Abstract
Astrocytes produce laminin, a key extracellular matrix guidance molecule in the developing brain. Laminin is bound to transmembrane receptors on the surface of astrocytes known as integrins, which are, in turn, bound to the microfilament meshwork inside the astrocyte. Previous studies have shown that T4 regulates the pattern of integrin distribution in astrocytes by modulating the organization of the microfilaments. In this study, the effect of thyroid hormone on the secretion and topology of laminin in astrocytes was examined. Linear arrays of secreted laminin were observed on the surface of the T4-treated astrocytes within 10 h after seeding the cells onto poly-d-lysine-coated coverslips and became an organized meshwork by 24 h. In contrast, little if any laminin was identified on the surface of either hormone-deficient or T3-treated cells until 36 h after seeding and then was restricted to punctate deposits. Secretion of laminin into the medium by hormone-deficient and T3-treated cells was significantly greater than that by T4-treated cells. Conversely, deposition of laminin into the extracellular matrix was significantly greater in T4-treated cells than in hormone-deficient and T3-treated cells. Thyroid hormone had no effect on the production of laminin by astrocytes. These data show that T4 regulates the extracellular deposition and organization of laminin on the surface of astrocytes and provide a mechanism by which this morphogenic hormone can influence neuronal migration and axonal projection in the developing brain.
THE CRETINOUS brain is characterized by severe morphological alterations that result from disturbed neuronal migration and deranged axonal projections, leading to attenuation of the number of neural circuits (1–3). For example, cerebellar granule cell death is markedly increased in the hypothyroid rat due to a failure of these neurons to complete migration and to form synapses with target cells (4–8). Axonal arborization and dendritic formation by the Purkinje cells in the cerebellum are also markedly decreased in the hypothyroid rat (3, 9), resulting in decreased synaptogenesis (10). Thyroid hormone supplementation during the first 2 weeks of life prevents the development of these disturbed morphological changes in the developing rat. Despite the identification of thyroid hormone as an essential regulatory factor in the brain developmental program, the biochemical and molecular events affected by this morphogenic hormone on neuronal integration remain elusive.
The formation of neural networks in the developing brain is accomplished by the migration of the neuronal growth cone down specific, preprogrammed pathways to its target region, followed by the projection of axons toward its target cell (11–13). Signals derived from the extracellular matrix (ECM) are essential to the guidance of the migrating neurite to its target cell (11–15). Of particular interest is the ECM protein laminin, which plays a key role in neuronal migration, synapse formation, and cell survival (11, 12, 14, 15).
Laminin is synthesized and secreted by astrocytes, both in vivo (16–20) and in vitro (21–25). Laminin is deposited into the ECM and fixed on the cell surface through binding to specific transmembrane receptors known as integrins (26–28). The regionalization of laminin on the astrocyte surface is determined by the clustering of integrins bound to the microfilaments into macromolecular complexes known as focal contacts (14, 29, 30). It is the organization of laminin into specific patterns on the cell surface that provides directional cues to the elongating neurite (11, 12, 15). Indeed, in vitro studies have shown that neurons readily and preferentially migrate onto laminin-coated surfaces (15, 31–34).
The appearance of laminin in the brain paranchyma is developmentally regulated and coincides with neuronal migration (11, 13, 15, 17, 18). Once the wiring network of the brain is established, laminin disappears from the brain paranchyma and is restricted to the basal lamina of the vasculature. We have shown that laminin is differentially expressed in the euthyroid and hypothyroid rat cerebellum (35). In the euthyroid rat cerebellum, laminin steadily increased from birth to postnatal days 8–10 and was concentrated in the molecular layer, through which the granular neurons must migrate to complete the cerebellum neuronal circuitry. By postnatal day 14, the granular neurons had all reached the inner granular layer, and laminin was restricted to the vasculature in the euthyroid rat. In the hypothyroid rat, laminin did not appear in the molecular layer until postnatal day 10 and disappeared by postnatal day 18 despite the presence of granular neurons that had yet to complete their migration.
