Genotoxicity testing of biotechnology-derived products

Mutation Research 436 Ž1999. 137–156

Genotoxicity testing of biotechnology-derived products Report of a GUM task force 1 Elmar Gocke a

a,)

c , Silvio Albertini a , Susanne Brendler-Schwaab b, Lutz Muller , ¨ e Willi Suter d , Friedrich E. Wurgler ¨

F. Hoffmann-La Roche, Pharma DiÕision, Department of Toxicology, CH-4070 Basel, Switzerland b Bayer, D-42069 Wuppertal, Germany c Federal Institute for Drugs and Medical DeÕices (BfArM), D-13353 Berlin, Germany d NoÕartis Pharma, CH-4002 Basel, Switzerland e Inst. f. Toxikologie ETHZ, CH-8603, Schwerzenbach, Switzerland Received 21 October 1998; accepted 4 January 1999

Abstract Various aspects of genotoxicity testing of biotechnology-derived products are discussed based on information gathered from a questionnaire which was sent to about 30 predominantly European companies. Feedback was received from 13 companies on 78 compounds, mostly recombinant proteins but also on a number of nonrecombinant proteins, which had been assessed for genotoxicity in a total of 177 tests. Four of the 78 compounds appeared to elicit reproducible genotoxic effects. For one of these compounds, the activity could be related to a nonpeptidic linker molecule. No scientifically convincing rationale for the other three compounds could be established, although, at least for two compounds, their activity may be connected with the enzymaticrhormonal activity. In addition to the survey, published reports on genotoxicity testing of biotechnology products were reviewed. The data are discussed relative to whether genotoxicity testing is a valuable exercise when assessing potentially toxic liabilities of biotechnology-derived compounds. It is concluded that genotoxicity testing is generally inappropriate and unnecessary, a position which is in accordance with the available guidelines addressing this area. For the ‘average’ protein, electrophilic reactions are difficult to envision. Indirect reactions via DNA metabolism and growth regulation seem possible for only very specific proteins such as nucleases, growth factors, cytokines. No information on testing of different types of biotechnology-derived products Že.g., ribozymes, antisense-oligonucleotides, DNA vaccines. has been received in the questionnaires. Discussion of their potential to cause genotoxic changes was based on literature reports. Even for those products for which concerns of genotoxicrtumourigenic potential cannot be completely ruled out, e.g., because of their interaction with DNA metabolism or proliferation control, the performance of standard genotoxicity assays generally appears to be of little value. All information, including also information on the occurrence of genotoxic impurities, has been utilized to formulate a decision tree approach for the genotoxicity testing of biotechnologyderived products. q 1999 Elsevier Science B.V. All rights reserved. Keywords: Biotechnology products; Genotoxicity; Compilation of industry studies; Review of literature; Relevance

) 1

Corresponding author. Tel.: q41-61-688-4797; Fax: q41-61-688-8418; E-mail: [email protected] Gesellschaft fur ¨ Umweltmutationsforschung.

1383-5742r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 3 - 5 7 4 2 Ž 9 9 . 0 0 0 0 2 - 2

138

E. Gocke et al.r Mutation Research 436 (1999) 137–156

1. Introduction

For many years the evaluation of genotoxic potential has become an integral part of the toxicological investigations of low-molecular-weight chemicals. The possibility of a ‘direct’ electrophilic attack on DNA has been—and still is—the main reason of concern. Indirect interactions with eg enzymes of DNA metabolism Žprecursor synthesis, replication, repair, etc.. and with the processes of chromosome distribution have also to be taken into consideration. Regarding risk assessment, the various modes of action might lead to different risk scenarios. The potential of biotechnology products, in particular the high-molecular-weight recombinant proteins, to elicit DNA damage in somatic or germ cells cannot be compared to that of the low-molecularweight chemicals. Electrophilic reactions are difficult to envision, while indirect reactions via DNA metabolism and growth regulation seem possible only for very specific proteins. The capacity of the standard mutagenicity tests to recognise such modes of action appears to be doubtful. For this reason a ‘case-by-case approach’ has been proposed. The relevant ICH ŽInternational Conference on Harmonisation of technical requirements for registration of pharmaceuticals for human use. guideline, which was enacted in the EU in late 1997, states that ‘‘genotoxicity studies, routinely conducted for traditional pharmaceuticals are generally not applicable . . . ’’. This assessment reflects a departure from the general three Žor four. test requirements for pharmaceuticals and will probably lead to much reduced or no mutagenicity testing of new protein medicines in the future. The basis for this new guideline appears to be predominantly theoretical considerations, in combination with some experimental evidence. Previously, it had been argued that the performance of at least part of the standard battery of mutagenicity tests would be necessary, not the least to detect mutagenic impurities in the formulated preparation. A unified approach did, however, not exist and the extent of testing was rather variable. It is not surprising that in 1996 a survey among European regulatory authorities regarding their policy for mutagenicity testing of biotechnology products revealed large differences between national authorities

Žpersonal communication, J. Sims, Medicines Control Agency, UK.. During the last years enough investigations of biotechnology products have been performed to merit the collection of data and thereby obtain a more satisfactory experimental basis for outlining a meaningful and harmonised testing strategy. The GUM ŽGerman speaking section of the European Environmental Mutagen Society. has constituted a task force to evaluate the current status. Here we report the feedback on a questionnaire sent predominantly to European pharmaceutical companies and contract laboratories. Furthermore, the open literature has been searched for studies dealing with genotoxicity testing of proteinaceous compounds. Other biotechnology products such as antisense oligonucleotides or DNA vaccines, which could conceivably interact with the genetic material by different mechanisms, have not been reported to an appreciable extent and are discussed only briefly.

2. Overview of the current regulatory guidelines Genotoxicity testing of biotechnology-derived products is addressed in a number of guidelines. The most recent and probably the broadest guideline on safety testing in the biotechnology area is the new ICH guideline ‘Preclinical Safety Evaluation of Biotechnology-Derived Pharmaceuticals’ ŽCPMPr ICHr302r95, adopted in the EU September 1997.. The following three paragraphs summarize the text of this guideline in its Section 4.7 ‘Genotoxicity Studies’: The range and type of genotoxicity studies routinely conducted for pharmaceuticals are not applicable to biotechnology-derived pharmaceuticals and therefore are not needed. Moreover, the administration of large quantities of peptidesrproteins may yield uninterpretable results. It is not expected that these substances would interact directly with DNA or other chromosomal material ŽNote 3.. Studies in available and relevant systems, including newly developed systems, should be performed in those cases where there is cause for concern about the product Že.g., because of the presence of an

E. Gocke et al.r Mutation Research 436 (1999) 137–156

organic linker molecule in a conjugated protein product.. The use of standard genotoxicity studies for assessing the genotoxic potential of process contaminants is not considered appropriate. If performed for this purpose, however, the rationale should be provided. Note 3: With some biopharmaceuticals there is a potential concern about accumulation of spontaneously mutated cells Že.g., via facilitating a selective advantage of proliferation. leading to carcinogenicity. The standard battery of genotoxicity tests is not designed to detect these conditions. Alternative in vitro or in vivo models to address such concerns may have to be developed and evaluated. Comments from the task force: This guideline is applicable mainly for biotechnology-derived proteins and not for antibiotics, allergenic extracts, heparin, vitamins, cellular blood components, conventional bacterial or viral vaccines, DNA vaccines, cellular and gene therapies or antisense products. For a number of the latter products, the approach for genotoxicity testing may be quite different than that outlined in the guideline. For years it has been argued whether it might be necessary to conduct genotoxicity tests for potential genotoxic impurities of biotechnologyderived products. It has been demonstrated via spiking of proteins with potent mutagens that, e.g., the bacterial reverse mutation test is able to detect these compounds at concentration levels sometimes below analytical detection limits. However, it is highly improbable Ži. that a biotechnology-derived protein is contaminated with a potent mutagen; Žii. that the extraction or purification procedure makes use of a potent mutagen; Žiii. that extraction and purification of a protein is so drastic as to allow for the generation of a potent mutagen. As biotechnology-derived products are very diverse and a single guideline such as the ICH guideline cannot cover all different types of products. The following lists additional existing or draft EU guidelines which also refer to genotoxicity testing of biotech products. Ø Note for Guidance on Preclinical Pharmacological and Toxicological Testing of Vaccines ŽCPMPrSWPr465r95, adopted in the EU December 1997..

