Cholecystokinin-8 protects central cholinergic neurons against fimbria–fornix lesion through the up-regulation of nerve growth factor synthesis

Abstract

In this study, we demonstrate that cholecystokinin-8 (CCK-8) induces an increase in both nerve growth factor (NGF) protein and NGF mRNA in mouse cortex and hippocampus when i.p. injected at physiological doses. By using fimbria–fornix-lesioned mice, we have also demonstrated that repeated CCK-8 i.p. injections result in recovery of lesion-induced NGF deficit in septum and restore the baseline NGF levels in hippocampus and cortex. Parallel to the effects on NGF, CCK-8 increases choline acetyltransferase (Chat) activity in forebrain when injected in unlesioned mice and counteract the septo-hippocampal Chat alterations in fimbria–fornix-lesioned mice. To assess the NGF involvement in the mechanism by which CCK-8 induces brain Chat, NGF antibody was administrated intracerebrally to saline- and CCK-8-injected mice. We observe that pretreatment with NGF antibody causes a marked reduction of NGF and Chat activity in septum and hippocampus of both saline- and CCK-8-injected mice. This evidence indicates that the CCK-8 effects on cholinergic cells are mediated through the synthesis and release of NGF. Taken together, our results suggest that peripheral administration of CCK-8 may represent a potential experimental model for investigating the effects of endogenous NGF up-regulation on diseases associated with altered brain cholinergic functions.

Nerve growth factor (NGF) is the founder member of a structurally related neurotrophic factor family termed neurotrophins that includes brain-derived neurotrophic factor and NT 3–4 (1–3). One of the major NGF effects is to promote the growth and differentiation of basal forebrain cholinergic neurons (BFCN; refs. 2 and 3) through the interaction with two transmembrane glycoprotein receptors: the high-affinity TrkA receptor and the low-affinity p75 receptor (for details, see refs. 1 and 4). NGF and its receptors are widely distributed in the brain, where the local NGF biosynthesis has also been demonstrated by the presence of a transcript NGF RNA messenger (5). A great deal of evidence indicates that the decreased availability of endogenous NGF or failure of NGF interaction with its receptive cells results in severe neurological (6) and behavioral deficits (see refs. 7 and 8). Low NGF availability is also known to cause a decrease of choline acetyltransferase (Chat) activity in rodent forebrain neurons, whereas intracerebral NGF injection prevents BFCN atrophy after fimbria–fornix transection (8). Because BFCN plays a crucial role in learning and memory, and their dysfunction contributes to age-related disorders and the development of dementia (9, 10), it has been hypothesized that NGF administration could be therapeutically beneficial in reducing the BFCN degeneration occurring in Alzheimer’s disease (AD; refs. 11 and 12). Indeed, clinical studies have shown that intracerebroventricular NGF injection in patients affected by AD promotes the increase of blood flow, nicotine bindings, and improvements of electroencephalograms (13). Despite these findings, the possibility of clinical use of NGF is limited by its inability to cross the blood–brain barrier and hence by the invasive neurosurgical procedures—such as NGF infusion into the ventricular space and brain parenchyma (10, 12, 14) or cerebral graft of neurotrophin-producing cells (15)—that are needed to deliver NGF into the brain. Therefore, substances with low molecular mass that can access the brain and stimulate the local NGF synthesis, when systemically administered, may be useful to devise methodological strategies for BFCN-associated disorders.

We have recently demonstrated that cholecystokinin-8 (CCK-8), a small neuropeptide widely distributed in both central and peripheral nervous systems (16, 17), affects brain NGF levels in a dose- and time-dependent manner (18). For instance, a physiological dose of CCK-8 (8 nmol⋅kg⁻¹) increased the NGF levels in mouse hippocampus (+60%) when injected i.p. in adult male mice. In the present study, we report that the same CCK-8 dose causes the activation of brain NGF biosynthesis and up-regulates the expression of cholinergic markers in the forebrain of normal and fimbria–fornix-transected mice.

MATERIALS AND METHODS

Animals and Treatment.

