The CAP superfamily was named after it was recognized that considerable sequence similarity exists among the cysteine-rich secretory proteins (CRISPs), antigen 5, and pathogenesis-related proteins.1 CRISPs are highly enriched in the male mammalian reproductive tract and in the venomsecretory ducts of snakes, lizards, and other vertebrates.9,10 The highly immunogenic antigen 5 proteins are abundant in the venom-secretory ducts of stinging insects,11 and the pathogenesis-related 1 proteins are up-regulated in plants after invasion by pathogens.12 Speculation about the functional relationships among these CAP proteins has led to the idea that they may actually be isozymes with distinct substrate specificity13 and overlap between the plant and human immune systems.14

The GLIPR1 gene was initially identified as being up-regulated within glioblastoma multiforme and astrocytoma tissues and in glioma cell lines.15 Shortly after publication of that initial report, a subsequent study identified the same gene in glioma cell lines and called it "related to testes-specific, vespid, and pathogenesis protein 1 (RTVP-1)".16 We also identified Rtvp-1/Glipr1 in a differential-display polymerase chain reaction (PCR) screen as a p53 target gene.3 Identification of GLIPR1/Glipr1 was followed by identification and cloning of multiple isoforms of additional GLIPR1 subfamily genes. Mammalian GLIPR1 proteins are a multigene subfamily that consists of three genes in most species and four genes in the mouse.1 We previously identified two human GLIPR1-like genes (GLIPR1L1 and GLIPR1L2) and three mouse Glipr1-like genes (Glipr1l1, Glipr1l2, and Glipr1l3) as members of the GLIPR1 subfamily.1,17 An important note is that human GLIPR1, GLIPR1L1, and GLIPR1L2 genes are closely clustered on human chromosome 12q21 and mouse Glipr1, Glipr1l1, Glipr1l2, and Glipr1l3 genes, onmouse chromosome 10D1.17 In addition, we found that all three members of the human GLIPR1 cluster are direct p53 targets. It is also important that we further identified and characterized multiple alternative transcripts for GLIPR1L1 and GLIPR1L2.17 The presence of a putative signal peptide sequence and extracellular protein signature motifs suggests that most of the GLIPR1-cluster proteins are located on the surface of the cell membrane or secreted. Many, but not all GLIPR1-cluster proteins also contain a transmembrane domain, suggesting different capacities for secretion.17 These results indicate that important regulatory functions are encoded in the GLIPR1/Glipr1 subfamily proteins.

GLIPR1 subfamily gene expression is also highly tissue specific. For example, in somewhat limited studies, we showed that GLIPR1 expression is relatively widespread, whereas GLIPR1L1 expression is highly tissue specific, with very high mRNA levels in the testes but only trace amounts in the bladder and undetectable expression in the prostate, kidney, lung, and bone marrow. GLIPR1L2, like GLIPR1L1, is highly expressed in the testes, yet relatively low but detectable levels of expression were documented in the prostate, kidney, bladder, lung, and bone marrow.17

Initial studies of GLIPR1 expression provided clear evidence of the potential for tumor-suppressor activities. Quantitative reverse-transcriptase PCR and/or in situ hybridization analysis showed that GLIPR1 expression was lower in primary prostate cancer cells than in normal prostatic epithelium.4 In addition, prostate cancer cells that were metastatic to lymph nodes demonstrated much lower levels of expression than did normal prostate epithelium or prostate cancer. Although GLIPR1 mRNA was predominantly localized in prostatic epithelial cells (among which the basal cells exhibited the strongest signal level), some isolated stromal cells also showed moderate GLIPR1 mRNA levels.

Further, immunostaining analysis of normal prostate, primary prostate cancer, and metastatic prostate cancer samples showed that GLIPR1 protein expression is significantly lower in primary prostate cancer than in normal prostatic epithelium; GLIPR1 protein levels are still lower, or even undetectable, in lymph node metastases.4 The same study also showed that the human GLIPR1 promoter is extensively methylated in prostate cancer tissues relative to its methylation in normal prostate and that such increased methylation correlates with decreased levels of GLIPR1 expression. These data led to the proposal that GLIPR1 is a tumor suppressor that undergoes epigenetic inactivation in prostate cancer.4

Gene-transfer approaches were initially used to demonstrate activities consistent with tumor-suppressor functions of GLIPR1/Glipr1. P53-dependent and -independent proapoptotic activities were demonstrated as a result of GLIPR1/Glipr1 overexpression in multiple prostate cancer cells and various malignant cell lines.3 It was of interest that the proapoptotic effect was considerably less in nontransformed mouse embryo fibroblasts than in malignant cell lines. That same study also revealed that gamma irradiation and doxorubicin induced substantial levels of GLIPR1 mRNA in both the presence and absence of p53, which is suggestive of p53-independent GLIPR1 tumor-suppressor functions. Moreover, we found that deletion of the GLIPR1/Glipr1 signal peptide significantly reduced the proapoptotic effects of GLIPR1/Glipr1 in vitro, suggesting that secreted and/or release or cleavage from the membrane is important for its biologic functions.3

These initial studies were extended in subsequent studies that showed that GLIPR1 overexpression led to significant suppression of colony growth and induction of apoptosis in multiple cancer cell lines.5 To test our tumor-suppressor hypothesis in vivo, we generated mice with an inactivated Glipr1 gene; these Glipr1^-/- mice had significantly shorter tumor-free survival times than either Glipr1^+/+ or Glipr1^+/- mice did in both p53^+/+ and p53^+/- genetic backgrounds.5 An interesting finding was that a wide spectrum of tumors developed in the Glipr1^-/- mice, including lung carcinomas and plasma cytomas. It was also notable that the progressive loss of Glipr1 in the p53^+/- genetic background resulted in progressive reduction of p53 loss of heterozygosity. These data supported previous in vitro data3,4 and showed that Glipr1 has independent tumor-suppressor activities under these conditions.

