Most midbody ganglia in the central nervous system (CNS) of the leech Hirudo medicinalis contain about 400 neurons. However, those in the fifth and sixth midbody segments (ganglia M5 and M6) are specialized for reproductive functions, and each contain several hundred additional small neurons. These neurons arise late in embryogenesis as a result of an innervation-dependent inductive interaction between the male genitalia and M5 and M6 and are therefore known as peripherally induced central (PIC) neurons. The results of a series of ablation and transplantation experiments show that the PIC neurons are induced during a 1 to 2 day period about midway in embryogenesis (E15). The male genitalia are not necessary for induction before or after this period, and their presence for only one day may be sufficient for the induction to take place. Heterochronic transplantation of male genitalia shows that the critical period of interaction is independent of the age of the inducing tissues. Since the inductive signal is available from E10 to postembryonic stages, both the beginning and the end of the inductive period are determined by the CNS, not the periphery.

Interactions between regions of the central nervous system (CNS) and the periphery often serve to match the size of a neuron pool to the requirements of a corresponding target by affecting neuronal survival. Such mechanisms are found in both vertebrates and invertebrates (reviewed by Williams and Herrup, 1988; Becker and Macagno, 1992). In some instances, the interaction is mediated by efferent fibers, in others by afferent projections. An example of an interaction mediated by efferents is that observed between vertebrate motor neurons and their target skeletal muscles: these neurons are vastly overproduced and only those that innervate the target muscles are maintained, through mechanisms that are still controversial (see Oppenheim, 1991, for a recent review). An interaction of the second kind is that which occurs between the compound eye (derived from epithelial cells) and the central regions that it innervates in certain arthropods: all laminar neurons, as well as some higher order interneurons, die when entirely deprived of afferent innervation by photoreceptor axons (e.g., Daphnia: LoPresti et al., 1973; Macagno, 1979; Drosophila: Power, 1943; Meyerowitz and Kankel, 1978; Fischbach and Technau, 1984; Steller et al., 1987).

In contrast to the many well-known cases of interactions influencing neuronal survival, there are few clear examples where CNS-periphery interactions regulate neuron number by affecting neurogenesis (see Williams and Herrup, 1988). One recent example is found in the arthropod visual system. In the fruit fly, Selleck and Steller (1991) have demonstrated that ingrowing retinal axons trigger proliferation of neuronal precursors in the optic ganglia. A similar phenomenon occurs in the crustacean Daphnia magna (E. Macagno, unpublished observations). A third example is the innervation-dependent neurogenesis observed in the leech CNS (Baptista and Macagno, 1988a; Baptista et al., 1990), aspects of which are the subject of the present study.

In hirudinid leeches, such as Hirudo medicinalis or Haemopis marmorata, most of the 21 midbody ganglia (M) contain about 400 neurons (Macagno, 1980). Midbody ganglia M5 and M6, which are associated with the male and female sex organs and are therefore also termed the sex ganglia, contain an additional complement of 300-400 comparatively small neurons (Macagno, 1980; Stewart et al., 1986). These additional neurons arise during the last third of embryogenesis (embryonic day 20-26; E20-26), long after all other neurons (Baptista et al., 1990). Their appearance depends on innervation of the male genitalia; ablation of these organs or disruption of the nerves connecting them to the CNS (the sex nerves) around E10 but before E16 prevents the production of these neurons altogether (Baptista and Macagno, 1988a). Since the inducing signal originates in a peripheral organ innervated by M5 and M6, these additional neurons are known as Peripherally Induced Central (PIC) neurons (Baptista et al., 1990). The nature of the signal and how it exerts a mitogenic effect are not known, but it is clear that its range of action is very short and that it is conveyed by some element of the sex nerves (for review on this topic, see Becker and Macagno, 1992).

In this study we asked, first, whether the time window of interaction is narrower than the period E10 to E16 defined in earlier experiments and, second, whether this time window is determined by the source or by the target of the inductive signal. To answer these questions, we performed ablations as well as homochronic and heterochronic transplantations of male genitalia in E10-E20 embryos and assayed a few days later for the production of PIC neurons. The results presented here suggest that (i) the PIC neurons are normally induced at E14-E16, (ii) the presence of the male organ for 1-2 days at E15-E16 is sufficient for the induction to take place, (iii) the inductive signal is available from at least E10 through the end of embryogenesis (E30) and up to postembryonic day 10 (P10), and (iv) the inductive period is determined by the ability of M5 and M6 to respond to the inducing signal between E12 and E16.

Animals and culture conditions

Leech embryos were obtained from a breeding colony of Hirudo medicinalis maintained at 23°C in our laboratory. Embryos were removed from their cocoons, staged by days of development after egg laying (Fernandez and Stent, 1982) and kept in sterile artificial spring water (0.5 g/l Instant Ocean, Menasha Corporation) at 23°C. Embryogenesis lasts about 30 days at this temperature.