Adult male CD-1 mice weighing 35–40 g (Charles River Breeding Laboratories) were used in these studies. Control, false-operated (see below), and brain-lesioned mice received i.p. daily injection (0.25 ml) with saline or 8 nmol⋅kg⁻¹ CCK-8 (Peninsula Laboratories) for 3 consecutive days and were sacrificed 1 day after the last treatment. After sacrifice, the brains were removed and dissected for biochemical and molecular analysis.

Surgical Procedures.

The fimbria–fornix lesion was produced in anesthetized mice mounted in a stereotaxic apparatus. A knife (0.7 mm wide) was introduced through a small slit in the skull 4 mm below the dura into the brain 1.8 mm lateral to the midline (bregma −1.08 mm) and was then moved 0.5 mm in both the anterior and posterior direction to produce a lesion from bregma −0.58 to bregma −1.58 (coordinated according to the atlas by Franklin and Paxinos (19). A group of anesthetized mice in which only the small slit in the skull was made (false-operated mice) was used as a further control.

Chat Immunohistochemistry.

Free-floating immunohistochemistry was performed on 20-μm cryostat sections from perfused brains (4% paraformaldehyde in PBS, pH 7.2). Fixed sections were treated for 1 h with a solution containing 0.1% phenylidrazine, BSA (1 mg/ml), and 5% goat serum, to block the endogenous peroxidase. Chat was localized by using mouse mAbs against brain Chat (Chat mAb17 kindly provided by Costantino Cozzari, Institute of Cellular Biology, CNR, Italy). The sections were incubated overnight with the mentioned antibodies and processed for immunoperoxidase with the ABC Vectastain kit (Vector Laboratories) following the manufacturer’s instructions. Staining specificity was assessed by omission of the primary antibody.

In Situ Hybridization for NGF mRNA.

Fourteen-micrometer sections from fresh brain were cut by cryostat and mounted on poly-l-lysine-coated slides. The slices were fixed in 4% paraformaldeyde in 0.1 M PBS (pH 7.4) for 10 min followed by repeated wash in 0.1 M PBS and dehydration by 70, 80, and 95% ethyl alcohol. After acetylation (25% acetic anydride in 0.1 M triethanolamine, pH 8.0), the slices were incubated at 42°C for 16 h in a hybridization mixture containing digoxigenin-labeled NGF probes (complementary to the sequences 5′-TCCTGTTGAGAGTGGTGCCGGGGCATCGA-3′) at a final concentration of 30 ng/ml hybridization buffer (50% formamide, 2× SSC, 0.1%SDS, 250 μg/ml denatured sheared salmon testes DNA). After washing, the slices were incubated for 2 h at room temperature with an 1.5 units/ml sheep anti-digoxigenin POD-conjugated antibody (polyclonal Fab fragment; Boehringer Mannheim). The immunoperoxidase reaction was detected by using standard diaminobenzidine (DAB) procedure (0.6 mg/ml DAB/0.015% H₂O).

NGF and Chat Measurement.

The dissected brain tissues were sonicated in extraction buffer (0.1 M Tris⋅HCl, pH 7.00/400 mM NaCl/0.1% Triton X-100/0.05% NaN₃/2% BSA/0.5% gelatin/4 mM EDTA/40 units/ml aprotinin/0.2 mM PMSF/0.2 mM benzetonium chloride/2 mM benzamidine) followed by centrifugation at 30,000 × g for 30 min. The supernatants were used for NGF and Chat activity assay.

NGF levels were measured by using a two-site immunoenzymatic assay (ELISA) as described (20). Briefly, polystyrene 96-well microtiter immunoplates (Nunc) were coated with a monoclonal mouse anti-NGF (Boehringer Mannheim) and mouse IgG for the evaluation of nonspecific signals. Purified NGF 2.5S dissolved in extraction buffer (see above) was used as standard in a range of 0.015 pg/ml–1 ng/ml. Fifty microliters of standard and samples were added in triplicate to the antibody-coated plates and incubated overnight at room temperature. The NGF content in the samples was determined in relation to a NGF standard curve. Data were not corrected for recovery of NGF from samples, which was routinely 70–90%, and was accepted only when the values were >2 SD above the blank. By using these criteria, the limit of sensitivity of NGF ELISA averaged 0.5 pg per assay. The results are presented as mean ± SD.

