Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Feb 24.
Published in final edited form as: J Invest Dermatol. 2022 Jun 1;142(11):2920–2928.e5. doi: 10.1016/j.jid.2022.05.1059

IFN-γ ELISpot in Severe Cutaneous Adverse Reactions to First-Line Antituberculosis Drugs in an HIV Endemic Setting

Mireille Porter 1,9, Phuti Choshi 1,9, Sarah Pedretti 2, Tafadzwa Chimbetete 1, Rhodine Smith 3, Graeme Meintjes 4,5, Elizabeth Phillips 6,7, Rannakoe Lehloenya 8, Jonny Peter 1,2
PMCID: PMC9952832  NIHMSID: NIHMS1869651  PMID: 35659939

Abstract

Severe cutaneous adverse reactions related to first-line antituberculosis drugs are associated with high mortality and long-term morbidity. Oral sequential drug challenge, as a form of drug provocation testing, helps to salvage therapy by identifying culprit drugs but is associated with risk and is costly. IFN-γ enzyme-linked immune absorbent spot (ELISpot), an adjunctive in vitro diagnostic tool, may help to guide risk-stratification approaches. To determine the diagnostic accuracy of IFN-γ ELISpot against full-dose sequential drug challenge, we analyzed samples collected prospectively at multiple time points in 32 patients with first-line anti-tuberculosis drug–associated severe cutaneous adverse reaction (81% HIV infected, 25 with drug reaction with eosinophilia and systemic symptoms, and 7 with Stevens–Johnson syndrome/toxic epidermal necrolysis). Sensitivity of IFN-γ ELISpot was 33% (4 of 12), 13% (1 of 8), 11% (1 of 9), and 0% (0 of 4) for rifampicin, isoniazid, pyrazinamide, and ethambutol, respectively (positivity threshold ≥50 spot forming units/million cells). Specificity was 100% for all the four drugs. Rifampicin IFN-γ ELISpot sensitivity increased to 58% (7 of 12) if a threshold of 20 spot forming units was used and to 75% (3 of 4) when restricted to samples ≤12 weeks after acute severe cutaneous adverse reaction event; specificity remained 100% for both. IFN-γ ELISpot offers adequate risk stratification of rifampicin severe cutaneous adverse reaction using acute samples and lowered threshold for positivity. Given the low sensitivity of IFN-γ ELISpot for other first-line antituberculosis drugs, additional optimization is needed to improve risk-stratification potential.

INTRODUCTION

Severe cutaneous adverse reactions (SCARs), including drug reaction with eosinophilia and systemic symptoms (DRESS) syndrome and Stevens–Johnson syndrome/ toxic epidermal necrolysis (SJS/TEN), are associated with high mortality, short- and long-term morbidity, and health care costs (Knight et al., 2019, 2014). In settings endemic to tuberculosis (TB) and HIV, the first-line anti-TB (FLTB) drugs, namely rifampicin, isoniazid, pyrazinamide, and ethambutol, are common offending drugs for SCAR, with the majority of these reactions occurring in HIV-coinfected patients (Lehloenya and Dheda, 2012; National Department of Health South Africa, 2021; Tan et al., 2007). Persons living with HIV are particularly vulnerable with a 2–100-fold increased risk of SCAR (Breen et al., 2006; Carr and Cooper, 2000; Coopman et al., 1993; Lehloenya and Dheda, 2012; Lehloenya et al., 2011; Marks et al., 2009; Meintjes et al., 2019; Roujeau and Stern, 1994; Tan et al., 2007; Todd, 2006; World Health Organization, 2020; Yee et al., 2003). Interrupted FLTB drug treatment, particularly with rapid TB progression in the HIV-coinfected population, can lead to more severe disease, dissemination, treatment failure, emergence of drug resistance, and death (Meintjes et al., 2019; Schaberg et al., 1996)

FLTB drug–associated SCAR is a diagnostic challenge, with the simultaneous use of multiple drugs, and is often associated with hepatic and/or renal failure (Nalitye Haitembu et al., 2021). HIV coinfection increases complexity with the use of sulfa antibiotics (for opportunistic infection prophylaxis or treatment) and antiretroviral therapy (ART), themselves known to cause SCAR (Kakande and Lehloenya, 2015; Meintjes et al., 2019). FLTB drugs are preferred agents for drug-sensitive TB owing to higher efficacy, better accessibility, lower side-effect profiles, and shorter treatment duration than non–FLTB drug treatment options (Corbett et al., 2003; East African-British Medical Research Councils, 1974; East and Central African/British Medical Research Council Fifth Collaborative Study, 1980; Grobbelaar et al., 2019; Hoosen et al., 2019; Joint Tuberculosis Committee of the British Thoracic Society, 1998, Schaberg et al., 1996). Thus, when managing FLTB drug–associated SCAR, rapid low-risk methods to isolate the offending drug and allow the expeditious and safe reintroduction of nonoffending drugs is critical (Lehloenya and Dheda, 2012). Our team has pioneered the use of sequential drug challenge (SDC) as a form of oral drug provocation testing with FLTB drugs, showing that full-dose challenge is safe and effective at both identifying offending drugs and allowing a rapid reintroduction of safe drugs (Lehloenya et al., 2021, 2020, 2011). However, although the benefits outweigh the risk, SDC can result in moderate or severe adverse events requiring specialized care and considerable cost (Knight et al., 2019; Lehloenya et al., 2012a, 2011).

Ex vivo assays such as IFN-γ enzyme-linked immune absorbent spot (ELISpot) assays are well-known adjunctive diagnostic tools successfully used to risk stratify patients with SCARs for future drug safety by identifying offending agents in delayed drug hypersensitivity (Copaescu et al., 2021a, 2021b; Lehloenya et al., 2020). They have been shown to be less complex and less likely to be associated with false positives than traditional in vitro methods such as lymphocyte transformation testing (50% sensitivity) (Maecker et al., 2008; Pavlos et al., 2014; Polak et al., 2013). However, IFN-γ ELISpot assays have not been optimized, nor are they available for high TB burden, resource-poor settings (Kakande and Lehloenya, 2015; Lehloenya and Dheda, 2012). There is currently no defined protocol for FLTB drugs IFN-γ assay without the use of T-cell clones and T-cell costimulation (Ye et al., 2017). The aim of this study was to optimize IFN-γ ELISpot for FLTB drug–associated SCAR in an HIV TB-endemic setting.

