The excellent results achieved with current immunosuppressive regimens in terms of graft function and patient survival after solid organ transplantation (SOT) are threatened by the development of complications such as infections or cancer [1, 2]. Therefore, the discovery and validation of biomarkers capable of informing on the net state of immunosuppression constitutes a research priority [3]. Many of these assays are designed to quantify the adaptive response against specific pathogens, typically cytomegalovirus (CMV) [4]. In addition, some non-pathogen-specific parameters have been proven to predict the occurrence of post-transplant infection [5] or cancer [6]. None of these approaches, however, provides a comprehensive assessment of the functionality of the innate and adaptive components of the immune system.

The innate immunity is triggered through various families of pattern recognition receptors (PRRs) that detect distinct evolutionarily conserved structural motifs present in microorganisms. Toll-like receptors (TLRs) are central actors in the orchestration of the innate immunity and its interplay with the adaptive arm [7]. The activation of TLR signaling pathways leads to the transcriptional upregulation of genes involved in inflammatory responses, such as proinflammatory cytokines or type I interferons (IFNs) [8]. Research efforts have been focused on the effect exerted by immunosuppressive agents on T-cell and B-cell responses, given their role in allorecognition and graft rejection. In addition, a renewed interest has emerged on the contribution of innate responses to post-transplant events [9]. Multiple studies have shown that polymorphisms in genes encoding for PRRs modulate individual susceptibility to bacterial, viral and fungal pathogens [10–12]. It may be hypothesized that the relative contribution of innate immunity to the host defense becomes more evident upon abrogation of adaptive responses by long-term immunosuppression.

The QuantiFERON-Monitor (QTF-Monitor) is a commercial enzyme-linked immunosorbent assay (ELISA)-based IFN-γ release assay (IGRA) intended to quantify innate and adaptive immune responses following incubation of heparinized whole blood with an agonist of TLR7/8 (R848 or resiquimod) and an anti-CD3 monoclonal antibody [13]. Despite the advantages of this comprehensive approach, only a few studies have investigated the usefulness of QTF-Monitor to predict infectious complications after KT [14, 15], liver transplantation (LT) [16] or lung transplantation (LuT) [17]. In addition, no previous studies have evaluated the potential application of this assay to evaluate the risk of de novo malignancies after transplantation. The pathogenesis of this complication is multifactorial, with the participation of host (older age, sun exposure, pre-transplant history of cancer, smoking and alcohol consumption, latent infection by oncogenic viruses) and transplant-related factors (such as donor-transmitted cancer) [18]. Nevertheless, the deleterious effect of immunosuppressive therapy on cancer immune surveillance and the assumed concept that post-transplant cancer acts as a marker of over-immunosuppression provide the rationale to investigate whether an assay able to interrogate innate and adaptive responses may be also useful to predict the occurrence of malignancy.

With these research gaps in mind, we have assessed the functional immune status of a single-center cohort of KT recipients by means of the QTF-Monitor assay performed at multiple points throughout the first 6 months in order to characterize the dynamics of IFN-γ levels and their correlation with the development of infection and cancer.

Most of the biomarkers proposed to determine the net state of immunosuppression after SOT share two limitations: the lack of functional measurements ―as is the case with immunoglobulin levels or lymphocyte counts [3]― and the sole interrogation of virus-specific adaptive responses [4]. The QTF-Monitor assay offers the possibility of a broader functional assessment by measuring IFN-γ release upon in vitro stimulation of the innate and adaptive arms [13]. In the present experience the assay’s performance was moderate at best, since no association could be demonstrated between IFN-γ production (either categorized as “low” immune responses or as a continuous variable) at different points during the first 6 months and the primary outcome of overall infection. Nevertheless, we found that the QTF-Monitor results obtained during the first weeks may still be valuable to specifically predict the occurrence of bacterial or opportunistic infection, although this finding should be taken with caution due to the non-negligible false positive risk in the assessment of secondary outcomes. On the other hand, no apparent associations were found for de novo cancer.

