c-di-GMP induction of Dictyostelium cell death requires the polyketide DIF-1.

Cell death in the model organism Dictyostelium, as studied in monolayers in vitro, can be induced by the polyketide DIF-1 or by the cyclical dinucleotide c-di-GMP. c-di-GMP, a universal bacterial second messenger, can trigger innate immunity in bacterially infected animal cells and is involved in developmental cell death in Dictyostelium. We show here that c-di-GMP was not sufficient to induce cell death in Dictyostelium cell monolayers. Unexpectedly, it also required the DIF-1 polyketide. The latter could be exogenous, as revealed by a telling synergy between c-di-GMP and DIF-1. The required DIF-1 polyketide could also be endogenous, as shown by the inability of c-di-GMP to induce cell death in Dictyostelium HMX44A cells and DH1 cells upon pharmacological or genetic inhibition of DIF-1 biosynthesis. In these cases, c-di-GMP-induced cell death was rescued by complementation with exogenous DIF-1. Taken together, these results demonstrated that c-di-GMP could trigger cell death in Dictyostelium only in the presence of the DIF-1 polyketide or its metabolites. This identified another element of control to this cell death and perhaps also to c-di-GMP effects in other situations and organisms.

Atg1 allows second-signaled autophagic cell death in Dictyostelium.

We investigated the role of Atg1 in autophagic cell death (ACD) in a Dictyostelium monolayer model. The model is especially propitious, not only because of genetic tractability and absence of apoptosis machinery, but also because induction of ACD requires two successive exogenous signals, first the combination of starvation and cAMP, second the differentiation factor DIF-1. This enables one to analyze separately first-signal-induced autophagy and subsequent second-signal-induced ACD. We used mutants of atg1, a gene that plays an essential role in the initiation of autophagy. Upon starvation/cAMP, in contrast to parental cells, atg1 mutant cells showed irreversible lesions, clearly establishing a protective role for Atg1. Upon subsequent exposure to DIF-1 or to more ACD-specific second signals, starved parental cells progressed to ACD, but starved atg1 mutant cells did not, showing that Atg1 was required for ACD. Thus, in the same cells Atg1 was required in two apparently opposite ways, upon first-signaling for cell survival and upon second-signaling for ACD. Our findings strongly suggest that Atg1, thus presumably autophagy, protects the cells from starvation-induced cell death, allowing subsequent induction of ACD by the second signal. ACD is therefore not only "with" autophagy (since it showed signs of autophagy throughout), but is also "allowed by" autophagy. This does not exclude a role for autophagy also after second signaling. These results may account for discrepancies reported in the literature, encourage searches for second signals in different developmental models of ACD, and incite caution in autophagy-related therapeutic attempts.

A second signal for autophagic cell death?

Dictyostelium cells in monolayers in vitro lend themselves well to a study of autophagic cell death (ACD). There is no apoptosis machinery in the protist Dictyostelium, no caspase nor Bcl-2 family members (except a paracaspase whose inactivation does not alter cell death), thus there is no apoptosis that could interfere with, or substitute for, nonapoptotic cell death. Also, Dictyostelium, a eukaryote, has a haploid genome, which facilitates random insertional mutagenesis.

Autophagic cell death in Dictyostelium requires the receptor histidine kinase DhkM.

Dictyostelium constitutes a genetically tractable model for the analysis of autophagic cell death (ACD). During ACD, Dictyostelium cells first transform into paddle cells and then become round, synthesize cellulose, vacuolize, and die. Through random insertional mutagenesis, we identified the receptor histidine kinase DhkM as being essential for ACD. Surprisingly, different DhkM mutants showed distinct nonvacuolizing ACD phenotypes. One class of mutants arrested ACD at the paddle cell stage, perhaps through a dominant-negative effect. Other mutants, however, progressed further in the ACD program. They underwent rounding and cellulose synthesis but stopped before vacuolization. Moreover, they underwent clonogenic but not morphological cell death. Exogenous 8-bromo-cAMP restored vacuolization and death. A role for a membrane receptor at a late stage of the ACD pathway is puzzling, raising questions as to which ligand it is a receptor for and which moieties it phosphorylates. Together, DhkM is the most downstream-known molecule required for this model ACD, and its distinct mutants genetically separate previously undissociated late cell death events.

