Survivin as a Preferential Target for Cancer Therapy

Cancer is typically a consequence of imbalance between cell death and proliferation in a way favorable to cell proliferation and survival. Most conventional cancer therapies are based on targeting rapidly growing cancerous cells to block growth or enhance cell death, thereby, restoring the balance between these processes. In many instances, malignancies that develop resistance to current treatment modalities, such as chemotherapy, immunotherapy, and radiotherapy often present the greatest challenge in subsequent management of the patient. Studies have shown that under normal circumstances, cells utilize different death mechanisms, such as apoptosis (programmed cell death), autophagy, mitotic catastrophe, and necrosis to maintain homeostasis and physiological integrity of the organism, but these processes often appear to be altered in cancer. Thus, in recent years developing various strategies for administration of cytotoxic chemotherapeutics in combination with apoptosis-sensitizing reagents is receiving more emphasis. Here, we review the properties of the anti-apoptotic protein, survivin, a member of the inhibitor of apoptosis protein (IAP) family and the clinical feasibility and anti-cancer potential of drugs targeting this protein. We also discuss some key points and concerns that should be taken into consideration while developing drugs that target apoptotic proteins, such as survivin.


Apoptosis, a Developmental and Defense Mechanism
In addition to playing critical roles in development, apoptosis is a natural defense mechanism to eliminate damaged and/or unhealthy cells that arise in many ways. In mammalian systems it is known to occur through two distinct pathways: the extrinsic pathway utilizing ligands and cell surface death receptors and the intrinsic pathway that includes the mitochondrion as a key component. Most conventional chemotherapies and radiotherapies target the intrinsic pathway. Choosing apoptosis and the intrinsic mechanism of programmed cell death as a target for cancer treatment has both merits and disadvantages for new drug development since some degree of apoptosis is necessary for organ function, but excess apoptosis, particularly in non-renewing cell types can obviously have detrimental consequences. Therefore, balance is critical and a detailed understanding of the different apoptosis pathways and their regulators, the signals triggering the activation and repression of apoptosis and, perhaps most importantly the consequences of altering apoptosis in tumors and surrounding cells, should provide novel targets for more efficient cancer therapeutics.
The ability to activate or up-regulate pro-apoptotic regulators and/or down-regulate or silence anti-apoptotic mechanisms has been facilitated by our past and current and understanding of the natural mediators of apoptosis. However, a major challenge remains in designing approaches that target efficiently and result in effective tumor cell killing but minimal effects on normal cells. Therefore, further understanding of different apoptosis pathways and their regulators, the signals triggering the activation and repression of apoptosis and importantly the consequences of apoptotic treatments in tumor microenvironments should provide novel and more efficient avenues for the development of efficacious cancer therapeutics. The major pathway that we will focus upon in this review is the mitochondria-mediated apoptosis (intrinsic) pathway.

Intrinsic Apoptosis Pathway (Mitochondrial Pathway)
The intrinsic apoptosis pathway is activated by various death-inducing stimuli, such as DNA damage signals that are often generated by chemotherapy and radiation. Several key regulators of this pathway belong to the Bcl-2 family of molecules, consisting of pro-apoptotic and anti-apoptotic proteins containing Bcl-2 homology domains (BH domains). The pro-apoptotic regulators have been subdivided into two groups: the Bax subgroup that includes Bax, Bak, and Bok proteins, having three BH domains (BH1, BH2, and BH3) and the BH3-only group that includes Bid, Bad, Bik, Bim, Noxa, Puma, and Bmf.
The anti-apoptotic group of proteins is comprised of Bcl-2, Bcl-X L , Bcl-w, Bfl-1, and Mcl-1 proteins, possessing four BH domains (BH1, BH2, BH3, BH4) [1][2][3]. One major determinant of whether a cell enters an apoptotic phase in response to cellular insults that activate upstream signaling pathways is the final stoichometry between pro-and anti-apoptotic proteins containing BH domains with anti-apoptotic members binding to, and neutralizing pro-apoptotic members. If the balance shifts past a critical value in favor of pro-apoptotic members, these proteins are able to bind at the mitochondrial membrane to decrease its integrity, resulting in membrane depolarization and release of mitochondrial contents into the cytoplasm. Cytochrome C, in particular, subsequently plays a central role in activating the apoptosome, triggering activation of the effector cysteine-dependent aspartyl-specific proteases (caspases) which then contribute to further the apoptotic cascade by digesting a broad range of cellular substrates [4].
In addition to the BH family, another major group of proteins called the inhibitor of apoptosis (IAP) proteins play a powerful role in regulating the intrinsic pathway of programmed cell death. Particularly well-characterized members of this family include XIAP (X-chromosome-linked IAP), cIAP1 (cellular IAP1), cIAP2 (cellular IAP2), Apollon, ML-IAP (melanoma IAP), survivin, IAP-like protein 2 (ILP2), and NAIP (neuronal apoptosis inhibitory protein) [5,6]. As some of these regulatory proteins have been implicated in the survival of tumor cells in response to cytotoxic stress induced by cancer therapies, they have provided an attractive target for cancer drug development that we discuss further, below.

