Diclofenac toxicity in susceptible bird species results from a combination of reduced glomerular filtration and plasma flow with subsequent renal tubular necrosis

Introduction

Old world vultures are purely scavengers reliant on carrion for their nutritional needs. For this adaptation, they are large sized for optimal soaring and energy expenditure; have excellent eyesight; and a high capacity haemoglobin function (Wink, 1995; Rodrigues, 2007). A notable adaptation is their large oesophageal diverticulum known as a crop. The crop represents a temporary storage area with capacity circa 20% of body weight. The crop allows for single large meal consumption and thereafter slow release to the proventriculus, longer than other raptors (Kierończyk et al., 2016). This feeding behavior, unfortunately, does predispose the bird to toxins in the food, as evident in Asia. Vultures on the Asian subcontinent have in the recent past been brought to near extinction through their inadvertent exposure to diclofenac in their food (Oaks et al., 2004; Swan et al., 2006a), with a resultant overall population decline of 99%. Species most effected were the white-rumped, long-billed and slender-billed vulture (Gyps bengalensis, G. indicus and G. tenuirostris) (IUCN, 2017).

Modelling of exposure suggests that toxicity resulted from an intake of 1 kg of meat contaminated with diclofenac equating to a dose of 0.8 mg/kg (Green et al., 2004). Clinical signs seen were a depressed habitus, inappetence and drooping of the head at 24 h post exposure, and death after 48 h. Pathological changes were marked hyperuricaemia and severe visceral gout (Meteyer et al., 2005). Since its description in the Asian vulture species, sensitivity has been confirmed in the Cape vulture (G coprotheres), African white backed vulture (AWBV) (G. africanus), Eurasian griffon vulture (G. fulvus) and the Stepp eagle (Aquila nipalensis) (Swan et al., 2006a; Rattner et al., 2008; Naidoo et al., 2009; Sharma et al., 2014), but surprisingly not the Turkey vulture (Cathartes aura). In an unexpected manner, the chicken was also susceptible to the effects of the drug, albeit at a higher LD₅₀ (10 mg/kg) (Naidoo et al., 2007). Since diclofenac was shown to be unsafe, a number of other NSAIDs, except meloxicam, have also tested as unsafe (Swan et al., 2006b).

While the pathological changes associated with diclofenac are well documented, the mechanism and apparent sensitivity of old-world vulture species remains unknown. Being predominantly carrion eaters, their natural exposure to higher uric acid from both the catabolism of purines and being uricotelic is expected to be contributory (Long & Skadhauge, 1983; Berger, Yü & Gutman, 1960; Dudas et al., 2005). It has, however, been speculated that diclofenac interferes with active excretion of uric acid at the proximal convoluted tubules (PCT) (Naidoo & Swan, 2009), or by interfering with blood supply to the cranial lobe (Meteyer et al., 2005). An ex vivo study with limited vulture tissue showed inhibition of para-aminohippuric acid (PAH) transport, indicative of organic anionic (OAT) channel inhibition (Naidoo & Swan, 2009). This is not unexpected as NSAIDs, including diclofenac, are OAT inhibitory in people (Sekine, Cha & Endou, 2000).

The aim of this study was to further expand the knowledge on the pathophysiology of diclofenac toxicity in vultures. To achieve this, the influence of diclofenac on uric acid excretion, glomerular filtration and plasma flow was evaluated in a domestic chicken model, previously validated as a suitable model for diclofenac toxicity studies (Naidoo et al., 2007). In this previous study, it was demonstrated that the time leading to toxic effects, clinical signs and pathological changes seen with diclofenac are identical between the chicken and the numerous vulture species (Naidoo et al., 2007). The other reason for using a surrogate species is the highly endangered status of vulture species (IUCN, 2017). More specifically, for this study chickens were exposed to diclofenac in the presence of PAH and iohexol (IOH) (Omnipaque) as per the model of Laroute et al. (2005). The former is a very specific transport substrate of the OAT channel while omnipaque is a non-radioactive radiographic marker of renal blood flow and glomerular filtration. Further, Smeets et al. (2004) showed that the Multiple Drug Resistance Protein (MRP) were not involved in PAH excretion in rat kidney slice cultures (Smeets et al., 2004). In combination, these two substances have been previously used in the dog as a means to non-lethally evaluate changes in renal functionality at the level of the glomerulus and the renal tubular basal transporters (Laroute et al., 2005).

