Differential hepatocyte toxicity of recombinant Apo2L/TRAIL versions

To the editor

Apo2L/TRAIL (Apo2 ligand or tumor necrosis factor (TNF)-related apoptosis-inducing ligand) was discovered by its sequence homology to TNF and CD95L (refs. 1,2). Recombinant soluble human Apo2L/TRAIL is a candidate for clinical investigation in cancer therapy because it induces apoptosis in a broad spectrum of human cancer cell lines but not in many normal cells, and exhibits potent anti-tumor activity without normal tissue toxicity in various cancer xenograft models3,4,5,6,7. Recently, however, a polyhistidine-tagged recombinant version of human Apo2L/TRAIL (residues 114–281, hereafter termed Apo2L/TRAIL.His) was found to induce apoptosis in vitro in isolated human, but not non-human hepatocytes8, raising the concern that Apo2L/TRAIL therapy might cause hepatotoxicity8,9. Here we show that different recombinant versions of Apo2L/TRAIL vary widely in biochemical properties and potential for cellular and whole animal toxicity.

Crystallographic studies show that, like TNF, soluble Apo2L/TRAIL has a homotrimeric structure10,11. Uniquely, Apo2L/TRAIL contains an internal zinc atom bound by the cysteine residues at position 230 of each subunit10,11, which is crucial for trimer stability and biologic activity11,12. As a clinical candidate, we developed a version of human Apo2L/TRAIL that lacks exogenous sequence tags (residues 114–281, termed Apo2L/TRAIL.0)3. Moreover, by adding zinc and dithiothreitol and maintaining neutral pH in buffers, we optimized large-scale bacterial expression and purification of the protein to maximize zinc binding and homotrimer stability. We prepared Apo2L/TRAIL.His using methods similar to those used by Jo et al.8 and compared it with Apo2L/TRAIL.0. Metal analysis (Fig. 1a) showed 0.46 mol zinc/mol trimer for Apo2L/TRAIL.His compared to 0.92 for Apo2L/TRAIL.0. Gel electrophoresis (Fig. 1b) and size exclusion chromatographic studies (not shown) together indicated that Apo2L/TRAIL.His was heterogeneous, with 79% trimer (56% of which contained intersubunit disulfide), and 21% free dimer (all containing disulfide). In contrast, Apo2L/TRAIL.0 was homogeneous, with 99% trimer (containing < 1% disulfide). Circular dichroism (Fig. 1c) demonstrated a less ordered conformation for Apo2L/TRAIL.His compared to Apo2L/TRAIL.0. Apo2L/TRAIL.His was relatively unstable in solution and formed insoluble aggregates upon incubation at 37 °C (data not shown). Moreover, Apo2L/TRAIL.His was less potent than Apo2L/TRAIL.0 at inducing death in three cancer cell lines (Fig. 1d), suggesting the importance of zinc and trimeric structure for this activity.

Figure 1: Characterization of recombinant soluble human Apo2L/TRAIL versions.

a, Zinc content as mole ratio to trimer. b, SDS–PAGE without (−) or with (+) reduction with 100 mM dithiothreitol (DTT), visualized by silver staining. c, Circular dichroic spectra. Thick line, Apo2L/TRAIL.0; thin line, Apo2L/TRAIL.His. d, Activity of Apo2L/TRAIL.0 (closed symbols) and Apo2L/TRAIL.His (open symbols) against Colo205 colon carcinoma (circles), SK-MES-1 lung carcinoma (squares), or BJAB B lymphoma cells (diamonds), as determined at 24 h by AlamarBlue assay.

Apo2L/TRAIL.His (ref. 8) or antibody-crosslinked CD95L (ref. 13) induced marked apoptotic morphology in isolated human hepatocytes within 24 h, while Apo2L/TRAIL.0 induced little morphologic change compared to vehicle (Fig. 2a, top). Flow cytometric analysis of DNA fragmentation14, a hallmark of apoptosis, showed that Apo2L/TRAIL.His induced substantial cell death at 0.1 μg/ml, while Apo2L/TRAIL.0 had a marginal effect even at 100 μg/ml (Fig. 2b, top). Similar results were obtained at 4 h (data not shown). Apo2L/TRAIL.His markedly increased cleavage of poly-ADP ribose polymerase (PARP), a substrate of caspase-3 that is activated downstream of caspase-8 in Apo2L/TRAIL-induced apoptosis3,15, while Apo2L/TRAIL.0 did not (Fig. 2c, top). Viability assays using Alamar Blue, MTT, Crystal Violet and Toludine Blue (data not shown) confirmed minimal induction of hepatocyte apoptosis or necrosis by Apo2L/TRAIL.0. Consistent data were obtained at Genentech, Immunex and University of Pittsburgh, with hepatocytes from eight donors, isolated either by two commercial sources or by S.S.S. Thus, despite isolation and culture in low serum, hepatocytes show little sensitivity to Apo2L/TRAIL.0.