The Structure and Function of IAP Proteins
The IAP family of proteins contain one to three baculovirus IAP repeat or BIR domains (Figure 1), which are essential to their anti-apoptotic function. Many of the cellular IAPs also possess a really interesting new gene (RING) domain at their C-termini, and some also contain a caspase recruitment domain or CARD [7].
Currently, controversy exists over the precise mechanism by which the IAPs inhibit apoptosis. However, there is evidence to suggest that XIAPs can directly bind caspases 3, 7, and 9 and inhibit their function [8] using the short regions highlighted in Figure 1. It has also been reported that this inhibition occurs through the ability of IAPs to promote ubiquitination and therefore proteasomal degradation of the caspase in question [9]. The IAP regulatory protein SMAC (Second Mitochondria-derived Activator of Caspases, or the mouse homolog DIABLO) is localized to the inter-membrane space of mitochondria under normal, non-apoptotic conditions and is released, together with cytochrome C, in response to apoptotic stimuli that affect the ratio of pro-and anti-apoptotic BH family members. As a consequence of pro-apoptotic BH family members forming dimers and permeabilizing the mitochondrion, SMAC is released and binds IAP proteins through an IAP binding motif. This displaces caspase-9 from the complex, resulting in disruption of the inhibitory activity of IAPs on caspases [10][11][12][13][14] and, thus, SMAC serves as an inhibitor of the inhibitors of apoptosis. The other three IAP proteins, c-IAP1, c-IAP2, and ML-IAP have also been reported to bind SMAC, thus, segregating SMAC from XIAP allowing it to remain active as an inhibitor. Finally, caspase-independent cell functions have also been identified for IAPs where studies in human, yeast, and C. elegans systems provide evidence for roles of IAPs in regulating cell division, particularly during cytokinesis [15][16][17][18][19].

IAP Proteins and Cancer
Undoubtedly, suppression of apoptosis is a hallmark of the majority cancers that typically become genetically unstable, which normally triggers an apoptotic response in non-cancerous cells [20]. Consistent with this idea, increased levels of different members of the IAP family have been reported in many cancer types [21][22][23] and over-expression of IAP proteins has been reported to enhance resistance to apoptotic stimuli in many malignancies [24][25][26]. Thus, a concerted effort has been mounted to further examine the precise role of IAPs in tumor development and to explore their potential as targets for cancer therapy. In this family of proteins, survivin has taken a center stage, due to its markedly specific expression in cancer cells.

