Metastatic renal cell carcinoma (mRCC) accounts for one-third of 73,750 new RCC patients at initial presentation in the US [1
] and 14,380 deaths are estimated in 2020 [2
]. mRCC is a group of heterogeneous tumor types characterized by distinct biologic patterns and prognosis [3
]. In the past 15 years, the introduction of targeted therapy, and more recently immunotherapy, have increased the therapeutic armamentarium available to treat metastatic stage and increased overall survival (OS) [5
]. In parallel, major technological advances have led to a better understanding of the molecular basis of disease progression and the identification of several prognostic biomarkers [8
]. However, well-validated prognostic biomarkers associated with tumor biology of metastatic kidney cancer stage are lacking.
Recent clinical trials have used the International Metastatic Database Consortium (IMDC) model for prognostication [11
]. This model is based on six clinical variables as risk factors for short survival (time from diagnosis to initiation of therapy of less than 1 year, Karnofsky performance status of less than 80%, serum hemoglobin below the lower limit of normal and corrected calcium, neutrophil count, and platelet count greater than the upper limit of normal) and classify patients into good- (0 risk factor), intermediate- (1 or 2 risk factors) and poor-risk (≥3 risk factors) groups [14
]. As a result, targeted therapeutics in metastatic RCC uses non-molecular based clinical prognostic variables for initiating novel drug interventions and combinations highlighting the need for enhancing prognostic markers in metastatic RCC [15
We have reported previously [17
] progastrin to be elevated in plasma of patients with metastatic kidney cancer. In physiology, progastrin is the precursor of gastrin synthetized by antrum G cells and processed into gastrin [18
]. Progastrin does not accumulate in G cells, by contrast to G34-Gly and gastrin [19
]. G34-Gly will generate gastrin upon full maturation. As a consequence, progastrin is barely detectable in the blood of healthy subjects even though few of them have been shown to be positive as observed the first time by Siddheshwar et al. [20
]. However, in line with the demonstration of the expression of the GAST
gene, encoding progastrin, in colorectal tumors as well as other tumor types, high levels of hPG80
(named as such when progastrin is released from tumor cells and detected in the blood) were reported in the blood of cancer patients [17
]. Moreover, in addition to the fact that GAST
is a direct target of the ß-catenin/Tcf4 pathway, activated in many cancers, including RCC [22
], a large body of literature supports the functional role of hPG80
in tumorigenesis [21
]. As a consequence, hPG80
is an interesting indicator of tumor behavior/activity and may also represent a factor for aggressive biology and clinical outcomes.
In the present study, we determined the prognostic value of hPG80 in clinically diagnosed metastatic RCC (mRCC) patients and examined whether hPG80 might improve stratification of the currently used IMDC model to predict overall survival (OS).
2. Materials and Methods
2.1. Patients and Control Cohorts
illustrates the different cohorts used in this study to determine prognostic value of hPG80
in metastatic RCC stage.
2.1.1. mRCC Patient Cohort
A large tertiary level, clinically annotated hospital registry with prospective and uniform blood/plasma collection from non-fasting metastatic kidney cancer patients between 5/2011 and 9/2013 and uniform sampling was used as has been previously described [29
]. All patients (n
= 143) provided written informed consent for research at the time of their blood collection, in line with international regulations and ICH GCP (International Conference on Harmonization- Good Clinical Practice) and on an Institutional Review Board approved study protocol (Mayo Clinic IRB #11-005855 00). Of note, the majority of the patients had the blood draw close to the time of the diagnosis of metastatic disease, with a median value of 6 days (IQR: 0–117).
2.1.2. RCC Patient Cohort
Blood samples from non-fasting RCC patients (n = 39) were obtained from Tissue For Research Ltd. (Spectrum Health System, Grand Rapids, Michigan, MI, USA).
