The BIRC7 gene is about 4.6 Kb (start: 63235883 bp and end: 63240507 bp; orientation: plus strand). There are two transcript variants deposited in the NCBI database (https://www.ncbi.nlm.nih.gov/gene). The transcript variant 1 (cDNA: 1334 bp) encodes the longer isoform (alpha; BIRC7α) (298 aa). The transcript variant 2 (cDNA: 1280 bp) presents an alternate in-frame splice site and encodes the shorter isoform (beta; BIRC7β) (280 aa). Isoform α blocks staurosporine-induced apoptosis and augments killing by natural killer (NK) cells, while isoform β blocks etoposide-induced apoptosis and protects against NK cell. There is an additional transcript variant reported in Ensembl (http://www.ensembl.org/), which contains a cDNA of 593 bp, and generates a protein with 193 aa.

The IAPs (Inhibitors of Apoptosis Proteins) are a family of proteins recognized, predominantly, for their role in preventing apoptotic cell death by directly or indirectly hindering caspase activity. Nevertheless, these proteins are further involved in other signaling pathways associated to cell survival, cell cycle and cell migration (Gyrd-Hansen & Meier, 2010; Owens et al., 2013; Plati, Bucur & Khosravi-Far, 2011). The presence of, at least, one BIR domain (baculovirus IAP repeat), a conserved sequence of about 80 amino acid residues with Zn⁺² in the center, which mediates the protein-protein interactions with caspases being essential for their anti-apoptotic activity, is the structural distinctive of the family (Budhidarmo & Day, 2015; Deveraux & Reed, 1999; Lopez; Meier, 2010). Eight IAPs have been identified within the human proteome - NAIP (BIRC1), cIAP1 ( BIRC2), cIAP2 ( BIRC3), XIAP (BIRC4), survivin ( BIRC5), Apollon/Bruce ( BIRC6), livin/ML-IAP (BIRC7) e ILP-2 ( BIRC8), and these may present either one or three BIR domains arranged in their N-terminal portion (Budhidarmo & Day, 2015; Deveraux & Reed, 1999; Lopez & Meier, 2010).
BIRC7 is a 39 kDa member of the IAP family, structured with a single BIR domain, added with a RING domain on the C-terminus portion, as firstly described by Vucic and colleagues in the year 2000 (Figure 1). Its BIR domain is globularly assembled, with four α-helixes and a three-strand anti-parallel β-sheet (Hinds et al., 1999). In turn, the RING-type zinc-finger domain, alike in other RING-bearing IAPs, has ubiquitin-ligase (E3) activity and, thus, is associated with the ubiquitination functionality; however further studies have demonstrated additional and yet unclear roles in addressing apoptotic activity (Ma et al., 2006).
A distinctive feature of this protein is that, unlike any other IAP, following a strong apoptotic incitement, BIRC7 is cleaved by CASP3 and CASP7 (caspases-3 and -7) at Asp52 to give a truncated form, which, paradoxically, promotes cell death (Nachmias et al., 2003). By such trait, rather than merely an apoptosis inhibitor, BIRC7 may actually be allowed to a broader title, one of a cell death regulator. Furthermore, through alternative splicing of the mRNA, BIRC7 has two splice variants, BIRC7α and BIRC7β, which have different subcellular distribution and distinct anti-apoptosis properties. Still, the BIR and RING domains are identical in both isoforms, and their only structural difference is assumed by the 18 amino acids present between these domains in the α isoform, allowing the formation of an α-helix linker (Ashhab et al., 2001; Li et al., 2013a).
For BIRC7, there are currently a number of unanswered questions concerning the protein interaction with caspases, the paradoxical activities it can undertake during apoptosis and, also, on specific functions and affinities of the α and β variants. Still, taken the range of activities and regulatory motifs of the protein, BIRC7 has been regarded as an interesting target for cancer therapy. As this IAP displays such unique pro- and anti-apoptotic properties, a treatment strategy could involve the promotion of BIRC7 cleavage, directing accumulation of the truncated protein, in an attempt to tip the scale towards the pro-apoptotic effect, to counteract apoptosis resistance promoted by IAPs and other disrupted signaling pathways present in cancer cells (Wang et al., 2008). Other therapeutic opportunities include targeting BIRC7 at the transcriptional level using antisense oligonucleotides, thus, reducing the expression levels of the protein. Interestingly, antisense IAP therapy is also under clinical testing for XIAP and BIRC5, with promising candidates (Xia et al., 2002; Hu et al., 2003).

