Penicillium polonicum K. M. Zalessky includes a group of fungi previously included within the broad species P. aurantiogriseum Dierckx, and historically within P. verrucosum var. cyclopium. Revision in the 1990s differentiated several distinct groups, for which refined cultural protocols for consistent expression of particular phenotypic characters, combined with some closely associated idiolytes, has created more meaningful speciation [1,2,3,4]. We have direct experience of occurrence of verrucosidin and penicillic acid [5], auranthine [6], anacine [7]. Other metabolites of an isolate of P. aurantiogriseum from the same Yugoslavian village as that used here are quoted as penicillic acid, fructigenine A, cyclopenol, cyclopenin, cyclopeptin, dehydrocyclopeptin, viridicatol, 3-methoxy-viridicatin, rugulosuvine, verrucofortine and normethylverrucosidin [7].

Concurrently, toxicological interest in some isolates of P. verrucosum var. cyclopium arose from the seminal findings in experimental rats of a nephropathy involving marked karyomegaly in the pars recta region several days after administration of cultured mycelium [8]. The findings were offered as of potential etiological relevance to the idiopathic human disease Balkan endemic nephropathy, since the fungus can readily be found as a stored-food spoilage mould in the region.

Further study has shown the rat nephropathy caused by moulds historically described as P. verrucosum var. cyclopium or P. aurantiogriseum or P. commune [9,10] to be confined to those now defined by the recent restoration of P. polonicum [4]. Chemical characterisation of the nephrotoxin is still incomplete, although it is water-soluble and amphoteric, and may be a peptide [11]. Weanling rats of both sexes are susceptible, responding to ingestion of contaminated feed within a few days by mitoses in proximal tubule epithelium progressing to striking diffuse karyocytomegaly distributed mainly throughout the outer stripe of the outer medulla [12,13,14,15]. In contrast, there was no similar response in the hamster [16]. A renal apoptotic component in early responses to oral administration to rats of cultured P. polonicum, or an extract, has been demonstrated [17].

Opportunity to make direct comparison of toxicological response to several days of treatment with a P. polonicum culture extract in rat and vervet monkey arose 20 years ago. Significance of the result was difficult to perceive at the time. However, there is recent suggestion of involvement of the fungus in porcine nephropathy in Bulgaria and South Africa [18,19].

The total cumulative dose given to the vervet over the 10 day period exceeded that given to a rat by a factor of 5.85 on a body weight basis, and by a factor of 2.3 expressed also on a daily basis. A further two-fold factor might even be applicable to widen the difference on a metabolic rate basis. Thus, the negative response in the primate is the more compelling.

Studies were made over 20 years ago on putative toxicity of P. polonicum to pigs because of its abundance as a food spoilage mould [11,20], and because the toxin causing renal responses in the rat resides in spores [10]. However, until now, the consistent negative findings were not easy to report. Relatively large amounts of cultured P. polonicum were used in collaboration in the UK with Prof. R.H.C. Penny at the Royal Veterinary College, London, UK and Dr. J. Heaton, Consultant Pathologist, Victoria Hospital, Worksop, UK.

In 1988, 750 g (wet weight) cultured mycelium of an isolate from a Yugoslavian village in which Balkan nephropathy was hyperendemic [9] was given to each of two 16 kg pigs, subdivided over five daily feeds (approx 30 g dry weight of fungus), in a slurry disguised with milk powder. No renal histopathological changes occurred. Similarly, daily dose of ~40 g (dry weight) of Yugoslavian isolate M2 [9] also evoked no renal histopathology. In 1989, administration over six days of isolate M2 to two adult mini-pigs (10 kg) also failed to reveal renal toxicity. In 1990, extract of yet another Yugoslavian isolate M6 [10], as also used here, was given in feed to another 15 kg pig over seven days; no renal histopathology was found. This contrasts with demonstration that only 1 g of P. polonicum spores caused typical acute rat renal histopathological change when administered in feed over four days.

Renewed published interest in P. polonicum stems from common occurrence of this mould in feed for pigs, not only in Bulgaria [19], but also in S Africa [18], where similar cases of idiopathic porcine nephropathy have been recognised. Isolation and partial characterisation of a highly non-polar metabolite of P. polonicum and other fungi [21] has led to speculation that this compound may be the P. polonicum metabolite which caused the striking rat renal karyocytomegaly illustrated here, and could be a factor in a putative multi-mycotoxin etiology for the porcine nephropathy. Unfortunately no elemental composition was proposed for the molecular mass of the metabolite (390.27701). The closest elemental composition is C₂₄H₃₈O₄, which fits several sterols; that type of compound could also fit the non-polar chromatographic mobility. Toxicity towards human lymphocyte cells was shown [21], but unfortunately without any control. Nevertheless, omission to cite [12] could account for not appreciating that the rat renal response to whole P. polonicum, or an extract designed to be cation-exchangeable and thus to exclude sterols and acids, could hardly be due to that highly non-polar compound.