2.1.3. Non-Cancer Age and Non Age Matched Control Cohorts
A young-aged cohort of plasma samples (18–25 years old) was obtained from non-fasting 137 healthy blood donors, from the French blood agency (Etablissement Français du Sang
). This cohort was assumed to be at very low risk of cancer in order to maximize the likelihood of the control group being cancer-free [31
]. A second, over 50 years old cohort of plasma samples (range 50–80 year old; median value 55 year old) was collected from fasting 252 healthy subjects, from the interim analysis from the PROCODE study (NCT03775473, https://clinicaltrials.gov/ct2/show/NCT03775473
2.2. hPG80 Level Measurements in the Blood Samples
The DxPG80 test is an Enzyme-Linked Immunosorbent Assay (ELISA) for the quantitative measurement of human hPG80 in EDTA plasma. The test is based on the principle of a sandwich ELISA to measure the concentration of hPG80 in plasma specimens that have been anticoagulated with EDTA. Briefly, a capture monoclonal antibody raised against the C-Terminus of hPG80 and that does not recognize active gastrin (gastrin-NH2) is pre-coated on the 96-well plate. hPG80 present in calibrators, controls and/or specimens added to the wells bind to the immobilized capture antibody. The test plate includes calibrators which are used to estimate the level of hPG80 in EDTA plasma samples. The wells are washed and a polyclonal antibody raised against the N-Terminus of hPG80 coupled with horseradish peroxidase (HRP) is added (i.e., detection antibody), resulting in an antibody-antigen-antibody complex. After a second wash, a 3,3′,5,5′-Tetramethylbenzidine (TMB) substrate solution is added to the well, producing a blue color in direct proportion to the amount of hPG80 present in the initial sample. The Stop Solution changes the color from blue to yellow, and the wells are read at 450 nm with a microplate reader.
The analytical performances were assessed following EMEA/CHMP/EWP/192217/2009. Accordingly, there is no cross-reactivity or interference if the variation in the percentage of recovery (when compared to control (vehicle)) is equal or does not exceed 20%, and there is no change in the interpretation of the result.
Cross-reactivity was assessed using human plasma samples spiked with native hPG80 concentrations ranging from negative to strong positive and a fixed concentration of each potential cross-reactants (i.e., KLH (Keyhole Limpet Hemocyanin), gastrin-17, Gastrin-Gly, CTFP (C-Terminus Flanking Peptide) at 2 µg/mL, CA125 (cancer antigen 125) at 2000 U/mL, CA15-3 (cancer antigen 15-3) at 100 U/mL, CEA (carcinoembryonic antigen) at 20 µg/mL or PSA (prostate specific antigen) at 210 mg/mL). For all the potential cross-reactants, the variation of the % of recovery was below 20%, showing no cross-reactivity.
Interference was also assessed using human plasma samples spiked with native hPG80 concentrations ranging from negative to strong positive and a fixed concentration of each potentially interfering substances (i.e., SN-38 at 60 µM, 5-FU at 3 mM, triglycerides at 0.05 mg/mL, cholesterol at 25 µg/mL, conjugated bilirubin at 0.5 µg/mL or hemoglobin at 2 mg/mL). For all the potential interfering substances, the variation of the % of recovery was below 20%, showing no interference.
Based on EMEA/CHMP/EWP/192217/2009, the within-run, inter-run and inter-operator variability is defined as being the mean coefficient of variation (CV) % value of each measured control sample. It is considered acceptable when ≤20%. The CV% were determined using hPG80-negative EDTA plasma spiked with three controls of hPG80 at 2.5, 12.5 and 22.5 pmol/L. For all the variabilities, the CV% was below 10%, showing acceptable variabilities.
hPG80 concentrations in pmol/L (pM) were calculated using the standard curve equation of the native hPG80 calibrators prepared in hPG80-negative EDTA plasmas. The limit of detection (LoD) and limit of quantification (LoQ) were calculated, as per EMEA/CHMP/EWP/192217/2009 and NCCLS EP 17-A vol.24 no.34, based on the standard deviation (CV) of the measured concentrations of n = 74 blanks as LoD = 3 × CV and LoQ = 10 × CV. DxPG80 have a LoD of 1.2 pM and a LoQ of 2.3 pM.
2.3. Statistical Analyses
Comparisons between groups were performed using two-tailed Mann-Whitney U-test for unpaired non-parametric variables. Spearman’s rank correlation coefficient was used to measure the association between between hPG80 levels and age. The overall survival (OS) was defined as the time from blood collection from mRCC patients with known date of death. The survival curves were constructed using the Kaplan-Meier method and compared performing a log-rank test on mRCC patients with full survival data. An optimal cutoff value of hPG80 was defined using the function of “surv_cutpoint” in R Package “survminer”, calculating the minimal p-value based on the log-rank method. R software 3.6.1 (The R Foundation for Statistical Computing) was used to perform all the statistical analyses. Prism software (GraphPad, La Jolla, CA, USA) was used to create figures. The level of significance was set at p < 0.05.
