• 2018 NAS Awards Nominations Now Open
  • Sign-up for PNAS eTOC Alerts

Tissue plasminogen activator regulates Purkinje neuron development and survival

  1. Richard L. Sidmana,1
  1. Departments of aNeurology and
  2. cMedicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02215;
  3. bDepartment of Anatomy and Neurobiology, Boston University Medical School, Boston, MA 02118;
  4. dDepartment of Neurology and Neuroscience, Weill Medical College of Cornell University, New York, NY 10065;
  5. eDavid H. Koch Center, the University of Texas M. D. Anderson Cancer Center, Houston, TX 77030; and
  6. fProgram in Stem Cell and Regenerative Biology, Sanford-Burnham Medical Research Institute, La Jolla, CA 92037

  1. Contributed by Richard L. Sidman, March 29, 2013 (sent for review January 18, 2013)

  1. Fig. 1.
    View larger version:
    Fig. 1.

    tPA regulates PN dendritic development and PKCγ activity in the cerebellar cortex. (A) Calbindin-d-28K (Calb)-stained PN dendrites (red) in the molecular layer (ML) were decreased in P15 nr (labeled “nr”) and tPA-injected (1 µg/µL at P10) WT (labeled “tPA”) cerebellar cortices as compared with untreated WT controls (labeled ”WT“). PSL, PN soma layer. (Scale bar: 20 µm.) (B) Thickness of the molecular layer was reduced significantly in the P15 nr and tPA-injected WT mice compared with the untreated WT group. Values represent means ± SD, n = 4–6 for each group. *P < 0.05, **P < 0.01 compared with WT control. (C) Dendrite length, branch points, and branch ends of PNs in the above samples (also see Fig. S3B). Values represent means ± SD, n = 6–8 for each group. *P < 0.05, **P < 0.01 compared with WT control. (D) Sholl analysis of the total length of PN dendrites within each concentric circle (0–10 µm, 10–20 µm, … 110–120 µm) in the above samples. The area under each curve represents an average value of the total PN dendrite length in each group. Peak values correlated positively with dendritic density. (E) PKCγ mRNA at P20, detected by PKCγ signal intensity and mRNA level, were expressed predominantly in P20 WT PNs, rather than in granule cells (GC). (F) Representative images of PKCγ protein (green) appeared similar in P15 WT, nr, and tPA-injected PNs. Cell nuclei were DAPI-stained (blue). (Scale bar: 10 μm.) (G) PKCγ protein intensity also was similar in PNs of P15 WT, nr, and tPA-injected mice. Values represent means ± SD, n = 4–6 for each group. (H) Protein levels of phosphorylated MARCKS (P-MARCKS) were increased significantly in P15 nr and tPA-injected WT cerebella compared with WT controls. Values represent means ± SD, n = 4–6 for each group. **P < 0.01 compared with WT control.

  2. Fig. 2.
    View larger version:
    Fig. 2.

    tPA and PKCγ mediate PN dendritic growth in cerebellar dissociated cell cultures. (A) Treatment with tPA+PL (20+5 µg/mL at 8–14 DIV) or PKC agonist, rather than a VDAC blocker or TrkB IgG, inhibited calbindin-d-28K (Calb)-stained PN dendrites in cerebellar cell cultures at 14 DIV. PKC inhibitor, rather than VDAC1 or BDNF, rectified the inhibitory effect of tPA+PL treatment. (Scale bar: 20 µm.) (B) Quantification of dendrite length, branch points, and branch ends of PNs in the above samples (also see Fig. S3B). Lv, lentiviral vector. Values represent means ± SD, n = 6–10 for each treatment. *P < 0.05, **P < 0.01 compared with control. (C) Sholl analysis of total length of PN dendrites in the above samples, as in Fig. 1D. Peak values correlated positively with dendritic density. (D and E) Effects of tPA+PL, a PKC agonist, or a PKC inhibitor, respectively, on PKC activity (P-MARCKS level) and on the levels of several important proteins (D) and of MAP2 and phosphorylated MAP2 (P-MAP2) (E) in the above samples. Treatment with tPA+PL or PKC agonist increased phosphorylation of MARCKS and MAP2 and suppressed PN dendritic growth, whereas the PKC inhibitor rectified these effects of tPA+PL treatment. Values represent means ± SD, n = 6–8 for each treatment. *P < 0.05, **P < 0.01 compared with control.

