Commonly used herbicides, including paraquat (PQ), glyphosate-surfactant herbicides (GPSHs) and 4-chloro-2-methylphenoxyacetic acid (MCPA), are responsible for many deaths and severe health effects particularly in developing countries where agriculture is the main source of employment. Nephrotoxicity is one of the main clinical effects of poisoning with these herbicides as these compounds are primarily excreted in urine via the kidneys. Acute kidney injury and irreversible chronic renal damage are observed in animals and patients with acute herbicide poisoning. Preserving kidney function in intoxicated patients may minimise toxicity of these herbicides.
The mortality rate of PQ is high as there is no specific and efficient treatment regimen. An ongoing cohort study on self-poisoning and a randomized controlled trial assessing the efficacy of immunosuppressive therapy on PQ poisoning were conducted in hospitalized paraquat-intoxicated patients. This cohort study showed there was no difference in survival rate between patients in placebo group and immunosuppression group. Nonetheless, acute kidney failure which is one of causes of death was not taken in to account. Consequently, one aim of this study was to characterize the toxicokinetics and toxicodynamics of paraquat in this population and to assess the effects of a candidate antidote on toxicodynamics.
A two-state approach toxicokinetic/toxicodynamic population analysis, by a non-linear mixed-effect (NLME) modeling approach using Phoenix NLME version 1.2, was performed in a subset of patients that had three or more consecutive blood samples. A two-compartment model was applied to fit the toxicokinetic data. Renal function, namely creatinine clearance, was the most significant variable to explain patient variability in paraquat clearance and volume distribution. Estimates of toxicokinetic parameters were 1.17 L/h, 2.4 L/kg, 87 h and 0.58 for the apparent clearance (CLPQ/F), the apparent volume of distribution (Vd/F), elimination half-life (t1/2 β) and bioavailability (F), respectively. This model suggested that rapid reduction in paraquat clearance occurred over 24 to 48 h, and afterwards the clearance was constant over time. Estimated concentration of PQ causing a 50% of maximum PQ induced renal function change was 429 µg/L. There was a trend towards improved renal function (an 8% improvement compared to placebo) with the application of an immunosuppressive regimen. The models may be useful as a prognostic tool to indicate patient status on admission. The low clearance and long half-life suggest further studies of extracorporeal elimination techniques, such as hemoperfusion, would be worthwhile.
Acute kidney injury was studied in terms of the pathological changes and biomarker patterns in a rat model after PQ, Roundup®; a commercial product of glyphosate surfactant based-herbicides; and Ospray® MCPA 500; a commercial product of MCPA; exposure at different time points and different dose levels. Currently, these novel acute kidney injury biomarkers are not accepted for use in human studies because although the markers have largely been validated in animals, there are no studies comparing animals and humans with the same biomarkers and same nephrotoxic agent. The early detection of acute kidney injury would be an important first step to enable the preservation of kidney function with herbicide toxicity. Thus far, there have been no studies using animal models with an extensive range of renal biomarkers to provide contrasting data for human studies.
Fourteen candidate biomarkers were examined in this study including cystatin C, kidney injury molecule-1, beta 2-microglobulin, clusterin, albumin, neutrophil gelatinase-associated lipocalin, osteopontin, vascular endothelial growth factor, tissue inhibitor of metalloproteinases, α-1-acid glycoprotein, calbidinD28, chemokine interferon-inducible protein 10, epidermal growth factor and α-glutathione S-transferase, to develop a non-invasive method to detect early renal damage and dysfunction and to compare with the conventional endogenous marker plasma creatinine. Furthermore, the Receiver Operating Characteristic (ROC) analyses of three methods were compared: absolute concentration, urinary biomarker normalizing to urine creatinine and biomarker excretion rate.
Male Wistar rats were dosed orally with 4 different doses of PQ or Roundup® or Ospray® MCPA 500 and the biomarker patterns in urine and plasma were investigated at 8, 24 and 48 hours after exposure. Biomarkers were quantified by absolute concentration; by normalising to urine creatinine; and by calculating the excretion rate. The ROC diagnostic performances of each method in predicting of acute kidney injury were compared.