Materials and methods

Results

Plasma and faecal uric acid, PAH and iohexol clearance

No changes in plasma uric acid concentrations were noted in phase 1 for any time point monitored, including before and after PAH administered, which indicated unconstrained excretory capacity. After the administration of diclofenac, plasma uric acid concentrations increased rapidly over a period of time, even though the concentrations at the start of each phase were within the normal ranges (0.217–0.403 mmol/l). The change in plasma uric acid for phase 2 was evident as early as 2 h post exposure, and for most chickens had peaked by 8 h (Times based on PAH/IOH administration, which was 2 h after diclofenac administration) (Fig. 1). The chickens (bird 9192 and 9194) that died 24 h after dosing had very high levels of plasma uric acid, which was more than 10 times the basal concentration as well as a more prolonged increase in uric acid concentrations (Fig. 1, Table 3). Concurrent to the plasma uric acid increase, faecal uric acid concentrations of the treated chickens and plasma uric acid clearance decreased, compared to their untreated values, with four diclofenac treated chickens showing up to a 99% decrease (Table 3), i.e., the increase in plasma uric acid occurred concurrently with a decrease in the uric acid clearance. Of the four chickens in which the excretion of uric acid had decreased, two had died. For these two chickens the uric acid faecal concentrations were reduced to almost 0. While Tm_uricacid had decreased for all chickens, it was virtually absent in three chickens including the two that died (Table 3). As for diclofenac concentrations, there was poor correlation between plasma and faecal changes. The area under the curve last (AUC_last) showed that the chickens that died had, at minimum, a two-fold higher total exposure to uric acid than the chickens that survived and a ten-fold increase to their healthy state. The cumulative result thus showed a strong relationship between death and absence of uric acid excretion, mainly at the level of the tubules.

For phase 1, PAH was not detectable in the plasma for two chickens (9190 and 9191) and for only two time points for the remaining four (9189, 9192, 9194, and 9195). As a result, it is concluded that PAH is rapidly cleared under normal conditions (Fig. 1). The clearance was estimated as presented in Table 3, using the LOQ as the first point value. The results were also highly variable in phase 2 between the chickens. The PAH clearance decreased from phase 1 to 2 with an average clearance of 32.34 and 8.99 L/h*kg respectively (Table 3). For five of the treated chickens, diclofenac induced an increase in the total exposure to PAH, as seen by the AUC_last (Table 3). The increase in PAH plasma concentrations was associated with a decrease in plasma clearance by 71.60 ± 16.54% (Table 3). For the two chickens that succumbed to the toxicity, PAH clearance decreased by 98% (Table 3). The plasma concentrations for IOH were more quantifiable (Fig. 1). The same trend was evident for IOH, however, with clearance decreased by up to 87%. The highest inhibition in the clearance of IOH was present in both chickens that died and for one that survived (Table 3). The tubular secretion of PAH (Tm_PAH) was reduced by more than 90% for all the chickens except one bird which was reduced by 45%. Furthermore, it was also interesting to note that with the chickens that died the Tm_PAH was reduced by more than 98% (Table 3). In all cases the changes evident in phase 2 of the study were significant.

