Circulating Soluble Fms-like Tyrosine Kinase in Renal Diseases Other than Preeclampsia

Abstract

Soluble Fms-like tyrosine kinase (sFlt-1/sVEGFR1) is a naturally occurring antagonist of vascular endothelial growth factor (VEGF). Despite being a secreted, soluble protein lacking cytoplasmic and transmembrane domains, sFlt-1 can act locally and be protective against excessive microenvironmental VEGF concentration or exert autocrine functions independently of VEGF. Circulating sFlt-1 may indiscriminately affect endothelial function and the microvasculature of distant target organs. The clinical significance of excess sFlt-1 in kidney disease was first shown in preeclampsia, a major renal complication of pregnancy. However, circulating sFlt-1 levels appear to be increased in various diseases with varying degrees of renal impairment. Relevant clinical associations between circulating sFlt-1 and severe outcomes (e.g., endothelial dysfunction, renal impairment, cardiovascular disease, and all-cause mortality) have been observed in patients with CKD and after kidney transplantation. However, sFlt-1 appears to be protective against renal dysfunction-associated aggravation of atherosclerosis and diabetic nephropathy. Therefore, in this study, we provide an update on sFlt-1 in several kidney diseases other than preeclampsia, discuss clinical findings and experimental studies, and briefly consider its use in clinical practice.

The vascular endothelial growth factor-A (VEGF-A or VEGF) is an important regulator of blood vessel formation and homeostasis through VEGF receptor 2 (VEGFR2) signaling. VEGFR1 or Fms-like tyrosine kinase (Flt-1) is a decoy receptor due to its high affinity to VEGF and low kinase activity. Its soluble form, sFlt-1, results from alternative splicing and ectodomain shedding of the membrane-bound receptor. By binding VEGF and blocking the membrane-bound receptors, sFlt-1 functions as a naturally occurring antagonist of VEGF. Soluble Flt-1 also binds placental growth factor (PlGF).^1–4

Despite being a soluble protein, sFlt-1 can be stored locally by binding to heparan sulfate proteoglycans on the cell surface and in the extracellular matrix through its heparin-binding domain. This property renders stored sFlt-1 displaceable by heparin (Figure 1).^5–7 Local retention of sFlt-1 may be protective against excess VEGF.⁸ In specialized pericytes, sFlt-1 exerts autocrine functions, controlling cell morphology and functioning independently of VEGF bioavailability. In podocytes, local production of sFlt-1 promotes cell adhesion and activates signaling pathways involved in cytoskeleton reorganization.⁹

Figure 1.

Soluble Flt-1 release in response to heparin. sFlt-1 results from alternative splicing and ectodomain shedding of the membrane-bound receptor (mFlt-1). It is stored on the cell surface and in the extracellular matrix via binding to heparan sulfate proteoglycans (HSPG). Because sFlt-1 contains a heparin-binding domain, it can be displaced from its cellular stores by heparin and heparin mimetics.^5,51 Importantly, sFlt-1 released by heparins seems to preserve its antiangiogenic properties.^51,7 Heparin does not regulate sFlt-1 at the transcriptional level but may induce shedding of the extracellular domain of mFlt-1, probably by activating proteolytic enzymes.^6,75 Several proteases able to cleave mFlt-1 have been described to be upregulated/activated in kidney diseases, including disintegrin and metalloproteases (ADAMs)^6,75 and disintegrin and metalloproteases with thrombospondin motif (ADAMTs),^6,30 along with others. Heparanase, a β-D-endoglucuronidase that cleaves heparan sulfate chains, can release HSPG-bound sFlt-1.^5,70,74 Manipulations of the heparin/heparanase system may be extremely useful to control circulating sFlt-1 levels. For instance, patients receiving heparin-based dialysis are presented with a marked increase in sFlt-1 levels after the start, returning to baseline at the end of the dialysis session, sFlt-1 levels remained unchanged in patients receiving heparin-free dialysis.⁵⁰

However, sFlt-1 can also be released into the circulation under pathophysiological conditions and affect remote organs by the indiscriminate neutralization of VEGF or PlGF, contributing to the impairment of endothelial repair, chronically compromising endothelial function.⁵ Of particular importance are sites of paracrine and autocrine loops, where constitutive VEGF signaling and bioavailability are needed to maintain vessel and oxygen homeostasis and endothelial fenestrations.^10–12

Circulating sFlt-1, the Endothelium, the Heart, and the Kidney: Basic Interactions

Endothelial Dysfunction

Endothelial dysfunction and microvascular disease are common in patients with kidney disease, with both conditions related to adverse clinical outcomes and high mortality.¹³ Several in vitro studies have shown that excess sFlt-1 in serum from patients causes endothelial dysfunction and leads to an antiangiogenic state.^14,15 Accordingly, serum from patients with stage 2–5D CKD displayed an antiangiogenic effect in the chorioallantoic membrane (CAM) assay, a robust technique used to study angiogenesis in vivo. Removal of sFlt-1 from the samples by immunoprecipitation prevented this effect.¹⁶ Serum from patients with ANCA-associated vasculitis also showed an apparent effect on the CAM microvasculature, with early blood flow disruption. Coincubation with excess recombinant VEGF or clearance of molecules with a molecular mass >10 kDa (which includes sFlt-1) avoided the antiangiogenic effect of patient serum on the CAM.

