Evaluation of the significance of changes found in studies on DINP (Di-isononyl phthalate) and DIDP (Di-isodecyl phthalate) in the rat.

Transcription

1 Expert Review: Reviewer: Sponsor: Evaluation of the significance of changes found in studies on DINP (Di-isononyl phthalate) and DIDP (Di-isodecyl phthalate) in the rat. Professor Sir Colin Berry MD, PhD, DSc, FRCPath Emeritus Professor of Pathology at Queen Mary College London, UK European Council for Plasticisers and Intermediates Avenue E Van Nieuwenhuyse 4, Bte Brussels Belgium May 22, 2012 Professor Sir Colin Berry Date

2 Evaluation of the significance of changes found in studies on DINP (Di-isononyl phthalate) and DIDP (Di-isodecyl phthalate) in the rat. Introduction This report provides an expert opinion on a lesion defined as spongiosis hepatis, observed in male rats exposed chronically to the high molecular weight phthalate esters, Diisononyl phthalate and Di-isodecyl phthalate. This opinion addresses the relevance of spongiosis hepatis to human beings, reviews available data on the lesion and evaluates the significance of spongiosis hepatis following exposure to DINP and DIDP. An appropriate no observable adverse effect level (NOAL) for this effect in rat studies is suggested, and an appropriate assessment factor for use in human health risk assessment is proposed. 1. Spongiosis and Peliosis Hepatis Spongiosis as a change found in rats Spongiosis hepatis is a spontaneously occurring lesion in the livers of ageing rats, appearing most often in the second year of life, with a strong predilection for male animals. The incidence of spontaneous spongiosis hepatis can reach 34% in male Fischer rats. The lesion is composed of altered sinusoidal lining cells which some authors have interpreted as hepatic stellate cells (Ito cells or fat-storing cells). The characteristic histological appearance is described by Bannasch et al (1981, 1986, 1997) in rats treated with carcinogens - spontaneous spongiosis hepatis is less likely to be multifocal and is less often associated with stellate cell aggregates. In my experience, there is no comparable human lesion, a view shared by expert human pathologists in this field. In Professor Sir Roderick MacSween s book the authors state to the best of our knowledge no human counterpart of spongiosis hepatis has ever been described. The authors use the term spongiocytic pericytoma but are considering the lesion we discuss here and Bannasch is a contributor to the volume. Further, there was no evidence of a lesion resembling spongiosis hepatis in a review of 163 human livers conducted by members of the Bannasch laboratory (Su et al., 1997) nor in my autopsy study of 1500 livers at autopsy. The broad consensus of pathologists appears to support the view that spongiosis hepatis is a degenerative change. From NTP studies, spongiosis hepatis is a lesion that appears to be confined to rats, particularly male rats, and teleost fish. In human terms, Peliosis hepatis (PH) is a histopathological entity which might be compared with rat spongiosis. PH has been defined in Man as a change in hepatic vasculature characterised by randomly distributed multiple blood-filled cavities throughout the organ. The size of the cavities usually ranges between a few millimeters to 3 cm in diameter and the change is commonly used as a macroscopic descriptive term. It has been reported in Man after exposure to drugs (including anabolic steroids, oral contraceptives, corticosteroids, tamoxifen, and Page 2 of 10

