Storm forecasting: additional lessons from the CD28 superagonist TGN1412 trial

Abstract

This letter is in response to the Comment article published in Nature Reviews Immunology by Thomas Hünig (The storm has cleared: lessons from the CD28 super agonist TGN1412 trial. Nature Rev. Immunol. 12, 317-318 (2012)) 1. In his commentary , Thomas Hünig attempts to explain "why the three sets of preclinical data … used to support this Phase I clinical trial had failed to predict the cytokine storm" in healthy volunteers. He describes some of the findings in these test systems and concludes that they "all failed to provide evidence for the toxic potential of the antibody for distinct and unrelated reasons". He further claims that "the disastrous outcome … may have made a positive contribution to the future development of immunomodulatory drugs", primarily by "teaching us the superiority of the MABEL [minimum anticipated biological effect level]-based approach over the NOAEL [no observed adverse effect level]-based approach". In sharp contrast to Thomas Hünig's opinions, we believe that the serious adverse events were predictable based on the known CD28 receptor biology, the intended pharmacological activity of TGN1412 and the preclinical safety assessment programme. The purpose of a preclinical safety assessment programme is to identify potential safety risks and to qualify and quantify these risks by using in vitro human data, in vitro and in vivo animal data, and translational and mechanism-based knowledge of the intended pathway modulation to 'predict' in vivo human responses. We feel that these criteria were not met for TGN1412. Specifically, the publicly available information in the investigational medicinal product dossier (IMPD) and/or the investigator's brochure demonstrates deficiencies in four major areas of the preclinical safety assessment. One, there was a failure to identify risk based on the intended target, mechanism of action and previous experiences with comparator products. Two, there was a failure to reconcile discordant preclini-cal data sets. Three, there was a failure to identify a pharmacologically relevant species and/or data set for the determination of the first-inhuman dose. And, four, there was a failure to translate the preclinical data to estimate the anticipated pharmacological activity of the first-inhuman dose. We shall discuss these topics in order. Failure to identify risk based on the intended target and mechanism of action. With TGN1412, there were strong reasons to proceed with caution based on the nature of the target and the intended and/or desired pharmacological activity: superactivation of T cells. Comparable superactivation of T cells occurs naturally when bacterial superantigens bypass normal antigen presentation processes and directly co-activate MHC class II molecules and T cell receptors (TCRs), resulting in fever, extensive cytokine release, toxic shock and death at plasma concentrations as low as 0.1 pg per ml (REF. 2). Historically, iatrogenic T cell mitogenesis, T cell superactivation and sudden , but prolonged, depletion of circulating T cells have been the hallmarks of immuno-modulatory CD3-specific monoclonal antibodies, including OKT3 (Janssen-Cilag), HuM291 (visilizumab; PDL BioPharma) and others 3. These reactions depend on the extent of receptor occupancy and can occur at remarkably low doses. Although TGN1412 was a first-in-class CD28-specific superagonist, it was intended to induce the superactivation of T cells, much like the CD3-specific antibodies that induce cytokine-release syndrome (CRS). While cursory mention of the possibility of CRS was made in the IMPD and investigator's brochure, insufficient consideration was given to the lessons already learnt from natural superagonists and comparator products. In fact, it was stated in the investigator's brochure that "TGN1412 is expected to be well tolerated in humans and not to elicit any adverse effects". Indeed, the design of the Phase I study, in which all the volunteers were dosed at 10-minute intervals 4 , suggests that no practical consideration was given to the possibility of CRS occurring. Failure to reconcile discordant preclinical data. The early decision by the Medicines and Health Regulatory Agency 5 to publicly release the IMPD, investigator's brochure, clinical trial protocol, assessment report and informed consent for TGN1412 (avail-able on the Citizens for Responsible Care and Research website) should be applauded because it facilitated independent review. Such a review shows that disparate results were obtained from in vitro and in vivo assessments of the superagonist activity of CD28-specific antibodies in humans, rats and monkeys. In vitro, T cell activation and cytokine release have been demonstrated for the monoclonal antibody 5.11A1 (a precursor to TGN1412) in human systems and for JJ316 (a rat CD28-specific monoclonal antibody homologous to TGN1412) in rat systems 6,7. Based on the publicly released information, it appears that no such demonstration of in vitro pharmacological activity of TGN1412 was presented for human or monkey T cells in the IMPD or investigator's brochure. In these documents, it appears that the only data describing the activity of TGN1412 in human systems were the results for in vitro T cell mitogenesis assays that were conducted with the precursor mono-clonal antibody 5.11A1 or with TGN1412 in the presence or absence of methotrexate. The concentration-response relationships for TGN1412 in human cell systems and their relationship (or lack thereof) to those for TGN1412 in monkey systems or JJ316 in rat systems do not appear to have been discussed , limiting the ability to model potential effects of the administration of TGN1412 to humans from the in vitro data. In rats, a single dose of 5 mg per kg of JJ316 resulted in a 3-to 6-fold increase in spleen size and a 20-fold increase in splenic regulatory T (T Reg) cell numbers by day 3; circulating T cells were not evaluated 8. This was the expected pharmacological outcome for a CD28-specific superagonist antibody in a rat system, and was consistent with the in vitro activity. By contrast, ascending single-dose pharmacokinetic and pharmacodynamic studies with 5 to 50 mg per kg of TGN1412 in monkeys resulted in the saturation of CD28 receptors on T cells, but caused only minimal (<3-fold) and delayed (by >2 weeks) increases in