139

Comments from the task force: The guideline which refers among other vaccines to antigens produced by recombinant DNA technology, simply states that genotoxicity and carcinogenicity studies are normally not needed. Ø Note for Guidance on the Quality and Preclinical Safety Testing of Plasmid DNA Vaccines ŽCPMPrBWPr611r97, draft guideline.. Comments from the task force: In this guideline draft, insertional mutagenesis with its potential to promote tumour formation caused by chromosomal integration of the DNA after inoculation is addressed. It is stated that the most sensitive methods available should be used to investigate integration of the DNA after inoculation. However, no specific methods are mentioned. Ø Note for Guidance on Gene Therapy Product Quality Aspects in the Production of Vectors and Genetically Modified Somatic Cells ŽCPMP 5863r93.. Comments from the task force: This guideline draft addresses mainly product quality aspects of naked nucleic acid, complexed nucleic acid, replication deficient viral vectors and genetically modified cells. However, insertional mutagenesis is identified as a possible consequence of unintended and unexpected gene transfer. Three potential problems of random integration of vector nucleic acid are specifically described: Ži. integration in the middle of a tumour suppressor gene, abolishing its expression; Žii. integration at sites which induce cis- or transactivation of proto-oncogenes or other growth promoting genes; Žiii. integration at sites which affect cellular responsiveness to exogenous agents, such as growth factors, cytokines or hormones. The guideline then goes on in stating that since oncogenesis is a multistep process, occurrence of single-site insertional mutagenesis may only carry a very low risk for the development of tumour cells. Ø Annex on Safety Studies for Gene Therapy Products ŽCPMPrSWPr112r98, draft guideline.. Comments from the task force: This guideline draft addresses safety aspects of naked nucleic acid, complexed nucleic acid, replication deficient viral vectors and genetically modified cells. Potential insertional mutagenesis is addressed as a problem. Detection and expression of the administered nucleic acid via PCR and RT-PCR methodology in tissues

140

E. Gocke et al.r Mutation Research 436 (1999) 137–156

other than the desired site of action is felt important. A sensitive method should be used to determine whether there is covalent linkage of the nucleic acid to genomic DNA. Ø Preclinical Biological Safety Testing on Medicinal Products Derived from Biotechnology. Comments from the task force: A section on mutagenic and oncogenicrcarcinogenic potential refers to the general inappropriateness of the ‘preICH’ guidelines on mutagenic and carcinogenic potential of the EU. It is understood that the new ICH biotechnology guideline largely replaces this guideline regarding content and recommendations. In conclusion, most of the currently applicable EU guidelines on biotechnology-derived pharmaceuticals describe traditional mutagenicity and carcinogenicity testing as inappropriate. In the rather new area of transfer of genetic material for vaccination or gene therapy, the problem of insertional mutagenesis is addressed. However, specific methods to investigate insertional mutagenesis are not prescribed. All guidelines and guideline drafts can be obtained from the internet at http:rrwww.eudra.org.

3. The questionnaire Table 1 shows a representative example of the information obtained by the questionnaire. It was asked to identify the test compound as much as confidentiality would allow it. Test results were to be summarized including dose range tested, observed cytotoxicity and genotoxicity, if any. Additionally, the reasons for performing the tests and the basis for dose selection were asked for. It became apparent that the questionnaire was not always completely answered, with the least information given for the description of the test compound, date of testing and completeness of evaluation. For instance, it was sometimes not detailed whether the compound was nature identical or whether modifications Že.g., regarding glycosylation. were introduced. Most likely, the exact properties were not available to the submittors. While this lack of knowledge is unfortunate it does not preclude the inclusion of test results into the files. A few studies were reported twice Že.g., from collaborating companies.. The secondary questionnaires were omitted from the files. For five com-

pounds which yielded positive results meriting more detailed discussions; further information on test results were sought by contacting the suppliers of the data in question. Reports for a total of 81 compounds were received ŽTable 2.. Most test compounds were recombinant proteins. Twelve nonrecombinant enzymes, e.g., lipase or glucanases were reported. As test results on these molecules are also informative they were included in the table. For one compound, i.e., glucagon, recombinant as well as nonrecombinant material was assayed. Fifteen monoclonal antibodies ŽMOABs. or components of antibodies constituted another large subset. For two MOABs, it was mentioned that submissions to regulatory authorities was done without genotoxicity evaluation Žno testing was deemed necessary.. Twelve compounds were identified as cytokines. Possibly some of the compounds described as growth regulating proteins could also belong to this group. One test sample was described as a ‘live fungus’. For comparisons this sample is also included in the table even though it is not directly relevant to the present discussion. Thus, for a total of 78 compounds subjected to mutagenicity testing 177 studies were performed. Ø Sixty-nine compounds did not yield any significant finding in a total of 152 genotoxicity tests Žsee also Table 3.. Findings for the remaining nine compounds are described shortly in the following. Ø For two compounds Žone hormone, No. 1 in Table 2, and recombinant as well as nonrecombinant glucagon, No. 44. artefactual increases in the AmesrE. coli tests due to histidine or tryptophan feeding effects were seen. Ø One test sample Žnonrecombinant lipase, No. 42. caused positive effects in the mouse lymphoma ŽMLrTK. test due to a low molecular weight impurity, which was not further characterized. Ames and CA tests were negative. Ø For one compound Žantibody, No. 70. irreproducible increases were seen in the chromosomal aberration ŽCA. test with V79 cells. The Ames test was negative. Ø For another antibody ŽNo. 14. it was mentioned that the highest dose of 5000 mgrml caused a borderline effect in the CA test Žincrease from 2.25% to 5% of aberrations; significant at 5%. without S9 at the extended harvest time only. Toxicity was

Table 1 Example of the questionnaire

E. Gocke et al.r Mutation Research 436 (1999) 137–156

141

142

Table 2 Compilation of the answers received to the questionnaire Substance

Description

Date

Assay

Dose ŽHDT.

Toxicity ŽLTD.

Genotoxicity

Notes

1.1 1.2 1.3 1.4 2.1 2.2 2.3 3.1 4.1

hormone hormone hormone hormone regul prot regul prot regul prot natural recomb. fungus

n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. live fungus

86r87 90r91 87 90r91 93r94 94 93 92 95

AmesrE.c. V79rHPRT CA ŽV79. MNT Ames CA ŽHL. MNT AmesrE.c. Ames

5000 5000r10000 3500 40 600 75 17 5000 5000

no 2560r10000 2500r1000 no no slight no ) 20 no

no no no no no no no no no

histidine effect recognized ŽqS9.

5.1 6.1 6.2 6.3 6.4

antibody no description no description no description no description

n.i. n.i. n.i. n.i.

96 94 94 96 96

AmesrE.c. AmesrE.c. MLrTK CA ŽCHO. MNT rat

20 200 100 170 250

no no no no 250

no no no no no

7.1 7.2 7.3 7.4 8.1 8.2 8.3 9.1 9.2 9.3 10.1 10.2 10.3 11.0 12.1 12.2 13.1 13.2 14.1

no description no description no description no description no description no description no description no description no description no description no description no description no description antibody IFN beta 1a IFN beta 1a IFN beta 1b IFN beta 1b antibody

n.i. n.i. n.i. n.i. glycos. change glycos. change glycos. change glycos. change glycos. change glycos. change glycos. change glycos. change glycos. change

90 94 94 91 95r96 95r96 95r96 95r96 95r96 95r96 95r96 95r96 95r96

Ames V79rHPRT CA ŽCHO. MNT rat Ames MLrTK CA Ames MLrTK CA Ames MLrTK CA

5000 5000 5000 500 5000 5000 5000 5000 5000 5000 5000 5000 5000

no 1250r1666 no 250 no no no no no no no no no

no no no no no no no no no no no no no

n.i. n.i. n.i. n.i.

95r96 95r96 90 95 89

AmesrE.c. CA ŽHL. Ames CA ŽHL. CA ŽHL.

40 5 600 150 5000

no no no no no

no no no no no

HDT s10E4 higher than therap. dose HDT s max appl vol HDT s max appl vol HDT s max appl vol Bacillus thuringiensis pesticide fungus formed colonies r not incl. In evaluationrsee text HDT s max appl vol; tox of vehicle HDT s max appl vol; tox of vehicle HDT s max appl vol; tox of vehicle slight decrease of b.w. at 250; iv application

dose request by sponsor, iv application

MOAB, testing not deemed necessary sporadic increases not deemed relevant

anti cytomegalie virus MOAB, at highest dose questionable effect

E. Gocke et al.r Mutation Research 436 (1999) 137–156

No.

rec. Factor VIII rec. Factor VIII rec. Factor VIII MOAB MOAB MOAB MOAB MOAB MOAB MOAB IFN alpha IFN alpha IFN alpha IFN alpha IFN alpha IFN alpha IFN alpha IL-2 IL-2 IL-12 IL-12 peptide

n.i. n.i. n.i. n.i. n.i. n.i. n.i. modified aa sequ modified aa sequ modified aa sequ n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i.

