There are multiple challenges complicating the study of the epidemiology of BEN. Data on the incidence of BEN have been controversial. There are no clear-cut data on the current trend for the incidence and prevalence of BEN. The studies carried out in different endemic areas have produced conflicting information. Some epidemiological studies reported an increase in the prevalence of BEN between 1967 and 1970, a steady state between 1970 and 1984 and ultimately a decrease in some endemic areas [7, 8]. Similarly, in another endemic area, a decreasing incidence over time was found during a follow-up period spanning the time period 1978 to 1997.
Assessments are frequently based on the number of BEN patients undergoing haemodialysis treatment. In Serbia, BEN patients represent an average of 6.5% (5–46%) of the haemodialysis population. The high prevalence rate of haemodialysed patients with BEN indicates that BEN is not disappearing.
It has been estimated that almost 100 000 people are at risk of BEN, whereas 25 000 have the disease. Despite intermittent variations, the incidence of new cases has remained stable over time. These differences may also be related to changes in the study design or true epidemiological differences between sequential time frames and endemic areas and can also be attributed to the natural course of the disease.
The clinical signs and symptoms of BEN are non-specific and often remain unrecognized for years. The initial asymptomatic period is followed by weakness and lassitude, mild lumbar pain, pallor of the skin and a copper-brownish discoloration of the palms and soles. At this stage, which occurs at an older age [13, 14], anaemia is associated with a significant loss of renal function and indicates the presence of chronic kidney disease (CKD). Blood pressure is usually normal, although in the advanced phase it may be elevated but at a lower incidence than in other forms of CKD. Intermittent proteinuria of the tubular type may be found early while in the uraemic phase, it becomes permanent. The urinary sediment shows sparse white and red blood cells. In the early stages of BEN, the loss of urine concentration capacity precedes the decline in the glomerular filtration rate [15–17]. Serum proteins and immunoglobulins are unchanged.
Kidney imaging reveals a variable decrease in kidney size with very small contracted kidneys in the end stage. Extreme kidney atrophy has been proposed as a criterion for the clinical diagnosis of BEN. Some studies suggest that the smallest kidneys are found in advanced stages of BEN, whereas others have reported that this occurs in the earlier stages. In BEN patients, it has been found that the dimensions of the kidneys depend on the parental status and the age of offspring. In the offspring of a mother with BEN, ultrasound measurements of the kidney cortex thickness seem to portend a prognostic value. In conclusion, adult offspring in BEN families can be characterized by shorter kidney length and an increased excretion of albumin, total protein and β2-microglobulin, in particular, when the mother had BEN [19, 20].
The pathology of BEN is characterized by a progressive atrophy and sclerosis of all structures of the kidney, and it shares similarities with tubulointerstitial kidney diseases (Figure 2).
The consistent renal histological findings characterizing BEN are extensive hypocellular interstitial fibrosis associated with tubular atrophy involving medullary rays, the attending outer medulla and the cortical labyrinth, where it decreases typically from the outer to the inner cortex.
Glomerular and vascular lesions are associated with periglomerular fibrosis, glomerular lesions, including ischaemic, microcystic, obsolescent glomeruli, occasional thrombotic microangiopathy-like lesions and focal segmental sclerosis-like lesions. Vascular lesions include arteriolar hyalinosis, intimal fibrous hyperplasia, occasional mucoid arterial intimal fibrosis and multifocal thickening and splitting of peritubular capillary basement membranes detected by electron microscopy [21, 22].
There are no diagnostic features which could be marked as pathognomonic of BEN but the set of epidemiological, clinical and biochemical data along with the pattern of pathologic injury, in the absence of any other renal disease, is highly suggestive of this entity. Initially, the criteria proposed by Danilovic et al. were widely used. Over time, these criteria evolved and contributed to a better diagnosis of BEN [24–26].
Updated recommendations developed during the ‘International Workshop on the Diagnostic Criteria in BEN’, held in Brač, Croatia, in 2008, and at a meeting organized in 2012 (Skopje, FYR of Macedonia) aimed at providing recommendations for the screening, diagnosis and therapy of patients with BEN. The consensus statement is intended to assist scientists in speaking the same ‘diagnostic language’, thus enabling the comparison of results obtained in different countries.
