Application of in Vitro T Cell Assay Using Human Leukocyte Antigen

Feb 13, 2018 - Application of in Vitro T Cell Assay Using Human Leukocyte Antigen-Typed Healthy Donors for the Assessment of Drug Immunogenicity ... P...
1 downloads 6 Views 1MB Size
Download PDF
Rapid Report Cite This: Chem. Res. Toxicol. 2018, 31, 165−167

pubs.acs.org/crt

Application of in Vitro T Cell Assay Using Human Leukocyte AntigenTyped Healthy Donors for the Assessment of Drug Immunogenicity Toru Usui,† Lee Faulkner,† John Farrell,† Neil S. French,† Ana Alfirevic,† Munir Pirmohamed,† B. Kevin Park,† and Dean J. Naisbitt*,† †

MRC Centre for Drug Safety Science, Department of Molecular and Clinical Pharmacology, University of Liverpool, Sherrington Building, Ashton Street, Liverpool L69 3GE, England

̈ T cells to drugs is detectable in healthy human donors expressing different ABSTRACT: It is unclear whether priming of naive human leukocyte antigen (HLA) alleles. Thus, we examined T cell priming with drugs associated with HLA risk alleles and control compounds in 14 HLA-typed donors. Nitroso sulfamethoxazole and piperacillin activated T cells from all donors, whereas responses to carbamazepine and oxypurinol were only seen in donors expressing HLA-B*15:02 and HLA-B*58:01, respectively. Weak flucloxacillin-specific T cell responses were detected in donors expressing HLA-B*57:01 and HLA-B*58:01. These data show that the priming of T cells with certain drugs is skewed toward donors expressing specific HLA alleles.

A

the intrinsic immunogenicity of candidate drugs at the preclinical stage. However, it is still unclear whether T cell responses are only detected in susceptible donors expressing specific HLA risk alleles. Thus, we recruited healthy donors expressing HLA-B*15:02, HLA-B*57:01, and HLA-B*58:01 (n = 3 per allele) and five healthy donors without these alleles and ̈ T cells from each donor to the attempted to prime naive compounds piperacillin, nitroso sulfamethoxazole, flucloxacillin, oxypurinol, and carbamazepine, concurrently. This approach allowed us to assess multiple drugs in a single experiment and find whether responses are only detectable in donors expressing HLA risk alleles. Blood donors (aged 18−60, with no medical complications, no history of allergy, and no current medications) were identified from our HLA-typed DNA bank, and sampling was conducted at the Royal Liverpool University Hospital with approval from the Liverpool Local Research Ethics Committee. Informed written consent was obtained from each donor. ̈ T cells were CD14+ monocytes and CD45RO negative naive separated from PBMC using magnetic beads. CD25+ regulatory T cells were removed and discarded. Monocytes were cultured in medium supplemented with GM-CSF and IL4 (800 U/ml; 37 °C/5% CO2) for 7 days to generate to generate dendritic cells. LPS/TNFα matured dendritic cells (0.8 × 105 per well) were then cocultured with autologous ̈ T cells (2−2.5 × 106 per well; 24-well plated total volume naive 1.5 mL) and 0.5 mL of the drugs (nitroso sulfamethoxazole [25 μM], carbamazepine [12.5 μM], flucloxacillin [1 mM], oxypurinol [50 μM], and piperacillin [1 mM]) in medium,

lthough toxic compounds are typically screened out during preclinical safety studies, in some cases, adverse reactions are detected in late clinical phases or during postmarketing evaluation. Immune-mediated reactions that target skin or liver are some of the most feared forms of adverse reaction because of their severity and difficulty in predicting patient susceptibility. Human leukocyte antigen (HLA) is the notable genetic biomarker of skin and liver reactions. Several associations have been reported,1 and drug-specific T cells have been detected in patients expressing the HLA alleles. This strongly suggests that the adverse events are immune-mediated and that the activation of T cells is a key initiating event. Drugs activate T cells by (1) the hapten-concept, which states that covalently modified drug−peptide adducts interact with MHC molecules to activate T cells (e.g., piperacillin,2 flucloxacillin,3 nitroso sulfamethoxazole4), and (2) the pharmacological interactions concept, which states that drugs interact directly with MHC molecules or the peptide in the MHC binding groove (e.g., carbamazepine,5 oxypurinol6). ̈ T cells from healthy donors are We have reported that naive activated with the drug metabolite nitroso sulfamethoxazole using a priming assay that requires the coculture of T cells with autologous monocyte-derived dendritic cells and drug for 8 days. Drug-specific T cell responses were measurable after the primed cells were restimulated with nitroso sulfamethoxazole and a second batch of dendritic cells. T cells responses to the drugs carbamazepine, flucloxacillin, and allopurinol were also detected separately with lymphocytes from healthy donors expressing either HLA-B*15:02, HLA-B*57:01, or HLAB*58:01, respectively.7 These findings imply that in vitro T cell systems using healthy volunteer peripheral blood mononuclear cells (PBMC) might be useful for prediction of © 2018 American Chemical Society