Surgical procedures

The primordia of the male and female genitalia are easily recognized under the dissecting microscope in E10 or older embryos. The male genitalia consist of the penis, prostate gland and paired epididymides connected to the testes via the vas deferens (see, for instance, Zipser, 1979).

In all operations, animals were anesthetized in 8% ethanol in sterile artificial spring water until all movements stopped. They were then placed, ventral side up, in a groove cut into a Sylgard-covered dish and pressed down with a small strip of coverslip held in place with fine tungsten pins. Male genitalia were removed using very sharp iridectomy scissors and forceps. For ablation, a small cut was made around the gonopore and the male organ was then pulled gently out of the embryo and the attached connective tissue, sex nerves, and tubes of the epididymides were cut off. This procedure left only a very small wound which healed within a few hours. For transplantation, the donor male organ was inserted, prostate gland first, into an opening made in the skin near the normal location of the male genitalia (close to the posterior margin of M5) and pushed in until the gonopore was flush with the host’s skin. A small glass plate was then placed on the graft in order to hold it in place. The ethanol solution was then immediately diluted to about 5% to allow the animal to recover gradually from anesthesia. Contracting muscles around the wound helped to hold the transplant in place after about 20-30 minutes, and the embryo could be transferred back to sterile artificial spring water until needed.

By E26 the graft had usually developed into what appeared to be normal male organs, often visibly innervated by either M5 or M6, or by both ganglia. A simple test for functional innervation at the level of motor neurons in M6 was to pinch either of the sex nerves with forceps. This usually caused the motor neurons or their axons to fire action potentials, thereby evoking visible movements of the muscular penile sheath (Zipser, 1979).

Administration and detection of BrdU

PIC neuron production was monitored by immunocytochemical detection of BrdU incorporation (Gratzner, 1982). E24 embryos were anesthetized as above and pinned ventral side up on a Syl-gard covered slide with a single tungsten pin through the upper part of the anterior sucker. About 2 µl of a solution of 20 mg/ml BrdU (Sigma) in leech Ringer’s (Muller et al., 1981) containing 0.05% Fast Green were injected through the mouth into the digestive tract of the embryo using a 50 µl Hamilton syringe with a 30-gauge hypodermic needle. Animals were then transferred back to artificial spring water and allowed to develop for 2 days before being processed for BrdU detection.

For immunocytochemistry, animals were anesthetized, opened along the dorsal midline, pinned in a Sylgard dish and briefly washed with Ringer’s containing 0.2% Triton X-100 to remove the yolk. After rinsing in Ringer’s without Triton X, the blood sinus was removed from around M3 through M7, and the skin was cut along the ventral midline to expose the CNS. The live tissue was then treated with a collagenase solution (2000 U/ml, Sigma, type IV) for about 20 minutes, to facilitate antibody penetration. Preparations were fixed in 70% ethanol for 1 hour, followed by rinsing in PBT (0.01 M PBS with 1% Triton X-100). The DNA was denatured by treating the tissue with 2 M HCl for 40 minutes. The acid was neutralized through a brief rinse in 0.1 M Na2B4O7, followed by a 1 hour wash in PBT. Tissues were then incubated overnight with a mouse monoclonal anti-BrdU antibody (Becton-Dickinson) diluted 1:20 with 2% normal sheep serum. The following day the preparations were first washed for 1 hour and then incubated for 2 hours with sheep anti-mouse serum (Cappel), diluted 1:20. After another 1 hour wash, the tissue was incubated with mouse peroxidase anti-peroxidase (PAP, Sternberger-Meyer) for 2 hours, followed by a final wash. Peroxidase was reacted with diaminobenzidine (0.4 mg/ml PBT), and the preparations were dehydrated in ethanol, cleared in xylene and mounted whole in Permount (Fisher). Ganglia were viewed and photographed on a Zeiss Photomicroscope.

Cell counts and criteria for induction

Induction of PIC cells was assessed by comparing the number of labeled cells in M5 and M6 with those in the neighboring ganglia, M4 and M7, in each experimental animal (Baptista et al., 1990). A few BrdU-labeled cells seen in every ganglion in late embryogenesis appear to correspond to proliferating microglia and epithelial cells of the ganglionic sheath (Baptista et al., 1990). Cell counts in M4 and M7 therefore gave an estimate of the background level of cell division occurring in midbody ganglia at these stages. Labeled cells ranged from 0 to 37 in M4 (5.12 ± 6.27) and from 0 to 33 in M7 (4.86 ± 6.17) in the 353 animals analyzed in these experiments, evidence of considerable variation in the number of labeled cells from animal to animal. This was probably due to variations in the size of embryos and the amount of BrdU injected, as well as the length of time the compound was available for incorporation. As the criterion for PIC cell induction, we required a number of labeled cells in M5 or M6 at least two standard deviations above the mean in M4 and M7.