Figure 2: Differential interaction of Apo2L trail versions with hepatocytes.

a–c, Effect of recombinant human Apo2L/TRAIL versions on isolated human (top panels) or cynomolgus (bottom panels) hepatocytes. a, Phase contrast images (×10) taken after 24 h treatment. b, DNA fragmentation (means ± s.e.m. of triplicates) in hepatocytes treated for 24 h with Apo2L/TRAIL.0 (), Apo2L/TRAIL.His (□) or Flag-tagged CD95L crosslinked with anti-Flag antibody (♦). c, Western blot analysis of PARP cleavage 24 h after treatment with vehicle (lane 1), Apo2L/TRAIL.0 at 10 (lane 2) or 100 (lane 3) μg/ml, Apo2L/TRAIL.His at 10 μg/ml (lane 4), or crosslinked CD95L at 1 μg/ml (lane 5). The top and bottom protein bands represent uncleaved and cleaved PARP, respectively. d, Binding of radio-iodinated Apo2L/TRAIL.0 (top) or Apo2L/TRAIL.His (bottom) to cynomolgus hepatocytes, each competed with its corresponding unlabeled version.

Previously, we have used the cynomolgus monkey as a preclinical safety model for Apo2L/TRAIL (ref. 3). Apo2L/TRAIL and its receptors DR4, DR5, DcR2 and OPG (ref. 16) showed 84–99 % extracellular protein sequence identity between cynomolgus and human counterparts (not shown). Cynomolgus DcR1 appeared to be a pseudo-gene; however, absence of this antagonistic receptor, if relevant, might be expected to increase sensitivity to Apo2L/TRAIL. Human Apo2L/TRAIL.0 bound with comparable affinity to Fc-fusion proteins of each human or cynomolgus receptor (not shown), demonstrating good cross-reactivity. As evidenced by morphology, DNA fragmentation, and PARP cleavage (Fig. 2ac, bottom), Apo2L/TRAIL.His induced substantial apoptosis in isolated cynomolgus hepatocytes, whereas Apo2L/TRAIL.0 had marginal effects, similar to results with human hepatocytes. Notably, Alamar Blue or MTT assay was not adequate for detecting death in cynomolgus hepatocytes despite clear evidence of apoptosis, perhaps explaining why rhesus hepatocytes appeared previously8 to be resistant to Apo2L/TRAIL.His by MTT assay. Neither TNF (0.1 μg/ml), nor interleukin-1β (0.1 μg/ml), nor cycloheximide (0.5 μg/ml) sensitized cynomolgus hepatocytes to Apo2L/TRAIL.0-induced DNA fragmentation (data not shown), further demonstrating minimal toxicity of Apo2L/TRAIL.0 toward these cells. Binding of radio-iodinated Apo2L/TRAIL.His to cynomolgus hepatocytes was not displaced by unlabeled ligand (Fig. 2d), suggesting irreversible interaction; binding of Apo2L/TRAIL.0 was displaced in a dose-dependent fashion, indicating reversible association. Together, these data demonstrate that the cynomolgus monkey is a valid species for preclinical safety investigation of clinical-grade recombinant human Apo2L/TRAIL.

Intravenous (i.v.) administration of Apo2L/TRAIL.0 in cynomolgus monkeys (0, 0.1, 1, or 10 mg/kg/d × 7 d, n = 3/dose) was well tolerated3. We assessed potential hepatotoxicity in a further study, by dosing vehicle or Apo2L/TRAIL.0 at 100 mg/kg/d over 5 d (n = 12/group). There were no notable differences between the groups in clinical signs, food consumption, body weight, serum chemistry, coagulation parameters, hematology, and urinalysis; importantly, there were no significant changes in liver enzyme activities, bilirubin or albumin in serum. Histology of all major organs including liver was normal. Serum Apo2L/TRAIL.0 peaked at –1000 μg/ml and remained for over 7 h above 0.1μg/ml, a dose of Apo2L/TRAIL. His that induced marked hepatocyte apoptosis within 4 h in vitro. We also studied Apo2L/TRAIL.0 in chimpanzees. Chimpanzee and human Apo2L/TRAIL, DR4, DR5, DcR1, DcR2 and OPG showed 97–99% extracellular sequence identity (data not shown). After i.v. injection at 5 mg/kg (n = 3), Apo2L/TRAIL.0 peaked in chimpanzee serum at 130 μg/ml and remained above 0.1 μg/ml for > 6 h; the animals showed no significant changes in clinical pathology, including liver enzyme activities. Thus, two relevant non-human primate models indicate that systemic administration of Apo2L/TRAIL.0 to cancer patients is unlikely to cause major toxicity to the liver or other organs. Further preclinical safety studies of clinical-grade Apo2L/TRAIL.0 are in progress to examine administration over longer time periods and to define doses suitable for investigation of a safe therapeutic window in cancer patients.