89 89 89 90 90 90 91 92 92 92 82r84 84 83 84 84 84 84 90 89 94 94 92

AmesrE.c. CA ŽCHO. DLT AmesrE.c. CA ŽCHO. UDS hepatoc DLT AmesrE.c. CA ŽCHO. UDS hepatoc AmesrE.c. S. cer. D7 CA ŽHL. MNT V79rHPRT UDS fibrobl UDS hepatoc CA ŽHL. MNT AmesrE.c. CA ŽHL. Ames

)100 Units 65 Units

22.1 22.2 22.3 23.1 23.2 24.1 25.1 25.2 26.1 26.2 26.3 26.4 27.1 27.2 28.1 28.2

GCSF GCSF GCSF peg GCSF peg GCSF peg OB protein peg IFN alpha peg IFN alpha DNase DNase DNase DNase IFN alpha IFN alpha IFN gamma IFN gamma

n.i. n.i. n.i. glycos. change glycos. change glycos. change glycos. change glycos. change n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i.

89 89 89 89 89 96 96 96 92 92 92 92 85 85 85 85

Ames CA ŽCHL. MNT Ames CA ŽHL. AmesrE.c. AmesrE.c. CA ŽHL. Ames MLrTK CA ŽHL. MNT Ames MNT Ames MNT

2000 2000 1500 2000 1600 533 533 5000? 2=10E6 Units 2=10E4 Units 10E8 Units 2=10E6 Units 2=10E8 Units 2=10E8 Units 10E6 Units 10E7 Units 250 20 5000

no 65rno no no no no no no no no no no no no slight slight slight no no no no no

no no no no no no no no no no no no no no no no no no no no no no

110 26 11 1250 50 5000 1125 50 5000 1000 2000 40 3 2=10E8 Units 200 20

no no no no no no no no no no no no no no no no

no no no no no no no no no no no no no no no no

murine anti TNF murine anti TNF murine anti TNF murine anti TNF humanized anti TNF humanized anti TNF humanized anti TNF

s.c. application

slight tox in preincubation test HDT s max appl vol HDT s max appl vol HDT s max appl vol

HDT s max appl vol HDT s max appl vol

E. Gocke et al.r Mutation Research 436 (1999) 137–156

15.1 15.2 15.3 16.1 16.2 16.3 16.4 17.1 17.2 17.3 18.1 18.2 18.3 18.4 18.5 18.6 18.7 19.1 19.2 20.1 20.2 21.0

HDT s max appl vol HDT s max appl vol HDT s max appl vol HDT s max appl vol, iv application

143

144

Table 2 Žcontinued. Substance

Description

Date

Assay

Dose ŽHDT.

Toxicity ŽLTD.

Genotoxicity

Notes

29.1 29.2 30.1 30.2 30.3 30.4 30.5 30.6 31.1 31.2 31.3 32.1

n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i.

94 94 85 86 85 85 87 86 87 87 87 89

Ames MNT Ames Ames CA ŽHL. CA ŽHL. UDS in vitro MNT Ames UDS in vitro MNT Ames

100 2 100 400 50 200 600 20 100 150 2.5 5000

no q? no no no no no no no no toxic no

no no no no no no no no no no no no

HDT s max appl vol HDT s max appl vol, iv application HDT s max appl vol HDT s max appl vol HDT s max appl vol HDT s max appl vol HDT s max appl vol HDT s max appl vol, iv application HDT s max appl vol HDT s max appl vol toxicity: deaths and bone marrow tox

89

CA

1% volrvol

no

no

HDT s max appl vol

89

UDS in vitro

1% volrvol

no

no

HDT s max appl vol

33.1 34.1 35.1 36.1 37.1 37.2 38.1 38.2 39.1 39.2 40.1 40.2 41.1 41.2 42.1 42.2 42.3 43.1 44.1

IFN omega IFN omega r-TPA r-TPA r-TPA r-TPA r-TPA r-TPA r-TNF r-TNF r-TNF galactosidase non rec glycos change galactosidase non rec glycos change galactosidase non rec glycos change pectinase non rec, lipase non rec tannase non rec tannase non rec xylanase non rec xylanase non rec xylanase non rec xylanase non rec glucanase non rec, glucanase non rec, arabinase non rec, arabinase non rec, lipase non rec lipase non rec lipase non rec lipase non rec lipase non rec lipase non rec glucagon rec

90 89 92 91 91 91 94 94 94 94 93 93 94 94 95 95 95 96

Ames Ames Ames Ames Ames CA Ames CA Ames CA Ames CA Ames CA Ames CA MLrTK MLrTK AmesrE.c.

5000 1500 5000 5000 5000 5000 5000 5000 5000 5000 5000 5000 8500 450r5000 5000 5000 5000 160r25

no no no no no no no no no no no no no 150r5000 no 2500rno 2500 120r15

no no no no no no no no no no no no no pos no no pos no no

44.1a

glucagon non rec

44.2 44.3

glucagon rec glucagon rec

32.2 32.3

n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. n.i. glycos change glycos change n.i. n.i. n.i. n.i.

AmesrE.c. 89 88r89

CA ŽHL. MNT

no 2500 200

464r1056 no

weak pos pos

HDT s max solubility

treat and plate assay yS9 extended harvest time only

pos. effect due to low mol weight fraction histidinertryptophan effect recognized qS9 histidinertryptophan effect recognized qS9 only yS9; hormonal effect? in males only; hormonal effect?

E. Gocke et al.r Mutation Research 436 (1999) 137–156

No.

88r89 96 96

MNT AmesrE.c. MLrTK

45.3 46.1 46.2 46.3 46.4 47.1 47.2 47.3 47.4 47.5 48.1 49.1 50.1 51.1 52.1 53.1 54.1 55.1 55.2 56.1 56.2 57.1 57.2 58.1 58.2 59.1 59.2 60.1

reg. prot reg. prot reg. prot reg. prot reg. prot reg. prot reg. prot reg. prot reg. prot reg. prot reg. prot enzyme enzyme enzyme enzyme enzyme enzyme enzyme enzyme enzyme enzyme enzyme enzyme reg. prot reg. prot reg. prot reg. prot reg. prot

96 96 96 96 96 96 96 96 96 96 96 96 96 96 95 95 95 91 91 95 95 95 95 96 96 96 96 96

61.1 61.2 61.3 62.1 62.2 62.3 63.1 64.1 65.1 66.1

reg. prot reg. prot reg. prot enzyme enzyme enzyme reg. prot enzyme reg. prot peptide

66.2

peptide

200 2000r850

no no nor1700

pos no pos

MNT AmesrE.c. MLrTK CA ŽHL. UDS hepatoc AmesrE.c. MLrTK CA ŽHL. MNT UDS ex vivo MLrTK CA ŽHL. CA ŽHL. CA ŽHL. CA ŽHL. MNT UDS hepatoc CA ŽHL. MLrTK MNT UDS ex vivo MNT UDS ex vivo AmesrE.c. CA ŽHL. AmesrE.c. CA ŽHL. UDS hepatoc

3000 300 300 300 5000 5000 5000 2000 2000 5000 5000 5000 5000 5000 ? 250 5000 5000 1000 2000 2000 2000 5000 5000 5000 5000 15

no no no no 60 no no no no no no no no no no no 250 no no 1000 no no no no no no no no

no no no no no no no no no no no no no no no no no no no no no no no no no no no no

93 95 94 92 92 92 90 93 95 ?

Ames MLrTK CA ŽCHO. MLrTK CA ŽHL. MNT CA ŽHL. MLrTK UDS hepatoc AmesrE.c.

5000 5000 5000 ? ? ? 5000 5000 1000 5000

no no no no no no no no 100 ?

no no no no no no no no no no

?

