Blockade of Endogenous Neurotrophic Factors Prevents the Androgenic Rescue of Rat Spinal Motoneurons

Perineal injections of trkB-IgG, trkC-IgG, and AADH-CNTF attenuate the androgenic sparing of SNB motoneurons

We asked first whether the rescue of SNB motoneurons could be blocked by localized, perineal injections of trophic factor antagonists. The anatomy of the SNB neuromuscular system and the design of the study are illustrated in Figure 1. Evidence suggests that by the day of birth, male rat pups have already been exposed to enough androgens to permanently rescue the SNB system (Breedlove and Arnold, 1983). We therefore took the approach of androgenizing newborn females, and simultaneously applying a blockade of trophic factor receptors.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Anatomy of the SNB neuromuscular system and experimental design. A, SNB motoneurons are located ventromedially in spinal lumbar segments 5 and 6 and innervate striated muscles of the perineum, including the bulbocavernosus (BC) and levator ani (LA). Motoneurons of the RDLN, which are located in the lateral horn of the same spinal segments, primarily innervate foot muscles.B, Newborn female rat pups received subcutaneous (SubQ) injections of 125 μg of testosterone propionate (TP) on postnatal days 1 and 2 (P1,P2) to rescue SNB cells. Control females received an equal volume of the sesame oil vehicle. In addition, trophic factor antagonists or control solutions were delivered into the perineum, directly above the BC/LA muscles on P1, P2, and P3. All females were then killed on P16, after the end of the cell death period in the SNB.

Signaling through CNTF receptors was blocked using AADH-CNTF (Di Marco et al., 1996), and trk-IgGs were used to bind and sequester endogenously produced neurotrophins. Seven experimental groups were initially compared: oil-treated control females receiving perineal injections of PBS (oil/PBS) and TP-treated females receiving either PBS, IgG alone (2.5 μg), trkA-IgG (2.5 μg), trkB-IgG (2.5 μg), trkC-IgG (2.5 μg), or AADH-CNTF (1 μg) at the perineum.

There was a significant effect of treatment on SNB motoneuron number (F_(6,61) = 14.5;p < 0.0001) (Fig. 2). As expected, androgenized females had many more SNB motoneurons than did oil-treated females (p < 0.0005; TP/PBS vs oil/PBS groups). TrkB-IgG, trkC-IgG, and AADH-CNTF, when administered to androgenized females, each significantly reduced SNB motoneuron number relative to that of TP/PBS females (pvalues ≤ 0.01 in each case). In fact, SNB cell counts in androgenized females receiving AADH-CNTF, trkB-IgG, or trkC-IgG did not differ significantly from counts in oil-treated, control females. Neither trkA-IgG nor human IgG alone had an effect on SNB motoneuron survival (p values > 0.50 compared with TP/PBS females). Although not shown here, we also examined the effects of higher doses of AADH-CNTF and trkB-IgG on SNB motoneuron survival (5 and 25 μg/d, respectively). When administered into the perineum, high doses of AADH-CNTF and trkB-IgG again reduced SNB cell number relative to counts in androgenized controls (p < 0.05 in both cases), but they were no more effective than the lower doses illustrated in Figure 2.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Perineal administration of trophic factor antagonists attenuates the sparing of SNB motoneurons. The number of motoneurons in the SNB of female rats receiving systemic injections of sesame oil or testosterone propionate (TP) and perineal injections of PBS, IgG alone (2.5 μg), trk-IgGs (each at 2.5 μg), or AADH-CNTF (1 μg). TrkB-IgG, trkC-IgG, and AADH-CNTF each significantly reduced the androgenic sparing of SNB motoneurons. * indicates significantly different from TP/PBS androgenized controls; † indicates significantly different from Oil/PBS controls. Dashed lines indicate the mean number of motoneurons in the SNB of the two control groups (TP/PBS and Oil/PBS). Numbers at the base of each bar indicate sample size.

The effects of our treatments on SNB motoneuron cell size are given in Table 1. The sizes of SNB somas and nucleoli were not significantly affected by treatment. There was an overall effect of treatment on the size of SNB nuclei (F_(6,61) = 2.68; p < 0.05), but none of the individual pairwise comparisons reached statistical significance in post hoc tests. We conclude that our treatments influenced SNB cell number, while having little or no effect on motoneuron cell size.

