In this study we tested uEGF/MCP1 as a clinical biomarker for the risk of CKD progression in ADPKD. The biomarker was negatively correlated with HtTKV and concordant with the stratification of patients based on the Mayo Clinic Imaging Classification in two independent cohorts. Unlike the Mayo Clinic Imaging Classification, the uEGF/MCP1 could not discriminate across PROPKD categories, while the combination of genetic data and baseline uEGF/MCP1 ratio demonstrated that uEGF/MCP1 remained significant in predicting renal outcome in the discovery ADPKD cohort. The two cohorts were then combined to assess the prognostic performance of uEGF/MCP1. To this end, we used Conditional Inference Trees to identify slow (Mayo Clinic Imaging Classes1A–1B) and fast (Mayo Clinic Imaging Classes1C–1E) ADPKD progressors and obtained a highly sensitive model with cut-off levels of uEGF/MCP1 according to age. Although the follow-up in this study was relatively short for a slow-progressing disease like ADPKD, we found that uEGF/MCP1 at baseline correlated with renal outcome in terms of delta eGFR.

The specific trophic effect of EGF on kidney cells is well known, as its key functions in cell differentiation and regeneration allows it to modulate tissue response to injury [9, 18]. EGF tissue expression and urinary excretion decrease after kidney injury [12, 18, 19], as well as in ADPKD [12, 20] where EGF tissue expression decreases in human PKD1 cysts [10] and the growth factor receptor system, involved in tubular cell proliferation, is imbalanced, suggesting its detrimental role in ADPKD [20, 21]. Moreover, EGF mRNA levels decreased in cystic kidneys of an ADPKD murine model, while a number of growth factor genes increased with disease progression demonstrating an involvement of EGF with the progression of cystic lesions instead of ADPKD cystogenesis [22]. uEGF excretion mirrors intrarenal EGF expression and, better than serum EGF, it showed its independent predictive value of CKD progression in several large cohorts [19]. Moreover, Ju et al. demonstrated that the addition of uEGF to standard clinical parameters substantially improves the ability to predict renal outcomes in a heterogeneous CKD population [19]. We previously demonstrated the decrease of uEGF in ADPKD patients as a prognostic marker of incipient renal insufficiency [12], here we confirmed, in a larger cohort, that lower baseline uEGF predicts eGFR decline in ADPKD patients.

In contrast to EGF and growth factor expression, the upregulation of Mcp1 precedes macrophage infiltration and promotes macrophage accumulation and cyst growth in Pkd1-knock-out mouse models. In addition, the double-knock-out mouse for Pkd1 and Mcp1 showed a significantly decreased rate of cyst growth and improved kidney function suggesting a substantial role of MCP1 in macrophage-mediated cyst growth [23]. MCP1 is a potent chemotactic factor for monocytes that plays an important role in inflammatory processes. Its tissue expression was up-regulated in the kidney of a rat model of PKD and reduced by an inhibitor of MCP1 synthesis, although it did not prevent renal cyst growth [24]. The authors suggested that the MCP1 synthesis inhibitor did not completely abrogate macrophage accumulation in the renal interstitium which could still contribute to cyst growth. In humans, MCP1 expression increases in PKD1 kidneys, [10] and indeed, high levels of uMCP1 in ADPKD patients have been found [19], which correlated with kidney cyst size, and were suppressed after tolvaptan treatment [25]. In a longitudinal study, [14] uMCP1 was investigated as a marker of disease progression because its baseline excretion increased in a cohort of 55 ADPKD patients prior to a substantial increase in serum creatinine concentration or urine protein excretion. The disease progression predictive power of high levels of uMCP1 was also indicated by a study on mice with PKD where uMCP1 increased earlier than serum creatinine and blood urea nitrogen [26]. It has been reported that a hypoxic environment induced by the pressure of growing cysts stimulates the expression and release of hypoxia-inducible factor-1α and the pro-angiogenic gene vascular endothelial growth factor, which in turn, can upregulate the expression of MCP1 [27].