CHOrHPRT

1000

?

no

in males only; hormonal effect? HDT s max appl vol HDT s max appl vol; pos only qS9 Ž2.5 to 3.5 fold.rboth exp HDT s max appl vol HDT s max solubility HDT s max solubility HDT s max solubility HDT s max solubility and toxicity

HDT s max appl vol Ž2=20 mlrkg. HDT s max solubility

weak pos at one dose not confirmed in repeat; HDT s max solubility

HDT s max appl vol HDT s max appl vol HDT s max appl vol

3 lots tested, 2 impure lots pos in TA98, TA1538 145

glucagon non rec reg. prot reg. prot

E. Gocke et al.r Mutation Research 436 (1999) 137–156

44.3a 45.1 45.2

146

Table 2 Žcontinued. Substance

66.3 66.4 67.1 67.2 68.1 69.1 70.1 70.2

peptide peptide reg prot reg prot antibody antibody antibody antibody

71.1 71.2 72.1 72.2 73.1 74.1 75.1

reg. prot reg. prot antibody antibody antibody antibody antibody

75.2

antibody

75.3

antibody

75.4

antibody

76.1 77.1 77.1 78.1 78.2 78.3 79.1 79.2 80.1 80.2 81.1 81.2 81.3

Fab Fab Fab reg.prot. reg.prot. reg.prot. antibody antibody reg.prot. reg.prot.

Description

rh-IGF-1 rh-IGF-1 chimaeric chimaeric chimaeric chimaeric

rhNGF, n.i. rhNGF, n.i. humanized humanized humanized humanized humanizedr qlinker humanizedr qlinker humanizedr qlinker humanizedr qlinker diff. linker diff. linker

humanized humanized n.i. n.i. modif. aa modif. aa modif. aa

Date

Assay

Dose ŽHDT.

Toxicity ŽLTD.

Genotoxicity

? ? 95 95

CA ŽCHO. MNT CA ŽCHL. MNT

2780 3000 487 97

? ? no no

no no no no

96 97 97

Ames Ames CAŽV79.

5000 5000 5000

no no somerno

no no no

97 97 94 97 94 96 93

AmesrE.c. CAŽHL. AmesrE.c. CAŽHL. AmesrE.c. CAŽHL. AmesrE.c.

3800 380 5000r3750 5000 5000r3750 500r1000 5000r3750

no no no no no no no

no no no no no no no

93

CAŽHL.

4500

no

pos.

93

MLrTK

2500

no

no

93

MNT

100

no

no

94 94 94 96 96 96 95 95 92 92 90 90 90

CAŽHL. Ames CAŽHL. Ames transg.mouse transg.mouse CAŽCHL. MNT CAŽCHL. MNT Ames CAŽCHL. MNT

4500 5000r3750 4500 1000 100 acute 10 rr 14 days 590 118 487 97.4 5000 153.5 50 Živ.

no no no

no no no no no no no no no no no no no

Notes

no testing deemed necessary

no no no no no 153.5 Žno.

at 20 h yS9 some reduction of MI, reff. attr. to solvent? some irreprod. Increases of CA frequency HDT s max appl vol HDT s max appl vol

HDT s max appl vol

pos at ) 756qS9 only; effect due to linker molecule shown in ext. testing

as No 75 but not linkered as No 75 but differently linkered as No 75 but differently linkered

HDT s max appl vol, solubility HDT s max appl vol, solubility HDT s max appl vol, solubility HDT s max appl vol, solubility

at 50 mgrkg some decreased activity

E. Gocke et al.r Mutation Research 436 (1999) 137–156

No.

E. Gocke et al.r Mutation Research 436 (1999) 137–156

147

Table 3 Summary information on types and numbers of tests performed, number of tests with positive or problematic observations Žfor abbreviations see bottom of Table 2. Test

Ames or AmesrE. coli CA in vitro ŽHL, CHO, or CHL. MLrTK V79rHPRT or CHOrHPRT MNT in vivo Žmouse or rat. UDS in vitro Žfibrobl. or hepatoc.. UDS ex vivo S. cereÕisiae D7 DLT transgenic mouse

No. of tests

57 54 16 4 27 11 3 1 2 2 177 tests Ž78 compounds.

No. of tests with positive result a Žcompound no. in Table 2. 3 ŽNos. 41, 44, 75. 1 ŽNo. 45. 2 ŽNo. 44.

No. of tests with other problematic observationsb Žcompound no. in Table 2. 3 ŽNos. 1 and 44. 2 ŽNos. 14 and 70. 1 ŽNo. 42.

1 ŽNo. 60.

6 tests Ž4 compounds c .

7 tests Ž6 compounds c .

a

Those tests which appear to represent valid positive results. Those tests for which irreproducible positive results, effects due to artefacts, effects due to impurities are observed. c One compound ŽNo. 44. appears in both categories; for the nine compounds in these two categories a total of 25 tests were performed Ž12 negative, 13 positive or equivocal.. b

lowrabsent. The effect was not thought to be relevant. No repeat and no other test were reported. Ø One regulatory protein ŽNo. 60. elicited a weak positive response in an in vitro unscheduled DNA synthesis ŽUDS. test with rat hepatocytes at one dose level only; an independent repeat was negative. No other tests were reported. Ø One compound Žregulatory protein, No. 45. caused a 2.5- to 3.5-fold increase in two independent

MLrTK experiments in the presence of S9 only, at the maximal dose level Ž850 mgrml. selected for evaluation. Toxicity was apparent only at a higher dose of 1700 mgrml. The Ames test and the micronucleus ŽMNT. test were negative. Ø One nonrecombinant lipase ŽNo. 41. was reported to be uniquely positive in the CA test at an extended harvest time in the absence of S9. Toxicity was apparent. It was suggested, that the effect was

Notes to Table 2: Abbreviations: For column ‘substance’: MOAB: monoclonal antibody; IFN: interferon; IL: interleukin: GCSF: granulocyte colony stimulating factor; peg: pegylated; r-TPA: recombinant tissue plasminogen activator; r-TNF: recombinant tumour necrosis factor; rh-IGF: recombinant human insulin like growth factor; rh-NGF recombinant human neurotrophic growth factor; Fab: fragment of antigen binding. For column ‘description’: n.i.: nature identical; glycos.: glycosylation; aa: amino acids. For column ‘date’: the yearŽs. of performance of the test is given. For column ‘assay’: AmesrE.c.: Ames and E. coli reverse mutation test; V79rHPRT: gene mutation test ŽHGPRT locus. with V79 cells; CAŽV79.: chromosomal aberration test with V79 cells, HL: human lymphocytes; CHO: Chinese hamster ovary cells; CHL: Chinese hamster lung cells; MNT: micronucleus test in vivo; MLrTK: mutation test thymidine kinase locus with mouse lymphoma cells; DL: dominant lethal test; UDS: unscheduled DNA synthesis test; S.cer. D7: genotoxicity test with Saccharomyces cereÕisiae D7. For column ‘DoserHDT’: the highest doses tested ŽHDT. are generally given in mgrml, mgrplate or mgrkg as appropriate for test system. If two doses are given the first indicates the HDT without S9, the second with S9. For column ‘toxicityrLTD’: if toxic effects are reported the lowest toxic doses ŽLTD. are given. For column ‘notes’: HDT s max appl volume: the highest dose tested was dictated by the maximal volume which could be applied. Shaded rows: not included in overview data Žsee text.. Further abbreviations: PCR: polymerase chain reaction; RT-PCR: reverse transcriptase PCR; ROS: reactive oxygen species; NO: nitric oxide; PCE: polychromatic erythrocyte; NCE: normochromatic erythrocyte; PHA: phytohemagglutinin; RET: reticulocyte; LPS: lipopolysaccharide; SCE: sister chromatid exchange

148

E. Gocke et al.r Mutation Research 436 (1999) 137–156

due to enzymatic action. The mode of action could relate to release of lysosomal enzymes, e.g., endonucleases, as a consequence of destruction of lysosomal membranes or general disturbance of the cellular and chromosomal architecture. Ø Glucagon ŽNo. 44. induced borderline effects regarding chromosome aberrations in human lymphocytes in the absence of S9. The effect was predominantly due to a tendency of increased polyploidy. However, effects were not fully comparable in two independent experiments and no final conclusion can be drawn. Toxicity was recognizable by reduced mitotic indices. Since the effect appears to be abolished by S9, which is known to cleave the peptide and abolish its hormonal activity, an involvement of the physiological activity was thought likely. In the in vivo MNT test there was a weak positive effect at the two highest dose levels at all three sampling times. Results were clearly significant, but historical control values were only barely exceeded. Effects were obvious in males while females responded at most marginally. Subsequently it was shown that nonrecombinant glucagon of pancreatic origin caused increases of the frequency of micronucleated PCE which were quite comparable to those seen with recombinant glucagon. No effects on erythrocyte proliferation was apparent. No pharmacologic rationale for the difference between sexes is apparent. There is no indication whether the polyploidy-inducing effects seen in vitro could be involved in the in vivo action. Ø A modified antibody fragment ŽNo. 75. caused a positive effect in the CA test with human lymphocytes in presence of S9. The Ames test, the MLrTK test and the MNT test were negative. In further experimentation it could be shown that the positive effect in the CA test correlated with the presence of a specific linker molecule: testing the Fabs without this linker or with a modified linker showed no activity, and the linker ŽOPDM, N, N X-1,2-phenylenedimaleimide, a bifunctional cross-linking agent., when tested separately, proved to be positive in absence and in presence of S9 activation. Thus, the genotoxic effect is a consequence of a very specific chemical modification of the protein. The apparent limitation of genotoxicity of the antibody fragment to human lymphocytes might be influenced by its targeting of a receptor present only on human leuko-