Survivin, a Crucial IAP Target in Cancer Therapy
Having 142 amino acid residues, survivin (also called Baculoviral Inhibitor of Apoptosis Protein Repeat-Containing 5 or BIRC5) is the smallest IAP, with the unique characteristic of having a single BIR domain (Figure 1). Various studies have suggested different mechanisms by which survivin levels might be regulated. A summary of the studies and processes proposed to control survivin expression, protein levels or activity is displayed in Table 1.   [43] It has been believed for a long time that the survivin protein is rarely expressed in healthy adult tissues despite its expression during fetal development. Although recent reports suggest roles for survivin in normal cells [44] including T-cells [45], hematopoietic progenitor cells [46], vascular endothelial cells [47], liver cells [48], gastrointestinal tract mucosa [49], erythroid cells [50], and polymorphonuclear cells [51], survivin expression is significantly higher in transformed cells suggesting a pathological role for the protein [52]. For example, clinical studies revealed that 60% of biopsy samples from neuroblastoma (stage3-4) patients [53], 53% of colon cancer [54], 50% of high-grade lymphomas [55] and 35% of gastric cancer [56] are survivin positive. The only exception reported is low-grade lymphoma in which no expression of survivin was observed [55].

Function of Survivin
Survivin was shown to inhibit apoptosis both in vitro and in vivo [57,58], perhaps via interactions with multiple regulators of both intrinsic and extrinsic apoptosis pathways. Survivin is negatively regulated by p53, both at the mRNA and protein levels [59]. In addition, over-expression of survivin rescues a p53-induced apoptosis phenotype [59]. It has been shown that survivin inhibits Fas (CD95)-mediated apoptosis by supporting caspase3/p21 formation as a result of interaction with cdk4 [60]. In addition, survivin was shown to suppress the cell death induced by TRAIL [61] and Bax [62]. Regarding caspase-dependent roles of survivin, various, and often controversial, data have been reported. Although many studies report evidence for interactions between survivin and initiator and effector caspases [38,[62][63][64][65], some suggest that this interaction does not result in caspase inactivation [66]. These conflicting data suggest that survivin may inhibit apoptosis by caspase-independent mechanisms under certain conditions.

Survivin as a Nodal Protein
Due to its role in many different cellular actions and signaling pathways, survivin has been described as a nodal protein ( Figure 2). In addition to a role in suppressing apoptosis, survivin is also a mitotic regulator involved in various cell division processes. One of the more remarkable functions of survivin revolves around its localization at the mitotic apparatus [67]. Survivin is a component of the chromosomal passenger complex (CPC) and thereby functions as a key regulator of chromosomal segregation and cytokinesis [68]. CPC localizes to centromeres and subsequently associates with central spindle midzones and the midbody. The association of survivin with two other components of the CPC complex, INCENP (inner centromere protein antigens), and Borealin, regulates the localization of the enzymatic component, Aurora kinase B, to kinetochores [68] and, subsequently, facilitates chromosome alignment, segregation and cytokinesis during mitosis. In addition, it has been shown that DNA damage-induced activation of the checkpoint kinase 2 (CHK2) results in rapid release of survivin from the mitochondria and consequently inhibition of cell death, helping to promote tumor cell survival [69]. DNA damage stimuli also stabilize p53, which in turn can repress the transcription of survivin and help balance the degree to which activation of CHK2 promotes survivin release and activation of caspases [29]. In addition, survivin associates with microtubules contributing to spindle formation [70]. Microtubule-assembled survivin is phosphorylated at mitosis upon association with CDK1. As a consequence, survivin is stabilized during mitosis, repressing cell death in mitotic cells [38]. In addition, DNA damage-induced activation of checkpoint kinase 2 (Chk2) results in rapid release of survivin from mitochondria, inhibiting cell death and promoting tumor cell survival. DNA damage also stabilizes wild-type p53, repressing the transcription of survivin and helping to balance this pathway.

Survivin Induces Chemoresistence
Early data indicated that survivin renders resistance to chemotherapy agents in tumor cells [71]. It this study, HeLa cells expressing survivin from an adenoviral vector showed resistance to apoptosis induced by the microtubule stabilizer taxol (also called paclitaxel). Subsequently, it was shown that forced expression of wild-type survivin suppressed cytotoxicity induced by taxol in human ovarian [72] and prostate cancer cell lines [73] and taxol-induced cell cycle arrest leads to increase in the expression of survivin [74].
As noted above, significantly high expression of survivin has been reported in many cancers including lung, colon, breast, prostate, bladder, breast, laryngeal, uterine, hepatocellular, soft tissue sarcomas, renal, pancreas, high-grade lymphomas, neuroblastomas, and gastric cancers, as well as in hematologic cancers, such as myelodysplastic syndrome and acute leukemias [53][54][55][56]75]. High levels of survivin expression have also been observed in colorectal cancer cell lines resistant to the tumor necrosis factor-related apoptosis inducing ligand (TRAIL) [76], as well as in prostate cancer [77] and thyroid cancer cell lines resistant to cisplatin [78]. Survivin was also suggested to induce resistance to flutamide anti-androgen therapy in prostate cancer cells [79]. These data suggest the exciting possibility that more efficient chemotherapeutic approaches can be developed in combination with inhibitors of survivin.

