Reciprocal developmental changes in the roles of Bcl-w and Bcl-x_L in regulating sensory neuron survival

Proteins of the Bcl2 family play a key role in controlling the activation of caspases, the proteases that dismantle the cell in programmed cell death (Adams and Cory, 1998; Korsmeyer, 1999; Vaux and Korsmeyer, 1999). Bcl2-related proteins fall into two groups that generally either repress apoptosis (Bcl2, Bcl-x_L [Bcl2l_L], Bcl-w [Bcl2l2], A1 and Mcl1) or promote apoptosis (Bax, Bcl-x_s [Bcl2l_s], Bak, Bad, Bid, Bik [Biklk], Bok [Bokl-pending], Bim and Blk). (Symbols in square brackets are the nomenclature according to Mouse Genome Database.) Although it is not fully understood how they do this, there is growing evidence that some members of the Bcl2 family control the release of cytochrome c from mitochondria which promotes a cascade of caspase activation by interacting with the adapter protein Apaf1 which in turn activates pro-caspase-9 (Li et al., 1997b; Qin et al., 1999). Pro-apoptotic members like Bax and Bak increase mitochondrial permeability allowing cytochrome c to pass into the cytosol, whereas anti-apoptotic members like Bcl2 and Bcl-x_L prevent cytochrome c release (Kharbanda et al., 1997; Kluck et al., 1997; Yang et al., 1997; Shimizu et al., 1999). Inhibiting the release of cytochrome c from mitochondria cannot entirely account for the means by which Bcl2 and Bcl-x_L prevent apoptosis because overexpression of these proteins can prevent apoptosis in cells in which cytochrome c is present in the cytosol (Li et al., 1997a; Rosse et al., 1998). Bcl-x_L can, however, promote cell survival by binding directly to Apaf1 and preventing it from activating pro-caspase-9 (Hu et al., 1998; Pan et al., 1998), and there is evidence that Bcl2 can regulate the activation of membrane-associated procaspase-3 independently of cytochrome c (Krebs et al., 1999). In addition to direct interaction with proteins of the survival/death machinery of the cell, it is known that members of the Bcl2 family can selectively heterodimerise and modulate one another’s function, suggesting that the ratio of pro-apoptotic to anti-apoptotic proteins in the cell governs whether it survives or dies (Korsmeyer et al., 1993).

There is growing evidence that members of the Bcl2 family play important roles in regulating neuronal survival. Whereas overexpression of Bcl2 or Bcl-x_L promote neuronal survival in vitro and in vivo (Garcia et al., 1992; Allsopp et al., 1993; Martinou et al., 1994; Farlie et al., 1995; Gonzalez-Garcia et al., 1995; Middleton et al., 1996; Parsadanian et al., 1998), reduction or elimination of endogenous Bcl2 or Bcl-x_L reduces neuronal survival (Allsopp et al., 1995; Motoyama et al., 1995; Michaelidis et al., 1996; Piñón et al., 1997). In contrast, Bax overexpression accelerates neuronal death following neurotrophin withdrawal (Vekrellis et al., 1997; Martinou et al., 1998), and reduction or elimination of endogenous Bax promotes the survival of neurons without neurotrophic factors in culture and prevents their death in vivo (Deckwerth et al., 1996; Gillardon et al., 1996; Miller et al., 1997; White et al., 1998; Bar-Peled et al., 1999).

Bcl-w is a recently identified member of the Bcl2 family, overexpression of which prevents the death of myeloid and lymphoid cell lines following cytokine deprivation (Gibson et al., 1996). Although Bcl-w has recently been shown to be widely expressed in the nervous system (Hamner et al., 1999), its potential role in regulating neuronal survival is not known. We have investigated this role by manipulating Bcl-w expression in developing sensory neurons obtained from mouse embryos and neonates. We have compared the role of Bcl-w in regulating neuronal survival with that of Bcl-x_L, an anti-apoptotic member of the Bcl2 family (Boise et al., 1993) that has already been implicated in regulating neuronal survival. Bcl-x_L is the major protein encoded by the bcl-x gene, the other isoforms being Bcl-x_S, which inhibits the anti-apoptotic actions of Bcl-x_L and Bcl2, and Bcl-xβ,which like Bcl-x_L has an anti-apoptotic function (Boise et al., 1993; Gonzalez-Garcia et al., 1994; Minn et al., 1996). Bcl-x-deficient embryos die around E13 with extensive apoptosis in the central nervous system and dorsal root ganglia (Motoyama et al., 1995), indicating that Bcl-x is required for the survival of sensory neurons during the early stages of their development. However, the extent to which Bcl-x is required for the survival of sensory neurons at later stages of development is not known.

