The process that transforms a normal cell into a malignant tumor cell requires several cellular alterations. Progressive loss of cell-specific differentiated functions and evasion of apoptosis are hallmarks of some cancers because defects in its regulators invariably accompany tumorigenesis and sustain malignant progression. Re-differentiation of de-differentiated cells will help them re-acquire normal functions to some extent, thereby enhancing treatment efficiency and limiting cancer cell progression. In addition, survival signaling is distinct from apoptosis resistance and both are major regulators of cancer cell survival. Thus, targeting only one of these pathways may not be sufficient to obtain therapeutic effects. The following two sections describe our identification of inducers for hepatocarcinoma cell differentiation and cross-talk of casein kinase 1α (CK1α) with retinoid X receptor in lymphoma cell death. Our discoveries have identified novel mechanisms in cancer cell resistance and will provide new approaches for cancer treatment.

Chemical-induced re-differentiation of hepatocarcinoma cells-differentiation therapy: Tumorigenesis and carcinogenesis accompany progressive loss of tissue-specific differentiated functions, making them become dedifferentiated and refractory to efficacy-proven conventional therapies such as radioiodine ablation therapy. Thus, identifying small molecules that promote dedifferentiated cancer cells to re-acquire cell functions is therapeutically significant. Retinoic acid is known to cause redifferentiation or to prevent further dedifferentiation of various tumor tissues. We demonstrated that retinoic acid is capable of inducing re-differentiation of human hepatocarcinoma cells. Importantly, we have found that selenium, in the form of selenite, can re-differentiate human hepatocarcinoma in vitro since treatment with retinoic acid and selenite affects hepatocyte specific functions, cell-cell or cell-matrix interaction (increase in ConA-bound but decrease in WGA- or LCA-bound N-linked oligosaccharides)[7,8], differentiation markers (α-fetoprotein, fibronectin)[7], and growth[9]. Moreover, combined application of suboptimal doses of retinoic acid and selenite showed additive effects on the reversion of malignant phenotypes[9,10], enabling us to avoid the side-effects associated with application of high concentrations of inducers. These identified re-differentiating inducers of malignant hepatocarcinoma cells will retard tumor growth and make tumors respond to conventional therapies via regaining cell normal functions.

Discovery of CK1 in regulation of retinoid X receptor-mediated cancer cell death: Retinoid X receptors (RXRs) including RXRα, -β, and -γ are unique in their capability of heterodimerizing with other nuclear receptors to form functional transcription factors. Despite their ubiquitous expression, RXR agonists are potent apoptotic inducers only in some cancer cells but many other malignant cells are insensitive to RXR agonists. Elucidation of cellular mechanisms that mediate apoptosis resistance is an urgent need. Dr. Qin and his coworkers have revealed a novel pro-apoptotic pathway induced by RXR agonists, which is negatively regulated by CK1α phosphorylation of RXR in an agonist-dependent manner. CK1α depletion renders resistant cells susceptible to RXR agonist-induced apoptosis, showing that CK1α can promote cell survival by interfering with RXR agonist-induced apoptosis[11]. The importance of this finding by Dr. Qin’s team lies in the fact that for the first time it was uncovered the molecular mechanisms thereby CK1α makes tumor cells resistant to RXR agonists and has made it possible to treat RXR agonist-resistant cancer cells with simultaneous inhibition of CK1.

To gain insight into important structural and functional features of regions of the RXRα that are required for RXR agonist-induced apoptosis, Dr. Qin has constructed RXRα mutants with deletions or point mutations (Figure 2). Dr. Qin’s group has found that over-expression of RXRα in DT40 B lymphoma cells dramatically increased sensitivity to rexinoid-induced growth inhibition. In contrast, cells expressing RXRα with a deletion of either the A/B or DNA binding domain (C domain) were resistant. Furthermore, his group has verified the importance of C domain integrity by point-mutating Cys(135) to Ser (C135S) that disrupts zinc-finger formation and subsequently fails to sensitize cells to RXR agonists[12]. Point mutating RXR Lys(201) to Thr and Arg(202) to Ala (KTRA) impairs RXR homodimer formation but does not affect RXR heterodimerization. Expression of KTRA mutant of RXR fails to increase sensitivity to rexinoid, but does sensitize to growth inhibition by retinoic acid receptor (RAR) and PPARγ agonists. Thus, Dr. Qin’s studies have uncovered the ability of RXR homodimers to mediate rexinoid-driven B cell apoptosis besides as a signaling partner for other nuclear receptors.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Schematic representation of Retinoid X receptor α protein and its deletion mutants. Retinoid X receptor (RXR)α consists of a variable N-terminal domain (A/B domain) containing activation function 1, a highly conserved DNA-binding domain (DBD), also named as C domain, a nonconserved hinge (D domain), a moderately conserved ligand-binding domain (E domain), and a ligand-dependent transactivation F domain (AF-2). The DBD consists of two cysteine-rich zinc-finger motifs through which nuclear receptors bind to a specific DNA sequence and a carboxy-terminal extension (CTE) that exists only in RXRs and RXR-binding partners. Point mutation of Cys135 to Ser in first zinc-finger disables the capability of RXR to bind DNA. KTRA mutations in the CTE disrupts RXRα homodimerization but does not affect RXRα heterodimerization. Ligand-binding domain contains a ligand-binding pocket for binding of small, lipophilic molecules, a cofactor binding surface, and a dimerization surface.