3 diethylstilbestrol) and to some toxicants including arsenic and thorium oxide. It may also occur as a result of circulatory change (Yoshioka et al, 1998) or infection (Garcia-Tsao et al, 1992) but in 20-50% of patients, no associated condition is identified. Any change in the relative tissue pressure exerted by hepatic artery blood flow, portal system flows or the egress of blood form the liver is a potential cause and peliosis is commonly found in long- standing debilitating disease. I do not believe this lesion has a relationship to rat spongiosis hepatis as defined in the documents here. For this and for other reasons there is difficulty in extrapolating a finding of spongiosis hepatis in the rat to Man. The nature of the blood flow in the liver, with afferents from hepatic artery and portal system and with central lobular venous drainage means that arterial and venous input may vary both normally and in a variable way in each component in disease. In many conditions where central venous pressure is raised (CHF- congestive heart failure) there may be considerable loss of gradient across the liver lobule. In CHF this typically gives a nutmeg liver appearance where the surface resembles a grated nutmeg with the dark areas representing dilated centrilobular venous areas a change so severe in some cardiomyopathies that it has been (wrongly) called cardiac cirrhosis. Circulatory change may cause thus large differences in appearance in man and these non-proliferative changes are responsive rather than primary. In addition human livers contain many naturally occurring vascular lesions and these may become modified by time with much structural change (see references below). Where detailed studies of human liver are carried out (this is uncommon) areas of vascular change with dilated vasculature and hepatocyte atrophy are not uncommon and may superficially resemble spongiosis in the rat I do not regard them as equivalent lesions. Although cell division may take place these are not proliferative lesions as Karbe and Kerlin (2002) point out in their consideration of the lesion in the rat. In considering 12 oncogenicity studies in rats with induced cystic degeneration in the liver, there was a marked sex predilection, with males more likely to develop either spontaneous or chemically induced lesions and there was an association with hepatocellular proliferation or hypertrophy. Vascular change in the liver is common and regulated. Where there is tissue damage and loss of hepatic lobular architecture new vessels are formed by the differentiation of endothelial cells (circulating bone marrow-derived endothelial progenitor cells are found in the blood of adults) and flow through these vessels (and other factors) results in the further differentiation of blood vessel wall with arrest of endothelial cell proliferation, reconstruction of a basement membrane around the neo-capillary, and investment and coverage of the immature capillary with pericytes or in sinusoids Kuppfer cells. In adults and in the absence of a disease process, transient formation of new blood vessels is only observed under certain physiological situations (during the menstrual cycle, in the placenta during pregnancy, or during wound healing), and occurs mainly through angiogenesis. Dysregulated angiogenesis is part of the pathogenesis of numerous diseases including vascular retinopathies, rheumatoid arthritis, and cancer (Folkman, 1995). In order to characterise these vascular responses to injury a great deal of work has been done with a mechanistic approach Gene knock-out experiments have emphasized the pivotal roles played by Vascular Endothelial Growth Factor (VEGF) VEGF receptor 2 (VEGF-R2) and VEGF-R1 in new vessel formation and Biscetti et al (2008, 2009a, 2009b) and Han et al (2008) have found that PPAR alpha and PPAR gamma activation stimulates neo-angiogenesis through a (VEGF) dependent mechanism. This emphasises that the pathology of vascular lesions in the liver is a controlled process, not analogous to neoplasia. Page 3 of 10

4 In many human individuals benign vascular tumours (haemangiomas) are present in the liver from birth. These are not true neoplasms but hamartomas tumours composed of tissues normally found in the tissue of origin but in an abnormal arrangement. These lesions are clearly developmental in origin and do not become malignant. They may resolve by thrombosis to leave fibrous scars (Berry, 1985) and may disappear with time. They are common; in one study Berry (1987) found an incidence of 2% of haemangiomas in the liver in 1500 unselected autopsies at all ages. These benign lesions bear no relationship to the changes found in the lesions commonly described as spongiosis or peliosis and they have no relationship to the rare human angiosarcoma. Most of the known causes of human angiosarcoma (vinyl chloride exposure, tumours following Thorotrast [thorium dioxide] administration in imaging, inorganic arsenicals) are clearly dependent on genotoxic effects. In other cases, including angiosarcoma associated with AIDS Kaposi s sarcoma and multi-centric Castleman s disease, infection with the HHV-8 virus (human herpes virus-8) plays a critical role the virus latency associated nuclear antigen is found in the cells of these tumours but not in juvenile haemangiomas (Schmid and Zeitz, 2005). Angiogenesis and endothelial cell proliferation is well regulated in Man (in the menstrual cycle there is extensive vascular proliferation and in wound healing vigorous vascular proliferation is necessary but vascular neoplasms are not seen). Hypoxic injury stimulates vascular proliferation but this too is well controlled and those who constitutively express angiogenic factors to excess (in Chuvash polycythaemia, say) do not develop vascular tumours or evidence of uncontrolled vascular proliferation. Reactive changes may occur in the liver vasculature as a result of changes in local pressure and flow within the lobules. As Karbe and Kerlin (2002) state risk assessment for man can be based on no-effect levels and safety margins, as for other non-neoplastic adverse effects that have no counterpart in man. As stated above, Han et al (2008) have found that PPAR alpha and PPAR gamma activation stimulates neo-angiogenesis through a (VEGF) dependent mechanism. These authors also showed that PPAR alpha- and PPAR gamma-induced in vitro and in vivo angiogenesis may be significantly decreased by inhibiting VEGF activity. Chintalgattu et.al. (2007) have shown that activation of the peroxisome proliferator-activated receptor (PPAR)-gamma by its agonists induces VEGF expression. In animals where PPAR s are clearly having an effect, there may be some vascular remodelling where circulatory changes are occurring. Vascular changes are produced by this remodelling but are clearly not neoplasic. Two studies in primates (oral administration of 2500mg/kg/day DINP to marmosets Hall et al [1999] and DEHP and DINP each at 500mg/kg/day in Cynomologus monkeys Pugh et al, [2000]) produced no evidence of peroxisome proliferation in these monkeys. 2. Evaluation of incidence and NOAEL of spongiosis hepatis in male rats 2.1 European Chemicals Bureau Risk Assessment A summary of the European Union Chemicals Risk Assessment report from the European Chemical Bureau on 1,2-benzenedicarboxylic acid, di-c8-10-branched alkyl esters, C9 rich and di- isononyl phthalate (DINP), makes clear that peroxisome proliferation was the major finding in terms of pathology. Studies were made in mice, rats, dogs, cats, rabbits and monkeys. Page 4 of 10