Interestingly, Cindrova-Davies et al. demonstrated that sFlt-1 increases endothelial cell-leukocyte adhesion upon TNFα treatment in vitro.¹⁷ They did not observe a significant increase in endothelial adhesiveness with sFlt-1 alone. Still, they showed that sFlt-1 and any kind of VEGF signaling antagonism renders endothelial cells more sensitive to a proinflammatory insult. Walshe et al. also observed increased leukocyte-endothelial interactions in mice receiving sFlt-1 adenovirus and increased endothelial expression of adhesion molecules.¹⁸ These findings suggest sFlt-1 involvement in endothelial activation and its contribution to the exacerbation of a proinflammatory microenvironment.

Microvascular Changes in the Kidney

The kidney possesses a highly organized vasculature and specialized endothelial cells, which are critical for its function. VEGF production by podocytes is required for glomerular endothelial health, whereas tubulovascular crosstalk by VEGF is essential to preserve peritubular capillary (PTC) networks.^10,19,20 VEGF is crucial for maintaining the renal microvasculature and stabilizing the renal function in renal ischemia-reperfusion, CKD, and preeclampsia. However, excess VEGF has harmful effects in some renal pathologies, such as diabetic nephropathy and common complications of CKD, including atherosclerosis.^21–24 Therefore, the role of sFlt-1 will reflect the specific role of VEGF on the pathologic condition in question.

Elevated circulating sFlt-1 levels or the neutralization of circulating VEGF by anti-VEGF antibodies in rodents leads to glomerular endothelial cell damage, detachment, and endotheliosis. Occasional proteinuria accompanied by podocyte damage and a reduced expression of nephrin, a key protein of the slit diaphragm of podocytes, are also observed in these animals.^14,25–27 The authors discussed that the effect on podocytes might have been mediated by the damage of glomerular endothelial cells due to the lack of survival signals from circulating VEGF.²⁵ Further experiments showed that sFlt-1 binds to lipid rafts of podocytes, then activates intracellular signaling, including nephrin phosphorylation. Thus, sFlt-1 has essential autocrine functions that control podocyte behavior.⁹ Whether excess sFlt-1 under pathologic conditions could disrupt such autocrine signaling remains to be elucidated.

Interestingly, increasing circulating sFlt-1 levels through sFlt-1 gene transfer decreases nephrin expression and accelerates proteinuria and the progression of glomerulosclerosis in rat models of crescentic glomerulonephritis. Besides, sFlt-1 also contributes to progressive glomerular and PTC loss in this model.²⁸ Similarly, the administration of recombinant sFlt-1 to rodents resulted in significant structural and functional changes in renal vasculature.²⁹ Observed morphologic changes, such as narrowing of the lumina, the presence of apoptotic cells, the protrusion of endothelial cells into the lumen, and fibrosis formation, suggest that increased sFlt-1 contributes to PTC loss by inducing capillary regression, even in the absence of a previous insult.

After an acute ischemic insult, such as ischemia-reperfusion, similar alterations in PTCs are observed, and an increase in circulating sFlt-1 levels and local expression.^29,30 Because angiogenesis is a critical mechanism that determines kidney recovery after such an insult,^23,24 repression of VEGF by elevations in sFlt-1 may relate to the lack of adaptive repair or stabilization after injury. Interestingly, neutralization of sFlt-1 by recombinant VEGF after ischemia-reperfusion injury in animals preserved capillary density and alleviated fibrosis formation.^23,29

Tight regulation of VEGF levels is necessary to maintain glomerular function, because both upregulation and downregulation of VEGF can cause kidney disease.^31,32 VEGF is considered a mediator of diabetic nephropathy. Therefore, inhibiting VEGF may be beneficial in treating renal complications under diabetic conditions.^21,33,34 Accordingly, overexpressing sFlt-1 in podocytes or increasing its circulating levels in diabetic animals will have protective effects on glomerular function by reducing mesangial expansion, matrix deposition, and macrophage infiltration and by normalizing albuminuria.^32,35,36

Microvascular Changes in the Heart

In the heart, VEGF inhibition, either by increasing circulating sFlt-1 or overexpressing sFlt-1 in the myocardium, results in cardiac phenotypes that include decreased vascularization, fibrosis, myocardial hibernation, reduced contractile function, and diastolic dysfunction.^12,27,37 Because these phenotypes are directly associated with reduced capillary density in the heart, the microvasculature is likely involved. In five out of six nephrectomized rats (a well-established animal model of CKD), we found approximately 2.5-fold higher serum sFlt-1 levels, reduced capillary density, reduced myocardial blood volume, and a marked increase in diastolic dysfunction. Neutralization of sFlt-1 with recombinant VEGF attenuated all cardiac manifestations in this model, preserving heart microvasculature and function.²⁷