Survivin Induces Radioresistence
There is also evidence that survivin suppresses radiation-induced cytotoxicity. In experiments using pancreatic cancer cell lines, survivin mRNA expression and radiosensitivity showed an inverse relationship [80]. In the same set of experiments, survivin expression was increased upon treatment with a sublethal dose of X-irradiation. This inverse relationship between survivin expression and radiosensitivity has also been observed in colorectal cancer, glioblastoma, and melanoma cell lines [81,82]. Taken together, these findings suggest that the inhibition of survivin has the potential to also enhance the effects of radiotherapy in cancer patients.

Survivin as a Cancer Diagnostic Marker
As mentioned above, expression of survivin is substantially different in tumor cells compared to normal cells. In the past few years survivin has emerged as a potential early predictor of malignancies. In a study on oral cancerous lesions, 33% of precancerous (without malignant progression) and 94% of precancerous lesions that evolved into full-blown squamous cell carcinomas showed expression of survivin [83]. Interestingly, tumors that transformed from these precancerous lesions preserved 100% of their survivin positivity. These promising reports suggest that survivin could serve as a marker for the diagnosis of malignancies at early stages.

Survivin as a Cancer Prognostic Marker
Survivin's over-expression at the mRNA and proteins levels appears to correspond with higher malignant grades and reduced survival rates in different cancers such as hepatocellular carcinoma [84], esophageal cancer [85], glioblastoma [86], lung cancer [87], B-cell non-Hodgkin's lymphoma, and breast cancer [88,89] patients. In a study of 55 patients with hepatocellular carcinoma after hepatectomy, the recurrence rate was significantly higher in those with survivin-positive tumors. Moreover, the 1-and 3-year survival rates for survivin-positive patients were substantially lower than for survivin-negative patients [84]. Likewise, reports in many other cancer patients are in agreement with the above-suggested correlation between survival and recurrance rates with survivin levels [89]. Interestingly, few reports suggest the localization of survivin as a prognostic biomarker. A report on patients with glioblastoma indicated that while cytoplasmic survivin levels did not have any effect on prognosis, nuclear survivin localization correlated with significantly lower survival rate than that of patients with low nuclear survivin levels [86]. On the other hand, nuclear survivin level did not show any correlation with survival in non small cell lung carcinoma (NSCLC) patients [90]. Thus, using survivin as a general prognostic marker needs further verification, however, so far findings in several cancers show clear promise for this potential biomarker.

Therapeutic Targeting of Survivin
Significant efforts have been focused on developing strategies to use survivin as a target for therapeutics in cancer. Developing drugs that target survivin might initially seem difficult because survivin is not an enzyme nor it is a cell surface protein. However, considerable progress has been made to achieve optimal efficiency in suppressing survivin. These strategies have been conducted using a variety of approaches as outlined in Figure 3. Among these the most promising therapies are discussed below.

