Pak1 and Pak2 are activated in recurrent respiratory papillomas, contributing to one pathway of Rac1-mediated COX-2 expression

Abstract

Recurrent respiratory papillomas are premalignant tumors of the airway caused by human papillomaviruses (HPVs), primarily Types 6 and 11. We had reported that respiratory papillomas overexpress the epidermal growth factor receptor (EGFR), the small GTPase Rac1 and cyclooxygenase-2 (COX-2), and have enhanced nuclear factor-kappaB (NFkappaB) activation with decreased levels of IkappaB-beta but not IkappaB-alpha. We also showed that EGFR-activated Rac1 mediates expression of COX-2 through activation of p38 mitogen-activated protein kinase. We have now asked whether the p21-activated kinases Pak1 or Pak2 mediate activation of p38 by Rac1 in papilloma cells. Pak1 and Pak2 were constitutively activated in vivo in papilloma tissue compared with normal epithelium, and Rac1 siRNA reduced the level of both phospho-Pak1 and phospho-Pak2 in cultured papilloma cells. Reduction in Pak1 and Pak2 with siRNA decreased the COX-2 expression in papilloma cells, increased the levels of IkappaB-beta and reduced the nuclear localization of NF-kappaB, but had no effect on p38 phosphorylation. Our studies suggest that Rac1 --> Pak1/Pak2 --> NFkappaB is a separate pathway that contributes to the expression of COX-2 in HPV-induced papillomas, independently of the previously described Rac1 --> p38 --> COX-2 pathway.

Figure 1. Pak1 and Pak2 are highly phosphorylated in recurrent respiratory papilloma tissues and cultured cells

a. Representative immunohistochemical staining of formalin-fixed, paraffin-embedded normal laryngeal and papilloma tissues, showing abundant phosphorylated Pak1/2. Arrows indicate location of the basement membrane, bar = 24 µm. Three papillomas and two normal tissues were analyzed. b. Western blot showing elevation of phosphorylated Pak1 and Pak2 in papilloma tissues compared to normal tissue. Bar graph shows the mean ± SD of the ratio of phosphorylated Pak1 and Pak2 to total Pak 1 and Pak2 in 5 papillomas, normalized to the mean ratio of 4 normal tissues (*: p<0.01). c. Representative western blot showing steady-state levels of phospho-Pak1 and Pak2 in papilloma cells and normal laryngeal cells cultured in serum-free medium containing EGF and insulin. The bar graph shows the mean ± SD of the ratio of phosphorylated to total protein, normalized to actin and expressed relative to normal cells, from 3 separate experiments (*: p<0.01).

Figure 2. Activation of Pak1 and Pak2 requires Rac1

Normal and papilloma cells were transfected with Rac1-specific siRNA and levels of phospho-Pak1, phospho-Pak2, Pak1, Pak2, Rac1, COX-2 and β-actin determined by western blot after 72 hrs. Luciferase (Luc) siRNA was used as a control. The bar graph shows the mean ± SD of the ratios of phosphorylated to total Pak1 and Pak2, normalized to actin and expressed relative to the ratios in normal cells treated with luciferase siRNA (*: p<0.01).

Figure 3. Pak1 and Pak2 signaling mediates COX-2 expression in papilloma cells via activation of IκBβ, not activation of p38

Western blots showing effects of specific siRNAs targeting either Pak1 or Pak2 on levels of COX-2, phosphorylated and total p38 and IκBβ in normal and papilloma cells. Luciferase (Luc) siRNA was used as a negative control. Levels of Pak 1 and Pak2 were measured to assess knockdown, actin measured to confirm equal loading. a. Representative western blot and bar graphs showing mean ± SD of 3 experiments with papilloma and normal cells, with levels normalized to actin and expressed relative to control cells of each type treated with luciferase siRNA (* p<0.01). b. Western blot of parallel experiment on papilloma cells using a second set of Pak1 and Pak 2 siRNAs (Pak1' and Pak2'), with the associated bar graph showing confirmation of minimal effects on relative phosphorylation of p38. c. Western blots and associated bar graph of papilloma cells treated with combined Pak1 and Pak2 siRNAs, and cells treated with the GroupI Pak1 inhibitor IPA-3, to assess effects on COX-2 levels of simultaneously reducing both Pak1 and Pak2 signaling (* p<0.05).

Figure 3. Pak1 and Pak2 signaling mediates COX-2 expression in papilloma cells via activation of IκBβ, not activation of p38

Western blots showing effects of specific siRNAs targeting either Pak1 or Pak2 on levels of COX-2, phosphorylated and total p38 and IκBβ in normal and papilloma cells. Luciferase (Luc) siRNA was used as a negative control. Levels of Pak 1 and Pak2 were measured to assess knockdown, actin measured to confirm equal loading. a. Representative western blot and bar graphs showing mean ± SD of 3 experiments with papilloma and normal cells, with levels normalized to actin and expressed relative to control cells of each type treated with luciferase siRNA (* p<0.01). b. Western blot of parallel experiment on papilloma cells using a second set of Pak1 and Pak 2 siRNAs (Pak1' and Pak2'), with the associated bar graph showing confirmation of minimal effects on relative phosphorylation of p38. c. Western blots and associated bar graph of papilloma cells treated with combined Pak1 and Pak2 siRNAs, and cells treated with the GroupI Pak1 inhibitor IPA-3, to assess effects on COX-2 levels of simultaneously reducing both Pak1 and Pak2 signaling (* p<0.05).

Figure 3. Pak1 and Pak2 signaling mediates COX-2 expression in papilloma cells via activation of IκBβ, not activation of p38

Western blots showing effects of specific siRNAs targeting either Pak1 or Pak2 on levels of COX-2, phosphorylated and total p38 and IκBβ in normal and papilloma cells. Luciferase (Luc) siRNA was used as a negative control. Levels of Pak 1 and Pak2 were measured to assess knockdown, actin measured to confirm equal loading. a. Representative western blot and bar graphs showing mean ± SD of 3 experiments with papilloma and normal cells, with levels normalized to actin and expressed relative to control cells of each type treated with luciferase siRNA (* p<0.01). b. Western blot of parallel experiment on papilloma cells using a second set of Pak1 and Pak 2 siRNAs (Pak1' and Pak2'), with the associated bar graph showing confirmation of minimal effects on relative phosphorylation of p38. c. Western blots and associated bar graph of papilloma cells treated with combined Pak1 and Pak2 siRNAs, and cells treated with the GroupI Pak1 inhibitor IPA-3, to assess effects on COX-2 levels of simultaneously reducing both Pak1 and Pak2 signaling (* p<0.05).

Figure 4. Reduction in Pak1 or Pak2 levels reduce nuclear NFκB and elevate cytoplasmic and nuclear IκBβ in HPV6/11-infected papilloma cells

a. Immunohistochemistry of papilloma tissues and normal laryngeal epithelium, showing high levels of NFκB p50 and minimal IκBβ in papilloma tissues. b. Western blot of tissue extracts, confirming high levels of NFκB p50 protein and limited IκBβ in papillomas. Bar graph shows mean ± SD of multiple samples of each type, normalized to actin and expressed relative to normal tissue. c. Representative effects of Rac1 siRNA on localization of NFκB p50 in papilloma cells, assessed by immunohistochemistry. Luciferase siRNA was used as control. d. Representative effects of Pak1 and Pak 2 siRNAs on abundance and localization of NFκB p50 and IκBβ in papilloma cells. Luciferase siRNA (Luc) was used as a control. Bar is 24 µm. The experiments in 4c and 4d were repeated 3 times.