Expression of Survivin Transcript

To elucidate a potential role for survivin during embryonic SMG morphogenesis, we evaluated survivin transcript levels in Pseudoglandular (~ E14), Canalicular (~ E15) and Terminal Bud (~ E17) Stage SMGs using primer extension analysis. As shown in Figure 1, survivin transcripts were detected in all stages investigated. There was no significant change in survivin mRNA levels between the Pseudoglandular and Canalicular Stages. By contrast, there was a significant ~ 26% increase (p < 0.05) in survivin transcript levels between the Canalicular and Terminal Bud Stages, this significant increase being concomitant with extensive epithelial cell apoptosis which results in lumina formation [1].

Determination of Antibody Specificity

Prior to determining the cell-specific distribution of survivin protein during embryonic SMG development, the specificity of the anti-survivin antibody was evaluated by determining its reactivity on Western blots of lysates of 3 cancer cell lines (MO7E, HL-60, and HELA) since cancer cell lines have been shown to overexpress survivin protein [5, 6 ]. Using the unpurified anti-survivin antiserum, survivin protein (M_r ~ 16 kDa) was detected in all cell lines but not in normal whole human blood cell lysates (Fig. 2); survivin was not detected by the control serum. After affinity purification, a single ~ 16 kDa band was detected in HL-60 human cells (Fig. 3) using this antibody; no band was seen in the affinity column eluate of the preimmune serum. Our affinity-purified antibody was then evaluated for its ability to localize survivin protein in tissue sections. Our observation of survivin immunolocalization in pulmonary epithelia (data not shown) is identical to that previously reported [4]. Based on the above results, we conclude that this anti-survivin antibody is specific for survivin protein.

Spatiotemporal Distribution of Survivin

We then evaluated the pattern of survivin immunolocalization in embryonic SMGs with progressive development. In the Pseudoglandular Stage, survivin protein is diffusely distributed throughout the cytoplasm of branching epithelia and is absent from the mesenchyme (Fig. 4A). This cytoplasmic distribution of survivin protein suggests that survivin may not have translocated into the nucleus to form a survivin/Cdk4 complex at this early developmental stage. In the Canalicular Stage, epithelial cell apoptosis toward the center of presumptive ducts results in initial ductal lumen formation [1]. At the sites of lumen formation, intense survivin staining is seen in nuclei of the columnar epithelial cells bounding forming lumina (Fig. 4B). In contrast, those epithelial cells not surrounding the forming lumen do not display nuclear-localized survivin protein (Fig. 4B). As development progresses through the Terminal Bud Stage, terminal buds hollow out until only a single layer of epithelial cells surround terminal bud lumina (Fig. 4C) [1]. Nuclear-localized survivin protein is seen in these lumen-bounding terminal bud epithelial cells destined to survive (Fig. 4D), as well as in lumen-bounding ductal epithelia (Fig. 4E). Although nuclear translocation of survivin has been shown to be required for survivin-induced initiation of the cell cycle in cell lines in vitro [11], this is the first report of survivin translocation into the nuclei of embryonic cells in vivo. Cell proliferation, but not apoptosis, is seen in this single layer of terminal bud epithelial cells which is adjacent to mesenchyme and extracellular matrix [1]. This observation is consistent with the previous report that survivin is primarily found in proliferating cells [6, 8].

Although survivin has previously been demonstrated in embryonic lung epithelia [4], this is the first report of qualitative and quantitative changes in survivin expression in any developing branching organ. Further, it is interesting to note that substantial differences in survivin protein localization exist between embryonic lungs and SMGs. Survivin protein is nuclear-localized in late embryonic SMG epithelia surrounding forming lumina (Fig. 4) whereas it remains diffusely distributed, and not nuclear-localized, in pulmonary lumen-bounding epithelia (data not shown). The likely explanation for this notable difference in these analogous branching organs is the different mechanisms of their lumina formation. The pulmonary lumen forms as an extension of the tracheal lumen into the lung bud; repeated branching of the lung bud results in an extensive epithelial tree which surrounds the pulmonary lumina [13]. Since apoptosis is not a key factor in pulmonary lumina formation, a pro-survival/anti-apoptosis signal is not needed for epithelial cell survival and, thus, the survivin pathway is not activated. By contrast, the SMG bud develops as a solid cord of epithelium, with SMG lumina being formed by the death of central epithelial cells, and subsequently each consecutive concentric layer of terminal bud epithelial cells, until a single layer of "surviving" epithelial cells surround the lumina [2]. Thus, molecule(s) mediating an apoptotic stop signal are required for these lumen-bounding epithelial cells to survive. Interestingly, Bcl-2 and NFκB, other pro-survival/anti-apoptosis factors, are absent from these Terminal Bud stage bud epithelia [3, 14]. Given that survivin translocation into the nucleus is necessary for Cdk4/survivin complex formation, induction of cell cycle entry, and inhibition of caspase 3-mediated apoptosis [11, 12, 15], our observation of nuclear-localized survivin protein in proacinar (Fig. 4D) and ductal epithelia (Fig. 4E) suggests that survivin may be a key factor for the survival of epithelial cells surrounding forming ductal and terminal bud lumina.