cytes, i.e., the pharmacologic action of the biotechnology product may be decisive for internalisation of the linker molecule. In conclusion, four of the 78 compounds appear to elicit reproducible effects ŽNos. 41, 44, 45, and 75.. Of these, for one sample ŽNo. 75. the activity could be related to a nonpeptidic linker molecule known to be electrophilic. No rationale for the other three compounds could be established, although, at least for two compounds ŽNos. 41 and 44., a connection to the enzymaticrhormonal activity was thought probable. One of these two compounds was nonrecombinant ŽNo. 41., for the other recombinant as well as nonrecombinant preparations caused the positive effects. A further four compounds ŽNos. 1, 14, 60, and 70. elicited increased mutation frequencies which are likely to be due to statistical chance occurrence or artefacts. For one compound ŽNo. 42., the effect in the MLrTK assay was traced back to an impurity.

4. Survey of published genotoxicity testing of proteins 4.1. Cytokines, growth factors Cytokines act as signals during the various steps of the inflammatory response w1x. Inflammatory processes are concomitant with production of reactive oxygen species ŽROS. by specialized cells such as the phagocytic leukocytes Že.g., neutrophils, monocytes, macrophages, and eosinophils.. ROS clearly have the potential to damage the genetic material. Thus, it is conceivable that after exposure to exogenous cytokines the induction of cellular responses involved in inflammation indirectly leads to enhancement of oxidative DNA lesions. A multitude of cytokines is produced by many different cells, they influence the synthesis of other cytokines leading to activation of the cytokine network. Thus, exposure of responsive target cells to an exogenous cytokine might trigger complex cascades of other cytokines. Furthermore, cytokines act as growth factors on many cells types and it might be hypothesized that the occurrence of ‘spontaneous’ mutations is modulated if the growth status of the cells is changed. Clearly,

E. Gocke et al.r Mutation Research 436 (1999) 137–156

it is of considerable scientific interest to understand these processes and to determine their effects on genetic stability—if any. Additionally, the potential for a promotor effect of growth factors in the tumorigenic process presents an obvious possibility. The twelve cytokine compounds reported in this survey did not induce gene mutations or chromosomal or DNA damage in a variety of standard test systems under the conditions of standard testing. Lazutka w2x recently reviewed the published literature of the genetic toxicity of cytokines. Most studies were performed with interferons in human peripheral lymphocytes and as genetic endpoint, sister chromatid exchanges ŽSCE. were investigated. Increases, as well as decreases of the SCE frequency have been observed. For IL-2 ŽT cell growth factor. Morris et al. w3x report, for instance, an increase of the SCE frequency by a factor of about 1.5 in PHA stimulated rat spleen lymphocytes. This effect was down modulated by concomitant addition of 2-mercaptoethanol. Chromosomal aberrations were investigated in six studies Žfor references, see Ref. w2x., four of these showed no effect while two showed increased values of aberrant cells. However, both these studies are of limited value for estimation of risk. The first w4x was an investigation of bone marrow cells in chronic myelogenous leukemia patients where other influences are likely to play significant roles. The second was a study with Epstein–Barr transformed cell lines from cancer patients and asymptomatic family members w5x. The conclusions drawn by the authors have been questioned by Lazutka w2x for various methodological reasons. Green et al. w6,7x, Delaney et al., w8x and Dunger et al. w9x have reported evidence for the induction of DNA damage, as determined by the comet assay in islet cells from unweaned rats, after exposure to several cytokines ŽIL-1, TNF-a, IFN-b .. These cells were chosen based on knowledge of cytokine involvement in the autoimmune disease insulin-dependent diabetes mellitus. The increased level of damage was at least partially attributable to enhanced production of nitric oxide generated by an inducible nitric oxide synthase. It has been established in recent years that inducible NO synthase is highly regulated by cytokines in a number of tissues such as macrophages, hepatocytes, fibroblasts and others w10x. Besides being an essential signalling molecule for

149

various physiological functions, this free radical is also responsible for the cytotoxicity of macrophages. The NO-radical might not only be toxic for invading microorganisms but also for the cells which produce it and for neighbouring cells. Interestingly, it has been shown that neutrophils actively accumulate vitamin C when exposed to pathogenic bacteria, possibly to enhance their antioxidative defense w11x. Ohshima and Bartsch w12x discussed the possible involvement of NO in the carcinogenicity of chronic infections and inflammatory processes. In the present context it may be notable that the standard process of stimulation of lymphocytes by mitogens ŽPHA, LPS. is also reported to induce the synthesis of the inducible NO synthase w13x. The potential to modulate mutation frequencies as a possible consequence of the growth factor-property of the cytokines, specifically erythropoietin ŽEPO., has been discussed by Yajima et al. w14–16x. They observed elevated frequencies of micronuclei in peripheral reticulocytes of mice at 48, 72, 96 h after single doses or after six daily doses of erythropoietin Žabout a 5-fold increase of spontaneous levels.. No mutagenicity in the Ames or in vitro CA tests and also no increase of chromosomal aberrations in vivo in bone marrow or spleen cells were seen. Surprisingly, no or only marginal increases in the frequency of micronucleated PCE were apparent in bone marrow. Notably, the RET ratio in peripheral blood increased about fourfold, while the PCErNCE ratio in bone marrow was not affected. The spleen weight was increased up to 5-fold. It is hypothesized that EPO perturbs a step in extramedullary maturation of erythrocytes, possibly the enucleation step. More extensiverrapid cycling might lead to enhanced rates of mistakes andror reduced repair capacity, rates of micronuclei formation might be different in extramedullary erythropoesis or it might be that elimination of micronucleated erythrocytes functions differently. Weak effects of extensive bleeding on the ‘spontaneous’ rate of micronucleated PCE in bone marrow have been reported w17x. Additionally, they observed an enhancement of the genotoxic effects in mutagen-treated animals after prior bleeding or EPO treatment. Similar effects in peripheral PCEs have been reported w18x for mice after bleeding, phenylhydrazine-induced hemolysis or splenectomy. Holden w19x discusses anemia as underlying reason for the

150

E. Gocke et al.r Mutation Research 436 (1999) 137–156

induction of micronuclei in phenolphthalein-treated mice. Clastogenic effects of pituitary or recombinant growth hormone ŽrGH. have been described w20x in Snell dwarf mice Ždoubling of micronucleated PCE frequencies. and CHO cells Ždoubling of aberrant cells.. Lymphocytes of short normal children undergoing treatment with recombinant growth hormone were reported to show enhanced chromosome fragility w21x. While spontaneous frequencies of aberrant cells were not increased after 6 or 12 months treatment, there was a weak tendency of increased breakage in bleomycin-, but not in aphidicolin-exposed cells. The authors claim to have also seen an increase in chromosomal rearrangements following rGH therapy. SCE frequencies were not affected. The extent of the experimentation leaves some doubt upon the validity of conclusions. A discussion about an increased risk of leukemia in children treated with growth hormone during their childhood has been ongoing for several decades. Most recently a reanalysis w22x showed that the incidence of leukemia is comparable to that in the general population of age-matched children. High levels of endogenous insulin-like growth factor-1 appear to carry an increased susceptibility to breast cancer in premenopausal women w23x, which could be seen as a concern for potential therapies with this specific growth factor. In principle, the potential of an excessive growth stimulation to contribute to tumorigenic processes is obvious w24,25x. Human therapy with cytokines or growth factors would generally be aimed at restoring deficient physiological functions which a priori is not to be regarded as posing an adverse risk. On the other side, it may not be possible to exactly simulate the normal physiological functions. Growth hormone therapy may use higher than normal doses and could last for extended periods which might substantially alter the chance of growth of initiated cellsrclones while cytokine therapy would generally last only short periods, which would be of lower concern. With regard to studying growth factors in rodent bioassays, it is to be noted that such animal experiments are usually ‘technically’ inappropriate due to species specific action and antigenic properties of the human growth factors which will rapidly abolish any pharmacological growth effects in rodents.

