Kuniaki Nakanishi, Department of Pathology and Laboratory Medicine, National Defense Medical College, Tokorozawa 359-8513, Japan. e-mail: firstname.lastname@example.org
To examine glucose-regulated protein 78 (GRP78; a major molecular chaperone at the endoplasmic reticulum, strongly expressed in several tumours) expression in urothelial carcinoma (UC) of the upper urinary tract (UUT) and to evaluate the diagnostic and progressive importance of GRP78 expression in UC-UUT.
PATIENTS AND METHODS
We investigated GRP78 expression (using immunohistochemistry) in 126 UC-UUTs to assess its relevance to progression. GRP78 overexpression was recognised in 23 (18.3%) of tumour samples.
There was no association between GRP78 overexpression and clinicopathological findings, except for an association with low grade in invasive tumours. GRP78 overexpression significantly improved the disease-free survival rate in all patients (according to univariate and multivariate analyses), but did not alter the overall survival rate.
The detection of GRP78 overexpression would appear to provide valuable information for the prognosis of UC-UUT.
Glucose-regulated protein 78 (GRP78), which serves as a major molecular chaperone at the endoplasmic reticulum (ER), is involved in the folding and assembly of newly synthesized proteins within the ER [1–3]. In normal conditions, GRP78 expression is maintained at a low basal level in many organs (such as the brain, lung, and heart). However, GRP78 overexpression can be induced by physiological stress that perturbs ER function and homeostasis, and such overexpression protects against tissue or organ damage under pathological conditions (such as neurotoxic stress, myocardial infarction, and arteriosclerosis) [3,4]. GRP78 has been found to be strongly induced in various carcinomas, and such overexpression correlates with higher pathological grade and aggressive progression [5–10]. However, it has been reported that such overexpression may lead to a more favourable prognosis in lung cancer and neuroblastoma [11,12]. Thus, the role of GRP78 expression in progression is poorly understood, and little is known about its expression in urothelial carcinoma (UC) of the upper urinary tract (UUT). In the present study, we used immunohistochemistry to examine GRP78 expression in 126 cases of UC-UUT. Our goal was to evaluate the diagnostic and progressive importance of GRP78 expression in UC-UUT.
PATIENTS AND METHODS
In all, 126 surgically resected specimens from patients with primary UC-UUT were examined. These specimens had been obtained at the Mutual Aid Associations’ Hospital, Tachikawa, and National Defense Medical College Hospital, Tokorozawa between 1970 and 1995. Histopathological stage was determined according to the criteria proposed by the International Union Against Cancer . Tumour cells were divided histopathologically into two grades using the criteria for urinary bladder tumours laid down by the Armed Forces Institute of Pathology .
The median (range) patients’ age at diagnosis was 67 (34–84) years. In all, 32 of the 126 patients died because of their tumours at a mean (median; range) of 24 (15; 1–132) months after surgery. The remainder survived for a mean (median; range) of 69 (64; 0–257) months after surgery. Among the 126 cases, 52 tumours (41.3%) were in the renal pelvis or calyces, 50 (39.7%) were in the ureter, and 24 (19.0%) were multicentric. Ten patients had simultaneous bladder tumours at the time of diagnosis, 26 had subsequent bladder tumours, and seven had an antecedent bladder tumour. In all, 42 (33.3%) patients had an associated bladder neoplasm. The initial management of the 119 patients who were not treated as having bladder cancer included complete nephroureterectomy with a bladder cuff (88 patients), nephroureterectomy without a bladder cuff (eight), nephroureterectomy with total cystectomy (nine), ureterectomy (two), and nephrectomy (12). Thirty-three patients received either adjuvant chemotherapy (22 patients), radiotherapy (seven), or both (four), in addition to surgery.
The tumours were divided into three groups (A, B, and C) based on tumour stage. There were 47 cases (37.3%) in group A of papillary, non-invasive tumours (pTa); 20 cases (15.9%) in group B of tumours invading the submucosa or muscularis (pT1 and pT2); and 59 cases (46.8%) in group C of tumours invading beyond the muscularis or renal parenchyma, or metastasizing the regional lymph node or a distant site (pT3 and pT4).