The performance of the QTF-Monitor assay to predict post-transplant infection has been investigated by a few groups, with variable reported accuracy [14–17]. In a mixed cohort of 137 SOT recipients, Mian et al. observed that IFN-γ levels measured between months 1 and 6 were significantly lower in patients that developed subsequent infection and proposed an optimal threshold of ≤10 IU/mL. Urinary tract infection and pneumonia were the most common syndromes during the early post-transplant period, with a shift to predominance of viral pathogens beyond month 3. No multivariate analysis was performed to confirm the predictive value of IFN-γ production [14]. In contrast, a cross-sectional study at a mean of 2.6 post-transplant years failed to show differences in IFN-γ levels between stable KT recipients and those with infection. A subgroup analysis revealed that patients with bacterial infection had a significantly decreased IFN-γ release. Such an association, however, was not confirmed after adjustment for steroid dose and tacrolimus levels [15]. In a single-center cohort of LT recipients, IFN-γ levels at week 1 exhibited a fairly good capacity to predict infection through the first month, with the majority of the events being classified as opportunistic [16]. Finally, a recent study recruited 80 LuT recipients in which the QTF-Monitor was performed at 2, 6, 12, 24, and 52 weeks. The presence of IFN-γ levels <10 and <60 IU/mL at weeks 12 and 24, respectively, was associated with the diagnosis of opportunistic infection (mainly CMV viremia and IFD). Similar results were not observed for earlier monitoring points [17].

The discordant results from the existing literature, including those reported herein, may be partially attributable to differences across studies in outcomes and definitions, as well as in the timing and frequency of monitoring. Taken together, they would suggest that the QTF-Monitor assay may perform better for predicting some specific types of infection ―particularly of bacterial origin [15]― and when performed early after transplantation. Indeed, we have only identified differences in the assay results obtained at week 2 and month 1 according to the subsequent diagnosis of bacterial and opportunistic infection, respectively (with the latter mostly represented by CMV disease and herpes zoster). These results are in line with those previously observed among LT recipients [16]. Of note, the discriminative capacity for both outcomes was low, as indicated by auROC values below 0.700. Sood et al. reported a slightly better accuracy for the results obtained at week 1 after LT (auROC of 0.740) [16]. To put these findings into context, our group has reported higher discriminative capacities for other non-pathogen-specific biomarkers, such as the CD8⁺ T-cell count at month 1 (auROC of 0.739) or the total lymphocyte count at month 6 (auROC of 0.820) to predict opportunistic infection [25], Torque Teno virus (TTV DNAemia) at month 1 for predicting opportunistic infection and/or cancer (auROC of 0.704) [22], or serum sCD30 at month 1 for predicting bacterial infection (auROC of 0.846) [26]. Therefore, the potential contribution of the assay to the existing prediction models for post-transplant infection ―such as the externally validated SIMPLICITY Score [5]― should be explored in future studies.

Our results align with the cross-sectional study by Margeta et al [15] in that the performance of QTF-Monitor assay decreases at late periods after transplantation, once the amount of immunosuppression has been stabilized in most recipients. No differences in IFN-γ levels beyond month 1 were observed for any of the outcomes analyzed. Interestingly, we found no association between the QTF-Monitor results and the development of post-transplant cancer, a complication that usually results from the long-term effect of sustained over-immunosuppression [27]. No previous studies have investigated the role of QTF-Monitor assay to predict de novo malignancy. Although the number of events was low (n = 14), this negative finding would point to a lower relative contribution to the assay results of T-cell responsiveness (as compared to TLR-mediated innate responses), taken into account the pivotal role of cellular immunity in cancer immune surveillance. In contrast, we and others have shown that certain immune biomarkers assessed within the first months are useful to identify SOT recipients at increased risk of developing cancer in the mid- and long-term follow-up, such as CD4⁺ and CD8⁺ T-cell counts [6, 28], monocytic myeloid-derived suppressor cells [29] or TTV DNAemia [22].