In this study, we examined possible mechanisms for the in vivo differential expression of laminin by determining the effect of thyroid hormone on the expression and extracellular distribution of laminin in astrocytes in vitro. We show that laminin is rapidly synthesized and secreted in astrocytes independent of the presence of iodothyronines. In T4-treated astrocytes, laminin is bound to the astrocyte surface after secretion and is organized on the cell surface in discrete linear arrays. In T4-deficient and T3-treated astrocytes, laminin is primarily released into the medium after secretion; the small amount that remains bound to the astrocyte surface does so in a diffuse, disorganized pattern. Thus, the T4-dependent regulation of the extracellular distribution of laminin on the astrocyte surface suggests a mechanism by which this morphogenic hormone can influence neuronal migration in the developing brain.
Materials and Methods
Materials
Pregnant (16–17 days gestation) rats were obtained from Charles River Laboratories, Inc. (Kingston, NY). T4, EHS mouse sarcoma laminin, rabbit affinity-purified polyclonal antilaminin IgG, and BSA were purchased from Sigma Chemical Co. (St. Louis, MO). T3 and rT3 were obtained from Henning GmBH (Berlin, Germany). Antirabbit IgG-horseradish peroxidase conjugate was purchased from Promega Corp. (Madison, WI), and antirabbit IgG-Texas Red conjugate and Hybond ECL nitrocellulose were obtained from Amersham Pharmacia Biotech (Arlington Heights, IL). The Lumiglo chemiluminescent kit and the TMB Peroxidase Development kit were obtained from Kirkegaard & Perry Laboratories (Gaithersburg, MD), and poly-d-lysine was obtained from Biomedical Technologies (Stoughton, MA). Immobilized rProtein A beads were obtained from Repligen (Cambridge, MA). DMEM, antibiotics, Hanks’ solution, and 0.25% (wt/vol) trypsin were obtained from Life Technologies, Inc. (Gaithersburg, MD), and defined bovine calf serum (heat inactivated) was purchased from HyClone Laboratories, Inc. (Logan, UT). Culture flasks were obtained from Nunc (Copenhagen, Denmark), and 6- and 96-well tissue culture plates were obtained from Falcon (Lincoln Park, NJ). All other reagents used were of the highest purity commercially available.
Culture conditions
Rat type I astrocyte cultures were obtained by enzymatic dispersion of neonatal rat brains (36). Cells were grown in a humidified atmosphere of 5% CO2 and 95% air in 37 C in DMEM, 15 mm sodium bicarbonate, 33 mm glucose, 1 mm sodium pyruvate, and 15 mm HEPES (pH 7.4) with 10% (vol/vol) defined bovine calf serum. Culture medium was changed three times weekly, and cells were subcultured (2–3 × 104 cells/cm2) when they reached confluence (7–10 days) (37). Confluent cells from passages 1–3 containing more than 95% astrocytes were used for experiments after incubation with defined medium containing DMEM, 15 mm sodium bicarbonate, 33 mm glucose, 1 mm sodium pyruvate, 15 mm HEPES (pH 7.4), and 0.1% (wt/vol) BSA in the presence and absence of 10 nm iodothyronines.
Immunocytochemistry
Confluent cultures of rat astrocytes were grown for 16 h in defined medium in the presence and absence of 10 nm iodothyronines, which achieves a free hormone concentration of about 60 pm and results in the maximal T4-mediated effect on regulated processes in astrocytes, including actin polymerization (38, 39) and type II iodothyronine 5′-deiodinase activity (39–42). Cells were collected by trypsinization, a monocellular suspension was made by filtration through a 20-μm pore size mesh, and cells were seeded onto glass coverslips (22 × 22 mm) coated with poly-d-lysine (10 μg/ml), an ionic glue that allows attachment of cells independent of cell surface receptors. After 3–36 h, cells were fixed to the coverslips with 4% paraformaldehyde. Total cell-associated laminin was visualized by permeabilizing the cells with iced methanol. Extracellular/secreted proteins were identified by restricting access of the antisera to the interior of the cell by staining unpermeabilized cells. Nonspecific binding sites were blocked by incubation with BSA blocker (2 mg/ml BSA in PBS) for 30 min at room temperature. Cells were then incubated with antilaminin IgG (1:500 dilution) in BSA blocker for 1 h, and immune complexes were visualized by incubation with an antirabbit IgG conjugated with Texas Red. A Carl Zeiss Axioskop microscope (New York, NY) equipped with an Olympus Corp. OM4 camera (New Hyde Park, NY) and Kodak TMAX ASA 400 film (Eastman Kodak Co., Rochester, NY) was used for image acquisition. Coverslips were initially examined in a blinded fashion to ensure unbiased evaluation of the immunocytochemistry results.
