Cell Surface Trafficking of Fas: A Rapid Mechanism of p53-Mediated Apoptosis

See allHide authors and affiliations

Science  09 Oct 1998:
Vol. 282, Issue 5387, pp. 290-293
DOI: 10.1126/science.282.5387.290


p53 acts as a tumor suppressor by inducing both growth arrest and apoptosis. p53-induced apoptosis can occur without new RNA synthesis through an unknown mechanism. In human vascular smooth muscle cells, p53 activation transiently increased surface Fas (CD95) expression by transport from the Golgi complex. Golgi disruption blocked both p53-induced surface Fas expression and apoptosis. p53 also induced Fas-FADD binding and transiently sensitized cells to Fas-induced apoptosis. In contrast, lpr and gldfibroblasts were resistant to p53-induced apoptosis. Thus, p53 can mediate apoptosis through Fas transport from cytoplasmic stores.

p53 is the most commonly mutated gene in human cancer (1). p53 is a sequence-specific transcription factor, whose transcriptional targets induce growth arrest and apoptosis (2). Although its tumor suppressor function requires both activities, some human tumor–derived p53 mutants transactivate p53-responsive promoters and induce growth arrest, implying that apoptosis is the more potent mechanism (3). Depending on cell type, p53-induced apoptosis either requires transcriptional activation (4) or occurs without new RNA and protein synthesis (5). The occurrence of mutants that transactivate p53 targets but are defective for apoptosis (6), or vice versa (7), suggests that p53 induces apoptosis through transactivation-dependent and -independent mechanisms, implying a structural and functional separation between the ability to induce growth arrest or apoptosis.

We examined p53-mediated apoptosis in untransformed human vascular smooth muscle cells (VSMCs) by expressing a conditional allele of p53, p53ERTM (8), encoding a chimeric full-length human p53 fused to a 4-hydroxytamoxifen (4-OHT)–sensitive estrogen receptor (9). 4-OHT addition to p53ERTMVSMCs rapidly activates p53 (10), caused translocation of perinuclear p53 to the nucleus (Fig. 1, A and B), and induced apoptosis (Fig. 1C), with apoptotic morphology first appearing within 60 min. Apoptosis was not inhibited by preincubation with actinomycin D (Act-D) or cycloheximide (CHX) at concentrations that completely block RNA or protein synthesis (10) (Fig. 1D). Neither Act-D nor CHX alone induced apoptosis (Fig. 1D), indicating that apoptosis of p53ERTMVSMCs was transcription-independent.

Figure 1

(A andB) p53 translocates to the nucleus upon activation and induces apoptosis. p53ERTM VSMCs were stained with mouse IgG to human p53 (500 ng/ml) (#14211A, Pharmingen) and FITC-conjugated anti-mouse IgG. (A) Control. (B) Thirty minutes after 4-OHT addition. (C) p53ERTM VSMCs 2 hours after p53 activation, showing membrane blebs (white arrow) and apoptotic body formation (black arrow). (D) p53ERTM VSMCs were transferred to 0% fetal calf serum (FCS) medium and 100 nM 4-OHT either alone or 1 hour after addition of Act-D (2 μg/ml) or CHX (10 μg/ml). Cells were observed by videomicroscopy and cumulative deaths were plotted against time (18).

Apoptosis through Fas (CD95), a tumor necrosis factor receptor (TNF-R) family member, is rapid and independent of new RNA or protein synthesis (11). Fas ligand (FasL)–Fas binding recruits an adapter molecule, FADD, through shared protein motifs (“death domains”), with resultant caspase activation leading to apoptosis. We therefore analyzed Fas, FADD, and FasL expression and Fas-FADD binding after p53 activation. No increases in Fas, FADD, or FasL protein were observed (Fig. 2A). However, FADD coimmunoprecipitated with Fas 30 to 60 min after p53 activation, but Fas-FADD binding disappeared by 2 hours.