In conclusion, a potential genotoxicrcarcinogenic risk via, e.g., increased levels of NO or ROS or changes in growth status after cytokine or growth hormone cannot be totally excluded at the present stage of knowledge. It appears possible that under certain conditions of treatment somatic cells could undergo inflammation andror transient proliferative stimulation processes which may indirectly lead to formation or fixation of DNA alterations. Based on these considerations, it is clear that performance of standard mutagenicity tests is not likely to provide an appropriate understanding of the complex interactions of cytokines with their targets, of the subsequent responses of the cells and of the toxicological relevance of any damage to the genetic material arising more or less indirectly as a consequence of the exposure. 4.2. DNA degrading enzymes Introduction of DNA strand scissions by nucleases has been established to lead to chromosomal aberrations. The groups of Obe and others have investigated the characteristics and frequencies of aberrations as a function of the cleavage patterns of various restriction endonucleases Žfor references, see Ref. w26x.. Folle et al. w27x electroporated nonrecombinant bovine pancreatic DNAse I into S-phase CHO cells and observed enhanced chromosomal breakage. Prerequisite for induction of aberrations is an efficient mechanism to enter the cells and reach the nuclear DNA. Treatment of cells in vitro with hypertonic solutions, streptolysin O, Sendai virus, or physical poration methods Že.g., electroporation, scrape loading. facilitates uptake of active enzymes. Of the proteins reported in Table 2, one is identified as a human recombinant DNAse ŽNo. 26, see Ref. w28x.. The enzyme is therapeutically used for the treatment of lung obstruction in cystic fibrosis patients. No genotoxic response was evident after exposure of bacteria, rodent or human cells in vitro or bone marrow cells in vivo under standard conditions for determining mutagenic potential. Small extracellular ribonucleases from Bacillus species have been shown to induce the SOS response and also mutations in bacterial test systems w29,30x. These enzymes are used in bioorganic chemistry. Studies with mammalian systems have not come to our attention.

E. Gocke et al.r Mutation Research 436 (1999) 137–156

In conclusion, nuclease enzymes are capable of causing chromosomal damage if they reach the DNA. Assessment of the likelihood of intracellular uptake during the intended medical use is of pivotal importance.

151

An investigation into the cytotoxic reactions experienced after vaccination of patients with DNA constructs containing regulatory genes of HIV-1 was recently reported w39x. 5.2. Antisense oligonucleotides

4.3. Miscellaneous nonrecombinant proteins Three lipases were evaluated Ži. in the Ames test, Žii. in the Ames, MLrTK and CA tests, and Žiii. in the Ames, MLrTK, CA, and MNT tests, respectively w31–33x with no adverse findings. Ishidate w34x reported negative findings in the CA test for acylase, b-amylase, lysozyme and naringinase while cellulase induced structural and numerical aberrations. A labile impurity or preservative present in one commercial preparation of b-glucuronidase induced mutations in the Ames test strains TA100 and TA1535 w35x.

5. Other biotechnology products 5.1. DNA Õaccines The prime safety concern for DNA vaccines is the possibility of integration into the genome of the host cell with the chance of inactivation of tumor suppressor genes or activation of oncogenes Žsee also section on biotech guidelines.. No genotoxicity evaluation for this class of ‘compounds’ has been described in the answers to the questionnaire. Standard mutagenesis testing is inappropriate to provide a relevant basis for risk assessment. The problem has been discussed in a number of publications w36,37x. Experimental determinations of insertion rates are feasible using modern molecular biology techniques such as PCR. It has been estimated w36,38x that random insertion in any given gene is likely to occur at frequencies which are several orders of magnitude lower than typical frequencies of spontaneous mutations. Further assessment is definitely needed. Since integration is probably influenced by the sequence of the DNA vaccine and may in effect not be completely at random, it is conceivable that insertions at specific locations may happen at much higher frequencies.

Use of antisense oligonucleotides for blocking expression of specific genes is promising to become a new modality of human therapy. Principally, translation is inhibited by site specific hybridisation of approximately 15-meric oligonucleotides to mRNA w40x. Two issues are of concern to the genotoxicologist: Ža. degradation products of the therapeutic oligonucleotide, possibly being nucleotide analogues, might become integrated into newly synthesized DNA which in consequence could cause mispairing and Žb. site-specific mutations might be induced by for example triplex formation of the oligonucleotide with the DNA fiber. Both aspects, in principle, might be accessible to standard mutagenesis testing when sequence specificity, uptake of the oligonucleotides and degradationrmetabolism are correctly addressed. Wang et al. w41x provided evidence that triplex formation of a 30-base oligonucleotide can lead to point mutation induction. They proposed a mechanism involving ‘gratuitous repair’ w42x when transcription is stalled at the triplex site. Shorter oligonucleotides possessed reduced mutation inducing activity, correlating with their reduced ability to block transcription. Such site specific action is of relevance to tumour formation if the potentially induced mutations occur at a critical region of an oncogenertumour suppressor gene in a susceptible somatic cell. Oligonucleotide sequences as large as 15-, 20- or 30-mers are supposed to act highly site specific thus probably excluding a risk of tight triplex formation and blockage at nontarget specific stretches of DNA. The possibility of triplex formation has to be discussed on a case-by-case basis taking into account the DNA sequence context and exposure conditions. Further investigations on the mode of action are needed. 5.3. Ribozymes Site specific cleavage of RNA molecules by ribozymes represents another new technology for

152

E. Gocke et al.r Mutation Research 436 (1999) 137–156

modulation of gene expression w40,43,44x. Specific ribozymes can also cause cleavage of DNA w45x. If therapeutic ribozymes would show residual interaction with DNA, similar considerations as for restriction enzymes Žsee above. would be applicable. No discussion of this aspect has reached our attention.

6. Impurities The employment of in vitro mutagenicity tests for detection of genotoxic impurities has been cited as one reason to not completely discontinue genotoxicity testing of biotechnology-derived proteins w46,47x. In the answers to our questionnaire one example was given where a low molecular weight impurity Žwhich could be removed by gel filtration. caused genotoxic effects of a Žnonrecombinant. protein preparation. It was not investigated what type of impurity was responsible. Notably, the Ames test, thought to be most appropriate to detect impurities, did not respond to the impurity. During synthesis of ‘traditional’ chemicalsr pharmaceuticals the use of reactive, low molecular weight chemicals is imperative. Such chemicals might remain as potentially genotoxic impurities in the final product Žsee ICH guideline ‘Impurities in New Drug Products’ CPMPrICHr282r95.. There are many—usually anecdotal—experiences where the genotoxicity of new products was traced back to contaminants of the tested material.

Conversely, cell lysis by physical or chemical methods, extraction, preparation and purification of biological macromolecules has to avoid harsh conditions. As such the use of reactive Želectrophilic. substances comparable to chemical synthesis processes is not likely to be found during downstream processing. Normal elution conditions during chromatographic purification are related to the stability of protein molecules and are therefore close to physiological conditions w48x. Impurities in the production of a biotherapeutic include proteins, nucleic acids, viruses Žin cell culture., endotoxins and substances derived from processing w49x. The first three types of impurities are clearly of toxicological relevance but would not be identifiable as mutagens by standard mutagenicity tests—even if they should pose a genotoxicr carcinogenic risk. Endotoxins derived from the cell wall of Gram negative bacteria are of relevance as pyrogenic substances. Because humans are extremely sensitive to endotoxins, endotoxin removal is essential and well established w49x. To our knowledge no genotoxic activity has been ascribed to such endotoxins. The remaining group of impurities Žsubstances derived from processing. comprises cell culture additives, detergents, stabilizing agents, leachables from membranes and chromatography media, and leachables from equipment surfaces w48x. A list of common materials is given in Table 4. None of the compounds listed poses special concern to the genetic toxicologist, with the possible exception of

Table 4 List of common process additives potentially present as impurities in biopharmaceutics Compounds added during fermentation

Compounds added during downstream processing

Leachables

anti-foam substances proteins, etc. for better growth

urea guanidinhydrochloride

antibodies from chromatography columns protein A from protein A columns

potassium rhodanide copper salts mercaptoethanol dithiothreitol glutathione triton detergents tween detergents cetyltrimethylammoniumbromide tris

phenyl derivatives from hydrophobic chromatography sugars from Sephadex and Superdex columns trimethylamine from Q-sepharose components of cleaningrstorage solutions ŽNa-azide.