Western analysis: secretion of laminin into the medium
Confluent cultures of rat astrocytes were grown for 16 h in defined medium in the presence and absence of 10 nm iodothyronines and collected by trypsinization, and a monocellular suspension was made by filtration through a 20-μm pore size mesh. Cells were seeded out onto six-well tissue culture plates (∼2 × 105 cells/well). Medium was collected from 3–24 h after seeding. Secreted proteins in the medium were denatured in a boiling water bath for 5 min, then applied to nitrocellulose via slot blot under vacuum. Blots were blocked for 1 h with milk blocker [20 mm Tris-HCl, 0.1% (vol/vol) Tween-20, 5 g/ml powdered milk, and 137 mm NaCl, pH 7.5], then probed for 1 h at room temperature with antilaminin IgG (1:1000 dilution) in milk blocker. Immune complexes were visualized by incubation with an antirabbit IgG conjugated to horseradish peroxidase (1:2500 in blocker) and developed with the Lumiglo chemiluminescent kit. Blots were analyzed by scanning densitometry, and laminin was quantified by comparison with results obtained with a standard curve that was run on each blot.
Enzyme-linked immunosorbent assay (ELISA) analysis: deposition of laminin into the ECM
Confluent cultures of rat astrocytes were grown for 16 h in defined medium in the presence and absence of 10 nm iodothyronines and collected by trypsinization, and a monocellular suspension was made by filtration through a 20-μm pore size mesh. Wells in 96-well ELISA plates were coated with poly-d-lysine (10 μg/ml). Astrocytes were seeded into the wells (∼1 × 106 cells/well) and grown for 24 or 36 h. Medium was then aspirated, and cells were burst by incubation with distilled water and washed with PBS with 0.5% (vol/vol) Tween-20. As shown by others, the remaining proteins attached to the wells after hypotonic disruption represent the astrocyte-derived ECM (43). Nonspecific binding sites were blocked by incubation with BSA blocker (2 mg/ml BSA in PBS) for 1 h at room temperature followed by incubation with antilaminin IgG (1:100 dilution) in BSA blocker for 1 h. Controls included astrocyte proteins incubated with either no antibody or only the secondary antibody, and wells were coated for 2 h at room temperature with either BSA (10 μg/ml; negative control) or EHS laminin (10 μg/ml; positive control). Immune complexes were visualized by incubation with an antirabbit IgG conjugated to horseradish peroxidase and developed with TMB solution. Absorbance was read with an ELISA plate reader at 650 nm.
Metabolic labeling and immunoprecipitation
Analysis of the effects of iodothyronines on the production of laminin by astrocytes was performed by incubating confluent cultures of rat astrocytes with defined medium in the presence and absence of 10 nm iodothyronines for 16 h. Cells were collected by trypsinization; resuspended in labeling medium containing the 35S label EZ-Tag (NEN Life Science Products-DuPont, Boston, MA), methionine-free DMEM, 0.1% (wt/vol) BSA, and 10 nm iodothyronine; seeded onto tissue culture flasks coated with poly-d-lysine (10 μg/ml); and incubated for 6 h at 37 C. The medium, containing secreted proteins, was collected. Cells were solubilized, and proteins deposited into the ECM were removed from the flasks by incubation with 1% (wt/vol) SDS, 10% (vol/vol) glycerol, 5% (vol/vol) 2-mercaptoethanol, and 62.5 mm Tris (pH 6.8) for 1 h at room temperature (43). Proteins from the medium, cells, and ECM were combined and further solubilized by the addition of 1% (vol/vol) Nonidet P-40, 0.4 m NaCl, 2 mm EDTA, 2 mm phenylmethylsulfonylfluoride, and 50 mm Tris (pH 8) and clarified. Solubilized proteins were incubated with antilaminin IgG (1:400 dilution) in the presence and absence of 50 μg EHS mouse sarcoma laminin for 2 h, and immune complexes were isolated with immobilized rProtein A beads (Repligen) followed by centrifugation. Isolated proteins were either counted or resolved on a 5% SDS-PAGE slab gel, and labeled proteins were detected by autoradiography. Specific laminin immunoprecipitation was determined by subtraction of 35S-labeled proteins immunoprecipitated by antilaminin IgG preincubated with 50 μg EHS mouse sarcoma laminin (nonspecific immunoprecipitation). Resolved proteins were also transferred to nitrocellulose for 2 h at 20 mA and analyzed by Western blot with antilaminin IgG as described above.