Discussion

Diclofenac toxicity in the vultures and chickens is characterized by major increases in plasma uric acid and subsequent gout (Oaks et al., 2004; Meteyer et al., 2005; Reddy et al., 2006; Naidoo et al., 2007; Jain et al., 2009). From mammalian physiology it is known that diclofenac has the ability to inhibit both the functions of the OAT channel, reduce blood supply to the kidney and alter glomerulus filtration (Khamdang et al., 2002; Anzai, Kanai & Endou, 2006; Katzung, Masters & Trevor, 2009). However, unlike in mammals where glomerular filtration of uric acid is very efficient, birds excrete more than 80% of their total uric acid using the OATs channels (Berger, Yü & Gutman, 1960). This aligns with the findings of this study during which five chickens were excreting more than 85% of their total uric acid using former channels, except for one chicken that was somewhat lower at 64.24%. Furthermore, for the efficient plasma clearance of uric acid by the OAT channels, adequate blood supply needs to be maintained to the renal tubules (Khamdang et al., 2002; Anzai, Kanai & Endou, 2006; Katzung, Masters & Trevor, 2009). Previous studies have shown that the glomerulus received only arterial blood supply, while the avian renal tubules received an equal admixture of venous and arterial blood as ascertained through PAH clearance studies (Shideman et al., 1981). With diclofenac known to inhibit both tubular excretion and blood supply, both these mechanisms could explain the diclofenac induced increase evident in the chicken and vulture. Specifically, for this study we investigated the pathophysiology of diclofenac’s toxicity using IOH and PAH as marker substances. PAH is a substrate for OAT channels which are located on the basolateral membrane of the renal proximal tubules and IOH is a marker for renal blood flow and glomerulus filtration (Laroute et al., 2005; Sweet, 2005; Sweet, 2010; Van Wert, Gionfriddo & Sweet, 2010).

For this study we intentionally chose high doses as we wanted to follow the lethal changes following exposure to diclofenac. As expected, following the administration of 10 mg/kg diclofenac as a single dose, mortality rate of 33% was recorded within 48 h, which were similar to previous mortalities of 40% evident in white leghorn after 12 days of exposure to 10 mg/kg of diclofenac (Jain et al., 2009; Reddy et al., 2006) and the LD₅₀ of 10 mg/kg reported after a single dose exposure in layer hens, and confirmed in Ross broilers of 4 weeks age (Naidoo et al., 2007; Locke, 2020). Furthermore, histopathology changes observed in the two domestic chickens that died following diclofenac exposure, were similar to those reported in the Leghorn and Gyps species (Oaks et al., 2004; Meteyer et al., 2005; Swan et al., 2006b; Jain et al., 2009; Naidoo et al., 2007). Likewise, the changes in uric acid concentration in these chickens were identical to that previously reported in vultures. It was also interesting to note that the one chicken which had mild urate induced renal necrosis did not succumb to toxicity within the 48 h monitoring period. This would imply that the difference observed in the former and other surviving chickens is subject to a degree of intra-subject variation which may be attributable to genetic, environmental and epigenetic factors (Fusco & Minelli, 2010; Frésard et al., 2013).

The half-life of elimination of diclofenac in the chickens was in all cases substantially shorter than previously reported in vultures (<2 h vs 14 h) (Naidoo & Swan, 2009). This was, however, very similar to the half-life of one hr reported in layer hens, albeit at a lower dose of 0.8 mg/kg (Naidoo et al., 2007). This would support previous reports that the unusual susceptibility of the vulture is related to metabolic constraints in the species (Hutchinson et al., 2014). Furthermore, in liver slice studies by Adawaren et al. (2018), the total protein content and thus cytochrome content, was deemed to be lower than in other avian species (Adawaren et al., 2018). Despite the short plasma half-life, the faecal half-life of elimination was in general longer. At present the exact reason for this is unknown, but we would speculate that the birds would have a degree of biliary retention as a result of the stress of being inducted into the study and single housed for the 24 h duration of the pharmacokinetic study (Harrison & Gibaldi, 1976). Also important to note, is that toxicity resulted a substantial time after diclofenac was no longer detectable in the plasma. This would likely indicate that the drug induced early cellular death, and that the changes in uric acid were likely a secondary outcome of toxicity. When comparing the chicken to the vulture, the maximum diclofenac plasma concentration for the vultures that died was 7.41 and 4.6 ug/ml respectively (Naidoo et al., 2009), which is above the toxic concentration of 4 µg/ml estimated for the chicken. This was evident despite the vulture being exposed to a dose of only 0.8 mg/kg as opposed to the 10 mg/kg in the chicken. Based on the difference in doses between the species, this would once again support metabolic constraint in the vulture, promoting plasma concentration that would inhibit uric acid excretion at a significantly lower dose.