Conversely, increased sFlt-1 is protective against renal dysfunction-associated progression of atherosclerosis. Increased atherosclerotic plaque formation and intensified macrophage infiltration into the plaque were observed in five out of six nephrectomised ApoE-deficient mice. Noteworthy, serum sFlt-1 was decreased—whereas PlGF was increased—in these animals, and restoration of sFlt-1 levels by repeated intraperitoneal injections of recombinant sFlt-1 significantly reduced the atherosclerotic lesions observed.³⁸ Accordingly, the knockdown of the sFlt-1 gene in ApoE-deficient mice further aggravated lesions in animals fed a high-cholesterol diet, independently of renal function, thereby indicating a reduction of sFlt-1 production itself could worsen atherosclerosis.³⁹ This antiatherosclerotic, cardioprotective property of sFlt-1 may be related to its ability to antagonize PlGF, known for enhancing intramural angiogenesis and monocyte recruitment.^39–42

Clinical Aspects of Circulating sFlt-1 Levels in Kidney Disease

CKD and Dialysis

The clinical significance of elevated circulating sFlt-1 in kidney disease was first shown for preeclampsia, the major renal complication of pregnancy.⁴³ Circulating sFlt-1 levels are increased in different types of diseases with varying degrees of renal impairment. These levels are mostly positively associated with proteinuria and inversely correlated with the eGFR (Tables 1 and 2).^{16,27,39,44–46}

Table 1.

Characteristics of included studies

Study	No. of Patients	Renal Status	Follow-up	Outcomes
Guo, 2009	185	CKD Stage 5D	31 months	Death
Di Marco, 2009	130	CKD Stages 2–5D	Cross-sectional	Renal function, CVD
Onoue, 2009	329	eGFR < 60	Cross-sectional	Renal function, CVD
Ky, 2011	1403	eGFR 33–98	24 months	Renal function, CVD, death
Yuan, 2013	211	CKD Stage 5–5D	29 months	Renal function, death
Matsui, 2014	291	CKD Stages 3a-5D	6 months	Renal function, CVD
Rambod, 2014	301	CKD Stage 2–4	32 months	Renal function, CVD, death
Di Marco, 2015	586	eGFR 25–94	6 months	Renal function, CVD, death
Mansour, 2019	1444	CSA-AKI	12 months	Renal outcomes, death
Chapal, 2013	136	KTx	3 months	Renal outcomes
Wewers, 2019	93	KTx	24 months	Renal outcomes, death

CSA-AKI, cardiac surgery-associated acute kidney injury.

Table 2.

Association of sFlt-1 with kidney function in patients with CKD

Study	eGFR
	Correlation Coefficient	Changes in sFlt-1
Di Marco, 2009	r = −0.463	..
Onoue, 2009	r = 0.32 (postheparin)a	…
Ky, 2011	…	Beta regression coefficient (95% confidence interval): −5.0 (−8.8 to −1.2) for a 10-unit increase in eGFR
Yuan, 2013	r= −0.21	…
Matsui, 2014	r = −0.323 (preheparin)a r = 0.537 (postheparin)a	…
Rambod, 2014	r = −0.121	…
Di Marco, 2015	r = −0.384	B regression coefficient (95% confidence interval): −4.1 (−7.2 to −1.0) for a 1-unit increase in eGFR

^aBlood sample collected before (pre) or after (post) heparin injection.

This association between kidney disease and high circulating sFlt-1 levels could reflect the effect of reduced kidney function on sFlt-1 regulation, the contribution of sFlt-1 to kidney function decline, or even the presence of additional factors that affect both kidney function and sFlt-1 simultaneously. The experimental data described above support a direct role of sFlt-1 in the development and progression of kidney disease. The renal microvasculature is undoubtedly a critical target of sFlt-1, but endothelial dysfunction may also play an essential role in the cardiorenal interactions.