  3. Fig. 3.
    View larger version:
    Fig. 3.

    tPA affects the PN synaptic ultrastructure and the BDNF level in the cerebellar cortex. (A) Transmission electron microscopy disclosed ultrastructural changes in asymmetric PF–PN synapses in the distal part of the molecular layer of P15 nr and tPA-injected (1 µg/µL at P10) WT cerebellar cortices compared with untreated WT controls. The heavily stained segment between two arrows shows the postsynaptic density. High mag, high magnification; Low mag, low magnification; pf, parallel fiber presynaptic terminals; s, PN dendrite spine. (Scale bars: 1,000 nm for low magnification; 500 nm for high magnification.) (B) Synapse number, presynaptic vesicle number, and postsynaptic density (PSD) length of the above samples were decreased significantly in P15 nr and tPA-injected groups compared with the untreated WT group. Values represent means ± SD, n = 11–56 for each group. *P < 0.05, **P < 0.01 compared with control. (C) BDNF (ligand) mRNA was found mainly in granule cells (GC), whereas TrkB (receptor) mRNA was found mainly in the PNs of P20 WT mouse cerebellum, detected by LCM/Chip (signal intensity) and LCM/qPCR (mRNA level). (D) As evident in the gel bands (Upper) and the bar graph (Lower), the protein level of BDNF, rather than TrkB, was decreased in P15 nr and tPA-treated cerebella as compared with WT controls. Values represent means ± SD, n = 4–6 for each group. *P < 0.05, **P < 0.01 compared with control. (E and F) Representative images (E) and quantification (F) showed that BDNF distribution (green in E) was reduced in the small residual external granular layer (shown in two apposed lobules, between the dashed lines) of P15 nr and tPA-treated cerebella as compared with WT controls. ML, molecular layers of the apposed lobules at the top and bottom of the images. Cell nuclei were stained with DAPI (blue). Values represent means ± SD, n = 4–6 for each group. **P < 0.01 compared with control. (Scale bar: 20 µm.)

  4. Fig. 4.
    View larger version:
    Fig. 4.

    tPA and BDNF regulate PN synapse formation in cerebellar dissociated cell cultures. (A) qPCR data (designations for each number along the x-axis are listed at the bottom of the figure) showed that tPA+PL (#4) (20+5 µg/mL at 8–14 DIV) or TrkB IgG (#11) significantly reduced postsynaptic density 95 (PSD95) mRNA levels at 14 DIV. BDNF (#12), rather than PKC inhibitor (#7) or VDAC1 (#10), rectified the inhibitory effect of tPA+PL treatment. Lv, lentiviral vector. Values represent means ± SD, n = 6–8 for each treatment. *P < 0.05, **P < 0.01 compared with control. (B) Representative images of calbindin-d-28K (Calb)-stained PN dendritic branchlets (red) and PSD95-stained postsynaptic sites (green) showed that tPA+PL or TrkB IgG decreased PSD95 staining in dissociated cell cultures at 14 DIV and that the addition of BDNF to tPA+PL corrected the concentration of PSD95-stained postsynaptic sites, although, as expected, it did not reverse the tPA+PL-induced reduction in dendritic branchlets. (Scale bar: 5 µm.) (C) Quantification of the intensities of Calb and PSD95 staining (from samples in B) and of the PSD95 (postsynaptic site)/Calb (dendritic branchlet) ratio supported the notion that blockage of BDNF/TrkB signaling by excess tPA/plasmin proteolysis inhibited formation of PF–PN synapses in vitro. Values represent means ± SD, n = 6–10 for each treatment. **P < 0.01 compared with control.