Transcriptional Inhibitors
Antisense oligonucleotides (AO) are short, single stranded RNA or DNA sequences that are complimentary to a specific RNA strand that act by hybridization to the target mRNA strand to suppress the expression of the particular gene. The first attempt for AO therapy targeting survivin triggered apoptosis in human melanoma cell lines [91]. Successively, subsequent studies using AOs against survivin confirmed that small chemically synthesized oligonucleotides or expression vectors encoding them are capable of specifically inhibiting survivin either at mRNA or protein levels. This led to decreased cell proliferation and increased caspase-dependent apoptosis in different tumor (lung, sarcomas, lymphomas, thyroid, head, and neck) cell lines [92][93][94][95][96]. Furthermore, down-regulation of survivin via AO was shown to enhance sensitivity to cytotoxic agents such as TRAIL [61,63], cisplatin [93], taxol [97], imatinib [98], etoposide [93,99,100], as well as to cytotoxicity induced by radiation therapy [101]. Moreover, treatment with survivin AO resulted in inhibition of tumor growth in murine models [102]. Although these reports using AO-therapy are very promising, caveats worth mentioning are the inefficient neutralization of the target mRNA and in vivo instability. Despite these limitations, the first AO drug, LY2181308, has entered phase II of clinical trials.
Ribozymes are small RNA molecules that cleave target RNA by their endonucleolytic activity. In different studies, transfection of human melanoma cell lines with ribozymes led to reduced expression of survivin protein and augmented caspase-9-dependent cell death and sensitivity to cis-platinum, topotecan, and radiation [81,103,104]. The potential caveat to use of ribozyme administration is that ribozymes degrade easily and cause aberrant cell trafficking. However, strategies to overcome these limitations and increase the efficacy of these drugs for use in vivo are being developed. Thus far, none of the drugs of this family have entered clinical trials.
Small interfering RNAs (siRNAs) are short, double-stranded RNAs that inhibit gene expression. SiRNAs are often more effective than other antisense approaches, rendering them better candidates for clinical treatments because lower concentrations of them is needed for each treatment resulting in less side effects compared to other techniques. The first RNAi-mediated inhibition of survivin resulted in delayed mitosis, misaligned chromosomes and accumulation in prometaphase in Hela cells [105,106]. Further, in preclinical studies that attempted to utilize RNAi technology (either plasmid vectors encoding short hairpin RNAs or chemically synthesized siRNAs) to target survivin resulted in decreased survivin expression, enhanced cell death, reduced cell proliferation and inhibited tumor growth [106][107][108][109][110][111][112][113][114][115]. Suppression of survivin by RNAi also enhanced the sensitivity of cancer cell lines to multiple treatments such as vincristine [116], 17-allylamino-17-demethoxygeldanamycin [108], doxorubicin [113,115], APO2L/TRAIL [109], TNF-alpha [113], and, as with AO and ribozymes, to radiation [113,114].

Hsp90 Inhibitors
It is believed that interaction of survivin with the molecular chaperone Hsp90 stabilizes the survivin protein [33]. Shepherdin is a cell-permeable antagonist of the Hsp90-survivin complex and acts through counteracting the binding of these two proteins. In addition, shepherdin was shown to destabilize many Hsp90 client proteins such as Akt, telomerase and CDK6. Moreover, shepherdin induces caspase-dependent and caspase-independent cell death in various cancer cells both in vitro and in vivo and suppresses growth in breast and prostate cancers in xenograft models [117]. Treatment with shepherdin has not shown any effect on normal cells except for a weak tocxicity on normal fibroblasts upon increase in drug dose. Shepherdin is currently in the pre-clinical phase.

Cyclin-Dependent Kinase (CDK) Inhibitors
CDK inhibitors can counteract mitotic phosphorylation of survivin on Thr 34 and thereby lead to the destruction of survivin and its function. Flavopiridol and purvalanol A are members of this group of inhibitors. Strikingly, treatment with CDK inhibitors in vivo, results in loss of survivin expression, escape from taxol-mediated mitotic block, mitochondrial-dependent and p53 independent apoptosis, as well as robust antitumor activity [74]. Interestingly, inhibition of survivin phosphorylation was reported to be the potential underlying mechanism that some of the CDK inhibitors act by to induce taxol-induced apoptosis in HeLa cells [118].

Promoter Inhibitors
The imidazolium-based compound YM155 has, so far, emerged as the most important in this group of low molecular weight antagonists. YM155 specifically inhibits survivin gene expression by suppressing promoter activity and subsequent gene expression in cervical, colon, lung, and other cancer cells. It also diminishes proliferation and tumor growth in NSCLC, lymphoma and prostate cancer xenografts [119]. YM155 is currently in phase II clinical trial in patients with refractory diffuse large B-cell lymphoma. At the same time, a phase I clinical trial is evaluating the feasibility of combination-therapy of YM155 with docetaxel and prednisone for patients with hormone refractory prostate cancer.