In one of the first original papers on BEN published by Danilovic et al., lead was incriminated as the causative agent of BEN, having been found in the flour used for baking bread in the affected villages. However, this hypothesis was not substantiated by later studies.
Metals and metalloids
Several publications have suggested the possibility that a deficiency of some essential trace elements, e.g. selenium, might be involved in the aetiology of BEN. The concentrations and the extent of selenium deficiency are well documented in rocks, soil, water, food stuffs and in serum samples collected from endemic and non-endemic regions of Serbia. However, no association between selenium deficiency and a high incidence of BEN and UUC in endemic areas was confirmed. The findings of a systematic 2-year follow-up study indicated that metals (cadmium and lead) and metalloids (arsenic and selenium) do not play a role in the aetiology of BEN.
Chronic intoxication with Aristolochia clematitis
Aristolochia clematitis is one of the old healing plants already in use by the ancient Egyptians and Greeks (Figure 3). The Greek name ‘Aristolochia’ is a compound word made up of áristos—fair, beautiful, and lochia—menstrual periods. It refers to the use of the plant in aiding childbirth and was prescribed by the Greeks to women following delivery to facilitate a resuming of their periods. The species name, ‘Clematitis' derives from the Greek 'klema' for tendril, the growth form of this species of Aristolochia. The English name 'birthwort' likewise refers to the plant's use as an aid to birth. Although herbal drugs derived from Aristolochia spp. have been known since antiquity and were used in obstetrics and in the treatment of snake bites (Rosenmund and Reichstein, 1943) in 1981, the use of Aristolochia spp. was forbidden in many countries due to possible carcinogenic effects. Peters and Hedwall in 1962 indicated that AA was known to be nephrotoxic in the rabbit as early as in 1892, in the horse in 1893 and in the rabbit and the mouse in 1958.
In 1969, Ivić proposed that the aetiology of BEN could be related to chronic A. clematitis poisoning in which seeds from these plants, which are encountered abundantly in local wheat ﬁelds, intermingle with wheat grain during the harvesting process. He speculated that human exposure to a toxic component of Aristolochia might occur through ingestion of bread prepared from ﬂour derived from contaminated grain. He demonstrated in rabbit models that flour prepared from A. clematitis seeds induced a nephropathy, which resembled the findings in BEN. Ivić even proved the carcinogenetic potential of the plant as rats developed sarcomas of the skin at the site of injection of aqueous extracts of A. clematitis. Ivić concluded that on the basis of geographical, epidemiological and laboratory investigations, there are strong arguments suggesting that endemic nephropathy is essentially caused by chronic poisoning with the seeds of A. clematitis. Although his classic paper presented the brilliant examples of drawing correct conclusions regarding the aetiology of a complex human disease from simple but well thought out experimental observations, this clever hypothesis has never been adequately pursued. His well-documented results attracted more interest from the scientific community many years later.
The ever-growing interest in unravelling the mystery of BEN and the role of AA was prompted by reports of a high incidence of CKD that developed in a group of otherwise healthy Belgian women. In 1990, a clinic in Brussels began prescribing capsules as part of a slimming regimen consisting of Chinese herbal remedies believed to contain, in part, Stephania tetrandra. Unintentionally, S. tetrandra (in Mandarin Han Fang-Ji) was replaced by Aristolochia fangchi (Guang Fang-Ji in Mandarin) since both of the plants are used in Chinese traditional medicine carrying similar names (Fang-Ji) [37, 38]. The Oriental Materia Medica specifies that the Chinese prescribe A. fangchi, Aristolochia heterophylla, Cocculus trilobus or S. tetrandra indiscriminately, obviously not taking into account the possible presence of AA.
The observed nephropathy has been ascribed to the ingestion of Chinese herbal remedies that have included species of the genus Aristolochia which were positively classified as carcinogenic to humans (Group 1) by the International Agency for Research on Cancer (IARC). The outbreak of the so-called Chinese herb nephropathy (CHN) in Belgium in 1993 affected more than 100 patients, mostly women, half of them requiring renal replacement therapy [41, 42]. It was found that, in the majority of cases, progression to end-stage renal disease occurred despite discontinuation of Chinese herbal remedies, and invasive urothelial carcinoma after exposure to Chinese herbal remedies containing AAs may occur even without severe renal failure.