Statistical methods
Results are reported as the mean ± se. Statistical analysis was performed using single factor ANOVA. Statistical significance was determined to be achieved at the P < 0.05 level.
Results
Immunocytochemical analysis of laminin in astrocytes
We initially examined the effect of thyroid hormone on the synthesis and secretion of laminin in astrocytes by immunocytochemistry. The antiserum used is a commercially available, affinity-isolated, antigen-specific antibody that is documented by dot blot assay by the manufacturer to specifically recognize laminin and does not cross-react with other common ECM proteins (Sigma Product insert). Confluent cultures of astrocytes grown in defined medium with 10 nm T4, 10 nm T3, or no hormone were seeded onto glass coverslips coated with poly-d-lysine and incubated for 3–36 h. Cells were fixed and stained with a specific polyclonal antilaminin IgG. Control incubations with nonimmune serum or with second antibody alone showed no specific staining (data not shown). Equivalent numbers of cells from each treatment group attached to the coverslips at all time points (data not shown), consistent with the prior observation that thyroid hormone has no effect on the attachment of astrocytes to poly-d-lysine (44).
In permeabilized cells (Fig. 1), which identify both intra- and extracellular proteins, specific staining was detected in the perinuclear space of astrocytes within 3 h of attachment in all treatment groups. These observations show that laminin is synthesized in all astrocytes attached to poly-d-lysine. By 10 h, immunoreactive laminin was diffusely distributed throughout the cell in all treatment groups. By 36 h, there appeared to be organization of staining into linear arrays (arrow) in the T4-treated cells along with continued staining in the perinuclear space. In contrast, no linear arrays were observed in any of the thyroid hormone-deficient or T3-treated cells. Instead, punctate clusters of staining were present in both the thyroid hormone-deficient and T3-treated cells at 36 h (arrows).
Immunocytochemical analysis of the effect of thyroid hormone on total laminin distribution in attachment-stimulated astrocytes. Confluent cultures of rat astrocytes were grown for 16 h in defined medium in the presence of 10 nm T4 or 10 nm T3 or in the absence of hormone (SF), collected, seeded onto poly-d-lysine-coated coverslips, and grown in the same defined medium. Total cell-associated laminin is visualized in astrocytes that were fixed with paraformaldehyde and permeabilized with iced methanol for increasing periods of time, as described in Materials and Methods. More than 10 sections on duplicate coverslips from at least 2 experiments were examined. Shown are photomicrographs of representative sections. Marker bar, 10 μm.
It is difficult, if not impossible, to differentiate which of the staining in Fig. 1 is intracellular and which is extracellular using routine immunofuorescence microscopy. To determine the effect of thyroid hormone on the extracellular distribution of laminin, immunocytochemistry was performed on nonpermeabilized cells to prevent access of the antilaminin IgG to the intracellular compartment. There was no detectable staining on the unpermeabilized cells at 3 h in any of the treatment groups (Fig. 2), confirming that the laminin IgG was indeed restricted from the intracellular compartment in the nonpermeabilized cells and indicating that the trypsin treatment removed any laminin attached to the astrocytes before cell collection. Further, these data show that all of the staining present in the permeabilized cells at this time point was intracellular (Fig. 1). By 10 h, there was detectable staining on the T4-treated cells in punctate clusters and in small linear arrays (Fig. 2, arrows). In contrast, there was no staining present on the thyroid hormone-deficient cells, and scant punctate staining was present on the T3-treated cells at 10 h. By 24 h, there was an extensive meshwork of staining on the T4-treated cells that was well organized into intense linear arrays by 36 h. In the thyroid hormone-deficient cells, there continues to be little detectable staining until 36 h, when staining is present in a few punctate clusters (arrows). A pattern similar to that for the thyroid hormone-deficient cells was observed in the T3-treated cells at both 24 and 36 h, with no linear arrays observed. These data indicate that astrocytes organize laminin into specific linear patterns on the cell surface in the presence of T4 but not in the presence of T3 or in the absence of hormone.
