Olivier Reelfs

Concomitant inactivation of p53 and Chk2 in breast cancer

The structure and expression of the human Rad53 homologue Chk2 was analysed in breast cancer. The previously described silent polymorphism at nucleotide 252 in codon 84 (GAA>GAG) was observed in 5/141 cases. Somatic Chk2 coding mutations were detected in 7/141 cases, these occurring in 4/18 BRCA1-associated breast cancers, 1/78 sporadic breast cancers and 2/25 typical medullary carcinomas. Each of the BRCA1-associated cancers with Chk2 mutations also contained mutations in p53, whereas the single sporadic cancer with Chk2 mutation was wild-type for p53. Expression of Chk2 was ubiquitously detected in normal ductal epithelium of the breast, but there was loss of expression in a significant proportion of breast carcinomas, and this occurred in cancers both with and without p53 mutation. A CpG island was identified 5′ of the Chk2 transcriptional start site, but there was no evidence of cytosine methylation in any of the cancers with down-regulated Chk2 expression. Analysis of the germ-line of 45 individuals with hereditary or early onset breast cancer revealed wild-type Chk2 sequence in all cases. Thus, despite the rarity of somatic mutations in Chk2 in sporadic breast carcinomas, our results nevertheless reveal that concomitant loss of function in Chk2 (via down-regulation of expression) and p53 (via mutation) occurs in a proportion of sporadic cases. However, consistent with other studies, we show that germ-line mutations in Chk2 are unlikely to account for a significant proportion of non BRCA1-, non BRCA2-associated hereditary breast cancers.

DNA breakage and cell cycle checkpoint abrogation induced by a therapeutic thiopurine and UVA radiation

The frequency of squamous cell skin carcinoma in organ transplant patients is around 100-fold higher than normal. This dramatic example of therapy-related cancer reflects exposure to sunlight and to immunosuppressive drugs. Here, we show that the interaction between low doses of UVA, the major ultraviolet component of incident sunlight, and 6-TG, a UVA chromophore that is introduced into DNA by one of the most widely prescribed immunosuppressive drugs, causes DNA single- and double-strand breaks (DSB). S phase cells are particularly vulnerable to this DNA breakage and cells defective in rejoining of S-phase DSB are hypersensitive to the combination of low-dose UVA and DNA 6-TG. 6-TG/UVA-induced DNA lesions provoke canonical DNA damage responses involving activation of the ATM/Chk2 and ATR/Chk1 pathways and appropriate cell cycle checkpoints. Higher levels of photochemical DNA damage induce a proteasome-mediated degradation of Chk1 and checkpoint abrogation that is consistent with persistent unrepaired DNA damage. These findings indicate that the interaction between UVA and an immunosuppressant drug causes photochemical DNA lesions, including DNA breaks, and can compromise cell cycle checkpoints. These two properties could contribute to the high risk of sunlight-related skin cancer in long-term immunosuppressed patients.

Identification of potentially cytotoxic lesions induced by UVA photoactivation of DNA 4-thiothymidine in human cells

Photochemotherapy—in which a photosensitizing drug is combined with ultraviolet or visible radiation—has proven therapeutic effectiveness. Existing approaches have drawbacks, however, and there is a clinical need to develop alternatives offering improved target cell selectivity. DNA substitution by 4-thiothymidine (S4TdR) sensitizes cells to killing by ultraviolet A (UVA) radiation. Here, we demonstrate that UVA photoactivation of DNA S4TdR does not generate reactive oxygen or cause direct DNA breakage and is only minimally mutagenic. In an organotypic human skin model, UVA penetration is sufficiently robust to kill S4TdR-photosensitized epidermal cells. We have investigated the DNA lesions responsible for toxicity. Although thymidine is the predominant UVA photoproduct of S4TdR in dilute solution, more complex lesions are formed when S4TdR-containing oligonucleotides are irradiated. One of these, a thietane/S5-(6-4)T:T, is structurally related to the (6-4) pyrimidine:pyrimidone [(6-4) Py:Py] photoproducts induced by UVB/C radiation. These lesions are detectable in DNA from S4TdR/UVA-treated cells and are excised from DNA more efficiently by keratinocytes than by leukaemia cells. UVA irradiation also induces DNA interstrand crosslinking of S4TdR-containing duplex oligonucleotides. Cells defective in repairing (6-4) Py:Py DNA adducts or processing DNA crosslinks are extremely sensitive to S4TdR/UVA indicating that these lesions contribute significantly to S4TdR/UVA cytotoxicity.

