Abstract
Amyotrophic lateral sclerosis (ALS) is a late onset neurodegenerative disorder affecting upper and lower motor neurons (MNs). The molecular mechanisms underlying ALS are poorly understood. Mutations in SOD1 is one of the known causes of ALS but occur only in a very small number of cases of ALS. Interestingly, mutations in human angiogenin (hANG), a member of the ribonuclease A (RNase A) superfamily known to be involved in neovascularization, have been recently reported in patients with ALS, but the effects of these mutations on MN differentiation and survival has not been investigated. We have used the well-characterized pluripotent P19 embryonal carcinoma (EC) cell culture model of neuro-ectodermal differentiation to study the effects of hANG-ALS variants on MN differentiation and survival. Here we report that P19 EC cells induced to differentiate in the presence of hANG and hANG-ALS-associated variants internalize the wild-type and variant proteins. The P19 EC cells differentiate to form neurons but the ability of the neurites to extend and make contacts with neighbouring neurites is compromised when treated with the hANG-ALS variants. In addition, hANG-ALS variants also have a cytotoxic effect on MNs leading to their degeneration. hANG was able to protect neurons from hypoxia-induced cell death, but the variants of hANG implicated in ALS lacked the neuroprotective activity. Our findings show that ANG plays an important role in neurite extension/pathfinding and survival providing a causal link between mutations in hANG and ALS.
INTRODUCTION
Amyotrophic lateral sclerosis (ALS), also known as Lou Gehrig's disease is a late onset neurodegenerative disorder in which upper and lower motor neurons (MNs) are selectively killed leading to progressive paralysis and death, generally between 3–5 years of diagnosis (1). Most cases of classical ALS are sporadic and ∼10% are familial. The causes and mechanisms underlying ALS are poorly understood and until recently, mutations in the SOD1 gene (2) were the main known cause of ALS. However, mutations in SOD1 account for only 1–2% of all cases of ALS and 20% of the familial cases. Besides SOD1, a few other human genes (five nuclear and two mitochondrial) have been linked to ALS or have been identified to predispose to ALS (six loci). The genes linked to ALS include the vesicle-trafficking protein (VAPB) in atypical ALS, ALSIN a putative guanine nucleotide factor for GTPase and a DNA/RNA helicase in juvenile ALS (3–5, reviewed in 6). More recently, Greenway et al. (7,8) have reported mutations in an angiogenic factor, human angiogenin (hANG) in both familial and sporadic cases of ALS (reviewed in 9) implicating this gene in the development of ALS. In an independent study, three new mutations in hANG have also been identified in a cohort of SOD1 negative ALS patients in the USA (Guo-fu Hu, personal communication). These findings suggest a hitherto unknown function for hANG in the nervous system.
The ANGs are members of the pancreatic ribonuclease A (RNase A) superfamily which induce neovascularization (10). hANG is a small protein with a molecular weight of 14.1 kDa and shows 33% sequence identity with bovine pancreatic RNase A. Key residues of RNase A necessary for the enzymatic activity including the catalytic site residues are also conserved in hANG (11–13). Functionally important regions of the protein include the catalytic residues, a nuclear localization sequence, a putative receptor-binding region and a region important for immunomodulation (14–22). ANG orthologues exhibit a high degree of conservation at the amino acid level (reviewed in 23) and in the case of hANG and mouse ANGs, the very weak ribonucleolytic activity (14) was found to be necessary for their angiogenic function (15,24).
hANG was originally isolated from HT29 conditioned medium (10) indicating that it is secreted by this tumour-derived cell line. It is believed that tumour cells secrete hANG which then stimulates endothelial cells leading to neovascularization. hANG promotes the organization of endothelial cells in culture into tube-like structures, induces secondary messengers and supports endothelial cell migration during neovascularization (25–29). hANG and murine ANGs-1 and -4 are all able to induce sprouting from thoracic aorta in culture (24). Endogenous hANG has also been implicated in cell proliferation induced by other angiogenic proteins such as VEGF. Down-regulation of ANG expression by RNAi decreases VEGF-induced rRNA transcription, and inhibition of nuclear translocation of ANG rescinds the angiogenic activity of VEGF (30). ANG has been reported to bind a 42 kDa protein on the surface of endothelial cells (31) as well as a 170 kDa cell surface protein suggested to be the putative receptor (20). Nuclear translocation of exogenous hANG in cultured human umbilical artery endothelial cells involves receptor-mediated endocytosis, microtubule and lysosome-independent transport across the cytoplasm, and nuclear import involving the nuclear localization signal (32).
hANG has previously been implicated in several pathological conditions such as progression of cancers, neovascularization associated with diabetic retinopathy and rheumatoid arthritis (reviewed in 23). Expression of hANG is also induced by hypoxia (33–36). However, until the recent report of the association of mutations in hANG with ALS (7,8), there had been no indication that hANG may have an important function in the nervous system.
We have recently shown that mAng-1, the murine orthologue of hANG is expressed at high levels in the developing nervous system both in the brain and in the spinal cord, predominantly in the neurons (37). mAng-1 is also expressed in the neural precursor cells and in neurons, but not in astrocytes in the P19 mouse embryonal carcinoma (EC) cell culture model of neuro-ectodermal differentiation. Differentiated MNs in the developing mouse embryo and P19 EC cell-derived post-mitotic MNs express mAng-1 very strongly. In the differentiating neurons, mAng-1 co-localized initially with GAP-43 and later with neurofilament (NF) in the neurites (37). NCI 65828, a small molecule inhibitor of hANG (38) was able to dramatically inhibit the neurite extension/pathfinding (37). These results provided the first clear evidence that ANG indeed has an important function in the nervous system which has implications for ALS.
In the study reported here, we show that neurons take up and internalize hANG and hANG-ALS variants. Based on this finding we investigated the effects of hANG and hANG-ALS variants on neuronal differentiation and neurite extension/pathfinding in the P19 EC cell model of neuro-ectodermal differentiation. We show that the hANG-ALS variants affect neurite extension/pathfinding and are also toxic to MNs. We also studied the effects of hANG and hANG-ALS variants on MN survival under hypoxia. hANG is able to protect P19 EC cell-derived MNs from degeneration under hypoxia, but the variants are unable to do so. Taken together, our results show that ANG plays an important role in neurite extension/pathfinding and survival of MNs and provides the basis for further investigating the role of ANG in neuronal differentiation and in the development of ALS.