The CHN reported in Belgium in 1993 presented as a rapidly progressive renal interstitial fibrosis leading to end-stage renal disease. Within a few years, it emerged that CHN patients developed a high risk of UUC. Urothelial malignancy of the upper urinary tract developed in almost half of the patients [42, 45]. Later, it was found that even patients who did not display the characteristic histological features of CHN were also exposed to a high risk of UUC. The outbreaks of AA-associated renal failure have been subsequently reported in several other countries, and the name was replaced by aristolochic acid nephropathy (AAN) [47, 48].
Following these publications, some research groups conducted pioneering studies on the molecular mechanism of AA-induced carcinogenesis [49, 50]. Later, methods developed in the course of this research were applied to the identification of AA–DNA adducts in tissues of Belgian women with CHN.
Exposure of CHN patients to AA that belongs to the family of carcinogenic, mutagenic and nephrotoxic compounds was substantiated by the identification of AA–DNA adducts by the method of 32P-post-labelling in urothelial tissue of these patients [42, 52–54]. AAs I and II are the most abundant of the AAs and are found in almost all Aristolochia species. Once established, AA–DNA adducts persist for years in the renal cortex, serving as reliable biomarkers of exposure to AA. Arlt et al. were the first to prove that this hypothesis is correct by detecting specific AA–DNA adducts in all urinary tract tissues from patients with CHN that they examined, as well as in invasive liver metastases and various human tissues outside the urinary tract.
Cosyns ﬁrst raised awareness of the unique renal histopathology of CHN with its striking similarity to BEN.
These fascinating reports renewed scientific interest and inspired new research that produced unequivocal evidence supporting the role of AA in the pathogenesis of BEN and UUC.
The similarities between CHN and BEN have led to the hypothesis of a common aetiological agent for both diseases. This hypothesis implies that BEN, CHN and AAN are the same disease and that dietary ingestion of AA, in conjunction with individual genetic susceptibility, accounts for all epidemiological, clinical and pathophysiologic features of BEN and associated UUC.
A recent publication by Grollman et al. [59, 60] presented results showing that the accumulation of AA–DNA adducts was present in the renal cortex and upper urinary tract of five patients with BEN from an endemic region in Croatia, but not in five patients with other forms of chronic renal disease or five patients with upper urinary tract transitional cell cancer living in a non-endemic area of Croatia. A study of the incidence of UUC and urinary bladder tumours (UBTs) associated with BEN in the 30-year follow-up revealed a total of 575 urothelial neoplasms during the 10-year period, 1989–1998, compared with 659 in the period 1969–1988. Remarkably, UUC had an 11.2-fold higher incidence in endemic than in non-endemic areas. However, this was far less than in the period 1969–1988, when UUC was 57.1 times more frequent. UBT were 2.3 times more frequent in endemic than in non-endemic areas, but 11.9 times more frequent in the previous period 1969–1988. It seems that the frequency of neoplasms is decreasing from the upper towards the lower urothelium.
Immunosuppression in BEN transplant patients comprises an even greater increased risk of developing UUC requiring nephroureterectomy. Thus, 33.3% BEN of patients develop UUC, compared with 0.67% affected transplanted patients with other causes of CKD.
Cytological screening procedures of UUC in AAN using fresh voided urine or cystoscopy with bilateral ureteral washing and brushing characterize the finding of giant multinucleated cells and/or nucleolar atypias. According to these results, the specificity of cytological analysis on voided urine and brushing samples is 100 and 96%, respectively, and the sensitivity 17 and 54%, respectively.
The finding by Grollman et al. of AA-derived DNA adducts in renal cortical and urothelial tumour tissue of patients with documented BEN, associated with the dominance of the A:T → T:A transversions in the p53 tumour suppressor gene mutational spectrum, was a breakthrough in the identification of AA as an aetiological agent of the upper tract malignancy observed in BEN. The TP53 mutation spectrum in AA-induced UUC displays an unusual pattern that is readily distinguished from all other spectra that have emerged to date from among the 27 000 tumour mutations in the IARC TP53 database. Unique features of this spectrum, including the predominance of A:T → T:A transversions found also in Taiwanese patients with UUC confirmed the hypothesis that all components of the AA signature TP53 mutational spectrum, established in the context of UUC associated with BEN, are similarly found in Taiwanese patients with UUC (Figure 4).