Much like most IAPs, BIRC7 has low or even no detectable expression in normal differentiated adult tissues, but is present in high levels in placenta, spleen, lymph nodes, developing/embryonic tissues and in a variety of tumor types (Li et al., 2013a). In fact, BIRC7 overexpression has been recorded in numerous cancers, however, it is best linked to the malignancy in which it was preferentially overexpressed when it was originally detected - melanoma -, justifying the attributed label of ML-IAP (melanoma inhibitor of apoptosis proteins) (Vucic et al., 2000). BIRC7 protein levels have been found upregulated in primary tumors as well as in various melanoma cell lines, whereas it was only detectable in negligible amounts in melanocytes or nevi. Still, in melanoma, high levels of BIRC7 have been associated to chemoresistance and poor prognosis, while intermediate levels (but not necessarily the total lack of BIRC7 expression) were related to a better prognosis - possibly due to the pro-death activity of the truncated forms of the protein (Gong et al., 2005; Lazar et al., 2012).
Gastric, prostate, bladder, breast, renal and liver carcinomas, along with neuroblastoma, leukemia, and lymphomas, as well as non-small cell lung, cervical, liver and pancreatic cancers have all been related to BIRC7 upregulation (Gazzaniga et al., 2003; Tanabe et al., 2004; Hariu et al., 2005; Kim et al., 2005; Wagener et al., 2007; Augello et al., 2009; Yuan et al., 2009; Wang et al., 2010; El-Mesallamy et al., 2011; Lazar et al., 2012). In neuroblastoma, bladder and gastric cancers, higher levels of BIRC7 expression could be pondered as a risk factor, once isoform α, but not β, was predominant in bladder cancerous tissue but neither isoform occurred in the healthy tissue (Gazzaniga et al., 2003). In gastric cancer, nearly half of the assessed patients expressed both isoforms in their tumorous tissue, however benign gastric lesions showed no detectable BIRC7 expression (Wang et al., 2010).
Moreover, Ashhab and colleagues (2001) assessed mRNA transcripts of BIRC7α and BIRC7β in a panel of human tumor cell lines and upregulation of both isoforms was detected in melanoma, colon, and prostate carcinoma cells. Nevertheless, when measured in normal fetal and adult tissues, different expression levels for each isoform suggests a specific pattern of splicing and expression related to histology. Notable levels of BIRC7β were found mainly in fetal kidney, spleen, and heart, whereas no BIRC7α was detected in fetal tissues. Adult tissues, such as heart, placenta, lung, spleen and ovary showed upregulation of both isoforms, while only the α isoform was detected in brain, skeletal muscle and lymphocytes (Ashhab et al., 2001).

BIRC7 can be found both in the nucleus and in the cytoplasm, and the localization of this IAP has been compared to that of BIRC5 (Kasof & Gomes, 2001). Nevertheless, full-length forms of both BIRC7 variants are present exclusively in the cytoplasm, while a peri-nuclear distribution is seen for the pro-apoptotic truncated forms of BIRC7 (t-livin), with prominent accumulation in the Golgi apparatus (Nachmias et al., 2007a). In fact, subcellular localization has been known to regulate the different roles of IAPs in the apoptosis process (Ambrosini et al., 1998). In this context, Nachmias and colleagues (2007) have demonstrated that the N-terminal region of t-livin directs the peri-nuclear distribution and this arrangement is essential, but not enough, to give t-livin its pro-apoptotic activity. The RING domain, whose purposes have been afflicted with much ambiguity among distinct IAPs, then fulfills the operative role in the properly localized t-livins pro-apoptotic functions.
Point mutations to the RING domain resulted in proper peri-nuclear distribution and further Golgi localization, but abrogated the pro-apoptotic effect of t-livin. However, RING-mutated full-length BIRC7 was found in both nucleus and cytoplasm, suggesting that RING domain, opposite to what was observed for the truncated forms, may affect the sub-cellular localization of full-length BIRC7 (Nachmias et al., 2007a). Moreover, while the occurrence of intact full-length BIRC7 in the cytoplasm directly correlates with resistance to apoptosis, the presence of t-livin in the nucleus is associated with increased cellular apoptosis.