RESULTS

ELISpot optimization

We have previously studied the highest noncytotoxic rifampicin and isoniazid concentrations for in vitro 18-hour drug stimulation (25 μg/ml for rifampicin and 50, 500, and 5,000 μg/ml for isoniazid) (Copaescu et al., 2021a). Here, we further evaluated noncytotoxic concentrations for ethambutol (50, 500, and 5,000 μg/ml) and pyrazinamide (50 and 500 μg/ml) (Supplementary Figure S1). We evaluated the use of two higher numbers of cells per ELISpot well (300,000 and 400,000 vs. 200,000), and no notable improvement in spot forming units (SFUs)/million cells was noted (Figure 1a). We compared ELISpot performance using fresh cells with using longer cryopreserved cells (0 days to 18 months storage in an −80 °C freezer). In three DRESS or SJS/TEN cases induced by rifampicin and isoniazid, we noted 2–4 times higher SFUs/million cells when using freshly isolated versus using cryopreserved cells; in two of three cryopreserved cells samples, SFUs were lower than the set positivity threshold (Figure 1b). In certain cases, it was not possible to use freshly isolated cells, and thus, where necessary, we still used cryopreserved cells as long as cell viability was ≥70%. Comparison of SFU/million cells by sampling timepoint was possible owing to multiple timepoint sampling. For analysis, patients were divided into two groups: ≤12 weeks (n = 30) (acute blood samples and early recovery) and >12 weeks (n = 22) (recovery-phase blood samples). Blood samples collected in the first 12 weeks after SCAR onset showed higher SFU/million cells after rifampicin stimulation (P = 0.0121) than blood samples collected >12 weeks after SCAR onset. No differences were noted after stimulation with isoniazid, pyrazinamide, and ethambutol (Figure 1c and d).

Figure 1. ELISpot optimization in FLTB drug SCAR.

Figure 1.

Open in a new tab

(a) The number of cells per well. (b) Fresh versus cryopreserved PBMCs. (c) Blood sampling timepoint by FLTB drug, and (d) blood sampling time point comparison of RIF response in six cases. ELISpot, enzyme-linked immune absorbent spot; EMB, ethambutol; FLTB, first-line antituberculosis; INH, isoniazid; PZA, pyrazinamide; RIF, rifampicin; SCAR, severe cutaneous adverse reaction; SFU, spot-forming unit; TEN, toxic epidermal necrolysis.

SCAR case inclusion

Between February 2019 and October 2020, 112 cases of cutaneous adverse drug reactions were enrolled in the Immune-Mediated Adverse Drug Reactions Registry and Biorepository (Figure 2). FLTB drugs were identified as suspected offending drugs in 50 (45%) of these cases, and 38 (76%) of these were defined as SCAR. SDC to one or more FLTB drugs was performed for 32 SCAR cases; this included 25 DRESS cases—13 definite (Registry of Severe Cutaneous Adverse Reactions [RegiSCAR]) score >5, 6 probable (RegiSCAR score = 4–5), and 6 possible (RegiSCAR score = 2–3) cases—and 7 cases of SJS/TEN (four definite and three probable cases) (Figure 2). SDC did not occur (n = 6) owing to death (n = 2), unconfirmed TB diagnosis (n = 1), FLTB drug continuance due to alternative offending agent identified (n = 2), and modified regimen use due to contraindication to challenge such as severe liver disease (n = 1). SDC included isoniazid in 32 patients; however, it included rifampicin in only 27 patients, pyrazinamide in 25 patients, and ethambutol in 29 patients. A total of 20 patients experienced SDC with all the four FLTB drugs.

Figure 2. Patient inclusion.

Figure 2.

Open in a new tab

CADR, cutaneous adverse drug reaction; DRESS, drug reaction with eosinophilia and systemic symptoms; FLTB, first-line antituberculosis; SCAR, severe cutaneous adverse reaction; SDC, sequential drug challenge; SJS/TEN, Stevens–Johnson syndrome/toxic epidermal necrolysis.

Table 1 shows the demographic and clinical characteristics of the 32 FLTB drug–associated SCAR cases who underwent SDC to one or more FLTB drugs. The majority of patients were female (66%). At the time of the SCAR event, 81% were HIV infected, with a median CD4 cell count (closest to time of SCAR) of 107 cells/mm3; 38% were on ART; and 50% were on sulfamethoxazole–trimethoprim. Naranjo scores where sulfamethoxazole–trimethoprim was suspected (n = 13) were 4 (possible) for 12 patients and 7 (probable) for one patient. One third of patients had disseminated TB, and 9% had a history of previous TB before the index event.

Table 1.

Demographics of 32 Patients with FLTB Drug–Associated SCAR Who Underwent SDC to One or More FLTB Drugs (n = 32)

Characteristic Value
Female, n (%) 21 (66)
HIV positive, n (%) 26 (81)
CD4 count, cells/mm3, median (IQR) 107 (59–175)
On ART at the time of SCAR, n (%) 10 (381)
On sulfamethoxazole–trimethoprim at the time of SCAR, n (%) 13 (501)
Previous TB, n (%) 3 (9)
Disseminated TB, n (%) 11 (34)
Positive drug reaction, n (%) 25 (78)
Single drug reaction, n (%) 18/25 (72)
Multiple drug reactions, n (%) 7/25 (28)
Any positive FLTB drug ELISpot, n (%) 7 (22)
Positive sulfamethoxazole–trimethoprim ELISpot, n (%) 1 (3)
Open in a new tab

Abbreviations: ART, antiretroviral therapy; ELISpot, enzyme-linked immune absorbent spot; FLTB, first-line antituberculosis; IQR, interquartile range; SCAR, severe cutaneous adverse reaction; SDC, sequential drug challenge; TB, tuberculosis.