The IOH clearance of 0.17 L/h*kg (2.83 ml/min*kg) for phase 1 of the study was similar to the IOH clearance of 2.94 ml/min*kg reported by Dhondt et al. (2019), which support the results from phase 1 being an accurate reflection of normal IOH clearance in healthy chickens. Following the administration of diclofenac, IOH clearance decreased in phase 2 of the study, indicating a decrease in glomerular filtration as IOH is purely dependent on glomerular filtration for its clearance. This finding was also in general supported by the radiographic analysis, which is a visual determination of glomerular filtration. This change in glomerular filtration rate (GFR) for chickens in phase 2, also indicates that diclofenac consistently decreased the blood supply to the kidneys in all chickens, and thus alters the rate of delivery of uric acid to the excretion points. This finding is not unusual as elderly people on oral doses of diclofenac (50 mg three times a day), were reported to have impaired renal blood flow and glomerular filtration rates leading to renal failure (Katzung, Masters & Trevor, 2009).This effect results from the suppression of the synthesis of prostaglandin through the non-specific inhibition of the cyclo-oxygenase enzyme in animals. Prostaglandins play a physiologic role in maintaining GFR. Prostaglandins predominately function to regulate renal blood flow primarily supplied via the afferent arterioles and as a renal vasodilator (Verlander, 1997; Oates et al., 1998). Using chicken venous tissue, Naidoo et al. (2007) were able to demonstrate that diclofenac was also a venodialatory substance. Based on the similarity between human and chicken COX2 enzymes, it is likely that COX inhibition drives this process in the chicken (Botting, 2006). It was also interesting to note that, due to the higher plasma concentration and despite the lower GFR, the predicted total uric acid filtered was higher than in phase 1 for some of the chickens.

For the four chickens in which PAH clearance was properly quantifiable in phase 1, their PAH clearance of 143.8 ml/min*kg compared well with Dhondt et al. (2019) who described PAH clearance of 117.5 ml/min*kg. The change for PAH noted in phase 2 was, however, not constant in the treated chickens and was very markedly changed in the two chickens that died with the decreased PAH clearance evident before 12 h, well-before the birds started showing clinical signs of toxicity at 24 h post-exposure. In terms of the general effect of diclofenac on PAH clearance, the finding was not surprising as both sodium salicylate and phenylbutazone have previously shown to inhibit PAH clearance in the chicken by 90 and 50% respectively, albeit at much higher doses (Berger, Yü & Gutman, 1960). The necropsy of these dead chickens further revealed urate deposits on kidney and other organs. Urate deposits were associated with very high plasma uric acid concentration which had multiplied more than 10-fold from its original concentration, reaching maximum levels of 6.196 mmol/l from 0.250 mmol/l in one chicken. With IOH clearance being constantly decreased in all the chickens, and PAH clearance being almost insignificant in only the two chickens that died (Tm_PAH and Cl_UA both reduced by more than 98%, furthermore with Tm_uricacid being absent), it appears that tubular excretion of uric acid was affected. In a study by Sun et al. (2021), using molecular techniques in the chicken cells, they were able to demonstrate that diclofenac likely induces kidney cell apoptosis, through interfering with purine metabolism and inhibiting the expression of OAT2. These factors in combination were speculated to induce increase the uric acid concentration in chickens leading to hyperuricemia and urate deposition.