The Chat activity was determined following the method of Fonnum (21). The Chat levels in the different brain regions were measured as units/mg tissue, and one unit is defined as a micromole of acetylcholine formed per minute at 37°C. The Kruskal–Wallis nonparametric analysis of variation with multiple comparisons was used for significance testing. P values <0.05 were considered significant.

RESULTS

In Vivo Effects of CCK-8 on NGF Production.

In the first experiment, saline or 8 nmol⋅kg⁻¹ of CCK-8—a dose that has the maximum stimulatory effects on NGF (18)—were i.p. injected daily in adult male mice for 3 consecutive days. The NGF levels in the hippocampus and cortex were analyzed by using ELISA as described in Materials and Methods. Consistent with our previous data, we observed that CCK-8 increased brain NGF levels when peripherally injected. In particular, the repeated injections with CCK-8 resulted in an increase of ≈50% and 25% in NGF levels in the hippocampus and cortex, respectively (see Table 1). Elevated NGF levels also were observed in the septum of CCK-8-injected mice.

Table 1.

Effects of CCK-8 injections on brain NGF levels

Condition	Hippocampus	Cortex	Septum
Saline	5,012 ± 657	1,726 ± 456	4,426 ± 431
CCK-8	7,330 ± 950a,d	2,155 ± 559a,c	5,804 ± 835
FF-lesioned + saline	4,458 ± 894	1,628 ± 429	2,924 ± 418a
FF-lesioned + CCK-8	6,554 ± 767c	2,404 ± 466a,c	9,045 ± 1,746b,d

All values are given in pg/gr tissue and expressed as mean ± SD. Kruskal–Wallis nonparametric test with multiple comparison was used to evaluate difference between groups. No changes in brain NGF were noted in the false-operated mice (see Materials and Methods) compared to controls and were thus omitted from the table. FF, fimbria–fornix. 

^*, P <0.01; ∗∗, P <0.001 (vs. saline). 

^†, P <0.01;  

^‡, P <0.001 (vs. FF-lesioned + saline). 

To assess whether the CCK-induced NGF increase was caused by a CCK action on NGF synthesis, in situ hybridization was carried out on brain sections from saline- or CCK-8-injected mice. The effects of CCK-8 on NGF mRNA expression were noted in both the hippocampal and cortical regions. As shown in Fig. 1, in the hippocampus of saline-injected mice, the specific NGF mRNA labeling was mainly localized in the dentate gyrus and CA1 field (see Fig. 1A), whereas in the hippocampal formation of the CCK-injected mice, NGF mRNA also was expressed in the CA2 and CA3 fields (Fig. 1B). In cortex, NGF mRNA labeling was detected in both the parietal and piriform areas. CCK-8-injected mice show a marked increase of the NGF mRNA expression over a cell subpopulation in the piriform cortex, although lower effects were noted also in the parietal area (data not shown).

Figure 1.

NGF mRNA detected by using in situ hybridization histochemistry in the hippocampus of adult mice receiving saline (a) or CCK-8 (b) injections. Note the increase of NGF mRNA labeling in the CA3 hippocampal area of the CCK-8-treated mice. Specific NGF mRNA expression was confirmed by specificity test including digestion of mRNA with RNase A before hybridization and hybridization with sense NGF probe, which resulted in the absence of hybridization signals (c). (Magnification ×15.)

Effects of CCK-8 Injection on the NGF and Chat Expression in Brain of Fimbria–Fornix-Lesioned Mice.

To evaluate the potential functional role of the CCK-8-induced brain NGF enhancement, the effects of CCK-8 on Chat activity and distribution in the forebrain of fimbria–fornix-transected mice were investigated. As demonstrated in Figs. 2 and 3, i.p. injections with CCK-8 promoted recovery of the lesion-induced cholinergic marker decrease. Compared with lesioned mice injected with saline (Fig. 2A), the mice receiving CCK-8 showed a more intense Chat immunoreactivity in the septum and Broca’s band (Fig. 2B). Lower but clearly significant effects of CCK-8 also were observed on the Chat expression in the magnocellular basal nucleus (data not shown).