1

Of HIV-positive cases within the group.

Oral drug provocation testing outcomes

A total of 25 (78%) of the 32 patients experienced one or more positive reactions to FLTB drugs SDC. The clinical phenotypes and timing of SDC reactions were varied and are detailed in Supplementary Tables S1 and S2. Reactions included immediate-type reactions (<6 hours) and delayed reactions and ranged from flushing and hypotension to worsening rash, eosinophilia, and liver functions. Rifampicin was the most common offending drug on SDC, with 12 of 27 (44%) positive reactions (10 of 20 [50%] for DRESS and 2 of 7 [29%] for SJS/TEN). Pyrazinamide was the next most common offending drug, with 9 of 25 (36%) positive reactions (5 of 19 [26%] for DRESS and 4 of 6 [67%] for SJS/TEN). Isoniazid and ethambutol were positive in 9 of 32 (28%) (6 of 25 [24%] for DRESS and 3 of 7 [43%] for SJS/TEN) and 5 of 29 (17%) (4 of 23 [17% for DRESS and 1 of 6 [17%] for SJS/TEN) SDCs, respectively (Figure 2). A total of 18 of the 25 (72%) patients experiencing positive SDC reactions showed a positive reaction to only one drug on SDC (single reactors): seven cases were positive for rifampicin, four were positive for isoniazid, six were positive for pyrazinamide, and one was positive for ethambutol. Seven patients experienced a positive SDC to more than one FLTB drug, with the commonest combination being reactions to rifampicin and isoniazid (n = 5) (Figure 3). One patient experienced positive SDC reactions to each of the four FLTB drugs, and one patient experienced positive SDC reactions to each of rifampicin, isoniazid, and ethambutol (pyrazinamide SDC was not done). Both of these latter two patients also exhibited intolerance to second-line anti–TB drugs, including moxifloxacin, linezolid, and clofazimine.

Figure 3. Proportional Venn diagram of positive SDC to FLTB drugs in 32 cases of SCAR.

Figure 3.

Open in a new tab

FLTB, first-line antituberculosis; SCAR, severe cutaneous adverse reaction; SDC, sequential drug challenge.

Seven patients experienced no reaction to FLTB drug SDC. In three of these patients, all the four FLTB drugs were tolerated on SDC, and sulfamethoxazole–trimethoprim was defined as the highest probable offending agent by Naranjo score. In two cases, rifampicin was not challenged because it was identified as the likely offending agent owing to clinical history, including a positive rifampicin patch test for one (for patient’s rechallenge characteristics, see Supplementary Tables S1 and S2). In these cases, rifampicin causation was diagnosed on the basis of exclusion with the successful challenge of the other FLTB drugs. Two additional cases were not fully challenged: one was initiated in the continuation phase (rifampicin and isoniazid), and the other was the one for whom rifampicin and pyrazinamide were excluded owing to severe renal and hepatic impairments.

IFN-γ ELISpot results

Each of the 32 FTLB drug–associated SCARs undergoing SDC had at least one FLTB drug IFN-γ ELISpot assay run, including 28 for rifampicin, 21 for isoniazid, 27 for pyrazinamide, and 24 for ethambutol. Five (18%) of rifampicin ELISpots were positive at a threshold of ≥50 SFUs. Isoniazid and pyrazinamide IFN-γ ELISpot assays only displayed positive results in one case each and none for ethambutol Figure 4).

Figure 4. FLTB drug sensitivity and specificity compared with those of gold standard SDC.

Figure 4.

Open in a new tab

The symbol † indicates ≤12 weeks from SCAR event, and the symbol ‡ indicates all the 9 cases of DRESS phenotype and threshold ≥50 SFU used. DRESS, drug reaction with eosinophilia and systemic symptoms; FLTB, first-line antituberculosis; MU, million unit; N, sample number; NPV, negative predictive value; PPV, positive predictive value; SCAR, severe cutaneous adverse reaction; SDC, sequential drug challenge; SFU, spot-forming unit; SJS/TEN, Steven–Johnson syndrome/toxic epidermal necrolysis.

Diagnostic accuracy of IFN-γ ELISpot compared with that of gold standard of SDC

To assess the diagnostic utility of IFN-γ ELISpot assay for the four FLTB drugs, its outcomes were compared with those of the gold standard of oral full dose SDC. This required both SDC and IFN-γ ELISpot assay to have been completed for each drug in question, resulting in rifampicin comparison for 23 patients, isoniazid and pyrazinamide each for 21 patients, and ethambutol for 23 patients. The sensitivity of ELISpot at a threshold ≥50 SFUs was 33% for rifampicin, 13% for isoniazid, 11% for pyrazinamide, and 0% for ethambutol, respectively (Figure 4). Specificity was 100% for all the four FLTB drugs, and positive predictive values were 100% for rifampicin, isoniazid, and pyrazinamide. In DRESS phenotype, rifampicin ELISpot sensitivity was 30 and 0% for the other FLTB drugs. In SJS/TEN phenotype, the sensitivities were 50, 33, 25, and 0% for rifampicin, isoniazid, pyrazinamide, and ethambutol, respectively (Figure 4). No difference was seen in sensitivity and specificity of IFN-γ ELISpot assay for each of the four FLTB drugs when DRESS phenotype was stratified into groups of possible, probable, or definite.

Acute sampling and lower cut-point to improve the sensitivity of rifampicin IFN-γ ELISpot assay

The sensitivity of rifampicin ELISpot increased to 75% when only samples acquired <12 weeks from SCAR event were included (Figure 4). These samples were all from patients with DRESS phenotype. The specificity for these samples was maintained at 100%. Diagnostic accuracy for IFN-γ ELISpot using different SFU thresholds is provided in Supplementary Table S3. If the threshold for positive IFN-γ ELISpot is reduced to SFU ≥20, 25, or 30, the sensitivity for rifampicin increased to 58% (DRESS and SJS/TEN combined) for any time of sampling, with specificity maintained at 100% (Supplementary Table S3). Among the seven single reactors to rifampicin, IFN-γ ELISpot sensitivity set with positivity at SFU ≥20 was 71%. The average value of SFUs in background control wells was 0–7 for the cases with <50 to ≥20 SFU/million cells.