In light of the persistent widespread use of Aristolochia herbal remedies in traditional Chinese medicine and recently published data that some crops can take up AA from the soil, one can raise the question whether AA could be responsible for a previously and currently widespread unrecognized global renal disease and UUC.
The finding that porcine nephropathy has many characteristics in common with BEN led some researchers to propose that ochratoxin A (OTA)-induced renal disease and urothelial tumours in humans are highly similar. It was one of the first well-elaborated hypotheses regarding the aetiology and pathogenesis of BEN that appeared in the literature in the early 1970s. Akhmeteli and later Krogh proposed that BEN was the result of contamination of the food chain in endemic areas by OTA.
Numerous surveys conducted in North America, Asia and Europe have revealed that OTA, a toxic product of molds that belong to the Aspergillus or Penicillium fungal genera, is a natural contaminant of plant products. Contamination frequencies of up to 40% have been encountered, at levels in the range of 5–500 µg/kg. OTA is a significant causal determinant of porcine nephropathy. Renal lesions in porcine nephropathy include degeneration of the proximal tubules, interstitial fibrosis and hyalinization of the glomeruli. The disease is endemic, and outbreaks have been associated with weather conditions.
OTA is considered to be a human carcinogen, an opinion based on sufficient evidence of carcinogenicity in experimental animals. Detection and quantification of DNA adducts related to OTA in human kidney tissues and urothelial tumours of BEN patients from Bulgaria, Croatia and Serbia were the subject of many studies, aimed at finding the molecular evidence for OTA role in the aetiology of BEN. Extensive field research analysing the OTA in foods consumed by inhabitants from the area with BEN aimed at providing results confirming the high exposure of this population to OTA, and thus strengthening the hypothesis of the involvement of this mycotoxin in BEN aetiology, has been performed. Although several authors published data on new molecular and field evidence for the implication of mycotoxin OTA in the aetiology of human nephropathy and urinary tract tumours [55, 68], no adequate human studies of the relationship between exposure to OTA, the human renal disease and cancer have been reported.
Analysis of tissue specimens from CHN for the presence of DNA adducts related to both OTA and AA exposure revealed that AA-specific DNA adducts were detectable in all five urinary tract tissues from five patients, whereas OTA-related DNA adducts were detectable in two kidneys and only one ureter. Furthermore, OTA-related DNA adduct levels were ∼50 times lower than AA–DNA adduct levels. In female and male rats treated with the same slimming regimen as the CHN patients, but with a 10-times higher level of Chinese herbs, AA–DNA adducts were found in kidney tissues but adducts derived from OTA were not observed. These results demonstrate that presumably OTA-related DNA adducts do not play a key role in CHN or CHN-associated UUC [52, 55] and consequently in BEN and associated UUC.
This hypothesis was proposed in 1991 by scientists from the US Geological Survey [69, 70] based on the geographical matching between the location of Pliocene lignite deposits in the Balkans and the location of endemic areas (Figure 1), as well as preliminary geochemical analyses of well water from villages located in endemic areas of former Yugoslavia, which showed the presence of toxic organic compounds not observed in well water from non-endemic villages [71, 72].
The lignite hypothesis is based on the assumption that toxic organic compounds in lignite, or in weathered lignite, may be released by groundwater and thus contaminates drinking water wells. Although the concentrations of these organic molecules in well water may be low, long exposure and/or accumulation in body tissues over time may lead to kidney lesions. The development of UUC in some individuals can also be explained by this hypothesis because most of these toxic organics are well-known carcinogenic factors [73, 74].
This hypothesis emphasizes the role of environmental factors involved in the aetiology of BEN, accounts for the peculiar geographical restriction of BEN and provides a basis for investigating the possible wider impact of coal-derived toxic organic compounds in groundwater on human health.
The familial clustering of the disease was indicative of the role of a hereditary predisposition in the aetiology of BEN and prompted many genetic investigations. The hypothesis implicating the multifactorial nature of the BEN aetiology assumes that genetic factors create a predisposition to BEN. The combined action of genetic and environmental factors may result in the development and determination of clinical and epidemiological characteristics and the progression of the disease.