It so happens that another mycotoxin that causes karyomegaly in rat renal proximal tubules is ochratoxin A, which was first recognised as a mycotoxin in South Africa [24]. Ochratoxin A is much less active in causing karyomegaly in male rats, requiring longer exposure and never achieving the magnitude of the P. polonicum toxin [12](Figure 3). Aneuploidy was recognised for both ochratoxin A and P. polonicum toxin and measured for the latter as incomplete multiples of the genome even beyond octaploid [25,26].

In a study in which pig feed contaminated with ochratoxin A at 800 µg/kg was given continuously for a whole year [22], only slight kidney hypertrophy ensued, which would probably have been unrecognisable in routine slaughterhouse processing. Thus the ~70 µg/kg measured in commercial feed for bacon pigs (for 6–8 months only) in two years in South Africa [18], hardly expected to be significantly nephrotoxic, would seem to require a very exceptional synergism with other natural toxicants to cause significant porcine nephropathy there. Consequently, the evidence [18] could equally be interpreted as that nephropathy not being a mycotoxicosis. The porcine nephropathy also seen in Bulgaria, associated with feed contaminated variously at ~100–200 µg ochratoxin A/kg in three years, and at 375 µg/kg in another [19,23], would still require considerable augmentation by other mycotoxins, and any contribution from P. polonicum would appear to be implausible. Nevertheless the matter could readily be resolved by experiment.

There is no suggestion that a karyomegalic nucleus is already en route towards renal tumourigenesis. However, demonstration of the aneuploidy in rat kidney in response to five weeks of continuous maintenance on feed contaminated with an ~20% component of shredded wheat moulded by P. polonicum has defined the extent to which the ‘polyploid’ nuclei illustrated in Figure 1 can be composed of a high proportion of incomplete genome multiples. Since the karyocytomegaly illustrated in Figure 2 persists for life [12], and aneuploid nuclei are generally recognised as having tumourigenic potential [26], the abnormal organelles may harbour genetic damage, including that resulting from mis-repaired adducts between DNA and ochratoxin A [27] or from any other putative mechanism proposed [28]. In the case of ochratoxin A, it is reasonable to consider that certain tumourigenic permutations of randomly-generated genetic change could reside in some of the karyomegalic nuclei, which are located in the rat kidney region within which renal tumours seem to arise.

The present situation with respect to nephrotoxicity of P. polonicum has echoes of a mycotoxicosis caused by Fusarium verticilliodes, described in its first report [29] as hepatocarcinoma and oesophageal basal cell hyperplasia in rats fed an artificially moulded dietary component. The finding generated wide interest, particularly in the USA, because the fungus readily causes spoilage of maize. There was also a regional hypothesis concerning putative involvement in hotspots of human oesophageal cancer in the Republic of Transkei. Subsequent studies in South Africa recognised the new mycotoxin fumonisin B1 as causing diverse toxicological responses in horse (leukoencephalomalacia), pig (pulmonary oedema) and rat (renal carcinoma). For the latter, a comprehensive 2-year toxicology study has been made [30], revealing male-specific renal carcinoma. Concurrently, Hard et al. [31] stated “carcinomas induced by fumonisin B1 were predominantly a rare and highly malignant variant of renal tubule tumour capable of infiltrative growth invading between tubules and glomeruli, sequestering these pre-existent elements as entrapped remnants within the tumour mass.” Much the same could be said concerning male rat renal tumours in response to ochratoxin A, with respect to envelopment of glomeruli (Figure 4). Hard et al. [31] also recognised sarcomatoid areas in the National Toxicology Program’s fumonisin renal tumours; a similar sarcomatoid component of mammary tumour has also recently been recognised in a female rat given the tumour promoter sodium barbitate in drinking water after chronic dietary exposure to ochratoxin A [32].

However, when a toxicology study was made during 13 years with the same fumonisin-producing Fusarium as that described by Marasas and colleagues in a non-human primate, no adverse histopathological renal or oesophageal changes were found [33], and current research does not easily sustain a causal hypothesis for fumonisin in oesophageal cancer. The present contrasting findings with P. polonicum, also in the vervet monkey, extend uncertainty about the rat as a good model for the human with respect to ochratoxin A as a potential renal carcinogen. This is compounded by the negative responses also in the pig.

Unfortunately, ochratoxin A seems never to have been tested for any renal histopathological changes in the primate. Toxicokinetics in the vervet monkey was studied [34] for up to 21 days after a single dose (2 mg ochratoxin A/kg), revealing a plasma half-life of that duration. However, although slow release from plasma proteins to renal elimination was occurring over that period, opportunity to study renal histopathology was unfortunately omitted. The failure also of hamsters to show any renal response, in an elegant study [16] of perfuse-fixed kidneys from animals given the same fungus as used here for up to five weeks, contrasts with the present striking response of rat 4 (Table 1; Figure 1 B, C). Thus, if the present contrasting findings might be repeated for ochratoxin A, assumptions about the validity of extrapolating from rat renal carcinogenicity data to risk assessment concerning this mycotoxin for humans may be questionable.