The rat tumour tissues were available from experiments on tumourigenic response to dietary OTA dose regimens, duration of exposure, latency, and both threshold and tolerable dosages. Tissues became available from old or ageing animals when euthanized due to clinical morbidity. No tissues were from decedents. Thus, tumours ranged from large aggressive metastatic carcinoma to cryptic micro-adenoma. It is assumed that in many cases the common developmental sequence from adenoma to carcinoma occurred, and that this may have involved additional genetic change, whether through direct genotoxic insult or through epigenetic influences. This is implied from the DNA ploidy distribution studies on these and other tissues [18
] showing diploid adenoma and aneuploid carcinoma. Nothing is known about the kinetics of rat renal tumour growth in response to OTA, bearing in mind that at least six months of continuous exposure to contaminated feed seems necessary in the first year of life at a daily intake of ~0.3 mg/kg to be sure of causing renal cancer in some individuals. Further, tumours have rarely been discovered in animals less than 18 months old, there can be a year’s latency between ceasing OTA exposure and discovering a renal tumour, and no time-course experiments have been performed. The latter would be a major undertaking, comparable in magnitude to the NTP study [3
], but modern knowledge could optimise yield of tumours in a hybrid rat [10
]. In our experience, protracted OTA exposure that is well tolerated is likely only to cause unilateral renal tumourigenesis, probably from a single focus or at most very few foci. In larger renal carcinomas the precise point of origin has been obscured, but the smaller neoplasms tend to centre in the outer medulla close to innermost cortical glomeruli. Thus the present tumour material is unique, and necessarily heterogeneous both in OTA exposure and histology.
The human tumours for the present study were chosen because their histopathology and DNA ploidy distribution had all been studied previously [9
] and ranged from diploid to marked aneuploidy lesions.
The clinical context for interpreting the present findings from immunohistochemical examination of rat and human tumours is that phosphorylated S6 ribosomal protein is a marker indicating activity of the mammalian target of the anti-tumour drug rapamycin (mTOR) in cell proliferation. Expression of p-S6 might therefore be expected in some malignant tissues and, indeed, marked expression has been illustrated [19
] in tumour cells lining some renal cysts in mice with mutation in the gene Tsc1
that forms part of the tuberous sclerosis complex. Illustrations differentiated between staining for p-S6 protein in cells (~100 m2
) lining one third of small (100–200 µm diameter) renal cysts. Staining was absent in others. However, intense staining occurred in >90% of renal cell carcinomas and cystadenomas, the cells of which were abnormally large (~180 µm2
Tuberous sclerosis complex is an autosomal dominant human syndrome with benign and occasionally malignant tumours in CNS, skin and kidney. Two human genes, TSC1
, are involved. The Eker rat, heterozygous for a dominantly inherited germline mutation in the Tsc2
tumour suppressor gene, is recognised as a valid model for human tuberous sclerosis complex [20
]. Intense illustrated immunohistochemical staining of renal tumour, contrasting with associated renal tissue, extends also in [19
] to recognising a small cluster of cells in a renal tubule as a potential tumourous neoplasm. Notably, the latter situation seems to be matched in the present findings. Mutations in either the Tsc1
or the Tsc2
gene can cause the pathology of the tuberous sclerosis complex [19
]. Wilson et al
] generated Tsc1+/−
mice with predisposition to develop cysts and then to progress to cystadenoma and renal cell carcinoma. They then identified somatic Tsc1
mutation in ~80% of these tumours but only in one third of cysts. A role for haploinsufficiency in Tsc1
in cyst formation was proposed. Tumours showed much stronger staining for p-S6 protein than did the cysts. We therefore conclude that consistent overexpression of p-S6 protein in rat renal carcinomas caused by OTA in the present study implies modulation concerning mTOR signalling.
It is particularly notable where, within animals, p-S6 differentiated between “normal” rat tumours (testis seminoma or subcutaneous fibrosarcoma that were not stained) and ochratoxin-generated renal carcinomas that were stained. The complete absence of p-S6 staining in the three small renal adenomas found in rats over two years old (OTA given only in the second year) may be demonstrating that genetic or epigenetic events associated with mTOR pathway dysregulation is a later step associated with progression to carcinoma. However, the precise status of these small neoplasms as benign or having proliferating potential remains unclear.