Paradoxically, pan-RAR antagonist (AGN194310) abolishes growth of both primary prostate carcinoma cells from patients and the prostate carcinoma lines with no apparent effect on normal prostate cells[13]. Importantly, the ability of AGN194310 to induce apoptosis of prostate cancer cells without apparent effect on normal prostate epithelial cells makes this compound a promising agent in the treatment of prostate cancer, as significant lower side effects on normal prostate cells are expected. Interestingly, synthetic retinoid-related molecule, 4-[3-(1-heptyl-4,4-dimethyl-2-oxo-1,2,3,4-tetrahydroquinolin-6-yl)-3-oxo-E-propenyl] benzoic acid (AGN193198), that neither binds effectively to RARs and RXRs nor trans-activates RAR- and RXR-mediated reporters, potently induces apoptosis in a variety of malignant cancer cells via activation of caspases[14], suggesting that AGN193198 may be an alternate choice in the treatment of prostate cancer as the compound works through different mechanisms.

Oxidative stress is an unbalanced accumulation of reactive oxygen species (ROS) due to a reduction in defense mechanisms and/or an increase in ROS production. Through oxidation of macromolecules lipids, proteins, and DNA, oxidative stress cause cell injury and mutagenesis that are associated with aging and etiology of many age-related diseases. Understanding the molecular mechanisms by which oxidative stress cause dysfunctions and injury of cells and tissues will provide novel therapeutics on unmet medical need.

Critical role of Syk in oxidative stress signaling in B lymphoma cells: Oxidative stress induces massive tyrosine phosphorylation of cellular proteins and Ca²⁺ mobilization, in which protein tyrosine kinases and phosphatases are involved in these biological processes[15,16]. Spleen tyrosine kinase (Syk), a non-receptor protein tyrosine kinase with two tandem Src homology 2 (SH2) domains, was cloned in 1991[17]. A rapid elevation in Syk activity is demonstrated and subsequent Ca²⁺ influx is necessary for the sustenance of Syk activity[15,18]. Thus, identifying roles of Syk and other tyrosine kinases in this process will definitely help us understand how cells fight against oxidative stress. The avian B lymphoma cell line, DT40 cell system, undergoes homologous recombination at very high frequencies and makes it attractive for phenotype analysis of single and multiple gene disruptions. Dr. Qin and his coworkers have taken advantage of this model to investigate the roles of Syk in oxidative stress signaling, with emphasis on Syk activation and calcium mobilization.

How does oxidative stress activate Syk? Mutagenesis study reveals that Syk activation is the combined effects of autophosphorylation and phosphorylation by other tyrosine kinases[19]. The initiating tyrosine kinase for Syk activation is Lyn, member of Src family protein tyrosine kinase in DT40 cells[20]. Moreover, mSH2(N) Syk, but not mSH2(C) Syk, in which the phosphotyrosine-dependent binding motif within the SH2 domains is mutated, showed a significantly higher activity than that observed in wild-type Syk, indicating that SH2(N) suppress Syk activity and SH2(N) and SH2(C) of Syk play distinctive functions in oxidative stress signaling[19,21]. Besides oxidative stress, Syk is activated by osmotic stress[22], interleukin-2[23], formyl methionyl-leucyl-phenylalanine[24], ceramide[25] and differentiation of HL60 cells into granulocytes[26]. Cell shrinkage that causes aggregation of cell surface receptors and ROS generation, rather than osmolarity, is responsible for osmotic stress-induced Syk activation[27]. Syk activation plays a protective role in mediating osmotic stress / ceramide-induced apoptosis[22,25].