In a cross-sectional study with 130 patients with stage 2–5D CKD, we showed that moderately, albeit chronically, increased sFlt-1 levels correlated with decreased eGFR and markers of endothelial dysfunction/activation, namely soluble vascular cell adhesion molecule 1 and vWF, and with a higher cardiovascular risk.¹⁶ Moreover, high sFlt-1 concentrations were associated with signs of heart failure and early mortality in a study cohort of 586 patients with varying degrees of renal impairment and angiographically documented coronary artery disease (CAD) (Tables 3 and 4; Supplemental Table 1).²⁷ No association was observed between sFlt-1 levels and the degree of CAD, suggesting sFlt-1 may contribute to micro rather than macrovascular disease in these patients. Accordingly, in a cohort of 1403 patients, Ky et al. also described higher sFlt-1 levels as strongly and independently associated with the risk of adverse outcomes and disease severity (e.g., New York Heart Association functional classes) in patients with chronic heart failure. Decreased eGFR and higher B-type natriuretic peptide (BNP) were each independently associated with elevated levels of sFlt-1, given the combination of sFlt-1 and BNP improved the efficacy of risk assessment when compared with BNP alone.⁴⁴ CKD and heart failure often occur together, whereas reduced kidney function contributes to poor patient outcomes with heart failure. To note, even in patients with the same degree of renal impairment, higher sFlt-1 levels on admission seemed to be a stronger predictor of outcomes, such as the development of severe acute heart failure after myocardial infarction.^47,48

Table 3.

Association of sFlt-1 with cardiovascular outcomes in patients with CKD

Study	sFlt-1 Levels/Cutoff	CVD	Outcomes and Effect Measures
Di Marco, 2009	…	Myocardial infarction	sFlt-1 elevated in MI group compared with event-free group (212 [125–530] versus 91 [68–127])a (P<0.001)
		Stroke	sFlt-1 elevated in stroke group compared with event-free group (171 [103–412] versus 94 [70–135])a (P=0.001)
Onoue, 2009	PlGF/sFlt-1	CAD	PlGF/sFlt-1 (post)b ratio increases (sFlt-1 decreases) according to the severity of CADc (P<0.05)
Ky et al., 2011	One-SD increase	Heart failure	Higher risk of Tx, VAD placementb HR, 1.14 (95% CI, 1.05 to 1.24) (P<0.01)
	>379 pg/mld	Heart failure	Higher risk of Tx, VAD placementb HR, 1.67 (95% CI, 1.06 to 2.63) (P=0.03)
	Q3/Q4 (>308 pg/ml)d	Heart failure	Highest quartiles associated with NYHA classes III+IV (P<0.01) and reduced EF (P<0.01)
	>308 pg/mld	Heart failure	In combination with high BNP, high sFlt-1 has a higher predictive accuracy for severe outcomes at 1 year than either marker alone AUC= 0.791 (95% CI, 0.752 to 0.831) (P<0.01 and P=0.03, respectively)
Matsui, 2014	>101.2 pg/ml (pre)b	CV eventse	No significant difference in the cumulative incidence of CV events between patients with higher and lower sFlt-1 (0.20 versus 0.13)f (P=0.10)
		CAD	sFlt-1 increases according to the severity of CADc (P=0.07)
	>239.7 pg/ml (post)b	CV eventse	No significant difference in the cumulative incidence of CV events between patients with higher and lower sFlt-1 (0.19 versus 0.15)f (P=0.06)
		CAD	sFlt-1 decreases according to the severity of CADc (P=0.03)
Di Marco, 2015	Q3/Q4 (>108 pg/ml)	Heart failure	Highest quartiles associated with NYHA classes III+IV (P=0.01) and reduced EF (P=0.003)
		CAD	No association with severity of CAD (P=0.77)

MI, myocardial infarction; SD, standard deviation; Q, quartile; VAD, ventricular assist device; HR, hazard ratio; 95% CI, 95% confidence interval; NYHA, New York Heart Association classification; EF, ejection fraction; AUC, area under the curve; CV, cardiovascular.

^asFlt-1 levels given as median (interquartile range).

^bBlood sample collected before (pre) and after (post) heparin injection.

^cSeverity of CAD is defined as the number of coronary arteries with stenosis over 75%. HR for adverse outcomes, including cardiac transplantation (Tx), VAD placement, and all-cause death.

^dTo note, the sFlt-1 assay used differs from the other studies (Supplemental Table 1).

^eCV events are defined as fatal or nonfatal newly developed CAD, sudden cardiac death, peripheral arterial disease, congestive heart failure requiring hospitalization, cerebrovascular disease, and aortic disease, including rupture and dissection of aortic aneurysm.

^fEstimated data extracted from the study. The italic text indicates trends only.

Table 4.

Association of sFlt-1 with all-cause mortality in patients with CKD

Study	Cutoff	All-cause Mortality
		Unadjusted HR (95% CI)	Adjusted HR (95% CI)
Guo, 2009	>141.5 pg/ml	1.75 (1.09 to 2.79)	1.93 (1.18 to 3.1)
Ky et al., 2011	>379 pg/mla	6.17 (4.3 to 8.86)b	1.67 (1.06 to 2.63)b
Yuan, 2013	>137.6 pg/ml	2.83 (1.32 to 6.06)c	2.33 (1.06 to 5.14)c
Rambod, 2014	>68 pg/mld	3.78 (1.39 to 10.28)	3.41 (1.49 to 9.51)
Di Marco, 2015	>118.1 pg/ml	2.87 (1.63 to 5.07)	…

HR, hazard ratio; 95% CI, 95% confidence interval.

