Fig 1.
A Comparison table of RR with loukoumasomes. B. Morphological variants of loukoumasomes found in pelvic ganglia. Pericentrin immunolabeling followed by confocal analysis (z-stacks, maximum projection) reveal rods indicated by open white arrowhead), toroids (indicated by solid white arrowhead) and intermediate forms. Some toroids in the first three panels are “side view” images and therefore lack the stereotypical “donut” shape. Intermediate forms (described as twisted/pinched tori) include loukoumasomes that are similar to intermediate forms of RR including “sewing-pin” (solid white arrows) and “figure eight” (open white arrows) structures. A representative “figure eight” (upper) and “sewing-pin” (lower) intermediate form can be seen in the higher magnification panels on the far right. Calibration bar represents 20 μm in low magnification views and 5 μm in insets.
Fig 2.
Morphological variants of loukoumasomes in retinoblastoma tissue, cell lines and sympathetic neurons as detected using two beta-III tubulin antibodies.
A-C. Y79 cells were immunostained with one of two beta-III tubulin antibodies (Tu-20, A, or SDL.3D10, B-C). Nuclei were counterstained with DAPI or Hoechst (Blue). Large, toroidal shaped loukoumasomes are recognized by both antibodies (see white arrows) and are often observed within DNA impoverished pockets of the nucleus that were dark under DIC (B). D-E. Loukoumasomes though rare were observed throughout the retinoblastoma tissue samples (see white arrows highlighting loukoumasomes). Loukoumasomes were immunolabeled with an anti-beta-III tubulin antibody SDL.3D10 (D) or Tu-20 (E); nuclei were counter stained for DAPI. Loukoumasomes were observed in both Rb-positive (D, E, right panel) and Rb-negative tissue (D, E, left panel). Insets represent a single slice of the stack data illustrating the fact that large loukoumasomes are typically within a DNA impoverished (DAPI negative) region. F. Loukoumasomes appear in both bi-nucleated and mono-nucleated rat sympathetic neurons. Arrows show two nuclei in a single adrenergic neuron immunolabeled with a beta-III tubulin and tyrosine hydroxylase (TH) antibody. G. Loukoumasomes also appear in multinucleated retinoblastoma cells. A single slice of the stack data from a multinucleated cell is shown following beta-III tubulin immunolabeling and DAPI staining. Calibration bars represent 20 μm.
Fig 3.
Loukoumasomes in retinoblastoma tissue and cell lines are composed of stabilized microtubules.
A-B. Y79 cells were immunolabeled with an alpha-tubulin antibody and costained with phallodin and DAPI. Alpha-tubulin recognized both larger retinal-loukoumasomes (B) and smaller, more numerous retinal-loukoumasomes near the plasma membrane (see white arrows, A). C. Sections of well-differentiated human retinoblastoma tissue were immunostained for alpha-tubulin; nuclei were counterstained for DAPI. Like loukoumasomes detected with a beta-III tubulin antibody, retinal-loukoumasomes detected with an alpha-tubulin antibody were observed throughout the tissue (see white arrows highlighting loukoumasomes). D-F. Y79 cells and retinoblastoma tissue were immunolabeled with beta-III tubulin (SDL.3D10) (D-F) and detyrosinated tubulin (D,F), and costained with DAPI. Calibration bars represent 20 μm. E. Representative images showing the interaction between beta-III tubulin and detyrosinated tubulin in Y79 cells as revealed by the in situ PLA together with beta-III tubulin immunofluorescence. The strong reaction product in the PLA indicates that the two proteins are in close proximity (i.e., interacting). The fact that some red fluorescent dots occur where loukoumasomes are visualized indicates that both proteins are in close proximity at loukoumasomes. Negative PLA controls obtained by omitting one of the two primary antibodies are found in the supplementary data (see S1 Fig).
Fig 4.
Loukoumasomes and RR are Antigenically Distinct A. Adult rat sympathetic ganglia were co-labelled with the beta-III tubulin antibody, (SDL.3D10) known to recognize loukoumasomes and the anti-RR/IMPDH antibody, (It2006) and stained with DAPI. A single large perinuclear toroidal loukoumasome can be seen within a large sympathetic neuronal cell soma. This neuronal cell is surrounded by glia (as evident by small intensely stained nuclei). Although the loukoumasome stained intensely with the beta-III tubulin antibody it was not immunolabeled with the anti-IMPDH antibody. B. Retinoblastoma tissue was also immunolabeled with a beta-III tubulin antibody (SDL.3D10) together with the anti- IMPDH antibody (It2006). Arrows depict a loukoumasome that is immunolabeled with an anti-loukoumasome antibody (SDL.3D10) but not the anti-RR antibody (It2006). RR immunostaining was not expected in untreated retinoblastoma tissue. However, intense cytosolic immunoreactivity with the anti-IMPDH antibody was observed in cells having condensed nuclei. This cytosolic staining can be seen encircling the nucleus but does not immunolabel loukoumasomes. C. Adult rat sympathetic ganglia were co-labeled with the gamma-tubulin antibody (also known to recognize loukoumasomes) and the anti-RR antibody (It2006) and stained with DAPI. Loukoumasomes though positive for gamma-tubulin, did not contain IMPDH.
Fig 5.
Induction of RR, but not loukoumasomes, in retinoblastoma cells with ribavirin.
