on only in the presence of atRA and G-CSF. Enhanced differentiation is accompanied by a more pronounced growth arrest (Gianni et al., 1994). Similar effects are observed in blasts obtained from a few cases of chronic myeloid leukemia (CML) and in some cases of non-M3 AML, but not in atRA-resistant myeloid leukemia cell lines. From a therapeutic perspective, one potential drawback of a combined treatment of the leukemic blast with atRA + G-CSF is a decrease in the secondary apoptotic response to the retinoid. Indeed, G-CSF is a well-known survival factor for the differentiated granulocyte (Colotta et al., 1992). However, at least in vitro, the apoptotic effect of atRA seems to dominate and mask the antiapoptotic activity of G-CSF. The molecular mechanisms underlying the potentiating effect of G-CSF are likely to be straightforward and the result of retinoid-dependent effects on the G-CSF signal transduction pathway. In fact, atRA causes a marked increase in the surface expression of the G-CSF membrane receptor (G-CSFR) (Gianni et al., 1994; Tkatch et al., 1995). As a result of this, G-CSF binding to G-CSF-R results in a more pronounced activation of the downstream signaling pathway, with increased phosphorylation of signal transduction and activator of transcription (STAT3) (E.G., unpublished observations). So far there are no reports describing a modulation of the nuclear RAR pathway by G-CSF. The enhancing effect of the cytokine is relatively specific, as other molecules of the same family, like GM-CSF, TNFa, and TGF-b, do not potentiate the differentiating activity of atRA in APL cells. As a matter of fact, TNFa seems to exert an inhibitory action on the atRA-induced granulocytic maturation of myeloid leukemia cells orienting them along the monocytic pathway (Witcher et al., 2004). It remains to be proved that this combination is therapeutically significant in preclinical and clinical models of myeloid leukemia. However, it is interesting to notice that there is an anecdoctal report on a complete remission achieved in a relapsed APL patient, following combined treatment with atRA and G-CSF (Shimodaira et al., 1999).

The superfamily of TGFs-b includes a large number of structurally related polypetides with prominent roles in cell proliferation, differentiation, and death. Interactions between retinoids and TGF-b have been the object of numerous studies. In particular, TGF-b is induced in a variety of cell types and the effects of combinations between the growth factor and atRA on cellular differentiation have been studied in a number of models (Cao et al., 2003; Roberts and Sporn, 1992; Sporn and Roberts, 1985; Sporn et al., 1989; Wakefield et al., 1990). In skin epidermis, the actions of TGF-b and atRA on normal keratinization are synergistic, whereas those on abnormal differentiation associated with hyperproliferation are antagonistic (Choi and Fuchs, 1990). TGF-b enhances neuronal differentiation of human neuroblastoma SH-SY5Y cells treated with atRA, as determined by the increase in the length and density of neurite outgrowth as well as the measurement of a specific maturation marker like tyrosine hydroxylase (Gomez-Santos et al., 2002).

TGF-b has been shown to activate the monocytic differentiation of AML cell lines (De Benedetti et al, 1990; Turley et al, 1996; Walz et al, 1993). When HL-60 cells are treated simultaneously with atRA and TGF-b, a mixture of granulocytic and monocytic cells is observed. In this paradigm, atRA and the growth factor compete with each other and commit cells toward two different types of maturation processes. Antagonistic interactions may be the consequence of a retinoid-induced phosphatase mediating the dephosphorylation and inactivation of SMAD2, a downstream effector of TGF-b (Cao et al., 2003). The growth of retinoid sensitive MCF-7 breast carcinoma cells is inhibited synergistically by combinations of TGF-b and atRA (Turley et al., 1996; Valette and Botanch, 1990). However, to the best of our knowledge, interactions between the two compounds on the differentiation of this type of tumor cells have not been reported. From a mechanistic point of view, it is interesting to notice that TGF-b induces the expression of AIB1 (Lauritsen et al., 2002), a well-known coactivator of the RXR/RAR transcriptional complex, suggesting that the growth factor has the potential to modulate the retinoid-signaling pathway in a positive fashion.