^aTo note, the sFlt-1 assay used differs from the other studies (Supplemental Table 1).

^bHR for adverse outcomes, including all-cause death, cardiac transplantation, or ventricular assist device placement.

^cIn combination with high IL-6 levels (>7 pg/ml).

^dEstimated median value extracted from the study.

However, in 301 patients with stage 2–4 CKD, Rambod et al.⁴⁵ did not find any association between sFlt-1 and cardiovascular disease (CVD) but circulating sFlt-1 levels above a mean value of approximately 69 pg/ml were associated with a >3-fold higher mortality risk (Tables 3 and 4; Supplemental Table 1). In patients undergoing hemodialysis, sFlt-1 was not directly linked to CVD either, but it correlated positively with inflammation.^46,49 In two cohorts of 185 and 211 patients with stage 5–5D CKD,^46,49 sFlt-1 levels associated with C-reactive proteins, leukocyte counts, and IL-6, well-known inflammatory markers, and with soluble intercellular adhesion molecule-1 and sVCAM.⁴⁶ Guo et al.⁴⁹ showed that increased sFlt-1 was an independent risk factor for all-cause mortality. In contrast, Yuan et al.⁴⁶ observed that only in combination with elevated IL-6 levels could high sFlt-1 levels enhance the risk of death in these patients.

Soluble Flt-1 is moderately increased in most patients with CKD, but heparin-based dialysis could aggravate this scenario. In a cohort of 48 patients with stage 5D CKD, Lavainne et al.⁵⁰ showed that dialysis sessions were accompanied by a marked increase in sFlt-1 levels, which peaked 15 minutes after the start and returned to baseline after 4 hours, with a mean peak concentration during dialysis of 2551 pg/ml. This 25-fold increase was associated with heparin use and was independent of the dialysis modality. Accordingly, similar heparin effects were observed in patients with CAD undergoing coronary angiography or percutaneous coronary interventions. In those patients, the increase in circulating sFlt-1 levels (50-fold) was accompanied by a reduction in VEGF levels, indicating a shift to an antiangiogenic state after heparin administration.⁵¹

The clinical consequences of increased circulating sFlt-1 during heparin-based dialysis remain unclear. Due to the remarkably high sFlt-1 levels reached during dialysis and to the repeated nature of the insult (thrice weekly for several months or years), it is plausible sFlt-1 may exacerbate the antiangiogenic state in these patients and aggravate pre-existing endothelial dysfunction, CVD, and risk of death.⁵⁰

However, in a cohort of 329 patients with various degrees of renal dysfunction undergoing cardiac catheterization or coronary angiography, Onoue et al.³⁸ showed that high circulating sFlt-1 levels are more beneficial than harmful, given a reduction in sFlt-1 levels was associated with the worsening of atherosclerosis that accompanies renal dysfunction. Moreover, a positive correlation was found between plasma sFlt-1 levels and the eGFR (Tables 2 and 3). This divergent finding can be explained by the fact that sFlt-1 was measured in plasma obtained from the aorta after heparinization. In the studies mentioned above, samples were collected in the absence of heparin. Interestingly, in a cohort of 291 patients with stage 3–5 CKD, Matsui et al.³⁹ observed that circulating sFlt-1 levels and eGFR were negatively correlated before, but positively correlated after, heparin injection. Thus, heparin loading induced sFlt-1 release in healthy subjects and patients with CKD, but the final postheparin sFlt-1 concentrations levels were lower in patients with CKD. At baseline conditions (preheparin), the authors assumed patients had elevated circulating sFlt-1 levels but reduced amounts of stored sFlt-1. As postheparin sFlt-1 levels represent the total amount of sFlt-1, their results suggest the production of sFlt-1 was lower in patients with CKD than in control subjects.

Moreover, they showed the preheparin PlGF/sFlt-1 ratio (high sFlt-1 levels) was not relevant to CVD incidence. In contrast, the postheparin ratio (low sFlt-1 levels) was a strong predictor of CVD incidence in the studied population. This reduction in sFlt-1 may reflect an impaired capacity to antagonize excess VEGF/PlGF, creating a proatherosclerotic milieu in patients with CKD and suggestive of a protective effect of sFlt-1.^38,52

CKD is recognized as a decisive risk factor for CVD, and it accelerates the risk of atherosclerosis beyond that predicted by traditional risk factors.⁵³ The possible adverse effects of elevated sFlt-1 in the microvasculature of the heart and kidneys of uremic patients are supported by several experimental studies. The same, however, is valid for the possible protective effects of sFlt-1 in atherosclerosis (Figure 2). Rather than contradictory, these findings support the diversity of actions of the VEGF system in kidney disease, as described above.²¹

Figure 2.