Role of intracellular labile iron, ferritin, and antioxidant defence in resistance of chronically adapted Jurkat T cells to hydrogen peroxide

To examine the role of intracellular labile iron pool (LIP), ferritin (Ft), and antioxidant defence in cellular resistance to oxidative stress on chronic adaptation, a new H2O2-resistant Jurkat T cell line “HJ16” was developed by gradual adaptation of parental “J16” cells to high concentrations of H2O2. Compared to J16 cells, HJ16 cells exhibited much higher resistance to H2O2-induced oxidative damage and necrotic cell death (up to 3 mM) and had enhanced antioxidant defence in the form of significantly higher intracellular glutathione and mitochondrial ferritin (FtMt) levels as well as higher glutathione-peroxidase (GPx) activity. In contrast, the level of the Ft H-subunit (FtH) in the H2O2-adapted cell line was found to be 7-fold lower than in the parental J16 cell line. While H2O2 concentrations higher than 0.1 mM fully depleted the glutathione content of J16 cells, in HJ16 cells the same treatments decreased the cellular glutathione content to only half of the original value. In HJ16 cells, H2O2 concentrations higher than 0.1 mM increased the level of FtMt up to 4-fold of their control values but had no effect on the FtMt levels in J16 cells. Furthermore, while the basal cytosolic level of LIP was similar in both cell lines, H2O2 treatment substantially increased the cytosolic LIP levels in J16 but not in HJ16 cells. H2O2 treatment also substantially decreased the FtH levels in J16 cells (up to 70% of the control value). In contrast in HJ16 cells, FtH levels were not affected by H2O2 treatment. These results indicate that chronic adaptation of J16 cells to high concentrations of H2O2 has provoked a series of novel and specific cellular adaptive responses that contribute to higher resistance of HJ16 cells to oxidative damage and cell death. These include increased cellular antioxidant defence in the form of higher glutathione and FtMt levels, higher GPx activity, and lower FtH levels. Further adaptive responses include the significantly reduced cellular response to oxidant-mediated glutathione depletion, FtH modulation, and labile iron release and a significant increase in FtMt levels following H2O2 treatment.

Thiothymidine plus low-dose UVA kills hyperproliferative human skin cells independently of their human papilloma virus status

The thymidine analogue 4-thiothymidine (S4TdR) is a photosensitizer for UVA radiation. The UV absorbance spectrum of S4TdR and its incorporation into DNA suggests that it might act synergistically with nonlethal doses of UVA to selectively kill hyperproliferative or cancerous skin cells. We show here that nontoxic concentrations of S4TdR combine with nonlethal doses of UVA to kill proliferating cultured skin cells. Established cell lines with a high fraction of proliferating cells were more sensitive than primary keratinocytes or fibroblasts to apoptosis induction by S4TdR/UVA. Although S4TdR plus UVA treatment induces stabilization of p53, cell death, as measured by apoptosis or clonal survival, occurs to a similar extent in both p53 wild-type and p53-null backgrounds. Furthermore, different types of human papilloma virus E6 proteins, which protect against UVB-induced apoptosis, have little effect on killing by S4TdR/UVA. S4TdR/UVA offers a possible therapeutic intervention strategy that seems to be applicable to human papilloma virus–associated skin lesions. [Mol Cancer Ther 2007;6(9):2487–95]

Efficient silencing of EGFP reporter gene with siRNA delivered by asymmetrical N4,N9-diacyl spermines.