A. A 24 hour exposure to ribavirin (an IMPDH inhibitor) increased the number of Y79 cells with RR but not retinal-loukoumasomes. Loukoumasomes were quantitated using an anti-beta-III tubulin antibody (clone Tu-20) and RR were quantitated using an anti-IMPDH antibody (anti-RR serum It2006). Data are represented as the mean ± SEM from 4 independent experiments with greater than 400 cells counted per treatment group. * p < 0.01, different from untreated control. Significance was determined using an ANOVA and Fisher’s (LSD) test. B. Y79 cells were immunolabeled with an IMPDH autoantibody It2006 (red) and counterstained with DAPI (blue) following exposure to ribavirin (1 mM and 2 mM) or control (0 mM) for 24 hours. Yellow arrows depict an example of RR in the 1 mM and 2 mM treatment groups. No RR were present in the control group. C. Representative figures of Y79 cells, immunolabeled with a monoclonal antibody against beta-III tubulin (clone TU-20; green) together with an IMPDH autoantibody It2006 (red). Nuclei were counterstained with DAPI (blue). This was done following exposure to 1 mM and 2 mM ribavirin for 24 hours or vehicle control. White arrows highlight loukoumasomes, while yellow arrows indicate examples of RR. D. A confocal stack of a Y79 cell (left panel) along with orthogonal views of the same cell (right panel) at a given z position (z = μm). Each successive orthogonal step (labelled 1–2) is 7.36 μm (for the yz orthogonal views) or 2.56 μm (for the xz orthogonal views) apart; broken arrows indicate the direction in which steps were taken. The orthogonal views indicate that the loukoumasome and RR are distinct structures in separate sections of the cell. Tubulin and IMPDH It2006 intensity profiles on this same cell were performed. Fluorescence intensities indicate that the loukoumasome (green fluorescence) and the RR (red fluorescence) are distinct structures. Calibration bars represent 20 μm.
Fig 6.
The localization of RR but not retinal-loukoumasome subcomponents differs following ribavirin treatment.
A. Western blots demonstrate a marked increase in nuclear IMPDH1 in Y79 cells treated with 1 mM ribavirin for 24 hours. In contrast, ribavirin treatment did not affect beta-tubulin III or GAPDH subcellular localization. B. Densitometric analysis was performed on Western blots run in parallel and then probed with either beta-III tubulin, IMPDH1 or GAPDH. Protein expression was averaged and normalized to the untreated cytoplasmic fraction within each group. Data are represented as the mean ± SEM from 3 independent experiments. * p < 0.05, different from treated fraction in the same group. Significance was determined using an ANOVA and Fisher’s (LSD) test. C. Y79 cells were untreated (0 mM) or treated with 1 mM ribavirin. Cells were then immunostained for IMPDH and costained for DAPI to visualize RR. D. Orthogonal views of the same Y79 cells at a given z position are shown. While RR were not evident in controls, fluorescence intensities indicate that RR (IMPDH immunostaining) occur within the nucleus (DAPI positive) of some cells following ribavirin treatment.
Fig 7.
Retinal-loukoumasomes reside in the perinuclear space in close association with lamin in retinoblastoma cells.
A-C. The perimeter of retinal-loukoumasomes are outlined by intense lamin B1 staining suggesting that they are perinuclear. A. Y79 cells immunostained for beta-III tubulin using the SDL.3D10 antibody (green), lamin B1 (red) and costained for DAPI (blue). The z positions of single slices from the z-stack data are indicated. White arrows highlight retinal-loukoumasomes, while yellow arrows highlight intense lamin B1 staining. B. Tubulin, lamin B1 and DAPI intensity profiles on this same cell was performed. Fluorescence intensities indicate that tubulin nestles within a lamin fold near an area of the nucleus lacking DNA labelling. C. Confocal orthogonal views of the same Y79 cell at a given z position (z = 6.4 μm) are shown. Each successive orthogonal step (labelled 1–4) is 1.6 μm apart (arrows indicate the direction in which steps were taken). An intimate association with lamin B1 folds and the loukoumasome is evident. Polymerized tubulin “tubulin tail” appears to be continuous with the toroidal loukoumasome. D. A large loukoumasome is shown in a Weri-Rb 27 cell immunolabeled for beta-III tubulin together with lamin B1 and stained with DAPI. As seen for the loukoumasome in A-C, a beta-III tubulin “tail” is also evident here. E. The lamin foldings around loukoumasomes were also evident in Y79 cells coimmunolabeled for alpha-tubulin with lamin B1 and counterstained with DAPI. F. Representative images showing the interaction between beta-III tubulin and lamin B1 as revealed by the in situ PLA. Punctate red fluorescence indicates beta-III tubulin and lamin are in close proximity.
Fig 8.
MAP2 colocalizes with loukoumasomes in retinoblastoma cells and retinoblastoma tissue.
A-B. Y79 cells were immunolabeled for MAP2 alone (A) or together with beta-III tubulin using the SDL3D10 antibody (B) and costained with DAPI. C. Colocalization of beta-III tubulin and MAP2 in Y79 cells was also observed using the in situ PLA. Red fluorescent signals indicate that both beta-III tubulin and MAP2 are closely associated in Y79 cells. D. Retinoblastoma tissue was coimmunolabeled for beta-III tubulin using the SDL.3D10 antibody together with MAP2 and costained with DAPI (blue). The arrows depict retinal-loukoumasomes.