Immunocytochemical analysis of the effect of thyroid hormone on the extracellular distribution of laminin in attachment-stimulated astrocytes. Confluent cultures of rat astrocytes were grown for 16 h in defined medium in the presence of 10 nm T4 or 10 nm T3 or in the absence of hormone (SF), collected, seeded onto poly-d-lysine-coated coverslips, and grown in the same defined medium. Extracellular laminin is shown in astrocytes that were fixed with paraformaldehyde and stained for laminin without permeabilization for increasing periods of time, as described inMaterials and Methods. More than 10 sections on duplicate coverslips from at least 2 experiments were examined. Shown are photomicrographs of representative sections. Marker bar, 10 μm.
Biochemical analysis of laminin secretion by astrocytes
Potential mechanisms by which T4 could affect the organization of laminin on the surface of the astrocyte include altering the secretion of the ECM protein or altering the ability of the cell to retain this ECM protein on the cell surface after secretion. We examined these possibilities by determining the effect of thyroid hormone on the secretion of laminin into the medium and into the ECM by astrocytes. Confluent astrocytes were incubated overnight in defined medium containing 10 nm iodothyronines or no hormone then collected and were seeded onto six-well tissue culture-treated plates and grown in the same defined medium for 3–24 h. As shown previously (44), attachment of cells to the tissue culture plates was equal in all treatment groups by 3 h (Fig. 3). Neither T4 or T3 had any effect on astrocyte proliferation over the ensuing 24 h (Fig. 3).
Effect of thyroid hormone on the proliferation of astrocytes attached to poly-d-lysine. Confluent cultures of rat astrocytes were grown for 16 h in defined medium in the presence of 10 nm T4 or 10 nm T3 or in the absence of hormone (Hormone-deficient), collected, seeded onto tissue culture-treated six-well plates, and grown in the same defined medium. Cells were collected, and DNA was determined at increasing periods of time. Results are presented as the mean ± se of triplicate points of a representative experiment, which was repeated at least five times.
Shown in Fig. 4 is the effect of thyroid hormone on the secretion of laminin into the medium by astrocytes. Laminin was detected in the medium within 3 h of seeding onto the tissue culture plate in all conditions. In the absence of thyroid hormone or in the presence of T3, the amount of laminin secreted into the medium increased linearly up to 24 h. In contrast, the quantity of laminin secreted into the medium in the presence of T4 was significantly less than that in the absence of thyroid hormone or in the presence of T3 from 3–16 h after seeding. From 16–24 h, the quantity of laminin secreted into the medium increased in all three treatment groups.
Effect of thyroid hormone on the secretion of laminin into the medium by attachment-stimulated astrocytes. Confluent cultures of rat astrocytes were grown for 16 h in defined medium in the presence of 10 nm T4 or 10 nm T3 or in the absence of hormone (Hormone-deficient), collected, seeded onto tissue culture-treated six-well plates, and grown in the same defined medium. Aliquots of medium were obtained and analyzed for laminin at increasing periods of time. Results are presented as the mean ± se of at least six replicates in a representative experiment, which was repeated at least three times. *, P < 0.05 compared with either T3-treated or T4-deficient cells.
The effect of iodothyronines on the deposition of laminin into the ECM was examined by ELISA analysis of the astrocyte-derived ECM after removal of cellular material following hypotonic disruption (43). Shown in Fig. 5 is the effect of iodothyonines on laminin deposition into the ECM. By 24 h, laminin had been deposited by astrocytes into the ECM in all conditions. The laminin content in the ECM of the T4-treated cells was significantly greater than that observed in the absence of hormone or in the presence of T3, although the differences were small (∼10%). Over the next 12 h, laminin content in the ECM of the T4-treated cells nearly doubled, whereas no significant change was observed in the absence of hormone or in the presence of T3. There was no significant difference between the amount of laminin secreted into the ECM in the presence of T3 or in the absence of hormone at either 24 or 36 h. These data correlate with the immunocytochemical observations (Fig. 2), which show the organization of laminin into intense linear arrays on the T4-treated astrocyte by 36 h and indicate that T4 increases the deposition of laminin into the ECM, presumably by increasing the binding of this ECM protein to the cell surface.