Based on the overall evidence obtained from this study, it offers insight into the mechanism of toxicity. Firstly, with the short half-life of diclofenac, it is unlikely that the drug had a primary inhibitory effect on the functionality of the uric acid transporter system to such a large extent, as we initially speculated in our hypothesis. However, delayed effect of uric acid excretion has been described in humans, where the uricosuric effect of diclofenac was still evident a day after a single treatment despite the half-life of elimination circa 2 h, the effect on uric acid seen in people was not as marked as in this study (Davies & Anderson, 1997; Tiitinen et al., 1983). This would indicate that acute cellular (renal) toxicity and apoptosis is the likely underlying mechanism of toxicity, as is evident from the cellular death shown by the necropsy on the two dead birds. From previous published work this has been shown to result from reactive oxygen species (ROS) induced cellular death, 2 h after exposure under in vitro conditions (Naidoo & Swan, 2009), the same time period required for the PAH increase evident in this study together with reduced OAT2 expression (Sun et al., 2021). However, since only two birds died, a degree of individual variability in protection against ROS is likely to be present. To further understand this mechanism, further studies can focus on direct markers of cellular death such as gamma-glutamyl transferase (GGT), pro-inflammatory cytokines and creatinine (Mohan et al., 2012).

With decreased uric acid and PAH clearance being the predominant change following diclofenac exposure, the results from this study also offers some insight into the sensitivity of the vulture compared to the chicken. The uric clearance in the chicken was at minimum two-fold higher than the vulture despite the high protein diet of the vulture when corrected for body mass. In general, it is accepted that the main source of uric acid in protein diet is red meat especially beef, lamb, veal and pork (Teng et al., 2015). As is evident in the feeding study, after a protein meal plasma uric acid levels are raised for a fairly long period in the vulture. With tubular damage resulting in an increase in uric acid, it is not surprising that the vulture’s diet predisposes them to toxicity. However, the unexpected finding was the difference in the uric acid clearance between vultures and chickens. What makes this a surprising finding is that with the vulture, being a carnivore and naturally exposed to a high uric acid load, one would have expected a more efficient uric acid excretory system. This would then suggest that a secondary protection mechanism has to be present. Evaluating the digestive system, the crop would be the likely protective mechanism. As a structure the crop evolved to allow for large scale feeding (Zheng et al., 2011) and slow release of food in the birds which—in the African White- Backed Vulture - was established as a 28 h period (Kierończyk et al., 2016). This lower rate of release would create the buffer, akin to a valve, allowing for slower increases in plasma uric acid concentration which would not exceed renal excretory capacity. The lower ability of the kidney to excrete uric acid is also likely to be the pitfall of not needing to adapt to smaller more frequent meals like other species with rapid GIT transit (McWhorter, Caviedes-Vidal & Karasov, 2009). Unfortunately, the true effect of the full crop on plasma uric acid concentrations could not be measured as fully engorged birds regurgitate as a natural defense mechanism when handled.

Conclusion

Diclofenac administration, together with IOH and PAH in the chickens allowed for real-time evaluation of changes. In untreated chicken, no ill heath was noted but in treated chicken, within 48 h, 33% of the chickens died and were thereafter diagnosed with urate nephropathy. IOH clearance of the untreated birds was fast and most cleared within 6h. However, for the treated group uniform IOH clearance decreased by up to 87%, leading to a conclusion that diclofenac interferes with renal arterial blood flow with resultant decreased glomerular blood supply. PAH clearance of the untreated birds was rapid and non-detectable with up to 99.90% excreted while for the treated birds, diclofenac induced an increase in the total exposure to PAH. Death, however, only results when there is near 100% inhibition of uric acid excretion from total glomerular and tubular inhibition. It should also be noted that toxicity occurred in the relative absence of diclofenac in plasma, whilst the drug was present in vultures for over 14 h. It should be noted that the following study intended to only look at the lethal changes induced by the drug. It is possible that lower doses, in the absence of overt toxicity, may have induced more subtle changes in physiological functioning, which may require further elucidation.

Supplemental Information