The close relationship between hPG80
and cancer progression has been previously reported in several studies [21
]. We have recently shown that tumor cells from many cell types overexpress hPG80
]. Its expression is related to cancer cell activity and represents a risk factor for tumor recurrence and therefore hPG80
may provide prognostic significance in metastatic stages of tumors which secrete hPG80
in measurable amounts. In this study, we demonstrate for the first time the prognostic value of hPG80
in mRCC. We also demonstrated the potential diagnostic utility of hPG80
levels in differentiating RCC and mRCC patients from healthy individuals. We established an optimal cutoff value for hPG80
(4.5 pM) in mRCC stage and were able to demonstrate that higher hPG80
levels are associated significantly with lower OS. We also explored hPG80
levels in mRCC patients for stratification into good and poor prognosis groups like the widely used IMDC model who classify patients into three risk groups (good, intermediate and poor) [14
Our results suggest that addition of hPG80
levels to IMDC might further improve the IMDC prognostication, and also refines the heterogeneity within the intermediate-risk IMDC group. Patients in the intermediate-risk group account for 40 to 50% of all risk classes patients [32
]. In our patient cohort, 56.6% of patients had IMDC intermediate-risk group. Study from Sella et al. has shown that patients with one-risk factor had longer OS than patients with two-risk factors, suggesting that adding new prognostic risk factors to actual IMDC risk model such as tumor and stage associated hPG80
levels could enhance the accuracy of prognostification [34
]. We observed that hPG80
levels collected and measured at the time of patients progressing to metastatic stage were able to refine the existing clinical factor based IMDC risk model prognostication of survival in patients with intermediate-risk group. In this group patients with low and high hPG80
levels were able to be identified with different survival time (median OS: 17.9 in patients with high hPG80
levels versus 29.8 months with low hPG80
= 0.0083). In line with these data, a recent study from Kunishi et al. has shown that combining C-reactive protein (CRP) value to patients classified in the intermediate group according to the IMDC risk classification might further refine prognosis in this patient populations [35
]. Altogether, although the optimal hPG80
cutoff value will need further validation in larger cohort, it supports the benefit of further studies using hPG80
levels to stratify mRCC patients and to improve the prognostic predictive power of the existing IMDC model.
Novel tissue and blood-based biomarkers including PD-L1 expression, CRP, VEGF and IL-6 serum levels have been proposed as prognostic biomarkers for mRCC [36
]. Other putative biomarkers such as miRNAs, circulating DNA and metabolic biomarkers have been reported to have prognostic relevance in mRCC [36
]. In addition, assessment of gene expression signatures has also shown to significantly improve the predictive power of the IMDC model and provided additional prognostic information in mRCC patients [37
]. However, most of these factors were derived from retrospective studies and none of them have been validated in larger prospective studies. Currently, no specific molecular marker has been shown to improve the accuracy of existing prognostic scores which are largely based on clinical and not tumor biology associated factors (e.g., IMDC) and their use is not recommended for clinical care [38
Cancer stem cells (CSCs) have been implicated in tumor initiation, progression, metastasis, multidrug resistance and recurrence [39
]. Several lines of evidence indicates that renal CSCs are involved in driving RCC progression and treatment failure [40
]. Previous studies have shown that hPG80
plays a major function in CSCs by regulating their survival and self-renewal ability [21
]. Thereby, we can hypothesize that elevated hPG80
levels are associated with poor clinical outcome in mRCC patients by promoting CSCs survival and resistance to antiangiogenic drugs like tyrosine kinase inhibitors.
Our study presents some limitations including the use of the prospectively enrolled mRCC cohort who were retrospectively analysed and the number of patients excluded from overall survival analysis for a variety of reasons. A third of our patients were lost to follow-up and this could have biased our results. Nevertheless, hPG80 levels in mRCC biology has been well reported and the ability to measure hPG80 levels in blood provides a significant advantage for its future exploration as a prognostic or predictive biomarker.