Effect of thyroid hormone on the deposition of laminin into the ECM by attachment-stimulated astrocytes. Confluent cultures of rat astrocytes were grown for 16 h in defined medium in the presence of 10 nm T4 or 10 nm T3 or in the absence of hormone (Hormone-deficient), collected, seeded onto poly-d-lysine-coated 96-well plates, and grown in the same defined medium. After 24 and 36 h of incubation, the medium was aspirated, and the cells were hypotonically burst with distilled water. The content of laminin remaining in the ECM and attached to the plate was quantified by ELISA as described in Materials and Methods. Shown is the result of a representative experiment that was repeated at least four times. The points represent the mean ± se of at least six replicates. OD, Optical density at 650 nm. *, P < 0.05 compared with T4-treated cells at 24 h; #, P < 0.05 compared with T4-treated cells at 36 h.
Effect of thyroid hormone on the expression of laminin in astrocytes
Still undetermined is whether T4 alters the synthesis and secretion of laminin in astrocytes, which could result in the observations of the differential release of laminin into the medium and deposition into the ECM shown in Figs. 4 and 5. However, the immunocytochemical data in the permeabilized cells (Fig. 1) suggests that the synthesis of laminin by astrocytes after cell attachment is not altered by thyroid hormone. To clarify this and to determine whether the altered release of laminin into the medium and deposition onto the cell surface were due to altered production of this ECM protein, we examined laminin protein expression by metabolically labeling the cells after attachment to poly-d-lysine-coated tissue culture flasks. Confluent astrocytes were incubated in defined medium containing 10 nm T4, 10 nm T3, or no hormone overnight, then collected, resuspended in labeling medium, seeded onto poly-d-lysine-coated tissue culture flasks, and grown for 6 h. Cells were solubilized and combined with the medium and then spun to remove the insoluble organelles. The supernatants were immunoprecipitated with an antilaminin IgG and analyzed by SDS-PAGE. Preliminary studies indicated that immunoprecipitation of EHS mouse sarcoma laminin standard increased linearly from dilutions of 1:1000–1:500 and then plateaued from dilutions of 1:500–1:100, indicating that excess antibody was achieved at dilutions lower than 1:500 (data not shown).
Laminin consists of three chains: an α (or A) chain of approximately 400 kDa and β (or B1) and γ (or B2) chains of about 200 kDa each (30, 45). Shown in Fig. 6 are the 35S-labeled proteins immunoprecipitated with the antilaminin IgG (lane 1). The predominant signal is a band of proteins of about 200 kDa. Immunoprecipitation of these labeled proteins was completely blocked by preincubation of the antilaminin IgG with 50 μg EHS mouse sarcoma laminin (Fig. 6, lane 2). Western analysis of the EHS mouse sarcoma laminin (Fig. 6, lane 3) and the astrocyte proteins (Fig. 6, lane 4) immunoprecipitated with the antilaminin IgG showed a signal at about 200 kDa (Fig. 6, lanes 3 and 4, arrow) that corresponded to the β- and γ-chains of laminin (22). The higher molecular mass signal (∼400 kDa) detected in the EHS mouse sarcoma laminin (Fig. 6, lane 3) and corresponding to the α-chain (22) was absent in the astrocytes (lane 4), consistent with the previous observations that astrocytes do not produce the laminin α-chain (22–24, 46).
Identification of laminin synthesized by attachment-stimulated astrocytes. Confluent cultures of rat astrocytes were grown for 16 h in defined medium in the presence of 10 nm T4. Cells were collected, resuspended in labeling medium, seeded onto poly-d-lysine-coated flasks, and grown in the same defined medium for 6 h as described in Materials and Methods. Immunoprecipitation of labeled proteins by laminin IgG was performed. Shown is an autoradiograph of 35S-labeled proteins from attachment-stimulated astrocytes resolved on a 5% SDS-PAGE gel. Lane 1, Proteins immunoprecipitated with antilaminin IgG; lane 2, proteins immunoprecipitated with antilaminin IgG preincubated with 50 μg EHS mouse sarcoma laminin. Western analysis of proteins immunoprecipitated by antilaminin IgG. Immunoprecipitated proteins were resolved on a 5% SDS-PAGE gel, transferred to nitrocellulose, probed with antilaminin IgG, and developed with the Lumiglo chemiluminescent kit. Shown is a representative Western blot. Lane 3, EHS mouse sarcoma laminin; lane 4, astrocyte proteins immunoprecipitated with antilaminin IgG.