Autophagic cell death: analysis in Dictyostelium.

Autophagic cell death (ACD) can be operationally described as cell death with an autophagic component. While most molecular bases of this autophagic component are known, in ACD the mechanism of cell death proper is not well defined, in particular because in animal cells there is poor experimental distinction between what triggers autophagy and what triggers ACD. Perhaps as a consequence, it is often thought that in animal cells a little autophagy is protective while a lot is destructive and leads to ACD, thus that the shift from autophagy to ACD is quantitative. The aim of this article is to review current knowledge on ACD in Dictyostelium, a very favorable model, with emphasis on (1) the qualitative, not quantitative nature of the shift from autophagy to ACD, in contrast to the above, and (2) random or targeted mutations of in particular the following genes: iplA (IP3R), TalB (talinB), DcsA (cellulose synthase), GbfA, ugpB, glcS (glycogen synthase) and atg1. These mutations allowed the genetic dissection of ACD features, dissociating in particular vacuolisation from cell death.

Necrotic cell death: From reversible mitochondrial uncoupling to irreversible lysosomal permeabilization.

Dictyostelium atg1- mutant cells provide an experimentally and genetically favorable model to study necrotic cell death (NCD) with no interference from apoptosis or autophagy. In such cells subjected to starvation and cAMP, induction by the differentiation-inducing factor DIF or by classical uncouplers led within minutes to mitochondrial uncoupling, which causally initiated NCD. We now report that (1) in this model, NCD included a mitochondrial-lysosomal cascade of events, (2) mitochondrial uncoupling and therefore initial stages of death showed reversibility for a surprisingly long time, (3) subsequent lysosomal permeabilization could be demonstrated using Lysosensor blue, acridin orange, Texas red-dextran and cathepsin B substrate, (4) this lysosomal permeabilization was irreversible, and (5) the presence of the uncoupler was required to maintain mitochondrial lesions but also to induce lysosomal lesions, suggesting that signaling from mitochondria to lysosomes must be sustained by the continuous presence of the uncoupler. These results further characterized the NCD pathway in this priviledged model, contributed to a definition of NCD at the lysosomal level, and suggested that in mammalian NCD even late reversibility attempts by removal of the inducer may be of therapeutic interest.

Analysis of autophagic and necrotic cell death in Dictyostelium.

Non-apoptotic cell death types can be conveniently studied in Dictyostelium discoideum, an exceptionally favorable model not only because of its well-known genetic and experimental advantages, but also because in Dictyostelium there is no apoptosis machinery that could interfere with non-apoptotic cell death. We show here how to conveniently demonstrate, assess, and study these non-apoptotic cell death types. These can be generated by use of modifications of the monolayer technique of Rob Kay et al., and either wild-type HMX44A Dictyostelium cells, leading to autophagic cell death, or the corresponding atg1- autophagy gene mutant cells, leading to necrotic cell death. Methods to follow these non-apoptotic cell death types qualitatively and quantitatively will be reported.

A UDP-glucose derivative is required for vacuolar autophagic cell death.

Autophagic cell death in Dictyostelium can be dissociated into a starvation-induced sensitization stage and a death induction stage. A UDP-glucose pyrophosphorylase (ugpB) mutant and a glycogen synthase (glcS) mutant shared the same abnormal phenotype. In vitro, upon starvation alone mutant cells showed altered contorted morphology, indicating that the mutations affected the pre-death sensitization stage. Upon induction of cell death, most of these mutant cells underwent death without vacuolization, distinct from either autophagic or necrotic cell death. Autophagy itself was not grossly altered as shown by conventional and electron microscopy. Exogenous glycogen or maltose could complement both ugpB(-) and glcS(-) mutations, leading back to autophagic cell death. The glcS(-) mutation could also be complemented by 2-deoxyglucose that cannot undergo glycolysis. In agreement with the in vitro data, upon development glcS(-) stalk cells died but most were not vacuolated. We conclude that a UDP-glucose derivative (such as glycogen or maltose) plays an essential energy-independent role in autophagic cell death.