Interferons (IFNs) are probably the cytokines or growth factors that have been studied more extensively in relation to the biological and pharmacological activity of retinoids (Bollag, 1994; Chelbi-Alix and Pelicano, 1999; Eisenhauer et al., 1994; Fossa et al., 2004; Smith et al., 1992). This is explained by the fact that interferons were among the first cytokines to be identified and characterized. In addition, type I IFNs in combination with 13-cis RA have been used in chemoprevention studies of head and neck cancer with encouraging results (Shin et al., 2001). IFN synthesis is activated as part of the cellular response to viral infection. Various types of IFNs have been described and classified as type I (IFNa and IFNb) or type II (IFNg). While IFNa and IFNb are synthesized by the majority of cell types, IFNg is produced only by cells of the immune system. IFNs act predominantly as autocrine or exocrine factors and bind to specific membrane receptors which are coupled to the JAK/STAT and mitogen-activated protein kinase (MAPK) pathway. In particular, binding and phosphorylation of IFN receptors lead to recruitment and activation of JAK kinases and subsequent phosphorylation of STAT1 or other types of STATs. Activated STAT1, in the form of a homodimer (IFNg) or a heterotrimer (STAT1-IGF3-p48, IFNa, or IFNb), migrates to the nucleus and acts as a transcription factor regulating the expression of numerous genes (Ramana et al., 2000, 2002; Stark et al., 1998). IFNs have also been shown to activate the MAPK pathway via p38MAPK (Kovarik et al., 1999). Type I IFNs exert predominantly growth inhibitory effects acting on various intracellular systems.

Numerous reports demonstrate cross talk between the retinoid and IFN pathway in various cell types, both in terms of cytodifferentiation and growth inhibition (Chelbi-Alix and Pelicano, 1999; Dimberg et al., 2000,

2003; Garattini et al., 1998; Gianni et al., 1996b, 1997; Lembo et al., 1992; Matikainen et al., 1996, 1997, 1998; Pelicano et al., 1997). The most interesting results on the cytodifferentiating effects of these types of combinations have been reported in the context of the APL or the AML blast. In NB4 cells, we demonstrated that atRA and synthetic RARa agonists are endowed with an IFN-mimetic action. Treatment of this cell type causes a rapid phosphorylation and activation of STAT1, which results in the transcription of IFN-responsive genes. STAT1 activation is followed by transcriptional stimulation of the corresponding gene and subsequent increases in the levels of the encoded protein (Gianni et al., 1997). A late event triggered by atRA is represented by stimulation of IFNa secretion, which suggests that retinoids may trigger an autocrine loop involving the cytokine (Gianni et al., 1997). Induction of STAT1 by retinoids is not cell specific, as it is observed also in breast carcinoma cells (Kambhampati et al., 2004), and may be mediated by rapid induction of the transcription factor interferon-responsive factor 1 (IRF1) (Percario et al., 1999) or activation of PKCd (Kambhampati et al., 2003). In breast carcinoma cells, atRA-dependent STAT1 induction is associated with reversion of natural resistance to the growth inhibitory action of IFNs (Kolla et al., 1996, 1997; Lindner et al., 1997; Moore et al., 1994). Combinations of type I or type II IFNs with atRA in myeloid leukemia blasts causes enhanced cytodifferentiation, at least in terms of expression of some granulocytic maturation markers (Chelbi-Alix and Pelicano, 1999; Gianni et al., 1996b). STAT1 and IRF1 induction is relevant not only for the observed interactions between retinoids and IFNs but also for the cytodifferentiating effect of atRA in APL cells. Indeed, selective suppression of STAT1 and/or IRF1 results in inhibition of the myeloid differentiation program activated by atRA in this cell type (Dimberg et al., 2003). Further points of contact between retinoids and IFNs are represented by the fact that both types of stimuli activate the p38MAPK and PKC pathways (see Section IV.C). Although largely unexplored these phenomena may also be at the basis of the observed additive or synergistic intercations between the retinoid and IFN systems. For instance, PKCd is required for the generation of the synergistic effects of IFNa and atRA on gene transcription. Such regulatory effects on transcription are mediated by the atRA-inducible, PKCd-dependent upregulation of STAT1 protein expression.