The effects of thyroid hormone on laminin protein expression in the first 6 h after attachment are shown in Fig. 7. There was no significant difference in the amount of laminin produced by astrocytes treated with T4, T3, or no hormone. These data indicate that attachment-stimulated synthesis of laminin by astrocytes is not affected by thyroid hormone.
Effect of thyroid hormone on synthesis of laminin by attachment-stimulated astrocytes. Confluent cultures of rat astrocytes were grown for 16 h in defined medium in the presence of 10 nm T4. Cells were collected, resuspended in labeling medium (∼4 × 107 cpm/well), seeded onto poly-d-lysine-coated flasks, and grown in the same defined medium for 6 h. Cells and medium were collected, proteins were solubilized and immunoprecipitated with antilaminin IgG, and isolated 35S-labeled proteins were counted as described in Materials and Methods. Results are the mean ± se of quadruplicate values obtained in four separate experiments and normalized to DNA content.
Discussion
Laminin is an essential guidance molecule for migrating neurites in the developing brain (11, 12, 14, 15) and is developmentally expressed in the brain parenchyma (11, 13, 15, 17, 18). Laminin is synthesized and secreted by astrocytes both in vivo (16–20) and in vitro (21–25). The pattern of this ECM protein on the astrocyte surface provides directional cues to the elongating neurite (11, 12, 15). The masking of laminin-derived cues with antibodies specific to the γ (B2) chain of laminin inhibits neuronal migration in the neonatal rat cerebellum (34) and in cerebellar explant cultures (33) and blocks the elongation of retinal ganglion axons (19). Thus, the absence of laminin-derived guidance cues results in disordered neuronal pathfinding in vitro and in vivo and would probably result in increased neuronal death caused by neurites failing to reach their target destinations in the developing brain, much like that observed in the cretinous brain.
In this study, we show that thyroid hormone, specifically T4, regulates the ability of the cell to deposit and orient the ECM protein laminin into specific patterns on the surface of astrocytes. In T4-treated astrocytes, laminin is secreted and deposited in linear arrays upon the cell surface. In the absence of T4, what little laminin that is deposited on the cell surface is restricted to disorganized punctate clusters, and the remaining laminin secreted by the astrocyte is released into the medium. Neither the synthesis nor the secretion of laminin appears to be affected by thyroid hormone; thus, regulation of the extracellular distribution of laminin on astrocytes does not appear to be a transcriptionally mediated event. These data suggest that the ability of the astrocyte to provide laminin-derived cues to migrating neurites is markedly impaired in the absence of T4. These data taken together with the observation that the expression and regional distribution of laminin are delayed and diminished in the hypothyroid rat cerebellum (35) represent the first demonstration of a mechanism of action for thyroid hormone that can explain many of the morphological derangements observed in the cretinous brain.
The T4-dependent differential patterning of laminin on the astrocyte surface most likely results indirectly from the T4-dependent regulation of cytoskeletal-integrin interactions within the cell. The deposition of laminin on the astrocyte surface requires interactions between the laminin-binding transmembrane receptors known as integrins (26–28) and the filamentous actin microfilament network within the cell that allow clustering of integrins upon binding to laminin, forming strong focal contacts (14, 30). We have previously shown that T4 and its metabolite, rT3, dynamically alter microfilament organization in cultured astrocytes by regulating actin polymerization via a novel extranuclear mechanism of action (38, 42). The transcriptionally active thyroid hormone, T3, is approximately 100-fold less potent in promoting actin polymerization in astrocytes. The T4-dependent regulation of microfilament organization in the astrocyte, in turn, modulates the ability of integrins to cluster into focal contacts upon binding to laminin (44). The regionalization of integrins on the astrocyte surface is diffuse in the absence of thyroid hormone, whereas prominent focal contacts are observed in the T4-treated astrocyte (44).