Figure 2.

(a–d) Immunohistochemical localization of Chat-positive neurons in the septum and Broca’s bands of fimbria–fornix-lesioned mice receiving saline (a) or CCK-8 (b) injections. (×50). c and d show the high magnification (×150) of septal Chat-immunopositive neurons of fimbria–fornix-lesioned mice injected with saline or CCK-8, respectively.

Figure 3.

Effects of fimbria–fornix lesion, Ab-NGF intracerebroventricular treatment, and/or CCK-8 injections on septal and hippocampal Chat activity. The bars represent the percent difference ± SSE with respect to the Chat values detected in the septum and hippocampus of the control mice (=100%). No significant changes were noted in the Chat activity in false-operated mice and after intracerebroventricular injection with saline (data not shown). ∗∗, P < 0.01; ∗∗∗, P < 0.001 vs. control. ⦵⦵, P < 0.01; ⦵⦵⦵, P < 0.001 vs. both control and fimbria–fornix-lesioned mice receiving saline.

In agreement with the morphological evidence, the Chat activity in the septum and hippocampus of fimbria–fornix-transected mice injected with CCK-8 were significantly increased compared with those receiving saline injections and similar to that measured in unlesioned mice (see Fig. 3). Highly significant enhancements of Chat activity were observed in the septum and hippocampus of unlesioned mice injected with CCK-8 (Fig. 3).

CCK-8 injections also affected the NGF levels in the septum and hippocampus of fimbria–fornix-transected mice (see Table 1).

Evidence for NGF-Mediated CCK Effects on Chat.

It has been reported that intracerebral administration of NGF antibody (Ab-NGF) alters the brain NGF levels and Chat activity and results in impaired behavioral performance in rodents (7). Thus, to assess whether the observed effects on Chat expression were caused by a direct CCK-8 action on cholinergic neurons, we have analyzed the effects of CCK-8 in mice receiving Ab-NGF. In this experiment, 30 μg of Ab-NGF were stereotaxically injected intracerebroventricular in adult male mice before injections of saline or CCK-8. As shown in Table 1, Ab-NGF administration caused a marked reduction of NGF in the hippocampus (Ab-NGF + saline = 3,171 ± 372; Ab-NGF + CCK-8 = 3,201 ± 400 pg/gr tissue) and septum (Ab-NGF + saline = 3,020 ± 379; Ab-NGF + CCK-8 = 2,112 ± 368 pg/gr tissue) of both the saline and the CCK-8 injected mice. Moreover, in both the saline- and CCK-8-injected mice pretreated with Ab-NGF, the NGF reduction was associated with decreased Chat activity in the septum and hippocampus (see Fig. 3).

DISCUSSION

Consistent with our previous observations, this study shows that CCK-8 is a potent inducer of brain NGF when administered i.p. in doses close to the physiological circulating CCK levels (22). In particular, the evidence that repeated i.p. injections with 8 nmol⋅kg⁻¹ CCK-8 give a parallel increase of NGF protein and its mRNA in the hippocampus and cortex demonstrates that CCK-8 affects brain NGF content by acting at transcriptional levels in the two regions that represent the major sources of NGF in brain (23). Compared with the effects of the single CCK-8 injection, which resulted in a transient increase of NGF concentration in the hippocampus and no effects in cortex (18), repeated CCK injections resulted in an increase of NGF levels in both the hippocampus and cortex that persists for 1 day after the last CCK injection. This indicates that long-lasting effects can be obtained by increasing the length of CCK treatment.