We performed the following operations on E10-E17 embryos (see schematic in Fig. 1): (1) ablation of the male genitalia, including the penis, the prostate gland and parts of the epididymides; (2) transplantation of male genitalia from a donor of the same age into E15-E17 hosts that had their male organs ablated at E10-E11 (in some cases, these implants were removed one day later) and (3) two sets of experiments in which male organs were ablated at E10-E11 and either (a) replaced immediately with a graft of an older age or, (b) at E16-E17, replaced with an E10 male organ.

Fig. 1.

Four different operations were performed on leech embryos at embryonic days (E) 10-17: (1) ablation of the male genitalia at E10 through E17, (2) ablation at E10, followed by replacement at E15-E17 with a male organ from a donor of the same age (homochronic mode), (3) heterochronic transplantation of older male organs into E10 hosts that had their male genitalia ablated during the same operation (a) and (b) ablation at E10 followed by replacement at E16-E17 with an E10 organ (heterochronic mode). All animals were injected with BrdU at E24 and processed for immunocytochemistry two days later.

Fig. 1.

Four different operations were performed on leech embryos at embryonic days (E) 10-17: (1) ablation of the male genitalia at E10 through E17, (2) ablation at E10, followed by replacement at E15-E17 with a male organ from a donor of the same age (homochronic mode), (3) heterochronic transplantation of older male organs into E10 hosts that had their male genitalia ablated during the same operation (a) and (b) ablation at E10 followed by replacement at E16-E17 with an E10 organ (heterochronic mode). All animals were injected with BrdU at E24 and processed for immunocytochemistry two days later.

Normal time course of interactions that trigger the birth of the PIC neurons

Baptista and Macagno (1988a) found that ablating the primordium of the male genitalia at E10 prevented the production of the PIC neurons. In contrast, animals that had their male organs ablated at E16 or later produced the full complement of PIC cells, as revealed by cell counts and immunohistological staining of juvenile ganglia with neuron-specific antibodies. To define more precisely the course of induction between E10 and E16, we performed a series of ablations at successively later stages (schematic 1 in Fig. 1) and assessed subsequent levels of cell division in the sex segments. Operated animals were injected with BrdU at E24, well within the period (E20 to E30) when the PIC neurons are born (Baptista et al., 1990), immunostained two days later and examined for the presence of BrdU-positive nuclei in M4 through M7.

Animals whose male genitalia were ablated before E13 were never found to have PIC neurons, but the percentage of animals containing PIC neurons sharply increased with ablation at later times (Fig. 2A). Thus, following ablation at E13, only 20% (6 out of 31) of the animals examined had large numbers of BrdU-labeled cells in either M5 (1 case) or M6 (4 cases), or in both (1 case). For ablations at E14, this percentage increased to about 65% (17 cases, 5 in M5, 8 in M6, 4 in both, n=26), and for those at E15 induction was observed in 94% of the animals (15 of 16 cases, 1 in M5, 3 in M6, 11 in both). Ablations at E16 and E17 had little effect on the appearance of the PIC neurons, as 33 of 35 (94%, 3 in M5, 2 in M6, 28 in both), and 25 of 27 animals (92%, 2 in M6, 23 in both), respectively, showed high levels of PIC cell induction. It is interesting that induction was not always observed in both of the sex ganglia of an animal, particularly in the cases of the ear lier ablations. Also, in some cases of early ablations, PIC neurons are observed in one hemiganglion only. The maximum number of stained PIC neuron nuclei after ablation at E13 was 200, which constitutes a sizeable fraction of the 300 or so additional cells per sex ganglion. This suggests that induction may indeed be complete by this time.

Fig. 2.

Results from all four types of experiments depicted in the schematic in Fig. 1. Bars give relative numbers of animals with high numbers (see Materials and methods) of stained nuclei in either one or both sex ganglia and so were considered to represent cases of induction.

Fig. 2.

Results from all four types of experiments depicted in the schematic in Fig. 1. Bars give relative numbers of animals with high numbers (see Materials and methods) of stained nuclei in either one or both sex ganglia and so were considered to represent cases of induction.