It is important to obtain structure-activity relationship (SAR) data across cationic lipids for the self-assembly and nonviral intracellular delivery of siRNA. The aims of this work are to carry out a SAR study on the efficiency of asymmetrical N(4),N(9)-diacyl spermines in siRNA delivery and EGFP reporter gene silencing, with comparisons to selected mixtures composed of symmetrical N(4),N(9)-diacyl spermines. Another important aim of these studies is to quantify the changes in cell viability, assayed with alamarBlue, as a function of lipid structure. Therefore, we have designed, synthesized, purified, and assayed novel cationic lipids that are asymmetrical lipopolyamines based on spermine. Flow cytometry and fluorescence microscopy in an EGFP stably transfected HeLa cell line, measuring both delivery of fluorescently tagged siRNAs and silencing the EGFP signal, allowed quantitation of the differences between asymmetrical cationic lipids, mixtures of their symmetrical counterparts, and comparison with commercial nonviral delivery agents. Intracellular delivery of siRNA and gene silencing by siRNA differ with different hydrophobic domains. In these asymmetrical N(4),N(9)-diacyl spermines, lipids that enhance siRNA uptake do not necessarily enhance siRNA-induced inhibition of gene expression: C18 and longer saturated chains promote uptake, while more unsaturated C18 chains promote gene silencing. These properties are efficiently demonstrated in a new nontoxic cationic lipid siRNA vector, N(4)-linoleoyl-N(9)-oleoyl-1,12-diamino-4,9-diazadodecane (LinOS), which is also shown to be comparable with or superior to TransIT-TKO and Lipofectamine 2000.

A powerful mitochondria-targeted iron chelator affords high photoprotection against solar ultraviolet A radiation

Mitochondria are the principal destination for labile iron, making these organelles particularly susceptible to oxidative damage on exposure to ultraviolet A (UVA, 320–400 nm), the oxidizing component of sunlight. The labile iron-mediated oxidative damage caused by UVA to mitochondria leads to necrotic cell death via adenosine triphosphate depletion. Therefore, targeted removal of mitochondrial labile iron via highly specific tools from these organelles may be an effective approach to protect the skin cells against the harmful effects of UVA. In this work, we designed a mitochondria-targeted hexadentate (tricatechol-based) iron chelator linked to mitochondria-homing SS-like peptides. The photoprotective potential of this compound against UVA-induced oxidative damage and cell death was evaluated in cultured primary skin fibroblasts. Our results show that this compound provides unprecedented protection against UVA-induced mitochondrial damage, adenosine triphosphate depletion, and the ensuing necrotic cell death in skin fibroblasts, and this effect is fully related to its potent iron-chelating property in the organelle. This mitochondria-targeted iron chelator has therefore promising potential for skin photoprotection against the deleterious effects of the UVA component of sunlight.

Design of novel fluorescent mitochondria-targeted peptides with iron selective sensing activity