Oxidative stress elicits an increase in [Ca²⁺]i by stimulating both an extracellular Ca²⁺ influx and Ca²⁺ release from internal stores in DT40 cells[18,20]. Oxidative stress-stimulated calcium release from intracellular stores is achieved through the activation of phospholipase Cγ2 (PLCγ2) via a tyrosine phosphorylation mechanism, but this Ca²⁺ response is partially abolished in Syk-deficient cells[18,20], suggesting that other molecules work in concert with Syk in mediating oxidative stress-induced calcium release. To this end, Bruton’s tyrosine kinase (Btk), one of Tec-family non-receptor tyrosine kinase with one pleckstrin homology (PH) domain targeting proteins to where phosphatidylinositol 3,4,5-trisphosphate is produced, is activated by oxidative stress[28,29]. Btk deficiency resulted in a significant reduction in oxidative stress-induced calcium response, which is correlated with the significantly reduced tyrosine phosphorylation of PLCγ2 and the complete elimination of IP₃ production[29], showing that tyrosine phosphorylation of PLCγ2 by Btk is required but not sufficient for oxidative stress-induced activation of PLCγ2. How do Syk and Btk cooperate to activate PLCγ2 in response to oxidative stress? Oxidative stress stimulates a Syk-, but not Btk-, dependent tyrosine phosphorylation of B cell linker protein (BLNK). Tyrosine-phosphorylated BLNK provides docking sites for PLCγ2 and Btk as well as enables subsequent tyrosine phosphorylation of PLCγ2 by Btk[29]. Thus, activation of PLCγ2 requires tyrosine phosphorylation of PLCγ2 by both Syk and Btk. Intact SH2 and PH domains are required for a full recovery of Btk-mediated oxidative stress-induced intracellular calcium mobilization, although mutation of the SH2 (Arg307 to Ala) or PH (Arg28 to Cys) domain did not affect the oxidative stress-induced activation of Btk. Interestingly, Syk SH2 domains play different roles on substrate selectivity as Syk SH2 domains has no effect on oxidative stress-induced phosphorylation of STAT3[30]. The overall results demonstrate that functional SH2 and PH domains are required for Btk to form a complex with PLCγ2 through BLNK in order to position the Btk, PLCγ2, and phosphatidylinositol 4,5-bisphosphate in close proximity for efficient activation of PLCγ2 and to maximize its catalytic efficiency for IP₃ production[29].

Interestingly, the phosphatidylinositol 3-kinase inhibitor, wortmannin, partially inhibites the oxidative stress-induced calcium release without affecting tyrosine phosphorylation on PLCγ2[28]. Oxidative stress induces Lyn- and Syk-dependent tyrosine phosphorylation of the p110 subunit of PI3K, which does not alter its catalytic activity[31]. Fractionation studies demonstrate that oxidative stress activates PI3K pathway signaling through recruitment of p85 PI3K to the particulate fraction where its substrate resides[31], revealing that PI3K and Btk target the activated PLCγ2 to its substrate site for maximal catalytic efficiency[28,31]. In summary, Dr. Qin’s studies demonstrate a central role for Syk in regulating calcium release in B cells upon oxidative stress and reveal a sophisticate network in efficiently coordinating oxidative stress-induced calcium response via both tyrosine phosphorylation- and membrane recruitment-dependent mechanisms (Figure 3).

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Proposed model for Syk in mediating oxidative stress-induced calcium release. Phosphorylation of Syk by Lyn and autophosphorylation is required for oxidative stress-induced activation of Syk. How Lyn senses oxidative stress in DT40 cells is currently unknown. Oxidative stress might inhibit tyrosine phosphatase, thereby activating Lyn; Activated Syk phosphorylates phospholipase Cγ2 (PLCγ2) and adaptor protein B cell linker protein (BLNK). Phosphorylated BLNK provides docking sites for PLCγ2 and Btk, enable Btk phosphorylates PLCγ2. Simultaneous phosphorylation of PLCγ2 by Syk and Btk is indispensable for oxidative stress-induced activation of PLCγ2; Activated PLCγ2 cleaves phosphatidylinositol 4,5-biphosphate, releases IP₃, triggering Ca²⁺ release via binding to ryanodine receptor. Oxidative stress also activates PI3K pathway signaling through recruitment of p85 PI3K to the particulate fraction, producing phosphatidylinositol 3,4,5-trisphosphate that targets the activated PLCγ2 to its substrate site for maximal catalytic efficiency.