These results demonstrate that, in a population of embryos, induction of the PIC neurons is not found earlier than E13 and essentially all animals show some degree of induction by E15. From E16 onward, the presence of the male organ is no longer necessary for these neurons to appear, confirming previous observations by Baptista and Macagno (1988a). Two interesting questions are raised by these findings: first, are the inducing tissues necessary only between E13 and E15 and second, if this is the case, what is the minimum time of interaction sufficient to induce PIC neurons? Because we measure induction with a single pulse of BrdU, we cannot assess whether all PIC neurons are induced in these ablation experiments. It is conceivable that induction of all ∼300 PIC neurons requires the whole period from E13 to E16.

Time and length of the critical period of interaction between the sex ganglia and the male genitalia

We defined the critical period of interaction by ablating the male genitalia early (E10-E11) and replacing them at a later time with a homochronic transplant (i.e., from a donor of the same age, see schematic 2 in Fig. 1). It had been shown earlier (Baptista and Macagno, 1988b) that male genitalia transplanted near M5 or M6 at E10-E11 can become innervated by those ganglia and induce in them a normal complement of PIC neurons (Baptista and Macagno, 1988a).

In this experiment, animals from the same cocoon were divided into two groups. One of the groups had their male organs ablated at E10-11 and restored later in embryogenesis with a transplant for which the other group served as donors. Since preliminary experiments had shown that restoring the male genitalia between E11 and E14 always triggered the birth of PIC neurons (data not shown), we carried out a series of replacements starting at E15. The data from this experiment are shown in Fig. 2B. At E15, PIC neurons were induced in 15 out of 18 experimental animals (83%). Although one might expect male genitalia restored at E16 to be ineffective in inducing PIC neurons because of the results from the ablations described above, 23 out of the 38 animals (63%, for an example, see Fig. 3A) so tested were found to contain PIC cells. Restoration at E17, how-ever, yielded only one out of 14 embryos (7%) with increased levels of cell division, and none of the 17 embryos with E18 restorations showed evidence of PIC neuron induction, although the male organs were clearly innervated by motor neurons from the sex ganglia (not shown). We counted maxima of 225 and 223 BrdU-stained nuclei per sex ganglion in animals that had undergone organ replacement at E15 and E16, respectively. The single case of increased cell division after restoration at E17 yielded 39 labeled nuclei.

Fig. 3.

Examples of E26 experimental animals stained for the birth of PIC neurons in midbody ganglia M5 and M6 at E24. (A) Large numbers of stained nuclei in both M5 and M6 (arrows) of an embryo that had its male genitalia ablated at E10 and then replaced at E16 with a graft of the same age. (B) Stained nuclei in M6 of an animal from the same type of experiment where the replacement at E16 was done with an E10 organ; note the smaller size of the male organ. (C) Transplantation of a 60 day (P30) male organ into an E10 recipient; no stained nuclei can be detected in either sex ganglion. (D) Transplantation of P10 male genitalia into E10; in this case, there are large numbers of BrdU-positive cells in M6 (arrow). Anterior is up; mo, male organ. Bar corresponds to 250 µm.

Fig. 3.

Examples of E26 experimental animals stained for the birth of PIC neurons in midbody ganglia M5 and M6 at E24. (A) Large numbers of stained nuclei in both M5 and M6 (arrows) of an embryo that had its male genitalia ablated at E10 and then replaced at E16 with a graft of the same age. (B) Stained nuclei in M6 of an animal from the same type of experiment where the replacement at E16 was done with an E10 organ; note the smaller size of the male organ. (C) Transplantation of a 60 day (P30) male organ into an E10 recipient; no stained nuclei can be detected in either sex ganglion. (D) Transplantation of P10 male genitalia into E10; in this case, there are large numbers of BrdU-positive cells in M6 (arrow). Anterior is up; mo, male organ. Bar corresponds to 250 µm.

These results demonstrate that the presence of the male genitalia is not necessarily required in the period from E10 to E15, since it is possible to induce PIC neurons in M5 and M6 by restoring the tissues at E15 or E16. How do these results fit with those of the ablation experiments which showed that by E16 the male genitalia are not required for the appearance of the PIC neurons in 94% of the embryos? A simple interpretation is that the sex ganglia are capable of responding to the inductive signal from E12 to E16, but are normally induced by E15. Since the results with organ restoration at E16 suggested to us that interaction might take place in as little as one day, we performed an additional experiment, in which male organs were removed at E11, restored at E16, and again ablated at E17. When eight of these animals were tested for BrdU incorporation at E24, we found one case with 22 labeled nuclei in M5, and one case with 64 labeled nuclei in M6. Although the success rate in this experiment was low, it shows that at least some induction can take place in one day, a remarkably short time considering that both rein-nervation and conveyance of the inductive signal must take place.