Shown in Fig. 8 is a proposed mechanism by which T4 could regulate the extracellular orientation of laminin on the astrocyte surface. In the presence of T4, laminin is fixed on the cell surface through binding to integrins, which cluster into macromolecular complexes known as focal contacts (Fig. 8, arrow) that are organized in a specific pattern (14, 29, 30). The integrins are bound to the microfilaments within the cell, which provide the locomotive force required for integrin clustering. Indeed, the distribution of laminin on the astrocyte surface of the T4-treated cell (Fig. 2, 24 and 36 h) coincides with the distribution of integrins on the surface of the T4-treated astrocyte attached to laminin (44). In contrast, the microfilaments are disorganized in the T4-deficient astrocyte (38) and are unable to allow integrins to cluster upon binding to laminin (Fig. 8) (44). The absence of laminin deposited on the surface of the thyroid hormone-deficient or T3-treated cells (Fig. 2) correlates with the absence of integrins clustered into focal contacts on the surface of the thyroid hormone-deficient astrocyte attached to laminin (44). Thus, the inability of the integrins to cluster into focal contacts after binding to laminin prevents the T4-deficient cell from holding onto secreted laminin and from forming linear patterns of laminin on the cell surface. Although the T4-dependent regulation of integrin-cytoskeletal interactions is a dynamic process, the regulation of laminin distribution and organization on the cell surface occurs more slowly, as the synthesis and secretion of laminin are the rate-limiting events and are not affected by either T4 or T3.
Proposed model for the regulation of the extracellular organization of laminin on astrocytes.
How may these in vitro observations account for the in vivo observations that laminin expression in the hypothyroid cerebellum is delayed and diminished compared with that in the the euthyroid cerebellum (35)? The ECM in all tissues, especially developing tissues, is continually being remodeled by the actions of ECM-degrading proteases and protease inhibitors that function to maintain the integrity of the ECM (47, 48). One possible mechanism is that the abnormal deposition of laminin into the ECM in the absence of T4 may activate such enzymes, resulting in the increased degradation of secreted laminin and, thus, alter laminin protein content in the cerebellum without altering laminin gene expression. Studies are ongoing in our laboratory to examine this potential paradox.
The T4-dependent regulation of the organization of the microfilaments in the astrocyte is a well characterized extranuclear action of this morphogenic hormone (38). T4 dynamically regulates the organization of the microfilaments via a mechanism that requires neither transcription nor translation (38). Indeed, astrocytes lack significant numbers of functional thyroid hormone receptors (TR) (49, 50), and the transcriptionally active thyroid hormone, T3, has little if any effect on microfilament organization. The predominant (>95%) TR isoform in astrocyte cultures is the non-T3-binding isoform c-erbAα2 (49), which has been shown to exhibit dominant negative activity in the presence of authentic T3-binding receptors (51–53), although the degree of dominant negative activity by c-erbAα2 has been questioned (54). However, only small quantities of TRα1 (49) and TRβ2 (55) have been identified in cultured astrocytes and, even if the dominant negative activity of c-erbAα2 is weak, the approximately 100-fold excess of this isoform in astrocytes (49) is likely to render the T3-binding TR isoforms transcriptionally inert. These studies suggest that actions of thyroid hormone in astrocytes are not mediated by TR (50).
If the T4-dependent regulation of laminin distribution plays an essential role in modulating brain development, the administration of T3 alone to a congenitally hypothyroid neonate would fail to restore normal brain development. While many T3-regulated genes have been reported (for review, see Refs. 56, 57), the data on the effects of T3 replacement alone on brain development are scarce. T4 is the iodothyronine of choice for thyroid hormone replacement in both animal and human studies. When reported, T3 is primarily used to make the animals thyrotoxic (58–60). Studies are ongoing in our laboratory to examine potential differential roles of T4 and T3 in brain development.
Consistent with previous results (22–24, 46), we found that rat astrocytes produce a variant laminin that lacks the α(A)-chain present in basement membrane laminin in other tissues (Fig. 6). The approximately 200-kDa laminin chains are rapidly synthesized and secreted by astrocytes after the cells attach to poly-d-lysine. The signaling mechanism that induces laminin transcription after cell attachment is unknown at present. It is clear that thyroid hormone does not alter this signal transduction pathway, as production of the approximately 200-kDa laminin chains is equivalent in the presence and absence of iodothyronines.
In summary, we have shown that T4 regulates the deposition and orientation of laminin on the surface of astrocytes. The T4-dependent regulation of the distribution of laminin on the surface of astrocytes during development provides a mechanism by which this morphogenic hormone can influence neuronal migration and development.
This work was supported by NIH Grant DK-49998 (to A.P.F). The contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.
References