Motoneuron counts also were determined for the RDLN. RDLN motoneurons are located at the same level of the spinal cord as the SNB and innervate intrinsic muscles of the foot. In contrast to effects in the SNB, neither TP nor any of the trophic factor antagonists influenced motoneuron number in the RDLN (F_(6,54)= 1.30; p > 0.25; data not shown). Body weights at time of killing also did not differ across groups (F_(6,67) = 1.01; p > 0.40; data not shown). Thus, the reductions in SNB motoneuron number after treatment with trkB-IgG, trkC-IgG, and AADH-CNTF do not appear to be attributable to nonspecific toxic effects on motoneurons or on the health of the animals.

Trophic factor antagonists do not alter SNB motoneuron number in oil-treated females

The findings described above suggest that endogenously produced trophic factors signaling through CNTFRα, trkB, and trkC regulate the androgenic rescue of SNB motoneurons innervating the androgen-dependent BC/LA muscles. Alternatively, it remained possible that the antagonist treatments reduced the survival of androgen-independent SNB motoneurons that survive even in control females and primarily innervate the external anal sphincter muscle. To test this possibility, female rat pups were treated with subcutaneous injections of sesame oil on P1 and P2 and perineal injections of either PBS, trkB-IgG (2.5 μg), or trkC-IgG (2.5 μg) on P1, P2, and P3. AADH-CNTF was not available for this experiment. As is evident in Figure3, the trophic factor antagonists did not reduce SNB motoneuron number in oil-treated females (F_(2,29) = 0.88; p > 0.40). It seems likely, therefore, that perineal injections of trophic factor antagonists reduce SNB motoneuron number in androgenized females by blocking the rescue of those motoneurons innervating the androgen-dependent BC and LA muscles.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Trk-IgGs do not decrease motoneuron number in oil-treated females. The number of SNB motoneurons in oil-treated females that were given perineal injections of PBS, trkB-IgG (2.5 μg/d), or trkC-IgG (2.5 μg/d) is shown. There were no significant differences between groups. Numbers at the base of each bar indicate sample size.

Trophic factor antagonists do not block masculinization of the SNB when administered systemically

Next, we asked whether trkB-IgG, trkC-IgG, or AADH-CNTF could block masculinization of SNB motoneuron number when delivered systemically rather than into the perineum. By comparing the efficacy of systemic and perineal injections, we might begin to address questions regarding the site of action of injected antagonists. We reasoned that antagonists injected at any site will spread to some extent throughout the body, but that the perineal muscles and SNB motor nerve terminals should have better access to antagonists delivered perineally than to antagonists delivered systemically.

Two adjustments were made from the previous experiments in an attempt to enhance the efficacy of systemic injections: doses of antagonists were increased 10- to 20-fold over those administered perineally, and injections were given twice per day. Newborn female rats were androgenized as above, and antagonists were administered as five subcutaneous injections at the nape of the neck given every 12 hr between P1 and the morning of P3. Then, pups were left undisturbed until killing at P16.

The overall effect of treatment on SNB motoneuron number did not quite reach significance (F_(3,29) = 2.77;p = 0.059) (Fig.4A), suggesting that systemic administration of trophic factor antagonists is ineffective or marginally effective in attenuating the androgenic rescue of SNB cells. TrkB-IgG was clearly ineffective when administered systemically (Fig.4A). There appeared to be a trend for slightly lower SNB cell numbers in animals treated systemically with trkC-IgG or AADH-CNTF; however, if post hoc tests were performed, none of the individual antagonist treatments significantly altered SNB motoneuron number relative to cell counts in androgenized controls (allp values ≥ 0.05). We conclude that systemic administration of trkB-IgG, trkC-IgG, or AADH-CNTF at the doses used here does not block the androgenic sparing of SNB cells.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Motoneuron number in animals receiving systemic injections of trophic factor antagonists. The number of SNB (A) and RDLN (B) motoneurons in androgenized female rats treated every 12 hr with subcutaneous injections of PBS, trkB-IgG (25 μg), trkC-IgG (25 μg), or AADH-CNTF (5 μg) is shown. There were no significant differences among the groups for either motor pool. Numbers at the base of each bar indicate sample size.