Four patients (three with DRESS and one with SJS/TEN) of the 12 patients who experienced positive reactions to rifampicin on SDC had rifampicin ELISpot results ≥50 SFUs (Supplementary Table S1). All four of these patients were single reactors, HIV infected, not on ART or sulfamethoxazole—trimethoprim at the time of SCAR, and with low median CD4 cell counts ≤121 cells/mm3. Of the eight patients with positive rifampicin SDC but negative rifampicin ELISpot ≥50 SFUs, three had SFU/million cells >25, three were single reactors, and five were multiple reactors. In contrast to the four patients with rifampicin ELISpot results ≥50 SFUs, half of this group were HIV positive with CD4 counts ≥142 cells/mm3.

DISCUSSION

A high burden of life-threatening SCAR leading to treatment interruptions occurs amongst vulnerable persons living with HIV. In HIV/TB endemic settings, FLTB drugs are common offending agents (Lehloenya et al., 2021, 2020). Rapidly identifying the offending drug and reintroducing effective anti–TB therapy as well as other important drugs such as sulfamethoxazole–trimethoprim and ART can be life saving in advanced immunosuppression. This is the largest study to examine the utility and optimization of in vitro IFN-γ ELISpot assays to aid the identification of the offending drug in FLTB drug–associated SCAR. The major findings of our study were that several strategies could improve the sensitivity of IFN-γ ELISpot for FLTB drugs, including (i) the use of fresh rather than −80 °C freezer–cryopreserved PBMCs, (ii) acute sampling (≤12 weeks after acute SCAR admission), and (iii) a lowered SFU cut-point for assay positive to SFU ≥20/million cells for rifampicin. Rifampicin IFN-γ ELISpot was the only one of the four FLTB drugs where assay sensitivity (compared with that of the gold standard of SDC) was sufficient to offer meaningful diagnostic utility. In contrast, even when applying optimization strategies such as the use of acute samples, IFN-γ ELISpot sensitivity was 0–25% for SCAR to isoniazid, pyrazinamide, and ethambutol. Both DRESS and SJS/TEN phenotypes were included in this study, with a predominance of DRESS cases (78%). Sensitivities were similar by phenotype for the FLTB drug ELISpots; however, the small number of SJS/TEN cases in this cohort limits meaningful comparison between phenotypes. Unlike in other published studies where there is an increased sensitivity of IFN-γ ELISpot seen in DRESS with other drugs, we did not see this clear difference in our cohort (Polak et al., 2013).

We have previously optimized the highest noncytotoxic drug concentrations for stimulation in our rifampicin and isoniazid ELISpot assays (Copaescu et al., 2021a). In this study, we included noncytotoxic concentrations for ethambutol and pyrazinamide, and we also showed that the use of fresh rather than −80 °C freezer–cryopreserved samples improved drug-specific IFN-γ–producing cells with rifampicin and isoniazid stimulation. Our findings are consistent with those of other studies that reported ELISpot sensitivity to increase with the use of acute samples in the early stage after DRESS onset (Polak et al., 2013). The reduced sensitivity using cryopreserved cells may relate to cell viability after storage in −80 °C freezers versus in liquid nitrogen. The comparison between fresh and liquid nitrogen–cryopreserved samples may be less significant; however, −80 °C freezer capacity is more widely available than liquid nitrogen facilities in low and low- middle-income countries (De Oliveira et al., 2018; Mendy et al., 2014). The use of an increased number of cells, which we hypothesized may be required in the context of HIV and CD4 T-cell depletion, increases nonantigen-specific responses and did not improve assay sensitivity. In contrast, adjustment of the threshold of SFUs (to ≥20 SFU/million cells) for a positive assay significantly improved sensitivity in the rifampicin INF-γ ELISpot, without compromising specificity. In some studies, the threshold of positivity is defined by measuring the frequency of antigen-specific INF-γ–producing T cells in drug-tolerant controls (Rozieres et al., 2009), setting the positivity threshold at twice the average value of background spots (Wang et al., 2007) and on the basis of the distribution of negative controls (Keane et al., 2014, 2012). In our results, the mean of the background wells ranged from 0 to 7 SFU/million cells. With an optimized number of cells to stimulate and incubation period, nonantigen-specific responses are limited, and with a lower limit threshold of 20 drug-specific IFN-γ–producing T cells, an ELISpot for rifampicin can be considered positive (Streeck et al., 2009). Diagnostic accuracy was also not compromised in patients with reactions to multiple FLTB drugs or second-line drugs, which is reassuring. In one patient with reactions to all FLTB drugs, IFN-γ ELISpot was borderline positive for all FLTB drugs, whereas in the others, rifampicin IFN-γ ELISpot was 0 SFU/million cells; the significance of these results on the underlying mechanism of these reactions—whether flare ups or multiple drug hypersensitivity—remains under further investigations. The low number of patients at this stage limits the ability to draw conclusions.

Rifampicin is a cornerstone potent and important treatment for drug-susceptible TB, with rifampicin-containing regimens found to be superior to those not containing rifampicin with sterilizing activity allowing 6 months treatment duration (Grobbelaar et al., 2019; Hoosen et al., 2019). Immune-mediated adverse drug reactions, including SCAR, drug-induced liver injury, and acute interstitial nephritis, are major treatment-limiting side effects of rifampicin (Jin et al., 2021; Tan et al., 2007). In this and other South African cohorts of SCAR, rifampicin is the commonest of the four FLTB drugs to cause SCAR (Lehloenya and Dheda, 2012; Lehloenya et al., 2011). Strategies, such as optimized rifampicin IFN-γ ELISpot, which can rapidly identify rifampicin as the causative offending drug, negating a need for SDC and allowing the fast-tracked reintroduction of the other three FLTB drugs, offers major clinical utility. The large number of FLTB drug–associated SCAR cases examined in our study compared with those in previous literature has enabled us to present the optimized sensitivity and excellent specificity and positive predictive value, further supporting the use of rifampicin IFN-γ ELISpot (Copaescu et al., 2021b). Further studies are now required to confirm these findings in other high TB HIV–burden settings.

