In the search for serum markers of hepatocellular cancer, several investigators have recently focused on Golgi protein 73 [GP73; also known as Golgi membrane protein 1 (Golm1) or Golph2; Gen ID 51280]. GP73 is a 400–amino acid, 73-kD transmembrane glycoprotein that normally resides within the cis-Golgi complex. Its messenger RNA was first identified in a search for up-regulated hepatic genes in a patient with syncytial giant-cell hepatitis.1 Subsequent studies revealed minimal GP73 expression in normal hepatocytes but marked expression in patients with acute and chronic hepatitis and liver cirrhosis2, 3 (Fig. 1A). This up-regulation occurred in response to viral and nonviral etiologies and thus appeared to be a general feature of liver disease. In addition to hepatocytes, GP73 was consistently expressed by normal biliary epithelial cells as well as hepatic stellate cells in injured livers.3 Further studies demonstrated constitutive expression in cells of the epithelial lineage, especially in the prostate, gut, breast, and thyroid, and within the central nervous system.1

thumbnail image

Figure 1. (A) GP73 expression levels in hepatocytes in the course of acute and chronic liver disease. Normal hepatocyte GP73 expression is minimal. Marked and reversible increases in the percentage of GP73-positive hepatocytes and their cellular GP73 levels occur in acute hepatitis. Progressive up-regulation is also observed in chronic liver disease and can be reversed by treatment of the underlying disease etiology (i.e., steroid treatment of autoimmune hepatitis) and resolution of chronic hepatitis.2, 3 GP73 expression in hepatocellular cancer cells is further increased (2- to 3-fold) in comparison with patients with cirrhosis.13 (B) Structural features of GP73 and sGP73. GP73 consists of a short cytoplasmic N-terminus with a myristoylation domain followed by a single TM and a large C-terminal ectodomain.1 The N-terminus contains two coiled-coil domains and is N-glycosylated. The C-terminus is highly acidic and of unknown function. Furin cleavage after aa 55 results in the extracellular release of sGP73.8 The epitopes for the capture and recognition antibodies used in the study by Riener et al.13 are shown. Abbreviations: aa, amino acid; ELISA, enzyme-linked immunosorbent assay; GP73, Golgi protein 73; HCC, hepatocellular carcinoma; sGP73, serum form of Golgi protein 73; TM, transmembrane domain.

Download figure to PowerPoint

Given its intracellular localization as a resident Golgi protein, its widespread expression in extrahepatic tissues, and its up-regulation in nonmalignant liver diseases, it came as a surprise when two reports identified a serum form of GP73 (sGP73) as a marker of hepatocellular cancer. Block and colleagues4 first demonstrated GP73 immunoreactivity in serum and found increased sGP73 levels in patients with hepatitis B virus–related hepatocellular cancer. In a subsequent study by Marrero et al.,5 sGP73 levels were significantly increased in patients with hepatitis C virus (HCV)–related hepatocellular carcinoma (HCC) in comparison with cirrhotic controls. The sensitivity and specificity of sGP73 for HCC were superior to those of alpha-fetoprotein, especially in early HCC. Similar results were reported in a Chinese study on patients with predominantly hepatitis B virus–related liver cancer.6 In response to these encouraging reports, GP73 was added to a group of emerging candidate HCC serum markers.7

The mechanism by which sGP73 reaches the circulation was worked out by Bachert and colleagues8 in cell culture studies. Despite its steady-state localization within the cis-Golgi complex, GP73 cycles through the distal secretary apparatus and transiently reaches the apical cell membrane, from which it returns to the Golgi complex via an endosomal retrieval pathway.9 A slightly smaller form of GP73 is present in the supernatants of several cell lines. This secreted form is generated by N-terminal cleavage of the molecule by the proprotein convertase furin after amino acid 55, resulting in the release of the large C-terminal ectodomain into the extracellular space8 (Fig. 1B). Using N-terminal sequencing, Gu and colleagues10 confirmed that the serum form of sGP73 is identical to the furin cleavage product identified in cell culture supernatants. This finding provides a mechanistic explanation for the appearance of sGP73 in serum.

One drawback of the initial studies was that sGP73 was detected by western blotting, a semiquantitative and laborious test. sGP73 sandwich enzyme-linked immunosorbent assays (ELISAs) were first developed by Gary Norman at Inova Diagnostics and by Gu and colleagues.10 The assays were tested in two large patient cohorts from Italy and China, respectively. The results were disappointing. In the first study, sGP73 was found to be elevated in patients with liver disease but did not distinguish between HCC, cirrhosis, and chronic hepatitis.10 In the second study, which was reported in an abstract form, sGP73 was surprisingly found to be decreased in HCC patients.11

One potential methodological issue in both studies was the use of antibodies directed against bacterially expressed GP73. Given the extensive glycosylation and hyperfucosylation of GP73 in eukaryotic cells,12 these antibodies might not recognize the fully modified sGP73. Furthermore, recently described GP73-specific serum autoantibodies might interfere with the ELISA analysis.11

