the attraction afforded in the latter by the cut ends of the nerves." The large size of the nerves in the transplanted limb as com])ared with the nerves connecting them with the center, may be due partially to the presence of the sheath cells transplanted with the bud, and partially to an abnormal number of dividing fibers. Braus does not exclude beyond doul)t the possibility of the latter. In any case the evidence for autogeneration of fibers could be regarded as crucial only if nerves having no nervous connection whatever with the center are developed in the transplanted part. This condition Braus has failed to demonstrate. While the facts necessitate our deciding against the validity of Hensen's view, as far as the question of primary continuity is concerned, it should be pointed out before closing that this view is in so far correct as in many instances the nervous connection between center and end organs is established when the two are very close together, and the long nerve paths originate in such cases by the moving apart of center and innervated organ after the establishment of the connection. The best example of this is seen in the lateral line. Here the ganglion is practically in contact with the rudiment of the sense organs when the first nerves are developed. The cell processes have merely to grow out for a distance less than the diameter of a cell in order to make connection. Yet by the wandering of the sensory epithelium from the head to the tip of the tail the lateral branch

^ W. H. Lewis, Proceedings Ass. Am. Anatomists. Am. Joiirn. of Anat., Vol. V, No. 2, 1906.

" H. Braus, Verhandl. d. Anatom. Gesell., Jena, 1904. Anatom. Anz.. Bd. XXVI, 1905.

'"Forssman (Ziegler's Beitrage, Bd. 24, 189S, and Bd. 27, 1900), lias shown beyond question that a tropism of this kind does play an important part in the regeneration of peripheral nerves.

Ross Granville Ilarrison 131

of the vagus is ultimately drawn out to this enormous length. The observations of Kerr" upon' the motor nerves of Lepidosiren are, in my opinion, capable of a similar interpretation and are a valid support of Hensen's view only in the above modified sense. In other ^^'ords, the nervous connection, though formed very early, is by no means primary.

The results of the foregoing may be summarized as follows : The axis cylinder of the nerve fiber is the outgrowth of a single ganglion cell, with ■which it remains in continuity throughout life. It grows gradually from the center towards the periphery establishing secondarily connection with its end organ. The other elements, the cells of Schwann, which are found upon the developing nerve have nothing to do with its genesis, though they may play an important part in the nutrition and protection of the fibers.

"J. Graham Kerr, Trans. Roy. Soc. Edinburgh, Vol. XLI, 1904.