5 As spongiosis hepatis is a lesion observed in aged rats (males predominantly), the evaluation below relies primarily on results from two chronic studies on DINP. It should be noted however that liver effects were observed in several sub-acute and sub-chronic studies in rats following treatment with DINP. Essentially, the findings were similar in several feeding studies in the rat, depending on duration for example a one week study in male Fischer 334 rats showed increase in absolute and relative liver and kidney weights at a dose of approximately 1700mg/kg/day, identified as an LOAEL. Other feeding studies gave results for NOAELs of mg/kg/day. In a marmoset 13 week gavage study there were no findings of note in animals given 0, 100, 500, 2500 mg/kg/day. The main value of the monkey study was to emphasise the difference in response of rodents and primates to peroxisome proliferators. Lington et al. (1997) and Moore (1998) reported that spongiosis hepatis increased in frequency with chronic DINP exposure in male but not female rats, but that there was no evidence of the lesion in male or female mice in these studies. However, while the frequency of spongiosis hepatis in male rats increased with dose, the severity did not. In fact, an independent review of the liver slides from Lington et al. showed that the lesions were primarily minimal to mild regardless of dose (Brown, 2000). Specifically, this review classified 75% of the lesions observed as being of minimal severity, 22% as mild and only 3% as moderate ; none of the lesions were classified as exhibiting marked severity. In light of there being no identified adverse effects of spongiosis hepatis, the low severity of the lesion in male rats, independent of dose, does not suggest that spongiosis hepatis is a serious health effect in rats. As concluded by Dr. Goodman after a review of these studies, [r]egardless of dose in either study, cystic degeneration/spongiosis hepatis remained a mild lesion affecting the liver. It is unlikely that such lesions had an adverse effect on liver function or the overall health of the animal. Similarly, Dr. Cullen reviewed Lington et al. and concluded it is evident from this review that there is no dramatic dose-related change in the character of the spongiosis hepatis in DINPtreated male rats and that these data do not provide evidence of a severe liver injury (emphasis added). Overall, it was considered reasonable to employ a NOAEL of 88 mg/kg/day based on liver changes documented biochemically and on renal effects. In my opinion, this conclusion is reasonable and supports the view that the document expresses in its conclusions where it is made clear that there is no need for further information, testing or risk reduction measures with regard to the environment, for workers, consumers, humans exposed via. the environment, or from combinations of exposures of this type. In the carcinogenicity study of Cho et al (2008) with DIDP no treatment related lesions were seen in internal organs, including the liver at doses up to 8000mg/kg/day for 2 years. Mononuclear cell leukaemia occurred at a significantly increased rate compared with the vehicle control but were within historical ranges in the controls. Pre-neoplastic liver lesions were decreased in the DIDP treatment groups. A slight increase in spongiosis hepatis was noted by the authors in all treated groups. The incidence ranged from no animals in the control group to three animals in the low and mid dose groups and five animals in the high dose group resulting in 0, 6.3, 6.1 and 12.8% incidence, respectively. The authors make little note of the finding other than to note its presence and the overall conclusion made in the paper is that The increases in the relative weights of the liver and kidney were not accompanied by any histopathologic lesions in those organs. There is little information presented in the paper to fully evaluate the lesion (i.e., correct diagnosis, information on severity, number of sections reviewed). As such, it would be difficult to utilize this endpoint as a point of departure in hazard identification and risk assessment. This is particularly the case, when as noted above, the changes observed have no relevance for human pathology. Page 5 of 10