Received: February 6, 2018 Published: February 13, 2018 165

DOI: 10.1021/acs.chemrestox.8b00030 Chem. Res. Toxicol. 2018, 31, 165−167

Chemical Research in Toxicology

Rapid Report

13 donors in the proliferation assay using concentrations of 25−100 μM (Figure 1). Good to strong nitroso sulfamethoxazole-specific responses were detected with primed T cells from

̈ T cells from genotyped donors with nitroso Figure 1. Priming of naive sulfamethoxazole and piperacillin. Primed T cells were restimulated with fresh dendritic cells and drugs and (A) proliferation and (B) IFNγ release were measured. (A) Dose-dependent proliferative response of primed T cells from all donors. (B) Representative ELIspot data from four donors (1 HLA-B*57:01+, 1 HLA-B*58:01+, 1 HLA-B*15:02+, 1 not expressing these alleles). NC, noncarrier donors not expressing three HLA risk alleles.

which did not contain additional antibiotics for 8 days. Primed T cells (5 × 104; 200 μL in 96 well plate) were restimulated with autologous dendritic cells (4 × 103) and drug (nitroso sulfamethoxaozle [25, 50, and 100 μM], carbamazepine (6.25, 12.5, and 25 μM), flucloxacillin (0.5, 1, and 2 mM), oxypurinol (25, 50, and 100 μM), and piperacillin (0.5, 1, and 2 mM)) and assessed for IFN-γ secretion (enzyme-linked immunospot [ELIspot]) as well as proliferation ([3H] thymidine). The ambitious plan to study multiple concentrations of five drugs with two readouts resulted in a trade-off that experiments were conducted in duplicate cultures per condition. Data are expressed as stimulation index (SI; mean cpm in test incubations with drug/mean cpm in control incubations with vehicle alone), and results are discussed according to the classification of Faulkner et al. (SI < 1.49 negative; SI 1.5−1.99 weak response; SI 2.0−3.99 good response; SI > 4.0 strong response).7 This classification describes experimental outcomes more accurately than statistical analysis of multiple replicates where false positives have been observed. Fourteen priming experiments were performed with nitroso sulfamethoxazole and good to strong responses were seen with

̈ T cells from genotyped donors with Figure 2. Priming of naive carbamazepine, flucloxacillin, and oxypurinol. Primed T cells were restimulated with dendritic cells and drugs, and (A) proliferation and (B) IFNγ release were measured. (A) Dose-dependent proliferative response of primed T cells from all donors. (B) Representative ELIspot data from four donors (1 HLA-B*57:01+, 1 HLA-B*58:01+, 1 HLA-B*15:02+, 1 not expressing these alleles). NC, noncarrier donors not expressing three HLA risk alleles. Oxypurinol experiments ̈ T cells. were conducted on 11 donors due to a limited number of naive 166