RESULTS
P19 mouse EC cells (39) differentiate to form neurons when plated directly on the mouse stromal cell line PA6 (40) in serum free medium. We have modified this method and co-cultured the P19 EC cells with PA6 cells in medium lacking serum but supplemented with knockout serum replacement (KOSR) and retinoic acid (RA) as described in Materials and Methods. Under these culture conditions, MNs are generated (37) and can be identified by staining for the MN markers—Islet1 (41) and peripherin. The effect of hANG and hANG-ALS proteins on neurons was studied using this model system. We also used this method to induce differentiation of P19 EC cells grown on Cytodex-3 beads as described earlier (37) and studied the effects of hANG and the hANG-ALS variants on neurite outgrowth and pathfinding. The Cytodex-3 beads provide a focal point from which the outgrowth of neurites can be monitored and the distance the neurite network extends can be measured easily.
hANG and hANG-ALS proteins
We assessed the effects of three hANG variants (hANG-K40I, -Q12L and -C39W) of the seven variants, reported by Greenway et al. (7,8) [see Fig. 1A for the location of the mutations in the 3-D structure of hANG (13) and Figure 1B for the amino acid sequence with the positions of the mutations] and compared them with hANG on neuronal differentiation and neurite extension/pathfinding in the P19 EC cell neuro-ectodermal differentiation model. These three variants were selected due to their location at the active site of the molecule. It is well established that the very weak ribonucleolytic activity of ANG is important for its angiogenic function (15). The three mutations we chose to study are predicted to affect the catalytic activity—C39W affects a cysteine residue which contributes to a disulphide bond, K40I affects the catalytic lysine residue and Q12L affects a glutamine residue which interacts with K40, the catalytic lysine. Since these mutations are not in the putative receptor-binding site of hANG, they are unlikely to affect the uptake of the variants by cells in culture.
(A) 3-D structure on hANG (13) showing the location of the ALS-associated mutations—K40I, Q12L and C39W (7,8). The figure was created using the program PyMOL (DeLano Scientific, San Carlos, CA). (B) Amino acid sequence of hANG. The residues affected by ALS mutations K40I, Q12L and C39W are highlighted.
The hANG-ALS mutations were introduced into hANG by site directed mutagenesis, the variants were successfully expressed in Escherichia coli BL21 (DE3) cells and purified by the method of Holloway et al. (42) as reported previously from our laboratory. The authenticity of the mutant proteins was verified by amino acid sequence analysis and molecular weight determination using electrospray ionization mass spectrometry (ESMS) as described earlier (43). Based on circular dichroism measurements it was established that all three variants were correctly folded with no significant changes in secondary structure. The ribonucleolytic activity (activity toward tRNA) was measured using a ribonucleoytic activity assay. All three variants showed substantial loss of ribonucleolytic activity compared with hANG (43).
Effect of hANG and hANG-ALS variants on PA6 cells and PA6–P19 EC cell differentiating co-cultures
We have previously shown that mAng-1 is expressed at low levels in PA6 cells in the differentiating co-cultures and is strongly expressed in the neurons and MNs derived from P19 EC cells co-cultured with PA6. mAng-1 is also expressed in neurons generated from P19 EC cells in the absence of PA6 cells (37).
In our present study, we found that the addition of exogenous hANG and hANG-ALS variants to the P19–PA6 co-cultures did not have an adverse affect on the morphology of the PA6 stromal cells in the co-culture. PA6 cells in co-cultures with differentiating P19 EC cells in media containing ANG and hANG-ALS variants remained viable and their nuclear morphology was normal and similar to that of PA6 cells in co-cultures with the differentiating P19 EC cells in KOSR media (control) over a period of 144 h. This indicates that the hANG-ALS variant proteins did not have a cytotoxic effect on the PA6 cells as can be seen from the DAPI stained cells at 144 h treatment (Fig. 2A(a–e)). The PA6 cells also retained their ability to induce neuronal differentiation. We also investigated if hANG and hANG-ALS variants had any effect on the proliferation of the PA6 cells. The PA6 cells grew at a steady rate over the 4-day assay period and no significant effect (either increase or decrease) on cell proliferation was observed over the negative control in the cultures exposed to wild-type ANG or the hANG-ALS variants—C39W, -Q12L and -K40I (Fig. 2B).
(A) Morphology of PA6 cells treated with hANG and hANG-ALS variants. PA6 cells in co-cultures with MNs neurons derived from differentiating cells treated with hANG and hANG-ALS variants were fixed at 144 h and stained with DAPI. (a) KOSR; (b) hANG; (c) hANG-C39W; (d) hANG-Q12L and (e) hANG-K40I. Bar represents 10 µm. (B) Effect of hANG and hANG-ALS variants on proliferation of PA6 cells. Histograms showing mean proliferation (±SD) of PA6 cells in response to the indicated proteins, as determined by the MTT cell proliferation assay. In each case, the negative control was 0.1% FCS. Angiogenins were assayed at 200 ng ml−1, all in the presence of 0.1% FCS.
hANG in medium or in circulation is taken up by cells (18,19). We used this property of hANG and incubated PA6 stromal cells alone or differentiating P19 cells co-cultured on PA6, with either wild-type hANG or the hANG-ALS variants to examine their effects on neuronal differentiation and neurite extension/pathfinding. The uptake of hANG by PA6 alone and by PA6 cells and P19 EC cell-derived neurons derived in the differentiating co-cultures (at 96 h of incubation, when neuronal differentiation is prominent) was monitored by immunostaining for hANG with mAb 26-2F. The anti-hANG mAb 26-2F is specific for hANG and does not cross react with mAng-1 (44) and so will specifically stain the hANG taken up by the PA6 and the P19 EC cell-derived neurons and not the endogenous mAng-1. We found that hANG is taken up by PA6 cells but at very low levels and localizes to the nuclei (Fig. 3(a–f)). hANG as well as the hANG-Q12L and -K40I variants were also taken up by the PA6 cells, and the P19 EC cell-derived neurons in the differentiating co-cultures and localize in the nuclei of both cell types (Fig. 4(A–L) larger nuclei-PA6 and smaller nuclei-P19 EC cell-derived neurons). In these experiments we did not test the uptake of hANG-C39W. However, in the co-cultures the level of the uptake by the PA6 cells was considerably lower when compared with the P19 EC cell-derived neurons. Low levels of ANG and variants were also detected in a few neurite outgrowths in P19-derived neurons (Fig. 4A–H).