Overview of possible mechanisms and effects of sFlt-1 on the kidneys (solid arrows) and the cardiovascular system (dashed arrows). Soluble Flt-1 may be harmful in renal pathologies such as CKD, renal ischemia-reperfusion, and preeclampsia, yet it may be protective in diabetic nephropathy and CKD-related atherosclerosis. VEGF and PlGF exert potent angiogenic factors through VEGFR2 and 1, respectively. Besides, PlGF regulates the expression of monocyte chemotactic protein-1 (MCP-1) and monocyte/macrophage mobilization. VEGF induces endothelium-derived nitric oxide (NO), an important regulator of endothelial homeostasis. By binding VEGF and PlGF, sFlt-1 regulates both factors' functions, thereby limiting monocyte/macrophage infiltration in response to elevated PlGF,^39,42 and controlling neovascularization in atherosclerosis. By reducing macrophage infiltration in response to increased VEGF in the diabetic kidneys, sFlt-1 can also protect against kidney damage in the settings of diabetic nephropathy.³² On the contrary, pathologic inhibition of the VEGF-signaling in response to high sFlt-1 may cause endothelial dysfunction,¹⁶ endothelial damage,²⁶ and lead to capillary rarefaction in the heart and loss of PTC in the kidneys,^27–29 for example, in CKD. Capillary rarefaction increases the risk of fibrosis and, consequently, loss of function. Soluble Flt-1 also increases endothelin-1 (ET-1) expression, an important hemodynamic factor.⁵⁶ However, whether ET-1 increases in response to VEGF inhibition itself or indirectly to activation/dysfunction of endothelial cells remains unclear (dotted arrows). Soluble Flt-1 may sensitize endothelial cells to inflammatory cytokines and angiotensin II, whereas pre-existing endothelial dysfunction may exacerbate sFlt-1 associated effects.^17,54,55

Furthermore, a possible synergistic effect between sFlt-1 and other vascular factors and conditions that normally go along with the decline of kidney function (e.g., angiotensin II, inflammatory cytokines, endothelin-1, and pre-existing endothelial dysfunction) cannot be excluded.^17,54–56

AKI and Renal Transplantation

In the Translational Research Investigating Biomarker Endpoints for Acute Kidney Injury cohort involving 1444 patients, Mansour et al.⁵⁷ reported an 8- and a 2-fold increase in circulating sFlt-1 levels on day 1 and day 2 after cardiac surgery (coronary artery bypass grafting or valve surgery) in comparison with preoperative levels. However, circulating sFlt-1 levels were higher in participants who developed adverse outcomes, including AKI, long AKI duration (≥7 days) (odds ratio, 1.75; 95% confidence interval, 1.09 to 2.82), and all-cause 1-year mortality (Table 5). In contrast, VEGF concentrations decreased on average after surgery and a more accentuated decline was observed among patients with higher odds of adverse outcomes.

Table 5.

Association of sFlt-1 with kidney outcomes and all-cause mortality in patients with cardiac surgery associated-AKI and kidney transplantation

Study	Changes in sFlt-1/Cutoff	AKI/DGF	1 year-eGFR	All-cause Mortality
Mansour, 2019	1-unit increase in Ln(log)a	OR 1.56 (95% CI, 1.31 to 1.87)	…	OR 2.28 (95% CI, 1.61 to 3.22)b
Chapal, 2013	250 pg/ml	OR 2.46 (95% CI, 1.5 to 6.09)	…	…
Wewers, 2019	1-unit increasec	OR 1.005 (95% CI, 1.00 to 1.01)	B -0.054 (-0.09 to -0.02)d	HR 1.004 (95% CI, 1.00 to 1.01)
	250 pg/ml	…	…	HR 5.804 (95% CI, 1.17 to 28.77)e

DGF, delayed graft function; HR, hazard ratio; 95% CI, 95% confidence interval; OR, odds ratio.

^aOne-unit increase of natural log-transformed sFlt-1 concentration.

^bOR for 1-year all-cause death.

^cOne-unit increase in sFlt-1 concentrations (not transformed).

^dLinear predictor (B) with (95% CI); correlation coefficient = -0.275.

^eHR for 2-year all-cause death.

In a study cohort consisting of 136 patients who had undergone a kidney transplant (KTx) Chapal et al.⁵⁸ described a >2-fold increase in circulating sFlt-1 levels throughout the first week after the KTx. A peak level ≥250 pg/ml was associated with an increased risk of delayed graft function, a manifestation of AKI that occurs in the first week of KTx, and impaired endothelial repair as indicated by a more pronounced PTC loss (mean 27.1%; range, 8.7–45.7 versus 18.9; range, 0.8–30.4) within the first 3 months after the KTx (Table 5). Donor age and the type of donor appeared to be significantly associated with maximal sFlt-1 levels in these patients.