As a concluding remark, it must be underscored that combinations of IFNs and retinoids have been the basis of numerous clinical trials in different types of neoplasia and have provided some encouraging results. In the CML context, where IFN is one of the treatment mainstay, some interesting data have been published (Egyed et al., 2003). The cytogenetic responses during the chronic phase of 11 patients with CML treated with atRA + IFN were compared with those of 9 other CML patients treated with IFN alone. The preliminary results suggest that the atRA + IFN combination may be superior in achieving cytogenetic remission in the first chronic phase of

CML (Egyed et al., 2003). It remains to be established whether the positive clinical effects reported for the combinations of retinoids and IFNs in this as well as other clinical contexts is related to enhanced cytodifferentiation of the neoplastic cell.


The intracellular second messenger, cAMP, controls many aspects of the cellular homeostasis, including proliferation, differentiation, and apoptosis. Adenyl-cyclases are coupled to many membrane receptors and control the synthesis of cAMP from ATP. The levels of intracellular cAMP are modulated in a negative fashion by phosphodiesterases, a large family of enzymes catalyzing the hydrolysis of cAMP. Elevation of intracellular cAMP results in the activation of the cAMP-dependent protein kinase (PKA), which phos-phorylates and controls the state of activation of a number of protein substrates. PKA is a dimer consisting of a regulatory and a catalytic subunit. Binding of cAMP to the regulatory moiety causes release and activation of the catalytic subunit. A second and less studied effector molecule stimulated by cAMP is EPAC, aGTP-exchanging factor, mediating some of the biological effects of the cyclic nucleotide.

Most of the knowledge on the cross talk between the retinoid and the cAMP-signaling pathways have been acquired in two systems: AML and neuroblasto-ma. In AML, stimuli capable of increasing the intracellular levels of cAMP potentiate the cytodifferentiating and growth inhibitory effects of atRA (Breitman et al., 1994; Imaizumi and Breitman, 1987; Parrella et al., 2004; Yang et al., 1998). Cotreatment with atRA and cell permeable cAMP analogues, like dibutyryl-cAMP or 8Cl-cAMP as well as specific phosphodiesterase IV inhibitors, like piclamilast, results in a more rapid and more efficient granulocytic maturation of HL-60 and NB4 cells relative to treatment with the retinoid alone (Altucci et al., 2005; Garattini and Gianni, 1996; Gianni et al., 1995c; Guillemin et al., 2002; Kamashev et al., 2004; Parrella et al., 2004; Taimi et al., 2001). This is accompanied by the expression of terminally differentiated granulocytic maturation markers, like leukocyte alkaline phosphatase, which are not induced by atRA alone (Garattini and Gianni, 1996; Gianni et al., 1995c). Interestingly, treatment with cAMP-elevating agents in the absence of retinoids is associated only with a mild growth inhibitory effect in NB4 cells. In contrast, HL-60 cells undergo a certain level of granulocytic maturation on treatment with cell permeable analogues of cAMP alone (Breitman et al., 1994; Imaizumi and Breitman, 1987). Combinations of dibutyryl-cAMP and atRA induce myeloid maturation in a subset of freshly isolated non-M3 AML and chronic myeloid leukemia cells (Gianni et al., 1995c). These data suggest that cAMP-elevating agents not only potentiate the activity of atRA in retinoid sensitive cells but also sensitize otherwise refractory blasts to the cytodifferentiating action of retinoids.

The potentiating effect of cAMP on retinoid activity requires the presence of active RARs, as similar phenomena are not observed in HL-60 and NB4 sublines made resistant to atRA and showing inactivating mutations at the level of the ligand-binding sites of the nuclear receptors, RARa or PML-RARa.