Uptake of hANG by PA6 cells. PA6 cell monolayer exposed to hANG or media alone for 96 h, stained for hANG with anti-hANG mAb 26-2F, visualized with goat anti-mouse Alexa 488 (green), nuclei counterstained with DAPI (blue). (a–c) media alone and (d–f) hANG 200 ng ml−1. (a and d) DAPI stained nuclei; (b and e) ant-hANG mAb 26-2F and (c and f) merge. Arrow indicates nuclei. Bar represents 10 µm.
Uptake of hANG and hANG-ALS variants by PA6 cells and P19 EC-derived neurons in differentiating co-cultures. PA6 cells and P19 EC-derived neurons in differentiating co-cultures were exposed to hANG and hANG-ALS variants at 200 ng ml−1 for 96 h. Cells were stained for hANG with anti-hANG mAb 26-2F, visualized with goat anti-mouse Alexa 488 (green) and nuclei counterstained with DAPI (blue). (A and B) KOSR; (C and D) hANG; (E and F) Q12L and (G and H) K40I. (white arrow) PA6 cell nuclei; (white arrowhead) nuclei of neuron and (red arrow) neurites. Bar represents 10 µm.
hANG-ALS variants affect neurite extension/pathfinding but not formation of neural precursors or neuronal differentiation
P19 EC cells grown on Cytodex-3 beads were induced to differentiate to neurons by co-culture with PA6 stromal cells in medium containing RA (37). Over a period of 96 h, neurite outgrowths extended out from the NF positive mesh on the beads in medium with 0.5% KOSR alone (Fig. 5A(a and b)). Cultures treated with hANG also put out neurite outgrowths which formed an extensive interconnected network with neurites derived from neighbouring beads, extending considerable distances from the bead (Fig. 5A(c and d)). In cultures treated with hANG-ALS variants, neurite extension/pathfinding was inhibited and they remained largely within the NF positive mesh on the beads (Fig. 5A(e–j)). The diameter of the NF mesh in cultures treated with hANG-ALS variants was also considerably smaller than in those treated with hANG with fewer extended neurite outgrowths (Fig. 5B). The hANG-K40I variant had the most pronounced effect followed by hANG-C39W and hANG-Q12L variants (Fig. 5B).
(A) hANG-ALS variants inhibit path finding of P19-derived neurons. Neurons differentiating on Cytodex-3 beads stained for NF with the mAb 2H3 at 72 and 96 h after treatment with the indicated proteins. (a, c, e, g and i) 72 h and (b, d, f, h and j) 96 h. (a and b) 0.5% KOSR; (c and d) hANG; (e and f) hANG-C39W; (g and h) hANG-Q12L and (i and j) hANG-K40I. (B) Diameter of the mesh of NF positive cells derived from P19 EC cells on Cytodex-3 beads in co-culture with PA6 cells at 96 h. Histograms showing diameter of the NF positive mesh after treatment for 96 h with the indicated proteins. *Significant values compared with KOSR (students t-test), hANG (p< 0.02, n= 10), hANG-C39W, hANG-Q12L and hANG-K40I (p< 0.0001, in all cases, n = 10 beads). Observations from two independent experiments. Bar represents 50 µm.
P19 EC cells induced to form neurons by direct plating on PA6 cells and treated with RA (37) were cultured with wild-type hANGs and hANG-ALS variants and were stained for NF, Islet1, and peripherin at 96, 120 and 144 h of culture to determine the effects on the numbers and survival of neurons and more specifically MNs. NF positive cells appeared in all cultures by 72 h. The neurites formed an extensive meshwork in the control (Fig. 6A–C) and hANG treated cultures by 144 h (Fig. 6D–F). However, in the cultures treated with variant hANG proteins, the neurite outgrowths were not well extended and remained in clumps at 96 h and the effect was most severe at 144 h (Fig. 6 G–O). Interconnections with neighbouring neurons were considerably reduced. By 144 h, there was some degeneration of the neurites as seen by the loss of NF expressing cells (Fig. 6I, L and O). Analysis of the extent and density of the neurite network was carried out by confocal microscopy on cultures stained for NF at 120 h of culture (Fig. 7A–E). Topographical projections of the intensity of NF staining show that the KOSR and hANG-treated cultures had a more even distribution of NF staining material across the field (Fig. 7K and L) as compared to cultures treated with hANG-Q12L, hANG-C39W and hANG-K40I (Fig. 7M–O). hANG-K40I variant caused the most severe bunching effect reflected in the intensity profiles of fluorescence taken at various cross sections along the horizontal axis of the field. Representative profiles are shown in Figure 7F–J.
Effect of hANG and hANG-ALS variants on neurites. Neurons stained for NF (mAb 2H3) 96 h (A, D, G, J and M); 120 h (B, E, H, K and N) and 144 h (C, F, I, L and O). (A–C) KOSR; (D–F) hANG; hANG-C39W (G–I); (J–L) hANG-Q12L and (M–O) hANG-K40I. Arrows indicate degenerating neurons. Bar represents 50 µm.
Effect of hANG and hANG-ALS variants on neurons (neurons at 120 h stained with mAb 2H3). NF staining showing clumping of neurites (A–E); intensity profiles of NF staining across the field (F–J) and 3-D projections of stained neurons in the entire field (K–O). (A, F and K) KOSR; (B, G and L) hANG; (C, H and M) hANG-C39W; (D, I and N) hANG-Q12L and (E, J and O) hANG-K40I. White line across the photo in A–E indicates the plane of the representative intensity profile shown in F–J. Topographical projections were generated for each treatment from confocal images from five independent and random ocular fields of NF-stained cultures using the LSM software.
Post-mitotic MNs as seen by Islet1+ nuclei, first appeared at 72 h of differentiation, peaked at 120 h and were clearly present at 144 h in control cultures. In the presence of hANG the percentage of Islet1+ cells was similar to 0.5% KOSR, both at 120 and 144 h (Fig. 8A(a–d) and B). Of the ALS variants, hANG-Q12L did not significantly affect the number of cells with Islet1+ nuclei when compared with 0.5% KOSR, but both hANG-C39W and hANG-K40I variants showed a significant reduction in Islet1+ nuclei at both 120 and 144 h (Fig. 8A(e–j) and B).
(A) hANG and hANG-ALS variants affect number of Islet1+ cells. Cells stained for Islet1 (mAb 40.2D6), 120 h (a, c, e, g and i) and 144 h (b, d, f, h and i). (a and b) KOSR; (c and d) hANG; (e and f) hANG-C39W; (g and h) hANG-Q12L and (i and j) hANG-K40I. Bar represents 50 µm. (B) Number of Islet1+ cells at 120 (white) and 144 h (grey). *Significant reduction in Islet1+ cells in hANG-C39W cultures (p< 0.05) and in hANG-K40I exposed cultures (p< 0.0005). n= 10 colonies, Observations from two independent experiments.