Limitations of this study include the low number of SJS/TEN phenotype cases; thus preventing comparison of the diagnostic use of IFN-γ ELISpot across SCAR phenotypes. The time points of sampling were limited owing to the inclusion of retrospective cases as well as prospective sampling. The ability to achieve consistent sequential sampling at all key time points would have strengthened this study, but the complex clinical management makes this difficult to achieve. Furthermore, follow-up data, particularly regarding the role of ART initiation on IFN-γ ELISpot assay findings, would have been a valuable addition.

The high HIV prevalence and concomitant use of ART and prophylactic sulfamethoxazole—trimethoprim seen in this study highlight the complexity of FLTB drug–associated SCAR in an HIV-endemic setting. We have also described a considerable occurrence of multiple drug reactions (over 25% of FLTB drug SDC reactors). Multiple drug reactivity is a complex phenomenon, and its pathophysiology is ill understood (Lehloenya et al., 2012b). This complicated environment limits the utility of the current gold standard of SDC, with particular reference to the variation in reaction phenotypes seen on SDC in multiple drug–reacting patients and the likelihood of nonspecific and transient immune responses that could be safely treated. The frequent lack of drug-specific IFN-γ–producing T cells to FLTB drugs in many of these cases may support this approach, although other factors may influence ELISpot performance, and further laboratory optimization is required for the other three FLTB drugs. This additional work includes (i) focusing on drug-specific T-cell reactivity to drug metabolites and (ii) the use of alternative cytokine readouts, including TNF-a, IL-2, and perforin. In the HIV-endemic setting of South Africa, the use of optimized IFN-γ ELISpot for the identification of causative drugs in FLTB drug–associated SCAR (with a predominant phenotype of DRESS) was examined. The diagnostic utility was found to be best for rifampicin, and the performance of rifampicin IFN-γ ELISpot compared with that of the gold standard of SDC could be improved by lowering the threshold to 20 SFU/million cells. Limitations in the application of ELISpot for isoniazid, pyrazinamide, and ethambutol highlight a need for exploration of alternative approaches using different concentrations, different drug metabolites, and/or different cytokine readouts in these assays.

MATERIALS AND METHODS

Patients, case validation, and sampling

The Immune-Mediated Adverse Drug Reactions Registry and Biorepository was developed through the Groote Schuur Hospital Combined Drug Allergy Clinic (Cape Town, South Africa)—the first interdivisional drug allergy clinic on the African continent (Peter et al., 2015). For this study, we identified patients within the Immune-Mediated Adverse Drug Reactions Registry and Biorepository who were admitted to Groote Schuur Hospital with FLTB drug–associated SCAR requiring treatment interruption and who required SDC toward the reinitiation of anti–TB drugs. All patients gave written informed consent, and the study was approved by the University of Cape Town Faculty of Health Sciences ethics committee (HREC R031/2018). Patients were enrolled both retrospectively and prospectively. Retrospective cases (n = 14) were identified through ward records and contacted to return for enrolment and sampling. Prospective cases (n = 18) included all SCAR admissions between January 2019 and October 2019 who were able and willing to provide informed consent. SCAR was defined and verified by two experts, including one board-certified dermatologist, as SJS/TEN (probable or definite)—on the basis of the international SCAR group consensus (Bastuji-Garin et al., 1993; Rzany et al., 1999)—or as DRESS (possible, probable, or definite) using RegiSCAR score for DRESS (Kardaun et al., 2014, 2007). RegiSCAR DRESS scores of 2–3 defined a possible case, scores of 4–5 define a probable case, and scores >5 define a definite case. Drug causality was assessed using the Naranjo and ALDEN scores for patients with DRESS and SJS/TEN, respectively, scoring each possible offending drug (Naranjo et al., 1981; Sassolas et al., 2010). For prospective patients, we collected blood samples into EDTA tubes at the acute stage of the reaction, before SDC, and on positive challenge reaction to any of the FLTB drugs. For both prospective and retrospective patients, blood samples were also collected where possible at 3, 6, 12, and 24 months and up to 10 years after SCAR onset.

Patients were included if they had undergone at least one FLTB drug provocation test. Oral full dose SDC was performed applying context-specific SDC protocols developed by the dermatology service at Groote Schuur Hospital (Cape Town, South Africa) (Lehloenya et al., 2020). Positive challenge reactions were clinically diagnosed on symptoms, clinical examination, and laboratory markers and on resolution with cessation of the offending agent. The diagnostic accuracy of IFN-γ ELISpot assay for each individual FLTB drug was measured against the gold standard of individual drug challenges as part of the SDC strategy.

IFN-γ ELISpot assay methods

Owing to the lack of a standardized strategy for these drug tests and variability based on incubation times, costimulation factors, and one or two cases tested (Copaescu et al., 2021b; Ye et al., 2017), we set out to optimize an IFN-γ ELISpot assay to detect anti–TB drug–specific T cells without driving nonspecific responses and T-cell clone generation. The IFN-γ ELISpot assay was optimized for (i) noncytotoxic drug concentration for drug-specific T-cell restimulation, (ii) fresh versus cryopreserved PBMCs, (iii) the number of cells per well, and (iv) blood sampling time point. Blood samples used for these optimization steps were from those within the FLTB drug–associated SCAR cohort and healthy controls.

