Our data demonstrate a significant inverse association between lower urinary CTGF concentrations and the development of CKD stage 3 in this case-control study. A similar, but non-significant, trend was observed in a complementary analysis of the ARIC study, and results were robust in meta-analysis of both studies.
CTGF belongs a family of matricellular proteins known as CCN, an acronym derived from three prototypic members: Cysteine-rich protein 61, CTGF, and Nephroblastoma overexpressed protein. These proteins interact with cell surface receptors in the extracellular matrix, growth factors, proteases, cytokines and extracellular matrix proteins35
and stimulate mitosis, adhesion, apoptosis, extracellular matrix deposition, and the migration of several cell types.36
In health, CTGF appears to play an essential role in embryogenesis,37
angiogenesis (via modulation of vascular endothelial growth factor [VEGF] function),37,38
wound healing and tissue repair,37
including mesangial repair following kidney injury.26
However, CTGF has mostly been studied in the setting of disease, particularly fibrotic kidney disease and diabetic nephropathy.39 In vitro
, CTGF is a highly up-regulated protein in glucose-exposed kidney mesangial cells.40 In vivo
, increased expression of CTGF is observed in the renal cortex of rats with streptozotocin-induced diabetes mellitus.41
CTGF is an essential downstream mediator of the pro-fibrotic effects of TGF-ß,42
and has been shown to mediate advanced glycation end product–induced epithelial–mesenchymal transition and kidney fibrosis in rodents.43
Furthermore, treatment with antisense molecules directed against CTGF has been shown to attenuate the development of tubulointerstitial fibrosis in animal models of fibrotic kidney disease.44
In cross-sectional human studies, urinary CTGF excretion correlates with disease severity in diabetic nephropathy45,46
and was shown to predict progression of albuminuria, but not decline in GFR, in a small prospective study of patients with diabetic nephropathy.47
The present study extends knowledge in this area in several key ways. First, we demonstrate a longitudinal relationship between urinary CTGF concentration and subsequent kidney disease in a population-based setting. We also demonstrate that, in this disease-free cohort, the direction of the relationship with risk of kidney disease appears reversed, with lower
urinary CTGF concentrations associating with worse kidney outcomes. These observations in the low and normal CTGF range may appear to contradict the prevailing understanding of CTGF biology in diabetic nephropathy, where levels are typically markedly higher,47
and merit further discussion.
Despite its name, CTGF is not a true growth factor, but rather a contextual modulator of cell function whose effects depend on the local cellular microenvironment, and whose functions are complex and cannot be easily generalized.38
For example, transient physiologic expression of CTGF in response to injury results in wound healing,48
and CTGF is released in response to a wide variety of injurious agents, including glucose, advanced glycation end-products, angiotensin II, endothelin, glucose, thrombin, stretch, and oxidative stress.48
Conversely, constitutive over-expression of CTGF promotes an environment where such stimuli may induce potent fibrotic responses,36
and such over-expression is associated with the development of fibrosis.49
In addition, the local level of expression of CTGF markedly influences its ultimate effects at a tissue level. For example, CTGF at low concentrations is a promoter of angiogenesis, whereas at high concentrations CTGF causes inhibition of this process via modulation of VEGF function.38
As such, CTGF may be pro-angiogenic, anti-angiogenic, or may not be involved in angiogenesis at all depending on the local level of expression. Many other concentration-dependent effects of CTGF also have been described, including its induction by TGF-ß, its release in response to mechanical forces, and its modulation by chromatin modification.38
Although the mechanism of the observed inverse association between CTGF and risk of kidney disease is unclear, a similar differential effect of CTGF concentration may be responsible. Our most robust findings were observed in participants without
diabetes, an understudied population in terms of CTGF biology, and must be interpreted in that context. It is notable that urinary concentrations of CTGF are typically 100-fold higher in patients with diabetic nephropathy when compared with unaffected diabetic controls47
and, as such, the literature on diabetic nephropathy may not apply to the present population. It is possible that at these basal levels of expression, CTGF is necessary for mesangial health via the mediation of mesangial repair following injury. This hypothesis is supported by in-vitro
studies demonstrating CTGF to be an essential pro-resolution factor following mesangial injury.26
These studies demonstrate that mesangial cells demonstrate little or no capacity to heal in the absence of CTGF, and that the pace of mesangial repair is a function of local CTGF concentration.26
This proposed hypothesis, whereby CTGF performs distinct functions in different contexts, mirrors exactly the `homeostatic challenge model' described for TGF-ß, a closely-related cytokine.50
Itself a powerful trigger of tubular cell injury and fibrosis, TGF-ß has recently been shown to also have an essential role in maintaining epithelial cell homeostasis, such that a deficiency in signaling paradoxically triggers tubular cell injury and fibrosis.51
Tight regulation of TGF-ß signaling appears vital for kidney epithelial cell structure and health, with the cytokine apparently performing an essential custodial function
under basal conditions.52
Studies with larger sample sizes, spanning later and more severe CKD, are needed to confirm whether a similar paradigm might explain the present observations, stemming as they do from this small, pilot study.
Of note, it is also possible that CTGF per se may not be directly responsible for these observations. For example, the results may relate to the form of CTGF that was measured. As CTGF is an adherent matricellular protein, tissue levels (which are the likely mediator of fibrosis) may not correlate with secreted/excreted protein. Further studies incorporating measures of plasma CTGF would help address this issue.
The findings from the present work suggest that CTGF levels may predict the development of CKD stage 3 in the general population and in participants without diabetes up to a decade before clinical diagnosis. However, this work also highlights the complex biology of CTGF, and efforts to elucidate the mechanisms underlying the association between higher basal CTGF expression and lower risk of kidney disease require further study. Furthermore, those involved in current efforts to develop anti-CTGF therapies for the treatment of fibrotic diseases should be aware of the attendant possibility that there may be an optimal level of CTGF expression, below which harm may occur, possibly due to inadequate repair of ongoing damage. This concern is supported by a recent study of TGF-ß blockade, which unexpectedly resulted in increased fibronectin deposition in renal tubular cells.53
This study has several strengths, including the long follow-up, the richness of the dataset and consistent findings in an independent, external analysis. Limitations pertain to the observational nature of the study, such that causality cannot be inferred from our results. Given that this is a nested case-control study, the results may not be generalized to the Framingham Heart Study as a whole. Also, although every effort was made to adjust for confounding by baseline proteinuria and eGFR, the possibility of residual confounding cannot be completely eliminated. Furthermore, results were stronger in the Framingham cohort compared to ARIC. It is possible this observation may be explained by a `winner's curse' i.e. upward bias of effect sizes in early studies. Nonetheless, further studies in larger, population-based cohorts are desirable to confirm the findings of this exploratory, pilot study.
In summary, the presence of lower urinary CTGF concentrations is associated with the development of CKD stage 3 in two independent study samples. Further work is required to fully characterize how CTGF influences risk of kidney disease.