Ten of the 47 patients with a papillary, non-invasive tumour (pTa) had either recurrence (six patients), metastasis (one) or both (three), but 37 developed neither recurrence nor metastasis within the follow-up period. The mean (median; range) of follow-up for patients with a non-invasive tumour who had a recurrence and/or metastasis was 72 (60; 14–124) months, whereas the mean (median; range) of follow-up for those without recurrence or metastasis was 79 (86; 4–203) months. Of the 79 patients with an invasive tumour (pT1, pT2, pT3, or pT4), five had metastasis at the time of surgery or the tumour could not be excised totally by surgery, 38 either had recurrence (nine patients), metastasis (17) or both (12), but the remaining 36 developed neither recurrence nor metastasis within the follow-up period. The mean (median; range) of follow-up for patients with an invasive tumour who had recurrence and/or metastasis was 37 (15; 1–175) months, whereas the mean (median; range) of follow-up for those without recurrence or metastasis was 59 (35; 0–257) months.
At the time of diagnosis, tumours were low grade in 76 patients (60.3%) and high grade in 50 (39.7%). On inspection of the pattern of growth of the tumours there were 89 cases (70.6%) with a papillary pattern and 37 (29.4%) with a non-papillary pattern. Of the 31 patients who died from their tumours, 12 (38.7%) had a tumour in the renal pelvis or calyces, 14 (45.2%) in the ureter, and five (16.1%) had multicentric tumours. Of these 31 patients, two were in group A, three in group B, and 26 in group C.
For the immunohistochemistry we used the polymer-peroxidase method (EnVision+/ HRP; Dako Cytomation, Denmark) on deparaffinized sections, using a rabbit polyclonal antibody against GRP78 (ready-to-use; Thermo Fisher Scientific Anatomical Pathology, Fremont, CA, USA). For the negative control, the incubation step with the primary antibody was omitted. For the analysis of immunoreactivity, the intensity and extent of staining were scored from 0–3 and 0–4, respectively, with 0 representing no staining. The intensity was scored as 1 indicating weak, 2 indicating moderate, and 3 indicating strong staining. The extent of staining was scored as 1, indicating ≤10% of tumour area stained; 2, 11–25% stained; 3, 26–50% stained; or 4, ≥51% stained. Those tumours in which the staining intensity was scored as >1 and the staining extent >3 were graded as positive. The evaluation was performed twice by investigator K.U. and twice by investigator K.N., with each of them being ‘blinded’ to both tumour stage and grade.
In addition, proliferating cell nuclear antigen (PCNA) was evaluated immunohistochemically, with the technique used and the results obtained in these same patients having been reported elsewhere . For the analysis of PCNA, the percentage of nuclei exhibiting a positive immunoreaction (PCNA index) was determined (based on the immunoreaction in ≥1000 tumour cells). The PCNA index was classified as high if it was ≥56%, a figure representing the median value for the carcinomas. Furthermore, we measured microvessel density (MVD) in 79 invasive carcinomas by immunohistochemistry, using a rabbit polyclonal antibody against factor VIII-related antigen (1:1500; Dakopatts, Glostrup, Denmark). MVD was assessed by light microscopy in those areas of tumour invasion containing the highest numbers of capillaries and small venules per unit area (neovascular ‘hot spots’), as previously described . To this end, the areas of highest neovascularization were found by scanning the tumour sections at low magnification (×40) to identify those areas of tumour invasion having the greatest numbers of immunostained microvessels per unit area. After the areas of highest neovascularization had been identified, the number of microvessels was determined within individual ×200 fields (×20 objective and ×10 ocular; 0.739 mm2 per field). Any immunostained endothelial cell or endothelial-cell cluster clearly separated from adjacent microvessels, tumour cells, and other connective-tissue elements was considered to indicate a single, countable microvessel. Results were expressed for each tumour as the highest number of microvessels identified within any single ×200 field. The MVD was classified as high if it was ≥70%, a figure representing the median value for the carcinomas.
For statistical analysis, disease-free survival (DFS) and overall survival (OS) rates were the two main dependent variables tested in this study. ‘DFS’ was defined as the period between the initial radical operation and the subsequent appearance of recurrence or metastasis. Recurrence was defined as UC occurring anywhere in the genitourinary tract. The end-point was either recurrence/metastasis of UC or the closing date of the study, whichever came first. ‘OS’ was defined as the interval between surgery and death; the end-point for this variable was either death or the closing date of the study.