Urinary kidney injury molecule-1, urinary albumin and urinary cystatin C elevations correlated with the degree of renal damage and injury development in rats dosed with PQ. Absolute concentration generally predicted acute kidney injury in the PQ-rat model better than normalised concentration and excretion rate. Using the absolute urinary biomarker concentrations, absolute urinary kidney molecule-1 showed the best performance (area under the curve (AUC)-ROC of 0.81 and 0.98 at 8 and 24 h) for prediction of histopathological renal injury (demonstrated by tubular necrosis) and outperformed plasma creatinine at these time points. At 8 h, the urinary kidney injury molecule-1 AUC-ROC had an area of 0.81 and the cut-off was > 326 pg/ml (sensitivity 79%, specificity 75%). Absolute urinary cystatin C and urinary albumin also performed equally well or better than plasma creatinine at 24 h. Plasma cystatin C was as good as plasma creatinine in reflecting renal function.
Absolute concentrations of plasma creatinine proved to be the best early biomarker in predicting histological changes (best cut-off point > 0.26 mg/dl, sensitivity 71%, specificity 70%) in MCPA-induced-renal toxicity. Plasma cystatin C, another surrogate of renal function, was comparable to plasma creatinine in its ability to predict GFR. All urinary biomarkers were generally less predictive, although normalization to urine creatinine greatly improved the performance of several urinary injury biomarkers. Normalised concentrations of osteopontin and albumin were best in predicting later histological changes as early as 8 h, with areas under the ROC curves of 0.95 and 0.93. Both absolute urinary kidney injury molecule-1 concentration and normalised concentration elevations also correlated well with the degree of renal damage. However, this partly reflects the fact that urinary biomarker concentrations normalised to creatinine greatly amplify the biomarker signal when there is declining glomerular filtration rate and steady state has not been reached.
In the Roundup® model, plasma creatinine levels significantly increased (1.7-fold time-matched controls) as early as 8 h. However, by ROC analysis of the selected biomarkers, urinary kidney injury molecule-1 best predicted histological changes at 8 h (cut-off point > 290 pg/ml, sensitivity 100%, specificity 58%). Plasma creatinine performed better than other biomarkers at 24 h (best cut-off point > 0.21 mg/dl). Using normalised biomarker concentrations normalised to urine creatinine levels or biomarker excretion rates did not improve the ROC performances of urinary biomarkers in predicting the development and establishment of acute kidney injury. Plasma cystatin C performed as well as plasma creatinine in indicating impaired kidney function in this model. Acute tubular and glomerular necrosis and apoptosis are likely to be the main cellular mechanisms of Roundup®-induced nephrotoxicity. To our knowledge, this present study is the first to demonstrate apoptosis induced by Roundup® in an in vivo model of nephrotoxicity.
In summary, urinary kidney injury molecule-1 outperformed other biomarkers in predicting histological changes, with areas under the ROC curves of 0.83 to 0.98 as early as 8 h in the PQ and Roundup® nephrotoxic models. The absolute urinary kidney injury molecule-1 levels greater than 290 pg/ml could indicate AKI development. Plasma creatinine was still the best marker for early detection of acute kidney injury within 8 h after MCPA dosing. Changes in urinary kidney injury molecule-1 levels also mirrored the effects of these herbicides on proximal tubule injury.
The performance of plasma cystatin C in mirroring renal function was similar to that of plasma creatinine; however, plasma cystatin C performed poorly in diagnosing acute kidney injury in this present study. Further study is required to compare changes in biomarkers with those seen in human poisoning. This translational research contributes significantly to the identification of more sensitive and specific markers for pre-clinical identifying of markers for predicting outcome after exposure to nephrotoxic agents and in assessing the extent to which animal acute kidney injury biomarker data can inform risk assessment in humans.