During neuro-ectodermal differentiation of P19 EC cells in the PA6–P19 co-culture model, peripherin positive cells first appeared at 96 h and were maximal at 120 h (Fig. 9A–F). The number of peripherin positive cells was significantly less in cultures treated with variant hANG proteins at 120 h (Fig. 9G–O) when compared with cultures maintained in KOSR or hANG. The most severe reduction was seen in cultures treated with the hANG-K40I and by 144 h of treatment the numbers were further reduced. The few neurites that remained either had varicosities or were fragmented. This effect was most pronounced in the cultures incubated with hANG-K40I and hANG-C39W, and less severe in the cultures treated with hANG-Q12L (Fig. 9I, L and O).
Effect of hANG and hANG-ALS variants on peripherin+ neuritis. 96 h (A, D, G, J and M); 120 h (B, E, H, K and N) and 144 h (C, F, I, Land O); (A–C) KOSR; (D–F) hANG; (G–I) hANG-C39W; (J–L) hANG-Q12L and (M–O) hANG-K40I. Arrows indicate degenerating MNs. Bar represents 50 µm.
We found that the hANG-ALS variants had two different effects on P19 EC cell-derived neurons. First, in the presence of the hANG-ALS variants, neurons differentiated but neurite extension/pathfinding was affected and these included both Islet1+ and Islet1− neurons. Second, in the presence of hANG-ALS variants C39W and K40I the number of Islet1+ neurons (MNs) were significantly reduced in differentiating cultures. The hANG-Q12L variant was milder in its effects.
hANG-ALS variants are toxic to MNs
Post-mitotic P19 EC cell-derived MNs were exposed hANG and hANG-ALS variants at 96 h of differentiation (when MNs are clearly detectable by staining for Islet1+ cells) for 48 h to determine if their survival would be affected. In cultures treated with hANG (Fig. 10A(c) and B) the number of Islet1+ cells increased significantly over 0.5% KOSR (Fig. 10A(a) and B). The number of Islet1+ cells in hANG-Q12L-treated cultures (Fig. 10A(g) and B) was reduced, which was significant when compared with hANG (Fig. 10A(c) and B), but not in comparison with KOSR. There was also a significant reduction in the numbers of Islet1+ cells in cultures exposed to hANG-C39W and hANG-K40I (Fig. 10A(e and i) and B) both in comparison to KOSR and to hANG. The effect on peripherin positive neurites was similar to that seen for Islet1. Cultures maintained in 0.5% KOSR or exposed to hANG showed healthy neurites staining strongly for peripherin (Fig. 10A(b and d)), whereas cultures exposed to hANG-C39W, hANG-Q12L and hANG-K40I variants all showed varying degrees of degenerating peripherin positive neurons (Fig. 10A(f, h and j)).
Toxic effect of hANG-ALS variants on MNs. (A) Neurons stained for Islet1 (mAb 40.2D6) and peripherin. Islet1+ cells (MNs) at 120 h (a, c, e, g and i); peripherin+ cells at 120 h (b, d, f, h and j). (a and b) KOSR; (c and d) hANG; (e and f) hANG-C39W; (g and h) hANG-Q12L and (i and j) hANG-K40I. Bar represents 50 µm. (B) Number of Islet1+ cells (MNs). *Significant reduction in Islet+ cells in hANG-C39W cultures (p< 0.05) and in hANG-K40I exposed cultures (p< 0.0005). n= 10 colonies. Observations from two independent experiments.
Angiogenin has neuroprotective activity towards neurons under hypoxia and hANG-ALS variants lack this function
hANG, like VEGF has been reported to be expressed in some cells in response to hypoxia (33–36) and there is also a preliminary report suggesting a neuroprotective function for ANG in MNs (45). Since hypoxia is suggested to be one of the causative factors of ALS, we investigated if ANG is neuroprotective to P19 EC cell-derived MNs under hypoxia and if this property of hANG is altered in the hANG-ALS variants. In order to test the ability of wild-type and variant hANGs to protect MNs from hypoxia, differentiated MNs (Islet1+, peripherin+) generated from P19 EC cells at 96 h of differentiation were subjected to normoxia and hypoxia for 18 h. Survival of neurons was assessed by staining for NF and survival of MNs was assessed by staining for Islet1 and peripherin. VEGF165 (VEGF165 promotes survival of MNs during hypoxia through binding to VEGF receptor 2 and neuropilin 1; 46) was used as a positive control and the background level of MN survival was monitored in media containing 0.5% KOSR alone.
P19-derived neurons maintained under normoxia survived and stained positive for NF, perpherin and Islet1. Neurons maintained under normoxia in 0.5% KOSR, VEGF165, hANG, VEGF165 + hANG showed a large number of neurite outgrowths which formed an extensive network as observed in earlier experiments (Fig. 11A, C, K and M). The neurons treated with hANG-C39W, hANG-Q12L, and hANG-K40I under normoxia clumped together and formed a sparse neurite network as expected (Fig. 11E, G and I). Under hypoxia, there was a degeneration of neurons in cultures maintained in 0.5% KOSR as well as in cultures exposed to hANG-C39W, hANG-Q12L and hANG-K40I (Fig. 11B, F, H and J). hANG, VEGF165 and VEGF165 + hANG were neuroprotective (Fig. 11D, L and N).
Effect of hANG and hANG-ALS mutants on survival of NF positive neurons under hypoxia. P19-derived MNs exposed to normoxia (A, C, E, G, I, K and M); hypoxia (B, D, F, H, J, L and N) stained for NF with mAb 2H3. (A and B) KOSR; (C and D) hANG; (E and F) hANG-C39W; (G and H) hANG-Q12L; (I and J) hANG-K40I; (K and L) VEGF and (M and N) VEGF + hANG. Arrows indicate fragmented neurites. Bar represents 50 µm.
Under conditions of normoxia, the number of Islet1+ cells in cultures exposed to KOSR was similar to that expected at 144 h (Fig. 12A(a) and B, Table 1). The number of Islet1+ cells was significantly higher in cultures exposed to VEGF165 and VEGF165 + hANG (Fig. 12A(k and m) and B, Table 1). Under conditions of normoxia, there was no significant reduction in the number of Islet1+ nuclei in hANG-C39W and hANG-Q12L (Fig. 12A(e and g) and B, Table 1) compared with KOSR since these cultures were exposed to the variants for only 18 h. However, even in the short exposure time of 18 h, the hANG-K40I-treated cultures showed a reduction in the number of Islet1+ cells, when compared with 0.5% KOSR (Fig. 12A(i) and B, Table 1).