This hypothesis was substantiated by family investigations in Bulgarian patients with BEN. A study by Toncheva et al. in 4077 subjects from 417 families affected with BEN led to the conclusion that all patients with BEN belong to certain families. Even residents from non-endemic villages diagnosed to have BEN were found to be members of BEN families that had migrated from their birthplaces. Furthermore, some epidemiological characteristics of BEN are indicative of the involvement of genetic disorders, i.e. the proportion of the affected offspring increases in accordance with the number of parents affected. Consequently, the risk of developing BEN is much higher in first-degree than second-degree relatives and decreases substantially in remote relatives.
Changes in enzyme activity
Norum and Gjone proposed in 1967 a familial deficiency of lecithin-cholesterol acyltransferase (LCAT) as a primary disorder. They showed that familial renal disease can develop secondarily to LCAT deficiency and lipid abnormalities. Familial LCAT deficiency is an autosomal recessive disorder.
LCAT deficiency is associated with the percentage increase of free cholesterol followed by the proportional decrease of esterified cholesterol and abnormalities in the structure of lipoprotein particles. As a consequence of lipid disorders, other organs can also be involved, such as the kidneys, cornea and erythrocytes, with the clinical manifestations of proteinuria, usually associated with renal insufficiency, corneal opacities and haemolytic anaemia. The gene encoding LCAT is localized in region q 21–22 of chromosome 16 which consists of six exons. In LCAT-deficient patients, several mutations in all six exons have been described. Biochemical and clinical manifestations of familial LCAT deficiency are highly variable. The finding of no or extremely low LCAT activity in affected patients suggests that expression of the disease is modulated by additional environmental factors and genes of minor importance [79, 80].
A study by Pavlovic in 1991 showed that a certain proportion of healthy subjects from BEN families had a peculiar form of lipid abnormalities associated with an abnormal LCAT activity, and a possible association between these abnormalities and the aetiology of BEN was raised for the first time. The contribution of genetic and/or environmental factors to these abnormalities in BEN has not yet been elucidated.
The hypothesis of a multifactorial aetiology of BEN anticipates that an interaction of polymorphic gene variants and various environmental factors causes an increased risk of renal disease and cancer. Xenobiotic metabolizing enzymes are known to play a role in the metabolic activation of environmental mutagens and carcinogens to exert their carcinogenic effects as well as detoxification by increasing their hydrophilia. It has been shown that genetic variants of these enzymes involved in the uptake, conversion and excretion of xenobiotics determine individual levels of detoxification and are modifiers of an increased/decreased risk of chronic diseases and/or cancer [83, 84].
Extensive studies conducted in experimental settings and in BEN patients have investigated the role of several genetic polymorphisms in a number of enzymes (CYP2D6, CYP3A4, CYP3A5, NQO1, GSTT1, GSTM1, GSTP1, NAT1 and NAT2) from a detoxification system [85–87].
Some studies presented results demonstrating that the CYP3A5*1 allele, previously reported as a marker for CYP3A5 expression in the human kidney, is associated with increased risk of BEN, while CYP3A4*1B and CYP2D6 genotypes do not significantly modify the risk for the disease. It has also been found that the GSTM1 wt allele associates with BEN. The significantly lower prevalence of the GSTM1 deletion homozygotes among BEN patients suggests that individuals bearing the GSTM1-null genotype could be better protected. Additionally, some results established that alleles NQO1*2 and NQO1*3, as well as lack of GSTT1 and GSTM1, did not influence the BEN risk.
It was hypothesized that the occurrence and frequent association of BEN and cancer can be explained by the chromosomal hypothesis of oncogenesis. The first cytogenetic investigation in healthy relatives of patients with BEN born in non-endemic areas was done in 1996. Characteristics of BEN No. 3 chromosomal anomalies in terms of extremely high frequencies of 3q25 homologue discordance, chromosome breaks at the 3q25 band, structural aberrations affecting the 3q25 band, unusually high frequency of acquired chromosomal aberrations and a family history with 1 or 2 BEN parents were identified in five relatives. It is proposed that they are at a high risk for developing the disease and that a genetic mechanism might be involved in the aetiology of BEN. The additional finding of a specific chromosome marker 3q in BEN was reported, characterized by a discordance in the banding patterns of the long arms, shortening band 3q25, faster fusion of sub-bands q26.1 and q26.3 and lack of differentiation of q24.