6 2.2 A report from Experimental Pathology Laboratories, Inc. entitled Histopathology peer review and Pathology Working Group review of selected lesions of the liver and spleen in male and female F344 rats exposed to di(isononyl) phthalate. The review was carried out in accordance with EPA Pesticide regulation (PR) notice 94-5 and covers selected histopathological lesions found in two studies, one carried out for Exxon Corporation and one for the Aristech Chemical Corporation. It is a carefully conducted study by a number of competent investigators who agreed on terminology of the lesions they discussed and who provided documentation of that agreement and a number of references with illustrations of the lesions that might be the occasion of note in the current discussion. Spongiosis hepatis was increased in the liver of male rats when they received more than 3000 ppm in the diet when compared with lower dose groups or controls; it was noted that differences in incidence between the two studies was likely to be due to a different pattern of sampling of the liver. An important point is made on page 16 of the document and refers to the occurrence of spongiosis like changes in foci of other tumours, making clear that in the opinion of the Society of Toxicologic Pathologists, the areas should not be scored additionally see later comment. Practically, however, the finding of lesions of this type indicates that the vascular change occurs as a result of local vascular response to changes in local tissue conditions. The expert group concluded that the appropriate NOAEL for DINP would be 88 mg/kg/day. While the European Chemicals Agency (ECHA) describes this report in great detail in the draft reevaluation of DINP and DIDP, they omit a key conclusion from the report in regards to an appropriate NOAEL for spongiosis hepatis, a key topic of the evaluation. ECHA is strongly encouraged to cite the expert panel s conclusion as it should carry some weight given that these were expert pathologists, re-evaluating the pathology slides and drawing conclusions from the actual studies. All other cited reviews did not re-evaluate the available data and instead perpetuated the opinion of the scientific committee review of This failure to cite the conclusion of the Pathology Working Group is a clear omission by ECHA. In ECHAs conclusions on spongiosis hepatis (page 78 of the ECHA draft report, (ECHA 2012)), they appear to conclude, on the basis of two human case reports, that there is no evidence that would lead to the conclusion that spiongiosis hepatis seen in studies with the rat would not be relevant for human risk assessment. The two human case reports are from Nime et al (1979) and Kaiserling and Muller (2005). In fact neither of these papers is reporting spongiosis hepatis or establishing that there is an equivalent lesion in humans. Nime et al 1979 is reporting peliosis hepatis which is not related to rat spongiosis hepatis (see discussion on pages 1 and 2 of this report. It is important to note that any change in the relative tissue pressure exerted by hepatic artery blood flow, portal system flows or the egress of blood from the liver is a potential cause and peliosis is commonly found in long standing debilitating disease). With regard to Kaiserling and Muller (2005), ECHA state in their draft report that cells appeared to be hepatic stellate cells (Ito cells). However, Kaiserling and Muller clearly state that markers for hepatic stellate cells such as synaptophysin were not found in the present case. Furthermore, the spindle cells in the present case did not express GFAP, in contrast to Ito cells. Both of these papers consider vascular and sinusoidal changes in the presence of tumours which appear to be associated with oral contraceptives; in the case of rat spongiosis hepatis seen with DINP and DIDP there is no evidence of a causal association with tumour formation. While it is understandable that regulators may use NO(A)Els for spongiosis hepatis from rat studies for risk assessment purposes it is important to take into account the human relevance of such effects, and to use assessment factors commensurate with human relevance in determining derived no effect levels. Page 6 of 10