Figure 2

(A) Effect of p53 on Fas-FADD binding. Protein immunoblots of lysates from p53ERTM VSMCs 0 to 120 min after 4-OHT addition were probed with mouse IgG to human Fas (100 ng/ml), FADD (1 μg/ml), or FasL (1 μg/ml) [#F22120 and #F36620 (Transduction Labs) and NOK-1 (#65320C, Pharmingen), respectively]. J, Jurkat cells. (Bottom row) Immunoprecipitation of Fas-FADD complexes isolated 0 to 120 min after 4-OHT addition (19). (B) Flow cytometric analysis of Fas in p53ERTM VSMCs or Jurkat cells. Profiles obtained with mouse anti-Fas (CH-11, Upstate Biotechnology) (black) are shown relative to an isotypically matched mouse IgM. (C to H) Golgi complex localization of Fas in VSMCs. p53ERTM VSMCs were stained with CH-11 (500 ng/ml) (C) or a rabbit polyclonal antibody to Fas (500 ng/ml) (Fas Ab-1, Oncogene Research) (D), followed by staining with FITC-conjugated anti-mouse or anti-rabbit IgG. Nuclei were counterstained with propidium iodide. (E) Initial treatment with BFA (5 μg/ml) for 1 hour produced weak, punctate cytoplasmic Fas staining, including at the microtubular organizing center (inset). (F to H) p53ERTM VSMCs were stained with Fas Ab-1 and a mouse monoclonal antibody to galactosyl transferase, then with Texas Red–conjugated anti-mouse IgG and FITC-conjugated anti-rabbit IgG. (F) Galactosyl transferase (red), (G) Fas (green), (H) galactosyl transferase and Fas superimposed (overlap is yellow). N, cell nuclei. (I) Subcellular localization of Fas. L, whole-cell lysate; H, homogenate after mechanical disruption before centrifugation; N, nuclear fraction; G, Golgi-enriched fraction (20).

We next examined both surface and total Fas expression in VSMCs. VSMCs expressed little surface Fas, compared with Jurkat cells (Fig. 2B), but total Fas (in permeabilized cells) was much greater than surface expression, indicating that Fas was predominantly intracellular (12). Fas immunocytochemistry demonstrated a compact, perinuclear distribution that colocalized with the Golgi marker galactosyl transferase (Fig. 2, C to H). Brefeldin A (BFA), a protein-secretion inhibitor that redistributes Golgi proteins to the endoplasmic reticulum (ER) or near the microtubular organizing center (13), disrupted Fas staining to produce an ER-like pattern (Fig. 2E), with fluorescent spots abutting the nucleus (Fig. 2E, inset). Thus, Fas localized to the Golgi complex and trans-Golgi network. Subfractionation of VSMC homogenates and protein immunoblotting confirmed Fas in the Golgi-enriched fraction (Fig. 2I).

To examine p53 regulation of Fas distribution, we measured surface Fas by flow cytometry after p53 activation. Activation of wild-type (WT) but not a dominant-negative p53 (DN-p53ERTM) (14) transiently increased surface Fas (maximum 1 hour, baseline by 2 hours) (Fig. 3, A and B). Cells expressing the ERTM vector alone were also ineffective. BFA, but not Act-D or CHX, blocked p53-induced increases in surface Fas (Fig. 3, C to E), although Act-D or CHX alone did not increase surface Fas. p53 activation, BFA, CHX, or Act-D did not increase total Fas expression (Fig. 3F) over 2 hours, indicating that p53 caused cell surface redistribution of cytoplasmic Fas.

Figure 3

Effect of p53 on surface Fas expression. 4-OHT was added to (A) p53ERTM VSMCs or (B) DN-p53ERTM VSMCs, and surface Fas examined over 0 to 120 min by flow cytometry. p53ERTMVSMCs were incubated in BFA (5 μg/ml) (C), Act-D (2 μg/ml) (D), or CHX (10 μg/ml) (E) for 60 min, then 4-OHT was added and cells were isolated over the subsequent 120 min. (F) For examination of total Fas expression, p53ERTM VSMCs were unpermeabilized (Un) or treated with 4-OHT, BFA, CHX, or Act-D for 60 min and permeabilized before Fas staining. (G and H) Etoposide (5 μM) was added to p53ERTM VSMCs or VSM-E6 cells and surface Fas was examined over 240 min. (I) 4-OHT was added to p53ERTM rat 1 fibroblasts, and cells were isolated over 120 min for surface Fas. (J) p53ERTM VSMCs were examined for total TNF-R1 in permeabilized cells (P) or surface TNF-R1 was examined 0 to 120 min after 4-OHT addition.