DFS and OS curves for all of the univariate analyses were assessed using the Kaplan–Meier method. Comparisons between two or more survival curves were assessed using Wilcoxon and log-rank tests. For multivariate analysis of the clinicopathological variables, the Cox stepwise-regression model was used. For comparisons of age, sex, stage, grade, and pattern of growth the Mann–Whitney U-test was used.
In normal urothelial cells, GRP78 was expressed very weakly within the cytoplasm. In tumours, a positive expression of GRP78 was recognized in 23 (18.3%) of tumour samples. GRP78 was mainly expressed within the cytoplasm of tumour cells, although it was sometimes localized to the plasma membrane of these cells (Fig. 1). However, there was no association between GRP78 overexpression and clinicopathological findings (except an association with grade in invasive tumours; P= 0.04), PNCA index, or MVD (Table 1).
Table 1. Relationship† between GRP78 immunoreactivity and other tumour characteristics (clinicopathological findings, PCNA index, and MVD) in 126 cases
All tumours (n= 126)
Non-invasive tumours (n= 47)
Invasive tumours (n= 79)
No of cases
IR for GRP78
No of cases
IR for GRP78
No of cases
IR for GRP78
IR, immunoreactivity; chi-square analysis used for all comparisons except stage;
Mann–Whitney test was used for tumour stage comparisons; the tumours were divided into three groups on the basis of tumour stage (A, papillary, non-invasive tumours, pTa; B, tumours invading the submucosa or muscularis, pT1 and pT2; and C, tumours invading beyond the muscularis or renal parenchyma, or metastasizing the regional lymph node or a distant site, pT3 and pT4).
The 5-year DFS and 5-year OS were 63.4% and 72.1%, respectively. Analysis of the DFS and OS curves (Fig. 2) showed that GRP78 overexpression significantly improved DFS in all patients and in patients with invasive tumours, but did not significantly alter OS in either group. In the assessment of DFS, 121 patients who had no metastasis at surgery and in whom the malignant tumour was excised totally were included in the analysis. In the assessment of OS, all 126 patients were included in the analysis. Our univariate analyses of DFS and OS showed that GRP78 overexpression had a significant effect on the DFS rate in all tumours (Table 2). Furthermore, stage, grade, and pattern of growth all had a significant effect on each of the two survival rates (except pattern of growth in the log-rank test of DFS; P= 0.14). In the final models of the multivariate analysis for all tumours, GRP78 overexpression and stage were shown to be prognostic factors for DFS (P= 0.025, P < 0.001, respectively). In the final models of the multivariate analysis for all tumours, only stage was a prognostic factor for OS (P < 0.001). In invasive tumours, GRP78 expression was a prognostic factor for DFS, while stage was a prognostic factor for OS (Table 2).
Table 2. Univariate analysis of DFS and OS rates
DFS (n= 121)
OS (n= 126)
DFS (n= 47)
OS (n= 47)
DFS (n= 74)
LR, Log-rank test; W, Wilcoxon test; P value was not determined because of one factor or because that particular analysis was not done.
In UC, differences in tumour anatomy lead to differences in survival. Although prognostic significance has been established for both stage and grade [16,17], it is important to identify prognostic markers that will predict which patients are likely to have disease progression. The purpose of the present investigation was to look for possible relationships between GRP78 expression and clinicopathological findings and/or clinical outcome in UC-UUT. In the present analysis there was no relationship between GRP78 expression and clinicopathological findings, PCNA index, or microvascular density (except that the incidence of GRP78 overexpression was significantly higher in low-grade than in high-grade invasive tumours). However, there was a significant correlation between GRP78 expression and DFS rate in both the univariate and multivariate analyses of all tumours, and also in the univariate analysis of invasive tumours. Thus, detection of GRP78 overexpression would appear to be of value in informing the prognosis in UC-UUT, such overexpression being associated with a favourable prognosis.
In the present study, GRP78 was expressed within the cytoplasm of normal urothelial cells and most tumour cells, and in addition, it was sometimes localized to the plasma membrane of tumour cells. Generally, GRPs are localized to the ER, and their induction has been widely used as a marker for ER-stress and the onset of the unfolded protein response [1–3]. In two in vitro studies in which cells were treated with ER-stress inducers, GRP78 was shown to be present on the cell surface and, as a secretion, in the extracellular space, as well as within the cytoplasm [18,19]. The present results are consistent with this data.