A joint work with Zhengling Qi (University of North Carolina at Chapel Hill and George Washington University), Ying Cui (soon to join the University of Minnesota), and Yufeng Liu (University of North Carolina at Chapel Hill), this paper establishes the statistical consistency of stationary solutions of a class of coupled nonconvex and nonsmooth empirical risk minimization problems. Derived from available data via sampling, these problems are the computational workhorse of a population risk model which typically involves the minimization of an expected value of a random functional; the stationary solutions of the former problems are practical approximations of a stationary solution of the latter minimization problem whose globally optimal solutions are computationally elusive due to two reasons: nonconvexity of the model and the presence of the expectation operator that cannot be exactly evaluated for general random distributions of the data. The considered model covers a broad range of emerging applications in machine learning, statistical estimation, and stochastic programming. While it is known that the global minimizers of the empirical risk asymptotically converge to a globally optimal solution of the population risk under mild conditions, a convergence result of this kind is practically meaningless because such optimal solutions cannot be computed when the discretized optimization problems to be solved are not convex; thus there is a gap between a minimizer-centric theory of statistical consistency and practically-oriented computations which at best can reach a stationary solution of some sort. The situation becomes more serious when nondifferentiability is coupled with nonconvexity because there is a host of stationary solutions of the resulting optimization problems. Typically, the sharper the stationarity solution is (sharp in the sense of least relaxation in its definition), the more difficult it is to compute. It is thus important to understand whether in practice, the focus should be placed on computing sharp stationary solutions (which distinguish themselves as being the ones that must satisfy all other relaxed definitions of stationarity) that potentially require higher computational costs versus computing some less demanding solutions. Our derived results show that the sharpness of the stationarity at the empirical level is preserved at the population level, thus favoring the former. By establishing the consistency of the computationally-tractable stationary solutions for the coupled nonconvex nonsmooth M-estimation, we are closing the gap between statistical consistency and practical computation. Download presentation slides here.Objective: The primary purpose of this study was to investigate the ability of some Gram negative bac-teria to acquire different sources of iron in order to identify the iron supplement with lowest bacterial bioavailability. Material and Methods: The effects of four different sources of iron (FeSO4, Fe(NO3) FeCL3 and Carbonyl Fe) were compared as iron source for bacterial growth in minimal medium with several bacterial spe-cies. The effects of nano-encapsulating some of these compounds was also tested. The ef-fects of these compounds on the formation of biofilms were also assessed. Bacterial growth was determined by measuring bacteria density at 600nm. The formation of biofilms was de-termined by using the crystal violet binding assay. Results Differences were observed between different sources of iron, especially with Carbonyl iron, which seemed to be poorest source of iron for all bacterial species. The effects of different sources of iron on the formation of biofilms were also noticeable. The encapsulation of some of the iron compounds had major effects on growth and biofilms formation. Conclusion The variation in iron bioavailability for bacteria with different iron sources suggests that iron supplement should be chosen not only based on their human uptake efficiency, but also by choosing iron sources with the poorest bacterial bioavailability. In addition, the results ob-tained following encapsulation indicate that this might provide a mechanism to further limit bacterial iron uptake, and therefore pathogenicity.

Deferiprone And Deferasirox As Promissing Candidate Ligands For Skin Photoprotection By Light-Activated Caged-Iron Chelators

A joint work with Zhengling Qi (University of North Carolina at Chapel Hill and George Washington University), Ying Cui (soon to join the University of Minnesota), and Yufeng Liu (University of North Carolina at Chapel Hill), this paper establishes the statistical consistency of stationary solutions of a class of coupled nonconvex and nonsmooth empirical risk minimization problems. Derived from available data via sampling, these problems are the computational workhorse of a population risk model which typically involves the minimization of an expected value of a random functional; the stationary solutions of the former problems are practical approximations of a stationary solution of the latter minimization problem whose globally optimal solutions are computationally elusive due to two reasons: nonconvexity of the model and the presence of the expectation operator that cannot be exactly evaluated for general random distributions of the data. The considered model covers a broad range of emerging applications in machine learning, statistical estimation, and stochastic programming. While it is known that the global minimizers of the empirical risk asymptotically converge to a globally optimal solution of the population risk under mild conditions, a convergence result of this kind is practically meaningless because such optimal solutions cannot be computed when the discretized optimization problems to be solved are not convex; thus there is a gap between a minimizer-centric theory of statistical consistency and practically-oriented computations which at best can reach a stationary solution of some sort. The situation becomes more serious when nondifferentiability is coupled with nonconvexity because there is a host of stationary solutions of the resulting optimization problems. Typically, the sharper the stationarity solution is (sharp in the sense of least relaxation in its definition), the more difficult it is to compute. It is thus important to understand whether in practice, the focus should be placed on computing sharp stationary solutions (which distinguish themselves as being the ones that must satisfy all other relaxed definitions of stationarity) that potentially require higher computational costs versus computing some less demanding solutions. Our derived results show that the sharpness of the stationarity at the empirical level is preserved at the population level, thus favoring the former. By establishing the consistency of the computationally-tractable stationary solutions for the coupled nonconvex nonsmooth M-estimation, we are closing the gap between statistical consistency and practical computation. Download presentation slides here.Objective: The primary purpose of this study was to investigate the ability of some Gram negative bac-teria to acquire different sources of iron in order to identify the iron supplement with lowest bacterial bioavailability. Material and Methods: The effects of four different sources of iron (FeSO4, Fe(NO3) FeCL3 and Carbonyl Fe) were compared as iron source for bacterial growth in minimal medium with several bacterial spe-cies. The effects of nano-encapsulating some of these compounds was also tested. The ef-fects of these compounds on the formation of biofilms were also assessed. Bacterial growth was determined by measuring bacteria density at 600nm. The formation of biofilms was de-termined by using the crystal violet binding assay. Results Differences were observed between different sources of iron, especially with Carbonyl iron, which seemed to be poorest source of iron for all bacterial species. The effects of different sources of iron on the formation of biofilms were also noticeable. The encapsulation of some of the iron compounds had major effects on growth and biofilms formation. Conclusion The variation in iron bioavailability for bacteria with different iron sources suggests that iron supplement should be chosen not only based on their human uptake efficiency, but also by choosing iron sources with the poorest bacterial bioavailability. In addition, the results ob-tained following encapsulation indicate that this might provide a mechanism to further limit bacterial iron uptake, and therefore pathogenicity.