The ex vivo 18-hour stimulation ELISpot assay was performed using IFN-γ antibody kit (3420-2H, Mabtech, Stockholm, Sweden), ELISpot plates (MAIPS451, MilliporeSigma, Burlington, MA), and 3,3’,5,5’-tetramethylbenzidine substrate (Mabtech). Cryopreserved or fresh PBMCs were cultured in R10 media containing 10% fetal bovine serum (Gibco, Waltham, MA), 1% N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid (1M, Gibco), 1% L-glutamine (200 mM, Sigma-Aldrich, St. Louis, MO), 1% penicillin/streptomycin (Life Technologies, Carlsbad, CA) in RPMI (Gibco). Ex vivo anti–TB drug stimulation was performed in triplicates, with 200,000 PBMCs added to each well. Cells were stimulated with rifampicin (≥97% [HPLC], R3501, Sigma-Aldrich), pyrazinamide (≥99.0% [purity thin layer chromatography], P7136, Sigma-Aldrich), ethambutol (≥99% [purity thin layer chromatography], E4630, Sigma-Aldrich), and isoniazid (≥99% [purity thin layer chromatography], I3377, Sigma-Aldrich) ); 25-desacetyl rifampicin (≥96% purity, sc-206558, Santa Cruz Biotechnology, Dallas, TX), acetyl isoniazid (sc-391344, Santa Cruz Biotechnology), pyrazinoic acid (≥95% [HPLC], CS-M-52650, Clearsynth, Mumbai, India), 5-hydroxy pyrazinamide (98.77% [HPLC], CS-P-02437,Clearsynth), and 5-hydroxypyrazine-2-carboxylic acid (98.77% [HPLC], CS-T-28539, Clearsynth); positive control (anti-CD3 antibody, Mabtech; staphylococcal enzyme B, Sigma-Aldrich); and negative control (unstimulated). The drug stock solution was prepared by dissolving the drugs in relevant diluents and stored at −80 °C for use within the stability tested period (Supplementary Table S4). The stock solution was serially diluted with R10 media to obtain working solution concentrations. After overnight incubation and the addition of detector antibody, avidin enzyme, and substrate (ELISpot assay kit, Mabtech), spots were counted using an AID automated reader (software version 7, AID diagnostika, Straßberg, Germany).

IFN-γ responses were averaged across triplicates, and results were presented as SFU/million PBMCs after subtraction of background (an average of no stimulant wells). A positive response was defined as previously reported ≥50 SFU/million cells. Results were analyzed using GraphPad Prism, version 9 (GraphPad Software, La Jolla, CA) (www.graphpad.com). Mann–Whitney test and paired t-test were used to compare between two groups and related samples.

Supplementary Material

1

ACKNOWLEDGMENTS

The Immune-Mediated Adverse Drug Reactions-Africa project is part of the European and Developing Countries Clinical Trials Partnership 2 program supported by the European Union (grant number TMA2017SF-1981). Research reported in this publication was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health (NIH) under award number R01AI152183. RL’s work is supported by the South African Medical Research Council and nonrated researcher support from the South African National Research Foundation. JP is supported by an NIH Fogarty career development award (K43TW011178-04). PC is supported by the NIH Fogarty PhD fellowship (5 D43 TW010559) and the South African Medical Research Council through its Division of Research Capacity Development under the Bongani Mayosi National Health Scholars Programme. EP reports grants from the NIH (R01HG010863, R01AI152183, and U01AI154659), grant support from UAI109565, and grant support from the National Health and Medical Research Council of Australia. She receives Royalties from Uptodate and consulting fees from Janssen, Vertex, Biocryst, and Regeneron. She is a codirector of the Institute for Immunology & Infectious Diseases Pty, which holds a patent for HLA-B*57:01 testing for abacavir hypersensitivity and has a patent pending for Detection of Human Leukocyte Antigen-A*32:01 in connection with Diagnosing Drug Reaction with Eosinophilia and Systemic Symptoms without any financial remuneration and not directly related to the submitted work. GM is supported by the Wellcome Trust (214321/Z/18/Z and 203135/Z/16/Z) and the South African Research Chairs Initiative of the Department of Science and Technology and National Research Foundation of South Africa (grant number 64787).

Abbreviations:

ART

antiretroviral therapy

DRESS

drug reaction with eosinophilia and systemic symptoms

ELISpot

enzyme-linked immune absorbent spot

FLTB

first-line antituberculosis

RegiSCAR

Registry of Severe Cutaneous Adverse Reactions

SCAR

severe cutaneous adverse reaction

SDC

sequential drug challenge

SFU

spot-forming unit

SJS/TEN

Stevens–Johnson syndrome/toxic epidermal necrolysis

TB

tuberculosis

Footnotes

CONFLICT OF INTEREST

The authors state no conflict of interest.

DISCLAIMER

The content of this paper is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

SUPPLEMENTARY MATERIAL

Supplementary material is linked to the online version of the paper at www.jidonline.org, and at https://doi.org/10.1016/j.jid.2022.05.1059

Data availability statement

No large datasets were generated in this manuscript. The authors agree to share data generated in this manuscript on request.