We have focused on two well-characterised populations of cranial sensory neurons that are supported by different neurotrophins during and beyond the period of naturally occurring neuronal death. The majority of mouse trigeminal ganglion neurons are supported by NGF in culture from E12 onwards and depend on NGF derived from their peripheral targets, for survival during the period of naturally occurring neuronal death in vivo (Davies et al., 1987; Buchman and Davies, 1993; Piñón et al., 1996). The majority of nodose ganglion neurons are supported by BDNF in culture from E12 onward and depend on BDNF for survival in vivo during embryonic development (Davies et al., 1993a; Ernfors et al., 1994; Jones et al., 1994). The total number of neurons in the trigeminal ganglion peaks between E13 and E14 and falls by half by birth as a result of naturally occurring neuronal death (Davies and Lumsden, 1984) with a peak of dying neurons in the ganglion at E14 (Piñón et al., 1996). Our studies of manipulating the expression of Bcl-w and Bcl-x_L in these neurons during development have shown that Bcl-w becomes increasingly important for survival during the late fetal and postnatal period whereas the requirement for Bcl-x_L decreases over this same period of development, differences which accord with developmental changes in these expression of these proteins.

Dissociated cultures of trigeminal and nodose ganglion neurons were established from CD1 mouse embryos at 16 and 18 days gestation (E16 and E18) and from postnatal day 1 pups (P1). The dissected ganglia were trypsinised and dissociated by trituration, and the neurons were purified free of non-neuronal cells by differential sedimentation (Davies, 1986). The neurons (>95% pure) were grown in defined, serum-free medium on a poly-ornithine/laminin substratum in 60 mm diameter tissue culture Petri dishes (Davies et al., 1993b). After an initial 12 hour incubation period with either 5 ng/ml NGF (trigeminal neurons) or 5 ng/ml BDNF (nodose neurons), the cultures were washed extensively to remove these neurotrophins and were injected intranuclearly (Allsopp et al., 1993) with expression plasmids. The pSFFV plasmid containing Bcl-xL in the sense or antisense orientation and the pcDNAIII plasmid containing Bcl-w cDNA in the sense or antisense orientation were used to investigate the role of Bcl-xL and Bcl-w in regulating neuronal survival. Expression plasmids without a cDNA insert were used to control for any non-specific effects of the injection procedure. The expression plasmids were diluted in 100 mM potassium phosphate buffer pH 7 to a concentration of 100 ng/ml and filtered through a 0.22 μm filter prior to injection. Some cultures were re-supplemented with the same neurotrophin 30 minutes after injection. The number of surviving neurons was counted at 24-hourly intervals after injection and is expressed as a percentage of the number injected. Between 45 and 100 neurons were injected for each experimental condition in each experiment.

The heads of E12, E14, E16 and E18 embryos and P1 pups were fixed in Zenker’s fluid (5% glacial acetic acid, 5% mercuric chloride, 2.5% potassium dichromate, 1% sodium sulphate in distilled water) overnight prior to dehydration and embedding in paraffin wax. Serial sections of the heads were cut at 8 μm and mounted on poly-lysine coated slides (BD) or Ultra stick Gold Seal Slides (Erie Scientific). For quantifying the total number of neurons in the trigeminal and nodose ganglia, the sections were stained with cresyl fast violet acetate and neurons were counted as described previously (Piñón et al., 1997).

For Bcl-w staining, the sections were cleared in xylene and dehydrated before quenching (3% hydrogen peroxide in PBS) for 20 minutes. Non-specific interactions were blocked using 10% rabbit serum in PBS with 0.4% Triton X-100 at room temperature for 1 hour. The sections were incubated with polyclonal goat-anti-Bcl-w (Santa Cruz) for 2 hours at 37°C. For Bcl-x_L staining, the sections were cleared in xylene and dehydrated before quenching (3% hydrogen peroxide, 10% methanol in PBS) for 20 minutes. Non-specific interactions were blocked using 10% goat serum in PBS with 0.4% Triton X-100 at room temperature for 3 hours. The sections were incubated with polyclonal rabbit-anti-Bcl-x_L (Santa Cruz) overnight at 4°C. After washing, the cells were labelled with biotinylated secondary antibody (1:200) and avidin/biotinylated horseradish peroxidase macromolecular complex (Vectastain ABC Elite Kit, Vector Laboratories). The substrate used for the peroxidase reaction was NovaRed (Vector Laboratories). After staining, the sections were washed in tap water before rehydration and mounting. Estimation of the number of labelled neurons was carried out as for cresyl fast violet-stained sections, however, only strongly positive neurons were counted.