In a longitudinal study with 93 patients who underwent a KTx, our group validated and extended the findings mentioned above.²⁹ sFlt-1 levels increased during the first week after transplantation, similar to the kinetics described by Mansour et al.⁵⁷ in patients experiencing AKI after cardiac surgery. After a sharp sFlt-1 peak on day 1 (5-fold increase), VEGF levels decreased slightly throughout the week. Also, high circulating sFlt-1 levels correlated with (r=−0.539) and predicted PTC loss within the first 4 months after a KTx (linear predictor B of −0.08; 95% confidence interval, −0.15 to −0.008) and were positively associated with delayed graft function. Furthermore, by using the same cutoff point of 250 pg/ml employed by Chapal et al.,⁵⁸ patients with higher sFlt-1 levels presented with an increased risk for graft rejection (hazard ratio, 3.54; 95% confidence interval, 1.64 to 7.66) over a 2-year follow-up period. We did not observe any significant association between circulating sFlt-1 levels and CVD during this time, but elevated sFlt-1 was linked to an increased risk of all-cause mortality. sFlt-1 levels during the first week after transplantation also predicted the impairment of 1-year graft function (Table 5). Patients in the higher sFlt-1 group were further characterized by increased proteinuria and decreased eGFR. Markers of endothelial dysfunction/activation, such as syndecan-1 and soluble vascular cell adhesion molecule 1, appeared to independently determine sFlt-1 in our study population.

Elevations in circulating sFlt-1 were found in patients with and without AKI (including delayed graft function), although the levels are of much greater magnitude in patients with AKI. These elevations were observed within hours after surgery, suggesting that sFlt-1 may vary according to the severity of the initial insult and contribute to disease progression. As mentioned above, VEGF is an essential factor for microvascular maintenance in the kidney.^10,21 Increased sFlt-1 levels within the first days after cardiac surgery or transplantation may favor a detrimental antiangiogenic state in the crucial initial phase of AKI/KTx-related ischemia-reperfusion, leading to PTC loss and impaired renal function.^22,59,60

These findings are limited to patients undergoing cardiac and KTx. Further studies are needed to assess whether these associations between sFlt-1 and poor outcomes exist in patients with AKI in various settings.

Primary Kidney Disorders Associated with CKD: Brief Remarks

In the studies mentioned above,^16,44,45 no differences in circulating sFlt-1 levels were observed as a function of hypertension status, although (positive and negative) alterations in sFlt-1 levels were reported in patients with uncomplicated hypertension.^61,62 Contrasting circulating sFlt-1 levels were also found in patients with diabetes.^63–65 In our study of 130 patients with CKD, univariate analysis showed a positive association between sFlt-1 and the presence of diabetes. However, a multivariate analysis showed only the influence of renal impairment and endothelial dysfunction on sFlt-1 levels.¹⁶

Circulating sFlt-1 levels were elevated in patients with proteinase 3- and myeloperoxidase–ANCA-associated vasculitis (n=20 and 10, respectively) during the acute phase of the disease in comparison with healthy subjects, but this decreased significantly in patients in the remission phase (n=20 and 13, respectively).¹⁵ Increased sFlt-1 during the acute phase of the disease may affect the renal microvasculature and contribute to the initial endothelial insult that often progresses to chronic vascular damage and irreversible organ damage, thereby leading to CKD. To note, 79% of patients had renal involvement, and sFlt-1 correlated positively with the degree of proteinuria in these patients.¹⁵ In 205 patients with IgA nephropathy, circulating sFlt-1 levels were also elevated compared with healthy volunteers and correlated positively with proteinuria, hypertension, and increased vWF levels.⁶⁶

Contrasting circulating sFlt-1 levels may reflect the limitations of the research design itself (e.g., a cross-sectional study), the different samples (e.g., serum or plasma) and assays employed, etc. Renal sFlt-1 expression has been found in biopsies of patients with CKD, including those with hypertension, diabetes, ANCA, and IgA nephropathy.⁶⁷

Sources of sFlt-1 in Kidney Disease

There are several pathways able to activate sFlt-1 expression.^68–75 Nonetheless, most studies have used placental tissues or cells as in vitro models of sFlt-1 release. However, in the context of kidney disease beyond preeclampsia, endothelial cells and monocytes should be more relevant.⁷⁶

Our group showed that CKD increases sFlt-1 expression in monocytes. Gene and protein sFlt-1 expressions were higher in peripheral blood monocytes isolated from patients with stage 2–5D CKD than those isolated from healthy subjects.¹⁶ Similarly, monocytes purified from KTx recipients secreted higher amounts of sFlt-1 in vitro than monocytes isolated from healthy donors.⁵⁸ The authors suggested the increased sFlt-1 production could be associated with activation of monocytes/macrophages in these patients.⁵⁸ Accordingly, stimulation of healthy monocytes with complement cascade products or monoclonal antibodies against ANCA autoantigens significantly increased sFlt-1 release.^15,77 Furthermore, serum drawn 1 hour after the start of a dialysis session induced sFlt-1 mRNA synthesis and late release (after 16 h) by purified monocytes, suggesting heparin-based dialysis contributes to the maintenance of the low-grade chronic increase of sFlt-1 in these patients. However, in patients undergoing dialysis, heparin is the main trigger of sFlt-1.⁵¹ As heparin is a well-known inducer of sFlt-1 release from endothelial cells, smooth muscle cells, and the placenta, endothelial cells/arterial walls are likely primary sources of sFlt-1 during dialysis.^5,39,50,51 To note, monocytes do not respond to heparin stimulation in vitro.⁵⁰