Other Low Molecular Weight Antagonists
Another low molecular weight antagonist of survivin is terameprocol, also called EM-1421. The basic function of this drug is believed to be the inhibition of Sp1 regulated proteins such as survivin and Cdc2, which subsequently inhibits tumor progression [120,121]. Terameprocol is in phase I clinical trials for refractory solid tumors and lymphoma and in a phase II clinical trial for leukemia.

Immunotherapy
The recognition of specific tumor-associated antigens by T lymphocytes forms the basis for cancer immunotherapy. Earlier reports using immunotherapy for targeting survivin, revealed that cytotoxic CD8 + T lymphocytes have significant cytolytic activity against specific survivin epitopes in vitro and in vivo. Interestingly, a strong rationale for developing survivin-based anti-cancer vaccines emerged when the above findings were confirmed in mouse models of lymphoma and pancreatic cancer by the identification of cytotoxic CD8 + T lymphocytes that were responsive to some of survivin's epitopes. These vaccines were able to induce tumor suppression in multiple animal cancer models including lung [122], lymphoma, neuroblastoma [123], and pancreatic and hormone refractory prostate cancers [124]. Importantly, no significant toxicity was observed in studies that have been reported to date. Promising results from initial survivin-based vaccines against different malignancies have led to many ongoing investigations focusing on finding the epitopes that generate the strongest immunodominant, immunoprevalent T-cell reaction against survivin. Many of these drugs are currently in phase I or phase II clinical trials.

Dominant-Negative Mutants
The use of gene therapy was one of the earliest strategies used to inhibit survivin activity. In an initial study, induction of apoptosis was observed upon transfecting melanoma cells with survivin dominant-negative expression constructs [91]. Subsequent studies reported that transfection with dominant-negative mutants of survivin led to suppression of growth and increased apoptosis in gastric cancer cell lines [125], and to suppression of tumor growth in breast cancer [126] and thymic lymphoma [102] animal models. In addition, survivin dominant-negative therapy increased sensitivity to cisplatin and 5-fluorouracil [125]. Interestingly, mice expressing dominant-negative survivin showed decreased probabilities of developing tumors or exhibiting tumor-associated angiogenesis [125].
The results from experiments with other dominant-negative mutants of survivin agreed with those of the first mutation used. They also showed that treating different cancer cell lines [126], as well as melanoma [127] and breast cancer [126,128] xenografts with dominant-negative mutants of survivin resulted in suppression of tumor growth and angiogenesis [128] and augmentation of apoptotic responses. Intriguingly, in vitro data showed no effect of these mutants on the growth of normal endothelial, fibroblast or smooth muscle cells [126].

ZFN, TALEN, and CRISPR/Cas-Based Methods, as Potential Methods for Targeting the Survivin Gene in Tumor Cells
In the last decade, the advent of novel "genome editing" techniques, which are based on site-specific-directed nucleases, has raised hopes for enhanced feasibility of targeting the main causes of many diseases. The potential for therapeutic application of Zinc-finger nuclease (ZFNs), transcription activator-like effector nucleases (TALENs) or clustered regulatory interspaced short palindromic repeat (CRISPR/Cas-based) RNA-guided DNA endonucleases [129][130][131] has enabled a paradigm-shift compared to conventional gene therapy methods.
In general, such genome modifying tools are composed of two main domains: (1) a designable and sequence-specific DNA-binding domain; and (2) a nuclease domain, which modifies the specific sequence of interest by leaving double-strand breaks behind, in the locus. The damage, in the form of double-strand breaks, will further undergo error-prone mechanisms of DNA repair, such as non-homologous end joining (NHEJ) or homology-directed repair (HR), thereby, impacting the locus by an extended variety of mutations such as gene disruption, insertion, deletion, inversion, addition or correction.
To date, conventional genome engineering techniques have been limited by several factors such as low efficiency, low specificity and labor-intensity. However, the emergence of the novel strategies in disease treatment, such as ZFN, TALEN, or CRISPR/Cas-based RNA-guided DNA endonucleases presents promise to overcome such limitations in future. This new technology, with further high-throughput validations of stringency and specificity, could potentially revolutionize cancer treatment. Performing personalized "gene-manipulation" therapies for oncogenes in cancer patients would be one possibility following from the current efforts to optimize targeted nucleases. For example, designing ZFN or TALEN to target the survivin gene in cancer cells, would trigger cell death specifically in transformed cells, while having minimal effect on normal cells. As with other gene therapy approaches, a major limitation is the likely requirement to target all, or a great majority of cancer cells for the therapy to be durable so the development of more efficient vectors may be the step limiting effectiveness of this approach.