In 1975, some evidence of a viral involvement in the aetiology of BEN was published, suggesting that a slow coronavirus infection causes BEN in humans. Data substantiating this hypothesis remained unconvincing. It was more recently documented when some authors were able to detect a novel corona virus, using primary kidney tissue cultures established as explants from tissue obtained at operations from five BEN patients with urinary tract tumours. Four out of the five biopsy specimens on extended culture yielded a coronavirus virus which was cytopathogenic for human fibroblast and Vero cells, but none from the control cultures. It was proposed that the microorganism was a novel coronavirus based on its cross-reactivity with the human coronaviruses OC43 and 229E, as well as a porcine-transmissible gastroenteritis virus.
Nonetheless, the coronavirus hypothesis was challenged when Vero cells infected with the BEN virus were additionally tested with various methods, including electron microscopy studies. It was found that the only virus detected in the cell cultures infected with the BEN-associated virus was unrelated to coronaviruses,since coronaviruses caused a cytopathic effect on infected cells differed from those observed after infection by the BEN-associated virus. Importantly, they highlighted that the involvement of a coronavirus should no longer be considered in BEN induction. Further studies are needed to clarify the nature of the 28.4 nm, non-enveloped virus particles found in the kidney cells of patients with BEN and to determine whether this virus is the causal agent of the disease.
This hypothesis supports the presence of an inflammatory pathway in BEN through the involvement of polymorphic enhancer hs1.2 influencing different binding complexes and consequently the 3D structure of the 3′ regulatory region of IgH. This study is the first that clearly links BEN to a gene involved in the regulation of immune response.
Miscellaneous multifactorial aetiologies
Pavlovic et al. tested for the first time the possible simultaneous role of coal-derived toxic organic compounds and decreased enzyme LCAT activity in the aetiology of BEN, using water concentrates from both endemic and non-endemic areas. It has been shown that well water from BEN villages contains higher numbers and concentrations of both extractable and high-molecular weight organic compounds compared with controls. The authors presumed that in this study, organic compounds contributed to higher LCAT-inhibiting activity of Serbian and Romanian drinking water samples from both non-BEN and BEN villages. On the other hand, the higher abundance of organic compounds in drinking water samples from BEN villages caused much higher LCAT-inhibiting activity compared with non-BEN ones. This finding indicates that there is presumably a clear-cut distinction between drinking water qualities coming from those two localities.
On the basis of these results showing higher organic compound LCAT-inhibiting activity, the authors proposed the concept of cause and effect events in the pathogenesis of BEN. Thus, the organic compounds in drinking water from BEN villages inhibit LCAT activity, which may cause plasma lipid abnormalities and consequent changes of intracellular and cellular membrane lipid composition that can trigger the pathogenic mechanisms responsible for the development of BEN.
Based on all of these previously published hypotheses, we think that a multifactorial aetiology is the best match with regard to BEN pathophysiology, with a genetic predisposition being the ‘condicio sine qua non’ for the development of BEN. Accepting the concept of a multifactorial set of causes seems to be a valid hypothesis, as this would help explain why the BEN aetiology has been so difficult to decipher for >50 years. It seems that this multifactorial character of BEN aetiology is associated with the multilevel distribution of causative agents within a pathogenetic chain of events. A significant proportion of the pathogenetic chain of events does not only take place in the renal and upper urothelial tissue as target organs. It is likely that the intrinsic environmental aetiological factor never reaches kidney structures in its original form but only after being metabolized in the liver.
The genetic predisposition is presumably responsible for providing the key circumstances for imposing the action of other environmental and endogenous aetiological factors. Further work aimed at exploring the contribution of factors mentioned in some hypotheses to a multifactorial aetiology and occurrence of BEN is necessary to determine the role of the environment in the aetiology of BEN and UUC.
In light of the worldwide distribution of Aristolochia spp. and the widespread use of herbal remedies in traditional Chinese medicine, as well as recently published data that some crops can take up AA from the soil, there is a possibility that diseases similar to BEN and UUC exist elsewhere as unrecognized disease entities.