7 This is particularly the case in this situation where an effect specific to aging rodents is being used for risk assessment of children. 3. Effects in Childhood Specific concerns have been expressed about possible effects in children and these are considered in the document Review of recent scientific data on di-isononyl phthalate (DINP) and risk characterisation for its use in toys and childcare articles produced by the European Council for Plasticisers and Intermediates. The standard arbitrary pattern of an order of magnitude change for both inter and intra species variation in making a risk assessment is difficult to apply rationally if the lesion taken as a starting point has no relevance for Man. There is no evidence that spongiosis hepatis occurs in primates: intra-species variation thus depends on extrapolation from a non-existent problem. The lesion can be produced by a number of compounds in rodents but does not seem to have been investigated in mechanistic terms. Mechanistic studies in Man suggest that proliferation of vascular structure related cell types is strictly controlled. The very large (1000 fold) difference between exposure and effects might well be calculated as 10,000 times. The toxicological evidence that is included in the ECPI review concentrates, to a significant extent, on data which might help to allow a realistic estimate of exposure in various circumstances to be made. Bio-monitoring data from real life exposures are available and these data inform the discussion. In general dermal absorption of 14C DINP decreases as the carbon chain length increases (Elsisi et al, 1989) but lies within the range of 2-4% of an applied dose in the Fischer 344 rat. DINP was rapidly metabolised and excreted once it was absorbed (McKee et al., 2002). Oral absorption of large doses demonstrates a saturateable process with increasing amounts of compound (DINP) being eliminated as dose increases. Urinary metabolites of DINP have also been quantified in several human studies with the hopes of using them as biomarkers of exposure. In Man, approximately half the ingested DINP is absorbed (unless the dose is high), rapidly metabolised and excreted in urine and faeces. Rapid absorption from an oral route of exposure is followed by metabolism to the mono-ester (MINP) which is further transformed into oxidative metabolites. A number of exposure assessment studies have been made in infants and children [and are reviewed (pp 26 35)]. In Table 10 of this document (pp. 40) the key studies showing exposure estimates to DINP from toys for children are summarised and in Table 11 data are presented that indicate that previous exposures may have been overestimated. I agree with the conclusion that there are margins of safety of at least 1000 based on the most conservative exposure data as concluded by the ECPI review. In general, when attempts are made to extrapolate from animal data to Man in terms of possible adverse effects (in the predictive sense) it is assumed that there will be defective modelling it is comparatively uncommon for data on mechanisms of action to be sufficiently Page 7 of 10

8 well understood for a direct read- over or for simple dismissal because of a well-defined mechanism not applicable to Man (see Dellarco et al, 2006 for example). Factors for inter and intra species variation are thus introduced to reduce uncertainty. Here it seems likely that a pathological change not observed in Man, occurring naturally with age and with many potential causes in the rat has limited relevance as a pathological entity on which to base calculations of safety margins. The limited exposure (both low dose and irregular) with poor absorption documented in human children make it likely that a conventional application of assessment factor guidelines (ECHA, 2010) will result in a significant overestimation of effect. Indeed in this instance, I do not consider that spongiosis hepatis in the rat has any useful predictive information to convey for human exposure. Based on this analysis total assessment factors of are appropriate for determining tolerable limits or derived no effect levels (DNELs) when using rat specific spongiosis hepatis as the point of departure. Page 8 of 10