Most IAPs are thought to directly interact with caspases and block their activity (Kasof & Gomes, 2001), and the integrity of the BIR domain is necessary to achieve this effect. In fact, BIRC7 has been initially proposed to bind and inhibit CASP9, but not CASP1, CASP2, CASP3, CASP6, CASP7 or CASP8 (Vucic et al., 2000; Vucic et al., 2002). However, subsequent work carried out by the same group demonstrated that BIRC7 established a weaker interaction with caspases 3 and 9 when compared to that of XIAP (Vucic et al., 2005) (Figure 2). Moreover, three critical BIR residues were found responsible for BIRC7s lower caspase inhibition potency - Ser150, Gln167 and Glu168 - while XIAPs corresponding residues- Gly326, His343, and Leu344- rendered the latter IAP a stronger inhibitor. Once the critical residues were replaced for those of XIAP, the mutant BIRC7 actually showed greater inhibition towards caspase 9 than native XIAP (Vucic et al. 2005). Nevertheless, BIRC7 was demonstrated to bind with high affinity to the pivotal pro-apoptotic, IAP-antagonizer mitochondrial protein DIABLO (SMAC) through the BIR domain, while BIR-mutations were shown to abolish BIRC7-SMAC/DIABLO interaction (Ma et al., 2006). In this scenario, the anti-apoptotic function played by BIRC7 is essentially revealed through the sequestration of SMAC/DIABLO, thus abrogating the anti-IAP effect exerted by such protein (Chai et al., 2000; Liu et al., 2007). Furthermore, Ma and colleagues (2006) showed that BIRC7, in fact, resorts to the E3 ubiquitin ligase function of the RING domain to target SMAC/DIABLO for degradation through the ubiquitin-proteasome pathway.
BIRC7 has prompted apoptosis blockage induced by a number of death receptors, such as FAS, TNFRSF1A (TNFR1), TNFRSF10A (DR4) and TNFRSF10B (DR5) (Vucic et al., 2000), and has been associated to other proteins that are within the apoptotic pathway, thus inducing further indirect caspase inhibition. Similar to XIAP, the ability of BIRC7 to activate MAPK8 (JNK1), a protective pathway against apoptosis induced by TNF (TNF-&alpha); and interleukin, was verified by the MAP3K7 / TAB1 (TAK1/TAB1) signaling cascade (Sanna et al., 2002; Chen et al., 2010). Moreover, BIRC7 may play a role in the WNT/ CTNNB1 (Wnt/β-catenin) signaling pathway, a key component of gene activation with outcomes on tumor development through the activity of TCF (T-cell factor) transcription factors (Uematsu et al., 2003). Indeed, Yuan and collaborators (2007) confirmed BIRC7 to be a target of the β-catenin/TCF complex, suggesting their transcriptional regulation by upon BIRC7 (Yuan et al., 2007). Recent studies have also demonstrated a role for BIRC7 in regulating the epithelial-mesenchymal transition in colorectal cancer cells, favoring metastasis by the activation of the p38/ GSK3B pathway (Han et al., 2017).
BIRC7 overexpression has been also associated with tumor aggressiveness, chemoresistance and reduced sensitivity to radiation, while silencing of BIRC7 has been shown to lessen such features, both in vitro and in vivo. When SMMC-7721 cells were transfected with BIRC7 siRNA, mRNA and protein levels of both splice variants, BIRC7α and β, were greatly downregulated. Moreover, transfected cells displayed G1-arrest and a diminished S-phase cell count, reduced invasiveness, and re-established a response to apoptotic stimuli (Liu et al., 2010). BIRC7-silenced SCG-7901 cells, in turn, regained sensitivity to cytotoxic chemotherapy drugs, such as 5-fluorouracil (5-FU) and cisplatin (Wang et al., 2010). In xenograft models, HCT116 tumors treated with BIRC7 siRNA presented reduced volume in a dose-dependent fashion, while mice maintained healthy body weight and no signs of toxicity (Oh et al., 2011). Animal models were also used to demonstrate the differential effects of BIRC7 isoforms in tumorigenesis. BIRC7α was shown to promote tumor progression, whereas those expressing BIRC7β inhibited tumor growth due to high degrees of cleavage, mediated by natural killer (NK) cells activity, of this variant into t-livin, which has a pro-apoptotic effect. Nevertheless, the expression of a mutated BIRC7β with lowered inclination to undergo cleavage restored a positive effect of tumor growth (Abd-Elrahman et al., 2009). In turn, it was also established that, while BIRC7α enhanced killing by NK cells, the β variant took on a modest protective effect against apoptosis induced by NK cells, however, in Jurkat cells, this action occurred alongside a concurrent inhibitory trigger, and not self-sufficiently. Nevertheless, when both isoforms were detected in melanoma cells, a low killing rate was observed (Nachimas et al., 2007b).
The paradoxical pro-apoptotic effects of BIRC7, although baring intact BIR domains, are exerted by the truncated forms of both variants; still, t-livin β was found to give a stronger, however less stable, pro-apoptotic effect than t-livin α. t-Livin occurs around the cell nucleus, although not sturdily bound to that, and accumulates in the Golgi apparatus. Resorting to mutagenesis and co-localization studies, Nachmias and colleagues further demonstrated that an intact RING domain and merely the first N-terminal glycine (G53) residue were sufficiently responsible for t-livins localization to Golgi apparatus and also for its pro-apoptotic function, while deletion of either of these regions resulted in restoration of anti-apoptotic effect (Nachimas et al., 2007a). Additionally, t-livin has also been termed a flexible inducer of cell death once Shiloach and colleagues (2014) recognized its capacity of inciting necrosis, like in 293T cells, or apoptosis, like in A549 and MelA1, and such distinction may be possibly linked to TP53 status. Moreover, once the BIR domain was deleted from t-livin, the prior necrotic effect observed in 293T cells was replaced by an apoptotic effect, regardless of its inactive TP53. Both effects, apoptosis and necrosis, were linked to activation of JNK pathway. However, once the BIR domain was deleted from t-livin, 293T cells failed to express JNK, suggesting a role for BIR in activation of this pathway. In MelA1 cells, when these were treated with a pan-caspase inhibitor, t-livin-induced cell death was only partially abrogated, implying an aptitude of t-livin to induce cell death in situations where the apoptotic process is compromised (Shiloach et al., 2014).