E. Gocke et al.r Mutation Research 436 (1999) 137–156

sodium azide. Extensive washing of the columns after storage effectively removes this potential impurity. Generally, a purification by gel filtration represents the final preparatory step in order to remove low-molecular-weight contaminants. It can be concluded that the use of in vitro genotoxicity tests for the detection of mutagenic impurities does not seem as justified for biotechnology-derived macromolecules as it is for ‘standard’ chemicalsrpharmaceuticals. It should be noted that if the product is prepared from ‘unusual’, not well characterized, organisms Ži.e., molds, plants, with potential to contain toxins. or if chemical modifications such as introduction of linkers Žas in sample No. 75. are employed in the preparation of the product, special consideration has to be given to impurities potentially present in the starting material or introduced by the respective chemical processes. 7. Conclusions To our knowledge no similar compilation of data from genotoxicity testing of biotechnology-derived products is available in the literature. Test data on 78 compounds from 13 mostly European companies were received. Testing was performed in the years 1982 to 1996. Discounting artefacts and positive results due to impurities, four of these 78 compounds appear to elicit reproducible genotoxic effects Žin CA, MNT, MLrTK assays.. One compound was a nonrecombinant enzyme, for a second one, nonrecombinant as well as recombinant preparations induced positive effects. Thus—of course not surprisingly—the process of genetic recombination cannot be implicated as giving rise to genotoxic properties of the product. Notably, one of the four positive biotechnologyderived products was chemically modified with a cross-linking agent and the genotoxic activity was related to the chemical modification. Thus, special consideration has to be given to proteinaceous compounds if they are modified with chemically reactive substances. For the three other positive compounds, no rationale for the genotoxic effects can be established although indirect, pharmacology-related modes of action are most likely responsible. Altogether, the fraction of positive compounds presented by this review of biotech pharmaceuticals

153

is much lower than the approximately 20% of compounds with positive genotoxicity results for ‘traditional’ low molecular weight pharmaceuticals. The latter estimate is based on personal communication by various colleagues from industry and contract labs and on the experience with submissions to regulatory authorities. For example 47 of 161 compounds Žs 29%. tested by an industrial laboratory in both the Ames and the CA test have yielded positive responses in at least one assay. In a review of 335 new pharmaceuticals submitted for registration to the German national authorities between 1990 and 1997 a fraction of 22% of in vitro clastogens was noted w50x. Riskrbenefit considerations in genotoxicity cannot be performed without appreciation of pharmacologic and toxicokinetic properties. This is true for any drug but especially obvious for recombinant proteins. It does not make sense to test these compounds in assays primarily designed to detect electrophilic substances capable of damaging the DNA molecules—either alone or after S9 metabolic activation. It does not make sense to evaluate cytokinergrowth factors in cells which are not responsive to their action but it also seems hardly possible to assess their risk of ‘indirect’ genotoxic action by experimentation with isolated responder cells. The complex interactions of the cytokine network make any prediction to the human exposure situation utterly naıve ¨ and simplistic. Further, it does not make sense to test enzymes capable of degrading DNA in assays without knowledge of their capacity to enter the cells and the nucleus. Even if such knowledge exists for in vitro conditions, this is likely of little value for the in vivo situation. The same holds true for ribozymes, antisense oligonucleotides or DNA vaccines. Standard in vivo genotoxicity tests are restricted to a few organs Žbone marrow, peripheral blood, liver.. If cells from these organs should coincide with the target cells of pharmacologic action Že.g., as for erythropoietin. a potential indirect genotoxic action could be investigated under comparatively more relevant conditions. Still, it remains questionable whether such information is in any way predictive of a carcinogenic risk. Due to the immunogenic properties of human derived products to rats or mice, the traditional two year bioassays for tumorigenicity are,

154

E. Gocke et al.r Mutation Research 436 (1999) 137–156

however, often inappropriate as well. New models using transgenic animals have been proposed as replacements of the standard bioassay. It is not obvious that these approaches could help in assessing the tumourigenic liabilities of biotechnology products. The question arises, whether performance of any currently available in vivo mutagenicity assay can be recommended for risk identification or assessment of a genotoxicrcarcinogenic liability of biotechnology-derived products. Using recently developed systems—such as the Comet Žsingle cell gel electrophoresis. assay or the Muta MouserBig Blue assays—one would generally be able to investigate DNA damage or gene mutations in the target organ for pharmacologic action. In case ROS are thought to be produced upon treatment with a biotech product Že.g., a cytokine. then the Comet assay might be the

assay of choice to detect ROS related genotoxic effects. However, results obtained solely with this strand breakage assay appear of questionable relevance for risk assessment. So there is hardly any doubt that the statement of the ICH guideline ‘Preclinical Safety Evaluation of Biotechnology-Derived Pharmaceuticals’ correctly judges that ‘‘genotoxicity studies, routinely conducted for traditional pharmaceuticals are generally not applicable . . . ’’. In summary, during the toxicological assessment of biotechnology-derived products, genotoxicity testing according to the routine procedure employed for small molecular weight organic molecules appears generally as an inappropriate and ineffective expenditure of resources. However, adequate mutagenesis investigations need to be considered, on a case-by-

Scheme 1.

E. Gocke et al.r Mutation Research 436 (1999) 137–156

case basis, for compounds which for example interfere with DNA synthesis andror growth regulation or for compounds for which modifications could lead to suspect chemical reactivity. This case-by-case approach is illustrated in the following decision tree Scheme 1.

w9x

w10x

Acknowledgements Thanks are due to the many colleagues from the various companies who returned the questionnaires and helped in discussing the test data. Dr. J. Sims, Dr. N. Yajima, and Dr. K. Weyer are specifically acknowledged for information on the response of European authorities, on results and views on erythropoietin testing and on potential impurities in biotech preparations, respectively.

w11x

w12x

w13x

w14x

References w1x J.H. Levy, The human inflammatory response, J. Cardiovasc. Pharmacol. 27 Ž1996. S31–S37. w2x J.R. Lazutka, Genetic toxicity of cytokines, Mutat. Res. 361 Ž1996. 95–105. w3x S.M. Morris, A. Aidoo, O.E. Domon, L.J. McGarrity, R.L. Kodell, H.M. Schol, W.G. Hinson, J.L. Pipkin, D.A. Casciano, Effect of interleukin-2 on cell proliferation, sisterchromatid exchange induction, and nuclear stress protein phosphorylation in PHA-stimulated Fischer 344 rat spleen lymphocytes: modulation by 2-mercaptoethanol, Environ. Mol. Mutagen. 15 Ž1990. 10–18. w4x F. Hild, M. Freund, C. Fonatsch, Chromosomal aberrations during interferon therapy for chronic myelogenous leukaemia, New Engl. J. Med. 325 Ž1991. 132. w5x B.B. Aggarwal, P.K. Gadhia, V.L. Hopwood, B.J. Dave, S. Risin, L. Owen-Schaub, R. Pandita, S. Pathak, Induction of chromosomal rearrangements by tumour necrosis factors produced by primary B lymphoblastoid cells derived from cancer patients, Lymphokine Cytokine Res. 12 Ž1993. 9–19. w6x I.C. Green, J.M. Cunningham, C.A. Delaney, M.A. Titheradge, J.G. Mabley, M.H.L. Green, Effects of cytokines and nitric oxide donors on insulin secretion, cyclic nucleotides, glucose metabolism and DNA damage: relation to nitric oxide synthase activity and nitrite production in Insulin secretion and pancreatic B-cell research, in: P.R. Flatt, S. Lenzen ŽEds.., Smith-Gordon Publ., UK., 1994. w7x I.C. Green, C. A Delaney, J.M. Cunningham, J.G. Mabley, M. di Matteo, M.H.L. Green, Cytokines, nitric oxide DNA damage and insulin-secreting cells, Environ. Mut. Res. Commun. 16 Ž1995. 245–253. w8x C.A. Delaney, M.H.L. Green, J.E. Lowe, I.C. Green, Endogenous nitric oxide induced by interleukin-1 b in rats islets