9 REFERENCES Bannasch P, Bloch M, and Zerban H. (1981). Spongiosis hepatic: specific changes of the perisinusoidal liver cells induced in rats by N-nitrosomorpholine. Lab Invest. 44, Bannasch P and Zerban H. (1986). Pathogenesis of primary liver tumors induced by chemicals. Recent Results in Cancer Res. 100, Bannasch P and Zerban H. (1997). Spongiosis hepatic and spongiotic pericytoma, rat. In: Monographs on Pathology of Laboratory Animals, Digestive System, 2nd ed. Jones T, Popp J, Mohr U. (eds.). Springer Berlin, Heidelberg, New York, pp Berry, C. L. (1985). Solitary "necrotic nodule" of the liver: a probable pathogenesis. J Clin Pathol 38: Berry, C. L. (1987). Liver lesions in an autopsy population. Hum Toxicol 6: Biscetti F, Gaetani E, Flex A, Aprahamian T, Smith RC,. Castellot JJ (2008). Activation of PPAR-alpha and PPAR-gamma induces angiogenesis in vivo through a VEGF-dependent mechanism. Diabetes Biscetti F, Straface G, Pitocco D, Zaccardi F, Ghirlanda G, Flex A (2009a). Peroxisome proliferator-activated receptors and angiogenesis. Nutr. Metab. Cardiovasc Dis. ; 19: Biscetti F, Gaetani E, Flex A, Straface G, Pecorini G, Angelini F, Stigliano E, Aprahamian T, Smith RC, Castellot JJ, Pola R. (2009b). Peroxisome proliferator-activated receptor alpha is crucial for iloprost-induced in vivo angiogenesis and vascular endothelial growth factor upregulation. J Vasc Res. 2009; 46(2): Boorman GA, Montgomery CA and MacKenzie WF (1990). Pathology of the fischer rat, Reference and Atlas, Academic Press, San Diego, CA. Chintalgattu V, Harris GS, Akula SM, Katwa LC (2007). PPAR-gamma agonists induce the expression of VEGF and its receptors in cultured cardiac myofibroblasts. Cardiovasc Res. 74: Cho W-S, Han B-S, Ahn B, Man KT, Choi M, Oh SY, Kim SH Jeong J, Jang DD. (2008). Peroxisome proliferator di-isodecyl phthalate has no carcinogenic potential in Fscher 344 rats. Toxicology Letters 178; Dellarco VL., McGregor D., Sir Colin Berry, Cohen SM., and Boobis AR. (2006). Thiazopyr and thyroid disruption: case study within the context of the 2006 IPCS human relevance framework for analysis of a cancer mode of action for humans. Critical Reviews in Toxicology, 36; Elsisi, A.E., Carter, D.E. & Sipes, I.G. (1989) Dermal absorption of phthalate di-esters in rats.fundam. appl. Toxicol., 12, European Chemicals Agency (7 May 2012) Evaluation of New Scientific Evidence Concerning DINP and DIDP in relation to entry 52 of Annex XVII to Regulation (EC) No 1907/2006 (REACH) Page 9 of 10

10 European Chemicals Agency (December 2010). Guidance on information requirements and chemical safety assessment Chapter R.8: Characterisation of dose [concentration]-response for human health Folkman, J. (1995) Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat Med. 1995; Garcia-Tsao G, Panzini L, Yoselevitz M, West AB (1992). Bacillary peliosis hepatis as a cause of acute anemia in a patient with the acquired immunodeficiency syndrome. Gastroenterology. 102 (3): Hill AB (1965) The Environment and disease: association or causation? Proc Roy Soc Med; 58, Kaiserling E., Muller, H. (2005). Neoplasm of hepatic stellate cells (spongiotic pericytoma: A new tumour entity in human liver. Pathology Research and Practive 201, MacSween R, Burt AD, Portman BC, Ishak KG, Scheuer PJ (eds). Pathology of the Liver. 4th edition. Churchill Livingston. Edinburgh Nime F, Pickren, JW, Vana J, Aronoff BL, Baker HW and Murphy GP (1979). The Histology of Liver Tumours in Oral Contraceptive Users Observed during a National Survey by The American College of Surgeons Commission on Cancer. Cancer 44: Su Q, Benner A, Hoffmann WJ, Otto G, Pochlmayr R, Bannasch P (1997). Human hepatic pre-neoplasia: phenotypes and proliferation kinetics of foci and nodules of altered hepatocytes and their relationship to liver cell dysplasia. Virchows Arch 431: Swaen G, van Amelsvoort L (2008). A weight of evidence approach to causal inference.j. Clin Epidemiology: Page 10 of 10