The possibility of controlling NGF production and/or release into the brain by changing the dosage and/or length of CCK treatment is particularly relevant considering the inability of NGF to cross the blood–brain barrier and the invasive neurosurgical procedures necessary for its delivery into the brain. Recently, Friden et al. (24) devised a method that utilizes NGF conjugated to transferrin receptor antibodies to permit NGF to cross the blood–brain barrier. Although this methodology appears useful in delivering systemically injected NGF into the brain, the relatively small amount of NGF that reaches the central nervous system limits its validity.

Likewise, the use of substances such as IL-1 (25), tumor necrosis factor-α (26), steroids (27), and glucocorticoids (28) to up-regulate the endogenous brain NGF may cause indesiderable side effects, thus rendering this type of strategy impracticable.

Although CCK-8 is known to affect behavioral functions (29), studies aimed at elucidating the effects of CCK-8 administration on food intake in animals indicate that CCK-8 effects are highly dose-dependent and are subject to tolerance resulting in unchanged food intake when CCK-8 is administered in the long term (30). Similarly, human studies have shown that CCK-8 or its analogues are ineffective when given in doses that elicited plasma CCK concentration in the physiological dose range (31). These observations suggest that, although the effects of a prolonged treatment with CCK-8 remain to be elicited, i.p. injections with low doses of CCK-8 can represent a reliable alternative to the described invasive neurosurgical procedure to deliver NGF into the brain (14, 15, 24) and to investigate the effects of brain NGF up-regulation on the BFCN-associated pathologies.

In this context, it is interesting that parallel to the effects seen on NGF biosynthesis, CCK-8 increases Chat activity in forebrain when injected in unlesioned mice and counteracts the septo-hippocampal Chat alterations in fimbria–fornix-lesioned mice. The effects of CCK-8 on brain cholinergic cells have also been described. For example, administration of CCK-8 or its analogues increases the release of acetylcholine in rodents’ brain (32) and can preserve the Chat activity in the cortex of forebrain-lesioned rats (33, 34). Despite this evidence, the mechanism(s) by which CCK-8 affects the central cholinergic neuronal population is still not understood.

Recently it has been shown that i.p. CCK-8 administration in adult rats results in enhanced brain expression of NGF-induced genes, including NGFI-A (35). Similar to what we have observed for NGF (18), the effects of CCK-8 on NGFI-A can be blocked by selective CCK receptor antagonists and by vagotomy. NGFI-A induction is involved in the NGF-mediated trans-activation of the Chat gene in neuronal cells and results in a dramatic increase of Chat activity (36). This evidence, together with our observations that CCK-8 effects on Chat activity can be blocked by the administration of NGF antibodies and that the parallel NGF and Chat alterations after fimbria–fornix lesion can be prevented by CCK-8 injections indicates that the CCK-8 effects are mediated through the synthesis and release of NGF.

One relevant implication of our findings is that the administration of CCK-8 may be useful in mitigating the cognitive disorders associated with loss of NGF and the degenerative events in BFCN. In particular, the impairment of cognitive functions in AD patients have been significantly correlated with deficits of brain Chat, acetylcholine release, and neurotrophin expression, including NGF (8, 10). In AD therapy, many trials with the purpose of increasing cholinergic transmission by synthetic cholinergic agonists and acetylcholinesterase inhibitors have failed to result in any dramatic changes to the memory impairment of AD patients. This failure may be partly because the cholinergic system is so severely damaged in AD brain and because there is no ability to respond to cholinomimetic drugs. The CCK system still survives even in advanced AD stages (37); therefore, CCK-8 might be effective. Moreover, the evidence that both NGF (8, 10) and CCK-8 (29, 38, 39) have been shown to improve memory when administrated in rodents and humans renders the possibility of CCK-8 use in AD therapy highly attractive.

In conclusion, our study presents clear evidence that brain NGF biosynthesis can be modulated by peripheral administration of CCK-8 and suggests a potential experimental model to investigate the role of endogenous up-regulation of NGF (and possibly the other neurotrophic factors) on the diseases associated with altered BFCN functions.

ABBREVIATIONS

CCK-8

cholecystokinin-8

NGF

nerve growth factor

BFCN

basal forebrain cholinergic neuron

AD

Alzheimer’s disease

Chat

choline acetyltransferase