The kinetics of IFN-γ levels measured by the QTF-Monitor assay was comparable to previous studies, which typically describe a sharp decline from the pre-transplant assessment followed by a progressive recovery through months 3–6 and a plateau thereafter [13, 14, 16, 17]. This pattern is in line with the accepted timing for immune reconstitution after SOT, as validated with other biomarkers such as TTV DNA load [30, 31]. The clinical factors influencing assay results have been only partially investigated. The association between the use of ATG as induction therapy and a lower IFN-γ production has been reported by other authors [14]. In our experience this effect persisted until month 1 and was supported by the inverse correlation observed between IFN-γ levels and T-cell counts. The impact of tacrolimus levels is less consistent, with studies reporting either strong [17] or borderline correlations [15], or even no apparent association [14]. We only found a weak inverse correlation with tacrolimus levels at month 6. Mian et al. also reported an association with daily doses of prednisone and mycophenolate [14], which were not recorded in our database. Although beyond the scope of our research, we found no significant association between the immune status measured by the QTF-Monitor assay at the different monitoring points and the subsequent occurrence of biopsy-proven acute rejection (data not shown). Patients with a low response at week 2 were more likely to have received pre-transplant dialysis and to have experienced delayed graft function (defined by the early requirement of renal replacement therapy). The deleterious effect of dialysis on the T-cell ability to produce IFN-γ after specific stimulation is well established for IGRAs used to detect Mycobacterium tuberculosis infection due to insufficient mitogen response and premature immune aging [32, 33]. Inversely, living donation was associated with a more robust immune response, which may be explained by the lower recipient age and the immediate graft function in this subgroup.

What may be the position of the QTF-Monitor assay for immune monitoring in the clinical arena? With the limitations inherent to multiple secondary outcome analyses and the lack of consistent associations at later points, our results would point out to the potential usefulness of the early assessment within the first weeks with the specific aim of predicting bacterial infection. By decreasing the IFN-γ threshold to <7.9 IU/mL we obtained a sensible improvement in specificity without a major impact in sensitivity, although the resulting estimates (66.7% and 68.2%, respectively) were far from excellent. Sood et al. proposed a clinical threshold of <1.30 IU/mL as the most discriminative to predict infection beyond the first week after LT, with a diagnostic accuracy in the line of our results (sensitivity of 71.4% and specificity of 63.0%) [16]. On the other hand, an alternative threshold (<47.3 IU/mL) at month 1 yielded a reasonable sensitivity (81.8%) to predict opportunistic infection, at the expense of a poor specificity (51.2%). Gardiner et al. also found a relatively low discriminative ability for different outcomes (overall infection, severe infection or opportunistic infection) and monitoring points after LuT [17]. In our opinion, any decision regarding the implementation of the QTF-Monitor assay in daily practice must balance diagnostic accuracy (which was found to be suboptimal in our experience), requirement of laboratory resources and economic costs with those of alternative biomarkers [3]. For instance, the observed impact on IFN-γ production of ATG induction and CD4⁺ and CD8⁺ T-cell counts would suggest that low responses may ultimately act as a surrogate for the presence of lymphocytopenia, which constitutes a well-established biomarker for opportunistic infection [25, 34–38].

Our study is based on a large cohort of KT recipients with regular monitoring, and it is strengthened by the assessment of immunosuppression-related complications which comprised infections and malignancies. We also provided an insight into the clinical determinants of the IFN-γ kinetics, including peripheral blood lymphocyte subpopulations. Nevertheless, a number of limitations must be noted, such as the relatively low number of some events, which may have limited statistical power. Due to logistical reasons, the assay could not be tested in certain patients at all the scheduled points. Although the minimum follow-up was set at post-transplant month 12, the last monitoring point was performed at month 6. In addition to budgetary considerations, the rationale for such decision was that most events would have occurred within the first 6 months, according to the classical timeline for post-transplant infection [39]. In addition, the overall amount of immunosuppression (i.e., prednisone dose and targeted trough tacrolimus levels) is usually stabilized beyond that point in most KT recipients. Therefore, it is not to be expected major changes in the results of the QTF-Monitor assay beyond month 6, as supported by the plateau between months 3 and 6 observed for IFN-γ levels (Figure 1). In addition, any conclusion on the potential usefulness of the QTF-Monitor assay for predicting bacterial or opportunistic infection should take into account that both events were considered as secondary outcomes.

In this cohort of KT recipients we found no significant association between IFN-γ production measured with the QTF-Monitor assay and the primary outcome of overall post-transplant infection. Secondary outcome analysis would suggest that the usefulness of this assay is presumably limited to the prediction of bacterial and opportunistic infection when performed within the first weeks after transplantation. Further studies are needed to establish the role of this promising method in the available repertoire of non-pathogen-specific immune monitoring biomarkers.