Selective glial cell line-derived neurotrophic factor production in adult dopaminergic carotid body cells in situ and after intrastriatal transplantation

Abstract

Glial cell line-derived neurotrophic factor (GDNF) exerts a notable protective effect on dopaminergic neurons in rodent and primate models of Parkinson's disease (PD). The clinical applicability of this therapy is, however, hampered by the need of a durable and stable GDNF source allowing the safe and continuous delivery of the trophic factor into the brain parenchyma. Intrastriatal carotid body (CB) autografting is a neuroprotective therapy potentially useful in PD. It induces long-term recovery of parkinsonian animals through a trophic effect on nigrostriatal neurons and causes amelioration of symptoms in some PD patients. Moreover, the adult rodent CB has been shown to express GDNF. Here we show, using heterozygous GDNF/lacZ knock-out mice, that unexpectedly CB dopaminergic glomus, or type I, cells are the source of CB GDNF. Among the neural or paraneural cells tested, glomus cells are those that synthesize and release the highest amount of GDNF in the adult rodent (as measured by standard and in situ ELISA). Furthermore, GDNF expression by glomus cells is maintained after intrastriatal grafting and in CB of aged and parkinsonian 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-treated animals. Thus, glomus cells appear to be prototypical abundant sources of GDNF, ideally suited to be used as biological pumps for the endogenous delivery of trophic factors in PD and other neurodegenerative diseases.

Figure 1.

Selective GDNF expression in adult carotid body. A, GDNF expression in the carotid body (blue-green dots) evidenced by X-gal staining in heterozygous GDNF/lacZ mice (17.5 months old). The inset shows the typical glomerular structure of the CB after TH immunohistological staining. B, CB section from a wild-type animal with complete absence of X-gal labeling after staining (8.5 months of age). The inset shows again the typical glomerular structure of the same CB. C-H, Lack of GDNF expression in the adrenal medulla (C, D; 5 months of age), superior cervical ganglion (E, F; 5 months of age), and retina (G, H; 7 months of age) of heterozygous GDNF/lacZ mice after X-gal staining. To confirm the stability of the tissues, the insets in C, E, and G show, respectively, the characteristic TH-positive immunostaining of adrenal medulla chromaffin cells, superior cervical ganglion neurons, and retinal amacrine cells of the same mice.

Figure 2.

GDNF protein content and release in CB compared with other TH-positive preparations. A, GDNF levels (in picograms per milligram) measured by standard ELISA in rat tissues. The number of separate measurements was as follows: six CBs, six SCGs, eight AMs, and three Zuckerkandl's organs (Zuck). B, GDNF release (picograms per 10⁴ cells in 24 h) measured by the ELISA in situ assay from dispersed carotid body, adrenal medulla, and PC12 cells. Data are from six CB, four AM, and four PC12 different cultures in three independent experiments. The statistical significance of differences among parameters was considered at ^*p < 0.05 (see Materials and Methods).

Figure 3.

GDNF expression in CB glomus cells. A-C, GDNF expression (blue precipitate, A) in a typical dispersed CB glomus cell that was TH positive (B, red fluorescence) but GFAP negative (C). D-F, Lack of GDNF (D) and TH (E) expression in a representative dispersed subtentacular CB cell. Green fluorescence in F indicates that the cell was GFAP positive. The cells are examples obtained from primary cultured CB cells removed from heterozygous GDNF/lacZ mice. Cultures were subjected to X-gal staining and simultaneous TH and GFAP immunofluorescent detection. G-I, Ultrastructural analysis of a CB from a 6-month-old heterozygous GDNF/lacZ mouse after β-galactosidase reaction. The blue asterisks (H, I) indicate the specific localization of the X-gal deposits within glomus cells (see the correspondence with the blue staining indicated by arrowheads in the semithin section in G). CB glomus cells are clearly identified by the characteristic large nucleus with fragmented chromatin (G-I).

Figure 4.

GDNF expression in intrastriatally grafted CB glomus cells. A, GDNF expression (blue X-gal staining) in a 1 month intrastriatal CB transplant (from a 2-month-old heterozygous GDNF/lacZ mouse) grafted into a 4-month-old wild-type mouse. B, C, Comparison of GDNF (green dots) and TH (light brown) expression and general phenotypic appearance of an intrastriatal CB graft (B; 1.5 months after transplantation) and a CB in situ (C; removed from a 5-month-old animal). In both cases, the tissues belong to heterozygous GDNF/lacZ mice. D-F, Ultrastructural analysis of the same CB graft shown in A. The arrowhead in the semithin section (D) and the pictures (E, F) indicates the correspondent X-gal deposit, showing the GDNF expression in grafted CB glomus cells. G, Ultrastructural analysis of a 15-d-old graft showing a typical CB glomus cell with some dense-core vesicles that are dispersed throughout the cytoplasm. CB glomus cells in D-G are clearly identified by the characteristic large nucleus with fragmented chromatin.

Figure 5.

Absence of endogenous striatal GDNF induction by CB grafting or nonspecific brain damage. A, B, Heterozygous GDNF/lacZ mice striatal sections ipsilateral (A) and contralateral (B) to a CB graft from a 1-month-old wild-type donor. Arrowheads indicate the presence of X-gal deposits (GDNF expression) in the host tissue, which was similar in the two sides of the brain. C, Heterozygous GDNF/lacZ mice striatal section of a sham-operated animal. Note the absence of X-gal deposits (indicating the lack of GDNF expression) within the bright yellowish macrophages (arrows) along the needle tract. However, X-gal deposits are clearly seen in the neighboring parenchyma. TH immunohistochemistry was performed in all sections after the X-gal staining.

Figure 6.

Maintenance of CB GDNF expression in aged and parkinsonian mice. A, Increase of GDNF expression (ordinate; estimated by the number of X-gal deposits per CB) in old heterozygous GDNF/lacZ mice (13-17.5 months of age) compared with young adults (2-4.5 months of age). B, Similar data as in A but normalized to the CB volume. C, Increase of CB GDNF expression in chronic MPTP-treated mice compared with saline-treated littermates. In the same animals, the application of MPTP induced a clear reduction in the number of TH-positive nigral neurons (D). The number of experiments is indicated in parentheses. The statistical significance of differences among parameters was considered at ^*p < 0.05 (see Materials and Methods).