A quantitative RT-PCR technique (Wyatt and Davies, 1993) was used to measure the levels of Bcl-w and Bcl-x_L mRNAs in RNA extracted from nodose and trigeminal ganglia. The levels of mRNA for the housekeeping protein GAPDH were also determined by quantitative RT-PCR allowing Bcl-w and Bcl-x_L mRNA expression to be calculated relative to GAPDH mRNA. The RT-PCR reactions were calibrated by the inclusion of known amounts of cRNA competitor templates for each of the mRNAs in the reverse transcription reaction. The cRNA competitor templates were synthesized in vitro from cDNA competitor constructs.

Total RNA (Chomczynski and Sacchi, 1987) spiked with known amounts of the appropriate competitor cRNA, was reverse transcribed for 60 minutes at 37°C with Superscript M-MLV reverse transcriptase (Gibco) in a 40 μl reaction containing the manufacturer’s buffer supplemented with 0.5 mM dNTPs and 10 μM random hexanucleotides. A 5 μl aliquot of each reverse transcription reaction was then amplified in a 50 μl PCR reaction containing: 1× Taq Supreme PCR buffer (Helena Biosciences), 0.1 mM dNTPs, 1.5 Units of Taq Supreme polymerase, 40 ng of primers specific for either Bcl-x_L, Bcl-w or GAPDH.

The forward primer for assaying Bcl-w cDNA was 5′-CGGAACATGGCTTGTAGCTC-3′ and the reverse primer was 5′-AATCCCATTCATCTAGTCGAG-3′. These hybridize 73 bp apart in mouse Bcl-w cDNA and 77 bp apart in the Bcl-w competitor cDNA. The forward primer for amplifying Bcl-x_L was: 5′-TCTGAATGACCACCTAGAG-3′ and the reverse primer was: 5′-GTTCCCGTAGAGATCCAC-3′. These hybridize 72 bp apart in mouse Bcl-x_L cDNA and 76 bp apart in the Bcl-xL competitor cDNA. The details of the GAPDH RT-PCR assay have been described elsewhere (Wyatt et al., 1997).

Bcl-x_L cDNA was amplified by cycling at 95°C for 1 minute, followed by 1 minute at 53°C, followed by 1 minute at 68°C. For Bcl-w cDNA, cycling conditions were the same except the annealing step was 1 minute at 56°C. The exact number of PCR cycles was dependent on the initial target cDNA concentration but was typically between 25 and 30 cycles. The RT-PCR products of the native Bcl-w, Bcl-x_L and GAPDH mRNAs and those of the cRNA competitor species were separated on 8% non-denaturing polyacrylamide gels. These gels were subsequently stained with SyberGold (Cambridge Biosciences) and the intensity of the RT-PCR products were determined using a gel documentation system (Biogene) with Phoretix software.

To determine if overexpression of Bcl-w protein is able to prevent the death of neurotrophin-deprived neurons, BDNF-deprived and NGF-deprived sensory neurons were injected with an expression plasmid containing bcl-w cDNA in the sense orientation. E16, E18 and P1 mouse nodose and trigeminal neurons were purified free of non-neuronal cells and were grown with either BDNF (nodose neurons) or NGF (trigeminal neurons) for 12 hours. After washing to remove these factors, the neurons were microinjected with the Bcl-w expression plasmid and their survival was compared at 24-hourly intervals with the survival of neurons injected with an empty expression plasmid and with uninjected neurons. Immunocytochemistry revealed that neurons injected with the Bcl-w expression plasmid were strongly stained for Bcl-w protein, confirming this protein was effectively expressed in injected neurons (data not shown).

At all stages studied, nodose and trigeminal neurons died rapidly following neurotrophin deprivation (Figs 1 and 2). The rate and magnitude of neuronal death was the same in uninjected neurons and neurons that were injected with the empty expression plasmid (data not shown), indicating that microinjection itself neither enhances neuronal survival nor accelerates death following neurotrophin deprivation. Overexpression of Bcl-w rescued a substantial number of neurotrophin-deprived nodose and trigeminal neurons at all ages studied (Figs 1 and 2). In BDNF-deprived nodose neurons, the effectiveness with which Bcl-w overexpression prevented cell death improved with age, from 43% survival in E16 cultures 72 hours after injection to 71% survival in P1 cultures. Although the level of survival in control cultures also showed a small increase during these stages of development, the survival of Bcl-w-overexpressing neurons also increased relative to the survival of control injected neurons (from 43% above control survival at E16 to 60% above control survival at P1). In cultures of trigeminal neurons, there was no obvious change in the effectiveness with which Bcl-w overexpression prevented neuronal death following NGF deprivation. By 72 hours at each age studied, all NGF-deprived trigeminal neurons had died in control cultures, and between 60 and 70% of the neurons were still surviving following injection of the Bcl-w expression plasmid.