BIRC7 has a high homology among different species (Table 1).
Table 1. Comparative identity of human BIRC7 with other species

% Identity for: Homo sapiens BIRC7	Symbol	Protein	DNA
vs. X. tropicalis	birc7	55.4	58.2
vs. D. rerio	birc7	60.3	62.7
vs. G. gallus	BIRC7	55.4	58.2
vs. D. rerio	birc7	59.1	65.0
vs. G. gallus	BIRC7	60.3	62.7
vs. X. tropicalis	birc7	59.1	65.0

(Source: http://www.ncbi.nlm.nih.gov/homologene)

Recurrent mutations in the BIRC7 gene are rare. Among the 48,467 unique samples reported in COSMIC (Catalogue of Somatic Mutations in Cancer; http://cancer.sanger.ac.uk/cancergenome/projects/cosmic), only 99 presented BIRC7 mutations (71 missense substitutions, 26 synonymous substitutions, 2 nonsense substitutions, and 1 frameshift insertion). Similar findings are reported in cBioPortal (http://www.cbioportal.org), in which, among the 10,967 cancer samples accessed, somatic mutations in BIRC7 were shown in merely 0.5% of the tested samples (corresponding to 59 mutations, of which 54 are missense substitutions, 4 truncated genes, and 1 other mutation). BIRC7 gene has been localized to the long arm of chromosome 20 on 20q13.3, a region frequently amplified in melanomas and other malignancies (Vucic et al., 2000; Ibrahim et al., 2014; Choi et al., 2019);