Potential Caveats and Alternative Approaches
Despite the considerable potential for altering survivin levels as a method for helping to selectively eliminate cancer cells, there are some cautions that have been overlooked to various degrees. For example, the heterogeneous microenvironment of tumors has strong influences on tumor progression and also on the effectiveness and/or toxicity of tumor suppressor therapeutics. Indeed, apoptotic cells can induce immunosuppressive and anti-inflammatory effects on their neighboring cells. Specifically, repression of pro-inflammatory factors such as TNF-α, IL-1β, IL-8, and IL-12, as well as production of TGF-β (as anti-inflammatory mediator), are induced through contact with apoptotic cells [132][133][134][135]. Cells juxtaposed with apoptotic cells also generate differentiation and growth signals. These signals include growth and even pro-angiogenesis factors, such as VEGF and hepatocyte growth factor [136][137][138]. In addition, FKN, a survival and proliferative factor, is secreted from apoptotic cells [139]. Furthermore, release of pleiotropic anti-inflammatory factors such as TGF-β, lactoferrin and IL-10 by apoptotic cells can result in growth and differentiation in neighboring cells, favoring overall tumor progression. This phenomenon is best described in vivo in Drosophila and is referred to as "compensatory proliferation". In this case, apoptotic cells stimulate the neighboring surviving cells to proliferate. Considering the possibilities of any type of "compensatory proliferation" in response to apoptosis-induction therapies would be crucial because it directly influences the outcome of the treatment in favor of either tumor regression or growth. However, this is obviously not a caution unique to survivin-based apoptotic therapeutics and indeed it is an important issue that should be taken into consideration in pre-clinical stages of development of any cytotoxic drug that activates apoptosis.

Conclusions
Developing strategies that cause cell death specifically in tumors is a central theme in cancer treatment. However, after many years of considerable efforts, fundamental challenges such as tumor specificity, sustainability and toxicity of treatments and durability of patient response remain. Recently, an additional concern has arisen regarding the potential of cytotoxic cancer therapeutics to have growth-promoting effects on neighboring normal cells. The initial studies that noted the idea of tumor repopulation after cytotoxic treatments originated more than 40 years ago [140,141], where tumor cell repopulation was observed after exposing tumors to radiotherapy. The mechanisms underlying increased proliferation after inducing cell-death in tumor cells have been elusive for many years. In this context, a recent study by Huang et al. [142] showed that caspase-3 activation in response to radiotherapy stimulates tumor repopulation. This has generated concerns with respect to some of the deleterious consequences of cytotoxic agents under certain circumstances. Thus far, adjuvant use of inhibitors of repopulation-stimulating pathways, along with cell-death-inducing therapies, is one possible approach to overcome this potential concern.
The differential expression of survivin in malignant versus normal cells is a strong rationale for development of survivin-based cancer therapeutics, from which many ongoing trials have ensued using a variety of methods to alter survivin or target it in cancer cells. One method, taking advantage of tumor-specific promoters like that of survivin, may allow future studies to potentially direct the expression of other relevant apoptosis-inducing genes or apoptosis-inhibiting mutants in tumors. For example, in a recent study [143], the survivin promoter was used to direct expression of other apoptotic genes selectively in tumor cells. Results from these and other approaches support the idea that the survivin gene and promoter might have continued utility for developing novel agents to target and attack tumors of various types.