  1. 1

    Wiley, S.R. et al. Identification and characterization of a new member of the TNF family that induces apoptosis. Immunity 3, 673–682 (1995).

    CAS  Article  Google Scholar 

  2. 2

    Pitti, R.M. et al. Induction of apoptosis by Apo-2 Ligand, a new member of the tumor necrosis factor receptor family. J. Biol. Chem. 271, 12697–12690 (1996).

    Article  Google Scholar 

  3. 3

    Ashkenazi, A. et al. Safety and anti-tumor activity of recombinant soluble Apo2 ligand. J. Clin. Invest. 104, 155–162 (1999).

    CAS  Article  Google Scholar 

  4. 4

    Walczak, H. et al. Tumoricidal activity of tumor necrosis factor-related apoptosis-inducing ligand in vivo. Nature Med. 5, 157–163 (1999).

    CAS  Article  Google Scholar 

  5. 5

    Gliniak, B. & Le, T. Tumor necrosis factor-related apoptosis-inducing ligand's antitmor activity in vivo is enhanced by the chemotherapeutic agent CPT-11. Cancer Res. 59, 6153–6158 (1999).

    CAS  PubMed  Google Scholar 

  6. 6

    Roth, W. et al. Locoregional Apo2L/TRAIL eradicates intracranial human malignant glioma xenografts in athymic mice in the absence of neurotoxicity. Biochem. Biophys. Res. Commun. 265, 1999 (1999).

    Article  Google Scholar 

  7. 7

    Chinnaiyan, A.M. et al. Combined effect of tumor necrosis factor-related apoptosis-inducing ligand and ionizing radiation in breast cancer therapy. Proc. Natl. Acad. Sci. USA 97, 1754–1759 (2000).

    CAS  Article  Google Scholar 

  8. 8

    Jo, M. et al. Apoptosis induced in normal human hepatocytes by tumor necrosis factor-related apoptosis-inducing ligand. Nature Med. 6, 564–567 (2000).

    CAS  Article  Google Scholar 

  9. 9

    Nagata, S. Steering anti-cancer drugs away from the TRAIL. Nature Med. 6, 502–503 (2000).

    CAS  Article  Google Scholar 

  10. 10

    Hymowitz, S.G. et al. Triggering cell death: the crystal structure of Apo2L/TRAIL in a complex with death receptor 5. Molec. Cell 4, 563–571 (1999).

    CAS  Article  Google Scholar 

  11. 11

    Hymowitz, S.G. et al. A unique zinc-binding site revealed by a high-resolution X-ray structure of homotrimeric Apo2L/TRAIL. Biochemistry 39, 633–640 (2000).

    CAS  Article  Google Scholar 

  12. 12

    Bodmer, J.-L., Meier, P., Tschopp, J. & Schneider, P. Cysteine 230 is essential for the structure and activity of the cytotoxic ligand TRAIL. J. Biol. Chem . 275, 20632–20637 (2000).

    CAS  Article  Google Scholar 

  13. 13

    Schneider, P. et al. Conversion of membrane-bound Fas(CD95) ligand to its soluble form is associated with downregulation of its proapoptotic activity and loss of liver toxicity. J. Exp. Med. 187, 1205–1213 (1998).

    CAS  Article  Google Scholar 

  14. 14

    Nicoletti, I., Migliorati, G., Pagliacci, M.C., Grignani, F. & Riccardi, C. A rapid and simple method for measuring thymocyte apoptosis by propidium iodide staining and flow cytometry. J. Immunol. Methods 139, 271–279 (1991).

    CAS  Article  Google Scholar 

  15. 15

    Kischkel, F.C. et al. Apo2L/TRAIL-dependent recruitment of endogenous FADD and caspase-8 to death receptors 4 and 5. Immunity 12, 611–620 (2000).

    CAS  Article  Google Scholar 

  16. 16

    Ashkenazi, A. & Dixit, V.M. Apoptosis control by death and decoy receptors. Curr. Opin. Cell Biol. 11, 255–260 (1999).

    CAS  Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Lawrence, D., Shahrokh, Z., Marsters, S. et al. Differential hepatocyte toxicity of recombinant Apo2L/TRAIL versions. Nat Med 7, 383–385 (2001). https://doi.org/10.1038/86397

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing