Increased levels of survivin, via association with heat shock protein 90, in mucosal T cells from patients with Crohn's disease

Abstract

Background & aims: Defective apoptosis of lamina propria T cells (LPTs) is involved in the pathogenesis of Crohn's disease. Survivin, a member of the inhibitors of apoptosis family, prevents cell death and regulates cell division. Survivin has been studied extensively in cancer, but little is known about its role in Crohn's disease.

Methods: LPTs were isolated from mucosal samples of patients with Crohn's disease or ulcerative colitis and healthy individuals (controls). LPTs were activated with interleukin-2 or via CD3, CD2, and CD28 signaling, and cultured at 42°C to induce heat shock. Survivin expression was assessed by immunohistochemistry, confocal microscopy, and immunoblotting; survivin levels were reduced by RNA interference. Cell viability, apoptosis, and proliferation were measured by trypan blue exclusion, annexin-V/7-Aminoactinomycin D staining, and uptake of [3]thymidine, respectively.

Results: LPTs from patients with Crohn's disease had higher levels of survivin than LPTs from patients with ulcerative colitis or controls. RNA knockdown of survivin in LPTs inhibited their proliferation and promoted apoptosis. Levels of survivin were low in LPTs from patients with ulcerative colitis and controls as a result of ubiquitin-mediated proteasome degradation. In LPTs from patients with Crohn's disease, survivin bound to the heat shock protein (HSP)90, and therefore was resistant to proteasome degradation. Incubating LPTs with 17-N-allylamino-17-demethoxygeldanamycin, an inhibitor of HSP90, reduced levels of survivin and induced apoptosis.

Conclusions: Levels of survivin are increased in LPTs from patients with Crohn's disease (compared with ulcerative colitis and controls) because survivin interacts with HSP90 and prevents proteasome degradation. This allows LPTs to avoid apoptosis. Strategies to restore apoptosis to these cells might be developed to treat patients with Crohn's disease.

Figure 1

Distribution and quantification of survivin-expressing cells in colonic mucosa and colon cancer. (A) Survivin-positive cells (dark brown) in the crypts of normal colonic epithelium but not the lamina propria; marked increase of survivin-positive cells (arrows) in CD lamina propria; modest increase in survivin-positive cells (arrows) in UC lamina propria. (B) Increased numbers of survivin-positive cells in inflamed CD, noninflamed CD, and inflamed UC mucosa compared with normal control mucosa. *P < .001 for inflamed CD compared with control and UC. **P < .01 for inflamed CD compared with noninflamed CD, and UC compared with normal mucosa. ***P < .004 for noninflamed CD compared with control (groups included 11 control, 10 inflamed CD, 10 noninflamed CD, and 10 UC). (C) From top to bottom: CD3-positive lymphocytes (red) in the lamina propria and survivin-positive crypt cells (green) in normal mucosa; abundant CD3 (red) and survivin (green) co-expressing lymphocytes infiltrating colonic CD; CD3-positive lymphocytes (red) in the lamina propria and survivin-positive crypt cells (green) in UC mucosa; CD3-positive (red), survivin-negative lymphocytes infiltrating survivin-positive neoplastic cells (green) in colon adenocarcinoma. Nuclei were stained with 4′,6-diamidino-2-phenylindole (blue). Representative of 10–14 samples for normal and IBD mucosa and 3 colon cancers.

Figure 2

Levels of survivin protein in fresh, activated, and rested LPTs. (A) Increased survivin and p-survivin in freshly isolated CD compared with control and UC LPTs (P < .001 and P < .04, respectively). (B) Increased nuclear p-survivin in freshly isolated UC and CD compared with control LPTs (P < .05). (C) Up-regulation of total and p-survivin in control, CD, and UC LPTs upon exposure to IL-2 (P < .02–.001 compared with day 0). (D) Subcellular distribution of survivin in IL-2–activated LPTs. Increased survivin in activated UC and CD LPT cells compared with control LPTs. (E) Up-regulation of survivin in control, CD, and UC LPTs upon activation with anti-CD3/CD28 or CD2/CD28 (P < .05 and P < .04 compared with day 0, respectively). (F) Spontaneous increase of survivin in fresh CD LPTs and disappearance of survivin in control and UC LPTs cultured in the absence of IL-2 (P < .03 and P < .02, respectively). GAPDH used as loading control for total and cytoplasmic survivin, and lamin B1 for the nuclear fraction. Immunoblots are representative of 12–18 samples in each control, CD, and UC group; bands represent a loading of 10 μg protein lysate/lane regardless of the total (T), cytoplasmic (C), or nuclear (N) origin of survivin. P values were calculated using the ratio of survivin to GAPDH or lamin B1 obtained by densitometric analysis and comparing the ratio values among control CD and UC.

Figure 3

Effect of survivin and HSP90 silencing on protein levels, proliferation, and apoptosis of LPTs. (A) Immunoblot showing the effect of survivin and scrambled siRNA on protein levels of IL-2–stimulated control, CD, and UC LPTs. Bcl-2 was used as specificity control and GAPDH was used as loading control (figure is representative of 4 experiments). (B) Upper panel: time-dependent suppression of anti-CD3/CD28 plus IL-2–induced LPT proliferation upon treatment with survivin siRNA (each column indicates 1 control, 1 CD, and 1 UC LPT isolate). Lower panel: lack of significant effect on anti-CD3/CD28 plus IL-2–induced LPT proliferation by treatment with HSP90 siRNA (each column indicates 1 control, 1 CD, and 1 UC LPT isolate). Percentage of suppression compared with that of scrambled siRNA. (C) Increased LPT apoptosis induced by survivin, but not HSP90, siRNA compared with scrambled siRNA (*P < .012; n = 3 for control, CD and UC, respectively, for both survivin and HSP90 siRNA). (D) Flow cytometric analysis of increased LPT apoptosis induced by survivin siRNA as assessed by annexin-V/7-Aminoactinomycin D. Red indicates double-negative cells and green indicates annexin-V–positive cells (representative of 3 experiments).

Figure 4

Effect of inhibiting proteasomal activity on time-dependent expression of survivin in LPTs. Western blots of IL-2–activated LPT extracts showing results with 3 different anti-survivin antibodies recognizing distinct survivin epitopes (anti–survivin-1, -2, and -3), one antibody against Bcl-2, and reciprocal immunoprecipitation of ubiquitin and survivin probed with antibodies against survivin (anti-survivin 3) and ubiquitin, respectively. Anti-survivin 1 is a 6E4 mouse monoclonal antibody, anti-survivin 2 is a mouse monoclonal antibody clone 60.11, and anti-survivin 3 used for immunoprecipitation (IP) is a rabbit polyclonal antibody (Supplementary Materials). Antibodies directed at distinct survivin epitopes ensured that all detected bands did contain survivin. Results displayed derive from a CD LPT isolate. Dimethyl sulfoxide (DMSO) is the solvent for the proteasomal inhibitor MG132, and E64 is a protease inhibitor used as control for MG132. GAPDH was used as loading control. Figure is representative of 1 control, 2 CD, and 2 UC LPTs.

Figure 5

Proteasomal activity of LPT and survivin sensitivity to proteolysis. (A) Immunoprecipitation of survivin in fresh control, CD, and UC LPT lysates followed by immunoblotting with an anti-ubiquitin antibody. Figure is representative of 6–8 samples in each group. (B) Differential levels of proteasomal activity in fresh control, CD, and UC LPTs. *P < .04 for CD compared with control LPT activity, and for degree of lactacystin-induced inhibition in CD compared with UC LPT (n = 17 for control, CD, and UC, respectively). (C and D) Effect of proteinase K activity on survivin levels in cytosolic and nuclear fractions of IL-2–stimulated LPTs. Immunoblots showing a dose-dependent decrease of survivin in the cytosolic fraction of control and UC, but not CD LPT lysates (*P < .02, **P < .004, by densitometry of the survivin:lamin B1, survivin:γ-actin, and Bcl-2:γ-actin ratios, respectively), and a dose-dependent decrease of survivin in the nuclear fraction of the same lysates. Immunoblots of the same lysates show a dose-dependent decrease in the cytosolic fraction of Bcl-2, which was used as control. Figure is representative of 4 experiments.

Figure 6

Association of survivin with HSP90 in LPTs and dissociating effect of 17-AAG. (A) Reciprocal immunoprecipitation of survivin and HSP90 in fresh control, CD, and UC LPT lysates. Immunoblots showing immunoprecipitated survivin probed with an anti-HSP90 antibody, and immunoprecipitated HSP90 probed with an anti-survivin antibody. Figure is representative of 6 experiments. (B) Effect of heat shock on expression of survivin, HSP90, and Bcl-2 by LPTs. Immunoblots showing time-dependent increase of survivin and HSP90 and a decrease of Bcl-2 in fresh control, CD, and UC LPTs. GAPDH was used as loading control. Figure is representative of 4 control, CD, and UC LPTs, respectively. (C) Effect of inhibition of HSP90 chaperone activity on LPT survivin levels. Immunoblots showing the dose-dependent effect of 17-AAG on survivin levels of IL-2–stimulated control, CD, and UC LPTs. MG132 was used as control to block survivin degradation through the proteasomal pathway, and dimethyl sulfoxide (DMSO) was used as solvent control for both 17-AAG and MG132. γ-actin was used as a loading control. Figure is representative of 3 control, CD, and UC LPTs, respectively. (D) Dose-dependent increase of control, CD, and UC LPT apoptosis induced by 17-AAG (n = 3 control, CD, and UC LPTs).

Figure 7

Association of survivin with HSP90 in LPTs. Confocal microscopy of LPT preparations showing co-localization of survivin (green) and HSP90 (red) in fresh, heat-shocked (6 hours), IL-2–stimulated (7–10 days), and alone (absence of IL-2 for 7 days) control, CD, and UC LPTs. Co-localization of survivin and HSP90 is indicated by the orange color generated by the combination of green (survivin) and red (HSP90) colors. Micrograph panel is representative of 5–10 isolates in each group.