DOI: 10.1021/acs.chemrestox.8b00030 Chem. Res. Toxicol. 2018, 31, 165−167

Chemical Research in Toxicology

Rapid Report

Funding

all 14 donors using the IFN-γ ELIspot to visualize cytokine secretion (mean SI 4.7 ± 5.4; range 2.1−19.3). With piperacillin, weak or good responses were observed in the proliferation assay with 5/5 donors not expressing risk HLA alleles, 2/3 HLA-B*15:02+ donors, 2/3 HLA-B*57:01+ donors, and 1/3 HLA-B*58:01+ donor (Figure 1). IFN-γ release was observed with primed T cells from 11/14 donors (mean SI 13.0 ± 12.4; range 2.5−29.0). One HLA-B*57:01+, one HLA-B*58:01, and one donor not expressing HLA risk alleles yielded negative results in both assays. Nitroso sulfamethoxazole and piperacillin are directly reactive and bind covalently to cysteine and lysine residues on protein, respectively. Our data clearly show that both compounds are intrinsically immunogenic and activate T cells in donors expressing a range of HLA alleles after immune regulation has been reduced through the removal of regulatory T cells. The next component of the project was to study: (1) the haptenic drug flucloxacillin, which causes liver injury in donors expressing HLA-B*57:01, (2) the direct HLA binding drug carbamazepine, which causes severe skin reactions in donors expressing HLA-B*15:02, and (3) the metabolite of allopurinol, oxypurinol, which activates T cells from donors with skin reactions expressing HLA-B*58:01 again via a direct binding interaction with MHC molecules, to ascertain whether priming ̈ T cells is only detected in healthy donors expressing of naive the relevant HLA risk allele. With carbamazepine, good responses were detected with primed T cells from 2/3 donors expressing HLA-B*15:02 in both the proliferation assay and IFN-γ ELIspot (SI 2.5 and 33.8) (Figure 2). With oxypurinol, a weak proliferative response was detected with primed T cells from one donor expressing HLA-B*58:01 and IFN-γ secretion was observed with two HLA-B*58:01+ donors (SI 3.3 and 8.7). Priming of naiv̈ e T cells with both carbamazepine and oxypurinol was not observed in donors without the relevant HLA risk alleles. Weak proliferative responses to flucloxacillin were observed with primed T cells from three donors (57:01− 3, 58:01−1, and 58:01−3), while a good response was observed with donor 57:01−2. These data suggest that flucloxacillin ̈ T cells from donors expressing preferentially activates naive HLA-B*57:01 and HLA-B*58:01. Interestingly, the two alleles are structurally similar, differing in only five amino acids. Furthermore, flucloxacillin binding to HLA-B*57:01 and HLAB*58:01 has been shown to activate CD8+ T cells from patients with DILI.3 To conclude, we have shown that it is useful to assess the immunogenicity of drugs with a panel of donors expressing HLA alleles associated with a range adverse events. Responses were detected in rare cases in donors not expressing HLA risk alleles, suggesting that drugs bind to multiple MHC molecules to activate T-cells. In the future, it is imperative that simplified test systems are established to permit the screening of a large panel of drugs associated with different frequencies of ̈ T cells from an increased number hypersensitivity against naive of donors.

■

The work was part-funded by the MIP-DILI project supported by an IMI grant (115336) and the MRC CDSS grant (G0700654). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS T.U. is a visiting scientist from Sumitomo Dainippon Pharma Co., Ltd. We would like to thank the blood donors for participating in this project and the nurse who collect the samples.

■ ■

ABBREVIATIONS ELIspot, enzyme-linked immunospot; PBMC, peripheral blood mononuclear cells; SI, stimulation index REFERENCES

(1) Pirmohamed, M., Ostrov, D. A., and Park, B. K. (2015) New genetic findings lead the way to a better understanding of fundamental mechanisms of drug hypersensitivity. J. Allergy Clin. Immunol. 136, 236−244. (2) Meng, X., Al-Attar, Z., Yaseen, F. S., Jenkins, R., Earnshaw, C., Whitaker, P., Peckham, D., French, N. S., Naisbitt, D. J., and Park, B. K. (2017) Definition of the Nature and Hapten Threshold of the betaLactam Antigen Required for T Cell Activation In Vitro and in Patients. J. Immunol. 198, 4217−4227. (3) Monshi, M. M., Faulkner, L., Gibson, A., Jenkins, R. E., Farrell, J., Earnshaw, C. J., Alfirevic, A., Cederbrant, K., Daly, A. K., French, N., Pirmohamed, M., Park, B. K., and Naisbitt, D. J. (2013) Human leukocyte antigen (HLA)-B*57:01-restricted activation of drug-specific T cells provides the immunological basis for flucloxacillin-induced liver injury. Hepatology 57, 727−739. (4) Castrejon, J. L., Berry, N., El-Ghaiesh, S., Gerber, B., Pichler, W., Park, B., and Naisbitt, D. J. (2010) Stimulation of human T cells with sulfonamides and sulfonamide metabolites. J. Allergy Clin. Immunol. 125, 411−418. (5) Ko, T. M., Chung, W. H., Wei, C. Y., Shih, H. Y., Chen, J. K., Lin, C. H., Chen, Y. T., and Hung, S. I. (2011) Shared and restricted T-cell receptor use is crucial for carbamazepine-induced Stevens-Johnson syndrome. J. Allergy Clin. Immunol. 128, 1266−1276. (6) Yun, J., Marcaida, M. J., Eriksson, K. K., Jamin, H., Fontana, S., Pichler, W. J., and Yerly, D. (2014) Oxypurinol directly and immediately activates the drug-specific T cells via the preferential use of HLA-B*58:01. J. Immunol. 192, 2984−2993. (7) Faulkner, L., Gibson, A., Sullivan, A., Tailor, A., Usui, T., Alfirevic, A., Pirmohamed, M., Naisbitt, D. J., and Kevin Park, B. (2016) Detection of Primary T Cell Responses to Drugs and Chemicals in HLA-Typed Volunteers: Implications for the Prediction of Drug Immunogenicity. Toxicol. Sci. 154, 416−429.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +44 151 7945346. Fax: +44 151 7945540. ORCID

Dean J. Naisbitt: 0000-0003-4107-7832 167

DOI: 10.1021/acs.chemrestox.8b00030 Chem. Res. Toxicol. 2018, 31, 165−167