As was seen after perineal injections, there was no effect of subcutaneous antagonist treatments on RDLN motoneuron number (Fig.4B) (F_(3,29) = 0.44;p > 0.70). There also was no effect on SNB cell size, RDLN cell size, or body weight (all p values > 0.15; data not shown).

Endogenous trophic factors regulating motoneuron survival

Trophic factors of the CNTF and neurotrophin families enhance the survival of motoneurons in vitro and in vivo(Elliott and Snider, 1996; Mitsumoto and Tsuzaka, 1999). Despite this, the developmental cell death of motoneurons is not affected in transgenic mice with deletions of the genes for CNTF, NGF, BDNF, NT-3, NT-4/5, trkA, trkB, or trkC (for review, see Elliott and Snider, 1996;Ip and Yancopoulos, 1996). The minimal effects of such knock-out studies generally have been attributed to redundancy and/or compensatory mechanisms that take place when a gene product is eliminated throughout embryonic development. In the current study, trophic factor interactions were normal until perturbed during the first few postnatal days. Under these conditions, signaling through trkB, trkC, and CNTFRα all appear to be required for the rescue of SNB motoneurons. These findings raise the interesting possibility that the acute deprivation of a single trophic factor may have very different consequences from the chronic absence of that factor or its cognate receptor. This might occur if, for example, previous exposure to endogenous trophic factors alters the trophic “dependencies” of a developing motoneuron, as has been described for sensory neurons in vitro (Vogel and Davies, 1991).

The present results also suggest that SNB motoneuron survival requires the simultaneous activation of at least two classes of receptors that use distinct, as well as overlapping, signal cascades. TrkB and trkC are tyrosine kinase receptors that activate multiple intracellular signaling cascades, including the ras/raf/MAP kinase pathway (Heumann, 1994). In contrast, CNTFRα is part of a three-part cytokine receptor complex that signals primarily through a JAK/STAT pathway (Ip and Yancopoulos, 1996). CNTF and BDNF have been shown previously to have synergistic effects on motoneurons in vitro andin vivo (Wong et al., 1993; Mitsumoto et al., 1994; Stoop and Poo, 1996). However, the molecular basis for the collaborative actions of neural cytokines and neurotrophins remains unknown.

The observation that AADH-CNTF attenuated the rescue of SNB motoneurons is consistent with our previous findings that exogenous CNTF increases SNB cell number in newborn female rats (Forger et al., 1993) and that deletion of the CNTFRα gene eliminates the sex difference in the SNB of mice (Forger et al., 1997). We conclude that an endogenous ligand of CNTFRα is required for motoneuron survival, including SNB motoneuron survival. That ligand is unlikely to be CNTF itself, however, as several lines of evidence indicate that a second ligand of CNTFRα exists and regulates motoneuron number during development (DeChiara et al., 1995; Shelton, 1996; Forger et al., 1997). Although not yet fully characterized, a molecular heterodimer has been identified recently that may turn out to be the long-sought, second ligand for the CNTF receptor (Elson et al., 2000).

The perineal administration of trkB-IgG or trkC-IgG also decreased SNB motoneuron number in androgenized females. The trkB ligand, BDNF, has been shown previously to regulate androgen receptor immunoreactivity of adult SNB motoneurons (Al-Shamma and Arnold, 1997; Yang and Arnold, 2000); the present results suggest that BDNF (and/or NT-4) is also required for SNB cell survival during development. It is somewhat surprising that trkB-IgG was no more effective than trkC-IgG in reducing motoneuron survival in the present study. The trkC ligand, NT-3, is primarily implicated in the survival of Ia spinal afferents (Ernfors et al., 1994; Klein et al., 1994), and exogenous NT-3 is a less potent trophic factor for α motoneurons than either BDNF or NT-4/5 in several assays (Sendtner et al., 1992; Koliatsos et al., 1993; Houenou et al., 1994; Fernandes et al., 1998; Wiese et al., 1999). However, the present findings, together with the observation that NT-3 is expressed at high levels in developing skeletal muscles (Funakoshi et al., 1995; Griesbeck et al., 1995), suggest that endogenous NT-3 may be a more important motoneuron trophic factor than previously recognized.