  5. Fig. 5.
    View larger version:
    Fig. 5.

    tPA influences PN mitochondria and VDAC in cerebellar cortex. (A and B) Transmission electron microscopy showed huge, round mitochondria (A) with increased diameters (B) in P15 nr and tPA-injected (5 µg/µL at P10) WT cerebellar PNs compared with untreated WT controls. mito, mitochondria. Values represent means ± SD, n = 20 mitochondria for each group. *P < 0.05, **P < 0.01 compared with control. No significant difference was seen between nr and tPA-injected PNs. (Scale bar: 400 nm.) (C) Intracellular distribution of VDAC1 in both mitochondria (MT) and plasma membrane (PM) fractions of the cell lysate (CL). Protein levels of VDAC1 and other organelle markers were measured in the cell lysate, plasma membrane, and mitochondria of cultured PNs. Na-K ATPase, a cell membrane marker, was detected in the plasma membrane. Calreticulin, an endoplasmic reticulum marker, was not detected in plasma membrane or mitochondria; cyclooxygenase IV (COX IV), a mitochondrial inner membrane marker, was found in mitochondria. The results verified the high purity of the cellular fractions and the presence of VDAC1 in both mitochondria and plasma membrane. (D) VDAC1 mRNA was five- to10-fold greater than VDAC2 or VDAC3 mRNA in PNs and granule cells (GC) in P20 WT mice as detected by LCM/Chip and LCM/qPCR. (E and F) Representative images (E) and quantification (F) showed that VDAC1 distribution (green in E) was decreased in the molecular layer (ML) but was increased in the PN soma layer (PSL) of P15 nr and tPA-injected cerebellar cortices as compared with WT controls. Cell nuclei were stained with DAPI (blue). Values represent means ± SD, n = 4–6 for each group. **P < 0.01 compared with control. (Scale bar: 15 µm.) (G) Plasminogen kringle 5 levels were increased significantly in P15 nr and tPA-treated cerebella compared with WT controls. Values represent means ± SD, n = 4–6 for each group, **P < 0.01 compared with control.

  6. Fig. 6.
    View larger version:
    Fig. 6.

    tPA and VDAC regulate PN mitochondrial morphology and function in cerebellar dissociated cell cultures. (A) Representative images of JC-10 dye-stained PN cultures showed normal ΔΨm (red) in untreated control cultures and abnormally reduced ΔΨm (green) in cultures treated with tPA+PL (100+25 µg/mL) at 8–14 DIV or with VDAC blocker at 14 DIV. Reversible formation of JC-10 aggregates was based on membrane polarization that caused shifts in emitted light from 520 nm (JC-10 monomer emission) to 570 nm (JC-10 aggregate emission). In normal cells, JC-10 concentrated in the mitochondrial matrix where it formed red fluorescent aggregates at high Δψm, but in apoptotic and necrotic cells with low mitochondrial Δψm JC-10 diffused from mitochondria in monomeric form and fluoresced green. (Scale bar: 50 µm.) (B) Quantification confirmed that tPA+PL or VDAC blocker significantly decreased ΔΨm, whereas the addition of VDAC1 to tPA+PL corrected ΔΨm. Numbers under the x-axis are defined at the bottom of the figure. Lv, lentiviral vector. Values represent means ± SD, n = 6–10 for each treatment. *P < 0.05, **P < 0.01 compared with control. (C and D) Mitochondrial morphology (C) and diameters (D) confirmed that tPA+PL or VDAC blocker induced mitochondria to become huge and round, as seen here in CMXRos-stained PNs at 14 DIV. Values represent means ± SD, n = 20 mitochondria. *P < 0.05, **P < 0.01 compared with control. (Scale bar: 2 µm.) (E) tPA+PL or VDAC blocker also significantly inhibited ATP levels in cultured PNs. Values represent means ± SD, n = 6–8 for each treatment. *P < 0.05, **P < 0.01 compared with control.

  7. Fig. 7.
    View larger version:
    Fig. 7.

    tPA and VDAC mediate caspase 3/7-independent PN death. (A and B) Representative images (A) and quantification (B) showed reduced survival of calbindin-d-28K (Calb)-stained PNs (red in A) in P20 tPA-injected (5 µg/µL at P10) WT cerebellar cortex compared with untreated or sham-injected WT controls. Cell nuclei are stained with DAPI (blue). PSL, PN soma layer. Values represent means ± SD, n = 4–6 for each group. *P < 0.05, **P < 0.01 compared with control. (Scale bar: 15 µm.) (C and D) Representative images (C) and quantification (D) of Calb-stained (red) and caspase 3 (Casp3)-stained (green in C) sections showed no double-stained PNs in P20 WT, nr, or tPA-injected cerebellar cortices. Values represent means ± SD, n = 6–8 for each group. (Scale bar: 30 µm.) (E and F) Representative images (E) and quantification (F) of Calb-stained PNs (red in E) indicated that tPA+PL (100+25 μg/mL at 8–14 DIV) or VDAC blocker, rather than PKC agonist or TrkB IgG, increased PN death in cerebellar dissociated cell cultures at 21 DIV. Lv, lentiviral vector. Values represent means ± SD, n = 6–8 for each treatment. *P < 0.05, **P < 0.01 compared with control. (Scale bar: 50 µm.) (G and H) Quantification of caspase 3/7 activity (G) and plasma membrane integrity (LDH assay) (H) suggested that tPA+PL or VDAC blocker did not change the caspase 3/7 level but did increase LDH release in cerebellar cell cultures at 21 DIV. Values represent means ± SD, n = 6–8 for each treatment. *P < 0.05, **P < 0.01 compared with control. (I) Correlation analysis of plasma membrane integrity (positive correlation, P < 0.01) or caspase 3/7 activity (no correlation, P > 0.05) with PN survival. (J) Correlation analysis of ΔΨm (positive correlation, P < 0.01) or mitochondrial diameter (negative correlation, P < 0.01) with PN survival.

  8. Fig. 8.
    View larger version:
    Fig. 8.

    Protection of nr PNs by tPA deletion and the proteolytic pathways controlling nr PN development. (A–C) Representative images (A) of calbindin-d-28K (Calb)-stained PNs (red in A), quantification in the sum of all cerebellar cortical lobules (B), and motor coordination behavior measured by rotarod test (C) in P40 and P120 WT, tPA−/− (tPA−/−), nr, and nr;tPA−/− (double mutant) cerebellar cortices. The results indicated that deletion of endogenous tPA in nr;tPA−/− double mutants significantly increased PN survival and restored motor coordination behavior of nr mice. Values represent means ± SD, n = 4–6 for each group. *P < 0.05, **P < 0.01 compared with control. (Scale bar: 100 µm.) (D) Diagram summarizing our overall view of tPA-based proteolytic pathways controlling PN dendrite and synapse development and the mitochondrial pathology before necrosis. Excess tPA produced by PNs and granule cells (GCs) in young nr cerebella activates PKCγ and inactivates MAP2, inhibiting PN dendritic growth. Excess tPA/plasmin proteolysis degrades granule cell-derived BDNF, decreasing BDNF/TrkB signaling in PNs and in turn impairing the formation and structure of PF–PN synapses. Excess tPA increases binding of the plasminogen catabolite kringle 5 to VDAC1, and the modified VDAC1 further alters mitochondrial structure and function.

Online Impact

    Related Commentary

    Commentary - Biological Sciences - Neuroscience:
    • Dan Goldowitz
    Excessive activation of tissue plasminogen activator makes a mouse nervous PNAS 2013 110 (27) 10882-10883; published ahead of print June 21, 2013, doi:10.1073/pnas.1309965110