Neuroprotective effect of hANG and hANG-ALS variants on Islet+ neurons (MNs). (A) P19-derived MNs exposed to normoxia (a, c, e, g, i, k and m) and hypoxia (b, d, f, h, j, l and n) stained for Islet1+ cells with mAb 40.2D6. (a and b) KOSR; (c and d) hANG; (e and f) hANG-C39W; (g and h) hANG-Q12L; (i and j) hANG-K40I; (k and l) VEGF165 and (m and n) VEGF165 + hANG. Bar represents 50 µm. (B) Number of surviving Islet1+ cells (MNs) under normoxia (white) and hypoxia (grey). *Significant loss of Islet1+ cells under hypoxia in KOSR (p< 0.007), hANG-C39W (p< 0.01) and hANG-Q12L (p< 0.05) when compared with normoxia; **Significant survival under hypoxia in hANG (p< 0.01), VEGF (p< 0.001) and VEGF+hANG (p< 0.001) compared with KOSR. n= 10 colonies in all cases. Observations from two independent experiments.
Comparison of the effects of hANG, hANG-ALS variants and VEGF165 on survival of MNs under normoxia and hypoxia. MNs stained for Islet1
| Protein | Normoxia | Hypoxia | Normoxia versus hypoxia |
|---|---|---|---|
| KOSR | Average of 94 Islet1+ cells present per 1000 cells‐control | Significantly fewer Islet1+ cells (39 per 1000 cells) than in normoxia (p< 0.007). | |
| VEGF | Significantly more Islet1+ cells than KOSR normoxia (p< 0.01) | Significantly more Islet1+ cells than KOSR hypoxia (p< 0.001). | Significant reduction in Islet1+ cells (p< 0.007). |
| hANG | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | Significantly more Islet1+ cells than KOSR hypoxia (p< 0.01) | No significant reduction in Islet1+ cells |
| VEGF + hANG | Significantly more Islet1+ cells than KOSR normoxia (p< 0.01) | Significantly more Islet1+ cell than in KOSR hypoxia (p< 0.001) | No significant reduction in Islet1+ cells |
| hANG-Q12L | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | Significant reduction in Islet1+ cells (p< 0.01) |
| hANG-C39W | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | Significant reduction in Islet1+ cells (p< 0.05) |
| hANG-K40I | Significantly less Islet1+ cells than KOSR normoxia (p< 0.03) | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | No significant reduction in Islet1+ cells |
| Protein | Normoxia | Hypoxia | Normoxia versus hypoxia |
|---|---|---|---|
| KOSR | Average of 94 Islet1+ cells present per 1000 cells‐control | Significantly fewer Islet1+ cells (39 per 1000 cells) than in normoxia (p< 0.007). | |
| VEGF | Significantly more Islet1+ cells than KOSR normoxia (p< 0.01) | Significantly more Islet1+ cells than KOSR hypoxia (p< 0.001). | Significant reduction in Islet1+ cells (p< 0.007). |
| hANG | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | Significantly more Islet1+ cells than KOSR hypoxia (p< 0.01) | No significant reduction in Islet1+ cells |
| VEGF + hANG | Significantly more Islet1+ cells than KOSR normoxia (p< 0.01) | Significantly more Islet1+ cell than in KOSR hypoxia (p< 0.001) | No significant reduction in Islet1+ cells |
| hANG-Q12L | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | Significant reduction in Islet1+ cells (p< 0.01) |
| hANG-C39W | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | Significant reduction in Islet1+ cells (p< 0.05) |
| hANG-K40I | Significantly less Islet1+ cells than KOSR normoxia (p< 0.03) | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | No significant reduction in Islet1+ cells |
Ten independent colonies averaging 1000 cells/colony were selected at random from Islet1 and DAPI-stained cultures for each treatment. Observations from two independent experiments.
Comparison of the effects of hANG, hANG-ALS variants and VEGF165 on survival of MNs under normoxia and hypoxia. MNs stained for Islet1
| Protein | Normoxia | Hypoxia | Normoxia versus hypoxia |
|---|---|---|---|
| KOSR | Average of 94 Islet1+ cells present per 1000 cells‐control | Significantly fewer Islet1+ cells (39 per 1000 cells) than in normoxia (p< 0.007). | |
| VEGF | Significantly more Islet1+ cells than KOSR normoxia (p< 0.01) | Significantly more Islet1+ cells than KOSR hypoxia (p< 0.001). | Significant reduction in Islet1+ cells (p< 0.007). |
| hANG | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | Significantly more Islet1+ cells than KOSR hypoxia (p< 0.01) | No significant reduction in Islet1+ cells |
| VEGF + hANG | Significantly more Islet1+ cells than KOSR normoxia (p< 0.01) | Significantly more Islet1+ cell than in KOSR hypoxia (p< 0.001) | No significant reduction in Islet1+ cells |
| hANG-Q12L | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | Significant reduction in Islet1+ cells (p< 0.01) |
| hANG-C39W | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | Significant reduction in Islet1+ cells (p< 0.05) |
| hANG-K40I | Significantly less Islet1+ cells than KOSR normoxia (p< 0.03) | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | No significant reduction in Islet1+ cells |
| Protein | Normoxia | Hypoxia | Normoxia versus hypoxia |
|---|---|---|---|
| KOSR | Average of 94 Islet1+ cells present per 1000 cells‐control | Significantly fewer Islet1+ cells (39 per 1000 cells) than in normoxia (p< 0.007). | |
| VEGF | Significantly more Islet1+ cells than KOSR normoxia (p< 0.01) | Significantly more Islet1+ cells than KOSR hypoxia (p< 0.001). | Significant reduction in Islet1+ cells (p< 0.007). |
| hANG | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | Significantly more Islet1+ cells than KOSR hypoxia (p< 0.01) | No significant reduction in Islet1+ cells |
| VEGF + hANG | Significantly more Islet1+ cells than KOSR normoxia (p< 0.01) | Significantly more Islet1+ cell than in KOSR hypoxia (p< 0.001) | No significant reduction in Islet1+ cells |
| hANG-Q12L | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | Significant reduction in Islet1+ cells (p< 0.01) |
| hANG-C39W | No significant difference in Islet1+ cell numbers compared with KOSR normoxia | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | Significant reduction in Islet1+ cells (p< 0.05) |
| hANG-K40I | Significantly less Islet1+ cells than KOSR normoxia (p< 0.03) | No significant difference in Islet1+ cell numbers compared with KOSR hypoxia | No significant reduction in Islet1+ cells |
Ten independent colonies averaging 1000 cells/colony were selected at random from Islet1 and DAPI-stained cultures for each treatment. Observations from two independent experiments.