Strong diffuse staining for p-S6 protein has also been shown in some human soft tissue sarcomas [21
]. The present finding of consistent overexpression of p-S6 protein in a rare rat mammary angiosarcoma associated with OTA exposure, suggests that genetic changes targeted by OTA also caused this tumour, potentially via change in the Tsc1
The present findings towards the end of tumourigenesis complement those from studies on gene expression changes in Eker versus
wild-type rats during up to two weeks exposure to aristolochic acid or OTA [22
]. Cell proliferation was assessed immunohistochemically for proliferating cell nuclear antigen (PCNA) by an anti-PCNA antibody on wax-embedded kidney sections. Differentiated responses suggested that aristolochic acid toxicity was Tsc2
-independent in both Eker and wild-type rats, whereas that of OTA was more prominently associated with deregulation of mTOR genes in the Eker rats.
For resolving the uncertainty of experimental models for OTA as relevant to human renal tumourigenesis, it is unfortunate that relatively so much scientific effort has focused on only short-term animal experiments and in vitro studies to predict tumour mechanisms, and so little has been devoted to comparing actual tumours in rodents and humans. Several mechanisms that avoid recourse to genotoxicity for OTA have been proposed, e.g., oxidative stress and aberrant mitosis [23
]. However, short-term whole animal experiments may not identify influences which fit the requirement for many months of OTA exposure, cultured cells in vitro poorly represent the complexities in kidney, and measurements on kidney tissue are difficult to apply to initiation of a single neoplasm in the outer medulla of only one kidney in a rat. A recent unsatisfactory proposal for aberrant mitosis [24
] attempted to apply experimental in vitro
findings to the in vivo
situation, but the authors omitted to recognise that most of the OTA in plasma is protein-bound, from which correct extrapolation to their mitotic aberration data from cell-cultures should actually be viewed in the no effect range, negating the perceived relevance to renal tumour formation by OTA. Genotoxicity of OTA is still a matter of debate although the structure of an OTA-DNA adduct has been determined [25
]. Finding DNA adducts in tissues is, of course, indicative of OTA exposure, but claims that adducts associated with tumours identify the tumourigen are unsustainable, particularly since adducts have recently been found in rat and human blood [26
]. Thus, adducts could reasonably be detectable in all well-vascularised tissues and the extent to which analysis measures adducts actually within the tissue parenchyma is obscure. Notably the concept of OTA-DNA adducts in blood, and disconnection of OTA from genesis of rat testis tumours in the present study, further conflicts with perception of OTA as a cause of human testicular cancer [27
Ultimately, the precise mode of action of OTA as a rat renal carcinogen may remain unresolved, and indeed may not matter if the key genome change (s) can be discovered and compared for relevance with those within the range of human upper urinary tract cancers. Meanwhile, the present findings weaken an assumption that the rat is a valid model for considering OTA as a human carcinogen. It would be interesting to study p-S6 expression in OTA-generated renal tumours in male mice [5
] since they are the other possible mammalian model, and also to explore citrinin-induced rat renal tumours (adenomas) [29
] since citrinin and OTA have structural and toxicological similarities as nephrotoxins [30
In the rat, OTA exposure elicits renal accumulation of aneuploid karyomegalic nuclei in the outer medulla where tumours seem to arise. However, any role of these unstable nuclei in carcinogenesis is unclear. It is unfortunate that the few experiments with OTA in primates have not included renal histopathology and so it could be reasonable to assume that the same karyomegaly as is seen in rats might also apply in a primate. However, experimental animals are not always perfect models for humans. Therefore, the contrast between the striking renal histopathological response in the rat to four days of dietary administration of extract of wheat moulded by Penicillium polonicum
, and the absence of any adverse response in a vervet monkey to 10 days of nasogastric administration of the same extract to a cumulative 5-fold greater dose than in the rat on a body weight basis, [31
] may be a relevant consideration.
For human relevance, some regard OTA as a cause of the Balkan endemic nephropathy and its associated tumours of the upper urinary tract, but compelling evidence of exposure and relevant extrapolation from experimental animal toxicology is lacking. Yet, for example, there is precautionary EU legislation concerning tolerable human intake of OTA from food, and necessarily sophisticated analytical monitoring of food components for traces of the toxin. Further immunohistochemical study of the present rat OTA tumours could contribute to debate about relevance of the rat model.