In this issue of HEPATOLOGY, Riener and colleagues13 report the development of a novel sGP73 ELISA that was validated with fully humanized GP73. sGP73 was measured in patients with HCC, chronic liver diseases without cancer, and bile duct cancer and in healthy controls. sGP73 serum levels in normal subjects ranged from 1.5 to 20 μg/mL. Significant (3- to 5-fold) increases over control levels were seen in HCC patients with underlying HCV infection (especially HCV genotype 1b) but not in patients with HCC unrelated to HCV. The overall difference between cirrhotic subjects with and without HCC was not statistically significant. The authors suggest that sGP73 might not be suitable as a general serum marker of HCC but might have utility as a marker of HCV-related HCC. The results of this study cast doubt on the diagnostic utility of sG73 as a serum marker of HCC.

However, a few methodological questions will need to be addressed before the authors' interpretation is endorsed.

First, the sGP73 serum levels reported by Riener et al.13 were approximately 80-fold higher than those reported by Gu et al.,10 with a median serum concentration in normal subjects of 4 μg/mL, which is well within the range of many classical plasma proteins.14 This raises questions regarding the specificity of the ELISA, especially because other investigators have previously noted the presence of smaller GP73 antibody–reactive bands in serum western blots (A. Mehta, personal communication, 2009). This issue could best be resolved by side-by-side comparisons of western blotting and ELISA measurements in the same samples.

Second, the authors' conclusion was based on a relatively small number of total and HCV-related HCC cases (29 of 62). A more definitive assessment of its performance characteristics will require measurements in larger and better defined patient cohorts, including the reference serum bank that was recently established by the Early Disease Recognition Network.

The second major contribution of Riener et al.'s article13 is the systematic analysis of GP73 tissue expression in HCC. Using tissue microarrays, the authors clearly established that GP73 was overexpressed in malignant liver tumors in comparison with surrounding noncancerous tissues. The degree of GP73 expression correlated with the tumor grade. In contrast, expression levels in benign liver lesions—focal nodular hypertrophy and hepatic adenoma—were not significantly different from those of the surrounding areas. These findings provide evidence that the increased sGP73 in HCC patients originates from cancerous hepatocytes, an important requirement for the validation of tumor biomarkers.7

Another novel finding is the marked up-regulation of GP73 expression in cancers of biliary origin: GP73 was overexpressed in 97 of 114 (85.1%) intrahepatic and extrahepatic bile duct and gallbladder carcinomas. High-grade GP73 expression was correlated with improved survival; this is an intriguing finding that remains to be further explored. An analysis of sGP73 in 10 patients with biliary cancers revealed a substantial increase. This preliminary finding is particularly interesting because there is a lack of clinically useful serum markers for this type of cancer.

Riener et al.'s data13 add to the growing number of studies indicating dysregulation of GP73 in different human cancers.15 Along these lines, three studies have recently demonstrated increased GP73 expression in prostate cancer cells.16-18 GP73 protein and messenger RNA were present in the patients' urine, and urine GP73 was diagnostically superior to serum prostate-specific antigen in prostate cancer.17 Interestingly, the urine form of GP73 represents the full-length protein, and this suggests a furin-independent release mechanism.

In light of the present study, it seems doubtful that sGP73 will ultimately be sufficiently HCC-specific or even liver tumor–specific. Given its expression in breast, bronchial, thyroid, and pancreatic epithelial cells, studies measuring GP73 and sGP73 in patients with corresponding cancers in these organs should be forthcoming soon. On the other hand, recent reports indicate that GP73 expression was not increased in renal cell cancer,19 and sGP73 levels were not increased in patients with colorectal cancer4; this suggests that sGP73 is not likely to become a pan-cancer marker.

One of the remaining challenges will be to unravel the biological role and biochemical function of GP73. GP73 might have structural roles in the Golgi complex, or it might be involved in the trafficking and modification of its secretory cargo. Unfortunately, its sequence reveals no unique functional domains. Its luminal coiled-coil domain is necessary and sufficient for its retrieval to the Golgi complex, but the function of its highly acidic C-terminus is unknown.8 Furthermore, it is unknown whether the furin-cleaved ectodomain has specific biological functions. The development of GP73-knockout mice might be helpful for answering these questions. However, as seen in the case of alpha-fetoprotein, even knockout studies may not easily yield a convincing phenotype. Future studies of GP73 may reveal multiple roles in different tissues, as highlighted by its recent identification as a susceptibility allele for Alzheimer disease.20 Ultimately, it may be harder to assign biological function to GP73 than to establish its diagnostic utility.


  1. Top of page
  2. References
  • 1
    Kladney RD, Bulla GA, Guo L, Mason AL, Tollefson AE, Simon DJ, et al. GP73, a novel Golgi-localized protein upregulated by viral infection. Gene 2000; 249: 5365.
  • 2
    Kladney RD, Cui X, Bulla GA, Brunt EM, Fimmel CJ. Expression of GP73, a resident Golgi membrane protein, in viral and nonviral liver disease. HEPATOLOGY 2002; 35: 14311440.
  • 3
    Iftikhar R, Kladney RD, Havlioglu N, Schmitt-Graff A, Gusmirovic I, Solomon H, et al. Disease- and cell-specific expression of GP73 in human liver disease. Am J Gastroenterol 2004; 99: 10871095.
    Direct Link:
  • 4
    Block TM, Comunale MA, Lowman M, Steel LF, Romano PR, Fimmel C, et al. Use of targeted glycoproteomics to identify serum glycoproteins that correlate with liver cancer in woodchucks and humans. Proc Natl Acad Sci U S A 2005; 102: 779784.
  • 5
    Marrero JA, Romano PR, Nikolaeva O, Steel L, Mehta A, Fimmel CJ, et al. GP73, a resident Golgi glycoprotein, is a novel serum marker for hepatocellular carcinoma. J Hepatol 2005; 43: 10071012.
  • 6
    Mao YL, Yang HY, Xu HF, Sang XT, Lu X, Yang ZY, et al. Significance of Golgi glycoprotein 73, a new tumor marker in diagnosis of hepatocellular carcinoma: a primary study [in Chinese]. Zhonghua Yi Xue Za Zhi 2008; 88: 948951.
  • 7
    Block TM, Marrero J, Gish RG, Sherman M, London WT, Srivastava S, et al. The degree of readiness of selected biomarkers for the early detection of hepatocellular carcinoma: notes from a recent workshop. Cancer Biomark 2008; 4: 1933.
  • 8
    Bachert C, Fimmel C, Linstedt AD. Endosomal trafficking and proprotein convertase cleavage of cis Golgi protein GP73 produces marker for hepatocellular carcinoma. Traffic 2007; 8: 14151423.
  • 9
    Puri S, Bachert C, Fimmel CJ, Linstedt AD. Cycling of early Golgi proteins via the cell surface and endosomes upon lumenal pH disruption. Traffic 2002; 3: 641653.
  • 10
    Gu Y, Chen W, Zhao Y, Chen L, Peng T. Quantitative analysis of elevated serum Golgi protein-73 expression in patients with liver diseases. Ann Clin Biochem 2009; 46: 3843.
  • 11
    Sangiovanni A, Romeo R, Prati GM, Manini MA, Iavarone M, Ronchi G, et al. Serum expression of lectin reactive alpha-fetoprotein, des-gamma-carboxy prothrombin, golgi protein-73 antigen and antibody, for the diagnosis of hepatocellular carcinoma [Abstract]. HEPATOLOGY 2007; 46(Suppl): 406A.
  • 12
    Norton PA, Comunale MA, Krakover J, Rodemich L, Pirog N, D'Amelio A, et al. N-linked glycosylation of the liver cancer biomarker GP73. J Cell Biochem 2008; 104: 136149.
  • 13
    Riener M-O, Stenner F, Liewen H, Soll C, Breitenstein S, Pestalozzi BC, et al. HEPATOLOGY 2009; 49: 16021609.
  • 14
    Anderson NL, Anderson NG. The human plasma proteome: history, character, and diagnostic prospects. Mol Cell Proteomics 2002; 1: 845867.
  • 15
    Lu Y, Yi Y, Liu P, Wen W, James M, Wang D, et al. Common human cancer genes discovered by integrated gene-expression analysis. PLoS ONE 2007; 2: e1149.
  • 16
    Wei S, Dunn TA, Isaacs WB, De Marzo AM, Luo J. GOLPH2 and MYO6: putative prostate cancer markers localized to the Golgi apparatus. Prostate 2008; 68: 13871395.
  • 17
    Varambally S, Laxman B, Mehra R, Cao Q, Dhanasekaran SM, Tomlins SA, et al. Golgi protein GOLM1 is a tissue and urine biomarker of prostate cancer. Neoplasia 2008; 10: 12851294.
  • 18
    Kristiansen G, Fritzsche FR, Wassermann K, Jager C, Tolle A, Lein M, et al. GOLPH2 protein expression as a novel tissue biomarker for prostate cancer: implications for tissue-based diagnostics. Br J Cancer 2008; 99: 939948.
  • 19
    Fritzsche FR, Riener MO, Dietel M, Moch H, Jung K, Kristiansen G. GOLPH2 expression in renal cell cancer. BMC Urol 2008; 8: 15.
  • 20
    Li H, Wetten S, Li L, St Jean PL, Upmanyu R, Surh L, et al. Candidate single-nucleotide polymorphisms from a genomewide association study of Alzheimer disease. Arch Neurol 2008; 65: 4553.