The sex ganglia do not respond to the periphery after E17

The results of the two types of experiments described above imply that the inductive period ends at E16-E17, a finding that prompted us to ask what is the source of this apparent shutdown. The inductive period could be terminated by the male genitalia ceasing to produce the appropriate signal. Alternatively, the sex ganglia might lose their ability to respond to a signal that is still present. To distinguish between these possibilities, we performed a series of heterochronic transplantations, in which male genitalia were ablated at E10-E11 and immediately replaced with male organs from donors older than the inductive period determined in the above experiments (3a in Fig. 1). The ages of the donors were E20, E30, 40 days old (10 postembryonic days; P10), and 60 days (P30) and older. The results presented in Fig. 2C demonstrate that E20 male genitalia were capable of inducing PIC neurons in 76% (13 cases, n=17) of all animals examined. E30 and P10 male genitalia were still capable of eliciting additional cell births in the sex ganglia, albeit at a much smaller frequency (27%, 3 cases out of 11, see also Fig. 3D, and 16%, 3 cases out of 19, respectively). At or after P30, an example of which is shown in Fig. 3C, however, none of the transplanted male genitalia induced PIC neurons in 14 animals, although a nerve connection between the graft and the sex ganglia could be demonstrated. We take the results of these series of heterochronic transplantations to mean that the male organ continues to produce the signal, and is capable of inducing PIC neurons, in at least some cases, up to an age of 40 days, a time when, in normal development, the generation of these cells has already been completed. Maximum numbers of BrdU-stained cells in M5 or M6 in this experiment were 145 after transplantation of E20 male genitalia, and 205 and 147 after transplantation of E30 and P10 male organs, respectively. Evidently, the end of the inductive period is determined by the sex ganglia rather than by the periphery, which leaves the question of whether the CNS also determines the beginning of this period.

Initiation of the inductive period is also a property of the sex ganglia

In particular, we asked whether the putative signal is present in the male genitalia before the interaction takes place. If this were not the case, it would be possible that induction simply starts as soon as the factor(s) that trigger appearance of the PIC neurons are available. In this scenario, one would expect a preinduction male organ not to be capable of inducing cell birth in the sex ganglia. If, however, a male organ younger than E12 were capable of inducing PIC neurons, this would mean that the period of interaction is also initiated by the sex ganglia.

To distinguish between these possibilities, we undertook a set of heterochronic transplantations in which we ablated male organs at E10 and replaced them at E16 and E17 with E10 organs (Fig. 1.3b). It was clear from the isochronic replacement experiments that the sex ganglia can respond to the inductive signal up to E16 and show little response on E17. If there were no signal present in male genitalia at E10-E11, one would expect no cases of PIC neuron induction. This is clearly not the case, however, as can be seen in the results that are summarized in Fig. 2D. E10 male organs triggered cell divisions in the sex ganglia in 64% of all experimental animals at E16 (9 cases, n=14, see also Fig. 3B), but only in about 6% (1 case, n=17) when transplanted into an E17 host. In addition, transplantations of 11 to 12-day-old male organs into hosts older than E17 never resulted in the induction of PIC neurons (columns 3 and 4, Fig. 2D). These results are very similar to those obtained from isochronic replacement transplantations (see Fig. 2B). A male organ, when transplanted into an E16 embryo that had its male genitalia ablated at E10, gives rise to PIC neu-rons in about 2/3 of all cases (63% versus 64%), no matter whether it is 16 or 10 days old. Only one day later, the probability of PIC neuron induction goes down to 1/14 (7% versus 6%), regardless whether the organ was 17 or 10 days old. At E18, no induction can be demonstrated in either experiment.

In Hirudo, the birth of a unique and segment-specific neu-ronal population in midbody segments M5 and M6, the PIC neurons, is induced by the male genitalia. In the absence of the male organ or of the nerves that connect it to these ganglia, the PIC neurons are missing (Baptista and Macagno, 1988a; Baptista et al., 1990). We are interested in how signals originating in the genitalia are conveyed to the CNS. In this work, we examined, first, the exact temporal characteristics of the PIC induction process and, second, how these characteristics are determined. The findings from these experiments have certain implications on how the mitogenic signal might be conveyed from the male genitalia to the CNS, which we discuss below.

Based on the available data, we propose the following model for PIC neuron induction. Soon after the sex nerves have been established between the male genitalia and M5 and M6 (E11-E12), a putative mitogenic signal produced only by the male organ is transported via these nerves. This signal causes precursors residing in the sex ganglia to divide and, eventually, to give rise to the PIC neurons. Two ways for conveying the inductive signal from the male genitalia to the sex ganglia can be envisioned: induction could be evoked by afferent fibers from the male genitalia arriving in the CNS (anterograde mode), or efferent projections from the sex ganglia could transport the signal back to the CNS (retrograde mode). In the former case, the afferents carrying the signal could directly influence the precursors of the PIC neurons in the sex ganglia, whereas in the latter scenario one has to postulate in addition that the signal is picked up by the terminals of central cells and is then transferred to the precursors within the CNS. Around late E16, either the PIC precursors lose their ability to respond to the signal, or the elements of the sex nerves crucial for induction are no longer able to convey a signal that is still present.