To ascertain the physiological significance of endogenously expressed Bcl-w protein in regulating the survival of developing nodose and trigeminal neurons, we injected plasmids expressing anti-sense Bcl-w RNA to interfere with the synthesis of Bcl-w. We also injected separate sets of neurons with an empty expression plasmid to control for the injection procedure. At all ages studied, the majority of control-injected nodose and trigeminal neurons survived for 3 days with their respective neurotrophins (Figs 1 and 2) and there was no significant difference in the survival of these control injected neurons compared with uninjected neurons (data not shown). These results show that the microinjection procedure does not affect the survival response of neurons to neurotrophins.

There were marked, age-related effects of expressing antisense Bcl-w on the survival of nodose and trigeminal neurons grown with their respective neurotrophins. In E16 nodose cultures, antisense Bcl-w had no effect on the survival of nodose neurons growing with BDNF throughout the 72 hour culture period. In E18 nodose cultures, antisense Bcl-w caused a small reduction in survival compared with control injected neurons that was evident at all time points studied. In P1 nodose cultures, antisense Bcl-w caused a marked reduction in survival that became more pronounced with time in culture, dropping by 66% in antisense Bcl-w-expressing neurons relative to control injected neurons 72 hours after injection. These results suggest that endogenously expressed Bcl-w is required for the survival of the majority of postnatal nodose neurons in the presence of BDNF.

Antisense Bcl-w had similar age-related effects on the survival response of trigeminal neurons to NGF. A decrease in survival of 19% relative to control injected neurons was already evident in E16 cultures 72 hours after injection, and became more pronounced in E18 cultures (31%) and P1 cultures (55%). These results likewise suggest that endogenously expressed Bcl-w is required for the survival of most postnatal trigeminal neurons in the presence of NGF.

Although Bcl-xL has previously been shown to play a key role in regulating the survival of developing sensory neurons in vitro and in vivo (Middleton et al., 1996; Motoyama et al., 1995), our demonstration of marked age-related changes in the requirement for Bcl-w in sensory neuron survival prompted us to investigate the relative importance of Bcl-x_L over the same period of development by injecting neurotrophin-deprived neurons with an expression plasmid containing Bcl-x_L cDNA in the sense orientation. Immunocytochemistry revealed that neurons injected with this plasmid were strongly stained for Bcl-x_L protein, confirming this protein was effectively expressed in injected neurons (data not shown).

We began by studying the effect of overexpressing Bcl-x_L on the survival of BDNF-deprived nodose neurons and NGF-deprived trigeminal neurons at E16, E18 and P1. Injection of an expression plasmid containing Bcl-x_L cDNA in the sense orientation rescued a substantial number of neurotrophin-deprived nodose and trigeminal neurons at all ages (Figs 3 and 4). However, in contrast to Bcl-w overexpression, there was a clear decrease in the ability of Bcl-x_L overexpression to prevent the death of neurotrophin-deprived neurons with increasing age. At E16, Bcl-x_L overexpression was more effective than Bcl-w overexpression in promoting the survival of both nodose neurons (82% survival in Bcl-x_L-injected neurons compared with 43% survival in Bcl-w-injected neurons after 72 hours incubation) and trigeminal neurons (80% survival in Bcl-x_L-injected neurons compared with 62% survival in Bcl-w-injected neurons after 72 hours incubation). By P1, Bcl-x_L was much less effective than Bcl-w overexpression in promoting the survival of both nodose neurons (53% survival in Bcl-x_L-injected neurons compared with 71% survival in Bcl-w-injected neurons after 72 hours incubation) and trigeminal neurons (45% survival in Bcl-x_L-injected neurons compared with 68% survival in Bcl-w-injected neurons after 72 hours incubation). These results show that the effectiveness with which Bcl-x_L overexpression is able to prevent the death of neurotrophin-deprived sensory neurons decreases with age during the late fetal and postnatal period.