Studies of RPE cell inflammatory responses and oxidative stress signaling: Dry age-related macular degeneration (AMD) is a multi-factorial syndrome characterized by outer retinal and RPE atrophy and choriocapillaris degeneration, accounting for 85% of patients with AMD[2,3]. Dysfunction and subsequent loss of RPE cells are believed to be central to the etiology of this disease. Cumulative oxidative stress due to high level of oxygen consumption and lipid peroxidation in the retina constitutes a significant risk factor for the disease[2,32]. Inflammatory injury to the RPE due to necrotic and pathological activation is another risk factor for dry AMD[3]. RPE debris and injured RPE cells, by recruiting macrophages and releasing inflammatory mediators, participate in amplifying inflammatory signaling, and consequently leading drusen biogenesis and RPE atrophy. The conditions that trigger RPE cells under oxidative stress and inflammation are summarized in Figure 4. However, the molecular mechanisms for the role oxidative stress and inflammation in dry AMD are poorly understood and no therapies are currently available.

Open in New Tab   Full Size Figure   Download Figure

Figure 4 Conditions that contribute oxidative stress and inflammatory injury of retinal pigment epithelial cells and subsequent etiology of dry age-related macular degeneration. Retinal pigment epithelial (RPE) cells can be injured by environmental stress, complement attack, and metabolic stress. Injured RPE cells undergo cell death and inflammatory-related drusen biogenesis, leading to atrophy of RPE cells. AMD: Age-related macular degeneration.

Sterile inflammation induced by necrotic cell lysates is used as an in vitro model to study RPE cell inflammatory responses by measuring cytokine production and increase in RPE cell permeability. We have successfully discovered that heat-shock protein-90 (HSP90) inhibitors including 17-N, N-dimethylaminoethylamino-17-demethoxy-geldanamycin (17-DMAG) can reverse necrosis-induced cytokine production and increase in RPE monolayer permeability. Critically, a cell impermeable HSP90 inhibitor DMAG-N-oxide (DNO) can achieve similar potency in inhibiting necrosis-induced cytokine production and increase in permeability. Furthermore, DNO also completely reverses interleukin (IL)-1β/tumor necrosis factor (TNF)-α-induced increase in permeability. This finding is very important since use of cell impermeable HSP90 inhibitors will reduce the side effects associated with classical HSP90 inhibitors, in particular in cases for long-term application. Activated RPE cells by pro-inflammatory cytokines TNF-α and IL-β are used as the second in vitro model to study RPE cell inflammatory responses[33]. RPE cells treated with TNF-α or IL-1β stimulated massive production of cytokines such as IL-6, IL-8, and MCP-1 as well as expression of adhesion molecule ICAM-1. Examining signaling pathways has pinpointed to S-adenosylmethionine-dependent methyltransferases that mediate IL-1β/TNF-α-induced RPE cell inflammatory responses[33]. These observations point to novel mechanisms for regulating RPE cell inflammatory responses and provide novel therapeutic approaches in the treatment of a range of inflammation-associated retinal diseases.

Photo bleaching of photoreceptors and/or excitation of photosensitizers in RPE lipofuscin in the eye trigger lipid peroxidation, producing lipid radicals and a number of stable but chemically active aldehyde-containing breakdown products such as 4-hydroxy-2-nonenal (4-HNE). Understanding how these active lipids damage the subretina will provide tools for protection of RPE and photoreceptors from damage. In a study into the regulatory mechanisms of 4-HNE-induced RPE cell death, we have demonstrated that knockdown of expression of AMP-activated protein kinase α2 (AMPKα2), but not AMPKα1 protects RPE cells from 4-HNE-induced RPE cell death[34]. RPE cells form the outer layer of the blood-retinal barrier that control exchange of nutrients and waste products between the underlying choroidal blood vessels and overlying photoreceptors. 4-HNE causes significant increases in RPE monolayer permeability that is blocked by knockdown of AMPKα2, but not AMPKα1 expression[34]. Knockdown of AMPKα2, but not AMPKα1 expression reduces the capability of RPE cells of phagocytizing photoreceptor outer segments, thus switching RPE cells to a self-protection mode upon stress: this is due to the fact that the process of phagocytosis is a chemical and physical stress to RPE cells[35]. Taken together, our findings reveal that AMPKα2 inhibitor might be candidate suitable for the treatment of retinal diseases associated with RPE atrophy and barrier breakdown.