REFERENCES

  1. Bastuji-Garin S, Rzany B, Stern RS, Shear NH, Naldi L, Roujeau JC. Clinical classification of cases of toxic epidermal necrolysis, Stevens-Johnson syndrome, and erythema multiforme. Arch Dermatol 1993;129:92–6. [PubMed] [Google Scholar]
  2. Breen RAM, Miller RF, Gorsuch T, Smith CJ, Schwenk A, Holmes W, et al. Adverse events and treatment interruption in tuberculosis patients with and without HIV co-infection. Thorax 2006;61:791–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Carr A, Cooper DA. Adverse effects of antiretroviral therapy. Lancet 2000;356:1423–30. [DOI] [PubMed] [Google Scholar]
  4. Coopman SA, Johnson RA, Platt R, Stern RS. Cutaneous disease and drug reactions in HIV infection. N Engl J Med 1993;328:1670–4. [DOI] [PubMed] [Google Scholar]
  5. Copaescu A, Choshi P, Pedretti S, Mouhtouris E, Peter J, Trubiano JA. Dose dependent antimicrobial cellular cytotoxicity—implications for ex vivo diagnostics. Front Pharmacol 2021a;12:640012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Copaescu A, Mouhtouris E, Vogrin S, James F, Chua KYL, Holmes NE, et al. The role of in vivo and ex vivo diagnostic tools in severe delayed immune-mediated adverse antibiotic drug reactions. J Allergy Clin Immunol Pract 2021b;9:2010–5.e4. [DOI] [PubMed] [Google Scholar]
  7. Corbett EL, Watt CJ, Walker N, Maher D, Williams BG, Raviglione MC, et al. The growing burden of tuberculosis: global trends and interactions with the HIV epidemic. Arch Intern Med 2003;163:1009–21. [DOI] [PubMed] [Google Scholar]
  8. De Oliveira L, Dias MAB, Jeyabalan A, Payne B, Redman CW, Magee L, et al. Creating biobanks in low and middle-income countries to improve knowledge – the PREPARE initiative. Preg Hypertens 2018;13:62e4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. East African-British Medical Research Councils. Controlled clinical trial of four short-course (6-month) regimens of chemotherapy for treatment of pulmonary tuberculosis. Third report. Lancet 1974;2:237–40. [PubMed] [Google Scholar]
  10. East and Central African/British Medical Research Council Fifth Collaborative Study. Controlled clinical trial of four short-course regimens of chemotherapy for two durations in the treatment of pulmonary tuberculosis. Second report. Tubercle 1980;61:59–69. [DOI] [PubMed] [Google Scholar]
  11. Grobbelaar M, Louw GE, Sampson SL, van Helden PD, Donald PR, Warren RM. Evolution of rifampicin treatment for tuberculosis. Infect Genet Evol 2019;74:103937. [DOI] [PubMed] [Google Scholar]
  12. Hoosen K, Mosam A, Dlova NC, Grayson W. An update on adverse cutaneous drug reactions in HIV/AIDS. Dermatopathology (Basel) 2019;6:111–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Jin HJ, Kang DY, Nam YH, Ye YM, Koh YI, Hur GY, et al. Severe cutaneous adverse reactions to anti-tuberculosis drugs in Korean patients. Allergy Asthma Immunol Res 2021;13:245–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Joint Tuberculosis Committee of the British Thoracic Society. Chemotherapy and management of tuberculosis in the United Kingdom: recommendations 1998. Thorax 1998/November/03;53:536–548. [PMC free article] [PubMed] [Google Scholar]
  15. Kakande B, Lehloenya RJ. Drug reactions associated with antituberculosis drugs. Curr Allergy Clin Immunol 2015;28:264–8. [Google Scholar]
  16. Kardaun SH, Mockenhaupt M, Roujeau JC. Comments on: DRESS syndrome. J Am Acad Dermatol 2014;71:1000–1000.e2. [DOI] [PubMed] [Google Scholar]
  17. Kardaun SH, Sidoroff A, Valeyrie-Allanore L, Halevy S, Davidovici BB, Mockenhaupt M, et al. Variability in the clinical pattern of cutaneous side-effects of drugs with systemic symptoms: does a DRESS syndrome really exist? Br J Dermatol 2007;156:609–11. [DOI] [PubMed] [Google Scholar]
  18. Keane NM, Pavlos RK, McKinnon E, Lucas A, Rive C, Blyth CC, et al. HLA Class I restricted CD8+ and Class II restricted CD4+ T cells are implicated in the pathogenesis of nevirapine hypersensitivity. AIDS 2014;28:1891–901. [DOI] [PubMed] [Google Scholar]
  19. Keane NM, Roberts SG, Almeida CA, Krishnan T, Chopra A, Demaine E, et al. High-avidity, high-IFNγ-producing CD8 T-cell responses following immune selection during HIV-1 infection. Immunol Cell Biol 2012;90:224–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Knight LK, Lehloenya RJ, Sinanovic E, Pooran A. Cost of managing severe cutaneous adverse drug reactions to first-line tuberculosis therapy in South Africa. Trop Med Int Health 2019;24:994–1002. [DOI] [PubMed] [Google Scholar]
  21. Knight LK, Muloiwa R, Dlamini S, Lehloenya RJ. Factors associated with increased mortality in a predominantly HIV-infected population with stevens Johnson syndrome and toxic epidermal necrolysis. PLoS One 2014;9:e93543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Lehloenya RJ, Dheda K. Cutaneous adverse drug reactions to anti-tuberculosis drugs: state of the art and into the future. Expert Rev Anti Infect Ther 2012;10:475–86. [DOI] [PubMed] [Google Scholar]
  23. Lehloenya RJ, Isaacs T, Nyika T, Dhana A, Knight L, Veenstra S, et al. Early high-dose intravenous corticosteroids rapidly arrest Stevens Johnson syndrome and drug reaction with eosinophilia and systemic symptoms recurrence on drug re-exposure [published correction appears in J Allergy Clin Immunol 2022;10:650] J Allergy Clin Immunol Pract 2021;9:582–4.e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lehloenya RJ, Peter JG, Copascu A, Trubiano JA, Phillips EJ. Delabeling delayed drug hypersensitivity: how far can you safely go [published correction appears in J Allergy Clin Immunol Pract 2022;10;651]? J Allergy Clin Immunol Pract 2020;8:2878–95.e6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lehloenya RJ, Todd G, Badri M, Dheda K. Outcomes of reintroducing anti-tuberculosis drugs following cutaneous adverse drug reactions. Int J Tuberc Lung Dis 2011;15:1649–57. [DOI] [PubMed] [Google Scholar]