Defining the beginning of induction

Irrespective of how the inductive signal is conveyed to the CNS, the relevant connection has to be established by E12, since the results of the series of ablations discussed here (Fig. 2A) show that induction may be complete as early as E13 in some cases. Induction does not appear to start at the same age in all animals, however, as suggested by the fact that PIC neurons are found only in a small number of animals after ablation at E13 (see results). Also in these cases, as well as after ablation at E14, PIC neurons are often found only in M5 or M6, indicating that the signal is not conveyed to both ganglia simultaneously. There is no apparent rule as to which sex ganglion receives the signal first, suggesting that, although M6 is further away from the male genitalia than M5, induction under these conditions is not delayed by the greater distance. We conclude that induction may start as early as E12-E13 in a given population of embryos. However, in most animals, it is well under way only by E14, and, as shown by Baptista and Macagno (1988a), is complete at about E16.

Termination of the sensitive period

Results from the ablation/transplantation experiments (Figs 2B, 3A) show that induction, although generally complete by E16 during normal development, can take place during E16 with relatively high efficiency, despite the absence of the male organ for 5-6 days prior to its replacement. Induction is no longer possible from E18 onward under these conditions. The experiment in which the male genitalia were present only from E16-E17 shows that the time during which interactions take place can be as little as 1 day at the end of the critical period. In contrast, induction takes about two days after the establishment of the sex nerves (i.e. from E12 to E14) at the beginning of the inductive period (compare Fig. 2A andB). Perhaps the signal-conveying cells or the precursors of the PIC neurons have to mature in order for induction to proceed at the earlier times. From E17 on, induction frequency declines sharply either because the pre-cursors fail to respond to the signal or because central neu-rons that normally convey the signal are no longer able to do so.

Anterograde versus retrograde interaction

Although there is no direct evidence that central, rather than peripheral, neurons convey the inductive signal, we favour this possibility for the following reasons.

First, the sex nerves are established by growing central fibers that contact the male genitalia by E12 (Jellies et al., 1987; Baptista and Macagno, 1988b), preceding the beginning of the inductive period by a day. Jellies and Kristan (1988), by using double labeling with antibodies to serotonin and tubulin, have demonstrated that the leading growth cones in sex nerve formation belong to the (serotonergic) Retzius cells. They found no evidence for sensory projections advancing simultaneously from the male genitalia to the sex ganglia. Nor can sensory cells in the male genitalia be demonstrated at this developmental stage with either of the two different monoclonal antibodies Laz 1-1 and Lan 3-2 that are thought to stain most, if not all, of the peripheral sensory neurons in the early embryo (Zipser and McKay, 1981; McKay et al., 1984).

Second, as shown by restoration of male genitalia at E16, regeneration of innervation must take place in less than one day. Efferent fibers, for instance of the RPE motor neurons, continue to grow in the absence of their target even after embryogenesis (Baptista and Macagno, 1988b), and around E14 are able to contact a dye-labeled implanted male organ within one day, presumably by regenerating only the most distal parts of their axons (unpublished observations). After ablation of the sex organs, the Retzius cells in the sex gan-glia are found to branch extensively in the body wall, thereby resembling the morphology of their homologs in other segments (Loer and Kristan, 1989). These branches could be envisioned to reinnervate the implant in a very short period of time. Afferents from the implant, in contrast, would have to regrow their axons the distance from the male genitalia to M6 (approx. 800 µm at E16), a task unlikely to be achieved in one day.

Third, an E10 male organ is capable of promoting induction when transplanted to an E16 host (Figs 2D, 3B). Since induction at E17 is virtually impossible, this male organ, at the time of induction, cannot have been older than 11 days. If the interaction were performed by sex afferents, one would expect its time course to be dependent on the out-growth of sensory fibers and hence on the age of the transplant. However, at E11, there is no visible connection between the male genitalia and the sex ganglia, yet an organ of this age is capable of induction to the same extent as an E16 organ (compare Fig. 2B and D). The only way that this is feasible in an anterograde interaction is by these younger organs having grown sensory projections prematurely within one day, reaching the sex ganglia and inducing the PIC neurons, an event that seems unlikely at best. We cannot, at present, exclude that the graft was influenced by its older environment such that it matured much faster than usual. Male organs in these heterochronic replacements, however, at the time when embryos were killed at E26, were considerably smaller than those in E26 animals from isochronic replacement experiments (compare Fig. 3A and B). The results from both heterochronic transplantation experiments suggest that, within the sensitive period, the age of the inducing organ is irrelevant, a finding one would not expect if the induction were dependent on the ingrowth or regeneration of afferent fibers.