A significant number (∼50%) of P19 EC cell-derived MNs failed to survive in 0.5% KOSR under hypoxia, as seen by the reduction in the number of Islet1+ nuclei (Fig. 12A(b) and B, Table 1). However, in cultures maintained in hANG under hypoxic conditions the number of Islet1+ cells was not significantly reduced (Fig. 12A(d)) when compared with cultures under normoxic conditions (Fig. 12(c) and B) representing 90% MN survival (Table 1). A substantial number of P19-derived MNs also survived when exposed to VEGF165 (Fig. 12A(l) and B, Table 1). Survival of MNs was severely compromised under hypoxia in cultures treated with hANG-Q12L and hANG-C39W (Fig. 12A(f and h) and B) and the number of Islet1+ cells were significantly reduced when compared to equivalent cultures under normoxia (Fig. 12A(e and g) and B, Table 1) as well as to cultures exposed to hANG under hypoxia (Fig. 12A(d) and B, Table 1). In hANG-K40I-treated cultures, the number of Islet1+ cells were very low in normoxic conditions itself (Fig. 12A(i) and B, Table 1), hence it was not possible to quantify the effect under hypoxia (Fig. 12A(j) and B).
VEGF165 and hANG were used in combination to assess if they had an additive neuroprotective effect. Cultures under normoxia exposed to a combination of VEGF165 and hANG showed an additive effect on the number of Islet1+ cells (Fig. 12A(m) and B, Table 1). There was no significant reduction in the number of Islet1+ cells in the presence of both VEGF165 and hANG under hypoxia when compared with normoxia, indicating that there was increased survival of MNs (Fig. 12A(n) and B, Table 1). The survival of MNs under hypoxia was also significantly enhanced over cultures exposed to either VEGF165 or hANG alone (Fig. 12B, Table 1) suggesting a synergistic effect. The mean area of cell colonies did not vary significantly between different treatments both under normoxia and hypoxia and neither did the total number of cells per colony.
Neurons treated with KOSR, hANG, VEGF165 and the hANG-ALS variants cultured under conditions of normoxia and hypoxia were stained for NF and peripherin individually and in combination. It can be clearly seen from Figure 11 that some NF positive neurons survived under hypoxia in cultures exposed to KOSR, hANG and VEGF165 as well as in cultures exposed to the hANG-ALS variants. The picture is quite different when the pattern of staining for peripherin alone was compared between cultures under normoxia and hypoxia (Fig. 13). Neurons cultured in KOSR alone survived under normoxia (Fig. 13A), but under hypoxic conditions almost all peripherin positive neurites were fragmented (Fig. 13B). Most peripherin positive neurons in cultures exposed to the hANG-ALS variants also showed degenerating and fragmented neurites with hANG-C39W and hANG-K40I being the least neuroprotective (Fig. 13F and J). hANG-Q12L was marginally neuroprotective (Fig. 13H). In contrast, a large number of peripherin positive neurons survived under hypoxia in hANG (Fig. 13D) and VEGF165-treated cultures (Fig. 13L) as well as under normoxia (Fig. 13C and K). The combination of VEGF165 and hANG was more strongly neuroprotective (Fig. 13N) under hypoxia than VEGF165 or hANG alone. The inability of the hANG-ALS variants to confer neuroprotection on MNs can be seen quite clearly in Fig. 14A and B, which show cultures stained for both peripherin and NF. The neurites staining for peripherin are considerably less in number and fragmented as compared to neurites staining for NF in cultures maintained in KOSR and hANG-ALS variants under hypoxia.
Effect of hANG and hANG-ALS variants on survival of peripherin positive neurons under normoxia and hypoxia. P19-derived neurons exposed to normoxia (A, C, E, G, I, K and M); hypoxia (B, D, F, H, J, L and N) stained for peripherin. (A and B) KOSR; (C and D) hANG; (E and F) hANG-C39W; (G and H) hANG-Q12L; (I and J) hANG-K40I; (K and L) VEGF165 and (M and N) VEGF165 + hANG. Arrows indicate fragmented neurites. Bar represents 50 µm.
(A) hANG protects MNs from hypoxia induced degeneration. Neurons cultured under hypoxia and normoxia were co-immunostained for NF (mAb 2H3) and peripherin. (a–f) KOSR; (g–l) hANG; (m–r) VEGF and (s–x) VEGF + hANG. (arrows) fragmented peripherin positive neurites. (B) hANG-ALS variants lack neuroprotective activity towards MNs under hypoxia. Neurons cultured under hypoxia and normoxia were co-immunostained for NF (mAb 2H3) and peripherin. (a–f) hANG-C39W; (g–l) hANG-Q12L and (m–r) hANG-K40I. (arrows) fragmented peripherin positive neurites.
DISCUSSION
Our finding from this study and from our earlier report (37) is that ANG has an important role in neurite extension/pathfinding. Treatment of neurons and MNs with hANG-ALS variants significantly compromises this function and also affects MN survival. Besides its function in neurite extension/pathfinding, hANG also acts as a neuroprotective factor towards MNs under oxidative stress and hANG-ALS variants are unable to perform this function. These findings have implications for ALS.
Angiogenic growth factors such as VEGF, involved in sprouting, growth and branching of blood vessels and capillaries (reviewed in 47) are now increasingly seen to perform similar roles in the development of neurons (reviewed in 48). The converse is also becoming true, in that ligand—receptor signalling systems involved in neuronal guidance are also being implicated in vessel guidance and patterning (reviewed in 49). An important aspect which needs to be investigated in this regard is whether ANG is secreted by neurons. If so, does ANG signal by an autocrine mechanism or by a type 2 intracrine mechanism and how this relates to the role the hANG-ALS variants play in the aetiology of ALS. Hu et al. (20) report a receptor for hANG in endothelial cells and suggest the mechanism by which exogenous hANG is translocated to the nucleus in these cells (32). Our experiments clearly show that exogenous hANG is taken up by neurons and is translocated to the nucleus. It remains to be established if ANG binds to specific receptors on the cell surface of neurons before internalization and if the mechanism of translocation to the nucleus is similar to that in endothelial cells.