Finally, it is interesting that extensive cell proliferation, but not apoptosis, is seen in Initial Bud and Pseudoglandular Stage SMG epithelia in the absence of survivin protein. To gain insight into which molecules likely protect these early embryonic SMG cells from apoptosis, we must turn to recent gene targeting experiments [16, 17] investigating FGF-10 and its receptor, FGF-R2(IIIb). Analysis of Fgf10 null and Fgfr2(IIIb) transgenic mice indicate that FGF10/FGF-R2(IIIb) signal transduction is essential for SMG branching morphogenesis [16, 17]; SMGs are absent in E14.5 and older mice. However the presence of an initial SMG epithelial bud which subsequently undergoes cell death in the absence of FGF-R2(IIIb) signaling [16], suggests that FGF10/FGF-R2-(IIIb) signaling mediates epithelial cell survival during the Pseudoglandular Stage. What remains unclear is which factor or factors protect the initial bud epithelium from apoptosis. Further studies are needed to address this question.

Tissue Collection

Female B10A/SnSg mice, obtained from Jackson Laboratories (Bar Harbor, ME), were maintained and mated as previously described [1]; plug day = day 0 of gestation. Pregnant females were anesthetized on days 14-18 of gestation (E14-18) with methoxyflurane (metafane) and euthanized by cervical dislocation. This research conformed to the Animal Resource Institutional Review and Ethics Committee. Embryos were dissected in cold phosphate buffered saline (PBS) and staged according to Theiler [18]. E14-E18 SMGs and E15-E17 lungs were collected and stored at -70°C or processed for histology.

Analysis of RNA by Primer Extension

Total RNA from E14 Pseudoglandular SMGs (64 glands/3 litters); E15 Canalicular SMGs (52 glands/3 litters) and E17 Terminal Bud SMGs (48 glands/3 litters) were extracted as previously described [3]. Primer extension was conducted essentially as described in Li and Altieri [10]. A msg-PE oligonucleotide 5'-GGCTCTTTCTCTGCT CAGTTT-3' was synthesized, gel purified and 5' end-labeled; a control primer was also radiolabled. For primer extension, 40 μg total RNA per stage was hybridized with an excess of 5'-radiolabeled DNA primers. The resulting end-labeled cDNA was separated on a urea-denaturing polyacrylamide gel, visualized, and quantitated by phosphor image analysis with ImageQuant software. Relative levels of survivin mRNA were tabulated as the percent transcript level relative to that in the Pseudoglandular Stage. Subsequent analysis was by t-test for the significance of differences between percentages based on the arcsin transformation; it employs a constant representing the parametric variance of a distribution of arcsin transformations of percentages [19].

Immunohistochemistry

A rabbit polyclonal antibody was raised against a 19 aa peptide [NH₂-MGAPTLPPAW OPFLKDHRI-COOH] representing amino acids 1-19 of the deduced human amino acid sequence [6]. Using anti-survivin antiserum and Western blot analysis, survivin protein (M_r ~ 16 kDa) was detected in the lysates of 3 cancer cell lines (MO7E, HL-60, and HELA), but not in normal whole human blood cell lysates; survivin was not detected by the control serum. The antibody was purified from the immune sera by affinity chromatography on a peptide column and specificity determined by Western blot analysis [3]. Extracts of the HL-60 were analyzed to demonstrate antibody specificity. No band was seen in control affinity column eluate of the prebleed serum. For immunolocalization studies, embryonic SMGs and lungs were fixed in Carnoy's fixative, processed, and immunohistochemistry conducted as previously described [3]. Immunohistochemical staining was performed using the Elite ABC kit (Vector Laboratories, Burlingham, CA).