In an ADPKD cohort, uMCP1 correlated positively with height-adjusted TKV and negatively with the eGFR slope. Furthermore, a multivariate model including urinary levels of β2-microglobulin, MCP1 and vascular endothelial growth factor improved the ability to predict the decline of eGFR in ADPKD patients compared with height-adjusted TKV alone [27]. Likewise, baseline uMCP1, as well as the β-2-microglobulin, were associated with eGFR decline in the ADPKD cohort, and, when it was added to a model containing conventional risk markers that explained annual changes in eGFR its performance significantly increased [13]. According to these findings, we observed that baseline uMCP1 alone is a better predictor of eGFR decline in ADPKD compared to uEGF, though we used the urine spot test instead of 24 h-urine samples [13]. The 24-h urine sample may be more accurate than the urine spot tests used in most other studies because of the circadian rhythm in the urinary excretion of the markers, but it is inconvenient due to the long collection times, which can sometimes be inaccurate and at risk of bacterial contamination. Instead, the use of second morning urine in our study, besides being more practical, ensures the quality of the urine sample as the specimen can be collected directly in the clinic upon the patient’s arrival, thereby significantly shortening the time between sample collection and processing.

We examined whether decreased uEGF and increased uMCP1 levels in our patients mirrored changes in their gene expression in human and murine cystic tissues. Indeed, we found a significant down-regulation of EGF and up-regulation of MCP1 in both human and murine PKD1 cystic kidney tissues from previously published microarray data, and confirmed this by qRT-PCR of additional cystic samples [10]. Taken together, the data suggest that uEGF and uMCP1 levels mirror changes in the cystic tissues, supporting the notion that they are pathophysiological biomarkers of ADPKD, although the exact mechanisms of EGF and MCP1 in cystogenesis have not been fully clarified.

Recent data confirmed an inverse correlation between uEGF and age [28]. These authors demonstrated the exponential decrease of serum and urinary EGF with age in healthy adults and children, emphasizing the importance of EGF in renal maturation and growth during the first years of life [28]. The age-related decrease of EGF could be ascribed to a reduced capacity of the kidney to regenerate and to recover renal function. Our data confirm a significant, negative correlation between uEGF and age, as well as for uEGF/MCP1; the central role of age in the ctree statistical model, which allows a rapid clinical use of the data, is also highlighted.

In previous studies, uEGF/MCP1 was found (i) to correlate better with renal prognosis in patients with IgA nephropathy [15]; (ii) to be independently associated with tubular atrophy and interstitial fibrosis severity in primary glomerulonephritis [23]; (iii) to predict complete remission in primary glomerulonephritis [29]; however, uEGF/MCP1 has not been tested in ADPKD. Our data showed that baseline uEGF/MCP1 is associated with the severity of disease, better than uEGF and uMCP1 alone. The stronger correlation of baseline uMCP1 levels with height-adjusted TKV and with the predicted variation of eGFR at 10 years (ΔT0–T10), compared to uEGF, may be due to the involvement of MCP1 in cyst growth, and likely, in the disease pathogenesis [8]. Based on association study data, MCP1 appeared to play a major role in ADPKD compared to EGF, which seems consistent with data reported in the literature [8, 14, 26].

Although EGF and MCP1 both correlate significantly with TKV adjusted for height, the uEGF/MCP1 shows a stronger, inverse correlation with renal volume in both the discovery and validation cohorts thus emphasizing its role in the clinical setting, not least in recommending a therapy.

The unsolved issue of CKD in ADPKD patients is the difficulty in predicting progression and the lack of specific biomarkers. Genetic analysis of PKD1 truncating mutations helps the prediction of ESRD [30], but is not always available. The PROPKD score takes into account clinical and genetic data to stratify the risk of disease progression in ADPKD patients, but it is limited to patients older than 35 years or, with regard to patients younger than 35 years old, only to those suffering from hypertension and with urologic complications [30]. Currently, height-adjusted TKV, used in combination with age and eGFR, is the best imaging biomarker to predict eGFR decline in ADPKD patients [31], to help in the selection of patients for clinical trials [32], and to guide the therapeutic approach. A recent study proposed the use of the urine-to-plasma urea ratio to predict the progression of ADPKD, together with other risk markers (TKV and PKD gene mutation). Although the combined risk score of these three risk markers predicted rapid progression of the disease better than each predictor taken alone, the laboriousness of the single risk marker measurement may still represent a limitation [33].

Some limitations of this study that might hamper data generalizability must be stated. Specifically, the relatively small number of participants that was included, which implies the lack of a validation cohort for the Conditional Inference Tree analysis. We are aware that to perform an accurate direct comparison between the prognostic performance of uEGF/MCP1 and the PROPKD score or Mayo Clinic Imaging Classification, longer follow-up and a larger cohort are needed. A novel study is ongoing to collect extended follow-up data towards a comprehensive model to test uEGF/uMCP1 with genetic and imaging data.

In conclusion, we demonstrated the role of baseline uEGF/MCP1 as a non-invasive pathophysiological biomarker that can be used with other conventional risk markers in the clinical setting for risk stratification in ADPKD patients.