Evaluation of the photoprotective potential of novel hexadentate iron chelators against sunlight-induced damage to skin cells

A joint work with Zhengling Qi (University of North Carolina at Chapel Hill and George Washington University), Ying Cui (soon to join the University of Minnesota), and Yufeng Liu (University of North Carolina at Chapel Hill), this paper establishes the statistical consistency of stationary solutions of a class of coupled nonconvex and nonsmooth empirical risk minimization problems. Derived from available data via sampling, these problems are the computational workhorse of a population risk model which typically involves the minimization of an expected value of a random functional; the stationary solutions of the former problems are practical approximations of a stationary solution of the latter minimization problem whose globally optimal solutions are computationally elusive due to two reasons: nonconvexity of the model and the presence of the expectation operator that cannot be exactly evaluated for general random distributions of the data. The considered model covers a broad range of emerging applications in machine learning, statistical estimation, and stochastic programming. While it is known that the global minimizers of the empirical risk asymptotically converge to a globally optimal solution of the population risk under mild conditions, a convergence result of this kind is practically meaningless because such optimal solutions cannot be computed when the discretized optimization problems to be solved are not convex; thus there is a gap between a minimizer-centric theory of statistical consistency and practically-oriented computations which at best can reach a stationary solution of some sort. The situation becomes more serious when nondifferentiability is coupled with nonconvexity because there is a host of stationary solutions of the resulting optimization problems. Typically, the sharper the stationarity solution is (sharp in the sense of least relaxation in its definition), the more difficult it is to compute. It is thus important to understand whether in practice, the focus should be placed on computing sharp stationary solutions (which distinguish themselves as being the ones that must satisfy all other relaxed definitions of stationarity) that potentially require higher computational costs versus computing some less demanding solutions. Our derived results show that the sharpness of the stationarity at the empirical level is preserved at the population level, thus favoring the former. By establishing the consistency of the computationally-tractable stationary solutions for the coupled nonconvex nonsmooth M-estimation, we are closing the gap between statistical consistency and practical computation. Download presentation slides here.Objective: The primary purpose of this study was to investigate the ability of some Gram negative bac-teria to acquire different sources of iron in order to identify the iron supplement with lowest bacterial bioavailability. Material and Methods: The effects of four different sources of iron (FeSO4, Fe(NO3) FeCL3 and Carbonyl Fe) were compared as iron source for bacterial growth in minimal medium with several bacterial spe-cies. The effects of nano-encapsulating some of these compounds was also tested. The ef-fects of these compounds on the formation of biofilms were also assessed. Bacterial growth was determined by measuring bacteria density at 600nm. The formation of biofilms was de-termined by using the crystal violet binding assay. Results Differences were observed between different sources of iron, especially with Carbonyl iron, which seemed to be poorest source of iron for all bacterial species. The effects of different sources of iron on the formation of biofilms were also noticeable. The encapsulation of some of the iron compounds had major effects on growth and biofilms formation. Conclusion The variation in iron bioavailability for bacteria with different iron sources suggests that iron supplement should be chosen not only based on their human uptake efficiency, but also by choosing iron sources with the poorest bacterial bioavailability. In addition, the results ob-tained following encapsulation indicate that this might provide a mechanism to further limit bacterial iron uptake, and therefore pathogenicity.