It is known that GRP78 inhibits apoptotic signalling and protects the cell from the apoptosis induced by ER-stress. In fact, in a study on mice in which GRP78 induction was suppressed by an antisense cDNA in fibrosarcoma cells, there was an increase in apoptotic cell death and an inhibition of tumour formation . This suggests an important role for GPR78 overexpression in tumour-cell proliferation, and indeed its overexpression has been reported to be associated with both tumour aggressiveness and an unfavourable prognosis in various tumours, including carcinomas of the breast, stomach, liver, colon, and prostate [4–11]. However, other evidence indicates a protective role for GRP78. For instance, induction of GRP78 is closely associated with cell-cycle arrest in the G1 phase , an effect that could serve as a cellular defence mechanism, as indeed has been shown in the case of the DNA-damage response [22,23]. Moreover, Cai et al. who used a human epidermoid carcinoma cell-line under a chemical-stress condition to examine the relationship between GRP78 and growth arrest, showed that induction of GRP78 induced a down-regulation of the epidermal growth factor-signalling pathway through the formation of a stable GRP78-epidermal growth factor receptor complex, and subsequently led to growth arrest among the cancer cells. In the present study, the incidence of GRP78-positivity tended to be higher among cases with a low PCNA index than among those with a high PCNA index, although the P value indicated only borderline significance (P= 0.052). Furthermore, GRP78 induction was associated with a favourable prognosis, as also reported for lung cancer and neuroblastoma [11,12]. Thus, some evidence suggests that GRP78 overexpression favours the growth of at least some tumours, while other evidence suggests a predominantly protective role.
With regard to the relationship between GRP78 expression and histological grade, there was a significantly higher incidence of GRP78 expression in low-grade invasive tumours than in high-grade, although we failed to show a significant difference between the low- and high-grades in all tumours. In a previous report on neuroblastic tumours, in which GRP78 induction was associated with a favourable prognosis , most tumours exhibiting a positive-GRP78 expression displayed a differentiated histology. Collectively, the above data suggest that in these tumours, a good prognosis may be favoured by the presence of a low-grade or differentiated tumour exhibiting GRP78 expression. Concerning, low-grade tumours exhibiting GRP78 expression (as in the present data), the above idea may be supported by the tendency for a low PCNA index to be associated with GRP78 expression (see previous paragraph).
With regard to GRP78 expression and angiogenesis, Dong et al., who examined a genetic model of breast cancer in Grp78 heterozygous mice, found that Grp78 heterozygosity prolonged the latent period and significantly impeded tumour growth. They further showed that GRP78 acted via three major mechanisms: (i) enhancement of tumour-cell proliferation, (ii) protection against apoptosis, and (iii) promotion of tumour angiogenesis . Their study also showed that a partial reduction of GRP78 expression in the Grp78 heterozygous mice substantially reduced the tumour MVD. They therefore suggested that GRP78 is probably required for the maintenance of the tumour vasculature. However, in a previous report on non-small cell lung cancers, tumours with a low MVD exhibited a higher expression of GRP78 . In contrast to both of those studies, the present study did not show an association between GRP78 expression and MVD. Moreover, although Dong et al. reported a protective effect of GRP78 against apoptosis, an induction of GRP78 mRNA in association with apoptosis has been reported for nerve growth factor-deprived sympathetic ganglion cells in primary culture . Although the explanations for the discrepancies among the above findings remain unclear, one possibility (among many) is that the effects of GRP78 may vary in direction and/or magnitude among tumours.
In conclusion, we have shown that evaluation of GRP78 expression provides information about the prognosis in cases of UC-UUT, with its induction being associated with a favourable prognosis. However, the unfolded protein response triggers multiple pathways that allow cells to respond to ER-stress [1–3], and the expression of Grp78 has been reported to be regulated by transcription factors such as activating transcription factor-6 and X-box binding protein 1 [28,29]. Therefore, future studies of UC-UUT (preferably of more cases) should examine other factors associated with unfolded protein response, and seek to elucidate both their functional roles in tumorigenesis and the relationship between their expressions and clinicopathological findings or clinical outcome.
The authors are indebted to Dr R. Timms for correcting the English.