Autophagy and autophagic cell death in Dictyostelium.

Autophagic cell death can be conveniently studied in Dictyostelium discoideum, an exceptionally favorable model not only because of its well-known genetic and experimental advantages but also because in Dictyostelium there is no apoptosis machinery that could interfere with nonapoptotic cell death. Moreover, autophagic cell death in Dictyostelium can be dissociated into a starvation-induced sensitization stage, during which autophagy is induced, and a death induction stage. We show here how to demonstrate, assess and analyze this autophagic cell death. This can be studied in vivo during the development of Dictyostelium, and in vitro, using modifications of the monolayer technique of Rob Kay et al. Methods to follow this autophagic cell death qualitatively and quantitatively are reported.

Clinical relevance of amphiregulin and VEGF in primary breast cancers.

The characterization of novel prognostic markers in breast cancer is necessary to improve the identification of high-risk populations. In our study, the prognostic significance of VEGF and amphiregulin (AR) was investigated and compared to conventional prognostic factors in primary breast cancers. The analysis was performed using enzyme-linked immuno-assay in a series of 193 patients, and univariate and multivariate analysis were performed in the overall population as well as in pre- and post-menopausal patients subdivided in node-negative (N-) and node-positive (N+) subsets. AR (median, 44.8 pg/mg protein) appeared strongly correlated with progesterone receptors (PgR) (p = 0.0018) in the premenopausal N+ population, and with uPA (p= 0.020) and VEGF (p= 0.0053) in the postmenopausal/N+ patients. Despite these attractive data, AR expression was not significant for recurrence or survival outcome. Data revealed strong correlation between VEGF and uPA, and PAI-1, in the N+ population. Moreover, patients with high VEGF levels displayed poor outcome, with an increased risk for N+ subset. These data were confirmed by multivariate analysis that presented histologic grade (HR, 10.55, p = 0.001) and VEGF (HR, 3.89, p = 0.03) as the prominent prognostic markers for overall survival for the N+ population. Furthermore, infiltrating ductal carcinomas (IDC) were shown to express higher levels of both uPA (p < 0.0001) and VEGF (p = 0.002) than intralobular carcinomas. This retrospective study reinforces the pejorative biological role of VEGF in the progression of breast tumors. Our data also suggest that VEGF and uPA might play particular role in the biology and progression of IDC.

Transforming growth factor beta-1 and amphiregulin act in synergy to increase the production of urokinase-type plasminogen activator in transformed breast epithelial cells.

Amphireguline (AR) is an epidermal growth factor (EGF)-related peptide that seems to play an important role in breast cancer progression. We have demonstrated recently that suppression of AR expression in transformed breast epithelial cells considerably reduced both size and neovascularization of tumors developed in nude mice. We show that the reduction of AR expression allowed to an important decrease of the levels of urokinase-type plasminogen activator (uPA) and transforming growth factor-beta1 (TGFbeta1). According to these data, exogenous AR (10(-10) M-10(-8) M) stimulated the production of uPA and TGFbeta1 in AR antisense-transfected A2-15 and A2-P17F25 cells. The addition of 2 x 10(-10) M TGFbeta1 into culture medium increased the level of uPA produced by AR-expressing parental cells but not by A2-15 and A2-P17F25 cell clones. Whereas AR alone stimulated uPA production to 200% of control, combined AR and TGFbeta1 treatment increased protease level in A2-15 and A2-P17F25 cells to 500-600% of control, demonstrating a synergism between TGFbeta1 and AR. This was accompanied by an important augmentation of the number of tumoral cells that invaded matrigel in vitro. The synergistic induction of uPA protein resulted of an early and transient augmentation of steady state mRNA level and was blocked in the presence of the MAP kinase kinase inhibitor PD098059, strongly suggesting that synergistic effect of AR and TGFbeta1 on uPA expression required MAPK pathway. This data demonstrates concerted action between AR and TGFbeta1 that may have profound effect on protease production and consequently on breast cancer progression.