The perineal administration of trkA-IgG did not affect SNB motoneuron number at the dose tested here. This indicates that the trkA ligand, NGF, is not involved in the regulation of SNB motoneuron cell death. In the absence of an activity assay for trkA-IgG in vivo, this conclusion must be viewed as tentative. However, the lack of effect of trkA-IgG is consistent with several previous demonstrations that exogenous NGF does not support the survival of either spinal or brainstem motoneurons, in vitro or in vivo(Sendtner et al., 1992; Houenou et al., 1994; Baumgartner and Shine, 1997; Wiese et al., 1999). In this respect, the trophic factor requirements of developing SNB motoneurons appear to be similar to those of other, nonsexually dimorphic motoneurons.

These observations naturally raise the question of how a blockade of endogenous trophic factors would reduce motoneuron survival during cell death in other, nonandrogen-dependent motor pools. Our antagonist treatments did not affect motoneuron number in the RDLN of the lateral horn. Although this observation was useful in confirming that the antagonists are not generally toxic, the perineal injections that effectively reduced SNB cell number were distant from RDLN target muscles, which are primarily located in the foot. In addition, motoneuron cell death in the lateral horn is complete before birth (Harris and McCaig,1984; Oppenheim, 1986), i.e., before the start of our treatments. To test the role of endogenous, target-derived trophic factors in the regulation of developmental cell death for most motoneurons, one would have to impose a trophic factor blockade at the relevant target muscles during embryonic development. This would obviously be technically difficult and may explain why, to the best of our knowledge, no previous study has attempted a similar trophic factor deprivation during the time of cell death for any other mammalian motor system. In embryonic chicks, Krieglstein et al. (2000) have recently reported profound effects on motoneuron cell death after the systemic immunoneutralization of endogenous transforming growth factors β.

Site(s) of trophic factor action

Converging evidence indicates that androgens act at the perineal muscles to indirectly spare SNB cells (Fishman et al., 1990; Freeman et al., 1997; Jordan et al., 1997). In the present study, trophic factor antagonists attenuated the androgenic rescue of SNB motoneurons when low doses were injected into the perineum, whereas much higher doses did not significantly reduce motoneuron number when administered systemically. Although we cannot rule out central sites of action on the basis of these findings, the observations are consistent with the idea that trkB-IgG, trkC-IgG, and AADH-CNTF prevented the sparing of SNB cells primarily by blocking trophic factor interactions in the perineum. Specifically, testosterone may act at the BC/LA muscles to prevent their degeneration and to increase the availability of target-derived BDNF, NT-4/5, NT-3, and/or a CNTF-like factor. Although the trophic factors produced by the perineal muscles have not yet been fully characterized, preliminary evidence indicates that NT-3 and NT-4/5 are present in the BC/LA of newborn rats at concentrations similar to those found in hindlimb muscles (J.-J. Park and N. G. Forger, unpublished observations). Other developing skeletal muscles produce BDNF, NT-3, and NT-4/5 (Koliatsos et al., 1993; Funakoshi et al., 1995; Griesbeck et al., 1995). Because developing motoneurons express the corresponding trophic factor receptors (Yan et al., 1992;Henderson et al., 1993; Koliatsos et al., 1993; MacLennan et al., 1996), such target-derived factors could presumably enhance SNB motoneuron survival via a classical, retrograde mechanism.

In addition, because CNTF, BDNF, and NT-4/5 each regulate synaptic strength in developing Xenopus muscle–motoneuron cocultures (Stoop and Poo, 1996; Wang et al., 1998), changes in synaptic activity could conceivably have contributed to the effects of trophic factor blockade seen here. Finally, the BC/LA muscles themselves express CNTFRα and trkB during perinatal life (Xu and Forger,1998; and our unpublished observations), and exogenous CNTF prevents BC/LA atrophy in newborn females (Forger et al., 1993). Thus, AADH-CNTF and/or trkB-IgG might influence SNB cell survival via an indirect effect on SNB target muscles. Although our preliminary analyses indicate that the trophic factor antagonists used here did not prevent the androgenic rescue of the BC/LA muscles, we are currently exploring the possibility of more subtle effects of trophic factor antagonists on muscle development.