The declining efficiency with which older male organs induced PIC neurons could have several different reasons: either, the inducing factor(s) are absent from male genitalia older than P10, or reinnervation by the crucial elements does not take place within the sensitive period. Although grafts were usually reinnervated by motor neurons very rapidly, the large size of the older transplants could impede this process.

The fact that male organs in ectopic locations do not trigger PIC neuron birth in ganglia other than sex ganglia (Baptista and Macagno, 1988a) could also be explained by assuming that efferents, rather than afferents, are pivotal for induction: if the conveying elements were neurons unique to M5 and M6 that connect to the male genitalia, like the RPE and LPE motor neurons (Zipser, 1979) or several other neurons that can be visualized by backfilling the sex nerves with dyes (e.g. Passani et al., 1991), then the inductive signal from this organ cannot be conveyed to a ganglion other than M5 or M6, whether the precursors of the PIC neurons are there or not. In conclusion, although the evidence is admittedly circumstantial, a retrograde mode of interaction best accounts for the data. Interactions between the male genitalia and neurons in the sex ganglia are known to affect the morphology of these cells only after target contact, as exemplified by the RPE motor neurons (Baptista and Macagno, 1988b) and the Retzius neurons (Jellies et al., 1987; Loer et al., 1987; Loer and Kristan, 1989). In the latter case, the beginning of these changes coincide with the beginning of the inductive interaction determined in this study. Whether the Retzius neurons in the sex ganglia have any bearing on the induction of the PIC neurons, however, remains to be shown.

Retrograde interactions affecting neuron number in the CNS have been known to exist in vertebrate development for a long time, and the nature of the conveyed signals is currently investigated at many levels. The most common view about the nature of these interactions is that the target releases some trophic factor for which the central neurons compete. At least some of these factors appear to be growth factors such as nerve growth factor (NGF), that support cell maintenance, growth and differentiation, and hence prevent cell death (for review, see Thoenen and Barde, 1990, Lowry and Vrbova, 1992). Growth factors such as basic and acidic fibroblast growth factor (bFGF and aFGF), insulin-like growth factor (IGF-I) and insulin have, in addition, been shown to promote proliferation of telencephalic cells from mouse, rat and chick in vitro (reviewed by Rohrer, 1990). Whether these factors promote proliferation in the intact vertebrate CNS has not been shown conclusively. The development of the nervous system of the leech could serve as a simple model to understand how, on a cellular level, interactions between periphery and CNS can affect neuronal numbers.

We thank Gabriel Aisemberg, Marty Chalfie, Nick George, and Darcy Kelley for their most valuable comments on the manuscript, and Nik Necles for help with the photography. T.B. would like to thank Carlos Baptista and Beatrice Passani for discussion and very helpful advice at the beginning of this study.