The ribonucleolytic activity of ANG (the enzymatic activity of hANG is miniscule compared with pancreatic RNase A) is important for its angiogenic activity, as blocking it by a potent small molecule inhibitor, NCI 65828, affects neovascularization (38). We have previously shown that this inhibitor is also able to inhibit neurite extension/pathfinding in the P19 model of neuroectodermal differentiation (37). The hANG-ALS variants that have been reported by Greenway et al. (7,8) not only involve residues that affect the catalytic activity of hANG, but also residues in other regions of the protein such as the nuclear translocation signal. Many of the hANG-ALS variants involve residues in hANG that are intolerant to substitutions (50). One can speculate that these mutations cause structural alterations affecting a function of ANG that may or may not be linked to its enzymatic activity. Our studies on the biochemical characterization of some of the hANG-ALS variants including the three used in the current report show that there is no substantial alteration in the secondary structures of the variants (43). Thus, it is unclear how the different mutations in hANG cause the degeneration of MNs in ALS. One possibility is that in the cellular milieu the hANG-ALS variants are misfolded, thereby leading to formation of aggregates which have cytotoxic effects like the SOD1 variants (51,52). It is well established that in several neurodegenerative disorders there is a deposition of aggregated protein in inclusion bodies enriched in NFs. This is also seen in the case of SOD1 in ALS (53). We have observed that hANG-ALS variants not only affected the neurite extension/pathfinding, but were also toxic to MNs, although endogenous wild-type ANG is present in these neurons. Since circumstantial evidence indicates that the hANG-ALS variants may function as autosomal dominant (7,8), this observation is not surprising. That the hANG-ALS variants are able to inhibit the function of the endogenous ANG, suggests that they may interact with wild-type hANG and block its function. The other possibility is that the hANG-ALS variants may counteract the function of wild-type hANG by binding to the substrate unproductively. These possibilities need to be explored.
Several reports link VEGF, a regulator of angiogenesis to neuronal development, neuroprotection and to the pathology of MNs in conditions such as ALS (46,54–56). hANG, like VEGF, is also expressed in response to hypoxia (33–36) and an earlier brief report indicated that ANG has a neuroprotective function (45). We found that hANG, like VEGF165, is neuroprotective towards MNs under hypoxia, but the hANG-ALS variants lack this function. hANG in combination with VEGF exerts a marginally synergistic effect on MN survival suggesting that VEGF and hANG may be acting via different pathways as far as the neuroprotection is concerned.
How do the different functions of ANG in the nervous system relate to the effects of hANG variants on MNs in ALS? It has been suggested that there is a possibility that MNs are selectively vulnerable to limited hypoxia or vascular insufficiency; overlaying genetic and environmental factors which may trigger ALS (57). hANG is known to be up-regulated under hypoxia (33–36) and is neuroprotective (this report). It is possible that under hypoxia, the levels of wild-type hANG produced from the single normal allele in patients with an autosomal dominant mutation in hANG are insufficient for neuroprotection, since the hANG-ALS variants lack this function. Additionally, an increased production of the hANG-ALS variant proteins leading to their accumulation would affect pathfinding by neurite outgrowths and also contribute to the degeneration of MNs by their cytotoxic effect.
We have previously shown that the expression of ANG is important for neurite extension/pathfinding (37). We now show that hANG-ALS variants interfere with this function and are toxic to MNs. We also show that hANG has a neuroprotective function in MNs and the hANG-ALS variants are unable to perform this function. However, molecular mechanisms by which ANG mediates neurite pathfinding and how this is affected by the hANG variants still need to be addressed. Biochemical studies coupled with gene knockouts and knock-in of the hANG-ALS mutations should provide some of the answers, but the knockout approaches may prove to be difficult as there are six members of this family and three pseudogenes in the mouse (58,59).
MATERIALS AND METHODS
Expression and purification of hANG and hANG-ALS variants
Wild-type hANG and hANG-ALS variants in pET-22b(+) (Novagen, Nottingham, UK) bacterial expression plasmid were used to transform E. coli BL21(DE3) cells. Recombinant proteins were prepared by the method of Holloway et al. (42). The recombinant proteins deposited as inclusion bodies were solubilized, refolded and purified by SP-Sepharose chromatography followed by C4 reversed-phase HPLC. Protein molecular masses were determined by ESMS as described earlier (43).
Ribonucleolytic activity assay
Ribonucleolytic activity towards tRNA was determined by measuring the formation of acid-soluble fragments as described by Shapiro et al. (17). Assay mixtures contained 2 mg ml–1 yeast tRNA (Sigma-Aldrich, Dorset, UK), 0.1 mg ml–1 bovine serum albumin (BSA) and 0.05, 0.1, 0.2, 0.3, 0.4 or 0.5 µM test protein in 33 mM Na-HEPES, 33 mM NaCl, pH 7.0. After 2 h of incubation at 37°C, reactions were terminated by the addition of 2.3 vol ice-cold 3.4% perchloric acid, the mixtures were centrifuged at 13 000g for 10 min at 4°C, and the absorbances of the supernatants were measured at 260 nm.
Cell culture
The mouse EC cell line P19 (42,60) and the PA6 mouse stromal cell line (40) were cultured in α-MEM-glutamax (Invitrogen, Paisley, UK) supplemented with 10% fetal calf serum (FCS) (Invitrogen) and 1% non-essential amino acids (NEAA) (Invitrogen) at 37°C in 5% CO2.
Cell proliferation assay
PA6 cells were seeded in 48-well plates at a density of 5 × 103 cells per well in growth media [α-MEM-glutamax, FCS (10% v/v) and NEAA (1% v/v)]. Cells were incubated at 37°C, 5% CO2 for 24 h and the growth media was replaced with low-serum media [α-MEM-glutamax, FCS (0.1% v/v) and NEAA (1% v/v)] supplemented with wild-type or mutant ANG proteins at a concentration of 200 ng/ml. Cells cultured in low-serum media alone was used as the negative control. Cell proliferation was measured every 24 h over a period of 4 days using the Promega MTT cell proliferation assay kit following the manufacturers instructions. Media (with appropriate protein) was refreshed after 48 h during the assay period. Results were analyzed using the paired students t-test with a two-tailed distribution.
Uptake of hANG by PA6 cells
PA6 cells (2 × 104 cells/well) were plated on gelatinized coverslips in growth medium [α-MEM-glutamax, FCS (10% v/v) and NEAA (1% v/v)] and allowed to grow overnight. Cells were washed with PBS and incubated in uptake medium [α-MEM-glutamax, 0.5% v/v KOSR (Invitrogen) and NEAA (1% v/v)] supplemented with 200 ng/ml hANG and incubated for 96 h with a change of medium (with appropriate protein) at 48 h. Cells cultured in uptake medium alone without any added protein served as the control. Cultures were fixed at 96 h in methanol at −20°C and immunostained for hANG.