To determine whether p53 that was induced after DNA damage increased surface Fas, we treated p53ERTM cells or VSMCs expressing human papilloma virus E6 [which lacks functional p53 (10)] (VSM-E6 cells) with the topoisomerase II inhibitor etoposide. Etoposide transiently increased surface Fas in p53ERTM but not VSM-E6 cells (Fig. 3, G and H), albeit delayed (maximum 2 hours, baseline by 4 hours) compared with 4-OHT activation. 4-OHT activation of p53ERTM in rat 1 fibroblasts (Fig. 3I) and WT mouse embryo fibroblasts (MEFs) also increased surface Fas, indicating that Fas trafficking may be widespread in mesenchymal cells. Surface TNF-R1 also increased after 4-OHT addition to p53ERTM VSMCs (maximum 30 min, baseline by 60 min) (Fig. 3J), indicating that p53 may induce transport of other death receptors. TNF-R1 in VSMCs was predominantly cytoplasmic.

To examine the requirement for Fas-FasL in early p53-induced apoptosis, we transiently expressed p53ERTM inlpr or gld MEFs, which contain inactivating mutations in Fas and FasL, respectively (15). p53-induced apoptosis was reduced in lpr and gldMEFs compared with WT (Fig. 4A). In addition, a dominant-negative (DN) FADD or crmA, which both inhibit Fas-mediated apoptosis (16, 17), also inhibited p53-mediated apoptosis (Fig. 4B).

Figure 4

p53-induced apoptosis requires Fas and FasL. (A) Wild-type, lpr, andgld MEFs were transiently infected with p53ERTM, cells were transferred to medium containing 0% FCS and 4-OHT, and apoptosis was observed by videomicroscopy. (B) p53ERTM VSMCs stably expressing DN-FADD or vector (Hygro) (21), or cells infected with an adenovirus encoding crmA (Ad-crmA) (22) or β-galactosidase (Ad-β-Gal) were transferred to medium con-taining 0% FCS and 4-OHT. (C) p53ERTM VSMCs were incubated in 0% FCS medium. An agonistic IgM antibody to Fas (CH-11) or isotypically matched control mouse IgM (IgM control; #M-5909, Sigma) (1 μg/ml) was added either simultaneously (IgM Fas) or 4 hours after 4-OHT addition [IgM Fas (4 hours)]. In addition, p53ERTMVSMCs were incubated with CH-11 but without 4-OHT. (D) p53ERTM VSMCs were incubated in medium containing 0% FCS and 4-OHT and an antagonistic IgG1 antibody to Fas, antagonistic IgG1 antibody to FasL, or isotypically matched mouse IgG (IgG control) (all 1 μg/ml) [#F22120 (Transduction Labs), NOK-1 (#65320C, Pharmingen), and M-5284 (Sigma), respectively] added either simultaneously or 4 hours later. p53ERTM VSMCs were also incubated with BFA (5 μg/ml ) 60 min before transfer to 0% FCS medium and 4-OHT.

To confirm that Fas transported after p53 activation could induce apoptosis, we incubated p53ERTM VSMCs with an agonistic immunoglobulin M (IgM) antibody to Fas (anti-Fas) (CH-11) or control antibody. CH-11 increased apoptosis only after p53 activation, although this effect disappeared if addition was delayed 4 hours after p53 activation (Fig. 4C). CH-11 did not induce apoptosis in ERTM or DN-p53ERTM cells. Consistent with these findings, neutralizing anti-Fas or FasL antibodies inhibited apoptosis of p53ERTM VSMCs only if added <4 hours after p53 activation (Fig. 4D), indicating that Fas-FasL death signals are completed within 4 hours of p53 activation. BFA inhibited apoptosis in low-serum medium only after p53 activation (Fig. 4D), further confirming that early p53-induced apoptosis of VSMCs requires surface transport of Fas.

p53-induced apoptosis clearly involves several mechanisms, with transcription-dependent or -independent pathways being determined by cell type and apoptotic stimulus. However, we have found that p53 activation can regulate sensitivity to apoptosis by allowing cytoplasmic death receptors to redistribute to the cell surface. Tumor cells lacking functional p53 will evade apoptosis induced by both p53 transcriptional targets and FasL. Intracellular sequestration of death receptors may cause resistance to apoptosis and insensitivity to chemotherapeutic agents for cancer. Conversely, therapy based on cell surface redistribution of death receptors may promote apoptosis through endogenous ligands.

  • * To whom correspondence should be addressed. E-mail: mrb{at}


Stay Connected to Science

Navigate This Article