  26. Lehloenya RJ, Todd G, Mogotlane L, Gantsho N, Hlela C, Dheda K. Lichenoid drug reaction to antituberculosis drugs treated through with topical steroids and phototherapy. J Antimicrob Chemother 2012a;67:2535–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lehloenya RJ, Wallace J, Todd G, Dheda K. Multiple drug hypersensitivity reactions to anti-tuberculosis drugs: five cases in HIV-infected patients. Int J Tuberc Lung Dis 2012b;16:1260–4. [DOI] [PubMed] [Google Scholar]
  28. Maecker HT, Hassler J, Payne JK, Summers A, Comatas K, Ghanayem M, et al. Precision and linearity targets for validation of an IFNγ ELISPOT, cytokine flow cytometry, and tetramer assay using CMV peptides. BMC Immunol 2008;9:9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Marks DJ, Dheda K, Dawson R, Ainslie G, Miller RF. Adverse events to antituberculosis therapy: influence of HIV and antiretroviral drugs. Int J STD AIDS 2009;20:339–45. [DOI] [PubMed] [Google Scholar]
  30. Meintjes G, Brust JCM, Nuttall J, Maartens G. Management of active tuberculosis in adults with HIV. Lancet HIV 2019;6:e463–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Mendy M, Caboux E, Sylla BS, Dillner J, Chinquee J, Wild C, et al. Infrastructure and facilities for human biobanking in low- and middle-income countries: a situation analysis. Pathobiology 2014;81:252–60. [DOI] [PubMed] [Google Scholar]
  32. Nalitye Haitembu BN, Porter MN, Basera W, Hickmann R, Dlamini SK, Spearman CW, et al. Pattern and impact of drug-induced liver injury in South African patients with Stevens-Johnson syndrome/toxic epidermal necrolysis and a high burden of HIV. J Allergy Clin Immunol Pract 2021;9:4483–5.e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Naranjo CA, Busto U, Sellers EM, Sandor P, Ruiz I, Roberts EA, et al. A method for estimating the probability of adverse drug reactions. Clin Pharmacol Ther 1981;30:239–45. [DOI] [PubMed] [Google Scholar]
  34. National Department of Health South Africa. The First National TB prevalence survey South Africa 2018 short report. 2021.
  35. Pavlos R, Mallal S, Ostrov D, Pompeu Y, Phillips E. Fever, rash, and systemic symptoms: understanding the role of virus and HLA in severe cutaneous drug allergy. J Allergy Clin Immunol Pract 2014;2:21–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Peter J, Dlamini S, Lehloenya R. A South African multidisciplinary drug hypersensitivity clinic. Curr Allergy Clin Immunol 2015;28:270–4. [Google Scholar]
  37. Polak ME, Belgi G, McGuire C, Pickard C, Healy E, Friedmann PS, et al. In vitro diagnostic assays are effective during the acute phase of delayed-type drug hypersensitivity reactions. Br J Dermatol 2013;168:539–49. [DOI] [PubMed] [Google Scholar]
  38. Roujeau JC, Stern RS. Severe adverse cutaneous reactions to drugs. N Engl J Med 1994;331:1272–85. [DOI] [PubMed] [Google Scholar]
  39. Rozieres A, Hennino A, Rodet K, Gutowski MC, Gunera-Saad N, Berard F, et al. Detection and quantification of drug-specific T cells in penicillin allergy. Allergy 2009;64:534–42. [DOI] [PubMed] [Google Scholar]
  40. Rzany B, Correia O, Kelly JP, Naldi L, Auquier A, Stern R. Risk of Stevens-Johnson syndrome and toxic epidermal necrolysis during first weeks of antiepileptic therapy: a case-control study. Study Group of the international Case Control Study on Severe cutaneous Adverse Reactions. Lancet 1999;353:2190–4. [DOI] [PubMed] [Google Scholar]
  41. Sassolas B, Haddad C, Mockenhaupt M, Dunant A, Liss Y, Bork K, et al. ALDEN, an algorithm for assessment of drug causality in Stevens–Johnson syndrome and toxic epidermal necrolysis: comparison with case–control analysis. Clin Pharmacol Ther 2010;88:60–8. [DOI] [PubMed] [Google Scholar]
  42. Schaberg T, Rebhan K, Lode H. Risk factors for side-effects of isoniazid, rifampin and pyrazinamide in patients hospitalized for pulmonary tuberculosis. Eur Respir J 1996;9:2026–30. [DOI] [PubMed] [Google Scholar]
  43. Streeck H, Frahm N, Walker BD. The role of IFN-gamma ELISPOT assay in HIV vaccine research. Nat Protoc 2009;4:461–9. [DOI] [PubMed] [Google Scholar]
  44. Tan WC, Ong CK, Kang SC, Razak MA. Two years review of cutaneous adverse drug reaction from first line anti-tuberculous drugs. Med J Malaysia 2007;62:143–6. [PubMed] [Google Scholar]
  45. Todd G. Adverse cutaneous drug eruptions and HIV: a clinician’s global perspective. Dermatol Clin 2006;24:459–72. vi. [DOI] [PubMed] [Google Scholar]
  46. Wang JY, Chou CH, Lee LN, Hsu HL, Jan IS, Hsueh PR, et al. Diagnosis of tuberculosis by an enzyme-linked immunospot assay for interferon-gamma. Emerg Infect Dis 2007;13:553–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. World Health Organization. Global tuberculosis report 2020. https://www.who.int/publications/i/item/9789240013131; 2020. (accessed March 17, 2022).
  48. Ye YM, Hur GY, Kim SH, Ban GY, Jee YK, Naisbitt DJ, et al. Drug-specific CD4(+) T-cell immune responses are responsible for antituberculosis drug-induced maculopapular exanthema and drug reaction with eosinophilia and systemic symptoms syndrome. Br J Dermatol 2017;176:378–86. [DOI] [PubMed] [Google Scholar]
  49. Yee D, Valiquette C, Pelletier M, Parisien I, Rocher I, Menzies D. Incidence of serious side effects from first-line antituberculosis drugs among patients treated for active tuberculosis. Am J Respir Crit Care Med 2003;167:1472–7. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1869651-supplement-1.pdf (85.3KB, pdf)

Data Availability Statement

No large datasets were generated in this manuscript. The authors agree to share data generated in this manuscript on request.

RESOURCES