w15x

w16x

w17x

w18x

w19x w20x

w21x

w22x

w23x

w24x

155

of Langerhans and HIT-T15 cells causes significant DNA damage as measured by the ‘comet’ assay, FEBS 33 Ž1993. 291–295. A. Dunger, J.M. Cunningham, C.A. Delaney, J.E. Lowe, M.H.L. Green, A.J. Bone, I.C. Green, Tumor necrosis factora and interferon-g inhibit insulin secretion and cause DNA damage in unweaned-rat islets: extent of nitric oxide involvement, Diabetes 45 Ž1996. 183–189. S. Moncada, The L-arginine:nitric oxide pathway, Acta Physiol. Scand. 145 Ž1992. 201–227. Y. Wang, T.A. Russo, O. Kwon, S. Chanock, S.C. Rumsey, M. Levine, Ascorbate recycling in human neutrophils: induction by bacteria, Proc. Natl. Acad. Sci. USA 94 Ž1997. 13816–13819. H. Ohshima, H. Bartsch, Chronic infections and inflammatory processes as cancer risk factors: possible role of nitric oxide in carcinogenesis, Mutat. Res. 305 Ž1994. 253–264. D.T. Efron, S.J. Kirk, M.C. Regan, H.L. Wasserkrug, A. Barbul, Nitric oxide generation from L-arginine is required for optimal human peripheral blood lymphocyte DNA synthesis, Surgery 110 Ž1991. 327–334. N. Yajima, Y. Kurata, T. Sawai, Y. Takeshita, Comparative studies in induction of micronuclei by three genetically recombinant and urinary human erythropoietins, Mutagenesis 8 Ž1993. 237–241. N. Yajima, Y. Kurata, T. Sawai, Y. Takeshita, Induction of micronucleated erythrocytes by recombinant human erythropoietin, Mutagenesis 8 Ž1993. 221–229. N. Yajima, Y. Kurata, T. Sawai, Y. Takeshita, Genotoxicity of genetic recombinant human erythropoietin in a novel test system, Mutagenesis 8 Ž3. Ž1993. 231–236. O. Hirai, Y. Miyamae, Y. Fujino, H. Izumi, A. Miyamoto, H. Noguchi, Prior bleeding enhances the sensitivity of the in vivo micronucleus test, Mutat. Res. 264 Ž1991. 109–114. G. Steinheider, R. Neth, H. Marquardt, Evaluation of nongenotoxic and genotoxic factors modulating the frequency of micronucleated erythrocytes in the peripheral blood of mice, Cell. Biol. Toxicol. 1 Ž1985. 197–211. H.E. Holden, Phenolphthalein: a sheep in wolf’s clothing?, Environ. Mol. Mutagen. 31 Ž1998. 103–104. P.P.W. van Buul, P. van Buul-Offers, Clastogenic effects of biosynthetic human growth hormone in Snell dwarf mice and CHO cells in vitro, Mutat. Res. 127 Ž1984. 61–64. B. Tedeschi, G.L. Spadoni, M.L. Sanna, P. Vernole, D. Caporossi, S. Cianfarani, B. Nicoletti, B. Boscherini, Increased chromosome fragility in lymphocytes of short normal children treated with recombinant human growth hormone, Hum. Genet. 91 Ž1993. 459–463. D.B. Allen, A.C. Rundle, D.A. Graves, S.L. Blethen, Risk of leukemia in children treated with human growth hormone: review and reanalysis, J. Pediatr. 131 Ž1997. S32–S36. S.E. Hankinson, W.C. Willett, G.A. Colditz, D.J. Hunter, D.S. Michaud, B. Deroo, B. Rosner, F.E. Speizer, M. Pollak, Circulating concentrations of insulin-like growth factor-I and risk of breast cancer, Lancet 351 Ž1998. 1393–1396. H.E. Jones, C.L. Eaton, D. Barrow, C.M. Dutkowski, J.M. Gee, K. Griffiths, Comparative studies of the mitogenic

156

w25x w26x

w27x

w28x w29x

w30x

w31x

w32x

w33x

w34x w35x w36x

w37x

E. Gocke et al.r Mutation Research 436 (1999) 137–156 effects of epidermal growth factor and transforming growth factor-alpha and the expression of various growth factors in neoplastic and non-neoplastic prostatic cell lines, Prostate 30 Ž1997. 219–231. B.N. Ames, L.S. Gold, Too many rodent carcinogens: mitogenesis increases mutagenesis, Science 249 Ž1990. 970–971. C. Johannes, W.D. Heller, P. Eke, G. Obe, Frequencies of polycentric chromosomes following combinational treatments of Chinese hamster ovary cells with restriction endonucleases and X-rays, bleomycin or DNase I, Mutat. Res. 373 Ž1997. 237–244. G.A. Folle, E. Boccardo, G. Obe, Localisation of chromosome breakpoints induced by DNAse I in Chinese hamster ovary ŽCHO. cells, Chromosoma 106 Ž1997. 391–399. J.D. Green, Pharmaco-toxicological report: pulmozyme rhDNAse Genentech Hum, Exp. Toxicol. 13 Ž1994. 1–42. O.N. Ilinskaya, O.B. Ivanchenko, N.S. Karamova, Bacterial ribonuclease: mutagenic effect in bacterial test systems, Mutagenesis 10 Ž1995. 165–170. O.N. Ilinskaya, N.S. Karamova, O.B. Ivanchenko, L.V. Kipenskaya, SOS-inducing ability of native and mutant microbial ribonucleases, Mutat. Res. 354 Ž1996. 203–209. M. Kondo, T. Ogawa, Y. Matsubara, A. Mizutani, S. Murata, M. Kitagawa, Safety evaluation of lipase G from Penicillium camembertii, Food Chem. Toxicol. 32 Ž1994. 685–696. A. Broadmeadow, C. Clare, A.S. De Boer, An overview of the safety evaluation of the Rhizomucor miehei lipase enzyme, Food Addit. Contam. 11 Ž1994. 105–119. M. Stavnsbjerg, A.S. DeBoer, B.F. Jensen, Safety evaluation of lipase from Mucor–Mihei for inter-esterification of edible oils and fats, Food Addit. Contam. 5 Ž1988. 251–265. M. Ishidate, Jr. Data Book of Chromosomal Aberration Test In Vitro, Elsevier, Amsterdam, 1988. M. Cerna, ´ Mutagenicity of b-glucuronidase in the Ames test, Mutat. Res. 143 Ž1985. 167–172. W.W. Nichols, B.J. Ledwith, S.V. Manam, P.J. Troilo, Potential DNA vaccine integration into host cell genome, Ann. NY Acad. Sci. 772 Ž1995. 30–39. P.D. Minor, Regulatory issues in the use of DNA vaccines, Ann. NY Acad. Sci. 772 Ž1995. 170–177.

w38x S.M. Galloway, Genotoxicity tesing of pharmaceuticals: the need for a rational test design, Presentation at the 25th Annual Meeting EEMS, Nordwijkerhout, 1995. w39x G. Calarota, S. Bratt, A.-C. Nordlund, V. Leandersson, B. Sandstrom, ¨ B. Wahren, Cellular cytotoxic response induced by DNA vaccination in HIV-1-infected patients, Lancet 351 Ž1998. 1320–1325. w40x G.M. Arndt, GH rank colocalization of antisense RNAs and ribozymes with their target mRNAs, Genome 40 Ž1997. 785–797. w41x G. Wang, M.M. Seidman, P.M. Glazer, Mutagenesis in mammalian cells induced by triple helix formation and transcription-coupled repair, Science 271 Ž1996. 802–805. w42x P.C. Hanawalt, Transcription-coupled repair and human disease, Science 266 Ž1994. 1957–1958. w43x B.A. Bunnell, R.A. Morgan, Gene therapy for infectious diseases, Clin. Microbiol. Rev. 11 Ž1998. 42–56. w44x L.A. Couture, D.T. Stinchcomb, Anti-gene therapy: the use of ribozymes to inhibit gene function, Trends Genet. 12 Ž1996. 510–515. w45x E.A. Griffin Jr., Z. Qin, W.J. Michels Jr., A.M. Pyle, Group II intron ribozymes that cleave DNA and RNA linkages with X similar efficiency, and lack contacts with substrate 2 -hydroxyl groups, Chem. Biol. 2 Ž1995. 761–770. w46x D.J. Kirkland, N.N. Kim, Special considerations for conducting genotoxicity tests with protein materials, Mutagenesis 10 Ž1995. 393–398. w47x S. Dean, Meeting report of the Industrial Genotoxicology Group. London, December 1995, Mutagenesis 12 Ž1997. 49–50. w48x G. Sofer, Biotechnology product validation: Part 4. Clearance of impurities from protein and peptide biotherapeutics, Pharmaceutical Technology Europe ŽMay 1994. 29–32. w49x G. Sofer, L. Hagel, in: Handbook of Process Chromatography: A Guide to Optimization, Scale-up and Validation, Academic Press, 1997, 149 pp. w50x L. Muller, P. Kasper, Human biological relevance and the ¨ use of threshold arguments in regulatory genotoxicology assessment: experience with pharmaceuticals, Mutat. Res. Žsubmitted..