Because overexpression of either Bcl-w or Bcl-x_L promotes the survival of most but not all nodose and trigeminal neurons in the absence of neurotrophins, we investigated whether overexpressing both proteins would further enhance the survival of these neurons. As shown in Fig. 5, co-injection of Bcl-w and Bcl-x_L plasmids into E18 nodose and trigeminal neurons promoted the survival of more neurons than injection of either plasmid alone. In the case of nodose neurons, 75% of the neurons overexpressing Bcl-w plus Bcl-x_L were surviving 72 hours after injection compared with 60% and 66% of neurons overexpressing Bcl-w and Bcl-x_L, respectively (compare with Figs 1, 3 and 5). In the case of trigeminal neurons, 79% of the neurons overexpressing Bcl-w plus Bcl-x_L were surviving 72 hours after injection compared with 70% and 61% of neurons overexpressing Bcl-w and Bcl-x_L, respectively (compare with Figs 2, 4 and 5).

To ascertain if there are any age-related changes in the physiological significance of endogenously expressed Bcl-x_L protein in regulating the survival of nodose and trigeminal neurons over the period of development during which Bcl-w becomes increasingly important for sustaining the survival of these neurons, we injected plasmids expressing anti-sense Bcl-x_L RNA to interfere with the synthesis of Bcl-x_L in E16 and P1 neurons. In contrast to antisense Bcl-w, antisense Bcl-x_L caused a marked decrease in the survival of nodose and trigeminal neurons growing with their respective neurotrophins at E16. There was a 70% reduction in the number of nodose neurons and 71% reduction in the number of trigeminal neurons 72 hours after injection with the antisense Bcl-x_L plasmid compared with control-injected neurons at this age (Figs 3 and 4). Also in contrast to antisense Bcl-w, antisense Bcl-xL was much less effective in killing P1 nodose and trigeminal neurons grown with their respective neurotrophins. There was only a 31% reduction in the survival of nodose neurons and a 27% reduction in the survival of trigeminal neurons in P1 compared with control injected neurons. These results indicate that endogenously expressed Bcl-x_L is required for sustaining the survival of many nodose and trigeminal neurons in the late fetal period but that it becomes much less important in this respect postnatally.

To determine if the developmental changes in the relative importance of Bcl-w and Bcl-x_L in sustaining sensory neuron survival are correlated with corresponding changes in the levels of expression of Bcl-w and Bcl-x_L, we studied the expression of Bcl-w and Bcl-x_L mRNA and protein in the nodose and trigeminal ganglia. Although the small size of sensory neurons prior to E16 precluded studying the roles of endogenously cultures 72 hours after injection of the antisense Bcl-x_L plasmid expressed Bcl-w and Bcl-x_L before this age, we extended our analysis of Bcl-w and Bcl-x_L expression back to E12. Competitive RT/PCR revealed a 2.5-fold increase in Bcl-w mRNA level in the nodose ganglion and a 6.5-fold increase in Bcl-w mRNA in the trigeminal ganglion between E12 and P1 (Fig. 5). In contrast, over this same period of development, there was an approximate 3-fold decrease in Bcl-x_L mRNA level in both nodose and trigeminal ganglia. Interestingly, although there was a large change in the ratio of the relative levels of Bcl-w and Bcl-x_L mRNAs during development, the overall level of Bcl-w mRNA was much higher than than of Bcl-x_L mRNA at all ages studied. However, this does not necessarily reflect differences in the relative levels of the corresponding proteins as there may be differences in translation efficiency and protein stability.

In accordance with the developmental increase in Bcl-w mRNA expression in nodose and trigeminal ganglia, no Bcl-w-positive neurons were observed in sections of these ganglia in E12 embryos and there was a clear increase in the overall intensity of Bcl-w staining in neurons with increasing age (Fig. 6). In addition to this general increase in Bcl-w immunoreactivity in developing neurons, a subset of strongly Bcl-w positive neurons became evident in both ganglia. The number of these neurons were easily and unambiguously quantified and showed a marked increase with age (Fig. 7). Control sections that were not incubated with primary polyclonal anti-Bcl-w were completely unstained (not shown). In contrast to the increase in Bcl-w immunoreactivity during development, the numbers of Bcl-x_L immunoreactive neurons peaked at E14 and progressively decreased at later ages. Although the total number of neurons in these ganglia also peaked at E14 and decreased at later ages as a consequence of naturally occurring neuronal death, the decrease in the number of Bcl-x_L immunoreactive neurons after E14 was more pronounced than the decrease in the total number of neurons. The proportion of Bcl-x_L immunoreactive neurons decreased from 97% and 98% in E12 nodose and trigeminal ganglia to 70% and 64% in these ganglia by P1. Control sections that were not incubated with primary polyclonal anti-Bcl-x_L were completely unstained (not shown). Taken together, these results demonstrate that the expression of Bcl-w and Bcl-x_L change in accordance with the changing importance of these proteins in sustaining the survival of sensory neurons with age.