The retinoid-sensitizing effect observed on elevation of intracellular cAMP augments neuronal differentiation of neuroblastoma, suggesting that the phenomenon is not limited to particular cell contexts and may be of more general therapeutic interest (Abemayor and Sidell, 1989; Holtzer et al., 1985). In F9 mouse teratocarcinoma cells, addition of cAMP-elevating agents to atRA reprograms cells along the visceral endoderm maturation pathway, suggesting that combinations of the two stimuli causes not only quantitative but also qualitative changes in the process of cellular differentiation (Rochette-Egly and Chambon, 2001).

The molecular mechanisms underlying the cross talk between the cAMP and the retinoid pathways are still incompletely defined. However, in myeloid leukemia cells, it is clear that potentiation is the result of a PKA-dependent and not of an EPAC-dependent event. In fact, only selective PKA inhibitors suppress induction of the granulocytic maturation markers triggered by the combination of cAMP-elevating agents and atRA. Whether this is related to the rapid and short-lived activation of PKA observed in HL-60 and NB4 cells treated with the retinoid alone is not yet known. A plausible mechanism at the basis of the cross talk is represented by the direct effects exerted by PKA on the nuclear RARs (Parrella et al., 2004; Rochette-Egly et al., 1995). RARa is phosphorylated by PKA and this phosphorylation seems to be necessary for the full activation of the receptor. In fact, cAMP-elevating agents, like piclamilast, cause a marked stimulation of the ligand-dependent activation of RARa. Similar effects are observed when RARa is substituted with the APL-specific fusion protein, PML-RARa. Relatively specific PKA inhibitors block the stimulating effects of cAMP analogues or phosphodiesterase IV inhibitors on the ligand-dependent activation of RARa or PML-RARa (Parrella et al., 2004). This is consistent with the presence of a PKA-sensitive and key serine residue (Ser/369) in the ligand-binding domain of the nuclear RAR (Rochette-Egly et al., 1995). Mutation of this serine to an alanine knocks down the potentiating effect afforded by picla-milast on the ligand-dependent activation of RARa (Parrella et al., 2004). If this mechanism of action is really operative, the prediction is that cAMP-elevating agents should have important effects on the expression of the majority if not the totality of RXR/RAR-dependent genes.

An interesting observation relates to the ability of cAMP to activate not only RXR/RAR heterodimers but also the RXR/RXR homodimers. Activation of RXR/RXR complexes by rexinoids is associated with induction of an apoptotic response in HL-60 cells (Nagy et al., 1995). This phenomenon is not accompanied by granulocytic maturation of the leukemic blast (Nagy et al., 1995). Surprisingly addition of cAMP cell permeable analogues to rexinoids induces cytodifferentiation of retinoid sensitive as well as retinoid-insensitive or resistant myeloid blasts (Altucci and Gronemeyer, 2002; Altucci et al., 2005; Benoit et al., 1999). The observation is of potential therapeutic significance, although the molecular mechanisms underlying RXR/RXR activation by cAMP are not well understood.

A key question is whether all these observations and findings can be translated into therapeutic effects in vivo. In preclinical models of APL, this seems to be the case. Indeed, we demonstrated that combinations of picla-milast and atRA are superior to the single components of the mixture in increasing the survival of immunodeficient SCID animals transplanted with NB4 cells. At present the contribution of cytodifferentiation to the overall antileukemic effect of combinations between piclamilast and atRA has not been determined. Nevertheless, administration of this association is well tolerated and does not produce significant toxicity. Although the effect is significant, the dosage and schedule of the combined treatment need to be optimized. Similar results were reported in a different model after prolonged administration of 8-Cl-cAMP and atRA using infusion pumps (Guillemin et al., 2002). The results obtained are promising and suggest that the approach leads to an increase in the therapeutic index of atRA. The approach needs to be tested in other leukemia and cancer models. This is particularly relevant as cotreatment with cell permeable analogues of cAMP and atRA sensitizes human neuroblastoma cells to the cytotoxic actions of chemothera-pics like doxorubicin, melphalan, and BCNU (Carystinos et al., 2001). The observation suggests that cytodifferentiating therapy with cAMP-elevating agents and atRA could be combined effectively to classical chemotherapy of the neoplastic disease.

0 0

Post a comment