As a member of the CAP family, GLIPR1 contains 11 cysteines that are somewhat concentrated at the carboxyl terminal of the molecule.17 Because cysteine residues within polypeptides can play important roles in redox homeostasis in mammalian cells,18,19 we hypothesized that GLIPR1 overexpression affects cellular ROS generation. The results of extensive analysis showed that GLIPR1 overexpression led to significantly increased ROS in various tumor cell lines, including prostate cancer cells.5 Additional studies showed that sustained JNK signaling resulted from GLIPR1-stimulated ROS production. Overall, increased ROS generation is required for GLIPR1-mediated activation of JNK and ultimately the induction of apoptosis in an inducible bladder-cancer cell model in vitro.5

These results provided mechanistic underpinning to the notion that GLIPR1 is a novel broad-spectrum tumor suppressor whose proapoptotic properties are exerted in part through ROS-JNK signaling.

As a secreted and/or membrane-bound proapoptotic tumor-suppressor protein, GLIPR1 (and potentially other GLIPR1 subfamily proteins) may have unique properties. GLIPR1 contains both an amino-terminal signal peptide and a transmembrane domain.17 In addition, the results of our previous studies showed that deletion of the GLIPR1/Glipr1 signal peptide significantly reduced the proapoptotic effects of GLIPR1/Glipr1 in vitro.3 Thus, it is likely that GLIPR1 is secreted and/or tethered onto the membranes of cells that express substantial levels of GLIPR1. Membrane-bound GLIPR1 may also undergo proteolytic cleavage, adding to the extracellular pool of GLIPR1. Although it is speculative, this biologic scenario would involve a pool of extracellular GLIPR1 with the potential for significant autocrine and/or paracrine activities.

To move beyond speculation, it will be necessary to directly test GLIPR1 protein under various conditions using relevant cell types that are present in the prostate cancer microenvironment. However, the results of previous studies are consistent with GLIPR1-mediated, multicell type-specific tumor-suppressor activities. We previously showed that AdGlipr1 treatment of orthotopic mouse prostate cancer resulted in reduced microvessel density and that AdGlipr1 also directly inhibited endothelial-cell sprouting in a rat aortic-ring sprouting assay.6 These data are consistent with antiangiogenic activities of secreted and/or membrane-bound GLIPR1/Glipr1 in vivo. In addition, the results of previous studies have shown that increased expression of Glipr1 is associated with macrophage differentiation.2 Both of these cell types are dominant, active components of the tumor microenvironment.20 In further support of the notion that secreted or cleaved GLIPR1/Glipr1, we have shown that a vaccine prepared with mouse prostate cancer cells, which were transduced with Glipr1 and irradiated, significantly reduced orthotopic prostate cancer "tumor take" and establishment of experimental prostate cancer lung metastases.7 Increased natural-killer cell and CTL activities were documented in those studies, suggesting direct systemic immunostimulatory activities. Overall, these data suggest that prostate tumor cell-derived secreted or cleaved GLIPR1/Glipr1 may both exert direct proapoptotic antitumor effects and suppress tumorigenesis and/or local tumor growth through antiangiogenic and immunostimulatory effects within the prostate cancer microenvironment. Further studies are required to confirm this hypothesis.

Elucidation of the unique tumor-suppressor properties of GLIPR1/Glipr1 led to the notion that GLIPR1 treatment may be effective for local and/or systemic control of prostate cancer and, potentially, other malignancies. Our preclinical studies, described above, showed that direct injection of AdGlipr1 into prostate cancer tissues using an immunocompetent orthotopic mouse model resulted in significant tumor growth suppression and longer survival of tumor-bearing animals.6 Analysis of AdGlipr1-treated tissues demonstrated increased prostate cancer cell apoptosis and significantly reduced tumor-associated angiogenesis. In addition, AdGlipr1-stimulated antitumor immune responses resulted in specific CTL activities in this model. Glipr1-related antitumor immunostimulatory activities were confirmed and extended in subsequent studies. Administration of a novel Glipr1 gene-modified tumor cell vaccine to mice had significant antitumor activity in a preclinical model of recurrent prostate cancer.7

On the basis of the results of our basic and preclinical studies, we completed a phase 1b clinical trial of in situ, adenoviral vector-mediated, neoadjuvant, pre-radical prostatectomy GLIPR1 gene therapy in patients with locally advanced adenocarcinoma of the prostate (IND13033). Preliminary analysis of the data revealed that intraprostatic administration of AdGLIPR1 was safe in men with localized high-risk prostate cancer before radical prostatectomy. In addition, preliminary evidence of biologic antitumor, systemic, and local activity was observed, suggesting a role for further development in the perioperative setting.8

Because the growth-arrest and proapoptotic effects of GLIPR1 are likely mediated to a large extent through autocrine and/or paracrine activities in prostate cancer, we have undertaken the development of GLIPR1 protein-based anticancer therapy. It is conceivable that GLIPR1 protein therapy would have important advantages compared with viral vector-based GLIPR1 delivery under specific conditions. For example, GLIPR1 protein may be delivered systemically and therefore be potentially effective against metastatic disease.