Tremendous effort has been made to understand the biological processes of RPE injury and inflammatory responses utilizing those models described above, thereby sorting out cyto-protective and anti-inflammatory signaling pathways in RPE cells. Revealing these signaling pathways leads to identification of AMPK, HSP90, IL-8, and methyltransferases as potential targets. Their contributions to RPE cell injury and atrophy are presented in Figure 5.

Open in New Tab   Full Size Figure   Download Figure

Figure 5 Potential roles of AMP-activated protein kinase, heat-shock protein-90, interleukin-8, and MTases in the development of retinal diseases. AMP-activated protein kinase (AMPK)α1 controls expression of antioxidant genes while AMPKα2 maintains assembly of tight junctions in retinal pigment epithelial (RPE) cells. Heat-shock protein-90 (HSP90) can function as a danger signal once released from injured cells or as a holding chaperone for assembly of inflammatory signaling complex. Interleukin (IL)-8 causes RPE cell injury via an autocrine feedback activity. MTases promote transcription of inflammation genes through methylation of proteins and DNA. MTases: S-adenosylmethionine-methyltransferases; PRRs: Pattern-recognition receptors; ISC: Inflammatory signaling components; TFs: Transcription factors.

AMPK: AMPK is a ser/thr kinase, comprising a catalytic α subunit and regulatory β and γ subunits, which senses cellular energy levels. When energy supply is compromised, activated AMPK limits energy utilization and promotes energy production to ensue cell survival. Intriguingly, AMPK has important functions in controlling phagocytosis, permeability, immune response, and polarity development in addition to cell survival. α2 but not α1 AMPK inhibition protects RPE cells from lipid-peroxidation (4-HNE)-induced oxidative stress and maintains RPE monolayer function under 4-HNE stress. Knockout of α2 AMPK also protects RPE cells from sodium iodate stress. Isoform-specific protection provides us a good opportunity to achieve targeted goals while minimizing potential side effects once the pathway is shut down.

HSP90: HSP90 has two major cytoplasmic forms, HSP90α and HSP90β. HSP90 is also expressed at the cell surface or is relocated to cell surface upon cell activation. HSP90 functions as molecular chaperone to promote the refolding of damaged proteins and to inhibit protein aggregation under stress conditions, or as “holding protein” in normal unstressed cells through binding to client proteins with unstable tertiary structures to help assembly and maintenance of conformational stability. Once released, HSP90 is shown to function as a cytokine, activating inflammatory cells. In a sterile inflammation model that mimics necrotic inflammation in vivo, occurring in dry AMD, we demonstrated that HSP90 inhibitors inhibit necrotic inflammation responses in RPE cells. Furthermore, HSP90 inhibitors block increase in RPE monolayer permeability induced by necrosis or proinflammatory cytokines. Intriguingly, these effects can be achieved by using cell-impermeable HSP90 inhibitor, revealing that inhibiting cell surface HSP90 is sufficient to block inflammatory responses, avoiding potential side effects derived from targeting intracellular HSP90.

Interleukin-8: Interleukin-8 (IL-8), a chemokine recruiting neutrophils, is a primary mediator of angiogenesis. IL-8 is released from degenerating RPE cells and associated with drusen in AMD. Increased level of IL-8 mediates LPS toxicity through an autocrine feedback loop. Polymorphism that promotes IL-8 expression modulates susceptibility to AMD and has been reported to be a risk factor for AMD development in certain population. We have detected increased expression of IL-8 in stressed RPE cells such as sodium iodate stress, hydrogen peroxide, lipid peroxidation, complement attack, and necrosis. These observations point out to IL-8 neutralizing antibodies as a potential approach for AMD therapy.

Methyltransferase: Methyltransferase carries out cellular methylation reactions by transferring methyl group from S-adenosylmethionine to proteins and DNA, which epigenetically control gene expression. The extent of methylation inhibition of methyltransferase by S-adenosylhomocysteine, a powerful competitive methyltransferase inhibitor, depends on the methyl donor S-adenosylmethionine availability, which might be influenced by inflammation. We have obtained experimental evidence showing that S-adenosylmethionine-methyltransferase mediates pro-inflammatory cytokine-induced RPE cell inflammatory responses. Inhibiting methyltransferase activity limits RPE cell inflammatory responses by reducing production of inflammatory mediators.