Induction of differentiation of P19 cells and treatment with variant ALS-hANG proteins
Differentiation by direct plating on PA6 feeders
PA6 cells (2 × 104 cells/well) were plated on gelatinized cover slips in α-MEM-glutamax containing 10% (v/v) FCS and 1% (v/v) NEAA in 12-well plates 2 days prior to seeding with P19 EC cell. P19 EC cells (5 × 103 cells/ml) were plated on the confluent monolayer of PA6 stromal cells in α-MEM-glutamax medium containing 0.5% (v/v) KOSR, 1% (v/v) NEAA and 5 × 10−7 M RA (Sigma). Cell morphology was monitored daily using a Nikon inverted microscope. Differentiating P19 EC cells were treated with hANG and hANG-ALS variants at 200 ng/ml on the day of plating on the PA6 cells. Differentiation media without ANG proteins served as control. Media with the appropriate proteins was replenished at 48 h intervals. Cultures were fixed in 4% buffered paraformaldehyde (PFA) (Sigma) or in methanol at −20°C at 24 h intervals over 144 h.
Differentiation of P19 EC cells on beads
P19 EC cells were grown on Cytodex 3 bead (GE Healthcare, Amersham Place, UK) in α-MEM-glutamax containing 10% FCS and 1% (v/v) NEAA. When confluent (generally overnight), the cell coated beads were plated on a confluent monolayer of PA6 cells on gelatinized coverslips in α-MEM-glutamax containing 0.5% (v/v) KOSR, 1% (v/v) NEAA and 5 × 10−7 M RA (Sigma) (as for differentiation by direct plating) to induce neuronal differentiation. The test proteins (wild-type hANG, hANG-ALS variants) were added at a concentration of 200 ng/ml on day of plating. Differentiation media without ANG proteins served as control. Samples were fixed in 4% buffered PFA at 24 h intervals over a period of 96 h.
Hypoxia
P19 cells were differentiated to MNs by culturing for 96 h on PA6 stromal cells in α-MEM-glutamax supplemented with KOSR (0.5% v/v), NEAA (1% v/v), RA (5×10−7 M), HEPES (25 mm) at 37°C under 5% CO2 and media was refreshed every 48 h under normoxia. The co-cultures were transferred to differentiation media containing the sample proteins (200 ng/ml for hANG and hANG variants, 20 ng/ml VEGF165, R&D Systems, Oxford, UK) or to differentiation media alone just prior to exposure to hypoxia at 96 h of differentiation. One set of cultures were transferred to a humidified hypoxia chamber (Billups Rothemberg, Del Mar, CA, USA) at 37°C under 94% N2, 1% O2 and 5% CO2 [chamber evacuated for 5 min at 20 l/min with gas mixture (94% N2, 1% O2 and 5% CO2) to a final level of 1% O2, confirmed with an oxygen meter] and incubated for a further 18 h. A duplicate set of co-cultures was kept at 37°C under 5% CO2 (normoxia control) for the same 18 h time period. Cells were fixed with buffered PFA (4% w/v) and immunostained for markers.
Indirect immunoflourescence of P19 and differentiated neuronal lineages from P19
Cell cultures on coverslips were fixed in 4% PFA in PBS in all cases, except for detection of hANG and hANG-ALS variants in the uptake studies, when the cells were fixed in methanol for 10 min at −20°C. Non-specific binding was blocked with PBS containing 0.1% gelatin, 0.5% BSA and 0.1%Tween 20 for 1 h at room temperature. The primary antibodies used for staining were mouse anti-Islet1 (1:5, mAb 40.2D6, DSHB), mouse anti-NF (1:10, mAb2H3, DSHB), mouse anti-hANG (10 µg/ml, mAb 26-2F, 44) and rabbit anti-peripherin (1:500, Chemicon, Hampshire, UK). Primary antibodies diluted in block solution were allowed to bind overnight at 4°C with shaking, washed with Triton X-100 (0.1% v/v) in PBS (3 × 10 min). Secondary antibodies were goat anti-mouse Alexa 488 (1:1000, Molecular Probes, Paisley, UK), goat anti-rabbit Alexa 488 (1:1000, Molecular probes), donkey anti-mouse Texas Red (1:250, GE Healthcare). Cells were washed with wash buffer as earlier and then stained with DAPI (0.1 µg/ml, Sigma) in PBS for 5 min at room temperature, rinsed in PBS and mounted in Mowiol. Staining was observed using either a confocal- Zeiss LSM 510 or a Leica DMRB and images captured digitally. Topographical projections were generated for each treatment from confocal images from five independent and random ocular fields of NF-stained cultures using the LSM software.
Estimation of diameter of neurite network (2H3 stained) on the beads
P19 EC cells coated on Cytodex-3 beads were co-cultured with PA6 in differentiation medium containing KOSR alone or ANG proteins for 96 h and were stained for NF and with DAPI. Well separated beads (10 in number) were selected randomly from each sample and the diameter of the neurite network that surrounded each bead was measured using a linear eyepiece graticule at the widest point. Statistics was performed by the paired students t-test.
Determination of Islet+ (MNs) cell numbers
Post-mitotic MNs were identified by immunostaining with mAb40.2D6, which stains Islet1. The cultures were also counterstained with DAPI. The number of cells in a colony was counted using DAPI-stained nuclei and area of the colony of neural cells was measured. Neuronal cells were distinguishable from the larger well spread PA6 stromal by their smaller nuclei. The colony area, number of cells per colony and the number of MNs (Islet1+ cells) in each colony was determined using the Leica Application Suite. Cell counts and area measurement were acquired from a total of 10 independent colonies for each treatment. Data were obtained from two independent experiments in all cases. Statistics was performed by the paired students t-test.
FUNDING
Medical Research Council, UK (G9713682 to V.S.); Wellcome Trust (UK) (083191 to K.R.A.; 073153 to K.R.A. and V.S).
ACKNOWLEDGEMENTS
We would like to acknowledge the Developmental Studies Hybridoma Bank, University of Iowa and Tom Jessell for the mAbs 2H3 and 40.2D6 and Guo-fu Hu for the mAb 26-2F. We are grateful to Susan Wonnacott for helpful comments on the manuscript, Ying Feng and Shalini Iyer for help with figures.
Conflict of Interest statement. None declared.