Baptista
,
C. A.
and
Macagno
,
E. R.
(
1988a
).
The role of the sexual organs in the generation of postembryonic neurons in the leech Hirudo medicinalis
.
J. Neurobiol
.
19
,
707
726
.
Baptista
,
C. A.
and
Macagno
,
E. R.
(
1988b
).
Modulation of the pattern of axonal projections of a leech motor neuron by ablation or transplantation of its target
.
Neuron
1
,
949
962
.
Baptista
,
C. A.
,
Gershon
,
T. R.
and
Macagno
,
E. R.
(
1990
).
Peripheral organs control central neurogenesis in the leech
.
Nature
346
,
855
858
.
Becker
,
T.
and
Macagno
,
E. R.
(
1992
).
Central neurogenesis in the leech
.
In: Determinants of Neuronal Identity
. (ed.
M.
Shankland
and
E. R.
Macagno
)
New York
:
Academic Press
, pp
79
95
.
Fernandez
,
J.
and
Stent
,
G. S.
(
1982
).
Embryonic development of the hirudinid leech Hirudo medicinalis: Structure, development and segmentation of the germinal plate
.
J. Embryol. Exp. Morph
.
72
,
71
96
.
Fischbach
,
K. F.
and
Technau
,
G. M.
(
1984
).
Cell degeneration in the developing optic lobes of the sine oculis and small optic lobes mutants of Drosophila melanogaster
.
Dev. Biol
.
104
,
219
239
.
Gratzner
,
H. G.
(
1982
).
Monoclonal antibody to 5-bromo and 5-iododeoxyuridine: a new reagent for detection of DNA replication
.
Science
218
,
474
478
.
Jellies
,
J.
,
Loer
,
C. M.
and
Kristan
,
W. B.
(
1987
).
Morphological changes in leech retzius neurons after target contact during embryogenesis
.
J. Neurosci
.
7
,
2618
2629
.
Jellies
,
J.
and
Kristan
,
W. B.
(
1988
).
An identified cell is required for the formation of a major nerve during embryogenesis of the leech
.
J. Neurobiol
.
19
,
153
165
.
Loer
,
C. M.
,
Jellies
,
J.
and
Kristan
,
W. B.
(
1987
).
Segment-specific morphogenesis of leech Retzius neurons requires particular peripheral targets
.
J. Neurosci
.
7
,
2630
2638
.
Loer
,
C. M.
and
Kristan
,
W. B.
(
1989
).
Peripheral target choice by homologous neurons during embryogenesis of the medicinal leech. II. Innervation of ectopic reproductive tissue by nonreproductive Retzius cells
.
J. Neurosci
.
9
,
528
538
.
LoPresti
,
V. A.
,
Macagno
,
E. R.
and
Levinthal
,
C.
(
1973
).
Structure and development of neuronal connections in isogenic organisms: cellular interactions in the development of the optic lamina of Daphnia
.
Proc. Natl. Acad. Sci. U.S.A
.
70
,
433
437
.
Lowry
,
M. B.
and
Vrbova
,
G.
(
1992
).
Dependence of postnatal motoneurones on their targets: review and hypothesis
.
Trends Neurosci
.
15
,
80
84
.
Macagno
,
E. R.
(
1979
).
Cellular interactions and pattern formation in the development of the visual system of Daphnia magna (Crustacea, Brachiopoda). I. Interactions between embryonic retinular fibers and laminar neurons
.
Dev. Biol
.
73
,
206
238
.
Macagno
,
E. R.
(
1980
).
Number and distribution of neurons in leech segmental ganglia
.
J. Comp. Neurol
.
190
,
283
302
.
McKay
,
R. D. G.
,
Hockfield
,
S.
,
Johansen
,
I.
,
Thompson
,
I.
and
Frederiksen
,
K.
(
1984
).
Surface molecules identify groups of growing axons
.
Science
222
,
678
684
.
Meyerowitz
,
E.
, and
Kankel
,
D.
(
1978
).
A genetic analysis of visual system development in Drosophila melanogaster
.
Dev. Biol
.
62
,
112
162
.
Muller
,
K. J.
,
Nicholls
,
J. G.
and
Stent
,
G. S.
(Eds.) (
1981
).
Neurobiology of the Leech
.
New York
:
Cold Spring Harbor Laboratory
.
Oppenheim
,
R. W.
(
1991
).
Cell death during the development of the nervous system
.
Ann. Rev. Neurosci
.
14
,
453
501
.
Passani
,
M. B.
,
Peinado
,
A.
,
Engelman
,
H.
,
Baptista
,
C. A.
and
Macagno
,
E. R.
(
1991
).
Normally unused positional cues guide ectopic afferents in the leech CNS
.
J. Neurosci
.
11
,
3868
3876
.
Power
,
M. E.
(
1943
).
The effect of reduction in numbers of ommatidia upon the brain of Drosophila melanogaster
.
J. Exp. Zool
.
94
,
33
71
.
Rohrer
,
H.
(
1990
).
The role of growth factors in the control of neurogenesis
.
Eur. J. Neurosci
.
2
,
1005
1015
.
Selleck
,
S. B.
and
Steller
,
H.
(
1991
).
The influence of retinal innervation on neurogenesis in the first optic ganglion of Drosophila
.
Neuron
6
,
83
99
.
Steller
,
H.
,
Fischbach
,
K. F.
and
Rubin
,
G. M.
(
1987
).
disconnected: A locus required for neuronal pathway function in the visual system of Drosophila
.
Cell
50
,
1139
1153
.
Stewart
,
R. R.
,
Spergel
,
D.
and
Macagno
,
E. R.
(
1986
).
Segmental differentiation in the leech nervous system: The genesis of cell number in the segmental ganglia of Haemopis marmorata
.
J. Comp. Neurol
.
253
,
253
259
.
Thoenen
,
H.
, and
Barde
,
Y. A.
(
1990
).
Physiology of nerve growth factor
.
Physiol. Rev
.
60
,
1284
1335
.
Williams
,
R. W.
and
Herrup
,
K.
(
1988
).
The control of neuron number
.
Annu. Rev. Neurosci
.
11
,
423
453
.
Zipser
,
B.
(
1979
).
Identifiable neurons controlling penile eversion in the leech
.
J. Neurophysiol
.
42
,
255
464
.
Zipser
,
B.
and
McKay
,
R. D. G.
(
1981
).
Monoclonal antibodies distinguish identifiable neurons in the leech
.
Nature
289
,
549
554
.