We have provided the first evidence for a key role for Bcl-w in regulating neuronal survival during development. Overexpressing Bcl-w in late fetal and postnatal sensory neurons cultured from the mouse nodose and trigeminal ganglia promotes the survival of these neurons in the absence of neurotrophins and reducing Bcl-w expression using antisense RNA kills many of these neurons in the presence of neurotrophins that would otherwise support their survival. The effectiveness with which antisense Bcl-w antagonises the survival-promoting effects of neurotrophins increases from E16 to P1, and this accords with our demonstration that Bcl-w expression increases in these neurons over this period of development. These results indicate that Bcl-w is an anti-apoptotic factor for developing sensory neurons and that its expression is required for sustaining the survival of increasing numbers of sensory neurons in the late fetal and postnatal period. These results together with the recent demonstration of widespread and increasing expression of Bcl-w mRNA in the mouse brain between birth and maturity (Hamner et al., 1999) raise the possibility that Bcl-w may play an important role in sustaining the survival of many different kinds of neurons in the mature nervous system.

We have also shown that Bcl-x_L plays an important anti-apoptotic function for nodose and trigeminal ganglion neurons. Bcl-x_L overexpression promotes the survival of these neurons in the absence of neurotrophins and antisense Bcl-x_L kills these neurons in the presence of their required neurotrophins. However, in contrast to Bcl-w, the importance of endogenously expressed Bcl-x_L for sustaining the survival of these neurons decreases with age. The effectiveness with which antisense Bcl-x_L antagonises the survival-promoting effects of neurotrophins decreases from E16 to P1, and this accords with a decrease in Bcl-x_L expression in these neurons during development. Although previous studies have shown that overexpression of Bcl-x_L in developing sensory and sympathetic neurons is able to promote their survival in the absence of the required neurotrophins (Gonzalez-Garcia et al., 1995; Middleton et al., 1996), no detailed developmental studies have been undertaken to see if the role of Bcl-x_L changes during development. The demonstration that bcl-x^−/− mice die at E13 with massive apoptosis in the brain and sensory ganglia clearly shows that Bcl-x expression is required for the survival of many sensory neurons during the earliest stages of their development. However, because Bcl-x-deficient embryos die at this early stage of development it has not been possible to ascertain the importance of Bcl-x for regulating the survival of neurons at later stages of development. Our study shows, however, that Bcl-x continues to play an important role in regulating the survival of sensory neurons into the late fetal period, although its importance wanes in the postnatal period.

In addition to Bcl-w and Bcl-xL, Bcl2 also plays an important role in sustaining the survival of cranial sensory neurons during development. Antisense Bcl2 RNA substantially reduces the number of embryonic trigeminal mesencephalic neurons that survive in medium containing BDNF (Allsopp et al., 1995). Likewise, the survival of trigeminal and nodose neurons of bcl2^−/− embryos is reduced in the presence of their respective neurotrophins during the peak period of naturally occurring neuronal death, and this impaired in vitro response to neurotrophins is associated with 25% reduction in the number of neurons in the trigeminal and nodose ganglia of bcl2^−/− embryos by the end of the period of naturally occurring neuronal death (Piñón et al., 1997). The sustained expression of Bcl2 in sensory neurons throughout life (Merry et al., 1994) appears to play an ongoing role in maintaining the survival of many sensory neurons as increasing numbers of dorsal root ganglion neurons in Bcl2-deficient mice die after the postnatal period (Michaelidis et al., 1996).

In summary, our study together with previous work on Bcl2 has shown that several different Bcl2-related proteins are required for the survival of developing sensory neurons. Each of these proteins acts preferentially during a particular stage of development which is correlated with its level of expression. Ascertaining how the expression of these proteins is regulated during development will be important for understanding how the appropriate number of neurons is controlled in sensory ganglia and other populations of neurons.

We thank Gene Burton of Genentech Inc. for the purified recombinant NGF and BDNF and Gabriel Nunez for the Bcl-xL cDNA. This work was supported by grants from the Cancer Research Campaign and Wellcome Trust.