Interestingly, Zsengellér et al.⁶⁷ reported on sFlt-1 expression by infiltrating histiocytes/macrophages in the kidneys of patients with CKD. Of 52 analyzed renal biopsies, 37 were positive for sFlt-1, and sFlt-1 expression was positively associated with the degree of inflammation. Interestingly, sFlt-1 expressing histiocytes were found next to areas of PTC loss. The authors suggested that sFlt-1 positive macrophages recruited into the kidneys during injury may promote a localized antiangiogenic milieu, leading to capillary repair inhibition.⁶⁷ Aside from monocytes/macrophages, renal epithelial cells may also account for local sFlt-1 production. Immunostaining of renal samples from patients with noncardiac surgery AKI evidenced a significant increase in tubular sFlt-1 expression compared with control samples (non-AKI tissue).⁵⁷ Curiously, angiotensin II stimulation increases sFlt-1 production in renal proximal tubule cells in vitro through a mechanism involving AT₁ receptor activation⁶³; and overexpression of sFlt1 induces angiotensin II sensitivity, at least in pregnant mice.⁵⁴

Whether renal sFlt-1 expression would contribute to elevated blood and urine levels is unclear. Yet, in a population of 126 patients with varying renal function, Onoue et al. showed that plasma levels of sFlt-1 from the aorta were positively correlated with plasma levels of sFlt-1 from the renal vein (Pearson's correlation coefficient, 0.70), with higher levels (>2-fold) in the renal vein than in the aorta, thereby suggesting the kidney is a significant source of circulating sFlt-1.³⁸

Clinical Practice and Therapy

The clinical and experimental observations described above raise the question of whether an intervention to reduce circulating sFlt-1 levels would improve renal and cardiovascular outcomes in patients with kidney disease.

Extracorporeal removal of excess circulating sFlt-1 could be employed; although, any dialysis membrane cannot filter out sFlt-1 (85–120 kDa). However, due to its electrostatic surface potential, sFlt-1 can be adsorbed by dextran sulfate apheresis columns.⁷⁸ Therapeutic apheresis applied to 11 pregnant women with preterm preeclampsia reduced circulating sFlt-1 and proteinuria without apparent, serious, adverse maternal or fetal effects.⁷⁸ Besides, because heparin is the main trigger of sFlt-1 release during dialysis, minimization of its use by employing alternatives to heparin or heparin-sparing dialysis modalities would be an option for controlling sFlt-1 levels in these patients.⁵⁰

Before discussing whether sFlt-1 has potential as a therapeutic target in kidney diseases other than preeclampsia, several questions should be answered. The mechanistic aspects of how sFlt-1 increases in kidney disease and the standardization of laboratory testing are missing. Moreover, any strategy to control sFlt-1 should consider the differences between its local retention and systemic release, and the importance in maintaining a delicate physiologic threshold for VEGF required to prevent its untoward effects.

The compelling observational data that link sFlt-1 to cardiorenal interactions have instigated mechanistic studies to understand whether sFlt-1 is purely a biomarker depicting a complex array of metabolic disturbances or if changes in circulating sFlt-1 may be harmful on their own. Its effects likely reflect the different roles of the VEGF system in kidney disease and its complications. Still lacking is the determination of the stoichiometric relationship between sFlt-1 and VEGF, conclusive studies exploring sFlt-1 functions independently of VEGF and its relationship with inflammation, and confirming the diagnostic and prognostic values of sFlt-1 in kidney disease. We look forward to further progress in understanding sFlt-1 in both health and disease.

Disclosures

H. Pavenstädt reports receiving research funding from Sanofi; and reports being a scientific advisor or member of Deutsche Forschungsgemeinschaft. All remaining authors have nothing to disclose.

Funding

This work is supported by a research grant from the Deutsche Forschungsgemeinschaft (SE2824/3-1).

Acknowledgments

Dr. G.S. Di Marco, Ms. A. Schulz, and Dr. T. Wewers reviewed the literature, conceived the report, and drafted the manuscript; Prof. I. Nolte, Prof. H. Pavenstädt and Prof. M. Brand provided supervision and critically revised the manuscript for important intellectual content; and all authors approved the final version of the manuscript.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2020111579/-/DCSupplemental.

Literature review methodology.

sFlt-1 assays and measurement properties.

Supplemental Table 1. Assays and biological samples employed.